Volumes 1-5, 7, 10, 11, 13. 14. 16, 17, 21, 22. 23-27, 29. 31 are out 01 print.

Pill MID
PHASE BEHAVIOUR OF PE:1RO LEOlvi RESERVOIR FLUIDS
All DANESH

6
8 9 12 iSa

15b
18a 18b t9a 19b 20 28 30 32 33 34 35 36 37 38 39 40. 41 42 43 44 45 46 47

Fundamentals of Numerical Reservoir Simulation Fundamentals of Reservoir Engineering Compaction and Fluid Migration Fundamentals of Fractured Reservoir Engineering Fundamentals of Well-log Interpretation, 1. The acquisition of logging data Fundamentals of Well-log Interpretation, 2. Tire interpretation of logging data Production and Transport ofOil and Gas, A. Flow mechanics and production Production and Transport of Oil and Gas, 8. Gathering and Transport Surface Operations in Petroleum Production, I Surface Operations in Petroleum Production, II Geo/ogy in Petroleum Production Well Cementing Carbonate Reservoir Characterization: A Geologic-Engineering Analysis, Part I Fluid Mechanics for Petroleum Engineers Petroleum Related Rock Mechanics A Practical Companion to Reservoir Stimulation Hydrocarbon Migration Systems Analysis The Practice of Reservoir Engineering Thermal Properties and Temperalure related Behavior of Rock/fluid Syslems Studies in Abnormal Pressures Microbial Enhancement of Oil Recovery Re~enl Advances Proceedings ofthe 1992 lOternational Conference on Microbial Enhanced Oil Recovery Asphaltenes and Asphalts, I Subsidence due to Fluid Withdrawal Casing Design Theory and Practice Tracers in the Oil Field Carbonate Reservoir Characterization: A Geologic-Engineering Analysis, Part II Thermal Modeling of Petroleum Generation: Theory and Applications Hydrocarbon Exploration and Production PVT and Phase Behaviour ofPetroleum Reservoir Fluids

Department of Petroleum Engineering Heriot Watt University Edinburgh, Scotland

______
~--

1998 ELSEVIER Amsterdam

Lausanne

-~

New York Oxford

Shannon

Singapore

Tokyo

CONTENTS
PREFACE NOMENCLA1URE l:l 1.2 PHASE BEHAVIOUR FUNDAMENTALS RESERVOIR FLUID COMPOSITION PhASE BEhAVIOUR Iurc (onipound (or rcspondi rig Stales Multiconiponent Mixture (LASSIFICATION OF RESERVOIR FLUIDS l)ry Gas Wet Gas Gas Condensate Volatile Oil Black Oil REFERENCES EXERCISES 3 4 It) IS

Library of Congress Cataloging in Publication Data A catalog record from the Library of Congress has been applied for.

1.3

22

1.4 1.5

24 25 25 27 28 29 30 33 34 34 36 40 41 42

2.
2.1 2.2

PVT TESTS AND CORRELATIONS

FLUII) SAMPLING Well Preparation Sample Collection PVT lESTS 38 2.2.1 DryGas
2.2.2 Wet Gas

ISBN: 0 444 82196 1

1998 Elsevier Science By. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Science By., Copyright & Permissions Department, P.O. Box 521, 1000 AM Amsterdam, The Netherlands. Special regulations for readers in the U.S.A. This publication has been registered with the Copyriglil Clearance Center Inc. (CCC>. 222 Rosewood Drive, Danvers, MA, 01923. Information can be obtained from the CCC about conditions under which photocopies of parts of this publication may be made in the

2.3

2.2.3 Black Oil 2.2.4 Gas Conclensate 2.2.5 Volatile Oil EMPIRICAL CORRELATIONS 2.3.1 Black Oil Rubble Point Pressure Gas in Solution Oil Formation Volume Factor Total Farina lion Vo/unre Factor Oil Density Oil Viscosity 2.3.2 Natural Gas Volumetric Data
(;~ Viscosity

52
65

66
67 68 70 70 71 73 77
79

2.3.3

U.S.A. All other copyright questions, including photocopying outside of the U.S.A., should be referred to the publisher.
No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods,

products, instructions or ideas contained in the material herein.

The paper used in this publication meets the requirements of ANSI/NISO Z39.48 1992 (Permanence of Paper). Printed in The Netherlands.

2.4 2.5

Formation Water Water Cr,ntent of hydrocarbon Phase ll drocarbon soluinlity in Water 5 Waler Formation Volume Factor Compressibility of Water Waler Density Waler Viscosity REFERENCES EXERCISES

80 83 86 87
90

92
92
93

93 95 99

3.
3.1

PHASE EQUILIBRIA
CRITERIA FOR EQUILIBRIUM Chemical Potential Fugacity Activity EQUILIBRIUM RATIO Raoults Law Henrys Law Empirical Correlations REFERENCES EXERCISES 270

105

I05
107 108 Ill III 112 114 116 125 127 129 130 131 132 138 140

6.4 6.5

Direct Application REFERENCES EXERCISES

241 247 249 253 254 258 260 260 265 266 270 270 273 277 279 281

7.
7.1

GAS INJECTION
MISCIBILITY CONCEPTS Miscibility in Real Reservoir Fluids EXPERIMENTAL STUDIES Slim Tube Rising Bubble Apparatus Contact Experiments PREI)IC1ION OF MISCIBILITY CONI)ITIONS First Contact Miscibility Vaporising Gas Drive Condensing-Vapor ising Gas Drive REFERENCES EXERCISES

3.2

7.2

3.3 3.4

4.
4.1 4.2

EQUATIONS OF STATE
VIRIAL EOS AND ITS MODIFICATIONS Starling-Benedict-Webb-Rubin EOS CUBIC EQUATIONS OF STATE 4.2.1 Two-Parameter EOS Soave-Redlich-Kwong EOS Peng-Robinson EOS Volume Sh~fl 4.2.2 Three-Parameter EOS Schmidt-Wenzel EOS Patel-Teja EOS 4.2.3 Attraction Term Temperature Dependency MIXING RULES 4.3.1 Random Mixing Rules 4.3.2 Non-Random Mixing Rules REFERENCES EXERCISES

7.3

7.4 7.5

141
141 45 146

8.
8.1 8.2

INTERFACIAL TENSION
MEASUREMENT METHODS PREDICTION OF INTERFACIAL TENSION Parachor Method Corresponding States Correlation Comparison of Predictive Methods WATER-UYDROCARBON INTERFACIAL TENSION REFERENCES EXERCISES

282
285 285 288

147
149 153 154 158 162 165 167 168 175 179 183 189 192 195 197 198 2(X) 201 203 206 209 210 210 215 8.3 8.4 8.5

289
292 295 297 301 302 302 308 310 314 316 318 319 320 322 323 324 325 325 327 330 330 331 333 334 338 340 345 349 353 385

4.3 4.4 4.5 5. 5.1 5.2 5.3 5.4

9.
9.1

APPLICATION IN RESERVOIR SIMULATION

GROUPING Group Selection Group Properties Composition Retrieval COMPARiSON OF FOS Phase Composition Saturation Iressure Density Gas and Liquid Volumes Robustness TUNING OF EOS Fluid Characterisation Selection of EOS Experimental Data Selection of Regression Variables Limits of Tuned Parameters Methodology DYNAMIC VALIDATION OF MODEL Relative Permeability Function Viscosity Prediction Implementation EVALUArION OF RESERVOIR FLUID SAMPLES REFERENCES EXERCISES

PHASE BEHAVIOUR CALCULATIONS

VAPOUR-LIQUID EQUILIBRIUM CALCULATIONS Root Selection Rapid Flash Calculations STABILITY ANALYSIS Stability Limit CRITICAL POINT CALCULATIONS COMPOSITIONAL GRADING Equilibrium Assumption Non-Equilibrium Fluids Heat of Transport Significance REFERENCES EXERCISES

9.2

9.3

5.5 5.6

6.
6.1 6.2

FLUID CHARACTERISATION
EXPERIMENTAL METHODS Distillation Gas Chromatography CRITICAL PROPERTIES Lee-Kesler Correlations Riazi-Dauhert Correlations Perturbation Expansion Correlations DESCRII9ION OF FLUID IIEAVY END Single Carbon Number Function Continuous Description

9.4

221

6.3

221 222 223 227 228 234

9.5 9.6 9.7.

APPENDICES INDEX
vii

vi

NOMENCLATURE a A b B lI~ It, B C~ C, f G h

Ii

PREFACE

Reliable measurement and prediction of phase behaviour and properties of pet rolciini reservoir fluids are essential in designing optimum recovery processes and enhancing hydrocarbon production. This book explains relevant fundamentals and presents practical methods of determining required properties for engineering applications by judicious review of established practices and recent advances. Although the emphasis is on the application of PVT and phase behaviour data to engineering problems, experimental methods are reviewed and their limitations are identified. This should provide the reader with a more thorough understanding of the subject and a realistic evaluation of measured and predicted results. The book is based on the material developed over many years as lecture notes in courses presented to staff in gas and oil industry, and postgraduate students of petroleum engineering. It covers various aspects of the subject, hence can he tailored for different audience. The first two chapters along with selected sections from chapters 3 and 5 can serve as the subject matter of an introductory course, whereas the rest would be of more interest to practising engineers and postgraduate students. Ample examples are included to illustrate the subject, and further exercises are given in each chapter. Graphical methods and simple correlations amenable to hand calculations are still used in the industry, hence they are included in this hook. The emphasis, however, is on the more advanced compositional approaches which are attaining wider application in industry as high computational capabilities are becoming readily available. I would like to thank Professor DII Tehrani for reviewing the manuscript and valuable suggestions stemming from his vast industrial experience - Also, I urn grateful to Professors M. Michelsen and C. Whitson for their helpful comments on sections of the book. Much of the material in this book is based on the authors experience gained through conducting research sponsored by the petroleum industry, at Heriot-Watt University. I am indebted to the sponsors, my students and colleagues for their contributions that made this book possible. In particular, I would acknowledge valuable contributions of Professor AC Todd, Mr F Goozalpour, Dr DII Xu, Mr K Movaghar Nezhad and Dr D Avolonitis. My son Amir cheerfully helped me in preparing the hook graphics.

H I1~ h 1 k k, k,~ k,, K K~ m M n N N~ P P, P1 P5 0 P~ R R, S l I~ ii

V ~

attractive terni parameter of equation of state dimensionless attractive term parameter of equation of state repulsive tcrm(co-volume) parameter of equation of state clinrensiontcss repulsive term parameter of equation of state gas formation volu mire f~ic tor oil I urinaliiiir volt: lire factor total formation volume factor gas isothermal compressibility coefficient oil isothermal compressibility coefficient fugacity Gibbs energy height molar entha!py total enthalpy Flenrys constant partial molar ent lIi~l PY peruieahility binary interaction parameter gas relative permeability oil relative permeability equilibrium ratio Watson characterisation factor slope in ~scorrelation with temperature molecular weight (molar mass) mole or carbon number number of components number of pseudo-components presstire bubble point pressure convergence pressure parachor vapour pressure universal gas conslant gas in solution specific gravity, relative density at 288 K (60 F) temperature irorrirat hoihiig point temperature iriiitar :nlcini.r! enci gy
imiolar votunie velocity

V x, y,
7.,

Z Z 55

volume mimic fraction mole fraction in vapour phase mole fraction compressibility factor Rackett compressibility factor

GREEK LETIERS
a temperature dependency coefficient of attractive term mean value parameter of r distribution function activity fugacity coefficient parameter of r distribution function calculated critical compressibility factor total number of phases chemical potential mass density molar density interfacial tension lowest molecular weight in F distribution functiomi acentric factor EOS parameter coefficient activity coefficient any phase

St JPERSCRI PtS F h o
5

Cj
~ ~y s~
K

V W

Iced, imrixliile hydrocarbon phase liquid phase reference state saltmration vapour l)hase water phase

j.t p Pu ~
t
(i)

SUBSCRIPIS
I)
C

LI ~

(I
It

base or bubble point critical point differential liberation process

gas

It
0

ACRONYMS hbl

r s w

hydiocarbirmi oil reduced property salt w,mter

value/valtmc at critical point

B IP CCE CGR CVD DL EOS GOR GLR GPA GPM

barrel binary interaction parameter constant composition expansion condensate to gas volumetric ratio constant volume depletion differential liberation equation(s) of state gas to oil volumetric ratio (Sc) gas to liquid volumetric ratio (Sc) Gas Processors Association gallon of liquid per thousand cubic feet of gas (sc) IFT interfacial tension MMP minimum miscibility pressure MME minimum miscibility enrichment PNA paraffins-naphthenes-aromatics PR Peng-Robinson EOS PT Patel-Teja EOS Sc standard conditions SCF standard cubic feet SRK Soave-Redlich-Kwong EOS STB stock tank barrel SW Schmidt-Wenzel EOS TBP true boiling point temperature VPT Valderrama-Patel-Teja FOS 7JRK Zudkevitch-Joffe-Redlich-Kwong EOS

xi

1 PHASE BEHAVIOUR FUNDAMENTALS

Petroletim reservoir fluids are composed mainly of hydrocarbon constituents. Water is atso present in gas and oil reservoirs in an interstitial fonrr. The influence of water on tIre phase hchavioumr and properties of hydrocarbon fluids in most cases is of a minor consideration. The phase behaviour of oil and gas, therefore, is generally treated independent of tire water plrase, unless water-hydrocarbon solid structures, known as hydrates, are formed. The behaviour of a hydrocarbon mixture at reservoir and surface conditions is determined by its chemical composition and the prevailing temperature and pressure. This behaviour is of a prime consideration in tire development arid management of reservoirs, affecting all aspects of petroleum exploration and production. Altirotmgh a reservoir litmiut may he composed of many thousands of compounds. tire phase behaviour fundamental s can be explained by exami 0 ing the behaviour of pure and simple mrrulticonrponent mriixiurcs. The behaviour ol all real reservoir fluiuls basically follows the saline principle, but to hicilitate tIre application of tire technology in the industry, reservoir fluids have heemr classified into various groups such as the (try gas, wet gas, gas condensate, volatile oil and black oil.

RESERVOIR FLUII) COMPOSITION

lucre are various hypotheses reg~ridiirgthe fornratiorr of petroleum from organic materials. Ihese views suggest that tIre coisrpositmon of a reservoir fluid depends on the depositionai environment of the formation, its geological maturity, and the migration path from the source to trap rocks Ii]. Reservoir gasses are mainly composed of hydrocarbon molecules of small and mediumni size,s and sonic light non-hydrocarbon compormnd.s such as nitrogen and carbon dioxide, whereas oils are predominantly composed of heavier compounds. Fluids advancing into a trapping reservoir may be of different compositions due to being generated at different times and environments. thence, lateral and vertical compositional variations within a reservoir will he expected dumring the early reservoir life. Reservoir fluids

1.

tlrase

Helun,ou r Fiinukun:e,,i~i!.c

I. 1. Reservoir Fluid O~nrposit:on

are generally considered to have attained equilibrium at maturity due to molecular diffusion and mixing over geological times. However, there are airrple evidences of reservoirs still maintaining significant compositiomial variations, particularly tatcially as tIre dill usi ye rrixing may require many tens of million years to eliminate coiripositmonal hmclcmogeiiuit cs I 2 I Furthermore, the pressure and the temperature increirse with depth for a tluid column iir a reservoir. This can also result in compositional grading with depth. For operatiomial puposes. this behaviour is of considerable interest for near critical lluids, and oils containing bight concentrations of asphaltic material. The compositional grading and its estimiration based on thermodynamic concepts will be discussed in Section 5.3. The crude oil composition is of major consideration in petroleum refining. A number of comprehensive research projects sponsored by the American Petroleuni Institute have investigated crude oil constituents and identified petroleum compounds. API6 studied the composition of a single crude oil for 40 years. ihe sulphur, nitrogen and orgtrmiomricttrlhic compremnds of crude oil strniptcs were investigated in projects API-48. API 52 ;inrd API-Sti respectively. API-60 sttmdied petroleum heavy ends. Nelson [31 gives a review of petrolcumni chemistry and test methods used in the refining industry. liighly detailed information on the constituents composing a reservoir fluid is not of very muchr use in exploration and production processes. Reservoir fluids are commonly idcmrtilmcd by their constituents individually to pentanes. and heavier compounds are reported as groups consrposed mostly of components with equal number of carbons such as C6s, C s. Cgs. All the 7 possess the same compounds forming each single carbon number group do not necessarily number of carbons as will he discussed in Section 6. I. Tire most common method of describing the heavy fraction is to lump all the conipounds heavier titan C and meport it as C +. 6 7 hydrocarbon compounds can be expressed by the general formula of Ci1II2Irf~with some sulphur, nitrogen, oxygen and minor metallic elements mostly present in heavy fractions. Hydrocarbon compounds are classified according to their structures, which determine the value of ~,. The niajor classes are paraffins (alkanes). oletins (alkenes), naphthenes, and aromatics. The paraffin series are composed of saturated hydrocarbon straight chains with ~=2. Light paraffins in reservoir fluids are sometimes identified and reported as those with a single hydrocarbon chain, as normal, and others with branched chain hydrocarbons, as iso. The olefin series (~=0)have unsaturated straight chains and are not usually found in reservoir fluids due to their unstable nature. The naphthenes are cyclic compounds composed of saturated ring(s) with ~=0. The aromalics (~,=-6) are unsaturated cyclic cotripounds. N~rphthrenes~mmr(h aromatics form a major part of C6-Ct t groups and some ofthem such itS methyl-cycio-pcntanc. bcnzene, tohuene and xylene are often individually identified in the extended analysis of reservoir fluids. For example, the structural formulas of tire above groups of hydrocarbons with six carbons are shown in Figure 1.1. As reservoir hydrocarbon liquids may be composed of many thousand components, they cannot all be identified and measured. However, the concentration of hydrocarbon components belonging to the same structural class are occasionally measured and reported as groups. particularly for gas condensate fluids. The test to measure the concentration of paraffrns, naphthenes, and aromatics as groups is commonly referred to as the PNA test 141Further information on the structure of reservoir fluid comnpounds and their labelling according to the IUPAC system can he found in [SJ. The compositional analysis of reservoir fluids anh their characterisation will be discussed in Chapter 6. Nitrogen, oxygen and sulphur are found in light and heavy fractions of reservoir fluids. Gas reservoirs containing predominantly N2, H2S. or c02 have also been discovered. Polycychic hydrocarbons with fused rings which are more abundant in heavier fractions may contain N, S, and 0. These compounds such as carboids, carbenes, asphaltenes and resins are identified by their solubihity, or lack of it, in different solvents [6]. The polar nature of these compounds

can affect the properties of reservoir fluids, particularly the rock-fluid behaviotmr, disproportionally higher than tlreir concentrations [7]. These heavy compounds may be present in colloidal suspension in the reservoir oil and precipitate out of solution by changes in the ~ tcmrrpcratumre or comnpos!tiomrs occurring durimig production. H H C H It H II II H It II

I I I I I I C C C C C Cli I I I I I I
II II U H It II
Norinnat Icx;ine

H H

HC CC C CH

ALKANES (PARAFFINS)

II

II

It

II

It

iso-lie ranc

II

It C~H
H tt

It

It

II C It

HH

It C C C C C

I
II

I
II

I
It

II C C C C= C

I I

ALKENES

I
II

I
II H

Ii

It

H H

I -Itexenc

3-Meltryt- I Pentene

/
H\
,,,,,~ -~

H /11 c\tt I/It

C~

I1\~ H~ ~
It

I
H/NC
~C\
11 Benzcne

/\

It

Cyclotrcxanc

NAIIIrIIENLs

AROMATICS

Figure 1.1. Structural formula of various groups of hydrocarbons with six carbons. 1.2 PIIASE BEHAVIOUR

Reservoir hydrocarbons exist as vapour, liquid or solid phases. A phase is defined as a part of a systemli whtcir is physically distinct from other parts by definite boundaries. A reservoir oil (liquid phase) may form gas (vapour phrase) during depletion. The evolved gas initially remaIns dispersed iii the oil phase before forming large mobile clusters, but the mixture is considered as a two-phrase system in botlr cases. The formation or disappearance of a phase, or variations in properties of a phrase in a multi-phase system are rate phenomena. The subject

of phase behaviotmr, however. focitses only on the State of equilibrium, where no changes will occur with time if tIre system is heft at tire prevailing constant pressure and temperature. A

1. Phase Behaniour Fu,rdamelrta/.l

1.2. !ha.ce liehaeu~ur

system reaches equilibrium when it attains its minimum energy level, as will be disctmssed in Chapter 3. The assumption of equilibrium between fluid phases in contact in a reservoir, in most cases, is valid in engineering applications. Fluids at equihihriumm are also referred to as saturated fluids. The state of a phase is fully defined when its conrposrtion, temperature arid prcssurre are specified. All the intensive properties for such a phase at tire prevailing conditions are tmxed and identifiable. The intensive properties are those which do not depend oil tire amount of material (contrary to the extensive properties), such as the density and tIre specific heat. The term property throughout this hook refers to intensive properties. At equilibrium, a system may form of a nunrber of co-exiting phases, wimlr all tIre fluid constituents present in all the equilibrated phases. The number of independent variables to define such a system is determined by the Gibbs p/rose rule descrihet as follows. A phase composed of N components is fully defined by its number of irrolcs plus two thermodynamic functions, commonly tenrperatumre anul pressure, that is, by N2variables. The intensive properties are, however, determined by only N+I variables a~ tire concentration of components are not all independent, but constrained by, (II)

critical temperature. I lence tIre vapour pressure values at temperatures above the critical temperatures, shtowrr by in Frgurrc 1.3, are not real, but simply extrapolated values.
Critical Poini

8
A

Solid

ii.

I)

A triple Ponni

Vapour

tennperatnnrc

>

Figure I .2. Pressure-temperature dragrrrmrr of pure suibstairce. where, x is the mole fraction of component i. Thums, for a systersi with K phrases, tire total 1 number of variables are equal to rc(N-f I). However, the temperature, presstrre. and chenrical potential of each component throughout all phases should be unifornrr at equilrhriunr conditions, as will be described in Chapter 3. This inrpuses (N2)(icI) constraints. lterrce. the nunrber of independent variables, or so-called the degrees of freedom, F, necessary to (lefirre a multiphase system is given by,: F = K(N+l)-(N+2)(K-l)
=

The line AR on Figure 1.2 is the solid-liquid equilibrium line, which is also krtown as tire mrrclting point curve. lhre intersection of the vapour_liqummd and liquid-solid lines ts the triple point. It is thre only poirrt where the three phases cart coexist for a pure system. Tire lure AD is the solid-vapour equnihihriunr line or the suhlitrratmon curve. The solid carbon uhioxiule (dry ice) vaporising into its gaseous form is a common example of this region of the phrase behaviour diagram. lire variation of saturated fluid ulensity with teirrperature for a pure compound is shown in Figure 1.5. The densities of vapour and liquid phases approach each other as tire temperature i iru-reases. Ihey hcconie equal at conuhitions kmiowir as tire Critical point. All tIre differences between tire phrases are re(Iuced as the system approacires the critical point. Indeed, the p/maces i,euonre t/iC .Sa?tiC uiiid iF,(IiStiprhi415/i(1ble at tire (ritical point. go me I .4 shows I Ire variation of saturated Ii uid chensr Iy with temperature for a nurmrrbcr of prime Irydroearbomrs - All tIre conmporrrrds slrow a sirmnilar tremid, that is, tIre vapour and liquid rlemrsiries bccorrre equal at tIre cnrtrcat point. OtIrer properties also show tire same tiend. Ilre en mciii tenrrperatrire. I~.,arid tIre critical pressure, P,. are tire niaxilriuni terriperature arid me at which ii proe conrpororil cirur fomnmr coexisting ilrascs. 1 The ternrs vapour and liquid are reterred to the less and the more dense phases of a fluid at equilihriunr, I leirce. a pure corripound at a temperature above its critical value canirot be called eitlrer liqumid or vapour, lire continuity of vapour and liquid is scherrraticahhy strown in Figure 1.6. TIre density at e~rclrpoint is sirown by the shauhing intensity, where tIre ularker shadmrrg corresponds to a iriglrer density. Tire discontirruity across the vapour-pressure curve becomes less significant as tire temperatuie increases arid vanishes above tIre cmitical poirrl. Tire superheated vapour F can he chranged gradumally to tIre compressed liqumid F, thiough an arbitrary patlr EG F, Wi tirout any abrupt phrase cirange.

K+2

(1.2)

For a single-component (pure) system, the degrees of freedom is equal to three nrinus thc number of phases. The state of the equihibriurm of a vapour-liquiuh mixture of a pure lluid, therefore, can be determined by identifying either its pressure or its tcnrperalume.

Pure Compound
The phase behaviour of a pure compound is strowir by tire prcssurc-teiriperutrrre di:rgrurnrr it Figure 1.2. All the conditions at winch tire vapour and umquid phases can coexist at cqtrrlmtamrrrir are shown by the line AC. Any fluid at any otlrer pressure-tenrperatuie conditions, is unsaturated single phase as required by the phrase rule. The fluid above and to tire left of the line is referred to as a compressed or under saturated liquid, wlrereas that heiow and to the right of tIre line is called a superhreated vapourr or gas. The line AC is commonly known as the vapour pressure curve, as it shows tire pressure exerted by the vapour coexisting with its liquid at any temperature. The temperature corresponding to the atmospheric pressure is calleut tire normal boiling point or simply tire boiling point of the compound. The boiling point, Tb, of some compotrnds foundt in reservoir fluids are given in Table A. I in Appendix A. Figure 1.3 shows tIre logarithm of vapour pressure plotted against an arbitrary temperature scale for some compouinuls. The scale, which is an adjusted reciprocal of the absolute temperature, has been selected so that the vapour pressures of water and most hydrocarbons can be exhibited by straight lines. This plot is known as tire Cox chart. A pure substance cannot exist as liquid at a temperaturre above its

1, Phase Belrannon,r Fwrda,,re,m,url.s

1.2. Phase Behaviour

t-i--~ ~

j~.-~J
-

L-

1000

-IL

~
no

2
!L~i~ H~ ~
-~

moo
mOO
no

V
I U,

:i~_~:, -~-~ ~

LE~L 7Z L

2
7
.O
U, U,

0~

IL.

~
~

~ ~

~4
200

~
00

iso

250

300

350

400

450 500 350 600

no.
700 ~ ~00 000

Temperature, F K=(F+459.67)/l.8 MPa=O.006895 psia

Rcpnnxiuccd

Temperature, F K=(F-f459.67)/ 1.8 MPa=Q.006895 psia McGraw.Hitt

Companies Copyright.

Figure 1.3. Vapour pressure of normal paraffins. McGraw-tlitt Companies Cnnpyrigtrn. homer 181 wiih permission.

Figtmre 1.3 (Coin). Vapour pressure of normal paraffins. Rcprnstuced fioi~n I~! s~iitrpermnnicsionn

I. Pha.me Rehannonrr Fnondamnre,rtat.r

1.2. Phase Behaviour

Temperature,K
200 250 I...., . 300 350 400 450 S00 550
600 650
A

San urmued I _rquiul

C I)

il Hifli It I U
Saunnr~niedVapour

tennnprrarure

f I I ~&I I

I 1f~ITU

uigure 1.5. Vamiatiomrs of saturated tluid density with tennperature.

10-44151

41401

!iUIU
________ H II fttiWH-ft1t-tt
fH+HIfH P-H-I-/ ftt1-iHH+~f+H1-I-H+H-fl rLIFHU 100 0
tOO 200

~tiIi~iJllhI!~ Ill~ftI~FI~IIt1 ~1f~-~IU U~I-~-tf~f~UIt II~ ~B ~ ~t Ii~A LflI

~Yt-14 ~kY

1
700

a
C,

1enrnpermninrre

>
Cnipyrngtii Reprnstnncent trim

30C

400

500

600

ll~llt~ll~ ii

Irginre I .6. (ontinltrity oh vapour anul hmquid. Mc(Ir~nw-Ititt(onnipannies IS! with pernnnnssionn.

Temperature,oF Figure 1.4. Saturated fluid density of pure compoumnds (curves iderrtilred by letters are related to binary and multicomponent fluids described in Reference 8). McCraw-ttilt Coninpanies Copynigtnn
Reproduced from 81 wiih permission.

TIne pressurevolunre diagram of a pure substance is shown in Figure 1.7. (onsider tire corrrpresseul hiquniul. Point A, at a tercrperature below tire critical temperature. lbe reductiomn of fluid pmcssure at constant teirnperature increases its volume. As tine liquniul is relatively incomrrpressiblc tine finn uI expansiomn is small until tire vapour prcssnrre is reaclred, at Point 13 where tire first bubble evolves. Further exparrsioin of tire systeirn results un chi~rmiginigtire liquid iran tire vapour phn:ise. For a pure substance Ilie pressure remlrai us constant mmd equnal to tIre vmrpourr pressure, a consequmence of tire phase rurle, until the last drop of tire liquid virponises, Point I). ~hlnis point, where tIre vapour is in equnilihrnumrr with an imnfinitesimnral anrnount of hiqtnid is called tIre utew poinn.

10

I. P/ia Cf Behan,,,nn, !nondan,ne,nials

1.2. Phase Behaviour

Ii

N T=Tc

where v is tire molar voltrmnie, P is (absoltmte) pressure, F is (absolute) temperature, and R is the umnivcrsal gas constant (Table A.3 in Appendix A). Hence one mole of any ideal gas occupies the same volume at a given pressure mmd temperature. In emrgimneering applications, gases at tire standard conditions can be treated as ideal. The occupieul volume of one mole of gas at various standard conditions, calculated by Eq.( 1.3), is given in Table 1.1.
T >Tc Single Ptiunse

T<Tc
A

Table 1.1. Molar volume of ideal gas at various stammdard conditions.

Unit ~perarurc
-

~ure,V_olnrnne

Frctd Mennic St

60(1 F 271. t S K 2)18 K

14.69 psia I ann Its) kImn

380 tm~IIhriIot 1 22.414 mnr /kgmot 23.95 rim ItcgumIInt

0.

As omre minnIe of a hydrocarbon gas amnd one mole of air occupy the same volume at the standard comrditions, thte speciFic gravity of gas relative to air (relative density), S, is simply determined by.
I)

Sg=MglMair whmere, Muir is the mrmolectmlar weight (mnrolar nnass) of air, equal to 28.96 kg/kgmol.
Two Phase Regrnnin

(1.4)

l)ue to imrtermolecrilar forces remit gases din mint behave iulcahly, particularly at elevated pressures. Fq.( 1.3) is extemnulcd to remit systemrrs by including a compressibility factor, Z, as, Pv=ZR~l
Volume
->

(1.5)

Figure 1.7. Pressure-volume diagram of pure lluid. The system bubble points at various temperatures form the bubble point curyc, whereas the dew points form the dew point curve. The two curves meet at the critical point and together identify the phase envelope. Any fluid within the phase envelope, Point M. forms two equilibrated phases with the vapour/liquid molar ratio equal to BM/MD. The bubble point and dew point curves appear as a single vapour pressure curve on a pressure-temperature plot for a pure compound, Figure 1.2. lhe chamrge of plnase from liquid to vapour is accortmpan;ed by a large inncrcmrse in volrime mit low temperatures (Figure 1.7). The expansion reduces ax the temnnpcrature approaclnes the critical point. Indeed the system changes front all liquid into all vapour, or vice versa, without arty change in tire mixture volume at the critical point. An isothnermal expansion of a fluid at a temperature above the critical temperature does not result in any phase change, Poimni N. This fluid is called a supercritical fluid.

Fine comnnpressihility factor can be deteurnined from various theoretical-empirical equations of state (Chapter 4), or determnnined from a generalised chart for gases as shown in Figure I .8. Note tinat tIre compressibility factor depends only unn the ratio of temperature to critical tcmnrperature (ahsolurte), tire reduced temperature, Tr. and the ratio of pressure to critical pressure, the reduced pressure,
15 r-

TIne above approach is based on a very inrportant concept, known as the corresponding sta:e.c principle, which states that suhstamnces helrave similarly when tlney are at the same relative proxirinity to their criticmrl poimnts. This inrphies tlnat all substances behave similarly at their critical points, hence, shouhul Inave equal critical compressibility factor, Z~, Pv

(1.6)

Corresponding States
All gases behave ideally when the pressure approaches zero. The pressure vokrnne relation for an ideal gas is,

lime real vunlue of critical coirnpressibihity factor, however, is not thne same for all compounds crumble A. I in Appendix A). lhne conmprcssihility chart, trowever, provides reliable estimates particularly for simpercritical gases and at low pressure conditions. Charts relating the compressibility factor to tIne reuluced pressure and temperature, similar to Figure I .8, hut specific to compounds such as rinetlrane, ethane, propane, have been produced to improve the accuracy of predicted values (101. Application of tlne correspondirmg states principle to the vapour pressure of pure compounds, follows a similar trend. The logarithm of vapour pressure of pure compounds approximately varies linearly with tire reciprocal of Iemsnperature as shown in Figure 1.3. Ii can he expressed, therefore, as

Pv=RT

(1.3)

12
-

1. !/na.n !te/nnnlioin, F=nnmnc/cznn,nqn0n/.r :~

~tm 2 , .. ~

/2. !-/no.ie l/e/nannonnr

13

:.:

.~;f.~
~ ?~T
~,

~
- .~

~
-

~r~. ~ 1~~U~ifi=~j
~.~---

...

~
, .

.-

log(P I P,)

~~

( I .7)

winere P~ is tire vapour pressure amrd ~t annul ~2 are constamnts for eurch substance. At the critical point PIPC=TIF,.= I , hence ~ t =
c2

anul, (1.8)

0.

-~/t.f.t .-.444.--P ~ ~~-i~z t~ 4t~f ~I. t~Pir~ I:: ,t t1=~. 4~ ~

5

T
.:t =:u
...

~.

..

U~

~ ~

-..

I. ,1 ,:Ii~ .: .~ I mi.
. ,

~.

tIm
1-

i_.-

~
-~ -

)1..~. ~t~t t.~:~f t~t-L_.j-- 1

,~ -.

,~
~i ~ ~

.~s-TT .~ .: ..u:~,.l.j

fl.
~,

I
8.

:~

~:

1::

~
.: a
.~ ~

~ t~

I
-.*

I
4~

log(P,~)=~ (l~_L) 1 If tIre correspomnhing states principle were exact, tire vapour pressure curves of all line commipounds, plotted inn tIre redunceul form, should have tIme sammne slope, tlrat is equal ~ n . falling oil the saline line. lmr practice, tins (hoes not occur. ~1 he nhevrat mm of mrnoulels haseuh omi fire two parunmincler correspon(lr mrg states pm nc Iplc rs due tic uhi Iferemnccs in nmroleculunr strinctunres oh various commrpoummnuls. resulting nit (litleremit imnicrmnroleu-unlumr Iorues. lIne imnchusiorm oh a tlnirul parmmrmlcter, aduhtiomnal to tire rcduceul tcmmnperalnmre mmml pressure, svhnichr corncnmr.s to thee rnnolecinlmur structnnic shrould imnrprove tIne reliability of the coriespomrding states primrciphe. Pit~,er LIIJ noticed thrat tIne redunced vapour pressure curves of sinnple spherical rnrolecules. snicln as argon, kryptomn and xemnon, iriulccd lie on tine samnne curve wmtln a reduced vapour pressure of I). I at tIne reduced temrrperature of 0.7. Hence, for otlrer substances he selected the deviation of 7 tire reduced vapour pressure curve frommn tlnat of spherical molecules at Tr=O. as the third pmrrannreter of tIne correspondimng states principle. amrd irntroduced the acentric factor, as, (fl=log(P/P~)(, . ~ 1.0 1 (1.9)

-~

-:~

U~J~ 4~t

41H

~
t

tJ

~7
.~

~~T~W: ~
0

The ahuwe definition gives arc acerniric factor of zero for simple spherical molecules, and positive values for otiner corurpourids except hydrogen and helium. The acentric factor generally imrcreases with imncreasing size of iromologuc irydrocarhons. The values of acentnc factor for sonne connpoimnds are given in Table A. I in Appendix A.
0 a

~_

4
~

Tire acentric finctom Imas been widely accepted as the ttmird parameter in generating generahised correlations, based omr tIre corresponndinng states principle, particularly tlrose related to fluid plnmse cqnmilibiia. For examnnple. lIme vapour pressure of pure comcepounds can be reliably estimirated using the Lee and Kesher 1121 correlation whniclr is based on the three parameter corresponding stales. P /P
=

~1

I ~

exp(f + ( fW)
0

(I. It))

wlnerc, lIt)). anul ph) are furcctiomns of the reduced temnrperature,

1.~

AI/~

~
a
0 0 0

1t~t~~

= =

5.92714

6.09648 /(T,)

1.28862 ln(T,) + 0.1 6934(T~ )5

)6

15.2518 l5.6875/(i~) - 13.4721 ln(T, ) -I- 0.43577(T,

E~T~b~
o
0 ii

L~n,nrjIe 1. 1.

Calcnniate the vapour pressure of normal hexane at 355.1S K. using: (a) hire Cox drum, (h) tire Lee-Kester equationn.

z -~n~i cli mqlscaldwo3

1

14
Solution:

1. //na.ce 1?elntn,iou n /nnnndnn,,ne,nan/s

/.

2. l/na,ce !?e/rannonnr

15

(a) From Figure 1.3, am T=355.15 K (179.6 F), the vapour pressure is remid eqniuml to 0.15 MPa (21 psia). (b) The critical properties of normal hexane are read from Table At in Appendix A, ummnd used in Eq.(I.lO) to calculate the vapour pressure as follows:
MPa 3.025 m 0.30t3 1, 0.69966 1 -2.306192 C -2.306921 P, MIun 0.1504

Tlre cylinder pressunre remnnains curnrstmnnn, equal to time irormal butane vapoumr pressure, as long as the mixture remains two ptna.ces at 393 K. The vapour pressure can be calculated from the Lee-Kester equation, Eq.(I It)), sinrihar to that in Example 1.1, which results in: P=2.2 160 MPa, In 393 K. The vapour density am the above conditions can be calculated from Eq.( 1.7). The donurpressihility factor, Z, no read fronmr Figure 1.8. at prevailing reduced values of: P,=P/I~=2.216/3.796=0.5838 and T,=0.9244, to he Z=0.67. The universal gas constant is read, from Table A.3 iii Appenmutix A, to be 0.(X)83 144 MPa.m/(K.kgnrol). I Icncc, v
=

T,,

P~,

507.6

The umse of critical compressibility factor as the third parurnnetcr for (levellmpinmg genemumlisi.d correlations to predict volumetric data has also proved successful. Ann exmmnrpie is tIne Racketl equation [131 for the saturated mnrolar volume of pure comnnpounds, v / v
=

/.I~I/ Jm = I

IX))

mmn/kgmrnot. ummrnt tIm vap~nnnruhemisity is

5 5 p =M/v =58. 123/I .00.1=57.95 kg/inn lIne minass balance resumlts inn, 5 nrr=Vn ~4 v p~ 1 t).5=V x425.3 -440.00235 I-V )< 57.95 Lnuinn nd butane
vol mm.

Z~UT.4

(1.11)

where v~,and v~are the saturated liquid and critical molar volunnes. respectively. A more reliable estimation ofthe liquid molar volume is expecteul front the modification of the Rackett equation by Spencer and Danner [141, where the critical compressibility functor hrmms hecim replaced by the parameter ZRA, known as the Rackett compressibility factor, v~ = (RT / P~ (1.12)

V =0.1)009902 tim

Multicomponent Mixture
Fire phrase helmaviotmr of a mrrultr-comrnponnent system is more elaborate than that of a pure cninmpound. llmc comnmplexity gemnerunlly conrmpounds as components with widely differemrt structures mmd mrnolecular sizes corriprise tIme system. Reservoir fluids are mainly composed of Inydrocarhons with smnmrlar structures. Tineir phase beiraviour, therefore, is not generally highly cnirmplex. TIre phase helnavioumr of a hinany system. although relatively simple, is very much similar to a curt imruilti-comnnponent reservoir fluid. It is, tlnercfore, an appropriate substitute for explaining tine qunalituntive betmaviounr of reservoir lrydnx_arhon mixtures, line pinunse nile inruircuntes that inn a hmmrary vapourliquid system, both the temperature and the pressure are imrulepeindemnl variables. TIme pressure-temniperattmre diagram of a binary mixture is sclnennatmcatiy slrowim inn Figumre 1.9. the phrase envelope, inside which the two phases coexist, rs bouiimulcd by tIme buibble poimrt mmmmul chew point curves. flme two curves meet at the critical point (C), where mill dnfferences between lIme two plnunses vanislr and the phases beconne nmnnlistmngtnishiatnlc. None tlmat tIre two plrases can coexist mmt some c(inditions above tire critical pnnnmrt . llme highest trressnirc (13) annul the higlmesl tenmrpcratumre (D) on the phrase envelope are emil lent tIre (rnu(n,n(/(,nlra,- mmmd tIre cri(o,i(je,nt/me,-m, rcspcclively. rlre piessure-voluimme diagram of a binary mixturre is schematically shown in Figure 1.10. Note tlnat tIre systemnn pressure decreases during ant isothermal expansion between its bubble and dew points, conntrunn-y to Ilumnt for a punme conrrpounnul. TIre plrase diagranr of a mixture is determimned by its comtrposition. Figure 1.11 shows the phumse diagram of etlmane-Imeptanie sysnenn. llce critical temperature of different islixtures lies hetwcenm tIne critical tenrmperatures of tIre two pure comrrpounds. The critical pressure, however, exceeds the values of boOm commipomnents as pure, in most cases. Tire locus of critical points is slrosvmn by tire daslned litre in Figures 1 . II - The greater the difference between the critical

The values of ZRA for some substances [l5j are given jim Table A. I in Appenulrx A. For oIlier compounds, it can he estimated from the Yamada-Gtrnn correlation 1161= ZRA=O.29O56-O.08775u (1.13)

The application of acentric factor and critical compressibility factor in developing generahiseul correlations will be described further, particularly in Chapter 4 dealing with equations of stale. Fixaunple 1.2. Calculate the density of saturated normal butane liqumid at 393 K, using mice Ruickeit equation. A cylinder contains I kg of satunrated liquid humane at 393 K. Wlrat is mIne volume of liquid butane remaining in tine cylinder after comnsunninmg 0.5 kg of botanic?
Solution:

Reading the critical properties of normal butane from Table A. I in Appennutix A ammul substituting them in Eq.(l.l2). at 393 K, we ohtaimr: M, kg/kgmol
58.123 T~,K 425.12 PC MPa 3.796 7~ 0.2730

r,
0.92444

v, nrn/mnnt

0.13665

Don siny, kg/inn 4253

where the density. p, has been calculated as, p = M/v The volume of cylinder, 5containing 1kg of the satunrated liquid butane, is: V=mnifp=I/425.3=0.0023 1 m

16

I. P/nose Re/mnn,nnnmr Fnnpnnla,nne,nt,zI.s

I 2 I/nase Aelmavionnr

17

points of the two components, the higher the mixture critical pressurrc earn rise as sinnnwn nm Figure 1.12. No binary nrixture can exist as a two-phrase systenn outside tire region boummdeul by the locus of critical points.

lime cnnrrespomnding slates prmrrciplc. described for pumre stibstances, is also used for mniuhitcomurponenrt systetnrs. Pseudo cmiiical values mire used, Imowever, instead of true critical properties in applying fluid niouhels develuiped for ptrre strbstances, such as those in Figure 1.8. amnd Eq.(l.Il). Pseudo critical properties of a mixturre are calculated by applying a mixing rule to the critical propertmes of its constituents. A rrumber of mnlixing nrles have been proposed, but molar averaging, also known as Kays nnixinng rule, is the most common rule, ,0~~ (1.14) is unmry pseudo critical property, such as tetnrperature.

B C

I)

wlrere /,, ns tIne mrroic inuncninnn, ~0

a
0

pressure. mmmcui volimmmre, mnnnd 0, is lice cnnticmul properry of comnponemmt i. Properties scaled relumnive to tine pseudo crnnical vunlumes are referred to IIS pseurdo reduced properties, such as, pseudo reduced tennrpenatuire: n-i; and, pseudo reuluced pressuire: ~ Temperature
>
= =

l/~T~

(I . IS)

(I . 16)

to
Figure 1.9. Schematic pressuretenniperature di~ngrammmof mm birnary mmrmxlunre.
9 8

1 I T2

C
Criiicat Poini 0~
C)

7
6

m,,te% citnanne I i()0.00 2 9685 3 88.71 4 77t)9 5 58.71 6 2654 7 790 8 1012 9 4.05 it) 000

a 5
A 0~ 4

1:1

Vapniur
0 U

.3 2

i/tnt

it
250

300

350

400

450

500

550

600

Voiunne

>

Temperature, K

Figure 1.10. Pressure-volume diagram of binary mmxtures.

Figure I .11 .

Phrase diagram of etlrane

normal hreptane.

Mcflraw.}tnhl Companies

C,ipyrig)ni

Reproclunceut frn,nni f81 winin l~enmnissonn

18

I. P/nose 8e/nnr,unnn Fnundunnnrenikn!.c

1.2. P/muse Belmanionnr

19

Tire true critical properties, however, are different from the pseundo values calcumlaled by averaging. The true critical pressure often shows the irighest deviation froirn tIme psen.mdo value, as evidenced in Figure 1. 12. The prediction of true critical properties will be ulescmiheul inn Section 5.3. Temperature, K 100 ISO 200
250

A typical phase diagram of multi-coisrponent system at constant composition is shown in Figure I. 13. Vapour and liquid phases coexist at any pressure-temperature conditions within the phase envelope. Tire liquid/mixture volumnietric ratios are sirown by the constant quality lines. Note ttrat tIne distamice hetweemr iso-volume or quality lines decreases as the critical point is mmpproached. Smmrmmll pressure or Iennperature changes at a region near the critical point cause mrlajor phase chamngcs.

300

350

400

451)
Crinical Poinmi

12
A

Ii)

0~

8
~1

in

tcnnnperamure

>

0~

as a
is

a) a
is

Figure 1.13. Plnmtse dimmgramn of a mnulticomponent mixture. An isothermal reduction of pressure for a vapour-like fluid, Point A, forms the first drop of liquid at tine dew poinit, Point B. Further reduction of pressure will result in further condensation, as indicated by tIre qumality lines. This plretromenon is known as the retrograde conden.cutw,n. Tlre cuindensation will cease at some point, Point D, and the condensed phase will revumporisc by fuirtlmer reduction of pressnire. TIme shaded region of the phase diagram. wlmcrc pressure reduciiomr resumlts jim comndensation is referred to as tIme retrograde region. Note mlnunt tIme mrhove behaviniuir occurs only if the gas temperature lies between the critical temperature mind Ihe cric(imrdentlnemnmr. Figure I - 13 slrows that there are two dew point pressures at any termmperatunre for retrograde gases. Tire upper dew point is sometinres called the retrograde dew point. Tine lower ulcw point is of little practical significance for most gas condensate fluids, Time relative posilion of tIre critical poimin to the cricotrdentherm and the cricondenbar on the phmmse emnvehope can lead to other retrograde plnenomena. Figure 1. 14 shows that an isobaric iircrease of tenipcratune from poinrt I to point 2 results inn condensation. This behaviour, which earn mnlso he cunllcd rerograde cuindeirsation, is of little interest in reservoir operations. It indicates, however. limat rmmisirmg tIne tcnmrpcralumre of a high pressure rich gas may not be a proper iirocedimre to avoid conulensation inn fluid haimdling. Tlre vaporisation of liquid by isobaric tennnperature decrease, simown in Figure 1.15, or by isotimermal pressure increase is known as retrograde vaporismution. TIre vapour-liquid phase diagrannm of In typical mnulti-connponent system, Figure I . 13, describes lIne helmaviour of reservoir fluids inn nnnost cmtses. Thmem are, however, exceptional cases. Weinaug and Bradly [171 observed an unusual belmaviour for a naturally occurring hydrocarbon ninixture as shown in Figure 1.16. Note that an isothermal reduction of pressure, e.g. at 160F,

0~

(a

0~

()

Tennperature.F Figure 1.12. Critical loci for binary mixtures. McGraw-Hill Conrnpunnies Copyrngirn 181 n~imtnperunissiorm.
Reproduced fronn

20

I. lI,use fle/nani,,,nr Iim,n,Ja,nie,ifa/c

/2

//iace

!/eIn,mC,ou,

21

results in an increase of the liquid volume after an imiitial norrnnuml helraviounr. A simrrilunm behaviour has also been reported [18] for a multieorrrponenrt Irydrocarhon oil, as showmr iii Figure 1.17. Note that the gas/liquid volumetric ratio increases initially below tIre bubble point, as expected. The trend reverses over a lmmnited pressunre range, prior to belraving normally again. The calculated gas to liquid ratio mn mrnoiar nermir is showtm also inn Figure I. 17. The ratio increases very gradually over line winnie tested pressunme rairge, wiilrourt airy peculiarity. The reason for the apparent disagreemennt between tlnc two plots, is nIne cIrzmmnge un molar volumes of the two phases.

1/

0%
t)ew tonnnl (n/rye

207,.

a
0

K=(Fm-459.67)/l 8

Mta=0.006895 psia

Figure I . I 6. llnasc uhnmngmmmmmi of a hnynlm ocmnn horm nnixtumre SPF Copyriglnn. Reprnxiuccd rrncnn Ii 71 wnllm
a~. I
/1

Sm,

fl

t4

Figure 1.14. Retrograde condensation at constant pressure.

Volnnnmnc Ranio

1.2
0~

-.-

But,hie Poinn 85 % 2.
70%

-5

S.

5%

Curve
SI

I I)

tni,,Ic R,~ri,,

Liquid %
0 U

0.8 2(0 Crnuical Poinn ~ ~ -o

0

265
trcssure

0,

ci

270
pa

27.5

Fngunne I . I 7.
(PyngIni

Varnuni ions of gins In liquid rmmtio by reduicimrg pressure below humhhlc poinnt Rnpr~~duccd tn,,nnn (81 wnltn pcnnnnissll/nn.

pp

Temperature

>

A single plnunse Iryuhrocarbonr reservoir flumnd may fornn rrnore (Iran two phases uhumimrg depletion. Sohnd. or seninisolnd pbrunses, such as aspiraliemies can formrn at sonume conulinions. A bnighn pressunre gums, rich in lrychrocarbotm connpounnds of different lnonrologous series, mnray corndense
two imninriscible liquid pirumses, eaclr rich with one structural type of molecules. Gas mixtures rich inn C02 or ii2S at low terirpermniures earn formrn a rich liquid phase imnniccihle svifin tine

Figure 1.15. Retrograde vaporisation at constant pressure.

hyulrocunrhomi ricin coundensate phase.

22

/.

//,ace Refnrn,n,,,r F,,,nthnnn,,nwla

I. .1 (Ia.uuifin-a iron of Re.uerroir Fluids

23

1.3

CLASSIFICATION OF RESERVOIR FLUIDS

A

1he typical phase diagram of a reservoir hydrocarbon systenni, shown inn Figunme I. I 3, cummm be used convenieirtly to describe various types of reservoir fluids. A reservoir cminnnuminrs gins if its temperature is higher tlran the fluid critical tennperatnmre. otherwise it comrtains (nil lIne depletion of reservoir will result in retrograde condensation inn tire reservoir if the reservoir temmnpermmturc lies between the critical temperature and the cricondentherm, whereas no hiqumid will forrmm if it is above the cricondenthcrm. The oil in a reservoir witln a tempermulunre close to its critical point is more volatile than that at a lower temperature. A smmiall redinction (if pressure hehow tine bubble point, in a reservoir with a temperature just below the fluid critical tennperatumrc. mumy vumporise half the oil volume. It is evident, therefore, that the location of reservoir temripemature (in tIne phase diagram can be used to classify reservoir fluids. Time temperature of a reservoir is determined by its depth. The phrase behaviotir of mm reservoir fluid is detennined by its composition. Typical comnnposmtmomns of variotms classes of reservoir hydrocarbon fluids are given in Table 1.2. Critical temperatures of heavy lnydrocminboirs ate higher than those of light compounds. Therefore, the critical tetrnperatunre of Irydrocarbonn mixtures predominantly composed of heavy connpoumids is higher tinmnn ttrc nnunnnmmmml mmnmrgc (if reservoir temperatures, and these fluiuls bebnave liqumid-Inke, i.e., oil. Wlnercas tIne teirnperatuimu of a reservoir mainly corniposed of urethane. witlr a critical lemunperaturre of I 91)6 K, will be higher than the mixture critical temperature. Table 1.2. Typical compositions of various reservoir fluids. Cuniponeni. Moie% D~yGrsGasConden.saieVotaintcOiI
N2

Critical
Poinl
a
0,

Iemnnperaiurc

>

Figuire I .18. IInasc dimugrutmos of segregumted oil and gas phases in (lie vicinity of gas/oil contact. lIne reservoir fluid is prouluceul annul nnieastnred at tine surface as tire stock tank oil and gas at stmmnndunrd conditions, as sitowmr sclnenimntically in Figure 1.19. As the material balance equations relate tire tnro(luicc(I liuuiuls to those inn the neservoir, tIre initial producing gas to liquid volumetric ratio is considcrc(l mrs lIre nm(nsl iunpiinlant innuhicator of Itme class of a reservoir fluid. The gas to oil rmmtio. ((.)R, is rnrost corrmtmn(nnrly nlcinimed as tIre nnurrmber of cubic feet of the associated gas produrced at stamndand comrditionms lien barrel of stock tmnnk (nil in tire Field umnits. For gas c(nnndemrsatc Iluids, wlrere tIre pnodrnced fluid is predomnninantly gas. the inverse of the above ulcfimnitiomi. kmnown as line courdensate to gas ratio, CGR, is often used, Gas Gas

C02 Ct
C2

625 234

81.13
7.24

029 1.72 79.14

7.48

012 1.50 69.59

5.31

i)kmck OnI 0 n6 0.91 36.47

9.67

C3 iC4 nC4
iC5

2.35 0.22 0.35 0.09

0.03

nCS C6 Cli

3.29 0.51 1.25 0.36 0.55

4.22 0.85 1.76 0.67

1.12

6.95 144 3.93 (.44

1.41

0.61 4.80

1.22

4.33

16,64

33.29
Receuv,nir t)iI

When the reservoir pressure falls below the saturation point, the phase diagramrm of tIne origimimtl rcscrvomr fluid is no longer valid. Gas and liquid phases are produced at a ratio differemrt to thmnt in the original combined state, resulting in changes of the overall comnrposition. llie gravitational segregation of the two phases with different densities will also inimihit the contact between the phases. hence preventing the achievement of equilibrium throughourt tIne reservoir. In a hydrocarbon reservoir consisting of a gas cap and Inn oil coiunnn two separate phase diagrams, one for each phase can he connsidered. The two phases are both suntunratcul. with tIne saturation pressures ideally equal to the reservoir pressure at tIre guts-oil contact as slnowim in Figure 1.18. 1-lence, when a saturated gas reservoir is discovered, an oil coiunnn below it is generally expected. Similarly a saturated oil reservoir may strongly indicate tine presence of a gas cap. Petroleum reservoir fluids can be classified according to various criteria. Althroumgh identifying a fluid as gas or oil is adequate in most phase behaviour studies, it is more comrnron to classify the fluid in accordance to its volumetric behaviour at the reservoir and surface comnditions. This approach yields a few set of formulations, known as material balance equations, which can be appropriately applied to each class of fluid for reservoir studies.

Figure I .19. Sclnenniatic diagranur of slabilising produced oil as stock tank oil and gas at standard conditions. llne stock tank oil gravity generally varies significantly for different classes of fluids, hence it cmiii also he used as air inrdicmrtor. lIne gravity is expressed as API degrees in field units, AIl = ( 14 ISIS,,)

131.5

(1.17)

where S is tire stock tank oil specific gravity, or relative density, to water at 60F(288 K). 0 Thre concentration of lremmvy fractionn, C ,in reservoir fluid correlates reasonably well with 5~ GOR. As tire stock tmmmmk oil is minosnly connrprised of tIns fraction, it can also he used as an inrdicmmtor of tIme reservoir fluid type. Figuire I .20 shows that an initial producing GOR of 570 v/v (3,200 SCF/SIB) and 12.5 nmole% (~ are valid boundaries for gas and oil systems [19], as simuwn un Figure 1.20.

24

L Phace fie/,a,i,,,,, F,,<v/a,anr,ntrnI,

/ .1

(/ar.n/ka/rorn of Ren ,,io,,

25

Wet Gas

too
Gas (o,n,ie,,sa(e U Ojt a U
0

700 500 em 30(t 7.5

iO0

A wet gins is mnnunrmiby eonnihnoseul of nnuctimumnc amrcl other light conipoinents witir its phrase emnvelopc louunled entirely over um tenrnpcrmn(uire rminngc below ihumt (nf lIme reservoir. A wel gas, therefore, will riot drop-out condens~tc inn line reservoir during depleti<nnn, (1)10 (2), as shown in Figmrre I .22 The sep~nruntor conditions lie, however, ;vi(hin tIre phase envelope, prodtmcimrg sonmne connulensate at tine surface. Gas iicld.s nnn line Southern Noriii Seun mire gounui exanmnples of this type of res&rvomrs.

ice I,,,m,i (urve ci

A

m2 5

50

75

20(1

225
(7ntinil/
iu,nit

C + nnroie % 7 Figure 1.20. C GOR relation for typical oil amid gas connulerrs~mte fluiuis. (~nIrrcsyul tOni 7~ Pnnbticanion Inc. Reproduced from 1191.
dl,

li
(
.

V~t~.III

2
Rcstrv,,i r

~ Sepsu,nnon The most common method of identifying petroleurnsn reservoir fluniuls is to elassily Ilnerin mms dry gas, wet gas, gas condensate (retrogrmnde guns), voluntile oil aund black oil. lenruperanunre

Dry Gas
Dry gases are predominantly composed of ninetiraune an(l nonn-hyulrocumrhions such uns mnitnogenr and carbon dioxide. Figure 1.21 sirows tire phase (Iiuugrunnnr of mr ulry guns. [Inc hillase enrveiope is relatively tight and mostly located below the anthienrt temperature. Note that the gas remrrainns single phase froni the reservoir to time separator conditions. Water, Irunwever, mmnuny conndennsc at the surface conditions dine to the gas cooling. Fyi tests in tine laboratory are iinnnitc(i to tine guns compressibility measurement.
Reservu,nr

Figunre 1.22. Phase diungramnn of wei gas. As inn connulenrsunle is mmmcd inn tine rcsenvomr, nnnaneuial b~nlumnrce equninlionns for a dry guns are equullly sunnt~nlnlei~nr un wui guns [Inc ninly lVi test ncquninud unl line reservoir comndmniuinns is the gins connrpressibilnty nnneumsurcrnrennn . Supunruutor tests urru gcmneral ly comnuluncled to detennnnnre time unnniounnnt unnd properties of tine condensed pirase at tine sumrf~ncecomnulntions. A wet guns reservoir is connnnnnounly produmeeci by siminpie halow-dowur, simnnilunr no a dry guns, ins rio cnnndcurs~nteis fornmtuci mm lime neservoir. Iroducinmg gins to conrdensate ratios are typicunlly above 10,01)0 v/v (50,000 SUI:/Sh13) mid rcnrrunini constunni dunring the entire life of tIne reservoir. FIre conrulenrs~ntecolour is tnsnrumhly w;nler-wlnite with a low specific gravity which rennrainrs ninebmuingeul (itmri ing lint reservoir pruidurci iorn life Gums (~ondeimsalt
A

( rni,c,nt
I,,nni

A iypnuunl guns cn,nndunnsuntu phunse drungmuntni is slmniwni inn Fngunrc .23. FIre presence ot lrnznvy I my dune ~nn hu,mrs exlnunni uls lIre pl rinse envelope re lint sc to a wet gmus, hence, n Inc reservunm tcnnrpem unitmue lies inctwcunn (lie crinical poini and tine crieonndemntlnernrr, The gas will (In 1i~ourn I iu uiid 1 by in2lmogiaule eunnndenisationn inn (inc reservoir, when fire pressunre falls below ttne dew point, fmomnn 1)10 (2) inn Figunre 1.23. Further connulemrsation fronni tire produced guts also occurs at sepumrumnor comiulmtiomrs ulnie to coolming.
///~~ Separator

0,

Temperaimnre Figure 1.21. Phase diagram of dry gas.

>

[lie uuinmoimnt of siternlinlly coniuiennsahle imydrocmnrhorrs irr the reservoir increases with lime 1 as Ineavy connpounrds shifi the critical temperature towards the reservoir ricinnress of tlne guns. temnnperiiture. WIncnc~nsu gas wiihn a cricu,ndenitlnernmn ntear the reservoir leniperature will behave very mmntnclr like a wet gas. Gas to Iiqumid ratios range helween 570 to 30,000 v/v (3,21)0 In 150,01)1) SCF/STB)l l9J. For practical punrposes In gins condensate reservoir with a GOR of mnhove 1(1.000 v/v (50,000 SCF/SFhS) earn he treated as a wet gas. lhe producing GOR initially remnnunimrs conrstamri unntii tire reservoir pressumme falls below tire utesv poinl and increases thereafter. For guises willr G()R of unhove 20,001) v/v (100,000 SCF/STB), time condensation in reservoir

26

I Ihone J(,/n,n,no,,r 1,u,u/,p,,,,mti,I.u

13 (l,,s.snfln,rrr,n,n of Reuem,,,ir f (nod,

27

has negligible effect on the properties of produced gas. hut it can noticeably reduce the gas recovery rate.

very seldom cmrrrieut out beemunse of shortage of gas. Partial pressure maintenance is more cninnimon to minimnnise the losses of condensate, where it is economical to do so. In recycling opcrmitionns inn(ennneuliumle mnnnd heavy connipounds of the produced fluid are separated and the reirraitmimig leaur gas is imr~cctcd hack inrio tIne reservoir. The recycled gas which is predominantly rnrellnmnnre, not only reduces tire prcssunre decline rate, hunt also makes the system leaner. The rennroval of a sufficient amount of heavy hydrocarbons from a gas condensate reservoir may iulcmnhly shift the entire phrase uhiagramni fartiner away fromni the reservoir temperature to form a wet guns reservoir. [Inc reservoir cann tlren he produced by blow down without much loss of vuultm~mhlcliquid. lInt tIne lack of conmiplete uiisplmmcemmnent mind mixing of the recycled gas with tIre inn-situ fluid limnnits tine stnccess of tlne ~nhuveoperatiomr. Ilowever, time liquid loss by depletion will he lower mufter recychnng.

a:

a Temperauunre >

Dew Point

:9
Figure 1.23. Phase diagram of gas condensate. lhe comrcentration of heptanes plus is generally less tinan 12.5 mnioie% inn gas conndemmsate fluids as fluids containing tniore than tlrat alnuost always behave hiqunid like in thne reservoir. Exceptional cases with condensates as high as 15.5 mrnole% and oils witir as low as 10 tnmole% of heptanes plus have also been reported 1201. The condensate colour can be water-whrte or dark. Dark condensates usually irunve relatively high specific gravity and are associated with higlr dew point gases. Comndeinsale specific gravity ranges between 0.74 and 0.82 (60 to 40 AP[),although values as high mms 0.88 (as how as 29 AP1)have been reported 1211. Material balance equations develuped for dry gases can be used for a gas condensate reservoir as long as its pressure remains above the dew point. A composilionnal material balance method should be used below the dew point. It is commonly assumed that the condensate fornned inn reservoir remains immobile due to its low saturation, and is mostly non-recoverable. Recent results [22]. however, have indicated that the condensate can flow even at very low saturations. Figutre 1.24 shows a common characteristic of gas condensate fluids. TIne liquid drop-out reaches a maximum and then decreases by vaporisation during pressure depletion. This behaviour may imply that when the reservoir pressure decreases sufficiently, Ihe condensate will be recovered by revaporisation. However, by the time the pressure falls below the dew point. the original phase diagram is no longer valid as tire system composition changes during the production period. PVT tests simulating reservoir conditions wilt he described in Cbmmpner 2. Condensation and loss of valuable compounds in reservoirs could be avoided by maintaining the reservoir pressure above the fluid dew point by gas recycling. lIn practice. Inowever, this is
a
-I
N

Pressure

>

Figunre 1.24. Liquid (Imp-out behaviour of gas condensate.

Volatile Oil
Volmmtile oils hmmve mnnlnury c(innnnsnon femniures with gas conrdensates, but as they contain more heavy connrponnnnds Itrey hehurve liquid-like unt reservoir conditions. The phase envelope of a volatile oil is relmrtivcly winIer nlnamn llnmit of mm gas contlennsate, witir a higher critical temperature due to its larger conccntrationn of heavy commnpotmnrcts. A typical volatile oil phase diagram is shown in Figure 1.25. Tine reservoir teninperaUtre is near the critical temperature, hence, volatile oils are referred to as near-criticuni oils. Note tlrat iso-volume lines are tighter and closer near the bubble poiurt currve. A snrmmnli reductionn of pressure below tIre bubble point vaporises a significant fraction of the oil. hennce the name volatile oil. Separator conditions typically lie on how quality (iso-volunne) lines. Initial producing guts to Iiquiul ratios (GOR) of volatile oils typically range between about 310 mmmd 570 v/v (1,750-3,200 S(FISTiI) (51. The GOR increases when the reservoir pressure falls below tire bubble point durinng the reservoir life. Tire stock tank liquid is coloured with a specific gravity umsununhly lower than 0.82 (higbner than 40 API). The specific gravity decreases (luring produmctioun below time hunhble point, particularly at high producing GOR, as a significant iiqsnid production is due to condensation of the rich associated gases. Saturation pressunres of volatile oils are high. Gases produced below the bubble point, tlmerefouc, mire quite riclm mind heirmmve as retrograde gases. The anrount of liquid recovered from tIne gas constituhes a sigmnifmcant portion of the total oil recovery. Compositional material balance nretlmods shroumld he applied generally to study volatile oil reservoirs.

28

1 P/nnn.s, Reluin,on,, I ,u,nd,n,,n,nnnnl.n

/ I ((ouirfin(mtnnn of Resernoir b/nod.,

29

(rinncal tnimnt

tPumnnrcurvt
A

99...
U

,....

~ 20

60
40

,
Scpar~nmnnr
0
.

lnntnunl prurducmng GURu mire less tlnumnn about 310 v/v (1,751) SCF/STB). The GOR may ulecreumse mmritn~mhlywiienr (tie reservoir pressure fmmlis below tire bubble point, as the evolved gas renmtumnnrs mmnmnrnn hrle at very low s~nIunmuninouns. The GOR. Ilrenm increases slnamply as tine gIns to oil nnmohn In ty rat no wit Inn nn lire reservoir varies inversely witlm tire viscosity ratio, winicir is typically of two nnmuiers ot inniugirnurde. In frunctinred rcservonrs, inowever, svhere Ihe fractunres provide a gounul comnltnii for (lie gins Inn risc by grinvi ty, line GOR uicui inres tlrrouglrount the produncinng life of tine field, mns mug as lIre pressure keeps dcclinnimng annul no gins corrinng imnkes place. The stunt-k tuunnk Inqunnd ns dark witlm In specn Inc gravity lmmginer Ihtmnmn 0.80 ( lower ihrmtmr 45 niAPI ) 12(11. lIne vumm iuntionr of tIne specific grurvniy is relmniivcly snmnahi, in comparison with that of volatile oils. dunring tIne reservoir proutumctionn life. lIne s~nlur~ntnumnr li~CS5OiCinf black oils is neluntively knw. Commlrihunfluin ~if inemrvy conrrponnmnnis present inn evolved gunses inn reservonr to tIne total hiquunl recover) is miot signnificairt t lennec vnihnnnnicnmnc nnma(eriai halunnicc eqnmintrunrrs, wInch trea( the reservoir fluid uts a twin conmnponent systcnnn. n.e., oil annul guns, mrnuny lie soiincienii for sonrne reservoir siunnlies. Imnuiceul. ~nstincne is nun de In in inc hounnrdunry betweein hi unck unnnd volunti he oils, the uncceponbi hity of resur its uihtmuinud by tire volunrmclmnc nnrefirout is~ip acincunl cm nieriomr for uhistnnguislninng between tIne two types.

I rc,ervn,nr
Terniperunnunre >

1.4. REFEREN(ES
I ltnmnt l.A. M Pet nolutninr Geunclrunrristry mnnd Geology, W .11. Preennnutn umnrd (ui, Sunnn
Fr~nnrci,scni (1979).

Figure 1.25. Phrase diagram of a volatile oil.

Black Oil
Black oils, or ordinmnry oils, mnrc the Inmost connnmnromn type of oil reserves. Flit niunnnne dunes riot reflect the colour, hunt to uhisimngunnsln nt fmorrn line vunlunt he on I. Tire iii I is gemncn ~nl iv uuninuposeul of more tharn about 20 nnole% heptanes amnd lreumvmer comnipounrds. its pimunse emnvulnnpe. nineneiniru, is the widest of all types of reservoir fluids, witlm ins critical iennnpermnturme well mnbnnve tire reservunir temperature. A typical black cml phuuse dimrgrmnnnn is shrowir mn Figure I .26. lIre qinumhity limes are broadly spaced at reservoir condmtrons witlr seputruntor connlmliuins lyinmg (inn relatively lnigbn qurinhity lines. The above characteristics lead to a low shrinkage of oml wlnen produrceul.

2. Enigiunnid. WA. unnid M~nckcnniic. A.S: Gcologiscire Runrnulschnuntn, 78, 214-237(1989). 3

.

Gennchenmnistry nif Petrolcunmi Reservoirs.

Nd snurn. W .1 .. 958).

Pcin unleunnmn Re finery Fnigrnreerinrg . 4th Ed., McGrawIl ill, New York

4. Inistniumie nil letnnulcnnmu: Methinids for Amraiysis,nnmd Iesiinng .Tine lnnstinntc nil Petnok.uumnn. Johmnn Wiley unnrut Sonns, New York (I 984). 5. McCuminm, WI): line lroperlies of Petrolennnrn Fluiuls, 2rrd Ed.. Penmnwell h3ooks.lumls~n (1990). 6 Suit-Inunnrenr. AN [Ire (Iicmmnncunl (omislmiunenits of letrnuleunnni, Reinihoid Pub. (o. (1945).

Critical point
A

((

7 I urn Icr, S. R , uund ApI inn, A.( : Rcservonr ( coulicnmnisiry : Men Irods. Appiic:nt mmmv umnid ()ppomtunniniics Inn l:nngiunnrd, WA. unird Cunhilt, J. (cdv) lire ( euncireimtistry nil Resemvonrs euil . Sine. lunhhicuilnoin ( 1994). 8. KaIz. I). ci ~ni: I luunulhook nil Natuirunl GaS Ennginicerinrg (1959). 9.
)klunironmiun (1980).
,

Sn a
I,

/
~40

McGrurwl Ii II Book Coin npanny I nil~ui,

20

0~

( ;a~ lrou-essors Suppliers Assnicmuitmuinn, cii: St Fmngiincer inng I )unlun ltunuik,

Separator
S

.e
,,..~

(it)

T reservoir
feinipm.riiture
>

hO. Browmr (1G., Kum(,, t).L., Oberfell (1G. and Alderm R.C:Na(mnrai Gasohimne and (lne Vumh~mliie I lydmocarhoirs, NGA of Antreric~n,lulsa, (1948).
II. Pit,en. KS.. l.uppnmrannmn 1)7.., Cmii, R.F. Jr., llumgginms (.M mmmd leterscnn, t).E: lhne Vohuunnieni it unnnullhemimnun&lyniumnnrie Prnipcrtics of Fluids. II. (onmipnessilnihity Fuiutors, Vunpumm hiressunnu. annul Ennt i nipy iii V~npumris~nnionr. J. of the Anrierieunn (lmenrricuui Society. 77. 34333441), (Jnniy 5. 1955).

Figure I . 2(n. Plrmnse (Iiuigrannn of mu hI mit k oil.

30

/ //na,ce Itehnnnn,,nnn

,nnn/,nn,nn,ntnnln

I ~

n (Prune,

.31

12. Lee, B,t. and Kesler, MG:A Generahised Thermodynairnics Correluntiomm Bursed on Three-Parameter Corresponding States, AIChE J.,2t No.4, 510-527 (Mmmy, 1975). 13. Rackett, H.G:Equation of State for Saturated Liquids. J. Chem. Eng. Data, IS No.4. 514-517, (1973).

1.5. Esttnmrmrte tIne eritmemnl tennnpcrature aird pressure of a mnnixture conmnposed of 55 rnole% ethane amrd 45 nmmolc% nornmnmii Ireptaune.

14.

Spencer, C.F. and Danner, R.P: Prediction of Bubble Poimit Pressure nil Mixtures, J.

Chem. Eng. Data, 18, No.2, 230-234, (1973). 15 Spencer, C.F. anti Adler, S.B:A Critical Reviesv of Equmationns for Predieimnng S~ntrmr~nted Liquid Density, J. Ctrennn. Eng. Data, Vol. 23, No. I, 82-89(1978). 16. Yamada,T. and Gunn, R:Smmturatcul Liqumnd Molar Volunmnnes: tire Rumckenl l/quiuu( mum, .1 Chem. Eng.,Data. 18 No.2, 234-236 (1973). 17. Weinaug. CF. and Bradley. 11.B: Phrase Behaviorir of um Natunral I lydroeunrbomr Systennr. Trans. AIME, 192, 233 (1951). 18. Danesh , A., Xum, D., mind [odd AC: An Evumlunationi of Cunhie Equnmutiumni of Stunle Our Phase Behaviour Calculations Near Miscibilmty Comrdnlions , SPE/DOF 20267, Pumper Prcsemrted at the Scvetrthm Synrposiumnr onr Eurhrmmnceul Oil Recovery, TumIs~n.Oklunlnoinuun, April 22 25 (1990). 19. MeCain Jr. W.D,, and Bridges B: Volatile Oils and Retnogrurule Gases-Wlmuits tlnc Difference?, Petroleurmn Engimrcer Imiternuutiourmnl, 3536 ( Junmi.. 1994). 20. Moses, P.L: Engineering Applications of Plrase Bebraviour of Crunde Oil ammd Conmdensate Systenrs, JPT, 7 15-723 (July, 1986). 21. Kilgren, K.H: Phase Behaviour of a 1-ligh-Pressurre Condensate Reservoir Flunid. JPT. 1001.1005 (Aug., 1966), 22. Daneslm, A., h-lenderson, GD. and Peden, J.M: Experimental lnvestigatinitn of Criticunl Condensate Saturation and its Dependence on Connate Water Saturation, SPE Res, Enig. Journal, 336-342 ( Aug.. 1991). 1.5 EXERCISES

1.1. Calculate the vapour pressure nil normal ulecane at 355 K. unsung: (a) tIre Cox chart, (0) tIne LeeKesler cquatiomr. (c) a Ii nrear re hat urn bet wcemi IImu iiig;in illmini nut vumpoum pressure and inverse of temperatunre eonncctnnng tire nomnniuml hoihinmg point mind lIne critical point. 1.2. Plot tIre vapour pressure vs. tcnnpcrattmre for thre following eonnpounn(Is mm lire redunced scales of (P~P~) and (lIF~): nietirane, normal tmexmnne. hcnzene, nnorniumml decunne. mmmd cicosmnnme. Suggest a physical property, such as the acenlrie factor, or critical cornrpressihiliny fmmctor, as the third parameter in a three-parameter corresponding state model for lIne vapour pressurre 1.3. A cylinder contains I kg of saturated liqunid butane at 385 K. pressure after consuming 950 g of butane, What will lie tine cylinder

1.4. A 5 litre cylinder contains 1.5 kg of propane at 393 K. Estinniate its prcssumre. Flow mrmumcln propane will be left in the cylinder when the pressure falls by half.

2
PVT TESTS AND CORRELATIONS
Aecnnraie mind neiiunble phrase helrunviuutmr mnnrd voltnmnuetric data are essential ehennemrts for proper nnramnagemrnemmt of petroleunnn reservoirs. lIre inforuriation is required to evalumate reserves, to develop line optinmnmmnir recovery p1mm, anul to determnimre the qumurlity mnncl qunaiity of produmced hiumiuls. Most reservuuirs mire produrced by depleton in wlmtch tine reservoir pressunre uleelnrmes Its fouls mire rceovemeuh. lire reservoir temnnperatumrc stays practically cotmstanr( inn nmrosl recovery nnnetlnods . tine nnmainr variable Ihumt ule(cnnuiinnes line huuIravnonrr nil t hunids, ummider reservour unnmnulitimnmns. uinnnimug dupleninnnu is. iimuneimnrc, (Ire nescnvnuim pressunne. Hems-c, rcla(nvtIy snnnmplc tests whit-in sinmnunlate mtumnveny himuinesses mine cnummdmmc(cnl by vuuryinmg tIme flnnnd pmcsstirc. lIne mmcmiii enmmpl mmivis is omr tIne vunlum mmmetrme ulmnlmm ~nttIre reservonr mnnnul vu riace 1cm n pcraturres. Imenree time nanmre (pressumrcvolumnnetennnper~mlmrrc)PVT datmn. 1mm tIre simnnplcs( appioincln of predietinng tire PVT (Immua, tIre reservoir oil is cnunsidtreul tim he connrpnuseui of two pseunulnu cnnmmrponremnts, i.e., gmns mmml tiil. These pseumulo conniponents, unnc iulemrtilied my flunstuimmg lIne rcservonr llummtl at lire stanrulumrd comrdnl ruins, and chnar~mcternsmmng tine sepanmmteul gmns annul mu phn~nscs by tlneir specific grunvity and mmnuulectmI~ur wengirt valunes. (onmrposiiiomruul d~ntmm on tIne procluncetl tIumids are nnaimrly determnnineul for timeir mnpplncationns in Irydrocarhorn processinng. lIne prune innfonnumntionm fronrr PVl les(s mire tine ratio of pirase volunnnue at reservuimr condihonns Inn that urt surface coniuhitionns, mind the solumbihity of guns in oil, lhe informationi is generally smnflicicnt inn stnrulies nil hlumck oil reservoirs, and the ~npproach is referred to as line black onl nnnethrod. Ctnnrnpositional stumdies, where uletailed innformaniomn on tIre fluiuh eonstitucnnls mire used to cslinnrmnte flumiti properties, uure oftern conulueted ftnr gas comndensmn(e mind volatile oml reservoirs. Onnly in special cases such as gas iniection or miscible ulmsplacennent Ore conrposrtionral approaclr is unseti for black oil reservoirs.

34

PV7 lesn.c

and Correinmnm,nn.c

2.!. Flnmid Sampling

35

A compositional phase behaviour model, in principle, is capable of predicting mmli tIne PVT dunta. using only tire composition of the original reservoir fluid. The models, however, are reqmimel to be evaluated and tuned against the measured PVT data prior to bemng used in reservoir studies with confidence, as will be discussed in Section 9.3. The compositionmnl method. which can provide reliable information rapidly using advanced conipumlers. ms heennnnninng mmmore popular. Empirical correlations annd charts, nrainhy renniniscenee of dumys wlneur inmmnul calculations were the norm to predict PVT data, however, are still used. In this chapter phase behaviour considerations related to the smmmnnplnnng of reservoir fluids unnc described. The most conmmonly conducted PVT tests mire detailed next - Selected ctrrpiriemml correlations, to estimate PVT properties fronr himnnited Inelul data, are also givemr. hlresc correlations have bccn genermmtcb over years, insinig labonuntomy dmmtun. lhey wene nnnnustly developed nriginally in grmupinicmui forums. Inn tIns honik tIre rrmmitlmennrmml ncmnl expnessnmnuns nil line correlations are presented in preference o tlneir originral graplmical brims, lIne eoinelatmunms unsc field units, and are reported as such in this chapter. A comnvemsionr table is giveun inn Iahle A.5 nun Appendix A. 2.1 FLUID SAMPLING

bulk as the pressure declines during produmction. As the depletion rate is low, the advancement of the two-phase region is slow. Ilence, it is reasonable to assume a quasi-steady-state condition around tire producer. witlr nnimnimal clnanges over a short period. At such conditions, Ore overall composition of the gas-comndensate mixture flowing into the wellbore is the same as tlmat flowing into the two-plnase regionn, as no condensate mnccumuia(ion occurs in that region. Ileurce the reservoir outflow, if collected properly, should represent the original single phase reservoir fluid,

well

I wnittm~nsc~ ~ Rcgmnnrn I
,~rTressur~

Intinuw

Gas 7~nc

Reservoir fluids should be sampled as early as possible durimig time produnctnuor Infe nil mm reservoir. When tine reservoir pressure falls below the inntimul smmturmmtion tiressumre tIre hydrocarbon phase forms two phases ol gas and hiqumul. TIre mnrole ratno of time two plnases flowing into the well is not generally equal to that formed in tIne reservoir. liennee, tire collection of a representative sample becomes a inighly dennianuhiung, and in mommy cases air impossible task. The sample can be collected either as a single phase at tIre bottonnr hole. svhenn mIne pressure is still above the saturation value, or at tine surface. Time bottonmr Irole sanmiplcs are usunmnlly collected during formation testing, prior to production. Surface sampling is conducted (in producing wells eitbner at the well head, as a sample representing the producing mixture stream, or as separated gas and liquid samples out of the separator(s). As long as the reservoir pressure has never been below its saturatiomn pressure. and mm snnnglc phase sample flows into the sampling bottle, tine chance of collectnmng a represenrtative sample is high. Producing fluids, however, are generally at two-phase conditions. Ihemnce. tIre sanvnplmng procedure should aim at collecting both phases at such conditions where tine suhsequnennt reconrhination provides the original reservoir fluid. Samsipling proceulures Inave becnr discuissemi in details [I-SI. First, it should be ensured tlrat representative fumitls are flnwimrg ount of tIre formnaniomn, by properly conditioning the well before sampling. Next, fluid sunmnnpies slnounld Inc collected from all coexisting phases. umnd recombined urn tIne prodrncimng ratio. Sunnnrplnmng fronun an oil reservoir, pamnicumlarly ann unuiersaturated orre, is relmmtively a mmmcli simpler tuisk tlnmnmr tlnumt from a gas condensate reservoir.

t)istmnurce Figure 2.1. Schetnmatic dimngram of two-phrase flow around wellbore, Tire validity of the above assumnmrption, can be evalunated by numerical simulation of the flow umemnr the wehlhorc unslung a comumposi(ionmtl model (6], as will be described in Section 9.5. Sumniden cimanges of suite will disttrrh tIre steady state conditions and the outflow composition. It is advismuble. ilnerefore, to unrmmintmmin the rmmte pmnor to samnnpling. Producing tire gas al mm low rate to unrmmintmmin the bottotin hole pressure above the dew point can ensure tine flow of single phase gas inrto tIne welibore. It is imperative, however, that the well flow rate remains above a nninimunn value for tIne continumal up-lifting of the condensate formed witbnin the wellbore. lire liquid plnase is trmmnrslcrred rip tire well partly as entrained drops in the gas core, and partly as a Imlnr on the wmnll by tire gmns slnemrrimng effect (annular-mist flow) , TIre transfer of liquid between time luhin mmmruh ulroplets is a countinumous process along the liquid path up the well. When the gmts flow ruute is reuluiced below a mnnimminrnumm value, tIne energy transferred to the liquid by the llowimrg gIns mmnumy nmnt be stifficient to carry the liquid. Then, the direction of liquid flow in the Inlmrn is reverseul amid tIme cmmtrmmincul ulrops frill hmuck, hour resultinrg in well flooding. The mmii iii mnnunnnm flnmw rmute 0mm unnmrti nnnmml menumovmnl of I iquids (eomidensmite or water) by tIre flowing gas cunnr be dctermniiired by minunlysing tIns fihmnn 11mw mnmrd time emrtrmmirncd drop movement. lumner et al. 71 ulcvclopeul a mmmeehimnunistic two-plnmmse flow mmdcl amid mupplied it to the removal of liquid in a gmms well. hIre mmumthnors cnnnmrparcd tIne nnrimniisnunnnn gmms velocity required to lift time entrained liquid svithr tlnumt for trumunsfemn inng tire Imhnnn un uward, mnmnh councluulcd thrat tIne former svas tire controlling 1 hinurit, Tine mrrajor forces whiicln detenniine tire velocity of a liquid drop are the downward gravity, and tIne umpward gmns drmrg. Tine gravity force is determined by the size of the drop and the liquid-gas demisity dilfereunce. wlneremus tIne drmmg force is domiurated by the gas velocity and the physical properties of tire two phrases. An inncmease in the gas velocity increases the ratio of tine drag force to tIne gravity force. Tunmer Ct al. balanced the two forces and derived the following relation for the nniniuinmmrrn gas velocity to tnnload ttre well,
gnn =

Well Preparation
In oil sampling, if the well bottom hole pressure has fallen below tIne oil bubble point, the well is generally conditioned by a period of reduced flow, followed by a shut-in period of about 1-3 days. This lowers the pressure draw-down and raises the oil pressure, possibly above its original bubble point. The method is not .sumitable for a gas condensate reservoir. The pressuire build-up may vaporise tire condensed liquid in the reservoir into tine gas plrmnse In) form a gmrs condensate even richer than the original fluid, Unless, the condensation wmms limited ounly within a small zone around the wellhore, allowing the disposal of (Ire ricirer gas over a reasonable period of conditioning, the collected sauniple will not be representative. The formation of condensate initiates arotmnd the wellbore. where the pressure is al its lowest value in the reservoir. Figure 2.1. The two-pirase region gradummilly grows into the reservoir

2.67 & (p,, pr)

/4

Pg

(2.1)

36

2. IVT Tn.ctn

anmd (orrelatinn,i.

2.1. Fluid Sannpling

37

where V~,,is the minimum gas velocity, rn/i, cn is the gmus-condensate inmterfmncial lension, mN/rn, and Po and Pg are the oil and gas density, kg/mtm, respectively, all unt lime well mcmi conditions. Turner et al.(7] used the following average values in tire above eqummmtion, tm 721 kg/inn (45 lb/flu) Condensate density Gas-condensate interfacial tension: 20 nnnN/nni Gas specific gravity: 0.6 0 Gas temperature: 322 K (l20 F) and proposed the following equation, Vg~=i, (l0~P)t/ 4
.

flummni properties arc unsed to mnemtstnne thre produnction rates, tine reported ratio shoumid be adjuisted usmng tIre vmmlues mnncmmsured in time laboratory, instead of tine approximate data unsed un tine field. Wlnemn time reservoir flumid is satunrated and the compositional grading within tine reservoir is nnminimal, see section 5.4., lIre pnesstmre-tennperatunre at the gas oil contact identifies tire smmtumrmmtnon point. Ilenice, tire nmnemmsurred saturration pressunre of time recommrhined fluid slnoumld be comurpmnre(l wmtin tire imelul dmnta. For mm recombined oil sanmnple. a mimalcir betweemr tine Iwo values indicates a representmmtivc sanurple. Whren the oil hunhhle point is known with confidence, it is advisable to adjumst tine reconmnhirrmmtiun rmuiio to achnieve it, instead uif relying omn the reported gas/hiqumid ratio, lire recomnuhimred sumnnrple is expected Inn reasonably represent the reservoir oil, mis (Inc hmmhble P~imntus semmsit ye to tIre gmns/hiqumid ratio ann increases nvitlr it, A nmrmitthm betwecmn tIre mm memmsmmred slew pnni trt in tine Immhommmtory amnd the field reported vuultie is desmratie, hunt shies not necessunrily inrdicale a represemntalive gas samnnphe. lhme dew poimnt 1 mmnmny irnerease mr (ieercmnsc by innenemnsinrg tine comndensurnc/gmms ratio, dependinng (in tine smnrumple. l ignnme 2.2 slum inns Ilit I iqun ii drumpounI lid nmmvinnumr of mm NontIn Scun gmns conrulcnsmmte at I Ire resurvoi lenmnpcimnttire prepared mml diffenennt recomnnhiinaiiomm rmm(ios, Note the nnrmmrked difference between lIre condensate dropout belrmmviomrr of different fluids, whilst tlmeir dew points are ainrost tine sattne. Ii is qumie evident tinmn( nmratchming line dew poimnt is mrot a reliable tmnclinod for reconnnhitning a gmms eorndemnsmnte samnmple. Time uric of ptrase hehmtviour models to evmrlurate and imnrprove tine fluid recommnhiiurmmtionn ratio is tlescrilicd irn Sec(ioir 9.5.

49

(2.2)

where P is the welihead pressure in MPa. TIme mininrrtmnr gmns veloenty can be cnnnrverte(l Inn tIne gas production rate, knowing the tubing inside dimutnicter mmmnul estinnnmitmmng line gums cnmuiuinnessibnlity factor, Section 2.3.2

Sample Collection
Surface samples are commonly collected from lest separmilors. lIre oil (condcnnsame) mmnnul gmis samples must be collected as single phase fluids. The production rate of eacln plnase shoumld be monitored over an extended period to ensure a steady mtnd stmth,le flow. flre sepmnrator temperature and pressure, along with the producing gas/liqumid volunmetric rmtflnu mnrc reported to the PVT laboratory. The information is umsed to evaluate the integrity of collected samrmples received in the laboratory, and to use in the recomnmhitmation process. The condensate carry over by gas in In separator cmmn sigrniiicumntly distort nIne nmremnsurrud condensate to gas ratio. The effect can be serioums for lemmn gmus comndensmmte systemmns. Ann alternative surface sampling method is the collection of blowuing plnmnses in tIne tnrhinng mit tIme well head. A narrow tuibe, with the inlet facinng tIre blosv uiiree(iomr, is inrserteui tn tIme cemnnrc nil tire tubing. A two phase sample, consisting of llne gmns mmnnd enlrmmnned drniplcts, is eohleeieul tlrmuuumgin the narrow tubing into the saniphing bottle. lIne samnrple flowing is collected winO a fluid velocity in the tube equal to tine average fluid velocity in tine turhimng. llmis is to uuvoid preferential collection of gas or condensate becmmuse of their duffcrenrt (lenrsities mmmmui mnmomnncnntumium changes due to changes in the fluid velocity. Jhe method, known mrs the iso-kimnetie smnmnmpling [8), relies on the assumption that the comndensate is homnnogeneoumsly distributed inn lire tubing flow. The homogeneity can be improved by insertiung a nmiximrg sectionr mmlremmd of tIne summnnplinmg tube. Samples received in the laboratory are evaluated for Ilneir integrity, primarily by mmneasuring tine opening pressure and comparing it withn the reported sampling conditiomns, As tine collecied samples are saturated fluids, they often form two phases in the smmnnrphinng bottle (lime to cooling. The pressure of collected liquid samples are often lowered puirposely below lire saturation pressure to form two phases for safety reasons, to avoi(i excessive inressumre dunning transportation in case they are exposed to high tennperature. Any leakage fromni mm sainnphumng bottle containing a gas-liquid mixture will change the sample conmmposition. A kiwer opening pressure does not necessarily indicutte a fluid loss, as it cnuld be dune to tine tlmenmnumi contrmiction. This may he examined by heating the bottle to tine sampling temperature. A phmmse helmavinunr model can also be used afterwards, when the fluid compositiomn mnnd PVT dmmtmm are known, to estimate (he expected opening pressure, and to adjust the flunid conrposilion if a fluid loss wmms indicated [9). Further information on the use of phase behaviour models to evaluate and improve collected samples are given in Section 9.5. Separator samples are recombined in the laboratory according to tIre reported gmts/hiqumid ratio recorded in the field during sampling. When flow meters with coefficients depending on tIne

ml)

Recoin. gas/hiq
vnii/vnml

245
1401 1509

C 0 0

12

.. ..

n.m
0

LL

8 a
0

a ci-

1(1

20 30 Pressurre, MPa

40

Figumre 2.2. Variations of c(undensm(ie (lropoumt witlr undjumsting rccounrhination rmutio. Smnnnuphes tested imn PV1 lmnhormmtories arc blinds collected mit the bottonnr mole or at lire sumrlmnce, amid mmrc not neecssmnrily tine smmmmme mis tlrose present witinin pores nnf a reservoir. Sigmnilncamnt differimnees inn eonmnpnisiiionm bctwcenn produiced ilui(ls,mn(l core extracts irave been reported 110,111. Cone extmmncts ofnemn imndicate a richer fluid in lneavy fractions particuularly sunrfacc active mrmmmneniais. TIns cmnn he mnrostly due to the mmdsorption of polar compounds onto the core surface, wlnielm mnnmuke tlncnmn iimnnnnobilc shining convemntional smmnmpling. The effect of adsorbed material on tIne mrnunlti-pinase flow belravioumr of oil-water in pores is probably moore significant than on tIne PVT properties in mnmosi cases. Tlrc samples collecteul from a flowing stream, however, may trot he suritable for special phrase behaviour studies such as asplraltene deposition. A small amnrounnt of utnrecovered mndsorbed material is not expected to sngnnifncantly change the saturation pressunre amid tIne guns-liquid vohunnmrctric fractiorn nil aim oil sanrple, but the effect on the gas

38

2. PVT Testn mind Cmnrrelasimi,rs

2.2. !VT lenin

39

condensate dew point may be very marked as its phase behaviour is dominated by lhe concentration and properties of the heavy end. When different samples are properly collected from the same reservoir, tIne sannnples mnre generally expected to be similar. Inn saturatcul reservoirs contmuinnmmg Inn nnml conhmmmmmn muun(l mm gums cap, the samples collected fromrr eaclm 7.one arc expected to) he reunsonrmntnly in equnihiinniunmmn witlm each other at the reservoir pressure and temperature. Compositional grmrding dune to gravity mmmn(l temperature gradient may however exist in a reservoir, with samples collecteti from different depths being vastly different. The compositional grading can be very severe. resunltunng in mm column of fluid changing from a gas at the top to oil at the bottom, without any phrase boundary. Phase behaviour models can be ursed to evalunate tIne extent (if comnnpositio)nnal changes due to grading in order to evaluate tine sannples. TIne eonmnpositininnunl grmmditng is described in Section 5.4. 2.2

numnierically rcconrhined, using the rmniio of tIne separated phases. The gas and liquid phases are commonrly atmalysed by gas chromrnatography and distillation, respectively. Details on connpo)sitional analysis and various techniques applied to characterise reservoir fluids are described in Section 6.1. lime umhinive unnmurlysis mmppmuimmchm, kmnmwmm as tIre blomn-d,nnn,r nnicllmod, can give reliable results for large sanniples of highr pressure liquids, wlrere tIne error involved in measurement of the two plnmmse ratio is relatively small. For small sannples or high pressure gases, where the condensate volume fornncd by blow down is low, tine technique is unnrehiable. Fun!! stream .sannpIium~,.wlncrc a snnnmill anrronunnnt of inigin pressure fluid is directly injected into a guns cimrmnnnnmntmngm unpIn. Inmms mccci ved 5mm ire altennt ion mis ruin mmitcmirrutive to the blow donvn method inn tIre punst deemmnlc. lIre hmmsic primneiple is to flow tine Iniglr pressure test fluid throungbn a special vunl ye tin trump mm sununli qunmmuntity nil (Ire smmmnmple imr tIne valve for injection. The trapped high pressure lltmiol is tbncmm cxpnnscd tin hunt flnrw nnf mm cmmnicr guns, which vaporises the sample unto mu gums clrromnnmitmngrapln eonlummnnn for amnumlysis. lIns valve itself is generally heated up to help elutimrg Iremuvy conistitunenls. lIre samnnphing valve with the flumid trapped inside may be physically rcmnnoved fronn time sampling port mmmrd transferred to a gas chromatograph (20,21], or just isolmntenl from bite colunihihmitmmmm ccli, mumnol Ilnenn heatcoh up. lhe vmuponised sample is directed to a gums elm onmnmttngrmtplr IInronuigln a incateul trrunnsfer litre (221. lIme mulca of lumhh stremmnrn (dim-eoi) snumnipliung is quite innlerestinmg, particumlarly for the comnrpositional amrmnlysis of eqtmilihrunted plnunscs in PVT tests, wlnere the removal of a large quantity of a phase will distunrh tire overall connnpnnsitioum. Certain (nperationmal probleurms, however, have prevented its wide apphicationn. Altinotigln lIre smmmnnple vohunne actually injected into the gas chromatograph is very smmnuitI, of tIre order s)f nnicrnnlitres, a large amrmoumnt of the fluid is required to fill up time smmmnrphtumg loop system, winicir inclunles a nmumher of issilating valves, and most ofit is lost wlren tIme smmmmrphimng valve is memnroved onr Ircmrted, All tIne limes irave to be cleaned and evacuated after emnclr injection to repeal the analysis.. The introdumctiomn of a Inighr pressure fluid into an evacuated hinrc gemrcraliy rcsumlts in plrmnse clnmmmrges, hence, a large volunre of the fluid has to be passed 1 through the oop to ensunre mu represenntative sanmple for injection. lIre Imirge lonss oil line lest fluiti slurring tIre sampling ansi the problems associated with the tnmunnsfer limes mmmd isonlatinng valves Inumve becmn avoided by designing special vapour-liquid cells in wlmmclr tIne sannnphinng vmmlve cmnmn I-me imrstaihed directly onto) the equilibrium cell 123]. A small vnnlummnne oil line test flumid enters tIme vmulve, locked in, ammd the valve is detached from the cell for tine fluniol to be tramnsfcnresl to a guts chnrounnalogrmmpbn. lIuc test fluid does not flow througir the mnbnunve valves and, as tIne expostnrc nil a Imiglm pressure fluid to an empty cavity is generally mmcconmpamnned by mmmc phase ehmmnge, rcntrovai of a representative sample cannot be ensured, Repemrted smumirplimng is mint an eunsy onr a safe operation in the above arrangement, as the sampling valve Iras to be assembled to a high pressure cell kept in a constant temperature environment, flre ireavy constituents of the samniple in all the ah,ove methods may be partially retained between tire smmmnmpiing valve and tbne guns clnroinnatograph colummn and, as the injection volume is very smrnuull, tIne cnnnecnntra0on of these cnnnstituents could be highly under-estimated, Therefore, lIne nnncthnod.s hate been unwire suicee,sslunl inn nIne analysis of gases tinan liquids that have very hiemmvy counslitunents. A olirect samnmplmmng tcelmnnnqume inn whniclm a smmuhi samnipie of a hniglr pressure fluid in a narrow tube is pminclned by mnmr aumxmiimnry fluiol (snnlvennb) at the test pressure has been proposed [24] to avoid tIne above problems. lIne flow 0)1 tIme solvemnt directs the slim slug of the sample into a high pressume valve wlnicim has replaced (Inc injector of a gas chromatograph. When the onnncomnmuunnninate(l smmnnnpte reunehnes line valve, it is then exposed to flow of a hot carrier gas whichn mnn~eetstine samnnple mnmto tue gas eimromrnatnngraplr. line preference of the direct compositional amnahysis, as condumctesl by tIre above nrethod, to the conventional blow-down technique, will be furtlncr discussed in Section 2.2.4.

PVT TESTS

PVT tests are designed to study and quantify the phase behaviour amnd properties of mm reservoir fluid at sinmulated recovery conditions. The effect of interstitial water oni tIne phrase belrmuvioumr of hydrocmmrhon fluids are ignored in most tests, and time PVT tests mmrc cormdmmcteui iii tIme absence nil water. lhe nrajority of tests arc depletiomn expeninncmrts, whncic tIne prcssunrc ml tIre siungle phrase lest fluid is lowered in successive steps either by imncremnsing tine flunitl volsunne nir reunnoiving pmurt of it. The reduction of pressure results in formation of a second phase, excepn urn dry mmnrul wet gas mixtures. lhe formation of a new phase is generally acconnipanied by sonic degree of snipcnsmmtumrmmtionn. lhc pressure of an oil sample can be lowered below its hunbhhe ponimit grmnulmnmnhly winilsi main(aimning it as a single phuuse fluid. Such a fluid, however, is inn nnnetastahlc eomnuiitionns, as further ulescribed in Section 5.2. Tests conducted in laboratories on liquid samples contained in a pornnus mediutnni, brave rcsunltcol in some degree of supersaturation, wmthr values as high as 5 MPa (12,131. hligln suipersaturahion has been observed no tests where tIne pressure hmts been lowered rmnpidly. hnr a reservoir where the pressure decline is slow, signilicmnnnt sunpersatnnration is trot expected 1141. Surface forces can be significant in tight pores, affecting the pinmmse bchnaviounr of fluids. Capillary condensation, where gas condenses in pores due to fluid-solid inicrmnetion. is a well known phenomenon (15,16). The effect would he of significance in pores typically less thmmmn 8 l0 m. Gas condensate reservoirs are generally assunnncd to be wInter wet, withn tight cmrvities filled with water. Hence, the capillary condensation effect may be ignored. Tests inn a cell packed with 30-40 mesh head have resulted in tine samrre dew point as tinmut msreasured conventionally in arm equilibrium cell (171. The above review suggests that the assumption of equmihihriunn betweemr nIne phases inn reservoirs, and neglecting the surface effect on fluid equnilihnia. is a remmsonable engimncerinng approach. This has greatly simplified experimenr(al amnd theoretical stundics of the phrase behaviour of reservoir fluids. In conventionnal PVT tests, tine flunids mire given amrnplc tune mnnrui agitation in equilibrium cells, to mrpproacln equilibrium. At certain coniditionms. mmdi mis in ruupiul pressure build-up near Ore weilbore or in higbn pressure gradient flow, line dcviumtininn fromnnn equilibrium may become significant. Should non-equilnhrmunmrn ioforunnation hecnnimnc innrportmmnrn tmi field operation, such as bubble nucleation in water invaded reservoirs dunring dcpleliomn 118,191, especial tests could be designed to generate the required data. An important test on all reservoir fund samples is tIre sletcmmnninmntiomr mnf tine fluniol comnmpnnsiniunnn. The most common method of counposilio,ral wralvsi,s til Inigin pressure fluuids is to ilaslu mu relatively large volume of the fluid sample at the atmospheric pressure to fomn gennerally two stahihised phases of gas and liquid. The two phases are innclividuahly analysed and then

40
2.2.1 Dry Gas

2 !Vl ic c;.c mud (nrrehmmio,i.c

2.2 IV1 lenin

41
t
Z ~, ~I~l

~,=.

As no phase change occurs for a dry gas, mIs connposition remnrmnins unmnehuminged duiniung production. The only PVT test required for a dry gas is tine pressurc-volummmne melmttion at tire reservoir temperature. A volume of the gas is loaded into a cell maintaimned at the reservoir lenrnperuutnmie. Ilne gums volume is measured at the reservoir pressure and a numnnrber of prcssumrc intcrvmmhs Iiehonw it. llre specific gravity of the gas, relative to air at 60 nup (288 K)t is deteninined by mmremnsurrimrg tIne weight of a known volume of the gas, or by using the guns nmoleculurr weiglnt kmrowimrg its composition. The gas specific gravity, Sg, and the molecular weiglnt, Mg, are related by lire followimrg relation, as gases almost obey the ideal gas law at the atmospheric pressure. Sg
=

V k ~l )~

(av~

(a7~

(27)

A typical guts fonmnrnainin volnmunne fmuctnr plot is showmn in Figunre 2.3.

o
Sn.

.0

Mg/Mair

Mg/

28 96 .

(1.4)

a
a.
0

The volume of gas at reservoir conditions required to produce mimic mint volummrns of guns mnt tIne standard conditions is defined as the gas formation volumnme fmmctor, B , 8 Ba =.~1(-=Z(~-)(3) (2.3)

a 0

tressmnre --------------.>

Figuire 2.3. Selrcnmrmmi mc vmmmnuniromns of guts fommnrmnl imimn vonlnnnnre factor willr pressure.
kuti,nninle 2. 1.

where VR is the gas volunre at reservoir conditionrs of pressure P unnnoi tennrpcrmmnunne T wilim tIne gas conmpressihility fact(nr of 7., mmmd V,~ is tIme gums voltmmnne uul Ilie stururulunmnI sommntmt nnnnns ol lcssmume ~ and temperature T equal to 0.1 MPun (I hmnr) or 14.7 psimm, mummnl 288 K or (.t5n)(n7+tiO) it, 5~ respectively. Substituting the values at the above slaundard conditionns inn Eq.(2.3),
4 Bgrr3. 7X

10~ Z (T/P)

(2.4)

liii) turn of a gums mi mime rtsemnnnin uommnli(ionrs of .180,3 K (225 nm : mmmd 2(1.79 Mlui (300(1 11 hnsi~)was bnrourglnn tui I Inc St mmniulurn d unmmrnl it ions, wtncrc tIre guns ocsnn pied a volnmmmre of I 8,531 ennn . [ire pmoiuiccui guns specific gravity is 0.65t). (alcmmlale Bg. 1.. mmmd the gums density at (Ire reservoir connnl itionrs. Solnuinn,r

where T and Pare in K and MPa, respectively. (Bg0.O 28 Writing Eq.(2.3), we obtain, 3 Z (T/P) T : R, P: psia) Br=Vs/V,,=( tOO/I 8533)=Z(380..(/288. I 5)x(O. h/20.79)=O.005396 The measured pressure-volumnrre data are employed to calcmnimmte tine compressibility factor, 7., and the gas formation volume factnnr Bg. umsing Eo~s.(2.4-2.(n) z=
Z=O.85(it)

Pv nRT

lint guns mnmmilecnitunm rntmgbni (2.5)

ns

c~mtcniIuntsmtfrnimmr Fut.( 1.4),

where n is the total nunniber of moles, calculated by dividumng tire tontunl nnmmuss, inn. loaded in time cell by the gas molecular weight, Z= PM(V/m)/(RT) (2.6)

M~=S M,,,=t).65Ox2tt.96=i 882 5 tisinmg Eq.(2.6). nIne olemnsity is cumlcnnlated, as, p=(PM )/(ZRT)=(20.79x I 8.tt2)/(O.8500x0.0083 I 44x380.3)= 146 kg/rn tm

(V/m) is the specific volume, and is equal to the inverse of the density. p. Tine value of R for different units are given in Table A.3 in Appendix A. The isothermal compressibility coefficient of the gas, Cg, can be calculated also umsing tIre variation of Z with pressure, Various siandards for nennperature. including 60 F, 0 ~C, IS C and 21(8 K, trunve been adnnpncd flue are 60 F, in Field Units, arid 288 K, in SI absolute scale, wlincln arc uccum inn mInus interchangeably.
nnmisn linnmnk

wlnerc time uimnvcrsunl guns connstmmmnm is tmmkemm from Tahle Al un Appennulix A.

2.2.2

Wet Gas

common values

PV1 tests fonr mu wet gas mt reservoir conditiomns are simirilar to throse for a dry gas. Scpunrmnne tests are. Inowever, needed to) uletemmrnimne tIne muminonumnt ansi properties of prouluced fluids at tbme surfunce conditions, The formation volunne factor of a wet gas, Bwg~ is defined as the volunmnc of tIne gums mit reservoir comnditions requmireul to produce onre untnit volume u-mI tIne stock-tank liqumid. In the Fuelol ummnnts, tIne guns voilummnre is ursumunily expresse(l un termnns of barrel in the above defimmitionnr.

42

2. IV! Te inn nnnd (~m, mhnii.nni.m

2 2 PV1 Te.orc

43

A more practical definition of the gas formation volunrme factor, currently uised in reservoir simulators, is the number of barrels of reservoir gas (inclumding time liquids dissolveoi inn it) mut reservoir pressure and temperature. per cunhic foot of the dry gas produiced fmnnmmn it mit stanrdunrd conditions. This is analogous to oil formation volume factor, described in Section 2.2.3. The molecular weight and specific gravity of produced condensate are nnreasuured in tine laboratory. The molecular weight is commonly deteninined by drssolving the liqumini in bcn7.enre and measuring the depression of its freezing point. TIne liqumid density, hence, ins specific gravity relative to water, is determined by a pycmnotineler or mmmi oscnilmutitrg tutne mlenmsitnmnnmeter. The apparent specific gravity (or relative deursity to air) of time reservoir gmus is de(ermmminneul by calculating the reservoir gas moiccumlar weight, and,, unsinng Eq.( 1.4). 1 laviurg nnrcmmsunrcd tIns mass mmd molecular weight of proxluccd gas ansi onil (corrulensmnle) plrumses. (Inc nrrixlunre nnmonleermlmmm weight can be detemminesl by material balance cumlcunlmmtions,

In the flash vaporisation test a sanrple of oil is placed in an equilibrium cell at a pressure equal or greater than time initial reservoir pressure. Tire pressure is reduced incrementally by expanding tine fluid volunsie. The total vohuime, V , is mrmeasured at each pressure stage. The test 1 is also known as fluush liberation, flash expamrsion, constant composition expansion, and pressunre volume relation, Typicuni PVt lest (lumla mum reporieol by a lmrhoralowy is given in Table 2.1. The presstmre-volumme olumluu mnf tine black nnil, wilin comnrpnnsitionmn mis in [able 2.1 A. is shown in Table 2.1 B. Tire data is mmlsnn plontled un Figunre 2.5. FIre pressure umt whmiclr tIne slope of pressure-vohumnie line changes is (Inc humhhle-point pressure. lire slope of tIne cumrvc above tine bubble point, Table 2.iC, is a nnreasunrc oil tire is(ntirermrr,nh compressihiimly ofoil,
~mn=

(~~)T

(2.9)

Mnr =(mg+mo)/(~L+~2~]

(2.8)

winere C,, is tIne minI rsnntlnummnnuml eninmmpressnlnmhmty coefficient, The system voluinne is commonly neiimirtcol my t Inc relatime so/maine, ololinreml mrs tIne runt inn nil lIme total vodumnre to the immit imul bubble p~mmnIvoiiunrme. Fumb-de 2.1, Selected luthnles froumn a typmcmml PVi report on black oil, Reproduced from Core Laboratories June. report witlr permmnission. Humid : Good Oil _______ Reservoir lcumnpcrutlumrc: 378 K (220 F) Originral Reservoir Pressure: 283.7 bar, (4100 psig)

Eunmpiricai correlations, Section 2.3.2, are also available 10 estimate the reservoir guns specific gravity from production data when some of the scpuurators data, partmcumlmmrly in mnnuilti-stunge separation. are missing.

2.2.3

Black Oil

The phase transition of an undersaturated oil during depletion is depicted in Figumrc 2.4. Away from time weilbore, zone A, where tire pressure is still above the bumhhle point, tine oil expands as a single phase liquid. The pressure in zone B is just below the huibble point ansI tire voilunine of the evolved gas is too small to allow its mohilisatiomn. In zone C, the evoilved gas flnnws towards the producer, but segregates from the oil dume to gravity and suurface forces, In tIme wellhore, the two phases are considered to flow togellner due to tine dominunmnu mnnixinng. it is assumed that the phuuses are at the equilibrium throughout as time pressumre depleliuin runie is lumile low. The above reservoir processes are simulated in line laboratory by (he equmilibrium flmmsh vaporisation test, for zones A, and B, and the differential vaporisation test, for zone C. All tIne reservoir tests are conducted at time reservoir temperature. A series of flash tests mit selected temperatures are also conducted to simulate the surface conditions,

lunble 2.1 A. (onrnpositionn snf rcsemvnnrr oml,

(nnirnpnnncnt
, -. - -_-.

.M~L7~_ ._-_-NnI (tOt

0.16 36.47 9.67 696 I 4.1 5,95 I..>.> Itt 4.51

W~tnt%

Ilydrmngcn SulFide
Cunrhmnn

Nil
0.43

flimixide Nimnnngen
Mum ionic ntn,nrnc Itnnnnane
liurtmnne

0.05
6.24 3.10 3.27 ((.89 2.44 I It .09 3.97

Sc paran n nr

rnltnni:nnne tnnnn:Innc
Oil (.rc iitn,nnmmnnc I tsXnnncS I,,~j~,nancs plus tepu:nnncc pills t~rn~nwn i cv mum cc.

33.29 77.41 ticmnsnry-~85l5 kg/nmm (34.5 AII), Mrlecutar Wcight=2t8

Reservoir
A

lIne lmuboirumtory slatum is of (em evurluated, snunoonhed, and extrapolated by a dimensionless function \. defined as,
~

=f(P~P)/P]/[(V, ~Vb)/VbI

(2.10)

svlncre tIne sunbscripi h, meters to (Inc bubble point conditions, A plot of Y function versus pressuire sinould yield a lime eillner s(nunighrt or very sliglmtly cunrved. Figure 2.4. Phase transition in an unnder saturated oil reservoir,

44

2. IV7 lo.nr.m imnmd Corrmlmntimi,nm

2.2. IVi le.st.s lahle 2.lC Volunrretric data of oil, I. Saturatinnn pressure (hniintitc.pniinnt tircss;mre): 2. Density at saiur:niion pressnnre:
3. 4.

45

Table 2.1B
Pressure volume relation at reservoir temperatunre, 220 F (378 K). Pressure Relative Volume Y Function Densiiy psi~ bar (1) ~ 5000 345.7 0~9639 0.6808 4500 311.3 0.9702 0.6763
4100 283.7 0.9756 0.6726

262(1 psig @ 220F (181.6 bar ? 378 K)

40.97 lb/fr (656.2 kg/mimi)

thermal expansion nil rcscrvoin nnil

4000
3500

276.8 242.3
207.8

0.9770 0.9845
0.9928

0.67 16 0.6665
0.6609

3000
2900

2800 27(8)
2620 2605

200.9 194.1 187.2 18 1.6

180.6

0,9946 mr.99.64 (m.9984

1 .OtX)0 1(1(121 1.0042

0.6597
0.6585

F (288K) (omnrnpressmhiimry nil snnturmnlcd nnil t? reservoir iennperanurc: (Voi/Vol)/psi ((Vol/Vmnl)/bunr): Froirn 500(t psng (345.7 hmmr) Inn 4000 lisig (276.8 har)= 13.5 x ttt~ (1.96 x 10 4) Prminm 4001) psig (276.8 bar) inn 3(X)0 psig (2(17.8 har)= 15.9 x tO~ (2.30 x 0~) Frniirn 30(X) ~i~g,,(2078~ (181.6 har)= 18.7 x i0~ (2.72 x I0~)
V (ii (no

@ 5000 psng (345.7

burr) =

V @ 220fF (378K)
=

1(18795

((.6572
1)6562

2591 2516
2401 2253

179.6 174.5 166.5

156.3

1.0157
1.0353 1.0648 2.497

2090 1897 1698 1477 1292

1040 830 640 472

145.1 131.8 118.1 102.8 90.1

72.7 58.2 45.1 33.5

1.1041 I.I(n27 1.2415 1.3611 1.5000 .7839 2.1709

2.7603 3.6961

2.418 2.325
2.22m) 2.122 2.033

Iin tine e/iff~,nn,tiaI mnmp(mIr.satuni or Iiliermitiomn test, tire mnil pressnmrc is reoluiced below its bubble lnmninmt ull I lie reservmmnr lennnpcrmntmnre by rxpummroiinng line systenmr vnilumme. All the evolved gas is then expelleol mit cnnnstumnnt tircssunrc by rcslumcmnmg (Ire Cnhuiulihriummin cell volunmnme, Figure 2.6. IInis pronce(lunme is repemnted iii 1(1-IS pressunre slunges olowmn 10 tire atmnrospireric pressure. A eumclr stage tire nemnmainimmg oil volunnne. the expelled gas volunmnre at tIme cell uunnd stanrdard conditions, and lIme gums specific grasrty mire nnieasunrsd, The ga~formnration volunurre factor is calculated fronr Fu .(2.3), hunt onftemn diviuled by 5.(nI coirvertiung ml to huirrel per stuumndard cubic foot (hbl/SCF) 1 unsung time Field unmnits uus time gums volumnire mnt reservoir commslitiomrs is to be added to the oil wlmemn vnnlummnme inn barrel to (le(erilninC the total hyulrocmurhon volunune.

1.911 1.1(1(1
1.718 1.637

(I)

Relative Volume: V(V,~ is barrels al i ndicamed pressure per barrel am saiurmnminnn pressurns.

40

no
tlurhhte Inninl 20

/
0

Figuire 2.6 . Sclneunuutic diungrmmm of ditferential vaporisation (liberatiomn) test. Time compressibility fuictor of produced gas is determined from, Z=(VPT c)/(VscPscT) 5 (2.11)

10

0 0.0

5.9

i.e

ii

n 2

ii

1.4

1.5

Relummive Vum(nmmmme

wincre V is (Inc expellesi gas voluinie mit tine test pressure P. mmd tenmnperatunre T, hontlr mr ahsoIurte scmnles, Time rennnmninninmg (nit vmnlummrre, at lire mnt mniospineric pressunre, mit tIne ensi of tIre test is converted tni One v(ilunmmmc mit 6O~F(288 K), conrnnrnunnrly ursing ann unvcrimge I hermmnal comntraet nmnnm cuici incicmnt oil 0(88)46 (v/v )/F, mmmd n c erred Ion mrs time mcsisiunmml oil. line vnnlnmrnrc oil oil at cureIn sluuge is reported by tire relmntive onil voltunire, B m&i, defimned as the rnttio of mmii volumnne/resisiumal volununne. Tmntile 2. It).
0

Figure 2.5. Pressure-volume plot of Good Oil at 220 F (378 K) to uietcmmmnimne its t,nmhble point pressure.

46

2.

