S

SPE 1488
864
M
Modeling
g Hydraulic Fractures in Fin
nite Difference Sim
mulators Using
A
Amalgamm LGR (Lo
ocal Grid Refinemment)
S
S. S. Abdelmo
oneim, ENAP
P Sipetrol; R. Rabee,
R Cairo University; A
A. M. Shehata
a, Cairo Unive
ersity/TPS; A.. Aly, TPS

C
Copyright 2012, Society
y of Petroleum Enginee
ers

T
This paper was prepare
ed for presentation at the North Africa Techniical Conference and Exhibition held in Cairo,, Egypt, 20–22 Februa
ary 2012.

T
This paper was selected for presentation by an a SPE program comm mittee following review of information containned in an abstract submmitted by the author(s)). Contents of the papeer have not been
re
eviewed by the Society y of Petroleum Engine eers and are subject to o correction by the autthor(s). The material ddoes not necessarily reeflect any position of the Society of Petroleu
um Engineers, its
officers, or members. Electronic
E reproductionn, distribution, or stora
age of any part of thiss paper without the wrritten consent of the S Society of Petroleum E Engineers is prohibite
ed. Permission to
eproduce in print is res
re stricted to an abstract of
o not more than 300 words;
w illustrations mayy not be copied. The abbstract must contain co
onspicuous acknowled dgment of SPE copyrig ght.

A
Abstract
HHydraulic fracturing
f allows
a num
merous, otheerwise unprroductive, low permeability hyddrocarbon
fformations to o be produceed. The interractions betw ween the fraactures and tthe heterogenneous reservvoir rock,
hhowever, aree quite com mplex, which h makes it quite
q difficuult to modell productionn from hydraaulically-
ffractured sysstems. Vario ous techniqu ues have been applied iin the simullation of hyydraulically fractured
wwells using finite
f differeence simulators most of these techniiques are lim mited by the grid dimenssions and
ccomputing tiime and harrdware restrrictions. Most of the cuurrent analyttical techniqques assume a single
rrectangular shaped
s fractu
ure in a sing
gle phase hom mogeneous reservoir, thhe fracture iss limited to tthe block
ssize and th he fracture properties are adjusteed using peermeability multiplier. The curreent work
ddemonstratess how to mo odel these systems
s withh a smaller grid block size which allows you to apply
ssensitivity to
o the fracturre length and d model the fracture wiith enhancedd accuracy. IIt also allow ws you to
sstudy the eff
ffect of reserrvoir heterog geneity on the
t fracturedd well perfoormance. It iis proposed to apply
aamalgam LG GR techniqu ue to decreasse the grid size
s to the ddimensions oof the hydraaulic fracturee without
ddramatically increasing thet number of grid block ks which woould cause a great increase in the coomputing
ttime and thee model sizee with no ad dded value. This
T paper eexplains howw the amalgaam LGR is designed
aand comparees between standard
s LGGRs and the proposed deesign and a case study is presentedd from an
aanonymous field
f in Egyypt to illustrrate how to use this techhnique to m model the hyydraulically fractured
wwell. The simmulation mo odel is match hed to availaable producttion data by changing frracture lengtths. Then
tthe model iss used to preedict future response fro om the wellss. The advaantage of this technique is that it
aallows hydraaulically fraactured reserrvoirs to be modeled w with less griid size whicch will lead to more
rrealistic moddels and more accurate predictions; however, thhe most useeful applicatiion of this ttechnique
mmay come in n the fracturre modeling stage. With this tool, vaarious fractuure geometriies and scennarios can
bbe tested in the
t simulato or, and the most
m econom mic scenarioss selected annd implemennted. This w will cause
bbetter fracturre placementt, and ultimaately greater production.
2 SPE 148864

Introduction
Hydraulic fracturing is the process of pumping a fluid into a wellbore at an injection rate that is too high
for the formation to accept in a radial flow pattern. As the resistance to flow in the formation increases,
the pressure in the wellbore increases to a value that exceeds the breakdown pressure of the formation
that is open to the wellbore. Once the formation “breaks-down”, a crack or fracture is formed, and the
injected fluid begins moving down the fracture.
DOE research has developed several alternative fracturing techniques designed to accomplish specific
tasks such as:
 Tailored pulse fracturing
 Foam fracturing
 CO2, sand fracturing
In general, hydraulic fracture treatments are used to increase the productivity index of a producing well
or the injectivity index of an injection well. The productivity index defines the volumes of oil or gas that
can be produced at a given pressure differential between the reservoir and the well bore. The injectivity
index refers to how much fluid can be injected into an injection well at a given differential pressure.
One of the major problems facing the reservoir engineers in modeling the hydraulic fractures using the
finite difference simulators is the wide gap between the grid size of the reservoir model and the fracture
dimensions.
The purposes of this paper is to model the flow of the reservoir fluid in hydraulically fractured reservoir
using finite difference simulators in a manner that would allow the simulator to mimic the actual fracture
geometry without dramatically increasing the number of grid cells and hence increasing the computing
requirements and time. This is achieved as shown in the paper by using amalgam LGR to decrease the
fractures dimensions to the size and dimensions required to achieve this goal and leave the number of
grid cells only slightly affected which makes a minor change in the required computing time and
capabilities. This design was tried on an anonymous field in the western dessert in Egypt, and the results
were compared with the actual production data which was recorded after the fracture to verify that the
model was capable of modeling the actual reservoir performance.
Hydraulic Fracture Mechanics
The theory of hydraulic fracturing depends on an understanding of crack behavior in a rock mass at
depth. Because rock is predominantly a brittle material, most efforts to understand the behavior of crack
equilibrium and growth in rocks have relied on elastic, brittle fracture theories.
However, certain aspects, such as poroelastic theory, are unique to porous, permeable underground
formations. The most important parameters are in-situ stress, Poisson’s ration, and Young’s modulus.
In-situ Stresses
Underground formations are confined and under stress. Fig. 1 illustrates the local stress state at depth for
an element of formation. The stresses can be divided into three principal stresses. In Fig. 1, σ1 is the
vertical stress, σ2 is the maximum horizontal stress, while σ3 is the minimum horizontal stress, where
σ1>σ2>σ3. These stresses are normally compressive and vary in magnitude throughout the reservoir,
particularly in the vertical direction (from layer to layer). The magnitude and direction of the principal
S
SPE 148864 3