PVT Incus

(nun,! Corre!aninn,rs

2.2. PVT le.uf.c

47

The total volume of gas evolved at each pressunre and all previnnuns pressunre stunges. mit tine standard conditions (se), is calculated amid commvcried Inn tine volummmue un tIne lest pnessumme. unsimng tire prevailing B . and is added to tIne oil volumunc to ohtaiir tine tsnlmml Uwo-Inimurse) voilumumre. lire tots] volume is 8reported by the relative (otal vtnlumnnne, Bi,t. delmnned uns tire ummtinn of totmul volumelresidural volunre. The evolved gas is reporleul by tIme solunlionmn guts to unil rumlmon, Rsnt. defined as the difference between the total gas evolved at tIre atmnnosplreric pressumrc (tIre iinnuul stage), and each pressume stage in SCF, divided by the residuuul oil volunnne, inn bunmrels, mns shown in Table 2.1 D. The differential relative volume data cuin be evunlurmilcd umn(i smnnoontined by plmnttitmg lmug( I -I versus iog(Ph-P). [he relation is expectesi to be limmeuur. Table 2.ID Differential vaporisation (liberation) test results. Retanive Oil Relative Solu0nnn (las/Oil Pressure Ram run (I) Vnntunsne (2) Tonal
Votunne(3~ p~ 1 1 n,nl/l nxttn)

Tire cnnmposmtmonr mutnd spccnfic grmmvity of flumslnesl gases aie urneasured, Table 2.1 F. The volume ummnsl tIme specific gruivily of the stock tmunk (nil at 60F(288 K) are also delennined. All the vniltmnumetric results mnie reported relative Inn tIme stock tank oil voiuunme. Time ratio of reservoir oil vmnlunmrrc to stock tank oil vonlumrre is givenn by the oil fnnrmatiomr volume factor, B , defined as the 0 units. The nuunnbcr of reservoir oil hmurrcls ton proolumce one stock tank barrel of oil in Field evolved guns is reponrtcd by tIre solumtmnnmn guns to oil ratio, Rcb defined as the volumetric ratio of lIne tn,(uml gums evolved (sc) to the stock tmunmk oil (SCF/STB in Field units). lIre nlnsiritiunt nun nil cmnnmrpninnemmls lnetweemr lIne proolumccoh guns airol stock tmmnk oil depends on the nmum mmrtns r ol sepmmnumt nun sIurges, mmiroI (Inc ~ uninml tcmnipcrurtturc of scpmura(oms. The (nplimumm unrrunmigemmnemnt is tIme omne wlmicln pinuluues unmore of Ore stock lank oil, considering also otircr cconomnic criteria. The stock tmnnk mnil generally contaiums only a trace of metirane and an nnnsmgmnilmcunnt amnoumnt of ellnamne, regunrdlcss of the separation conditions. The concentration of (7~ in the gas phrase is very small in most cases. It is time relative distribution of (he intermnnedimnle frunctiouns bctwccun line plrumses thrat detennmimnes tire optimuumrr separation conditions. lime effect of scpurrmntor arrmingemnmcnt beconmes mnnore significant for volatile oils. Fmulnle 2.1 F. Sepmnmator test results
Snpniinnur lrcscnnne

Oil
Dcnsniy g/crnn 1)6562 0.6655

C ormupres. Gas FmnrimiumI lmncrenmneninat (macunit Votunnnnc ( ;~~s ~nclnsr (4) (hasiny
--

2620
2350

SCF/bblvnnl/vo 18 I .6 854 152

bar 63.0 763 136

-~

I .600
1.554

I .601)
.665

1)846

0.00685

1825

21(51 11150 16(5) 1150 tOO

1150

6(8) 350
159

145.8 28.6 III 3 94.1 1611 59.6 42,4 25.1

12.1)

1.0

684 612 St4 479 416 354 292 223 57 0

122
109

Gravity of rcsr~ual oil = (I) Volumnie nif gas at nbc smandard comndinions per volume of residual unit. (2) Volume of oil an indicaned pressure and nemperamure per voiunnic of reonslual oil an line snanndmmrd cnnnmlnmimns. (3) Vniluinne oil oil plus liheramed gas at indicared pressure and remperarure per vnnlunmie of rcsnduat nil an Ihe siandard conditions. (4) Volume mif gas at indicalest pressure and temperature per volume am nIne slanidand connulimions. In the .ceparalor test, a known volume of the reservoir oil at its bubble point is flaslncd generally in two stages, where the last stage represennts the stock tank as shown in Figure 2.7. For oils with high gas in solution, more tinaur one iniennnedma(e separator is often used. A field avsrimgc lempermntuirc is selected for the separator tests. Tine test is usunally conrducleul at mm nunrber oil separator pressures to determine tire optinnmuunn field sepmuration cotnslitinnns, rahlc 2.1 E. The stock tank pressure is always atmospheric.

1.32(1 1.283 1.244 1.075 35.tAPI @ 60F

97 85 74 63 52 40 28 1)

1.515 1.479 .445 I 412 1.182

I 351

1.748 1.859 2.016 2.214 2 593 3 I (i9 4.254 6975 14.693

(1.6731 0,6508
06889

(1.851 1) 859 0.872

1)887 ((.9(13 1)922

(1.18(771 000882 0,01051

11(11 2.~S (101552 ((.02042

o,6t69 (1.7044
1) 7121

(1.818 0.797 0.7~t 07)4

0811) ((.831 (1881

itt

75 F (297
0 ;n~/( )nI Rumu no (I) .SCI/tntnt 715 41 637 91 Ours I )nl Rmnm Inn (2) .51.1/S Ut 737 41 676 Snuck Inmnk 50 (nuT Fnirnenatimimm (3) vnnl/vot Sepunrunmmir Factnnr (4) vol/vol 1.031 1(1117 1.062

0 ;rmnvmmy Att Vmulunnnc Finctuir Vmiluiunne

Sped tic ( ;~~iuy of Flashed Gas

l5
5(1 Ii

Inam
4.45 (1 7.9 I) 14.8 1) 21.7 t)

1)7198 1)7291 0.7182 1)7892

(1941 1)965
0.954

(((12951
()0S1%,S

t).840
1.3311 0.786

010851

1)988 .215
2.039

411.5

1.481

11(0
ins

1) 2(8)
Inn t) 31)0) (1 (I) (2) (3) (4)

92 6(12 178
549

40.7
-

1.474

1.007 1.11 2

1.363 0.732 1.329 0.704 1.286

542 177
478

40.4

1.483

1.007

1.148
40.1

245

246

1.495

1.007

(rums/oil

Ours

(;as

Rumilmu nn cnnlnrc leer mnf guns 51 (n and 14.65 psia per barrel of oil @ indicated pressure and 0 1~ icnnnpcrature. Gas/Ott Ratio nun cubic feel of gas (~160F and 14.65 psia per barrel of stmnck tank oil @ 60F. Pn,rniianim,n Vmilunrnne Jmacnnnr is barrels ml saturraled oil 50 2620 psig and 220F per barrel of stock tank oil 50 60t~. Separanmnn Oil Vnnlunmne Facnnnr is humrrcts nil onil 50 indicarcul pressureand memperamure per barrel of stnrck tunnik mini 51 6I(t~,

lIme results of sepmuruunor mests fmir the oil given inn Tumble 2.1 F are shown in Figure 2.8 [25). [he
oipinnsmuirnr scpmurmntor pnessumre is amboumt 1(8) psig winere tine fornration volume factor (FVF) is nnninrmnnumunn umursl tIre nnrunxinnnumnru slock tmmmrk nil is produnced. The crude oil gravity, OAPI, also

motauns mIs mrnaxinnmunnor vmmlume at line oipnnnmnuro pressure wimercas the gas to oil volumetric ratio (GOR) rs umt its nnmninnrummim. All tlrese indicators point to a Inigirer accumulation of intermediate connnponncnrt.s iii tIre oil plnase with mu separator pressure of about 100 psig. Operational lnrmnitumlmomrs mruuuy, Inowever, dictmute otlrcr pressure conditions in the field. Figure 2.7. Schematic diagram of separator test.

48
mill

2. IVT Tent.n

ntis! (u,rrrknI/mu,i.c

2.2. !V7 7e,ct,s

49

&
4 0

W
U ~ C ~

n
1.413

lhc oil viscosity is conomonly mntemnstnred by a rollirrg ball viscometer at tire reservoir temnperatuire amnsl a nummrnher of pressunre steps above amrd below the bubble point. The pressure below tIre hunbhle point is achievesi by depleting lIme visconnmeter fluid clnamber and expelling the gas. Fine produmced gums viscosity is 011cm calcunlated umsing a prediction nretlmod, Sectimiar 2,3.2. lIre results for tIne Good Oil umrc simnmwn inn Fable 2.1G. Tumble 2.1G. Viscosity of (nil at reservoir tenmnperature.
Prcssmnre Oil Viscmisiny liar cenlipnnise 0.450
(1.434

on

~ I

s~
C

; E ~

: ~.
0

8
1475

it 0

(atcutaned Gas Viscuisily 1691

ccntipoise

()nt/Cas

Viscnnsrnv Ralio

psig
lb

Separamor pue~aur.. plig

5)11)0 451%) 41)1)11

345.7
311.3 276 II

Figure 2.8. variations of oil properties witim separator pressuirc. Repriinncnl frnimun 1251, cmimnumesy ml PennwWell Publishing Cnnmpany. Table 2.IF. Compositional analysis of senarator gas at 100 psig and 75F (7.9 bar mind 1 297 K). Lior.~Dcs~/cinn Component Mol % (1PM MoI.W. Hydrogen SulFide 0.00
Carbon Dioside 1.67 44.0111)
28.013

35111) 3(1181
2818)

2620 235(1
2100

Nitrogen Methane
Ethane

0.32
71.08 15.52 7.36 0.92 1.98 0.33 0)26 t).27 4.128 2.017 .299 .621 . 121) .094 .104

Propane
i-Butane n-Butane i.Pentane n.Pentane Hexanes

16.043 30.1(70 44.097 58.123 58(23 72. 5(1 72. 15(1 84

Heptanes plus Calculamed gas gravity (air

0.29
=

.128

103

.8172 .8086 .2997 .3558 .51)65 .5623 .5834 .624 I .6305 .6850) .7370

1851) 16(8) I 350 111%) 85(1 600 351) ISO 0 I cennnmpoisc= I rrnPa.s

212 3 21(7 8 94 I 11(1.6 I 63.0 .15.8 128.6 111.3 94. I 76.8 59.6 42.4 25.1 12.0) 1.0

(1418 11.41)1 0 385 ((.379

1)373 (I 394

1(4(6 11.440 (1.469 t).502 1)542 ((.590 0.653 1)742 (1.854 1.29

11.0(96 (I.t) 183 0.01 73

0.0164

2)). I
22.7 25.5 28.7 32.2

0)0156 0.0149
(1.01 42

36.4 41.6
48 6

0.0134 0.0) 125 (1.01 i6

59.1
73.9

FIne belmmrviour of mm reservoir suil slumrinng slepletiour is siinmurlated by a combinatiomm of all tinree
types nil tests discussesl above. tIne reservoir oil rennains single phmmmse as burg as tIme pressuire is unhove its buhhle poinmt, mind its helnaviour is simuilated by Ihe sirrnple isotlrerminal expannsion inn Ihe pnessure-volumnre lest. TIne guns evolved jun51 belsnw tIme bubble point initially remmrains inrnmmrohile in pores. Eiemnce, time pressnrrevoiuinrne test (Iluishm vaporismntion) almmnost describes tIre process. althonigh pant of lIme liquid plnuuse is recovered wlnilst the gas is immobile. The evolvesl gas hegimms to unmove away froimm the oil as the gas saturun(iolm exceeds a critics] value. Tire process tinemr becomnmcs mrnore smtnnilar to the dilferetrtimnl vmnporisumtion. A part of the gas, however, rcmmnuniuns inn comntact wntIn lIne oil connlraiy to tIne sliiferemrtial vaporisation test. The flaslr sepmirmnniomr siunntnlmmtes tire llnnw nil guns ummnsl (nil in tIme well lnnnre and tlneir suibsequment separation in lire scliumn uilmmr. Ilre vurlucs of nntn uumnsl R , delemmnnirned by tIne separator test represent tIre original reservoir fluid 51 belrunvioumr mit lime initiuml hunbhle poini. l3nnthr variables, tlnurt is, time oil formation volume factor, mmmd (Ire solumtiomnn gums Isi oil runtion, decrease as tIre pressunre fmulls below the bubble p(nint. The differcunliurl libcramiomm lest is cnnursideresl Inn simumiate lire evorlution of gas and the associated slrrmmmkurge nil muil in line reservnnir below tine hurhble poimnl. 1mm materimml halazrce calculations, the prnnpci(ies oil humid prmnduiccd at tine surface are relalesi to lirose at reservoir conditions by tIme mesunlls mnf scpmmrummnnr Icsts, mmmd mnol Ihmnnse of differcmrtimnl hilneralinun. As ~ and R d, dctenrninred 5 by ulmflcrcnrlial hiberatinimn, mit pmessunres below tIne inilial huibNc point are available in PVi repnnrts, Ilnese vmilunes are often nmnistmnken as time fornration volume factor and time solutiorr gas in mnrumterial halammce calculations. FIne differetntial liberation test data are based on the residual oil un the reservoir, whereums tire volume factor and solution gas slata based on the stock tank oil nrust be used in immutterial halammce calculations. The corresponrding values by the differentimtl test are alnrnosi unlways higirer unnd earn lead to errors of 10 to 20% in the calcunlated oil in place and
t3

1.000) = 0.786

Processing of separator gases to liquefy intermediumlc lnydrocarhoms unnay be economrnicunlly feasible. The amount of these compounds is expressed in ternnns of galionrs per tlnousand standard cubic feet of gas ot GPM in field units. The GPM of a component is cumleulated from: (GPM) 1
=

(l0001380)y M (7.481/p ) 1 1 1

0.3l5y M 1 1

/S1

(2.12)

where, Yi and M~.are the mole fraction and lime molecular weiglnt of time comnmlnnummennt m inn line 3 produced gas phase, respectively; Pi and S~ are tire density, lb/It , and the specific gravity of the component i, as liquid, at standard conditions (Table A. I in Appendix A). Example 2.2. Calculate the Iiqumid n-bulanne content of the gums prmdiiced fromnm tIme separuuimnr nil tIne (Imnund Oil at 100 psig amrd 75 F.
So! ofio,m:

The specific gravity and Molecular weight of normal butane are read frommn Table A. I equal to 0.5840 and .58.12, respectively. Hence, GPM=0.3l5X0,0l98X58,l2/Q.5840~0.62l gallon of liquid per thousansi ft(sc) of gas

50

2. PVJ Te.qn mind Cm nelnlio,ns

2 2. IVT Te,ni.s

51

recoverable oil 1261. These confusions could have been avoided mostly if tine residumal oil irasl heeun reported at the reservoir lenrperatunre. and separator tests haul heenr comndumcted on liberated (nil samples as well as line original reservoir oil. The reported data by conventional PVT tests can be combined, however, ton deternmniune tine required data for reservoir studies [27J. Tine main assuuumnptmons are: (I) Tine gas in solution at reservoir conditions below the bubble poimit llnmlt will be Iihermmlcsl mit the surface by flash vaporisation is equal to time difference between tire original guns inn solumlmon ansi the liberated gas by slilferential liberation at the reservoir pressuire. (2) The relation between tine FVF of flashed aund differeurnially litneruited smmnnmples mcmnummmmns constant over the entire operating pressure. llre oil fonrnratidnn volunrnme factor B
by.

hunhble poimmt is equmal to the origiunmnl gas Rcdh tininnuis tIre remaining gas R d by the differential 5 prnsccss. lIne unnit of Ihis gas is inn SCF per hunrrel of (Inc residunal nil. The evolved gas per SIB
is (lmemn enjunuml tmn.

R,)-~55

(2.14)

I lemrce, I Ire gums in sol umtnonn R is equrunl to,

=

R, , 1

(R,,ns

R,,n)

(2.15)

11

at any pressure below lIre hurbblc pnnmmnt is llnemr surlcuniumlcd

B = B~ 0

(2.13)

lime cmmlculmitcul R lnnm I lie mint inn lunblc 2. I Is slnmuwnn in Figunre 2. 10. lhe above two assummnptions inn cumInvu iii mng (Inc5ml 1 Isrcmnt imml I ibenuit in inn olumtum Inn I Ire selnumm mutmir slatum tiecommme less rcliumhle as more gums is Iiberumned lo nnnn line unit. line sumlcuilun(cd duntum mime unmmcccptahlc near lire residual oil cnnnrolilionun~. however. mus tlncre is very billIe emngineerinng application for the near residual data, tue nnretlnnnd is used widely for Inlack oil systemins slune to its simplicity and acceptable accuracy.

where the sunbscripts h, and d refer to (he initial bubble pount, aunsl tine olrflercnmtimnl test conditions, respectnvely. Tine adjusted formatnon factor for tine oil in Table 2. I , below line btmhble ponmmn) Figuire 2.9.

is slnowur nun

>

0 0
0

C
>

5
I)) lrcssure, MIa

IS

20

Figure 2.10 Adjustuunent of gas inn solutisin.

In

It)

IS

2))

Pressure, MPa

lIre ratio of molal guns mmnid oil volunme unt tIre reservoir conditions to the stock tank oil volume is expressed by (lie lontmrl volumme functor. B . it may be cleimned as the reservoir volume occupied 1 and its associated gas. Hence, by one unnit volume of tIme stock tunnk oil
B

Figure 2.9. Adjustumrent of oil relative volinme factor. The gas in solution below tine bubble point by flasir test can also he cmmlcurlated by comrmbnnmiung tIne differential liberation data and flash test results of time original oil. TIme gas evolved below tine

(2.16) 1

II tire solution gas in Field umnits is ulescribed by ft lbbl, tire gas volume must be divided by 5.61,

52
B,
=

2, lVj im,!., mmmi,! Im,, reIn,iim,mm.c

2.2. lVi le.ti.u

53

+ B

(R,bR,)15.61

(2.16mm)

Table 2.2.
Selected tables fronr a typical PVT report on gas condensate, - Fluid A Nsnrmln Sea Gas Condensane Reservoir lenmnpcrunlunre: 394 K (250 F)

Using the differential liberation data, we can write,

Bm momBOb .~___+B~(Rsdb_Rcd)j__

(2.17)

Rcseivmiir Prcssure:49.64 MPa (7 911 psia)

and in Field units, B,

lmmhle 2.2A. 1)etmnr led__comrnpcnsmt monnun I ummnmnlysrs nil Ilne web I slreamrm.
Coinnpninenis MniI Ye
b~t~

B,,b BOd.

Bg

~~imtu~

(2. l7um)

Nilnungen
an bonn ulim,simlc Mcllrunnnc

0.298
I .721) 71) I 39

The variation of B with pressure and its coinnpumrisomr witlr 13,, unre sinmuwmn in Figumne 2.11. 1

Fllnanc
linmpannc

7 4113 3.293

iBu(unnic

nBumianc
i.Pcnnmunncs n-Icnmiune it Icsunnies

((.5(5 1.255
0.359 1)551 0.282 (1.334 11.111 1)271
II 31(9 1)2)5

n-tlcxumnie
0
LI.. 0

n-I lupnznnes l(emn,mnnc

(yctunnes C, nil ttpn inc 1 )uruinmcs tmitumrnnc Cyclumncs C, o Occinnu i Nm,nnummncc

C
0

I). 45
0. I 5)) 1)253 1)1 (i8

a
0

Arnmiiizmtics 1., CyuIumnes C, in Nmnnanne

n~t)cuunnnus

(115% 1)143
1) 061

((.113
(1.176

10

20
Pressure, MPa

30

4)1

Figure 2.11. Variatiomrs of oil amrd total forrrrumtinnmn vonbummmmc fumsknns wulin lessnums. 2.2.4 Gas Condensate The compositional analysis of gas condensate fluists is conducted gencrmnlly inn nnrore details thamn that of oil. The compositional data are used oftenn in plnase behmnvinurr nroolels. pumr(icunlunrly in reservoir simulation. The fluid is comnrmonrly alrmulyseol by flumslnmmng ml mnl tIme umnnmrosplieric pressure and measuring the composition of tIre stabilised gas and liqumisl plmase.s, mis described inn the blow down method. The fluid heavy fraction is analysed to islentify nnajor componnenrts, and also to characterise it by extended carbon grounps, as the results of phase helnaviour mmrodels are very sensitive to the heavy end description of gas condensate systems. Selected tables fronnn a PVT laboratory report on a North Sea gas condensate sample arc presented ins Tables 2.2. whereas Table 2.2A shows typical measured connpositiomrmul data. lahle 2.213 describes tIre distribution and properties of lneavy conlpnincnnts uns simmglc cmmrbmnnn nnmimnnber gnnnunps deternnnimncul by distillation, Tire properties of single carlnonmm nnumrrnher grnnuups mmreuusuiicoi inn lIne iinlmnid lulninse are considered lobe Ihe same as those in tIne well stmeummnm. l)etails on testimng lire listnniol frumctiomn to characterise the fluid are given in Section 6.1.

Armmurmaiics C, nn~l)cuunne Unnnlccaiies Dmxtccanes lnintecunncs Icirunnlccunnics

lcuniunmlc,aincs

0)1154 1)084 (1.318

0273
0) 253 0.225
(I. 1711 0. 144 ti 126

I lcxumdun,innics I lnpiunducanes (klunnlccuuncs

Nnnrnuiutcuunnics

(1.127
(1063

t/.icnnsmnnes.ptmis 0.553 Munlecurtar Weiglrn 27.3 t0cnmsunnesplmms clnuiracierisinns. Mu)cculunr Weiglni = 353 l)ensiiy ai 21(8 K= 852.1 kg/inn

TIre two mnnost connmmrnon tests at tIne reservoir temperature are tIme constant composition expuumnsio)ur, CCE, unmnd time constmnnl votunne depletion. CVD. In CCE, or the pressure-vumluimne test, a knnown mumoumnml of gas coundennsmn(e is loaded into a visumal cell above lime initial rescrvnnir pressumne. lIne systcunn pressumme is lowered stepwise -by incrementally expansiing the cell vnnluinnre. Gums anol conndcnrsmute vnnlumnmres unre recorsled unt cumcin pressumre step mrs well as time onhscrvcd ulew poimut, iumble 2.2C. A Iypicuul pressuirevolumume curve, witim lIne data reported in Tumble 2.2C, is slrowmn in Figure 2.12. Arm abrupt clramnge of slope at time dew poimnl ulnncs noit gemrermnlby occuir. Ilennce, the dew point camnnot he measunmed accuralcly by nnooitoring pressure-

54

2. PVT Tens mnum! Correhijimnn.r

2.2. IVi Tm,si.m

55

volume changes. in rich gas condensate samples, i.e.. close to tlneir critical tennnpcrmnlulie.s, tIne dew point is manifested by fornnation of a large amount of conslensuste. Tine nnneasurcd slew point is, therefore, qumile accunrate. Tlnese fluids may also slrow mm grmmdumunl reversible eolnrumr change, getting darker, mis tlne dew point is uupproimclmed. hr nnnosl curses tine innitiunl Iioinniol hmmildum n 1 is gradual, which nrakes the measured dew point qunile subjective. FIre volummnre of csnnndcmnsuuI liquid in Ihe above test for the Nortlm Sea gas, Table 2.2, is shownn in Figunre 2. 13. Note tlnunt the condensate fraction is defined relative to line total volume. Table 2.2B. Distillation results of the liquid fraction. Coniponent Weighm Range nil ulismillununonn
Ye Carhnmn Dio,uiule Meihamne tnlnunne Fr, panic Itun)anncs Pennunnec 0 00 1)1)0)
0,05

(I) (2) (3)

(4)

(5)

Salinrunniunnn pressure am nnmticamcsl nennpcrannnrc. I nmnm nI rcscrvmnmr pressnmne. V,,, = s,ulunrmc ni liuki nI Salultaiinnmn prussnnrc and indicaicd nerunperalure. 7, = PV/nRI (n imniuml mnninmihcr ,,i mmim,(us) (Vumlunne mnl reiru,grade liquid am indicated pressunre)/(Iolat volunmnne an saluralionn pressure)xtOO

(no

ncnnpcrumnuire, C

Dcnnsiiy mn 288 K kg/inn

Mm,kunnlunr wn.igIni 50 40 30

I texancs
Ilcpnunnes
Os.thncs

((.4 I I 0) 2.44 3.56 7.70

8. Il

363) 70.1)
100.0

-.

7)).0 ((81)1
127.))

711 5 79).))
749.4

81, 89
0(5

21)

Nnunanes

Decanes
Undecmmnnes l)nxlccanes

Tndeca~s
Telrantecanes Pcnnadccanes Itesaslecanes Hepradecanes Ocnadecanes Nmnnundccanes Ficosanesplus

7.04 5.31 5.89 5.48 5.51 5.29 4.50 3.88 3.60 3.88 203 23.93

127.0 152.0 175.5 197.0 219.0 236.5 254,5 271 5 288.0 303.0 318.0 332.0

152.0 175.5 1970 2(9.0) 2365 254.5 271.5 2880) 303.0 3111.0 332.0

764.1 776.6
71(5.7

121 138
sI

10 0.8 II) 1.2 1.4 1.6 1.11 2.0

796.9 810.5 81.1.4 822.5 8295 832.2 835.7 838.1 852.1

164 178
192
206

Relamive Volume Figurre 2.12 Iiessuievmdumnne cunn ye br gums cotndcmnsumtc umt 394 K.

22)1 234 2t9 263 353

5

2
10

Undecancsplusclmaracmerisrics: Molecular Weigh(=23t Table 2.2C.

Densily an 288 K=834.6 kg/nnn3

>

11 6

Pressure-volume relatisnn of tine reservoir fluid at 394 K.

Pressure ~ RelaIn cc v,,tnnnn,e
-

a
C, nninprcssnhnlnIy Vm,Ini,,nm n nO rcirm grumnk _Iunc or. ~ Inqunl.(5J I 2866 I 2429 1.2158 1.2074 1.1991 1.1876 11718 I I (~60 1.153)) 1.1302 1.1039 1.0)661. I 0195 ((1764 0.9401 ((.9094 ((.8894 00)) 0)0)1 1)1)0) 1)1)0 0.00 0.00 0)00 ((.3)) (1112 (.73 2 92 4.79 7 18 9.41) 11.0)2 12.40) 13.15 ~ 1-

Spec,Inc

_____________________- , ________~YLY~1~)..(.~)__ volume

551.0 521.0 51)1.0 (2) 496.4 491.0 481.0 (I) 471.5 466.0 456.0 441.0 421.0 391.0
351.0

7975 7540 7250 7183 7105 6960 6822 6743 6598 6380 6)190) 5655
5075

311.0)

271.0
231.0 181.0

4tOS 39(5 3335 2610

0.9395 0.9599 0.9765 0.9787 0.9826 0.9935 I 0000 1.0)11611 1.0(74 1.0312 1.1)550 1.0971 1.1687 1.2632 1.3959 1.5841 1.9773

2.8012 2.8620 2.9113 2.9181 2,9297 2.9620 2.9815 3,0)1)17 3,0(333 3.0747 3.1456 3.2710 3.4844 1.7664 4.1619 4.7229 5.8953

It)

20

30 Iressusc, Mta

40

50

Iugumre 2.13 Liqunid bumilul_urp clmrve for gas condensate at 394 K.

II is cminnnunnnumnly umssnmmnncsl timunt (Ire uunnrdemnsumle dropped ount in pnnres remains inrmnnobite. Tire dc )lclionnm process is. tlnerefunme. sinnnmnl~nlsdmy CVI). lIre lest comrsists of a series omf expansion 1 bollniwed by cxpcllimmg line excess guns mit conmstanrt prcssumrc in suclr a way that tire cell volume renimuniums cosnnstaot mut nine end oleumclm stage. as shown in Figure 2.14, The expelled gas at each pucssumre stage is collecled mmnd Os composition. vobunuc and compressibility (deviation) factor

56

2. PVT Tesi.s and Corre!aiio,i,s

2.2. PVT Teds

57

are determined. The condensate volume is also measured. As the gas connposition remains unchanged above the dew point during depletion, the test can be simplified by jumst expandiung the cell volume without removing any fluid from it (pressure-voiunme test), TIre compressibility factor is then calculated, using Eq.(2.5). Time volumurme unt tIre dew point is comnsiulercet mis tIne reference (constant) volume in this procedure. TIre results of CVD test oim a Nortlm Sea gins condensate are given in Table 2,2D,
S.

Table 2.2D. Constant volume depletion test results at 394 K.

Cumnmnmlalcd

Iressmnre
551.0 7975 7540)
7250

Prnxtuciion. (3)
(1.00 0.00
1(00

Speeilnc gravity (relalive to air) of produced gas 0.943 (1.943

0.943

Comnnpressihilimy faclcir Volunme of of produced gas, Z retrograde liquid, (4) 1.2866 1.2429
1.2158

0.00
0 00
(1.111)

521.0)
51)1.0

(2) 496.4
491.0) 4111.1) (1)471.5 401.0
341.1)

7183
7105 6960 61(22 5804)
4930

0.00
0.40 1.51 2.17 9.67
17.66

0.943
0.943 0.943 0.941 (1.889
((.845

1.2074
1.1991

(1.1)0 0.00
0.00
0.114)

1.1876
1.1718 1.0767
1(1056

271.1) 211.0) 141.0 GAS:: :GAS~

Hd~~sH

GAS,:

GAS

(I) (2) (3) (4)

m,nmlemnsuniu

29.89 0.797 09479 42.90 0.760 0.9176 60.29 0.737 0.9171 81.0) 1160) 76.17 0.728 0.9476 Sumnuratinnn pressure uni inslicated temperature. Irnilial reservoir pressnnre. (Mmnies nil wet gas prnxlniccd/mnnmilcs cit fluid am initial reservm,rr pressure)xt00 (Vmnlumnne m,l remrungradi liqunid an indicaled pressure/luntal vnnlunnmne an samurannmnn pressure)x I (Xl

3915 31)45 2031)

4.31 7 53 ID 111 11.211 11.32 10.49

nindemisane 1 ~> dew

dew

~onnoIemnsaie nnnmlensame

t 1<< otew a

16
S S

Figure 2.14, Schematic diagram of constant volume depletion.

12
0

0 S

Figure 2.15 shows the liquid drop-out volume in CVD test, as the fraction ofcell voiunre at tire
dew point, which is taken as the reference volume, The liquid volume produced in CCE lest is also shown in Figure 2,15 for comparison. The liquid drop-out 1mm both tests has been defined as the ratio of the liquid volume to the volume at the dew point. Note that the accutmumlated condensate volumes are very much the same during tire condensing region. The dropped out liquid vaporises back into the gas phase at lower pressure conditions, It should be remembered that the transfer of components between the phases always occurs in both directions, and it is the overall result which exhibits itself by the condensing and vaporising regions. The liquid volume, as sbmown in Figure 2.15. incre.tscs witlr tire pressure rcolnmelionmr below tIne dew point at a significant rate. PVT test data often show a liquid build-up tunil, wlmerc time condensate build-up below the dew point is insignificant over a considerable pressure ramrge. Exceptional cases, with a liquid build-up tail extending over 10 MPa 1281. lnuuve been repsirtesl. The tailing behaviour in the liquid build-up curve inums been tire sumbjeet of conrsisleruuhle imnterest with diverse views on its cause. The initial gradual and small build-up of tIre condensate plnuuse has been attributed to the contamination of collected samples with hydraulic fluids from various sources during drilling, production and sampling. The test procedure can also affect the observed dew point and the liquid build-up beimaviour, A 2 MPa tail was successfully eliminated when the dew point was approached over weeks instead of the conmmon practice of over hours or minutes[291. There is no firm evidence, however, tirat the tailing cairnot be tIre true characteristic of a real reservoir fluid, Indeed, the presence of inimobile interstitial water and marked differences between the solubilities of different compounds in it, and the adsorption of surface active and heavy conmpounds on reservoir rock surface may contribute to the above behaviour.

0 0. 0 c:i 0 U

s
0

CVD

0) 1)

10

20

3))

40

50

Pressure, MPmm

Figure 2.15. Lislumid drop-out bclnaviour of the North Sea gas ccnndensate at 394 K in CCF. and CVD tests. TIme produced gas deumsity and csrmposition, Table 2.2E, are commonly nreasured by flashing the gas at the laboratory conditions, and analysing the collected fluid phases. Properties of time condensate accumulated inn tine equilibrium cell during depletion are not measured, except in the last pressure stage wlnemm tire eonrslensate is also expelled fronr the cell, Table 2.2F. The condensate properties at other pressure stages are calculated by time material balance [30J. The condensate density, Po at the jUn depletion stage is calculated by,

58

2.

Ptl Te.ut.s

mmnnd (flrre/(rfio,n,c

2 2. IV7ip,si.s

59

[min_~m~_P~(v_VoJ) /V

0~

(2.18)

where, nnin
~mg
= = = = =

Reliable ressnits fur the condemrsunte phmssc by malerial balance calculations can only be obtained by mnmn acciurate amnalysis onf prodmnccsl guises. The loss of Ireavy connpounds in the produced gas resunits mm umnn umurreunlistisunlly lnigln coinmdcnsumle slemnsily. amnd higlner concentrations of lighu connrponunctnis inn line cnnnrdcnnsmmte plnumss. It rummy eveni lead to negative mass of liglrt counrponents imn

Pg
V

initial mass of gas condensate total mass of gas produced density of gas at equilibrium with condensunte
totuni cell volunme

tIne condemmsuntc, suncln mns nitrogen. willn knw concenlrations in the mixture. Drohnn et al, 1311 sltndiecl (TVE) dumlus of 80 experiuriemnts anrd foumid uregative couurponnent fractions in 71 cases and condensate densities lncumvicr (burn wun(er in 45 cases, line above problem can be alleviated by
direct sannnplinng of nIne guns umnrd conulemrsau(e pinases at equilibrium. Additiomrai to tIre conmsrpnisitionnal anmnlysis of line Inigln tnrcssrmre sanrnple. tine slensity can be measured insitu by miscillatiung tribe densitonnrctcr 1321. Mmntedud hunlumnce cmilcumlations can be uses! then to evaluate tIre meliunhility of sIuituu innstcumsl nil cumlcnilmmlinng lIne pmnnpemlics of time c(nndcnsalc plrumse and tire :mssnncnumtcol eo uuilnbu nmnnnu runt mis. 1

V,,

condensate volunnie

The condensate cotrnposituon can be cunlculated by the cominponnenl urnurncriuml bumlumnrn.c mis.

Xjj =

[ninYlin ~n~yjj _(V_Vn,j)p~yi

[nm ~nri _(V_V,,~)p / M~] 61

(2.19)

Fumble 2.2F. (onnipositioinnalunmmmilysisof_renrraitnimngdl8l bar in constant volume depletiomr test.

(nnnn,tinnnncnrs ,,r rae)i,n,ns Rennnanmnnng gas
~mm4.%)
______ -.

wlnere: nm n 8 Mg
Yn Yn,nn

Remaining oil (moi.%) 0.00 0.03 1.29 28.06 4.94 4.03 0.80

= = = = = =

initial number of moles number of moles of gas removed

nnolecumlar weiglmt of gas at equilihriuumm with comislemmsun(e

mole frumction of connponent i in condennsate nnuole fnactmonr of comnnpomrcnt i in gums url equnilibriunurn willm conukmns:nle initial trumnber of unmoles of comnnponent i

lhe eqrmilibriununn ratios at eumch pressunre step caun then be delerumninmed smsimmg tIne calcunlumled condensate connpositmon and tIme niemmsured gas compositiomn. Table 2.2E. Compositional analysis of produced gas in constant volume depletion test uut 394 K.
bar Pressure 49303915 Ilydrusgen Sulfide Ninrnngen Cuurhmnnt)ro,uide Menlnane
Fnhumnn~

471.5

401 .0

34 I .0

271 .0)

211 .0 31(45

14 I .1) 203)) 0) ((1) (1.33 1.75 (02 71

7.64 .0,22 ))4m)

I Iyulrmngcni Snullisle Ninrmmgen (uirtmnmnm Dioxide Mcnlnunne llhuminc I, mmnmuunnc t)nniunnme nnIlnniunnnc i-le,nluun,cs n Pcnnunnne Ilexunncs I lepianes Octannes Nnnmnanncs Dcuuuncs

~m(n~cum9s-,plus,~
0.00 01.311 1.72 79.17
7.48

0)00 0.30 1.71 79.93

7.44

Propane i- Bunianne n-Bumane i-PennIunnes n-tenniane ttcxancs Ilepmanes CXianes N,nnanes Decaines Undecanes-plus Molecular weight Gas specific gravity

3.29

0.52
1.25 0.36 01.55 0.62 1.1)0 0.71 0.47 0.31 2.25 27.3

3.22 0.5 I 1.23 ((.35 05.1 0)58 0.90 0)68 ((.46 0.28 1.87 25.8

0.00 0.3 I 1.71 80.77 7.41 3.21

10.5))

0)00

((.32 1.72
111 (ml 7,I6

(1,0(0) 1) 32 Ill
82,31 7 5.1

Mu,lccuutar wcigtnn Guns specnlic grunviry (retamnve tnn ant)

Msnlccutar wcigtnu ml 0 (ndccunnes-plus

I) 01) (1 32 1.77 82 58 7.79 1(2 ((.51 1.22 11.33 0.5)) 0.46 ((64 0.34 0. 13 0.07 0.02 2I.I 0.728 66

2.36
0.85 I .42 2.85 6.28 5.66 4.75 3.50 33.18 116.0

3.2))
(0,50

3))
II 49

1.21
(1,34

(1.52 01.55 (1.84 0.61 0.4)) 1)25

1.37

1.18 ((.33 ((SI)

0) 52

I IS 1) 32
1)48

I IS 11.32
0.48 ((.46

0..))

0.76 0.53 (1,32 (1 20 085 23.1

(1,7)) 0)43 (1.24 11.14 1)45 22.1) 0.7W) 180

0.64 1)37 0)17 03)9 I). IS 21.4 (1,737 174

tIre munterunuml commsistcncy of tire nnmeumstnrcsl comnmpnsilionai dmmta can also be cvalruated by plotting tIre edlunilihrituunm numlio vs. a punrunurncler which indicates lire volatility of tire components. Figure 2. Rn sinnnws the nrnost connnunnonly umsed plot, known as tIre Hoffmann 1331 plot for a North Sea guns cuinnclensunlc. lIre compnnsitiomr of both phases lmas been measured directly in a CVD test at 373 K. Faclr data point refers to un counnponenl identified by a function which depends on its crilicuml pressunre, hn.~ critical tcunrpermmlumre, T,, and the boiling point temperature, Tb at tine alnnrosplncric pressure 1)1 P,. The equmiiibriuun ratio data are expected to fall on a straight line, A rcumsonnahie deviation fromnn tire straiglnt line for non-hydrocarbon compounds, such as nitrogen ansi canhonr dioxide, and compomnent groups containinrg unusually high fractions of aromatics or naphllmencs is expected. Time cunnlnposition of prodsnced gas as measured during CVD test by the conventional method is gerrerally fumr inferior to tIme rcqmmired accuracy for describing a phase which is after all time prosIunct. Fine repoutcd cournpositionmmml information is seldom used in evaluation and tuning of plmurse helmaviour models due to tine lack of reliability.

24.5

(relative

(ci

air)

0.943 231

0.889 207
,

(1.845
202

((.797
190

MmIccnmlar weighl of

tindecanes-ptus:

60

2. PVT Tests (inn,! Correlnrrisn,m.s

2.2. PVT le,sts

61

100

0. 0

MIr~~

0. N 0..

l0
mz.
U

0.

211 -4 -3 -2 -I t) I 2

40

611

80

lnnual lrmndmncuinnn, h1nnle31 in Ilunec

(LnngPc-lungPa)(I T1b- I (I)( I Ills I[Ic) Figure 2.16. plot. Internal consistency cineck of measured compositiomr in CVD lest by lloffmnmannn Figure P.2.3. Variations of PIZ with total number of moles of produced fluid. Itn a reurl reservoir in the mmhsence of water advmincement, the volunre can be treated as constamm(, neglecting compaction. When the produced flumid is only gas, line produced number of moles, n~, is proportional to the gas volume at standard conditions (I kgmol=23.95 mm, Table 1,1). Hence the above trend can be shown in terms (nf gas vol umurme produrceci. Exannple 2.4. The conslant volunmnne depletion test results of a gas condensate fluid are given inn Tables 2.21)-P.. (mm) t)eterimmimne the density of conrdensate phase by nmaterial balance calculations results. (b) Calculate the conmposition of condensate am 81 bar by material balance, and compare the resmnlns winh tlne omeasunred values.
So(nmtion:

Euamp!e 2.3.

Calculate the two-phase compressibility faclor, Z. of the cell Conteni in the constant volume depletion test reported in Table 2.21). Piol PTZ vs. lIne uotal produictionn unmnnJ comment on the observed trend. Solution: The two-phase Z is calculated as.

Z=(PV)/(nRT)
where V is the reservoir volume and remain.s constant dunring depletion. The volume of one kgmoie of gas, molar volume, at the initial reservoir pressumre cmun he calculated, using the measured Z for single phase gas. V= I .2074x lx 0.0083 l44x394.3/49.64=0.07974Q m Substituting the volume in the above equation produces the two-phase Z as shown in the following table, P. MPa 47.15
,.

(mu)
Basis: I t)0 kgnnroies u)f gums run the inilial reservoir pressure of 49.64

MPa. with

nrm,=100x27.3=2730 kg. The (constnmnmt) reservoir volume is calculated, using tIme measured compressibility factor, equnal tcn 7.974 inn (Example 2.3). Usinng the reported number of moles and the comnnpressihilily functor at the dew point results in, V=nnZRT/P=97 .83x 1.171 Sx 0.0083 144x394147. 15=7.971 m The differemnce in caicrulated reservoir volume by the two methods is due 10 the accuracy of reported clala. An average value of 7.972 m will be used in the material balance calculations as follows. Table Table lahle Tumble 2.21) 2.2F.-F 2.2D 2.2D P, Mia M 5 Zr lOOx(V,,/V) 47.15 40.1 27.3 25.8 1.1718 1.0767 0 4.31 27.1 21.1 14.1 , 8.1 24.5 23.1 22.0 21.4 21.1 .1)056 0.9479 0.9176 0.9171 0.9476 7.53 10.18 11.28 11.32 10.49
34.1

Number of Motes 97.83 Z .,,,,~ 1.17227 5 P/Z 40.22

40.1 90.33
1.07977 37.14

23.113 1.00731 0.94017 0.8988(1 0.86365 0.82676 33.85 28.82 23.48 16.33 9.80

34,1 82,34

27.1 70.11

21.1 57.1

14.1 39.71

8.1

The linear variation of P/ Z and total production is expected as,

=

(~i)(n.

np).

and (RT/V) is constant.

62 Table 2.21) I00.n~ nm. cum. prod, mole, n~ mole in cell, n mole gas rem..
gas density. kg/mn mass gas rem.. imn~. kg rn,,-Em , 5 V( I -V,~/V) V-Vn p, V 1 2.17 9.67 17.66 97.83 90.33 82.34 2.17 7.50 7.99 335.07 293.0) 253 .4 2

2. PVT Te.nj,c ~nnnd Corrm!nmnioni,s

2.2. PVT Te.ni.s

63
gums

298) 42.9 70.11 57,11) 12.23 1301 200 45 154.31

76. I 7 23.83 7.39 15,1(8 10036 55.02

6)1.29 3m) 71

cmnnd.
11.111(2 (1,1(68 I 453 1) 260 O.2Om)
1)1)45 (1,119

nmnsle % mole %
-0.1)4

Km

nnmnn,
rn/V,, (h)

tonal nnnass in cetl.m. kg gas vol in cell, Vr. rn 7971 7627 7371 con. vol. V,,, m 0.0 343.5 (not).2 mass gums in ccll,m~ kg 2670.8 2235.5 186711 muuss C(ifld. in cell, nm,, kg 0.0 241.7 413.7 703.7 689.2 Consl. den.,p,,. kg/rn

59.2 1935 195.8 2825 286.2 172.1 315,1 2671).8 2477 3 221(1.5 1999,)) 1712.8 I 34)1,6 1)11(5 (n 7159 7072 7068 81(4 8991 n%)n ( 1442.2 1091.2 7094 5561) 686 I 62)5 (m9l 3
(0) 2 (,qn) 6

Nitrmngcin arFsm I)nmnside Mellnanc Unlnumnc trmpunine

l)nnnmnnc nnltnn)umnnc n lcnnmunnnes

1)0(6)) ((329 1.5 362 I 4.r9 ((.61 8

1)1(95 (1 227

71(5
(n) 2.5 611))

733.2

in tennirnnnc I Iexumnncs tteplunnics 1 )ciumnncs

Nn,imun,nm,

I) 061 00) 3
0.086

(1.1)47 (11)77 11.153

1)32% I).293 1) 2,li

(1,119
(1.0i63 1(1)21

The number of moles of gas phase in the cell at any stage is delernniined ms li,=flim./M,, with the number of moles of condensate calculated as, n~ =n n~. IIne mnmninnlier of mrmoles 1~ of each component is then calculated as,: ~ ninyjin ~n~y

I )ecumvmcs Jnmlecaincs phms

lnnnal

(LIII 3
1(00(4

(1.18))
I .75 I 5.227

18.603

.3)) 27.79 4.96 4.01 085 2 28 1)91 1.46 2.93 6 27 5.60 1.76 3.44 3 3,49 I 00

0 (13 1.29 281)6 4.94 41)3 IL 81) 2.36 (1,8 5 .42

-7.34125
1.36668 2.97 164 1.569 1)

K, 10.66667 1.37209
2 .94298 1.57692 0.82382

168.8 0.4 1.11 In.5

(1.82852 0. 59849
((.53437 11.36439

0.63750
0.5 1695 (1.38824 O.352t I 0. 16141) 0. 10191 0.06(107
0.02717

-0.6 6.1 -3.4

6.1 3.1) 2.7 -(1.2 ().l -1.7
t).9

2,8S
628 5,66
1.75

1)34151 ((.157)02
0.10210

11.06076
((1127)4 1)1)2(135 (1.0(106))

3-5))
33.18 100

03)20111) 0.0)0060

1~

TIne results are as follows: Pressure, MPa rnmle gas in cell mole cond in cell 49.64 47.15 40.1 34.1 27.1 21.1 14,1 11,1 18.6113 5.227 ((.057 ((.397 16.815 I 709 0.827
(1.139 1)3.11,
0.09

The CVI) test can be considered to siunulumle tine fluid behaviour in tire reservoir hulk, where the connslemnsale cmnnr be rcmmsonnunhiy assurunned imnnnrrohiie. At conslitions near the producer within the condensate ring. wlneme qumasi-steundy state conditions may he assumed, the CCE test simulates the helnavioumr mrrorc clnisely mis Ilne connslemnsate flows willn the gas, and time mixtuire composition
rennnunitns umluriost connstmnmnt,

100
0 1)300

97.830 0.00()
0.293

86.648 3.682
0.271

Cmimponent Nilrnngen Carbon Dioxide Methane

P.mtnane Prnmpane iBuniane
n

76.238 6.102 kgrnol

0.246

62.434 7.676
0.207 1.2(17 55.023 5.255 2.329
0.369 OXOfl

49.6(0 7.499
0.165 0.982 44.312 4.274 1,914
1) 306 1)71))

33.149
6.51st

1.720 79.170
7.48(1 3.290
0.521)

1.683 77.452
7.318 3.219
0 509

1.554
71.457 6.760 2.977
1)470

1,418
65.01)4 6.168 2.721
0.431 1.113-I

0. (0)11 (1.678 29.928 2.946 I 354

1) 22)) II ~.l))

t3ni)mnne

i-Penianes n-Penlane

1.25)) 1)360 0.550

1.223 0.352 0.538

1.131 0.326 0.498

Hexanes
1-teptanes Oclmnnes Nonanes Deeanes tlndecanes-plus Tonal

0.620 1.000
0.710

0.607
0.978 0.695 0.460

0.563
0.911 0.644 0.425 0.282 2.061 90.331)

0,299 1L456 0.5 19 0.844

0.258
1)395

1)2(7
1)332

1)161
((.249

llrc sleposited conslensunte in a reservoir can punrtially be recovered by injecting a lean gas into Ilme reservoir, and producing time gas enriclned by the vaporised condensate. The produced gas is stripped from its imrternmrediate annul heavy componeumts at tire surface prior to being recycled huuck inrto tine reservoir. rime recyclinng process below lIme dew point can be simulated in the lmnbormmtory by iunitiunlly comrducting a CVD down to tire recycling pressure prior to introducing line recycled gas into tire equilibriunmm cell. TIme recycled gas is introduced at constant pressure by mullownng tine cell voltnnne to expumnd. The cell is retumrnred hack to its initial volume by expelling the equilibruitesl gas at comrstant pressure. TIne produced gums volume and comirposition, unnsl tine shrinkunge of cnnnndensate mIre nsneumsurcd. Figumrc 2. 17 slnows the results of methane cyclnnng of a Nornh Scum gums condensumte. Note that line hiqumid volume fraction decreases by 20% unIter connl mmcliung twin lnnnre x~ mlninnies nil imnjccled nsnctlnunmne. IIc flow of Iluinds lm)wmurds tIne wellhnnrc will establish ur pressure gradient within the reservoir, As tine guns volumnnre irrcreunses at lower pressures, and the flow area is reduced near the wellbore, lIme pressure grumdiennt svill increusse shutrply as tine well bore is approached. This will result in a nunpid flow of ricin guns unto tine near well region and a liqumid build-up in pores at a fraction much Iniginer Ilnan Ilnat mnneumsunresl in CVD test, uns tine rich gas keeps flowing towards the low pressure regnon sleposiling urmore condensate. Winen tire well will be shut-in or the rate reduced, the pressunre will huuilsl-urp agurin. lIne belnaviour of tire mixlure composed of a high liquid fraction coumlsl be liqrnid-inke, inmslead of gas-like. Clearly the CVD data cannot provide adequate murfornnmatuonr to evalumunte mmmd tune the plrmnse behaviour unodel for sinmiulation of such a buildup test. A laboratory test siminiluur to tinat of guns cycling, hut using the original gas instead of the lean gums. follnnwed by stepwise pressure increuuse can simmrulmmlc the above process and provide phase tnutmunvumnnnr dmntmm for uls munodelliung. Figunre 2. 18 slnows tlnc variatiomr of fluid saturation pressuire mncmrr tIme wcllbsmrc fnnr mu Nnrtlr Scum gums cmnmndeursate flumid in sumclm a test. The fraction of pore occunpied my conslensumte iurcreascd irounn 28 % to 63% due to the inflow of 1.1 pore volunre of gas. Note tinust tine accunrrulation of condensate did not significantly change the mixture

11.171)

0.456
0.751 0.530 (1.354 0.238 1,848 70.110

((.392
0,661) 0474 ((.323 (1.220 1.789

(1.3 I 2
1)548 (1.40)

(1.239

0.595
0.393 0.262 1.951 82.340

0.470 (1.311) 0.303 2.250 2.201 100.000 97.830

0.293
0.204 (.758

57.11)0

39.71))

11.447 0.356 ((.273 (1.193 1.754 23831)

At each stage the number of moles determined as, n~=n~yj.The number phase is then deternnined as. n ,=n-n ,. 0 0 nhe following table, and connnpmmred svmtin

of each component in the guns phrase camn he of moles of each connponenni inn the condensate The calctntunussI resmnlls am mhc lumsi sluugn mire givcmn inn line unucasur rud cmi mmcc. ltmc per de?mmumgc I tmV I rut minim In I calcnmlaled equilibrium ratios fronnr mIne nnneasnrred valures imndicuiics (Ire reliumtirlnny nil experimental data, except for nitrogen.
Component mole in mole in cal.
Xm~

meas. n calculanect

nrnea.sured

Km desnmntn,,n

64

2. PVT Tests wmd (o,rjIani,nnc

2.2.

PVT Te.nis

65

saturation point for this tested rich gas condensate system. The systenr, imowever, changed from a gas-like to liquid-like fluid, showing a bubble point at the last tested slage. instead of a dew point.

immcreasing the dew point. i-hence, it can promote condensation, if the reservoir pressure is not far above the initial dew point, instead of preventing it, However, it will be limited only to tire 7.one near tire inrjeclor, wimere metinmmne is rnnixed with the original fluid. The overall nnixlure will becourmc progressively leunner, resnnlling in a lower liquid drop-out during late stages of produiclimnmn, winen tine injeclion will he stopped unnml line remaining guns produmced by depletion.

40 35 30
35

30
25

CI Vol./iniiial Vol. at 34.47 Cuinnulamive (I linjecled DO 0033

25
~ a
C

20 . 15 10

I,

20

LI

50

100

51)

200

C a o is > a a a. no

o
0

0068
0.133 0.169 0.401 1.080

o
A
X

Cum. Vol. CI Added/urinal VnnI.%

5

1.749

Figure 2.17. Reduction of condensate volume in methane cycling at 373 K aurd 27.58 MPa. 32.30
0 60

25

I U

30

35

40

45

50

Saiuration Pressure Liquid Fraclion 0

Pressure, MPa

32.25
0,
rl

Fignure 2.19.

Varimn(ions of dew point and liquid fraction in CCE test with methane injection.

;~! 50
~.

32.20
32.15 40
:
-

ii

a Sm
0.. C

Altirouglr laboratory datur gemnerated on gas condensate fluids can be used directly in reservoir studies, Ilmey are often used to time a phmmse belmaviour nnodel, Section 9.3. The model is then used convenmiently mm siirninlation of time recovery process. 2.2.5 Volatile Oil PVT tesls oim volatile oil sunmnnples are imot well detiimed aund slocunneunted. Tests simrmilumr In) tinose described for black m)ils mure commnurmonnly comrducted on volatile oils. As tire cvolvcd gas plmase below tIme bumhhle poiunt almnnost imnmmnrediately becomes immobile, tire differential test seenms to sinnmumlate ttre process. however, tine immobile gaswlnicir is produced with the oil heiraves as a ricir retrogrumsle gas ummmd contributes significantly to the collected liquid at the surface conditions. Time production of volatile oil by depletion is not an effective method for optimum oil recovery. As tine pressure falls below tIme bubble point, a large volunnc of the gas is prosluiced which unnay altuninr mr nnnobility exceeding Ihat of the oil, resulting in a large gas production and leaving the oil belminrcl un tire reservoir. Tirerefore a constant volume depletion lest, similar In) that for gas condensmute is sonnetinnmes eourdunctcd. None of the pressure depletion tests commonly conducted in laboratories can sinrulale lire fluid heimuuviounr mrs occuirs inn lIre field. llne tests should, however, provide sufficient conrpositional uumrd vnlunmmretric dumlmm for Iumnirng of a phase behaviour nrodel. The constant comrnposition expumnsion test at tIre reservoir tcunrpcrature provides most of time required data our the oil behaviour at reservoir conditions. The anmount of condemrsate collected from produced gases in separator and differemntial liberation tests should also be measured and reported.

cm 32.10) a
(I, -

30

Bubble Pmiinn

>

32.05 32.00

zmi

50

100

150

Total Gas inflow, Vol.fPore Vol. % Figure 2.18. Variations of the fluid saturation pressure due to gas inflow neuur time wellhsnre. It is a common practice to inject sonmre lean gas into a gas commslensumle reservoir to cmmptlmre mis much as possible of the pressure declimne and possibly nmumintain its presscrrc atovc time slew point to avoid the liquid loss by deposition in reservoir during pressure depletion. Figure 2. 19 shows the results of a laboratory test, simulating pressure maintenance of a rich gas condensate reservoir by methane injection. Note that the addition of methane to the rich gas has resulted in

66

2 JVT Tests mnnnm/ (,nrrm!mnlim,nm.c

2. .1. !~nnnpnn meal (mnrrelanna,n.c

67

An on site simple pressure-volumme test of tIre collected oil smnnnrplc is un umscfunl gnnidc Ia iulemntify the oil type and decide on the required tests. Tine clrmnnmge of slope mO hunfiblc poiurl is less pronounced for volatile oils in comparison with limat for hlumck unil s:nurnples. TIne slope clnmmnges tm so gradumally for very near crilicmnl oils tinumt tine huibble poinnl n nnmmy nrot Inc delesteul. 1 mnr sunc hr mm fluid, a visunal method. sinmilar to tinat for guns cmnnrdcunsuuns. is pncfcrmunl. Figumne 2.21) smimrnpmmnes the pressurevolume behaviour of a North Sea volunlile oil willn tlrunt of tIme black oil, slescrilncd in rable 2.1.

stimulies (34-38(. A few of tIne nrmorc widely used correlations are givemr in this chapter. Table 2.3 provides infnrmnnmntioun our time raunge of data used in tire correlations to help selecting a correlatioum for a specnlnc cunse. Iuutnle 2.3. Ranrecs of data used inn bluick oil correlations,
(nmtnchniim,n

Ret N,n: l(nnlit,Ie I,nnnmi tnmssurc. psiun lcnnnpcrum)mnre, ~I tm,nnnn.Vm,I tuac,, 1,61/Sn II Guns/Oil Raiinn, .S(F/Sllt
lank Oil (Iraviny. API Guns Specilnc (iravimy 0, a

Stumndinig 39 13(1-7(118) 1011-258 1.0242.15 211(425

l6.5-(,3.l0 0.59-)).95

1.unsunlcr

tI) 48 57811 82-272 32905

17.9-51.1

i~gg~,_J~laso 141 42
5.6055 162-181) 1.1)28-2.226
-

165-7142 8tn-280 1.025-2.588

90.2637 22.3.48.1

Murrhoun 43 130-3573 74-240 1.032-1.997

II-2t99 15.3-59.5 60-565

26.1602
19,4-44.6

1)574-I .22
IS 605

1)51 u-l.351 76- 151)

Sepumrunnnmr lrcssnirc. psnrn Scn,mmralnnr lenmnns

265.465 101)

36- 106

0.650-1.276 -, 415 125

0.752.1.367

lIne sclcclcdi correlumtionrs are presenmtesl inn this section using fneld umnrits as follows: P : Prcssumre, psimi

: Temnrperuutnmre, skgrees Fumlrremnlneil. F (= I .8K-459.67), in oil and water correlations., degrees Rmnmnkiune, R (=1.8K), mr gums cmmrreluutiomns: Eqs.(2.58-76). v : Mini umr voltmmnre. ft /Ibnniol (=0.062428 unr /kgmol) p : I )ensity. Ihnnm/fI (= I 6.() IX kg/Inn) I~,:I )icsnnlveul nnr litneruulcst gums. S(F/hbl (=5.6146 nnn/nnn) Vmnlunnc, cm 3

2.3.1

Black Oil

Figure 2.20. Comparison of pressure-volume behaviour of volatile oil and hlmuck oil. 2.3 EMPIRICAL CORRELATIONS

Blmnck (nil comrcluntionns (remnt line oil ums comrmposed (nf two connnponeumts, i.e., the stock tammk oil and tIne collected dry gums at stamrslard coirditiouns. Euncim comnrponent is clmaracterised by its specific grmmvity. Ann accurate prediction of tIre phase behaviour of complex multi-component systems witir only a few vuuriumhles slronntd not. liunwever, be expected. Black oil correlations simoulcl cununlioumsly he usesl for volatile oils. lineue arc ml Imurge ununrrber.of correlumliniurs to detemmnmine properties of a typical black oil. All tIre corrcluntionrs unse tIne reservoir temperunturc, guns and oil specific gravity, and the solution gas to oil rmulio (0 de(ernnnimne nIne properties of smutunratecl oil . .Several autlrors have provided correctioum fmncnnins to iunclunde tIne effects nil nnnnnlnydrocmmrhonr counipoumrds and separator conditions. All time mrsnllrors lmumve unsed un large nunnnber nil experinrmentunl clmmta to regress the parameters of theiir proposcul correiuitisnurs Inn umniuninnise tIre slifferences between tIre predicted and measured values. Stunnidmug 3~) tusenl mm lunlmil of 11)5 ulmmtmn pmninnts on 22 dnfferent crude oils fromnr Califonmna to develunp Inis cmnrrelmnnioums. Lmmsumler (40( prcsernted mm htnhble poinrt correlation usimrg 158 measured bunfihle point dmuta orn 137 crude oils frornr Canadms, Western and Mid-Continental United States unnsl Sourtir Amrrcricun. Vasquue7. and Bcggs (411 developed correlations for the solution gas to oil rmntio mnmncl fornniuntioun volunurre faclor ursing 6004 data points. Glaso [42] umsed data fronn 45 oil cmnlniplcs mnroslly frminrn tIne North Scum megiuinn Inn develop Iris correlamions, Marlmoun (43] used 160 inuibble poiurt dmmtui nnnm 69 Middle Fmmstennr crunde sumnrmples to develop a bubble point pressrurc ceo mel mit our. Ammmcd 1441 unsent tIre cmnmnnhi ned rcporteml duun mm of Glaso and Marlnosun to develop a cuinnclmnliomn Ior sletermnninniung the oil fnrumnmu ron volunnune fmncnor, Asgarpour et al ,[45J. i..abedi [461. mnnnd tetrosky-Furrslnmmd 1471. unsesl sluntun Inn lluids ~ronrm reservoirs inn Western Canada, Africa, umsi Iexuns-Losmisimmmnun, respectively Inn develop vunrious cnrrelationms. De Ghetto Ct aI.[38] used mnbourt 370t) nnnemnsumne(l (luntun psiiurts nor 195 crusie oil smmmnnples from the Mediterrumnean Basin, Africun, Pcrsiunn Gunlf. umnsl Non-lb Seum. Ia evumluale pumblisired correlatuons, and modified some of tlnemin Inn innprove predicted resiults.

Mumny investigators Inave umsed PVT lahormmtony test resunlts, unnsl field ulumlun. to devclnnp generalised correlationrs for estinnnating properties of reservonr fluids. thnc numinn properlnes whiclr are sietennnined from ennnpiricmnl correlatiouns unre ilne hunhblc poinnl. gas solnrhility. vnilurmnnc factors, density, cousnpressihility, and viscosIty. TIne correlations ypncmnlly nrnuntclm lIne eurnploycsl experimental data with an average deviation of less (Iran a few percent. It is not uninsunuml, however, no observe deviations with an order of rnmagnitude higher when mnpplied to nlher flumids. fine correlun(ions can he classified broadly into two groups. Fmrst, those wlnichn consider oul, gas. and water as three pseudo components, annd treat a reservoir fluid as composed of tlrese pseudo components. The second group consists of those correlations which use tIre flumisl conmiposition, typically identified to C by discrete eoinnpounds ansI tlmc rest uns C +, to estininmnle 6 7 the fluid properties. rhe first approurcin, is the connnnnour orre. Tine reliumbility mif tlnese correla(innns sigunilrcantiy depends on the reservoir flurid cinaraclerislncs. If lIne flunnd is lypicmml. and falls withiun the range of tested fluids insed in that particuitmur corrciationn, mini acceptable accuracy can be expected. There are many fluid property correlations. A number of tlnese correlations Inunve unsed dala of certain localities, hence, their application is limited. Sounne correlatiomns hunve received lnigimcr attention and wider acceptability than others. The correlatiours inave been reviewed ammsl compared by several investigators, resulting in no clear superiority order amongst time correlations. Sonne of them, Inowever. have slrownn tincnr reliumbility in vmnrioums comnrpunrativc

68

2. IVT le.st.n amid (orrelatnn,,m.s

2..). Fnipiriea/ (arrnlatian.u

69

The main application of these correlations is the estimation of reservoir fluid properties uusirmg field data. The gas evolved at the stock tank is often vented and not mimeasurecl. As lIne annnoumrnt of vented gas could exceed 10% of the total dissolved gas in the reservoir oil, its vunlue slnould be estimated and added to the gas volume evolved in tire separators. Time solutiotr guns vented umt the stock tank can be estimated [48] from the information on tire last separator hclire tire stock tank, log(R ) 51
=

A deviation of about 15% is expectesl fronr the above correlation (37,38]. Vassiuez and Beggs [411 point (nut that the gas gravity depends on the sepunrator conditions, Hemrce, the authors tised the gas gravity normalised to a separator pressure of 100 psig. [(C R, 1

,se,

0.3818 -5.506 log(S ) +2.902 log(S ) +1.327 log(P ) -0.7355 log(T )

0 85 5 5 T <l40 un}m 5

(2.20) a

) (I

~yn

(2.23)

whnere,
=

-C (API)/(T+460) 3

where, R t is the stock tank vented gas in SCF/SlB, ~ is tine stock tumumk oil specific grmtvmly. 5 Sgs~ P~and T , are the separator gas specific gravity, pressure, amnd lemnnperumlnnrc (F). 5 The gas gravity used in lire correlations is the avermuge vunlunc of mmli snillecled guises frommn nIne separators.

atusl lIne v,nlures of tlre coefficreurts are:

n met ticie,nn

AIu<3( )

27.62

A Il >0)0 56.18 (0.84246 0.393

S =~R, S~I~R 8 1 5

(2.2!)

11.914328 11.172

5 Sgn is tine guts nororahised specific gravity related to tine separator gas gravity, g by,
Sgin
=

where j refers to the separation stages, including the stock tammk oil if infornraliomr is available.
S Example 2.5. 8 5 [t+s.~ 12 (or ) (API~l~ lnng (P / 114.7)] 5

(2.24)

Estimate the evolved guts from Good Oil at mIre stock lzmnnk coumnlitisnnis svitli selnuurzitmnr pressures equal to those reported in Table 2.lE. Compare the results wimln mnneasured values, Solution: The evolved gas is estimated using the separator pressure and lemperalure (75 F), ansi gas specific gravity data, and the stock tank oil specific gravity as follows: P~ 65 I IS
215

wimene t.~ mmnrul Is mire (Ire muctumuul sepuurmulor pressure, psiun, mmmd tcmnrperrtture, F,respectively. Example 2.6. Eslinmnate tIne hmmhlnle point of Good Oil at 220 F.using the measured separator test dala at 1(X) psig, Table 2.IF. Salrmtio,n:

API 40.5 40.7

40.4

S, 0.823 0.822
0.823

S~ 0.840 0.786
0.732

R, calc.

R~ incas.

% dcv,

45 80
148

41 92
178

tO -12
-16

IIme notal gas in solunlionr. amrd gas specific gravity, umsimng Eq.(2.2 I), are calculated R,=2~R,=676+92=768 SCF/STII
S,= ~R,S /~R,=(676X0.786+92Xl 5 ..3631/768=0.855

315

40,1

0.825

0.704

217

246

-II
lIne bubble poinil ms cstunnnmuisd using tIne Stuumnulnmng cminneluiuion, iudl.(2.22), 9 mn=t).000 I x220-0.t) I 25x4t).7=-0.30855

Bubble Point Pressure Standing initially produced a graphical correlation (39] for determining lIme humhhle l)oilrt pressure, and later [49] expressed the graph by the following correlation,
083

1 (1o)~ 1.41

P =2503 lsimn 5 Alternatively, tIne Vmmsques-I3eggs correlation, Eq.(2.23). can be used to eslimate the bubble point. As the selnarator pressure is 100 psig, S~,=S~=0.855

~nn=

18.2[(R~/s8)
a Pb R 5

(2.22)

where =0.00091T - 0.0l25(API) = bubble point pressure, psia = solution gas to oil ratio SCF/STB = Temperature, F

hence,

a=- i0.393x40.7/680=-0.62205 16

P =274 I psia 5
The nneasumresI value is 2635 psia.

66

2. !tl

Tnciv mmnnd ( nmrnmlatimm,n.s

2.3. Ennnpinca! (orrelamno,n,n

67

An on site simple pressure-volume test of the coliectcsl oil smnnnmplc is mn umsefunt gnnisle nun idcmmtify the nil type and decide on tIne required tests. The clnannge (if slope u(t buninble poiunt is less pronoumnccsl for volatile oils inn comparison witin that for black nrl sumurnples. FIns slope clrmmnges so graulual ly for very nemnr criticunl on Is thunt tine humtmhlc pun mm nnnmmy mnot inc duIectnul. I nir ,smnutn mm fluniul, a visunal nretinod, siunni tar to that for guns coniderisalc, is preferment. Figure 2.21) cmnnnnlnmmres time prcssure-volurrne behaviour of a North Semi volumtile snil wi(ln tlnuml of Ilne black (nil, descrilnesl in Iahle 2.1.

134-381. A few of time mnmorc widely used correlations are given in this cirapter. Table 2.3 providcs inforunmumtiomn our tine ramnge of datum used in tIme correlations to help selecting a correluntioum fmnr a specific case.
slundics

Tumble 2.3. Ranmges of datum used irr black oil correlatnomns. (narcluniiminn lumindning . I,unsr 4)) lt.t. N,,: 19
I(ni)liIc tnmnnn( Incssnnre. p.nun Iennnpcrunminre. fu tm,nnmn.Vn,l Fac., htI/SFl) 13)1-7)8%) 11%)- 258
1.024-2 IS

48 5781)
82.272

Va.squcd.Bcgg~Olaso 41 42 15.6)155 165.7142

162.18)) 80-280

43 13)1-3573
74.240

(iuns/OmI Rmmuinn. SCF/SII3 lank Oil Gravimy, API

(las Spccilnc Gravily

Separaimr Ircssnirc. psia

0.

21)1425 16.5.63.8 059-1)95 265.465

3-2905

1.1)28-2.226 0-2199

I 7.9-51.1
(0.574-1,22 15-605 36.106

15.3-59.5
0.511-1.351 6))-565

1.025-2.588 90-2637 22.3-48.1

0.650-1.276 415

1.032-1.997 26-164)2 19.4.44.6

0.752-1.367

Sernarator Icnmnp., F

1(8)

76-ISO

125

lime selected correlumtionns mire preseurled inn (Iris section usiumg fueisl utmils a,s follows:
a-

I : Pressure, psna degrees Funirremnlneil, F (= I .8K-459,67), tennnpermununre,

in oil and water correlations.

90

tIll)

I III

12)) 1

1.0(1

(1)1

degrees Rmnmrkimne, R (=1.8K), mn gas correlmmtnons: Eqs.(2.58-76). v : Mnlmmr volonnne, ft m/Ibnmnol (=0.062428 unnIkgmol) p : I)smnsity, Ibumn/fI (= I 6.018 kg/urn) I )nssnnlvcul nnr libermmtcul gums. S( I :fIilnI (=5.6 146 urn /mnn )

Vn,tunsnc, cnn

2.3.1

Black Oil

Figure 2.20. Comparison of pressure-volume behaviour of volatile oil and hlumck oil.

2.3

EMPIRICAL CORRELATIONS

Blmnck oil corrclunlionrs treunt lIre oil uns conrnposed of two conrmponeirts, i.e., the stock taimk oil and tIne collected dry gums at snandumrd comndilinnns. Each counporrent is characterised by its specific grmtvriy. Aim accuraue prediction of tire phrase behaviour of complex multi-component systems willr only mm few variumbles shnourld Imot. Inowever, be expected. Black oil correlations shouidL caunliously he umscsl fsr volatile ouls. lhcrc are un luurge nnununrher cnf correlumtionms to detenmnine properties of a typical black oil. All time coruelunlions unse tIme reservoir temperunturc, gums and oil specific gravity. and the solution gas 1cm oil uumtio 10 sleternmnitne lIne properties of sumtuirated oil . .Several authors have provided correctiorm fumc(nirs to iunclumdc tine effects nil nnnrr-Iryslnocarhonn compounds and separator conditions. All the unuthnors Inave umsed mm large nuinnrbcr nnf cxperinmrental clmutum to regress the paranreters of their pnoposed correluitiours to unninninnnse tire slilterences between time predicted and measured values, Sturund imng 3i) umsed mu tnnluml nnf 11)5 duu(mi points ann 22 different crude oils from (Talifornnia to develop Iris cmnrrelum(ionrs. Lumsunter (4(11 presented un buibble poinnt correlation using 158 measured bubble point dmmta (nun 137 crude oils fronn Canmasla, Western and Mid-Continental United States ansI Soumtin Aunrericum. Vassitnez unnd Beggs (411 developed correlations for the solution gas to oil ratio anmsl fornimmtion volunnrme factsnr unsing 6004 data points. Glaso [42] used data fronim 45 oil smmnnrples nnnosnty frnnnmn tIne Nomtln Sea negiour to develop his correlations, Marhoun [431 used 160 buibbie poimil dumma mm 69 Middle Emistermr crumde sunmniples to develop a bubble point pressure eorreluuliomr. AInunrcd 44( uuscd tIre cnmnnrhimmcnl reported slmutmu of Glumso and Marhoun to (lCVCIop a corrctuntion tunr delernmniuninng Ihe unit funnnnnmmlion voln.iunc fmrc(or. Asgarpounr Ct al.[45J, Lahedi [46(, unnnst lelrosky-Funrslrmmsl (47(. unsesl slmrtmu on flunds fronnn reservoirs in \Vestern Canasla, Africa, mnnnul lexmms-Loumisiumnimn. respectively tnn develop various correlationrs. De Ghetto et aI,(38] used umhiniunl 3700 mnncasnmmc(I dmntun pnni nts mm 195 crude (nul suuu(rples fromsm the Mediterranean Basin, Afriemm. Pn.-rsiumnn (4nmlt, umumul Nnnulhn Ssmu. no evmmlurale pumblislmed correlations, and modified some of I lmc urn Inn Inmnprove predicted results.

Many investigators have used PVT laboratory test restults, ansi field dluntmr, 1(3 develop generahised correlations for estimating properties of reservoir flinids. The nnumin properties which are determined fronm ennpiricmtl correlations are tIne hunhhle point, gas sunlunbility, volsnmnnc lmnctors, detmsity. compressibility, and viscosity. Time correlations typically nnuntclr the euniplnnycd experimental data with an average deviation of less (Iran a few percent. It is not umunustimmt. however, to observe deviations wutin an order of tnmagnitu(Ic higher when umpplied to ntlmer fluids. lhc correlations can be classified broadly into two groups, First, Ilnose which conmsider oil, gas. and water as three pseudo components, and treat a reservoir fluid as coumnposcd of llrcsc pseumdo components. The second group consists of those correlations which urse tIre flunid composition, typically islentilned to C by discrete compounds and the rest as C +, to estinnnmnle 6 7 the fluid properties. rhe first approacin, is the conrmon one. Time rehimnbihity nil tlnese correlmn(ions smgunifncantly depends on the reservoir flurid cirunracterislics. If time flunid is typicmmi, amnd falls ~vithmnthe range of tested fluids used in that punnticulmnr correla(iomr. ummn acceplunhle accuracy can be expected. There ane many flm.nid property correlations. A number of these corrclatiomrs Inurve nnsed slumnum of ccrtainm localities, inence, their unpplicmmtiomn is iiuinited. Sninnme cmnrrehumtionns Inmuvc m eceiveul Iniglier atnentnonm mnumd wider acceptability than others. line conetalions hnunve bccnn reviewed mnnnsl comtnared by several- investiguntors, resunlting in no cleuir sumperiority order mmmnnmngst tIne correlations. Sotnne of tirem. however, have simown tlreir reliability in varioums conmnpuurmmtive

68

2. IVT le.cn.n aiim! ( n,r?e!alun,m.u

2..1. EnipiraI (orrelatiou.c A deviatisnn of about 15% is expected frornn the above correlation [37,38].

69

The main application of these correlations is the estimation of reservoir fluid properties using field data. The gas evolved at the stock tank is often vented and not measured. As tIre arnoumnt of vented gas could exceed 10% of the total dissolved gas in the reservoir oil, its value slmould be estimated and added to the gas volume evolved in tire separators. Tire solutiomm gas vented mit the stock tank can be estimated [48] from the information on tIme lmmst separator before mIme stock tank, log(R t) 5
=

Vasquez aurd l3eggs [411 point out that lhe gas gravity depends on the separator conditions. Hemnce, the authors used the gas gravity normalised to a separator pressure of 100 psig.
}Ci
=

0.38 18 -5.506 log(S ) +2.902 log(Sgs) 0

+ 1.327

log(P ) -0.7355 log(T ) 5 5

(2.20)

i(~nR, is8,) (io)~

3 (API)/(T460)

(2.23)

where,
a = -C

T <l40 F 5 where, R is the stock tank vented gas in SCF/STB. Sn, is lIme slock lammk (nil specific grunvity. 51 Sgs. P~and T , are the separator gas specific gruuvity. pressure, amrsl temnmperuntunie (mnrm), 5

arid lime vuilnues of line coefficients mire:

mel tncicnnm (2 C 3 Al 1<3)1 A I >0)1 56.18 1)84246 10.393

The gas gravity used in the correlations is the average valtne of mill cm,Ilected gumses
separators,

frommn lime

27.1n2
0.914328 11.172

S =~R, S~/~R 8 1 5

(2.2 I)

wherej refers to the separation stages, including the stock tank oil if inforunnationr is umvailahlc. Example 2.5. Estimate the evolved gas from Good Oil at tIne stunck mumunk comndilmunmns svilli sspmirmltnr pressures equal to those reported in Table 2.IE. Compare tIre resunlms with nnneasured values. Solution: The evolved gas is estimated using Ihe separator pressure and lemperalure (75 F), and gas specific gravity data, and the stock tank oil specific gravity as follows: P,
65 115

Sgn is the guts normalised specific gravity related to tire separator gas gravity, Sg, by.
Sgnn
=

[1+5.912

5 (or ) (API)l~log

(Is /114.7)]

(2.24)

wlmere P~ mmmd Is mire tIme mmcluru,I sepumumitor pressuire, psium, mmmd temperumture, F,respectively.

Example 2.6.

Eslimnnumte the hunhlnle point of Good Oil at 220 F,usiung the measumred separator test data at 1(X) psig, Table 2.IE.
Salmmtio,n

API
40.5 40.7

S,,
0.823 0.822

Sr 0.840 0.786

215
315

40.4
40.1

0.823
0.825

0.732
0.704

R, calc. 45 80 148 217

R, incas. 41 92 178 246

Sim dcv.

II) -12 .16 -II

IIme tolmni gas in solummion, arid gmns specific gravity, using Eq.(2.2 I), are calculumted R,=2R,=676+92=768 SCF/STB

S,= 2R,S~I2R,=(676x0.786+92xl .363~/768=0.855

Bubble Point Pressure Standing initially produced a graphical correlation [39] for detenniining tine bubble poinl pressure, and later [49] expressed tire graph by lime followimmg correlurtion,
0.83 =

Ilns tiulihlm. Pnnnimi us cslinmmumlcd nnsimrg tIre Slunnndunig cnnm nclmmiinnnr, Fq.(2.22), uu=O.0009 I x220-0.() I 25x40.7=-)).30855 l~=2503Insimi

18.2[(R~/S ) (10)a _1.41 8

a Pb R T5 =0.0009lT - 0.0l25(API) = bubble point pressure, psia = solution gas to oil ratio SCF/STB = Temperature, F

(2.22)

Alnenranively, tIne Vmnsqures-Beggs correlation. Eq.(2.23), can be used to esmimane the bubble point. As (Ire separator pressure is 100 psig. S~,=S~=0.855 Hence, a=-IO.393x40.7/680=-0.62205I6 P =274 I psia 5
tIne mnieursumred vunlue is 2635 psimu.

where

70
Gas i,r Solutio,n

2. !VT if, lu

,rnl(!

( ,~?!nhIti(nnS

2.

1.

/~n~~,nrin a! ( n,rrelaiionn,c

71

All the bubble point graphs and correlations can be used to estimate the announnt of guns in solution at a given saturation pressure. de Ghetto et al. [38] comrrpared tire corrclationrs for preducting the solution guns, and foumnd the correlationr (nf Vursqumez-Bcggo, 1:ml.( 2.23). moore reliable tinan others with a standard deviation of 29,5%. Tine Iiberaled guns is cumlcunluited mis line difference between the gas in solution at the originnal bubble poinmt nnninnuns that ml lIne opermmlinng pressure. Oil Fornua ham Volunie
Fin c/or

is Ihic oil fornmuilnonr vohunnine fumclor mul tIme ~mrcssunre p. and C is tire average oil isothenrmal 0 sonnrpressuimmlnty coellmciennl over lIne pncssumre range of P , to P. 1 line valunc (if tire isotliennnmml connmpressrhihily coefficient, C , can be estimated from 141], 0 5 C,, = (1433.1) + 5.0 R + 7.2 1 - 1180.0 Sg + 12.61 APi) / (10 P) (2.30) 5 lime unbove correlumlmoum us helneveul to getmeraily unmmdeupredict tire coummpressibility, particularly at Inugir pressures [37], wrtln urn average unbsolute deviation of about 25%, Whmunm time preusunre grmmuliennt of mm stmll ic nnil (uilumnninn wit Ininn tIme reservoir is knowmr. the oil density is cnnnnvennicmntly cuutcunlmilcd froini, 144 (dp/dln) (2.31)

The nil formalion volunme factor of saturruntesl oils lnas been cmirrelunleul by mm numnnrher nnf investigators using tIne gas in soluntion R , (gas to oil runnmo), guns grmnvny. nnil grulvily muurd 5 reservoir temperature as the correlating parameters. Standing iniliunlly produced a graphicunl conelunn ionm (39] for Cs) immrmnlinng tine oil fnnrnmnum( mum voluninic faclor, and later [491 expressed tine grapln by the follownng correlatnnnnn, B, = 0.9759+0.000l20[R.(S /S )+ l.25T] 0 0
2

wlnen e tIre pressurne gimidicmrl , dp/uhhn. is nm ps/fl, mumnsl Pun is tire oil density umt the prevailing
pismnn~

umunml Icurnl~urre inn I1mm/f t~.

(2.25)

tIne miii fm mmin:mt mum vohmnnne fund m nr, I lie mm, cuinn he stclernnii med by the trnunlerial balance equnation for mimic slmick tummnk barrel mnf mill, rcstnllinrg inn,
=

The Vasques and Beggs [411 correlmmtion, whiclr accounts fur time scpumruilnir prcssnmre is mms follows, B =l.0+C R,+(T60)(API/S )(C +C,R,) 0 1 80 2 (2.26)

(62.4 S,, +0.Ol36RsSg)/P,,

(2.32)

Ltainnple 2.7.

where Sgn is the normalised gas gravity, Eq.(2. 24), anmd the values of tile coefficicmnts unre: C,ncITncient
C1 C2 C3

Fstmnimum(e lIne isn)tlierinmuih connipressihility coefficient (if Good Oil at 220 F and 4500 psig. (omumpare I Ire resin hr wi nh lIne imnemisinned smmluue. ,no/nmiio,i : FIne nsolhermnnmnl coninpressibmhnty cmiefficiemnt is estimnated using Eq.(2.30) with the data cunlcunlaled iii Exunminple 2.6. R,=7(n8 S(F/SIE3 S,= 0.855 AtI=40.7 T=220 F P=45 IS psia

APR30 4 4.677x10

API>30
4.67OxI0~
I.I00sl()~

5 I.7Slxtt) 8 -l.8IixlO

I.337x10

lhe oil fornmnation volummne factor earn he cslinnnuulcsl witln mn slevimutmann less IIran 5% fanmnn time mmimn,ve
correlationrs (37,38].

(,,=l.26x10

~nui

The Arps correlaiic,n (50] can be used to roughly estimate the oil formation volunnme faclor wlnemm the properties of gas and oil are not known. B 0
=

tine result of pressinre-vohnmunre test ml tine reservoir conditions, Table 2.1 B, can be used to curlcur late mine 1)11 cnnrnpressihilily.
(,,

1.05

0.0005 R

(2.27)

)V/,~P)/V)~, = (AV/V)/AP

TIme oil formation volume factor of an under.caiurated oil is calculated by correctinmg lire estimated formation factor at the saturration pressure for its conmpressihihity urn tIne reserv(nir (emperalurre. Tine oil isotlnermal commnpressibiiily coefficient, C ,, is nclunlcsl to the oil fornnmatiuimn 1 volume factor as, C, or B,,~= B,, exp[C , (P 1 where,
=

fisunig tine pressnrrc-volnime durta mmt 40()0-5000 psig, we ohiain,

ii

4000-4500 psig

urn 4501). 5000 psig :

C,,= I .28x I)) psi 3 C, = I . . 6x It) psi

(2.28)

lire mnvcrmuge value unt 4515 psiun is esiunuml to I .32xlO psi T~ta/(Tw(,-P/ra,5e) Fonnnatjw, Va/wire Factor

(2.29)

lIre tinlunl fomnnnumtnmnmn v,ilunrnie funcuor of oil unt a pressure below the original bubble point pressure cant be eslininated frnmnn a nunmniher nmf correlations. Glaso [42] proposed the following correlation,

72
2

2. PVT Test., aiim! (..orrelatio,u.c

2.3. Empirical Corre!aiio,ns

73

log Bu

0.080135+0.47257 logB+0.l735l(logB~)

(2.33)

Tine esliurrated valine by rhe Vunsqiuez-Beggs correlation. Eq.(2.26), wilh R,=410 SCFISIB is equal to 1.279. The nnemusured value is deuerrnined fromni Eq.(2.I3), B,, 6 ,,ul =1 .445x 1.474/I (00=1.331 B Bm~lh

where B~is a correlating number defined by:

5 3 B~= R T S~/ (s~ pr.m089) 5 and C = 2,9/l0~~~

The Marhoun [43] correlation is, B~=0.3l4693+0.l06253x l0~A+0.l8883xl0A where, 2

(2.34)

B,, =

Ic)
lIre loImnI fnrmnmmnn arm vnilimmnne fmmclunr cumin he estiunnumted frunnmn (he GI mmso correlat ion, P . (2.33), 0 with R,=768 SCF/SIB.
1.79902831 I1:=2..02277288 B,= I .889 TIne predicmn.d vuilmie by tIne Muirliunnnmr cuirreluml man, Fq.(2 .35). is as tolknws:

(2.35)
A= 128671.275

13=1.994 Tine nnemmsumred vatume is deteiummiimud fronmi Eq.(2.16).

=

0644516 79 4 76 9 A = RS 0.724874 (T + 460)2~062I i(s~ tp~ l I~)

Note that R, in the above correlations is the original oil gas imn solution, witlm partly still dissolved in oil at the prevailing pressure and the rest evolved as the gas phase. Example 2.8. Estimate: (a) the gas in s~Iution,(b) oil fornnalion voltnme factor, and (c) tolmul formnnmmumon volume factor of Good Oil at 1600 psig and 220 F. Assume a 100 psig and 75 oj; separator. Compare the estimated values with the measured data, Table 2.IE-I). Solution: (a) Using the Standing correlation, Eq.(2.22). at a saturation pressure of 1615 psia. estimate the gas in solution.
~

Bun + Bg (R,l,. R )/5.6I=I .33l+0.0l034x(768.482)/56I=I .858

Oil I)e,msitv TIre demnsity ofsattnrumted oil cunnn he estmmnrmnted Iromnn Eq.(2.32) using time calculated oil formmratiomr volunmrmc fumctom from unny of time gcmreralised correlumtions, The isothemmal compressibility coefficient can be umsed to adjust the calculated saturated oil density due to compression for an ummder satm.mrumted oil,
P
=

6 wlrere p,, is tine oil deursity unt pressure P.

0~~

Pun explCmn(P P ))

(2,36)

R,456 SCF/STB The estimated value by the Vasquez-Beggs correlalion. Eq (2.23), is eqummml to 410 SCF/STB. The measured value is determined by commibining the reported differential hiheratiomr ummnd separator test data, as described by Eq.(2. 15),
=

Tine oil density can be estimmmunted also by calculating the mass and volume of the oil at reservoir conditionns. Inn this approunch tIme effective volunnme of tire evolved gas, as a hypothetical liquid, is estimrmated aurd adsied to tine stock taumk oil volummme at the standard conditions. Katz [SI] sievetoped a clmmsrt for estimrratimng lIme apparent liquid density of natural gas, which was later described by the followinng correlation 149]. Pal = 38.52 x l0~0o326 Att + (94.75 33.93 log API )logSg (2.37)

ROb

(R ,db

R ,dJ B.~ , 1 B~b

winere p,, is tIne appumreol density of guns

inn the liqunid state, Ibm/fe.

R,=768-(854544)xl .474/1.600=482 SCF/STB (b) Using the Standing correlation, Eq.(2.25), with R,456 SCF/S1E3, we esniunnunme time oil formation volume factor, B,,=I .309

Fine appmnrent denmsity of tIme oil, Paun. inclunding the dissolved gas, at tire standard conditions is determnnined by dividing tire total mass by the apparent volunre for one STB of oil, as,
Pun,,
=

[(Rs/380)(28.96Sg) + 5.61 x 62.4S )]/[5.61 -l-i(Rs/38O)(28.9fiSg)/Pan

1
(2.38)

wlmicir ucdtmces to,

Pam,
=

[0.07621 RsS

+ 1St) S,, 115.61 +

(0.07621 RsSg Pal)]

74 3

2 !VT luui.u amid (u,rr,!aui,,,.s

2.3. Enrpirical Correlatia,,s

75

lIme mnmetlnod lmrcdicts iniglily mn~uunnmnlc ulemrsily sluntum fmnr lIne stock tannk oil and low humhhle poimrt where Pa,, is in lhmmn/ft , The apparent oil density is corrected for pressurre, and therm temnperatunre uns folluiws,
= +

API

APT

(2.39)

pressure oils. lIne rcmlsmrmrs for its stncccss unre line dimcct usc of urmemmsurcd C + density, which 7 assumption in ollen counmprises hulk of tire mnnmmss, mind tine validity of the additive volume imydrocarhon liqutid mixtuires. Tire nmethod also provides a reliable estimate for oil samples with lnigin bubble point pressure. Pedersen et mit. [53] uuiso presenled a set of correlations representing the Standing-Katz gm unpin icmnl nnmetinod. A simmni luir mmmcllnoui, hmnscd ann tine volumnmnc adclitiomn at the stamrdard cotrditions mnnnsl correctiours for puessurre annd ennnpcruunnnre has hcenr proposed by API [543.

where tine correction due to pressure is,

Ap~ = (0.167 + 16.181 x I ~ and tirat due to tcmperalure rs,

APT
=

)( P / 1000) 0.01(0.299 + 263 x I tr ~

)( P / 1000)2 (2.4)))

l:namnnple 2.9.
Fun immune (Inc (naomI Oil (Fumlnle 2 I )

ulcmnsity at is hnntnhte poimml. rising the Katz equivalent

[o.o I 33 + I S2.4(p~,,+ Ape)

](T

1,0) - [8. I x I 0~ - 0.1)622 x I)) (uo7Mu

1

i,,,

I ns~mmmdvunlmmmmmu nnmuItrmml
,lrutin,r:

(2.4 I ) Eqs (2.40-41) were proposed initially by Slandnng and Katz 352] as workitrg chumrts, annd lumtcr nunmerically by Stansling [49]. When the oil composition is known, Standing and Kat7. [52] proposed Ia calcuiluinc time unppunremmt oil density by (he following method.
it is assumnnned that hydrocarbon comnnpounds heavier nhnuun e)inumne (includimrg II S) uelumnTn their

lIre umppmurcnml density of ihe dissolved gas in hiq,nid state is calculated from Eq.(2.37).
6 p,,=25. 5 Ibm/It

lIne muppmnrenrt density of lIne unil, immcluchimng the slissolved gas. at nine standard conditions is euulculaied frunun Pc .(2.38), p,,,=44.66 Ibm/fl 1
1 tnc dcnnsnly carredimoims mimic nun lnrcssmirc annd temnnlneratnnre are ealeumlated from Eq.(2.40). mmmii Fq.(2.4 I), rcspectmvely, as fu,lluiws: APr=)).92 Finn/it

inmdividuiumi voliumnies as purre in tIre unnixture. determined from.

1C7,
\

Timercfnre, Ilie dcnrsi)y unf mm

(,,

nnnixtmnme is

ici,

Ap,=4.55 11mm/fm

Pc;

~ u

x,M,

if
~,,

~
ci

xM /p,

(2.42)

wlnicln resurlls inn, p,,=4 1.03 Ihmmm/fI lIme estimrmmmlcd vmnlure mnnatclmcs lime mnneasunrcd value of 40.97 Ibm/It (0.6562 g/cnm) within

where x is the mole fraction of component i, and p~is the density at the stanmdard conmdi(ions 1 ([able Al in Appendix A). Time measured density of C~,fraction is used in this unretirod. Ilre contribution of ethamme and mnretinmutie to tire appmnrcnt (nil density is linen cunlcunluntcd un turin l)y considering their effeclive volume as dissolved in the liquid,
=

nIne expernnnemmtuml uuccmnrmmcy (Tumble 2.11)). Wlmenn lIme oil conrnposutiomn is kmmowmn, lIme mnrethod of Ala,mi mid Kemredy [55] can be used witim cunnnfrdcurce to predict dcnnsily evemn for highly volmntile oils. The aurthors used experimental density data to devekip a cubic equmutiummr for time oil molar volume, v (R(l+460)
~,,---.-~

p,.,,(l 0.01 386w~

0,000082w~2) + 0

379

wc

0 0042w~

(2.43)

and P.,
=

+h v

+ miv / P mmb / P

(2.45)

Pc,.

P~ ,( 2

0.0l2w~,0.000l58w~.,)+0.0l33w~., + 0.0(X)58w~.,
ibm/ft tm
Wc,< 16

(2 44)

wlneme, v
= = = =

mnnuntmmr vunlmmunne, ft/lbmnmnnl

lcnnnpera(urrc, I ~ psium

40< ~

w,.,< II)

where w~ .and w~,are the weight percent of C mr C +, ammd C in C (tnlal mnmixtunre), 2 2 2 1 1~ respectively. The apparent density is lInen adjusted for pressunre mnnnd tenniperumlure uns inn Eqs (2.39-41). If CO2 is present at a low concentration, it can be treated our tlne additive voliumnne basis, umsmng an effective specific gravity of 0.420. The mass of nitrogen, if present at low concentrmntion, is added to we,,

10.7335, (psnun)(ftflhmnol)/R

mu mmnmsi b clepemrd aim lIne connrponmemnt mnmmul tennmperatunre, anrd for pure conrpounds are given by, 4 a = X c (2.46) h = tnm(T + 460)+ C where time comnstants in Esls.(2.46-47) are given in Table 2,4. (2.47)

76

2. PVT Te.st.c mumd correlaiimn,n.m

2.3. Enmpmrnca! Corrdatno,n,,

Cnnmnpnunent IOOx

77 a I,
ft7342478 0.8727392 1.0551130

Table 2.4. constants of Alani and Kennedy equation Component X C (70.300F) 9160.6413 1 Cl (301-460 F) 147.47333

Cl
(2

36A7
9.67 6.95

10033.56 25767.48 26784.27

ax hs 3659.240 t).2677802 2491.7 iS 0.0843939

1861.507 1)0733304

M 16.1143 30.07
44.096

Mx

585.1 290.8
306.5

~ 61.893223 3247.4533

~._

3.3162472

- 14(172637

(1.51)874103 I 83261,95
(1.52239654 0.62309877 0.90832519

(4 mr-C4
CS

1.44 3.93 2.85 4.33

((.16
0.91

(9 41)932.8(1
39082

l.331M69 1.3141589 1.5857214 1.8442812

5(n2.791 1)0191757 1608.659 00516464 1640.404 1)1(451931 3279.81)0 ((.0798574

58.123 58.123 72.15 86.177

44.01

85.7

228.4 21)5.6 373.1

7.0

C2 (100-249 F) 46709.573 - 404.48844 5.152(1981 C (250460 F) 17495.343 34.163551 2.8201736 C2 20247.757 190.24420 2.1586448 3 i.C4 32204.420 131.63171 3,3862284 mn-C4 33016.212 146.15445 2.91)21257 C5 37046.234 299.62630 2.1954785 C6 52093.006 254.56097 3.6961858 132S* 13200.0 0 17.9(8) N2 4300.0 2.293 4.490 CO2~ 8166.0 126.1)0 1.8180 The constants for these compounds were calculamed tamer by l,nnlnrcin,. en ml. 1561 For the C 75 fraction: 1

C6

57558t(3 7574596

N2 C02

1.1(113834 1.1168144 1.4364289 1.59294(16 0.3945 (1.3853 03872

C7+ lm,nal

31.29 100)10

43 n4.52 0.6901,2(10 6.91)3 0.001 050 9828.37 0.510824(1 89.438 0.0046485 176019.35 36344215 58596.841) 1.2098989

28.01 218

73797 297

I 8370294

25.5 7257.2 936

2 45 Substituting tIre mnnixtnire parmnmnneters inn Eq.( . ). witir tIme pressure and tennperuilumrc equal nun nlmose at the bubble point, restults in lire following cubic equation: v-4.60696491 v+2%.0065645v-5 I .448883=0
lIme uihmnve cnulnic eqnnunlrain Imuis u,mme mcmiI oat ( Appeindis C), v=2.266 Ii /Itnnnnuil

a~ += exp [3.8405985 x lO Mc + 7 7

9.5638281 x l0~~ M~ /S.~ 15 7~ +l0.7535l7J

248)

line densuiy is cmmlcnrlmuicd p=M/v=93.6/2 266=41 32 thnm/ii)690 kg/m) TIme estinunmited vmulune deviates almomnt 0.8 % fronn the measured value of 40.97 Ibm/fl (0.6562 g/cmrn, Tunhle 2.1 D). Oil Viscosity

+ 2.6180818 x 102 /(T+460) +7.3104464 x l0~M~.

2 4 b~ + = 3.4992740 x l0 M~ 7.2725403 S~ + 2.2323950 x lO (T+ 460) 7 7~ 7 2 1.6322572 x l0 M~ +/S~ + 6.2256545 7 7 The values of a and b for mixtures are determmnined by molar averagimmg, a b
=

(2.49)

lhe live oil viscosity is often estimrmated fronrn correlations whicir account for the effect of dissolved guns and pressure omm time viscosity of dead (stahilised) oil.

~a

(2.50) (2.51)
The viscosity of gas-free crude oil cunn be estimated frotmi correlations of Beal [57], BeggsRobimrson [58], Egbogah-Ng [59), or Labedi [60], to name a few. Beggs unnd Robinson 3583 correlated viscosity data of 600 oil samples within a wide range of pressure and temperalunre as follows,

1 When Eq.(2.45) has more than one real root, the lowest value is taken as the Iiqunid demnsity.

~b

it should be noted that the Aiani-Kennedy equation of state can be used only to determine tIne hydrocarbon liquid density, and no other thermodynamnnic properties, as its pumrameters Irave been optimised to match only the density data. Clmapter 4 provides a detuuiled coverumge of cubic equations of state. All those equations can be used to predict tire oil and gas detrsity wincn the fluid composition is known.
Example 2.10.

l.n,,~

10

(2.52)

wlrere, logA = 3.03240.02023API 1.163 logT (2.53)

Calculate the Good Oil density at its bubble point, using tIre Alani-Kennedy method,
Solution:

and ~

is the dead oil viscosity in cp at temperature T in F.

Eghogunlr amrd Ng [591 modified tire expression for A, as

logA
=

The two parameters for C~-C are calculated from Eqs.(2.46-47), with iC, and nC, added together. The parameters for C6 , are calculated fromni Eqs.(2.48-49). The resuilts are as given in the following table, with the 7 mixture parameters amid nmolecuinr weight caicuilmmted by mmiolar averaging,

1.86530.0251)86 (APt)0,5644i iogT

(2.54)

Tire correlations wlnicln estinmate tire dead oil viscosity fronr the oil gravity and temperature only are not very reliable and errors over 25% are expected from tire above correlations [38].

78

2. PVT Tm un c unnnd (orru!atnu,nu.s

2.3. Ennpnruea! (rnrre!aumoni.c

79

D=0.40886446 Beggs and Robinson [583 proposed tIre following correlation 10 cstntnnuntc lIne cifect of (Inssdnlvcd
(n,,=0.46 I cp

gas,
=

C ft,,~

2.55

lire reported i.uuhcdi cmirrclunnuonn, Fq,(2.57), cslmmnnumtes lIne nnrcasuired value is 0.45t) cp. rime viscosily at 5015 psia equal to 0.422 cp.

where, 2.3.2 Natural(as C B

=

10.715 (R

+ 100)
~

-070

5.44 (R. + ISO)

Whereas tIre pinutse behaviour u)f black (nil is controlled uniainly by its content of light cuunnnlmunurcnnts (guns), tIre bchavirurr nsf rich guns slepends strongly on tire concentration ansi slistrilnnmlim,mr m,f its iremuvy coinrponemmts. Ilcnee. reliable estimtnation of phase change and time umssmnciunlesl Inrunpcrties. nnsiimg tire sununic umlmpi(iadin as tlnal for black oil, cannot be expected. Single Inimumse gums prmmpcitrec, lnmnwevcr, cumn be cslmmnnalesl reunsomrumbly using empirical correlations. (nuts is generunily clnaracterised, in tine emnnpirical correlations, by its .specific gravity relative to air uru anne atmirosplreme mmmxl 520 R (60 F). Its nnrolecuniar weight can be calculated simply from F.q.( 1.4). lii all //ne ,ga.c eorrelaiio,m,c jim i/mis section, i/re temperature is in the absolute sea/u of Rwmkioe, Riclr gurses Iunrnnr cmnnndensuile mil Ilne stunndumrcl (lunhormmtory) cmnmnditions. Hence, their meastmred ujxcuiic gravily sln(nunlsl be adjusted lny inmclumditng lIre condensed pirase. In laboratory tests, (2.5(n) when themmmolecumlmur scpmmrumtcd gums mmmd comndemnsuntc mnre amnalysed and the overall composition is known, the nrmixtnmre weiglrl cmli lie calcunlumted by molar averaging, and its specific gravity is (leten mnnimmeul fromnm t:q.( 1.4). Wlnemm lIne c(nm(nposition is nnol known, butt the mass and molectular we i glut of guns mmmxl cuimidemrsatc urIc k nn(iWnr, Eq. (2.8) P~ umvmsles lire mixture nrolecsu lar weigirt
vuu lure,

t anui I oh is time sumturui(eul oil viscosity at its huiblmle pomirt pressnnrc. An average deviation of abouut 20% [381 is expected fronmn the unbove cOnTeluiliomi usnng the measured (lead oil viscosity. The deviation will be much lnnglrer due to comirpounrdnng of errors when estimated dead oil viscosity is used.
Vazque7.-Beggs [41) proposed the followinng correlation to accosunt for the effcc( of pressunme
Oii

the oil viscosity above nts humhhle point pressunre Pt~~ It.
=

It,,, (P / ~

)ii

where, 1)
=

2.6 pm

is?

exp(l I.5I38.98x 10 P) ITn lire absctrce of nmneasunmcd courdcunsurte mmnsnlecuilunr weight, Eq.(2.58) may [611, M,,
=

be

used to estimate it (2.58)

In the conrparative study of Dc Ghetto [38], the correlation of Labedr 1601, performed mnrore reliably than others with a standard deviation of 13%. The Labedi correlation, whrclm ms based on linear changes of viscosity with pressure ah~ove the bubble point as reported often experimentally, is as follows,
=

5954 /(AP1 8 811) = 42 43S,, /(1 .008

S,,)

+(pIp~ i)(lO-2455Itoxonspo6nsi

/ 100057

~Ari)

(2.57)

Smmbstitumtiimg E:q.(2.38), weFq.(2.SX) ohtmmimn, in Eq.(2.8). mmmd writing it for one STB of condensate similar to [(R / 5 380 28 ( .OfiSg 5.61 x 62.4S,,)

Example 2.!!.
=

1/

(R~/3X0)(28.96S )

Calculate the Good Oil viscosity at 5(8)1) p~igaird 220 F.

So/uiio,u:

. .

8.96S

0 42.43S /(1 .008 S~~)j

0

5.61x62.4S

svlnmc In will reulunce to,

Using the Beggn-Robinson correlation, the dead (nil viscosily, l:ul.(2.52), us cumlcnnlumleul ums 1.020 cp. The Egbogah-Ng correlalion, Eq.(2.54), results jim a deumd oil viscosily uif 1.152 cp. lhe measured value is 1.29 cp. The viscosity at the bubble point is calculated using Eq.(2 55), wittr R,7(n8 SCF/SI C=0.328590l3 B=0.542l982 ft,,~=O.3SS [~
0.373

[0.0762 IR S f 350 Smul/F(Rs /380) + 8.25(1.008 s )j (2.59) 5 8 0 line ncservoir gums specific grunviy cuumm he cmmlcunturtecl simply by dividing its molecular weight by lhmmt snf muir (28.96). When uusinng fielul produrclimnur datmn, Ilne gas evolved in the stock tamrk is often non nnicasurred. Gold Cl uml. 1611 cxaimniuncd experimental data on 234 gas samples, and proposed tIre follmiwing correlugion to estirsraue gas specific gravity, when only the first stage separator
Mmmi

The reported measured viscosity at the bubble point pressure of 2635 psia is

~,.

guns slmn(um is avunrlable nun mu tlnree stunge sepunrumtion process, including tine stock tank,

The effect of pressure. P=501 5 psia. on the oil viscosi(y can he estinnmated using tIne Vasquez-Beggs correlation, Eq.(2.56).

Sg

(RstSgm

+ ~a ~

4600S,,)/R

51

V~)

(2.60)

80

2. !VT Te.c~cand (orretaiio,n.s

2.3. !Dmmpirieni (orre!a,jnn.s

8i

where R and Sgi are the first stage separator gas to oil ratio (relative to the stock tank oil) and 0 specific 1 gravity, respectively, a is related to the mass of gas produced from lire stock tank,
and the second stage separator if present. V~ is the volume of gmms produnced fmom the stock tank, the second stage separator if present, and the volunre of the stock tank oil if it were gas. 0 The values of a and Ve are estimated as,
= PStUDO REDUCED ~n I 33 P;tuoo StDUCED TEMPERAnURE
.~ .

pntss

R
6 7

A~ (P B 0

, )Ai S~n(~ AP1)~ (1 14 65

460

)An (T

46

Q)Am

(2.61) (2.62)
iC

+ BnPIR2Sf(0API)R4(Tm

460)

8~

(T

460)86

~
_______

__________
-~ ______.

____

____ _____________ _______

where P and T are the first stage separator pressure. psia, mind lcimnperunlumrc. F. respeclivcly. 1 1 and T is the second stage separator tenrperatumre, if presetnt. llmc vuuiuies oh comnstmimils umrc mis 2 follows: Three-stage separation, A = 2.99222 1 A =-l.l9605 5 B = 535.916 0 B = 1.20940 4 Two-stage separation, A = 1.45993 1 A =-0.934460 5 B = 635530 0 B = 1.58124 4 A 2
=

~.

f,,,

0,,
.

~ .

.::.
. . .-

m,

~.

A = 0.970497 2 A = 0.553669 6 B = 2.62310 1 B =-0.849l15 5

A = 6.80491 3 B = 0.793183 2 B = 0.269869 6

A = l.0791(n 4 B~= 4.66120

= 07

...~

:~i
_____ ..,, _____ .~ ~

~
.~. .,. ___________

~
~... ~

____

6 ,7~ p~.....,..
~

~
.~ - .,~ ~.

5
,

~ 1.33940 A 3
=

u~

:;~~
.

7.09434

1.14356

______ ~ I7~

B = 0361821 1 B =-0.79l301 5

B = 105435 2

R = 508305 1
,

____ ~

~
~

~
i3

The estimated specific gravity by tlre above tumethod is expected to be witlnin 2% of laboratory determined value, increasing to a deviation of 6% when noun-hydrocarbon contcrrt of the gas is between 2 and 25 mole %. It is not recommended for gases containing more than 25 mole % non hydrocarbons Volumetric Data The equation ofstate Eq (I 5) relating ttre pressure volunnie amnd tensnper ntumre is adequu ute to provide all the requnred volumetric information suclr as Ihe gas fornnatuon volume factor density and isothermal compressibility coeffictent The key parameter is the connpressibmlity factor Z which can be estimated using a generalised chart Figure 2 21 shows the clrart for sweet natural gases as prepared by Standing and Kati 3623 TIre cin-mrt wns developed by using data on methane binary mixtures with etlrmnne, propmnne, mmnd hurtanc unnsi otlrer nuuunruul guises over a wide range of composition with a maxnmumn molecular weight of 40 The success of tire cinart has motivated many investigators to reproduice it by numericunl correlations. Takacs (63), compared eight correlations amenable for counrpnntcr cunicumlatiomms, both for accuracy and computational effort. The correlation of Dranchuk and Abou-Kasseimm [64] was found to reliably reproduce the data with an average absolute deviation of 0.3%. The correlation is basically the eleven parameter Benedict-Webb-Rubin equatiomm, modified by Starling [65] as will be described in Section 4.1, with the parameters determined by fitting the equation to the chart.

~ ..,.:. .,

~.

: :
,.._~ ~ ~__ ~ ~. -

~
.

. J .
. ~ .

n
_______

H~4,
~ -

..2~.. 4

=~

u2

1o~ r,~1:
-

I ~
~

ii

~ri
-.

.-

_.

-~
~ ~ - _ ___

=
.
0

== =
:~=,

~
=

\ :

= =

no

~
I

=
ii 2 3 4 nm~ ~

PSEUDO REDUCED PRESSURE

2 ~

I. Compressibility factor of natural gases. SPE Cmipyrigtnm. Repnsndnnced

frmnm

1621 wimlr

82

2. PVT Te ,n c annd Cunrrelauio,n.c

23. Ennpnrieal Corre!atio,i.u

83 (2.68)

1=
(2.63)

A (A /~T,+ A /~T,~ )p~ + Ano(i + A p~ )(p~/~T,~)exp(_Auufr~) 9 7 8 1~

(~ ~ )E ~
~t~

+ y

where pOr the pseudo reduced density, is defined as,

25

(l

Yin ) F) 5

(2.69)

1 Inc irumlunruml gums cmnmnmpmcssmbnl ny functumn, mis cuilcun luited frmmmnm Figure 2.22, irray be increased by I

~p,=0.27I~P,/(Z CT,))
and the constants are, A = 0.3265 1 A = 0.5475 6 A = -1.0700 2 A = -0.7361 7 A = -0.5339 3 Ag= 0.1844

(2.6.4)

Imnr cunclm 5 nnmnnlc % unilrogcn inn lIre guns 251. ku(rnmip/e 2,12.

A = 0.0 1569 A =-t).05 165 4 5 A = 0.1056 A = 0.6134 A 1=0.7210 9 10 1

Fstmmnimutc lIre coinn~nncssihiIimyfunclor of tIne gas coindein.smute, reported in Table 2.2, at the n eservuni r teimmpcrunium ne mind l=55t) hmmrg, ursi ing mIne genmeral ised clmart. .SO/imtmOn lIre gums s~nec Inc gi mm viny is r initially curl cnn lined to csn i rnnunle its critical lemrrperalure and pressure, misming kml.(2.65) anmd Eq (2.66), respectively.

Eq.(2.63) is valid over tire following ranges: and 0.2~P,<30 0.7<~I,l mnTmd ~ <I

TIre pseundo critical temperature and pressure are often calcmmlmnted by mmrolmmr averungimng of lIne critical properties of tire gas conrponcmnts, Kays muxinrg rule Eq.( 1.4). ()tlmcr nnrmxunmg miles nun calculmute the pseudo properties for estimating time connnpmessnhmlmly funclor mmmc urlso uuvummlurlnle. Sumnnon 166J used datum omn 264 fluids to hack catcuilumte ttreir pseundo crnlic:ml pimmpcrties lunr imnmproving predicted 7 usimrg tIne above equmalioun, aurd proposed to use nIne fohlowning correlations,
=

S,=27.3/28.96=0.9427
l,=(2504.59.6)/432.9t

l,=432.91 R
=1.639

P,=630.l psia
t,=7990/630. 1=

12.68

TIme unlxnve redmmu.ed values resumli inn 1=1.31, unsung Figure 2.21. Substitution of the rculnnced vaknes mm Fq.(2.63) resmnlns inn 7=1.2996. tine estimated value deviates only by I 7, frnnmnn nIne mnncuisurrcd valime mnf I .286(n. The corrcctiunn rnf gas critical properties due to N2 amnd C02 cmnmmlumnt svuus igmiunred in tIme unhove examnple.

756.8 l3l.0S~ 3.6S~ 169.2 + 349.5 S~74.0 S~

(2.65) (2.66)

Ga,c Vi,cca.s ~ti lIne guts vmscosmty gemrerally mnrcreases svithr pressuine. TIre increase of temperature decreases the Inutuid viscosity, wlnereuns it increases tIre guns viscosity at low and nroderate pressures. At high tinessuire. tIme gas vIscosity hehmunvmourr unpprounches thrat of liquid as slnown in Figure 2.22 [681, l.ee et al. 1691 ummeunsumrcd tIre vnscosity of four nalunral gases over a temperature range of 5608(X) ~iR, uip to 80(X) p51mm. mund proposed tIne following correlation,
=

TIre umbove correlations inn St units are,

=

5.218 ~ .

0 90325

0.0248S~

(2.65mm) (2.66ui)

rTc

94.00+194.2 S~ 41.1 S~

4 10 un cxp lb (Pg /62.43)c

(2.70)

wtrerc the pressure anut temperature are in MPa mmmd K. respectively.

svlierc
mu In
=

Even wlren tIne gas conniposition is known, tIre unse of above correlatiouns t(r cslirnrumle Ilme pscmnlun critical properties is meconnmended in preference to ursing airy rmniximng rurlc. It simununlsl be empinasised that the calculated pseudo critical properties from tIre above correlmitiomis shunuilsl be used only in calcunlating the reduced values for estimating 7 from Eq.(2.63). or Figunre 2.22. The above approach results in eslinnation of the conmpressibility factor with a deviation less thunum 2% 137,661. Nanural gases which contain significant quaumtities of sour gases, belnave differently Iliumur timunt stmowmi in Figure 2.22. Wichert and Aziz [67] defined a critical tenrrperunture adjuslnmiemnl fumctor, F, winch is a function of CO and F1 S concenlratiunns in time unixtunme, 2 2 E = 120 (Ymm~s+ Yc,n2 )05 (y~, + Ycom )i 6) + IS (vnums~ 5
tIne pscumdo-critical properties, as, 1
YH~S) is ursed

(9.379+0.0l60M) I /(209.2 ~ I9.26M +T) 3.448+0.OlOO9M +(986.4/T) 2.4 0.2h

rs lIme gas viscosity (cp) urt time unbsolute temperature of T (R), M is the gas molecular 1 weigbml. ansi p~is lIme gas density umt prevailing pressure and temperature, in Ibm/ft .

(2.67) to umdjunst

lime correlumt morn iii Cmmnr en mul. 17t)I ns mmttern usesl to esl iunmuute the natural gas viscosity, particularly ior gumses conrnumnimiing srgni ficumnnt mmnrnounnts of nounhydrocarbon conrpmnents. It initially estimates lIme guns vnscosiny attire mutmnnosplicric pressurme amnd lIre pmevumiling tennperature,
06
=

wlnere y is the commmpomrenmt mniole fractiorm inn tire msnixtuine. lIre correcliunmr factor

11.709 xt()

2.062 xi()

S~ lIT 460) + 8.188 x

l01_6,i5

x l0~logS~

(2.71)

84

2.

IV7 7ent,s nine! (orrehrtrmn,n.s

2.3. Ennrpi real (orrelacimnns

85

For guises containing non-hydrocarbon compounds, the following corrections to the calculated untmnospheric viscosity mnrust be included, Itm
=

Pm, +

x~42+ X(nn~+ ~mm1c

(2.72)

wimere,
~N2
=

y~ x 10 19.59 + 8.48 logS~

~< l0(6.24 + 9.08

(2.73) (2.74) (2.75)

= )c,n~

logS~)

1(1 (3.73 + 8.4~)logS ) 1 umnrsl y is time mnnole Frunctiomr of mron-lmydrocumrtnon couurpomncmmt in tIne gas.
~nins = Yim,s x

lIne calculated viscosity at the atmmrospheric pressure, i l~ is then adjusted for pressure, using the guns pseudo redruced temnmperature and pressumre, over ranges of 1-3, and 1-20, respectively. mis.

lnn[r1,~~]=mum,+urnrP,+uumpP,2+uulrP,~+rTr(un +asrP,+a rPr2+alpP,1) 4 6

0 moO 200 .l,IIliuIIiIlIiIlnI!I~~i 300 400 500

(2.76) 1 , P,)

re.,,Deron,,,e, 459 F

+:l.i (mm + mr, 6

,~, + mm ,

10

P,~4- uu~ r P, )-t.rT,(mui, + an ,~,P,

3 + urn ~~+ mu

wlncre I~mmmnd , P mire tIme pseuidlmr teduccd temnrpcrumlure and pressure, respectively, anti tine vuulunes of tIne coefficients mire un,,=-2.462i l820E-0() ur4 = 2.8t)860949E-00 umg =-7.93385684E-0l uil2= 8.39387 l78E-02 mit = 2.97054714E-00 un~ =-3.49803305E-00 a,j= l.39643306E-00 a~ ~=-I .86408848E-0I a =-2.86264054E-Ol 2 a6 = 3.60373020E-0 I a =-l.49144925E-Ol 10 aI4= 2.03367881 E-02 a~=8.05420522E-03 a~ - 1.0443241 3F-02 a~ ~= 4.4l0l55l2F-03 a =-6.09579263E-04 15

TIre correlation was originally given inn graphical forms by Carr et al. [70), and was converted to Eqs. (2.71-75) by Standing [49J, aund to Eq.(2.761 by Deummpsey. [71).
E.sannple 2.13.
>

Estinmmmte the viscsnsity of tIme gas condensate described in Example 2.12. H++M+l++4++Hw-mi+f+t4t.~.:rd~m1st61.rs*v1 Solnmtimnpn:
Tine gums viscosily cumnn he eslinnuntesl using [ igure 2.22 at 251) F and 7990 2 psia.

tm

linmerpolmulinng between the ctnarrs at S~=O.8ansI S =I .0 for the gas with S~=0.94 7.s~ 1 ohtaimn a guns viscosity nnf (1.0372 cp. Ann mmlt~mmralivemmretlnosl is tine Lee et al correlation, Fq.(2.70). The three pmmrmnmeters are emmlculated mit T=(250m-459.6) R aurd M=27.3, as follows:
11=128.464652 h=5.11353592 c= 1.37729282

moo

Te,,,peronrnR, 4e9 F

200

300

400

500

K=(F+459.67)/l.8
permission.

MPa=0.006895 xpsia

innPa.s=cp

rIme gas dennsity is calcunlated nisinng Z=I .2866,

p=MP/(ZRT)=27. 3x7990/( I .2866x I 0.732X709.6)5r 22.262 lbmnn/ft. Hence, i.i~=0.04347cp

Figure 2.22. Viscosity of natural gases. McGraw-thu Companies Copyriglnm. Reprom.lmiced froinn 1681 wulh

86

2 IV/ I ,.s Is ,s,eI u

,,

n mlati,,i.,

2..t. /:trn,snn ns nil I ,,pn ,latmn,m.s

87

2.3.3 Formation Water

The mutual solubuhities of wunter and Irydrocarhonis are smmnmnll. umuisl inn nnnmnst cmmses nIne hydrocarbon phase belraviour can be studied inndepenrdcmntly of tIre wumler plmumsc. As time temperature increases the volatility of water increases, and mIs contriburtionr 10 lIne gums plnumse inn the reservoir becomes significant. A thermodynamically consistent approaclm ~5 10 lremmt water uns just another component, along with hydrocarbon ansI other non-Irydrocarhon connmpoirenrts, amid to determine the system behaviour. This approach will be discunssed in Seclionr 4.3. unsung equations of state. Contnate water in a petroleum reservoir can be mnssuuned to he in equuilihriummn withn Imyulrocunmhonm phases. When water, which is not in equniiihriunnnn with time reservoir inydruncunrhunmr plnunse encroacires into a reservoir, sunch as in a wunler injection process. tIre dissoliutimn of lmghi hydrocarbons from the oil mb the water will, given enougln tune for dnffumsnotn. reduce tIre oil hurhhlc point. Figuire 2.23 stnows lIre reulunctiomn of lIme bubble poinnt pressure mni mm Nmnrtlm Semi black oil at 293 K, wlmen contacted with wmnter.
.c

16
0.

. Is . 14 3
S

a0.
.0 .0 0

em

S
S

TEuprpA.4u.~9E F S p

12

Fignnre 2.24. Ilydrunle phrase boumurdumry willm natural gas. Rcprodurccml fromnn 1681 svnnlm pcmnnssnsnn.
5

McGraw.HiIl Companies Copyright.

.4

Waler/Oil, v/v Figure 2.23. Varialiomms of oil bubble pount pressurre contumcted witln fieshr wmnter. Under certain conditions of pressure mmnd temmmperuut4nne, wmlter mmmid somnre guises mommy Imrnni smnhiul
crystalline conrpounds known as ciatlnrate guns lmydrmnlcs, mr sinmrply Inydrumtcs. Figure 2.24

Waler Co,mlunt of IInu/roearl,o,m Ilma.se lIre sunltmhiliny of water un liqinid Inydrocmnrhomns utl tlmeir vapour pressures is shown in Figure 2.25 73J. lime solimlnility imncmcmisc,s willm lcnmperature. Fire effect of pressure on liquidliquid
eulimihibmiun is gemmerumbly srnmunll.

I Ins wmntcr vumtnommi cinniteint mnf mimmitmiumt gases inn eujumiIihniinnnnr witir waler is commonly evalunated
I mommm Figumne 2.26, i mrclunding cmrn mscn ionms fnnr lIme nnrurlecumlmmm weight of gas and salinity of waler

shows the hydrate fornmation conditions for natunral guns-waler systenrrs. lIne prcssunre rcquuircul to form solid hydrates increases with teunmperalunre. ilydrumtes are nnot reviewed un this book. Sloan [72] covers tire subject comprehensively. In conventional applications, where the irydrocarhomr-waler nrutural sm)Iuhility is smnnall. simnrplc emrrpirical correlations can be used to estimate the water plnase properties, amnd tIme water content of hydrocarbon phases. Water formed by condensatioim fmounr lire gums plimmse is smnlt free. lime reservoir formation water may contain salt, fronir less Ilnan Ihunt of the numrnruul scum wmnter, no almost being saturated with salt. Althoungim various utrits are enmnpioyed to describe the salt content (25], the weight percent of salt/brine is often used inn the correlahions. lhe presence of salt reduces the mutual soluibility of hydrocarbon-water. As tire solurhility ml Imydrocarhon compounds in water decreases with increase of water salinity, sonne investigmmtors ignore tine dissolved gas, and propose to use brine physical property correlmmtismns for tlne lunmurnuntimum water.

1731

line mnmmnlc frmmclimnmr nil ssumlcr mm tIne guns plnumse cmnmr he estimated by dividing water vapour pressure, mit lime prcvmmilimmg teuimperunlmmme, by time prcvailiung pressuire at low pressure conditions (Runounlls law. sec Section 3.2). TIme vunpotnm pressure nf puire water, from its ircezing point to criticunl poinnt. can he eurlculaled by lIne following relation (74),
=

exp[t. + B/(F

+ 459.6) +

(l,n(T +459.6) + D(T + 459.6))

(2.77)

wlrere T is inn F, P inn psia and lIne cunnstumurts are,

A=69. 103501

131)64.76

C=-7.3037

D= I .2856E-06

E=2

88

2. PVT Tees and C(nrrelatia,m.o

2.3. Ennmpirieal Correlatio,n.m

T.mnn,.ru,.. F
60 nO 20

89

20

no

~o

60

04

,20,a0160.40200

2~O 250

Co,,e,r,on Io,,elale d.n,,r,

0.7 06

800000

~
__________

he, ab.i 0 1450 pa,. F = )C n Ml+32

______________
60000

.,
--.\l-

600 000

071

,,-------~-----

04

:::
00000_~

~
MoI~cnjIarens.

________________

20000

__________________________________________ ,00(,0

j
H cosOrcIor
~OP SOnItiT6

moooo~ ~

I~~ii
I _____TI

C 0 0 Li 0 0

J~
0000

~v

~-

~ S.

s_

S.

~.