sstresses are important

i beecause they control the pressure
p reqquired to creeate and proppagate a fraccture, the
sshape and veertical exten
nt of the fraccture, the diirection of th
the fracture, and the streesses trying to crush
aand/or embed d the propping agent durring producttion.

.
Fig.1:
F The loca
al stress state at depth for aan element of formation

A hydraulic fracture willl propagate perpendicula

p ar to the minnimum princcipal stress ((σ3). If the m
minimum
hhorizontal sttress is σ3 th
he fracture will
w be verticcal and, we can computte the minim mum horizonntal stress
pprofile with depth
d using the followin
ng equation.

= − + +
1−
PPoisson’s rattio can be esstimated from acoustic log l data or ffrom correlaations basedd upon lithollogy. The
ooverburden stress
s can bee computed using densitty log data. The reservooir pressure m must be meaasured or
eestimated. Biot’s constan nt must be leess than or eq m 0.5 to 1.0. The first
qual to 1.0 aand typicallyy ranges from
ttwo terms on n the right haand side of th
he equation represent thhe horizontall stress resultting from thee vertical
sstress and thee poroelasticc behavior off the formatiion.
PPoroelastic theory
t can be used to determine the t minimuum horizontaal stress in tectonicallyy relaxed
1, 2
aareas. Poroelastic theo ory combinees the equatio ons of linearr elastic streess-strain theeory for solidds with a
tterm that includes the efffects of fluid
d pressure in
n the pore spaace of the reeservoir rockks.
TThe fluid prressure acts equally in all a direction ns as a stresss on the forrmation matterial. The ““effective
sstress” on thhe rock grains is comp puted using linear elasttic stress-strrain theory. Combining the two
ssources of sttress results in the total stress
s on thee formation, which is thee stress that must be excceeded to
initiate fractturing. In addition to the t in-situ or o minimum m horizontall stress, othher rock meechanical
pproperties arre important when desig gning a hydrraulic fracturre. Poisson’ss ratio is deffined as “thee ratio of
llateral expan
nsion to long gitudinal conntraction for a rock underr a uniaxial sstress condittion”.3
TThe theory used
u to comp pute fracturee dimensionss is based uppon linear elaasticity. To aapply this thheory, the
mmodulus of thet formatio on is an impo meter. Youngg’s modulus is defined aas “the ratio of stress
ortant param
3
tto strain for uniaxialstres
u ss”.
TThe modulus of a material is a meeasure of thee stiffness oof the materrial. If the m modulus is llarge, the
mmaterial is sttiff. In hydrraulic fracturring, a stiff rock
r will ressult in more narrow fracctures. If the modulus
is low, the fractures
fr willl be wider. The
T modulus of a rock w will be a funnction of thee lithology, porosity,
ffluid type, annd other variiables.
4 S
SPE 148864

F
Field Casee Study
11. Field Ch
haracterizattion
WWe conducteed a fracture reservoir sim mulation stu
udy for well A-1 over Abbu-Roash G reservoir inn the field
((A), in the western
w desseert of Egypt.
TThis field prroduced ligh ht gravity oill (37° API) from Abu-R Roash G sannd at an average drilled depth of
55400 ft TVD DSS. Only tw wo wells weere drilled inn the area annd had been reviewed inn the field sttudy. The
interpretation n of the well logs sho ows hydrocarbon beariing in midddle A/R G formation w which is
ssubdivided into
i two saand bodies. The two sand bodies were perfoorated and ttested; theyy showed
pproduction with
w initial rate 370 BO OPD by N2 2 lifting wiith traces off water andd no gas prooduction.
PProduction started from well A-1 wiith ESP yield ding a produuction rate oof 350 BOPD D and 35% w water cut,
tthen the welll production n rapidly decclined to 1300 BOPD andd water rate started to deecline. The E ESP then
ffailed severaal times and has
h been rep placed with sucker
s rod ppump. The laast static fluiid level measurement
sshowed averrage static reeservoir presssure of 1247 7 psig at a reeference deppth of -53000 ft TVDSS. No PVT
ssamples were taken from m this field, so calculatio
ons were donne using botth correlatioons and PVT T samples
ttaken from th he nearby producing fieelds. An estiimation of thhe oil in plaace was donee using bothh material
bbalance and volumetric, showing reeserves rangiing from 100 to 15 mmS STB. Only ccompressionn velocity
wwas available from the sonic
s log and d the shear velocity
v wass calculated using the foollowing corrrelations
ffor both the sand and sh hale layers. A static mod del was buillt using the available daata with appproximate
ccell dimensio ons of 50 X 50 m and a dynamic model m was ssuccessivelyy prepared too be used too test the
ddifferent fraccture scenariios.