_______________

0 0

~~~~______

o ~o~e_______________________________ 0000
____-

~ o

a
0 .4

a,

000

a
6
000

Wnrning~Dashed lines are menastable equiiibrium, Acnuai equiiib,,um sown, wane, oonnenn, Aeg~e 5 a injection of cemposnn,00,

J1~ 1
Z~

S
____a00 0

6 S a

800
U!
600

1=

600

!
0O04~

20

~/ ...
00____

i:: ____________

0.0O3_____.~~/____._

p
0.002
-____________.

~
30 35 40 Adapted no St by GPSA Original Irom Dr. John J. Mcketia University oi Tenas 45 50 55 60 65 70

.
75
trmnmnn

:
-50 -nO

________

Wehe Hyd,on.,60, P~oces.,,9, 0,.qost 1555

0 OOi

no

15

20

25

80 -30 -20 0

Temperaisnre. C

in

,I,.n,n

in,,

20

nO

60

80

so

mrs

ins

Figure~2.25.Water content of liqunid hydrocumrhons. permission.

(WA

Cnnpyrighi. Rcpmnnduceml

1731 wirln

Figuirc 2.26. Water vunpounr conmtcimt of mmaturmul gas in equiilihriuimmm with water. (WA Cmnpyriglim Rcunrmnmtuuccul frommm 1731 wnilr permimnssimsni

90

2. Il/T ic no ann(I (a, rehntjo,n.r

wmnler

2. .1. Finnpm rmcmnl (onrmhmlun,n.u

91

As one lhmnole of gas occupies a volume of 381) SCF, Table 1.1, lIre amoumnnl of ulelermmnined as, mnr~=( 18/380) P~/P lhnmISCF The above correlation mr SI units, P in MPa and F in K is as follows,
=

in guns is A 0
=

4 2 8.15839-6.l2265X1() ~T+l.9I663Xl0 T
1.01021 x 10

1 3 2.l654x10 T

A, A (2.77ur)
Ccc-7.3037 D,=4.i653 x 106

7.44241 x 10 F

3.05553 x It) T

7 2

2,94883 x IOmOT

106 exp(A+B/T+CLnT+DT

7 4 2 3 9 4 10 (9.02505_0.I30237F+8.53425X l0 T 2.34122 x l0~T +2.37049x I0 T ) l000<P<I0,t)00 ansI 100 oF<T<340 OF

A=7 3.649

B-7258.2 1

F2

wincre F is inn F. lIre presence nsf smut

inn wumler rcdinecs

lIne gins solumbility. line correlation of McKetta-Wehe. as

m~=(l8/23.95) P~/P kg/m (se) lire cxtrumpolaliomn of tIne smmhimmity corrcctimmmr Imuelmnr inn Fugmmme 2.27 tin high smmlt cmnmiscmitummlnmnmms is believed to underpredicm the wunter vuupounr commtent of mm gums mm cu1umi Iihritnmnn witin hnriure. tIme graphical correlation of Katz (68) for the sumlinity correction factor is reconmnmemmd insteumd. The graphical correlation, developed from water vapour pressurre depression sluc to smmlt, cain be expmesseul as (25). 5 4 qn =l4.920xl0 w l.7672xl0 w~ 1 5 where w is the weight percent of salt in brine. 5 F.marnple 2.14. F.stimate the waler conlent of tire above gas condensate umi its dew poiirn rn cqnmilmhrmnnmnn with watercontaining 10%, by weight, saIl.
Solution:

pmc.ssnrtcuh by Mc(mmmmr 1251, is ums Imnllinws.

lmng(R,,, / R )

.=

(I.IIHII)SSw,l

(2.80) and w <30% 5

70 F<1 <251) F

(2.78)

The water vapour content is read fromn Figure 226 an 250 F (121 C) mnnmml (n8 ~7 psium, (47.14 MPa) to be 430 lb water per nnilliomr It of wet gas at the stunnrdard connditiomns ((n.89 g/m) prior to any adjumstment for the gas molecular weight ansi the wmmuer sumliniry. Figurre 2.26 shows a correction factor of 0.96 for the gas with molecmnlmnr weighn of 27.3. A 2 78 correction factor of 0.93 is obtained from Eq.( . ) for the salt content of 10%. Iieumce, \V,=430x0.96x0.93384 lb water/ mnnillion ft of wet gums mit the snuummulutrd comnditimuns (6. I 5
g/nn5.

I II

z
0

I0

llvdrocarho,u Solubility in Water The soluhility of hydrocarbon gases in water increases witir pressure umnd decreases witln lemmrperalure to a nminimurur value before increasing, as slmownn in Figure 2.28 funr rnnetlrmmrre (74]. Ilre gas soluhility decreases with increasing carbon number. Time soluhilily ol Imvdnocarbonc in svunter cmun he estimated by applyiurg henrys Imuw for sliluute solumniuins. as described inn Section
3.2. Figurre 2.27 is often used to estimate the soluhilrty of nuntunral guns inn water. represented within 5% mmccuracy (25) by, R~= A
+

it
5-

-i

lIne clmmurt cmsnr be

2 AP+A P (279) 2 3 wlrere RW is time ft gas (sc) shissolved inn a bunrrel of water unt pressure P inn p.siun. lire coefficients depend on nine temperalunre, mms. 0

TEMPERATURE, m K=(F+45 ).671/1.8

MPutO.006895 x psia

Figure 2.28. Solunbilily of tnmethunnme inn water. SPE Cmpyri8lmm. Reproduced from 1751 wimh permission.

92 Water Formation Volume Factor

2. PVT Te.ut.s and Correlaiio,is

2 .4 Empirical Correhrtio,ms

93

Water 1)e,m.sitv lhe fonumalion water density at the stunmrdard conditions can be estinnated fronn [25],
ow 62.368 + 0.438603w,
+

The volume of water at reservoir conditions when brought to tire surface gemmerumily decreases. unless from highly undersaturated reservoirs, due to the combined effect of liberated guns. thermal compaction, and pressure expansion. McCain 125] proposed the foilowiirg correlmmtiunur, B =(1+,~V~~)(l+AV_T) 0 (2.81)

1.60074 x l0~w~

(2.87)

where ~wP and AwT are the volume changes due to pressure and temperature, respectively,

Neglectiung tine nmunss of dissolved gas in water at reservoir conditions, the water density cuun be calculated, as,
i~ =p 60001 /B,,.

as follows:
= (3.58922 x I0~ -I. 1,95301 x 10 T)P

(2.88)

(2.25341 x

Ion

wimeme B
+

1.72834 x 10 nnl.)pr (2.82)

60

is the forrmiation volume factor mt the prevailing conditions,

Water Viscosity
ow 1.0001 x l02
+

7 2 1.33391 x l0~T+ 5.50654 x 10 T

(2.83)

lIne viscosity of brine (ep) at the atmnosplreric pressure cmnnm be estimmnted fronr 125], I~~ = (109.574 8.40564w~ + 0.3 133 l4w~ + 8.72213 x l0~w)T i000F<fcz400 wlmere, m1 F anmd w <26% 5 (2.89)

The correlation is vmulid at T < 260 F and P < 50(X) psimu. over a wide ummnge of omit concentration, as presunnahly. the effect of sumlt on tlrcnnmmnl expumnrsiomm nsf wumtcr no cuunnceIheul by its effect on the gas solubility in water (37].
Conmpressihility of Water

2 5 1)=I.l2l662.6395lxl0 ws+6.7946lxI0~w~+5,47hi9xlQ wl.55586xl0~w~ 601 )cmumr be cuulctnlmutcul fromnm 1761.

lIne effect of presum~rru~ 0mm tIme binmre viscosmty is estimated, as,

The isothermal compressibility coefficient of gas free water (C

2 l0~(C +CmT+C T ) (2.84) 0 2 1 where Cwf is in psi- , at temperature T, F,and the coefficients depend on the pressure. as,

ic

/~t,~ 1

ow

0.9994 + 4.0295 x 10P+ 3.1062 x 10P 86F<T<l67F and

(2.90)

14,000 psia <P

C =3.85460.000134P 0 7 C =0.010524,77xl0 P C =3.9267xlOS_8.8x10OP 2 where P is in psia. The dissolution ofgas in water increases its compressibility, as,
3 C~=C~i(l+8.9Xl0 R ) 0 (2.85)

E.ranmple 2.1$. Tine gas condensate, reported imr Tunble 2.2, is urt esluilibriummn with the reservoir waler. lIre water sumlinnity, WS, is 10%. Fstinnunte: (a) the anrotnnt of dissolved gas in water, and (h) mhe ismnmtnernrral commipressibility cuneffieieint, (c) density, mmmnui (d) viscosity of wumler an the gas In ydrmncunrbunii (Ic w inmmm inn cmnumd 1 nomns. ,nu,Irutionn
( ui

Assnmmmning the imunturral gas solmihiliiy in water is apprsnximrrately the same as that of methane,
Figunre 2.27 slrows lhe nmnumie frmnctiunnr of gas in water equmal to 4.1 xl 0. The uiissolved gas damn be coirvemned mr let iris of SCF/hunrrct as

where R~isgas water ratio in SCFIbbI. The following multiplying factor to correct the isothnerinal conmmpressibility coefficient due to salt has been proposed [77], 6 2 9 5 = I +(0.052 + 2.7 x l0~T1.14 x 10 T + 1.121 x lO T )w, where T is in F,and w, is the weight percent of salt un brine. (mnnole gas)/(imrunle waner)=0.004 I /( 10.0041 )=0,004 12 1 R_ =(38() SCI gums/murole gas)x(nrmnle wmmter/I 8 lb water)x(62.4 lb water/ft waier)x(S.61 in/hhl)x(t).004 12 mmnole gas I mmmoie water)=3t).4 SCF/tthh (5.41 nn/mn) The dissolved guns cumin attemnunlively he calculated fromnr Eq.(2.79) Au 250 F the vmutues of tIne cmreflicicnmts mmre as follows: A,,=l .447265 A,=0.0(.)598559 A,=-2.483E-07

(2.86)

94

2. FVi ic.sls minim! (,n,rc!atiorn,c

2.4. Rnfm,e,icr.c

95

resulting in R,,=30.76 SCF/bbl at 6837 psia.

The reduction of sotubihity due to the salt content is estinrated umsinmg F.q.(2.i40). RjR~=0,6707

2.4 REFERENCES
Remideilnurber, F.0: Sepurrator Sampling of Gas Conslensate Reservoirs, Oil and Gas J., 138-140 (June, 1954). 2. Reundelhumber, P.0: Better Scpunralor Smnmpling of Crude Oil Reservoirs, Oil and Gas J., 181-183 (Nov., 1954). .1. Rcnmslcllmulrer, 1:0: Suimmmplinrg lrmmcedumrcs for Oil Reservoir Fluids, JPT, 15-18 (Dec., 1957). 4. Aimnerismmmm Pen mmnlemnmmm lmnsl tunIc: AII Recmnnnminemnrled Practice for Sampling Petroleumurr Reservoir Fluniuls, APt 44 (1966). 5. I learmr. R.S. (cmnmnrpiied by): Smnnnrplnng mmmd Analysing Gas/Condensate Reservoir I1unids. LJKOOA Report (Marcim, 1986). 6. McCusinn Jr., WI). usnnsl Alexannder. R.A: Sammrpliirg Gas Conmdensate Wells, SPE Res. Enrg., 358-362 (Aung., 1992). 7. lurnrcr, R.G., llunhbuurd, MG. unnsl Dukter, A,E: Anmalysis and Prediction of Mininnuinn Flow Rumle fm)r tire (ommtinuoums Rcnmrovumi of Liquids fr(nmmr Gas Wells, JPT, 14751482 (Nov., 1960). 8. Flail7.. J.M. umnrst Pumrks, A.S: Sunmrmphing Gums Conmdensate Wells, Trans. AIME Proc., 46. 232-245 (l9tl). 9. Kikumnni. J. uunmul Rmnlulowski, J: Cotrsistcnrcy Cimeck and Reconciliation of PVT Data frommn Sumnmrples Obtuninred wihir Fornrmunhion Iesters tisinng EOS Models, SPE 36743, Proc. of SPE Ann. Coirf., 6 15-622 (Oct., 1996). 10. l.eyllmmneunser, I). unmnd Runcklucrmrm, J: heterogeneity of Oil Conrposition within a Reservoir mIs mm Rciicctuunnce of Accummnunlatnon ilistony, Geochemrmica et Cosmochimica Ada, 53, 21192i23 (1989).

Hence, R,~=20.6 SCF/bbl (3.67 nnIm) (b) The isothermal compressibility coefliciemmt of the gums tree w;mler i.s csnmnmm~mnmnt In mmnn Eq.(2.84). At 6837 psia, ihe coefficients are cuutcrilmilcd as follows: C~ow2.9384 Cow.0.007259 6 C =3.325x 10 7

resulting in C~,=3.20xt0 psi at 250 F. Applying Fq.(2.85), mine cmulculunted compressibility coefficient is adjusted for R =20.6SCF/bbl, resrrllinmg inn (~=3.79x 1(10 5~ psi. The reduction of comrrpressihitity dune to salt is eslimnateci from tiq.(2.86), W,0.6 I76 iiemnce. 6 C_,=2.34x10 psi(3.39 xlO MPa) (c) The wmmter densim y am ttre stanrdumrd coinditionns. Fq.(2 87) equal 10 66.91 lbnn/fi.
inc lund

i ng I 0% sumlt, is cmnlcintmmied mm si imp

rIme water fonnnaiion viniume factor, at 250 F unnul 6837 psiui, is cumtcnitumiesl mnsing I3q.(2.8i), with the volumisne changes due to pressure and tenmiperuilnnre ums follows,
AV,,,=-0.0I83454 AV,,,=0.05776263 B~1.0383

I lence, p=66.9 1/1.0383=64.44 Ihmmm/ft (1(132 kg/inn) (d) The viscosity of formnmation water at the atmospheric pressunre. tempermntuire of 250 F. munot l07c sam is calculunned using Eq.(2.89), D=0.9648084 p~~=0.3 1854 cp The effect of pressure on viscosity 6837 psia is estimated. usinng Eq.(2.90) 1.42
Hence,

II. Ilorstunsi, 1., I,umnlcr, SR., t)y~mvik,II.. Aagunarsl, P., Bjornvik, AM,, Johamrsen, P.E. mmmd Fm mksenr. 5: I)egnuuduution umnn,I Munlunrily Controls on Oil Field Petroleum Column I lelesogemrcity un 11n ( ;~ Ifmuks I 4elnl. Nmnrwcgimmmn Nortir Sea, Orgumnic Geocimcnnnistry. 1 6, I 3. 49/-5 It) (1990). I 2. Keimmrcdy, I it. ummnd Olsinmr, R: tJtrhlrle Foriunmtliomn in Supersmnturated Hydrocarbon Mixlunnes, lruunns. AIMF, 195, 27278 (1952). 13. Wmelansl, i).R. annul Kennnedy, lIT: Measurement of Bubble Frequency in Cores,
Trunmns. A1MP, 210. 122-125 (1957).

4. linmmmniurhuuii. A., Ottescmn, B. mmmd Mikkelsen, M: Mea.surenrsent of Supersatuiration and Crntncurl Guns Suutumruulionm, SPE tornniuutioim Evaluation, 337344, (Dcc., 1992), IS. Ycin, G.C., Sinalr, MS. and Yeh, B,V: Vapoumr-Liquid Equilibria of Non Electrolyte Soluntions in Snmrusil Capillaries. I. Experimsmenntal Determination of Equilibrium Compositions. ACS. 2,90(1986).
16. Yein. G.C. rind Vein. h4.V: Vapounr-Liquid Equmilibria of Non Electrolyte Solutions in Smrnall Capillaries. 2. Theoretical Calculations of Equilibrium Comnnpositions, 60th Colloid and Surrface Science Symnposiunm, ACS. (Junne, 1986).

p.=0.4523 cp (mPa.s)

96

2. IVT Tm.sic mmml Cm,rrnkmmjmn,n,s

2.4. heferm,iee.s

97

17. Sigmund, P.M., Dranchuk, P.M., Morrow, N.R, mind hunrvis, RA: Retiogimude Condensation in Porous Media, SPEJ, 93-104 (April, 1973). 18. Kortekaas, T.F.M. and van Poelgeest, F: Libeiuutiomm of Soluniiomr Guns l)unmimmg Pressuire Depletion of Virgin and Watered-out Oil Reservoirs, SPE Trans. 291, 329-335 (Aug., 1991).

34. Cinierici, CL., Giucci, G.M. mmmd Sclocclmi. M: Iwo-Phase Vertical Flow in Oil WellPrcdictiomn of Pressuire 1)rop, JPT. 927-938 (Aug.. 1974). 35. Ostcrmmmainn, RI)., Flniig-Econmnmnni(les, CA. and Owolahi, 0.0: Correlations for the Rescrvmir Fltmid Prmmpertics of Alaskunrr Crudes, SPE 11703 presented at the 1983 California Regionmul Meeting (Mmnrclm, 1983). 36. Sutton, R.P. mmcl Farslnad, F.F: Evaluation of Empirically Derived PVT Properties for

19. Moulu, iC. and Longeron, D: Solution Gas Drive, Experiments and Simnmmnlalion, Pnoc.
of 5th Europ. IOR Symp., Budapest, 145-154 (April, 1989). 20. Jacoby, R.H. and Tracht, J.H: Fluid sampling: a better tecimnique, Hydrocarbon Processing, 101-102 (Feb., 1970). 21. Jacohy, RI!. and Traclmt, J.il: Collection of Sammrples unnuler Pressumre for Chromatographic Analysis and a Systenn for Ilamrdlimrg Gas Comndemrsumlc lype Fltmids, J. oF Chromatog. Science, 13 (I), 44 (1975). 22. Varotsis, N., Stewart, G., Todd, AC. mnnnd (imuncy, M: limmuse h3clrumviomir oi Systemnns Comprising North Sea Reservoir Flumids amrd Injection Gmuses, JPh, 38 (II), 1221-1233

Gulf of Mexico Crude Oils, SPE 13172, Proc. of 59th Ann. Teem. Conf, (1984),
37. McCain Jr., W.D: Reservoir Fluid Property Correlations, SPE Res. Eng., 266-272 (May. 1991). 38, l)c Ghetto, Ci., Pmuomne. F. mmmd Villa, M: Reliability Analysis on PVT Correlations, SPE 28904, Proc. of Eumrmn. let, Coin., L.ondon, 375-393 (Oct., 1994). 39. Stunnmchiimg, MB: A hmessunre-Volumnnme-lemnmperumtuire Correlation (muliformnian Oils and Gmuscs, t)milI, mntnd Prod. Proc., 27.5-287 (1947). 40. for Mixtures of

(1986).
23. Legret, D., Riclnon, D. and Renon, hi: Vumpour Liquid Equmilibriun tip to lOt) Mhuu: A New Apparatus, AIChE, 27(2), 203-207 (1981). 24. Danesh, A. and Todd , AC: A Novel Sampling Metirod for Compositional Anunlysis of High Pressure Fluids, J. Fluid Phase Equilibria, 57, 161-171 (1990). 25. McCain Jr., W.D: The Properties of Petroleummi Fluids, 2usd Ed., Pcmmnwell Books. Tulsa (1990). 26. Moses, P.L: Engineering Applications of Phase Behaviour of Crude Oil mmd Condensunte Systems, JPT, 7 15-723 (July, 1986). 27. Dodson, CR., Goodwill, D. and Mayer, E,H: Application of Laboratory PVT Data to Reservoir Engineering Problems, JPT, 287-298 (L)ec., 1953). 28. Robinson, D,B: The Interface Between Theory and Expcritnretmt, J. Fluid Phrase Equilibria, 52, 1-14 (1989). 29. Reservoir Fluid Studies, 1993-1996 Final Report, Report No: PVT/97/I. Dept. of Pet. Engng., Heriot-Watt University (Jan., 1997). 30. Bashbush, J.L: A Method to Deternmmine K Values frotnm 1.umhormitony Dmmta mmtnd Its Applications, SPE 10127, Proc. of 56th Ann. Conf. (Oct., 1981),

Luisunler, iA: Bubble Point Pnes.suire Correlmmtioni, Trans. AIME, 213, 379-381(1958).

4 I . Vunsquiez, M. mind l3cggs, 11.1): Correlmntknns for Flunid Flrysicmml Prmrperty Prediction, JPl, 968-970 (June, 1980). 42. Glaso, 0: Gemreralised Pressure-Volunne-Temniperature Correlations, JPT, 785-795 (Mmuy, 1980). 43. Muirhoun, M.A: PVT Correhurtion for Middle East Crude Oils, JPT, 650-665 (May. 1988). 44, Almnred, F: h-Iydrocarhomn Phase Behaviour, Gumlf Publishing Co., 1st Ed. (1990).

45. Asglnunrpoumr, S.. McLauchmiinr. L., Womrg, I). and Cheummmg, V: Pressure-VolunrreTemnipermiture Correlations for Western Canadian Gases and Oils, Pet. Soc. of CIM, Paper No: 88-39-62 (1988). 46. Lumbedi, R: Use of Production Data to Estinnate Volummue Factor, Density Conrpressihility of Reservoir Fluids, J. of Pet. Sci. and Eng., 4, 375-90 (1990). 47.
of Mexico Crumde Oils, SPE 26644, Proc. of 68tir Ann. Conf., 395-406 (1993).

and

Petrosky, G.E. uund Farsirad, F.F: Pressure-Volunne-Temperature Correlalions for Gulf

48. Rollins, J.B., McCumin Jr. W.D. and Creeger, ii: Estinnation of Solution Gas-Oil Ratio of Black Oils, JPl, 42. 92-94 (Jmsmn., 1990). 49. Slunmrdimng, MB: Volunnrrelric mmmuul Phrase Behavimmtnr of Oil Field ilydrocunrhon Systenns. 9mln Irimrtinrg, SPE, l)utlluns, iexums (1981). 50. Prick, r.C: Petmoheumnr Prosluiction llanndbook, Vcrlume It, SPE, Dallas (1962),

31. Drohm, J.K., Trengove, RD. and

Standard Gas-Condensate PVT Symposium, Dallas, Texas (1988).

Goidlhorpe, Wi!: 0mm tIne Qurmility of Dmsla fronmn

Experiments.

SPE

17768,

SPE

Guns Icchmmrology

32. Danesh, A., Todd, AC., Somerville, J. and Dandekar, A: Direct Measurement of Interfacial Tension, Density, Volume and Compositions of Gas-Condensate System, Trans. AIChE, 68, 325-330 (1990). 33. Hoffmann A.E., Crunmp IS. and I-locott CR: Equilibriunmm Coinstamnts for GasCondensate Systems, Trans. AIME, 198, 1-10 (1953).

SI. Katz, DL: Irediction of tire Slrrmnkage of Cnnde Oils, Drill, and Prod. Prune., API, 137-147 (1942).
52. Snunnnulimrg. MB. mmd Kmrtz, D.L: Deirsity of Crude Oils Saturated with Natural Gas, AIME Trains,, 159-165 (1941).

98

2. FVT Tm., n.m mn:nd (

,,,

re/a(n,,nn 5

2.4. Refnrennces

99

53. Pedersen, KS., Tlrommnssen, P. atnd Frcsicnmsl timid, A.A: lImcrmmnmndyni:mnn miss mnf 1 n.troleumnnm Mixtures Containing Ilyclrocarbons. 2. Flash and PVT Calculatiomrs witin tire SRK Equmumtiomr of State, md. Eng. Clrenmr. Proc. Des. and Dcv., 23, 566-573 (1984), 54. Annerican Petroleum Institute, Technical Data Book, Petrolcurnm Rcfinninmg, API, New York (1982). 55. Alumni, 11G. mind Kennedy, LIT: Volumes of Liquid Temperatures and Pressures, JPT. 272-273 (Nov., 1960).
ilysirmncummhmnnns mmt I ligln

71 . t)cnnrpscy, J.R: (omnmpuitcr Runmutine Trcmmts Guns Viscosity as a Variable, Oil and Gas I. 141 (Aug.. 1965).

72. 73. 74.

Slommnm. FE): Cluntlrrmnte hiyulrmmlcs of Natturunl Gases, Marcel h)ekker Inc (1990).

Gas Processors Suppliers Associmmtion: SI Engineering Data Book,(1980). Dmnnhcrl, i.E. mmcl I)mmnncr, RI: DIPPR Dumta Compilnmtionr Tables of Properties of Pure

56. Loiirenz, J., Bmmry, B.C. ansi Clark, CR: Calcunlmmninmg Viscosities of Reservoir Fluids from nlmeir Compo.sitionrs, JPT, 1171-1176 (Oct.. 1964). 57. Beal, C: Time Viscosity of Air, Water, Naturmml Gums, Crude Oil mmmd nts Associmnted Gmmscs at Oil Field Temmmpcralunres amid Pressures, Trmnns. AIME, 94, 115 (1946).
Beggs. hI.!). mmd Rohinsomm, J.R: Estimnrurtinmg tIme Viscmnsity of Crude Oil Syslemnrs, JhI, 27, 1140-1141 (1975). 58.

Conmnpounsls, AICIrE, New York, (1985). 75. (unlherson. 0.L. mmmd McKcttum Jr., ii: Phumse Equilibria in 1-lydrocarbonr-Water Systenrrs Ill - Solmmlmility of Menhnummre mr Wmntcn mit Imcssuircs to I O,(XX) psimm. Trans. AIME, 192. 223226 (1951).
7(n.

Mccliumnr, I).N: A (mnrielmntiomn for Wmnler (omnprcssibilnty, Pctroleunn Engineer, 125-126 Nmnv., 1981))

77. Nummnmlnere. I).. lIrigirumnnn. \W. mmmn(I Snuunrulinrg, Ml): Correlmstions for Phmysicmnl Properties of hchroleuimi Rcservunir Br immcs, PR! Report, Stumirlmnrd thmniversity (Nov., 1977).

59. Egboghah, E.O. and Ng, iT: An Improved Temperature Viscosity (unrmelmmtiomn for Crude Oil Systems, J. Pet. Sci. and Eng.. 5, 197-200 (1990).
2.5 6t). 1 muhcsli, R: I improved Correiumtions for Predictinng tIre V iscunsily unf I .iglmt (mimIcs

EXERCISES

I. Id r. 2. I. ftc PVF lest mesunlts (slum reservoir (nil mire givemn inn (Inc following tables. Calculate the gas inn solunlruin. (nil fcnrnnnmnt imnmm volummnie fmmctunr, mmmd totmnl formrnation vohunme factor at 2277 psig. with
tIne irrtcrnnncuiiumlc sepmmimmlmr opermmtnnng at tIne optimnnsmmrr pressttre. Reservoir temnnpermmlsnrc = 195 F.

Sci. and Engng., 8,221-234 (1992).

Golsl, D.K., McCunin Jr., WI). umnrd Jennnimrgs, J.nV: An Imnnprovcsl Method fnrr tine 1 t)eternnnination of the Reservoir Guns Specific Grusvity for Rclrogrmnde (immscs, Jh T, 41, 747-752 (Jurly, 1989). (nI .

Inuitimml Resemvmnir lncssmnre

5712

isIg

62. Standing, MB. and Katz, DL: Density of Natural Gases, AIME Trumns., 146, 140-49 (1942). 63. Takacs. G: Conrrparisons Made for Computer Z-Faclor Cmtlcimiationms, Oil mind Guns I., 64-66 (Dcc., 1976). Dranclmuk, P.M. mnnd Abou-Kassemni, J.l-1: Cunlculumtion of Z-Functors for Nmrttnrmnl Gutses using Equmation of Slmute, JCPT, 34-36 (Jumly-Sep., 1975).
64.

Reservoir fluid connmposntnon. (onnmpmnnntmnis -. MnlX Nnmnmngcn 1).)))

(mmmt,mm,, dimnxmmlc
Mcnlimnmnc

I 19
51.54

t~iImmnmic tnm mmmnnc 1

6.57 4 1(3
(I (,8

65. Starling, K.E: Fhnid Thermmrodynunmrnic Prmnpcrl es finn I .igirl hetm mnlernnrm Systcmmns, ( imnl I iuh. (1973).
66. Sutton, R.P: Compressibility Factors for hligln Molecurlar Weigint Rescrvmnir Gases,

mmn,nmnn~ mm- timmm.mm,e tmnnl,minc

SPE 14265, Proc. of 60th Ann. Tech. Conf, (Sept.. 1985). 67. Wnchert, E. and Azr7., K: Calculate Zs for Sounr Gases, Ilydmocambon Processing, 119122 (Mumy, 1972).
68. Katz, D., et al: Handbook of Natural Gas Engimreering, McGrumw iIill (1959).

mm lcn,immn, I tcxunnnes I t,pnunmmes ()mn.nnncs N,mnnunmnts I )ecmnmies I nnmlccmmmnes l),~tecmmnncs+

(~,,

2 (n 0 0 n 17 2 Ii 41. (II 3.96 I .95 t (6

.1)5

3.82

69. Lee, A. and Gonzalez, Ml!. and Eakin, BE: The Viscosity of Naturmnl Gases, JPT, 997-1000 (Aug., 1966). 70. Carr, N.L., Kobayashi, R. and Burrows, D,B: Viscosity of l-lydrocarhon Gases Under Pressure, Trans. AIME, 201 (1954).

(.Tlnarunclcrmsi cs:

M=265

S=O.883

Prcssure-volunnrc test mrt 195 F. Ircssunre Relanive i)ensiiy

_,~psig vnlun~c_~j~/m,~

57 I 2

0.98)))

645

100 5427 5137 4850 4566 4544 4492 4382 4281 4100
31169

2. PVT le.nsm tiiiml (.orrelm.mti,m,ms

2.5. E.nerem,re.c

101

3618 3314 2940 2641 2331 1924

0.9849 0.9897 0.9947 1.0000 1.0011 1.0038 1.0099 1.0159 1.0278 1.0456 1.5)694 1.1055 1.1660 1.2328 1.3242 1.5077

642 639 635 632

iwo slmuge sepuirator test mnn 84 F. Ircmsmjrc 1st Sep. (014 i,,tal Gold Fmrummatjon Smock lank p,~rg SLI IS II) ~_S~i~t B Vmllmctor Oil SCi 6(X) 15)76 1358 1.678 (1.835 40(1 1147 1342 I .645 0.831 2(X) 1245 1361) 1.659 0.833 1474 474 1.759 (1.840 * Onne smungc separanimn (smock rank onuly). Conniposition of separator gas (nnole%).
Iresourc. psig
Nimn,,gtin (mmr(nmmmn ,Immmsj,k

(mill)
I (,(, 2.27 84.1) 7.21
I

1(10
1.58 2.311 82.Sti 8,07 1,7) (1.31 (1.8,5 (1.15

200
1.12 2.36 (OmIt 15 82 4 65 ((5) 1.11 1)22

(m
1.24 2 24 72.2)5 ).2)( 6,9(1 ((.86

Mmitmam,c
Ir~ mmmiii, lt,mnm,m,e millumimmnc

II

0.26
(1.67 (1.13

Differential vaporisation at 195 F. Pressure Oil Vol. Sol. Gas Compres. pug factor SCFIbbh Factor 5712 1.798 5427 1.807 5137 1.815 4850 1.825 4566 L834 1541 Ph 3983 1.695 1261 0.1189 3556 1.614 1092 0.865 3136 1.542 939 0.845 2707 1.480 806 , 0.839 2277 1.422 680 (1.843 1849 1.370 564 0.852 1415 1.320 451 0.873 986 1.271 344 0.897 566 1.222 238 0.931 214 1.166 135 0.960 0 1.059 0 1.000 Residual Oil SO = 0.844 (36.0 APt)

2.9.5
(1.83

iPcmnmumm,e
nPcnuummnc

t). 8

llexumnes
Hepnmmncs +

0.13
0. II

0.20 0.17
0. II)

(1.26

MW SO

19.52 0.674

19.95 0.689

0.21 0. 14 21)39
((.704

I .2 I 1.06

I . 17 24.57 0.848

Reservoir oil viscosity at 195 F. Iresstnre, psig VmsLumsmny. cp

640(1 (,045 11.605 ((.58)) i).56))

5689 5334
4978

0.540 0.52))
(1.500 0.5(15 0.520 (1.565 0.6 51) 1)725

4566 4267
3556

Composition of liberated gas (mole%) in differemmtial liberation test. 566 2t4 (1 Pres. psig 3983 355 3136 2707 2277 1849 1415 986 N2 2.00 1.97 1.89 1.75 1.59 1.32 15)2 0.75 0.44 0.17 1)11(1 C02 1.99 2.00 2.05 2.10 2.18 2.25 2.39 2.59 2.97 2.97 1.84 Methane 77.72 79.37 80.46 81.29 81.58 81.58 80.72 78.31 72.58 55,72 6.68 Ethane 6.95 6.90 6.93 6.97 7.20 7.48 8.17 9.38 12.14 18.15 19.36 Propane 4,25 4,10 4.00 3.89 3.82 3.94 4.20 4.99 6.74 12.73 28.40 0.46 0.42 0.40 0.4t) 0.42 0.49 (1.67 1.29 4.21 i-Butane 0.60 0.52 n-butane 2.03 1.70 1.48 1.32 1.25 1.21 1.27 1.48 1.94 3.98 14.57 0.44 0.38 0.34 0.32 0.32 11.37 0.47 0.95 3,88 1-Pentane 0.67 0.54 n-Pentane 0.93 0.76 0,64 1)55 0.50 (1.47 t).48 0.53 1)66 1.33 5.14 Hexanes 0,97 0.78 0.64 0.54 0.49 11.46 0.46 0.51 (1.63 1.23 3.51 1.01 0.79 0.65 0.57 0.55 0.60 0.76 1.23 2.41 Heptaimes + 1.89 1.36 MW 23.39 22.38 21.71 21.24 21.01 20.93 21.13 21.79 23.36 28.52 44.77 SG 0.807 0.773 0.749 0.733 1)725 0.723 0.729 0.752 0.806 0.984 1.545

2845
2 133 1422

711
284
(1
*

0.895
1.085

t.4i2

lliihttc pmimii

2.2. Cmilculunte tire lrqtrid propamme content of time prcndumced guts in Exercise 2.1, with tIne optimmrummnm separator pressure. 2.3. lIre laboratory sluntmu is often evahumuted and smnoothmed by tine siiniemmsionhess funcmion Y, definned in Eq.(2. 10). A plot of Y funnction versus pressuire is expected to yielul eitlner a straight lure or very slightly curved. What is the implicit assummption in tine above approach? 2.4. lIme results of consturnml volumimne depletion test otn mm gas condeimsate sannple, as reported by mn lumborunmory. mire mrs follows:

102 Conunposition of reservoir gas. Comnmpnnnenn MoleS

Niirmgen Uartxmn dnmi,imic Methane Enhanc Propane humane
in-Buuane

2. PV1 lest., mend Cm,rre(mriio,m.c

2.5. Exercise,, I) us mmmc mm- Immune n I e I(( muincs mm Iemmnamme I Ics,mm,cs I lmpnummmms (cimumnes N,mmmmmn,s I ),cmnmm,s nmdccmmnc. plums
~

103
(1.43 I-I I (1.17 0.49 067 .2)) 0.8.5 ((.5)) ((. 2) I .35
________

Density. kg/nm

Mol. Wciglnm

n-Pcnmanes nPeniane Ilesanes tlcptanes

0.386 2.524 78.585 7 029 3.360 0.440 1.114 0.396 0.502 0.661 1.277

0.42 I .07 (1.34 0.47 0.5) 1.11 0.74 (1.40 ((.20 ((.90 24 I
-

0.41 I .04 0.33 ((.45

(1.54 0.96 (1.62 0.3(1 0. 14 0.47 22.8 178

0.40 I .00 0.31 0.38 0.46

(1.80 0.48 0.20 0. 10 0.30 22.0 176

0.40 0.96 0.31 0.39 0.42 0.66

0.38 0.12 0.5)6 0.16 21.4 169

M,,tcculmnr wcmgtnn Mmnlcculzmr wcigtmm

25 5 97

686.4 744.5

Octanes
Nonanes Decanes Ijmmdecancs Duxiccanes Tridccancs

0.902
0.588 0.337

761.6
775.9

86 89 102 118

Comimpositnnnn unf mciuiaininrglluidsumn7ihumr. ~ (mmmnmpmmnnenns m~rlrmmcmmmns Renmmmmnmnnnng gums Remmmaimnisg miiI

____-

783.6
789.4

136
147

~mpm~) 0)8) (1.40 2.51 82.56 7.42 3 315

(I .i 2

(~LS~ 0.00 (1.05 0.90 15.31 3.79 4.39

((.82

0.345 0.256 0.245

0.186 (1.163 0.120 0.098

798.0 814.2
83(1.1 811.6 8348 83.5 2 8360

162 176
190 204 21)5 2.52 246

Tetradccanes Pennaulccancs
Itcsmxiemmmncs lleptadecanes Ocnmxhanes

Nunnadecamnes
Eicosanes pIus

0.091 0.065
(1.330

I Iyulrmmgcmm Smnltidc Nimrmgcn (mmmtMmn Diunxidc Meilmane Pitnumnc Irmpmmnm humane mm ll,mmumnc i Ikinn;nmnec
n tenmnmnc

I ((.1 (1.12

846.1
870.4

260
335

((.39
((.41 (1.5) ((.28 ((II ((.06 (1.12 2 (.2

C,,, Propernies: Molecular Weigtnu=335

Density = 870.4 kg/inn

Constant volume depletion test results at 373. 1 K (212 F).

Pressure
__________

______

________

Curnutmmm,ve Gas.

_~,__t~_.~ ___P691
(I) 389.4

(2) 373.0 32m.0

271,0

221.0 171.0
(I)

5632 5394 4640 3915 3191)

Productnomn J m0lS~ 9 1)00 2.33

-

S~c Inc Gravmiy Compres.smlmmlmmy Pmmct,mr Volumnre of of Produced Gas of Produced g~s Rcii~gradeLiquid 0.923 1.037
-

}lcsmnmnes I teplamnes Ocn,inus Nmnammes I)cLmmmnes Unmlec,nmnes- ,Ius 1 Molcumnlar weiglrl C,,,, M,nleculuur weiglmi

67

2.90 1.57 2.2(1 4.1)4 1(1.39 9.23 7,90 4.84 31.67 It (.9 204

((.00

(1.923
0.879 (1.831 0.787 0.758 0.737 ((.732

1.017
0.956 0912 (187.5 (1.873 (1.880 0.914

000
3.58 6.52 8.61 909 8.41)

(mm) (alcumlurte tIne lwtr-plimmsc comsmpicssihility fuuctor of the cell content, and plot P/Z vs. the Iotuul volumnre of produiccsl gums.
)b) Dctcnnmnmre (he denrsity mmmd Commipm.nSition of condensate phase at Pr 71 bar by material

9.91 19.49 31.58

I)mm Immmmce cmulcun lat iomns. 2.5. lIne results oi a constmunt volummme depletion test Imave been umsed to calculate the equilibrium
nmutios by trnanermmml halanmce equnanromrs. The K vs. P plot shows that the equilibrium ratio curves of C3 ansi rC4 crmmss at mm pressure vmmlue. Is Ilmis physically possible?

2465

45.63
6I.23 76.88

121.0 1740 71.0 lOIS lnniial reservoir pressure.

(2) Sanuration pressure an indicated menmperammirc (mlew poinu). Composition of produced gas. Pressure bar 321.0 psig 4641)
tlydrusgen Sulfide Ninrogen Carbon t)ioside Memtnane Emlnunne Pr~~pane 0.00 (1.39 2.44 79.58 7.05 3 35

2.6. An umndersmntumraned oil ns produced llnrouglr one intermediate stage separator at 200 psia mumruh I 20 F, witlr mu GOR = 600 SCF/STB. The stsck tank oil API=40, and S =0.7. The 5 reservoir pressunre mmmrd tcmsmperalure are 65(X) psia and 210 F,respectively.
271,1)

39(5
0.0(1 (1.39 2.48 81)57 7.5)2 3.30

22 n 0 319(1 0.00 (1.40 2,49 81.53 7 04 3.28

171.5) 2465 0.00 (14(1 2.5)) 82.34 7.12 3.21

121 0 I 740 ((.0(1 ((.40 2 5)1 82.78

(a) (b)

Estnnnnur(e tIne out bubble poinnt pressure at reservoir temperature. Cuulcurlmnte fire slumluc oil pressure grmudiemrt inn Ilne reservoir.

7.23

(c) Wirurl will be tIme oil mind tunlal formnration volume fumctor whets the reservoir pressure falls 4(X) psi bekrw tIre bubble poinrt.

104

2. IVl 7..,r., moim! ( mm,rtImili,,n.c

lOS

2.7. Estimate the density of oil reported in Exercise 2.1 at its btmbhle point. tnsitng hIre KatzStanding and Alani-Kennedy methods. Compare the results witlm the turemisuired vmuitie. 2.8. Estimate the viscosity of the above oil, and conmmpumre it witlr mIme mnmemrstmred v.mlunc. 2.9. A rich gas is produced and stahilised in tlmree sepmmration stages. Iime first stmmge sepmmrumlor is at 1500 psia and 140 F with a producing gas to oil ratio of 80(X) SCF/Sil3. and tine gmus specific gravity of 0.73. The second stage sepmmrator temperature is 100 F. rime produced condensate specific gravity is 0.760. E,stirtrate time reservoir gas mnrolectmluur wemglrl. 2. I0.Estimate the compressibility factor of tIme gums conrdensmnte, reported in Exencise 2.4, unt tIre initial reservoir condition, umsinng tine gcncrmil used clmmurl, 2.11 What is the minimum flow rate of mu guns witlm mm specific gravity of 0.6 to conntiirtmmntisly remove condensate in a well with a wellhead pressure of 2(XIX) psium mmt 120 F 7 TIre ttmbe nominal diameter is 4 in. 2.I2.The viscosity of a North Sea gas conslensatc at 100 F, ummrd 55(X) psmuu huts beern measured equal to 0.0588 cp. The gas nmolecular weigint is 28.7. Estimmmatc tIne gmms viscosity at the above conditions, using the Lee et a!. correlation. 2.13.A gas reservoir is at 300 F and 6()00 psia. TIne gums specnlic grurvnty is ((.6 mind lire connate water salinity is 200,000 ppm. Time first stage separator teirmpcrumtumre mmmd pressure mure 50Fand 1500 psia. respectively. (a) (b) I-low much liquid water is produced by condcnsmutiomn inn tine sepunrurlor. Is there a possibility of hydrate formuruution urt time uibove conrdmtiomns!

3 PHASE EQUILIBRIA
Productiomn of reservour tiumids often is mucconnipannned wills vmmriationms inn conrnpm.nsition, pressure unnnsl Iemnrpermutsnre. llris leads not only to ctranmges in Ilurid properties, hunt also t(r formation of

2. l4.A lean gas is at equilibrium with water at reservoir pressure of 6527 psnmm amrsl 279 F. The water salinity, WS, is 10%. Estitnate: (a) the anmoumnt of dissolved gums inn wuiter, mind (t) tine isothermal compressibility coefficient, (c) density, and (d) viscosity of the waler uut tire mubove conditions. 2.1 5.A sample of the reservoir oil described in Exercise 2.6. is broumght into equilibrium witin equal volume of fresh water at the reservoir con(Iitions to simulmmte a water drive process. Estimate the reduction of the oil bubble point pressure at the above tempermmnurre due to its contact with water.

new piimmscs. or elimrninrmutuon msf sunmnme of tire existing plnases. As clnunnges within time reservoir mire oflein quite slow, it is rcumsomruuhlc to mtssunne thmmt mill lIne co-cxitimng plnases, at any point in llre reservoir, utre ins equmililnriunmm. hlenrcc, tIne prohlemnr humsically redurces to detennination of tine equilibrium commditiomms for a imnullicomnmponment system.
Fluid equnrlibrium, amnd lIme missocimmted emngineerinng applications, are of innterest inn mnmamuy fields, witlr

well estumblishnesl primrciples. limermnodynmunnics Irave 10mg beenn used to inrvestigate fluid equnlibrius, mmnd to reduce gcmiermml criterimu and Imuws to prmuctical tools. llris subject has been extcmnsively covered mm nrsmimnerouis text hooks amrd pumper collections. References l-2~prmsvide buisrc tinermnmodynmuumnnic concepts of Iltnid plmase equnilil)ria. In tIns chmrpter. ccntmminr concepts. sIc finn it ions, mmnrd tlnsrnnmmdynnumi mnic reluit ions winch mmre fm.mndummmnemrtmmi to Humid equmi I iimria are reviewed. llmese will brims tIre foundations for mull tIne methmmsuis used to dcternmninc fluid behaviour in tine renrnmmining clmmnpters of tins book.

3.1

CRITERIA FOR EQUILIBRIUM

Iknr mu stmnscd sysndmnm, i.e. mnur( exclrumnngimng mmrumss winh ins sunrrusumnidi nrgs, tIne cirmmngc msf syslennn total energy F. stunred mrs (ire iintcrnnmml emnergy U, potemmtiunl energy E~,and kinetic energy Fk. is only dune to rammsfer of treat Q, unmmd womk W, across its boundary as stated by the first law of
tlremmimodynuuinmics.

AI~A(I+ Al)~ + Atuk

(3.1)

where tire lmemmt gnvcun to lIne syslemnn mmtud work dome try nIne sys(emnm Inave been assigned positive signs in tIne umbove equmltiomm. Wlnemr sucim a systenmi unndergoes ann ideal but unreal reversible

106

.1

9 I uu.c Fq,i,

mi, i,m

.(. I. ( rileria for Eqimilibrirmmmn

1 07

process, with no changes in the kinmetic ansI potential energy url unnnifusrmrr trcssumrc P uunrul temperature T. the combined first and second laws Of thermoulyimainnics stunte.s 2. dU = TdS PdV where S and V are the system entropy, and volunnue respectively. If the process is irreversible, the clmange of entropy is higher than tlnat in tire umhm,ve eqnmmmtionr. leading to. dU <ldS

A = Ii 1S

(3.12)

(3.2)

lIne tmnimnimnnmnmnr (iihts energy. mis the gemncral criterion of equilibrium, is often used to derive w(nrking expressionns. llne Ilelnmrlnohl7. energy isray, Imowever, lead to simpler expressions in ssnme applications, as described in Sec(non 5.3. It should be notcsl that Eqs.(3.X-)) are necessary. hut not sufficient conditions. At equilibrium, time system energy is at its lowest value amongst all possible conditions, including all possible nminimmma occurrimmg in lIne calculated Gibbs furnmction. Hetnce, the equilibrium should he dcncrtniincd by semurching for line global nnininruinr value of Gibbs energy. Further infornrruulion omr tIne Gibbs energy mmninimnuismution to idcnmlify equihihriunm conditions is given in Seclnons 5.2 Chemkal Po(eritia~

PdV

(3 1)

Since all real processes are irreversible, the above immeqsnmnhity ,stunlcs limit f(nr un process unt cm(msslmmnrt S and V, U tends to decrease as tine state of equnmlnhrmunnnr ns mnpproactned.
Tlrernnmuxlynaismic relations can be emnployed to develop similar s(atenmnennts anmnomrgst ollrcr sImile

properties. The Gibbs energy G, is defined as. (iwhllS

where II is tIre system enthmnlpy,

A clmsesl systemmm counsisling (Sf mm nunmnhcr of pirases in contmrct, called a heterogeneous closed systcnnn, can he tremuted a.s us collect imnnm of unpen systems. wlnere each phase is considered to be a lmommn(ngCneOUS one. cxclmmtngnnmg trruuss wills otiser open systems. (3.4) Inn umnn openr syslcnnr lIre clruummgc mnf Gibbs energy cannot he expressed by Eq.(3.6) as the energy cmii) vmury by comnrponrcnrns of mIss systenur crossimrg tIme plsmrsc boundary. itence,
N

si(

Ssli

VdP

X(mJG /dmi, mm,,,, sims,

(3.13)

11w Ii

+ PV

(3.5) winere n is tlne nuninsher of nnnoles of emmcln coimnponeint, with (he subscript nj referring to all mole 1 imuinubers except nj, mmnnsl N is tine lolal nnunrsher of components in tIre system. (3.6) lIne dcrivumlivc of mmnr exlcmnsive property relative to tIre number of moles of any conmponent at comrstmnnmt pressure, tcmmnperatunre otirer mole numbers, is defined as the partial molar of thrmtt connponemrt. TIre partimulmmnnd inrolmmr Gibbs ennergy is called clnennicai potential, Ii property (dG /~rr,)~ ,,. 1 (3.14)

Substutunting Eqs.(3.4-5) in (3.2-3), we obtain, dGSdT+VdP

which states Onat, at constant 1 and P the Gibbs energy tends to decreurse mis remsl processes, aurd remains constant in a reversible process, (aG)~, 0 (3.7)

Tlrat is, in the equihibrimnm stale, whicln is tine ultitmmate comrditnon of mmny rcmul process, lIme systenrn (Iihhs energy is tnminminnmunn, i.e., =0 and m (d G) r >0 (3.9) (3.11)

It cmunr he slnmswnr Ill, tlmunl,

(i, =(0G/)nr,)nm.,

=(4A/m~(nn,)~ ~,,,,

=(~UI~n~),~

(3.15)

Sumhstitutinmg Eq.(3. 14) its Eq.(3. 13), we obtumin, (IG

=

SsiT+ VdP+ ~t~dn~

(3.16)

Simmiitar cx pressiomns using anotimer energy IC rio. tlmumt is, lire I Id nmmlnoltz. cue n gy. cans umlso be derived as the requirement for equilihriunnr. (~A), and (amA) ~
>

For mm clmrsed sysncnnn cumnrsislinng msf 0 plmumses, tIne mmhmnvc equnation cmun be written for each phase. lIme clruunrge of totmml U ihhs ennen gy md tIme closed systennn is, tluereforc.
0 0
+

(3.10)

0
~( ~fi

s1G = ~(S),~dT + ~(V) dP ,n-~l t=t 6 0 (3.11) wlncre s denotes emmch plrase.

,slnr, ),~

(3.17)

15=1 i

where, A is the Ilelmholtz energy defined as,

108

.5. Pinas Eqimilibrm

.5. 1. Crmterma for Eqmmiiib~iunm

I 09

At uniform and constant temperature and pressure conditions, the general requirement of equilibrium, given by Eq.(3.8), leads to,
0

Integmumtimrg Eq.(3.25) at cotrstant temnrperuuture, we obtain, I.

~t

Ri ln( PiP)

(3.26)

(dG)Tp

)~ =

(3.18)

TIre mrhnve cqummlmorm prn~idesmm simmnple relation for the ctmmtnge of clsemnmical potentimml of a ptnre idcmml gums wlsemn its pressunre clsmmnrges Ironnn P0 to P isotlnerniral ly.
,

As the total system is closed with no chennicai reaction, the total number of nnroles of eumcin
component remains constant within the system,
8

l_esvms (I genrerumlmscd Fq.(3.2(n) for milnlmhicumlioms 1mm neuml systemnus, by definimig ur corrected pressure fummsction f. cmilled fim,m~wiiv(escunping temndcncy) as follsws,
o,
-.

~ ~dn~)

i= I ,2,...N

(3.191

p~ =

RU tmm(i, / I,)

(3.27)

Considering Eqs.(3. 18) and (3.19), and since changes in mole nummnbcrs of eaclr cmsrnpoiretrt mire arbitrary, we obtain, I.h~
(I)

wlsere Iii ummsml 1~ mime tine chmennicmml pmnnentimml and fugacity of tIre commmponemmt i, respectively, at mm ictcmemuce sImile. For mum ideush gmms, linerefore, lIne fmmgmncmty is eqummi to its pressinre, and tIre fungacity of cads conmmponenst us eqummml to its parliuml pressure. lire rmmtio of furgmucity to pressure is defimned mis tire fumgmmciiy coefmmciemnt ~. For a tnntnlticompornent systdmnn,
I
=

1 (2) hu

(3) i~i

(0)

i= 1,2, . .N

(3.20)

The above general requirement, that is, the equality of chennical polenntnal of emucls coinrponemmt throughout all the co-existing phases at equilibrium, beeotsmes mm practical emsgimseerinmg tool if (Ire chemical potential can be related to nneasurahle quantities. TIns is mmclnieved try expmessimng lIre chemical potential in terms of auxiliary thermodymnamnic fumnctions, suds mis fugacily or uudlivity.

~ 1)1,)

(3.28) Simmce mill systdtrms belnave mrs isIeuml guises al

Fugacity
As relations amongst state properties are independenrt of tire process puntlr [21. Fq.(3.6) fmr mi reversible process can be used to express the Gibbs energy clnange. hence, tlnc cisenmicmul potential,

winerc ij is tIre nnmmnle Irmmcluonm of tIns comiuponent i. senv low pressnircs.

wlmemr

(3.29)
nomrideuulity of the

dG = SdT+ VdP
For a pure substance partial nnnolar properties are the sanse as molar properties. change of chemical potential of the pure substance i, is given by, dIl =dg =5 dT+v~dP 1 1 1

(3.211
Hence, tIre (3.22)

tire depuurlure of fsnguucmly coeffrciemuts fronrm unmrly us, tlmenefore, a nsmeursure of systemir.

Writing Fsj.(3.27) for lIme connnpomnerml i, in emmch plnuise of a lmeterogemmeous system, wills mmli referemsce stmites unt lIre sumnnre temnperatumrc. tIne equnmulily of the clsemniicmul potential at csiuilihriumiur givemr by Eq.(3.20). lemmds to.
=

where g. s ansi v are tIne moluur Gibbs cncmgy, mmnm)immr cnunrmrpy mmmd mnroluir vmsluuminc respectively. At constumnt temperurture I Ire above eqummilion rcslumccs to,
(~11~ ~)T
=

~(?

=~~=-

=t-~~

i=l,2,...N

(330)

llnuil us, tIns fmngumcily unl cmmclm dnmmnpmnnncnst slmorilci he eulnmmml thmromrghmmain mmll the plummses helen ogeureonis sysne iii mu ruin nit tin mmiii

mm

mm

Vi

(3.23)

Eq.(3.30) is mms genmemumi as Fq.(3.20) for rclmmtinrg the properties nil equnilibrumned phrmuses.

In.

which leads to a simple expression for the chemnnmcal potential of an ideal gums, witlr lIre pressurevolume relation as,

Pvj=RT
that is,

(3.24)

Irowever, Inums I Inc umdvmmmrtumge of cnnnploynng mu function wlr ichm can be mnrore easily umndc rstood mind evmulunmmtcsh as mmcormected pressure. Fugmmcity dunn he immnuugined uus mu mmmeasure of tIre escmupinng tcmrsicnucy of inmolectiles fronnn one pimumse to arm asljmucent plsmuse. I leurce, in a mnntnlticonnnponetnt systems. ml the fumgacmty of a commmpoirenrt in the two asljumcent phases is the same, the two phases will he us cqinnlnbrismmrm witlm no met Irmmnsfer of mmiolecules fnonss one plsmmse to anollner. iIue tnmgumemly cumin he relumled rngomounsly to mmnemrsimrmmhle properties umsimrg tinem mmnodynmummnric

relumlnomssl IJ, (~I.Lt ~~)T = RT/ P where R istheuniversal gasconstant.

(3.25)
lmm~, =
-~

Rriv

L~~)
~l ~

R~fVdVinZ

i=l.2,...N

(3.31)

r.v,,,,,,

110

:1 ih,,,ce Lqmmi/nh, i(i

3I (

riUrul

f,r Eqmmiiihrimom,
4 2 7.~x l0 P /2l:~= 0.6875

Ill

where V is the total volume, n~is time nummmher of mrrolcs of comnspommenrl i. mmml
conspressihility factor given by, 5 Z=t V/nuRl whsere, n, us the total nutimber of moles in the mixture with N compommemrls,

Z,

is tIre nrmixtumie

2 lmm~= ~6.5 x 10 p m~=O. 5028

(3.32) f=m(o<i=5.028 MPui

Activity
(3.33)
Eq.(3.3 I) simply yields the fugacity coefficient, provided nhnmut the vmulunc of (01 I dir, )m v over tine whole range of tIre imntcgra(e, i.e. V mul P, Or V= infiruily urn P=0 , curms he evumlmmmutcd. this is urn ;mcimmevumhlc tursk, wills msn muuccptmmhle cmngmneernnsg mmscmmrmmcy, using ssnmnr-ennnjnirmcmil snlmiumnionrs of stmutc mis described ins ( Iumspler 4. The unrsler of Integralmoms mmmd differenntnmutnon mrs Eq.( 1.1 I ) nnmmy he rmitcrclnuuurgcsl, resumlni nng

Fntlner of ft or 1 uslay he selected mumhi(rmnrily as the reference stale property in Eq.(3.27). The furgacity is most often the chosen reference properly. llre substance i. as pure or in a mixture, mit umnry pmc.ssmnrc cummn he selected mis lIre rcferemuce stumle. line nmmly limunitation for selecting tIne relerenncc sImile is IlnumI its lcnmipcrmmtmmrc mmntisl be equnmml Ims (hue equilibrium temperature. It i.s often seleuned mus (he pumme sumhstumnce mit tIre systems) totmsl prcssumrc, or at its vapour pressunre, at lIne tine summl n mg IC mnmpem mmii) re - A ~inm 95Cr ulrmnmcc of nIne mc Fre,sce sImile fmncihitaies the application of then nimmsdyn;mumnnc ruImmlimnnns Inn cirgilucsrinrg prolslensis.
Inc n mit nun (ni tIne fsngmms nly mm I Inc s(mmte of innIciest to tlnmut mm (Ire referenice state is called (Inc activity.,

mu.
(3.34)

In~ =

~IJ

dn,L

1~

~ ~.RU V)

~dVl
Jrvn,,,

In Z

i= I ,2,...N

I, / 1,

(3.36)

line immtegral is the residsnal, defined as the actunal property mmrinius that calculmuted by musstrnruinrg tIne

lIme mdtnvity cmmn, tlmerefore. he connsmdcrcd as mu meassure of fugacity contribution or activeness of

mu conapomment in a nnixlnnme. I,
=

ideal gas behaviour. i-Ielmholtz energy divided by (RT), The aslvmunlage of Ire mmhove frmrmn is tlsunt tIne immtegrationm Immus to be done only omnce, usnrst mrll properties euro Os cmmlcunimmtcul mis mierivuutives of (lie residual 1 Ichmnslsuihtz fumnctiois 131. lIrc use of dcsci itiimig lIre tnchrmmvnunmmr mn! mm systemnr by evmmlummtmng its Iiclnmrholtz energy, in preferetrcc to the Gibbs cnnergy, will lie disctnsscd inn Section 5.3.
the fm.ngacmty coefficient of mm pure compoumnd cumin he dclernnrinncsl by inncorpormmlimmg Fq.(3.32)

u,f,

(3.37)

into the general expression for the fugacity coefficiemmt, Eq.(3.3 I), ln~= J(~_~)dP = (Z I) - lnZ + P)slv (3.35)

II is m intuit m ye (hrumt time fungurcily of cumuli cmnmnnponmenrl slnoum Id depend on its concentration in lIne mumxl tine. I Icmice, Ilne unbmnvc eqummul mmmi) stmggcst5 Ihiurt the mmcfi vity slsounld be closely reluuted to the somscumnln at ion. lIne rurl nun of mmcl r vity lo commcc rnlrmntiomn, often in mole fractions, is called tire (I(tIu,13 (oeffjcne,n( (~,
(-), =

c/s~

(3.38)

llcmscc,
wIreme v is tIre msroimur vunluimme. Depcmrdmng omm tIme fornnr of tIre eqummul ions of stumle, ommc nil lIme uulnmsve two cxptcssiomss for tire fugacity can be sinmrpler to usc.
=

(~x,b,

(3.39)

F.~aoupIe.5.1.
The connpressihiliiy factor of a pure gums at 290 K cain he rctmmied 10 ins ~ us,

lime mmclivnty cmncffmcmcmmt is mu very umsefiul mmumxilimumy functions to describe liquid phases in phase Cdli nil m hr idunn duntcum I mit ions. Es tenisi ~ sttidies lnmuve been conmiucted ho relate fine activity duel I rcmenit hr millrer I lncnmmnmrshymnmmnmnic funnnct iotrs, munch numnncrosns nsrodels have been reported to (helenmnnmmne ml vmiltne for liqunid connrnonmetmls Ill.
3.2

Z=l6.5x10

4 2 P7.5xl0 P
Calculate the gas fugaciny an 10 MPa.

P<15 MPmm EQUILIBRIuM

wbmere P is in MPa. .So(m, (join:

RATIo Eq.(3.3O) for such a

l,et urs :or~smder two ptmumses of liqumisl. t.. murrul vmupounr, V. at equilibrium, ,systenmm is. i =i~

Substituting the above expression of Z in Eq.(3.35), we obtains,

i=h,2,.,,N

Inm~ = ~(~_!)dP

2 j(_6.5 x lO~ P

(3.40)

AIipiymnsg Eq.) .3.28) to hurtin phmumses, we ohstmnin: I, =x,P~r

i=t,2,.,,N

(3.41)

112
fVyp4~v i=l,2....N

3. P/nose Equilibria

.5.2. Equilibrium Ratio

113

(3.42) For vapotnr and liquid phases at equilibrium Eqs.(3.40) and (3.46) leurd to,
=

Hence, i=l,2,...N (3.43)

xf,~,,,,,,

(3.47)

where Ki is called tlse equilibrium ratio and is defused mus the rustio of irsole frmuctimnir of component i in the vapour phase Yi~to that in the liquid phase Xi. A general mmmd rigorous approach to determine tine fugacity coefficient of a component in both phase.c fromrs volumetric information, using an equation of state, is given in Clmapter 4. The lack of success of some equations of state in describing tIre volsuirmetric helnuuviour of complex liquid mixtures has lead to the use of tIme activity concept to determrmine fugacilies in tIre liquid phase. Using Eq.(3.37), instead of Eq.(3.28), to obtain the fugumcity of cmucln conmpouremrt in the liquid phase, we obtain, K =O f~/P~ 1 1 (344)

Asstunnminsg thsmmt time vmmpmrunr is mini irlemul guns, we obtuninn,

=

(3.48)

tlre dIed of pressuure on fungmmcnny of mu eonslensed pimase at low pressm.rre is small III anti can he tucgledlcd. llnc fugmucity oh mu ptmme liqn.nnd rut krw pressunre can, therefore, be assuunnred cqtnmml to its fugmucuty at (Ire suuturuutu(rms pressunre. line funguucmties of sumlurnatesi vapour and liqumid are cqummml, uns time Iwo plrutses umre mit equmnlihrnunmum. Funrtinermore, lhc vmupouur fugaciiy at low pressure canm be mussumrrcd equutl to ils pressure. 1-lenee, the hmqumisl ftngacity can be tmukeii equmuul to tine vapounr presstmre of time stubstance mit tine prevaihinrg temperature, ~I.

pure

(3 49)

The above approach, known as the split mcthmod, cmun give relimuble re.suill.s fmrr sysleunis wills vapour and liquid phase properties far apart. The split imnetlsod proposed by Clnmmo-Seadcr 141 has been extensively used in the oil industry to predict the phase behumviour of gmms-oil systenrs, particularly at transfer line conditions. At Imigh pressures, especially close to critical conditions, where the liquid and vapour phases behave sininilarly, the mmhove approacis is not recoumnimiended. Phase behaviour models which use an equation of state for tIre gas uumnd an activity model for line liquid phase may also miss the retrograde behaviour of gas condensmute systemrms. In getrermul, the use of a single equation ofstate to describe all fluid plsumses slmould he adequmutc in mulnnost mull petroleum engineering problems, as described us Chrupter 4. The fugacity can also be evaluated by simple methods, employing limiting musstmnmptions. hunt still useful for engineering purposes at a variety of conditions. An attractive ursstmnunpliomn is tlruut the fugacity of each component in mmmixture is limnearly proportionmul to ils comncennlrmution. TIns assumption, known as tlme ideal solution, is gemrerally valid for immixtuires coimn moscd of similar 1 components, or for dilute solutions, Ij
=

wiscrc Pt us tlme suutunrmutmonm (vmrpounr) pressure of the psure compounsl, i. Suhstituiing Eqs.(3.48) and (3.49) into Eq.(3.47), we obtain, y~P=x~P~ tsr K=P~/P (3.51) (3.50)

Eq.(3.5 I) us knowus mrs Ruuoults lmnw. Considering the above assummmptions, it is only valid at low pressure for .simmrple unmixtumre.s, Euanmple 3.2: (mmlcuulanc eqnuilihrimnumm rmutios of (I, mmmrsl nC 10 in a vapour-liquid mixture at 344.3 K. and 6.895MPun. musing Rmiouults lass. Solnitm on:
The esluilihriunnn rmntmos are calculuited fronru Eq.(3.5 I). where tlne vapour pressure at 344.3 K is estmnrmmied unsung lIne t_ee-Kesler equmaiiomn, Eq.( 1.10)), simsnilar to Example 1.1

(345)

wisere Xj and C~ are the proportionality constant and the concenlrmulion of tIne comrrpomnent i in the mixture, respectively. Depending on the definition of tIne proportionality comrstmmnl. two widely used methods, Raoults law and Henrys law, will result. Raou!ts Law First, consider the case that Eq.(3.4.S) is valid over tIre whole range of cotncemslrmution, including C~=t. i.e. pure component, the proportionality constant slsould, therefore, he equuml to the fugacity of component i, as a pure substance,
f=zf,,,,~

(3.46)

(.~onmpmmnenm C nrC,,

_~__

I 7.68167

-4.25215

(.757116 -5.0053k

l,N-lta I 16111.1)7 0.0025552

where the concentration is expressed by the nmrole fruiclion ~i The above equation is known as tIre Lewis fugmucity runle, Cmmnnnpumriimg Fu .(3.46) witln 1 all activity Eq.(3.39), it can be concluded tlmat the Lewis nile is valid for mixtures witbm coefficients equal to unity.

line e,stnmnimuled vunluic of rnretinmmmse vmmtiouur pressure, 11681 Mia, by ilne Lee-Kcsler equnafion is mom micceplunhte. tIme prevaihning temmmperummuure ol 344.3 K is well above the methane u imicuml memmmperummunre of 190.56 K. uunnul pure mneiiianrc cmmnmmot exist as liquid rut this icimipermuiuume. I Icuuce, tIne cmnlcmmlmmied vmmimour pressure by ulme Lee- Kesher equation, or ansy onhner vzmpour pressurre correlatioms, is an unrreal value mmusd just an extrapolation of time vapour pressure curve above tIre criticmuh temperature. Sinssple correlations, such as tIre Cox chart, Figure 1.3, gennermully provide more reasonable values comparing with cousmplex

114

.5 lfmosm Eqm,i!il,ria

.1.2.

6qmmmlil,riunn

Ratio

I I 5

fumnctions that mnclunde parainmeters adjusted by mumtclning experinuenmtmml suupouur prcssumre data.
Figumre 1.3 shows muir extrapolated vapour pressure of 48.26 MPmn (7000 psimi) for nmrcttrmmmre

Ilenmrys comrstmmmmt of mu coumnponenl in a solvent is considered to he independent of its eonccnrtrumtionr. html a fuinctionr of tenrrpcratunre, and to a lesser extemnt pressure. Henrys constants for gmmseouis comnrponrenrts of reservoir hluids inn water, at low pressure, are given in Figure 3.2

at 344.3 K (160F),
Suibstitsiting the estinrated vapour pressure valumes inn Eq.(111.51). tIne cmmlculancd cqmrilihriumnn ratios are as follows:

13000 12(800 110(8)

.~rur~cn~

K,.=7.000,

Knci,,=0.0003706

Experinrental values K,.=3.998, and K~(.,,=0.00271271. Henrys Law Ihe proportionality of component fumguucity to its comrcentrmution, as assuimmncd ins Eq.(3.45) is valid for components mmt low concentrations in most liquid msmrxtures, lm
=

l(X)00-

Pro mumme

9(88)(3.52)
cm 0 cm a
U(5

/
-

H X 1

8000

Hydrogen

.~

where I1~is called flenrv.s co,n.cta?i/, which is experimentally de(ernmnnned. The concentration of component. i, is generally expected lobe less tlrmmnm 3 tniolc e for tIre muhove equation to be valid Ill. It is, therefore, a tmsefuml equnmulion Its slcternnninc tire sotumhility of hydrocarbons in wmuter where time smlsnhilily is gemnermully low. At low pressure, where the assumption of ideal gas is valisl, Eq.(3.48) can he unsesl to describe fugacities in the gas phase. 1 which is known as Henrys law. Hence. K =H~/P 1 Py~= iI,x (3.53)
7

70(X) 60(8)

T1~
F thane
.

Methane

cm a cm
0

5(88)

4(8)0.

y
cm
5)

(3.54)

Figure 3.1 compares Henrys law with the Lewis rule, Eq.(3.46). Note that wlnilst the fiuguncity of a component is proportional to its concentration across the whole rmmnge of comicenmtration for the Lewis rule, the proportionality is only linmited to the low concentration ramnge for ilenrys law. The proportionality constants are generally different.
It,

Lewis rule Counnpositimnn, x,

Figure 3.1. Comparison of Henrys law with Lesvis rule.

280

3(8)

320

340 360 380 Temperature, K

400

420

440

Figunre 3.2 llennrys constants for soltuhility of hydrocarbons in water. Reprinned winh permission 161. Copyright (1953) Airrerican Chemunical Sociemy.

116

.~

lha.se Equilibria

.C 2. Eqmmilibrinmnm Ratio

17

The dependency of Henrys constant on pressure can be determined rigorously by

thermodynamic relations as, = H~exp[vr(PP)/RT] (3.55)

where Hmis Henrys constant at ~0, and v is tIre puurlial nmolar vmniuirmc of comnulxrtneml( i ins mIsc solvent at infinite dilution, assumed constant over time prevailing pressure uunnd comnspositioum ranges. Eq.(3.55) is known as the Krichevsky-Kasarnovsky equation t51. Limited information on v~ofcompounds in water are available in the literature [61. The partial molar volume varies with temperature, and becomes pressure dependent near time critical point. 3 An average value of 35, 40, 55, and 80 cm /gmol, can be used for nitrogen, nurethuitne, ctlnamse. and propane respectively. Example 3.3. Estimate the solubihity of methane us water at 373 K, amid 65 MPa using ilemurys Iumw. Compare the result with the value shown in Figure 2.28. Solution: The Henrys constant for methamse at 65 MPa is calculated fromnr Eq.(3.55): 3 At T=373 K, H0=6.5xi0 MPa/mol fraction (Figure 3.2), and V=~.IOMPa. Hm.m(6.5X 10 MPa/mol frmmciion) exp[(40X hO uns/kgurrol )x) 65.t)0-O. I 0)M Pmm/ 5 (0.00831 44x373 MPa.m /kgmol)J
=

components at a certain pressure known as the convergence pressure. Figuure 3.3, This appemurs to suggest that, the consposition of both phases should be the same at the convergence pressumre. As similarity of hotim phases only occurs at (he critical point, it implies that for any petroleum fluid, the critical temperattmre may be selected arbitrarily, which cannot he correct. lire fact us thnat tine comrvergence prcssuume does not pimysically exist, unless the prevailing tennperahuure is the nsnixture critical Iensperature. At other tensperatures, the nnixture will be an under satunated single phase at that pressure. It is however, a very useful parameter to con-elate experimentally determined K-values with pressure at any given temperature as an end point.

me

0 a

E
I .4378x I o~ MPaimoI fraci iomn.

}~J~t{1 ~ .1-~ ~ ~fl1~ -~

~ ~-.

~\1lffl1~

.ii
~
-

-r~JJI4--~-~
~.

EEE

~
-

s,_

..

cm 00

The solubility of methane is calculated usimrg Eq.(3.52), 1~m~ = lYcu4~i= The gas can be assumed as pure methane due to low volatility of water relative to methane: Ycn= The fugacity coefficient of methane vapour nit the prevailing conditions can be calculated by an equation of state as appliesh in Example 4.1. Assunniung ~ we obtain, f~=P=65MPa= (h.4378x10 4 MPa/mol fraction) x
ac

~~I1ll~ Ii~NJi! ~
0
---~
~-~

~\1H}~J ~ f_
~ ~

~L~E
0

.~E
~, -

-~

~ ::,
-

mO

io~

mole fraction of rrsethane in water.

Pressure, psia

The solubility value is read from Figure 2.28 equal to 4.3xlO. Empirical Correlations In spite of recent developments in theoretically based phase behaviour models, empirical K-value correlations are still used in vapour-liquid equilibrium calcuilmulions, pmurticularly at low and intermediate pressunre conditions, where the K-value cams he mrssumssed independent of tIre mixture composition for reservoir isydrocarhon liumisis. It is well established that, for a multicomponent mixture, graphs of experimentally detemmined K-values versus pressure at constant temperature tend to converge to K = I for all the 1 Figure 3.3. Equilibriunmim mmutio.s of a hydrocarbon mixture at 322 K (120F). SPE-AIME Copyrigbni. Reproduced fnnnn 171 winlr permission. lhe preslicted equmilibriuurn ruttios, usiung Raoults law, Eq.(3.5l). are also shown in Figure 3.3. As the systemrm tenrpermmlurc is conss(usnt, hence, tire vapoumr pressure of all (Ire components, the lmsgumri(lnmnrnc plmrts of tIne Kv:ulumc.s witis pressure are all parmullel straight lines witis a slope of I, as deterrmmimmed by Eq.(3.5l). lIne deviations of predicted values by Raoultx law grossly increases at high pressure commditions, particularly where K-values tend to increase with pressure as opposed to Raoults law.

118

1 P/name Liinninb~ ma

.5.2. b.qnuihhrium Ratio

119

For a binary system, the convergence pressure is tire critical pressure of a nrnixltnre wlniclr lrmrs mu critical temperature equal to the system tenmperature. The compositioim of sunclr us umuixture is different from that of the system under consideration, utiless it is at its critical teursperaluire. Figure 3.4 shows the pressure-temperawre diagr~mof ethane-nrormmsi Ireplmmnie nnnixlumres mit different compositions, with the locus of the critical points identilned by the slotted cturve. At 450 K, for example, the convergence pressure is deterrsnined to he equal to 8.23 MPmr for all C -nC mixtures regardless of the composition. The composition of equnilibrated phases. 2 7the equilibrium ratio, at a pressure-temperature condition do not slepetid on the irnixtuic hence, compositions for binary systems due to the plnase rule, Eq.( I .2). lIme genmerumlesl K-vuulimcs nm any binary mixture are, therefore, valid for all compositions.

wisere KCI and KC2 are the equilibrium ratios of methane and ethane in the mixture respectively. lIse Iladden rtrethod of calculating the convergence pressure treats the mixture as a pseudo hiisary system composed of the Iiglrtest component, and all other components grouped as a simrgle iseavy pseudo comnipomrent. The pseudo component formed by grouping will be, therefore, different for vapour mund liquid phases. Iii (Ire GPA method the convergence pressure is the (Sine calculated for the liquid phase. hence, the calculation of convergence pressure is iterative, except for bubble point calculation, because the liquid composition is not knrown in aslvance. lIre cuilcuilmution proceduure, as sumggestcd by GPA 1101. is as follows:

no
9

,,monesn cilia,,,4
mlii me 2 5585 .5 887! 4 7709 S Ssin (, 25.54 7 00

(I) (2) (3)

8 7

Assunnmne ttse lis unid plrase cotmsposition, or use tIne feed cotnmposition in the first trial. 1 Select tIne liglutest luydrmstumrhotn compssmlent. mmrcttrane alnmrost in all cases., which is Cmslculmrte the weighted urvermuge critical temperature and critical pressunre for tine

presemrt umt least 0. I nuole % in One liqunid plma.se.

remaining hemsvier commrponents to form a heavy pseudo component. (4) Irmuce tIre crinit-mul locums of lIne hinmuury consisting of tIre light and pseudo heavy connuponent on Figumic 1). I (Appendix I)), by interpolating between the neighbouring critical loci.
(5) Read the convergeumcc pressunre mut time prevaiiinmg tensiperature.

0.4

((r) Obtain K-vmultmcs for tlue systenr components from the K-charts corresponding to the estimnrmitesl convergence pressure. (7)
300

Calculate line equilibrium conditions and determine the liquid composition.

350

400

451)

500

550

1500

Tcuu,pcranucc. K

(8) Repeat steps 2 throsngh 7 uumtil the assumed and calculated convergence pressures check wi(hin an acceptable tolerance. Winen tIre calculalcsl consvcrgenncc pressure is hetweemr lime values for which charts are provided, interpolmutioms between charts uurmuy he tueccssmmry. Clearly, a reasonable interpolation can only be cxpccted wlrens tine rpcmumting pressure is lower Iluans (Ire convergence pressure of lire two charts umscml for interpolmmtioim. Plsmsse eqmnilibriunnr calculatnomrs using Kvalues will be described in Sections 5.1. (itA K-vuulnnes ismuve been fitted to funsclionrs of vmmrioums forms for computer calculations 111.1. rIme K-valumc of tIre C fraction (nf mm oil irsixture earn be estimated by various methods. A rule 71 of thumb [7~ smuggests mu vahime equmul to 15% of time C . It can also be estimated as equal to the 7 K-value of a Imydrocuirhon conspounmmd, or (hal of a single carbon group as described in Chapter 6, wills tIre smmnne specific gravity or molecunlar weight as that of the C . 7~ In Clmmrp(er 6 on fluid clsaracterisation, various nsethods to describe the C fraction, and to 74 eslimmmuntc its properties are proposesl. It is nmsore appropriate to describe the C fraction by a 74
nunnrher of pseudo c(rtrrp(rnents. However, tIre following correlations may be used to estimate nIne critical properties of the C frmsction.

Figure 3.4. Detemmination of the convergence pressure for a hinmurry mixtunme. McGraw-IIIII (oummpmunics Cm~pyriglnn. Reproduced from 171 wimh perunnissionn. K-vmmlm.mcs in multiconssponent hydrocarbon systems mut inigh pressure depenrsl on comnmpositionr. A comrmmon assumption is, thuut the convergence pressure alone can remrsomsmuhly describe tIme mmhovc depcnsdency. I knee, equilibrium ratios are often correlmuned mus ftunctioin.s usf lIre pressure, teimuperaturre and convergemmce pressure. Tlnerc are a nummsher of corrclmutioiis lun cslinsmuu(e time cquunlihriunm ratio 181, and different methods to calcunlate the convergetsee pressunre of multicomponent systems in the literature. Care slsould he takcnn to use the smmnne metlsod of convergence pressure calculation as used in the development of the K-valume correlmrtiun. TIme Gas Processors Associmmtion its 1957 presented K-value graplsicmul corrclmutiomis for nmrraffirs,s 1 from metlnane to decanme, ethylene, propylene, nitrogen ansI carbon dimnxisie, using tIme convergence pressure calculated by the 1-ladden nnetlmod [9]. The K-charts lrmsve been revised since frequently. The 1976 revised cimarts for convergence pressures of 50(8) psiut (34.47 MPa) [10] are given in Appendix D. The equilibrium ratios of CO at low concentrations canm 2 be estinmated, within ann accuracy of10%, as,
=

(KC,KC. )03 5

(3.56)

7~ Standing [121. represented tIre grmmphical correlation of Mathews, Roland, and Katz. [13] as,
=

338 202 x log(M~, 71 .2)+ (1361 x IogM~ 211 I)1ogS~

(357)

120
=

3. P/mace Equilibria

8.191 2.97xlog(M~, 61.l)s-(S~ 0.8)[15.99 5.87 x log(M 5 5

121 3.2. Eqnmilibriuuun Ratio it and tinat of C-nC, (kensol), wInch results in a coirvergence pressure of about 41.37 7 MPa (6(8)0 psimu) at 325 K (125 .3l). K- Value.s at luter,nediate Pressures At pressunres below 7 MPa (1(8)0 psium), the effect of mixture courrposition on equilibrium ratios of isydrocumnhonss is not significant, mlmercfore, K-values can be correlated in terms of pressure and temperalture onmly. Stuundimrg 1161 correlutted (he experinusemstuml K-valuues of Oklahomsma City crude oil/natural gas suuummplcs gcmncuumlcd Iry Kunli, uunr(I I lmuclnminulls 1171. musing Fm .(3.59) proposed by Hofftnann ci nil. 1 hl8l.
log KP
= =

where T~, and P~are in K and MPa, respectively. Other correlations to estimate the critical properties of the C fraction arc also available 7 [14,151. Example 3.4. The composition of an oil sample and its equilibrated guns. at 10.45 MPa armul 125.0 K. mire given in the following table. Component CI C2 C3 1C4
nC4

mole Gas 82.14 11.22 4.06 0.42 1.07

0.24 0.30

Liquid 27.36

r~ +
.f

[~

ha (

1~~

h 2

(3.59) (3.60) (3.61)

IC

nC5

10.93 8.56 1.46 4.73 1.77

2.77 3.87 38.56 S=0.8399

0.96

6.53x 10 I + 3.1 6x 10 I~

0.890 2.46x10 7 P

7.36x10

tm 2 P

C6 0.21 C~ 0.33 C7 Characteristics M=210 4

wlmere P is pressure in MPa, and T , (normal boiling point) and T are in K. a is time slope of 1 tire strusiglst line connecting tine criticuul point and the boiling point at atmospheric pressuure, Pa. 1 on a log vapour pressure vs. (T) plot.
a =

Calculate the convergence pressure of the liquid phase at the above conditions.
Solution:

[log( P / P, )J // [I / Tb I / T ] 5 5

(3.62)

The critical properties of C -C are read from Table A.l in Appendix A, and mlnsmse of C~, 1 5 are calculated using Eqs.(3.57-58). It should be noted ihat C is not msormal Inexane, hut a 5 group of compounds with boiling points between those of normal pentane mumrd msormal hexane. This subject is described in Chapter 6, where more appropriate critical properties for the hexane group are introduced. Methane is selected as the light conntr(rnent. with the rest grouped as a pseudo heavy component based on mass weighting, as follows: Component I2quation C 1 C C 3 IC 4 nC IC 4 3 nmC, C C4 75 Total
2

1~ 1 The plot of log KP vs. a ( fb- /T) for conrmponents of a systemsi at a given pressure often formsms a straight lure for inlernnmedimnlc atrsi heavy frumctuons, as showms in Figuure 3.5. Standing nmodilied vmslues of a mmnsd l~,for nmrclhane mmmd etimmune munsd unon-irydrocarbon coismpoumnds to lit tIme same straight line mus other comsnpounenrts. The values of a annd Tb are given in Table 3.1.

M
16.043 30.07 44.096

T,, K 190.56
305.32 369.83 408.14

P~.MIa 4.599
4.872 4.248 3.648

w, xM,I~. x,M, 2 0.03344 0.1)3842 0.0(1864

r K 1,.,w,.
1(1.211 14.209 3.526

MPmu 1~,x,
0.16293 0.16321 0.03151
0,

(1.2736
0.1093 0.0856 0.0146

0,

58.123 58.123 72.15 72.15 86.177 2t0.0

425.12 460.43 469.7 507.6 691.2

3.796 3.381 3.37 3.025 1.812

0.0473 0.0177 0.0277 0.0387 0.31156

1.0000

1)02801) 0.01302 0.021)34 0.03392 0.82422

.00000

11.904 5.993 9.553 17.218 569.732 642.35

t).I(s629
0.04401 0.06854 or

(1.10261 1.49363 2.173 a(rrr -nrr) 5 Figure 3.5. Equmilihriunmrn rmutios mut 5.52 MPms and various temmmpermnlurcs. SPE Copyrmglun Repmodumccd tronru Il 6j wimh perurnussummmn.

The heavy pseudo component with T~~642.35 K (696.56 F) and P,=2.l73 MPa (315.2 psia) is slightly heavier than flCn,. as located in Figure CI. The locus of critical poimsts of methane and pseudo component is drawn parallel to that of CnnCni. interpolating between

122 The values of a and Tb for C + fractions can be obtained from: 7 2 a = 563+ l8On - 2.364n
=

.5 P/mace Equilibria

.5 2. Equmhhrusun Raiuo

123

TIme Mollerup equation[211 to estinnrate K-values, as initial guesses in flash calculations, is equivalemrl to the above, hut for a constant acentric factor of 0.01. (3.63) (3.64) The Wilson equation generally provides reliable estimation of K-values for sub-critical consponents, huut overestiurmates those of the supercritical components (24]. The equation has been extended to higlser pressures [22] as, K
=

167+33.25n -0.539n

where n is the number of carbons of the nornmal paraffin that has the sanrme K-value as tlmmul of the C + fraction. It can be estimated by cotnmparing tire molecular weight of tIme C frmuction 7 74 with those of normal paraffins, Table A. I in Appendix A. Standiimg correlated n for tlse Oklahoma City crude oil samples by. n
=

(P

/P

)5

(P / P)exp~5.37A(l+ 00, )(l 5

T5

IT)]

(3.67)

wlmere Awl
-1(1-

3.85+ 0.0I35T

0.2321P

(3.65)

P~)/(P,-

where T is in K and P is in MPa. Table 3.1. Values of a and Tb for use in Standings equilibrium ratio correlation. Tb, K Cummpound a, K Nitrogen 261 61 108 362 Cartxnn Dioxide Hydrogen Sunlphide 63l 184 Menhane 167 52 Euhane 636 168
Propane

mmnsd Pk is tIne comrvergensce pressure, mis correlmmted by Stmmnding [12]. P 1 44 0. l M 29.0 (3.68)

wlncre ik is ins MPa. rise expurnremnt n varies between 0.5 and 0.8, depending on the fluid, with us sicfmuult value of 0.6. The urhove modification to the Wilson equnation may lead to unrealistic values resulting to non-convergemrce in plmase behaviour calculationrs.
!~xatun pie 3.5.

iso-Butane
n.Rutane

999 1132 1196

231 262 273

iso-Pennanc n-Penuane iso-tlexanes n-Hexane Hexanes (lumped) n-Hepnanre

n-Ocuane

1316 1378 1498 1544 1521 1704

1853

301 309 335 342 339 372

399

Estinmrane cquuilitmrimunn ratios of tIre guns-oil system described in Example 3.4. using the Standinug nnnetiuod amid the Wilsuin equation. Connpare the results with the experimental values.
,Solum (tam:

n-Nonane
n-Decaume

1994 2127

424
447

The critical properties of C,-C are read from Table A.I in Appendix A. The properties 5 unf C~, are calculated as follows. Staumding Correlauion Suhsmimuilinng tIme pressure amnut tempermnluire in Eq.(3.65), the equivalent carbon number of 46 2 C,, is deternmnimucut cqnnal to 10.66. wlniclr results in (X22l 3 K and Tb= O. K, using Eq.(3.63) mmnd Fs .(3.64), respectively. 1 The coefficiemrns of Eq.(3.59) at 10.45 MPa are calculated as. r~=0.067465, umsimrg Eq.(3.60)
3
=

Alnhouugir the equilihriuimm ratio correlation was based on a limited nunrrber of dmmta, it Irums been sirown to be reliable for crude oils from various regions of the world, sonune constaining ssmbstantial amounts of non-hydrocarbons [19]. Other generalised K-value correlations, eilher neglecting the effect of mixture consmposition [20, 211 or using a single parameter such as the convergence pressure to characterise the mixture 122.231 have also been developed by various investigators. These correlations would be mainly valuable in generating initial guess values for using in an equation of state to calculate fugacities, as described in Chapter 5. Wilson [20] proposed the following equation to estimate the equilibrium ratio below 3.5 MPa (500 psia): K,
=

0.5526, usinrg Eq.(3.62)

The results are given in ihe following table, Wilson Equation flue esninnmation of criuical properties of a pseudo component, using its specific gravity and nrrolcculmrr weight, is described In Section 6.2. A simple approach is to represent C , with a 7 weight nrornral alkamre with tIre same nisolecunlar weight. In mInis case, C, . with a molecular 1 of 212 is considered to reprcsenrt C ,. The critical properties of C.,. are, therefore, 7 esimated equal ~mT,=736 K, P =l.340 MPa, and w=0.7697. 5

(P / P)exp[5.37(l 5

w,)(i

T , IT)] 5

(3.66)

where 00 is the acentric factor and F~ansd P~are the ahsolunte criticuul teurspcrmttunre mmnd pressuire respectively. The Wilson equation basically uses Raouilts law, with tIre vapour pressure related to the critical properties using the definition of the acentric facor, Eq.( 1.9).

124

.~.

l/,a.ce Fqniilmbrum

.5.2. Equmi!ihrimmnn Ratio

125

The convergence pressure for the modified Wilson equation is cimlculmuted frurnn Eq.(3.68). equal to 57.94 MPa, with the value of A=0.644. The calculated equilibrium ratios using Ihe Standinng menirod. K,. mIne Wilsous dtummnlion, K~, and the modified Wilson equation, K,,~. mire cormrpmired with tIre e.uperiusseuntmul vuilues, K,, un the following table, Component Equation C, C 2 C 3 iC 1 nC 4 iCm nC 3 C 4 C , 7 a, K 167 636 999 1132 1196 1316 1378 1521 2214 T. K 5 52 168 231 262 273 301 309 339 460 K 5 3.59 3.4601 1.1454
0.549(1

K~ 3.66 4.1626 0.6666

(1.1731

K,,~ 3.67

K,

lime equmilibriumnn ratios of consponemsls of am oil measuired in a test, sinrsulumting oil vaporisuition in reservoir by usmethumne iunjection us tlnrcc contuuct stages at courstant pressure aumd iennperature. ale shown in Figunre 3.7. TIne vmuriation of K for each comsrponent is solely due to clsammges in the 1 ursixture composition by gas c(rnlaclinmg oil, TIme results clearly indicate thrat the correlations igimoring tire compositional effect cmumummot provide relimmhle estimates of tine equmihihritnmnr rmltio at isigis pressure conditions. lIme measumred dmsta at umny overall consposition can, irowever. be represensted reasonably by a straiglmt line.

to

4.60993.01)22 1.4170 .0274 (1.5947 0.474)

0.3245 0.2727 0.1686 0.1478 0.1)874

0(1088

0.1)693 0.1)499 0.0207 0.1)162 0.01)57

(1.00(8)

1)32911 11. 2669

(1.1517

1)2855 (1.2268
(1.1351

0 ci

0. 1292 0.0661
0.00(15

0. 1111)1 (1.0539
0.0086

me: C
0 .0 0

.1

...

0 0 o~ q cm 0 1
00

PG
Contaci Smage

or

Note that although the pressure is above the working range of the Standing correlation, it predicts the results more reliably than others. TIme modificatious Immus inrpnsved the Wilsurn equation in. general, except for predicting the equilibrium ratio of nretlsamrc. All the above K-value correlations can be used to check the interrsmml consistetrcy of mmreasumred equilibrium ratios. Figure ~.6 compares the Hoffnmann plot with that of mnodilned Wilson plot. where log K~ has been plotted vs. (I + (0, )(I T,, / T) at coisslmmnt pressuire Ior initenmscdimmte compounds of an oil sample. A straight line can clearly fit the data, including non-paraffins, by the Wilson method, which uses the acentric factor, more closely than that of the Hoffmann method.

.01

.u

02
01

.001

-2

l
(t+tO)(ilITr)

Figure 3.7. Measured equilibriums rmmtios in a test simrmulating oil vaporisation by methane at 34.58 MPa and 373K. Altisouglm tIme data at Inigln pressure mnmay umot fall on mm straight line, one would expect them to follow a nronotonous trend. Therefore, any mmseasured point tisat is completely outside the general trend could be suspect. The iloffmann or Wilson plot is thus a good tool for checking the integrity of the mmmeasuremenrts. Other linear plots between logK and component properties 1 2 sumch as id , ThI [25] mind the mniolecular weight 1261 have also been proposed which can be used to evalumate nnscmusured datum amsd estinrnmmte equilibriuns ratios.

0
,~

Hoffmann Plot

C
.0 .0 0

0 n11

0.4 -0.8

-0.6

.0.4 -0.2 0.0 log(l-TblT)/(l-T~)

0.2

0.4 3.3 REFERENCES

Prausniti, J.M., Liclmtentlmaler, RN. and de Azevedo, E.G: Molecular Thermodynamics of Fluid-Phase Equilibria, 2nd Edition, Prentice-Hall Inc., NJ (1986).

a:
.0 ,0

~3

Modified Wilson Plon

2. Snurimln, J.M. munrd Vmuus Ness, I iC: listrodunction llmernnirdymnaunsics. ihuird Edit ions, McGrawI lill (1975). 0.4 -1.0 3. -0.9 -0.8 -0.7 (t#0))(llIlr) -0.6 -0.5 -0.4
4.

10

Clmennicurl

Fnginseering

Miciselsen, ML: Private Conrsnsmunications.

Clmmro, K.C. aimd Seader, 3D.: A Gemmeral Correlations of Vmipour-Liquid Equmilibria in Ilydrocarboim Mixtures, AICInE J., Vol. 7, No. 4, 598-605 (Dec.. 1961), 5. Krichevsky, lR amnd Kmmsumrmsovsky, JS:Thernrodynansical Cuilculations of Soluhilities of Nitrogems and Hydrogen in Water at High Pressures, 3. Anu. Clsensm. Soc., 57. 2 168(1935).

Figure 3.6. Equilibrium ratios of intermediate compounds of a gas-oil mixture at 20.79 MPa and 373 K.

I 26

.5. lha.ce F.q,uilibria

3.3.

Referent-e.c

127

6. Kohayashi, R. and Katz, DL: Vapour-Liquid Equilibria for Binrary ilydrocarhoms-Wurler Systenms, tnd. Emsg. Chem., 45, 440 (1953). 7. Katz, D., etal: Handbook of Natural Gas Engineering, MacGraw hlill (1959).

24. Miclselsen, Ml,: Phase Equmilihriuunnr Calculationrs, What is Easy and What is Difficult?, Comrupniners Clmenm.Fngnmg.. I 7 No. 5/6, 431-439 (1993). 25. Cmnmnrphcll, J.M: Gas Conrditiunmninrg ausd Processing, Vol. Oklaluonnmu (1976). 1, Cannpbell Petroleum,

8. Lawal, AS. and Silberherg, IF!: A New Correlations of Vmupour-Liqunid Equnilibriummns Ratios Internally Consistent with Critical Behaviour, SPE 10287, Proc. of 56lIs Ann. Connf.. (Oct., 1981). 9. Hadden, S.T: Convergence Pressure in Hydrocarbon Vapour-Liquid Equilibria, Clmem. Eng. Symp. Series 49. No. 7, 53-66 (1953). 10. Gas Processors Suppliers Association, ed.: SI Engineeriumg Datum Book, Oklahoma (1980). lulsa,

26. Dammesh, A.. Xn.n, D.-H. ammd Todd. AC: A Grouping Method to Optimise Oil Description for Conrposilional Siniunlations of Gas-Injection Processes, SPE Res. Eng., 343348 (1992), 27. Sage, 3.11. amrd Lacey, W.N: Pirase-Equilibria us Hydrocarbon Systems. Ind. Eng.
(tsemsr., Vol. 34, No. 12. 1526-1531 (Dec. 1942).

3.4 EXERCISES 3.1. Prove Eq.(3.35)

tine vuupotnr uunnd Iiqu i,I plummscs of ur hitrmury nrsixttmre are at equmilibriuns at T=300 K and 12 MImm. I Inc nnnolc lrmmcl ionr o I cm nnrnlrtsnucnnl I mnr line equmiI ibrumtcd gas ammd liquid piruises is 0.9990 munsd 0.2500, rcspectivclv. Ihc connn~nrcssihiIityfmnctor of comrrponcust I as a pure gas at 300 K cmmn be related to its pressure as, 3.2.

I I - MacDonald, R.C: Reservoir Simn.ilationm wills Interplmunse Mass Inmunskr. Report No. UT 71-2, The University ofTexas at Austin (1971). 12. Standing, MB: Volumetric umnrd Phase Belrmrviour of Oil Field llydrocmnmhons Systemsss.

9th Printing, SPE, Dmmllas. Texas (1981).

13. Mathews, TA., Roland, C.H., and Katz, D.L: High Pressure Gas Measurennient, Proc. NGAA,41 (1942). 14. Rowe, AM:lntemally Consistent Correlations for Predicting Phase comrrpositions for Use in Reservoir Compositional Simulators, SPE Paper 7475, Pres. at the 53rd Ann. Fall Tech. Conf and Exh. of the Society of Petroleum Engineers. l!ounston, Texas, Oct. 1-3 (1978). IS. Ahmed, T: Hydrocarbon Phase Behaviour, Gulf Publishing Courspany, llounston (1989). 16. Standing M.B: A Set of Equations for Computing Equilibrium Ratios of a Crude Oil/Natural Gas System at Pressures Below 1000 Psia, JPT, 1193-1195 (Sept., 1979). 17. Katz, DL, and Hachmuth, D.H: Vaporisation Equilibrium Constants mm a Crude OilNatural Gas System, I & EC,29, 1072-1077 (193.7).

2 2 Z=l5.45xI0 P6.35xl0~P where P is in MPa. Calculmste tlre fugacity coefficient of component I in the liqunid phase. 3.3. A mnnixture of C, mind nC is fimnslscd at 423.1 K amnd 7.093 MPa. Use Raoults law, tIme Stamsding metlnod. tIre Wilson 4equation and the GPA chart to predict the equilibrium ratio of muretlnmmnre aurd nornnruul clecanme at tIre umbove condutmons. TIme measured mole fraction of methane in the equilibrated gums and liquid ptrases is t).949 amsd 0.229, respectively. [Kohn, J.P., Bradish, W.F., J. Cheiss. Emsg. D:uta, 9, 5, (1964)). 3.4. The measured couurposition of equnilibrated vaporur and liquid phases at 10.45 MPa and 324.8 K are given in tIre following table.
Con~pomncnu,nnotc%Gas~qwd C02 7(1.98 5(1.15 N2 0.59 (113 (I 18,18 726 (2 5.16 39) 245 378 (4 (1.31 072 nC4 I) 88 2.41 CS (523 059 ,,C5 ((32 5 St (6 ((.31 271 (7+ (5.59 26.69 (las conrp. Iacu,mr 0.578 Lig. dcnsiny. kglm 765.9 C7+ Ctuaracuerisnics, M~208. S=0.8382

18. Hoffmann, A.E.. Crump, iS. and Hocott, C.R: Equilibrium Constants for GasCondensate Systems, Trans. AIME, 198, 1-10 (1953). 19. Glaso, OS. and Whitson, Cli: The Accumruucy of PVT Paramsieters (.mulcuuluuted fronrm Computer Flash Separation at Pressures Less Tlsumn 1000 Psia, JP~F 1811-1813 (Aug., 1982).
20. Wilson, G: A Modified Redlmch-Kwong EOS, Application to Getreral Plsysicmd Duuma Calculations, Paper No 15C, presensted at tlre AtClrE 65tIm Nationrmul Meelinsg (Mmuy. 1968). 21. Michelsen, ML: Calculation of Phase Envelopes and Critical Points for Mimlticotnsponent Mixtures, J. Fluid Phase Equilibria, 4, 1-10 (1980). 22. Whitson, C.H. and Toup, SB: Evaluating Constant Volume Depletions Data, SPE 10067, Proc. of 56th Ann. Conf. (Oct., 1981). 23. Voratsis, N: A Robust Prediction Method for Rapid Phase Behaviour Calculations, SPE 16943, Proc. of 62nd Ann. Conf. (Sept., 1987).

Predict the equilihriuum ratios at tine above conditions using GPA charts, the Standing method annd the Wilson equations. Counpare the results with measured values.

128

129

3.5.

An equi-molar binary gas mixture composed of etlrane and C02 is us equilihriumss witln water at 350 K and 30 MPa. Assume (he gas fugacity coefficients equal to one, and use Henrys law to estimate the gas solubility in water.

4 EQUATIONS OF STATE
TIme equuality of funguncily of emiclu conrnpomnemst tlsroughount all pisases was proved, in Chapter 3. to imc tIre requireunsemsl for u,Incmsuical equilibrium ium mntulticonsspomrensl systems. lIre fumgurcity coefficiersi, 4r , defiured mis tlsc rumtio of fumgacity to pressure, of each component in umtly pisase is 1 related to pressure, temperature ammd volutne by Eq.(3.3 I),

ln4.

.j

.~

Ri V l,,,~)n,)~vnj.,

RT/V dVlnmZ

i=1,2,...N

(3.31)

The fumgacity coefficient can, therefore, be determined from the above with the aid of an equations relatimmg pressure, temperature, volume and conspositions, that is, an equation of state (EOS). Ins genermsl, any equations of state which provides reliable volumetric data over the full range of the integral in Eq.(3.3l) can be unsed to describe the fluid phase behaviour. Several types of EOS have been suuccessfully applied to hydrocarbon reservoir fluids.

130

4 Iqs~i(icns of .Sh,(e

4, 5. Virial EUS and ill Modmfleotio?u.c

131

The simplest, and higirly successful equation, us tIme semi-enrpiricuul vans dcr Wmrmuls type EOS
with two or three parameters. Since 1873, when vats der Wauuls iussproved tine ideal gas equal ion by including parameters that represented the attraclive anrd repulsive inslermnnolecun lumr forces, the equation has beets revised and nnodified by numimerous investigmmtors. Otlncr equations with many parameters have also been tused to describe (Ire plrurse heirmnviotmr, sotrse with reasonable success, Amongst tlsese equations, tire Bencdnct-Wehh-Runhins (BWR) typeE I], which is an empirical extension to tIme virial EOS, can he uupplicd to boIls liqumid amsd vapour phases of reservoir fluids, These equnalions provide no additionsal reliuuhility its plrmmse hetsmmviour studies, its spite of their conmplexity. in conruparison wulir tire vmrnr der Wmmals type EOS. They are, however, valuable tools to describe volu,mrretric hclu:mviotmr. pmurnicumlmurly of pumrc conmpounds, due to their large numrnher of paransseters, hence, lmngim llexnbiliny. Fqunationc of state mmrc hasicmully dcveloped for pore curnrnl)ourcnsl.s. 1mm :mppliud to nrruulnmcmsnusponenl sysnemnrs by ennnploying sonrmc nnnixnnrg mules to mlctcrnnuinre tlucir pmmnanmuttcnc for nmixlumres. lIne mnsixinng runlcs mrre comnsiulcmid 10 dn.scmitrc tIme putvmnilinmg lonucs liuIwcumr musolecules of duflerctst sunhstances formsning tire mixtumme. Sinsuple nlumxinsg muiles, stmclu mis tisose Ilsal unssummne comnpounuls are ruumsdonrrly distributed wntlsun tIre mnnixtuure. mmmc (lumIe mrdcquumnle Ru descrnbc irydrocarbomm mrmixtures of reservurir flunds. More conmiplex msnnxnnmg rtmles. lsowcver. urre required to represent the iruteractuon between lmydrocmmrhons mmnd musynmrnnclric conmspoumusuls suuclr as water, which is present in reservoirs, or nmetlmanmol which is sonmetumnres mudded to rcscrvsir fluids as a lsydrate inlsihitor. Al tlmouuglm tlsermodynaimsics ngoroumsly describe tine equilihriuunss conmdnl ours ummrd relmulc tluennm to volumtrrctric data, ums given by Eq.(3.3 I), it is tIre cmmpurhility of EOS mmmsd tIre mmssocimutcd mmsixinng rules Ilnat determines tine success of phase equuilihriunsr prediction. mms will be described ins lbs
clrunptcr.

valtne to reservoir fluid studies, wirere a single equations of state is to describe the behaviour of both vapour and liqunid pisases. It is, however, very useful in showing guidelines for applying setni-etsipirical EOS to mnmixtuures, as the virial coefficients can be described rigorously for mrmixtures using statisticuml nrsechmumnics. Ilmis subject is described in Section 4.3.

Starling Modification of Benmedict-Webb-Rubin EOS (BWRS)

lIme l3cncdicn .WehhRtnhinm EOS (E1WR) [I) is an empirical extemssion of the virial EOS. A nsrm-,dificumliomr of lIme Benedict-\Vclrb-Rubin EOS as proposed by Starling [3) witis II pmrrmmmmseters Imuts heeur mupplied successfunlly to petroleuumn reservoir fluids,

51

RT

(lt,l~l.- A,,

)p~,+ (lI~I.-. mm

u(a

(4.3) +~s!.(I yp~)exp(yp~,) 2 T wlucre PM is tIre molmmr dctrsity uunrd tIne II coefficients can he evaluated from the following genre rmnl i scd equnumt n ours: p~U,,
= 0.443690 +

0.115449o

1.28438

0.920731(n)

R1~. 4.1 VIRIAL EOS AND ITS MODIFICATIONS

=

The virial equation is based on theories of statistical mssechmmnnics )2), anrd canm he expressed ums
an infinite series of either molar volume (molar density), or pressure,

RT,
=

0. 356306

1.7087 ho

Z= I

B/v

C/v

Div

(4.1)

0.0307452+0.179433o

(Z=l+BpM+Cp~+Dp~,+ or, 2 3 Z=l+BP+CP +DP +.... (4.2)

t).006450 0.022143(m) cxp(3.8(0) p~h = 0.528629

=

0.349261(i) 0.754130o

where Z is the compressibility factor, v and p are the molar volmunne and tIme nusolar density. respectively, and P is tIne pressure. B, C, D. etc., are called tine secommd, tlsird, fourth, and so on, virial coefficients, and depend only on temperatunre for eaclm compoumnd. The coelficjenst B muccounnuls for tIne itrteractiomr between Iwo mmmolecrules. svlncnemms ( is tluan for tirree nrrolecules and so onm. For exausmplc, if tIre effect of a tlmird ummolecule (run tIme prevailing forces between two molecules can be ignored, tlse third and Imigiser terms cuun he neglected. As the fluid becomes nmore dense, the higher terms become more significant muumd cannot he ignored. The equation reduces to Z=l, that is tIne ideal gas equation, wlsen pressumme approaches 7,ero. Nurmrerous theoretical and experimental studies on deterimiinatioms of virial cocfficiemsts, mostly the second virial coefficient, have been reported. As higlr order coefficieusts arc hard to determine, the equation can be applied to the vapour phase only. It is, therefore, of little RI,.

0.484011

11.0732828+ 0.463492oo (1.0705233 - ().044448m

0.504087
+

Rl

1.322450)

0.544979 -0.270896(0

132

4. Equations ofState

4.2. Cubic Equations

of State

133

where T~, p~and Co are the critical temperature, critical molar density and acentric factor, respectively. The above critical properties, including v~=( I/PM~)~ for pure compounds are given in Table A. I in Appendix A. The application of BWR type equations demuunds a higlm counpulaliomsal tunic ummrd effort, due to their high powers in volume and large number of parameters, hence, unsuitable for reservoir fluid studies where many sequential equilibrium calculations are required. More importantly, for multicomponent systems each parameter must be determined using a mixing rule, which at best is quite arbitrary. The choice of mixing rules often has a more pronounced effect on the predicted results, than EOS itself. Although acceptumble phase behaviour results can be obtained by BWRS [41, it has been surpassed by the siusspler, yet more reliable van der Waals type cubic equation of state.

corresponds to that of liquid. The predicted volume within the two phase conditions, v2 (or Z~), is of no physical significance.

T2
T=Tc

Single Phase

4.2

CUBIC EQUATIONS OF STATE

0 .3

\Critical Point \\
\

van der Waals improved the ideal gas equation by considering lIne inlermsrolectulusr attractive and repulsive forces, and introduced his well-known equmution of stmmte irs 1873.

N
N

0~

(P+-)(vb)=RT

(4.4)

P1

N
N, N N,

where a,V2 and b represent (he attractive and repulsive ternns respectively, and v is tIme molar volume. As the pressure approaches infinity, the molar volume becotnes equal to h. Heumce. h is mulso considered as an apparent volume of the molecules and called co-volumsse. It sisoumld he always less than the molar volume v. The above equation in terms of volume or compressibility factor takes mm cubic forums as follows:

N N

RT ~+()v=0 a ab v (b+)v P P P or 1 1 Z (I+B)Z +AZ-AB=O where the dimensionless parameters A and B are defined as, A aP 2 (RT)

vi

52

vi

(4.5)

Vortunne

>

(4.6)

m I iguune 4.1. Voluimmctric tsclnmtviotir of l~ coisspounmd as predicted by cubic FOS of vumms der Wusals type.

The predicted maximnunin and minimunrn volumes within the two-phase region, however,
(4.7) immdicate time pressure limits within svlnich the fluid can be compressed or expanded whilst it remnmains a mclastahle simrgle phrase fluid. This behaviour will he described furtlmer in Section 5.2. TIre difference hetweenm lIme limmnits reduces as the tensperature increases mmd vanislses at the critical poimst. At a tensmperature above tIme critical point, e.g.. at T on the Figure, tIme 2 equal ions provides only one plmysically acceptable roust: For the critical isotherm, a horizotntusl inflectioms pniumt slson.nld, tlrerefore, exist at the critical point,

Bsu-~-

RT

(4.8)

Hence, van der Waals type EOS are often referred to as cubic EOS. A typical volumetric behaviour of EOS of van der Waals type is shown in Figure 4.1. For a pure compound at temperatuures below the critical temperature, e.g. F , tIne equation 1 may give three real roots for volume (or Z) at pressure P as simown in Figure 4.1. TIme 1 highest value, v I (or Z ) corresponds to that of vapour, whereas the lowest value, v~(or Z ) 1 3

(~ =1~~!l=0

(4.9)

Applying tIme umbove requmirennuents to time van der Wamuls equation, tIme values of a atsd h are deterunniured mis:

134 9 2 2 27 ( R T ~

4 Equations ofState

4.2. (uhic Eqimation.c of Statr

135

RT~v,=

(4.10)

Suuhmsnimuitimsg the 2 umbove vunluics inn Eq (4.6), resmults inn the following cunhic equation for Z, /.-I.90257 Z +I .55625 Z- 1.40463=0 willr smu ly one reuml roust

I IRT h=~-v =_(..._u~)

. where the suuhscrmpt c, refers to the values at tIme crnlncal poInt.

(4.11)

Z=l.49069 Snuhstitmnnimrg nIne above vumlnic of tIme comnupressihilimy factnsr in the fougurcity expression results in.
4n=Q97779

The van der Waals equatuon of state (vdW) gives a cmulncunl comrrpressnhulnty fmmctur of 0.375 for all conspounds, whereas very few connpounds, sunch mis quummntuurr gmnscs, lummve A grcmutcm Itnmnmm 0.3.
Example 4.1.

I lcnscc tIre c,sirccnsurmnniunmr of dissu,Ivcd nscmtn:nnre inn wmnter. corrected for ilre fugacity coefficient,
us.

In Example 3.3. the solumbility of methane in water was calcumlaled by asstnminrg tIne n,mctlrmmnre fugacity coefficient equal to one. Use vdW to eslinmate tIne fugacity csefficienrt amrd imnmprove - the accuracy of predicted gas solubmhmty.
Solution:

f,Y1 =Pxn*=65 x(I.97779

x

Ml~m= (I .4378x10 Mta/mnnol fraction) xx,:

=4.42x 10

nuuolc frunctnonr of urrethamre inn water.

Substituting the pressure in the fugacity expressionr for pure compounds, F.q.(3.35), onsing vdW, we oblain. ltrm~= (Z I) In Z + I ~( RF RT v P)dv
~

J~

A simple equnation, such as vdW, eannmot accurately model tine behaviour of dense fluids particumlarly Ilmat of cotrsplex fluid mixtures. Nunmerous modifications have been made to imnmprove ifs capahilily by flso(lifying the attractive amnd repumlsive terms. The two parameters of mm mrmsd h us tire origimsal vdW c:nn ire determsnitred simply from the boundary conditions at lIre criticmml poinmt, wlrercmus in nrnosdificd versions, mndditional to the above, experimental data on punre flunids Irurve umlso hecir used gennermully to deterissine tIre parameters. These equations are, Ilnercfone, setnni-emrrpnmicmul. hluurd splnere fluid usmordels Irunve hcemn selected [5) to describe repulsive forces. TIne equation proposed by Cartrmrlsmmn anrd Starlinmg [6] imas been used extensively to develop new forms such mis perturbed Isard clsmmin [7], and cinumims of rotators [8). In spite of recent efforts [9] to simplify tine earlier modifications in order to mnrake them more practical, they have not received muclm unllenmtusn as emmgnncernng tools. Although vann der Wmumls considered that Iris representation of repulsive forces, expressed by a constant b. required mmnore imniprovement than that of the attractive term, in practice the mnmodification of tIre latler ismus beers nusorc rewardimng. Almost all popular van der Waals type EDS lrmrve improved tlmeir cumpahilities by modifying the attractive term. They can be expressed by tIme following gemrermml form,

(Z

I)

In Z +

_Lj(.~I ~--~-- -~--~dv 2 RI ~. v vb v )

+

Integration of the above equation results in, lno~ = (Z 1) In Z +

_!_IRT In ~ RTL vb

vj,.,

lrnpienrentnng the Inmnmts, and making the equanmon dmnremssmonlcss, using Eqs (4.78), we o lain, ln~ =(Z I) In(Z B) A/Z The parameters of vdW are calculated, using Eqs.(4.9- It)) amrd mncthane critical properties given in Table A.l in Appendix A. as. 2 27 (R T2 ~ a = [ s.- j=(27/64)x(0.0083t44xl90.56)/4.599=0.230274 Mlmn.(m/kgnnoI)~ 64 ~ P~ j b I RT (-~-)=(l/8)x 0.0083 I44x 190.56/4.599=0.304254 m/kgmol 8 ~ .

~.~--. ~--

(4.12)

vli v -I- nuv w Inn mu two-pmsrumiimeter fosrnsm of mIsc equmulion in and w are related to h whereas in a tlmree-p~nranneterforum on, and w are related to) h, and/or a third parameter c, In a fourparumrureter nrodificalion u and w more related to b and/or c and a fourth parameter d. FIme above general eqlnmmtion in terms of tIre conspressihility factor is, 2 2 2 2 z(l+lt)z +(AB~JtJw )z(ABr3w w )=o

(4.13)

with the dimensionless values, defined in Eq.(4.7-8). as follows, A=I.55625I B=0.902574

where the dumnmensmomsless paranneters A and B are the same as those defined in Eqs.(4.7) and (4.8), respectively, anrd

136

4. F?qmuatuomn.r of State

4.2. ( nmbic &/mmati.n?i.u (nl .St~ir

137

RT RI

(4.14)
(4.15)

(4.20)

Witlm tIme exception of Redlicln mmmxl Kwong [10) who originally proposed the temperature 5 depensdency of a in 1948 as u=Tr recent investigators have uised vapour pressure data to

The two-parameter EOS are the most popular equations, where the parameters mure expressed by, R T 2
c

tleternmmine a
Fine above discussion lmighhiglrts tire siurnilarity between all tire comssmmionly umsed equations of
state and suggests tlmat tlsere is very little fundamental difference between them.

(4.16)

RT
b
=
~

(4 17)

All the lwo-paransretcr equmations husve selected a fomm of Eq.(4. 12), assumed, a, to he tennperatnrc dupendemmt, ammd lmmmve detcrnmntned lime tenmperalurc dependency by matching tIme vapotir pressuire dmmta. lIme mudditioms of tIre Ilmird parammmeter hums increased the flexibility, where tIme inrvcstigmutosrs Imurve dctermmnined it gcnmemmmlly by rmratciring saturusled liquid data, Fhe van der oWmnuuls type E()S Immuve bccmm cosinnpuclucnrsively reviewed in literature 111161.
Euannnple 4.2.

Note that the expressions for the parameters ins tine mnodificd equumutioums are siunnilmur to llutssc orf 1 the original vdW, but the coefficients lmave been generalised as ~ a and ~ iIme ollmeu parameters, in EOS which use more than two, are generally of co-volume usunture. Imcmmcc, expressed by an equation similar to Eq.(4. 17), hut with different coefficients, The substitution of Eq.(4.l2) into the expression for fugacity of a pure suuhsluunce, Eq.(3.35). results in the following generalised expression, using the same approach as in Exanmple 4.1, lnmi~ =(Z

Prove Ihat the mureas hetweemm tine saturation pressure lure and the predicted volunme isotherm Imy a cmnhic equations of state mrre equal for a pure substance. The above equality, known as the Mmrswell equmml murea nile, is counsidered equivalent to equality offugacities of saturated vapour mmmd luqumud ilsmuses. lnemrce, mupplmcmmble in determmsimming tIme parameters of an empirical POS.

I) ln(Z B)+
,1U2

I 2Z,t,U_,[U2+4W2 n

i~i-j3-:~: Jii~4W2

(4.18)

There is hardly any theoretical foundation, or strongly convincing arguinments. for selec(iisg a
particular form of EOS amongst nmany described by the general form of Eq.(4.l2). Time success and popularity of certain equations are nsore due to) featmures other than tine selected have a higher impact on the predicted results, than the mathematical form of the equation. Although EOS primarily provides volumetric (density) data, its msuijor contrihuition as an engineering tool is through its coumpling with thermodynamic relations in predicting phunse belmaviour and physical properties of fluids. As tIne parmmmrmeters of mm sermmi-cmnnpiricmul BOS mire determined by matching its prediction to experinmeumtal data, time inchusion of nmmore pmmrameters in EOS makes it more flexible. Whereas a two-parameter EOS would suffice to predict tIre vapour pressure, hence, phase equilibria, the inclusion of a tlmird paratmmcter will gemsermnlly

form in most cases. For example, the method used to determine tIme EOS putraimmeters could
3.)

nvn.
S

.~:::::::::::::::.~ t:::::::::::::~:s

vv

improve the prediction of density along witls reliable vapour pressure. Two (Sr tlsrec
parameter EOS are considered adequate for all applications in the pctroleuums itmduistry. The main departure from the original vdW, which has resulted in the success of tniodified equations, is not the revision of the attractive termsm functional forums, hut treating it as mm

Volume

temperature dependent parameter, a=a,a (4.19)

Figure E4.2. Pressurevoluunse isottsermrs of a pure fluid as predicted by cubic EOS.

Solumt,o,n:

where a expresses the dependency of the parameter, a, on temperature and ac depends only on the critical properties of the compound as given by,

Integrating Eq.(3.22) from the satnmrated liquid to tire saturated vapour, we obtain, Jd~=JsdT+JvdP=~
~L ~

138 because the vapour and liquid are urn equilibrium.

4. Eqomotuonms o~f State

4.2. Cubic Equations ofState

139

The integral over dT is equal

J

0) 50 to zero at the isotlmerm. ttemsce. (0.45

Acenrnric Facnor

vdP = Jd(Pv)_

Pdv = 0
(1.40

or,

tt

P(Vv _VL)_JPdv=0
Which is only satisfied when the two shaded areas mmre eqummrl.

0.35 a. no Ii 25
0)201

4.2.1 Two-Parameter EOS

Redlich and Kwotmg [10) modified the attractive tert3m of vdW as,

P=RT/(vb)a/[T~v(v+h)] and 0.08664 respectively, using Eq.(4.9).

(4.21)

015 000)

00.2

04

moo

os

1.0

The values of ~a and ~b were considered to he constatnt, imence, detcrnsnimmed to he 0.42747

Reduced remi,erauure

Zudkevitch and Joffe [17), and Joffe et al. [IS] assumed tlmat ~a and ~b in Ilse RedlichKwong equation of state (RK) were temperatumre-dependent. lIme values of ~a and ~h for each pure substance at any temperature were obtained by matching the predicted liquid density data to the measured value, and equalisation of saturated liquid and vapoumr phase fugacities. Above the critical temperature, the parameters were taken as constants and eqummtl to their values at Fr = I. The above approach was necessary as two equilibrated phases cannot exist for a pure compound above the critical poinmt. Figure 4. 2 shows lIme varimmtion of ~a and ~ with Tr as derived and correlated with the acentric factor by Yuurhorougim [19). rhe sharp change of both parameters near tIme critical point clearly rnnled ounl the extrapolation of the relalion above the critical point, an approach which is generally used in otlner EQS. Manchming of the predicted data to mnmeasured vmmluues at saturmut ion to (Ictcrnmsinse E()S pmnr:mismelers has been used almost by all recent investigators. Wlnereas Zumdkevitch-Jotfe suggested determining the parameters when required at tire prevailing conditions, others Isurve generated generalised correlations for parameters by applying tire method. Iime use of correlations to determine the parameters definitely simplifies the calculation task, hut reduces the accuracy as any generalised correlation is bound to have some deviations fronri line cmrrelated duntmu. The mmiellmod suggested by Zudkevitch and Joffe does not significanmtly inncreases lIme calcumlational effort when applied to a petroletmm reservoir where the tennupermntulre is treated mostly constant, and tIme parameters are calcomlated only once. Ansy reliable correlation in literature [20), suclm as those of Lee-Kestler, Eq.(l.lO), and modified Rackent, Eq.(l.I2), can be used instead of experimental data to calculate vapour pressure mmmd saturated liqumid density respectively.

Figure 4.2a. (Iemrermulused values of U in ZJRK. Reprinned with permission [191. Copyright(1979) 5 Ammuerucan Chemnrical Srcicuy.

00, tO Accnnr;c Iacnor 0,09 006 0) 07 oh

(00)1,

005 (0(0 01(01 002 0.0

0)2

0.4

0.6

0.8

1.0

Reduced Temperamnice

The approaches of Rediich-Kwong, and Zudkevitch-Joffe to imnrprove vdW, that is nnaking

the parameters of EOS tenrperature dependent and using the smsturated data to determmsine time parameters, have been adopted in all the successful modifications of vdW. A few of the cunrrenntly prevalent EQS will he presented here. Figunre 4.2h. Generahised values of Ub in ZJRK. Reprinted winlm perummission 1191, Copyrighu (!979) Anncrncan IIncunuicat Sotieuy

140

4. Eqmmaticn,u.c of State

4.2. Cubic Equation.s of State

141

Example 4.3. A class of equations relating volumetric properties to temperature and pressure, is that based on the corresponding states principle, which considers that fluids behave identicmmlly at conditions of equal reduced properties. Reduce the Redliclm-Kwomng EOS to a corresponding states form of ZZ(Tr, Pr) Compare the result for T=l .5, over a P range of 0.5-3, with that 1 ofthe generalised compressibility chart shown in Figure 2.22.
Solution:

Grahoski and Daubert [23] used the API vapour pressure data and modified Eq.(4.24) to improve pure-conspomsent vapour pressure predictions, m=0,48508 +1.55171 to- 0.15613 ~2 SRK in terms of time compressibility factor Ztakes the following form.
3 2 2 Z -L +(A-B-B )Z-AB=O

(4.25)

(4.26)

The Redlich-Kwong EOS in terms of the conmpressihility factor is tIme smmuime as Fq.(4.26). wiiti the following expressions for A and B according to Fs s.(4.7) and (4.8), respect ivtIy. 2 5 2 2 71 A = aP/(RT) =(0.42747T,. R T~ / P~)I/ R i 0.427471, ~I~ B = bP/ RT
=

wlmcre tIme definitions of A umusd B umre given ins Eqs.(4.7) umnd (4.8) respectively.
(oil mpuurm ung SRK witIi I luu gemrcmuml I ~()S , I ~q.(4. 12). I Inc val ties of tm=h mind w=0.

SRK is ulummtc capable of predictimug vapoumrIiquid equilibrium, hut it does not provide reliable (0.08664RT~ I P,~ )P/ RT = O.O8664T,~mP~ liquid density. Ieng-Robin.so,u EQS (PR) Peng and Robinson 1241 modified tine attractive term mainly to improve the prediction of liquid density in comparison witir SRK, P=RT/(v -b)-a~u/[v(v+h)+b(v-b)J where, 2 ac = 0.457235 R T~2 /p and Soave-Redlich-Kwong EQS (SRK) Soave [21] replaced the temperature dependency of the attractive termmm un RK, TrMS, by mm more general function a: P=RT/(v-b)-aca/[v(v4b)] where
=

Substituting the above two expressions in Eq.(4.24), results in,

5 [(o.42747~ / F~ 0.08664P,

IT,

(0.08664P, / I, )2 ~Z 0.037036I~2 / T~=0

Substituting T,=l.5 and various P, values in the above equation results in a ctmhic equations, with the results as follows. The comparison with the values of Zread from Figure 2.22 is also shown. P, Z, RK Z, Fig. 2.22 0.5 0.952 0.950 1.0
0.907

(4.27)

0.903

2.0 0.834 0.823

3.0 0.798 0.778

T 15 b = 0.077796 R c / c They used a sirniiuur form

of a as proposed by Soave, Eq.(4.23), but used vapour-pressure

data Irons the nornmmal boiling point to the critical point, and correlated m as, ins
=

(4,22) 0.37464
+

1.5422o - 0.26992u

(4.28)

2 2 0.42747 R T~ / P~

The correlation was later modified to improve predictions for heavier components [25),
in
=

0.3796

1.4850) - 0.1 fl44~2 + 0.01 667&

(4.29)

b =0.08664 R T~ / P~ and

PR in terms of the commspressibility factor Z takes the following form,

-

(l-B)Z

2 2 3 (A -2B -3B ) Z - (AB-B -B )

(4.30)

a=[1 +m(1

5 2 Tr )1

(4.23)

PR is obtained by substituting u and w in Eq.(4.12) by 2b and h, respectively. Volunne S/u uft A comparison of time predicted liqunid molar volume by leading two parameter EOS with experimmmemmtnui data of pumre consspounds generally slsows a systematic deviation. The deviation is almost constant over a wide pressure range away from the critical point. Hence, subtracting the predicted molar voslume by a constant correction term can improve the

The function a was selected, and m was correlated with the acentric fmmctor by equating fugacities of saturated liquid and vapour phases at Tr =0.7. m=0.480+ I .574o-0. 1 76& Soave et al. [22], later suggested to divide tIme value of m determined from equation by 1.18 to improve the results. (4.24) the above

142

4. irquaoonn.c ofState

4.2. Cuuhin- Equation.c ofState Jlmumveri and Yonnngren

143 sinsril:mrly to Penelounx et al., applied the volume shift concept to

predicted liquid density. The effect on the predicted vapour volume is generuslly insugnificant due to its large value relative to that of liquid away from the criticmul point. Peneloux et aI. [26] were the first who iumtrodumced the voltunrie slunfl comnrcepl. u. slnnftunmg tIme volume axis, and applied it to SRK, vcor=v_c (4.31)

1271,

iR, munrd relumted c to lIne paraumseter Is. by defining ml dimensionless shift parameter, S . 5 S t~h (4.35) 5 was determined by matching tIne predicted and nmmeasured molar volumes for various hydrocarbons. The shift parammseters for light compounds are given in Table 4.1, Table 4.1. Vmulues of slsift parummmneter inn Peng-Rohinsomm equmation of state. cnnrnnpunnenn Ci C2 (3 IC nC4 4 S~ .01540) ~0.IO02 ~0O65Ol -0.07935 -0.06413

where vCor is the corrected molar volume, and c is the correction terums delermuned by matching the measured and predicted saturated liquid volumes at Tr = 0.7. EOS are applied to multicomponent mixtn.nres by introducing fluxing runles to determniime mixture parameters, as will be described in Section 4.5. The following mnxmnmg rule is used to determine c for mixtures: c
=

iC5 nC~ -0.04350 -0.04183

_________
-0.01478 (4.36)

lIne unutlsors correluutcd tine slnift pmrrunnrieter to the molecumlar weiglst as, Sn.
=

~x,c,

(4.32)

where, x , is the mole fraction of component, m, in the nrmixturc. 1 The inclusion of the third parameter in EOS changes the calculated fiugacity coefficient using Eq.(3.3I). as,
ncor =

wlmere s~i and Table 4.2.

are positive coefficients. Stmggested values for the coefficients are given in

imthlc 4.2. Cocfficienmts of slnift parunmnsctcr correlation, Eq.(4.36). Connuponncnmmjy,pe . NL_. ..

larmuftins Nmn uhotnenscs 1 Atonnnummtics 2 258 3 0)004

~uexp

(-

c P / RT) 1

(4.33)

where ojs~Cunr and ~i are the nnmodified and original fugacity coefficiensts of componcimt n. respectively. Wimen the fugacity of each conmpomment is calculated by lIne sammue EOS ins both vapouir and liqunid ptmases, the above modification will not affect the predicted equmlibriuni conditmons. It only multiplies the fugacity of each component in both phases by an equal ansmoumnt. resulting in the same value of equilibrium ratio, Eq.(3.43). Hence, the thmrd parameter can he emnployed merely to adjuns( the predicted vohumme. and need not he mnciunded in EOS for calculating the vapour-liquid equilihriumsm ratio. Time volumsse shift generally improves the predicled liquid density, and lrums a mninnimnmal effect on the vapour density at low and immoderate pressunres as its molar volummie us relatuvely large compared to the value of c. At high pressure condition, the inclusion of c puirmuimneter may nsot necessarily improve the predicted gas density as it is just a correction ternn fosr the liqumid density. However, it is advisable to adjust tIme gas phase volume by the third parameter to maintain consistency, particularly near the critical point where properties of the two phases approach each other. Penelounx et al. correlated the volumne translatnon parammieter c as,
=

2.516

0.162.) u 2321 0.2(6)8

Tire value of for lmemmvy fractiomrs of a reservoir flunid can be used as a tuning parameter to murtcts tIne predicted to mmmeasumred satumraled liquid densities, as will be described in Section 9.3. Matlsias et al. 1281 pointed out that umpphication of the above method to PR raises the calcuulumted liqomid voluusnc above lime cxpcrinssemital value for almost all tested pure compounds mmbosvc mm reduced tensrpcrumturc of umrotund 0.85, and tire uleviatioms reaches its mnaximum at the critncmml point. Therefore, an additionnutl voluusne correction term scaled according to proximity of the prevailinsg conditions to tine criticuml point. was proposed, v c (437)

0.40768(0.2944 I

ZRA) R.T~

(434)

wimere ~3, is the volume correction, additional to c, to match the critical volume by EOS, deternsmined using experinmmental data, and ~. is a dimensionless distance to the critical point. ~, is a cusnslmnmml, determinmed by regrcssionr of saturated vapour and liquid volumes, and found to he 0.41 for PR. lIne dimnren.sioirless distance, A, was related to the slope of the prcssuure-denrsmty isotlrerisu as,

A
wiscre ZRA is the Rmmckett compressibility fuuctor mrs develolsed by S memrcer mnnnul t)mnmsner ins tIme 1 modified Rackett equation, Eq.(t.I2). 3 v =(RTC/PC)ZRA (1.12)

(I / RI.) (~)P/nJp)r

(4.38)

A sinmnilmur uuppromncln, to mnnnprovc tIme lnrcdicted demnsity near time critical point by EQS. was also .suuggesled immdepemsdently by Cisoum and Prausnsitz [29]. who applied it to SRK and found a value of 0.35 for ~. The proposed method was extended to binary mixtures, with an additional term related to derivatives of time molar Helmholtz energy. It was, however,

144

4. l~qmuutio:n.sofState

4.2. Cmthic Equations of State Stibstituting the above two valunes of the compressibility factor in the fugacity expression results in, 4,=op=0.729704 m~n~4n=0.746fs 1(1 For a pure compounsd the eqtuality of fugacity reduces to the equality of fugacity coefficient. The comparison of the calculated ftmgacmty coefficiemmts indicates that the assumed pressure is chose to the sumturaliomn pressure. hut requires improvenment. The next pressure nmay be estimsmated as,
~(r+O) =

145

concluded that the near critical contribution did not appreciably affect the restults for imsixtures in most cases. Example 4.4. Calculate the vapour pressure of normal hexamme at 477.6 K using PR. What are tIne predicted values of the saturated vapour, and liquid density? Solution: At thesaturation point, the fugacities of hexane as vapour and liquid shotmld be eqtu~tI.Hence, a pressure is assumed and the fugacities are calculated, using PR. The pressure is iterated until the two calculated fugacities become equal. Substituting u=2b and w=b in the generalised fugacity espressiour for pure conmmpoumnds. Eq.(4.l8), results in. ln4 =(Z I) ln(Z B)+~lnZ+(l ~J~B 2B-f~ Z+(l+..f2)B The parameters of PR are calculated, using Eq.(4.27), and imormuml hexmmume criticuml pusperties, Table A.l in Appendix A, as,
=

[P(~,~V

)j
The

where r is tIme itenaliomn mmuirniser.

TIme aimove mnppromuch restuits un mm pressure equal to 1.9031 MPa, for the next step.

iteruutiomn converges to, 1.9458 MPa 071716

Ihe cstiummalcd valune by lhc 1_ce-Kesler equmulioum, Eq.( 1.10), is 1.936 MPa.

2 2 2 0.457235R Tc 1Pc =0.457235x(0.0083 l44x507.6) /3.025=2.ts92273 MPmm.(mmr/kgmmrol)~

The cubic eqnummtmonr at tine mmt,ove pressure is as follows, b

=

0.077796 RTc/Pc = 0.077796x 0.0083144 x507.6/3.025=0.108539 nm/kgmnol

m Z-0.9468 152 Z +0.2337603 I Z- 0.015562=0

The temperature dependency factor of the attractive term, a, is calculated from Eq.(4.29). and Eq.(4.23), for 0)=0.30 13 at T,=477.6/507.6=0.94089. m
=

with the following roots: Zu~0.60089 Z,=0. 10958 Z,=0.23634 Rejccnnmng the ummlernsnedimute root, amd cmmlcoilalinrg tine mrmolar volume, Eq.( 1.5), we obtain,
v=ZRTIP Vu=0.22362 mmm/kgmnot v=l.22623 nm/kgmol

0.3796 + 1.485w -0.1 f44w

+ 0.01

667&= 0.812562

a=(I +m(I Hence,

05 2 Tr )] 1.049349

a=axa,= 2.825 135 MPa.(m/kgmol)

Fire volnumnme shift fusr unonmmmml Irexanne is cmilctmlmuied, Fq.(4.35), ums, c=S,.ts=.t).0l478xt). It)8539=.0.0()ltu04 mnsIkgmol wtnucin resulls in tIre knlkswinng corrected ummolar volunnes, i~q.(4.3 I), v.=~=0.22523ns/kgnmol TIre demmsmties of tIme suutuurated plmases are:
p=M/v p)826 kg/nun 5 p =70.IR kg/nun

Assuming a saturation pressure (sI 1.86 MPun, musing Figure 1.3 or Fq.( I It)), tIme two dimensionless parameters, defined by Eqs.(4.7-8). are calcimlunted as. A=0.33324353 B=O.0508396

v~=l.2279 m/kgniol

which results in the following cubic equation for Z. Eq.(4.6), 5 2 Z -0.949l604 Z +0.22381034 Z- 0.0142259=0 The above equation has three real roots, Appendix B. Zna:0.62954 Z =0. 10557 Z2 =0.2 1405
3

1Ine unmeasured vmuluues, Figure 1.5, more p=423, and pV= kg/mn. The modified Rackeit equation, Eq.(I.12), predicts a saturmmted liquid density of72 424.3 kg/rn. 4.2.2 lhmee-Parameter EOS

The intermediate root is rejected, and Zn and Z, are assigmred to the vapour amid liquid phase. respectively.

A two-paranmeler EQS predicts tine susmmse critical commipressibility factor, Z~, for all sumbstances, i.e. 0.307 ammd 0.333 by PR mnmsd SRK respectively, whereas Zc varies within a range of 0.2 to

146

4.

Equuatiunru.c unf.Stuuta

4,2. ( shun !qunatnuu,u.c of State

147

0.3 for hydrocarbons. Although the inaccuracy of predicted volummmme mit the critic,sI point, trot necessarily leads to unreliable volumetric data at all conditions, it dcisionstruites tlnc inflexibility of two-parameter EQS for matching both time vapour pressure mmmd volunrme. FIre inclusnon of a third parameter relaxes the above liminmmtismi. fIre tlnird parumumuetcr is gcnrerunlly determined by employing volumetric data.
Sc)nnmidt- We,izel EQS (SW)

wlrere mi is lIne critucal comsnpressibilily fumctor, as predicted by Eq.(4.39), and is related to the correlating paruuusletcr q. by ii
=

I / 13 (I

q(o)[

(4.42)

mmmd q. defnmmed as b/vt, is tIne sminallest positive root (sf lire following equation, 3 2 (6m+ l)q +3q +3qI =t) (4.43) Figure 4.3 shows the deviation of liquid density at T =0.7 predicted by SRK mind PR for mm 1 reliable for suuhstanrccs witlm smnmmll number of pure substances. Note that SRK is more acentric factors, whereas PR gives reliable data for conmrpounds wills urcentric fmuctosrs umrotnnmd (1/3). Based on the above observation, and amn error analysis of tlue genscruul EQS. Eq.(4.9), Schmidn and Wenzel [30] incorporated the acentric factor as the tlnird pumramnneter in tIre attractive term as, P
=

witlm an approximnsunte value of,

=

0 25989

0.0217(0 + 0.00375(02

RT / (v

b) - a~a / [v

(I

3m) by

3mb ]

(4.39)

Sclrnnnnuit umurd Wcmni.cl selecteol tIme sumnluc fosrurn of a mrs proposed by Soave, Eq.(4.23), html curium. Ian ed, inn. w tIm I Inc mcemrtmc lunclu nr un rd reoltuced Ic unuperatuire by msmatching vuipour pressure (luntun of ~nnrccomsmpnsunmds. urn nnu mm 1 2
=

Substituting acentric factor values of zero and ~/ in the Scbnmsmidt-Wenm7,eI EQS (SW) will 3 reduce it to SRK and PR respectively, where these equations predict lIne liquid demmsity reliably. SW can, therefore, be considered a general form sf SRK urnd PR. 20 o tO SRK
00 0 0 0000

mnnn m,,

0.01429 (STr - 3mm,,

1)2

form form

0.4 0.55

(4.44) (4,45)

no

+ 0.71

(r, - 0 779)2

wlsere, non,
=

t).465 + 1.347w - 0.528os2

for

(0

0.3671

~O

nmi,,=0.536l +0.9593w

form>0.3671
<

I
I IS

mmmd for unmtcrmsncdiumte values of 0.4 mrs =[(0.55 -cm)) / 0.15) inn 1
+

ons <0,55

1(w-O.4)/0.51 m

C
> -~

(4.46)

PR For supercritical consupotunds, cx

=

(t),4774

1.3280)) In T~

(4.47)

-20

00

0,1

02 Acennmnc Fzucnor

03

0)4

I
I

Figure 4.3 The deviation of predicted fromn nseasured liquid demmsity by SRK uund PR [30). The authors used the boundary conditions at the critical point. Eq.(4.9) ts determine use, and b,as, a~ RT,
~,

lIre nmrclumsiots of m in EQS as tIne Ilsird paraunneter by Schmidt and Wenzel resulted in a vumrimmhmlc calculated critical cornurpressihility, according to the value of acentric factor. The predicted values are, inowever, mihout 15% hmigimer than tine true valumes. This was known to the umutlrors, Isut wmrs uuccepted as tIre price for an overall optimum accuracy in predicted volumes, 2 2 Sunhslituting um=(I+3cn))h. and W =3(n)h in tIme generalised EOS, Eq.(4.i2), will reduce it to SW. Patel-Teja EQS (PT) tmmtcl utund lcjui [31[ mnnodilmed tIre attractive ternr by incltndiumg a more flexible third parameter. c, mrs. P
=

2 1~ac = [I
-

51(1

P
~.

qfl

(4.40)

RT / (v - h)

a~a/ (v(v

h) + c (v - h)]

(4.48)

= ~

La,,

~q

(4.41)

where tIne paramrmeter c is defined by,

148

4. Equations of State

4.2. Cuthic Equation.s ofState nu

=

149 for Tr> I

c=Q

(449) cPc
,

an d
=

3~

(4.50)

Time uthove correlations were developed by matching predicted and experimental data without being restricted by time conditions mit the critical point, Eq.(4.9). This appears to he a major step mr igusoring time muctumul behmmviour of puire fluids in favour of an overall improvement imm predicted vaslucs by EQS. It has, isowever, acimieved its objective as demmmonstrated in Section 9.2 wimere time pcrforimmance of various equations are conmpared. The volume shift concept can also he applied to three-parameter EOS. The improvement, if any, will be expected to he minimsnal, as time saturated density data have been generally used to deternmine tine third puiraimmeter. It was poimmted ount, imowever, that the prediction of these C(ltimutim)mss deteriorates near time criticuul point. Timis deficiency can hc corrected using time aupproumch described by Eq.(4.37). hut witIsout including the conventional shift, C.
.

1 is an adjusted critical compressibility factor, determined by nmatcIming the predicted aisd measured saturated liquid densities. It was correlated with tine acentric factor as.
=

0.329032 0.076799w

0.02119470)2

(4.51)

The authors found that the use of true criticunl comnpressihility factor will resuult iii the overuslI loss of accuracy in predicted density, a conclusioum also reacimed by Sclmmsmidt umird Wcni.cI. Applying the condition at the critical point, Eq.(4.9), the other coefficients were derived as,
1 +

v+.

(2 - 351) ~h

2 +

21~ 3 3~ h -1 = 0
.

(4.52)

fire muhove correction cmrmm he mmppiicd to VPT, hut the distaisce parameter A is not zero at the criticuml poimmt, as VPl does mcml smmtisfy time houmndary comsdition of Eq.(4.9). Hence a correction tcrns for time distance pumrammmeter needs to he inclumded [34).
. -

~b is taken as the smallest positive root of the above equation with an approximate value of
=

0.32429~- 0.022005

(4.53)

)
.

and ~ac
=

Time dimensionless distance given by Eq.(4.38) represents the approach to the critical point for pure fluids. Its extension to multi-component systems, however, is questionable. Fuurthernsore in nmost reservoir engimmeering problems, the critical point is approached by coimnpositionuul variations, and not by changes of pressumre and temperature. Hence, all the state points along the path have diffement compositions and different critical points. An alternative dinnucunsionless distance to the critical point, which is as valid for variable composition cases is the relative value of the equilibrium ratios of the mixture components [34]. At the critical point, all K-values are equal to one, hence the proximity to (lie critical stuste cumn he expressed hy. A
=

312

3(1

51)~h+ ~b +(l

3m1)

(4,54)
, .

The Patel-Teja EOS (PT) reduces to PR or SRK by substituiting the value of 0.307, or 0.333 for ii, respectively. Note that these valumes are predicted by the two equations mis lime critical compressibility factors for all substances. Hence, PT can also be considereul a general form of SRK and PR which will reduce to either of themn at their prevailing constant critical compressibility factors. The temperature dependency function of (he attractive term in PT is similar to Ihat proposed by Soave, Eq.(4.23). The authors determined irs, usmng time vapouur pressure data of pure compounds, and correlated it with the acentric factor, as, m
=
.

~
.

(Kt/Ktn)

(4,61)

where Ki, aurd KIm are time equiIibriumsm ratios of the lightest and the lmeaviest components of the mixture at mummy .state noint, Figure 4.4 shows tine insmprovemnent iii predicted density of a binary mixture by including the near critical corrections [34] in VPT. Both definitions of the distance parameter. i.e., the pressure derivatives (PC), and the equnilibrium ratio (KC) have been used. It sinould be noted tlnmml time inclusion of variable density correction term in EOS changes time predicted equilibrium conditiomms, contrary to the constant volume shift, and increases the complexity of the matlsematical expressions of the fugacity coefficient. Hence it is advisable to use it only for adjusting the predicted density by the original EOS, instead of including it mnEOS. 4.2.3
I

0.4524 13

1.30982w - 0.2959370)2

(4.55)

2 Substituting u=b+c and w rrcb in the generaiised EQS. Eq.(4.I2), will reduce it to PT. Valderrama and Cisternas 132] and later Valderrama [33] modified PT by using the critical compressibility factor, Z~, to correlate its parameters, ~ac
=

0.66121

0.76 l05Z~

(4.56)

Attractive Term Temperature Dependency

0.02207 0.57765

o.2o868z~
i.87080Z,,
.

(4,57)
(4.58) 2 for T~ <1 (459)

0.46283

3.58230wZ~ + 8.I9417(wZ~)

As imsdicated by Wilson [35]. a reliable prediction of vapour pressure of pure compounds by amy EQS is a prerequnisite for its reliability in estimating vapour-liquid equilibria of mmsuilticoumnponcnt syslenms. Tlmis lsuus been achieved, alnsost with no exception in all recent mrmodificmutiomns of vdW, by adjumstiusg time relationship between a mind the reduced enmmperature to mmnmrtch tise vunpour pressure data.

150

4.

Equa:no,u.s ofStat,

4.2, Cunbic Equaiionus ofState

151

~i + C,(l

~l~) +C,(l

~)2

+C,(l ~)5]2

(4.63)

where tIre cosefficicmrts C , C , and (~, are to he determined for each compounmmd by matching 1 2 its vump(sur pressure.
0,

a
F.xperunueona) XC ucanung nC sc.uhuug

Somuve [22] reviewed lemm differemmt functions for a, proposed by various investigators, and preferred the original form. as given by Eq.(4.23). lwtn ci url. f 38} poimmlcd out mhmmt hue use of lIre conventional a fuumction, Eq.(4.23), with the slope, on, correlated by power tirrcc of lIre acentric factor, makes a a sixth order function of the mucentric furctor. ilence. the extrapolation of function to heavy compounds with high vumltncs of lIne unccntric factor earn lead to large deviations, particularly at low reduced Ic unnlsem multincs. tIme mount hors cvmml ummntcd a in PR aurd showed that It varied linearly with time utcentric factor mmt constusnt reduced tcnmspcrature. Figure 4.5 shows variations of a at several redsuced temmmperatumrcs as deteruirined for PR by nmmatching the msieasured and predicted vapour pressure of pure comnsposmnds.

0.002

N,,c,,ncuI,o,,

0.000

0.004 0.006 0.008 Murlar Densiiy, gumvrl/cnn3

0.onO

0.0 n 2

Figure 4.4. Effect of near critical volume correction on predicted density of etlrurmmc-propcure mixture at 311 K. lime most direct approacln is tlsuut of Zudkevinclm-Joffc. mis cxplmninrcd inn Sectinnmn 4.2. I. tIne common approach has been, irowever, to imrtrodsnce mm geircrumlnscd cosmnelmmtnomn fnsr a mum conjunction with the proposed equation of state. The most consusson fumnctionmumi form is that proposed originally by Soave, Eq.(4.23). Although the same form has been used by most of the leading EQS. it is not unncommon to find different funirctionuml formirs, or correlations for their coefficients, for the same equation of sturle. Time popularity (sf an EQS nmnmmy he juidgcd even by the number of modifications to its temperature depenmdency fmsctor! The a function attains generality by relating its coefficients to sonnme properties of compounds. The acentric factor is the popular choice. It is reasonable, however, to expect that a generalised equation in terms of the acenmtric factor, sunch as Eqs.(4.25, 29, 52). may not he adequate to describe all conmpoummnds of vastly different characteristics. Iiuc a pmumanssctcr for asymmetric compounds can he related individually to the reduced temumperature, e.g., Eq.(4.88). to improve the predicted results for complex systems. However, a single generalised correlation should be adequate generally to correlate lsydrocarhsn conmmponcnts of reservoir fluids. Correlations as those given by Soave are in general quite reliable to predict tlre vapour pressure of relatively light compounds, particularly at high values of reduced tctrmperatures. To improve their capabilities over wider ranges. higlmer order polymxsmmmials of uucentric fumctor, and reduced temperature have been used. Examples are: m=0.378893+i.4897l53w-0.17131848w +0.0l96554co 2 3 (4.62)

us

aC,

00

0.2

04

0.6

0.8

1.0

Acenriric Facior

Figuire 4.5. Variatinrnms nsf a witlm muccurlric factor in PR at constatrt reduced temperature [38]. Bussed on tIme above observation, the autlmors suggested time following function. a = a
+

co(a a)

(4.64)

whmere a, mmnnd a mire related onmly to tIne reduced temsmperature. Tlmc above approaclm clearly reduces tIre risk of extrapolating the a function to compounds wit Ir Irigim umccnstric factors. rwun et al. [38] correlated a of stmh-critical components for PR as.

as proposed by Stryjek and Vera [36] for PR or Eq.(4.29) proposed by Peng-Robinson later, in preference to their original correlation, to inrprove the muodel for Ineavier comrmpotunds. The following three-coefficient functional form was proposed by Mathias ansd Copeman [37], in preference to that of Soave,

152
atou = T,~~i7n8u3 exp[0. 125283(1 T
~~
)~

4. hqunatnomn.c of SlaIn

4.2. (un/ui, Eqniation.s of Stair

153

(4.65)

a0) =T~~exp[0.51 1614(1 _T,aasmml)J

and for SRK [39),
amom
=

lise above approach was applied [41,42)10 PR, as the most widely used equation in the industry. Over 5,000 vapour-liquid equilibrium experimental data of binary systems containing a stiper critical compomment, with hydrocarbons ranging from C to nCl2 were 1 pressure ofthe umsed to develop a correlation for supereritical components. The bubble point
liquid phase and the compositiomr of tIme equilibrated vapour phase, as predicted by PR, were matcimed to experitmiental data by adjusting the paranseter a of the super critical components. The optimum value of a for suiper critical hydrocarbon components was found to he reasonably expressed by Eq.(4.29) replacing, m, with, m, where m=l.21m

T~o.20n ussexp[O

141599(1 T,

229526

)]

(4.66)

a~mu= T,~

m45 exp[0.5003 15(1 T~ 6Suss)]

(4.67)

The average absolute deviation of predicted vapotnr pressure for pure hydmrcarbons by PR from the triple point to the critical point was found to be 3.28%, 8.2 1%, msnrd 12.08% using the Twu et al.s proposed correlation, the Stryjek-Vera correlation, Eq.(4.62), amid time origin~ml

correlation of PR, respectively. The average absolute deviation by SRK usitmg the proposed correlation, Eq.(4.66). was found to he 3,37%, conmpurrabic wills that of Twu Ct mil. for PR. in spite of apparent difference between tine two equations. The results clearly demonstrate the importance of the selected functional form of a, and lime vmipour pressure tiusta tused in correlating it. The success of EOS in predicting phase behaviour using various functiommuni forms for time temperature dependency of the attractive term, all correlated by mruatching vapour pressure data, raises an important question. How reliable their extrapolations are for supercriticuml compounds? Light components of reservoir fluids, particularly methane which constitutes a large fraction of reservoir fluids, are generally at temperatures well mthovc tlmeir critical points, where no vapour pressure data exist to be used iii correlamiing the pmiruumnuelcrs. lime prevailing approach is to assume, altnost in all the leading EOS, tisat time correlation is also valid for super critical conditions, Whereas Zudkevitch and Joffe assumed valumes of ~a and ~b for supercritical compounds to be the same as those at the critical point in the RedlicisKwong EOS. A few investigators have suggested different correlations for supercritical compounds [30, 38,40]. There is no limitation in evaluating the reliability of any a correlation at super critical conditions if volumetric data is to be predicted. There are abundant density data on super critical compounds. The a function, however, is the tool to improve the vapour pressure prediction. An alternative to using vapour pressure data of pure compounds to correlate a, is emmsploying phase behaviour data of binary systems comprising of one super critical component. This approach extends the temperature range of relevant data, and it is only logical to expect higher reliability when employing any correlation within its correlated domsmain, The approach has an additional practical advantage. As binary data are used in correlating the parameters of EOS, the interaction between pairs of non-sinnilar molecumles and/or tine deficiencies of EOS for binary systems, arc taken into account to some extent. Hence, the need for the use of binary interaction parameters (BIP) in mixing rules, described in Section 4,3, will be reduced. This will allow for a significant simplification in phase behaviour calculations, see Section 5.1, resulting in reduction of the computational requirement for mixtures described by a large number of components. The reduction of computing time spent in flash calculations is highly desirable in compositional reservoir simulations, where many millions of flashes may be performed in a study.

Time predicted dew poimmts by time tmnodified PR, using the above correlation without any BIP, as well as those by the originuil PR, with amid ~vitimout BIPs, are cotnpared for a 5-conmponent mmsixttmre in Figure 4.6. Note that tire predictions by lime nmodified a function (shown by rnPR are superior to timose of the origimsal witim and wittmout time use of BIP. Fhc mmhove cxaimmplc clcumrly indicates the immmpact of tIne temperature dependent term of EQS
on predicted results, ammd the success of using binary data to determine its correlation for

supercritical compoummds. The umpproach may he immnplemnented in any EOS, resulting in modified correlations for super critical components [39).

.5)

a-

32
0 0 .-

30
aC

0~

a S 0

28

PR(kijl

25
20 40

Pro

a
..,._._.

Eng.

60 81) Tempcranure.~C

100

120

Predicted dew point of a mixture with composition: C =82.05, C =8.95. 1 3 nCs=5.00, nCj~p 1.99. mind nCl6=2.Ol mole %. Figure 4.6.

4.3

MIXING RULES

Equations of state are applied to multicomponent systems by employing mixing nmles to

determitme their pamarneters for mrmixtumres. The parameters of EOS are considered to represeumt

time attractive and repulsive forces between the tiiolecules. Hence the mixing rule sisould describe the prevailing forces between molecules of different substances forming the mixture.

1 54

4. Equnation.c ofState

4.3. Musing Run/es

155
.

4.3.1

Random Mixing Rules

a,)

(4.73)

As pointed out in Section 4.1, the coefficients of virial equmation (Eq. 4.1) describe time non-ideal behaviour of a real fluid due to interaction amongst various comnmhinatiomss of molecules. Statistical nsechanics can be employed to derive miximmg rules for lIme coefficicnmts of virial equation. It can be argued that the nsixing rule tnsed in umny EQS shounld attaimm tine

~
,

x Im
, q

~
,

(h
u

h )/2

~x h

(4.74)

same form as that of the virial equation at conditions wisere both equations urre vuulicl.
For gases at low pressures, the third and higher virual coefficmemils can he mncglecled. Tbse second coefficient, which represents the interaction between two neighbouring molecules, is sufficient to describe the volumetric behaviour. Tine mixing rule for tine second coefficient, B, is of the quadratic form, A mnmixiurg rule simmnilar Its tlmmrt of h is also used for otlmer pumrameters in EOS that contain more than Iwo parameters, wlsen the additional parammseters are of the co-volume characteristic, c=~x,c, (4.75)

B=~~x,xB,
wlmere B is the coefficient due to interaction between urroleculcs m mmmd 1~ Employing Eq.(4. I), the second coefficient is deternsined as,
j.

(4.68) Sinnunhmmrly, tIne mmtsunvc mnnolmmr nnrixinng ruulc is,rlso umsed Err auxilmuury paranreters suich as tIne vsl mm nrrc crrrcct mum , er,. dc fi mmcd mr I ~qs.(4.37). tIme suggested nmsi x inrg rmnle for tire aeermtric tmuctmr, used mrs lIre tlmird pmmrmummuctcn ins tIne Sclmnrmidt-Wcnrzcl EQS. by the authors is,
On) ~

(4.76)

B=Iim(aZ/ap) Using a van der Waals type equation to describe Z at low pressures, tIre above equation results in, B
=

lime urhove mixing rules, known as tire van dcr Waals mimixing rules, treat all the components similuurly, imence, referred to as time random nnixing rules. For reservoir hydrocarbon fluids the random mixing rules (whicim oumly consider time interaction between pairs of neighbouring mrrolecules mmd neglect interactions between three or more molecules) are adequate.
.

lim(dZ/~p)= b(a/RT)
p-.O

(4.69)

hence, the mixing rules for a and b, at least at low pressures, should he compatible witlm that in Eq.(4.68). i.e., it should be of quadratic form. The attractive force between molecules i and j. represented in EQS by paranmneter. aij. which is of an energy nature, can be expressed in a simple geometric average fornsr [43] as, a,
=

It is comssmon to incorporate air additional parameter in Eq.(4.71) to express the attractive termms between pairs of nnommsimssnlar immolecules, 2 a,, = (a,a,) (l k,,) (4.77)

wlmere k is kirown mus tIne hinurry interaction parameter. 1~ tJsinng hIre umbove descriptions, lIme rumnrdomns mixing rule of the attrachive term becomes, a
=

(a a,)

(4.71))

lime repulsive force between molecules i and j, represented in EOS by parulmsreter h . winch 11 has the characteristic of volume, can be determined by arithmmmetic umverage, hnj
=

~
, ,

x (a a

)~(l k )

(4.78)

(b +h~)/2 1

(4.71)

Eqs.(4.70) and (4.71) describing the interaction between a pair of differennt msmolecules are nsore intuitive than rigorous. Otlmer forms, perhaps with equally valid argunirents, cats uslso be considered, For example, considering the distance between the two molecules, immstcad of averaging their volumes results in.
h
+

TIne cisc of lsinary insteraction paraunneter for the repulsive term, particularly in mixtures with Imiglu conscentratioun of C02 [44i, lsas also been suggested, but has not gained popularity, b~ =

Rb

+h,)/2Rl-k~p

(4.79)

wlnerc k~ are the repulsive BIP. 1 I llre binary intermuchion paraummeter (RIP) is generally determined by minimising the difference between predicted anmd experinmnental data, nnainly time saturation pressure, of binary systems. A RIP slsotnld, mherefore. he considered as a fitting parameter and not a rigorous physical ternnm. ilenmce, tIre imnterumction paramsnctcrs developed hor any EOS should generally be used only for that EQS.

h ~

(4.72)

Applying the quadratic mixing rule for the parameters of EQS , we obtain,

156

4. Equoiion.c ofState

4.1. Mixing Run/es

151

As the effect of third and higher molecules on the interaction between two molecules is assumed to be insignificant, the binary interaction parameter so determined is coissidered to be valid for multicomponent systems too. The interaction paranmeters between hydrocarbons with little difference in size are generally considered to he zero, hut the values of k for 1~ non-hydrocarbon-hydrocarbon components and also light-heavy hydrocarbons are non-zero. Values of BIP for the EOS described in Section 4.2 are given in Tables A.4 in Appendix A, where k,~ = k~ , and ~ =0. 1 Correlations to estimate BIP for specific EQS, such as SRK[45] and PR[461, as well as general ones [47,48, 49) have been suggested. The most commonly used correlation [47] is, (vn/avuhl) 2~ + ~ ~ (4.80)

where, ~ and. ~2. are constants equal to I and 0 in SRK, and Ii respectively.

and

i-~J~ in PR,

Prove that the fugacity of each connponent in a mixture, using the above EOS and the random

mixing rules is given by,

lno~=~-(Z- l)-ln(Z-B)(E4.5)
.SoInitnon:

11

0, I

lhc fnngmmcity coelficicust is lui~=~-J Rl

~
v

cuulcuilmmted fnssnnn Eq.(3.3 I),

where the constants ~ and 0, are determined for cads EOS using time availumhle binary data, or adjusted in tuning of EOS for a particular fluid system, as will be described inn Section 9.3.
A default value of 0=6 may be used [50). There is no doubt that the inclusion of binary interaction parameters in EOS mixing rules will provides more flexibility, and in most cases reliability at least within a limited working range. It is particularly a powerful tool to tune (calibrate) EOS for a reservoir flumid against the available experimental data. Additional flexibility can also be obtained by making BIP

-RT/V dVInZ

(3.31)

~mr )r.v.,,,,, 1

wlnere V is the totmul volume. ilence, the equuumion of state is written in terms of total volutnre by suhstitcutimmg v=V/us, where n is the total number of nmshes,
Is = ~lii

temperature [46], pressure [51], and composition dependent [521. It shosuid he noted that making BIP dependent on pressure or composition causes additional comnmplexity in the expression for fugacity of each component as the pressure derivatives in Eq.(3.3l) are only at constant temperature and total volume, and not at constant pressure or composition. The flexibility achieved by inclusion of BIP, particularly variable ones, can he quite misleading, as excellent results can be obtained for binary systeumis. That, however, only demonstrates a successful curve lilting. The results for mumlticomponent systemns particularly within wide ranges of temperature and composition may he quite disappointing. A comparative study of ten EOS [53] indicated that the Patel and Teja equation as modified by Valderrama, without any BIP was more successful in modelling of the phase behaviour of reservoir hydrocarbon fluids than others with BIP. An improvement in EQS or a more thorough fluid characterisation should reduce the need to use BIP for hydrocarbon fluids which do not contain compounds of vastly different characteristics. An example on improving EOS instead of using BIP was shown in Figuire 4.6. TIne reduction of computational time for flash calculations in tine absence of BIP will he described in Section 5.1.
The application of mixing rules in EOS, will allow tine calculatisn of commaponent fugacity

We obtain.

msRT

2 n mm

V_mlh(V+~nimb)(V+~,tsh)

The derivuutivc of pressure at eonstuuumt totmml volunmme, pressure and all mole numbers except n, is calculated as, 2 2 ~(n a)/an(n a) + (V + ~~umh)(V + ~ nh)

~~

)T.VM,,,

RT V bin

nRl[nJ(mmh)/an ] 2 1

(V

nmh)

2 {~rn~s[~(nts)2 /~n(mmh) ]+ (es,

[(V + ~nt1b)(\ +

coefficients, as given by Eq.(3.3l). The expression for fuguucity coefficietst, using the generalised EOS, Eq.(4.I2), and the random mixing rules is given its the Appeirdix C.
Example 4.5.

~32nsb)12
1

Applyimrg tire rmnumdomru mnniximmg tunIcs tr calculmmte a and b, Ec .(4.78) and Eq.(4.74) respectively,

2 n a=

The Soave-RediichKwong, and the Peng-Robinson EOS are the most widely used equations in the petroleum industry. It is common to express these equations by the following general form, ~RT a

2 ~n x x~a, i~u j~u 1

~ =

= ~

i~n

~nn a~ 5

mb

~nx b~ 1

vb

(v+3,b)(v+~ b) 2

lime derivmmtives of tIre two paraunnelers umre obmaitmed as,

158

4.

Fq,uatiours of State

4,3, Muxnnng Rules

159 x,)x,x,(a,a,)

a
[~(nb)/~fl~1TV ~,
=

(4.82)

where l~j=-l~i.s tine himruury intcrurctnon coefficient for time asymnmmetric term.

[a(n2a)/anjl
will result in.

2n~x~a~

Substituting the above calculated terms in Eq.(3.31) and integrating it hctwcems thue two limnnins

the above mixing rule is quite flexible, partncsnlarly wmth tensperature dependent interaction coefficients, and capable of describing tIme behaviour of unulticomponent mixtures containing higlmly asymmetric comnponemrt.s wlsen umsed in us cubic EOS. Flme tssixing rule, Imowever, is not consistent with the quadratic form of the second virial coefficient .Sc lmwmurtzentruher and Rcnonm 1551 have sirown that the above inconsistency can he unvoidcd by i nit rsdnncinng ann add it isnumm I lmmmrmmmmmchcr. y~. inn tIre general EQS. Eq. (4. 12), as, (4.83)

ln~ =lnZ(lnb/V)+~~~+ Vnh

~u RT(~m~)h

1,,

2~nrum. /nuuh /b

I n V+~ V + mn~h nb RTb(V+nmmb)(V+ nmnVh, nb)

vb

v4unv~w

wills I lse furllowing nrr ix inrg rmnIc for tIre a(ldnIionnuul parannrcter of EQS. Suhsnituting
=

(4.84)

2 an = P (V+~nnh)(v+~ nb) 2

nRT Vnb

i.e., the eqsuatnon of state, ansd V= nv

nhZ/B

and time conventional random mixing rules for the remaining parameters. llse mmhove mmmodification sahisfies tine mixing rules of the virial coefficients, as ajj expresses the interaction between two nnolecules by a quadratic form, and Xujk represents the interaction between three molecules by a cunhic form, similar to parameters of B and C of time virial equation, respectively. It also appeurrs to have mmnaimntained the dependency of the attractive ternim to power three of concentration rcqsmired by the local composition mixing rule, a
=

in the above will result in Eq.(E4.5). Non-Random Mixing Rules The van der Waals mixing rules are quite adequnate to describe isydrocarhoum nriixtsurcs of reservoir fluids. They cannot, however, represent the interaction between lsydrocarhomns and asymmetric compounds such as water, or methanol which is often added to reservoir flumids as a hydrates inhibitor. Although additional flexibility that is achieved by increasing time number of coefficients in binary interaction parameters, may provide acceptable resumEs for binary nmixtures containing these compocmnds, tine model cans fail conrupletely for multicomponent systems [541. The assumption of random mixing in systems c(ntumining higtmly polar mund asymmrmcnric compounds is not justified as the existence of pamticsular forces bctwccmm snrnrnc unsuslecculcs, such as those due to permanent dipoles, may restult ins non-unmiform distrihcutiomm unt tIme molecular level. Local composition mixing rules address Ilmis beImavitsuir by rclmmtitrg lire attractive term in EOS to composition with a higher order polynromials thats quadratic. The majority of mixing rules for the above term can be represented by the following form, a=aC
+

X/(v b)

(~~~,x,x,xkx,k )/(v

b)

(4.85)

the utbove eqsummtion bussically beionsgs to a class of mixing rules, known as the density depemmdemmt mixing rsnles. Time asymmmnmelric term becomes negligibly small as the pressure mrpproachies 7.ero, with time voltumne uuttaimring a large value, reducing the mixing rule to the conmventional ramsdourr mrmixing rule amsd consistent with that of the virial equation. The cubic formmn of conscentrmmtinrn is, inowever deceiving, as the volumme of a mixture depends on the comrccnstrmution of its conmrponcmrts, tummtcss their partiuul nmnolar volsmmes are very sinmilar, wimich eIfcchivcly reduces lIre asynnmnssctric tcrumm to mm qtmurdrumtic forum [55j. This has been observed in prunctice by varioums immvcstigators [56,57). It is perlmaps a sounder engineering practice to use nnnixing rumles of lire type of Eq.(4.82). arrd not he quite consistent with the virial mixing rules, tlmamm lose reliahilily of results by adherimmg to them. TIne mixing nile, Eq.(4.82), with time cubic dependency on composition suffers also from the invariance condition 1581. that is, if one of the components is divided in two or more comnuponenmts identical to it, differemmt expressions for tire attractive term are obtained, This defect is of practical insterest for umnixtures containing significant amounts of very similar commlponenhs. Matirias et al. [59) tested the mixing rule of Panagiotopoulos and Reid [60], wlmiclm is of ur fornrs simmsilar to Eq.(4.82), amid demonstrated that this deficiency causes cnrssmnemrmms prcdictionrs of wmmhcr soluimility in henzcume-cyclohexane nnixture. Modifications of tIme cunhic mnmiximmg rules ho avoid this deficiency for mmmulticomponent systems., whilst reducing tisem to tIne sammse formms as Eq.(4.82) for binary systetmis. have been ~roposed[55. 59). Avlonitis et al. [57], proposed a mrmixing rule similar to Eq.(4.82) as follows:,

as

(4.81)

where the attractive terns is separated into two pumrts. use, wimich is tire conventionsal raurdomss mixing term given by Eq.(4.78), and ~ which is,tlme asymnmnsretric tcrnm clime to polarity. Various expressions have been proposed and successfully tested to himsumry systcmsrs br tIne asymmetric term. Schwartzentruher and Renon [55] have simown that most of them can he expressed by the following general form,

160

4. Fqunasionn.r a/Stale

4.3. Miximig Rides

161

aA

(4.86)

I anmd water 2. The predicted composition of the non-polar-rich phases are only very slightly affected by the invariance condition, since 1mm this case tire terms in Eq.(4.86) containing polar component nmole fraction products contribute negligibly. fhe musing rumle ~rf Eq.(4.86) has a nunsher of important advantages for application to reservoir lluids. Commmpulation tiumne is short and is commsparable to that of the random mixing rules, as additionuml smmnnmmmatiomrs of nsole fraction produmcts are only for polar components, typicmmlly just a few. This also avoids time dilumtion problem occurring with the mixing rule in hq.(4.82), that is, vaunislming of time cubic terimi wisen Ilme number of components increases. Fturtlmcrmmrore, no puirtictular commsptmtumtionmnl nmemnory atscl space reqtuirements are imposed for uilmplicumliots of time proposed nnmodcl, sitsce tine nummmher of binary paramsieters is limited to only three. Kmrumulscmm Ct uml. 1541 cvmmluated tIne capability of different leadimig mnmixing rules to describe time pimumse hehnaviouur of msnixtures eomstainmimng asymmstnetric commrpoummds by coniparing predictiomm with cxperiummcmmtal data. lhe hltmron-Vidal mnmodel [611. based rum the principle of minimisation of excess Gibbs energy, overally performed nmore reliably than others, whereas the random mixing rule with foumr binary parameters failed.

I-Iere the subscript p refers to the index of polar consponcnls, amrd is tine biurmury iustcruiction coefficient, The binary interaction coefficient should be a decreasing fuimctiomm of temmnperatture as the asymmetric non-ideality reduces with temperature [431. The above mixing rule, generally, should not he affected uudversely by the immvmmriancc condition when applied to reservoir fluids, The authors, however, eliminmutemi lIme prohletmi 2 completely by substituting xp with S~, Xp in Eq.(4.86), wimereS~= ~ ilcmmce,

aA =
P

(4.87)

For a binary polar-non-polar system the modified form is identical to Eq.(4.8(r), lmcmmcc, time same binary interaction coefficients can be used for both forms. For mirixtures consisting of polar components only, S~,=1, and the term aA becomes identical to aCwitlm polar-polar interaction coefficients equal to zero. The authors [57] applied the mixing rule to tine Valderransa mnnodification of Puitel-Teja EQS to model the phase behaviommr of reservoir fluids including water, mind nmnctlnmumsoi which is commonly used as hydrates inlnibitor. To furtlscr immnprove tine muccuummucy of puc(lictiouns of both the vapour pressure and the saturated volunmes of pure polar commmponemsts of water and methanol, the constants in the correlation of a(T ), Eq.(23) were determmmined by regressing 5 pure compounds data, instead of using the generalised correlation, as a(Tr)= (1
+

m[I

t (Trf J j2

(4.88)

fumble 4.3. Effect oh inmvuuriunmmcc comm(liti(snm (so predicted pimase equilibria of C iI SC0 ll Q at 1 2 2 2 i=3 11.1 K, smmsd P=6.26 MPa 1571. ______________________________________ (onsnpuuncnt wmnucr 2 /wancr I Mcnluurne }lydr. sulfi. Carh. diunx. Wamcr~ Ieum.I 0.1151)4 0.39116 0.0503 (r.5t$uS Wamer-riclm liquid -t3uupn. 0.000490 0.0284 0.00350 0.9677 Calctd. mv. 0.000402 0.0269 0.00326 0.9696 Cmnhend. 0 0.000402 0.0269 0.00326 0.9696 1/9 0(100104 t).0l62 0.00151 0.9823 1/7 0.000016 0.0082 0.00052 0.9915 I/I 0.000008 0.0065 0.00036 0.9933
Expt.
-

where for methanol m=0.76757, !=0.67933, and for water m=0.723 18, Y=0.52084. The binary interaction coefficients were expressed by, I ~ =10 ~l~(T273) (4.89)

Catcid. mv.
C,mlcnd.

0 1/9
3/7

0.0653 0.0602 (1.0602

0.8197
0.8391 (1.8391

0.1049
0.0894 0.0894

0.t)lt)I
(tot 13 0 1)113

()(16() I
0.06(11 0.0601

(1,8385
0.8383

l/t Expn. Cmnlcnut. lnv. Cmnkud.

.

0.8382
0.5028 0.5248 0.5248

t).090 I t).09tu3 1)0903 0.1739

0.1517 0.1517

(1.1)114 tIlt 114 0.0114 0.00)214

0.00194

where l and Ut are dimensionless constants, and T is in K. The above function will change sign at some high temperature values where it can be set equal to zero. The binary interaction parameters and coefficients were obtained by forcing agreement of the mmsodel to binary data, with the results given in Table A.4.6 in Appendix A. The detrimental effects of the invariance condition can be demonstrated clearly for a mixture shown in Table 4.3, by subdividing water in two identical components, demnoted as water I and water 2. The three-phase equilibrium of the four component mixture has been predicted by VPT using two expressions for the asymmetric contribution, Eq.(4.86), and the invariant version Eq.(4.87). Note that if only one polar component is present in the mixture, the mimiximig rule amsd its corresponding invariant version lead to identical results, since S~, =x~. In mill other cunscs tire predicted composition of the water-rich liquid depends very strongly on tIne ratio of time identical water components. A maximum effect is observed at equuimmiolar annotmnts of water

0.3213
(1.3216 0.3216

(1

((.3209 0.5244 3/7 0.3207 0.5243 I/l 0.3207 0.5243 *Cmnlculmmmed waier is the suurn of tIme aurmomunus un water I and wmmncr 2

n19

0.1527 0.1530 0.1531

0.00194 0.00197
0.00198

0.00199

fine local comnsposition miximmg rules relying on description of non-randonn forces by iimcreasing time order of concentration polynomials are more sumccessful in modeliimsg tine vumpcrur-liqumid eqtuilibrium than the liqumid-hiquid equilibria. Time prediction deteriorates puirlicumlarly for syslcmsns near the plait point, equivalent to the vapour-liquid critical point

1591.

162

4 Eqnuatnonr.c of.S(ate

4.4. References

163

4.4

REFERENCFS
19. Ymurborotugln, F: Applicuuhiomm of a Gencralised Equation of State to Petroleuim Reservoir Eltuid, Eqtiamiomrs nsf SImile in Engimmcering, Advances in Chemistry Series, Chao, K.C. and Rohinmson, R.L. (eds), Amriericumn Chemical Soc., No. 182, 385-435 (1979). 20. Reid, R.C., Prausmritz, 3M. and Sherwood, T.K: The Properties of Gases and Liquids, 4tIm ed, McGraw Ilill (1986). 21. Soave, Cu: Equihibriumims Constunmsts from a Modified Redhich-Kwong Equation of State, Chem. Eng. Sci., 27, 1197-1203, (1972). 22. Soave, C., Barolo, M. and Bertucco, A: Estimation of High Pressure Fugacity Coefficients of Pure Gaseomis Fluids by a Modified SRK Equation of State, 3. Fluid Phase Eqnnihihria, 91,87-100 (1993). 23. Grumhssski, MS. ummnd I)uuuuhcrt. IF: A Modified Soave Equation of State For Phase Eqtmulmhrnum Calculmntionms. I. Ilydrscarhon Sysiemsms, md. Eng. Chem. Process Des. Dev., 17(4), 443-448 (1978). 24. Peng, DY. aircl Robimrson, D.B: A New Two-Constant Equation of State, md. Eng. Chem. Fundam., 15(1), 59-64 (1976).
25. Rohinsoms, 1)8. unurd Peng. t).Y: The Characterisahion of the Heptanes and Heavier Frunclions for tire (ilA lcmsg-Robinmsusns Programs, GiA Research Report 28, rulsa (1978).

Bendict, NI., \Vehh, 0.8. and Ruhin, L.C: Ams Eumspiricuul Eqnnumtions for Tirennmsodymrannsic Properties of Light Hydrocarbons and Their Mixtures, Methane, Etlmane, Propane aurd n-Butane, J. Chem. Phys., 8, 334-345 (April. 1940). 2. Mason, E.A. and Spurling, T.H: The Virial Equation of State, mm The hnhcrnmmitioumah Encyclopedia of Physical Chemistry and Chemical Physics, Pergamon Press Inmc (1969). 3. Starling, K.E: Fluid Thermodynannic Properties for Ligist Petroleumim Systenns, Gulf Pub. (1973). 4. Starling, K.E: A New Approach for Deternmining Eqmiahionm-of-Statc Punrmnnmreters t.lsinsg Phase Equilibria Data, SPEJ, 363-371, Trans. AIME, 237 (t)ec., 1966). 5. Thiele, E: Equation of State for hard Spheres, 3. CIrcus. Plryt., 39, 474-479 (1963).

6. Carnahan, N.F. and Starling, K,E: Intermolecular Repumlsions mind lime Eqtmalion of Slate for Fluids, AIChE, 18, 1184-1189(1972). 7. Donohue, M.D. and Prausnitz, 3M: Perturbed Hard-Chain Tiseory for Fluid Mixtures, Thermodynamic Properties for Mixtures in Natural Gas and Petrolemnnmi Tecinnsslogy, AlCinE, 24, 849-860(1978). 8. Chien, C.U., Greemnkorn. R.11. atmd Churo, K.C: A Clsmmiin of Rst mutors I~limit non nsf Smmute. AIChE, 29,560-571(1983). 9. Lin, H.M., Kim, U., Guo, TM. and Clnao, K.C: Cubic Chsumin of Rotumnors Equations of State and VLE Calculations, J. Fluid Phase Equilibria, 13, 143-152 (1985). 10. Redlich, 0. and Kwong. J.N.S: Qn tIne Thermnmodynammiics of Solumlioums, Chemicurl Review, 44, 233-244 (1948). II. Abbott, M.M: Equations of State in Engineering and Research, Advuun. Chremir. Ser., 182 (1979). 12. Airmed, Y., Sugic, II. and Lu, B.C.Y: Comsmparuutivc Stundy of Fight h:(Immuuti(rns of State for Predicting hydrocarbon Volumetric Phase Behaviour, SPE 15673 (1986). 13. Firoozabadi, A: Reservoir-Fluid Phase Behaviour and Volumetric Prediction with Equations of State, JPT, 40(4), 397-406 (1988). 14. Martin, J.J: Cubic Equations of State - Which?. md. Emsg. Chiem. Fundumos., 18(2), 8197(1979). I 5. Anrderko. A: Equal noun of State Methods fmrr tIre tvlodc Ii mmg Hmmid Phrase Equuihibrimu, 61, 145-225 (1990).
ot

26. luunclotnx, A.. Ruumni.y, F. unnud Freze, R: A Consistent Correction for Redlich-KwongSoumve Volunmmses, J. Fluid Pirumse i:qnuilihria, 8. 723. (1982). 27. Jinaveri, B.S. mmmd Youmrgrcnr. OK: Tlrrec-Paramneter Modification of time PengRobinson Eqtmaluomr of SImile to lnmmprove Volumetric Predictions, SPE 13118 (1984). 28. Munthmias, P.M., Numiseiri, T. amid Oh, EM: A Density Correction for the Peng-Robinson Fquuumtion of State, 3. Fluid Plumuse Equilibria, 47, 77-87 (1989). 29. Clnomn, OR. anrd Prummnsnitz, J.M: A Phemiomenological Correction to an Equation of Slmste for tine Critical Region, AICIuF, 35(9), 1487-1496 (1989). 30. Sclmmmnmdt, G. munsd Wenmzcl. II: A Modified van der Waals Type Equation of State, Clucums. Fng. Sci.. 135. 1503-1512 (1980). 31. Pmutel, NC. mmmd Tejir, A.S: A New Cubic Equmation of State for Fluids and Fluids Mixtuurcs, CIrcuss. Emmg. Sci., 77(3), 463-473 (1982). 32. Valclcrrauurur, 3.0. urnsd Cislernas, l..A: A Cuubic Equation of State for Polar and Other (omnnplcx Mixtures, J. Fluid Plmunsc Fqnnihihrimm, 29, 431-438 (1986).
33. Vmmlderrmmmnrmu, J .( ): A (cnrcn mu I iscut Pmutcllejmn Fqnnuumion of Stuute for Polar and NonPolar Flunds urird lIscir Mixtures, J. Cincmrn. Fnrg. Jutpumn, 23(1), 8791 (1990). t)mumncshr, A.. Xum, t). umumul Todd, AC: Ann Evumiumation of Cubic Equations of Slate for Piumuse Beimmnvioumr Cmrlcmuluuniomns Ncmnr Miscibility Conditions, SPE/DOE 20267, Proc. of 7th Symnup. urn FOR (April, 199(1). 34.

lhummsc I ~qumi I ilsrimm,

3.

16. lsonopoulos. C. mmd IIeudman, iL: Fromrr Rcdlmchs-Kwong ho tIne Frcsemsh, Phase Equnihibria, 24, 1-23 (1985).

Flmuid

17. Ztudkcvitch, D. amud Joffe, E: Correlation and Prediction of Vapour-Liquid Equilibrium with the Redlich-Kwong Equation of Stale, AIChE, 16(1), 112 (1970). IS. Joffe, J., Schroeder, G.M. and Zudkevitch, D: Vapour-Liquid Equilibrium Redhich-Kwong Equation of State, AIChE, 496-498, (1970). with the

35.

\Vulsomm, C. M: Vapour-hiqtuid Equilibrium, Correlation by Means of a Modified

Redlnclm-Kwong Equation of State. Advan. Cryog. Emmg.. 9. 168-176(1964).

164

4. Eqiiation.c of State

4.4. References

165

36. Stryjek, R. and Vera, 3.11: An Improved Equation of State, ACS Symnp. Ser. 300, Washington DC (1986). 37. Mathias, P.M. and Copeman, T.W: Extension of tine Peng-Robmnson Equuition of Stumle

52. Bjorlykke, OP. and Firoozahuuchi, A: Measurement and Computation of Near-Critical Phase Behaviour of a C h/nC24 Binary Mixture, SPE Res.Eng., 271(1992). 53. Dumnesh, A., Xu, D. and Todd, A.C: Comparative Study of Cubic Equations of State for Predicting Phase Beimaviour aind Volumetric Properties of Injection Gas-Reservoir Oil Systenss, J. Fluid Pisase Equmilibria, 63, 259-278(1991). 54. Kmmudsemm, K.. Stemrlmy, Eli. umnmd Fiedenslund, A.A: A Comtmprehmensive Comparison (if Miximmg Rules for Calculations of Phsase Equilibria in Complex Systens.s, J. Fluid Phumse Equilibrium, 82, 361-368 (1993). 55. Sclmwartzentruher, J. and Renon, 11: Equation of State: How to Reconcile Flexible Mixiusg Rules, lire Virial Coefficiemrt Constraint, and the Michelsen-Kisteunmacher

to Complex Mixtures: Evaluation of the Various Forms of time Local Comnposition Concept, 3. Fluid Phase Equilibria, 13, 91-108 (1983). 38. Twu, C.H., Coon, J.E. and Cunningham, J.R: A New Generalised AIplm~rFunction for
a Cubic Equation of State, Part 1. Peng-Robinson Equation, J. Fluid Phase Equmilibria, 105, 49-59 (1995).

39. Twu, C.H,, Coon, J.E, and Cunningham, J.R: A New Generalised Alphimm Fnmnction for
a Cubic Equation of State, Part 2. Redlich-Kwong Equation, J. Fluid Phase Equilibria, 105, 6 1-69 (1995). 40. Kubic, W.L: A Modification of Martin Equation of State for Calculaling Vumpouir

Syndronme for Mnlticotsnpommemmt Systenns, J. Fluid Phumse Equilibria, 67, 99-I 10(1991). 56. Melhiemim, GA., Suuinsm, R. mind (ioodwitn, B,M: A Modified PemmgRohmnsomn Equation oh State, J. Fluid Phase Equilibrium, 47, 189-237 (1989). 57. Avlooitis, D., Danmeshn, A., Todd, A.C: Prediction of VL and VLL Equilibria of Mixtures Commhuninimsg Petroleummi Reservoir Fluids and Methanol with a Cubic EOS, 3. Fluid Pimase Equmilibrimu. Vol. 94, 181-216(1994). 58. Mictielsen. ML. mumd Kuslemnummaciser, H: On Consposition-Dependent Interaction Coefficients, 3. Fluid Phiumse Equilibria, 58,229-230(1990). 59. Matlrias, P.M., Klotz, hl.C. amid Prausnitz, J.M: Equation of State Mixing Rules for Mullicounmpomment Mixtures: the Problem of Invariance, J. Fluid Pimase Equilibria, 67, 3 1-44 (1991).

Liquid Equilibria, 3. Fluid Phase Equilibria, 9, 79-97 (1982).

41. Reservoir Fluid Studies, Final Research Report, Vol. 1. Depmmrtmncumt trf Petroleumusm Engineering, Heriot-Watt University, UK, Report No: PVT/93/2 (July, 1993). 42, Danesh, A., Xu, D.H., Tehrani, D. and lodd, A.C: lmsmproving Predictions of Equation of State by Modifying its Parameters for Super Critical Components of Ilydrocmmrhon Reservoir Fluids, J. Fluid Phase Equilibria, 112,45-61(1995). 43. Prausnitz, J.M., Lichtenthaler, RN. and de Azevedo, E.G: Molecular Thermodynamics ofFIuid-I~hase Equilibria, 2nd ed., Prentice-Hall Inc (1986).

44. Turek, E.A., Metcalf, R.S., Yarborough, L. and Robinson, R.L: Phsase Equilibria ium Carbon Dioxide-Multicomponent Systems, Experimental Data and Improved Prediction Techniques. SPE 9231, Proc. of 55th Ann. Conf. (Sept., 1980).
45. Elliot, J. and Daubert, T: Revised Procedure for Phase Equilibrium Calculations with Soave Equation of State, mnd. Eng. Chem. Proc. Des. Dcv,, 23, 743-748 (1985).

60. Panagiotopoumlos. A.Z. umnd Reid, R.C: New Mixing Rules for Cubic Equations of State for Iligimly Polar, Asymmsmsmetric Mixtures, ACS Symp. Ser, 300,571-582(1986),
61. Huron, M.J. and Vidal, 3: New Mixing Rules in Simimpie Equations of State for VapourLiquid Equilibria of Strongly Non-Ideuml Mixtures, J. Fluid Phase Equilibria, 3, 255-271 (1979).
4.5

46. Varotsis, N., Stewart, G., Todd, A.C. and Clumney, M.: Pimase Behaviour of Syslemiss Comprising North Sea Reservoir Fluids and Injeclion Gases, JPT, 38(11), 1221-1233 (Nov..
1986),

EXERCISES

4.1. A (sine litre cyliirdcr comstmmiurs 16(1.43 g of nmethunnse at 373 K. Calculate its pressure uusimrg BWRS.

47. Chueh, P.L. and Prausnitz, J.M: Vuupour-Liquuid Equilibria utt lligls Pressumres, Calculation of Partial Molar Volume in Non-Polar Liquid Mixtures, AICIrE, 13, 6, 10991113 (1967). 48. Katz, D.L. and Firoozabadi, A: Predicting Phase Behaviour of Condemmsate/ Crude Oil Systems Using Methane Interaction Coefficients, JPT, 1649-55 (Nov., 1978).

4.2.

Prove Eq.(4.13).
~uu

4.3. Sisow that tIne vuilumes of respectively. 4.4. Puovc Eq.(4. 18).

and

in RK are equal to 0.42747 and 0.08664,

49. Teja, A.S: Binary Interaction Coefficients for Mixtures Containing lIne n-Alkanes,
Chem. Eng. Sci., 33, 609-610(1978). 50. Whitson, C.H,, Anderson, T.F. and Soreide, I: C , Characterisation of Related 7 Equilibrium Fluids Using Gamma Distribution, in C , Cimaracterisatioms. Clssrn, L.G. amnd 4.5. Calculate time vapour pressumre of normal isexane at 477.6 K using SRK. What are the predicted values of tire saturated vapoumr and liquid density? 4.6. Calculmute time vuuhue of criticuml compressibility faclor as predicted by PR.

Mansoori, G.A., Eds, Taylor & Francis, 35-56(1989). 7

51. Voros, N.G. and Tassios, D.P: Vapour-Liquid Equilibria in Nonpolar/Weakly Polar Systems with Different Types of Mixing Rules , 3. Fluid Phase Equuhmbrua, 91, 1-29 (1985).

4.7. Reduce the PR to a tlnree parameter correspotsding state form of Z=Z(Tr. ~r, (0). Courspare tIme calculated values of Z at T =1.5, P =2 and o~0and, also o=0.6, with those of the generalised comsmpuessihilihy clrutrt sho.vn in Fi~ure 2.22.

166 4.8.

Equuatwuu.c ofSioe

167

Derive an expression for the critical compressibility factor as predicted by SW in terms

of the acentric factor.

4.9. Compare the value of attractive term temperatunre coefficient, a, us PR for ilC,u, mmt T~=0.8as calculated by the original and modified correlumlions by tine untutlnoms. uunnd uulso by tire Twu correlation. 4.10. Prove Eq.(4.69). 4.11. Prove Eq.(4.33). 4.12. PR predicts the density of a nmixttnre of C -nC (50-50 mol %) equnuml ks 52(1 kg/urn. 1 10 Inirprove time predicted result by including the volume shift correctiomn. 4.13. Predict the density of a single plsase mixture csninposcd of (~=59.30%,C ~ mmnsd nC =3.24% (nnolar) at 311.0 K and 12.07 MPa, using PR, SRK and VP1.(Mcmmstured m 1 value=297 kg/m ). 4.14. Derive the expression for fugacity coefficient of a componcmrt in mm mmmnxtunrc usimrg the random mixing rules and the generalised EOS. (Answer mn Appendmx B). 4.15. In Exercise 3.5, the soluhility of etbnane and C02 in water was calculurted by uissunsning the gas fugacity coefficients equal to one. Use PRIo estmnra(e tIne fugacmty coefficiennts, mm order ho improve the accuracy of predicted gas soltnhmlmty.

PHASE BEHAVIOUR c ~& L CU L AT!ON S

lhnasc cqtmiiibritunsr cunlcnnlmmtionms for pchr(slcummmi reservoir fluids irray in general involve the Imeatmunenh of a mmtmmnrhcr of flumid mund solnd plsases. Wisen displacing oil in a reservoir with CO at 2 low hcmnperatlnre two liquid phases. onne hydrocarbon rids and one CO rich, can be in 2 eqtmihibriutum wills the vuupour phumse. TIne appearammce of two distinct liquid hydrocarbon phases formrred by retrograde condemssumtionm has also beets reported. The formation and deposition of solid-hike nuateriuml of asphaltic mmumttmre resultimmg from compositional changes in miscible displacemmrcnrt or vutriatiouls in pressure mmnd teusiperature are well docunnmented. Water in general is alwumys present in reservoirs as us separate phase, and it can also form solid phases of Inydrunlcs mmh cerluninr conn(hihmons of presstnne mind tcmsuperature. All co-existing fluids and solids in gemnermul should u ltimnruutcIy attains cqui Iurnitnmns, given so ffieicnt ti nme. Iypicuml calcunluumiomns of cqtmilibriunnm conmdilions can be classified in Iwo categories. In the first cmmtcgory. time conrposiliomn and properties of the co-existing phases at a given set of temperature annul pressure are required. lur line second case, tIme saturation condition, either temperature or pressure is seuurclmcd for a given commupositioms and pressunre, or tensperature. The main interest inn dcumhimrg with solid-like phrases, such as mnsphaltenes, waxes and hydrates, is the determination of tlreir fornmnuntiomn consdilions. Ins sunclm cases the overall composition of fluid phases remains uuuuchrmunsgcd mmnrd tIme ImInuse uqnnitihn inunnn calcnmlationms for fluid pimases can be performed generally imndcpenmdcmrh of lIre sirlith ImInumse. Also. hccunuse the effect of water on the hydrocarbon phase hmeluuuviosur cams lie ureglecled ins inmost cmscs, thre majority of phrase equnilibrium calculations are unruled ommly to two phases, that is, vapour-liquid equilibria. The uniformity of fugacity of each comnnponent thsroughirnut all phases, solids and fluids, as the requirement for chemical cqumilibrmtnmmn cams be eusrployed, Isowever, to detemmine equilibrium conditions regardless of the nunnnher of plmurses. Malimennmmticmml mmrehhnods for calculating vapour-liquid equilibria can also be extemnded to ansy nunrber of phrases. TIre mrinnmher and state of the phases at equilibrium may not he known, Imowever, in advance. This can be determined by minimising the Gibbs energy, as will be described.

168

S IIua,ce Relmanmu,uar (alemmhnlsunm.c

.5. I. Vapour I.iquid Equtiiihriunnm Cafru,huikuns

169

Fluid samples collected from various locations and depths within a reservoir ollen show somise discrepancies. In some cases, the observed differences are due to imsiproper smtnsphitmg. Time reservoir could also be compartmental, with limihed comnmumnicmmtiomms betweems uhiffereimt sections. The lack of reservoir fluid maturity cumn also caumse conrpositiomnmul vmnriuntitsmms willrimn mu reservoir as mentioned in Section 1.1. Moreover, uts tIne lemsuperatture mmmrd prcsstmue chrummige with depth in reservoir, a degree of compositional gradimmg should alwumys he expected. lime compositional gradient with depth is often negligible. and the vunrimitiomis utre witlmims tIme experimental accuracy in collecting samples. It can, however, become qumite significant for neuir critical fluids, and reservoir oils containing high concentrations of asphaltic nmaterials. The variation of fluid composition with depth may be estimated by applying tlmennodynamic concepts introduced in Chapter 3. The fluid column can be assunmed at equilihriunn, ignorimig the heat flux due to the temperature gradient, or be tremrted uut slcmidy sImile comnditiomis with compositional grading controlled by the balammce of clmeirricutl mind tlncmimmal fonecs. usiimg umumnequilibrium thermodynamics. Both approaches will be described in tins chapter.

K.z
y

I + (K l)nv

(5.5)

Simmmi lumr cqumuut i(rums cunni unlso he hcri vcd imr tcrmmms of is iussteumd of n. For kmmowtm values of Ki, airy of tIre above two equmrliomrs cams be stmhshituted inn Eq.(5.3) hum delermimse the value of ms~ (or eu.) An iterative msmelhmod is required to solve time resulting eqtmuution. Tire followimmg equation, kurown as the Raclsford-Rice LII equation, is generally the preferred form, as its vairue nmomnotommically decreases with increasing n,
N N z(Kl) 0 x,) = ~l+(K,_l)un

f(n) =

(5.6)

5.1

VAPOUR-LIQUID EQUILIBRIUM CALCULATIONS

Let one mole of mixture be flashed at pressure P and nn.+nv..l

teunperatuire T into

mm~ immtrlcs of liquid aumd (5.1)

lire uubovc cqnmttionm yields a plsysicmilhy correct roust for mm hetweens 0 mind I, provided thuit,
I

n moles of vapour. The total material balance for the system is,
.

(5.7)

with material balance for each component, i, as, z 1

=

and

x,nL +y n 1

i=I,2

(5.2)
~z /K
>

(5.8)

where z , xj and y are mole fractions of the component i, mm the misixture, liqunid anul vmnpour, 1 1 respectively.

lime mmiixlure is at its hobble poimnt when mm approacines zero. llence Fq.(5.6) reduces to, I (5.9)

(5.3)

where N is the total number ofcomponents in the systemn.

and At equilibrium, the fugacity of any component, i, in the vapour is equal to thmtt in the liquid. The equality of fugacity can be expressed by the equilibrium ratio, K , mrs given by F.q.(3.43), 1 K =y /x i=l,2 N (3.43) 1 1 1 The material balance equations, Eqs.(5.h-3), and the equilibrium requircusscnt, Eq.(3.43)
y
=

K,x,

(5.10)

At umumy tcmnmperumlurc time bubble poimst pressure can he deternsined a.s the pressure at which K-values. s.ntisfy Eq.(5.9). TIme bubble point is nsost sensitive to time nsixture light components, whmichi exlsibit l,mrge K v,ulues. At time dew point. n~approacises I. Ilence Eq.(5.6) redunces to, ~z,/K =I 1 amid (5.11)

provide the required 2N+2 independent equations to determine the 2N+2 umnkusowns of x , y~, 1 and 0. The number of variables can be reduced, however, by comnhimming time above equations.
Substituting the equilibrium ratio K 1 Eq.(5.l) results in,
=

y /x into Eq.(5.2), and solving for x amid y using 1 1 1 1

x.= and

z. l+(Kn_l)nV

(5.4)

x,=y /K =z /K, (5.12) 1 1 1 Time dew point prcssmnrc is thruit urt wlnicim K-vuniucs satisfy Eq.(5. ii). The dew point is most semmsitive to tIme unixtrure heavy comnspomnents, which exhibit smssall K-values.

170

5. P/ruse Be/,a,,,,ur cohuhutuo,ms

.5. I. Vapour Liquid Equilibrium calculations

171

Eq.(5.6) can be used to identify the state of a fluid mixture at a given pressure and lenmmperature. If a physically unacceptable root of either n >1 or n <0, was found the mixture nrmay be consm(lcred to be eitbser all vapour, or all inqumid, respectively. The ushove idcnmlilmcuition is valid if reasonably accurate K-values are used. A more rigorous approachm to identify ttse slate of a mixture is given in Section 5.2 using the Gibbs energy mninimisation metlnod. Example 5.1. Ii is often a convenienl practice, yet reliable in most appticaiionis, Its replace mu reservoir fluid by a binary mixture in sinnulumnimng certain reservoir processes inn tIne huhorunnory. A reservoir hydrocarbon fluid has been mvuodchled by a nnixnurc of (:1 nod nrClfl (60-40 msuoie%). The reservoir temperature and pressure are 377.6 K mmnrd 27.58 MIa, respectively. The nil is produced through a one stage inher,rncdiatc scpmnrmmnor nt 344.3 K and 6.895 MPa. (a) What is the shame of mire fluid at reservoir ctsumdimiomis? Use lIme (WA lK.chrmuris givcmm in Appendix I). (h) Calculate the bubble point pressure. (c) Equilibrium flash equations for a binary system can be solved anahynicmrthy, when using K-charts. Derive the appropriate expression, and calculate the gas mid hiqunid mole fractions, and the phase compositions, at the separator conrditions.
Sol,miio,n. Component I: Ci (a)

ihe experimenlal valtne is 23,50 MPa (3408 psia) [2].

Note that ~zK~

strongly

depenmds sn tire K-valmne of nmetlnane, due no its high volatility and concentration. A ressonrable ininnumil guess for a reservoir oil in most cases could be the pressure at which (Kz)C,=l.
(c) For mm himumury .sysncnru Eq.(5.6) redinces ho:

nV =~z (K K )/(l K ) lI/(K, I) (E5.l) 1 1 2 1 lIne dcgnccs of [recdsunr Dir mm hi mary vmupour Iiqmu id system an equil ibriu in conditions nrc misty two, uccordinrg to tIme Gibbs phrase rule, Eq.( 1.2). Hence at a given tcnruperrlumre mmd presstnrc, tIne K-vmrlnucs are constann and independent of the overall comnqsm eon non.
Using Figures T).2 umnid I). 13, at 344.3 K amnd 6.895 MPa (1000 psia),

K =3.8, K =0.0029, (experimemmtal values K,=4~0t)5.and K =0.0027 E49}). 1 2 1 Eq.(E5.l) results inn, vunpoumr mole fraction: v liquid mole fraction: ~L543 0 0457 Eqs.(5.4-5) gmve Ihe Cofliposimi(rtn of equnihibrated phases as follows, s=0.263. x =0.737. 2 y,=0.999, y,=0.00l

Connpomrenrm 2: nC ID

Ttse convergence pressure at 377.6 K (220 F) is estinnnntcd from Figure 1)1 000 (Appendix D): PkS psia (34.47 MPa). The equilibrium ratios of Ci and nCl0 are then read fronm Figures D.2 mind I). 13 (Appendix D), respectively, an 377.6 K and 27.58 MPa (4000 psia): K=l .4 Checking ~z,K,. K =0.l3 1

(x=0.2496,

x,=0.7504,

y=0.9980,

y =O.O020, experimental values). 1

At low and nusderute pressures, whneme the dependency of equilibriuns ratio on phase commnposition cams he msegiccted. flaslr calculations are relatively simple, as K-values are known. lii general, K-values vary with commrposihion, Isence, the solution is reached by iteration. line calcumhatiomm damn hegimm by umsmtiahismng K-values in Eq.(5.6) estimated from Raoults law, witlm an appropriate correlatiomm for lIne vapour pressure, e.g., Eq.(l.I0), or the more widely used Wilson equalion, Eq.(3.66). Time solulioms of Eq(5.6) yields the compositions of the two phnumses unsing Eqs.(5.4) and (5.5). The calcunlunted connpositions are then used to re-evaluate Kis, usumng aim equihmon of state (EOS) or K-value correlations, to be substituted in Eq.(5.6) for the next round of iterumtnon. The iterative calculations are complete when values of check ftnmmcnmons air all smnrmuller tlman certain pre-set tolerances. The above procedure, using EOS, is sclnemsnumtmcahly slnown in Figumre 5. I. A sinnple sumccessive substitution iteration method has been used in time flow chart. Successive substitution methods may prove to be very slow in counvcrgnng to a soltutiusn, particularly in the critical region. Various methods to imnprove the convergence rate, such as promotion in equilibrium ratio updating, and the use of Newton type metlmods, winds have a faster convergence rate than successive substitution methods, have been reported in the literature [3-81. II slrouid he noted ttnan tIne soluntiomm of xj=yi=zi always satisfies the set of equilibrium flash equuuuhiomrs, Eqs.(5. 1-5). The unhove trivial solution is a major problem in phase equilibriumn calculatiomns when equations of state are used in deternnining the fugacity coefficients. This occtnrs often in tine critical regions, where the compositions of equilibrated phases are very close. The trivial solution sirounld non he mistaken as time critical point predicted by the equation of sturte. TIne use of cquurlionrs of sImile ho estimate hhe critical point is described in Section 5.3. lIme trivjmml soiutiomm, lruswever, nnrmuy he tIne oniy mnmatlmeumnatical solution at conditions such as tine eqtuihihriuumim flash its single pinase region.

z,K, =0.6xl .4+0.4x0. 13=0.89<1. Hence, the fluid is a compressed (undersalurated) liquid. For an undersamurated vapour, ~z /K 1 1 Eq.(5.7) and Eq.(5.8) should be satisfied.
(b)
<

I. whereas for a nwo phase systemmi both

At the bubble ponnn Eq.(5.9) must he satisfied. The K-values are read frons the charts at 377.6 K by iterating on pressure:
P. psia, (MPa) K 1 K 2 e.K 1

zK 2 2
(1.02 0.0!

~z,K

3500, (24.13)

1.60

3000. (20.68) 3400, (23.44)

1.80 1.64

0.06 0.03

096 .08

11,98 I 0)

0.05

0.98

0.02

1.00

172

5. Phone Re/nau,oumr (a!culotio,m.c

5. I. Vapommr Liquiinl Equilihriimnm Calcuulathnu.c

Solution.

173

The approach is sinnilar to that for flash calculations as shown in Figure 5.1. The liquid conmpo.sition renmmains unchanged in bubble point calculations, therefore, the vapour comnmposimion can be calcuiaied fronr Eq.(3.43) instead of solving material balance equationus, Eqs.(5.4-6), mis in flasln calculations. However, the pressure is not known inn bubble point calculuntions mmd nmust be estimsmated and iterated in converging to tIme solution.

(I) lhe properties of Compommemnt I, CI, and Component 2, nC It), are read from Table
A.l in Appendix A.
Nnunnnt,cr Conumpt~ineni

MW .~/!.8!irol..

ic
~

Pc

amccntric

I
2

Mcmlruuiue
n-lkummmnc iemrg-Rtsbinnsonn

16.043
142.21(5

l9(t.56
617.7

4.599
2.1 It)

0.0115
0.4923

lIne

EUS lsmnrmunmnetcrs for fluid conmpoimemmts at T377.6

K are
h
4.27

calculated as follows,
Connp. ~ Fquamion x, a, /~gmot) 4.27 inn
~

a 4.23

MPa.(rn~
4.29 a, a

I 2

0.6
tI.4

tI.249575l7 5.71576(176

0.39665578 0.702753115 1.07281059 .52284853

t).l7538971 8.70423787

0.02680134 0.18935786

Tine liquid mixture paranneters. b and a, are calculated uisiumg the mixing rules, Eqs.(4.74) and (4.78), respectively. The binary interaction parameter between unethamne and n-decmmmme is read from Table A.4.3 in Appendix A: k =k =0.0500, and 12 11 k=k =0. 25 h
=

~ x h =0.6 X 0.0268013440.4 x 0.18935786=0.1)9182395 m/kgrnol 11 5~

~~x,x (a, 1

ms ) 1

l k,~)= 5

0.6x0.6x0. 17538971 x 1+ 0.6x0.4x(0. 17538971 x8.70423787) x(i-0.050)+

0.4x0.6x(8.70423787 xt). 17538971 )~X(l-0.050)+0.4x0.4x8.70423787x I 2 2.01923838 MPa.(mmm Ikgmmmusl) line ii mnmml resu It slusnuId mon depennd on the initially selected value.

(2) A bubble poimmt isressure of 27.58 MPa (4000 psiuu) is assunnned as the initial guess.

(3) The Wilsomm eqnnmntionr, Eq.(3.66), is used to estinmsmmne the equilibrium ratios at 27.58 MPmm, mind 377.(s K: K =2.457, mind K =0.0004684. 1 2 (4) The vumisonur cnrmriposimmum is cunlctulated nusilug Eq (3.43), y,= K,x,, rcstm Ii ing iii 1 Y= .474, amid y,=0.0lX)l 1(74. Note ttmat ly, is mmot equal to I wtmich only occurs mit time correct bubble poinrh pressure. Figure 5.1. Flow chart of flash calculations using equation of state.
Exanmple 5.2.

(5) TIne Pemmg.Rohiunsonm DOS, Eq.(4.25). is set-up for both phases. The dimensionless values of DOS punrammiemers are calculated from Eqs.(4.7-8).
Liqnmid Phase:

Calculate the bubble point of the fluid inn Example 5.1, usiimg the Penmg-Rohinssomn equation of state.

A=5.650l, anrd B=0.8t)67, whicln results in the following cubic equmatiomm for (he liquid coummimressihiiity fmmc(or, Eq.(4.3t)):

174 Z-O. I 9332Z-2.08487Z-3.38239=0

5. /lra.re Re/urn tort, CaIrulritro,z.c

.~.

I. Vapour 1_iquid Equilibrium Cenlculasion,c

175 (8) Now wills the nrew pressure and equilibriunmn ratios, steps (4) to (7) are repeated. The results of a few initial, intermediate and final iterations are given in the following tables.
lrcs.,MPa 2624 25.59 24.3! .~.,Yi..... 0.8932 0.9329 0.9774 ~,Ya...,.,.. 0.051(29 0.0426! 0.02225

The above equation has only one real root (Appemndix C), Z = I .0985 Vaponur Phase: A procedure similar to that of hiqumid results in A=h .0661 mund B=0.34tn8 for tIre vaposur phase, with only one real root for its compressibility factor ctuhic eqsnatisns. Zv=0. 89802. (6) The fugacity of each component is calculated in both phmnses. usinng Eq.(E4.5), 2~xa. hn~.=~1(Z_l)_In(Z_B)~ A b B(2.J2) a i J
~

Z
1.1)5 13 1.0285 0.9828

2 3 14

29 tner. No. 2 3 14 29
ff.

24.294
1 v

0.97777
1 ~ .

0.91638 0.92134 0.9289

0.022228
fV

0.98213 K, 1.5548 1.5841 1.6292 1.6296

0.92877 K, 0.10653 0.082677 0.055615 0.055569 ~mr 7.43E-02 5.05E-02 4.97E-07 3.580-12

hn(~~__~2)B) Z+(l+,J2)l)

MPa 21 86 21.52 21187 20.861

MPa 20.92 21.13 211.87 20.861

MPa 1)113383 1103246

002989

0.029834

MPa 0.04628 0.04182 0.2991

11.029834

(E5.2) where a, =(a,a,) (I-k,,). 1 The calculated values of fugacity coefficients. fugacities, and equilibrium runtios mire urs follows: x, Comp. P. MPa Equation I 27.58 0.60000 2 27.58 0.40000
.

y,

q~-

,,

4s~
05.2 0.9157

fi. MPa
$LXP

fV

MIu.

K,

05.2
1.474 1.3643

22.56

t~y,P 37.22

4m
1.4899

0.004)1874 0.003345

0.02295

0.03690

0.000186 0.14575

I)uring the iteration of Eq.(5.6), values of n >1 or n <0 maybe obtained at conditions where htsth phrases pbrysicalhy exist, (hue to umnreliahle K-values. The common method is to reject Ilmcse values and constimsume tIre ileruitioun wihhm new physicumhly acceptable values. [however, this camm disturb time convergence trend towards the solutiomi increasing the number of iterations required. this advnsahle ho continue tine iheration witim the obtained values as realistic data will generally follow after a few iherations, if tIne system is in the two-phase region. The values of 1 in> (l-Kmrmin) and n < (I-Knmrax) should, isowever, be rejected as these will yield negative phase compositions. This approach known as negative flash has been addressed by a number of authors li0J and investigated by Whitson and Michelsen [Ii]. The negative flash approacir is mnost effective for high pressure conditions, or complex msmixtures, where equilmbriummm ratios, Kms, depend on plrase composition and the iteration process can become quile extensive. Convemstional flasls cumlcuulahiomn procedures, suclr as tIme one described above, are for known tennpcrature and pressumre coirditions. Wlren Ihe phase change is due to a sudden or a steep pressure redunctmon, sunclm ms vaporisuttnon of oil or condensation of gas through wellbore or pipe restmnclnons, tIre systenmr enhlmurlpy nusuny he assumed.constant due to low heat loss. In such connditiomrs, tiasIm cmnlctmlatiomms mul us given pressuire enthalpy, instead of pressure-temperature, by invoking energy hmnlamrce equations arc requmired. Alternatively, a series of conventional isolhrernmnal flasir cumlcsmlatitsns mit variosus hemmmpcratures can be conducted, followed by calculation of nIne systemrm entlnalpy, ho find Ihe tctsmperature at wlmichm the system enthaipy is unchanged. Agrawmml et nil. 1121 comsrpared tire uuhove approach with the method which solves material halanmce annd energy equations simultaneously, and concluded that the former approach was mrrore robust, albeit slower.

Clearly the fugacity of conmponents are non equal in the Iwo phases al lIne mmtrove selected pressure. The resulting error value of ~( I
.

f / fV)?

ID is far remote

from the objective value of <It)

(7) The pressure is adjusted for the nexm iteration as follows:

At equilibrium, ft.
~yi

fV

=4~y P. Hence, 1
=

~fL/~vp=(h/p)~fL/4sV

That is, the pressure can be adjusted as,

1I~:

ir)

/~

Root Selection
=

p[~x~~i~,,

(ES .2)

where r is the iteration numnber.

Other methods to iterate the pressure wilts faster

Cuhuc equations of state (DOS) can he comnveniently solved by analytical or numerical methods. Tlrc analytical solution of tIme geumeralised EOS, Eq.(4. 12), is given in Appendix C. Both approaches are almost equally effective when optimally implemented [13]. Tire mathennatical hehaviounr of EOS for mrmultieomponent systems is the same as that for a pure comnrpoummmd. At a lcnmperahunrc unhove tIme psesudo critical temperature, it will provide only one rool for the molar volumsme, or tIne conmnpressihility faclor, at a given pressure. At temperatures below Ilne pseudo critical temperature, tlnree real roots for the molar volume at any pressure

convergence rates have also been suggested

191.

Applying the pressunre updating mehhod guvenn by Eq.(E5.2), the mmcxl pressure is. P= 22.56/0.9l57m-0.0369010.02295=26.24 MPa

176

5. P/nate Behariour Calculations

.5.1. Vapoumr LiquidEqailibriunn Caleulatio,ms Solution: The paranselers of DOS are deternmined for the equi-molar mixture of C, (Component

177

may be obtained. The pseudo critical condition can be determined using the same criterion as that for the pure compound. That is, (aP~
~aV)T_T

(a2P~
I~av)TT

(4.9)

The above set of equations provide the same relation between the pseudo critical pressure, temperature, volume, and the equation of state parameters, as described in Chapter 4 for puire compounds, i.e. Eqs.(4.lO-ll). It should be noted that, Ihe calcunlated horizoustal immflectiomm point by Eq.(4.9) does not occur at the true critical point of the mixtuure. At pressures corresponding to the maximum and tire minimum niolurr voiummmrcs predicted by
EOS, Figure 4.1, two of the roots will be equal. With some EOS. such as the Penmg-Rohinson EOS, it is possible to obtain a molar volume smaller than tIre co-voluimme. fIns shotuld be rejected as the co-volume, b, is regarded as the molar volume when time pressumre approaches infinity. The general behaviour of typical EOS has been detailed in [1 4j. When EOS, written in terms of compressibility factor (or molar volume), has three real roots, the intermediate root is ignored, and the lower and the higher values of compressibility factor (or molar volume) are assigned to the liquid and vapour phases, respectively. It is ohviouis that for multicomponent systems, where compositions of vapour and liquid phases urt equsihihriumn are different, only the root assigned to the phase whose commrposition has been used in determining EOS parameters, will be of interest.

I) and nC~(Componnent 2), at 396 K, (.ouuip. s, mm,. mrs a a b Equations 4.27 4.29 4,23 a, a 4.27 I 0.5001) I.0I77t)302 0.6(12(11108 0.95856673 0.97553625 0.05631263 2 0.5000 1.50486716 0.6704416 1.04728482 1.57602453 0.07243918 NIi aturd lsmurmunmieturs. Is minrd mu. mire cmnlcun hummed musing time rmmiitloimi mmmising rules. Eq.s.(4. 74) mind (4.78), respectively, with k,,= 0.18)33 from Tumble A.4.3 in Appendix A. h
=

~x,h,

0.064.3759

mn/kgmnuol MPa.(m/kgmol)

a,=a,,= (I-k, )(a,a,) =1.23.585538 2 a

=

~~x x a =l.2558l788 1 5 1~

NIPa.(m/kgmol)

The above values result in the foihowing~ dimensiommless param~etersal 3.86 MPa:
A=0.447h5879 and B=0.07547i77 4 Substituting tIre rsarameters in Fq.( .30) results in the following cubic equation, t Z-0.9245282Z +0.279 I 2728Z-0.027622=0 ihne mibirve cquuuutimumn Inns tlnrcc rcuul rtssts: 4=0.394179 Z, =0.280758 Z,=0.249591
J~

In some cases it may not be straight forward to identify a fluid as vapour or liquid. Wlscmr three roots are found for such systems, the intermediate root is ignored, and tine one which gives the lower Gibbs energy from the other two is selected. The selection of the smmiahler root identifies the fluid as the liquid-hike, whereas that of the larger root indicates mm vunpour-hikc fluid.
The change of Gibbs energy can be calculated using Eqs.(3.14) and (3.27), with fugacities determined by EOS. For example, using the Peng-Robinson or Soave-Recliich-Kwong EOS

with component fugacity coefficientsas,

A b ~ (I ~
i=n
N

Rejecting the intcrnmednate root. 4 and substitutiing o =i+~, and ~2=IEq.(5. 13) to obtain the expression for the Peng-Robinson 1 EOS, we obtain,
(0,, G,)IRT =-0.Ot)046

in

ln4,

or using the total fugacity coefficient givems by Eq.(4.l8), the systeimn mumolar Gibbs energy
difference at the two roots Zh and Z is determined as, 1 (Gb G )/ RT (Zh Zn)+ In(~~) A 1 llme deumsity is cmrlcuu Imuted mrs.

/a_bj/hI

(E4 .5) Z+~ B 2

z+&iB

Ilence, 4 rn.presemnns the stumble pimumse with a lower energy level, and the fluid is vuilxrur-hikc.

loll z~ +~iB)1Zh

p,,,=IItZRT)=3.86/(t).394 I 79x 0.1)0831 44x 396)=2.9742

kgmol/mnm

Zh-B

B(&,~ ) 1

(5.13)
RZn4~aBAZh~nB

M=~x,M,=5 1.109
p p~

kgmmrol/mol

M=152.0l kg/or

If the above is positive, Z; is selected, otherwise, Zh is time correct root.

Example 5.3. The Peng-Robinson EOS is used to predict the density of a single phase equimotar C and nC~at 396 K and 3.86 MPa. Apply the mninimummm Gibbs energy 3 the proper root. concept to select Whnen at a selected temperuuture-pressure, EOS gives one real root, thumt root will be expected to be tIne Correct root for time phase under commsideration. Phase behaviour calculations using EOS is an iterative process as compositions of all or some of the phases, hence the parameters of EOS, are not knowms in advance. The imsitially estimated composition for a phase may provide a wrong single root, as shown sclnenrmaticaliy in Figure 5.2.

mixture of

178

.5. Phase //elulr.iorur Calculations

.5. I. Vapour l.uquirlEquilnhriu,mr (~olculatjo,n.c

179

The pressure-volume relation, at a constant temperature, for the correct conrposihiomr of a vapour phase, as predicted by EOS, is shown by the solid line in Figure 5.2. Cheurrly the correct root for the vapour phase is Z , which is the highest of all. Time pressure-volumsre 3 relation for the vapour with an initially estimated compositions is shown by tIne dmo.Iucd line. As only one root is obtained, Z , it will be assigned to the vapour phase. The root is, however, 4 far away from the correct value. Tine problems associated with improper roots during iterative phase helnaviotir cumlcnilnmtions, dume ho poorly assigned distribution of components between the phases, lsave been addressed by several investigators e.g., t5, 161. At temperatures below the pseudo critical teirrperatsnre. where three real roots may exist, the correct rool cans he identified by comnlpulninrg Ilsenrr wills tire pseudo critical volsinre 116]. If time prcdicled root is less tisnuir tIme pseudo cut icmnl volrune, it is that of the liquid phase, otherwise that of the vapour phase. Pmhimig et un. 1151 lsunvc mulso stnidicd tlse above problem, including the spuriouns (lerivatives nmeuur time predicted two-plnunse 7.onrc botnndary. and Imave suuggcsted emarpirical criteriomr ho mivoid sckctinng wromsg r,s,rl.

Solution.

lire nratineunumtical behaviour of mm cubic DOS for a single phase mtmlticomponent mixture, wlnere its conmnposihiomn rcnrains nunchamngcd, is the same as that for a pure connmpounsd. Ilemnce, tIme nennmpcrumlumre at which the cubic EOS shows the inflection point is tIre mnaximsurn temperalure for nirree real roots. This can be obtained from Eq.(4.9), snrnuluur to that for a pure compound. The calculated value, however, is not the crihical temperature of nine mnxture, hum it can be regarded as the DOS derived pseudo critical
lenrrperuuunn re

(~

~
(~

Setting tine unirove Iwtr deruvatuvcs eqummul to zero for the Peng-Rohinson DOS results in two eqnuuuninnuis simnmmlmur to t:ri (4.27), as, mm,,, b,,
=

If compuirison of time predicted volsmnrse witin Ilnat of psesudo crmlicunl vunlue mndncmnned time pmcsemnce
of a wrong root for the phase under consideration, tIre estimniated vumlumes used inn cquilihriummmn calculations should he adjusted. For flash calcn.mhations mit a givcmm tcnrpcratnnre.prcssurc, increasing or decreasing the concentration of light compomment.s slmould pronmrole tine identification of that pirase by EOS as vapour or liquid. 1mm the near critical region, often only a single root is found for cads plmase. BoIls roots mrnuny well be liquid-like or vapour-like which may correct themselves (luring fimrtlmer itcratioims. Interfering in the iterative process by adjushing roots, mimuly have an adverse effect oms tlnc convergence [13.1.

0.457235R

0.o77796R(,,i~/~l,)

Connnisinnng tire twrr equannomns, uund clunninating tIme isseudo critical pressure results in,
rT,
=

0.1 7014t46(ur,,, I Rh,,,) 6~

(E5.4)

NoIc thunt mu,, is unlso a fnmnnction of lcnnnperuntnmre:

Tone Connp~msrlion

~~x x (us,a,, 1 3 where,

ml,,
=

a,um,.,)

l k,)

us,

[I

mnn,( I ~~(T /-i

)l

unmsd, m is detcrnn,immcd Ironnn Eq.(4.29).

4 Estimated Com~,silion~.~
/1
I

6 determnninned in Exummnple 5.3 omn a,,, rn, bm=0.0 43759 nmn/kgnnrol, R=0.0083 144 MPmi.nmsIkgmnrol.K. uund substihuting them in Eq.(E5.4), results mr run eqnnuutuonr wnltn tIre psisudo reduced lemperatunre as the only unknown, which
Usmnmg the rmrfornnimut non gives: =397.95 K

Jhe corresponrdnmmg pseudo crinical pressure is. _1_J________

z4zt

___________
~

_____

r~=3.9985

MPa

Comprcssitilily factor

Rapid Flash Calculations

Figure 5.2. Improper root selection due to inaccurate phase composition. Mumthenmatical naselhods and associated prohlenms in solving equilibrium flash equations have (recur extensively sttndicd 117, 3-8]. Punnticuilar considerations have been given to development of robust and efficient algorithms for applicationm in compositional reservoir simulation. Ins the numerical simsrumlation of reservoir processes, finite difference methods are commonly employed wisere the reservoir is described by many equilibrium cells. The fluid conditions in cads cell at each time step is detemnsnimred by equilibrium flash calculations, using compositional

Example 5.4. What is the maximanum temperature at which the Peng-Robinson DOS provides three real roots for a single phase equimolar mistsnre of C, amid nC,.

180

5. Phase Relnoviour Calcn,latio,ns

5. 1. Vapour Liquid Eqnnilibrinnnnm Calcm4latio,m.c

181
(5.17)

models and complex iteration procedures. For a large reservoir, the total mmtmmber of equilibrium flashes may exceed many hundred thousands, resulting in expetusive compuntation. Hence, the reduction of computational time of flash calculations is an imniportant consideratiomm in compositional reservoirsimulation.

a~~=nVaV+(l_n~/)at

The number of equations describing the equilibriums between tire pinases imscrcmmscs with tine

number of components.

refer to tine feed, vapour and liquid phases respectively. Si on larly for punranmeter, h, we ohlumi mm.
where indices F, v and L. h~=
+

Reservoir fluids are conmposed of msmany commmpoumnds. The connmmnomi method of reducing the computational time is to group fluid commmponcmrts (Scctiom~9. I), describing the fluid by a few pseudo conrponeflts. TIme mmnaims drumw back of unix mmnctlsod is time loss of detailed compositional information on produced reservoir fluids, winch is often required in the design and operation of fluid processing plants at tire surface.

I
I

(I

(5.18) paraimmeters of the liquid plnase can be detemnitsed fromni,

(5.19)

I lence,
a
I

for a set of uuv muumd h, the EOS

Michelsen [18] has shown that the number of equmations to be solved in flaslm emmiculmmtions, using a cubic equation of state, can be reduced to three for two parameter DOS, with little additional complexity, regardless of the nunsher of comniponenhs whmcmn inn hiuimmry intermictiomi parameter is used in the mixing rules (Section 4.3). For example, tire imsipleummenmiatiomn of tine method using the Peng-Robinson EOS is as follows: With no binary interaction parameter, the random mixing rule for the attractive term Eq.(4.78), reduces to,
N N

a ri n v,v a

and
= in

nvhv

a,

a = ~~z z~(aa )
1 j~i a 1

I is
Tine computational procedunre, as suggested by Michelsen, winch imsdepcndent variables aV. b and ~V, is as follows:

(5.20)

~za~
i=t

1 )
=

(a)

involves only tine three

(5.14)

where:

~z,a,

ammd a,

a~

Substituting the above in Eq.(E4.2), the fugacity coefficient can be expressed as, ln~ 1
=

qa, +q b~ 2

(5.15)

(I) At the given T amnd P. evaluate a~and h for emmch component and calesilate a~and h. 1 Dstiunrumtc time vumltics oh umv, 1 V umnd mv. t 5 (2) Calculate aL and b - from Eqs.(5. 19-20) and evaluate ~ L and ~V from Eq.(5. IS). 1 Calculate K. = 4,~/4~, and x and yj from Eqs.(5.4) amrd (5.5). 1 (3) Evaluate check fuumnctiomss,

where q , q~, q2 depend on1y on the mixture properties, a, b, T and P as, 0 q =lnl 0 Ip(v_b)] I RT a ql~( 12~ ,) Iv+(~+I)h1 L~+(l_~h)bi
.

e~ = e 2
=

~(y,
_____

x,)
h

~y,a,

a
-

e~--~-i

1 ( Pv ~ a q _~1)4~q 2 The molar volume, v, is determined conveniently using EOS at given values of a, b, T atnd p~ Multiplying Eq.(5.2) by ai, and summing up for all the components, we get, ~ ~ or, (5.16)

If any of time above values is more than its selected tolerance, perform an iterative co~ctionof a, b amnd n~, then return to step (2), otherwise the solution has reached. Figure 5.3 equilibriums shows the variation ofof the CPU comntacted time vs. the number components for predicting conditions a computer blac~oil when with a richof gas at 373 K and 20.79 MPa in fouur counsecutive stages, simulating the advancement of injected gas in a reservoir [19]. TIme number of components describing the 25 component mixture has been reduced by groupimig them. The figure demonstrates that the CPU time required by the conventional fornnnniation, using a quasi-Newton method, is about two orders of magnitude Imigimer than time Micinelsen unsethod with no BIP when the mixture is described by 25 conspomsetuts. Note tinat tIne commiputmitional time decreases sharply for the conventional nniethod when the number of conmponents describimmg the fluid decreases, whereas the reduction for time Micimelsen nsetlnod is insignificant. Ihe reduction in computational time by employing the

182

.5. //r,r ~r !?c/,,rr,,,u, ( ,rl,rul,nii,,,rv

~ 2 .Stolr,li~.fl,niuln~ 01

183

Michelseum method depcnmds also on the matisematical methods used to solve tIne set of eqtuatiouis. The inmproveunent could be less striking wills mnretimods more efficienst tirmrmn tIne omme tused ins tire above cxunnmple.

5.2 sIARILJrv ANALYSIS

lIne rcqsuiremsment of minninrnsnmrn Gibbs eneugy, at eqtnihihriuum conditions leads to the umniformity of lIme clmcmmsicunl potential, Imence, unifornrnity of tine fugacity, of each component throughout all co-exishimng plnases, as described in Clnapter 3. A set of nnmaterial balance equations and equmaiity of fumgacity, similar to those given iii tIne above section for vapour-liquid equilibria, are in general undequahe ho dehemmine equilibrium conditions in tniost petroleum engineering problems. Tine imumber of phunses at equmhihriumm. Irowever, nsnuist he known. It us not generally difficult to idenlify mm reservoir fluid as a liquid-like or a vapour-like mixture froni its cousnposilions, with the exceptiomm of near critical flunids. Hence, the satuuration pressure cult he cmilcunluuieul rut mu givemm Icnnnlrenuntune. mis slunswmn inn tIre rmhovc section. The fluid is conmsiu Ierud ltu turin u t svo Inhmurses, except br lcann gums nnmixl tires, below its satsmrumtion pressure. ihuis uuppromnclr simonnld ssnfficc for most fluish calculuutionns. I lowever, one can consider many cases wlncrc mrmore llrmnms two plnunses uunc present at eqnuilibriumnn. The occurrence of three pimase systcmuus in riclu ~ fitmids, umnrd prccilsilation (if unspinmultennes mmn(l waxes are commnnnon examples. Wmuher is alurnost umlwurys present no ecervusirs as a liqumid phase sepmlrurte fronn line hydrocarbon ricim liqtmid plrutsc. tire prcscmmcc of waler mit low tennrpcrunlsmre csnditions can lead to fommation of a trsnmmmbcr of solid plnunscs of Imydrates mnund ice. hlemmce, at a given temperature and pressure, the nmtuinmher of pirases mnmmuy not alwusys he known in advance. TIre Gibbs phase rule, discussed in Section 1.2, clsses nust imnrposc mnny pruuclicuml limitation omn the number of possible phases, in real reservoir fluids, as they mure Comnrpose(l of nrranny consrpomnents. FIne rigor(ssis mmretluoh of delcrnnninnimrg tIme cqumilihriuuumm mit mm guven prrs.cure and temperature is to find the conditions ut whuicir tIne Gibbs energy of tine system is at its global minimum for all possible co,,nhipnaius,n of plna.se.c and component dislrthulion. This was expressed mnuallrenmnuthically inn Clnuuptcr 3. as (~JG)~.~ =0, mind (~~~)n.r >0 (3.9) (3.8)

00
1(0
Mi,hetscn Cnnvcnrional

-. -.
. -

a
,j

60

40

20

tO

15 Number of Conrrprrnenus

20

25

Figsirc S.3. Comparison of computatiomnal time between the cooventionuil uund Micliehscnn metisods. The proposed method of estimating the parameters of the vapour phase to initimnlisc cmnlculations, described in Step (I) above, nnay head to lack of convergence for gas comidensunle systcnmss. In tm flasln calculation of vapour-like fluids, it is advisable to estimate a~-, b insleundl of tlnsse for vapour. Then the valumes for the vapour will he calculated fornn equurtionrs eqtnivmulemnt to Eqs.(5. 19-20) as follows,
aV=

aF_(l~~nV)aI

(5.21)

ilne umnimnimimisurtion of (iihhs cnmcrgy to delcnnsine tIme equnilihriunm conditions, and to avoid false snsltmnionns, cams he illtnstruuted gcoumietrically by exannining tine Gibbs energy surface for binary systemsns 1211. Comrsider a himmutry systeurn wmtls lIme conrspositionn-pressure diagraisi, at constant temperature, as shown ims Figtmre 5.4. Note tlrah dcpendiimg on the mixture composition, xF, and pressure, P. tIre sysleos can formsr ommc, two, or three pimases of vunpoumr V. Liquid LI, and Liquid L2, at eqnmilihriuin. Else clnemmnicurl polcmrh imnI of emucir consup(nnnemrt wutinimmmi phase is defined mis,
=

and

by

hr _(l_nv)bI

(5.22)

The cireck functions in Step (3) sirould also be chanrged accordingly, mmd cvuuluualed for the liquid parameters. The above mnodification innproves the rohustnmess of tire Michrclsen mnnetlrod

for gas condensate fluids 1201.

The implementation of zero binary imrteraction paratmletcr for isydrocarhomr-hrydmocarhon compounds of reservoir fluids is quite reasonable, as pointed out ~nSection 4.3. Ilowever, for fluids with significant concentrations of non-hydrocarbon compomsemrts, suds as C0 , tire use 2 of binary interaction between non hydrocarbon-hydrocarbon is required. In stucir cases, for any non-hydrocarbon component two additional equations will be included wInch must he solved simultaneously with the three equations. The advantage of the simplified method is expected, therefore, to be minimal in such cases.

(~G I ~mr )nt,

(3.15)

Ilcnnce, tIre nroluur Gibbs energy, g. of cmuclr phuumsc cans be calculated from the above definition as,
g
=

x p, 1

n 2~ 2

(5.23)

wlmere, x , is the mole fruuctnon,

184
=

5. Phase Relnauiou,r (aleudatio,u.c

5.2. Sen hilit A,malvsia

5

185

n /n 1

(5.24)

Substituting tine ctmemmsicah potential by its equivalent, that is the derivative of Gibbs energy, in Eq.(5.23), we ohtaium,
=

where n is the total number of moles in each phase, and g = G / n

g

The chemical potential of each component, relative to a reference value, can be calculated by thermodynamic relations, Eqs.(3.27) and (3.31), using an appropriate DOS.

x (g/~x ) 2 2

(5.25)

and

p~ =gx (r~g/ax,)
1

(5.26)

Thmut is, a line tangent to the Gibbs energy curve, for exaumiple at point A, intersects the Gibbs energy axis (erected at x,= I) at a valume eqsnal to tIme cheimsical potential of component i at the point of tangency, as illustrated in Figsire 5.5 by ~ and ~ I lemmcc. Ior tine litre Imningeunt to mire Gibbs energy surface mit two points. B and 1). we get.
U U

(5.27) and (5.28) Tlnat is, time hiqsuid phase LI at point B, mind the vapour phase V at point D, Figure 5.4, are at equilibrium. Composition,
xi

Figure

5.4.

Schematic pressure-composition diagram of a binary mixtuire at constant

temperature. The molar Gibbs energy of the binary system, as a hypothetical single phase fluid, at pressure P1, calculated by EOS, is conceptually shown in Figure 5.5.

Now, conmsider a imsixture at point F ins Figure 5.4. If the mixture remains a homogeneous single phrase, its Gibbs energy is equal to gFas shown in Figure 5.5. If it splits into two equilibrated phases of B and D, the mixture Gibbs energy will be reduced to the value at F, tinat is gF~.

~gF

~ttgR

0 9 n g

(5.29)

Since the mixture F emits attuuin a lower level of emnergy by splitting into two phases of B and D, it nmunst be unstable as a single phase. lhe nunsber of moles of each phase at equmilibrium can he determined by immaterial balance (lever nile),
4

mm

mm
is(n

(5.30)
X~)

mnmn(x~
SQ U

x~ )=

(5.31)

02

The mixture A, Imowever, is a stable sinrgle phase because it cannot split into the two phases of B and D, lowering its energy level to mlmat at A, due to the material balance restriction, nun(x~5 x~)= nm2(x~ x~) (5.32)

which cans only be satisfied if the mole number of one of the phases is negative, a physical impossibility. Composition, 5
I

Figure 5.5. Variations of calculated Gibbs free energy with composition at pressure P1.

The above observation can he used to eliminate non-existing phases in flash calculations. II was indicated in Section 5.1 that during vapour-liquid equilibrium calculations, phases with negative mole numbers may appear tennporarily during iterations prior to converging to the final solution witim all of tlmemim positive. however, an ultinnate convergence to a solution with a negative phase mole imndicates tlrat the system is an under saturated single phase at the given

186

5. i/,ase lleha,iuur Cale,n!atio,ms

5.2. Stability Analysi.,

187

temperature and pressure. When iterating on the vapour mole fraction, ns, a negative value indicates an under saturated gas, a value higher than one indicates an under satunrumled liquid. The calculated molar Gibbs energy of the binary mixture, assummed single phase, at pressuire P2 (Figure 5.4) is shown in Figure 5.6. A single line can be drawn tuungemnt to tIme energy curve unt points E, M and N indicating the Three equilibrated phases of LI, L2, and V respectively. Amy mixture with a composition between those of E and N will split into the llrree equilibrated phases with the amount of each phase determined by material halansce. For example. tIre mixture F will lower its Gibbs energy from gF to gt~, by formninrg tlmree co-cxistiurg phrases. Note that a wrong assumption of two-phase equilibrium for the F rnixtumre, ssnchi as E-M, or FN, will also lead to the same overall Gibbs energy level of gF . Ilcncc tIne Gibbs energy minimisation cannot identify the nunmiber of phases mm thus cuisc. For a bimrary sysncnru witln mInce equilibrated phases, the degree of freedoun is one, thence, the three plnumse cquihiirritunrn will occur only at P1, fixed by the selected temperature, as a snniqume solution.
A

so

LI) .0
.0

5-)

Cn,mrrNrsituon. 5~
A

Figure 5.7. Variations of 2

calculated

Gibbs free energy with composition at P3.

so
>~ so 1, C

gr gi

ma

TIme hcternsination of tIme Gibbs energy surface and all tine associatedtangent points is relatively mm simple task for himiary mnnixtunrcs, hut trot a practical propositmomn for multicomponemrt systems. line hammgcmmt plamne criterion has. Irowever, been umsed by many investigators to develop nummmerical imirptcmnscnhuutuusnn of stumhihihy analysis[2224]. Tine mnnctimod developed by Michnelsems 1221 cain be successfuulIy applied to various multiphase cqsuilihriun prohlenms. line tmnetlsod is sclncnrnumhicahly as follows: :~t.ul
0

I. A plamre is drawn hangenrt to tIme Gibbs energy surface at tine feed composition. 2. A seconmd phase is assumed to be present. Its compositiomn is so determined that the tangent plane mIt tlmat point nm tIme Gibbs surfuice becomes parallel to tine first plane. 3. If tire second plaume, for all possible compositions. was found to be above the first one, the original nnixture is stable, otlscrwisc it is considered to split into two phases, and flash cumlcumlatiomrs are sumhsequenhly pcrfornmed. 4. For omme of tIne plmases ohtaiured by equilibrium flumsh calculations, the above steps are rcpemmned until all tIme phmuses are fosmmrd to be stable. As the previously determined phases share tIne smurne tangemnt plunnne, lIre snurhilily ununmilysis should yield identical results irrespective of the selected phrmise. time crucial clemsmemrt in time above procedure is the second step, that is determination of the second tansgent plamne. It was dcnnonnstrmuted in Figure 5.5 that the tangent line intersects the (;ilmhs eurcrgy mixis unt a vuiltue equmil to the clrcmnnncal potcmntial of that component at the point of tuunugency. The (hislunnmce hctweenn tIne two parallel tmmngent lines, therefore, is equal to the diileremnce between tIne clrenmnical polcnntials of each consponent at the points of tangency. Note thrat tIne difference is tiue sanume for both conmponeumts mis the lines are parallel. The same conclusion is valid for nuulticonnponent nnixtures, that is, i=l,2 N (5.33)

Composition, X;
Figure 5.6. Variations of calcunlated Gibbs free energy with composition at pressure P2.

Figure 5.7 shows tine mixture molar energy as a hypothetical single plnase at presstmrc P3 (Figure 5.4) for different conspositions. A feed witim composihion F can split imrlo two different two-phase systems at equilibrium, as shown by the Iwo tangenrt lines drawns to the Gibbs ensergy curve. Both ideotified two-phase systemns, satisfy material halansce mind equality of cisemical potential requirements. Although in hotim cases the nmsixlurc energy is reduced relative to tIre lrypotinct ical xi ogle phrase cond it ions, only tIre ptruiscs mut K uurd Q, mmd iemut Inrg I - I ml mn(I I .2 I uqsu md pluuuses, represent tIne tnme solutions, witls nhuc nmnixtsmrc cnrcrgy at its lowest possiislc vmuluc of g~. All tire above example.c lead to this consclusion tinat a mixtunre rcnniuuimms a slmnhle sinrglc pInuise when tIre Gibbs energy surface is concave upwuird at all connceotruilionss. Otlnerwise, tluc mnxture nsnay split into equilibrated phases indicated by tIme points os tIme Gibbs energy curve with a connunon tangent. Amirongst all the tangent points which suntisfy nruunlerial balance equuitiorns, only tlnose by the tangent which identifies the lowest energy level mit tIne mixture compositioru correspond to the true solution. The above conclusion, reached by examining a binary mixture, is equally valid nnulhiconrponent systems. The geometrical illustration, however, would he innpracticuml, as Gmhhs energy curve becomes a hyper sunrface with tangent hyper planes. Neverhlselcss, analysis indicates that tine tangent plane criterion is a necessary and sufficieumt conditnsn
equihihritunir.

for time tire for

wlrere, X, is tIme distance between the two parallel planes tangent to the Gibbs surface at the two pirases of y, and x. If the nrixture x, is to be stable, the value of X, must be non-negative for mull possible comnmposilions of nnixture y.

188

5. Phase Fk/mausuuir (a!cunluutiunus

5.2. Stability Analysis

189

Substituting for chemical potential, Eqs.(3.27) and (3.28), in time above,

RT[hn(Py 4~)_ln(Px 4~)J=X 1 1

i=l,2

(5.34)

mixtures. Examples can be found where suuch initial guesses nmay nniss a possible solution. Furtlmer informatioms on tIme selection of initial estimates and associated numerical nmethods of solving Eq.(5.37) are given by Miciselsen [25]. An additional benefit of applying time above procedure is that when the hack of stability is verified by converging to a solutioum with > 1, the converged K values may provide 1 appropriate initial estimates for suhsequment flash calculations. This provision could help equilibrium calculations near the critical point where close estimation of initial equilibrium ratios becomes essential. Ilue umbovc alrpromuch of stabihihy uumruulysis winch starts withm a single phase may become quite tune cinmrsulmnniung whucmr minute thmuimn two islnmmscs mire expected to exist at a given pressure and teunrpcruim tire. I or e suumn ilile, iii I ranspoitmitmuir of iumm~miocessed well stream fluids through a pipchimmc. uit lemust llmrcc ptruuses of vmulmounr, Imydrocarhon liquid, and water are generally pre.seimt. Al low tcmmnpcmumtumre conditions, suds mrs at time semi bed lcmnnperatumre in offshore production, lime formnnalmum of lmydruute, wax mrumd musphmahtenc phunses mime a strotig possibility. An equihibriumns stumuly of Ilsis problems umccessitmutcs time existence of at least four phases in typical pipeline conditions. An alternative uipprouncln 1261, based on the tangent plane criterion, iS to assume a reasonable mnuiximnsum nummnher of plnases at equilibriums, and then search for the non-existing plmases using materiumi humlance restrictions, as discussed in the negative flash approach.

where y,, and x,, are the compositions of the searched mixture, y, amrd tlre feed, x, with fugacity coefficients 4~~ and n~X,respectively.
4

Expanding the above equation, and defining a new variable, Y , mis, 1 Y~ =y exp(~./RT) 1 where, yl ~,Y/~Y we get, hnY +ln4u~Inx,ln4~=O 1
The mixture x is stable provided that indicates a non-negative ~. At ~
=
,.

(5.35)

(5.3(n)

i=l,2

(5.37)

Stability Limit
I at all solutions of the above equation, as it Tine fonnation of a new phase is generally preceded by some degree of supersaturation. The bubble nucleation in a liquid at a pressnr.~below its bumhble point value can be inhibited to a large extent by expammdimmg the liquid gruidually, avoiding fluid agitation, and ensuring the lack of umsinumte gas pockets inn the liqunid prior to the expansioun. Such systems are metastable with an energy level which will be redunced by fornming a new phase. Figure 5.8 shows tIme variultioru of total volume with pressure for a pure fluid at temperatummes below its critical point. At Ienrperature TI, the reduction of pressure results in a stable expansioum of the undersaturated fluid down to A, i.e., saturated liquid. Further expansion should result in gemnermution of an additional phase B, a saturated vapour, fornming a stable two-phase system at equilibrium with liquid. The formation of tine vapour pinase at point A mmnay. however, be inmhihited or delayed, resulting to further expansion of the liquid as a metastmnhle phase, sinown by time (lotted line in Figure 5.8. A similar behaviour can be eurvisaged by startimmg with a siungle plmase vapour and compressing it to a volume lower than that at B. nm

1, X=0, and the two tangent planes are identical, that is, the mixture, x, is at its saturation conditions in equilibrium with an infinitely small amount

of phase, y. The method of solution of Eq.(5.37) is similar to the conventional two-phase equilibrium flash, as it can be presented by, Y /x =~/u~= K 1 1 1 (5.38)

It is not, however, restricted by material balance equations. Note that Y is not the mole 1 as described by fraction, but can be interpreted as a measure of the mole number in phase, y, Eq.(5.36). Locating all the solutions to Eq.(5.37) requires a global search. For a two-phase vapour-liquid analysis, the composition of the new phase tsnuuy be estimated from,
=

K,x

(5.39)
A

or Y 1
=

x /K 1 1

(5.40)
0,

for liquid-like and vapour-like mixtures, respectively. For near critical fluids wisere the nature of fluid cannot be clearly identified, both estimates should he considered, as one could

converge to the trivial solution of y=x.

The Wilson equation. Eq.(3.66) can be used to estimate K values for vapour-liquid equilibrium, but for liquid-liquid, or solid phases otherestimates are required. Michelsen [251 suggests that the initial estimate of the new phase composition is not crucial, and pure phases only composed of the lightest or the heaviest components can be selected for hydrocarbon

Vohnmune

>

Figmure 5.8. Stability limnmit for mu punre fluid.

190

5. Ihase Be/nanwur Caleunlatio,ns

5.2. .Stab,hiy Ana/v.crs

191

Assuming a continuous transition from one phase to another, winch will not plnysically occtnr except at the critical point, a minimum and a maximum pressure, at poimits D and F respectively, will be expected. This behaviour is also predicted by a cubic equation of state, when time phase change is neglected, as showmm in Figure 4.1. The expaumsnonm of tine liquid phase passed D should result in increase of pressure which is not nncchuusicmulhy feasible. A similar argument is valid for compression of thue vapour passed E. Tlnese points nrc reIcrred to as the hunnits of intri,n.sic stability as timey indicate thue houundary for ml urnetuu,stuulmle simrglc tluase fluid, line curves formnncd by the two limits at various temnnpermntumres below tine criticurl temperatumre, as shown in Figumrc 5.8, are called tine .spnooc/al comes. Note nlmunh tIne gmup between the upper and lower imntrinsic stmnhiIity limits decreases witln temoperuittume unnrd vuimnishnes attire critical point. Indeed, tine slunhilily hinnit nod lIne phnunse houmnmdmiry coinrcide mrt tIre critic-md point. The muhove descriphion of mnecinanical sturhility for ml pure sumhslanrce is mm spccimnl curse nsf tIne general criteriotr for shuut)ilihy mrs givemn by Fq.(3.9). line immnrinrxic stuulsitmty liuuuil , is tIne cunuditiomn

Exanmple 5.6. Prove tlnat 11w unccinanicat stability hinsrit for a pinre compound, as described in Figure 58 cair he derived by tine genreruil energy concept. Find the stability hinrit of normal inexumnne al 473.0 K, using uhe Soavc-Redtich-Kwoing FOS (SRK).
,So/mution:

Dcxc ruhunug nine snmntrnlnuy tm,nsut cm uncruonn, Eq.(~.4 I), urn tcrimns of mIsc Helmholtx emmergy, Iat.( 3. I 2). wmntn vunrnuuhlcs ot tenssperuuutnre unnd voiuumnre, we obtaiur,

where,

(IA at which the reduction of Gibbs energy isjust to be violated. tlmat is, 2 tI G = 0 (5.41)

Suhl

ldV + ~~t~Inn,

For a punre comnipound at coimshaunn tempcratunre lIne above reduces to, dA~ishV
I

Figure 5.9 shows the stability limit for a binary system on lIne Gibbs energy plot mit crsmmstamnt pressure and temperature. Note that while the nnixlure F mmiush form two phmurscs of B uunsd A at eqtnihihrium to be stable, it can remain as a mnelastable simmgle phase. Ann increase of x will force 1 the mixtunre further into the metastable region till it reaches poinst N, whncre mm iimtlcctksms point on the g curve exists. That will he the hiusmit of intrinsic stability, mmd hIre nsmixtcmrc nnuumst split into two phases of B and A. The above criterion cans be expressed in ternsms of other energy [saruumrncters, sumuin mis lIne Iheimhohlz or the internal energy. They are all equivalent to describing the tlnermrrodynmummnic stability limit as the bound outside which entropy will decrease in real processes, a violation of the second law of thermodynamics, see Section 3.1. Tine detemninmation of intrimnsic stability hinnnit provides immformation on time limsnit of supcnsatunratiour of a reservoir fluid in a depletion process. The condition of intritssic stability is a suhcase of the tangent plane condition. Whereas in the latter, a global searcin is requuired. tIne fornnier cutn be identified simply by evaluating the g surface at the feed conditions. A nnmuin applications of determining tIme inntritssic stability limmnit is inn determuuinmulnonn of tine criticuul point by an equations of state. It was noted in Figure 5.8, that the hinodal curve and tIre phrase envelope meet at the critical point. This feature has been used succcssfumlly to detcrimninse time critical point of nsulti connponent systemns, as both the hinodunl cuurvc and tine plnmusc envelope camm be expressed by energy terms, sinsilar to those inn Eqs.(3.8), amid (5.41), mmmd rigorously calculated using thermodynamic relations.

Ilcnrce, 2 2 (d A/~V )~ =
_~

(t)P/t)v)~ =0

mm I Iran us. ntuc stuuhilnny humus for the vunpoumr aurd lidiniud phases of a pure compound lie at lIre nrrmnxnnmiumnm and nmnnnnmruumnl pressunre values, respectnvely. on the isotherm as described by FOS. Calcunlatimng lime derivative of pressunre with respect to volume at constant temperature, umsing SRK, we ohtminn. 4 v +(2Ir

2uu/Rl)v

2 + (h +

3mnh/ RT)v

ah

tm

/RT

lIne FOS p~nrunnrmdlcrsbr nucsnnnn,uI lucxmmne are cuilcumlated at 473.0 K as follows: Tc

l~rluiatnOn

Pc

on
. -

m~ m ~1I~.(m~ ~mnno~)~_, 4 22 2.5l70t2 4.25 0.938436

a
.

a
a, a 2.683527

h
4.22 0.120877

5t)7.6

3.1)25 1)2659

4.23 t.066I55

Snnbstittmlinrg nIne vmulues sf a mnnrch h inn the above equation results in,
-..t .

v- 1.1229661 v.i-0.26205757v.O.OOl 20518=0 wittr the roots as:

~

LI

A
.0

.55
1) x 5

v=-0,06014 v~=(>08275 nn/kgmol nmrVkgnmol v,= 0.30412 mmm/kgnnmol v = 0.79623 no/kgmnmol 4 TIne firsl two roots are non acceptable, i.e., one negative and the other smaller than h=0. 120877 nsn/kgmrsol. whereas the tinird and fourth roots represent the volume limits

Conrposilion, 5 u Figure 5.9. Intrinsic stability limit of a binary mixtuure al constanl pressure and lemperatn.nre.

192

S. /lna,ce flu/nainonj, (a/snu/atinu.s

5..?, (~,uIs(al!oi,nt Cah-uilatioums

193

for liquid and vapour phases, respectively. The associated pressures umt lIne stmmhihity limits can simply be determined from SRK by substituting tine values of vnnlumne ammd temperature, 1 5 P =0.6995 MPa P =2.l48 MPum 5.3 CRITICAL POINT CALCULATIONS Variations of fluid compositions in a reservoir mmny result in tIme critical stumte. Practical examples are miscible gas displacement fluids (Chapter 7), and flumid colummmmns ins reservoirs exhibiting severe compositional grading with depth. which will he discussed imm Scctiomm 5.4. There exist numerous conshinations of mixture comnslitunents wlnicin pioduice crimicmmh tiuids mit mmmsy given set of pressure-temperature in multicomnnpomment systems. The critical pressure mmmd temperature of a mixture with fixed composition, however, are unique and can be ktcnmtilied relatively reliably both experimentally and also by variotus predictions rnuetlmxls. Determination of the lrue critical point of a reservoir fluid mmrmmy nol be of mnntnciu imntcrust to ml practising engineer, but it can he quite useful in studying the belnaviouir of nncar critical volatile oil and gas condensate systems. Calculations of vapour-liqtuid equilibria for these flumids neuur the saturation pressure is difficult, particularly if the equation of state describing both phases is to be calibrated, or so-called tuned (see Section 9.3), against experimentud datum. A kumowledge of the critical point can be used to identify problenn areas and avoid unnecessary numunnerical complications. The pseudo critical properties of a nnixture are conveniently calculated by mniixing the critical properties of its constituents. The most common approach is mnolmmr mixing (Kumys rule), ~0, =~z O, 1 (1.14)

I-Ieidcnmrmrmmn mmmd Kimahil 1311 proposed applying tine criticumi point criteria as stated by variations of I Iehirnlnoltz ennergy, immstead of time Gibbs energy, becumuse it leads to less comnplex equatiomss when pressure explicit cubic EOS are used:

A A L= A
21

A , 1

...

0 =0
(5.42)

01

A , 0

...

00 A
A

A
M=

0
=

(5.43)

55

where,

(a2A

~Jms.
J

(5.44)

linumt is, ~ mue time second chcrivmntivcs of tine total Elelnnmholtz energy with respect to moles at constunmnt teurtperatumre mmtnd total volume. The definition of the lleltsihoitz energy is given in - Eq.(3.l2), and it can he calcunlated, simnilar to the Gibbs energy, using EOS. Later Miclrcisen mmmd Iheidensumnits [32] proposed a more commnputuitmonally efficient procedure to solve tine above equnurtioums pamticumlarly when all binary interaction coefficients are set equal to zero. Tine comnputurtionah effort in determining the critical point by EOS is comparable to that of an equilibrium flash [13). Model fltukhs, oflen binary systems, are frequently tested in laboratories to investigate fluid hehuiviour. Estiunsurted critical properties of these siunple mixtures are useful in designing mixture compositions amid test conditions. The critical temperature is particularly valuable in designing tests to simnuhurte liquid-like, or vapour-like behaviour. Simple empirical correlations arc adequate in most curses. A sinniple nnethod to estimate tire trn.me critical temperatture is to use the Lis mixing rule (33):

where z~is the mole fraction, ~O, is any pseudo critical property, studs as temperature. pressure, or volume, and O~ is the critical property of component i. Tine true critical
properties are, however, different from the pseudo values.

The concept of excess critical property, defined as the difference between the actual value and the calculated value by the ideal mixing rule of Eq.(l.14). has been used to predict criticurl
properties. Teja et al. [27] used a modified Wilson equation successfully to describe the excess properties. Mixing rules, incorporating experimentally determined interaction coefficients, have also been suggested [28] to predict critical properties. The main disadvmuntage of all the empirical methods is that the predicted critical properties may not be consistent witir the fluid

phase envelope as predicted by EOS. Therefore, they fail the whole purpose of tire exercise.

Rigorous thermodynamic methods, using EOS, can be applied to predict the critical point. Tine approach is not only more reliable than using empirical correlations [29), hut also provides consistent data with the phase envelope predicted by time same EOS. The Gibbs criteria for the critical point lead to the semrrch for a stumble point wlnicin lies omi time stability limit as schematically shown in Figure 5.8, and expressed by Eq.(5.4 I). Peng and Robinson [301 applied the method by calculating the Gibbs energy derivatives using the Peng-

T,

~XT

(5.45)

time cifcclivc cotmecnmtrunl ionr, X , is deIimmc~as, 1 (5.46)

Robinson EOS. The method was evaluated by predicting the critical pressure and temperature
of 32 multicomponent mixtures, which resulted to the average absolute deviations of 2.33%

and 1.3 1%, respectively, in comparison with the experimental data. The metinod requires the
evaluation of third derivatives of the Gibbs energy with respect to concentrations of components, which fornn elements of a number of high order determinants, hence,

computationally quite demanding.

194

5. Phase Fleluii iou, (ithulatjsou.c

5 1

(iou-al Point (aleu!atio,m.s

195

where z and v are the nmohe fraction and critical molar volume of component i, respectively. 1 1~ TIne method was tested for 135 binary hydrocarbon mixttnres, with aim average deviuution of less than 4 K, irncreasiumg to IlK for mulhicommnponetml systems [34) Empnrtcal correlahions generally fail to predict the criticuil presstmre of multicommiponemrt syslenrns as its relation with the concentruution of fluid components is highmly non-hioemur. Figure I . I 2 clearly demnonstrates such conrplexity as the mmnixtrmre critical presssume is mnnosnly Iniginer thrmmmu those of tIme conrnprising connmposundls. A simplified correlation of Kreglcwski and Kay 1351 can be used to cstinnumutc tlne criticurl pressure as follows:

Snnhstilnunirng tIre above values ins Eq.(5.47), we get P~=8.22 MPa which is in reasonable umgrecmsment witin the reported vunlue of about 7.80 MPa [36). The pseudo critical pressure of 4(12 MPmi is nrnumrkedly deviated. Applyimrg Pr=Z,RT,Jv, and using the nruolumr unvdragl comrnpressihility faclor of 0.2720, uunnd lIne critical volume of 0.2585 rn/kgumroh result inn an unacceptumble value of P,=4.03 MPa. 5.4 COMPOSITIONAL GRAI)ING

I)~ =~

(5.808 + 4.93o)[~~
1

I]]

(5.47)

wimere p~c and p c uure the mixture pseodlo crinucuml pressure anrd tcnrupcrurhn.nme cmulcmnlumted by Eq.( 1.14), respectnvely, and (I) us Use nnsoluur avcruuge acenntnic factor. The correlurtion estimateh the critical pressure of 967 mixtures with ats averurge error of 1.3 bar [29]. Mixtures containing methane were not inclunded in the cvmiltnmntiotr. hecurumse of lIne lmrck of reliability of the correhatnon, stmilar to other available correhuitions, for such syshemmns. line locus of critical points of binary mixtunres as shown inn Figure 1.12 or Figtnne 1)1, Appcmrdix I). cunn be used for rough estinualiotn of line eniticuul pressure of mmnixlurcs commtmninnimsg usnellnmuine. The simple molar averaging of critical compressibility factors of flunid componmcnls gcnrcruilly provides reasonable estmmates for hydrocarbon mixtures. The estininated critical comuspressibilily factor, temperature and pressure can be used to predict the nsixtumre critical voltnmrme,

It wmus ~ si msted omit IInuit lateruil ammch vertical cotnnpositmonral varimutions witlsin a reservoir are expected dumring tine early reservoir life. Oume mmiight expect time reservoir fluids to have attained equilihriunii nut nsatlmrity due to molecular diffusions and mnixing over geological times. itowever, tIme dhiffUsive mnmixing may requnire many tens of million years to eliminnate cornspositi(snal heterogenumities. Whemm tIre reservoir is comnsidered mahuire, it is often assumed nlsunt fluids ase at equilibrium. At tnnnifortni hcmsmperatuure anmd pressure, that is, when the thermal mnmnd unmeehanicumi eqtuilihuiimtn mmre esnurislishned, the above mmssumphion leads to the uniformity of tunguneihy of curds counnpurnrennt I Irronmgirsmmt all co-cs ntinng plsases mrs the requmreutrent f(sr the chnenmnicmil eqtnihihriunnnn (Section 3. I). Fusr a single plmase fluid, tine umnifommity of fugacity is eq iiivurlcnmt to tIne munmi fornnn ity of Cource mutrunt ions.

The pressure umnh lcnumpcrunltmnc, luswevcr, umre not tmnniform tlmroughmout a reservoir.

The lennpcrmuhuumc inucreumses wills deptlr wills mu gradiemmt of umbotit 0.020.03 K/meter, and even mumch luighier ins cxtrenssc cunscs. lise prcsstume also clnaunges according to tIre hydrostatic head for fimmids mnn rest. Ilserefore. comrupositionrmnl vunrialiomms witlnins u~reservoir, particularly those with a tall eoltu munnr oil fbi id shromn Id he expected.

Tumble 5. I simows tIme lluuid comunposilion of a Nomth Sea reservoir at dnfferent depths [37]. Note Ilnmrt tire nniethmmmne cusnmccntration iuas decreumsed from 72.30 nnole% to 54.92 mole% over a depth imrtervuml of only 81 mnmetcrs. Sucin ur nsajor clsange of cotmipositioun cannot be ignored as it strongly affects tine estimation of neserve, ansd production planning. In gencrurl, tIne mniixturc is expected to get niclmer in ineavier compoumnds, containing less of light components, sumch as methane, with depth . The variations of composition and temperature of fltuid wihh deplim resumlt in clmmmngcs of smutinrmnliomm presstnrc with depth. The oil saturation pressure generally decreurses witlu dccrcumsimrg mnretluumne coumcentration, whereas the gas condensate dew p(nimnt imucreumses wihin inrcrcasing Iremlvy fractionms. Tmuhlc 5.1. Vmrriations of fluid c~,U~nsmtLoms with deptis iii a reservoir.
Fluid t)epnlr (owner sutssea) Nilriigen (mnilniir t)ioxidc I), Welt I
1136

v~ =Z,RT,/P,
Example 5.7.

(5.48)

Esmumate the crntncal properties of a mixture of ethane-normal heptane (6))~4()nniole%), using the critical properties of pure constituenhs and reasonable unixing rules.
Solution:

The properties of pure conmiponenls are reund from Table A. I in Alslsemndix A, mumnil average values are calculated as follows: x

I) (iS 2 56

C. Welt 2 3 I 56 059 2.48

B, Welt 2

31)11
0.6(1 2.46

A, Well 2 3217 0.53 2.44

Mcntummnre
cornponerutT,,K

72.30
8.79 4.83 1)61 1 79 (1 75 086 113 1)92 054 3 28 5 49 33.1 260 0.8480

64.18
8.85 5.61) (168 207 (1.94 m 24 2.14 2.18 151 0.91 6 00 43.6 267 0.8625

59.12
8.18 5.50 0.66 2.09 11)9 t.49 318 2.75 .88 1.1)8 9.25 55.4 285 0.8722

etinane n-heptane

Pc,MPmnv ,nn/kgmnm~ 5 305.32 4.872 0.1455 0.2793

-.

540.2
~I

2.74

0.428 xT, 183.19 216.1)8

399.27

0,261 t

1)0995 (3.6000 0.3377 0.3495 0.4000 0.6622

l~utraiuc tbnspa ire n- It ulanc in-13 ninmmne

n-tcnnmrnc

54.92 9.02 6.04

0.74 2.47

1.33
1.71 3.tS

etim~ne
n.l~pi.ane

T, 103,11 357.76
460.87

2.923 t .096
4.019

x,w,x,L~ 0.0597 0.16758 0. t 398 0. t0444

0.1995 0.27202

0.0873 0. t 712 0.2585

Sum

I tc.sa,ics Itepumines (knauies N on mu lies )ecajnes I ,n,tccmines Plus

2.96
2.03 1.22

The calculated critical temperature

of 461 K agrees wilts the reported value of 460 K

Sicitectitmir Wenghin th~asiicsiaj.hamjsJn~osLucs

Motceutar Weigtrn Specific graviny

10.62 61.0

[36), whereas the pseudo value of 399 K does non.

290
0.8768

196

5. Ihase ite/narinnmr Cnlcuulation.o

.5.4. (onnpo,rilional (;i,mdung

197

The pressure in reservoirs increases with depth. If the saturation pressure remains below fine reservoir pressure at all depths, no gas-oil contact exists in the reservoir. The fluid can, however, behave oil-like and gas-like at the low and high sections of time reservoir, respectively, by going through a local critical point. If the saturation pressuure at any point becomes equal to the reservoir pressure, the gas oil contact is expected to appear at that point, with compositional grading in both phases, as slmown in Figure 5.10.

pressure change, due to the hydrostatic head, on the equality of chemical potential as the criterion of chemical equilibrium. A more rigorous approach, however, is to relax the assumsuption of equilibriuumrn and apply irreversible thermodynamic concepts by considering the coupling of heat and mass diffusion. Both approaches are addressed in this book. It shroimld be emphasised that tine ummetinods are applicable to static amid non-depleted reservoirs. In producing reservoirs, tine pressure and temperature gradients depend on flumid flow characteristics and cannot be expressed by static gradients. Furthermore dispersion emma stromigly affect the (ruinslen processes in these reservoirs. In reservoirs depleted below the saluration pressure, the forced ammd natumral convections, due to time developed density variationms within eacis fluid column, can becourme the dominant factors in controlling tIme conspositional variations in cotnparison witln tlnose itmmposing therniodynamic equnilibrium. I~quilehr.unmn Asstmnniphion Assignsiung null avcrumge lenupenuil Lime t(5 tIme Iluid coiuuuumnn murder comssidcmuut ion, the cmnergy cqtnat mum. Eq.(3. I), for a fluid nut dE= dU + d(mgh) umccclenuitksn.

~
~.

\
A

Gas

tmiuti.ici

~ (lit

Oil

nest reduces to,

(5.49)

where on, is tire nmsunss of tinc humid element nut tIre height h. and g, is the graviturtional
Pressure Pressurc

Figure 5.10. Phase variations in reservoirs with compositional graditmg.

The variations of properties of the reservoir fluid, described in Table 5.1, wills deptis are given Table 5.2. Note that the hydrocarbon mixture has a bubble point of 37.3 MPa at a deptin of 3181 meter subsea, whereas it shows a dew point of 37.8 MPa at 25 meter urhove that poiumt. The reservoir pressure was above the saturation pressure at all depths, and no gas-oil contact was observed in that reservoir. The estimation of compositional grading insight help evaluating the reliability of fluid samples taken at different depths. Furthermore, the fluid composition and properties at hifferent depths can be predicted when this inforniation is availmible only within mm limimited depth imrtcrvmml. line possibility of existence of an oil column under mm gas reservoir, or tIme prcscnscc rf mm mmummmbcr ol isolated reservoirs mistaken as a single reservoir can also be invcsligmmlcd.

Applying thernnodyumummrmic relations, similar to the mmpproach used us Section 3.1, will result in equiv.rlcnt equnrtiouis, hut witln mmmi additional energy lerns due to time gravity. The uniformity of cheusnicuml potential oleumcln conspoTment througlnout the system, Eq.(3.15), will he replaced by,
+

1t Mgh

~t~

M,gh

i.t~ + Mghi>

(5.50)

wlnere,

M is time nnohecular weigint, and the superscripts refer to various heighls with a total number of selected points designated by t. Ikncc for two points along the flunid colunsn, we oblmiin, 71 ~t~ = (Mgim Mgh )

(5.51)

Expressing tIne (lifferensce un chuenumical potential witlu tinat in hermsms of fsmgacity mis given by lAI.(3.22), we get. 2 RI Inn) f I f I or,
fiO = f2l~~p[~5~__
=

Table 5.2. Properties of fluids at different deptIns ins lIme Nortlr Sea reservoir.
Fluid Deptir(meuersubsea) Measured Reservoir Pressure, MPa Measured ReservoirTemperature, K Density at Res. Pressure. kg/rn Saturation Pressure. MPa
Saturation Point

t), Well I 3136 44.93 384.2

400.4

C, Well 2

39.0
Dew Point

Densimy at Sat. Pressure, kg/rn Separator Pressure, MPa Separauor Temperature, 5 C Separator GOR, m /m Tank Ott Specific Gravity

397.4 6.5

285.4
1005.0 0.7)177

3156 44.89 379.8 53t).8 37.8 Dew Pu,jnnt 503.0 1.6 308. I 611.0 0.8 170

0, WetI 2 3I8t 44.4 I 380.9 557.7 37.3 Dub. Point

540.0

A, ~VcII 2 217 45.35

382.)) 573.4 33.))

Mghln

In I

(5.52)

Im~1

iwl .2

(5.53)

Rub. Ioimrt 546.2

1.2

[,

RT

1.7 3m t).9 390.)) 0.8254

which reduces to Eq.(3.24), that is, equality of fugacity, when the two points are at the

same

290.9
304.0

0.8 185

The variation of composition with depth in a reservoir can he predicted by assuming

equilibrium within the reservoir. In this approach the coupling of thermal and compositionurl gradient is ignored. and reversible thermodynamic relations are mixed to nrccount for the effect of

level. W(nemn tIne composition of a fluid, hence, time fugacities of all its components are known at a point witinin the reservoir, the comnposition at any depth coumld be determined by using time above equation. The fumgacities can be calculated by Eq.(3.25) using EOS. When using a two-parameter EOS witin the volume shift, Section 4.2.1, time shift parameter should be

I 98

.1.

[hose Fle/,acioi~r(a/tu/atj,,i.s

.1.4. Conn ,asuio,ia! G,adi,ig

199
(5.55)

included in calculating the fugacities. This is required, as the fugacity coefficient nmiultiplier duue to the shift, Eq.(4.33), is different at the two levels. Monleh mind (noucl [38) used the Peng-Rohinson EOS to eslinsnuute lIme eonnmpositionnuih gruudinmg ut tine hiund column reported in Table 5. I. The estmmated clsanmge inn nnnchluuimne comncennnmmmtiomn (lute to the gravity over the ahove deptin was only 6 mole%. compared to tIne repontcd vumhmne of 17 mohe%. They attributed part of the deviuution to the deficiency of EOS to nsnohcl tIre plrumse hclnunviour. Eq.(5.53) suggests that compositional graditng hecousmes unore sigtnifieant wlnens tIne mnnixtinnc is cotsmposed of inolecuhes with widely different molecular weigluls. TIns is of purnticuluur significance in oil systems with large comncentrumtiouus of hemnvy comnipomremnts suncln us ursplnmmllennes. The trsain problem witin applying time above nmetlrod to tlncsc flui,~mis lIne Imuck of mudeqummule fluid charactenisation. The heavy ends umre genrenally reported mrs pscunuo-counpouncnnts on groups, curclr comnuposcd (If mummy connrpouimds, witln mon unvermnge nrrolccnulan wenghun unssignsu.h nut cmxli gniruill. When consnpositionurl grmuding occuurs, lIre conccmmlrmmtiomn of connupinunmds wilhninr eurchm pseudo cooupomrcnst chmunges. unnnd lire unse of a fixed value of nnrobeculumr wenglml reduces lIne reliurbil ily urt tine prehicncd results. In tall columns, wlmere assigning a mmnnfonm average temnlpcnmrturc lu tIne winole fluid colunrunr will not be justifmahle, lIne reservoir cmmmm he divided mnlo segnruents. cumcln mit uunn uuvcmunge tennmpcrumnnmre mocreursnng downward. Tlme connpositioumal gradttng cumin tiren he cvmilunmntcch for cuuclr secliomm wutlr calculaled composition an nhe bottom of emrch section used mis thud of lIne lop in tine uud~mrcemnn lower

l.,~= wInch is relerred to us thtc Omrsuugcr rccipnocuml rehumtiomn,s.

For mu nnnixhure wills N commnpotrenls, tIne hiux of courmponent i, amnd heat q, are given by the plrenonnnenological equuuutions as,
N

~
k-I

+L,X~

(5.56)

unmud
N

3,,

~ I

.,

X~-+ I

(5.57)

1 I )cinnnnnng,
=

(5.58)

we ohlaimn,
3,
=

sectioum. NonEquilibrium Fluids

lire equilibrium criteria, Section 3.1, prolnibit tIme existence of any gradient or flux wunlmnn a fluid system at equilibrium. As the reservoir temperature increases with depth. producing Inca) flux, the assumption of equihihrnmmm for a flmuih colunrnnr is not valid. I towever. if chcviatioums from equilibrium locally are not very large. wlmnch is the curse un pctrolemmirr reservoirs, irreversible tlnermodynamic concepts nmay be applied to analyse coniupositionunl variations. For such systems, it can be assumed that the equilihniummn exjsts locally, Ircnce, tIre saunme thermodynamic functions relating state properties of equmihihriunr systenus will he valid. Omssagcr lras presented a theory for time systemurhic innvestigation of irreversible processcsl39l. According to his theory, all shriving forces wilhin a syslemms, suclm as tennupcnmntune annd concentration gradients, can cause fluxes of different nnatures, stmcls as (Imose of lncmmt umtnsl nnuass. Any force can give rise to any flux. The above statement can he expressed by tIre gencruil phenomenohogicah rclmution as,

ki

~t~(X~

+Q~X,)

(559)

wiscre Q is cmmlled tIne /1(01 of trwm.s/roi-L

Conmsbitning the mmhovc equati(rmms for thne fluxes of meat and mass, it can be shown that at the isollrcmral condition,
N
=

~Q,J

(5.60)

TIsmul mx. Qi tine nmnummc

is tine

huemut lruutnsporncd by ouse umoit qunantity (ummass or tsiole) of component i, hence

meat of tia1n.sport.

As Ihe hluid cousnpositioui remniumimns mnnchmmnged with time, tire umet flux of each component, J~, mnunmsn Ire iero. I hcnmcc
N

J,

~LJ~X~ i,.l

j=t,2

E,

(5.54)

~L,k(Xk
k~i

+Q~X ) =1) 5

(5.61)

where i,j is the flux j, Xk is the driving force k, ~ ns the total number of driving forces, and L k 1 is called the phenomenological coefficient. For example, the coefficients l.~jjcan be the heat conductivity amid tIne diffumsions coefficient for the meat flux and tIme mass flux, respectively. TIne coefficients l.jk wills jk express the cross or interference phenomena, suds as that of mass flux due to thennrnuil gradienmt (Soret effect), and that of meat flux due to concentruition gradient (Dufour effect). Making proper choices of fluxes and forces, as suggested by Onsager. tIme matrix of phenomenological coefficients becomes synnmetnicmul.

Thre dnivnng forces for tnmass, Xk, and heat, Xq. can be determined by combining continuity equatnoims nsass, ennergy.produnction anmd momentum, and employing tinermodynamic relations to (lcteninnnne for the nale of emntropy describing irreversible processes [391, I., mm in d
--

Tgrad~m~ ITs

(5.62)

Xq = (gradT)/T

(5.63)

200

5. Phase Re!mau,onur (alciula:ion.r

5.4. Compositional Gradiing

201

where F is the body force acting on unit mass, and g.t is the chensical potential. The chemical potential is a function of fluid pressure, temperature, and composition. Hence, gradl.I~= (~1k
I aT)~,gradT +

where Ii,, and, ls~ are tine partial eusthalpies of the consrponent i, un the mnuixture amal in the ideal gums state, respectively. Oosl et al. 1411 applied the urbove expression, which was developed for liquids, to dense gases witlm mehiumblc resumhis. llmercforc. it lnmiy he applied to describe tlne heat of transport in oil aund
gums

(a1.~,I ~P)1,

gradP + ~

~)x

~p,r.u grurdx
,,

(5.64)

commdensuute sy.slemnns.

Applying thermodynamic relations to evahunate vutriuutions ofchenmmicuul potenmtiurl witlm teurmpcnmrtmure and pressure, and taking gravity as the only body force actimng on the fluid, tire following

There us very lilt he inn fonmmrmrtionm omn IIme Incat of tramnsporl in unulticoumiponent syslemnms. A simple
uupptounchm is to reduce nmnulticomnmpoumemmt umnixtunres to pseudobiumary systems, as

equation is derived by combining Eqs.(5.62-64),

~~(~I1k /aX )~ , gradx,

xi,, = I 1 1
=

(5.70) (5.71)

(M~ PVk )g

(gradT/T)

(5.65)

I.,, = ~
Q

where vk, is the partial molar volume of tIme conunponent k, p is tIne mnnixmure dcnmsity. mind

is

-i,k

defined as,
v,~= ~v~x~/x , Q;=Qkhk
t

(5.66)

n~k

(5.72)

where, hk, is the partial molar enthalpy. Q is called the pure meat of transport, as it does not include the energy transferred by the mass, expressed by the entinuilpy term,
Neglecting the thermal gradient, Eq.(5.65) reduces to time conventional equationn expressing the compositional gradient due to gravity effect,

wlsere tIne subscript mm, refers to tIme pseudo componemnt composed of all componeurts except k.

Hence, Eq.(5.68) will be as,

(5.73)

2v
~

vi,,

v~

(M~pv~)g

(5.67) line partial heat of vaponisation mrmnd minolar volumes can be cahcunlated by any relimible equation of state. Appendix C. Significance Although

which will reduce to Eq.(5.52) when applied to two points along the fluid cohn.uimmn.

Eq.(5.65) relates the compositional gradient to the temperature gradiemnt at any direction, where the vertical variations are only of signifmcance and interest, The density, partial mnmolar volumes, and the variation of chemical potential with composition, can all be determined using an equation of state, Appendix C. The information on the heat of transport, however, is sparse.
Heat of Transport Transfer parameters ~an generally be estimated by nmethods usiung concepts of statistical mechanics. Bearman et al. 140) proposed a method lo estirmsmmte the heat of Irmunsport for hiusuiry liquid systems. Broadly, the pure heat of transport was presented mis tine sum of an equnihibrium

some

degree of conmpositknnal grading dune to non~equilibriumis expected in all

1 lIne stability limit curve, Eq.(5.4I), the determinant of the above derivafives is equal to xero and profound commmpositionah grmndimrg shounld occur [42).

reservoirs, Eq.(5.65) shows that the effect of gravity and temperature gradient is nsore sigmmificant when (~ktk /ax )i.T.xjj is small. At the critical point, indeed at all the points on

term and a non-equilibrium contribution, where appropriate expressions were developed for each term. Neglecting the non-equilibrium term and assuming the liquid as a regular solution, the expression for Q* is as follows,
Q/x 2v~v
Vu

Fine mumea.snnred mmmd piedicled, using EOS, critical point of (he flumid reported in Table 5. I at time deptin of 3i56 umieter subsea, were 293 K & 40 MPa, and 363 K & 40 MPa, respectively 138). Figure 5.11 connpares the predicted variation of C of the above fluid with depth at various 11~ tennperammres when only the grading due to gravity has been considered. . Note that as the mixture temperature increases, approaching the critical value, the compositkmal grading with depth becomes more significant. . The compositional grading, particularly for nearcritical fluids, have been strAdicd by a nunmber of investigators(38,43,44]. Although the coupling of compositional and thermal gradients Imave been nrcknowledged by most of the authors, the lack of adequate information to evaluate time heat of transport has resulted ins ignoring the thermal effect in most studies. Clmmrhmmck [45) conmparcd the effect of thermal gradient to that of the gravity for an equi.nmolar mixture of nmmelhnuumne mind normal butane at a typicunl reservoir condition of 378 K and 10.34 MPa. Usimmg a value of 333 kJ/kg for the heat of transport of methane 146) and a thennal

(5.68)

where v~, and v2, are the partial molar volumes of connponents I mind 2, respectively, and v is the mixture molar volume. L is the heat of vaporisation, L, =In~ h~
(5.69)

202

5. F/iase Re/i,iiioii, CaIu-uulaiion.c

4 (ompo.mitio,mal (;radiiig

203 Tumble 5.3. Counposition of Inigli mmmd low aromatic oil sarmuphes. (ompiincnn Iliglu Low nni,lc% Arorn. Aronnn, N2 5 5 C02 5 5 Cl 52 52 C2 8 8 C3 5 5 iC4 It) no
nC4 4 4

gradient of 0.02 K/m, the thermal compositionaF~gradientwas estinmnmrted to he abotul Irutlf of thnat due to the gravity, and operating at the opposite direction.
Is
Teiuper.flii~ K

______________________

~0i 3798
388i ..~-

3731
-

5l 75 C

405!

a
13

(5 nCS
C) no

I I
I S

S
5150 ~ ~ithIhum,,i h,,,,i,i

nC6 hen,.cne
ni(7 iiituucune

cym cyc.trexmmne
n(.lO 6
5~70

5 5 5 7

2 5 2 3

h,~h .na,, .6!,

10..,,,,, -,

hi~h,n,,,,o~~iy p.OOy

nCt4
3200 3250
De mti Subsea. m 0

2 2

5 5

25

50

75

100

Oepih Reiow Refrmnsce Poinn. meici

Fngumrc 5. 12.
in

Vmu-immtions of nsictlrunne concentration with depth, due to combined gravity and

Figure 5.11. Variations of C 1 tenmperatures.

1+

of Fluid C, reported

Table 5.1, witln heptln at vmmnious

llnerniuuul effects, amnd onuly gravity effect for low aund high aromatic fluid samples.
As hIre thcnntmal amnd giunvnlationunl effects gemnerally oppose each other, it is conceivable that a fluid nununy nnnutnmntuniun Ilne suttmmc counsposition with deptis. Figumre 5.13 shows the predicted cotrcemrirmrtuon of tnnetlnanc wulln dcptls al various lemnperature gradients. Note timat at a timermal gradient (ml muromnund 0.025 K/urn, tIne domnnpositionmal grading in tIne above higlr aromatic oil is nmrsignificaumt. At Iniglmer thnernrmrl gradieusts, nnetliane concentration may even increase with depth.

Eq.(5.67) shows that the compositiomnal grading due to gravmty is nsore pronnounced for mixtures with components of vastly different sizes. Tine ligintest mind tine iremrviest counlpomments of a Inonrohogue group tend to grade more significammtly than intermmnediate ounes, withn tIne imeavies more concentrated with depth. Shulte [47) conducted a sensntivnty analysis of cotnipositional grading due to gravity, using equations of state, and conclunded that aromatics play a major role. Comparing two reservoir fluid mixtures which were only different in tIme ummount of aromatic contents, he showed the tendency of these connpounds to concenlrmntc witim depth, increasing the concentration of light components at the top. As aromatic compounds are relatively much deunser than paraffins and naphthenes with similar molecdmlar weights and volatility. such a conclusion should be expected. IioIt et al. [48] studied the compositional grading by considering both the gravitational and tisermal effects and reported similar behaviour for aromatic and paraffin oil systenns cotntrary to the SInulte~sconclusion. The reason appears to be due to neglecting the themnal effect by Shulle. The Ineat of transport, hence, the tinermal effect, depends on tine latent meat of vaponisation. Aromatics, due to strong bonding, have higher values of latent ireat tiran paraflins with similar molecular weights. Hence, mmeglecting thse tlscnnuil effect for arommiatics is a umrtncls more sever assumimption tlman that for paraffins. As tlmc thiermmnah urmsd gnavitatiousal effects gemnerunlly operate at opposite directions, such an assunmption will exaggerate tIne counposiliomnunl grading for aromatics. particularly dune to their highm density as discussed above. Table 5.3 shows the composition of two oil samples with (hifferent aromatic contents, very similar to those studied by Shulte[47]. The predicted compositional grading of methane, at 361 K and 34,5 MPa, with and without the thermal effect, at 0.02 K/m, for lIme two fluids is shrown in Figure 5.12 149]. The three parameter Peng-Robinson eqummtion of state, and the heat of branrsport given by Eq.(5.73) have been used to evaluate the conunpositional grurding. Note a significant increase of methane grading with depth by gravity due to the additional aromatic content, as observed by Shulte. Flowever, when the opposing th~ninal effect is included, the grading decreases and the difference between the two fluid becomes inmsigmmificamst.

0525

-~

0520

o s,s

:
0
2o iiopuli

RO,,i,o

Fngumc 5.13.

Vurrnuntmotrs of unretlrummne concentration witim depth due to combined gravity and

thrermnsal effects, at various tensperature grmmdients.

5.5 REFERENCES I. Ruicltford, hill. mrimh Rice, ID: Procedure for Use of Electronic Digital Computers in Calcunl.utimsg Flursh Vaporisatiots I tydrocarboun Equilibriunn, Petrol. Technol., Sec. I, 19, Sec. 2, 3 (Oct., 1952).

204 2.

5. Phase Rehuni-ionsr (.a!cudations

5.5. Reference.c

205

Sage, B.H., Lacey, W.N: Volumetric and Phase Behaviour of Hydrocarbomns, Gulf Publishing Co. (1949). 3. Fussell, D.D. and Yanosik, J.L: An Iterative Sequence for Phase Equilibria Calculations Incorporating the Redlich-Kwong Equation of State, SPE J., 173-182 (June, 1978).

19. Danesh, A., Xii, DII., Tehrani, D.H. and Todd, A.C: Improving Predictions of Equation of State by Modifying its Paranneters for Super Critical Components of Hydrocarbon

Reservoir Fluids, J. Fluid Plrase Equmilibria, 112,45-61(1995). 20. Gozalpoimr, F., Danesh, A., Todd, AC. and Tehrani, D,H: Integrated Phase Behaviour Modelling of Fluids in Reservoir-Surface Processes Using Equmation of State SPE 39630, Proc.of time SPEJDOE ltsrprovcd Oil Recovery Symnmposiumm. Tulsa, Oklahoma, 19-22 April
(1998). 21. Baker, L.E., Pierce, A.C. aund Luks, K.D: Gibbs Energy Analysis of Phase Equilibria, SPE J., 731-742 (1982). 22. Miclmelsen ML: Tine lsollnermal Plasm Problem, Part I, Stability, J. Fluid Phase Equilibria, 9, 1-19 (1982). -\ 23. Micinelsemm M.L: The lsotimermal Flasim Pmobleun, Part II, Phase Split Calculation. J. Flmnid Phase Equilibria. 9,21-40(1982). 24. Nghienn, LX., Li, Y.K. amsd Ileidennanum, R.A: Application of the Tangent Plane Criteriomm to Saturmrtiotr Pressure mind Temunperature Computatiomms. J. Fluid Phase Equilibria, 21, 39-60 (1985). 25. Miclncisens, MI,: linase Pqumilibriumri Calculatiomrs, What is Easy and Wisat is l)ifficuli?,

4.

Nghiem, LX., Aziz, K. and Li, Y.K: A Rohunst Iterative Metinod for Flurshm Calcmulmntions Using the Soave-Redhich-Kwong or the Peung-Robinson Equmurtiour of State, SlE J., 521-530 (June, 1983). 5. Mehra, R.K., Heidemann, R.A. and Aziz, K: An Accelerated Successive Substitution Algorithm, Canadian J. Chem. Eng., 16, 590-596 (Aug., 1983). 6. Pedersen. KS., Thomassen, P. and Fredenslund. A: Tiserniodynmrrtmics of Petrolesim Mixtures Containing Heavy Hydrocarbonns; 3 Efficient Flash Calcmulurtion Procedures Using time SRK-Equation of State, I & EC Proc. Des. Dev., 24, 948-954 (1985). 7. Gosset, R., Heyen, G. and Kahitventeff. B: An Efficient Ahgoritimusn to Solve Cubic EOS, J. Fluid Phase Equilibria, 25, 51-64 (1986). 8. Abhvani, A.S. and Beaumont, D.N: Development of aim Eflicient Aigoritlnurm for 2-Phase Flash Calculations, SPE RES.ENG., 695-702 (Nov., 1987). 9. Xu, D., Danesh, A. and Todd. A.C: An Accelerated Suiccessive Substitution Method for Calculation of Saturation Pressure of Multicomponent Fluids Fluid Phase Equilibria, 72, 1524. (1992).

Computers Chenss.Ensgng., 17, No. 5/6, 43 1-439 (1993).

26. Avlonitis, D: Thermodynamics of Hydrate Equilibria, PhD Thesis, l)epartnnent of Petroheumsi Engineerimig, tleriot-Watt University, Edimnburgh (1992), 27. Tejui, AS., Garg, K.B. urnul Smnith, R.L: A Method for time Calculation of Gas-Liqumid Critical Temperatures and Pressunres of Muiticomponent Mixtures, md. Eng. Chenn. Process Des. Dev., 22, 672-676 (1983).

10. Li, Y.K. and Nghiem, L.X: The Development of a General Plnuise Envelope Construction Algorithm for Reservoir Fluid Studies, SPE 11198, Proc. of 57th Ann. Conf.
(Sept., 1982). II. Whitson, C.H. and Michelsen, ML: The Negative Flasin, J. Fkmid Phrase Equnilibria, 53, 51-71 (1989).

28. Chueln, P.L. amid Prausnitz, J.M: Vapour-Liquid Equilibria at High Pressures: Calculation of Critical Tenniperatures, Volumes and Pressures of Nonpolar Mixtures, AIChE,
13(6), 1107-1113(1967). 29. Reid, R.C., Prumusnitz, J.M. and Poling, B.E: The properties of Gases and Liquids. 4th Ed., McGrmiw-Ilihl Book Conmnpany, 121-136 (1987). 30. Peung. D.Y. mmmd Rohimnsomm, D.B: A Rigorouus Metinod for Predicting the Critical Properties of Mutticouunpoumemmt Systenmms from an Equmation of State, AIChE, 23(2), 137-144 (1977). 31. lleidemanmn, R.A. and Khahil, A.M: The Calcumlation of Critical Points, AIC1mE, 26(5), 769-778 (1980).

12. Agrawal, R.K., Li, Y.K., Nghiem, LX. and Coomnbe, D.A: Mu.mlti-Plsase MultiComponent Isenthalpic Flash Calculations with a Cubic Equation of State, JCPT, 30-3. 69-75 (1991).
13. 14. Michelsen, ML: Private Communications. Edmister, W,C: Applied Hydrocarbon Tlmernnodynannics, Gulf Punh. (l9l~4).

15. Poling, B.E., Grens 11, E.A. and Prausnitx, J.M: Thermnodynmmnmic Properties from a Cubic Equation of State, Avoiding Trivial Roots and Spurious Derivatives, lund. Eng. Chem. Proc. Des. Dcv., 20, 127-130 (1981). 16. Asselineau, L., Bogdanic, 0. and Vidal, J: A Versatile Algorithm f(rr CalcuIating~ Vapour-Liquid Equilibria,). Fluid Phase Equilibria, 3,273- (1979). 17. Rijkers, M.P.W. and Heidemann, R.A: Convergence I3ehavioumr of Simmglc-Stage Plasm Calculations, Equations of State Tlseories and Applications, ACS Symposiums Sermes 300. 476-493 (1985). 18. Michelsen, M.L: Simplified Flash Calculations for Cubic Equatiomus of State, Imid. Eng. Chem. Process Des. Dcv., 25(1), 184-188 (1986).

32. 33.

Michelseun, ML. amnd lieideunann, R.A: Calculurtion of Critical Points fronru Cubic TwoLi, C.C: Critical leunperature Estimation for Siuniple Mixtumres, Can. .1. Clmem. Eng., 49

Constant Equation of State, A1ChE, 27(3), 521-523 (1981).

(5), 709-710, (1971).

34. Speurcer, CF., Dauhert, T.E. and Danner, PD: A Critical Review of (orrelatiomis for tIne Criticmrl Propenlies of Defined Mixtures, AIChE, 19, 522 (1973).

206

5. Ihasr !.khaiioiir (aleulations

56. Exereise,s

207

35. Kreglewski, A. and Kay, W.B: The Critical Constants of Conformuml Mixtures, J. Phys. Chem., 73, No. 10, 3359-3366 (Oct., 1969). 36. Katz, D., et al: handbook of Natural Gas Engineering, McGrurw-Ilihl Book Comiupatsy (1959). 37. Neveux, AR. and Saktlnikuniar, S: Dehimucation and Evaluations of a Nortln Scum Reservoir Conlumining Near Crilical Fluids, SPE 15856, Proc. of SPE Emropean Petroleunns Conference, London, (Oct., 1986). 38. Montcl, F and Couch, P.L: Prediction of Comnspositionnuml (unuudiurg inn mu Reservoir Fhtmid Cohninirun, SPE 14410, Proc. ol fiOtIn Aumur. (omit. (Sept., 985). 1 39. 1k ( root, SR hlnerumrodynummmuics (ml Irmeversible lrocessei., Nurutln --I un I:uurd Pimlrhishninig (a., Auunsncrdumnnm (1959) 40. Bearmumn Ri., Kirkwood, J.G. amid Fixuisan, M: Slurlistical-Meclrumnnicuul Theory of Transport Processes, ttne Heat of Irurmrsport in h3nmsany Liquid Solmmtiouns, Advuunuces inn Clncmmmicmml Physics, I, 1-13 (1958). 41. Oost, WA., Los, J., Cauwenbergh, H. and van Dale, W: Thcrmnmurl I)iffmusion in Moderately Dense Gas Mixtures and the Pair-Correluition Fumnctiomr, Physicun, 62, 409-426 (1972). 42. Lira-Galeana, C: Discussion of Treatment of Vmuiations of Connpositionn witlm Depth in Gas-Condensate Reservoirs, SPE RES.ENG., 158 (1992). 43. Whreaton, R.J: Treatment of Variations of Compositioun witln Depthr in Gums-Condensate Reservoirs, SPE RES.ENG., 239-244 (1991).
Ilmrscberg. A: Role of Asplsaltenes in Compositional (irmsdinrg of a Reservoirs Fluid Column, JPT, 8-94 (Jan., 1988). 44.

5.2. TIme coniposulnon of an oil sannmple is given in tine following table.

Cump~jnok%

Ct
C 2 C. iC 4 nC4
ir(6

27.36 10.93 856 .46 4.73 77

7 77 11)1

38.56 7u (tiuii;uincui,inc M - 21)).!)

N:))

Si))

The oul is flurslned mit 3 MPur aund 32S() K. Use the Stmnrmding nmetlnod to estimate the equilihriumn runtuos umnnd cuilcu mute tlsc conmmposu t noun umnid nurolc fruich four of equ ilibrunted gas and Ii qmnid plmascs. 5.3. Most reservoir, svellhore, mmd surface processes can he simulated by a series of cqimilihrmumin IhasIr cuulctmluutmons. Write ann algoritiunmm to simulate a constant volume depletion test tusing FOS. 5.4. Whnat us the. mnmmxmmrnmmns pressure at which tIre Peng-Robinson EOS provides three real roots for an equinniolar fluid comrnposcd of C -nC at 380 K. 5 4 5.5, Calculate tIre (hew point of a guus uunixhtnre connsposcd of 90 mol% methane and 10% normal decuumie urt 377.5 K, tmsinrg SRK (nircasmnred value P =33.72 MPa). 4 5.6. Derive an expuession to select tine proper root of generahised EOS for a fluid, based on tIne nsmimsimmrummn Gibbs ennergy. 5.7. TIne oml in Exercise 2 ms to be stahihised by one internrediate separator at 325 K. Find the ophinnnuurm separalor pressure 10 ohtmnun tIre mnnaxinnumn (nil volume at the stock lank conditions (0.1 MPuu mimic! 288 K).
t3stimmnuute tIme crinicuml pnopernucs of us mnnixtmure of methane ammd n-decane, 40-60 mole %, mmsmnrg tIne cruticuil properties of pmmrc consstituents and reunsonable nnixing rules. 5.8.

45. Chaback, J.J: Discussion of Treatmrsent of Variations of Cominposition willr Depth in GasCondensate Reservoirs, SPE RES.ENG., 157 (Feb., 1992). 46. Rutherford, W.M. and Roof, JO: Calcmmlahionn of Thcrnnruml 1)iffmnsiomr Factors for the Mctlmumne-n-Butamie systems in tire Critical and Liqmuid Regions, AICIiE, 841 (1963). 47. Slmulte, AM: Cormmpositiomnal Variations within a Hydrocarbons Colummn Dune SPE 9235, Proc. of 55th Ann. Conf. (Sept., 1980).
to

Gravity,

48. Holt, T., Lindeberg, E. and Ratkje, S.K: The Effect of Gravity aund Temperature Gradients on Methane Distribution on Oil Reservoirs, SPE umussohic. paper 11761, (March, 1983). 49. Danesh, A. and Movaghar Nezhad, K: An lnvestigumtionn on Compositional Gradient with Depth in Petroleum Reservoirs, Journal of Faculty of Engng., Tehran University. 30-I, 59-68 (1997).

is consnnmnommly plotted as a function of tIre reduced nemnmperurmtnre. T,, ummrd reduced pressunre, 1,, as stmowmi inn figure 2.22. Prove that the extension of amny tuongent line to conslmumnt redince terniperature curves sirould not cross the Z axis at a negative vurlue.
5.9. TIne comrspmcssihnhity fmmctor, Z, 5.1 0.Cunlcinlumte (Inc reservoir connrpositionr urt 500 nnmetcr below that given in Example 5.1 assmnmtninsg urns average lcnnrperurlure of 377.6 K for the fluid colunrmn.

5.6 EXERCISES 5.1. A fluid mixture consists of methane (30 % mol) and normal heptane (70 % mol) at 338 K and 6.5 MPa. What is the state of the flunid 7

209

6
FLUID CHARACTERISATION
Seunmi-empirical cumbic equations of state (EOS), though developed using experimental data of pumre compounds, are sttccessfully applied to predict phrase behaviour and volumetric properties of multiconmipomment systenss by emmmploying mixing rules, as described in Chapter 4. Real meservoir fluids, however, counld he composed of thousands of cotsnponents which pose two rmnajor restrietiotns: (I) A full descriptiorn of tIre fluid by idctifying mill its constituents nmiay not be possible. (2) Pimurse hchiaviouur curlcuulatmoums for systenus defummed by a large number of comnsponenmts are timmre connsumnming aumd particularly immnpractical in compositional reservoir simulation. A reservoir oil or cousdensate is commonly described by discrete hydrocarbon components up to C and the non hydrocarbon gases, such as N , C0 , 11 S and hydrocarbon groups for 6 2 2 2 Ineavier fractions. Tine concentration of certain major uron-paraffin constituents, within time C 6 to C groups, sucir mis, hcnzenc, toluene, cyclolmexamues and xylemse, nmay also be identified. The 9Imydrocarhon groups are gemmerally determined according to their boiling poimrts by (listihlation aumd, or, gas chirournatograplny. lisc distilled hydrocmrrhomm groumps are clmaracterised by measuring some of tlseir properties such as (lie average boiling poimmt tennmperatumre, molecular weight and density. These properties are used in generahised correlations to estimate other properties, such as the critical properties and the acentric factor, wliicin are required for EOS application. In some cases an extended distillusnion of heavy fractions may Irave not been conducted, hence, the required data need to be

estimnmated by othser methods.

210

6. Fluid C/,araeteri.(ation

6.1. Exper.nmental Met/nods

211
,, Ia

6.1 EXPERIMENTAL METHODS

Samples of reservoir hydrocarbon mixtimres, collected mit bottom lsole or sepmmrmrlor conditions, are generally flashed at laboratory conditions atsd the conmnpositionrs umn(l properties of tire separated gas and liqunid phases are mneasuured. TIre comnsposmtionral anmurlysis durhum of tIne separated phases are tinen nummerically recomrmbincd in (Inc surface proportiomus of gums uuurd Iiqmid to detennine the composition of the original reservoir fluid. The errors inn cousnpositioural anumlysis of high pressunrc reservoir fluids associated with the flumstnimrg hcclruniqcue, pumrticutlarly for gums condensate fluids, are discussed in Section 2.2.
The gas composition is determined by gas chromatograplry (GC) ]lj, in firrmmr of discrete comnponents. The nnixhunre average misohecular wciglnt, M. curn be cuulctnlumtcd try lIre nnuuihuur ummiximig rule, M=~y~M~ (6.1)
0

600
C 0~

0 .~

t , (~ 1 400

where Yn is the mole fraction of component i in the gas mixtuure. The gas density at laboratory c(nnditions cuin he nnieasunred, by ~veugInmnng a knownn guns volumnne, or calcuihated, approximately from the ideal gums law (by missumnnimsg Z= I),

20

4))

60

80

tOO

Cunurutuumive Dim/lIed Volume, %

Fmgumre (n.h. True bonhiung point (IBt) distillation curve of a North Sea condensate sample. ilse liquid phrase courhaimns mriununy comnuponnennts witlm properties varying in smmrll increments. Itence, time frmnctnommaliomr of hiqumid into pumre compounds is unfeasible. Each collected fraction comsspriscs mm large nuniber of comuiponents with close boiling points. Fractions are commonly collected witlrin tIre temperature range of two consecutive nominal alkanes, wlrcre each cut hegmnms uuumd ends mit lIme boiling point of mrormal C~. t and normal C~,respectively, and is mclemued to by line cuirhroun urunmurher mm. For cxmmumnple, tine reported C fraction, nonancs, 9 comnrprises all time counrl)ounds collected as distillate. withini the temperature range of normal octamne and normal nonnane boiling poinls, as shown in Figure 6. 1. The fractions are called, Incncc, siurgle cmmmhon number (SCN) groumps. In practice the boundaries are selected about 0.30.7 C. depending our time distillation unit and the fraction, above the normal alkane boiling poitnts, nnaunrly to c(,unlermrct time hiqinid Isold-up in tire apparatus for imnproved purity. Tire puruty of SCN grounps camu further be improved by using a more efficient distillation apparumtuns. TIre tnse of a 90 theoretical equmihibniuni plate unit, instead of the standard ASTM 15 plale unnut, is gainiung popimlarity. Even then, 5-30% of the C~fraction could be lighter than the unormsrurh (in- I [5]. Each cml is elnaracterised by its uiroheculmrr weight, density and normal boiling point. The l)oihing pounrt is taken mit its mid-vurlumune, Figuure 6.1. As tine boiling points of the neighbouring normal alkanes are chose and tlne distillate recovery is almost linear over a SCN range, the midvohuuunre temperature is about the satire as the arithnnetic average of the boiling poiurts of the two nonmnmrl paraffuns at the boundaries. The measured properties are used in some generalised eorrclumtuonns, Sectuomr 6.2, to detennmine the critical properties and tIme acentric factors. The nesudunc is reported mrs Cnm+ e.g., C3))~ wiren tIme hasn drop of distillate is collected at the boiling poinrl of nnC . TIne urvemage hoilnurg poimrt of the residue, if reqn.nired, may he estimumted from the 29 correlations given in Sections 6.2. TIre density of each cunt is mrseasumred by either weighing a known volume of the liquid, pycnonrietery, or by tire more rapi(I, yet rehiumble, oscillating tube densitometer. TIre latter mrseasuures hIre lneriixl of oscillation of mm tumbe tilled with tIne fluid, which depends on its mass, lsennce. mIs ulenrsity. A cuuhihrr,med nunit slsould provide density data with an accuracy of better than 3 0.001 g/clms .

Pg =(PaMg)/(RT ) 0

(1.3)

where P is the gas density at the atmospheric pressure P~and tIre standard tenmiperalure T . 5 of universal gas constant, R, for different sets of uunits is given ins lable A.3 0 The value in
Appendix A.

The oil composition can be determined by gas chromatogrunplmy, or nnnorc comnnonly by distillation [2] and reported as liquid fractiomss. The heaviest frmrclion, wisich fornmrs tIne residue in distillation, can he analysed by liquid chronrratogrmrpliy techniques 131.

Distillation
The liquid phase is generally charactenised by fractional distillation and nrieasturing the properties of the collected fractions. Tine distillation is commonly conducted using a column with IS theoretical equihibriunm stages at a reflimx ratio of 5 and is known as the true boiling point (TBP) distillation. The standard niethod is fully described inn ASTM 2892-84 ]2]. A pot is loaded with the liquid and heated up. vuiponising its cotnponents accordiung to tlmeir boiling points. As light compounnds vaporise, increasing the concentrurlion of heavier fractions in the liquid, the temperature is gradually increased. The boiled-off fractions mire collected as distillates, each within a temperature band at the column top. The distillation begins at the atmospheric pressuire, but tine cohunnn pressure is lowered stage-wise to vaporise heavier compounds to avoid higls temnrpcrmutures which can cause hydrocarbon cracking, hence, compositional changes. The distillate lenmperuuluure is converted to the normal boiling point, that is, at the atmospheric pressure, and thmc resunlts mire given as tIme percentage volume distilled versus the normal boiling point, as shown in Figunre 6.1. As reservoir hydrocarbon liquids generally contain very heavy comsipounds, such mis t aspinaltenes, a certain amount of the loaded saniple will not boil-off, mmd will he left in the po mis the residue. Methods, such as using empirical correlations 1 4], have been stnggested to extrapolate the TBP curve to 100% distillate. There are, however, more reliable unrethods In describe tIne residue by a number of fractions for phase behaviour modelling, as will be described in Section 6.3.

212

6. Iin,nd Cluarauleri.ruutio,n
d)f

6. I. Experimental Met/nods 11.0< K~ 12.5 8.5< K~ 11.0

213

The average molecular weight of each cut is often deermmiined by trremistmring tire depression

freezing point of a solvent, e.g., benzene, by dissolving oil at a concemntrmmtion (ml about 0. IS

mole per kg of solvent. A deviation of about 2 units of molecumlar weiglnt cumn typically be expected in a carefully conducted test. If the distillate is accumulated in a receiver, instead of collected as isolated fractions, the properties of each SCN group cannot directly be determined. In such cases, unaterial balance methods, using the density and molecular weight of the whole distillate and the TBP distillation curve, may be used to estimate the concentration and properties of SCN groups 161.
Katz and Firoozabadi [7] extended the data of Bergman et al. 18] our the average boiling point,

Naphtimenes Aronnatics

Tire cisaracterisation factors of generahised SCN groups are given in Table 6.1. Table 6.1. Avermmge normal boiling point, specific gravity, uniohecular weight and Watson characterisation factor of single carbon nunmber groups [9].
SCN
-

Boiling Pin/nt

Spccifmc Gravity

molecular weight and density of SCN groups of a large nunsber of reservoir flumids. Their data, revised by Whitson [9], to improve consistency in the reported unrrlecular weight, are given ins Table 6.1. The properties, known as tire generalisecl siurgle emtrhon numnmsber ulatmu, mire umsed when the measured data on a specific fluid is not avaihmubhe. lime cuulcmmlumted critical properties of generahised SCN groups, using correlations described un Scctiomn 6.2. mmmc giveum inn lumble A.2 in Appendix A. Haaland [10], Osjord et al. [III and Ronningsen ct mrl. (121 huive mmlscs reported average SCN group properties for North Sea oil and condensmile sanriples, winds differ somewhat from those given in Table 6.1. The properties of paraffins, naphthenes and aromatics, present in each SCN group are different. Hence, the properties of each SCN varies according to the relative eonmcentration of the comprising homologues. Table 6.2 shows the paraffins. napinthenes and aronsatics (PNA) content of a North Sea stabilised crude oil and their properties over the C -C range. Note 6time9 putrmrffinr group that, for example, the density of naphthene group in C is higher thumn tliurt of 6 in C . All the hsydrocmmihon commrpoumsds witinin eumchn SCN gromip do nnot lmuuvc tire sumunue mrtuummlwr 9 ofcarbons. Indeed aromatics with the samnne caubon ntmmsnber as puiruiflins will uulmpeumr inn tIme next higher SCN group due to their lower boiling points. For exanisple bemrzemrc. toluenc and xylenes are counted as C , C and C groups, respectively. 7 8 9 Due to an uneven distribution of hydrocarbon Imomnologumes in SCN groumps, all time properties of SCN groups should not necessarily follow the same trend. Figure 6.2 shows the variation of SCN group density in different samples. The plot clearly demonstrates that the density of a SCN group can be lower than its preceding neighbour. The mmnolecuhar weiglrt, imowever, is expected to increase with increasing carbon number. The PNA analysis of SCN groups is not generally required for modelling of vapour-liquid equilibria using equations of state. However, detailed infornmatinn ors the content of each SCN group may be required in special cases, e.g., when two hydrocarbon liquid phases or liquidsolid hydrocarbons are formed. Methods, relying on material balance and empirical correlations, have been proposed [4, 13] to estiniate the PNA content, using the specific gravity and molecular weight of each fraction. A measured detailed mrnalysis would be more appropriatein such cases instead of estimating them from the correlations. The overall characteristic of hydrocarbon fractions, is commnronly described by tire Watson or UOP (Universal Oil Products) characterisation factor, K~, as follows: K,, ~(l.8Tb)~/S

K
(6 (7 (8 C9 C II) 337 366 391)

rel. dens. am
288K
0690

Molecular We~Iu kg/kgmol

84 96 1117 121

Watsinn Char. Fact.

____

0727
0749

4ft
439

1)768 0.782

CII
C12 Cl3

461
482 51)1

0793
t).S))4 1)815 1)826

I 34 147
161 175 190

1227 It 97 1187 1182 II .82

1185
1186 11.85 11.84

Ct4 Ct5
C16 (17 (18 (19
C20

520 539
557 573 586 598 612 624 637 648

0.836
0.843 0851 1)856 0.861 0.866 1)871 0876 t).88I

206
222 237 251

1184
11.87 11.87 11.89 11.91)

263
275 291 300 312

1193
1193 It .95 11.95 11.96 11.99 12.01) 1201) 12.02

(21 (22 C23

C24 C25 C26 C27

C28 (29

659
671
681

0.81(5
(1.888 0.892

324
337 349

691
7(11

0.896
0899

360
372

709
719

0.9(12
0.91)5 0.909 0.912 (i.9)5 0.917 0.920 (1.922 t).925 0.927

382
394

1203
12.04

C30 C3t C32

C33

728
737 745 753 760 768 774 782

404
415 426 437 445 456 464 475

(2.04
1205 1205 1207 12.07 1208 12.07 12.09 12.09

C34
C35 C36 (37 C38

(6.2)~

(39
C40

788
796 81)1 807 813 821

0.929
0.931 0.933 1)934 0936 0.938

484 495
502 512 521 531

12.11
12.11 1213 12.13 12.14 12.14

where Tb is the boiling point in K and S is the specific gravity. For pure hydrocarbons the above definition of chiaracterisation factor results inn: 12.5
<

C4m
C42 (43 C44
-

K~ 13.5

Paraffins

C45

826

0.940

539

214

6 1/mid

( jlariiite,j.cUtjo,s

6. , f~efmr,?nrei:fa/Ales/rods

215

Table 6.2, Paraffins, naphihenes and aromatics content of single carbon nutriber grommps of C6 to C9 of a typical North Sea oil [III. -. Component Wcngtn(% - Motc% Votumric% MoIW liens. gui Hexane Group Paraffmns 0647 1.886 0.836 862 t)663 HexaneGroupNaplrthenes 0.052 0,185 0.059 70.1 0750

___________

not, however, comiunmionly available for petroleum fractions. Other characterisation factors, not widely used, have also been proposed 1141. Ilnc Wumtsorn cismrrumcterismnlion factor cuimn he related to properties otlser than the boiling point and specific gravity, using correlmmtiomms given in Section 6.2. For example, it can he related to the mnolecuhar weight (M) and specific gravity (S) [9], using the Riazi-Daubert correlation, Eq (6.4), K
=

Heptane Group Paraftins Ilepiane Group Naphihcnes Hepuane Group Aromanics

Ociane Group Paraffmns Ocnane Group Naphthenes Ocuane Group Aromatics Nonane Group Paraflins Nonanc Group Naphitncncs Nonane Group Arornatics

0.713 0930 0355

0870 1.404 0.958 0739 0646

1.787 2.682
1.140 1.912 3.435 261(1 1416 13i1

1)889
1)34 0343 11)54 I 556 0941 1)877 1)699 (122

11)0.2
87.1 781 1142 11)26 92.1 128 5 22)) 11)6.2

1)686
1)769 0884 1)7(17 1)772 1)871 1)721
))

4. 5579M ~s

84571

(6.4)

TIne above relatiomn is particularly useful for the last fraction, referred to as the plus fraction, where its boiling poinrt is trot known. TIre above equation becomes less reliable at M>300.
nuuisst cases.

FIne variations nih K~ ns ueIuutivchy smnmmmll lou dnfferenrt fractions, particularly for Incmmvy fractions in 5 It cann Ire nssnnnnnd I. lhucrcIusre. mm conmstuntrl for lreumvy frmuclions to evurhmmmute tIne

1042

2464

1)872

inuterunuil coursislenscy ut uuiemusuircil (lanun, or no estinmruute nmnissing iunfnrunatinn, as will be described nun Seclinsin (r3. (;as Chromatography Tine guns comnrposimionr is determmnimned, innvuirimuhly, by gas clmromnmatography (GC). Recent mmdvatmces ins gas chromnialograplny have enabled laboratories to extend the nnethod to oil analysis with a conrpuirahle mrccuuracy. Wtnilst urn extended oil analysis by distillation takes many days miund requires relmslivchy mm lunrgc vohiuumse of sample, gas chromatography cams identify components u.s ineunvy mis (5)) 115! inn ui urumnller of Isomurs using only a smunll fluid sample.
Tine smiusiplc is injected iuilo mu Ircumted inure, vaporised, and transported by a carrier gas, usually

0.90 0.88

0.86
en

084 082 0.80 078 076

0 74

beliumxn, into a columnun packed or iurtenmally coated with a stationary liquid or solid phase, rcsuluinrg ins partitioning of tIne inn jected sample eonstituuents. General purpose columns partition conrponenhs mostly umccording to thucir boiling pOints, Imence comsipounds are eluted in a similar order mis in distillations. TIne elunted conrpoumnds are carried, by the carrier gas, into a detector wlscre tIne conmnpomsent coniceumtratiour is relurted to tire area under tire detector response-time curve as shownr inn Figtnre 6.3. htrdivtchunmnl peaks msnay be idenstified by comparing their retention timires iunsi(hc tine coltinnnmn wills those of kunowus compounds previously analysed at the satire GC
comnditiomss.

Figumre 6.2.

Densities of SCN groups of Carbon fluid samniples Number fronn vuurious Northu Sea reservoirs.

Huthmig-Fachnveriage Copyright. Reproduced fronir [Ill wills pernsnission. The characterisahion factor for a mixture can be estimated by time weighted average mnnixing rule, K,,, ~ where w~ is the weiglnt fraction. A mixture of arornmntics and paraffins, therefore, may appear as naphtinemse evalunated by its K~. ilowever, it is a useful single factor describing the characteristics of pelrolcuurs factions. A more reliable characterisation factor, especially for complex fluids, may he obtained by including a third physical property, such as the viscosity or the refractive index. Tirese data are (6.3)

lIre Iwo most conrinsnounly umsed detectoms are lhc flunnure ionisation detector (FID) and the thermal connduichivihy clclector (TCD). TIre FID respourse is almost proportional to tIme mass comscentrmrtioum of tIne ionnisech connupounnnd. In, however, cannot detect nonhydrocarbons such as N uuusd CO . llennce. 1(1) is ofnems umsed for urnmnlysis of gaseous mixtures tlrat contairs non2 2 Iiy(lr(rcarh(nmr connupouncurts. tuuckech colunnimurs, witlu ann efficiency ruungimrg fronnn lemns to hundreds of eqmmilihrium stages, are nIne unuost versmrtile unund frequenrtly imsed devices. Tinese columns are capable of base line sepmmrumtiomr of gaseirtus comsnpirnumnds, hensce, dchcrnnining their concentrations as discrete courspolnn(ls. TIne iuntcumxrediurlc ununh lrcuuvy lrurctionns arc ehrmtcd, however, as a continuous stream of oven Imnppimrg comnrpouuirds. Fugnunc 6.4. ilmis is very simiunlar to the fractionation behaviour in a distillation mnnit and lreated simmuilumrly. lhmmt is, all tIre counponents detected by GC between the two ureighrhouring trormrral paraffimrs are comrimonhy grouped together, measured and reported as mm SCN eqnnunl to llral of the IrigIser noniral paraffin. The GC operating conditions may be adjusted ]h6]. Ihummt is. its efficicuicy lowered, ho siustunlale the 15 tray TBP distillation. Tire resulls, known as lire simmiuluiled distillation, are quite comparable to tlmose generated by tine TI3P nneihnod [2], as slmown in Figure 6,5. The percemntage area under the FID response curve luas been tmikeun to he eqimivaleurt to time percentage volumne distilled [17].

216 cm

6 Flair! (/n(una(gerjc,:fjo,u

6. 1. Erpe,iinerusal Metluod.c

217

800 Ce

ci
A
C 0

701)
. ci,, so
C 01 0

600
0 OS 0 0 0

0 I

8 C
0

a
0

nC4

a nCS

501)

z
404)
0S

C,C Sinrniaied

m)isn,nn4ni,n

nd?
uu~ mine
Renenuson Time
> 301)

20

Figure 6.3. Gas chromatogram of a gas sample, using a packed coluumsnn. The accuracy in compositional analysis can he improved by cahibruiting GC, that is, determining the detector response relative to the concentrmrtionm of emmch counpoument mit tire operating conditions, known as (he response factor. The courmimmon mnetirod is to ammurlyse a gravimetricaily prepared mixture of componeists witis known concentrations, as time staindard.
Normal alkanes are often used to represent SCN groups. It is known that the response of detectors to paraffins and aromatics are different. Hensce, the unse of typical SCN groups, instead of normal alkanes, in preparing standards appears to be more mupproprimmie. Time effect, however, is minimal [5] in most cases.

40 60 Cuunnulauive Volume, %

80

Figuure 6.5. Comparison of TBI distillation curve amnd simulated distillation restult, using gas chrounatography, of aim oil sarirple. A nrajor drawback of GC analysis is time lack of information, such &s the molecular weigist and density, on the identified SCN groups. Tine lack of molecular weight data is quite limitinmg as the response of FID, used for oil analysis, is proportional to the mass concentration. Molecular weight data are needed, therefore, to convert the mass fraction to molar basis required for conrpositionai studies.
Time very high hoilimmg constittmemrts of a reservoir oil sample cannot be eluted, inence, tisey canmmot be detected by GC, Time comnson umietlsod of estimating the non-eluted fraction is to unse an internal standard, winere one or a few fully detectable compounds, preferably not present in tIre oil, are added to the oil at a knsown mass ratio [161. The comparison of nnass ratio as detected by GC wittm tinat of graviusretricahly prepared mixture, gives an indication of tine ansmount of noms-eluted fractions. TIne nirethod, known a.s spiking, relies on certain hiunitimig asstumrmptions

nC

9 nCuo

winds nmnuy lead to large deviatiomrs in measured concentration of non-eluted fractions [5(.
4)

C 0 0~
4,

0
(I

A rplicustiomm of mm eomrtimmtmotms fumnction to describe tIme componeunt distribution (see Section 6.3) 1 extemnding tine mrrcursumred comrcenntrationn of eluted fractiomss to uleterumsine time non-eluited puurt is mimrd mrlso an option [5]. highs tensmperature columns [IS] are capable of almost connnplete elumtion of light conndensate fluids. However, the concentration of tine plus fraction (last reported group)

determined by GC, sinould always be treated with caution.

Capillmrry columns, equmivalent to many thousands of theoretical equuilibriunm stages, can be used in preference to packed columns, to improve separation and peak recognition, as showmn in Figure 6.6 for a North Sea condensate. Table 6.3 presents the components as identified by the peak nunsbers in Figure 6,6, aird tineir concentration, molecular weight and density.

lire molecular weight and density of components ideustifued in each single carbon group can be
used to estinrate the properties of that group, by the following material balance equuations: Time
>

Figure 6.4. Gas chromatogram of a North Sea oil sample, umsing a pumcked colunnum.

(~w~))/~(w fM) 1

(6.5)

a a a a t.. ~ NJ NJ NJ 9 .0 so -4 0
4, NJ

a 4.
so
-J 0 5~

9
.4,
.9

N) .0 NJ NJ NJ NJ NJ 9 .0 0% 0

NJ NJ NJ so 4 0 La .9 ~ NJ 9

so -.4

0
NJ

a
T3 0 (5

;~;:~ ~, 09 ,,.~.2: ~
~_~JJ .

It

NJ ~

~:;~
~

2,
~4.

4/~NJn1~ ,r NJ

r~
0
,~

Za~ a 7) 1%
-~~u9~~ ~< ~ a
n
%j~ 4. I~i .~ ~-Jnrir~ aa~L, ,,.a...J~a ~ 9

~ ~ 0

,,,
C
.4

a) 0 ~ L,.4< N,)

~.
~

a~-~r~ NJ ~) .aa

~
~
0 0 0 ()

~
0 , ~ 0.

cm.
~4, La ~
4. ~ 0

~
o

~) 7.) 7.)
r~
La

~ QQ~ ~ ~ a ~) .9
~:

~.

~ ~
1.
H ~
3 0
-

~
.~

-I La

La

-~
()u~- co~

~
..<

em
3
Co
02 1. 9333999T0 NJ.395~ 00 so so La 033999399990099

B
La003CNJ-J

.0 NJ *

3
~4
0 NJ

0999
NJ~~J__
.9 ~ .0

NJ -JC

0 c~ La NJ NJ .9 flNJ3.9JLJ~soLaC,...soNJC-400

3-3O39~39~3.0 0
..4

.0 C
.9 90 NJ La ...NJt4,NJNJ

999999993393 a099---a.,~.. -.~9 4. La 9 NJ S so ~LflsoNJONJ0~3~4.LflJ,*

939939 999

-4

5 a 0 - C
~

.9 9 0 (5 0

~0 (5
M ~0~NJ~~

00 C

em
~<
99~~333393*39~9_ ~.J 993993393

~4.NJNJ39

3333399!2, 4,
aNJ33 0,.9.,J,a at..tLa so

9993NJ33C3033333 00 0 0. = 0, NJ
NJ 4, NJ4~~,J499sor.4990.

~
07.9
-.4
NJ C NJ.9939~.9,,J 3 tO .9i = 4, 4. ~,aJ~,4,.03N.3soL..,CN,

,33333333 3e~ so NJ 0, 0. .0

3 C
4 .4 4 NJ 040soNJ..Jl()

,
soOc-soso--JOC
C-C0030 0
..., 4-J..44,4,4.L,J

a 0

0.

0 NJNJNJ,.4,,).9.9,0.,4NJNJaaaJosoNJso
NJCJNJDJN,,N4NJNJI.4NJ,JNJNJNJN.,

so
RC0C90_4so

.4-.4.4.
C0C0La.9

so

000%0ONJtJNJ0%.9ON4NNJNJ

so
NJ

C a

.,..Ssososotso50,soSJ,, NJ?4r.)t.4NJ 9 = so so 00 = so 3 0 Ct 0
CCCJ,GJ,NJ.9LaM..,

00 .0. 4 0. .0

-4 -.4 -4 -, .~ -.4 sosososo.0so

3 .0 C,)

NJ4NJsoso0..~r 3=
.a La a,.j C .9.9-43

93099

33~39933
.,4 La

3
-J ... 00

.4 .1-40% 0
0-)

=
a -4 -4

.4004,.4..J0-4.4-4.J .4 Laa .0 NJ-)

0 05 -.j a S S S so 3.0 0 04. 0 a ~o

.d~-4 NJ.9 0 .4 to a NJ

99993~ 0%0 0J0--4so-0-J so 00 La La a .c ~ so Cc so-~ ~ 00 LaL4.C .0 NJ..a -4

9 .0 NJ.,) La 9.0

9 0 0
C,) 0.
so -4

0000-40

La i, La NJ0

0 0 La 0 4. so .J0 so so -J NJ

4.

00,a4,LaLaa.9

.i .9 9.0000
0 4.

.9 9 .9 0 0. .9 -4 0 0 -4 4.90

so

9,

0%

0) C 3.
C)

.4

C
C) C) 0)

20

30

40

50

04

Time, Minutes

Figure 6.6. Gas chromatogram of a condensate sample using a capillary column. Peak numbers correspond Huthig-Fachveriage Copyright. Reproduced from [11] with permission.

to components in Table 6.3.

NJ .0

220 Table 6.3 (Cont.).

Individual components identified by peak numbers in Figure 6.6. Peak No. Component Weight % Mole % Votumnne% 43 DM.CyC6 0.031 0.069 0.034 44 1.trans-2-DM.Cy-C6 0.089 0.199 0.098 45 nC8 0.434 0.954 0.526 UnspecifuedC8 0.086 0.190 0.105 Octanes total 46 47 48
49 50
.

6. F! mild Cluaracierisatio,m

6. 1. Experionenla! Methods

221

Mol. W. 112,216 112,216 114.232 114.232 101.978 126.243 28.259 128 259
112.216 I t8.uuoO 126 243 128.259

Dens., g/cnn
0.7700

0.7799 0.7065 0.7000 0.7791

0.79(8)

3.231 0.047 0.009 0.017 0.024

0.281 0.047

7.957 0.094 0.018 0.033

0.t)54 0.599 0.1)93 0.034 0.tX)6

3.551 0.1)5 I 0.011 0.020

0.1)26 t).105 1)1)51 0.020 t).004

Tumble 6.4. Conrmparison of single carhomr nuumrher group properties measured by distillation and capillary (ic analysis. Disnillation GC Weigtnu % Mol. W. ~ens~ -- ~ Wdgta% ~ Den /mn~ 5 0.886 65 621 0.792 63 597 6 0.737 82 695 0.699 85 669 7 2.371 91 751 2.000 89 754 11 2.825 103 778 3.237 1(12 779 9 2.539 116 793 2.429 116 799 #0+
10 90.642 306

Unspecified naphtlrene 2,2.DM-C7 2.4.DM-C7 I,cis-2.DM-Cy-~6

E-Cy-C6#t.l,3.TMCy-C6 Unspecified naphthenc

0.7144 0.7192
0.801)3 0.79(8) ().790() 0.7262

869
798 803 107

90.846 2.437
2.191

31)0 134
148

868

2.479
1.916 2.352 2.091 3.667 3.722 2.(u34 4.135 3.772 3.407

112
t41 163

It
12 13 14 IS 16 17 18 19

81)1
803

51 52 53 54 68 55 56 57 58 58 58 59
60 61 62 63

2.523
3.106

162
175

3,5-DM-C7 2.5.DM-C7
Ethylbenzene

0.017
0.003

28.259
106.168 126.243 106.168

0.7208
0.8714 (1.7918) 1)8683

Unspecified naphthcne m-+p.sylenc 4-M-C8 2-M-C8 Unspecified naphthene Unspecified naphthene Unspecifuednaphthene Ortho-xylene
3-M-C8 I.M.3-E.Cy.C6 I-M.4.E-Cy-C6 Unspeciluednaphthenc

0.114 0.027
0.697

0.270
0.t)54 1.649

0.112
0.029 0.687

0.020 0.054 0.009 0.082

0.007

0.039
0.106

0.230
0.023 0.078 0.034 0.006 0.004 0.471 0.124

0.018 0.163 0.014 0.545

0.045 0.155 0.068 0.013 0(107 0.923 0.243

0.024 0.t)64 0.010 0.089

1)008

0.223
0.027 0.1)83 0.037 0.01)7 0(8)4 0.559 0.148

128.259 128.259 126.243 126.243 126243

11)6.168 128.259 126.243 126.243 126.243 126.243 28.259 128.259

0.7242 0.7173 0.7900 0.7900

0.7900 0.8844 0.7242 0.8000 0.7900 0.7900 0.7900 0.7214 0.7200

175 190 21)5 215 237 251 262 426

836 843 849 863

844

846 857 885 -

3.124 3.984 3.383 4.244 3.201 3.523

190 205 218 235 250 261

812 827 840 845 851 842 845 854 888

20+

61.057

59.130 422 ________________________

6.2 CRITICAL PROPER1IES Time critical temperature, Tc, pressuire, P~,volume,

Ve, compressibility factor, Z~,and tIre acenrtuic fuuctor, w, of siungle carbon number groups and th last (plus) fraction of reservoir fluids, siusnilar to tirose of discrete components, are required for phase behaviour modelling using EOS. These properties are determined from generirlised correlations in terms of the specific gravity, S. boiling point, Tb, or the molecular weight, M. of single carbon groups.

63
64

Unspecified naphuheune
nC9

UnspecifiedC9 Nonanes total

2.427

5.241

2.598

116.277

0.7995 Several nnet.hsods to calcunlmtte tire critical properties of petroleum fractions are available. Time metinods have mostly used measured critical properties of pure compoumsds to develop correlations, either in graphical forms or as equationns. Tire majority of these correlations are reported in L h9~.Tine most widely used, or pronmising methods, are reviewed in this section. lime correlations in threir original forms use Field Umnits, and are given as such in Appendix B. hun ilsis section, tIre correlations urre presented wills SI units. The units of P. T and v are MPa, K amsd nm/kgnnol, respectively. Lee-Kesler Correlations
so

A comparison ofthe calculated properties of single carbon groups using GC data and the above method, with those determined by 90 equilibrium tray distillation on a condensate is shown in Table 6.4. The differences are of the same magnitude as typical deviations in measuring the properties. A highly useful column for hydrocarbon reservoir fluid analysis is the wide bore capillary column, also known as the megabore column. The column which provides the versatility of packed columns, whilst maintaining a high resolution capability, can be used successfully in compositional analysis of live fluids, without any need for flashing the fluid into gas and liquid samples for GC application [18]. The preference of direct sampling and analysis of high pressure live samples relative to the conventional blow down method is discussed in Section 2.2. The aromatic content ofan oil may also be determined by gas chromatography using a column containing a strongly polar stationary phase column which elutes the components according to their boiling points. The combination of results from polar and conventional columns are then used to determine PNA [8].

[20,21] 74 S)Th +(o.l44l

189.R+450.fiS-1-(O.4244-4-0.l l
=

5 l.0069S)x l0 /T 5

(6.7)

humP.

7 3 3.3864 0.0566/S_(0.436394-4.l2h6/S+ 0.2l343/S )x lO T

+(o.47579 + 1.182 / S +t). 15302

is1) x I (Y6T~(2.4505 + 9.9099/Sn) x I 0T~

(6.8)

222 o =(ln ~hr

6. Fluid (ha,mlerj.caiiom

6.2. (, ,i,cal Propet lus

223

5.92714 + 6.09648/T,+ 1.28862 lflTbr 0. l69347T~)/

(6.9)

0=a [exp(h9

(15.2518l5.6875/T+, 7.904 + 0.1352K

13.4721 lnT~+0.43S77T 0.007465K~~ + 8.359Th for T , >0.8 5

1~

) forTh, ~O.8

+c0

+d0105)}0~0

(6.15)

wlmcre, a to 1, arc comnstumunls for cads properly as givems in Table 6.5. Fable 6.5 Constants in E2c .(6. 15) for T~, P~, v~,M and Tb. 1
() c t~ P~ P~ (vcfM) (vc/M) M 1)~ 01 lb M l~,
M

+(l .408

0.0l063K,,)/Th,

(6.10)

where Phr=PbIPC. Tt,rTt/rc, Ph is tIre pressure uut which h is measured, e.g. tIme uronrial boilimmg point at 0. lOt 3 MPa (1 urtrirosphcre) unnid K~us tIme Wunlson -Imuirmn&Icrismut ions factor. Eq.(6.2). Ihe estimmsatcd acentric factors by tIne above two corrclaliomms at l~,,=0.I~ duller umbomit 2%. The acentric factor correlation, Eq.(6.9) is simply a re-aiTangeunneunt of lInc Lee-Kesher vapour pressure correlation, Eq.(1.10). The correlations of Cavett [22], given in Appendix B, for T~ and P arc also often used in phase behaviour modelling of hydrocarbon systems. The Edmister0 correlation 1231 for time acentric factor is commonly used with tlse Cavctt correlatnon, w={~[log(~/P )]4(1yrb)_ l]}_t 4 (6.11)

~2 S S S
S

I)

9,523312+00 3 1510011+02
C 1)5102+114

3.)
6 0-19012.05
-

I~~S M S 1 h S
M
S

7 5211812-04 I .1) C 2 I 11+03 3.7659121%)

9.3 t4OF-04 I .1478b.04 85oS(it~l0 I .8078E-03 2.642212.03 2.657012-03 9.775412-04 3.774)12-03

0544.14 - ()(,l(,4 I - 4.81)14 - 0.31184 .0.2641)4 0.52117 - 9.533114

-

d 6.4791E.04
0.00001ij38)

0.811)67 0.2998
.

0.53691 1.0555
4.1)846 1.6015
- 1.2028
-

5.7490l~.03

0.000t)E+00
l.97t0E-03

0.4844 0.8063 0.7506

2.601212.03
-

0.21)3711
0.97476 0.401673

1.3036
6.51274

2.984036

t.9990I3.03 4.252912-03

1.58262

(71) < M <3(11)

3(111< I~ <61)) K)

Perturbation Expansion Correlations

Tlnese nsretlsods imnitiumlly correlate lIme properties of rrormal paraffins as time reference, and then cxtemmd these e(srmelatioums t(r pclroletmmmm fractions. lIme correlations developed by Twum 1261, wlmo nised lIne difference bctweeir Ilme specific gravity of time Irydrocarhon fraction umimd tlmat of tine mmomummal P~ a ff1us wit lu I Ire same I uoi Iinsg poinil mrs I Ire coru elating l)aruumeter arc as follows.
Nor,,,ul /araJ]rns

5 where 13 is time atnunosplmcric pressure, 0. It) I 3 Mt un, mit wlmuch time nsommmmal hom lmnmg pounil . i~,. mu memustmred. Time correluni ion is derived by crmbiisiumg tIne vapour pnesstune relmul imruu. Iiq.( I .8), munsd tire definition of acentric factor, Eq.(I.9). The above methods do not provide immformation on the critical volume or compressibility factor. i~he critical volume can be calculated from. P v =Z RT 0 0 0 0 with the critical conmpressibility factor estiusmated from the Pitzer correlation,
Z 0
so

flie properties of rmornmmal paraffins are correlated wills tine normal boiling point temperature,
=

140.533272 1)343831(11) )i, +2.526167(10

)T~ (6.16)

(6.12)

I.65848(I0 )i~ + 0.0460774 1(1, / I 0O)J

)2

0.290l0.0879tma

(6.13)

2 P~= (0.318317 + 0.099334t~ + 2.89698r~t + 3.00546W + 8.651 ~

(6.17) (6.18) (6.19)

Riazi-Daubert Correlations Riazi and Daubert [24] developed a simple two parameter equatmors for predicting physical properties of hydrocarbon mixtures. Oa Oh n 0< 2 (6.14)

v,.~,= [0.82055 + (1.71 Sr 0.843593 0. I

54 8 6 W + 2.21 266v~+ l

34

Il.Iry~J~
i

286 4 2 W

3.36159w

fl

7495111

wlsere tIre stmhscript p refers to propcrtmes of normnal paraffins and, I Tb/TCP The ummolecnshunr weight of paraffins is givemr by the following inmplicit relation,
=

(6.20)

where 0 is the property to be determined, and 01 and 02 cams be any two parameters characterising molecular forces and molecular sizes of a component. Any pair sucin as (Tb. M) or (Tb, S) may be used for 01 and 02. Properties stuch as tIre mrnolecular wemght, refractive index, critical properties, density, heat of vaporisation urnd llremnal conductivity were successfully correlated by the above equation 124]. The authors later [25] improved the correlation as,

exp[(s. 1264(1 + 2.71579 lnnM~ 0.286590(tn M~)

39.8544/(1nM~)0.122488/(ln M~rJ (6.21)

13.7512 InM~ + l9.6I97(InM~i

224 which can be solved iteratively using the following initial guess, M~= Tb/(5.8000.0052Tb)
Petroleum Fractions

6. Fluid (imaracteri.cation

6.2. Critical Properties

225

The above method, because of its reliability, has been used to calculate the critical properties of

(6.22)

generalised SCN groups, using the reported boiling point and specific gravity in Table 6. 1, with the results given in Table A.2 in Appendix A. Riazi and Daubert [25] compared predictions of their correlation, Eq.(6.15) using Tb and S, with others for i38 pure compounds, with the results given in Table 6.6.

The properties of any petroleum fraction are estimated by adjusting the calculated properties of the normal paraffin with the same boiling point as follows: Critical Temperature: T~ = T~p[(l + 2fT)/(I 2fT)1
2

Table 6.6.

Comparison of several niethods for prediction of critical properties. N~cmlm~s~ %Dev.Crniicaimrmu ~

(6.23) Riazi-Dumubert
lwnj KestcrI.ce

Abs. Average
11.5

11.6
((.7
~ ~ .~

Maximum 2.2 2.4 3.2

~

Abs. Average 2.7 3.9

4.1)

Maximum 13.2 16.5

12 4

AST[~270159hT~00398285_0706691~)~T]

~uvcui

5.5

31.2

so

exp[5(S~ S)]

Critical Volume:
=

0~

[(l + 2f )/(l 2f )1

Tine saturation pressure and density, predicted by tine Soave-Redlich-Kwong EOS 127] for a process where metlsane was imncreunentally added to an oil, are shown in Figures 6.7 and (r.8, respectively. Time Irropertics of time oil heavy fractiomrs were delermiuseci by various issetlmods for pimase bclsavmour curlculations. ihe results clearly demmnomrstrate the major immmpact of the correlation used to cumlculuoe time critical properties on predicted results by EOS.

(6.24)

so

AS40.347776/T~+ (_o. 182421 + 2.24890/T~)As,

40

ex~4(S~ - S2)] -1

35

~:::

Da~e4

Critical Pressure:
=

,,,,.,-~

~ AS~[(2.53262_34.432l IT~ .-t).002301931b)+(_I

(625)
+t).()0414963Fh)ASP]

~~_

I.4277+l87.934/l~

us

EnS ~, = exp [0. 5(S Molecular Weight: mM

~,

S)]

to 0.0

0.1

0.2

03

0.4

0.5

Aided Me(iiane(Oil. Mole Ranto

(lflMP)[(l+2fM)/(1_ 2fM)1

(6.26)

Figure 6.7. Vutriurtiouns of bubble point pressure predicted by Soave~Redhich-Kwong EOS usimmg dificrent correlurtiomms to estimate SCN group properties. TIme correlations are for single carbon number group properties, and their application to very wide boiling range fractions, such as c +, is not recommended. These fritlons should be 7 characterised initially as SCN groups, or by a continuous function as described in the next sectiomm.

EnsMfrlI+(_0.0175691+0.l43979/Tb)ASMI 1

1 = 0.0 123420 0.244541 /Tn,

~M

exp[5(S~S)]

226

6. F/uio!

(/timruilerj.c(j(iopm

6.2. Critical Properties

227

Ow

As the reduced boiling point temperature is equal to 520/708=0.734, Eq.(6.9) of the LeeKesler method is used to calcnmlate the acentrnc factor, which results in,
(u)=O.536

065

lhc cuilculaled values for single cusrhotm nummiher groups, C -C , using the above method, 0 44 are given in Table A.2 in Appendix A.
S

Riazi-l)auhert Correlations lime critical properties cumum he estinrated by eiiher using measured T -S, or M-S data of C,~. 0 lIme resunlts u if bittIn mnpproaclmcs, unsung liq (6. I 5). mure given in Ihe following tumble.
I, K l~.S M. s 711) 711
v, ntn/kguummii

060

I,, MPa
t~,S 1.9112 M, S 1.1135

0.55
00 01 02 03 0,4 1)

t)aiasnsiit
Added Meiliane / Oil. Mote Rali,,

m~.S
1)75(1

M, S 11.75 I

[lie unmeunsunred values mmf spccituc grmmvnly and hoiliumg point, as the most readily available dmmtn, are ctrnmmnmnorm Iy mm sed lo esi unumute tIme criticunl properties. Figure 6.8. Variations of saturated liquid density predicted by Soave-Redlnch-Kwong EOS using different correlations to estimate SCN group properties.
Example 6.1. Ilcirmg tIme i2dmislcr correlatioum. 11q.(6.I I), along with the above calculated T,. and 1,,

results in, (m)000 .4 116

Calculate the critical temperature, pressure, volume amid tIre acentric factor for C~awith properties as reported in the generalised property table. Table 6.1, uniung tIne nmrellsods of Tsvu (I..ee-Kesler for the acemitric factor) and Riazi-1)aubert (l1dnrmsler for lime acentric factor).
Solution.

6.3

DESCRIPTION OF FLUII) HEAVY END

Although numturuilly occurrimmg reservoir Imydrocarhons are commonly described by a nunsber of discrele components amid comnpomscnt groups, they can be more thoroughly expressed by courtitnuous descriptions. Tire TI3P cuirve. Figure 6.1, and the gas chromatograuns, Figure 6.4, umrc exaumrples of such conlimnuily. lime conlimnumous dcscriplion of a flund mmnixture has two major applications: (un) II can be used to iursprove amid cxtemmd fluid clmaraclerisation through describing the plus fr:mction by mm usmmmnmher of siusgle unmsd mmrunltiple carbon numsnher grotups, particularly in the absence of cx pcriimmcnmtuml dumla. (h) the commtimsuomis dislribuliois functious nnay be umscd directly in pisase behaviour models, munstead of d iscuele colmipouseust datun.

The specific gravity, boiling point and molecular weigist of C,, are read as 11.826, 520 K and 190 kg/kgmol, respectively. fronn Table 6.1. Twu Corretalioums Itne properties of lime norumsal alkanne witls lIme smmmnnc bonlnmmg t~~ mis that 521) K, are initially calculated as follows:
T , K 5 Table A.2

itt (~, that is

Equation

T,~.K 6.16

t-T /1, 1 0 6.20

P, ,, MPa 1 6.17

v 01 6.18

S~ 6.19

521)

686.8

0.2429

1.6348

0.8135

0.7635

Next, the above calculated normal alkane properties are adjusled, bussed on the difference between ihe C~specific gravity, 0.826, and that of the normal alkane calculated above,
0.7635.

ilme conncemslrurlions of SCN groups in a North Sea oil sample is. ,shown in Figures 6.9. Althougln complex funnctiomss mnmay be found to describe the concentration distribution in the whole mnixlure, it is more advantageotus to limit the continuous description to the heavy fractmomms, wlnere relatively sinmple functions would suffice. Tlnis approach, which describes light comsiponents by discrete compounds and heavy fractions by a mathematical function, is somrnetimes referred to as semi-continuous description. TIne conrtimsuous description of um flumid is commonly used for compounds heavier than nC , that 6 of us for C +. The Cg+ could he a ummore appropriate choice hr most cases, as the concentration 7 C fraction is getmerally mime highest amongst tine SCN groups, hence, simple decay functions 11 can adequately describe time fluid. A compound iisay be represented by its carbon number,

AST

I~

T,, K
6,23 708

AS

Equation
.0268l 0.003812

I,

v,,

-0.3276 .001399

mfkgniol AS~ f~, P~,MPa 6.24 6.25 0.727 -0.t>3073 0.004386 1.953

228

6. Flutid Characterization

6.3. Description ofFluid Heavy E,nd

229

molecular weight, boiling point, or other properties. The concentration can be expressed in terms ofmole, weight or volume fractions. The distribution of SCN groups shown in Figure 6.9 is typical of most reservoir hydrocarbon liquids. There are, however, mixtures with component distribution vastly different from the normal trend. Figure 6.10 compares different types of North Sea oil saumuples. Fluids
Normal Parallinic Oil

containing high concentration of aromatics and naphthenes often point to bacterial activity in the reservoir. Biodegradation generally reduces the alkanes and, to a lesser extent, the aromatics. Clearly the distribution of SCN groups in non-comnventional summmmplcs cuunnnot be represented by simple distribution functions.
lix)

en

It

Waxy 0/i ~O C3
-

CS

111nC4
C 62

tIJ.lt

..~

,J

Single Carbon Number

29

Aroiltumnic It itxlegraried (lit

Figure 6.9. Distribution of SCN groups in a Nortls Sea oil.

Single Carbon Number Function

A simple, but very useful, approach is to use a functions to describe the comncentration of SCN groups. Various functional forms have been suggesled and applied 128-301, with reursonahie success. The simplest of all is, that of Katz 1281 for the C7~frurclion of commdensumte systenns, as expressed mathematically [191 by,
Naphienic Biixieg,aded 0~i
so

I .38205z~,exp(0.25903n)

(6.27)

where z~is the mole fraction of single carbon number group C~. A linear relation between the SCN and logaritlnm of concentration is generally adequate to
describe heavy fractions of most reservoir fluids, Inz~ =A+Bn (6.28)

where A and B are constants for each fluid. Pedersen et al. 1311 proposed tlnc above equation, and evaluated it for a large number of North Sea reservoir fluids with measured compositional analysis to Cgo. The above simple expression was capable of representing the measured data so well that the authors did not see any advantage in having measured compositional analysis beyond C in preference to calculated data from Eq.(6.28).

Figume 6. 10. Gas chromsratograms of four different types of oil sannples showing line distribulion of varioums components. Reprinted wiih permission 1121. Copyright (1989) American
Cltcunical Sticieny.

20

230 replaced by some physical properties. nunnber by,

6. Fluid (luaracterisatiopu

6.3.

Description of Fluid 1/eats Eumd

231

In phase behaviour calculations the carbon number is umot directly used, lmenmce, it nmtmst be
TIre mnmolecular weiglst is oftemr relaled to tire carbon

Its x~ = 0.4665

0.020056M . 1

4 Ml n

(6.29)

where depends on tire chemical nature of the SCN group. A value of ~=4 is mm rcuisonumhhe approximation in most cases [30].

M~ = 14n 4

(6.30)

a
L .1
V

The abo~ecorrelation, suggests that Eq.(6.28). can equurlly be written iii lerimss of tIme molecular weight instead of the carbon number,

S . . S

lflZ=A+BM~
Obviously, the constants in the two equations have different values.

(6.31)
.0) 80 100 120 140 160 180 200

The above exponential function is also as valid when the concentrurtion is expressed in terms of weight fraction instead of mole fraction. The expression in weiglmt basis mury eveur be nsore appropriate for some fluids. The advanntage of weiglmt basis is tine lack of need for lIme

220

Molecular Weight Figure E6.2. Relation of unmolmmr coutcennration wnilt umiolccular weight for SCN groups. TIne unmolecuulunr weight amid specific gravity of SCN fronmr C , to C .,. are assurnrcd to be the 1 2 sammme mis those mm tic geusermutised table properties. Iahlc 6.l..Suhstituting time molecular weights inn I Ire mnhmive relmml tin rennurns tIme results as given in the following table, SCN
(7 (8 C)

molecular weight data of SCN groups, which are not avurilable for very iscavy frmmctiouns.
Wlsen partial analysis of the C + is available, the constants cams he determined by regression, 7 mininsising the sum of squared differences between the cumictulated mnmnd mmremmsummed comsceurtrumlion of known SCN groups.
E.rample 6.2.

M
.4 117 126 40 153 16.5 81) 07
209

The total concentration of C7+ fraciion of a gas condemnsale is 3.92 niote% with the analysis as follows. Extend the analysis to C3O+ by SCN groumps. Table E6.2.
Composition and properties of C7+ fraction,

CtO
CII Ct2 CII (14 C IS C tO C17 (18 (19 C21) (2 I (22 (25 (24 (25 C26 (27 (28 (29

~ C,
(~

C 1 C 12 C 11

211.2)) 21 .4 I 12.11 9.23 7,17

5.68
427 3.05 2.43 4.45

14
0

94 11 7 126 140 153 t65 180 197 209 374

0.730 0.754 0.769 1)785

0.799
0.806 0.820 0.843 0.844 0.909

222 237 251 263 275 21)1

300

Soluaio,n

Fmgure E6.2 shows the relation between the nroiar concemmirumlioum nnd nimolectular weiglml of SCN groups in this example. Note thimt the uusxumtmstntion of a hinucuir relunliont heiween tine logarithm of mole fraction and the molecular weight, Eq.(6.3 I), is reasonable for this fluid. The two parameters of A and B of Eq.(6.31) cars be determimmed by tine least squuare fit (excluding C,,), resulting in,

312 324 337 349 3611

372
382

S 0.730 ((.754 11.769 0.785 1)799 11.806 ((.82)) 0(143 (1.844 1)843 (1(1St (1.856 ((.861 0,866 0.87 I 1)876 ((.881 II (4(15 (1.888 1)80)2 11.8% 11,899 0.902

xM,

sM/S

1)2(121) Ii 2141 0.1211 0.0923 11.0717 0.0568 ((.1)427 11(13(15 (1,0243 11.1)1 86 11.0137 0.0104 (1.1)082 (1111164 0(1(147 (((11)39 (1(11)31
((.111124 (1.111119

18.9880 26.01 25.0497 33.22 5.2586 9.84

12.9220
10.9701 9.3720 7.6860 6.111)85 5.0787 4.1237 3.2586 2.6062 2.1467 1.7645 I .3546 1,1(159 1)9532
11.7781

16.46
3.73 11.63 9.37 7.13 6.02 4.89 3.83 3.04 2.49 2.04 t .56 1.33 1.1)8
0.88

Tmtnat

0.18115 0.11012 0.188)9 (1.1111(18 (1.933

0ft236 1)51)76 (1.4200) 1)341 I 0.2867 131.7

(1.71) 0,57 (1.47 (1.38 (1.32 166.99

232

6. Fluid (Iuarncter,satio,m

6.3. Description of Fluid heavy End

233

(a) The C~,mole fraction is calculated as,

C , 2

Example 6.3. Ihe mole fraction, molecular weight and specific gravity of the C7.~.fraction of a gas condensate sanmple are 0.0392, 165 and 0.8 15, respectively. Describe the C7+ fraction by SCN groups extended to C20+.
Solution:

Xe~, r1_~xc,
C,

=1-0.933=0.067

(b) The C , fraction molecular weight and specufic gravity shotuld renmimin the suuumme when 7 described to C,~,. Hence the molecular weight of C,, is deterimniimcd ilS, the fluid is c,,, c,,
=l65.4=~x(.M(.
C~

Rewriting Eq.(6.31), we obtaium, ,s.,

=

M~,= ~XC.M(.
C,

+ xU,,MU,.

so

131.7+0.067
(~

M~,,,,= 503
(c) The volume of C , fraction can be considered equal to the surma of voluuummes of mmli its components. Hence2 a similar approach to that of imrolccuuiar wcigimt cmmmm inc used t~ calculate the C,,~,specific gravity. Me,, /S~ , 7 S,,O.81 5 The volume balance for C ,, results in, 1
C , 2 M~, ISO, 2~XcMc ISe, +XcsoMe,, /S~ =166.99+0.067Ix503xM~,,, /S~, C,

so

~Z(~
C,

eA~e~0 C,

Sunurularly for Eq.(6.33), we obtain,

C
so

9 ~zC,MC,
(~

UN

=eft~~Mc,etu0mcn
(~

~
C 7

/S~,,=202.86w l65.4/S(. , 1 I lermce, M~, timat is,

C, =~

/c, ~M(.ci~t~

)/(~enm~tn

\
j

202.87 S~ ,= 0.940 35 The molecular weight and specific gravity of C , could huuve been cmmlctilaled itm tins 5 C ,, fraction only. im~stcmud of time C,, example by the mass and volume balance for the 1 fraction. When little or no compositional analysis of tire C .,. is available, tine two constants cain be 7 determined by solving the following two material balance equatiours:
Cn C~, .
I I

~
C 7

soO

(E6.3)

i3q.(i36.3) dernmonm.slrmutes tlrat lIne slope of the SCM group dislribution line, B, depends onmly out the irmoiccuilar weight of C , . The puurmmnremer A affecls only the mole frurction of 1 (7, in mine nnixturc, by simiftimng tIme line up or downs in Figure E6.2.

A so Inn. (I

~e~~t(n
(

~ )I
7

(E6.3)

Norumnumhising tire di~tributionnof SCN groups mm time C , fraction, by making Zt.,,=l. resmmlts

~
C,

so~exp(A+BMc)=ze,
C, C,

(6.32)

I Z~ Mc , 7 7 (6.33)

A so irs z

.,

~eBMtn c,

~ j

=~exp(A+BMe,,)Mc,
C 7

lImmul is, jusl the ummolecular weight of C , fractions is suufficiermt to describe the distritaution

CN is the heaviest carbon nunnber assunned to be present in the mixture. Values of 50 1321 to 80 [311, have been suggested as the cut-off carbon number, whereas larger nrolecules with higher carbon numbers are generally present in oil and gas condensate systems. The choice of the cut-off carbon nurmber, however, has very little effect on predicted results by EOS in most cases, as the contribution of very heavy fractiorns, due to tiseir low concentrurtions, is minimal for practical purposes.

of its coulmprisiung SCN groumps.

Assumimming SCN groups wills nrolecular weights equal to those in the geuserahised table, Table 6.1, and C as the tmeaviest fraction present in the mixture, Eq.(E6.3) resunlts in, 11 l3~0.013141 8.

234

6 F/tool ( l,a,a~term,calio,u

6 1 /)ese, iption of Fluid End cummbomm ntnrnmhcrs. trm hleamu ntmurtlncnraticuml terms, tIne funchions provides lIme value of concentration 235

Substituting time B vmmhumc in Eq.(E6.3). time valnuc of A is deterussimned equal to .5.81)1)8671 Eq.(6.31 ) is then used to calculate the mole fracliomm of SCN groups as given in tIne folhowiisg

table, SCN (;roup

Mole S

initegral between Cui.l and C . 5

so

C7
0.6273

C8
0,5429

C)
0.4516

C 0
0.3807

CII ().3209

Ct 2 (1.2670

C I 3 ((.2221

(I 4 I). 11124

(I 5 ((.1478

C 16 0.1198

id/.,

(6.34)

SCN Group
Mote %

C17
0.0983 C27 0.0195 C37

C18
0.0818

CI)
0.0699

C20
0.0597

C2I
0.0484 CIt 0.0110

C22
0.0430

(23
((((367 CII 0.0082 C43 ((0024

C24
0.0313 (:34 (1.4)1)71 C44 0.002!

C25
((.0264 (35 (((((1(94 (45 11.0(119

C26
0.0226 (:36 0.0055

where i refers to all tIre commmponenls counrprisiumg thue SCN grotup n. A more appropriate approurch is time continuous description of the fluid, where tIme distribution of all its conslitueusts, instead of carbon groups, is aimed. Indeed the carbon groups reported by lmrhoramories mire detentnrined by integration of corsipotunds comprising the groups. For exmnmmnple, Ilme area umuuder time curve in Figuure 6.4 between nC and nC is taken as fine 9 10 comsccnrlralion (If ~ grotip. the connhituunnini~uiusci jIlt ilium nctl((Is 11w trure nmmuluine of reservoir firuids hinal urre corssposed of nummiuty ctimiiliiummmds, with priifs.n tues vmim ymung so gu udumally himmut (10 mnol allow (lnstilmclive iulennt i ficumi i) iti lime contitruorns dislrihutiomr of comrponnenls can he expressed by a function, F(l). suds that,

SCN Group
Mole % SCN Group

C28
01)167

(29
0.0146

C30
0.0125

C32
0.0095

C38
0.11043

C39
0.0038

C40
0.01)13

C41
(1.0030

C42
0.0026

M,,le %

0.005(1

The C,,,, fraction is determined by sumniming up C, C, mole fracniouus, i.,.~,=0.4074%,with 0 tue unolccumlmrr weighst dcteruusinied mis:
(:4,

M~

51

so

~XC,MC, C ,, 2

=i3~

Assnuusning tbmat SCM groups have tine samsne specific gravity as those in the genmirmmhism,d tumble, the specific gravity of C , is deternniumed siunnilar to that in Example 6.2. 21
C,, ZC ,MC , IS .,,,, 51 210 5
=

5 F(I)dl=z

(6.35)

/S~.,

wlrcrc z is time totuml councemntration of all line connpouscnts, represented by I. within the integral boundaries. If all tIne coinrporrcnts of a fluid is described by the continuous description, then,

S,.,,,=0.892 The normmrahised mole fractions of SCM groumps iii the C , fractious cmulcuilaled in this 7 tbnose calculated in Exmumusple example are given in the following table for comparison with 6.2, llmmrt is, for a fluid with the sanme C,, fractioun properties. SCN
Cl

J F(l)dl = I

(6.36)

In lIne nsmorc pruiclical setuni-coumtiunuious description, only the heavy part (say > C,) is described by line contimstmous fumrctuorm,

C8
C9 CI)) CIt (I 2 Ct 3 C 14

P.6.3 0.1604) 0.1385

0.1152 0.0971 0.0819 0.068 t (1)1567 (1.0465

E.6.2 0.2020 0.2141

0.1211 (1.0923 11.0717 0.0568 0.0.527 (1.0305

F(l)dl =

(6.37)

svliere ~t) is line cormceuntraliomm of Ibte lseumvy frurcliour of tIme fluid described by the continuous ulcscm iptuour. llrc distrihmmtroun ftuuicliomu is generally selected sucln that the value of integral in Fq.(6.37) becoimmes eqmimil to I, Ibmurl is, it describes line relative concentration within tine conml tusumouns purt . I Ic m5cc , urn lime seitmiconl i nuotis uupprourcls the calculated Concerslration of tine comnttunuloums coussl ituiemsls slmouuld be inntultiplied by zI) to mm(srunalise it in the total mixture. lhre musteumsity chislrihumliotn, F(l), or lIre prohmnhihity of occurrence in statistical terms, is often expressed try nnolmur dislrihtnluomm, aind so used ins hhnis hook. time mass or voltuirme distributiomr, uss obtumined in gas cisrousrustograpisy or distillation, respectively, cars also be used to describe the concentration, Tire variable I, can be the carbon number or any property, such as the urnolecumlar weigist or tIme boiling point. characuerisiumg the compounds comprising the fluid. II shsould be umoled tlmal the conlinstuouns distrihulioun is valid at all the values of I, within the idenlnfued range of commnponeurts, contrary to the SCN group function which is valid only at tIre discrete carbon numbers. The insole fraction of SCN group n, can be determined simply by
integratimrg the distrihutioum function between (n-h) and n,

CtS
(tO

(1.0377
0.0306

0.0243
(1.1)1 86

C17 C18
CI)

0.0251 0.0209
0.0178 0.1(139

0.0137
0.0104 0.0082 0.0936

C20+

Continuous Description The above approach of describing the concentration of SCN groups by a function, nsay appear as a continuous description of tIre fluid. but it is basically a discrete representation. The finnctiomi describes the heavy part by a number of SCN groups, atsd is only valid at the discrete

F(l)dI so

(6.38)

236

6.

Fluid C/uaracterisaiio,u

6.3. l)e.seription ofFliii,l Fleas, Enrol I) =

. .

237 (6.45)

Similarly, the mole fraction ofany carbon group, or pseudo component, can he determined by incorporating the appropriate boundaries in the above integral. When the molecular weight of each component, M, is used to represent it ins llue distrihutiomm function, that is I M , the above equation becomes,
.

cM.
F(M)dM zc (6.39) which is the area under the curve of F(M) between M . and M . 51 5

rime duslruiuuutnon fimnctnomr us to getnerutlly to describe tIme C + fraction, witlm itsdata, parameters delermmrined by regression mnmatch used tine available SCN group experimental using 7 Eqs.(6.39) and (6.40). The valune oft should ire between 86 and 100, tlme molecular weiglnls of nC and nC , respectively, based on the defiumition of C .,.. It can be treated, irowever, as a luning 6 parannreter 7 un tlse regressiorr, or assumed eqtmurl to the 7 mid-value of 90 in line absence of unmeasured SCN group data. TIme function so developed can tisen be used to exteind tire fluid descruptmon to heavier compounds. tire vuulume ofy, ranging approxiumnately between 0.5 to 2.5 for typical reservoir fluuids, conlrois the distribution skewness. Figuire 6.11 shows typical distributions with different values of y kr a C .,. fractious witim MC7 =20() ammd t=92. Values of yequal to. or less tiran, one represent 7 5 mmmixttmrcs wilh cotstimstnotrs decliume in commceumtratioum, winereurs valucs more than orre demmmonstrmnte a mmmaxiunsurms in concermtratiomn. lime peak will smut towards Iseavier fractions by increasing the y value.

The molecular weight of SCN group n, is determimimmed by,

JF(M)MdM = M~z~ (6.40)

The most widely used distribution function is the gurummmsra probability fumnclioun, mrs proposed by Whitson [9], using the nsolecular weigint as the chumnmcterisatiomr variable, F(M) =

t)~i

exp[(M

t)113]I/[IYF(y)I

(6.41)

where r(y), is the gamma function, t is the mininnumsi molecular weigint included in the distribution, and y and ~ determine the shunpe of the distribution fumsctioms with lime mean and the 1~32 variance equal to (~3+t)and y ,respectively. Hence,
so

F(M)

(M , c)iy 1

(6.42)

wlmere MD is the average molecular weiglmt of tlre continuous part, courmprised of compounds with molecular weights beginning fromsn r and exlemiding to imsuinity. The gamma function is, 100 (6.43) Figure 6.11. Distribution of commnponents represented by the ganmrma probability function with different values of y. SPE Copyrighl. Reproduced from 191 with permission. (6.44) As described by Eq.(6.39), the fraction of area under the curve between two molecular weights, identified by the sinaded area in Figure 6.11, demonstrates tIne mole fraction of a pseudo component commuprised of all compounds with molecular weights between M . and M . 51 5 The value of p I, reduces tire gannma distribution futmction to, F(M)={exp[_(M t)i~3]}/I3 winds is a sinmple expourential distribution, l2(M)=~i~_Li~exp(_M/I3) (6.46) 200 300 Molecular Weight 400 500

r(y) so
which can be estimated from the following expressions [331. r(y)so1+~A(y1) Iy 2

where A represents the parameters in the approximate polynomial, 1 A 1 A 2 A 3 A 4

= -0.577191652 so so

0.988205891 0.918206857 44

so -0,897056937

A = -0.756704078 5 A = 0.482199394 6 A so -0.193527818 7 A = 0.035868343 8 ), tire recurrence fonummunlum rsnumy he used to evaluate

For y values otutside tIme range in Eq.(6. the gamma function,

(6.47)

238

15.

Fluid (Iuara,teri,mtjolt

6.3.

Deccriprionm of

Fluid F/e,uuv End

239

Substituting the above distribution function in Eq.(6.38), mind nntegralimng it, we ohluuumn,
so

_exp(t/~)texp(_M, I~) exp(M, .i

(6.48)

The mole fraclions of .SCN groups can be determined from Eq.(6.49), with M,=l4n+2 for nornmah umlkammc houun(laries, [ ( 14 ~) lnz~ =lnlexpi II [ ~M,,t) wtnicim results un,
=

where M, and M, are the upper and lower molecular weights of the comnipommcun)s presciml irs tine SCM group n. The molecular weight of SCN groups is determined by stubstituntiung line distrmhumlion function mu Eq.(6.40). _~eXP(t/~)ft~~ + IJexP(_M, i~)+ lJexP(M i /I3)]/zc 9 (6.49)

________

(l4n+2)t M~ t 1

7 z~, cxp(0.55 I 8(m I

M, / 79)=0.0392xexp(-0.577 18-0.17721 Sn)

Note ttmmrt lIme normnnmuhised mole frmuctions. calcunlated by Eq.(6.50), have been unnultiplied by

Eq.(6.48) cmmn be wrilteun in the simsnphe logarilismie fonss of Eq.(6.3 I) by rclumling M,, arid M, 1 logether. Assuming tlnat M,- M, =l4, we obtain, 1
lnz ,
~ so

lIne C,, unmolc lrmuctuonn. lIme rmmohe frmrciion, oh czncis SCN group cars he deterrmnimme(I frousm lime above equation. The simm of nsole fructuous hecouuses equnal to i,,, omuly when carbon nuumbcrs up to infinity are rnclumded. lime cormtrnhuulion of very hucmnvy connpoumnds to the total mole fraction, however, heconumes insignuificanl, mis stuinwuu mum line following table for various cut-off carbon numnsuhers.

14 ~\ M t Ini expi IIl t~,M t} j M t 0 0

I ( [

(6.50)

which is similar to Eq.(6.3 I), with the values of A and B as follows:

A= In

expl 14 ~M~,c)
1

M cj 0

(6.51)

~
(6.52)

I.asmcarhonnum,u,hcr 30 41) 50 61)

[.ast u,iiulccular weighi 422 St,2 702 842

Ti,tatnoruunaliscdnmole_fractions 0 1)858 0.9976 0.9996 0.9999

Eq.(6.52) can be used to calculate am initial guess for B to solve Eq.(E6.3). Note that M, in Eq.(6.50) is the highest molecular weight present in tire SCN group and M~,inn Eq.(6.3l) us the molecular weight of SCN group.
Exannple 6.4.

Describe the C7+ fraction of the fluid in Example 6.3 by mm countmnnuours function run lerms of the molecular weight, arid uuse it to estimate tine mole fraciton of SCM groumps comprising the C7+ fraction. Solution: The value of y is assumed equal to one, due to lack of partial analysis of the C,, fraction. The value of f3 can be calculated from Eq.(6.42) after seIecling the lowest molecumlar weight present in C,,, that is, t. The simplest approach is to select the normal alkane molecular weights as tIne SCM group boundaries. Hence, normal hexane will be the component with the lowest rsnolecn.uiar weight compound present, t=86, which results in ~3=79. Iherefore, mhe distribution 47 function of time C,, fraction in terms of the mnrolecular weight, Eq.(. ), becomes, F(M)= 0.0376 exp(-Mfl9)

tum phase hehmaviomur calcumlatioums, time properties of identified SCM groups must he known. Aluhounglm the selection of ummolecnuhar weights of norumnal alkatnes as the boundaries agrees wiibm the convenijoimmul definnitious of SCN groups, that is the fraction collected between two coissecmntive norrnrmul mulkarme hoihimig points, it results to carbon groups wills average nmolecumlar weights sigumificantly different fromir Itnose of generalised values. For example, mire C, group of this example Immus, by the above approach, a nmolecurlar weight of 625, 9 that of tine genrerahised SCN group has a molecular weiglnt of 539, Table A.2. whmercuss [Inis is dune tui preseunce of arousmatics and naphtlmeuses in the sanmple under consideration. ilensce, the unse of SCN group properties for the above characterisation is not justified. 1mm order to cimmuraclerise the C,, fruictioun with SCM groups similar to those in the genermilised tumble. their unnottcutar weights sisould he used to identify the group houuumdmnries. IImc rmmolcculmur weiglut houmundaries cumn he determined by solving Eq.(6.49)
for lire geunerahised SCN groups. The dependeuncy of calculated boundaries on the C,, ummolectulmur weiglmt amid selected vmulume of t, is insignificant. The calculated results will not typncmrlly deviate rnmore tinmun OliC unit from tire values determined by simple linear

unveruigiung, M, =(M~,

M .)/2+M . 1 1

Ilic SCM grouup nmolecuuhumr sveiglnl hounnsdmuries, and the values of group molecular weight calcumlated by hq.(6.49) are given un Table P6.4, and compared with the generalised vmnhrues. Mote that the valuues are cquuaI willmin the accunracy of measurement. The given ununlecuular weigini boundaries are germeral and earn he riced for other fluids in the absence of numeunsuired ,mmokcnnhmnr we ig ImI

240

6. Flmmuol (ha r,neteun.catuotn

6.3. l)e.ceriptionu

of Flm,id

F/eaui Eiud ClO CII C12 CI) CI4 CI5

241
CI6

Table E6,4. Generalised SCN group molecular weight boundaries. SCN Upper Molecular SCN Group Molecular Generatised SCM Group Weight Boundary Weight, Eq.(6.49) Molecular Wg~lmt, Table 6.1 kg/kgmol kg/kgmnmol kgFkgumu,uI Cl 102 96 96 C8 113 107 1117 C9 128 120 121 ClO 141 134 134
CII 154 147

SCN Gromup Mole % SCM Group Mole %

SCM (irmnip Mmutc % S(N Grump Mole %

C7 (8 (9 0.5573 1)5162 0.4689 CI7 (18 C19 0.1073 0.0800 0.1)625

C27 (1.0164

0.3784 0.3293 0.2843 0.2436 C20 C21 C22 C23

0.0614 C30 0.0101 0.0459 C31 00083 C41 0.0331 C32 0.01)76 C42 0.0325 C33 0.ttt)65 C43

0.2132 C24
0.0288 C34 0.0049 C44

0.1784 0.1401 C25 C26

0.0244 C35 0.0043 C45f 0.0191 C36 0.00311

C37

11.0034 1)0022 0.0018 0.0018 0.0016 0.0013 0.0104 tIme untIe trmucuu,n oI (,, I,mus hccum adjnusucul no nuiakc the total mmmu,lc equal to 1,92%. i.e., the nn,,te fraction of C,, in tIne original fluid.

C28 li.tt135 (311 11.0031

C29 0.1)116 (39 0.0027

C40

C12 C13 CI4 CI5 C16 C17 C18 C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33 C34 C35 C36 C37 C38 C39 C40
C41

168 182 198 214 230 244 257 269 283 295 306 318 331 343 355 366 377 388 399 409 421) 432 442 450 460 470
479

161 175
19(1

147 161

175
191)

206 222 237 250 263 276 289 300 312 324
337 349

2)16 222 237 251 263 275 2)t

300

\Viren sufficient inrfortmmuutioun our the distributions of SCN groups in Cy+ frmmclion and their moleculmu weight is avuuilahle, tIme putrausneters of tire distribution function can be oplinnniscd by mnmmrtclming tIme unneastired group (latun, instead of urssunmimsg yso I ausd using generanlised nmolectular weight datmu mis in time above cxanmmplc. Whitsomm Ct al. [341 proposed a procedure where the measured unass fraction of each SCM group is used to regress the three parameters of tIne gamma function while adjusting the molecurlar weight houundaries of SCM groups. They applied the imsethod to 44 gas condetssate umnd oil smumnples, and deternnmned time parameters of the gamma prohurhility distribution function for theism. Fire arutimors conclrmded tlmumt rand ycould he reasonably correlated by,
so I l0~ I 11(1 + ,~mm72l~1

3t2 324
337 34) 3(n()

(6.53) witlm t related to y by tine above

360 371 3142 393 404

414

I leusce time unptinmnisalioim cams onmly he cousducted for y and eqrmat ion.

~,

372 3142 394 41)4

415

lii tircir proposed mnmetirocl a value f(nr y is initially mrssunred, Ca. p1. and values of t arsd ~ are calculated from Eqs.(6.53) amid (6.42). respectively. Tlren the upper boundary of the first group is assumed atsd its mass fraction is calcunlated by,
so

426 437 446 455 465

474

426 437
445

zc M~/(t + )j~)
1

(6.54)

456
464 475 41)4 495 5i)2

490
499 507

484
494 503

C42 C43 C44 C45

517 526 535 544

512 521

512 521

where um and M are the umnole fraction and molecular weight of tlre selected group using 5 Eqs.(6.39) and (6.40), respectively. If the above calculated mass fraction does not matcln the experinmermtal valn.ue, witlmin a toleramnee of ~ the upper boundary of the molecular weight is adjunsted. llse urssummmption mmmd uudjustnsent of the upper msmolectnlusr weight boundaries are cointinued for otimer carhoum groups sequentially, up to the plus fraction. The sum square deviuttioums of curlctulmmtcd mmsolecunlunr weigints of the groups, umsing Eq.(6.40). from the measured vumlumes mnre Ilmeun evalrmamed. The parammmeters of y and ~1are optinmised by minimising the above stmumm. After lime oplirnniscd values are determined the reliability of the model is checked by comsspariumg tire calculated mnoleculmur weigints and nnole fractions with the experimental data. Fine use of SCM group data to tietermnnine the paraunreters of any continuous description is only reliable wincis the conceustratioms annul nunolcctnlar weigirt of sufficient number of SCN groups are known. Winitson et al. [34], evaluated their proposed uuiethod using different numbers of SCM group data. In general time mnnelhod was found reliable when information on more than 15 groups were used in optimisation of tine parameters. Direct Application The continuous distribution fumrction cain be used in phase behaviour calculations similar to discrete comtmponents. The distribution functiomn, as expressed by the integral in Eq.(6.35).

530 531 ____~_9 _______________________________ 539

Selecting r=90, the value of ~ is calculated fronn Eq.(6.42) equal to 75, with the distribution function as,

F(M)= 0.04427 exp(-M/75)

The mole fractions of SCN groups can be determined from Eq.(6.48), with the molecular weight boundaries given in the above table. The results are as follows:

242

6. F/ui,! ( !,,uu,m,ferisar:o,m

6.,). l)e.r,rupti,nu of Fluid I/eats E,td

243

replaces the concentration of individual discrete consponents. TIme propertres of time components, such as the critical properties required in equations of state, are mulso expressed by contunuous functions of the variable I, which represent comprising coumnpounds. For example, the two parameters of a vain der Waahs type eqinunlmon of slate, tmsnmsg time random mixing rsules of Eqs.(4.7374), for a mnxttnre described by lIme senmmi-coumtmnntnous umnetlmod, are calculated as,
SN on

umunusuericutlly even for simple distrihutiots functions. Anurongst the nummerical melimods applied sticcessfully 1361. the nnetlmod of Guuussian quadrature has proved to be reliable and practical. lire nnrethod approxumutles the vuiltue of integral numerically by ad(Iing the weighted values of a ftuusctioum usl mm tmunnher oi specified vuuhiues of the vumriahle. called the quadrature points,

Jf(l)cXp(I)

d(I) =~wkf(Ik )

(6.59)

a so ~~x x,(a, 1
N

a,)

I 5 +IJF(l)(a(l)) dl + Lu J

(6.55)

t where root, k.wk is tire wcnglst and f( t) is lIme vumlue of the function f(l) at the quadrature point, or Selccllmrg line gausrurma prohunhulmly functiomr, Eq.(6.4 I), to describe the eonlinuous fraction, the imrtegral in Eq.(6.37) becomes,

2~x a 5JF(I)(a(l))~dI 1 3 where the first term accounts tIne atlrmsclionn hctwecmn time discrete couunpouuumds. tIme sccomnd is that of the continuous fraction, and the tisird is dime to uumleractmots hctwecmn lIme discrete aumd tIme continuous parts. Similarly for the repulsive term, or co-volume, we get,
N

c (M
it

tY

exp[_-(M

t)I[IJ
=

fPE(y)

(6.60)

sinounid he unoted Ihat lime fuinctuoms provides lIme distrihumtioms of the constiluenls witlmin the

contimmumotus pmirl. imcnice. willn mu total valuue of I.

Defining a mew variumble,

b so ~x b~ + JF(I)b(l)dl

(6.56) X(M-t)/~ reduces Eq.(6.60) to, XYICxP(_X)dX F(y)

so

(6.61)

The terms a(I) and b(t) are souse continuous functiours of I with values equal in pusratsmelcrs a and b, respectively, for compounds described by the comrtimrruouns description. Exuimples of such functions are those given by Cotterrnan [351, for lime Somnve-Redhiclm-Kwoumg EOS, whrere he uses the molecular weight as tlne parameter representing compouunds. The equality of fugacity of eacir component in co-existimsg pinases mit equilibriuunn is expressed for the discrete and the continuous parts, as follows,
fV so

(6.62)

wlniclm cats he shown as, J,f(x)exp(x)dx I (6.63)

~i

(3.30) (6.57)

f(l)= fn.(t)

where ~(~)
I
~--.

lIme unmaterual hunlumusce equumitions requn ired for plmase helium v our ca cmi lumlions, Sectlout 5. I, utecd to he umnodilued also by incltuding tIme continumous description. For mu mmtixttmrc (lescrihed by the semicorstinuous rsnethnod, the umimulerial bumlance equations for time total unnixlunrc mind tIne (liscrete comsrponents remains the same as Eqs.(5. 12), wilhm mmum additional equnaliors for tine countinu(nus fraction, nt Fr(l) so nLFu(i)+ nVFV(I) (6.58)

(6.64)

r(y)
Suhsluturtionm of lIne couumpositionsush (histrihution of Eq.(6.63) for 5 F(I)dI , in Eqs.(6.5556), reduces tine instegrumls inn the expressions for a amid b, to tine fol1owin~ general form,

wlrere FF(l), F(I) and Fv(1) are the distribution fummctions for the continmmomus fractions of feed, liquid mmmd vapour respectively. It has been shmowmm tlnat whets a slistnihumtmon function, sumcls as Eq.(6.4 I), describes the continuous fraction of a pliunse. tine suuunsc type of fuunction also describes tine oIlier equilibrated pisuise 1361. TIme range of cousnpounds described by mill three functions is the same. lhnat is t in Eq.(6.4l). hut otimer paraunieters. y and ~, are generally different. Tire use of the distribution function, and the resulting derivatives fromrm Eq.(3.3 I), to detennine the fugacity coefficients leads to complex integral equations which can ously he solved

L f(X)exp(X)dX

(6.65)

I
I

Ohvnously time functions ~(x) is duffcretnt to that inn Eq.(6.63) and depends on the expressions for uuM) umund h(M). lire roots utund werghts, however, are the same for the integrals as they all belong t(n the sammie class ((lauss-Laguerre) of functions. Applytusg tIne quadrature integration nmetlnod to Eq.(6.63), we obtain,

244
,~

6. Fluid Cbnara,ierLcagio;n
i-i

6.3. i)e.criptiouu ofFluid tjeau,~Eund

245

wkf(lk) = wkf(Xk) = Wk ~?

(6.66)

where,
~zkl

those developed by Alsmnned [191 using tine generalised SCN group data, can be used to estinsate these properties, Table A.2 in Appendix A. A more specific approach for each fluid is to assume that the characterisation factor of all tine pseudo components are the same, and use Eq.(6.4) which relates the molecular weight to specific gravity. K~=4.5579M i.Si78~-O
8457i

(6.67)

(6.4)

and (6.68) That is, the values of root and weight at each quadrature l1oiust cats idemnlify a pseudo component, or a carbon group, k. with the molecular weight and nmole frauction given by Eqs.(6.68) and (6.66), respectively. The calculunted mole frmucturns mire relative to tine continuous part and should be nornialised by multiplying tlncm witls ~l)~ Therefore, replacing the continuous part with a number of pseudo comsrponents equal to the number of quadrature points, with the molecular weights and mole fractions as given above, should lead to the same values of a and h, as those by the continuous approach, Eqs. (6.5556). Behrens and SandIer [321 were the first who suggested to use tine pseudo componemsts determined at the quadrature points of the feed distribution function in vapour-liquid equilibrium calculations. They assumed a linear hogmnrithmic distribution fumnction, y=I iii Eq.(6.41), with C as the last carbon number to describe a nunnber of oil mutmd gas condensate 50 samples. They used two point quadrature integration procedure and deunmonstrated the capability of the proposed method. The accuracy of calculations increases with increasing the number of quadrature points. Any large molecular weight can be selected as the cut-off point of tIne distribution fumnctionm with little effect on the results. For mathematical sinnplicity, the distribution cams be assunssed to extend to infinity. The roots and weights for 2. 3 and 4 point integration, with tIne continuous function extending to infinity, are given in the following Table. Table 6.7, Roots and weights for Two (r~,1r,.,~r. P,s,n,., Root,X Weight. Wk Three Quadrature Poinis Root, ~ Weight. w~
op~rat~mrePoints Root, ~

In this approach, the Watson characterisation factor is calculated first for the total continuous part. e.g. the C fraction, using tine measured values. The same K~is then used for all the 7~ tu determssine timeir specific gravity. pseudo comimpommeusts WImil sinus et mil. 137 ohseuved thaI rusiung mu conrstaunt Wmnlsoim cimaraclerismnt iou lmnctor for mull lrsCti(Io componetsts leads to specific gravity values which do not correspond to the specific gravity of the mixture, calculunted by, S
so

1~zkMk]F[~i. M /S~1 5 5

(6.69)

Hence they solved tIne above eqtmation sinsuitaneously with Eq.(6.4), and introduced a different characterisation factor, C,
N

C = [o. I 6637Sn)~zkMk

/(7. M )] 0 0

(6.74))

and proposed the following relation between the nroieculan weigint and specific gravity, so 6.0l08M~ un 4m C
i 18241

(6.71)

The calculated specific gravity and tine nnolecularweight of each pseudo component are used in generahised property correlations, stmch as those proposed by Twu in Section 6.2, to estimate properties required in equations of state. When the same type of function is used to describe tine feed, vapour and liquid phases, tire roots and weights at the quadrature points are the same for all the phases. The associated molecular weights of the pseudo components in each phase, however, are bound to he different. They will become the same only whemn ~ is the same for all phases, as required by Eq.(6.6I). As yr) represents the mean of the distribution function, that is tire average nnolccular weiglnt wimicim is different for tine two equilibrated phases, the value of ~3,is generally differeint for the vapour, liquid and the feed. Figure 6.12 shows the distribution of SCM groups in time co-existing guns mind condensate phases of a North Sea reservoir fluid at II MPa below its dew point. Clearly, m~ single distribution function cannot reliably describe both phrases. Iiemncc depcmrthiumg on tIme distribution used, tlmal of feed, liquid, or vapour, the selected pseudo components will be different. Furthermore, tine selected values will he only at the qumrdnature points for time phase whose 1~ value has been used to calculate the group nnmolecular weights. Time above treatnsent, however, does not significamntly immspair the results in most cases. The use of pseudo components at correctly selected quadrature points is as effective as describing time fluid by twice as mauny pseudo components randomly [361. It is a valuable approach to reduce the number of components describing the fluid to speed up calculations, as described further in Section 9.1.

the quadrature method. 0.5851)

0.8536

3.4142
0.1464

0.4158 0.7111
0.3226

2.2943 0.2785
1.7458

6.2904) 0.0104
...

Weight,

Wk

0.6032

0.3574

4.5366 0.0389

9.3951

0(8305

Note that even with only four quadrature point integration the weight of the last point, representing the concentration of the last pseudo component, is quite small. Higher numbers of the quadrature points hardly improve the results, particularly as the molecular weight of the last pseudo component will become excessively high for which physical properties cannot be estimated reasonably. The above approach identifies each pseudo-component by its carbon number, or molecular weight only. Approximate expressions relating critical properties to carbon number, such as

246
0.3

6. Fluid (!mauncterj.c,mtio,i

6,3. De.ceriplioru of F lmn,d 11cm, End 247 Suhslitmntiung the above value in Eq.(6.7 I) provides the specific gravity of pseudo

componenls as given in the above table. Evuihuating the calculated values of specific gravity, by
0.2
0
U 0

Me,, /Sc,, = ~zkMk

5

5 k

U.0 8~1

we nhiain, S,,=0.X 149, which is the same as the mixture value.

01

The normalised uusohe fractions of the pseundo conmrponents calculated above should be nmiultiphieil by z,,=(l.0392, to obtain lime mole fractions in the total mixture. TIne critical 1 lununliertims of due identified ps<unr mi comnmpommemsts cmnmm he cumhcmuimnted from Eq.(6. IS), using I hueir mumolccunImur we ighit mmmd specific grmnvuty vmshues.

6.4
(10 5 10 IS S(N (;~~ 20 25

REIEREN(TES

Perry. A.J: hntmoc.huictioui to Ammumlylicuni (las Chnrotnatograplmy, Dekker (1981). 2. ASTM: Distihluulion of Cnude Petroleum, Designation D2892-84, Anmnual Book of ASTM Standards, 821-860 (1984). 3. Varotsis, N. mumnd Guieze. P: Microdistillation. a Method Combining Gas and Liquid (hiromatograpimy to (Immiraclerise Pelroleunms Liqui(hs . J. of Clmromatography, 403, 159-170

Figure 6.12. Dislribtutioms of SCN groups in equihihratesh gas and condetmsatc phrases of a Mortls Sea reservoir fluid. ksanuple 6.5.
b)escrihe the C7+ frmuction of tIne flmuid inn F.xausuple 6.4 svithu 4 t~etndo connipuini&umts (using I Inc quu adrat tire noel hod.

(1987). 4. Pedersen. KS.. Ihnomnimnscn. P. amid Fredeumsinnmnd. A: limermo(lynamics of Petroleum

Mi xliii cx (onulaimi inig I Icmuvy II ymhmm ucam hssuis. I : Plmuise Fmmvelopc Cuulcn.mlationns hny tuse of tIme SomuveRedlichsKwouig Equumtioiu of SImile, Imnd. Emsg. Cimetn. Process Des. Dev., 23, 163170

Solutuo,i: As ysoh, we obtain, F(y)sol from Eq.(6.43) munsd, z,sow~from Eq.(6.66).

The molecular weighi of pseudo-componeunts are cmuhculated frommn Eq.6.614, as,
M(soX~x75+90

(1984). 5, Reid, Al.: Chimtracterismmhionr of Nortin Semi Reservoir Fluids, MPhii Thesis, i-leriot-Wmntt Umniversity, Edinnhtnghn, UK (1992). 6. Miquel, J., llerumaundcz, J. amid Castells, F: A New Method for Petroleum Fractions and Cnnde Oil Charactenismntion, SPE Res.Eng., 265-270 (1992).
zM~0 S 36.26 (I 7675 eM/S 89.75 94,08 18.22 0.40 202.45

with the results as given in the followiung table.

Pseudo-comp.

Room. ~ Weight.w=z t).3226 0.61(32

M 114.2(1

iM 68.88

7.
8.

Katz, [).L. ami(l Iiroozahuudi, A: Predictiung Phase Behaviour of Condensate/ Crude Oil

Syslcmnis Using Methiunuse luntermsctioun Coefficiemits, JPT, 1649-55 (Nov., 1978).

2
3 4 Total

1.7458
4 5366 9.3951

0.3574
0.0389 0.0005 I

220.94
430.25

78.96
t 6.74 0.40 164.98

38.02
7.36

(1.8393
0.91145 0.9981 0.8149

Rergnuiuums, t).F.. lek, MR. mind Kale.. DL: Retrograde Comndensation in Natural Gas Pipeliumes, Monnogrmuphn Series, AGA, NY (1975). 9. Whitson Cli: Chiaracterisimig Hydrocarbon Plus Fractions, SPE J., 683-694 (Aug., 1983). I 0. I lmummlaurd, 5: Chuumrmnclcnisatiots of Norlh Sea Crude Oils and Petroleum Fractions, Thesis, N II, I Jnmiversit y of Irommdhseitnn (May, 1981). II. ()sjord, ElI., Rouninimrgsen. lIP, mind ruin, L: Distrihution of Weight, Density anel Molecular Weight ins Crtmde Oil Derived fronn Coniputerised Capillary GC Analysis, J. High Res. Chronnatograpimy. 8, 6,S3~690 (1985). 12. Romsningseus, H.P., Skjevrak, I. amid Osjord. E: Characterisations ofNorth Sea Petroleum Fractions: Hydrocarbon Group Types, Deussity aurci Molecular Weight, Energy and Fuels, 3, 744-755 (1989).

794.63

0.16
81.81

None utiat the cmulctmtated mohecuular weight of the mumixture is mnlmost equmut to thmmt of the unmeasured vahune, and the coimtrmbulion of the tasl pseimdo commspouncnui to it is qnuine summall.
The specific gravity of each component characterisation facior defined in Eq.(6.70),
N

is determined

by initially calculating the

c so [o. 16637 x 0.81 5~zkM~SMSS 1(165)]

=9.80826772

248

6. F/ink! (lnaracier,,ca,bum

6.4. Refere,mces

249

13. Robinson, D.B. and Peng, D.Y: Time Charmuclernsation of the lleptmmsnes amid Heavier Fractions for the GPA Peng-Robinson Programs. GIA Research Report 28. luI.sa (1978). 14. Yarborough, L: Application of a Generuilised Eqummtion of SImile to Peliolcunr Reservoir Fluid, Equations of State in Engineering, Advmmnces inn Clmemsnistry Series. (hmuuo, K.C. amsd Robinson, R.L. (eds), American Chemical Soc., No. 182, 385-435 (1979), IS. Curvers, J. and van den Engel, P: Gums Chrommiatograpimic Metlmod for Sinmmn.ulumlcd Distillation up to a Boiling Point of 750C Using Temperature-Programmed Jnjectiots anmd High Temperature Fused Silica Wide-Bore Columns, J. High Resolution Clmronnatograpisy, 12, 1622 (1989). 16. McTaggart, N.G, Glaysher, P. and Harding, A.F: Correlation of Simntmhumted True Boiling Point Curves by Gas Liquid Chromatograpimy unrmd IS Plumte L)islillation l)mmta, Calculation of Properties of Petroleum Products, ASTM, STP, 577, 81-94 (1975). 17. Chorn, L,G: Simulated Distillation of Petroleum Crude Oil by Guts Cinrommnuntogr;mphy, Characterising the Heptanes-Plus Fraction, J. (Inro. Sci, 22, 17-20(1984). 18. Danesh, A. and Todd, AC: A Novel Sampling Method for Compositional Analysis of High Pressure Fluids, J. Fluid Phase Equilibria, 57, 161-171 (1990). 19. Ahmed, 1: Hydrocarbon Phase Behaviour, Gulf Publisining Compummmy, lloustotn (1989). 20. Kesler, M,G. and Lee, B.I: Improve Prediclions of Enllnumlpy of Fractions, 1-Jydro Proc., 153-158 (March, 1976). 21. Lee, 11.1. and Kesler, MG: Improve Vapour Pressure Prediction, Ilydro Proc., 163167 (July, 1980). 22. Cavett, RH: Physical Data for Distillation Calculations, Vapour-Liquid Equilibria, Proc. of 27th API Meeting, San Francisco, 35 1-366 (1962). 23. Edmister, W.C: Applied Hydrocarbon Ttmertsnodynaunics, Part 4, Compressibility Factors and Equations of State, Pet. Ref., 37, 173-179 (1958). 24. Riazi, M.R. and Daubert, T.E: Simplify Property Predictions, Hydrocarbon Proc.,
59(3), 115-116 (1980).
,

29. Loimrenz, J., Brmny, Ba. unnel Clark, C.B: Calculating Viscosities of Reservoir Fluids fromn tiseir Conuposmitioums. JPl, 1171-76 (Oct., 1964). 30. Alrnmed, T., Cmudy, G. mind Story, A: A Generalised Correlation for Characterising the Hydrocarbon liemuvy Fractions, SPE 14266, Proc. of 60th Ann. Conf. (Sept., 1985). 31. Pedersen, KS., Ellilie, AL. and Meisingset, K.K: PVT Calculations on Petroleunn Reservoir Fluids tlsinsg Meastured mnmrd Eslinrsaled Compositional Data of the Plus Fraction, lnnd. ling. Cisemsr. Proc. Res., 31, 1378-1384, (1992). 32. Behrens, R.A, and SandIer, SI: The Use of Sennicontinuous Description to Model tine C , Fractions in Eqsialion of State Calculations, SPE/DOE 14925, Proc. of 5th Symposium on 7 EOR, [tulsa (April., 1986). 33. Ahrummmnowitz, M. mnmmcl Slegtnn, l.A. (Eds.): handbook of Malhematicutl Functions. Dover Puhlieaiions Inc., New York, 256-257 (1972). 34, Whitson, CII., Anderson, F.F. anmd Soreiele, 1: Application of the Gamnma Distribution Model to Molecular Weigint ansti t3oihin~Point Data for Petroleunn Fractions, Chemn. Eng. Consmrm., 96, 259-278 (1990). 35. Cottcrmnman, R.L: Plrase Equilibrium for Systems Contaimring Vemy Many Components, Developmmrcnt umund Application of (ontimnuous Thermnmodynnamics, PhD Dissertation, Univ. of Calif. at Berkeley (1985). 36. Cotiermmman, RI.. mmmd Pu uuusmmitz, J.M: Flasin Calculations for Continuous (Sr Senmicontintmous Mixtures Usimmg mtus Eqtuation of Slate, I & EC Proc. Des. Dcv., 24, 434-443 (1985). 37. Whilson, CII.. Andersomm. T.F. and Soreide, I: C , Characterisatiomn of Related 7 Equilibrium Fluids Using Gammsmnna Distribution, in C , Characterisation, Ed: Chorn, L.G. 7 and Mansoori, GA., Taylor & Francis, 35-56 (1989).

,
I I

6.5

EXERCISES

25. Riazi, M.R. and Daubert, T.E: Characterisation Parameters for Petroieumn Fractions, I & EC Res., 26, 755-759 (1987). 26. Twu, C,H: An Internally Consistent Correlation for Preelictimrg lime Ciiticunl lroperties and Molecular Weights of Petroleum and Coal Tar Liquids, J. Fluid Phmuse Equilibria, 16, 137-ISO (1984), 27. Grahosku, MS. mind Datuhert, T.E: A Modmficd Somuve l:qtmmttuour of Simile I~nuPhmmuse Equilibriuns Calculations I Hydrocarbon Systemsns, Iimd. Emmg. (Incus. Process Des. Dcv., 17(4). 443-448 (1978). 28. Katz, D: An Overview of Phase Behaviour of Oil and Gas Productioms, JPT, 12051214 (June, 1983).
~

6.1. lime commmpositionn of us live samniple Imas becur dctenssimred by flumslning time fluid mit line lmtbormttory conditions, mmmd mirnunlysing the collected, gas and Ilttiid phases by gas cisroimnatograpimy. lIme commmposituon of gas, in mole has~smeasured by a TCD detector, and liquid, us ussass basis nneumsured by a FID detector, are given in the following table. Calculate the mixture composition in mole basis.cccc
Couniponemml

Guns Mote% ~~t~ 19

Liquid Weiglmt% 0.0(1

(I
(4 I
i5

6851
1401)

(1(12
II 115 (114 tt 55
tI5l

t 68 2 81 I I) 1.84

III)

CS C6 C7 (8

1.1(1

0.75
0.37 0.17

2.17
3.41 5.28

250 C9
CII)

6. Flmmid Cho,acteri.cario,u

6.5. Exefl,.ce.c
i-IItmianne

251
0.5 15

0.05
0.01 0.00 0.00 0.00 0.00 0.00

4.63
4.37

n-ttuuane

1.255
0.359

CII 4.20 Ct2 3.17 C13 4.46 Ct4 3.38 ClSi. 62.57 Liquid Phase Molecular Weiglnt= 2(16 6.2. One method of estimating properties of SCN groups, or evaluating the uneasmired values, is to relate the molecular weight and specific gravity by assumiming a cotmstant (UOP) characterisation factor. Is this a reasonable asstmmption? 6.3. Calculate the critical 6 2 temperatrure, presstmre, volume and acentric factor for a pseudocomponent with Tb= I K, S=0.866 and M=275 kg/kgumnol, usirng lIne mnrcllmochs of Cmmvett (E.dmister for lire acemnl.ric factor), and Riazi-Dauhcrt (Lee-Kesler for lime micemitric fmuctor). Compare the calculated results with those of C~reported in Tumble A.2. 6.4. The concentration of C7+ fraction of a gas condeimsate is 4.77 I mole % with tIre umnalysis as follows. Extend the analysis to C30+ by SCN groups. Composition and properties of C7 fraction. Component MoIe% Densily, kg/mum MoI. Weight Ilepimuumes 2 I .09 739.0 89 Ocianes 15.01 749.4 10.5 Nonunuies 9.96 764.1 121 Decauses 6.58 776.6 138 tinmdecamses 6.67 785.7 151 t~idecaneun 5.71 796.9 164 Tridecanmes 5.30 810.5 178 Tetr~iecanes 4.72 814.4 192 Pentadecanses 3.73 822.5 206 Ftexadecaunes 3.02 829.5 220 Hepiadecanes 2.64 832.2 234 Demadecanes 2.66 835.7 249 Nonadecanes 1.32 838.1 263 Elcosanes plus I 1.59 1)52.1 353 6.5. The mole fraction, molecular weigirt amid speciFic grunvuly of C7.f fuaction of a gums condensate are 0.04771, 165 and 0.8012, respectively. Describe time C7~ Irumctioum by a continuous function in terms of the molecular weiglst. and use it to estimnale time ninole fraction of SCN groups comprising time C7+fraction. 6.6. The consposition of a gas condensate sample is as follows.
Components Niurogen Carbon dioxide Meuhane Enhane Propane

i- Pennanes n~ Pentane Ilesanes tlcpt7nncs+

0.551 0.6 16
4.771

TIme C , description of the fluid is time saumme as lhmtt given ins Exercise 6.4. 7 Describe tire C , fraction by two and mnlso fottr psetmdo connponents, using the quadrature 7 time dew point and saturated density of time gas condensate at T=394 K using a nuretlmod. Predict ptmase bclnaviommr umodel and describing the C , fractious by a single, two and four groups 7 (Measured values in Tumble 2.2C).
(s.7. lIne commiposil ionn of unn oil satrrple is mis follows 13 I ~immponctum mote Iraci n,,mm (1.0069 MW S(

N2 C02
(I

(1(8)12
0.4709 1)0569 ()()439 (1.0095 (1(1242

(2
(3

i.C4
(4

-(5
CS (6

((.0111
0.0146 0.0226

C7
(8

0.0393
0.0452

91.9
105.2 121 (I

C9 CII) CII C 12
(13 C14 CIS

0.0323 0.0230 0.0203 0.1)188

0.0162 0.0176

0.735 0.745 0.784

1331) 48.0 1(7311

1771) 190.0 204.0

(1.789 0.794
0.806 (1.819 0.832 (1.834

C16
Cl 7 Cl8 (19
C20

0.0139 00103
(1(7122 0.0085 (1(1)197
11.18)32

217.0
235(1 248.0 20(1(1 269.4 282.5 297.7 31)1.1 321.8 332.4 351.1 370.8 381.6

0.844
((.841 1) 847 0.860 1)874 (1.870 (1.872 (1.875 ((.877 (1.81)1 0.886 0.888 (1.895 (1.1)98

C2 I C22 (23 C24 (25

0.0081) 0.0053 0.0044 1)0(134 0.0048

Mot% 0.298 1.72

79.1 39

C26 C27
C28

(1.0039
0.0031 0.0031) 0.0024

7.48 3 3.293

C29 C30

0.0294

393.7 612.1)

0.935

252

6, FIn,id C!naracierj.ca,ion

253

Describe the C , fraction by the gamma probability function. Use the quadrature method to 7 , fraction by two, and also four groups. Predict the oil bubble point at represent the C 7 a phase behaviour model and describing C ,, by a single, two and four T=345.8, using 23~7 groups. Compare the results (Measured value ~b= MPa). 7

GAS INJECTION
I Itnjecting gas iisto ~inoil reservoir to imncrease time oil recovery, has long been applied. It can iunnprove tIre recovery tinrougim mnaitmtaining the reservoir pressure, displacing oil, or vaporising time insternnediate and imeavy fractions of the oil. As the injected gas is not initially at equilibrium with the reservoir oil, the commtact between the phases results in mass transfer, hence, changes in properties of the two phases. The displaceunent of oil by gas becomes highly efficient when tire properties of the advancing gas and displaced oil become similar. That is, the two phases aclnieve complete immiscibility amid tire vapour-liquid interface vanishes. At the pore level, the immiscible displaeemmment is practically 100% efficient, as time lack of interface eliminates the

-c ~

retainment of the oil hr pores.

It is reasonable to assuumre (mull tine eqimilibriuns is reacired at the gas-liquid interface. Hence, pirase belsaviour concepts and nnodehlimmg methods previously described can be applied to investigate gas injection processes. The concept of miscibility in gas injection, and the experimental and theoretical metisods to evaluate and design miscible displacement, however, nmerit particular considerations. In this chapter, miscibility concepts are initially de~cribed for simple mixtures, and then applied to real reservoir fluids by presenting methods to stimate miscibility conditions.

254

7 Gas limjeetio,m

7.1. Miscihilin Coiucepec

255

two-phase regiouns, hut the diagramir shown in Figiure 7.1 resembles most of the practical cases, and is quite adeqtmale for describing tIne miscibility commeepts.

7,1

MISCIBILITY CONCEPTS

Miscibility concepts cats be expressed couivenmicumlly by exmtmmliliimmg I lie lilmuise lmclumivmour of mm ternary system, simulating the injection gas-reservoir oil mixture, using a triatsgtular diagram.

The phase behaviour of a three component mixture (L: Light, h: Internnedimrte. H: 1-leavy) at a
constant temperature and pressure is shown in Figure 7.1 by a ternary diagrammi. Each corner of the triangular diagram represents a component as pure, 100%, whilst all binary nnixtures are on the lines connecting the two corners, e.g., Point D. Any point witlsmn the diagratss represents a three component mixture, e.g., Point M and its composition is delernmsinred by its posution relative to the corners. When two fkuicts of different compositions mire mmxcd. lIme overall mixture lies on the line connecting the two fluids, whiclm is called the dmhmmlnorm or operaliing !i,me, and its position can be determined by the lever rule.

I ignmne 7. I shows I Immmt lime hu usuu uc S ummmnlc of I. mmmd I, amid simmiihumrly of I and II, forumi single plnase mmmixburcs ~vlmen mmmixed mit ammy proportion, whcremms, L mind ii forum a single phutse fluid whsems mixed only withninm a Iimssited ratio. lwo fluids are considered to be miscible, wlmen timey form a single phase at all proportiotrs, at constant pressure and temperature. It is evident tlmat any two fluids with thse operating line not crossing tIme two phase region within the phase ensvelope are miscible. Figmnre 7.2 slnows (mat mm ins jection gas conrprised ommly of! is nniscihle when conlmmcted with Oil B, wlmereums Gas A is mint. It cami hecomime, hiowcver~miniscible eilimer by enrichmmiemmt wills I to A or by rumising lIme systenin pressure to shnrink time phmase envelope mis shown by the clashed phase etsvelopc. Wimeus lire ins~eclmon guns ammd reservoir oil, miiixcd at any ratio, forns a single phase. they mmmc cmuhled firs, eo?mIa(l nn,,scil,Ie. First conlmucl umuiscihility camm be achieved only for higisly rich gases, or at very highs pressumrcs for lean systenms.

100%

10% I) 711% i

Gas A

0 11100% 80 60 40 20 0

1. 100%

11 100 % Figure 7.1. Ternary diagram presentation of fluid plmase equilibria at constant pressure and temperature. The phase envelope is slnown by the curve ACB in Figure 7.1. AIry nsmxlmnre, F. insude the two phase eumvelope formnis a vapour phase, Y, and mu hiqumid plimise, X. mml equnmtubrmuunmm, lying on time phase boundary curve. Time line XY, connecting line two plnases at cquilmhriummi is kmnowms as the lie line. llemsce, time left hand side of the curve, AC, represemils smntsmraled liquids, i.e. the bubble point curve, whereas tine right hand side. CB. represetrts smslumrmnled gmnses, i.e. lIme chew poinrt curve. The two pmnrts of tire curve converge mit mIsc crilical poimsl C, umlso kmmowmm mis tine plait point. Any nnixturc outside the pinase envelope is mm siusgle phmuse umnder-saluruikd flmnid. Plnase envelopes, and the associated tie lines, at diffcremmt pressures can be shown (mn time sanre diagram, where increasing the pressure generally results in thm~ shnrinkage of the phrase envelope. A ternary system may form more than two phases, or Irave a number of isolated Figtmre 7.2. First coustact miscibility.

L 100%

Ann imijeclions gas whiiclr is miol immiscible wills mun oil mit first couslact, may achieve miscibility during ummimlhiple couilmnets by gcltimmg etsriched tlmrough vaporising lIme internnediate fractions of oml. The process, kmsown mts tIne sapori.silmg ga.s drive (VGD) is conceptually shown in Figure

7.3.
time immjeclioun gas. I.., commiprised mit the light Ilimich onuly, after contacting Oil A, fornnms two equilibrated phases of hiqumid X mind gas ~l~ witlnanm overall nmixture F~. Note that the gas pirase, Y , is the original gas L after it inas picked up some intenniediate and heavy fractions 1 from tIne oil phase. Tire gas phase, Y , moves forward and makes further contacts witls the 1

256

7.

Ga.s Imujecthnu

7. 1. Miscibiliis Commeepl.s

257

fresh oil and progressively becomes richer particularly in tine internnediates. as simowmn by Y , 2 Y .... The gas ultimately becomes miscible with oil at C, that is, where the tangent line at the 3 critical point, which is the critical tie line with zero length, gcnes through the oil consposition. It is quite evident, that the compositional path must go through the critical point, as it is the only condition that equilibrated phases lose distinctions, and a continuous transition fromnm gas to oil can be achieved without any phase boundary.
The above injection gas, pure L, however, does not acirieve mumltiple contmtct nmmiscihility with Oil B, as the enrichment of the advancing gas is limited by the tie lime X Y thimrmiting tie line) 2 cammnot 2 which, if extended, goes through Oil B. It is evident that the nmiscibility he acisievecl when the oil composition and the phase envelope are at tire same side of tine critical poimit tangent line (critical lie line extension). The vmiporising gas drive immiscibility iimr oil B cmnn he achieved, however, by raising the pressure sufficiently to shrink tine phase envelope, as shown by the dotted boundary. The pressure at which the critical tie line extension goes through the

conceplm.ualiy in Figure 7.4. Tine process. which is called the.condensing gas drive (COD), is
described below.

00%

GasA

oil is the minimum required pressure to achieve miscibility. lncnce, called the

mini,uu,n Gas II

miscibility pressure (MMP). At MMP, the limiting tie line beconmes the critical lie hue as tIme gas phase enriches through multiple contacts with the original oil attaimuimsg the critical composition.
I iOO%

00%

H it)l)%

Figure 7.4. Conulcumsiusg gmms drive scimetnatic pimase dimmgramn at nnimmirnsum miscibility pressure. The ricim gas A forms two plnases. gas, Y , and oil. X , in equilibrium after contacting the 1 1 reservoir oil. rise gums pinase unroves forward and leaves the enricined oil X beisind to be 1 cotstumcted further witln tIme fresh gas A. resulting inn an oil even riclser in the imstenmmediates as shown by X . X ... . This process goes on and the oil is enriched to the extent tlnat it finally 2 3 aquires the conmpositiomm of time critical oil at C. At thsis point it will be miscible with gas A. The pressure mit which the criticuil tie lime extension goes through tire gas conmposition, is the minimum required pressure to achnieve miscibility (MMP).
I. mIRm% H 100%

Figure 7.3. Schematic phase diagram of vaporising gas drive at minimum miscibility pressure.
In the vaporising gas drive, the nmiscibility is aclsieved at the fromnt of tine advuminciung gas. The gas composition varies gradually from that of the injected gas till rcaciminng tire critical composition. Then it miscibility displaces the original reservoir oil in a pislon-type manner.

At MMI, tIne oil plsase etiricired in time intermediate through multiple contacts witim the injection gas uittmmins time criticuml eommmposition, with the limiting tie litre become the critical tie iimse. lime inn jectioms Guns B, wInds is lemitier ins tlmc intermediates tlman Gas A, does not form immiscibility, at tine ciurrent piessuure, mis tIme enriehmnent of the oil is limited to the composition of the tie line extenditsg tisrouglm lire injection guns composition (limiting lie line). The miscibility, however, can be achnieved by raisimsg the pressure to shrink the phase envelope as shown by the dotted curve. Time origimnal oil coumnpssitiomm Inas urn effect on achieving tine ntniscihility slate in tIre condcnsiumg gums drive, as it is comsluolled by tIre insjedtion gas cumrmposition. Ilence, instca(l of raising the pressure tsr achmievc mmmiscihility, mIne injection gas mummy he enriched. Tire emmruclminment level mnt mninimmrutnm miscibility enrichment (MME). In tine condensing gas drive, tine uniscibility is achieved at the injection point. The injection gas displaces the criticud fluid, in a piston type manner, with tine liquid composition varying gradually to that of the original oil. No phase boundary exists witlsin the transition zone.

winieR tIme critiemmh tie line extension goes through the injection gas composition is called the

No phase boundary exists within the transition zone. The gas composition appears to have no effect on achieving the miscibility state in the vaporising gas drive as it is fully controlled by the oil phase, as demonstrated in Figure 7.3. A rich gas, not forming first contact miscibility with an oil, can, however, achieve multiple contact miscibility through condensing its intermediate fractions to the oil as shown

258 Miscibility in Real Reservoir Fluids

7. (a.s l,,jecsio,u

7. 1. Mi.oibmlmiy Concept.c

259

at time tail. Tisis proces~.cunllesh lIne eommden,si,ng/vapori.cing gas drive, was reported by Zick El)

in 1986, and detailed by Stalkimp in 1987

50 --5 40
.

E21.

The ternary plnase diagrammn of a multicomponent reservoir humid is ofteun expressed by representimsg the fluid with three pseudo components. It is commnnon to group C , mnuiul N as Ihe 1 2 light (L), C02, 11 S amrd C -C as the instermediate (I), and C mis the ineunvy (II) fraction. 2 2 6 7~ Pseudo-ternary diagrams imave been misrused, however, mn descrihinig line hcliavimmr of real reservoir fluids in gas injection processes, partucularly in estimating the ophinisuusr operating conditions such as MMP and MME. The conceptual discussion on multiple contact miscible processes, umsing line lermnamy diagram, is not strictly valid for real reservoir fluids, umund tine dimngramin shotmid mint he smscd genermully ins Ihse design of real processes. Tine hmisic idea of multiple contact miscibility throingim minass exelsange between the phases, and tire requirement of attaining the critiemsl corsipositiots. mmmc mill valid for real systems. However, the existence of ml large miummrmher of comnnpomncmsls inn ml real reservoir hluid provides middimiomnmml possibilities for comnmpositionmmil vmurialimmns, mind mmsliucviuig nnmissiinility. Inn tIne following discunssiomn, time unniscible cotrditmoun us icferrcd In time. conidutmomu wlucis mumuscihilily cumun just be unchieved, thuit is, at MMP. At liighrer prcssuurcs, tine niniscmhnlmty will ohviosnsly he achievable. The nruultiple conlact miscibility, can be aclsieved ounly wlren the comrmposilmotnah pmuths goes through the critical state. As tine critical compositiots, hence tine critical tie lime, for a ternary system at a given set of temperature and pressure is uunique, tine miscibility is dclermuned only by the two limiting lie lines. When the critical tie line coiimcides wills time linmiliung lie himme going lhroughm the originuml oil counposilion. the nniscihulity us achsieved by line vapmansiung process. When the critical tic line coineides witir tIne limnnihing tie line gommig lisrotigin 11mm injedhioms gmus composition, the miscibility is aelnieved by the condcnsmmig process. In a real system, it is possible tlmat neither of the limiting tie lines goes timrough tine critical point, hut miscibility still is achieved. This will occur if the fluid attains a critical state not at the leading or the hailing edges, hut sonmewhiere within the transutmomn zone. Suds a possibility exists for mixtures with more than three components. Indeed lire prevailmng nnnechsanism for achieving miscibility in rich gas injection is often, if not always, the above memmtuoned case and not the condensing mechanism described for a ternary system. Tine injected rich gas sines not generally contain heavy fractiomrs whicis arc present in tIme oil. I lence, winihst tine injections gas enriches the oil iun light internnncdimmte ramrgc. it strips tine ineavier fractions. The reservoir oil in contact with tIre fresin gums iuiimialiy bccouines lighter, but as it contacts more gas and loses only some of its lighter heavies, overall it tends to get enricimed in very incavy fractions and tirus becomes less similar to the unjection guns. Figure 7.5 shows the varialion of measured compomment groups in the oil phase at tire inn jedhion point for mu North Sea oil. As the oil is contacted witln additional rids gas, the cotucentralion 1)1 (.,, is decreased, apparently lightening the oil in its path towards achieving tIme condensing miiiscibility. An exannination of the heavy end, e.g. C , , however, shows llsmnt Ilmis fractiomn Iras increased 25 heavies. This oil eamnmsot becoumne miscible with the markedly due to vaporisation of the lighter frcshn injection gas. The phase envelope, as determinred by mneasurimrg tine compositions of equilibrated phrases at the injection point and also at tIme gas front in a laboratory test, is shown in Figure 7.6. Note tinat the bubble point and the dew point curves initially converge, demnonstrated by sisortening tie line lengtirs, aumd then diverge. As the forward moving gas becomes richer in heavy fractiorns, it vaporises less of these compounds whilst losing intermediates to the oil. It is conceivable that at favourable conditions the combined vaporisationlcondensation process results in a state withiur tine tramnsition zone where the compositional path goes through the critical point, acisieving miscibility. This can be envisaged as a combination of the condensing process at the front, and the vaporising process
0

Lighu

Flea~y
30
---- --

0 nnIc,,nctniamc

20

I0

(71),

10

Vol. Gas/ Vol. Oil Figure 7.5. Varimstioni of component grouups in comntacted oil at injection point wills tine ratio of

injected guts volumme to contacted oil volcunne.

50 60 70 inj gas
20

80 90
L

50

40

30

20

10

Figure 7.6. Phrase dimmgrmtnn of a Nortim Sea oil and rids injection gas determined experimentally

during comntact experimnnent. line mnnulhiplc contmndt unmiscihility inn nssumlticonssponenl systetmss is mncimieved in a dynamic process, tmence. it can be affected by olhcr factors additional to the fluid phase beimaviour. The coumnpositional paIls depends our other mechanisms such as multiphase fluid convection and dispersion inn porous unedia. Tine two limiting critical tie lines, that is, those with extensions goinrg timrough thre original oil and the injection gas, depend only on the original fluids. I-hence, time nniscihilily conditions can be deternmined by phase behaviour considerations only, if the nimiscihihity is achieved by either vaporisimng or condensing gas drive mechanisms, Otherwise, the critical lie line, hscnce mniscihilily, depends on tine local fluid mixture composition influenced

260

7.

Gas !?ujec:ion

7.2. E.uperinuentol Studies

261

by flow factors. Therefore, a proper investigation of the multiple contact mrmiscibihty should involve the simulation of fluid phase and flow behaviour as closely as possible to that in the reservoir. There are ample publications on mathematical simulation of multiple contact immiscibility as developed in a one dimensional flow, Tire reports by Orr and co-workers 13-71. tmtilisimng mm analytical method to solve the governing phase and flow behaviour eqsmations with no dispersion, for a four component system provide valuable insigint into tine rmrechanisnms of multiple contact miscibility. The authors demonstrate that tine displacememsh in a multicomponent system, can be expressed by a series of pseudo-ternary diagrams, where the compositional path progressively moves from each diagram to the next. Wisems tine two phases of vapour and liquid at equilibrium are present, the transfer from one tcrnmury diumgrummnm to time next must occur at a common tie line between tine two diagrams, called the cross over tie line. The number of the pseudo-ternary diagrams increases with the number of crmnponents, witln the number of cross over tie lines equal to the number of consponents mmniumsms 1111cc. For example, whilst the compositional path in a four component systemsm camn be depicted by two ternary diagrams, with a conmmon cross over tie line, there arc two cross over tic lines connecting three pseudo-ternary diagrams in a five component systems. Depending on the fluid composition, pressure and temperature, any of the cross over tie lines or tine limiting tie
lines can become the critical tie line. The miscibility is then achieved at that point.

The tube is initially saturated with the oil at reservoir temperature above the bubble poimst pressure. The oil is then displaced by injecting gas into the tube at a constant inlet, or more often outlet, pressure controlled by a hack pressure regulator. The pressure drop across the slim tube is generally small, tlrerefore, the entire displacement process is considered to be at a single comnstant pressure. The shun tube effluent is flashed at the atmospheric conditions, and time rmnte of recovery, density and commlposilion of produced fluids are measured. The gas break through is detected by comitinuously monitoring the effluent gas composition, and/or the producing gas to oil ratio. The miscibility conditions are determined by conducting the displacement at various pressures. (Sr injection gas enriclnuncurl levcls. and monitoring line oil recovery. This can also be aided by visumul observatiomi of tIne how through a sight glass placed at the tube outlet, The aclsievemmnenl of miscibility is expected to accomimpany a gradual clsange of colour of (Ire flowing fluid fronts that of the oil to clear gas. Whereas, observing two phase flow is indicative of an immiscible displmnceuncnt.

Considering the above, the estimation of miscibility conditions based on the liunmiling tic lines is not generally adequate for real systems. When tire process is known definitely to be
vaporising, such as methane displacing an oil composed only of hydrocarbons. Ihe oil tie line can be used to detennine the miscibility condition. The estimation of miscibility in rids guts

drive, using the injection gas tie line, is expected to be unreliable in most cases, if Isot in all.
When the key tie line which controls the miscibility is one of time cross over tie lines, the estimation of miscibility conditions is not as straight forward, Even if the key tie line cams be determined for a dispersion free, one dimensional flow, as suggested by Johns et ai. [6], it

Pump

may not represent the real process. It has been demonstrated that dispersion, viscous fingering
and gravity segregation can impair the achievement of miscibility, causing two pimase flow in reservoirs in which they would be miscible otherwise [8,9]. 7.2 EXPERIMENTAL STUDIES Back Press. Regulator Guts Meter

Gas injection experiments are conducted with several objectives. Most tests have been designed to directly measure the minimum miscibility pressure or enrichment. Tests are also conducted to generate volumetric and compositional data for specific studies, such as oil vaporisation by gas injection, or evaluation and tuning of phase behaviour models for

Figure 7.7.

Schenmmmtie diagramst of slim tube apparatus.

numerical simulation of the reservoir performance. Displacement of oil by gas through a porous mediumsi, simulates time gas injection process more closely than other tests, and it is considered as the definitive test. The displacement is
conducted either in a core, extracted from the reservoir, or more often in a long and narrow sand pack, known as the slim tube, Static tests, where the injection gas and time reservoir oil are equilibrated in a cell, are also conducted to determine tlne mixture pinumse behaviour. Although these tests do not closely simulate the dynamic reservoir conditions, tisey do provide accurate and well controlled data which are quite valuuihle, partieulummiy for ttmnimrg tine equumtion Slum tubes of differcmnt sizes mumnd orientatiomms have been used at various displacement conditioums. Tine effect of tube geometry and flow paranneters on fluid recovery has been systensaticahly investigated [10]. An ideal tube should provide a one dimensional dispersion free displuicemeirt of oil by gas. This is not to suggest timat such a displacement sinnulates tine process in a real reservoir. It is nmrerely a well controlled experiment, with valuable results for phase heimaviour studies imncludimrg miscibility evaluation. The actumal displacement in a reservoir is influemrced by variouns nmechaunisrns, such as viscous fingering, gravity over ride ausd dispersion. Reservoir hetcrogeuseity, at different scales, strongly divert time flow and affect the recovery. As it is imnnpo.ssible to sinnulate all these inter-related tmrechanisnms in a slim tube, or even us a core, it is logical to avoid them all in plnase behaviour tests. lhese factors, all immrportanmt to the success of a gas injection schenme, could then be studied in otlmer experiments

of state used in modelling the process.

Slim Tube Slim tube, a one dimensional nnodel reservoir, is mm nmnrrow tube packed witin smnnd, or glass beads, with a length between 5 and 40 m. A schematic diagram of tine tube aurd tine auxiliary equipment to allow displacement tests, is shown in Figure 7.7.

and umsing compositional simulators.

Probably tIne simmmplcst test usinig a slim tube, is determination of time nsiscihihity comsditiotss. Altsmost ammy shins tube, regardless of the design, can be used In estimate the effect of pressure

262

7.

Ga.c Iiujeeimoiu

7.2. Kmpe,inue,mtaI Stm,dies

263

and gas enrichment on miscibility. Operating conditions, however, should be selected based on the tube design, and the prevailing mechanistsi which results in miscibility.
A tube length of 12 metres is adequate in most cases, and allows miscibility mireasurement within a reasonable test period and fluid consumption. The length is to provide sufficient contact between the phases to achieve miscibility, if ustluninable, Thneorelicunhly. nniscihility is achieved almost at the tube entrance whets iso dispersion exists [4,111. Its practice, tubes as

80
000 0 0 0 0 0 0 0.

~0
0

(ill
~ S 4(1 I S
0

2S2SMPa

long as 36 meters have been recommended for displacement with nitrogens 112.13).

The

improvement in oil recovery, by increasing pressure or gas enrichnunsent, is ofteni nmnonsitored to determine the miscibility conditions. ilence a long ttmhc, whiclu Inigimlighnts cinmnnmges inn line oil recovery, provides a more definitive means for tine jlmdgemimenl (nf msmmscihrilily conimliliomns. The tube diameter should be selected small enotugh to suppress viscous fingerimig by traumsverse dispersion along the tube, that is, the shorter the tube, tire snmaller time diaurseter. A 6.3 tnmm (0.25 in.) diameter tube is often used, and is considered adequate. inn ltmhes of mml lemmsl 10 mrs

~.1mSMu~

E
0

>
0 50.

20 0

5~ 0 U 3

long, for justifying time assumption of onc-diunrensional flow iii lIne shiunm ttmhc. To usmimniunnise wail effects, the mnazimum grain diameter of time pmickung material shrould he less tinmiti I/It) of the internal tube diameler [14). Sand, or glass heads, of 100-200 nneshn size mire often used in packing the tube.
As a long tube cannot be placed in a straiglnt line, it is in be coiled preferably cyliundrical than flat, with gas injected from the top to promote gravity stable displacement. A low gas injection rate can improve the contact between the phases in a short shiumn tube. it however, improves the recovery efficiency, particularly at imumsiscible commditions, imence reduces the recovery gain when achieving miscibility. This cats reduce line accuracy in detenisining the miscibility conditions as measured by chammges in the oil recovery. A gas injection rate with an advancensent velocity of ahotut 1.5-2.5 muir, getserally shtoumld provide

50

iOO

n50

Pore Vol. lnjecmed. %

Figure 7.8. Oil recovery by gums iunjedtion at two different pressures.

I )J)~5flfl~

reliable recovery data within a reasonable experimental tumne. The displacement is often terminated after injecting 1.2 pore volume (PV) gas. Time recovery at that point is referred to as the ultimate recovery. Tire test nmay also be terminated at a pre-selected high producing gas to oil ratio, around 8000 vol/vol (40,000-50,000 SCF/STB),
with the ultimate recovery determined at those conditions.

;2.
0

24.15 Mla 28.25 MPa

0 0

>
0

10000
S

F
0 ~O 0O 0 0

The volume of produced stabilised oil in the separator ms generally converted to that at reservoir conditions, using the volume ratio measured on the original oil. Al high pressure conditions, particularly in rich gas injection, a significant ainosunt of lIme liquid collected in time separator can be due to condensate drop-out frousm the produced gas. The volunsre fumctor for suds a liquid is different from that of the original oil. Furthermore, the liquid recovery at suds conmditions is
not totally by displaceunent. Figure 7.8 shows typical oil recovery plots at two pressuires. At 24. 15 MPmm line gas breaks through at 38% pore volume (PV) gas injected, with an ultitnate oil recovery of about 50% PV. The volume of recovered oil prior to the gas break through is almost equal to time injected gas volume. It is generally slightly less than tire injected volume due to the shrinkage of the total gas-oil volume in contact. The oil recovery after break through drops sharply and can he quite small, particularly in efficient displacements. Tine gas to oil ratio (CUR) increases sharply after the breakthrough, as shown in Figure 7.9. The low recovery at 24.15 MPa is indicmntive of aim immmniniscihle displacemnmerrt. At tire Inigher pressure of 28.25 MPa the gas break through occurs qumile late. Note thmnt thie recovery, both the break through and the ultimate, lsas improved nrarkedhy at tine iniginer pressure, indicating an approach towards miscibility. Although at miscibility conditionns tIme displmiccunent is quite efficient, a 100% oil recovery is not generally expected, unless the condensate recovery in rich gas injection is included, where the complete liqmnmd recovery over I PV can be achieved.

0 001) 0
0

20

40 60 Pore Vol. Recovered, %

80

Figure 7.9. Producing gas to oil ratio in gas displacement at two different pressures. Differeust recovery levels, such as 80% at the gas break through [15], or 90%-95% ultimate recovery [16,171, lnave been suggested as the criteria for miscible displacement. The oil recovery, however, depemids our tIne lube design and operating conditions. A slim tube may ounly deliver 80% oil necovery unt msii.scihle conditions. The evaluation of recovery changes with dnslmlmnceulmennt prcssunue. or gums cnmriclnmmncnt, is nimore appropriate to determine miscibility condihioums tlmunnm seumrcimimng for a lmighn recovery. TIme rumost acceptable definition is tire pressure, or enrichmntnemsh level, mnt tine bremnk over poiust of the uuhlinnate oil recovery as sinown in Figure 7.10. Tlme recovemy is expected to imrcrease by increasing the displacement pressure, hut the additional recovemy above MMI is gemmerally nsininmal.

264
108

7. Gas Injeciio,m

7.2. Fnperinuenuta! Stnudues

265

0)

90 80

I) > 0 U C) C) C

10 60

Wireur ussiscnhuhity is umchnueved by time vaporising gums dnive, flow factors, should not. al least theoretically, affect tmniscihihity conditions as explained in Section 7. 1. Whems msmiscihihity is achieved within lIre transition zone, the critical stale is expected to depend on flow and dispersioun factors. Ilme use of reservoircores, is considered to include sonse of tinese factors. An arramngemenl comnsposed of mm slim tube ahead of a core, where tIme slim tube provides enough colmtuncting lenglin between tine pimases, to introduce miscible fluids into the core, is often counsidered. TIne advmmntuuges of suds refinensenmls, however, are open to questions.
too 1
40 45 50
0 0

50 0 40 20 25 30 35

MMP

00
0

1)isplaccnuncnn Pressure. MPmu

Os

0 40

S 0 0

Figure 7.10. Determination of MMP by plotliimg ultimate recovery versus pressure. The dispersion and compositional vamimttiomns, particularly un micin gums iumjectiomm, create concentration, and saturation banks along the tube, wInch can conveniently be detecled by fluctuating producing GOR, as shown in Figure 7.11. The banks rich in immterrnediate and heavy fractions contribute more to the oil recovery. The relative location of tire banks are determined by the composition of the injected and reservoir fluid.s mind di.splmncemmsent conditions. line recovery of a rich bank may be delayed by enriclmimmg tine immjectiomn gas, isemncc, it mmnay be mistaken as a reduction in oil recovery [181. If Ihe displacememst us tentnimnunled before major liquid producing banks appear, the change in liquid recovery ninuny not properly reflect the impact of the varied pressure or gas enrichment.
10000

Ntrogtn

S S

20
0 09

5 S

~7
25

),

oo
60

00

00 120

40

Pore Volunuic lunjecmcd. %

Figrure 7.12. Conceustruntion of ummetimane ansd nitrogen in the produced gas. Tine sums tube cmtul provide very useful pimase heinaviour infortnnation, additional to MMP or MME, for evaluating pinumse belsaviour models when applied in gas injection processes. This

mnpplieatiomn will be described in Section 9.3, where tuning of equations of state will be addressed. The MMP detemuined by shun tube displacenrent does not necessarily correspond to the thermodynamic miscibility, that is, the achievemeunt of the critical state. Adverse effects such, as dispersion can prevent or delay thermodynamic msniscibilmty [9, 20], but the prevailing low interfuicial tension betweemr the two plnases still provides a lsighnly efficient displacement.
Rising Bubble Apparatus

0
Cl

1000
S

S S S

0
0

100 0 20 40 60 80

Pore Vol. Injecled,%

Figure 7.11. Variations of producing gas to oil ratio with injected rich gas volunse. The compositional variation of produced gas in an inmmssiscible dispimtceumscmsl of an oil by nitrogen is shown in Figure 7.12. At the gas break througis, the msmethamme comncentration exceeds that in the original oil. This is due to the high volatility of methane which results in its vaporisation into the advancing nitrogen. An absence of methane bank [19], and a smooth profile of concentration of intermediates [hO], have also been suggested as measures of miscible displacement. A combination of the above criteria, including sight glass observation, often is used to identify miscibility conditions.

Tire observation of a gas bubble hchavk~ur, rising in a visual high pressure cell filled with the reservoir oil, lnmns heeum suggested 1211 as a quick mimethod of measuring MMP. The apparatus is simowmm in Figure 7. 13, svlncre mm smimahl gas bubble is introduced into time bottom of oil column throtugis a water pinumse, acting as a buffer between gas and oil containers. The gas bubble coistimmuumlly comstacts tire oil tlsrougis its upward journey which results either in reaching equilihriummmm wills tine original oil, or achieving nsiscihility depending on the lest pressure. A series of tests are conducted am dilTerenmt pressures, unnd tine bubble simape is nnotnitomed mis it rises up. At pressures, far below MMP, the bubble retains its almost spherical shape, but its size is reduced as the gas is partially dissolved in the oil. At or slightly above MMP, the bubble develops wavy tail with tine gas-oil interface vanishing from the bottom of the bubble. At pressures higher tinan MMP, the bubble disperses very rapidly and disappears immto tine oil. A

266

7.

Gas InjeCtion

7.2. Experi~neiutal Studie,s

267

gas htmhhle not achieving nmiscibility, will also disappear into ami unmdcrsunturated oil. html will uiot disperse.

tire equilibrated plrumses below tine saturation pressure provide information on fluids with various degrees of proximity to time critical point, albeit at different pressures.

Dew Point A Oil

cj
EJ

C)

(~~

20%
~

0 If)
C

40%

Liquid volummnie Fraction

80%
60% 50%

Needle Figure 7.13, Schematic diagraums of rising buhbl~apparatus, slmowmng humid pirumses at start of experiment [211. 0 Metlmane Added to Oil >

Figure 7.14. Variations of nmixture saturation pressure with added methane to a light oil. The method is suitable only for the vaporising gas drive process, where time enricimumnent of advancing gas creates tine miscible fluid. The measured MMP by the above mnnetbmod has been shown to agree reasonably willn tlmat by displacement using a slimin tumhe 1211. Contact Experiments Slim tube tests do not generate all the volumetric and compositional data required for evaluation and calibration of phase behaviour models. Therefore, gas immjeetion processes are often simulated in batch type tests, known as multiple contact experiunents in PVT cells. In this method, finite volumes of reservoir oil and injection gas are repeatedly contacted, and shrinkage or swelling of the oil, and tine density and composition of equilibrated oil and gas are measured. Batch-type gas injection experiments are designed to generate phase heinaviour data, particularly for calibration (tuning) of equation of stale models used in sitnulalion. The most common experiment is the swellioug test, or single contact gas injection. A knnown aunrount of oil is loaded into atm equnilibriuns cell and the injection gas is progressively added to the oil stepwise. After each addition of the gas, time mnixture saturation pressure mmmd volume are measured. A constant composition expansion test can then be conducted on lIne mmsixtuire, prior to the next gas addition, to generate additional infornmation. Figure 7.14 shows, the variation of mixture saturation pressure with added methane for a light oil. The bubble point pressure increases with addition of gas. The incremental increase of methatie content of tire fluid results in a critical state, point C, wisere time nnixtume heisaves gas like with furtimer injection of methane. Tire dew poimmt initially increases, and lInen decreases. witim increasing metlnane. Tine lures showing constumnt liquid volumumie frmsctiouns witlsinm tire two phase envelope converge at time critical poinsl. Allhoingh the single contact experiment, does not simulate lire commtinuous contuict between the phases, as occurs in gas-oil displacement, it provides valuable durta for tuning of EOS. It covers a wide range of fluid composition, witim little experimmmetstal effort. Although the composition of the mixture does not follow the compositional path in a gas injection process,
Ilme static cquililrrituumr tests which closely sinruulate contimmuous contact of injection gas and reservoir oil are nnnuuhtiplc commtuncl expcrinmncnts. The forward multiple contact lest, simuiates

conditiouns tnt time its jection frommt, inn wlmicim oil and gas are contacted at the reservoir pressure and temmnpermuture. Time equmilibrated gums in each contact is used in the next contact with the original reservoir oil, simulating gas advancement in reservoir (Figure 7.15). The volume, density and commnpositioms of tIne equtilibrated plnases are measured in each conlact. The above procedrmre is conliuttued until Ilne injections guns citimer beconres immiscible with the original oil, or attains equm i I ihriummn witim it (Ii mnitiusg tie Imuse).
Liujecti~nu ( ;.ts

4
Reservoir Oil

Figuire 7.15. Flow diagramms mm usnunltiple forward contact experinment. Fumble 7. I shnows lIne constpositiomm of equilibrated guns mmmd oil in a forward mumltiple contact test of mu volatile oil unnd umnetisumnuc tnt 373 K and 35.26 MPa, where the gas has become progressively richer by contunctimig oil. Time counsposition of equilibrated oil in the last contact is unimnuosl lime sumumie as thrat of tine originimul oil. indicating an approach to the limiting tie line with no nmsiscihility adlmicvedl.

268

7. Gas lnjectiouu

7.2. F.uperiennetntal Sluu!ie,u

10

269

Figure 7.16 shows the variation of equilibrium ratios of the mixture counpomments due to compositional changes in different contacts. Each component has been identified by a plsysical property group as described in Section 3.2. A straight line can be drmmwmn through the equilibrium ratios in eachcontact. Note that the slope of tine line decreases progressively with each contact. For achieving miscibility, the line should isave become horizontal with K-values equal to I for all the components. Figure 7.16 is also a revealing example of lime effect of fluid composition on the equilibrium ratios, as the temperature and pressure are constunnl. The backward multiple contact test, simulates the injection zone tmiil, timat is. what lunkes place at the injection point. It is similar to forward contact test, but the equilibrated oii un each contact is used in the next contact with the fresh injection gas (Figure 7.17). Table 7.1. Composition (mole%) Contact No. 0 Component Orig. Oil C 57.53 1 C2 10.16 C3 5.83

a
C .0 0 1)2

__.,0o~, 0000

-0_0

- 000 ,..

C,,ntacn Stage

em
02 03

.4

i-C

nC4 i.C5 nC5 C6

C~

1.22

of equilibrated oil and gas in a four stage forward nmumltiple contact test. 2 3 4 Oil (las Oil Gas Out ~ Oil (;us 57.87 78.24 57.1)3 74.46 57 II) 72.25 56.73 71.41 7.87 7.57 9.00 8.91 9.71 9.72 10.01 10.1)1) 4.89 4.04 5.40 4.69 5.69 5.06 5.81) 5.22 1.06 0.79 1.15 0.92 1.19 0.99 121 1.1)2
1.85 0.95 1.62 1.28 0.59 0.97 1.98 0.99 1.68

01 4

. 4

-2

I I ou)( I ITt)

Figure 7.16. test,

Variuutions of eqnnilihriuunm rumtio of fluid counponenls in mntmlliple forward contact

~l

2.06
1.01 1.70

.49

2.03

1.61)

Cg C9 C Cj 0 C12 C 4 1 Cl6 Cu Cl 8 Ci9 C20+

1.40 2.16 2.55 2.00 1.55 1.10

1.00

1.41 2.26 2.76 2.23 1.69 1.36

1.14

0.75 1.04 1.12 0.82 0.58 0.36

0.33

1.41 2.32 2.65 2.19 1.66 1.30

1.07

0.69 1.12 0.85

1.23

1.01 1.71) 1.41

2.27

0.74 1.20 0.92

1.33

2.06 1.02 .72

1.65 0.76 1.24

munjccniot (,as j

1.42
2.27

0.95
1.36

~Oas~

~j

~Gas

1.26 0.98 0.69 0.41 0.41

0.36

2.55 2. II 1.52 1.16 1.09

I .05 (1.8! 11.92

1.36 I .1)5 0.77 0.51 0.39

0.47 0.22 t).3 I

2.58 2. II 1.47 1.18 1.08

I .1)3 0.82 (1.91)

1.42 I .08 0.74 (1.54 (1.41

(1.41) (1.31 0.32

r
I Reservoir (SI

Oil

Oil

th1

0.99 0.78 0.85 0.72

0.49

1.19 0.96 1.05 0.86

0.66

0.29 0.23 0.23 0.18

0.11 0.13 010

1.10 0.87
0.97

Figure 7.17. Flow diutgram in mnnultiple backward comnlact experinreni.

In

0.27
(1.28

0.74
0.63 0.67 0.61

0.21 0.15
0. 16 0.13 (1.33

0.76 (1.54
0.65 t).56 4.17

0.24 0.16
0. 18 0.15 0.38

0.75
0.54 t).64
1)55

(1.25 (116
0. 19 0.15 (1.39

0.60
0.5!

0.76
0.65

3.81

4.92

(1,26

4.57

4.13

Figure 3.8 shows the variation of equilibrium ratio for the same sysleun as shown in Figure 7.16, but in backward contact test at the same temperature and pressuure. Time slope of tire line increases with each contact, indicating that tire properties of time Iwo pirases diverge progressively. Even, with rich gases the multiple backward countact test canisol generally provide miscible fluids in real cases as discussed in Section 7.1. The experimlmemmt, however, provides valuable infomsalion on the vaporisation effect of tIme iusjcctioms guns. Tine test is particularly valuable to simulate gas recycling of putmtially depleted guts commdensamc or volmmtile oil

~~

.0 0 0~

ujO .0!

00

U.)

Contacu Snage
St

02
03

reservoirs.

1)1)! -2 -l 0 (I +om)( I tfF~)

Figure 3.8. Variation of equihihriuimn ratios of fluid conmponents in multiple backward contact test.

270 7.3 PREDICTION OF MISCIBILITY CONDITIONS

7. Gas t,ujectuon

7..). Iredjction of Mi.ceibility Coudiz joins

271

The values of MMP or MME may be estimated using empirical correlations, or us compositional phase behaviour model. Empirical correlations, mostly developed uusimrg simm lube displacement data, provide rouglm estimates of the nsinimum uniscuhihity puessumre or enriclnnsent. They may be used for preliminary screening or feasihilnty sludies, html sinotild not he relied upon. Phase behaviour models, which provide informalion on tlrertnrodynaunic nsiiscibility, can be used with confidence, after being tuned to relevant experinnental dumla (Section 9.3), 10 predict the miscibility conditions. Even when the miscibility conditiotms for a specific fluid system are experimentally known, tuned phase behaviour models are often required for compositional reservoir simulation of gas injection processes. First Contact Miscibility

nitrogen has been studied by several investigators (13, 23-26]. Firoozabadi and Aziz [23J used experimental shun tube data, and proposed a correlation to estimate MMP for VGD processes, though it was considered more reliable for methane than nitrogen injection.

30

25

20

The injection gas and the reservoir oil should foutts a single pluase flunid whets mmnixe(l mit tiny ratio in this process. This can happen eitimer wills very rich injections guIses, or UI very Inigh pressures. The injection of suds a highly rich gas to displutce oil is not mnornmnuully ccononnnical. It is also often beneficial to deplete a Imigir pressure under saturated oil reservoir, mmd displmmce oil miscibily by vaporising gas drive at lower pressures thaum by first co.mtact miscible displacemenmt at high pressures, because high pressure gas injection is also usually a very expensive operation. Tire single contact (swelling) test provides time first contact MMF data: it us time umnuixiummunun pressure on the phase envelope boundary as shown Figure 7.14. Hence, it can he cslimsnatcd by simulating the above test using a phase behaviour model. Intermediate hydrocarbons, such as propane, butane and luquefied pelroleuunn gases (LPG). known as solvents, are usually first-contact miscible witlm oil at typIcal reservoir eondilions. As time solvents are expensive, they are generally injected as a slug, driven by un lean gas such as methane or nitrogen. The lean gas and the solvent slug should also be miscible for an efficient displacement. Figure D. 1 in Appendix D shows the locus of critical points of various binumry mixtures. At any temperature, the associated critical pressure is MMP for the selected gas-solvent. Figure 7.18 may also be used to estimate the minimum pressure required to achieve first contact nmiscihility between propane, or butane, and several other injection gases 122j. Note tlsat all the curves for different gases coumverge at time solvent critical poitnt. At tenipcrulllmres above its critical point, the solveunt can never fortn two pisases wlren ussixed witis the gas regardless of the pressmre. It unlay, however, fornnm Iwo plmases with tIne reservour oil. Figuire 7. 19 slsows counceptunally lIne mmmiscilmility heimunviomnr of mu simnglc counmlxnuscuil immjectiomm guns, a single component solvent and oil system. The gas and the solveust msre simosvn by tiseir vapour pressure curves, whereas the reservoir oil is identified by its critical point. lire loci of the critical points of mixtures formed by mixing gas and solvent or solvent and oil at various ratios are shown by the dotted curves. Any point above the curves is considered to he a single phase fluid. Hence, when time reservoir temperature, Td, is above the critical point of tIre solvent, the achievement of miscible displacement is controlled hy the solvent-oil belmaviour. At such conditions, M~MP can be estimated by simulating the single contact test between oil and solvent using a pisase behaviour model. At a temperatuire below the critical temperature of solvent, T , 1 the gas-solvent behaviour controls MMP, which can be estinmated fronm Figure 7.19. Vaporising Gas Drive Methane (lean gas). or nitrogen can displace oil very efficiently by developing a miscible bank through vaporising the oil intermediates. The minimum miscibility pressure for methane or

Is

I))

0 3(11)

350 Teunmperatmurc, K

400

450

Fmgure 7. 18.
Reproduced lromn

First contact nmmmnintntjm immiscibility pressure for gas-solvent .SPF.Copyrigtnt. 1221 wiltm rcrniniscioui

Pt
0
0)

--- ----

Locus ofCrimicat Poinls

Ri

0~

~
Ti

Td

Tensperature Fugture 7.19. Connceptuuml phase belsaviour diagranr of first contact minimum miscibility pressure for gas-solvent-oil systeusm.

272
1mm

7. Gas Injec(iouu

7.3. Prediction of Miscibility Co,udjtiopus

273

65.o4 I.296x lO5Xe

c I(Me,,(l.8T_460)023)+ l430x l0~[x(. I(M<., , (l.8T_46O)~25)]

4~

(7.1)
where,
=

Wlmen immiscibility is achnieved by tIre vaporising process only, such as displacing a hydrocarbon mixture willn metimamne or pure nitrogen, MMP can he calculated quite simply by a phase behaviour model. In tlsis process, the critical tie line exteinsion goes through the oil composition, which is independemnl of the injection gas composition. Most methods that advocate using a phase behaviour nnodel to predict the miscibility pressure, propose conducting the forward mnultiple contact test. The test is to be simulated by successively raisimmg tine pressure and monitoring the approach to tIme critical state. A simisple, yet quite rigorotus, umnelhod is linunt of Jemmsenn umnd Miclrelsen (281, winich is based on the ideun of mmcgmnlive liunsln cumlctulullioun described in Section 5.1. As sirown conceptually for a Ilirec colulpoumemnt nmmixltuuc inn Figtnnc 7.3, lime liumuiliung lie liunc anmil lime fiinunl opcrmttiutg line are line .saumne br Oil Ii. I lemncc, for exumnnmple, winemm line gas Y is incrementally added to time undersattnrated Oil B, the nnixlure composition enriches in the2light fraction by moving towards time connrposilion X . nun tire tie lime, which is the same as the operating line. Fine mixture 2 X unt tine hobble point condilion. Furtlmer a(lditi(rn of the gas will lake uultaiiss the composition 2 tIme mixture into the two-pima.se region, wimere time equilibrated gas and oil are sisown by X . 2 and Y respectively.
2

MMP, MPa

XC2CS

= =
=

mole fraction of intermediates in oil, ethane to pentane incluusmve

molecular weight of heptane plus Temperature, K

Mc + 7
T

Eq.(7. I) was found as tIme most reliable MMP correluutioum for lcuunn guns ulnnd nnmlnmngcmm iunjcciiots, with a standard devimmtion of 11.5% mind 25.3% respectively, mum a commnpumnmilivc sttm(ly oi MMI correlations [27]. The above correlation wlmich provides the same vunltue of MMP, rcguurdlcss of tine imnjcction gas composition, relies on the vaporising gas derive comncepl, where mmsiscibility is coinlrolied by tine original oil composition only. The view that MMP for nitrogen and mnnetlnaise is tine same, is also strengthened by the fact, that injected nitrogen vaporises tine Imigimly voluitile metinane to such an extent that the advancing gas front is very nnuch dominated by msretlsane inslead of nitrogen. Koch and Hutchinson [24] showed, however, that MMP declined wlsen a lean gas was added to nitrogen. The MMP for nitrogen is generally higher than timal for nsnethane, hut the difference decreases for oil samples with high bubble point pressures. linumt us, Iluids witlm a high methane content. Hudgins et al. [13] concluded that the oil methane content is an imnportumnmt pusraumneter f~r achieving miscibility in nitrogen injection and proposed the following correluntmoms to estimate

The above analysis imndieate.s that amsy nmixture on the tie line extension, wlnen flashed at time prevailing pressure amnd temnipcrature, will produce equilibrated plnases with counipositions as tirose al the two ends of time tie lime. hut with a negative vapour volume. As time pressure increases and tine trmiscihihity is approached, the limiting tie line length decreases approaching zero tnt the critical point. lnvokiimg tire eotmsponncrrt nnmateriuml bumlance in flash calculations, Eqs.(5. 1-2) we get, 7.,=x(I_nV)+y,mmS~ i=I,2 N (7.5)

MMP for nitrogen.

Pm =

38.39 e~X + 25.10

5 e 2

(7.2)

where,
=

792.06x~.. /(M~,(l.8T460Y~) 1~
2.I58x l0 x~ /(M~,, (I .8T .~ 460)0 25)

(7.3)

52

wlmere z~, x~and yj mIre llme mmiole fractions of conmponent i, in the initial oil, equilibrated (nil and equilibrated gas respectively. mns is the vaporised fraction of the feed (original oil) wisicin will be negumtive in this case. As the tie line lengtls decreases and nhiscibility is approached, the value of n decreumses approaclning nnninus infinity. hence, MMP can be determined by flashing time oil at the reservoir tenmperalure and successively increasing the pressure above the bubble poimst. Time calculated vumpour fraction is monitored, searching for a large negative number. As a pruuctical rule values less than -It) immdicate proxumity to MMP[281. Condensing-Vaporisimig Gas Drive lInc coundennsimng_vmm~sorisiungmntuscibmlmly is mmclmieveml geunerumily at lower pressures llmann Ilnose of time vaporising guts drive. As poimnted out in Sectiomr 7. I, time achievement of immiscibility by condensing gas drive alone is not generally expected for real reservoir fluids. Benlmummin et ,ml. (29( missurned tlnumt tine 3 critical tie line is parumlhel to the L-Il uuxis in time pseudo-ternary diaguarn In develop MM! correlatiomns for rich gums immjection. Tine results were presented by a set of curves correlating MMP with the concentration and molecular weight of the inntertmnediumte fractions of tine injection gas, the mmmolecular weiglrt of tIme reservoir oil C +. 5 anrd tire temrsperature. Glaso 1251 fitted curves to the results of Benlnam et7ai., as follows: 53 4 S 258 1198(1 Mt t \ 1 = 43.74 11.1 752M (32 2311.1 27M )y + ~t).777xtO M c 181 46(1)

(7.4)

The symbols are as defined in Eq.(7. I). TIme muccuruney of tine uthove corrcluulmomn ins predictiirg MMP for nitrogems was founmd (271 to he comsnparuible to Ilmuui of 12(137.1).
Dindorm.nk et al. [7] reviewed the literature on miscible displacement with unitrogen, and suggested some explanations on the apparently conflicting views presented by various investigators on the effect of injection gas and oil compositions onr MMP. Fhme umulhors studied the problem by analytically solving the governing equations for displacing a four component mixture with nitrogen-methane in a dispersion free one dimensional porous medium. Tlmey concluded that the oil tie line becomes tine critical tie line for low nitrogen to uruethane ratio, hence, MMP is independent of gas composition. For gases with a hugh concentration of nitrogen, a cross over tie line becomes critical, hence, MMP is affected by time gas composition. For pure nitrogen injection, the oil tie line controls the miscibility, simnmlar to low nitrogen gases. The above conclusions, though generated on a four component mixture, are quite informative in explaining the apparent diverse views expressed by various investigators, and also helpful to select realistic methods of estimating MMP.

)(

For M~

2~6

=34

(7.6)

274

7 Gas l,i,iectiopu

7.3. Predictio,u of Misc jbj!jty Conditip,us

275

= 38040

58 l326M (55.790.I88M)y +(t.l72xl0~ lM3.730el356.7yM~ )(I.x.r_4~ 44 16 5 t~ M 520e21706YmM ){l 8~ 460 (7.7)

For M~ ~ 2
=

)/IM . (l.wr 460)J 2~5 5 15 where y is the mole fraclion, and P~ and J~ tire in MPa and K, respectively.

6 ~=(l y~.

51 180 1772M (50690 t47M)y +(3 392xIO

Time predicted MMP by the utbove correlation was compared with slim tube data on II oil reservoirs.of5%. The predicted values, ramsgimmg generally fromnm 10 overpredicted to 40 MPa, slnowed an average deviation Time Benham correlation the results with an absolute average
absolute deviation of 37% [311. Al reservoir conditions, CO is often a super-critical compound and behaves as a strong 2 solvcnml extruncting connsponenls as Ineumvy uss C , even at moderate pressures. Hence, MMP for 31 C0 , is generuilly quite low. TIne solvency of CO iuncreases with its density. Tiserefore, 2 2 higimer pressures are requmired at higher tennlperalmures, to increase lIne CO density for achieving 2 two liquid phases, or miscthmlity. tun low temnmperunlure reservoirs, tine C02-oul mnixture may form tlnree phases of vunpour, ((~2rids liquid ummnd imydrocarhon rich liquid. Miscibility by cumrhon dioxide iuujectionr has been studied extensively, and various correlations to estitnmalc MMP imavc been proposed. Ennick et al. 1321 reviewed 17 correlations proposed by different inmvestigumtors. Alston et a!. [33] developed a correlation based on a large number of MMP data for pure CO , 2 P,,=6.05x hO 6(1 .8F~4fi0) M~ (x, , I x~) 0)16
6 8

(7.8) 2~6 where ~m is MMP in MPa, y is the mole fraction of methane imr the insjection gums, T is the temperature in K, and Mc2.C6 is the molecular weight of the C to C fraction in lIre mnjectiOmm 2 6 gas. .
=54

For M~

As it is known that non-paraffin oils achieve miscibility at lower presstmres, (ulaso inehinded lime effect of oil type on MMPby calculating time nrolecunlumr weigint of time stock taunk oil C +, M, mis, 7 M=572.7/(S~,.~570) where S~ i s the specific gravily of C + fraction. 7 7 Kuo [301 simulated the backward multiple experiment, usmng the Peng-Rohmumson EOS. tO predict the miscibility conditions for a number of oil and ruch gas systems. The results were correlated as,
y~ ME~FrMcO~
=

(7.9)

(7.12)

18.46- ~ (I 8T_46O)TIM~,,

(7.10)

where y~is the mole fraction of methane, MC2.4 is the mnoleculunr weight of C to (4 frunclion us 2 the injected 1m gas, and Mc5+ is the molecular of C5+ in tine oil. Tine minimmnultr miscibility pressure, m is in MPa, and F is inn K. lhe values of the constants in tIre above equation are: A= 0.7807248 C= 1.7 138599 E= -0.9909715 B=-0.00 17204 D-l 0695591 F=-0.00l0102

whmeuc xi,, ultm(i x mIme lIne untole fmunctioun of volatile (C , N ), and iumternnediate connupommenls (C 1 bubble 2 2 C~,C0 , ll S)1 its line (nil, 0t36 respectively. For oils witis point pressures less tlsan 0.35 2 value 2 of (x~/x ) MPa, the is taken as unity. 1 Contammminalion of CO witin gases of Irigher volatility. i.e. N and C , increases MMP, whilst 2 connponents such as H S, C , and 2 1 MMP. The estimated addition of less voluntile C lowers 2 3 MMP for CO can be corrected for time effect of 2 impurities by empirical correlations. Alston et 2 uul. [33J proposed the followiung correction factor, dip, to he multiplied to the estimated MMP of pure CO . 2 ( 87.8 ~ 87.8 hmi41~l.935j IInn~ I (7.13) I~ 460) 1\ I .8~ Ic _460) where ~ in K, is tIme pseudo-critical temperature of the impure CO gas, calculated by the 2 weight averumge mixiung nule, T( ~wi(. (7.14)

The above correlation was demonstrated to be superior to those of Benham et al. aund Glaso in a comparative study [27], with a standard deviation of 13.3%. The validity ofthe above methods, which are based on the criterion of cntical inlection gas tie line, for real reservoir fluids is in doubt. Pedrood [31] simulated the displacement of oil by ricim gases using a one duussensional compositional model. An extensive sensitivity analysms of the parameters affectmng the miscibility resulted in tire following correlation: I~49.l50.68630+2.482X10 0 0.20
t 2 54 W 2
gas

(7.11) umuid reservoir onl. nespectively, as

where 0 and s~, represent propertues of tine injected follows,

=

wlserc w is tine weiglnl frunclious of conmspomuent, i, inn gas mixtrure, and Tc is time critical 1 1 meunnpcrunlunrc of coninpomicunt i. Tine umtntirors sruggestecl to use a value of 325 K for the critical leusspcrumtuures oil 1 S mmmd elinamse, instead of their actm.muml values to inmnprove the results. 2 Otlner correlations, using tine molar average mixing rule for calculating the pseudo-critical tenrsperature Inave also been proposed [34].

100

(Yc~ +O.8y~,+0. y~ ,~

2 02

and

276

7. Gas lnjectio,n

7.3. Prediction of Miscibility C~onditio,us

277

Phase behaviour models can be used to estimate miscibility conditions in rich gas and CO 2 injection. As the miscibility is determined by a cross over tie line, the proper simulation of the compositional path is essential. The simulation of simple batch type experiments, such as the backward multiple contact test, cannot generally produce the miscible bank. The most common method is the use of one-dimensional flow coummpositional rnnodels sinrulaliung displacement in a slim tube. MMP can he found by monitoring the commnpositiommal vunriations along the tube, or plotting the predicted liquid recovery similar to using slim tube experinsental data. Any reliable EOS may be used to predict the phase behaviour of injection gas and reservoir oil. The model also requires correlations for calculating the relative pensieability, viscosity and interfacial tension. The selected mathematical parameters to solve the governing equations, such as the number of cells describing the tube and the time step. also affect the results by introducing numerical dispersion [14]. Further details on using a one-dinmnensional compositional simulator are given in Section 9.3.
Example 7.!.

0=1124 resulting in,

s~i~522I5

MMP=35.87 MPa The umneasured vunltue usinmg shuns tube data is 39.99 MPa. The predicted valume assuming miscibility by vaporising gums drive is expected to be higher than those by the condensing process. However, time Firoozahundi-Aziz correlation, Eq.(7.l), predicts MMP=36.51 MPa. which is only slightly higher tlnann the answer by Pedrood method, and much lower than the answers by Glaso and Kumo metlmods.

7.4 REFERENCES
I. 7.iek, A.A: A Conishinsed CondensingIVaporisimmg Mechanismmn in the Displmmcemenl of Oil by EirricIned Gases, SIE 15493, Proc. of 61st Ann. Conf. (Oct., 1986). 2. Stalkimp, Fl: l)isplacemrmenl Behaviotnr of the Condensing/Vaporising Gas Drive Process, SPE 16715, Proc. of 62nd Ann. Conf. (Sept., 1987). 3. Monmroe, W.W., Silva, M.K., Lutrse, L.L. and Orr Jr. FM: Comnposition Palms in Four-Component Systems, Effect of Disolved Methane on 10 CO Flood Performance, SPE 2 Res.Eng., 423-432 (Aug., 1990). 4. Orr Jr. F.M., Johns, R.T. atmd Diimdoruk, B: Developnnenl of Miscibility in Four-Commnpotsemmt Gas E)rivcs, SlE 22637, Proc. of66th Ann. Comnf. (Oct., 1991) 5. Jolmns, R.E., Diumdonnk, B. and Orr Jr. FM: An Analytical Theory of Combined Condensing/Vaporising Gas Drives, SPE Advanced Technology Series, 1(2), 7-16(1993). 6. Johns, R.T. and On Jr, F.M: Miscible Gas Displacement of Multicomponent Oils, SPE 30798, Proc. of 70th Ann. ConI., 985-998 (Oct., 1995). 7. Dindoruk, B., Orr Jr, F.M. and Johns, RI: Theory of Multicomponent Miscible Displacement with Nitrogen, SPE 30771, Proc. of 70th Ann. Conf., Res. Eng., 725-738, (Oct., 1995). 8. Pande, K.K. and Orr Jr. F.M: Interaction of Phase Behaviour, Reservoir Hetrogencity and Cross Flow in CO Floods, SPE 19668, Proc. of 64th Ann. Conf. (Oct., 1989). 2 9. Jolmns, R.T., Payers, F.J. and Orr Jr. F.M: Effect of Gas Enrichment and Dispersion on Nearly Miscible Displacements in Condensing Vaponising Drives, SPE Advanced Technology Series, 2(2), 26-34 (1993). 10. Flock, D.L. and Nouar, A: Parametric Analysis of the Determination of tine Minimum Miscibility Pressure in Slim lube Displecements, JCPT, 80-88 (Sept.-Oct,, 1984). II. Dunnore, J.M., Hagoort, J. and Risseeuw, A.S: An AnaIyti~~ Model for One-Dimeumsional. Three Component Condensing and Vaporizing Gas Drives. SPE J., 24, 169-179 (April, 1984). 12. Glaso, 0: Miscible Displacement, Recovery Tests with Nitrogen, SPE Res.Eng., 6168 (Feb., (990). 13. Hudgins, F.M., Llave, F.M. and Ch~mng, F.T: Nitrogen Miscible Displacement of Light Crude Oil, A Laboratory Study, SPE Res Eng., 100-106, (Feb., 1990).

Estimate MMP of the following reservoir oil and the injected rich gas at 367 K. Component Cl C2 C3 iC4 Oil, mole % 54.50 8.09 5.82 0.78 Gas, mole% 84.63 8.81 4.11 1.40 C7+ Properties: M=209 S=O.8323
Solution: The adjusted molecular weight of the C , fraciion by Eq.(7.9), usimsg its specific gravity, is Glaso (Benham et al.) Method:

nC4 2.17

iC5 0.94

nC5 1.65

C6 C7+ 2.39 23.66

C02 1.05

calculated as 205.9. The molecular weight of the C -C is determined as, 2 6 M~ _~=[~XCMC]/~XC 2 6 =36.84

The calculated MMP for ~ of 34, 44 and 54 at 367 K, using Eq.(7.6), Eq.(7.7), and Eq.(7.8), respectively, are as follows: P ,,=56.62 MPa P,,,~ =56.27MPa P , =53.32 MPa 0 4 04 36 84 Interpolating between the above values for Mcm.c,= . , we obtain, MMP=54. 16 MPa Kuo Method: The calculated MMP by the Kuo correlation, Eq.(7.I0), with M~,,= 186.19, and Mcm.c =
4

37.33, is, MMP=40.54 MPa Pedrood Method: The two parameters representing gas and oil in Eq.(7.I I) are calculated as.

278

7 (;~ Injection

7.4. Ref,re,ue.c

279

14. Dulhien, F.A.L: Porous Medna, Fluid Transport unud Pouc Structimre. Acadcuumie Press, London (1979). 15. HoIm, LW. and Josendal, V.A: Effect of Oil Comrmpositions (1mm Miscible-Type Displacement by Carbon Dioxide, SPE 3., 87-98 (Feir., 1982). 16. Jacobson, I-IA: Acid Gases and Their Contribution to Miscibility, JCPF, 57-59 (AprilMay, 1972). 17. Graue, 0.3. and Zana, E.T: Study of a Possible CO Flood in Ranmgcly Field, JPT. 2 1312-1318 (July, 1981). 18. Tajdar, R.N., Danesh, A. and Gozalpour, F: An Investiguntion of Oil l)ispiulcemsmcnmt by Enriched Gases Usnng a Compositmonush Model arid Its Applicmitionn to Kuirunnj Reservoir, Proc. of 2nd Internationuul (onfcrence otm Clscmntistry un Inndunstry, A(S, Ituilsrmiini (Oct., 1994). 19. Sibbald, L.R., Novosad, Z. and Coslunims, T.G: Metlmodology for tIne Slrceificumlion of Solvent Blends for Miscible Enriched-Gas Drives, SPE Res. Eurg., 373-378 (Aug., 1991). 20. Walsh, B.W. and On Jr. F.M: Prediction of Miscible Flood Performstammce, tue Effect of Dispersion on Composition Paths in Ternary Systems. In Situ, 14(1), 19-47 (1990). 2!. Christiansen, R.L. and I-lames, ElK: Rapid Mea.surennent of Minimmmumsn Miscibility Pressure with the Rising-Bubble Appratus, SPE Res. Eng., 523-527 (Nov., 1987). 22. Stalkup, FL: Miscible Displacemetst, SPE Mousograpir Vol. 8 (1983).

32. Enik, R.M., iloldcr, G. mmmd Morsi, B: A Thernmsodynamic Correlation for time Minimum Miscibility Pressure in CO Flooding of Petroleum Reservoirs, SPE Res. Eng., 8 1-92 (Feb., 2 1988). 33. Alstonm, RB., Kokolis, G.P. and Jamrres, C.F: CO Minimum Miscibility Pressure, A 2 SPE J., 268-274 (April, 1985). Correlation for lnrpure CO Streanmis annd Live Oil Syslenns 2 34. Sebastian, H.M., Wenger, R.S. and Renner, T.A: Correlation of Mininnum Miscibility Pressure for Impure CO Streaunms, JPT, 2076-2082 (Nov., 1985). 2 7.5 EXERCISES

7. I . Use mu plmmmsc bclnmmviour nnno(ldl to develop a tcrnnuury diagrann for C C rmC at 377.6 K mmmd 28 M lui. Wlmuut is I Inc mmmlmitutuimnum reqtnmed cotrceunlrmntnon of propane1 in 1 a guns11~ consposed of ( -(i Inn mmiisci hI ~ (1151)1 ace mini oil cotumposed ol (~(60 umsol%) ullm(l nC ~ (40 moi%) at time above
courditiours.

7.2. The comlmposiliomn of a reservoir oil us given ins the following table. The reservoir pressure amsd teminperunlure are equnal to 82.75 MPa and 387 K, respectively. The oil is to he displaced by unmclisuune. What is your recomusrended pressure for gas injection.
CouiuponenuCl

,_cL.J~_Jc~4,c~_~C~,,_C6
4175 7,61 4.77 M=l7i 0.99 2.6!

Out, mdc %

(6+ Properties: 7.3.

N2 1.05 1.51 35.41 0.33 S=0.8527

C02 4.37

23. Firoozabadu, A. and Azuz, K: Aunalysis mmmd Correlaliomr of Nilrogenn amid Leanr-Gas Miscibility Pressure, SPE Res. Eng., 575-582 (Nov., 1986). 24. Koch Jr., HA. and Elutcinson Jr., C.A: Miscible Disphacenments of Reservoir Oil Using Flue Gas, Trans. AIME, 213. 7-10 (1958). 25. Glaso, 0: Generahised Minimum Miscibility Pressure Correlation, 927-934 (Dec., 1985). 26. Boersma, D.M. and 1-lagoort, J: Displacement Characteristics of Nitrogems Flooding vs. Methane Floodirsg in Volatile Oil Reservoirs, SPE Res. Eng., 261-265 (Nov., 1994). 27. Yurkiw, F.J. atnd Flock. D.L: A Consparative tnvestigumtion of Miuninsmummsm Miscibility Pressure Correlationms for Enhanced Oil Recovery, JCPT, 33, No. 8, 35-41(1994). 28. Jensen, F. and Mnchelsenm, ML: Calculatuoun of First mind Mtmiliple (ountuict Minimnsum Miscibility Pressures, In Situ, 14(1), 1-17 (l99t)). 29. Benham, AL., Dowden, W.E. amid Kuniznnnan, Wi: Miscible Fltuid Displacement, Prediction of Miscibility, Trans. AIME, 219, 229-37 (1960). 30. Kuo, S.S: Preduction of Miscibility for Enriched Gas Drive Processes, SPE 14152, Proc. of 60tln SPE Ann. Conf. (Sept., 1985). 3 I. Pedrood, P: Prediction of Minimum Miscibility Pressure ims Rich Gas Injections, M.Sc. Thesis, Tehran University, Telrran (1995).

(uulcumluule MMI inn lime above exercise, umsumig lIne nmegative fluislm nnethod.

7.4. Nornnnal huntunnic IS 10 be nmscd mis tlte solvent slug in a tnnellmane gas injection process. Assunmsinsg lIre reservoir oil behaves as normal decane, estinsmatc the minimum miscibility pressuire at 360 K. Winat is MM! uml 410 K? 7.5. Estinnmunte MMP of the reservoir oul, described its Exercise 7.2, and the injected rich gas described mr time followiumg table. Commnpare various correlations (measured value=32.75 MPa).
Coin~nentCI Oil, mmuotc % 74.60

,~
9.9(1

C3
3.9!

j4

0.57

4 ~~5~C~6+ 1.12 0.26 0.27

N2

C02

0.17

0.30

8.90

7.6. The mescrvoir oil, described its Exercise 7.2, is to he mniscibly displumccd by a gas commnposed of 90 nmnolX (02 annd tO numol% N2. Estitmsmmte tine mimsinsnum miscibility pressure.

281

8
INTERFACIAL TENSION
Stnrface forces affect fluid phase equilibria. A tension always exists at the interface of fluid phun.scs, due to unhuuiumnccd nnnolcculumr attractive and repulsive forces. For a pure compound, the vmnpour pressure over mn nneniscus inn a pore whiclr is conmcave toward the vapour plnase is snmmallcr than tisat over a flat surface and decreases as the radius of curvature decreases. Capillary condensation, where an unsaturated vapour forms condensate in tight pores is a manifestation of tIme above effect. It, hnowever, becomes only significant in very tight spaces. It is generally neglected in reservoir engineering studies, because the rock in majority of gas condensate reservoirs is water wet, lmence, tIne tiglmt corners are filled with water and not open to the hydrocarbon. Surface forces, lmowever, affect the onset of formnration of new plnases and also play a major role ins multiphase flow ins hydrocarbon reservoirs annd in pipelines. A quantitative index of the molecular tension at the interface is tIre interfacial tension (1FF), defined as the force exerted at the interface per unit length (mN/nm dyne/cm).
~,

TIme capillary pressure is the concept wlnicln is often used in reservoir studies to consider the effect of surface forces on the fluid distribution within a reservoir, The capillary pressure is related to the interfacial tension aurd tlse pore characteristics [II. It has been established also timat tIme relative penneahility, winch describes tine mnultiphase flow behaviour its the reservoir rock, ninny strongly depend on tine interfacial tension [2]. The application of (FT depemident relative penunneability un time dynnamimic evaluation of pinase behaviour models will he described in Section 9.3. Tine evaluation of gas-liquid interfacial tension is of a nnajor interest in gas injection processes where the relative magnitudes of surface, gravitational and viscous forces affect the recovery. The gravity drainutge, commtroiled by tIme balance of gravity and surface forces, is a drive nseclnannisnn whicls is well recogunised inn the oil recovery. It is a common assumption that the liquid drop out by retrograde condensation in reservoir pores is imnirohile, hence, notm-recoverable unless re-vaporised. Recent studies [3J have imndicated tirat tire condensate recovery by gravity drainage can also be quite significant when the gas-condensate interfacial
teunsion is small.

Tine variation of guts-oil interfacial tensiomn witis pressure for a nunsber of reservoir fluids is shown in Figure 8.1. wisere the interfacial tension increases with decreasing pressure [22]. An equation describing the reported data is also shown. The interfacial tension is very small for near critical mixtures and approaches zero as the critical point is approached. Hence the effect of temperature on 1FF depends on the relative position to tlse critical point. For a gas

282

8. lnierfacial Tension

8. 1. Menus renuen! Metlmod,c

283

condensate, IFT is expected to decrease by decreasing temperature, where tine opposite is expected for an oil sample.
20

where, g is the acceleration due to gravity and pL and pV are the liquid and vapour phase (mass) densities, respectively. 1, (lie drop shape factor, is a function of9~ = d~ ide, where d~ i~ time equatorial diamnmcler, or the maximum horizontal diameter of tine drop and d~ is the diameter of tIme drop measured at tine imeighnt d~above the bottom of the drop, as shown inn Figure 8.2. Tuihunlated vaimnes of (, deternsinred by relating tine pressure difference across tine interface to tIme insterface curvature. vs. 9~, reported by several investigators [4,5], are given in Table 8.1. Tine pcmrdamnt drop nnetlmod can also be mmpphicd to mrieasunre the interfacial tension of hydrocarbonwater systetns.
!muisle 8. I

is C
tO

(-0 03417Rt)

IFT= 9.099* in

Values of time dmot slnmll)c functor I [5~.

I t) .90(74 .89822 8(,7i1 .86719 83471 .83154 .81)375 .80074 .77434 .77148 .74639 .74367 .71981 .71722 .69450 .6921(4 .67(140 .66805 2 .8947t .860(,7 .82859 .79774 .76864 .74097 .7(465 .68959 .66571 .5 .89122 .85716 .82525 .79477 .76581 .73828 .71208 .687 IS .66338 4 .88775 .85407 .52213 .7918(1 .76299 .73560 .70954 .88430 .85080 .81903 .78886 .76019 .73293 .70700 0 .88087 1 9

.84755
.81594 .78593 .75740 .73028 .70448

.87746 .84431
.81287 .78301 .75463 .72764 .70196

.87407 .84110
.80981 .78011 .75187 .72502 .69946

.87069 .83790
.80677 .77722 .74912 .72241 .696911

0.69
0.70
0 0 10 20 30 40 50

Pressure, MPa

0.7! 0.72 t).73 0.74 0.75

0.76
Rcpruxliuccd

.68472
.66107

.68230
.65876

.67990
.65647

.6775 I
.65419

.675 13
.65192

.67276
.64966

Figure 8. I. Variations of interfacnal tension of gas-oil witim pressure .srF from 1221 wiih permission.

.64741
0.77

.645)8
.62336 .6(1254 .58265 .56364 .54547 .5281(8 .51142 .49546 .481)15 .46545 .45134

.64295
.62123 .60051 .58071 .56179 .54371) .52638 .50950 .49390 .47865 .46401 .44996

.64073
.61912 .59849 .57878 .55994

.63852

.63632

.63414
.61282 .59248 .57304 .55446

.63196
.61075 .59050 .57114 .55264

.62980
.61)868 .58852

.62764
.60662 .58656

Copyrighi

0.78 0.79
0.80 ((.8 I

Time interfacial tensions between fluid phases at reservoir and surfuuce comrditmonns mire nmneasunred by various tcchmmiquues. Prediction metlmods, witln mm acceptable engiunecriumg mmcctmrmmcy, arc mulso availmmhle. The most widely used and reliable methods are described imn tins chsuiplcr.

.62551) .6(1458 .58460 .56551

.61701 .61491 .59648.. 59447 .57686 .57494 .55811 .55628

.56926
.55084

.56738
.54904 .53150

.54725
.5297% .51306 .49702 .48165 .46690

.54193
.52469 .50818

.54017
.5230(1 .50656

.53842
.52133 .50496

0.82 0.83
((.84

.53668 .51966
.50336

.53494 .51800
.50176

.53322 .51634
.51)018

.51470
.49860

8.1

MEASUREMENT METHODS

.49234
.47716 .46258 .44858

.4908(1
.47568 .46116 .44721

.48926
.47420 .45974 .44585

.48772
.47272

.48620
.47126

.48468
.46980

.48316
.46834

Time gas-liquid interfmncial tension at high pressures is conrmnmomnly nmreasured by mm peurdanrt-clrop apparatus. In this technique, a liquid droplet is allowed 10 hang frommi the lip of a cumpillaty tube in a Inugh pressure visual cell filled with its eqtmilihraled vapour, as slnown schiennuaticmilly in Figure 8.2. The shumpe of liquid droplet at static comnditiotss, controlled by the halmmrmcc of gravity and stmrface forces, is deternmined and related to tIne gas-liqunid innterfmncial tensious 141 by,
rube

0.85 0.86
0.87

.45832
.44449

.45691
.44313

.45551
.44178

.45411
.44044

0.88 0.89 0.90 0.9!

0.92 0.93 0.94

.45272 .43910 .42600

.41338

.43777 .42472

.43644 .42344
.41091 .39882 .387(6 .37588 .36495 .35473 .3439% 33386 .32930 .31394 .70379

lip

0.95 0.96 0.97 0.98 0.99 1.00

.41214 .40001 .38946 .38831 .37810 .37699 .36711 .36603 .35641 .35338 .3461)4 .346(11 .33587 .33487 .32588 .32849 .31594 .31494 .30586 .30483

.435 12 .42216 .4(1968

.39764 .38602 .37477 .36387 .3532% .34296 .33286 .32290 .31294

.43380 .42089 .40846

.39646 .38488 .37367 .36280 .35224 .34195 .33(86 .32191 .31 94

.43249 .41963 .40724

.39528 .38374 .37256 .36173 .35120 .34093 .31086 .32092 .31093.

.4t512t

.43118 .41837 .40602 .39411

.38260 .37147 .36067 .35016 .33991 .32986 .31992 30992

.42988 .41711 .40481 .39294

.38147 .37037 .35960 .34913

.42858 .41586 .40361 .39178

.38035 .36928 .35854 .34809

.42729 .41462 .40241 .39062

.37922 .36819 .35749 .34706

.33890
.32887 .31893 .30891

.33789
.32787 .31793 .30790

.33688
.32688 .3t694 .30688

L.uau;mple 8.!.

~=~~(pi._pV)

(8.1)

Figuure 8.2. lET measurement by pendant drop issetinod.

The vapour-hiqunid inmterfacial tension of a methane-normal decane mixture is measured by nIne pendausl drop method, The gas and liquid densities at the test conditions of 377.6 K anmd 23.59 MIa are eqmnaI to 143.5 and 544.7 kg/mu, respectively. Calculate the interfacial tensions when the liqunid droplet dimensions are deO.600 and d =O.472 nmm. 5

284 Solution:

8 !uuIutfucn~nlTe,u.m,o,n

8. 1. Mea.uuuennen: Alei/uods

285

For a calculated value of ~R=0.472/0.600=0.787, a shape factor of ~=0.5905 is read from Table 8.1. Hence Eq.(8.l) results in, 2 c~9.8 I x(0.0006) x(544.7- I 43.5)/0.5905=0.002399 N/m a=2.4 mN/rn At very low IFT values, the hanging liquid drop becomes very snmsall, requiriumg a very nnarrow tube to remain stable. Time use of a thin wire, under the ttube, for tire drop to hmnog from its tip, instead of the tube, is a more practical arrangement. At conditiomms close to linc critical point. where the interfacial tension is close to zero, tIre pendant drop msrcthmod mummy not he applicable. Laser light scattering techniques have been used 16.71 to Ineasture tire propagations of tlscrmmmunlly excited capillary waves at the vapour-liquid interface, determining very low interfmmcial tension values (0.001 mN/rn). Low interfacial tension values have been detertmiincd successfully by nneasurinrg the gas-liquid interface curvature in an equilibriunn cell (8(. The interface between the phases is curved due to surface forces, as depicted in Figure 8.3. This behaviour is indeed omse of time sources of error in determining phase volumes in equilibrium cells by measuring iluid interfaces and assuming them flat. The interface curvature in a visual, or windowed, equilibrium cell, mmppears as a hand with a finite thickness between the phases, due to light scattering. Tire thicknness of time hand increases with 1FF as the curvature becomes more pronounced at lower pressures.

wirere 0, is the contact angle, Figure 8.3, and nray be assumed zero at low interfacial tension as the liquid completely wets the window (9]. Hence, ~y=(pi._pV)ghl/2 (8.3)

The measurement of liquid rise on the equulibriutn call window provides accurate interfacial tension data with almost no extra effort during conventional PVT tests. It can be applied successfully to measure low IFI, down to values of the order of l0~mN/rn, where the interface between the vapour and liquid loses its definition. 8.2 PREDICTION OF INTERFACIAL TENSION

The interfacial tension between reservoir fluids can be predicted by several methods. Although the methods rely on some timeoretical foundations, they require experimentally determined paratneters. The vapour-liquid interfacial tension of pure compounds has been related to various fluid properties, such as tine density (10], compressibility [Ill and latent heat of vaporisation [I 2J, by vmnrious - investigators. Tine relatiomr between lET and. density has been extended to nmnulticonnpomnemnt systems successfully, resulting in a number of practical and widely used metimod.s ins the oil mnnd gas industry. Ans interesting approach is to counsider the interface, as a tlnird pInase with properties varying between those al the hulk of tine two pinases. This approach, known as the gradient timeory of inimonsrogemmeous fluid, um,ses equihibriunmi timermodynatsnic concepts by employing an equation of sonIc to cmdctmlmute tire required properties from volunmictric datum. The msietlnod has hecmr successfully applied to binary systenrs [13]. It. inowever, requires experimentally detennined paranmeters for all the conmponents and their bimmaries. Because of its elaboration and lack of immmproved resunlts relative to other mnethiods, it has not received nnuch attention in the industry. The two mnsost widely used nmsetlmods of predictinmg the interfacial tension in time petroleum industry are the paracinor metlmod, followed by the scaling law relying on the corresponding states principle.

Window

Gas id Phase

Band

L~~L;~uid
Film

Parachor Method
Lmqu Meniscus It was first reported by Macleod [9]. tlmat the vapour-liquid interfacial tension of a pure connmpoumnd is related to the density difference between time phases, as, (8.4)

Figure 8.3. Gas-liquid interface curvature near the wall of a flat window. At static conditions, the interface curvature can be related to the surface and gravity forces, resulting in differential equations which can be readily evaluated with the hotmndmury conditions depending on the geometry of the equilibrium cell [81. The rise of liquid on a limit window, h, is given by: wimere p~ and p~ are the molar density of the liquid and vapour phase, respectively, in 1 gmnol/cmn and ~ is in mN/m. lime proportionmuulily coumstmmnt, P , known as the panwIuor, has been extensively addressed by 0 Stmgdcn( 14], as mm pmmrmmnmmeter representing the molecular volummne of a connnpound under connditions where tIne effect of temumperature is neutrahised. It is considered to have a unique value for each consrpotmnd immdepcndent of pressure and temperature. Tine parachor values of various pure connpounmds lsave been determined from measured innterfacial tension data, usiung Eq.(8.4) known ins time Macleod-Sugden equation, reported by several investigators amsd reviewed l)y Ali 1151.

F2y(1_smno)12

(8.2)

(pL~pV)gj

286 8 1,ireuf~meiolieuu,cjo,u Parachor values of homologous lsydrocarhons slsow an alnnost limmear relmmtiomsslnip willn tIne molecular weight [16-17]. A relation for the parachor of pure norunsmsl paraffiuns is ums follows, P =21.99+2.892M 0 Parachor correlations in terms of critical properties Irave also heeur pumblisired 1181. I? 4 P =0.324T, v~ 0 3 where the critical temperature T~ is in K and the critical volumme v~ is in cmTn /gunsol. (8.5)

8.2. Preducu:oiu of Inlr;fneia! le,i.uion

287 (8.8)

~~z~z~P , 01

where z is the insole fuactiomn of connnponment i in tine hiqunid or vapour phase. P~jjis the average 1 pusrachor of commnponmeumt i aund j, =
~-

(8.(m)

(P

01

(8.9)

where C~is a temnlperunturc dcpcnndent interaction paraurseter, determined experimentally using lET dmuta our binary nrixtuures.

Paracisors of a number of pure compounds, to he used in Fq.(8.4), mire given ins Fable 8.2 1191. The values for compounds heavier than nC have been estimnated frousm Eq.(8.5). 5 Table 8. 2. Parachor and ~ values of pure conmnpoummds. (.ouuuponenl Paraclnor t value
(02 N2 78.0 4t.0 77.0 3.505 3.414 3.409

The sinimple nmsolar unvermiging as proposed by Weinatug and Katz, Eq.(8.8), is time nsmethod widely
u.used ins petroleutsu indiustuy. Tine paraclsor value of a componsent in a mixture is the sausne as that wlsen puure. The exceptiomn is tlsmit ut nutrogen in reservoir fluids, as given in Table 8.2.

Cl C2
(3

108.0
150.3

3.630
3.681

lIne rclmntiouu hetweenr tIne pmnrunchor mmmd line mnnolecuiar weigtmt of imydrocarbon groups, such as siunglc cmmrhxnnn ntununbcn grnntnps. is conssidencd to deviate signilncanlly from linearity 121,221. This trennd simouuld Ire expected mis tine pmnraclsor-nsmolccular weiglst relation varies for various hydrocarbon houmnologuues. rIse ratio of paraffins, aronsratics and cyclic compotunnds in various SCN grotups are not tine suutnme. resulting in a nonm-hinear relation. Firoozahadi et al. (22] detemunmined parmucinors for crunde oil frmnctions of various molecular weights annd proposed the following equnationn.
=

iC4
nC4 (5

181.5 189.9
225.0

3,597* 3.687
3.682*

-11.4 -i- 3.23M

0.0022M

(8.10)

nCS nC6
nC7

231.5
271.0

3.695
3.726

It

312.5
351.5

3.748*
3.852

simounkl be noted tlsmml genserahised correlations are nsot expected to provide a reliable parachor vmmlue for time oil imeavy eurd, winicls geuserally contaimss a highs concentration of asphaltic and surfmmce active mrmaterials. It is advisable to determine it experimentally.

nC8

nC9
nuCtO

393.0
433.5

3.865
3.855

nCtl 474.1 3.641 nCt2 514.7 3.815 nCt3 555.2 3.872 nCl4 595.8 3.820 nCtS 636.4 3.795 nCl6 676.9 3.822 The given value is for ninrogen in hydrocarbon misiurcs. The value for pure nitrogen is 60.0. * Calculated values from Eq.(8.18). The Macleod-Sugden equation, Eq.(8.4), has been extended to mlmixtures by incorporating various mixing rules. Weinaug and Katz 119] proposed simple tnrolar uuvermuging for lime parachor, ~ ~ =~P (x p~yp~) 1 (11.7)

The insterfacial tenrsion describes lIne nature of nnolecular forces at the interface, whereas the density or tIne mnolecumlar weighmt are hulk properties. In general all prediction methods which relmntc lET to sommne huulk properties, sucin as the denrsity, are not reliable for nnixtumres with component distribution at tIme interface different than tisat of the hulk.
Lnaiuuple 8.2.

A reservoir oil tsmns beeun umnodelled by a mnnixtuure of Cl and nC~~ (60-40 nnole%). The nunixture bubble poinst pressure mit 377.6 K is 23.59 MPa. The properties of the oil phase at the mubove coimdilionns and its equilibrated gas, are as follows:
tluasc Oil
Gas density, 5/cuss

0.5447

methane unmole fraction 0.6000

0,1435

0.9825

Esniunianc tIme gas-oil uuutcrfmucitut lcussnomn at mIsc above conditions, using tIne parachor nsicthod. Tine unicasuured valunc is 2.4 mssN/m. ,Soluuimo,m

01

wIrere x and y~are time insole fractions of counnpommeusl i ins the hiqunid amid v:ulsouir pinmuse, 1 respectively. P~is the paracinor of counponent i. Ilugill and van Welsenes 1201 proposed a quadratic nssixing rule for estimationi of tIme pimase parachor,

lIne

unsiutar dennsuty of tsunt Is

plumuse.c mire calculated

as

M =~xM,=0.6x16 043 u-0.4x 142.285=66.54

g/gmoh g/gnsmoh

0 M =Zy,M,=0.91125xl6.043-t-0 Oh 75xI42.285=l 8.25

288
~=0.008I86 gmol/cm m ~=0.007862 gmol/cun

8. Iuurerfaeial Turn.cjoui

8.2. Predicsio,n ofhuferfneial Ten.ciouu

289

where ~ and~ mire tIne nnohardemnsities of the liquid and vapour phases, respectively, and the exponent (dO) is equal to (45/I76)=(l/3.9l11). The coefficient P~ is equivalent to the parachor in Eq.(8.4) and is given by. P =A~ 0 45~76 v~/ ~ (8.17)

The parachor values are read from Table 8.2 and used in Eq.(8.7) to calculmute lET.
1,0

2S...77(~6x0

0081 86-0.9825x0.007862)+ 433.5(0.4x0.008 186-0.01 75x0.007862)

a=I.708 mN/m Corresponding States Correlation (Scaling Law) According to the corresponding states principle, Sections 1.2, fluids lrelnmuvc siummilmtrly wiucnm scaled properly relative to their critical points. For a pure fluid, time vapotur-hiqunid imslerfacial tension decreases with increasing temperature and heconmmes zero at tine critical point. hence tine following scaling can be considered, IT (I

wlnere v, is the minurlmur critical voluunsse and time value of ~ is estiminated from IF! data ots piure comrnpounnmds, as givens mum lmihle 8.2. It Inns been rclmnted also to the critical conmsprcssihihity functor, mis 1251.
=

1.854426 ,~-it.524lt2
7

(8.18)

Tr)

(8.11)

almInougln tine correlmstiuun does mmml midcqtuumtely nnuitcis lIme reported valn.nes. For nsni5turcs, unnolar avermugi ung us used to calculate vmu lues I~ c Ic I c v C amnd ~, for cutch phmnse to deternnni ne its pmnraclmor froism Eq.(8. 17). The immterfacial tension is Linen calculaled from Eq.(8. 16).

where T, is the reduced temperature and the exponent 0 may he estimmnated experinsnenntally or theoretically.
213 3 Brock and Bird [231 incorporated the dimensionless interfmscial tensions grornp of c~I(P~ Tc ) in the above scaling, adopted theempirical exponent value of 0=11/9 reported by (Iuggenheinn [241 and proposed,

Comparison of Predictive Methods An examinatiour of time above two methods revemmis timmst they are basically the sanme. Both unse experinsnenrtmnlhy deterumnined paraclmor coefficients, directly in the former and indirectly related to tine criticuil property by ~, in tine latter. Usinng a function relmiting the parunchor to critical properties. suds mis Fq.(8.6), ins line first method, will make the two even more sinnilar. The first nmnetlnocl, however, applies nmmixing rules directly to calculate the mixture parachor, wimereas tine pseudo mixture properties are initially calculated in the scaling law to determine the mixtunre parachor. The main difference is the value of exponent inn tine 1FF-density difference relation. This, Inowever, does not himil any of lire two methods to specific exponent values. Hough and Stegenneier [27] proposed an eqn.uation similar to Eq.(8.4), but with an exponent of 1/3.67, deteumnineil using lET dumlum of propmune and butane in tine critical region. It has been shown [6) tisat tire ilough-Stegeummeicrequmulion is nmore accurate at low interfacial tension conditions ( < 0.05 usrN/nn). [he IF! density difference relmitiour can be generahised as,
vv (T/m = (P,~p~I~ ~1 where i~ inas beens uussunmsmed to he constunmst in time above methods.

Ac (1 - Tr) 1/9

(8. 12)

where, (0.1 32f~~ - 0.279) (P c 2/3 Tc t/3~ I (8.13)

I~c is the slope of the reduced vapour pressure curve, plotted vs. tIme reduced tenmmperature, at tine critical point and can be estimated from, ~

=09076I~n(~
(1 T~ /T~)

(8.14)

where Tb is the boiling point of the substance at lime atinmosphmeric pressure P. . 1 Eq.(8.I2), developed for pure fluids, cannot be used for inmixtures as time conmmhx)sition of vapour and liquid changes with pressure and temperature. Lee and Chien [251 replaced the temperature scale with the liquid-vapour density difference,, as it also vanishes at the critical point, (pL pV) (T~ ~T)E (8.15)

(8.19)

Figmnme 8.4 shows IFI of a Noutim Sea gas. condcnrsate measured during several tests, including tine coumstmtunt connposition expansion, constammt volunne depletion and methane cycling at 383 K. All the unmeasured points follow time same trend when plotted against the liquid-gas density difference. The relation is linear on the logarithmic plot, but the slope clearly changes around IFT= I mN/rn. Tine change of slope for other systems has been reported also by other investigators 128]. Tine deviation of predicted lET by the Weinaug and Katz correlation from experimental data on seven hydrocarbon binary systems, including 213 data points, is shown in Figure 8.5 [29]. Note that the nmetimod generally uumder predicts at low lET conditions, whilst it over predicts at

They assumed e=~/I6as theoretically determined by Fisher [26] and proposed,

(TriO
=
(pLL

PM

~vv ~ PM)

(8.16)

higln lET conditions, tine unost accurate results around IFT=I mN/m. A sinmilar trend was also observed for the with Lee-Chico method [29).

z 290
8 I,iiiii~ic,alI(~i.(lOII 8.2. Iredieuio,u of hue ufaeia! Tepi.uio,i

291

!inc above muno(Iilications reduced tIme average absolute deviation from 23% to 14%, wlnen tested against 65 data points on tnulticonumponent model fluids. A similar improvement was also observed for real reservoir fluids iumcluding gas condensate and oil samples [30]. Schechler and Guo conducted us comprehensive survey [16] of reported 1FF and density data and clearly demonstrated tlnat time optinnunss value of E, particularly for hydrocarbon systems at low lET conditions, is less tinaum 4. Based on experimental data and molecular models, the authors concluded that a sinsgle conmstant value of E=3.88 in Eq.(8.19) should suffice to predict IF! witluinm ml range wimere tIne critical point scahinng is applicable, that is less than I mN/m.
(TI/iSO

~P ,(xp~, 0 1

y p~) 1

(8.21)

As clummnsginmg tIme vmultnc of exlmonsennt rcsunlts mr clnminnge of lIme parachors, lhney calculated approprimnle parunclnors for a large mnununhcr of punre connpomnnds, using experimcnntal lET data. Ilmeir cuilcuuluiied vmulumes mire giveun ins the plnysicunl properly table, lahhc A. I, in Appendix A. ihe it lain s unhso kvekm xuI I inncmnr com reluntionss betweuur tine parmmchor and nsnolectnhnr weight for 1 vminious inydrocunrhonr isonnsologumus. lime correlation for alkanes is as follows,
I)cnsniy l)illcrcnce, g/cuuu3

l~=l l.73+2.9871M

(8.22)

Figure 8.4. Measured gas-condensate 1Ff of a North Sea reservoir fluid mit 383 K.
tOO

A linear rclmstiomr between puiracimor vunlues and nnolecular weigirt was also suggested for all types of hydrocarbons, witim a deviation of 22%. as follows: P =3.72+2.95l9M 5 lIme above correlatiour ntay he mused to estimate parunchor values for SCN groups. (8.23)

S V S. V 0. C 0

!~~oniph 8.3.
.

Predict tine gas-liquid IF! of tlse ntmixture described inn Example 8.2. using the LeeChienm n,nd SclsumclsicrGnno mmci Imods.
l~

Suuhuuine,,u: L.ee-Chien Method

TIne counmponent criticuil properties are read fronn Tumble A.l in Appendix A. The values of ft for each coussponennt is calcumlated using Eq.(8.I4),
-

too
OOt
01 .1 iO 01)

ft ,-I330947

3 ,~

17.29700

tvr, nuN~m
Fugunre 8.5. Deviations of preducted IFT by the Weinaug-Katz mumetlmod for binary nmnixtuues. Botls correlmntions cmuo he imsmproved by nmmiking tine IF! expomsenit, l/E, a functions of mIne liquid vapomur deursiny difference 129]. The optiurrised value for tIme Weinsaung-Katz cuimrelation was determined by regressung E to ummiuninmuise the devimstions of tine predicted hinsmmny results mis, E=3.583 + 0.I6(~-p~) (8.20)

Vunlues of ~ for nnuethmmne and nsornnsal decane are read from Table 8.2, equal to 3.409 and 3.855, respectively. TIme values of v,. T,, P,, ft anscl

of each phase are calculated by molar averaging and siubstitunted in Eq.(8. I .1) muusd Emq.(S. 17) to determine A, and P,, , respectively.
l5uaw v, cuuu/guuuot

c1

I. K

I,, Mts

(1

Liquid
Vapour

299.16
07.37

36t.42
198.04

3.6034 4.5554

14.9045
13.3793

3.5874
3.4168

A,, mN/rn 28.2673

19.6633

195.970 64.104

3 wisere p us mn gmol/cmn . The contribution of time second terism imr the above correlation is 51 in most cases, resulting in a constant exponent in practice. negligible

Suihstuniting tIne mnlmovc values of t~ aund nmnol mur dcnmsities, calculated in Example 8.2, in Eq.(8.l6). we obtunin,

292
(Ti4vnmu)_

8. Inierfacia! Tension

8.3. Water.flvdrocarbon iumteufacia! Tension

293

l95.970x0.008l86-64.IO4x0.OO7862~l.10024 Connprehensive data omr multiconnponent hydrocarbons-water, particularly real reservoir fluids are scarce. For reservoir gases, time trend is very nsuch similar to that of methane, Figure 8.6. witim tine presence of heavier connpounds in the gas reducing 1FF. The effect of pressure on oilwmuter IF! is mininnmul above tine oil htnbhle point. The reduction of pressure below the oil bubble point generally reduces tIme oil-water lET due to release of gas out of the liquid nmixture. The oil-water interfacial tension is expected to decrease with increasing tenmnperature, but resmnlts contrary to that at some conditions have alsobeen reported [35).

a=l.453 mN/rn

Schechter-Guo Method The parachor values for methane and normal decane are read from Table A.I, equnal to 74.05 and 440.69, respectively, and substituted in Eq.(8.7), but with the lET exponent of (1/3.88), ~>=74.05(0.6x0.008 I 86-0.9825x0.007862) 440.69(0.4x0.008 186-0.01 75x0.007862) a=l .864 mN/rn

a z

555

8.3

WATER-HYDROCARBON INTERFACIAL TENSION

S

The interfaciai tension of pure hydrocarbon-water has been measured over wide ranges of temperature and pressure by various investigators [31-36]. ihe reported data often show substantial discrepancies as reviewed by Firoozahadi and Ramey [37]. The variation of methane-water IFT with pressure and temperature is shown in Figunre 8.6. fhe imsterfacial tension decreases with increasing pressure andlor temperature over the tested regiomn.

595K

ID

20

/0

40

50

Pressure, MPa

Figmnmc 8.7. Water-tn-octane inmerfacial. Pcmroteurn

wimtu joeruuuissi In.

Suucicmy umi Cuiunatla

Copyrighu. Rcprixiunced

fruinsu

1361

E z E

Altlnough paraciror values, about 52 [25, 38), have been reported for water, time use of vapomnrliquid correlations, Eqs.(8.7) and (8.16), to estimate lET of hydrocarbon-water is not recomnsmend. Timese correlations can pnoduce higlnly erroneomns results, even for sinsrple binary systeunns [371.
l~xauumpIe8.4.

Estiusmale the interfmmcial tension between tIme gas inn Exanmuphe 8.2 anmd water.
Solu,iouu:

Connsidering that the gas is predominantly composed of methane, Figure 8.6 may be mused to estinnuate gmns-wmmter lET (~,=43mN/nnn
mit

20

40

60

81)

ttX)

377.6 K and 23.59 MPa.

Pressure. MPa Figure 8.6. Methane-water interfacial tension. SPE Copyright. Reproduced from 1351
with permission.

Figure 8.7 shows the variation of n-octane-water lFl, witlm pressure and tensnperature [36]. The interfacial tension decreases with increasing temmnperature. Ann iurcremise in pressure, however, increases lET, but the effect is small. Tirere are reports sinowing sliglnt redmuction of IF! by increasing pressure [33], contrary to that in Figure 8.7. Tine effect of pressure on lET of pure liquid hydrocarbon-water is generally small and can be neglected in most cases.

Firoozabadi and Ramey [371 denmonstraled that IF! between water and pure hydrocarbons. over a wide range of temperature-pressure, can be described by a single plot shown in Figure 8.8. Tire 1Ff frunctioms is as follows: II I Iu.uinctnonm
E

li2S t,~u,W(p

. .5 03125 P S )j( .1/It

294

8.

(ml frrfacial

Te,uciomu

8.3 tVoiur-Flwlroearbo,n /,ute,Jacial Tensio,u

Sohuuiio,u.

295

where, ahW is in mN/rn, pW ~ ph are the water and hydrocarbons phase deunsity its g/cnnn , respectively, mmd T h ns the hydrocarbon pinase critical temperature. 1 The reliability of correlation was demonstrated for various compoumnds ranmging fronn nnethanme to n-dodecane. The plot can be represented almost by the following eqmuatiois. (8.24) whicin is also shown in Figure 8.8.

Neglecting the solumbility of Imydrocuirbon in waler, tIme water density at 377.6 K and 23.59 MPa can he estimated from Eq.(2.88), as in Example 2.15. AV,,~.0.0057RI p~=l.040 g/cnn i\V~ =0.046034 0 t3,,=l.03999

Tmnkinng the uniolar averuige critical tenssperature of mime hydrocarbon vapour and Iuqnnnd pluunses as calcuulmstcd un Exausiphc 4.3, neglecting the solubility of water in hydrocarbon pliumses aund using I:q.(8.24), tIme interfacual tension us calculated.
ttydruucarlxin

20
. -. .--. =.., ... ... . .1 . -

p.

g/cuui

g/cuuu

pW.p0~

g/cuus

F,. K

1,

1FF, nuN/rn

viupulur

tuqumid

1.1)4

04

(1)435
0.5447

0.8965

i).4953

198,04 361.42

1.9070 1.0448

44 51

16
. . . -. . -

TIne interfuicial teunsions hetweenm the hydrocarbons liquid phase and water is expected to be lower nlnan tlnat of tine Vuiponir plumuse, cuuuntrary to the above calcnulated values. Using the original plot mu Figure 8.8. unsiead of Eq.(8.24), results ins a calculated vapour-water imuterfacimil temssion of 49 nurN/m.

12
U
. -..;.

Equiauion 8.24
.

. . .

:
.

-. .

8.4 REFERENCES
. .

1 I

.
-.

- J. . - .

.j~

I. Itmimduumn, C. amid I .uunsgcmoun. I ).(: Imntluuenncc (it Very I_ow Imnterfacnal Tenmsnon on Relatnvc lernunemubility, SPEJ.. 391-401 (Oct., 1980). 2 D munesin A Kruuius D lhcnsderson G D ansd Peden J M Vnsuah Investugatton of Retrograde Pirenousmenron and Gmms Comndensate Flow in Porous Media, 5th European
Synmsposiuusu on Improved Oil Recovery. Budapest (1988).

. -

..

- .-

.:

i:

1 Andreis J M 11 truser E A md Tucker W B Boundary Tension by Pendant Drops Ircsensled ml time SOtIm (olloid Sympociuns Cannhrudge Massaclsmnsetts (June 1918)
00.0

~ 0.2
0.4 0.6 0.5

-04-,--

I .0

Densumy Dufference, g/cisr Eugmure 8.8. Generahmsed correlation for wunter-Imydrocarhon nnterfmucual lemsslon of Canada Copyrigini. Reproduced fruin 371 WIl ti peru ussioni
tcurotcuunu

4. Niedcrlnmiuser, 1)0. mnnsd llmumtcll, FE: A Conected Table for Calculatiour of Boundary Tennsions by Pendmrnst t)rop Metlmod. Researcim onm Occurrence and Recovery of Petroleum, A ConnIrihuutionr frourn API Researcin Project 27, 114-146 (Mar., 1947).
Society

5. lIaniff, M.S. uunsd Pearce, A.J: Measuring Interfacial Tension in a Methane - Propane (imns-Corudensmsle Systeun tJsimng mi Lumser Light Scattering Technique, SPE Res.Eng. 589-594 (~ , I 990) 6. l)orsinow RB: TIne Siumunltmimseouns Measuretneunt of Interfuiciai Teursion and Oil Viscosity for Prumdhoe Bay Fluids by Surface Laser Light Scattering Spectroscopy SPE 22633, Proc. of 66tln Ann. Conf. (Oct., 1991).
mit Resem-voir Comsdiliours

The mshove correlmituon was evalmnated for estimating the oil-wmnter IF!. wisicis uusdicmnted ernors exceedumsg 30% (37]. The presence of surface active connpommnds iii oil nsay proisuhut tlse use of ammy generalised correlation wlnicim does nsot take smucts effects into mmccount. Wiseur tine 1Ff value at a siurgle point, often at the atmospheric conditiomss, is known, mm parallel crmrve to Ilsunt in Fmgumre 8.8 may be used to estinniate 1Ff at other pressure-teurspermltuure vmtlues 1371. The presence of salts in water may alter its unterfacnal tension with oil shugintly at typical reservoir conditions. TIne 1Ff may decrease [39), or incremsse (40) by uncreasing the salt concentration according to various reported data.
E.mauuuple 8.5.

7. Pauresls, A , Todd, A.C , Sonsnerville, J. and Dandekar, A: Direct Measurement of Innterfmuciai Tension, Density, Voluunsne, and Consnpositions of Gas-Condensate System, lrans. tChensnE, 68, Pant A. 325-330 (July, 1990). 8. Cmnhn, 3 W Cuiticah Point Wettinrg, The 3. Chem. Physics, 66, No. 8, 3667-3672, (April, 1977). 9. MacLeod, D.B: 0mm a Relation Between Smnrface Tension, Density, Trans. Faraday

Estiunnate tine oil-water and gas-water interfacial teunsion for fluids described mu Exununuple 8.2, using the generalised hydrocarbon water correlatioun.

Soc., 19, 38-43 (1923).

296

8. FunfrufaniF Te,n.ciinn

8 5. Ererni,ses

297

hO. Mayer, SW: A Molecular Paraisreter Rclaliomsslnip belweenr Surfmmcc hemmsionm mmnd Liquid
Compressibility, J. Phys. Chem., 67, 2160-2164 (1963). 11. Sivaraman, A., Ziga, J. and Kobayashi, R: Correlations for Prediction of lnterfmmcial Tensions of Pure Alkane, Napthenic and Aromatic Compounds between their Freezing and Critical Points, J. Fluid Phase Equilibria, 18, 225-235 (1984). 12. Carey. B.S., Scriven, L.E. and Davis, U.T: Semiempirical Theory of Smurface Tensionm of Binary Systems. AIChE, 26, No. 5, 705-711 (Sept., 1980). 13. Carey, B.S.. Scriven, L.E. and Davis, H.T: Semiempirical Theory of Surface Tension of Binary Systems, AIChE, 26, No. 5,705-711 (Sept., 1980). 14. Sugden, Sr The Parachors and Valency (Romntledge, 1930); A List of Pmnrumchors, Brit. Assoc. Report (1932).
15. Ali, 3K: Predictions of Parachors and Petroleum Cuts anti Pseudocommmpomrents, J. Fluid Phase Equilibria, 95. 383-398 (1994).

28.

llunygens, R.J.M., Ronsdc, II. and ilagoort, J: Interfacial Tension of Nitrogen and

Voluntile Oil Systems, SPE 26643, Proc. of 68th Annr. Conf. (Oct., 1993). 29. Dmmnesh, A., Dmmnmdekar, A.. Todd, AC. and Sarkar, R: A Modified Scaling Law and Paraclnor Method Approach for lnmmproved Prediction of Interfacial Tension of Gas-condensate Systems, SPE 22710, Proc. of 66th Ann. Conf. (Oct., 1991). 30. Dandekar, A: Interfacial Tension and Viscosity of Reservoir Fluids, PhD Thesis, Heriot-Watt Unniversity, Ediumburgim, Scotlammd (1994). 3 I. Elomnglm, E.W., Rzursa, M.S. and Wood, B.B: Interfacial Tennsion at Reservoir Pressures and Tennperatures, Apparatus and tine W~ter-MethaneSystem, Trans. AIME, 192, 57-60 (1951). 32. Micinmmcles, A.S. mmnsd llmumuser. E.A: lnmterfmmcial Tension at Elevated Pressinre and Ieuumpermmtmmmes II, J. Ihys. ( hncnnm.,55, 408421(1951). 33. Ilmrssans. ME., Nielsour, RE.. Callnomun, iC: Effect of Pressure and lennperatmure on Oil-Wmmter hnmterfacimul Teumsiours for mu series of llydrocmmrhons, Trans. AIME, 198, 299-306 (1953). 34. Jennings Jr., 1I.Y: TIne Effect of Tenniperature mind Pressure on Interfacial lension of Benzene-Water and Nominal Decane-Water, J. Colloid & Interface Sci., 24, 323-329 (1967). 35. Jennings. lI.Y. arid Newnnan, Gil: The Effect of Temperature and Pressunre on tine Interfacial Tension of Wunter Aguuinnst Metlnaune-Normmral Decane Mixtures, SPE J., 171-175 (Jmmnse, 1971). 36. McCmuffery, F.G: Measmnrenmnent of Interfacial Teissions and Contact Angles at Iligln leunnperature and Pressmure, JCPT, 26-32 (July-Sept., 1972). at 37. Firoozahadi. A. anal Rannney Jr., H.J: Surface Tension of Water-Hydrocarbon Systenos Reservoir Conditions, JCPT, 27, 41-48 (May-June, 1988). 38. ilmnkinm, D.l., Steinberg. D. ammd Stiel, LI: Geuserahised Relationship for time Surface lennsion of Polar Fimnids, lnd. Eng. Cheun. Fund., 10, 174-175 (1971). 39. Bmtrtell, F.E. mmmd Niederhmmmnser, DO: Film-Fornning Constituents of Crude Petroleum Oils, Fmnndaunenlmul Researclr on Occmnrunnmce and Recovery of Petrolemntin. 1946-1947, API. Marylmmnd (1949). 41). Aveymnrd, R. ammd I Imuydomm. D.A: iimennsodynannic Properties of Alipinmmtnclnydrocmmrhonm Wmuter lustcrfmnces. Frmmns. Farmudmmy Soc., (ml, 2255-2261 (1965). 8.5 EXERCISES

16. Schechter, D.S. and Guo, B: Paracisors Based on Modern Plrysics mmmd Their ttses inn LET Prediction of Reservoir Fluids, SPE 30785 Proc. of Ann. Commf., l)mtllmms, USA (Oct.. 1995). 17. Baker, 0. and Swerdloff, W: Calculation of Surface Tensions. 3: Calculunlion of Surface Tension Parachor Values, Oil and Gas Journal, 53,87(1955). 18. Fanchi, J.R: Calculation of Parchors for Compositional Sinnulation, Aim Update, SPE RES.ENG., 433-436 (Aug., 1990). 19. Weinaug. C.F. and Katz, D.L: Surface Tension of Metirane EC, 239-246 (1943).
-

Propane Mixtures, I &

20. Hugill, J.A. and Welsenes, Van, A.J: Surface Tension: A Simple Correlationn for Natural Gas + Condensate Systems, 3. Fluid Phase Equilibria, 29, 383-390(1986). 21. Ahmed, T: Hydrocarbon Phase Behaviour, Grulf Publishning Conimpany (1989).

22. Firoozabadi, A., Katz, DL., Soroosh, 11. and Sajjadian, VA: Surface Tension of Reservoir Crude Oil/Gas Systems Recognising time Asphalt in the Heavy Fraction, SPE Res. Eng., 265-272 (Feb., 1988). 23. Brock, J.R. and Bird, RB: Surface Tension ansd the Principle of (orrespoirding States, AIChE, 1, 174-177 (1955). 24. Guggenheim, E.A: Thermodynamics, 3rd Ed., North-Holland Pub. Co., Anuisterdam (1957). 25. Lee, S.T. and Clnien, M.C.l-l: A New Multiconsnponent Surfmuce Tensionn Correlation Based on Scaling Theory, SPE/DOE 12643, EOR Synmmposiummr. Tmmlsmn (1984). 26. Fisher, M.E. and Scesney, P.E: Visibility of Critical Exponeunt Remrornnalization, Physics Review A, 2, 825-835 (1970). 27. 1-lough, E.W. and Stegemeier, G.L: Correlation of Surface and Interfacial Tension of Light Hydrocarbons in the Critical Region, SPE 3., 259-263 (Dec., 1961).

8.1 . The gas-liquid inrterfmucimuh temnsion of nmsetlnane-nonunmal hmmtane nmnxture at 311 K and I 2.07 MPmm is mmicasured by tine peurdant drop usnethod. Calculate the interfacial tennsion wlmen the droplet dimnenmsions mire d~=0.228 and d =0.325 mnn. 5 8.2. Tire vapomnr-Iiqumid equilibriumnn dunta of a five-component gas condensate usmnxtmnre at 353.1 K ins a constammt conmnpositiour expamnsion test are given in the following tables. Conmrpare tine ummeumsmnred IF1 values with those predicted by the original Weinmaug-Kali. nnetimod annd its modifications Eq.(8.21) by Scheclnter-Guo and Eq.(8.20), and the Lee-Chien mnetinod.

298 Gas condensate composition. Component Methane Propane Mole Fraction 0.8205 0.0895

8. !uulerfncial Tension

8 5. Erercises

299

n-Pentanne 0.0500

n-Decaumen-Hexadecane
0.0199 0.0201
_____________

Liquid_phase_properties_and_gas~1iguid_interfacial_t~nsion. Presuure Methane ~j nPentane_~caim~ MPa Mole Fracmiour 30.43 0.68130 0.1073 0.0772 0.050) 29.06 0.6640 0.1115 00827 0.0564 27.68 0.6414 0.1133 00886 t).0634 26.30 0.6187 0.1186 0.0940 0.0689 24.92 0.6003 0.1208 0.1001 0.0743
Vapour phase properties. Pressure Methane

Connpare the measured 1FF vunhues witlm tlrose predicted by the Weinsaug-Katz nnetlmod as modified by Scheclmter-Guo, and tine Lee-Chien method. It has been suggested to improve the prediction by adjusting the parachor value of time plus fraction to unatch the experimental data. Find the optirnised parachor valune of the plus fraction.
Densiny 1FF

~
0.0774

g/cnis3
0 480

mN/rn
(it 18 0.229 ((.39)) i).536 (1.738

8.4. Estimate the interfacial tension helween tine gas in Exercise 8.2 and water, at the gas isydrocarhon dew poinl of 31.98 MPa and 353. I K. using the generalised water-hydrocarbon correlation. Connpare lhe results with the value obtained by assuming the gas as pure methane.

0.0854 00933

i),0998
0.1045

(1497 (1.513 (1.524 (~535

Pruuuani~ .._n~~nuan

inc 0.0167 1)0145 1)0127 0.01(18 0.0095

un-t-tcxmudccauic 0.1)137 0.0(199 0.1(1)73 00054 0.004)

Density
g/cnnn3

MPa 3042 29.04

27.66

Mole Fraction 0.8362 0.8448

0.8512

0.0869 0.0863
0.0857 0.0844 0.0841

26.29
24.91

08580
0.8625

0.0465 0.0445 0(143) 0.0414 0.0398

1)298 0.278 026) 1)246 li.23 I

8.3. The interfacial tension, consnpositionm and dennsity of gas miund condenismile plnutscs were measured during a constant volunse depletions tests. fine results at two puessuircs are mis follows: Pressure, MPa C02
Cl C2

34.47 2.42
78.06 6.71 Liquid 0.4) 1.72 49.82 612

20.68
V~~q9~_ 0.65 1)26 2.59 1.39 78.88 35.36 6,90 5.43

Conm~p~nent,Mole%V~ppur N2 0.65

C3 iC4 nC4
iC5

2.99 0.60
I .25 0.48 0.59 1)73 1.10 1.01 0.68 0.47 0.33 0.25

3.30 0.74 I .68

0.75

3.04
(1.59 I .25 0.48 058 (1.84 .02 0,95 0.65 0.35 0.32 1)22

3.36
0.82

I .94
0.95 1.23 21)9 3.38 4.03 3.29 2,68 2.26 2.19 2.09 2,27

nCS
C6 Cl C8 (9 ClO CII C12

0.95
I 48 2.25 2,53 1,96 1.51 1.22 1,17

Ct3 C14 Cl5 Ct6

C17

0.22 0.26 0.19

0.12 0,11 0.10

1.11
1.21

0.14
0.16

1,16
1.01 0.77 0.87

0.13
0.06 0.04 0.03

2.11 1.63
1.51 1,47

Ct8 Cl9 C20+ Molecular Weight

densiny,g/crn Measured IF), mN/rn

0.09
0.57

0.84
15.42

0.02
0.11

1.35
16,91 120.62

28.05 97.44 0242 0.593 0.372

25.32 0.143 1.389

0.628

301

APPLICATION IN RESERVOIR SIMULATION

Phase beinaviour nmsodels are used extenisively in pctroleunsn insdustry. A nmiodel can be used to evaluate tIne consistency of memssured PVT data, or to gennerate such data as input in black oil reservoir simulation. The nrodels are also used in pipeline and weilbore multi-phase flow cmmlcmmlmntions. Ins tIme design and operation of surface facilities, phase behaviour nnodels are ennsployed to determine the properties and the amount of equilibrated gas and oil. The main application of phase behaviour models, based on determining the fugacity of components in both plnases by EOS, is however in conmmpositional reservoir simulation. in a simmrulator, the reservoir is couumuusonly divided into a number of grid-blocks, or cells. The fluids witlsin eacln cell mire considered to he in equilibriuns at the cell pressure ansd temperature. Time clnmunmge of reservoir conditions with time is investigated by determining average values in each cell during successive susnall tinme steps. The equilibrium condition over a time step is determined by flaslm calculations in eacls grid block. As reservoir calculations are generally itcrmmtive. nsnore than one equilibriums flash calculation per each grid-block at any time step is reqmnired. For a large reservoir, tire total number of equilibrium flashes may exceed many milliouss, consunning a large comnputumtional time and making the simulation very expensive. As the number of equations in conventional flash calculation increases with the number of components, see Section 5.1, tIne nmnmber of components characterising the fluid is conmimonly redmnced by grouping to reduce line computational time. Ann innnpoutminnt conmsideruution ins applying a l)lmase behaviour model to reservoir studies, is wide ranges of composition and pressure wimich are to be modelled by EOS. In tlre integrated nrodellinsg approach, where the phase behaviour model is to cover reservoir, wellbore, pipeline, scpmmrators, etc.. in rune sinrnuluutinn, time task becomes even more stringent and tIne model often immi Is Inn provide smil isfuiclony results. ~IIme connnmmunnm mmppromncin is t(n calibrate, or tmnunc line nsmodel. agaiunst experinnental data. lime selection of required data and parameters of EOS for adjustment

in tIne tuning process and the use of

engineer.

relevant methods constitute a major task for reservoir

302

9. App/ieariouu inn Receunoir Siuuuuilatiouu

9. 1. Grouping

303

9.1

GROUPING

Exanup!e 9.!. Describe the C7~fraction of tine oil in Exannple 6.3, identified to C , by a number of 45 pseudo-components (using tine Whitson nmethod. S(,!ustzo,n: The last carbon nunnssher describing the C , is 45. Hence the number of pseudo7 connponcnmts for lime C , frmmction is calculated from Eq.(9. I), as 7 45 N~=Integer 11+3.3 log( -7)]=6 The componenls connprising each psemido-component are identified by calculating the molecular weight boundaries of each pseudo-conmponent. Using Eq.(9.2), with the generahised msnohecmular weigint of single carbon number groups, we obtain the upper molecular weight of first pseudo-comsnponent, M~=96 {exp[( l/6)ln(539/96)J
J = 128

The concept of grouping has long been eunployed in fimukl description, html this is done nrostly due to limitations in the compositional analysis. Time nrost convenmlionial nnsetinmnd is to describe the hydrocarbon mixtmnre with discrete componeurts to nornral penltmnne and imexminses emich as a single carbon group and lump all the heavy fractions as the heplanes plmus (C ,). Tins is not an 7 efficient method of describing a reservoir fluid, particularly in conmnpositionnal siumiulation studies, where it is desirable to minimise the number of consmponents winile still retainming tIne reliability of predicted vaimmes by phase behaviotnr models. It is expected that the fluid description requirennrent varies witim tire complexity of tine process which is to be modelled. The phase beisaviour of a reservoir fimmid unnder pressmnre depletion may only be modelled by two components [II, whereas more than len eonmmponents nnay be required for miscibility studies. The key points in component grouping are: 1. The number of groups required and the distribution of components witInin emscln group. 2. The estimation of group properties required in pimmuse hehmmviou.nr nniodclhunig. 3. The retrieval of fluid description in terms of the original componeurts when needed.

I iemncc, C , C mmmd (~. wills time last Immiving a usnolecumlar weigist of 121, are assigned to 7 5 tine first psctndo-connmponmenl (gromnp). Similarly, for otiner groups we obtain, (;runut) No. k
~(pperM honinduury, M~ 28
Ill

Comptunenis in Group
(~ (
,~

Group Selection Many investigators 11-13] have given recomnrendmmtions on selecting tIne nuimmmsbcr of pseudo components (groups). In general, 4-JO pseudo-conmponenrls are considered adequate for simulation purposes. A simple approach is to add nitrogen and carbon dioxide, at low concentrations, to nnethane and ethane, respectively, amnd to combine iC with nC and iC with nC . The C frumction is also 4 4 5 5 7~ characterised by a numunher of pseudo-conrponensts mmnmd iuscludcd. Whitson [51 proposed representing the C , fraction of a mixture by N~psetndo-connponents 7 calculated as follows, N~ = lnteger[I + 3.3 log(N 7)] where N is the last carbon gromnp number. The boundary betweenm the consecutive groumps are based our tire molecular weights Mk, given m M {exP[+]lts(?J]} k = I, 2 N~ (9.2) 7 (9.1)
4

227 303 404 539

C cCus C1 -C~ 17 2 C C,u C 2 .C s 71 4

The mipplicunlions tnf coustinstnomns descriptions to select 35 psetmdoconmponents (nnntnlti carbon groups), unt (luadrmulnurc poimuls, discunssed in Sectionm 6.3, is time reconnnnnended method of dcscrnl)insg the incavy ensd. line selected pscmndo-comnuponenls will be as effective as choosing twice as unany, selected randomly or at equal intervals I l4J. Selectmng 4 psemndo-connponents by (lie quadrature method to describe the Ci+ fraction and following line siunnple uupproacim of addinsg isonners to normal hydrocarbons, described above, wnll redmuce lime tolmil nsmnnmnher of coumnponennts describing a fluid to around ten. lhe number of grounps, however, cmnnm further be reuluced particularly by grouping the iiglmt fractions. Lm cI mnl. Ill J proposed to group compousents on the basis of their volatility, using equilibrium runtuos obtainned by flmusIminng line fluid at reservoir tennperature and the average operating pressure. Tlmcy prcseuuted dnffercnut conTcimoiomns for groupinng ol highl and heavy components. Pedcrsen et at. 1151 suuggesleul tim grouup tine cousupouncunts hunsed ms nnass, llrat is, each group containing muppuoxnnnmmmleiy tIne saumnc umumuss fumiclion. A ntnnsnher of fluids were described by 41), 20, It). 6 mmmd 3 psemmdsmconnnponnents mmmd their smnttmratiu)n points were predicted usinug EOS. They consclmmded tlnat a 6-connsponsetnt rcpucscnntation was .cuflncient for an accmmrate prediction of the smmtunruutiomm point. Cotternnsan and lraumsnitz 114) used the criterion of equal mole fraction to select groups. A representative fluid descriptions is expected when tire groups are formed by due consideration to the volatility and the concentruntion of comimponents in time nnixture. Newley and Merrill [12], suggested a method of grouping based on minimising the difference, L~,between the apparent

The components of the original fluid with molecular weigints falling witimin Mk1 to Mk are included in group k. MN is the molecumlar weight of last carbon group.

304

9. Application in Re.e,noir Si,,nu!alio,u

9. I. Grouping

305

equilibrium ratio (K-value) of the pseudo-component and those of the original components comprising it, A(KIKk) In tin) K (93) 1

using different grouping schemes. Several EOS were used to simulate the test results. The average absolute deviations of predicted lesuits by different methods are shown in Figures 9. I and 9.2. Note that the reduction of the number of groups down to an optimum value did not impair the predicted results. Minor improvements by grouping observed at sonne conditions are due o the cancellation of errors. It appears that describing the oil by 4-6 groups is sufficient to model gas addition processes. Note tisat lowering the number of groups below the optimum
value results in a drastic innpairment of the results. The above conclusions were independent of

where K~, is the apparent K-vaiue of the pseudo-component k. defined as, 2 K~=!~ __ ~xi
i)k)

the employed EOS and grouping nrethod [13].

(9.4)
a

C C

y and x are the mole fractions of component i in the vapour and liqunid plmmnscs mmt line salmuration 1 1 respectively, and i(k) denotes the components assigned to the pseudo-eomponment k. pressure, In the above proposed method, the mixture saturation pressure and the connpositions of equilibrated phases at the saturation point are calculated by the phase behaviour model, musing the full compositional description of the fluid. The components are then ordered mmccording to their K-values and grouped initially by the equal mole criterion. The grouping is then adjusted by moving components within the adjacent groups to minimise the objective fmnnsction A. The volatility of fluid components at high pressure conditionns, dcpenrds run line nmmixtnurc composition at a given temperature and pressmure. Altimougls time cquilihriunmn rmmtio varies witis the composition, the order of relative volatility of components rcnnainms (lie saumme. ilence, some component properties, such as the boiling point, critical temperature, molecular weight, or their combinations can be used to describe the relative volatility of components in mu mixture. An example ofsuch a trend was shown in Figure 3.8. A grouping method based on the concentration and the molecular weight, representing the volatility, of components has been proposed [13). The original components are arranged by the order of their normal boiling pointtemperatures and grouped togctlser in an mnsccnding order to form N~ groups so that the values of ~ z~ lnM~ for all the groups bcconsnc nearly equal. It can be expressed mathematically as,
N

0
C

4: Si,
V >

4:

Number of Groups

Figmnre 9.1. Effect of tIne nsunnsbcr of groups describing fluid on predicted results by tine Valderrama modification of Patel-Teja EOS, using grouping method of Eq.(9.5-6).
30

.
25
C 0

- ~

ttl,uck

Oil)B) & Memtsane

~-O
--

Black 011(13) & Carbon Duoxjde

Black ()it)A) & Rich Gas

~z InM~(k/Np)~e InMj 1 i(k) 1

k=l,2

Np

(9.5)

2))
0
V C C .0

~z~tnM
i(k)

(k/Np)~z1tnMj 0
i

k= I, 2

Np

(9.6)

where z and M~ are, respectively, the molar concentration and the molecular weight of 1 component i in the mixture, fully described by N components. The last component in the group k, would be either, ore + I, depending on which inequality, (9.5) or (9.6), is smaller, respectively. Methane, due to its high volatility in comparison witin other hydrocarbons, should not be grouped with others except nitrogen at low concentrations. A number of 25-component oil mixtures, Table 9.1, were subjected to single and multiple contact tests, simulating gas injection processes experimentally using different gases[ 13]. The fluids were described by various numbers of pseudo-components ranging from 2 to 25 groups

4:

Number of Groups

Figure 9.2. Effect of the number of groups describing fluid on predicted results by tire PengRobinson EOS, musing eqmual unole fraction grouping.

306

9. Applicaiiauu mu Re.orcoir Siuuuulalion

9.!. ~

307

A comparative study [13) of various grouping methods, using 4 or 6 groups, indicated a preference for the metlmod described by Eqs.(9.5-6) to that by equmal mole frmrclionm, with lirose of equal mass fraction and Li et al. [Ii) as the least reliable metimods.
Example 9.2 Describe the oil reported jim Table 9.4 by three pseudo-compoineunts for a numeilmanne guns

Equunl mole mmnetlsuxl: The total mole% of (irouup-lh anid (6roump-tll is equmal to 52.802%, with an objective value for eacim =52.802/2=26,401 . Addinug coruspounents from C downwards, Gronmp-l will 7 with 29.711 unrole%, in conmsist of C,-C,. wills 23.091 nsnole% and lime rest, C -nC, 4 0 Grouup-llt. 1~qunmilweight (mnrass) umnethod: TIne wemglnt of cads comnponneunt is calcnlaned as x,M,, with the total weight of C -nC,,, for 100 kguunotes of oil ~ s,M,= 4692.2 kg. lhc objective weiglnt of each 2group is =4692 2/2=2346. I . Aildimsg counmponeunms frousi C, downwards. Group-I will consisn of C,-nnC,, with a weiglnl of 2219.1 unmnd misc rest, nC, -usC with a weight of2473. I, in 7 75 Grouup-Itt. h~ituumuI xlnnM, nuseltuod: lIne vuitune for cacti consmpuunncuut us dmnlcimlancd. wutln tIne total valume for C,nC,,,, ~ x, mM, = 223.53. The ohitcljvc v;nlune of cacti gruumup =223.53/2=111.77. Adding components 0 unsu (7 downnwards. ( ronnp- I wull conmsn Si of ( nC,wntln ~ x , InM, = I I I .66 uunsd the rest, 7~ ui(,us(,,, winlu mu value iui 111.87. mum GriumupIll. Newley-Merrill nmielhod: Tine consposition of vapouir mit tIne bubble point umrunst be calculated using a tuned phase behaviour model, as suiggesled by the mmulhors. In this example, the measured comnnposilionm of the vapour, as reported in Table 9.4 is used instead. TIne conmnponenmts
of C,-mnC,,, are initially grouped inn two by equnal mole and K and A are cmulculumted unsiung t:qs.(9.4) mnusd (9.3), respectively. Tine first coursponent of Gronnpllt is

injection study am 373 K, using the methods of equal nsass. equnmnh numole. eqnu;mt ,,lusM, anud Newley-Merrill.

Table 9.1.
Compositions and properties of model fluids mit 373 K. nuuruponenn Black Black Votauile Vutauulc kuclu ( uuu,,1c9 011(A) 011(13) 011(A) 011(0)
~t mus

4680
8.77

C2
(3

36,14 12 17

74.17 5.32

73.33
5.35

69.82
I 5.0)

nC4

7.44 4.01

nC~
nC6

McI Cyct Pent

Cyct Hex

2.56 .77 2.25

2.20

8.05 5.13 t 4.79

4 (,7 2,58 0.97

4.71 2.62 1.00

11.1)1 5.99

3.81 1.43
.45

0.69 0.88
0.86 0. I 8

((.71
0.91 ((.89 0. 19

nC7
Mci Cyct Hex J,,luene

0.46
2.36 0.72 I .02
1.79

0.36
2.49

0.94
0.28 0.4 I

0.98
0.30 0 42

nCg oXylene
nC 9 nC~ nC 1 nC 12 nCI3

1.66 2.73 237 2.04 1.77

0.76 I .08 3.16 2.29 1.91 1.74 1.47

0.72
0.66 1.11 0.96 0.83 0.73

0.75
(1.70 1.17 .1)2 0.89 0.78

nakens innno (Jm,uui~II unnud line corresponding valnne of A is re-calculated for the new grouping, tmunlil tIne lowest value of A is unClnieVe(I. llse calculation results for tIme first and lIne Imusn few iteralionss unre slmowns in tIne following Tumble.
Coun u.. 1 C 2 (3
nC
11,0?

5.

K,

G.tm, A,

G-III, A, G-It. A, G-ltl, A, G-lt, A G-ttt, A, G-tt, A, (3-lit. A, .89574 0.220180.744140.1296II 0.73600 0.12357 0.72016 0.1 1514
0.07038 0.07471 0.08350

nCI4 SCIS nC 16 nCl7

nCtg nC g 1 nC2O nC24

1.53 1.34
1.15 099 0.87 0.75 t).6S -

1.32 1.22
((.95 0.85 0.67 0.57 1)48 5.03

0.63
0.56 0.4% 0.42 0.36 0.32 0.27 -

0.68
0.60 052 11.45 0.39 034 0.30 -

II.~I 81 .767 1.012132 0.75992 It 473 9 1)41 (1 713802 1.97594

7.059 4 34! 4 1) 61496

(1.00310
0.41211 1)04412 0.302711 (1.27036 (1.197280 77178 (1 t 4480 1.19S93 0.1036 1.67476 ((.101)76 1.70778 0.05687 2.47649 0.04446 2.77883

0.00436
0.03874 0.25335 0.73610 1.14637 1.61043 .64246 2.38868 2.68242

0.00742
0.02927 0.22184 0.66912 1.05297

nC~ uuCt, Meu CycI Peal

CycI tIes iC7

(.295 (1.634 11 48)S13 0.982 (1.389 039613 .297 (1.461 I) 35544

1.31)1 0.279 m 463 0.448 0.648 0 422 0 090 1)423 1) 125 0 174 0247 1) 32437 0.32258 0 28913 (1 279(12 (1.26852 022212 0.151(61 0 12427 0.10420 0.09762 0.06581 0.05402 (1.04626

1.48893
1.5 m 907 2.22241 2.49976

Mel (yci mmcx

i,,lncne nC 8 oXytcne

Sat. Pres., MPa Sat, Denns., ,g/cnnm

PreSsU!~MPa

20.28 15.43 32.78 32.70 0LQ~_~.0.)2_8....

Densuiy,,glcm 0.542 0.550 0.557 0.564 0.569 0.593 0.601 (1.192 0.217 0.237 0.255 0.271

.199 0.2154 0.22018

0,03240 0.00000
0.00008 0.03979 0.21343 0.59566 1.23910 1.57626 3.70233 6.00615 9.46080

20.79 24.23 27.68 31 . 1 3 34.58

n(9 unClO nCl u

~~t2

0.26767 3.03090 2.82908 0.16927 0.19254 5.15626

0.17352 0.08645 0.01946 0.00183 0.05942 0.10728 0.St997 .06303 1.95706 0.19686 0. t0683 0.03223 0.00003 0.03457 0.07065 0.41096 0.87721 I .65741 0.23198 0.13896 0.05546 0.00540 0.01103 0.03220 0.27994 0.647t6 1.27992

1.112

1.923 0 353 0.18357

1,73) 0.261 1.545 0 192 .382 0.t44 t.2t9 0.119 (1.956 006t 0.833 0 045 0.715 (1.034

0.613

0.390

0.399

nCt3 nC 14 nC 1~ n(t
nC ~ nC1g 1

1,0139 0.082 0.0751)1

Solution:

m4.1364
21.991% 31.0331
94. 12547

3.24556
5.51658 8,22194 32.40257

2.79316
4,80937 7.22300

2.21718
3.90131 5.93394 32.5 1408

Methane is selected as Group-I. particularly as tine grouping is for a nselhmunle iunjcction snudy and the rest is grouped into Iwo pseundo-componensts.

SCI9 nC2O
Toial A

((.646 0.025 0.1)3870 0.567 0019 0.03351

32.0 1332

308

9. Applncatioin jim Reservoir Sinuuilation

9.1. Grouping

309

The

lowest value ofA is achieved when components C,-nC~ are included in Group-Il, with Xylene-nC~ in Group-Ill. Group Properties

b~
ihkl ilk I

(9.14)

en, Several methods have hen proposed to calculate the properties of pseudo-compomments required in EOS[4,5,lO,l2,l5].

~
ilkl

~x x (c,~ 1 1

+ c ) / x~

1~

(9.15)

The most common method is molar averaging.

k ~ (in)

where a~, hk and ck, are the paratnieters of EOS for the pseudo-cotrmponent k. Time binary interaction parameters between the pseudo-components of k and q,. are determined from,
= ~

(9.7)

where xk is the mole fraction of group k in the mixture,

xii
ilk)

kq

(9.16)

(nh

(9.8)

9 represents the component property, such as the critical properties (Ta, P~,v~,Z..~),acentric

A comparison of time above methods did not indicate a clear preference for any of them (13]. The results depended on the selected EOS and the number of pseudo-components used to describe the fluid, probably due to the cancellation of errors, as inrprovements relative to those predicted using tIne full composition wer~also observed at some conditions. Grouping is commonly conducted according to the concentration of componemmts in the feed, oflen tine original reservoir flmuid. This approach is jmnstilned in predicting the single phase volumnnclric bclnavhntnr mnnd to sonnme extent the saturatioun pressure. It, Imowevcr, dcteriorales inn flasin calculations wisere tine distiii,ution of components in each pseudo-consponent will be different in the two plnases. Tinis can cause problems, particularly for gas condensate fluids

factor, or the molecular weight.

Pedersen et al. [15] suggested using the mass frmmction iimsteumd of tIne tmmolc frunctionn inn Eq.(9.7), whereas Wu and Batycky [10] proposed to calculate group properties by a connhinnation of mass

and mole concentrations. The Lee-Kesler averaging method for critical properties [16,17] has also been used in the
industry,
=

where the retrograde condensate phase is over predicted, as the properties of the heavy pseudo
comsnponents are those of tine original gas and not the fornimed liquid. Newley annd Merrill [12] suggested to calculate the critical properties and the acentric factor of pseud~-connnponentsusing a weiginting factor, based on splitting of the original fluid at its saturation point,

.!.~~xmxj(vJ +
8
i(k))(k)

(9.9)

T, 5
ZCk

j ~
cii
ilk) ;hk)

(9.10) (9.11)
(9.12)

0,

(9.17)
ilk)
(k)

0.2905 0.085

winere tIme weiglnting fmictor, q~,is defined as. 2 p =(x~y~)= 7.~(K)~ (9.18)

Z~RT~ / van,
acentric

The molecular weight and the method.

factor are calculated by molar averaging in the above

The calculated properties of pseudo-components arc not directly incorporated in immost EOS, but are used to calculate the parameters of EOS. hence, applyiumg mixing rules directly to tine parameters of original components to calculate the parameters of pseudo-counpouments seems a

where, x~, y, are tine predicted ussole fractions of connponent i, using the full description of the fluid, in time liquid and vapour at the saturation point, respectively, with the equilibrium ratio of K and zj is the mole fractioun of component i in the feed. The authors also suggested using the above weiglnting fmnctor, instead of time mole fraction, in Eq.(9.16) to calculate the binary immtermuct ion punrumnnnelcr.s between the psemndo-components. Ncwley mmmd Merrill 1121 connnpmmred Ilmeir proposed grouping metinod with equal urnole fraction, eqxnal mmmuuss fractious and tinmmt of Winitsons. lhey applied Eq.(9. 17 ) to tlneir own nuetisod and time itnolar averaging to otlners to estiunnate the properties of groups. The study for a lean gas condensate denmmonstraled time stuperiority of their method.

reasonable approach [13],

=

~,~x,x (i 1
ilk) jIkI

k1)(a,a1)2 /x~

(9.13)

310
E.eanuple 9.3

9. ..lpp!i,,ilii,iu I,, I,, ,,,,oi, ,5j,,i,,/,,ij,,i,

9. I. (rounping

311

Calculate the critical properties and acentric factor of Group-Il in Exausnplc 9.2 comprised of C,, C, and nC, Solution: The comnponent properties are read from Table A. I imm Appendix A and umiolmur averaged, using Eq.(9.7), as shown in the following table.
Componenl (2 (3
nC4

1mm K =co+c (l+wi)(l_1J 1 1

(9.19)

where K, (n) and Tr mire tine e.qruilihriunmm ratio, acentric factor and the reduced temperature respectively, and c,, aind c are constants. The above function, is a modified form of the 1 Wilson equation 18], Eq.(3.66).

Group-Il

nsolc% 11.61% 11473 7.059 30.15

s (1.38534 ((.38053 0.23413 1.18)000

Mx, It 587 16 780 13.608 41.975

Tx,, K t~,s, Mta

117.65
14(1.73

18.51
ISIS

7.5 11.11(8
11)05

(1(1181
(1(158(1

99.53 357 92

8.77 43.25

0.1164 (1.277

((.1(469 ((.1432
0

---a
em
3:

Tise method of Newly-Merrill is the samnne as line above, hint wilts tIne weighliiug factors calculated from Eq.(9. 18), insstead of nnole frunctions, uns follows.
Cousmponenst K, 1.01282 (178802

(K)s,
0.38289 0.42867

up,
0.34491 0.38615

T~, up. K
105.31 142.81

P~, up. MPa

6.58 16.19

Z~,up,

w~, p ((.0343 (1.0588 11.0538 (1.471)

.1
0 .01 ua On o0 i~ Contact Stage
02 03

C2
C3

0.096 (1.11)7

nC4
Group-lI

0.61496

0.29856
1.11012

0.26894
1.001(00

114.33
362.45

1(1.07
42.85

0.1(74
(1.276

000

Composition Retrieval
In some processes, suds as gas-oil displacement and gas cycliung. wlsere tIne relative cotncenntratiomm of fluid components within each gromnp varies significmsnmtly, the use of group properties generated frons the original commsposilmon umsay be iuma(Ieqlnmnle for mm unccurunle prediction of the phase behaviour. Table 9.2 shows tire variation of concentration ratio of some couiiponcnts of mu North Sea oil. when contacted with ten volumes of an injection gas in a backward multiple comitact test. For example, consider a group formed by the three eoussecutive single carbon nunnbers of 7, 8 and 9. The properties of the group using tIme original compositions will be equally weiglmted for the properties of C and C . Clearly such properties will not be representative of Ilnat group, after 7 9 contacting the gas, where the concentration of C in (he group is reduced to 42% of C , 7 9

.01 II

-l

0
(I+m~~)(Ih/Tr)

Figure 3.7. and 373K.

Equihibriuimn rmntios ins un test simulmntinng oil vaporisation by mettmane at 34.58 MPa

Table 9.2.
Molar concentration ratios of oil components befpre unn(I umfter cormtmicting gums. Molerauio C /C ClO/CI2 C /C 7 9 1~ 15 I .76 I .49 Original oil 1(10 After conuact 0.42 1.36 1.06

An inverse groupinsg ussetlsod. to retrieve the fluid description using Eq.(9.I9) has been developed 1131. Ins (isis procedure. equilihriumn rmmlios of time groups, obtumined by flash calculuutions, are used to detenisiimre the conslannts of the above equation by tIme least square imsetlnod. Tine equnalion is tiseum eunployed to calculate the equilibrium ratios of tlsc original consnponsents and retrievimng tine full description of both phases by material balance calculations, Eqs.(5.4-5). If fn.urtiser flmnsim cmmlcmnlmmliomss are required, time imnixture detailed compositional description can be mused lo fornnn unew gnotnps. Clearly tine above melirod would be practical only for rapid groimpiung umnetlnods, such as lisose descrnhed by Fq.(9.2) and Eqs.(9.5-6). Otlrerwise tine rclnicvmsl umund regrusupinng caur k(uule excessnvely Iinnse consunisiung. beating the purpose of the exercise. Time sinsulaled results of hydrocunrhon recovery by gas cycling at two different pressures a.re sisownm ims Figure 9.3. Note thmml tine conmposition retrieval improves the results even for systems described by mm large number of groups at the higiner pressure where phase compositions changes markedly. Time effect of compositional retrieval at the lower pressure becomes only significant when less thanr six groups are used to describe the fluid. Time conspuitationuml (CPu) timnne for simulating tIme above process can be reduced by 75% due to groupiumg mns shown ins Figure 9.4. It also exhibits tlmat tine full composition retrieval and regrouping do not siguiificumnhy increase tIne compulational time.

The above problem can be alleviated by retrieviung time fluid description ims lerms of line original components at some stages of cell to cell calculations in reservoir sinmulation and forming new groups aecordingly.
Figure 3.7 shows the equilibrium ratios measured in a multiple test where Black Oil (B), Table 9.1, was vaporised repeatedly 373 K. The variation of equilibrium ratio for each component mmnixture composition. Note that the log of equilibrium ratio function as, backward contact gas injection by tsmethane at 34.58 MPa and is only due to clmatnges in the can be expressed by a linear

312

9. Applucatio,u inn Rem nrnoir Si,unulat join

9.1. Grouping

313

Solution:

80 60
(a a, 0
U U

lhe critical teussperature and acenntric fmsctor of the three groups are calculated by molar averaging and Ihe coefficients of Eq.(9.19) are determined. Group I (nuicilnamue) Il (C,-nC ) 4 III (nC,-nC2O) T,, K 19(1,56 357.92 626.67 Tr 1.95739 1.04214 0.5952 I acentric facior 0.0115 0.1432 0.4861 (l+w)(l-IIF,) Ink 0.49474 0.920282 0.04623 -0.2181145 -1.01066 -2.894982

*
~

Winln Retr. 34.58 Mls without Reur. 34.58 MPa

40 20
g

+ en I-)

~~a Wiuhoui ReIr. 20.8 Mta

*--.-.-~-e 0 5 IS Numnunher of Grinups
10

wIth Reir. 20.8 MPa

lnK~=_0.334l6+2.534l(I+Wi)(l_~~~_) Smnhstimniing tIne muccintric factor and redunced tcrnnpcruntmurc of e~ncin conuponent in line above equnalion resu Ins inn mime cqniilibriuinn man ios of tine original coumnponemmts. lIme conmnpiscncnmt huiluuumce fur emuch connnponent results inn,
7.

0 21) 25

Figure 9.3. Effect of composition retrieval on predicted recovery of C . . fronms Volatile Oil A by 31

methane cycling at 373 K.

Ix,

nL +(t nL)K

Substituting nu. =0.56239 in the above.

7.,/X~

=0.56239+0.43761K,

the commmposition of equilibrated phases are cumlculated as shown in the following table.
0 C 0

a,,

U
U

I-.
0.

~~~~ wimh Reir. 20.8 MPa

5
10

wiihouni Reir. 20.8 MPa

Is Number au Groups

20

25

Connponent. mole 5 C C1 2 C 1 nC 4 nC nC5 6 Met Cyct Pent Cyct flex nuC 7 Cyct lies Met riuluene
nCg o-Xyteume nC 9

(t+w)(l-tIJ,)

0.49474
(1.19950 (1.01)979 -(1.16771 -(1.32445

K, 2.508 19
1.18696 0.73393

x, 0.28657
0,10824 (1.13087

y 1 0.70604
0.12620 0.09435

-0.46958 -0.52701
.0.511654

0.46807 0.31463 0.21781 0.18831

0.16 194

0.09273 0.01864 0.01505 0.02027

0.0207 I

0.04264
0.00576

0.00322 0.00375
0.00329

.1160492 -11.65952 -(1.74148

-1)73432 -0.91)576 -(1.85759 -((.97900 -1.1)9131

0.15457 0.13460 0.10936

0.11136 0.072 12 0.0814%

0.00446 0.02373
0.00740 0.01069 0.02035

0.00068 0.00314
0.00079 0.00117 0.00144

Figure 9.4. Computational time of simulating nmiethane cycling.

Example 9.4. The oil reported in Example 9.2 was flashed mmm the reservoir temperature of 373.0 K. The oil was described by methane and two component groups, using the equal zlnM isnethod and molar averaged properties, as given in Example 9.3. The predicted results by a phase behaviour model, using the above fluid description, are given in the following table. Calculate the composition of equilibrated phases in terms of the original components.

0.01874
0.03293 0.03001

0.00150
0.00194 0.00133

nC

nC10 11

0.05991)
0.04506

Group
l(methamme) II (C -nC ) 2 4

Feed Comp.
0.47198

Oil Comp. mole fraction

0.28419

Gas Comnmp. 0.71332 0.26533 0.02135

0.30150

llt(nC -nC2O) 0.22652 5 Liquid mole fracnion=0.56239

0.32964 0.38616

Equilibrium Ratio 2.5100 0.8049 0.0553

nC nC12 nC11 . 14 nC~ nC 5 nC16 17 cC 18 nCi nmC

9

-l.2t)449 -1.30953 -1.40954

-1.51451

0.03383 0.02592 0.02012

0.01542

0.02698 0.02428 0.02151

0.01929

0.00090 0.00062
0.00043 0.00029

-l.(,115o -1.72225 -1.81626

-1.91179

0.01206 0.00911 0.00718

0.00563

0.01697 0.01482 0.01310

0.01153

20

-2.1)1937

0.00429

0.01013

0.00021) 0.00013 0.00009 0.00006 0.00004

314 9.2

9. ,4pp/ucoli,nu in Ri,s,,oij; ,Sipnu,hutj,,,u

9,2, (o,mupari.smi of IA)S

315

COMPARISON OF EOS

(3) An sonne snages nhc pressure was lowered below nhe bubble poiunt and Ihe properties of nbc equilntmraled phases (e) were nicasured.

The capabilities of tine equations presented in Section 4.2 and many other van der Waals type equations, have been evaluated by several investigators. The studues have resulted unainly in a general conclusion that none of them can be singled out as the nmnost stuperior equmation to best predict all properties at all conditions. A number of conmpmnrativc sltmdies 19-221 Imunve, however, shown Ihat certain equations exinihit a higiner overall muccunracy. The accuracy of predicted results by any EOS depends not only oun lIne nclimmhihity of thmml equation, but also on the mixing rules applied to its parameters, fluid clnaracterisation, estimated critical properties of the fluid components. etc. Therefore, proper considermmtioun of all pertinemst factors when evaluating EOS is essential. For hydrocarbon fluids, line rmnmmdonnn usuixing mimIcs. described in Section 4.3, are considered adequmite. Tinese nriximsg mimics hmmuve bceun used in almost all the comparative studies of EOS. Any phase behaviour model is hotnnd to carry errors intro(htnced by inmmnccumnmmIc imnptnl (Immlun, particularly the properties of single and multiple carbon groups, inmto tine predicted rcsunims. lIne problem can be alleviated, in comparative studies of FOS, by unslung dala 11mm umnodcl fluids, instead of real reservoir fluids. The use of many componenrts with remnlistic c(utsceuntrumlionms, such as replacing eaelm single carbon group with its ptnre cqunivmmlent its tIne nun(sdcl fluid, dunn produce reliable data witin conclusions applicable to remnl fluids. The performance of a number of leading EOS. namely the Zumdkevitcin-Joffc umsodiincmition of Redlich-Kwong equation (ZJRK) [23], the Soave-Redlich-Kwong equmation (SRK),124J mmumd its three-parameter form with the volume shift (SRK3) 1251, the Penmg-Rohinsous eqdnation (PR) (26] and its three-parameler form with the volumnme shift (PR3) 1271, the Sclsnnnielt-Wcnzel equation (SW) [28] , tine Patel-Teja equation (PT) [291 annd its nnodilmcation by Vmsldcrranra (VPF) [30], are compared in this section, Tinese eqinations were described in Sedlion 4.2. Tine above equations have been selected either because tlmey are widely rused in time indiustry, or they have been shown to be reliable in reported conimparative studies [19, 22]. Figures 9.5 and 9.6 compare deviations of predicted saturation pressure and smmturumtiomn volunne from experimental data, respectively, ium a swelling experinment wlnere Rich Gas was progressively added to Black Oil (A), botim described in Table 9. I. The experinnental datum arc given in Table 9.3. TIre results indicate thaI each EOS could be nnore successful dependiumg on tIne compositional range. Hence, in a general cotsipmmrative study, the evakiation should cover wide ranges of composition and temperature. Time experiummermlmml data sisouuhd be obtained preferably in tests simsrulating various pertinent reservoir process, such as nmuubtiple contacts and gas cycling for gas injection studies, in addition to conventional PVT tests. Table 9.3. Experimsncntal data on umddition of Rich Gas to Black Oil (A) at 373 K.
~ Phase State V d ,Cin 5 4 t ,MPa 1 .cm V 5 P~.MPa Oil Gas 1(11.73 9.85 2)5% II)) .16
.

IS

SRI< lR

1:

:...:;::

~
A
A

::

en

At
A A A

vpr
ZJRK PT

It) 0 I

Mole Injected Rich Gas/Black Oil A Figure 9.5. Cousmparisomr of errors in predicting saturation pressure at 373 K by various EOS. IS.

5)

E > em
C

ho

tR3 PT O SW SRK3
O

A A 0

VPT

ZJRK

cc, C
0

S
U

0 0~

a
a
0~9

em
5)

1)

0 5

Oil

Gas

Oil

Gas

101.73 34.9)) 1(11.73 (14.9(1 23.61 2463 134.40 163.43 20.57

Oil Gas Oil u;~. 5) 07 47 58 51(17 62.58 24.80 24.81 11.1.7) 97.64 24.58 24.69

-s
(I 2 Mole Inujected Rich Gas/Mole Black Oil A 3

Vc,cunui
Pe.MPa1 V~.c[ns
P~.MPa

11)9.96 36.36

83.72

14.61 85.93 28.29 Figunre 9.6. Connmpumrisomn of errors in predicting saturation volume at 373 K by various EOS, llie evaluations of capumbilities of van der Waals type EOS for reservoir studies is of a more inlerest in tlnis sectiomn, than sclectinmg a particular equation. The presented data are given as typicmml exumummples on the perfornsnmmmrce of Ihese equalions as reported in [19, 22, 31], where nnany fmunndrcds of dmntui poinls on various nrsodel and real reservoir fluids, generated at simulated

V~cisu

Eq.Dcns., g/cuss

24.34 24.04 77.34 21.61 (~7.8l 48.36 25.1%) 61.52 57 14 0.464 0.264

(I) Added oil and gas volumes, V dd, were measured xi 20.79 and 31.13 Mls. rcspccuive)y. 5 (2) Ph and Vh are ulse nsisuure bubble point pressure and v,,tumc, respectively.

316 9. Application ui Re.aersoir Sipnudaiion reservoir conditions, were used in the comparative studies. Otimer examples can be found in the
literature [19-20].

9.2. Comparison of LOS

317
10

Exp.

Phase Composition
The predicted composition of equilibrated phases is not only important in deternnmining the phase behaviour in subsequent flash calculations, but also in predictimmg otlner properties such as the viscosity, interfacial tension and the density. All the above leading EOS, gemmerally predict the composition satisfactorily. For example, the composition of equilibrated plnmuses in tIme first 3 3 contact of a test, where 120 cm of Rich Gas wums added to 60 cmn of Black Oil (A) mtt 20.79 MPa and 373 K, are given in Table 9.4. TIne equilibrium ratios mnt lime muhove conndilionms predicted by various EOS are compared witin time expcrinncnrlai clmilmi inn Figure i)], Note Ilnat all the equations predict the equilibrium ratios similarly. Figure 9.8 imiglnligints line reluntive errors of predicted equilibrium ratios in the same test, where each connmponemmt inmus lmecmm idetmlilied by its reduced temperature. The concentration of ineavy consnponenmls in tine vapour pimmuse is relatively low resulting in a high relative error band experinnentmmlly. This cans produce mm large deviation between the measured and calculated values. Neverilneless, tine percemmtage deviations of the predicted equilibrium ratio by.all the equations, increases systennatically, positive or negative depending on EOS, for heavier compounds. This trend slnould be expected for a number of reasons. The heavier a compound is, the furtimer its behaviour deviates from tlmat of a compound with simple spherical molecules, on which EOS models are based. Fumtimernnore, the parameters of EOS, particularly the attractive term, have been correlated using vapour pressure data biased towards the light components as described in Section 4.2.3.
0 1

E
as

- SRK PR . SW p~ VIS
-.

ZJRI(

.05 -1 (I +w)( I lrrr)

Figure 9.7.

Consparison of predicted equilibriunm ratios by various EOS witim experinmnenlal 60 a PT SRK O PR

O S

data.

Table 9.4. Composition of eguilibrated phases at 20.79 MPa and 373 K.

Component, mole %

Cm C 2 C 5 cC 4 nC 5 6 Cyct Pent Met Cyct Hex nC 7 Cyct Flex Met Toluene nCg
o-Xyteune

Oil 47.198 11.618 11.473 7.059

Gas

nC

1.295
0.982 1.297 1.301 0.279 1.463 0.448 0.648
1.199

70.287 11.767 9.041 4.341 0.634 0.389 0.461 0.422 0.090 0.423 0.125 0.174
0.264

40
0.
U

en

SW 7.JRK

a vii~
2t)

0.
ad
5)

0.

A- .~-0

t)
q ti~.

nC

9 nC~

nCmu nCi 2

nCun

nCm 4 nC~ nCi 6 nCii

nCug

1.112 1.923 1.733 1.545 1.382 1.219 1.089 0.956 0.833

0.735

0.247 0.353 0.261 0.192 0.144 0.119 0.082

0.061

0.4

0.8

1.2

1.6

2.0

Redmnced Tenmperature Figure 9.8. Comsnpmmrison of em-rons in predicting equilibritum ratios by various EQS. TIme average absolunte deviation of predicted equilibrium ratio by EOS, for a large number of commnpositional data 122] ame slnown in Table 9.5. The comparison of average percentage deviation of predicted equihihriummnn ratios can be misleading, as tlmey are strongly influenced by the lmsrge values of lneumvy counmponents with tow equmiiihriummn ratios. A nmore unseftul commmparison is thmmt of the average error in predicted composition, as also shown in Table 9.5. The deviuntiouns of predidled compositions are quite acceptable and ins most cases comparable with error bands of experinniental data.

0.045
0.034 0.025

flCn 9 nC 20 3 Equt. Vol., cm tm

E~ui.Dens., g/cm

0.646 0.567 63.06

0.4939

0.019 110.60
0.2238

318

9. App(ueotio,i

in

Rise, ,oir Sunuualao,,,i

9.2. (onupari.son of ADS

319

Ahmed [19] compared the performance of eight EOS for predicting lIne pinmmse belniviotur of ten real gas condensate systems after matclming the dew point of cuiclr fluid by adjuslinmg tine interaction parameter between methane and lIme plums fractioun. All line mubove lcmmdimsg EQS predicted the concentration of major componmenls of tise vapour pisase un counstannl voltunune depletion tests within a deviation of 2%.

Conrsidering Ihe reliability of EOS for prediction of equilibrium ratios of light consponents. tIme sumccess of these equmunlions in predicling tIme bubble point is expected, as the bubble point jsrcssnnre is nununiunly counlnulled by lise lrehimuviounr (mf light colsnpounermts. IligIn deviations in rcdiclcd cquiliimriuunum mat iuss oh lscmmvy conunponneusls should, imowever, lead In tnnreliahle 1 estimmnmulmon oh the dew point. TIse deviumlionns of predicted dew point pressure frommm experimenmtal data, by time leading EOS. can exceed 20% [31], eveur for well defined synthetic model fluids. TIne deviation can be nsuch lniglmcr for real fluids due to tIne presence of very large nsolecules, which strongly affect the dew poinsl, evens at low concen(ratiours. ilne behaviour of Ilsese compoumnnds are not ommly difficult to mnmodel by E()S, hut tiscir idcntiimcuntiotn anmci characterismution are also quite demaunding. Ahmed [19) evumlunaled line relimuhility of lemiding EQS 10 predict the dew point of a nunniber of gas conilensunle mixtures in swelling tests. Although tIne dew points of original fluids were initially nsnatclnccl by tuniung EA)S, tIme devimitionm of predicted values after adding gas exceeded 30%. Density lablc 9.7 slsows the mivermuge devuustioni of predidled saturation volume and liquid and gas dcmtsilics in vmmrionns gums injeclions teds )22). A sinnihmtr uncctnraey is expected for gas condensate systcmsms. Note tlsmul ins (sr(lcr to produce rehimmhle denusity of equilibrated phases, EQS slmould predict hotim lIne phumse cousspositiomm mnnnd time nmnuslmmr volunnme of a fltmid with known cotmrposition reliably. All tIne 3-parmnmsneter equimumions. wlscre tIme timird parusnsmelcr is included for improving density data, mis discussed inn Sectioms 4.2.2, mire moore uchimible tluusum tIne 2-paranmueter equations. line exception is 1.1 R K. A ltinommgiu it is mu I am u Pammmnmndter equmit ion, it uses (Ienmsity datum to delenninne EQS parmnnnidters. Devimmtiotss, sup to 25%, were nmoticcd using SRK. hut tIme inclusion of the voluune trmnmrslmitiotm, SRK3, emsltmmmnccd its cmnpmmbility to one of tine leading equations [22]. Fable 9.7.
Avermige msbsoltule devimutiouns of predicted liquid smmturation volunme,

Table 9.5.
Average absolute deviations of predicted equilibrium ratio and composition from experimnmental.
Equit. ratio Equation % OnI mokfracuon Gas

SRK SRK3 PR
PR 3 SW

16.27 16.45 18.35

18.56 17.17

0(1021
0.0021

0(1018
0(8)17

(1.0019
0.0019 0.0020 0.1)02(1

0.0011
0.0011 0.0014 0.0012 0.0011

PT VPT 7JRK

22.70 21.72 14.61

0.0014 0.0014

0.0012

Considering typical errors no measurung compositions in tests suclm as colmstmnmrt volunsme depletion, differential liberatnon and gas cycling, the predicted phuuse compositions by EQS for properly characterised fluids could he as reliable as line cxpcrimsscnnlal dmita. 1Iuc errors assocumsted wulin nrueasum-ed conspositional datum of equilibrumled phases uising poor Practices wend described inn Clnmupter 2. In suds cases nt ms probmmbly nnorc bcunelicimnl to chiicct lIme ellorl towmsrds characterising lhe origunal fluid aund genermnting reliable PV1 dmmta, and (men muse mi tuned phase behaviour unodel to prcdmct the produced fluid composition..

Saturation Pressure
Table 9.6 lists the average deviation of predmclecl bubble point pressure for mm vmiricty of oil satnmples. imscluding Ilnose with non-lnydrocarboun gases added to tlmenus in swclliusg tests 1221. The equations are generuslly capable of predictung tisc bubble poimml pressure willninu 5% devimitioum for hydrocarbon systems over the whole range of phase envelope itncluding nemur critical condmtions. The deviuntuons are generally iniglmer for flunids witlu isigim concentrumliorns of non-imydrocarbon gases. The VET and ZJRK appear to be overall tniore accurmmlc than (lOners, with a deviation of about 2%. TIne deviation of predicted values by VII for CO ricin syslenns 2 us relatuvely high . It should be noled tlsat no binary inleraction pmlrmsusselcr was used in VPF. All EQS generally requmre bunary unteraction paransselers for hydrocmmrhorm-C0 . 2 Table 9.6.
Average mmbsolute d evialion s of predicted saturation pressures.

munich gmms mmnnd liqtuid densities mit equulibruuumn.

Smulnirmnui(,uu Voluuuuuu I .i Iu~u~ I )eunsily ( mis l)encity

I Suiat ion

Average At usotulc I )cviaui,un 8 6.9) 3.34 16(i) 4.55 10)16

SRK
SRK3 PR IR 3
SW tt VPF

6.73
4.83 4.57 3.44 2.45

8.19
5.94
685 3.53

6.78 2.61
2.31 3.71 2.44 3,33

7JRK

2.3)

2.80 2,81

2.18

tnj Gas Flybocmnrbon N2 C02 Overall E~iaii,,n Average Absoluue Devnalion %

SRK

thse gums deunsity is gcmncrally predicted nmnore reliably tlman that of the liquid by two-parameter

SRK3 PR
IR3 SW

4,77 477

4,22
4.22 3)3

1205 12.05 2.29 12 29 9.91)

11.02 11.02 4.80 4.80 2.77

6.16 6.16 4.35 4.35 3.05

equnmnliouns. TIme perfornnnmsmmce is eqtmaily well for both phases witlm thnee-paramelcr equations. Ilnis is unuoslly dume tin time use of smihuraled liquid detssity data in correlating time third parameter in Ilucsc EQS. It is not summustimml to himid tlmree-paratmrctcr EQS predicting the liquid density more rehiumbly (humus tismit of its cquilihrmited gas. particularly for gas condensate systems 1261. As tIne tlnird pmsranmnelcr imums been generally correlated using saturated liquid volumnies. EOS may predlucl erronneotns deunsity for highly tinder-saturated liquids. It is advisable to calculate tIne smitsnrmmted liquid demnsity by EQS amsd then adjust it for conrmpression due to the excess pressure mmhove tine bubble poinrh. Ihc isotinermal compressibility coefficient, described in Section 2.3, emits he used to estimssale tIme incrcunse in liquid density by pressure. Alternatively, empirical

PT
VPT ZJRK

6.52
1.04 2.57

362
1.91 1 39

23.87
8.20 2.36

10.38
2.63 2.52

320

9. Appheaiioun in Re.ceruojr Sunmm,I~mlin,,n

9.2. conuporm.soun of LOS

321

methods of estimating oil density, presented in Section 2.3, may be used to calculate the

density of under-saturated liquids. Gas and Liquid Volumes The errors involved in predicting phase composition and density are couinhimmed to nmake tIne
calculated phase volume in flash calculations as lime lcasl relimubic predicted imnfiunmnmntionm by almost all EOS. Table 9.8 demonstrates the average absolute devimmtion of predicted gas anmd liquid volumes for a large number of data generated in various simulated gas injection processes [22]. Note that the error in predicted phase volume ratio is the higimest in all the cases. EOS generally predict tine total volume more reliably than the volume of each phase at equilibrium. Hence, an over estimation of one phase is generally acconnpanmied by under estimation of the other, resulting in a large deviation in predicted phase volume ratio.

Therefore, the overall error in a multiple contact sinnulation is expected tobe significantly higher than that in flash calculations of the original fluid.

Table 9.9.
First contact data of tmmettmane-Voiatile Oil (A) at 373 K and 34.58 MPa. (las Conimponent, Mi~% - - Di
(I 72.266 811.887

C2 5

4.479
4.1)75

4.292
3.721) 2,007 ((.742

nC 4 n(
unC,,

2.398 (1.958
1)712 11,944 (1.948 I). 199

11.515
((.650 (1.627

Mcm CydI teumm

(yet hex cC,

Table 9.8.
Average absolute deviations of predicted volumes at equilibrium. Single Coniact
Gas Liquid GaslLiquid Gas

Met (ycl tIcs

luuluicuue

1.11611
(1.325

11.135 11.676
0.204 11,293

Mutniple Conumuct I.iqmuid GaalLiquuid 32.91 35.11 32.50 33.29

32.67

n( 5 ii Xytenc
cC 5 usC,,, nC,~ nnCi 2 nnC 15 nC, 4 nC~ 5 nnC, 6

(1.462
((.876

(1.501

Equation SRK SRK3 PR

Average Absoluie Devimutiiun % 37.49 16.47 20.12 27.10 30.24 25.68 46.25 41.05 36.40 30.75
19.71

PR3
SW P1
VPT

14.18
11.59 9.68 11.51 4.54

24.70
14.65 24.27 11.19 10.31

41.41 31.56 32,21 31.84

29.13

48.59 46.77 43.64 44.64

42.06

4)791) 1.374 1.242 1.117

1.014

((.47I (1.767 11.651

11.54 8
(1.470 t).402 (1.348

0.918 1)830
0.744

ZJRK

44.77 27.77 14.02

11.53 14.84 21.95

10.71 17.81 25.96

5.53 19.94 32.38

nC 17 nC,g
fl(iq

0.661

0.594
(1.533

0.295 0.249 (1.213

1)181

nC2u)

(1.481) 80(10 17.02

The error in predicting phase volume increases sharply when the critical point is approached. The results of first contact between Volatile Oil(A), Table 9.1 aumd nrethanre at 373 K and 34.58 MPa are shown in Table 9.9. Note the severe usmass exchange between the pinmuses wlmere the oil/gas volume ratio of 4 prior to the contact changed to 0.2 at equihibraim, with ahunrost no change in the total volume. The deviations of predicted equilibriums voltnnmes by vumriotus leading EOS are shown in Table 9.10. Note that errors of over 100% are qtiile conmsmnnoun. Smcim higln errors near critical conditions should not be surprisiumg as a snmnall prcssn.nrc rcdticmioun below tine bubble point can vaporise almost lmatf the liquid volume. hience, for examnnple mu I 00% error in predicted gas/liquid ratio could be equivalent to only an error of less than 0. I % in the predicted bubble point pressure. Whilst such an error in predicting the saturation pressure by EQS is highly encouraging, its effect on the volume ratio is totally unacceplable. Time inmiprovement in predicting phase ratio, by inclusion of the near critical density correction. Sectioum 4.2. I, has been found to be negligible [32]. Whereas tuning of EOS to experinmental data, generated in the critical region, can significantly improve the results. In compositional reservoir simulation, where the reservoir is described by a number of equilibrium cells, the predicted results in each cell provide the input data for time neiginbouring cells in the flow direction. This generally results in compounding errors. Figure 9.9 shows the deviation of predicted phase volume by various EOS at the front of a forward moving gas in a reservoir described by four cells. Although VET was found to be reliable for the original oil, it resulted in a significant deviation of the predicted gas/oil volume ratio in the final stage.

Va~~.dunn 5 V~,cuns
Eq.l)euns.,

11.154 20.01)
81 .90

5/dunn

0.4327

0.3310

lmmlrlc 9. It). lercenmtmmge error inn predicting pinase voluusmc by various EOS. EOS ZJRK SRK SRK3 PR PR3 SW Oil Vol. -118 -182 -152 -126 -126 -135 Gas Vail. 26 28 31 27 25 26

Pf

.53 9

VPT .51)

II

lIne cmmpmihility of EQS in predictiung the plmase volumnne of gas condensate sysmeumus. particularly witlnins tIme retrogrmsdc region, is generally inferior to that of gas-oil systems. Such a beimaviour is expected a~tlme mmccuracy in munodlelhng the behaviour of heavy compounds, which dominale time lic umid fornmation, is generally insferior. The volumetric behaviour of a gas condensate, with 1 tIne commnpositiomm givems in Figtmre 4.6, as predictect by several EQS is shown in Iigtnre 9.10. Note Ilnat tire predictions of all EQS approach tine experimental values quite closely within tine vaporisiusg regioum, where tine systenn behaves oil-like. Time error in predicting time retrograde liquid volume below time dew point can be redumccd uinarkedly by tuning EQS to tsmatch the dew poitst.

322
20

9,

ppli( (ltiOii

jul

Reseii,,ii .Si,,uuihuijo,n

9.2. ( olnparisoiu of LOS

323

0
V

E
>
0

. -20 a

O ~4O

o
0

solutions. All cubic equations generally exhibit a similar convergence behaviour, they either canunot easily converge, or converge to the trivial solution of equal partitioning of components between the phases, at conditions close to the critical poinl. Convergence problems may be observed unlso at conmclitions near Ihte rnumxinnsunm pressure and time nmnaximum tennpcrature of the plmase boundary. EQS which locate time prevailing conditions away from the difficult conditions, such as time critical point, often converge, wlmilst others may fail. Hence, the converged equatiomss are not necessarily more reliable, or applicable to that system.

vmr
-60 81)
.5W 0

o 0

7JRK PR)

S
0

9.3

lUNIN(; OF FA)S

too

SKK3

Black Oil A/Rich Gas, Mole Ratio

lhc inmhucrcnnl deficiencies of F.()S, lsmirlictmlmlrly for nmntnlticonnnponcnt nnixtures, were (IeSCrihC(I in Inmulster 4. tlmmmsc hclumuvionnr nmiode Is bmmscd nun these cilumml iumnms nmmmmy predict Iniglnhy erroneous mestulls, puurlictuiumrly for unear cm it icmnl I Ituids. cvcnm for well clmunracterised nmodcl fluids uns shown in Scctiomn 9.2. Real reservoir fluids, comniposcd of Ihousmtnmds of connnpounds, are described by a limssiled nuunrbcr of pure subslatnccs umn(I carbon groups. Time compositional analysis of these fluids mire not umiways very reliable mmmd tIne carbons groups are not fully defined. Gencrahised correlations, often with signilncumntIy diverging restults mnmongst thenmselves, are used to estimate tIne crilical properties of (lie carbon groups required for EQS calculations. All these factors further deteriorate predictiomis of EQS for real reservoir fluids. The current approach in the industry to enmcounter tIne above deficiencies is to calibrate, or ttnne, mm EOS nnodeh agmninrst experinncmmtmul dumla generated mit pertinent conditions for specific field sttmchies. Ahtbmoungh time industry Imums not adapted a single standard method for tuning, the various approumchmes are hmnsicahly siuinilmtr; sommse uncertain values of input data to the phase behaviour tsnodel are adjusted In nsminimmnise tIne difference between tIne predicted and measured values. As the model is to prcchicl tine pimumse belnaviour and various fluid properties within wide ranges in conmsposilionnmml reservoir sinnnulumtion. mm Immrge nunnber of experinsmental data are often used in ttmning. The exercise is basically to rmnininsiise aim objective function, defined as the sum of weiglmied squmarecl devmalionss.
N,,~,

Figure 9.9. Error in predicting phase ratio by variotus EQS at frount of Rich (imns mudvanscinmg in Black Oil (A) at 373 K and 20.8 MPa.

16

E
>
0

12

08
V

(WNOn(X

)v~

ia
C U ~0
C

A= ~

wj~

(9.20)

04 (-3

where cunchu elensmenmt of Ilse ol~ective fanunclion expresses tine weiglmted difference between tire prcclicled aund experinnmeuntmml vmslues, mFIpreil mind lftxP. respectively; w is the weighting factor and Nalalmu expresses Ilne mnuusnber of nsnemnsured data points t(n be fitted; X~ designates tIme regression (Iunncdl) vmmriables.

0 10 15 20 25 30 35
Pressure, MPa Figure 9.10. Comparison of predncted condensate/gas volutmietric ratio in a constant composition expansion test at 383 K by various EQS.

Tine optirmrunnn values of variables are obtaitned by nminuimnising time function A. Altlmoughs various mmnethods [33-35] isave been suggested for solving multi-variable regression problems, none can he gtmmsranlied to scrlve tIne problem in all cases. A nmodification of the Levetnberg-Marquardt nnethmod 1351 is often used to minimise the value of A in Eq.(9.20). The innportanmce of mm property is ensnplmasised by multiplying its deviation with a high weighting fumctor. TIne satuuration pressure is perimaps the most innportant property of a reservoir fluid for pimase behavioumr sludies. Furthernmiore, it needs a high weighting factor if it is to beconme effective, as generally the nutnber of data points on the fluid saturation pressure is much fewer than tlnose of otiner properties. Iligln weighting factors could also be assigned to more reliable experinnental data. l..ess reliable data, sucin as the consmposition of equilibrated phases, should receive low weiglmtinmg factors, or preferably not rmsed at all. Table 9.11 provides weighting factors as a rouglm gunide 133].

Robustness
The robustness of a phase behaviour model in converging to a solution is more dependent on factors such as the matlmematical metinods of solving lhe governing eqtmuntions and inilial guesses tused in iterations, than EQS characteristics. The pertinent comiditions of composition, temperature and pressure cain also have profound effects on lime convergence of EQS to a

324 Table 9.11. Weighting factors of properties in tuning ofEQS. Property Bubble Point Density Volume Weighting Factor 40 20 10

9. App!iearion in R~.senpir Sinmuda,ion

9.3. Tuning ofLOS

325

Composition I

Although the dew point is an important parameter and its accurate prediction is desirable, assigning a high weighting factor to it may increase time deviation of predicted retrograde condensate volume. Many gas condensate samples show a liquid drop-out tail during depletion, as described in Section 2.2.4. Matching tine dew poinl gencrumhly restmlts in over prediction of the liquid volume during the early depletioms stages for these fluids. As tine measurement of dew point is quite subjective, tuning EQS with a higlmerenmmpim~isison the liquid volume, instead of the dew point, is preferred. The deviation between the predicted and experimental data is not only due to deficiencies of EOS, but mostly due to the input data, excluding the near critical conditionms. llennce, tIne luninmg process should primarily be conducted to evaluate and innprove tIne inrput (luila, inmslcmud of modifying EOS parameters indiscriminately to mnmatch the experinnentuml data. (icumerumlly a severe tuning could indicate overlooked problems. Furthermore, the tuning simould unot be regarded purely as a mathematical regression problem. Time parameters to he regressed need to be selected based on physical concepts and varied within reasonable limits. Fluid Characterisation

The eharacterisation of single and tmmultiple carbon groups has a major innpaet on the results predicted by EOS. Guidelines on time tntumber and selection of groups and the estimation of group properties were given in Section 9.1. Occasionally, insproper characterisation, such as describing the heavy fraction witlm too few pseudo components, may lead to lower deviations of predicted results tinan timat of an appropriate method. This can be due to the cancellation of errors at some conditions and should not be adopted. It is more logical to use proper characterisation and then attempt to improve other shortcomings, than relying on uncontrollable cancellation of errors. Tuniumg of [OS can be conducted with the fluid described by any nmunmber of conlmponennts. In genermul describing mine C with 4 groups, using tine quadrature tsmelhod. and mill line discrele connspotmnds mis reported, 7~ slnould be mmdcquate in most counpositional models. Time eomunponetnts witlm their optimised properties could be grouped again to reduce their nuimrher, if required. An additional tninor tuning of the new group properties, depending on the grouping method, may be necessary. Selection of EOS Forcing EOS to match certain datms by excessive adjustnnent of its paranrmeters, nnay lead to imigimly unreliable infommation at other conditions winere experinnental data is lacking. In general any leading EOS whicln predicts phase behaviour data reasonably well without tuning, would be tlme tnost appropriate choice. Tuning sinould not he comndiucted without considerations to capabilities of EOS. For example. tunning of a two-paraumneler EOS. kumown to be weak inn predicting tine liquid density, to nnatch experimemn(aI density dlmmta, may lemmd to serious problenms with prediclion of otimer data, or even lurtiner deterioratioun of predicted deunsity outside the raisge of available experiusnental (laIn. Aithougim usnost of tIme vami der Waumls type EQS are basically very siusnilar. certain equations may he preferable to others. As reliable volumninetric data are also required in reservoir fluid studies, three-parameter EOS simould he selected in preference to the two-parameter equations. Certain equations, such as the Valderrama nnodifieation of Patel-Teja EQS. which have consistently demonstrated their reliability, cotmid be considered annongst the first choices. Experinaental l)ata

A proper analysis and

characterisation of the reservoir fluid is tIme Inmost insnportanl step in successful application of a compositional model to determine the fluid belnaviour and properties. Pedersen et al. [36] provide many examples, where a proper characterisation of real reservoir fluids has resulted in reliable predictions by phase beimmuviour ummodels without any tuning.

Comparative studies, where the same fluid Inas been sent to diffeneust Imuhormllories for compositional analysis, have revealed striking inforusmalion on tine dispmmrily of results, particularly forgas condensate systems. The use of high pressure analysis techniques, to avoid loss of compounds in the flash (blow down) method, Section 2.2, is recomnnended. This is particularly valuable for gas condensate fluids, where tine flash mnetinod results in an analysis often leaner than the real fluid due to the loss of collected condensate. When compositional data generated by different metlmods, such as distillation, gas chromatography, high pressure analysis and mass spectromsretry, are available, liney all should be used to determine the most probable composition of the fluid. The capabilities of each method should be considered in driving the final analysis, rather than averaging the reported compositions. For example, the most reliable information on the relative concentration of light components is obtained by gas chromatography of the flashed gas, whereas distillation provides reliable data on heavy components, particularly the plus fraction. The high pressure compositional data on intermediates and lighter heavies are generally more rehimnble than those by other methods. The high pressure analysis also provides valuable infonnation for evaluating the reliability of the vapour to liquid ratio used in deriving tine overall composition by the flash method. The concentration of components is alnmnost always measunred in munass (or volumnmre) basis in distillation and also in gas chromatograpimy. fine results are generally reported un immolc basis, either by using the measured, or the generahised single carbon group. msmolecular weiglnts. It is always advantageous to work with the compositional analysis in mass fractions. Time nnolecular weight ofheavy fractions, particularly the plus fraction due to its low reliability, may be varied as atuning parameter. Working in mass basis will retain the original compositional data, when questionable molecular weight data are adjusted.

All reliable experirunensluml data should be used jun tumsing of EQS. The experinmental data, irowever, seldommm cover all prevailing conditions. Conventional PVT data may not be adequate generally for tuning of EQS. which is often used in simulation of reservoir processes other tinan simple pressure depletions. Experimental data should be generated at conditions closely sinnulating reservoir processes. For example, if gas injection is to be modelled by EQS. multiple contact test data are highly valuable for the tuning. The swelling test with a rich gas, particularly covering conspositions around the critical point, provides useful information for miscible displacemeunt processes. Time type of experinsenlal data required for tuning has been addressed by several investigators [37,38]. In general, the data should cover the pertinent range of composition, pressure and teunnperumtture. Tests unre genmerumily c,nnmducted at Ihe reservoir temperature to simulate reservoir Iii ~csses mmnd mmt lIne sepmurmulor tenumperature to sinunulate sunrface coumditions. Connipositiornal data onm eqn.uilihrated phase are known In be genermmlly unreliable, hence, seidouin used in the tuning. When the estimnation of produced fluid composition by gas cycling in a reservoir is the tnain target of the study, such a treatment of compositional data will be

326

9. Applieatio,u in Re.c,rroir Sj,,nulntjo,u

9.3. Tuning ofLOS

327

unjustifiable. In such eases, reliable compositional data, by methods such as high pressure fluid analysis, should he generated, evaluated and improved prior to being used in tuning.

Selection of Regression Variables line putrannelers thmrt uue ofteum used in humming are bitsary inmteraction parameters, properties of I)setn(lo-conmrponetsts. particularly time critical properties and parameters of EOS. An effective, html usot mnecessuirily (lie unsost appropriate, approach is to select and adjust those paranneters upon wlsich lime predicted properties are usnost Sensitive. TIne tuning is then achieved with nninor clsmmnges imm origimmal himmrmunnnelers. lhse relmntive effectiveness of various paranmneters nmay depend nun tIne fluud type. In nnuili-variahle regressions, tire nmrmuthnennruitical routinne mssay be designmed to rely onr adjusting the parmnnselers winich show high values of derivatives of time objective function relative to them. Aguurwmml et unl. (411 proposerl a nmmelhmod winere lIme nmmost effective paraimneters were selected (lynmmmsmicahiy frommm mm large set of paranneters dtmring tine regression process. Gani and Fredenslund (421 suggested a tumming procedure based on establishing the sensitivity of the prcdictedl resimlls, (lepcnding on tine fltmid amnd tIre prediction problem and selecting the most effcctive vumrimnhles for regression. Imu a nnunnumher of tesled cases, (Ire binary interaction paransmeter (HIP) wmms selected mis omne of tIme usnost effective paraumneters. line usmost connmnsoun approacls is to mndjtust BIt between tIne Iiglmtest componenl, present at a signnificamnt conmcenntrmnliotn in time mixture, and lime heavy end fraction [33]. The values of BIP bclwcemm tIme higlnlesl unmid lIne rest of cotnspomsenrts, (Sr between all lime eornponennts, umnay also be midjunstcd Iry regressiung tIme pmmmaunsctems of mm generalised correlation for BIt, such as Eq.(4.80). lime selection of BIt as a regressions varimmhle is mainly based on lime view that BIP is more of a hiltitng parannseter tlnuun mm pinysicmml property. It is also very effective in cisanging the predicted results of [oS. lluis unppromncln cmmun. however, dmvcrt mittenlion from adjustinmg oIlier tnneerlain iunprml parumunnclers of [OS, stucin mis iropcrties of lime pseudo colnnponents, winich nunay unctually require innprovenmenl. TIre critical properties mmmd tIme aceustric factor of pseudo components are probably tIme least mucctnrmmte inrput data, lsenrce, nrsay he unsed in tinning. Time critical properties are often estimated fromsm tIne specific grmmvity aund tine boiling poinnt. or tlse nmolecular weight, of fractions using generalised correlations preseunted ims Section 6.2. Deviations as high as 6%for Ihe critical temperalure amid acentric factor and 30%for the critical pressure of hydrocarbons may result frotns tlmese correlatiouss. rime adjuslmmient of critical properties affect the predicted results linrough clnanginrg pmmrmmnnsetcrs of [OS. Section 4.2. A direct regression of EQS parameters, or 1 their cocfficieumts Umn mind ~ b. hunve also been suggested [33]. rime scmnsitivily of predidtedl satinruntion pressure, equilibrated phase volumes and densities to the properties of psetudo coursponcumis in various processes, such as swelling and multiple contact tests. wmus evaltumiled for mm wi(ie rmnmngc of flimidls [381. lime adjustable parameters of the pseudo connuponrcmnts were varied in line rmuuige of -5% to +5% of tlreir original values to study their effects onn hsucdhictcdl plimnse hcluaviouur mmdl properties. F(ir example, Figtnres 9. 1214 dcuuuonsslruulc lIme ci lcd of mmdjustimig time pmnrmnmuneters (sin deviations of predicted pr(ipertiCs in a nsnultiple forward coustmucl test of mm hlumck oil with nnsclliummne. The experimental data are given in lable 9.12. Tine phase hehnavioinr was predicted using PR with the oil heavy end fraction described by otmly ouse pscudo-conssponent of C . Note that changing the specific gravity of 6~ line pseudlo comnmponncnt, rcsumltimmg ins adhjustutnent of critical properties calculated by the Twu msiellmod, Sectioms 6.2. had a profounusul effect on the results. The predicted properties, however, inst their scussitivity to lIme spcciiic gravity, or tIme lren(l reversed when it was increased by immore llmmmnn 3%. A cluamrge in sensitivity cmmn also be observed for tine EQS co-volume parameter, b. (omutrary to lIne connnnnnomm view, time parammmcter h can beconne more effective tinan tine mnttrumdtive ten-Inn parameter a in pmcdidlimmg phrase behaviour of high pressure fluids. Amongst line evumlumusted properties of pseudo commsponenls, including time molecular weight, specific gravity. composilion mind parameters of [OS. the specific gravity was found generally to be the usmost effeclive pmnranseter in tuniumg of [OS [38].

Material balance calculations are the most popular method of evalrmatirig experinmmenlmn] data. it
must be ensured, however, that such calculations have not been previously iunmplensmented to smooth or even generate data by the laboratory. TIne accuiracy of repouled phsmnse conmlposition can be evaluated by comparing the number of moles of cads conmmpounerst nn tIne feed with the sum of those in the produced streams. Tine total and conmponenst hunlansce equmnlnons mine essentially those used in flash calculations,
nL+nv=nF

(5.1)

and
z,nF = x,nL + y,n i=l,2 N

(5.2)

The component balance equation. Eq.(5.2), can be presented grapimicahly (39] mis slnown in Figure 9.11 for the data reported in Table 9.4. The deviation of any point fronm tine straigint line of ordinate+abscissa=i, identifies the error associated wmthm tine measturcd rhata of tlnat component. Random deviations generally show errors in the conrposilnonal munumlysis, whereas systematic ones can be due to the error in nmemtsuring the aurnotumsls of phases. Cenlammi plots, such as the Hoffmann plot (40] or tlme modified Wilson equation plot, dlescruhc(l nun Sectioun 3.2, can also be used to evaluate the internal consistency of comsnposutiommal datmm.

C
U

ci cm >
0
5)

C
U

0.0

0.2

0.4

0.6

0.8

1.0

usiole in equi. hiq./mole in feed

Figure 9.11. Material balance plot of the compositional datum reported imm Table 9.4.

Smsnootlmitsg data may soumretimes mask certain ummcomnmusonm fealunres of mt pmmitncuimmr fluid. Ats exansple was givenn in Section 1 .3 where the liquid phase volummise nnscreumscd wutlm decreasinsg pressiure over a short pressure range below the oil bubble point. Clearly if suds data had been smoothed by time laboratory, the correctly predicted behaviour by EOS coumld have been regarded as a flaw.

328

9. Applaauiiuii inn Ru.se,n(nr Sinuu!agion

9.3. Tuning of LOS

329

Table 9.12 (Con.)

Forward-contact experimental d ata at 37 3 k and 20 .79 MPa. I 2 Phase_________ Oil Gas Oil Gas 3 90.00 90.00 135.00 45.00 Added Vol., dunn 77(14 99.01 137.79 Equni. Vol., cuun 36.16 m 0.6896 0.1621 0.6491 0.1939 ~ui. Dens., grannm/cm
fin

fIne equnilmhratcd gas Iron the iirst smmuge was contacied wilh the frcsln oil in the second stage.

~1

10

~0~~~ SC) 2%

e
~ ~

~~~~0~~ dOME a-- suw s-- muir

inna,nct,, in mmcm r~,,,nIc, bin us unrnn,nnnncn SOS

ni
0

a ~6
.4

.i)

-2

>

~2 us
0

Change in Propeumy.%

IS

Figure 9.12. Effect of adjusting various heavy end properties on deviation of predicted liquid density.

Table 9.12. Composition and properties of Black Oi 1(C).

Component C 2 C 3 i-C MoIe%
34.00

-6

-4

Adjusimenm% Figure 9.13. Effect of adjusting various heavy end properties on deviation of predicted liquid volume.
-

inC 4 i-C 5 nC, C 6

12.45 8.42 1.29

4.56 1.60 2.98

is) 0 50
~ ._......_._.....

-n--- CuMr

nfl 505 i~,n,,,nwn i in

i,,nn,nI~,
~

C 7 C~ C 9 C~ C~
Cm

Cmi Cn4 C 0 C~,

2.45 3.66 3.64 2.97 2.34 1.96 1.63 1.59 1.31 1.36 11.80
3

84 94 117 126 140 153 165 180 197

209

Sat. Pres., MPa

Sat. Dens., g/cm S (C~)
M(C, ) 5

Pres., MPa
20.79 34.58

374 17.91 0.6553 0.838 194 3 Density, glcm 0.6599 0.6784

0.694 0.730 0.754 0.769 0.785 0.799 0.806 0.820 0.843 0,844 0.909

ii

~uw

sir

c in SOS

-6

-4

-2

Adjusunuenu,%

Figure 9.14.

Effect of adjusting vunrious heavy end properties on deviation of predicted gas

volume.

330

9. Applications mu Rescs upur ,Si,suuhntio,u

9.1. lunnning of LOS

331

Concentration of the plus fraction and its properties, are probably lime icmmsl reliable immptut information, hence, their adjustment us quite justified. The plus fractioum propeuties strongly affect predicted properties of condensing fluids. Sensitivity of tine prcdictedl results to artjrns(immg input data generally increases when the criticuul poimit is approundtscdh. Limits of Tuned Parameters it is reasonable to adjust mneasured parameters within their error haunds in tumnimug. Wider linsnits for tuned parameters provide higher flexibility for matchrimmg expermnssentmml dlatmi. iIsal cormld, however, lead to unrealistic values for the tuned paramelers. tIse muccuracy of nmmeasured msmolecular weiglmt is mmsucls less tinmmui lhsmul ol lIne specnluc grunvity. Typical error hands for measrured molecular weigin, boiling polunt mmmdl time specific gravity of pseudo components are about, 1%, 1%amid 0.2% respectively, line above hmmnnds, pmmrticulumrly for lIme average hoihimng point, msre unnucin wider f(ur the plt~mfmact monu. ilue mud mnslnnscnmt of the mmhove properties witiminm Ilneir error bmnnds, however, nnmmuy nuot be suit mc cml lii mis: Inmeve tIne required tuning. Pedersen 1361 suggested adjusting tine ismohecular weight by as inngin uns 0%. As these experimental data are employed to calculate tine criticmnl properties of pseudo comniponents, musing generahisedl property courelalionrs wlnicim introduce mmd(hitmonnmnl errors mlmlo [OS, wider adjustments are acceptable. rIme deviatiouss of Iluese coricimilionss, reported previously, can be regarded as the hitmsits for adjusting the properties. Met hodology

Although adjustment of BIt ium tunming of EQS is quite common, a highly effective tuning can be achieved, without resorting to BIP. by just adjusting the properties of carbon groups. This will allow rapid flash calciulations in connrpositional reservoir simulation as described in Section 5.1. A cotnspreimermsive dmntmm set on a voluntile oil, was used by several imnvestigators in a comparative Running exercise (391. All lIne participants used time Peumg Robinson EQS (PR), applying different in-house tuning methodls. lucy all, however, used BIP as a tuning paranmeter. For exammnple, the results of tIne tuned unodels for some properties are shown in Figures 9.15 and 9.16. Tlse results of a ttmnimmg urmetlnodl usiung PR hut with no BIP, are also shown in tlnese Figmures. In tins method [431, tIne specific gravity of time plus fraction and also its nsmeasured commceuntrmitioni, mis tIme lcmmst reliable comisposiliomnai inforunatioum, were unsed as time regression pmmrmmuuuelers. lime smut pmmrummsmeters oh carbons gromnps were also mndjuusted. Time results clearly denusonstrmnte that effective and probably unsore physically based tuniitmg. can be achieved witlmorut uusinsg HiP. Tise suitability of the nusetirod for guns condensate systems has uulso been
demmnonnstratcul

1441.

>

It is advisable to reduce tIme nunnber of variables in the regression to avoid nrnusscricmml problensis
and innprove the searcin for time global nrminrimum of tIne objective fiurnctiomm. I-lenncc. For exmmmsmple, regressing the parameters of time BIP correlustion, Eq.(4.80), amid those of lIne volume shift eorrelatiomm, Eq.(4.36). are preferable to adjusting a large nuummher of HIP vmulmmes aund shift paranseters. Alllmormglm sinmullanseous mndjtnstnmment of regressed vmuriahles ussay lead to smitisfactory resmmlts, tIme unsuiltistage tuning. wlnere selected parannmeters are adjusted nut turn cmunm be unnore mippmopriate. For example, when using a two-parameter EQS such as PR arid SRK, the denrsnty data can be initially left oul of regression. Tinen the volume shift fmmctors of pseudo connnponcnmts can be adjusted to match lime density dmmta, prior to goiung back to the otlscr regressedl pmirausmctcrs for fine tuning. It is inmportaunt to mnaintmmin the consistency of regressed pmiramnetcrs whsenm usnorc Ilnams us single variable is used to tunme a phase behaviour model. lhe critical tennpermmture. uncentric fmmctor and the boiling point temperature shotnld generally increase witim the molecular weiglnt or carbon number of pseudo components, whereas the trend should be opposite for tine critical pressure. The above critical properties are calculated from generahised correlations based nun tine specific gravity and normal boiling point temperature of carbon groups. Specific gravity is one of time IrlosI effective parameter in adjusting the predicted results of EOS. It camm be selecled as tIme tuning parameter, with the boiling point related to it by nsaintaimrimsg lise Watson cinaracterisation factor, K~, constant equal to its original value, as tine variation d)f linis factor is relatively small for carbon groups. K~w[(t.8T )~]/S 6 (6.2)

a
>

am

iso
F,cssuire. MPa

Figure 9. 15. Conmparisoun of relative vohuuumme in differential liberation experiment predicted by

variomis lmnnned nniOdlels. [Inc nmmaium drunsvhack of rising onsly tIne properties of pseundo components, particularly of the plus fraclioun, is thmnt tine guns pmnase propertmes, such as the density particularly at low pressures, are hot very sensitive to themnm due to tineir low concentrations. A powerful tuning parameter is tIme tennmpcralure coefficient of the attractive ternr in EQS. This coefficient and its adjustment for sumper-critical colmnponnents, were described in Section 4.4.3. It can be readily used as a tuning variable, along with heavy ernd properties.
A connnlmimnmmtion of rchimmhie EQS mmmdl properly characlerised fluid data should lead to predicted

results close to cxperiunmcnlal valuncs, hence, very little need for tuning. A phase behaviour nmsodcl wilh its input pumrmuumnetens umdjtmsted widely woumhd lead to unrealistic resmults at conditions otlner Ihmln tlrnse tested in tine ttmmsinrg. As experirumeumlal data set covering all possible conditions witlminr a reservoir is not fortlsconming in nsiost cases, severe tuning should be avoided. Pedersen Cl unl. (36] have reviewed the danger of tuning by considering various cases. 9.4 I)YNAMIC VALII)AlION OF MODEL

As all the critical properties and the acentric factor are calculated frorns time specific grunvily and boiling point data, they will be adjusted consistently when the specific gravity is varied to tune EQS.

Typical laboratory uncasuremneunts mused imm tuning include conventional PVF data, swelling and nrumltiple countact vumpour-Iiqsuid phase equilibrium data. in 1985. Kossack and Hagen [45]

332

9. Applmcaounm jim Reservoir Sinnulation

9.4. I)y,manmic Va/ida,jo,u ofMode!

333

studied the capability of EOS, tuned against static experimental data, in simulating the gas-oil phase behaviour in slim tube displacement tests. They concluded that an EOS tuned to the static PVT data was not adequate for simulating fluid displacement. A different set of EQS parameters was required to match both PVT and displacement data. A similar conclusion was reached also by Mansoori et al. [46].
0.65

kk AP (9.21) and V,, kk At (9.22)

0.55

ModelA Modetli Modei C

NoBIP

where, V is tine fluiul velocity, APIL is tIne pressure gradient along tIre tube, ~t is the viscosity and k is lime tube absolute pernmmeahihity. kr is time relative penneahility wlmich depends on the fluid saturation, the interfacial tensions (1Ff) and velocity at the displacenment conditionss.
U Esp - i.iquid t~p - vtnnr

0.45

0.35

Fqs.(9.2 1-22) clearly dennmonstrmnle tinmnl tIne ratio of lIne flowing pinumses. wimicln dctcrunnines tIne nummxlnnre cnmunilsumslmm(mnn. Imenuce, lime l)Imuusu hclnuuviotmr, dcpenmds nmnu lhue relmilive peunnncmnlmilimy cormelmml moms emnpioyedl inn lIne sinumtnlationn mnsodcl mnnmul line viscosity of hot in plmmnses. Fhmesc paraummeters, timerelore, shnounldl be dcleumnmimncd relimihly inn uudvamnce. Relative Permeability Function

0.25

50

Khazam et al. 1471, investigated tIme relative penmiemihility of gas-oil in a slim lube musing binary
fluids. Two phase nsnixtures equihibrmuted at tine test tenrperature and pressure were prepared.

Contacted Oil Volume! Iunj. Methane Vinlumnnc

Figure 9.16. Comparison of gas and oil density in a forward contact experiment predicted by various tuned models. The slim tube, described in Section 7.2, is the simplest apparatus that can be used to plmysically simulate compositional changes resulting from the conlinuuomms contact belween lime inmjection gas and the reservoir oil. The choice of an almost one-dinnensionmml flow mm ut properly designed tube displacement is reasonable, as time effects of dispersion, viscous fingering, gravity override and heterogeneity~ which are significant in a large three dimensional system. are minimised. It is, therefore, reasonable to expect that a reliable phase behaviour model, which is to be used in a reservoir simulator to study gas injection, should be able to predict the fluid

Tine slimnm tube was pumeked witlm tIne liquid and displaced with the eqmnilihrated gas, at alunmost no nmiass transfer cotmriitions. fine test was conducted at different pressures over time iumterfacial tension range of 9.8 (00.04 mN/m. [he results are shown in Figure 9.17. It is evident that as 1FF approaches zeuo, thmmt is approaching miscibility, the residual oil saturation, the immobile oil left behind, decreases towards zero and tire relative permeabilities increase. The investIgators conducted displacemenmt tests, using different fluids, includiumg real reservoir samples and concluded that a single set of relative permeability-saturation curves is adequate in describing the flow hehunviour of all fluid systems which lmave the same 1FF value.
1.0

conditions in such a simple displacement. Hence, the comparison of the displacement data with simulated results of a compositional model using the tuned EOS can be ennployed to evaluate
and, if necessary, further tune the phase behaviour model.

0-S

0.6

The experimental data used in tuning of EQS sinould cover the conmposilionunl range occurring in the displacement process. Generation of suncin data for soumme processes, smmchn as rids gas injection, where miscibility is not achieved at the two leading and truuilinng edges of time transition zone, is not practicable in static tests. Therefore, it is advisable to further evaluate the phase behaviour model, that is tuned to all the relevant static data, by checking its performance in predicting slim tube displacement data. The test could also indicate unexpected phase changes, such as asphaltene deposition, which may not be evident in statistic tests. The flow parameters and numerical methods incorporated in the sirnumlation model can strongly affect the prediction. Hence, these factors need to be carefully isolated and detennined, if the phase behaviour model is to be evaluated against displacement data. It has been demonstrated [47] that, after proper implementation of the above factors in a numerical sinmuimnlor, an accurate prediction ofthe displacement can be expected froun a reliable phase behaviour model. The flow of gas and oil in a slim tube is described by the L)arcys equation,

C0.4

am
0.2

0.0

20

40

60

80

100

Oil

Samuraumon. 55,

Figure 9.17. Variation of gas and oil relative permeability with interfacial tension (tFT). As tIne unmeasurenmemmt of relative pennieability in a slim tube at various 1FF values involves a major effort, relative permeability correlations may be used instead. It is. however, essential to

334

9. Applscasuons so Re.u e,uour Si,,usikuio,m

9.4. Dunanuic Validation of Model

335

measure the relative pernneabihities at a single high IFI condition (base curves) as lhsc starting input data. Displacing a binary oil with its equilibrated gas, and rtreasuuring the production rate and the differential pressure across the tube, provide all the reqmnimed data to detenusnine the relative permeability using a graphical method [48]. There is no useed. however, to usseasure the change of relative permneabilities with lET extensively. One set of relustive pcrnsscmmhihity curves at a low 1Ff value and the base curves are sufficient to define tine parameters of a gemmeralised
relative permeability correlation. A nuunmbcr of correlmitions (47, 4951] lsave been (IeVCl(ipcd to miccounusl for tine el fed of 1Ff our

lire second utpproacln is based on time analogy between viscosity-temperature and specific

volume-temperature beimaviour. Cuuhic equnations, similar to van der Waals type equations, but with vokuunne replaced by viscosity, have been proposed (56, 57]. There is very little reasoning behind linus approunclm, apmmrt from the similarity of pressure-volume-temperature amud pressurevuscositylensspcrature plots. The thirdi approacis uses the concept of residual viscosity, which is defined as the difference between viscosity mst prevailiung coundilions aund linmil at low pressmnre whmere the viscosity depends unmshy on tIne Ilmerusnal cumergy. line rcsiduuml viscosity cmiii he related to tIre fluid density, whereas lIne viscosity (sf gmmses at low pressunre can he reliably deteruimined by time kinetic Iheory of gases. This mnpproach, as imsrplcuncumleul buy I.olsrcnz-Bray-Clark (LBC) [58], is mused widely in tine p.lnolemmmnn imudiustry. p;uumucsmlanly ins rescrvunnr smumnmmlmntiomn. lime nmnellnod is hasicunily tbnmil of Jossi et ~mi. 1591 for ptnre commupomiuuds, extendled In lnydrocarhron reservoir flmnids, as described below.

relative permeability. One of the earliest atteusmpts to correlate gas-oil relmulive peumsuemuhiility witim lET is that of Coats [50]. The Coats correlation is based onr the commcepl lhsmml is tine interfmmcial tcursiimn between tIne two plmmsses appromicines zcno nmcmnr line critical point ( immiscible conmnhit ions), tIne residuuul phmnse saturation values decrease towards zero umund line relative penumucuitumluty curves vs. saturations approacim straight dIiagonal lines. I he suggestedl to cstiumnmulc tIme rclmulive pcmnnmcmuhihnty mit aumy 1FF vmnlue, by interpolation betwceun the base cmmrve dlelcrnnmincd mnt luigin lii mmmd tIne slraiglsl diagonal line,
=

line knmnctme theory of guises slniuws thaI time viscosity is inversely proportioumal to,
X F~M i~ (9.25)

F,,k,, + (I

F,, )k,,,

(9.23)

wincre kr is tlse relative pernneahihity to gas or oil, annul F is the scmnlmmsg functor between time bose 0 relative permmseabihity krb and the miscible relative permeability ~ F is a function of the interfacial tension, 0 F,,
=(~~b)
~,

Jossm et al (59] nnsultiplicd tIne residual viscosity by ?~ to unmake it dimensionless and correlated it

witim tine reduced density, P

p/pt.

for pure connmpouncls as,

as follows:
[(p

~u~)x

x + l0~

(9.24) wisere mn = 0.1023(1 1 mm = 0.023364 2 a = 0.058533 3 mu = -0.040758

a + a p + a p~+ a p~+ a p~ 1 2 1 1 4 5

(9.26)

wisere
(52].

55

h is the base IF] and mm depends on time porous usrediuuni type witlm mm dlcfmmmuit vmslue of 7

The value of k~ cams be assumed equal to tine phase satunruilion. Altisouglu tisis simple interpolation approach, Eq.(9.23), will result in zerur rcsiduual oil mit mill 1Ff conditions, its effect on the evaluation of phase behaviour model, using slinmi tube dispiaceusient dumla, is insignificant. Viscosity Prediction The viscosity of a single phase reservoir fluid increases witin pressure, except at nmear critical conditions. The increase of temperature, decreases the liquid viscosimy whilst it increases the gas viscosity at moderate and low pressures. At high pressures, lime behaviour of gas viscosity is more liquid-like (Figure 2.23). Hence, those correlations developed either for gas or for liquid. Section 2.3. may not be suitable for reservoir condiliotis, particunlarly for gas injection processes. It is required to use a single method to predict the viscosity of both plmases at such conditions, especially when miscibility is approachedl and the properties of vmnlxmur and liquid become similar. A nunmber of methods are applied to both gas and liquid in reservoir studies, which can be classified into three groups: First, tine corresponding states methods, where the redunced viscosity, defined mis lime rmnlio of time fluid viscosity to that at the critical point, is relaled to two redunced sImile properties, sunclm as tine reduced pressure and reduced temperature, Pedersen et al. [531. or time reduced teusiperature and reduced density, Ely and Hanley [54]. As reservoir fluids cannot be modelled accurately by the simple two parameter corresponding states principle, some correction factors are included in these correlations. The correction factors adjust tine deviation of time predicted result by comparing it with the viscosity of one reference fluid, methane [53-54], or two reference fluids, methane and decane [55]. Although the viscosity is not a thernnoclynanmic state property, the above approach predicts acceptable data in most cases.

4 a = 0.(X)93324 5

amnd ~.m is line low pressure viscosity whicln cmmmm be determined as, fn=34x10 T, 1X l7.78X l0~(4.58T l.67) /X
4

Tr 1.5 i~>1.5

(9.27)

Note that lime units of T, mind P, iii Eq.(9.25) should be K and atm (MPa/0.l0I325) in order to obtumimi the viscosity un nstPmm.s (cp). l,olnrcnz Cl mml. 1581 extended lime mihove In nnnixtmmres. by prusposing the Herning-Zipperer mixing mule (601 for lIne low pressure viscosity mind tIne mnmolunr mixing utile for other properties as foIlmiws,
N
(1 =

~x~p~M?
=1

/ ~x~M?

(9.28)

X =(txT~JtxM)~(txP J~ 5

(9.29)

336
1

9. Applicalion in Reservoir Sinmnlatio,n

9.4. I)ynuanuic Validaiimm ofModel

337

PM =(v ) 5 5

=(u~XiVci)~~t

(9.30)

p, = v / v
5
52. U

1.4
1.2 1.1)

where PM is the molar critical density and v is the critical molar volume. The amnthors used the 5 to back calculate and correlate tIme critical molar viscosity 0data of a number of reservoir fluids volume of C .~. as, 7 3 (v )~, = 1.3468 + 9.4404XlO~Mc l.7265lS~, + 4.4083x10 M~,Sc, (9.31) 5 where M and S are the molecular weight and specific gravity, respectively annd tine estimated 3 critical volume is in m lkgmol. The method of Lohrenz et ai. is quite sensitive to the flumid density, as apparent mm Eq.(9.26). Hence it should be used only in consbination wills EOS winich are known to pmechict gas annul liquid density reliably. Figure 9.18 shows the deviation of predicted viscosity by Eq.(9.26) for pure connmpounds. Clearly the correlation loses its reliability for heavy compounds. Ileisce, the unnctlnod becomes unreliable for dense fluids witin reduced densities over 2.5 [61]. It unray predict oil viscosity with deviations exceeding 100%. 40 20
?4~ 4
,~ 0 0 U

52,

a
U

a 0.8 >
0.6

0.4

I----

o2 to
20 Pressure, MPa .1)
41)

Figure 9.19. Consmpmmrison of predicted viscosity by LBC nnctlmod with experimeuntumi dama of mm North Sea oil at 371 K.
Example 9.5. Estimate the viscosity of a liquid

Menlumnne
Eilnannc

mixture conmposed of Ci=59.30, C~=37.46 and nCg=3.24 molc% at 311 K mind 20.68 MPa. The liqmuid density at the above conditions is 3 0.368 g/enn . The nmeasured viscosity is 0.05 10 nnPa.s (0.05 hO cp).
S(,huuiolm:

.2(1 -40
-60

A A

Prop;iuie n-Iiummnne un-Pcuuuan,~ nuIkoanie n-Osuminu

n-I)crznune

The properties of pmmre conmnpounenmts are read froun Tumble A.i us Appendix A, with time critical pressure converted to atom (divided by 0.101325) and their viscosity is calculated fronnm Eq.(9.27), mns follows. Conump. x M, g/gmunol Ic, K Pc, atm Cl 0.593 16.043 191)56 45.4 C3 0,3746 44,096 369.83 41.9 nC8 0.t)324 114.231 568.7 24.6 v,, cnm/mnoh 98.6
200

t~s -80

n-Dusujcc;nuue uuPeuuu,uict,iuue un-I Ieptmuutesauue

486

T, 1.63203 0.84092 0.54686

~t, mPa.s

0.04706 0.03343 0.03 186

0.01134 0.00864 0.00605 given mn

Fine usnixtmnre properties are then calculated using the nnixing rules Eqs (9 . 28-30)
0 I 2

4 Cl
C3 nC8 0.02693 0.02150

Reduced Densimy

xM 2.37519
2.48752

xi, 1 l3.0()
138.54

xM 9.513
16.518

xP,

xv

26.915
15.705

58.47
74.92

Figure 9.18. Deviation of predicted viscosity of pure hydrocarbons by the method of Jossi et ai. It is common to tune the nmodel by adjusting the critical volmunnue of tine C + fractions to nimatch the measured data. The above approach improves time predicted results 7 markedly in processes where the heavy fraction remains almost intact. A successful example of sucim practice, where the measured and predicted viscosities have been matched only at the bubble point, is shown in Figure 9.19.

total

0.00210 0.05052

0.34629 5.20900

18.43 269.97

3.701 29.733

0.796 43.417

15.75 149.14 The

Ihc valume of 1 for nnmixlunre is c;ilctulmnleil from Eq (9.28) to he 0.0096992 mPa.s.

valune of X for immixluire is cmnlcnnlated fronmm Eq.(9.29) to he 0.0377423.

reduced density is cmilculated froums Eq.(9.30), as.

The mixture

p,=(0.368/29.733)xi49.14=h .84583

338

9. Application in Rcse,uoir ,Sinuuulaiion

9.4. lknanuic Validation of In! ode!

339

4 Substituting the above values in Eq.(9.26) results in, p_~n=O.O37l 5 usnPa.s. tluuuce, 4 84 p=O.O 6 mPa.s The predicted valume deviates by 8% from the measured vmscosmty. Implementation Most reservoir simulators available today obtain solutions to flunid flow equuatuomms (non-linear patlial differential equations) by replacing the derivatives with finite-difference mippmoxinmations. The use of these approximations introduces truncation errors and nmuusnerical dnspersion. The simulation results are, therefore, sensitive to (lie numusher of grid-blocks mnnd Imusme-step suz.e usc(h to model the slim tube displacement. Tlmese paraunetcrs should be selected so tismul tine nunscrical
dispersion becomes close to that of the physical dispersion in tIme lube.

of 311 K. The Peng-Robinuson EOS was ttmned to tine vapour-liquid equilibrium experimental data of Sage and Berry (63] on the above ternary system and multiple forward contact test data (64], at time test temperatture amid pressunre. Tine mtmltiple contact test at 20.68 MPa imndicated the mmclsicvcnrsent of umsiscihihily. IIne ltuused mrsodcl nnatchcd tine volumes of vapour and liquid phases in equilihriium tests wills an average absolute deviation of about 3%, with the composition and density reliably predicled. Tine tuned plnmnse helmaviouur umnodel was inscorporated in a one-dinnmensional numerical unodel, along with nmeasunred relative pemnmeabihity curves and the optimum grid-time step sizing, to predict tIre above, IInc simnuhumled displacenssent resulls at 13.79 MPa are shown in Figure 9.21. TIse unodel tuned In tIme static dala clearly is capable of predicting the dynamic experimental resunils. Time resunils mit 20.68 MPmu were mnlso qmuite reliable, demonstrating miscible (hisplaceunnent (64]. Io indicate lime sensitivity of disphumcennent results to tlne phase behaviour nruodel, tine aunthors dclihcrumleiy usuistumned the model aguminst tIne static data, which resulted in a
significminl deviuml ion betweeim tine predicted and nsieasured displacement data, particularly the gas

The degree of sImm tube packing homogeneity and time associated plnysical dmspcrsion, cmmn be identified by conducting miscible liquid-liqumid displaceusmenl [471. For mummy nsununher of grid-blocks, the optimunu time step, whiclr yields the best nsatch between tine predicted and tIme experimental effluent profile, can be identified as shown in Figure 9.20. For the opluumsmscd grid size-time step, the numerical dispersion can be conmsidered equnivumlenl to lIme plnysmcmul dispersion. In the slim tube displacement the pimysicunl dispersion is generuilly sinumull mnusd cams he mmssmutned zero in most cases for lnonmmogeneous sand pmmcks. TIse umnetlnund proposed by I .muunl/. (62] cumnn be used to determine the optinmmum tinme step-grid block ~iIlmzero dispersions, uunstemudl of counduicting liquid-liquid displacement. For a typical sliuss tube, 100 grid blocks are generally needed to achieve a stable numerical solmution. At conditions approaching ussnscihihnty. iundlncmmled by low lET regions, more grid blocks are required due to sharp changes of fluid properties.

hremmk llsrougim mind prodtncing gums to onl ratio values.

so
C
40 0 20

0 Q0ii~0t52.0
1 E

LC
Li C
0
CI C) 5) 0

0
0

a C-)
C
0)

LL~

mm
0.8 t).9 1.0 1.1 1.2
0.

Pore Volume injected (fraction) Figure 9.20. Variation of simulated effluent conceustration profile witls Immune step size its a shun tube modelled by 100 grid blocks. 7 The above approacis was applied by Khazam et ai.(4 ] to a liquid nnmxture of CI/C3/nCl() displaced by methane in a slim tube at two pressures of 13.79 and 20.68 MPa and tennnperature

00

05

1,0

1.5

2.0

Pore Volume tnjecied. Fraction

Figunre 9.21. Memisured aund predicted displacenmnent data.

340

9. Applieatioin in ke.s.rnoir Si,,nndaiirnu

9.5. Euaiiwtion ofReservoir Fluid Sanuples

34 I

The slim tube displacement data are certainly valuable for evaluating a tuned phase belmaviour model which is to be used in reservoir simulation of gas injection. in immiscible displacement processes, however, tuning of the phase behaviour model to relevant stalic equilihriuumrn data should generally suffice and there is very little need for relatively tinme consuusnimmg displacenrnenmt tests. The swelling test, covering both sides of the critical poinrt, will be higimly unseful to generate static data for miscible processes [12]. 9.5 EVALUATION OF RESERVOIR FLUID SAMPLES

in

a saturated reservoir, tIme saturation point of coexisting gas and oil phases should be equal to the reservoir temperature-pressure at the gas-oil contact. Hence, neglecting time compositional grading in space, it is expected that the nreasured bubble point of the oil sample, or time dew poimnt of time gas samumple. be close the above value. In practice, however, the compositional variations with depth aund area are rarely negligible. The bubble point pressure of oil is a unonotonic fumiction of the gas to liqmuid recombiumatious proportion, i.e., tine bubble poinnt pressure increases with increasiung gas to liquid ratio (GLR). i-lensce, it is a reasonnahie practice to ignore time measured GLR during oil sampliumg aumd take oil and gas sanmples fronun time separalor and recombine timeun to achieve tine target bubble point. The dew point pressure, however, may increase, decreumse, or remain almost uncimanged by increasiumg GLR, Figumre 2.2. As time GLR-dew point pressure curve is dome-like shape in gas
condensate nnixtmnres, il is possible to obtain time saumme dew point pressumre wills two different Ci LR s.

The collection of a sample that reliably represents the reservoir flmnid is essential ium any phase behaviour studies. After all, a model tuned to experinnental datum of a saunpie, with properties significantly different to those of the reservoir fluid, will be of little value inn studying that reservoir. Challenges in obtaining representative samples from gas condeunsalc mmmd volatile oul reservoirs have been well acknowledged by those involved. Souse of time key issumes and pitfalls in fluid sampling were described in Section 2.1. Any fluid produced from a reservoir should, in principle, provide vunlununhic muslomnnmunlion nun tIne reservoir fluid. However, it may not have tine same composition and propertIes mis the ormgmnuuh reservoir fluid. The collected sample may have gone through certain umnwanted processes. resulting in changes of its properties. If these processes can be reasonably ideumtificd, it may be possible to trace back the original fluid from the collected sample. Phase behavuoumr mmnodels cmtnm play an important role in helping the above task. Evaluation and improvement of a collected sample generally benefit from a combined experimental and numerical modelling effort. Some processes. such as a snmmgle equuilibriumnnm flash, can be physically simulated by simple experinnents. Pinuuse heinaviour nmodels are required to simulate more demanding equilibriummm tests. Processes occmmrrinmg witimumm lIme reservoir generally need to be modelled by a reservoir siumnulator. The main source of error in sampling is the phase transition and collection of co-exitiung fluuids at an improper ratio. Reffstrup and Olsen [65] studied fluid conmrpositional clmanges during surface sampling under non-ideal sampling conditions. They used a nnodified black oil simulator to produce from a low permeability lean gas condensate reservoir amud an EOS nmodel to simulate the recounhination of separator samples. The authors showed that the dew point pressure of a wellstream (recombined sample) was lower than the initial dew point of reservoir, but higher than the bottom hole pressure. They recommended a method to hack calcunlate the initial reservoir fluid composition by matching the initial reservour dew point pressure. Fevang and Whitson [66] extended the Reffstrup and Osiens mnnethod to cover otlmer types of reservoirs. The authors conducted an extensive investigation of susmumpling Irons (lepheled reservoirs to determine the original reservoir flinid mmsinrg comumpositmonnmml smunmuimmlioun. Ilney proposed experimental methods to obtain the original fluid fronnn collected sammnplcs, based on their simulation results. A key recommendation to obtain the original reservoir fluid in saturated reservoirs was to equilibrate the samples collected from the gas cap aurd the oil zone at the original reservoir gas-oil contact, pressure and temperature. The equilibrated oil and gas phases, then, represent the original reservoir fluid in oil and gas condensate reservoirs,
respectively.

~Flnebelmaviour of mm typical North Scum guns condensunte. with a dew point of 31.94 MPa at 383 K, was simuimmted by flaslminmg it uml vmmrious pressuires. Ihe equilibrated gas and condensate
plnumses were tlsen rccouumhinmed nil differcmmt rumlios at cads pressure. Figure 9.22 shows time predicted dew point of tine differenst recombined fluids. The results clearly show timat tine higlrer

the separator pressure, the higimer is the difference between the two recombination ratios which resmult in tine same dew point. Therefore, ntis less likely to select the wrong reconnbination ratio at high pressures when mminning to matcim time dew point.

~13

CI cm.
5)

32 31 3(1 Separamor Pies.

MIa

C)

cm.
C

29

a cm. Is

(1.69

3.31
211 0)34 27 26 0.1)
. I

689
13.79

20(u( 0.1 0.2 0.3

0.4

Mole Liq. Add/ Mole Sep. Gas

Figure 9.22. Predicted dew point pressure of reconnbined separator gas and liquid samples. The plot also denmonstrates that the dew point becomes less sensitive to the recombination ratio as tIne separator pressure increases. As tine target dew point has a certain error hand, the lack of seumsitivity affects line recomunhination ratio markedly mit Imigh pressures. However, as time two pimases nit Inigh pressumres mire more siuusilar than at lower pressures, the effect of deviated gas to liquid ratio in recounhination on the total fluid composition and its behaviour is less significant at isigher pressures. If time separator pressure is equal to the saturation pressure, no condensate will form. Theoretically, if tine condensate formed at the dew point is added to the saturated

A main concern in surface sampling is the recombining proportion of the collected liquid and gas from the test separator. Any uncertainty in the measured gas to liquid rmtlio in the field directly affects the composition of recombined sample and its properties. imperfect sepmnration of the phases also causes either some liquid to be carried over witin the gas from nun upstream separator to the next (carryover) or some gas to be produced with the liquid (carry througin), disturbing the produced gas to liquid ratio.

342

9. App!ieaOon in Reservoir Si,nulatio,u

9.5. Eu.alima(no,u ofReservoir Fluid Sample.r

343

gas, regardless of the phase ratio, the gas composition remains unclnanged at Ihe dew point

pressure.
Sauur.uion

to

The predicted liquid drop out of tine mixtumres in the above exercise recoussbinned at time atmospheric pressure, with a false GLR selected, is shown in Figure 9.23. Tlse mumixtunre clearly lacks the true behaviour of the reservoir fluid. The predicted results with -3% error its tine target dew point pressure are also simown. Note that mmltinoungln tIme -3% error inn dew poinnt corresponds to a much higher deviation in the reconnhinmitioum rmulio at tine higImer pncssunre (Figure 9.22), the results are more acceptable than those at Ilse lower pressure. Distaist~c 10

Original

a
CI
5)) 0) CI C 0)

cm

a C a
0

Fignure 9.24. Gas-condensumle distribution arouuumd a producer. Figmire 9.25 shows time growtlm of line two phase region by depletion as simulated numerically for a typical Nortln Scum produmcer. lime overall connnpositioun of the produced fluid with time is simowmm in Table 9.13. Note tlsat tIme composition changes very littlc with time and it is almost the sane as that of the original reservoir fluid, in spite of the significantly extended condensate ring. I tencc collecting tire produced flunid can provide a reasonable representative sample of the original siumgle phumse gums. If lime rmute is decreased to reduce the draw down, similar to the nnnetlso(l used in oil smnmnsplirmg. tIme resullinsg pressure bdmiid-up not only vaporises the condensate iuslo lire gas plsase. html mnlso dunmrps sonsse conmdemmsatc into the well, as a lower condensate smmlumrmmtiOn is requmired to unmmuintmiin tine rcdunce(l condensate fractional flow. Both actions may lead to mm smmnmnpie mssunciu richer thunun tIne original reservoir fluid.

8 a > a a aa 0

Recom. Pies. 20.7 MPa

False GOR 6

a.

is

Pressunre, MPa
Figure 9.23.

Predicted liqunid drop Out in constant composition test at 383 K for various reconnbined fluids. The rumain imnpcding factor in collecting a representative fluid sample is the pinase lrmunsitiomm due to pressure reduction aund the variable mobility of the phases witimin the reservoir. Some recommendations to alleviate the above problem were given in Section 2.1. A long flow period to slahilise the well and its drainage zone is often advocated. it was noted, Isowever, thmnt for a gas con(lensate reservoir a representative sample may he obtained (luring nonsnmml operalioun if a quasi-steady State zone around the producer can be acinneved.
Tine condensate initially is formed around the weilbore, when the pressure falls below line dew

U.

>

(It

to
Time. [lay

tOO

000

point and the two phase region, referred to as the condensate ring, grows into tIne reservoir hulk by continual production, Figure 9.24. The condensate saturmnlionn at army locatiorn increuiscs
due to the local reduction of pressure and the inflow of rich gas towards the proiluncer. Tire increase in condensate saturation increases time condensate relative permeushility mind decreases the relative permeability of tine gas, Figure 9. 17. Tisis results in amm increase of coundensale to

Figmure 9.25. Growtir of commdcnsmute rimmg witin linme for a typical North Sea producer. Tine collected sausuple cminn he innrprovcd fuurtlncr winenr line reservoir gas dew poinmt is known, mnsiung lIme comnclumsiuimm ohtuummcd iun tIne mecombinuutionm exercise described in Figunre 9.22, as follows. The two plnumsc sulnnulslcs cuullccted at lime sumrfmmce mire rcconsbined at the nnicasured GLR to obtumin line welihmead stremmun. lIne usnixture is linen brougist to equilibrium at the average reservoir flowimrg pressmsre (or line imottonsn hole pressure) and temperature and the remaining liquid plnunsc is removed. Ann nmdeqummmle volumunse of line rensnoved liquid is added back to the gas to mnatcln its dew point pressmure to the initimml reservoir valume.

gas fractional flow out of that region as described by Eqs.(9.2l-22). Hence, line condensmule satuura(ion increases only to the value winch nnaintmsins (tie associated frmictiomumil flow. As tine condensate accumulation diminisises, nun unpproximnsate quuasi steady sImile msmmmy bc eslmublisimed ins thuit regiour, with the overall coumnpositioms of the ountflow beinng lIne sunmnse as timmmt tlowinrg imsto tIme region. However, if line above arguument was strictly valid lisrougluouut tine two plsasc regiusmi, the region should nol grow at all winich is not tine case. Nevertlseless, it is a rcasonmahIe assunmption for practical purposes.

344 Table 9.13. Variations of produced fluid cornp~tionwith Comp., mole % Original I day 10 days N2 1.024 1.026 1.026 C02 2.088 2.089 2.089 CI 75.543 75.646 75.652 C2 7.375 7.373 7.373 C3 3.764 3.760 3.759 1C4 0.534 0.533 0.533 NC4 1.366 1.363 1.362 1C5 0.441 0.440 0.440 NC5 0.613 0.611 0.611 C6 0.832 0.828 0.828 Cl 1.405 1.396 1.396 C8 1.400 1.389 1.389 C9 0.854 0.846 0.846 ClO 0.541 0.535 0.535 Cli 0.384 0.379 0.379 Ct2 0.296 0.292 0.292 C13 0.246 0.242 0.242 CI4 0.306 0.300 0.300 CIS 0.221 0.216 0.216 CI6 0.160 0.156 0.156 C17 0.108 0.105 0.105 C18 0.095 0.092 0.092 CI9 0.078 0.075 (1.075 C20+ 0.326 0.307 0.306

9. Application mum Reservoir Simulation

9.5. Evaluation of Reservoir Fluid Samples

345

time. 100 days 1.026 2.089 75.655 7.373 3.759 0.533 1.362 0.439 0.611 0.828 1.395 1.388 0.846 0.535 0.379 0.292 0.242
0.300

0.216 0.156 0.105

0(192

1000 days 1.029 2.090 75.767 7.372 3.754 0.532 1.359 0.438 0.608 0.1(23 1.385 1.377 (1.837 0.528 0.373 0.287 0.237 0.293 0.211 0.152 0.102
(i.089

with fluid sampling, such as flunid conntamination [671, which can be evaluated and rectified by applying phase beinavmour models. so

520 a

15

a 2

ma

Oriqinai
Ihead

a a a. -J

~%Wet

-~-

error OP
(rotact nieth.

0%

Cootact Seth..

2%

(oetaot C4eth.. -2% - Separator Seth., 0% A- Separator Seth. , -2%

tO

15

20

25

30

35

40

Pressure, M1a

Figure 26. Predicted liqumid drop-oumt (if various samples in constant composition expansion test at 383 K.

0.075 0.306

1)072 0.286

9,6 REFERENCES
I. Coats, K.Fl: Sinrmumlalion of Gas Condensate Reservoir Performance, JPT. 1870-1886 (Oct., 1988).

Figure 9.26 shows the predicted liquid drop-out from flumids prepared by recombinming the collected separator gas and condensate after producing the reservoir for 10(X) days in the above example. The wellhead fluid refers to the recombination based on the measured producing GLR. As expected it provides a leaner fluid compared with the original one, Table 9.13, due to loss of condensate within the reservoir. Applying the contact method, described above. provides very reliable results, when the dew point is accurately knownm. Ignoring the measured GLR during sampling and recombining the two phases at the low separator pressure conditions to match the dew point, results in a sample which is inferior to the wellstreunun sample. An error of 2% in dew point impairs the results for both adjusted fluids, with time contact method affected less severely. Although matching the bubble point by adjusting the phase ratio dunring tIme reconmhination process is adequate o improve the oil sample in most cases, the contunc unmethod is tine preferred option, particularly for volatile oils. In the contact nnetlmod for oil mill lIme remmmmuinminrg eqnilihriunn gas is removed at constant pressure. Omen adequate voluumne of it is mudded buick to time liquid to match the bubble point pressure. It should be mentioned that if the recombined sample remained single plnase at the contact pressure, most probably due to improper collected phase ratio at the surface, the conntact pressure could be reduced to form two phases. Then the removed phase is added to a portion of the remaining phase to match the saturation pressure. In this section the application of phase behaviour models in alleviating a number of impeding factors in fluid sampling was described. One can easily identify other problems associated

2. Jacoby, R.l1., Koehler, R.C. and Berry, U.J: Effect of Consposition annd Temperatiure on Pimase Behaviour and Depletioum Performance of Rich Gas-Condensate Systenms, Trans. AIME, 216, 406-411(1959). 3. Lee, ST., et al: Experinnemntal and Theoretical Studies on the Fluid Properties Required for Sinmnulation of Tinermal Processes, SPE J., 535-50, (Oct., 1981). 4. IIommg, K.C: Lunmmped-Connmponent Characterisation of Crtude Oils for Compositional Siusnumlatioms, SPE/DOE 10691. presented at time 3rd Joint Symposiunsn on FOR, Tulsa (1982). 5. Whitson. Cii: Cisaraclerising Hydrocarbon Plus Fractionss, SPE J., 683-694 (Aug., 1983). 6. Scinlijpcr, AG: Siuinuluntion of Conmmpositionnml Processes, the Use of Pscudoconmmponents inn Eqsmmmtion of SImile Calcuulmitions, StE/DOE 12633, presented at the SPFJDOF 4th Synsmposiummmm on EOR, fsulsmn (April, 1984). 7. Montel, F. mind Gouel, P: A New Lumping Scheme of Analytical Data for Composition Studies, SPE 13119, iroc. of 59tlm Ann. Conf. (Sept., 1984). 8. Behrcns, R.A. and Sandier, SI: The Use of Semicontinuotus Description to Model the C Fraction in Equation of State Calculations, SPE/DOE 14925, presented at the Stim 70 Symposium on EOR, Tulsa (April., 1986).

346

9.

Applimatiouu in Reservoir Si,,uu,latjo,n

9.6.

Reference.u

347

9. Gonzalea, E.. Colonomos, P. and Rusinek, i: A New Approacim for Clsuuracterising Oil Fractions and For Selecting Pseudocousrponents of liydrocmnrhonms, JCPT. 78-84 (MarclmApril. 1986). TO. Wu, R.S. and Batycky, J.P: Pseudo-Connponcnt Characterismutiour lor llydrocmmrhon Miscible Displacement. SPE 15404, Proc. of 61st Anmn. Conf. (Oct., 1986). I I . Li, Y K., Nghicnmm, L.X . antI Siu, A: Pinase Flelnmnvioumr Comiupuml mmtionns Ior Rescuvoir Fluids: Effect of Pseudo-Components on Phase Diagrams and Sinnuiationn Results, JCPT, 2936 (Nov-Dec., 1988). 12. Newly, T.M.J. and Merrill Jr. R.C: Pseudocomponent Selection for Comnspositiounmsl Simulation, SPE 19638, Proc. of 64th Atm. Conf. (Oct., 1989). 13. Daumesh, A., Xu, D. and Todd, A.C: A Grouping Method to Optinnisc Oil i)cscriptionm for Compositional Simulation of Gas Injection Processes, SPE Res.Eung., 343-348, (1992). 14. Cotternnan, R.L. and Prausnitz, J.M: Flasln Calculatiorus for (onmliumuous or Semicontinuous Mixtures Using an Equation of State, I & EC Proc. Des. Dcv., 24, 434-443 (1985). 15. Pedersen, KS., Thomassen, P., Fredensluumd, A.A: Tinerurnodynmunmssics (11 Petroleummmm Mixtures Containing Heavy Hydrocarbons. I. Phase Envelope Calculmmtions by Use of the Soave-Redlich-Kwong Equation of State, md. Eng. Chem. Proc. Des. Dcv., 23, 163 (1984). 16. Kesler, MG. and Lee, B.!: Improve Predictions of Ennthmulpy of Frmmctions, 1Iydro.
Proc., 153-158 (March, 1976).

25. Pemmeloux, A. aumd Raumzy. F: A Consistent Correction for Redlich-Kwong-Soave Volununcs, J. Fisuici Phmmse Equnihibria, 8,7-23(1982). 26. Rohimmsonm, DR. mmmd Pcusg, l).Y: tine Clmarunctcrisumtioum of lIne iteptanes and 1-leavier Fractions for the GPA Pensg-Rohimmsomr Progranmss, GPA Research Report 28, Tulsa (1978). 27. Jhavcri, B.S. umumd Youngren, G.K: Tisree-Paranuieter Modificalion of the Peng-Robinson Equmntion of Stuite to Inmnprovc Volumnmmctric Predictionns, SIE 13 118 (1984). 28. Schmidt, G. and Wenzel, H: A Modified Van der Waals Type Equation of State, Clmcnr. Eng. Sd., 135, 1503-15 12 (1980). 29. Patel, N.(. anmd Tejmm, AS: A New Cubic Equmations of State for Fluids and Flumids Mixtures, Cinenmu. Eng. Sci., 77(3), 463-473 (1982). 3t). Valderrannnmn, JO: A Gemmeralised Pumtcl-Ieja Equatioum of State for Polar and Non-Polar litmuls amid their Mixluures, J. (lucumu. Fusg. Japmmn. 23(l). 8791 (1990). 3 I Condensate PVI Stumdies, 1989-1990, Final Report, Report No: PVT/9 I/I, Dept. of Pet. Engng., hleriot-Watt University (Jan., 1991). 32. [)nmnrcsim, A., Xtm, D. and Todd, AC: An Evaluuation of Cubic Equations of State for Pinase Behaviour Calculations Near Miscibility Conditions, Proc. of the SPE/DOE 7th Symposiuun on FOR, 915-924 (April, 1990). 33. Coats, Ku. and Snnnmnrt, G.T: Application of a Regression-Based EOS PVT Program to Laboratory Data, SPE Res. Eung., 575-582 (Nov., 1986). 34, Watson, ST. ammd Lee, W.J: A New Algorithun for Automatic History Matching Production Data, SPE 15228 preseumted at the 1986 SPE Unconven. Gas Technology Synrmposium, (May, 1986). 35, Marquam-dt, D.W: An Algoritlnnsn for Least Square Estimation of Non-linear Parameters, J. Soc. md. AppI. Math., 11(2), 43 1-441 (1963). 36, Pederscn, KS., Tlmouinasscn, P. aund Fredenslund, A: 0mm the Dangers of Tuning Equumtion of State Paranmmctcrs. Clncnn. Eng. Sci. .43(2), 269-278 (1988). 37. Merrill, R.C. and Newly, F.M.J: A Systematic Investigation into the Most Suitable
1)ala for tIne 1)cvelopussent of Equnations of State for Petroleum Reservoir Fluids, J. Fluid Plnmnsc Equuilibrimm, 82, Ii) I lIt) (1993).

17. Lee, B.!. and Kesler. MG: Improve Vapour Pressure Prediction, Hydro. Proc., 163167 (July, 1980). 18. Wilson, G: A Modified Redlich-Kwong Equation of State, Applications Physical Data Calculations, Paper ISO, AIChE Natiou)aI Meeting, (May, 1968).
10

Geumeral

19. Ahmed, Y.,Sugie, H. and Lu, B.C.Y: Comparative Study of Eight Equations of State for Predicting Hydrocarbon Volunmetric Phase Behaviour, SPE Res.Eng., 337-348, (Feb., 1988). 20. Firoozabadi, A: Reservoir-Fluid Phase Behaviour and Volunmetric Prediction with Which?, Ind. Emmg. Clneunm. Funsdzmumm., 18(2), 81Equations of State, JPT, 40(4). 397-406 (1988).

21. Martin, i.J: Cubic Equations of State 97 (1979).

22. Danesh, A., Xu, D. and Todd, A.C: Comparative Study of Cubic Eqimultiouns of Stale for Predicting Phase Behaviour and Volumetric Properties of hnnjcction Gas-Reservoir Oil Systems, J. Fluid Phase Equilibria, 63, 259-278 (1991). 23. Zudkevitch, D. and Joffe, E: Correlation and Prediction of Vapour-Liquid Equilibria with the Redlich-Kwong Equation of State, AIChE, 16(1), 112(1970). 24. Graboski, M.S. and Daubert, T.E: A Modified Soave Equation of State For Pirase Equilibrium Calculations. I. Hydrocarbon Systems, Ind. Eng. Cheun. Process Des. Dcv., 17(4), 443-448 (1978).

38. Reservoir Fluid Studies, 1990-1993 Final Report. Vol. I, Report No: PVT/9312, Dept. of Pet. Enigung., 1 lcriot-Wmitl Utsiversity (July, 1993). 39, Merrill, R.C., Itarlminan, K.J. ausd Creek, iL: A Comparison of Equation of State Tunnming Metimods. SPE 28589, Iroc. of 69th Ann. Conmf. (Sept., 19.94). 40. lloffusmann, A.E., Crummmp, IS. and hlocott, CR: Equilibrium Constants for a Gas-Condensate System, Trans. AIME, 198, 1-10(1953). 41. Agarwal, R.K., Li, Y.-K. and Nglmiem, L: A Regression Technique With Dynamic Parameter Selectiomm for Phase-Behaviour Matching, SPE Res.Eng., 115-119 (Feb., 1990).

348

9. Application i,m Reservoir Sjrnula(jo,u

9.6. Refere,mr-e.a

349

42. Gani, R. and Fredenslund, A: Thermodynaunics of Petroleuunm Mixltmres Containing Heavy Hydrocarbons: An Expert Tuning System, Ind. Eng. Chem. Res., 26(7), 1304-1312 (1987). 43. Danesh A., Gozalpour. F., Todd, AC. and Tehrani, D.H: Relimmhle Tunmminsg of Equmumlion of State with No Binary Interaction Parameter, Proceedings of the lEA Conference, Australia (1996). 44. Danesh, A., Tehrani, D.H., Todd, A.C., Tohidi, B., Gozalpour, F., Mmmlcolusn, K., Reid, A., Bell, K., Elghayed, K. and Burgass, R: Phase Behaviounr And Properties Of Reservoir Fluids, Proceedings of the UK DTI EOR Seminar, London , Eunglaund , Jumne 19-20 (1996). 45. Kossack, CA. and Hagen, S: The. Simulation of Phase Behaviounr mind Siitsm Tube Displacements with Equation-of-State, paper SPE 14151 presented at tine Olln SPE Annual Technical Conference and Exhibition, Las Vegas, NV, September 22-25 (1985). 46. Mansoori, I., Haag, G.L. and Bergman, D.F: An Experimental and Modelling Study of the Miscibility Relationship and Displacement Behaviour for a Rich-Gas/Crude-Oil Systenni, paper SPE 20521 presented at 65th Annual Technical Conference and Exhibition of the SPE, New Orleans, LA, September 23-26 (1990). 47. Khazam, M., Danesh, A., Tehrani, D.H. and Todd, AC: Dynamic Validation of Phase Behaviour Models for Reservoir Studies of Gas Injection Schenres, Proceedings of Ihe Society of Petroleum Engineers 69h Annual Conference, New Orleans, (Sep. 1994). 48. Bardon, C. and Longeron. D.G: Influence of Very Low lnterfnnciai Tenmsion mm Relative Permeability, SPE 1., 39 1-401, (Oct. 1980). 49. Nghiem, L.X., Fong, D.K. and Aziz, K: Compositional Modelling witim an Equation of State, SPE J., 688-698, (Dec. 1981). 50. Coats, K.H: An Equation of State Compositional Model, SPE J., 363-376, (Oct. 1980). 51. Amaefule, J.0. and Handy, L.L: The Effect of Interfacial Tensions on Relative Oil/Water Permeabilities of Consolidated Porous Media, SPE J., 371-381, (June 1982). 52. Bette, S., Hartman, K.J. and Heinemann, R.F: Consnpositiomiuil Modelling of lunterfuncial Tension Effects in Miscible Displacement Processes, J. Pet. Sd. Eng.. 6, 1-14, (1991).

57. Wang, L. and Gino, T: A Unified Viscosity Model for Hydrocarbon Gases and Liqunids Based on Transposed Patel-Tcja Equation of State, uluagong Xuebao I Jourumal of Chemical Industry and Engineering (China), English Edition, Vol. 6, No. I, 38-49 (1991). 58. Loinrenz, J., Bary, B.G. and Clumrk, CR: Calculating Viscosities of Reservoir fluids from Tlmeir Compositions. JPf, 1171-1176, (Oct. 1964). 59. Jossi, J.A., Stiel, LI. aumd lhodos. 0: The Viscosity of Pure Substances in the Demmse Gaseoums mind Liquid Phases, AIChE J., 8, 59-63, (1962). 60. I lernning, F. and Zippcrcr. L: Cmulculation of tIme Viscosity of Technicuni Gas Mixtures fronsm the Viscosity of Individual Gases, Gas U. Wasserfach, No. 49, (1936). 61. Dandekar, A., Dumumeslm, A., Tehrani, DII. anmd Todd, A.C: A Modified Residual Viscosity Metis(xI for lnnproved Prediction of Dense Plmase Viscosities, Presented at the 7tin European improved Oil Recovery (IOR) Syusmposiumn in Moscow, Russia, October 27-29. (1993). 62. Lusmrtz, RB: Quuantitmntive Evalumalion of Numerical Diffusion (Truncation Error), SPF J., 315-21) (Sept. 1971). 63. Smmge. 13.1 I. mmd Berry, V.M: Puhlicationr, (1971).

Plmuusc Equmilibrimi in Hydrocarbon Systeimms, API

64. Khazaunn, M:Applicmmtionm of Pimmm.sc Behaviour and Flow Models to Gas innjection and Gas Condenmsuite Recovery Processes, PinE) Iisesis, Ileriot-Watt University, Edinburgh (1994). 65. Reffstrunp, J. ammd Olsen, Fl: Evaluation of PVT Data from Low Permeability Gas Condensate Reservoirs, North Sea Oil and Gas Reservoirs - lii, 289-296, Kluwer Academic Press (1994). 66. Fevang, 0. antI Wlmitson, Cl!: Accurate Insitu Conmmpositions ins Petroleunn Reservoirs, SPE 28829. Presented at the European Petroleum Conferensce, London, 25-27 October (1994). 67. MmncMillan, Di.. Ginicy. G.M. mmnd Dennmhicki, Jr., Il: ilow to Obtain Reservoir Fluid Properties frons an Oil Sanmple Contamtniunated witin Synthetic Drilliimg Mud, SPE 38852, Presemmted at the 1997 SPE Annual Technical Conference and Exhibition, San Antonio, 5-8 October (1997). 9.7 EXERCISES Tine reported cousnpositioun of a reservoir oil is as follows:
CI C2

Pedersen, KS. and Fredenslund, A: An improved Correspondimng States Model for Prediction ofOil and Gas Viscosities and Thermal Conductivities, Chens. Eng. Sci., 42, 182186, (1987).
53.

9.1.

Component

Ely, J.F. and Hanley, H.J.M: Prediction of Transport Properties. I. Viscosity of Fluids and Mixtures, I&EC Fund., 20, 323-332 (1981).
54.

Oil, umiole %
C7+ Properties:

54.5t)
M=209

8.09

C3 5.82 S=t).8323

C4 0.78

nC4
2.17

iC5

nC5

C6

C7+

0.94

1.65

2.39

23.66

55. Aasberg-Petersen, K., Knudsen, K. and Fredenslund, A: Prediction of Viscosities of Hydrocarbon Mixtures, J. Fluid Phase Equilibria, Vol. 70, 293-308, (1991). 56. Little, J.E. and Kennedy, H.T: A Correlation of the Viscosity of Hydrocarbon Systems with Pressures, Temperature and Composition, SPE J,, 157-162, (Jun. 1968).

Describe time oil by tinree pseudo components for application in simulation of a lean gas immjectionm process. 9.2. Estimate the viscosity of a gas mixture composed of 90 moi% Cu and 10 mol% nC al 10 377.5 K and 34.47 MPa. The unmeasured viscosity is 0.052 mPa.s (cp). 9.3. E.stinnumte time viscosity of tIre reservoir oil sample described in Exercise 2.1 at its bubble point using tIne LBC nsmetliod.

350

9. Apphcatron in Resenoir Sinu;ilation

9. 7 Exercises

351 1.6424
561 kg/mum.

9.4. The reservoir oil in Exercise 9.5 was flashed at the reservoir teunperature of 373.0 K and pressure of 20.79 MPa. The oil was described by methane and two component gronmps, tusing the equal zlnM metirod and molar averaged properties, as given in Fxamnmple 9.3. tlne predicted results by a phase behaviour model, unsing the above flumid description, are givemn in the following table. Calculate the composition of equilibrated phases in terms of tlme original components. Group/mole trac.
I (unmeulmane) II 111 Liquid mote

2.0934 7.00 2.9968 * Summuraled liquid wimtn a tcnsiuy

13.89 lt).45

42.56
3t.57

17.31
of

Differential liberatioum test at 373.1 K.

Pressure MFa 34.92 R d Solution 5 Gas/Oil Ramin 405 356
0 ,j Retamivc 0 Volunnc Oil

Oil 0.446 I I 0.28748 0.26641 Iraction=0.62975

Gas 0.79450 0.20083 0.00467

cud, Relative Tonal Volume

2.342 2.368

Z, LiberatedGa.s Compressibiliny Factor 1.040

B . Gas 8 Formations Volume Factor 0.00406

33.54 32.t6
30.44 28.72

2.342 2.168

121
286 258 23)) 203 177 138 lIt

2.t)5t
I 940 1.854 1.759 1.689 1.622 1.5)1 1.441

2.395
2.429 2.487 2.554 2.66! 2.787 3.220

t.t)07
0.956 0.948

0.00410
(1.110411 0.1)0433

9.5. The following set of experimental data is avumi)umhle on a volatile oil. lmuse a pimase behaviour model to the measured data, and consmpare the predictions of tuned and untuned models with the experimental results. Compositional analysis of volatile oil. Component Moie% Molecular Weiglst Cl 57.53 C2 10.16 C3 5.83 i-C4 1.22 nC4 2.06 i-CS 1.01 nC5 1.70 C6 1.40 85 C7 2.16 96 C8 2.55 104
C9 ClO Cli 2.00 1.55 1.10 116 131 148

26.3t) 24.24
21.48 17.34 13.89

0.916
0890 0.818 0.836

0.01)456
0.00481 0.00511 0.00639

Specific Gravity

0.45
7.01)

86
64 38 25 4 1)

1.377
1.329 I 252 1.212 1.168 I .076

3.767 4.823
7.1 12 13.777 27.571

0.839 0.862
0.907 0.927 0.969

0.0079! 0.01082 0.01699

0.03422 0.06956

3.55 1.83
0.79 I

63.435

0.963
1.000

0.15944

0.67 I 0.726 0.756

0.776 0.781 0.791

3.0352 grams of condensate was collected Iromnu Ihe liberated gas at 288 K and 0.1 MPa of mercury pressure.

Sepmmrator test at 293.6 K. Prcssure (his / Oil Ralio

5,27 0.79 0.10 225 44 17
lIne slock

Formnalion

Voluuuie

Separator Volume 1.132 1.039 1.004

Cl2 Ct3 Cl4 ClS

C16 C17 Ct8 Ct9 C20

1.00 0.99 0.78

0.85

162 75 187
201 218 229 243 260 419

0.798 0.813 0.832

0.831 0.837 0.828 0.839 0.847 0.903

romal
Densiuy
of

0.72 0.49 0.60 0.5) 3.81

286 1.764 mask oil at 288 K = 81)) kg/sun

(I) Separamor flashed gas volunnne (sc) per virliunune of smock lank oil.

(2) Votunise of samurated oil an 34.92 MPa and 373.1 K per volume of stock tank oil.
(3) Voluunc of oil at separalor pressure and meunperauure per volume ofstock tank oil.

Constant composition expansion tests at 373.1 K. Pressure Relative Vol. (VfVsam) Liquid Vol. Frac.
MPa 35.96 35.82 % 09954 Iorwmmrd conntmmct cxperinmmemnlal datum witln mmncthanse at 373.1 K and 35.26 MPa. llnasc 100.00k Oil 57.87 787 4.89 Gas 78.24 7.57 4.04 0.79 1.28 Oil 57.03 9 fit) 5,4)) t.tS I 98 Gas

0.995 8
0.9970 0.9978 1.0000 1.0394 1.0872

35.6)
35.34

Oil 57.10
9.71

Gas
72.25

Oil 56.73
10.01

Gas
71.47 10.00

34.92 31.13 27.68

24.24 20,79

._c3~ponenm
C 1 C 2

83.66
75,00 68.14

74.46
8.91 4.69 0.92 1.49

1.1508
1.2470 1.3948

17.34

60.10
52.54

C~ i.C 4 nC
4

5.69
1.19 2.03

.06
1.85

9.72 5.06 0.99 .60

5.80 1.21 2.06

5.22 1.02 1.65

352
i-C 5 nC 5

9. Application in ReservoirSimulation

353

C C6 C7 8

2.19 9 Cno 1.66 Cnn 1.30 0.41 1.16 Cn2 1.14 0.33 1.07 0.41 1.09 Cns 1.19 0.29 1.10 0.36 1.05 11.27 1)81 ~ 0.96 0.23 0.87 Cns 1.05 0.23 0.97 0.28 (1.92 Cm6 0.86 0.18 0.74 0.21 0.76 Cmi 0.66 0.11 0.63 0.15 1)54 C 0.76 0.13 0.67 0.16 0.65 C15 0.65 0.10 0.61 0.13 0.56 19 0.33 4.17 C , 4.92 0.26 4.57 20 3 V~d, cnn 90.00* 22.50*9 40.52~ 90.00 22.50* 3 67.98 43.71 Vequ). cm 67.98 43.71 34.39 3 61)4.3 312.0 Pc kg/rn 604.3 312.0 580.9 * Volume of fresh oil contacted with a density of 563 kg/ni. ** Volume of methane contacted at 35.26 MPa.

0.95 1.62 1.41 2.26 2.76 2.23 1.69 1.36

0.59 0.97 0.75 1.04 1.12 0.82 0.58 0.36

0.99 1.68 1.41 2.32

2.6.5

0.69 1.12 0.85

1.23 1.26 0.98 0.69

1.01 1.70 1.41

2.27 2.55 2.11 1.52

0.74 1.20 0.92

1.33 1.36 1.05 0.77

1.02 1.72 1.42

2.27 2.58 2.11 1.47

0.76 1.24 0.95

1.36 1.42 1.08 (1.74 0.54 0.41

0.51 0.39 0.47

0.22 (1.31

1.18
1.08

1.03
(1.82 0.90

0.41)
11.31 (1.32 0.25

0.24 0.16 0.18

0.15 1)38 25.01) 27.77

0.75 (1.54 t).64

11.55 4.13 51)00* 48.96

11.16 m).19 11.15

1)39 2t).tR) 211.92

350.5

564.8

358.5

APPENDICES

Appendix A: Tables A. I A.2 A.3 A.4 A.5 Properlies of Ptmre Counspounds Properties of Sinmgle Carbon Nuumnber Groups Universal Gas Constarmt Binary Interactiomm Parameters Conversion Factors

Page Nunmher 354 356 357

358 363
364

Appendix B: Critical Property Correlations in Field Units Appendix C: Eqtuatiomm of SLoe Expressuomms Appendix I): Equilibriumusm Rumlios

368
372

Equilibriums Ratio CInumrts at 501)0 psia (34.47 MPa) Convergence Pressure (WA Copyrigtmt. Reproduced willn permission from: SI Lngineering E)ata Book (1980).

354

Appendux A

lable.s

355

Table Al. ~~~rties of pure compounds.

Name Methane Ethane Propane i-Bumane MW kg/kgmuul 16.043 30.070 44.096 58.123 1, 1~

Tumble A. I (Cont.). Properties of pure coummpotnnds. K

190.56 305.32 369.83 408.14

K
111.66 184.55 231.11 261.43

P v MPa umm/kgnunot
0.0986 0.1455 0.2000 0.2627 0.2550 0.3(158 0.3036 0.3130 0.3664 0.371 0.428 0.486 0.544
0.601)

Z,

acenlruc factor

R,uckcu
~

virmctnu ur

SpUr. 0 3()0()

Naunme Llhylenc Propylene I -Bulene cis~2-t5ulcne urans-2-Bulenc Propadiene I 2lltnlaulicnc I 3.tlulaslicnuc l-lcuuuene cis.2.Pcntenc tuimns-2-Penmeune 2.Mentuyt-l-Buulcnc IMeslnyl I -l(uHcuue 2 MelIlyl 2.l)nnicuuc 1-tlesenc I-HeplCume (yctopcnn~unmc

MW kg/kgunol 28.054 42.08) 56.107 56.11)7 56.107 40.065 54.092

54(192

T, K 169.47 225.43 266.9 276.87 274.03 238.65 284

26874

I, K

P~ v MPa m/kgmol t).129l 0.1810 (1.2199 0.2340 0.2382 (1.1620 t).2l90

(1,2208

4
0.2767 0.2753 0.2765 0.2717 (1.2742 0.2711 0.2670
0,2704

n-Butane
i~Pcntanc Neopennane n-Penlane 2-Melhylpcnlanc n-Flexanc n-Ilepnane n-Ocuane n-Nonanc
n.Dccane

58.123
72.150 72.150 72.150 86.177 86.177 100.204 114.231 128.258

142.285
156.312

n.tlnslrvaoe n-Dcsiecane n.Tridecaune n-Tetradecane n-Penmadecane n-Hesadecane n.tlepuadecane n-Ocmadecane n-Non~Jecane n-t2ico.sane n.Heneicosane

4.599 4.872 4.248 3.648 272.65 425.12 3.796 300.99 460.43 3.381 282.65 433.78 3.199 309.22 469.7 3.370 333.41 497.5 3.010 341.88 507.6 3.025 371.58 540.2 2.740 39883 568.7 2.490 423.97 594.6 2.29(1 447.3 617.7 2.110 469.08 639 1.949

0.2862 0.01)5 0.2894 I 0.2793 0.0995 (1.28128

0.2763 0.1523 0.27664

74.05

112.91
54.03
185.32

0.3562
0.5070
05629

0.2824 0.1770 0.27569 0.2739 1)2002 0.27331 0.2701 (1.2275 0.2706 0.2693 0.1964 0.27570 (1.2701 0.2515 1)26853 0.2666 1)2781 1)2662 0.2659 11.3013 0.26355 0.2611 0.3495 0.26074 0.2559 0.3996 0.25678 0.252)) (1.4435 I) 25456
(1.2465 (1,4923 1)25)17.1

(93.90 229,37 236.01) 216.18) 269.15 276.71 118.44

(1.659
0.775

17t).338
184.365 198.392 212.419

489.47
508.62 526.73 543.83 560.01 575.3

65%
675

1.820 0.716
1.680

(1.2419 (1.5303 (1.24990 0.2382 0.5764 (1.24692

0.2320 0.2262 (12235 0.2199 0.24698 (1,24322 0.2303* (1.2276 0.2190 (17697 0.21431 0.6174 1)6430 0.6861 (1,7174

0.5840 0.6247 0.5974 11.6111 0,657% 0.6638 (1.6882 359 13 0.7070 39u) 57 0.72 9 .1.1) (.6~ (1 7142 482.181 0.7445
522.26 0.7527

282.36 5.032 364.76 4.612 4(9.59 4.020 435.58 4.206 428.63 4.103 391.15 5.470 444 4.50))
425.37 4.331)

aceintric Rackelt paraclmor factor Z , 9 0.0852 0.28054 101.53 0.1424 0.27821 143.02 0.1867 0.27351 (1.2030 0.27044 0.2182 0.27212 0.1596 0.27283 0.2509 0.2685
0.1932 0.27130

Sp.Gr.

0.5000
0.5210 0.6005 0.6286 0.6112 0.5997 0.6576
0.6273

1.570 0.830 1.481) 0.889 1.400 1)944 226.446 1.34)) 1.000 240.473 1.270 1.060 0.2168 (1.8114 (1.22917 771.95 254.500 268.527 1.210 1.120 0.215)) 0.8522 1)2158 813.85 851.67 282.553 1.160 1.170 0.2126 0.9069 0.22811 296.580 1.147 1.198 0.2114 0.9220 1)2(197* 897.64 1.101 1.253 0.2095 11.9551) 11.2068 939.55 n-Docosanc 310.610 n-Tricosane 324.630 1.059 1.3(17 0.21)78 0.9890 0.2038 981.43 n-Tetracosane 338.680 664.4 810.4 1.019 .362 (1.2061 1.019)1 0.2011 023.41) Z,A Iroun 11.151 except those idcnlitied by whictu are calculated frousi lIne Yannada-Gunmn courclanuon. Eq.(l.l3).

693 708 723 736 589.86 747 603,05 758 616.93 768 629.7 781.7 641.8 791.8 653.4 801.3

563,77 606.05 647.43 (u8%.50 730.05

0.7617 0.7633 (17722 (1 7772 0.7797

0.7820

Mcmtnytcyclopennane
Cyclotuesanc Me0nylcycluuhexanc
Lltuylcyctopenmane

1)7869 ().7924
0.7954

1)7981
0.8(8)4

0.8025

Parachor values are no be used only ins Eq.(8.2t).

Ellmylcyclohexaunc ttcnicne Iotmucne Hlmythenneenc cu-Xylcne nu-Xyleuuc s-Xylcnne Niungen Osvgeus Cambon Monoxide
Carbon Dioside Hydrogen Sulptuide

70.134 7(1.114 70.114 70 114 7)) I It 7)) I 14 84,161 98.188 70.114 84.161 84.161 91) 18% 98.188 112.215 78,114 92.141 11)6.167 106.167 1(16.167 106 167 28.1)14 31 999

101 Il 31(1.1)8 309.49 1)1.1.3 291 2) 311.71 136 61 366.79 322.4 344.96 153.87 374.08 37662 404.95 353.24 383.78 4(19,35 417.58 412 27 411.51 77.35 9(1.17

464.78 475.93

3.529 3.654

0.2960 0.3021 (1.2920 (1.3(121 11.292(1 0.3540 0.4130 0.2583 0.3189 (1.3079 0.3680 0.3745 0.4500 0.2589 0.3158 (1.3738 0.3692 ((.3758 11.3791 0(191)1 0.0734 0.0940 0.0985 0.1220 0.0642

0.2703 0.2329 0.27035 0.2790 0.2406 0.2694

0.6458 0.6598

475.17 3.654 0.3021

465 3.400 45(1.17 3 5)6 .17) 1,4)8) 504113 3.140 537.29 2.810 511.76 4.502 532.79 3.784 551.54 4.075 572.19 3 471 569.52 3.197 609.15 3.040 562)6 4.898 591.79 4.109 6(7 17 3.609 63(1.37 3 714 617.05 3.541 616.26 3.511 I 26.! 3.394 154.58 5.043

0.2793 0.2373 0.2697

(1.2568 0.2287 0.2705 (1.2817 (1.2286 0.2705* (1.2535 0.2767 0.2663 0.2653 t).2800 0.2660 0.2616 0.3310 0.2615* 0.2733 0.1943 0.26824 0.2725 0.2302 0.2704 0.2726 0.2118 0.27286 (1.2685 0.2350 0.26986 0.2687 0.2715 0.2667* 0.2701 0.2455 0.2690 (1.2714 0.2108 0.26967 0.2637 0.2641 0.2639* t).2629 0.3036 0.26186 0.2630 0.3 127 0.2620* 0,2594 0.3260 0.2620 0,2598 (1.3259 0.2870

0.6524
0.6563 0.6322 t).6683

0.6769
0.7015 210.05 0.761)3 0.7540 247.89 0.7835 289.0(1 0.7748 0.7712 328.74 0.7921 210.96 0.8829 252.33 0.8743 292.27 0.8744 08849 0,8694 0,8666

0.2917 0.0403 0.28971

0.2880 0.0218 0.28962

61,12

0.8094
1.1421 0.8 180 0.8014 1.3946

28.010
44.01)) 34.082 64.065 2.1)16

81.7

132.92

3.499 0.0931
7.382 8.963 7.884 1.313

0.2948 0.0663 0.28966

(1.2744 t).2843 0.2686 0.3053 0.2276 0.27275 82.00 0.0827 0.28476 85.50 0.2451 0.26729 -0.21500.31997

Sulptucr Dioxide Ilydrogen

Waler
*
* *

194.67 304.19 212.8 373.53 263.13 430.75 21)39 33.18

*

~l 8.1)15

373.15 647.13 22.055 0.0560

7.,, irons II 151 excepl mtnose idenlitied by

w)uicln are calculaued froun Inc

0.2294 0.3449 Yamada.Gunn correlation,

1.0000

tarachc,r values ire to he used only in Fq,(8 .2 I

1.11)

356

Appendix A

Table.s

357

Table A.2. Generalised single carbon number group properties.

SCN C6 C7 C8 C9 ClO CII CI2 CI3 CI4 C15 CI6 Cli C18 Cl9 C20 C2l C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33 C34 C35 C36 C37 C38 C39 C40 C41 C42 C43 C44 C45 MW 84 96 107 121 134 147 161 175 190 206 222 237 251 263 275 291 T, K 337 366 Sp.Gr.

0.690 0.727

390
416 439 461 482 501 520 539 557 573 586 598 612 624 637 648 659 671 681 691 701

0.749
0.768 0.782 0.793 0.804 0.815 0.826 0.836 0.843 0.851 0.856 0.861 0.866 0.871 0.876 0.881 0.885 0.888 0.892 0.896 0.899 0.902 0.905 0.909 0.912 0.915 0.917 0.920 0.922 0.925 0.927 0.929 0.931 0.933 (1.934 0.936 0.938 0.940

I K 510 547 574 .603 627

Table A.2 (Commt.).

7., Acen.F,ncu.
-

Rackelt

Generalised siisgle carbon nunmber gromup properties.

649
670 689 708 727 743 758 770 781 793 804 815 825 834 844 853 862 870 877 885 1(93 901

300
312 324 337 349 360 372 382 394

709
719 728 737 745 753

404
415 426 437 445 456 46-4 475 484 495 502 512 521 531 539

907
914 920 926 932 938 943 950 954 959 964 970 974

760
768

774
782 788 796 801 807 813 821 826

MPamVkgrnol 3.271 0.348 3.071 0.392 2.877 0.433 2.665 0.484 2.481 0.532 2.310 0.584 2.165 (1.635 2.054 (1.681 1.953 0.727 1.85.3 ((.777 1.752 0.831) 1.679 0.874 1.614 0.915 1.559 0.951 1.495 0.997 1.446 1.034 1.393 1.077 1.356 1.110 1.314 1.147 1.263 1.193 1.230 1.226 1.2(10 1.259 1.164 1.296 1.140 1.323 1.107 1.361 1.085 1.389 1.060 1.421 1.039 1.448 1.013 1.480 0.998 1.502 0.974 1.534 0.964 1.55(1 0.941 1.583 0.927 1.604 0.905 1.636 0.896 1.652 0.877 1.680 (1.864 1.71)1 0.844 1.733 0.83~L__.J.749

-(1.268 0.265 0.261 (1,257 0.253 0.251) (1.247 0.244 ((241 11.23% 0.235 0.233 0.231 0.229 0.226 0.224 1)221 0.22(1 0.217 (1.215 0.213 (1.211 0.2(19 0.207 (1.2(15 (1.203 0.21)1 0.199 0.197 0.196 (1.194 0.191 (1.191 0,19(1 0.188 0.187 (1.185 11.184 0.181 0.180

4,
(1.269 0.266 (1.263 1)261) 11.256 0.253 (1.251) 0,247
11.244

((.251 (l.28() 0.312 0.352 (1.31)9 0.429 11.467

(1,61)1 (IS (6

Ahrneds ConTelation: 2 1 0= A-m-A (C,)+ A~(C,,) + A (C,,) +A /( C,) 2 4 5 where, 0: Any plsysic;ml property
(,,:

Cuirhon grotip nmtnmnnher

(1.571 0.610 (1.643 (1.672 (1.698 0.732 0.759 0,789 (1.815 (1.841 11.1)74 (1,897 1)944 (1.96% 0.91(5 1(11)8 1.026 1.1146 t.(t63 1.1)82 1.095 1.114 1.124 1.142 1.154 1.172 1.181 1.11)5 1.2(17 1.224 1.232

(1.24(1 0.237 0.234 0.232 0.229 0.226 0.224 0.221 0.219 0.217 0.214 0.212 0.200 11.198 1)196 0.194 0.193 (1.191 (1.189 0.188 (1.187 (1.185 (1,184 (1.11(2 0.181 0.180 11.179 11.17% 11.177 0.175 0.174
-

w itim connslusnmls as follows:

----~.--.-.------,-.~-L
.~

~&.__

M 1~ R S
Cu)
T~, R

131.11375
434.38878

24.96156
50.125279

0.341)79022

2.4941184 E-3
7.0280657 E-3 2.49433118 E-8 1.4663890 E-5 5.8675351 E-3

0.86714949
0.50862704

3.414341)8 [-3

8.7(8)211 [.2
41.421317

psia v ,ft/Ib 1
P,,

915.53747 275.56275 5.223458 E-2

0.9(127283 2.839627 [-5 1.8484814 E-5 0.75861(59

12.522269 7.8709139 E-4

0.29926384

2.8452129 E-3
1.7547264 12-i

1.9324432 E-5

.468.32575 401.85651 1.1627984 1.8518106 1.3028779 El 1.7117226 E3 4.4017952 E-2

Time correlation gives the calcunlated critical properties by Whitson using the Riazi-Daubert correlaliotm, Eq.(6. 14), and not tlmose given iI~ Table A.2.
*

Ahmed, T: Elydrocusrbon Phase Belnaviour, Gulf Publishing Conmpany, Houston (1989).

Tumble A.3.

Universal gas constani values.

P. Pressur numb bar MPa psia I.,IJ_~r~l2r~,,~ 4 K cuui/guunuul K lil/gmot K m/kgmol R ft/lbmol R 1(2.0567 0.083144 0.0083144 10.732

Ic, Pc. and ye: Calculaned from Twu correlations. Eqs.(6.23-25).

Zc: Catculaied from Pcvc=ZcRTc Acenu.ric factor: Calculated from the Lee-Kesler correlation, Eqs.(6.9-l0). ZEA. Calculated from the Yamada-Gunn correlation, Eq.( 1.13).

358

Appendix A

TabIe.c

359

Table A.4.l. Binary insteraclion paranseters for Zudkevitch-Joffe-Redlicls-Kwonsg c,~,atmommof stu~c. ~..-, 3 ~,,,4 , 6 ~ 8
I 2 N 2 (02 C 1 I2lhylcnc ((000 .0601) .0280 1(1(10 .06610 (XXX) .0760 - (8)0

Iumhle A.4.2.
_______

10-24

3
4

5 6
7 8

.1090 .13(10
.1370 . I 30(1

.11(8)1) 1)000 .0(8)0 -.0031) (8)111)

.11(8%) .1)0211 .0010 .1)040 .1819(1 .0380 .0350 (111)1)) .11(118) .0005 (111)6 .0(11(1 .0(110

Binsary insteraction pumraummelCrs for_S~mvc~Rcdlnclm~Kwofl~e~Iuat~.?n_of state. 2 3,~4 (.onn,puuieuuu 5 6 7 (8)11(1 I N 2 (611)1) 011)1)1 2 ~ 1)278 111)7 0)11)0 3 ~l (((0(1 .1(118) ((18) .00(1) 4 Pn)uylene .0407 .1163 -(8)78 ((026 .0(1(111 S ~2 6 7 8 9 Propylene C 3 iC 4 nC 4 ((80)1
.

_______________
8 9 1(1-24

9
It)

Propylene C 3 iC 4 nC,~
iC~ Neopent~ne nC 5 nC 6 Met Cyc Pent Cyc Hex nC 7 Met Cyc Flex Totumene o-Xytene nCg

.1(11)0 .1240
.1200

.I6)()
16(8) .1600 .1870 1900 1900 1900 1900 1900

.13)81
.1(1110 .101)0 .1000 .1000 .1000 . 1(1(11)
- 10(8)

.0(16(1 .0050 .0190 .0230

.0230 .0190

.111)01)

11)1%)

.0289

.0000

1)200

.00(1)) .0000 .0000 .0000

.0(88)
.0(1(10

(1)8%) .0001) 1)1(1)) .006)) .0004 .000(1 (1(3(1(1 .0(8)0 .111)00 .00)1) .00(8) .0000 .040)) .0)100

.0763
.0944 .0700 .0867 .0870 01(78 .141)))

1(81)1
1001) 111)8) .1(1)8) .10(1) .I(1(l() .1(1)81

.0080
.0241 .0(156 -.0078 -.0078 .0019 .1)374

.00(1 .11(81) .00(10 .0000 .00(1)) .0011) .0)1%) .0(1)))) .0(8)0 .001))) .0000 .035)) .0000 (1111)1) (8)00 .111)01) .1)00(1 1)300 .0001) .1811)1) .0000 ((1))))) .11)18) 1131(1) (XXI) .1881)) .0000 .11(11)0 .01)01) .0(100 (1001) .00(8) .0(11)0 .0001) 31000 .11(1%)

.0(180 -.0220 .0033 (8)00 .0900 -.0010 -.0144 -.010

. 1(1(1))

.181(111
.0101) .0100 .0)01) .011)0 .01(10 .01(8)
.011)0

(1067 .0050 .0050 .0056 -.0156 .0230

.0000 .0000 .0000 .0050 .0050

.0(88) .0(178

.0000 .0000 .0000 .0000

I)) II 12 13 4 IS lb 7 8 19 21) 21 22 23
24

iC

.1)120 .0120 .0120 .0140

II
12 II 14 IS

.0190
.03(10 .0130 .1)130
.0131)

.0185
.0380 .1)100 .1)100
.1)101)

.01)01)
.18)00 .000(1 (81(10 .0001) .01)00 .0000 .0400 .111)00

Ncopenlnmne n(y

.0078

nCti Men (ye Penn (ye Hex

.0000 .0000 .0230 -.03(8) .0204 -.0022 .0000 -.0111 .0030 .0030 .0000
.0005

.141)0
1400 .1422

.1(1)10
.

.1)418)
.03)17 .1)5(11) .0978 .1(11)0 .11448 .1(448 .0551) 1)600 1)650

.0140 .0330 .0144

.0150 .0300 .0250 .02(10 (1201) .0301) .04110 .0450

111)11)

.0333.1)151)

.0050 .0050 .0100

.0100

.0000
.0000 .0000

.0000 .000(1 .0001)

.0000 .0000

16
17 18 19 20 21 22 23 24

nC~
MelCycttCX Toli.mene o-Xyleuue nC1(
nsCq nCIO-n(14

.10(1)) toOl) .11)111) .1(1(10 1(10(1 10(11) 1000 . 11)110 .1000

.0411
.0230 .0900 .0500 .1)170 .01 70

.11)00
1000 .1000 .1(1(1(1
.

1900
900 .1900

.02(8) .0200 .0270 .0150

.0100 .0100 .01)10 .01)1(3

.0101) .0100 .0201) 1)100

.1450
.1501) 1500 ISO)) ISO)) I 51)1) 15(1(1 .1500

.0300 .0300
.0101) .1)101)

.0044 .0005 .0050 .0005 .0300 .0200 .0300 .0200

.0040 .0040 .0015 .0015

.00(8) .0100 .0100

.0(11%) (101)0 .001(1 .0010 .0015

.1)000 .0000
.0000 .00(10 .0(8%)

nC 19(10 . 000 .1)15)) .1)1(10 .1)1(11) .0000 .1)1)11) .0(11)1) .0220 .0000 9 nC n( 9(11) . 0(10 .0100 .0060 .008(1 .18)0(1 .0000 .0001) .1)001) .00(X) 10~ 4 uuCIy-n(,t .1900 .1000 .0101) .0080 .0000 .0000 .0001) .000(1 .0001) .0000 9 nC -nC .1900 .1000 .0000 .0000 .0000 .0000 .001)0 .00(11) .0000 .0000 20 24 Fronn: Yarhorosmgh. L. : Applications of a Generalized Equation of Slate In Penroleuun Reservoir Fluids, Equalions of Stale in Engineering. Advances inn Cliemislry Series. Ediled by K. C. Chnuo auud Robinson. R. L., Aunmerucan Clnetniical Society, Washungmon. DC, 182. page 385-435(1979).

.0200
.035)) .0400

.0150
.0250 .0300

.0040
.0005 .0010

(8)20
.0025 .0050

.0000
.0000 .0000

nCls-nCl9
nC Q-uiC24 2

Prouun Knapp, II and Dorimmg. R. : Vapour-liquid F.quitihrua for Mixtures of Low Routing Substances, tserticns. 1) and Eckcrsuiani R.. Eds(l)ceheunsa Chemusustry Data Ser.), Part I- Binary System (1986).

360

Appendix A

Tables

361

Table A.4.3. Binary interaction parameters for Peng-Robinson equation of state. No. Component 2 3 4 5 6 I N .00(8) 2 2 C02 .0000 .0000 3 C .0311 .1070 .0000 1 4 Ethylene .0500 .1200 .0215 (88)0 5 C .0515 .1322 .0026 (8)89 (11)00 2 6 7 8 9 10 II 12
I) 14 Propylene

Iahle A.4.4.

Binary interactions parausseters for_Patcl-Teja equation of state.

7 8 9 10-24 No.
Coumipoiucnt

I
.111)11(1

2 .00(11) .11930 .11(11) .128)1 .131)1) .12811

. I 270

10-24

N 1 4 5

C 1 [tluylenne

.11600 .0320 .1(401) .06)))) .0800 .074)) .0540 .1)311)

(((1(1(1 11080 .11050 3)090 .0040

.1111(11) .18)11)

.11(8%) .0050 .0(8)2 .111)1)) .01)10 -.0010 -.111111) .1)0(11) .0(1(1(1 .11(811) .0(11%) .0001) .001)1) .0(11)0 .0450 .0000 .00(1(1 0000 11005 .0()06 .0010 .01)11) .18)18) .11(8)1) (8100 .1)000 .0(8)0 .0000 .0000

C~ iC4 nC~ iC 5 Neopentane nC5

nC6 Mel Cyc Penn Cyc Hex

.0600 .0852

.1300 .1241

.0330 .0141)

.1001) .0711
.1000

.14(8)
.1333 .1400

.0256
.0133 -.0056

(1)1(1)) .011)0 .021)0 .02()0

.111)89 .18)1111 .111)11 .111)10 -.18)67 .18)8(1 .0096 .1)080 .01)80 .0078 .1)140 .1)141 .1)141 .1)150 .0160 .0080 .0080 .0(180 .0100 .1)11)) .1)12(1 .0121) .0141) .0150 .02 10

6
.0(101) -.18)78 .1100 .01)33 .11111 .1)111 .1)1 2() 0267 .0270 .0270 ((561) .0580 .0600 .0(100 -(11)4 -.111)40 .1)1)211 .1)21(1 .11242 .0242 .1)25)) .025(1 .0(11(t) .1)171) .0171) (1171) .1)1 74 .0)80 .1)1 SIt .019(1 .0200 .0001) .00(111 .001)0 .1)0110 .1100(1 .1)00(1 .0000 .0001) .0001) 7 8 9 10 II 12 3 14 IS lb

Propylene

C~ iC 4 i( 5

nC,s

.0250

OIl))
(1111) (8)1)1)
11100

1150 .1251)
.

.11201) .030(1 3)02(1 .0301) .0020 .0300 ~0t11)) .031)0 .1)11(0 - .1(1 (Xi ((360 .1137(1
.11371)

3)000 .00(11) .0100 .0060 ((004 .11)1(11) .181%) .18)0(1 .1501)0 .01)00

.1111(1 .1)000

.0001)
.11(811) .0001)
.11(81))

(81(11) .0000 (1(8%) .111)11) .111118) (88)0 (XXX) .0(100

.0188) .1111(1) (XXX) .1111)8) (1(1111) .111(8) .111)01)

.1(100
1000 1496 150(1 .1500 .1441 .1500

.1400
.14(8) . 1450 .1450

-.0056 .0250
.0236 .1)422 .04511 .025(1 .1)3(8) .1)31(1 .031(1 .03))0 .0300

Nciupcnttanc n(5 nil 6

. ((XX) .I 350
.14181

IS 16 Ii IS 19 20
21

.1450
.1450 .1450

.0450
.0352 .0450

Mel (ye lenn (ye lIes

012)) .1)14(1 .0160 .0181) .03(11) .1161(0 .1)400

.14(11) . (401) .141)1) 1401) .1500 .16(X)

.

.0400 .0401) .0400 .0401) .0400 .0400 .0400

.0000 .00(1(1 .1111(1(1 .0(100

nCi Met Cyc Hex loluene o-Xylene nC 1(

nC9

17
I 8 19 20 21 22 23

.1700 .1500
.1500 .1550

.1800
.1400 .1400 .0145 .0145 .0145 .0145

.0600 .0470 .0470

.0474 .0500 .0601) .0700

.0400 .0200 .0300 .0160 .0300 .0160

.04(10 .0451) .0500 .060(1 .0190 .0300 ,()4(10 .0500

.01St) .0150
.0200

.0590 .0590
.01)71) .02(11) .1)251) .03(11)

.03(8) .026(1
.0260 .01)611 .01(8) .0150 .1)21(l)

.011(1 .012(1
.0120 .1)1(1(1 .111)1(1 .1)1)1(1 (NtIS

nC7 Met Cyc Hex l3uluemme

o-Xylcne

.0371) .0380 .0620 .0620 .0600

.0400
.0501) .0400

(XXX)
(XXX) .00(11) .01111(1 .01)181 .00(11)

.1)400
.11118)

.0000 .0000 .0400

.0001)

nC

1(

.1400

.001)1) .0350 .001)0

.0000 .0000 .03(11)

.01)01)

.000(1 .0(8%) .0004)

.00(10

22 23
24

nClO.nCl4
nCl5-nCl9

.1550
.1550

.0250
.030(1

nC2(XnC24 .1550 .035)) From: Knapp, H. and During, R. : Vapor-Liquid Equilibria ftsr Mixtures oIl.ow Rolling Subsuances,
Berhens. D. and Eckerman R.. l2ds(Dechema Chemistry Data Ser.), Part I- Binary System (1986).

24

nC .0450 . 4(8) .1)11)11 .0400 .0000 .011(8) .00(11) .0001) (811%) (XXX) nC -nC 14 .0500 .1501) .1)4311 .0600 .0000 .11001) .0000 .0000 (XXX) (1811) 0 nC ).nC 9 (15511 .18(10 .)t8t)0 .0800 .000() (8)00 .0000 .0(101) .0000 .11(11%) 1 un( -nC24 .061)0 .2(100 .128(1 .100(1 .11(8)0 .0000 (1000 .0000 .0000 .0000 20 Fronni Willnnnmn. 1) 1 nnnd Tejnm. A. S. : Conminuouus Thermodynamics of Phase Equilibria Using a Multivarame Distrihuiion Funclinin nmuuul an Equation of Slunle, AIChE J, 32(12), 2067-2078 (1986).

362

/tppendin .4

lable.s

363

Table A.4.5. Interaction parameters of C02. N2 and l-12S binaries for Valderrama-Patel-Teja equation of state.
Cornponentjfl~arbonsDdcN~~ Metttane 0.092 0.035 Ethane (1.134 0.038 Propane 0.128 0.070 u-Butane (1 I 26 1). I 34 n-Butane 0.138 ((114 n-Penlanc (1,141 (11)88 n-Hexane 0. 11(1 (1 ISO n-tleplane (1 I 0) (1. 142 Carbon Dioxide -0.036 ,j~~oen All hydrocarbon-hydrocarbon DIP=0 I-lydr. Sulfide 0.08)) 0.095 0.088 (1.050 0.050 0.047 0.047 0.047 0.081) 0 176
-

Table AS.

SI nmetric conversion factors. API - aImn(sld~ x

~
I .1)1:3 250* I .0~ .589 873

feAPI)
12 + 05 12 + 05

=~/cm = Pa
= Pa = nss

bar hhl (US) hhl/I) Iliti

1

x x

1.589 873
1.055 056

12-01 12 -01 E + (8) 12+00 E +00 12- 13 E - 1)2

[200

CI) (ccnti poise) cSI (cenli Stoke) d (darcy)

dyne dync/cnmi ft 1t II II /bhl

x x

C+273.l5 I 1)*
1.0*

Os/S kJ =K = umPa.s
= = = rntn/s = =

x
x
<
x x x

9.869 233 1 0~
1.0

m nmN

=rnN/m

3.048k 9.291) 304*

9.290 304

E-0l 12 02
12 + 04

=rn = ma
=

Table A.4.6.

nsmm/s

Interaclmon paratnleters and coefficients of usnellnanol and water hiusarics for Vaklciratmta-Patel Ieja equation of state with non-randonn mixing rule, Eqs.(4.86-89). ___________
Methanol
Conuponenu (I) Methane

k,
((.2538
-

Waler l~~
-

5 fI /]hmin 1~~
49.0)) 45.4)) 39.30
17 4))

x
><

l
0.7319 0.0519 0.0779 0.320) ((.29 17 (1.7908 0.5607 0.4592 0.5331 0.0700 1.0440 (1.1133 ((1)001) 0.1)1)111)

k 0.5021) ((.4974 (1.5465 (1.5863 0.581(0 0.5525 (14577 ((.4165 0.3901 0. (96S (1.4792 0.1382
**

I~
1.1)11)1) 1.4870 1.6070 1.7861 I 6885 1.61118 .5710 1.521(1 1.5200 0.7232 2.6575 0.380) 11(1)8)))

2.831 685 1.781(17(5 6.242 796 (F-- 32)/1.8

(F +

12 - 02 12 (II 12- 02

= bin

nls/m 1 =rn /kg = C

=

Ethane Propane
-Otitane

00(37 (1.1233 ()l46S ((.2528

n-lltutane n-Pentane

n-llexane
nHcptane

0.2245
0.1461 0.1403 0.1(510 0.2484
00694

n-Octane Carbon dioxide

Nitrogen Hydrogen sulfide

6.88 21.70 0.01) 17.6(1 0.111) 58.28 17.54 27.17 36.91 11.56 7.22

33.57 23.72 3 I .4 I 3S.21 35.11

2174

0.01)
0.111) 0(1))

Methanol
Waler

1) 0000

64.46 I 3. 2t
(1)1) ((.1)1)

ol propaneof

0.00(1)) 0.1(000 nietlian,,I ueiniperaturc ulcpendeuiu: kpin=O 0278+11(1111(911 (I 271Y

waler-uiietltasot us tcunperatunrc ulepettulcnu kiumw-0. 111(1 + 0.00(11115) I 271)

1 =m >< 2W =rn =un~ ~< fi,45( fi* 1 jun. x 1.638 706 = mm hf x 4.448 222 =N hf. s/rn x 6.894 757 = Pa.s lhnns x 4.535 924 =kg 3 htmi/ft x 1.601 846 =kg/m mmml =rn~ x 1.0* nsmuisl Ig=torr x 1.333224 =MPa psi ~< 6.894 757 ~~MPa =K x SI (s)oke) >< 1.0* [2 1)4 = mnm/s t ~SIt usml l,cltxcs: nnn(uniullt)=E 0.1 k(knlt)=l: + (1.1 M(nsnega)=13 + 06 * (ouuvenu,ious laclor u.s exacl gal (US) i Is. inn
x

459.67)/1.8

3.785412

E-03 12-02 E-04 12 05 12+00 12+03 E-0l 12+01 E-06 E-02 12-03

364

~t,ie,udit R

(ritical lropertie. Correlal,alu.s

365

CORRELATIONS IN FIELD UNITS

FOR

ESTIMATING

CRITICAL

PROPERTIES

Riazi-Daubert Correlations 0=a [exp(b0 +cO, +d0 O,)J0~O~ 1 1

I

The units of temperature, pressure and volume are Rankine, psia. and ft /Ihnmol, mespeclively, in all the following equations. The specific gravity, S. is defined relative to water at 60 F. The correlations in SI unit are given in Section 6.2.

_________________
I~Tt,.S M l,. S 554.4

where, a to f, are constants for each property as follows: _________ _________________________________________________ 0 ~l ~2 a c d e

106443
-

~
1.3478
5

-0.544445995k I(t~ -((.6164 I -4.8(114 O,3O84 -026404 0(1

l0-~

0.81067 - 0.53691 (1.2998 0,4844 0.8(1663 0.750(1

-

1.0555 4.0846 1.6015 1.21)28

It, ~ M S

6.162 x Itl~ 4.621(3 s ml)~ (.23) s I ((.4 .204, x 1(12 581.96 6.77857 -.

-4,725 5 1(1.1 -1.1(078 lO~

-l .4679 s 10 ~

3.1939 x p~3 ((.0 1.095 x l0-~

Lee-Kesler Correlations

I vIM) l~, S

341.7 + 81 IS + (0.4244+ 0.1

I74S)T~,+ (0.4669 3.2623S)

x~

/T

(vIM) M S M l~, S S (71) < M < 30(1 0.1.5

-2.657 x tO (1.5287 2.6(112 x lO~ (1.2(0711 -1.3036 543076x I))~ -9.53384 1.111)56 x lt)-~ 0.97476 6.51274 3 3.77409x1(f 2.984036 -4.25288 5 l0~ 0.401673 -1.58262

lnP,

8.3634 0.0566/S (0.24244+2.2898/5+ 0.1 1857/5) xIO

1~

54(1< l~, < II IO~R)

+(l.4685 3.648/S + 0.47227/S)x lOTb (0.42019+ l.6977/S)x 1(1

927 4 6 9648 l + .O /Tb, + 1.28862 lnTt,, 0.169347T~~,)/
/Tb, 13.4721

Twu Correlations W=(lflPbr S. TIne mellmod inilially con-elates the properties of fractiomms normal paraffins as the reference. calculated values are tlmems ad~usledfor petroleum using the difference between TIne tIne
boilinsg point as lIne correlating parameter.

(15.2518

6875 IS.

InTb, + 0.4357T~) for Tbr

0.8

specific gravity of lIme hydrocarbon fraction and that of the tsormal paraffin with the same

w = 7.904

+ 0.1352 K,,

0.007465K~, + 8~359Tbr

Nornnal Paraf/n,ns: TIne properties of normal parallnmns are correlaled with the normal boiling poinl temperature,

+(l.408o.01o63K~)/T~,

for Tb. >0.8

f
=

where Phr=PbPc, Tb .=T~/T~: ~b is the pressure at which Tb is nmeasured, e.g.. the nomnal 1 boiling point at 14.69 psia, and K~ is tIme Watson clnaracterismmlion faclor. 11(1(6.2).
Cavett Correlations
T,
=

10633272 + (1191017 ~ I 0~1Tt,+0.779681 x

tt T + 0.969468 x 102 /(I~ /100)111

7 I0~ T 5

0.28437611 10
2 6 x I0 (Tb 459.67) +0.3889 x 10 (T

768.071 + 1.7134 (Tb 459.67)0.10834

459.67)

48 P~,= (3.83354 + 1. I9629ip~+ 3 . 888W + 36.19525V + 104.193W~)

0.89213 x I0(T~ 459.67)API + 0.5309511 I0(T~459.67) API + 0.32712 x Ill ~(1~ 459.67) API 3 5 logP~ = 2.829 + 0.941211 I0 (Tb 459.67) 0.3047511 I0 (Tb 459.67)2 11 7 +0.15184 x I0 (T~, 459.67)~ 0.2087611 l0~(Tb 459.67)API + 0.1104811 I0 (T 459.67) API 5 9 0.4827 x 10(Tb 459.67)API + 0.1395 xl0~ (T 459.67)API 5
=

[i (o.~ 19869 0.505839w

I .66436w 9481

S~

= 0.843693 (1.1

8624

3.361691V

13749

wlnere tine subscript prefers to properties of norunal paraffimns and,

where API=(141,5/S)- 131.5.

IT

IT,,

366

.4 ,j,endi v I-I 1

(ru(ical !roperlues (.orrek,tions

367

The molecular weight of paraffins is given by the following imisplicil relation,

T 5
=

11) = ASp[12.53262_46.I955/7~_(t.(%o(271)1)5Fh)+
(I I.4277+252.I40l1~+0.(X)23O535ThJ~Sp]

exp[(5.714l9+ 2.71579 lnM~0.286590(ln Me) _39.8544/(lnM~)0.1 22488/(ln Mp)I

24.7522

mM

+35.3l55(lnM) AS~ ex~10. 6( S~ S) Molccuular WcighI Iu~ M = (Imn M~, )I (I f 2 f,,~1(1

which can he solved iteratively rIsing the following itnilial guess. Mv

= F,, /(11). 44

0.0062 -a,,)

Pptroleu,,m Fractious.

--

2l~) ~

The properties of any petroleum fraction are estiuim~ntedby adjusting line calculated properlies of the normal paraffin with tine same boiling point as, Critical Temperature:
F,

~uu

A5si~I + (o.om

75691 + 0.193168 /1~ )AS

T~,1(1 + 2fr)/(l 21) )I

I = 0.0123420

0.328086/T,~ I

.1. 1 T =A.St 41.362456/ T~4 0.03982850.948125/ T~ AST

AS,,,= expI5(S~ S)j

AS.r=CXPF6(Sv

S)J

Critical Volume:
Vc = v~pt(l+ 2f,)/(l

2f,)j

/ T +( 0.182421 + 3.01721/ T~)AS .1

AS,

exp[4(S~ s2)]

Critical Pressure:
Pcp(Te/Tcp)(vcpIvc)~I+2fp)/(I_2rp)~ 2

368

(-

flquazun of State E.mpre.t.cio~us Immmpiemnsemnling line raundom mmsiximsg mrules for the tnnixlure EOS parameters (Sectiotm 4.3.1),

369

EQUATION

OF

STATE

EXPRESSIONS

a The general cubic equation of state, Eq.(4. 12), ~RT vb a v+uvw

b= ~

u = ~xu

and w =

we ohlaimm, B(b./h) A
+

takes the following dimensionless form:

2 Z~_(1.4~B_U)Z+(A_BUUW )Z(ABBW-W)=O where
A~~, (RT) B~-~-, RT

lnu~, =iui(ZB)+

,j___I2~x~a L U +4(w~ +4W /w)W1 1/a (u /u)U 11

t ~1

mi

2Z

+ 1) ,f~3~+4W21 A[~(2z+U)(w /w)W +(UZ 2W)(u /u)U]

[2Z+ [I + ~JtJ2.F4W2j

(Z

+ lIZ W)(U +4W )

-j

~
RT

W~!,
RT

and
RT

Solution of Cuhk Equation 1 Z -t-a Z +a,Z+a =0 1 1

The above equation results in the following expression for the fugacity coeJ/Ic:euut 0.10 pure compound: ln4=(Z I) In(ZB) A VU24W2 In 2Z+U+,JUc4W2 2ZU-TV+4W

Let,

Q = (3a,
II D>0,

a~ >/9

(9aua, 27211

2a )/ 54

1 D=Q -fJ

lIme eqtmation inas ounly one real moot: +(J

Z~ = (J + \/D) The expression forfugacity of a component in a mixture is:

,fD)t/~ a /3 1

If D<0, tine equmatiomn Imas three real rools:

ln4 =In(ZB)+-~~--+

ZB

1 A A, ,Ju2+4w2L

1 uu +4WW2 jln}I2Z+U_,fU2+4W21 U-)-4W L2~~+tj+~4~i

.~

Z =2,f 1

1~

cos(O/3)a I3 1

Z, =2,JQcos(0/3-l-120)at/3

(Z+UZW AE2(2z+U)W;W2

l~(UZ_2W2)UU1 2 2 )(U +4W)

2,/Q cos(0/3 + 240)a /3 1

where, the dimensionless derivatives ofEOS parameters are defined as,

where,

0=cost(J/~J_Ql)
A.

E..anL ..1

01 ~1

~n

. B~

~.

Jn,~,.T

b[,

i[~rmu>~
ut ~)n 1

~ufl(iW

in , ,T 11

Jn,, .T 1

Ut

Id(nw)1 ~)un

in,.~,T

If D=0. One equmaliomu Isas tinree meal roots, at least two of tlnern are equal:
=

2J

with the total number of moles, n, defined as,

fl

a /3 1

370

.4 1~

f~endInC

l~.qt4ahiotu of State Expressuon.i

371

THERMODYNAMIC PROPERTIES USING PENG-ROBINSON FOS (PR)

h, (4.27)

2~2h a

)lnI h,\ (v(~ _i)hJ + v (~ l)b v +(~+ i)b ]+ Pv 1 1 1 1 b ~v+(f2+l)h v(/~l)h v+(12+I)b 1

RT/(v

- b)

a~a/[v(v+ h) + b(v -

b)]

wlnere v, is the partial molar volutsne.

~=[l

+m(l

5 2 Tr )1

(4.23)
(4.31)

When using the volume shift, tIme corrected partial molar enthalpy is given by,
h~=h c P 1 1
Partial Molar Volume

vcor=v~c

Molar Enthalpy
v

~)v ~

The total enthalpy is calculated from the following tlnernnodynauinic relation,

~i)r.r

n,.

H =

P _T(~)

}mv +

PV + ~

Converlimig the molar volume in PR

differentiating it, we obtain,

10 total volume,

by multiplying

it with in, and

where the last term is the total internal energy at low presstmre and prevailing Ieulspcralure. and it is determined by summing the internal energy of individual pure comnnponents.
N

(RT+ h,P)(v + 2hv

1 ________________________________________________
h)

h)

+ [2h RT 2~x~a 2h P(v

I

11 1 ___________________________________________________________________________

h)}v

b) + b~a

Applying the Peng-Rohinson equation of state to calculate ~,j~j) using the randoumm mnnixuung wles, Eqs.(4.74) and (4.78). and dividing the obtaiuned expression by the total number of
moles, us, result in.

P(v2 ~2hv

+ 2P(v b)(v+b)

2RT(v + b) + a

a (v(,J~--l)b N InI ~+ Pv+~x,U,, 2J2b ~v+(f2+l)h) I

where,
NN

xx,(l

k, )a~mm~[mT,~ + x~ 1

12

43 When usitng the voiutnne shift concept to correct tIme predicted minolar volumne by PR, Eq.( . I), the corrected molar enthalpy is given by, h =ImcP Partial Molar Enthalpy h,

~ an,

Multiplying tine molar enthalpy. h, derived above, by n and differenlualimng tine obtained expression of the Iotal entlmalpy. we obtaimn,

-3 53

_,_ l40
_~ -.0 2~ nOn

-2

~ ~ ~ 010 20 000 3~ 20 .00

,o.nfl,* ,~
C

30000

100=0
30=0

30000 .00=

.1

\-u,

00=0
-~ ~-. -~ . .-.

P
?~-.~

I ~

I[17.-~ 4
-~i~..L

-~
J~~!
~

I
~-:Th-~

a
S S

Ion

//:
~/H
30,1

___________
.

:1
~

/~ ,47

--

-~

~~Jo

305

:~~ ~
0% 00 0 00 210

300

~O

300

300

00

0%

000

Figure Dl. Convergence pressure of hydrocarbon mixtures for use with GPA equilibrium ratio charts.

a,

374

PRESSURE, PSIA
0
t ~I 1.

Fq,oiI,I~,lr,,;:fuji,,

Eq,ul,/n,

it,,,,

U,,:,,, 30
, 00 ,
I 0

,~0

100

PRESSURE, PSIA Y~ . Sam

~I:.
.1

375
I i.~

3.~

m3~~io

. n

..

~( . , IL . 4 :,
~ . ~.

I
, ,

00
.

~ ,C
~

.,t

~5.~
,-,

.
~

:,
. .

;~ is

00

~
. ,

~
~v

:
, 3.
, I S .

_i
:.

.,
~

.
. ,

4 .f.s . p
., ,

0
,, .~i,It . ~ ;,: ~ , , ~

.~
,
~

;-~ . ,,~, ~ ~ . ,.( .,, !,,,

.. ~

. .
.

V~

~.:::s(~ ~ ~,.,..v.
I ~

.. ~ _ ~ .

~
.

II
0

Jv ~ . : . ._...:c ~ ~~i1:~~.:. -, ~ ~ap,,t_ : ~. ; ~ Ii c~ ,. -:: - I . lili t~ 4

_~

.
.

., . .. . ..
--

i~:~ l.~ ~ . , ~_i .01 oSno

.

~
0

.
-

ii

I~

~o-~

~ ~

.
,

I.

,. , .
., . ~ .

,.,

to
I

~ ~ ~ K =Y/ 5
~ 0 ~ ~ -~----:-~-:, ~-= _~ _

~
~

.-

~
~
. _ .-

~
.~ .-

v~~i~ ~o
-0~
~-. .. .. .

:~
:._
~ _

C~

K =Y/x

-.

-.~

. ; ~ ., . . ~

.
0

~
,;~ , ., . . ..~ ~. I

.:.. ~it~
i~ . .-. ,. .

o
~0

~
~. ~

,.
. ~

It 3

~_.~Tt ,.:;..
- -~. ~

, ~~
~

::I~5.

,,,

~.

, 0 0

--

.-~

_:~:a. ,~~-EH~~

.. ~

~1t..):.

~ ~ ~. :0:.

4-.=.o-

~
00

~
~

~
~

~~~-Hl

, ;th~

~1 j: , ~~I ~ F ~~
,.,..,.,,,

~
~
, .

.0I

:.~

~~,-_-

,~

(It

0
0%~

.~~rtIso:iij~.

(1

):i:I5~. :iia~~ .

-~i~~ .. _ _ 6

4:i~-_

~.

i~: ::;,.-.~: , ~a,;:

~:vv-(

0.

~:

loom

Pressiure, kPa Fig,orn, 0+

Equilibrium mau,,, am 34.47 MPa (5(XX) ps,a) co,,ver~cnce pressure. I- ,g,orc I) 4

t~q,.,I ,hr,um

ral,,, ,,I

1~~ .c~ ~rn~ , -3 0 0~ PROPANE I ressure, kI a 115 14 47 ~tll,,I ))) r~,o) c,,nvcrseoce prescore

~-L~ ~1

376 PRESSURE, PS1A

300 ~ ,..1.0

!~,,,0,I,!,r:~,3pt koathi

Ij,0i/i/~~ jti.i,~ Riot,,,

377

Pressure, kPa Figure 0.5. Equilibrium raujo at 34.47 MPa (501)0 psia) convergence pressure.

Iressrmre, kP:m
i,g,ore i) (, Oqiii)~Iriut,~ ,.,:i,, at 34.47 NI Pa I515)1) psit~),~0,uvergeucepress,,Ie.

BUTANE

378
00 I0~ ~ ~ 3 3

PRESSURE, PSIA
moo aoo .

.,,00~/o)3o~o~~J R,oto, . I.~

Iq,o,l,I,iu,,, Ru(,i;

379

PRESSURE, PSIA
500

m,~

,tOO!!1100

05

10
. . .

~..

... ~

. ~

,si).
?(~ ...,

..,.

...4.

,. 3

S 0
,

a.
~4AI ~2

I,

,
:
~

., 3.

Ii~4~ = ,~I,,h.
j~~ &4~ : ~ ~ 0 fli.5:.r
~

.j~
~

II

:1
L
~

:~ .--~~-

~:
. -:.

,,

~~~-:

.. ~

mu

~
~

,.?,

_ ~ ... ~

:
. ..... . .,.. . . . -. ,.

..

10

:
I
I .

~?
_5 .

.:..::

.~ ~ ....... ::

c...~ 5,, a.
~

. ., . .

~..:
~ _ ...

. .

. --.:

8 K- /

. 5 ~

~_

.,.,

~e
~

~ ~

30) r, _ ,,~,, ,

_~4 0 ~ , ..,st

i.~o

nit~ ~

_ i50~)

~ I

~,

1 mm Pressure, kPa
I~ I ) I) l~,I,0, I,l,r,,,,t, ri) ,i , at 74.4 7 r~i ta ( 5(55) psia) co~nvergcncep3cssoome.
-

Jrcssume. kP:u
I tgure I) 7

PENrANE

II-PENTANE

.3

I ,goiI.)ri~or.,,;ii,~ ,) 1.1.11 Ptll;.

)S)$1() u~.) c~ns~-t),r, e

380 PRESSURE, PSIA

Iti)~Ill~),3i~0,,, Rmi,o

F~,;,,o/,I,,,,,,,,Riot,,,
- -

PRESSURE, PSIA
u

181
S I

Pressure, kPa
Figure 13.9. Equilibrium ratio at 34.47 MPa (5000 psia) convergence pressure.

Pressure, kPa
Figure I). 0. i,qoi,I,t,rtutit rat,,, at 34.47 Mta (5)))))) psia) convergence pressure

IIEPTANE

382

E,j,oi(ihriu,,o 8,0(0

Rut,,,

PRESSURE, PSIA
so m
I

PRESSURE,PSIA

383

3.aso I

~ Pt,o.~ 3,0,, )947 1flb0i0~i000, uf 0 P.o~,. Un.,.o,,5~ of 302,5. -~1~,, E.,o#o0o,o~ n,,d d~u-n~j

--

. I 00.4 I ro,u 947 1,bootn~iuo, of G. P,ow,., Uuio.,nto, of Uiuf,on. Euiro nntnt.d nod 3,u~ 3 54. 0 *907 - 0, Corp.

Pressure, kPa
Figure 1) I)
-

OCTANE l,g,,t,- I) 2 l,1o1),t,r,,ot,,

at~,,a) 54

Pretsrure, kPa
I/ Mt,, (5)HH) pun) c~inverget,ceprcssoorc

NONANE

l-qu,I,hr,un, rat,,, a) 5.) 47 MPa (5)5)0 psoa) convengenuc pRssi~re

384
I ~
0 I I S

PR ESSU RE, PSIA

~
~
I

1q~o,Iohr,,o,,,Rootot,
,

100
- . -

300

5f~

1,000
.

3,~

Pon.d ,oo, 947 ob,,Io,non. Os 0 5o~n. U ~ y of E,~.opuOoOod ond 4,0,,,, bj~ Tb. tO,, . Co,p 130 ,n ~957

gon

of

385
INDEX

:
I I

-I,

~(
L.

acentrtc factor,

13, 221, 222, 352-357

actuvuty c,,eflocient Ill

.
.. .. ,

l.~I
. ,, ..,

(.~i

I~I
i~ ~ ,, ~
~. ,. . ,

~ 1 ~ ~ h ~~
t
.

~I~~~

I
m.o
S i S. I ., i ~

I
,,

~ ~
~b.
0~

~ I~ l~~ ~ ~

to
,

,,~ :
,,

.,.

~
~4 ~ 4~ (t~/~,

~.. .5. ~, .

~
,, 2

,a -4n

...,.,

C
_~-1.

~i .)j)~((~
I
5

~ (~
.

: )) L2.~. U ~ 0~ .J
iiI

12 i~~u ~

U!F

~
jL~fl,.
,.,~.

-.
....

~0,,~F5a .
,

Ii
U
.

ml
.~) ~ S

Atano and Kennedy equatittn, 75, see also density alkanes, 2 API gravily, 23 apparcni oil densimy, 73 appareni liqiuid rlensily onFnatural gtis. 73 aru,itoau cs, 2, 213, 214 aspltaliemnes, 2 ads,,rhcnj untuleritml ,,ut r,uit. 17 ;tuuracfivc lenin, 132. 149, I 54, sue abai ef;uattn,n ,,l .sl;mfe

~ Ii

T .

~,

...

~
~ .,,~c,
~ .

t 5 ,,i,2 rE.,t, ( 410...

d1~~ E ~~3h
:~...II:1~-.
. ~ ~, ~ -

(ii ~,~i1:5 . _~:I:.t... ~

, ~

~s
,0, 9.ac
C

51

~
)

1~~
,,

ItP~

It,
,

0_I _

..__ -, . ,~._._. ,.-=i,:.

,.

~
,

. .,-

. . .

_c-i=

~,,

,, . (11.
~~O.

.
~

. ,

.,,

1~. .

. ,, . ~ 0

..,,,,,

o~I

:~
_~,,

~=-

~o

~ e~
C
,. , .

=~,~Ic

K~/x
I

~T=-
::.

.... _~ - -~

-~~t

~ r

~ .0,t,
~

01

__

r,,
:

-= .. ~7..R ~ . .. 1 . ~ : u ~~ ,,,O

.4::_.
.

.- - .

~,

_ _

I~
~

I~

jt ~

~: .~:

, .

~
3

:~
,
0

~-:,~:Hf .
:
~ I,, ..

~ mu

~,-:

,, .

_
.

~:

~=

__

.e0~.-:~...

~
_5_

~
-)

=
2

ce~
19

ist~s-j~,~ ~0

-.

-.Y

~io..7~E((~i/.:(.t,
2t

ut

I)
~

t
,

c
.

t
...

4-)(~.).

~ ~ Ii
,

: ~
,

(22)

r
l~

1 1

(~o1

.001

~~

sot

8,,,

-ii

.. ~
.

~.atv,.

~-, ~iy~
,

.( . h, (~, j~,4((5s tj~g

16...

~
60
~

),,b
LI, 5

5:
~

i)~

H~

~.
~:

.
~~

40.

(~)
soot

.osoi

,.

..

.,. 3.000

.
i0,~

,,,
DECANE

100

30)

500

0.~

Pressure, kPa
Figure 13.13 Equilibrium ratioat 34.47 MPa (5000 psia) c,,nvergeuce prcvsure

l~ickwarol uunuliiple clunusucl, 268, ~ see ntso, gas injection iests hum ling point, noruunal, 4 mruc, 211) Benedici.Wehh-Ruhjn EOS. 131 Slarlummg modification, 131 binary interacuion pimraunelcr. 152. (55, 181), 327. 331, see tulsjt mmsmxing rules black oil, 28-29 corrctamiorms, 67-79 le.sis, 42-52, see ills)) oil lesls bubble p00mm pressure, 64, 86 conlculamion, 169, 173 corretamiorms, 68 (7 j- clisoracmerj,,~im ion, See aI~t, cf,tfi Iinu,,uis descripiion crotucal propermy correlaliuns, 119 criiteal mnnolar volume, 336 eapullary condensation, 38 carbenes, 2 carboids, 2 carbon group criiical properties, 221-227, .156-157 Caveim corretamion, 222 Frlmusmer corrclalj),un 222 I_cc- Kesler (tim cat eorrel,otj, ,tts, 22 I peruurhinliotmu expansoon c,,rrelalio,ns 223 Riajj-[)auhcru coru-elalions 222 properuics by runiximug, 31)8-31)9 single carbon muuuimhcr grotomp, 211 nsotlecular weiglut. 213 msno,lecuolar wciglsl boundaries, 239 norunsal ho,iling poi sm, 21 3 specific gravimy, 213 Caveti eorrelatuon, 222, see also earhouu group clmaracterisatiomo lactonr, see Walson characuerisaui,,n lactor clseun coil poiemiui:il, ((1711)8. 183, 197 gradieni, 200 cla.ssi licai,,,n oI rescrv,,ir Ihoitls, 22.29

composilional grading, 2, 38. 195-203 aro,unatic effecm. 2(12 heal of Iransport, 199,200-201 non equitibrittin , l98.201 ()unsager relati,ons. 199 sigutilicance, 2111 Ilterustiol gradient, 2(11 Commup,osotional analysis, 38 hto,w-down 39 (Jal,) evaluaui,,n 126 (toll s)rcauts (olirscu 1 sauuspli nip. 39 coinup,usflto,n reirievtnt. see inverse grmuping comiipre.ssthlluly facio,r, see gas commpressihilimy factor coniinu,,us description 234.246, see also carbon grottip exp,,nentual dislrihumioin, 237 g000uoma prohahiliuy disirihumion 236, 241, 241 critical properuies hmnsrfes, 18, see also critical poini pure connpounds, 353-354 sungte carbon groups, 355-356,
condensing gas drive, 257-258, see also gas injecuion, minimum miscihiliiy pressure conndensiitg/vaporisung gas drive, 259, sec also gas in.ltCllon. mmnimusum miscibility presaure convergence pressure, Ill, see also equilibrium

rami~, corresponding slates, 10 Cot charl, 4, see also vapour pressure crico,ndentfnei-nm IS, see also gas condensate critical cuunpressmhiliuy factor, II. see also critical

properties
criiical pu~inm, 5. 134, 143 calculation. 192-195 Kreglewski and Kay unmeihod, 194 Lis mixing rule, 193 v,,lume 194 critical tie tone, 256-26(1, 272.273, see also riuuuluunnummo miscihilily pressure cross over lie hines, 26(1, see also nninimmnlun

mniscmhulimy pressure Darcys equalion, 332 degrees of freedom. See Gibbs phase nile
density

prcducm ion
Alanu and Kennedy e(luaii,,n, 75

FOS, 319

Siandiung and Ksoi,

valuramed pure compoion,Is. 8 uuieuloosj, 73.74

386
dew point

1,t,Iex

Index

387

, 10, 19, 37, 56, 64, 341, 343, see also gas condensate calculation, 169

forward multiple coniacl, 267-268, 273, 329, see

Gibbs phase rule, 4

Grouping, 3(12.314 Dtune,sli Ct al., 3(14 eo ioal iioo,le Iuacli,,io 306 1 equal weight (iunass), 307 Gaussian quadrature, 243 inverse grouping, 31(1-311 Newley and Merrill, 303, 306, 307, 309, 310 optimum number, 305 selection. 302-307 Wlnius,on 302 I ce - Kesler crulical pru~periy ci,rret,ot i,,uos 221, sec ,olsio carho,tn gioutop I leliiil,,,lti energy, 106. I 91, 193 I lenry.s law, 114, 116. see als,, gas soluhility I bit luiitiosn plol, 124, 326, sec also equi libriutti ral iii
loytlrates, 1(6 unlerfacial tension

dislillalion, 2 1(1-215 dry gas, 24 Edrntoer correlation, 222, see also carbon group enuhalpy

consuanu flash, 175

definition, 106 equaloon of state comparison, 3 14-323 general van ,Ier Waals iype. 135 la)el Feja. 147. 3)15. 314, 325, 347, 148 leng- Robinson PUS. 141, 156, 172, 178.

314, 339 predocooin reloahility phase composition, 316 volumes, 320 saturation pressure. 3)8 Redlicls-Kwong, 138, 314, 346 rohtostness, 182, 322 Schmidt and Wcts,el PUS. 146, 155.114 selection . 325 sensitivity, 327 Soave.Redliclt-Kwong, 140, 152, 156. 314, 346, Grahoski and Daubert rosodification, 141 Starling.Benedict-Wchb-Rubin, 131

also gas injection tesis fugacity, 105-110 I .ewis rule, 1(19 fugacity coefficicunt. 109, 129. 157, 174 pure substance, 136 gamnsa probability distribution. 236, 241. 243, see also continuous description gas density, see gas compressibility factor formation volume factor, 40 deal gas vo,Iuoiooe. I toi,,lecular weight. 79 specific gravity. 68 viscosity, see viscosity

local composition 159 nonrandom 159-161 polar.p,ilar inieracijon coefltcienus, 160 raisdom usiising rules, 159 iuoiioiuitntui miscibility enrichment, 257 mununsum muscibility pressure, 256-257
prediction methods

gas chromatography.

215.221 capillary c,,lunons, 217 comparison wo lIt distil lati,,to, 21 5 non-eluted fraction. 217 packed c,olutttns, 215

Henham ci al., 273 carbon dioxide, 275 Firroozahadi and A~iz, 271, 277 first coniact rriiscihiliiy, 270 (ilasii, 274, 276 Iliodgins el al,, 272 Jensen and Michelsen 273 Ku,,, 274, 276 Peclrood 274, 276
utiunlecular weight cottdensame 79 measuremeunl, 212

tuneasurennenl 282-255
Inlerface curvaloore, 283-284 pendant (Imp, 282-283

tuning, see tuning

Valderrama-Patel-Teja PUS. 148, 160 van ,krWaals PUS. 132, sinaI, 130-131 voolume shtfi, 141 Jhaveri and Youngren. 143 near critical poinl. 143. 149 Peneloux ci at, 142 Zoudkevilch -Joffe PUS. 13$, 314 equilibrium ratio, 111-125,171, 310 estimation 116-125 PUS, 317 OPA K-value, 118 intermediate pressures, 121 Mollersop equaloon. 123 Wilson eqtoation, 122 internal consistency, 59, 124 split meth,d, 112 equtlihrn,in flash calculations 68- 0(3 coutopulatiuinal I floe, 17) 152 negatise flash. 175. I 89 robustness. 182 root selection. 175 irivial solution, 171 first Contact miscibility. 255 flash calculations, see equilibrium flash calculations formation water, 86-94

gas compressthility factor,45 clnart, Ill) Dratuclnuk and Ahou- Kassetto c,,rrclation, 80 H2S and C02 effect, 82 gas condensate, 25.27 gas condensate tests. 52-65 constant cotssposo)uon expansioon. 53 cu,nslant vottoinne depletion. 53 colour change. 54 dew point. 56, see also dew point gas cycling, 2o,.~, sec ols, gas it, ject 000 uonaterial hala,,ce, 59.f, I pressoure btoilil.uip lest, 63 gas foonation volume facttir, 46, gas hydrates, 86 gas injeclion leSis backward tutultiple contact, 268, 276. gas cycling, -263 forward multiple c,ontact, 267-268, 273. 329, rising bubble apparatus, 265 swelling, 266 single Contact, see swellung slim tube. 260-265. 332. 338 gas in solution, 50-SI gas recycling, see gas injectioto tests gas soluhility in water
Krucl,cvsky.Kasarnovsky colioatton, It 6 llenrys law, 114, 116.

rclatoi,n willn density olilference. 289 preoliction metlt(xJ br lnydrocarl,,,ns Lee and Cinieti, 288 Macleod-Sugden, 285 scaling law, 288 .cclteclnier and Gui,, 291 Weonaug and Kati, 286 water.lsydrocmj-,on iuietlnane.waler 292 predic) i,,n incubi ,uI. 292, 293 salts in Water, 294 waler nonssal octa,oc, 293
inoruuosuc stability, 190. see als,, stability invariance condition l59, 16I, see als,, mixing

mixtures, 245 rich gas. 42 utnuttiple contact tests, 304, injecluon tests naplnilsenc, 2, 2 13.214 (oil compressibility, 43 oil densily. 71. 73 oil c,,rrclati,,n bubble poini, 69

314, 321,

see also gas

formation volume factor, 70

usollicrunnal conuunpressihility coeffmcjenu, 70 t,itat forisiatjinio voilujme factor, 7172 sisciisiiy, see Viscosity oil tests cotnohunalion ofdam, 49-52 (lofferentual liberalion 45 differential vaponisamjon 45 separator, 46.49 oil specubmc gravity, 23 olefmns. See alkenes paraffuns, 2, 212, 214, 223 Pamel-Teja EOS, 147, 305, 314, 325, 347, 348, see also equamion of stale Peng.Rohjnson POS, 141, 151, 156, 172, 1711, 314, 339, see also equation of slate perturhaluon expansion critical properuy courrelati,,ns 223, see also carbon group phase uliaprauui etluauoe.bocpmane, IS gas condensate, 26 inulticotsiponent mixture, 19 pure substance, 9 10 PNA, 220 pseudo critical condilions, 176

gas to oil ratiu,, 23. 340 gas viscosity, see viscosity generalised single carbon group, see carbon group Gibbs energy clnange. 176 definilion. 1(16 miuiinisation, 83 stability, 1(4, 185

rules Kay.s tuotsing rule, 17 I,ee.Kesler vapour pressure eorm-etati,in 13, see also sapoour pressture Lce-JKesler critical property correlati,,ns, 221, see oils,, carbon group l.ewus nile, 109. see als,, fugacity limiting tie liune, 256.258. 260, 267-268, 273, see ats,, tiiscihility Ititui,l-liqooid displacement, 338, see also, tuning I .i,tunen, Pray-Clark ri,rrelatf,,ui 1 4S see its,, vtscirsiy iiielioing point , 5 iutiscihility couucepts, 254.26(1 espcrttonental studies, 260-269, see also gas injection real reservoir fluids, 258-260 mixung rules, 142-161 tnvarmance condimi,,n 159, 161

pseudo crimjcal values, 17

pseudo reduced properties, 17. 192-194

388
quadrature points, 243-244, see also coniinuous description Rackett compressibility factor, 14, 142 Rackets equation, 14 Raoults law, 1(12 Redlich-Kwong EOS, 138, 314, 346, see also equation of state reduced properties, II

Index

experimental data. 325

fluid characierisation, 324 grid-time step suing, 339 limits of paramoeters regression variables, 327 weighting factors, 324 Valdenrama-Patel-Teja POS. 14$. 16(1, see als,n equation of slate van der Wa~bs EOS. 132, see alson equatioit of stile vaporisong gas drive. 255-256, see also gas injection, uniniussum miscibility pressure vapour pressure Cox chart, 4 Lee-Kesler correlation, 13 water, 87 Vasquez-Beggs correlations, 67, 69. see also black oil correlations virial EOS. see also equation of state Benetlict-Wehh~RuhinFOS. 131 Starling modification. 131 virial coefficients, l30, 154

relativepermeability. 333
miscible, 334 relation with IFT, 334 relativevolume, 43-46, see also black oil

correlations repulsive term, l54, see also equation of state residual oil, 45, see also oil tests residual viscosity. 335, see also viscosity resins, 2 retrograde condensation, 19, 321, see also gas condensate retrograde vaporisation. 19 Riazi-Daubert correlations, 222, see also carbon
group

viscosity
prediclion

sampling condensate ring, 342 evaluation of reservoir fluid samples, 340-345

gas condensate,
340

corresponding states, 334 Ely and Manley, 334

gas,83 I-lerning-7.ipperer. 335 Loltren,.- tlray-Clark, 335 low pressure viscosity. 335 - oil, 77-78 water, 93 tuning, 336 volatile oil, 65, 27, see also oil volume factor, see relative volume volume shift, 141, 314, 330, see also equalion of

iso-kinetic sampling, 36

minimum gas well flow. 35 recombination, 36. 341 sample collection, 36-38 surface sampling. 340 well preparation, 34 Schmidt and Wenzel EOS. 146, 155, 314, see also equation of state Soave-Redlich-Kwong, 140, 156, 314, 346. see also equation of state solution gas to oil ratio, 46, 47 vented at stock tank, 68 stability analysis , 183-192 limil, 189-190 Michelsen method, 187 Standing oil correlations, 67, 68, 69, see also black
siock tank

state
water

conspressihiliiy, 92 content of lnydro,carh,sn phase, 87, 89

conteuni of tiquid hydrocarlsuons. 88 density. 91 formation vo,bouooe factuir, 92 hydroicarhon solluhility in water, 90-91, see also gnssoluhility in water vapour pressure, 87 viscosity, 93 Watson characterisation factor, 212-214, 330 wet gas, 25 Wils,sn equation, 122. see alson equilibrium ratio Y functiois, 43, see also oil testing Yainada-Gunn correlati,sn, 14, 354-356, see also critical compressibility factinr Zudkevitch -Joffe EOS. 13$, 314, see also equation of state
-

oil correlations oil, 47

sublimation curve, 5 swelling test, 266, see also gas injection tests temperature dependency parameter, 136 Mathias and Copennan correlation, 150 super critical hydrocarbon components, 153 Twu et al. correlation, 151 ternary diagram, 254 tie line, 254 triple point, 5 tuning, 323-331 consistency of regressed parameters, 330 dynamic, 331-340