Fig.2: Overviiew of the stattic model

FFor the purpose of this case

c study sttatic and dyn
namic modeels were connstructed, thee dynamic m
model has
ggrid size of 50X50 m, consisting
c of
o 72,240 ceells (80 X 433 X 21). 399 LGRs werre used to m model the
hhydraulic fraacture scenarrios in three wells; well A-1 (currennt well) and the two prooposed wellss A-2 and
AA-3. The LGGRs were am malgamed in three groups each at thee location off each well aand were dessigned so
aas to match the
t actual geeometry pred dicted from the rock moodel as previiously mentiioned. The PPVT data
wwas taken from
f the nearby operatting fields anda reservoiir propertiess were popuulated basedd on the
uunderstandinng of the depositional environment to match thhe existing ttwo wells. T The model sshowed a
ggood match with
w the orig ginal production and pressure data.
SPE 148864 5

2. Fracture Design
In this section we discussed building the rock mechanics model to design the fracture and how to
optimize the fracture design.
Building the rock mechanic model
The rock mechanic properties have been calculated using the open hole logs (sonic, density and neutron
porosity logs). Although only compression velocity is available from the sonic log, the shear velocity
has been calculated using the following correlations for both the sand and shale layers.
Shear velocity prediction: Results from this limited depth range study show that shear velocities vary
linearly with compressional velocities for both sands and shale. A linear relationship between and
has also been observed by Castagnaet al. (1985) and Williams (1990). The equations for sand and
shale shear velocity predictions are given by
= 0.7149 − 2367.1, …………. Sand
= 0.6522 − 1902.2, …………. Shale
Where all velocities are given in ft/s. The rock mechanic properties (Poisson’s ratio & Young’s
modulus) have been calculated using the following correlations:
[(∆ /∆ ) /2 − 1]
=
[(∆ /∆ ) − 1]
( /∆ ) [3 − 4(∆ /∆ ) ]
=
[(∆ /∆ ) − 1]
The closure pressure for M AR/G and the stress profile for the layers have been calculated using the
following correlation:

= = − +
1−
The fracture toughness has been calculated using the following correlation
K = 0.313 + 0.027 ∗ E
Another method for calculating the Poisson’s ratio has been used, using Fig. 3 which showing the
relationship between porosity, sonic transit time and Poisson’s ratio. This method is considered less
accurate compared with the method discussed above.
6 S
SPE 148864

Fig. 3: The
T relationship between po
orosity, sonic ttransit time an
nd Poisson’s rratio

TTable-1 show ws the summ mary of the rock mechaanics calculaations for thhe M AR/G and the layeers above
aand below itt. This table showing alsso good alig gnment betw ween the calcculated Poissson’s ratio uusing the
ssonic data annd the porossity charts fo
or the Shale layers but foor the sand llayer there is large diffeerence. In
oorder to coveer these unccertainties arround the Po
oisson’s ratioo, different sscenarios weere used in tthe stress
pprofile calculations. Apppendix-A preesent the diffferent stresss profile scennarios.
Tab
ble 1: The Rocck Mechanics Calculations

OOptimize thhe fracture design

d and pump
p sched
dule
SSeveral hydrraulic fractu
ure designs were
w evaluatted to determ
mine the opttimum fractture design ggiven the
uuncertainty of the streess profile. Appendix--B summariized the reesults of thhe different designs
ccomparison.
AAfter this sennsitivity anaalysis we fou
und that massive hydraullic fracturingg treatment iis required ffor A-1 in
oorder to achhieve long half-length
h and
a high fraccture conduuctivity. Thiss massive fr frac will incrrease the
pproduction rate
r by 2.5 to
t 3 times. The
T fracturee half-lengthh that could be achievedd using the pproposed
ddesign rangiing between n 600 to 80 00 ft. The fracture
fr condductivity thaat could bee achieved uusing the
pproposed dessign ranging g between 10 0,000 to 20,,000 ft MD deepening oon the propppant type thaat will be
uused. Coarsee proppant type t is prefeerred (12-18
8) medium sstrength whhich will givve high condductivity.
S
SPE 148864 7

TThe amount of proppantt required to o achieve thiis design is ~ 360,000 L Lb. In most of the casess the frac
hheight will propagate
p do
own ward toward the carbonate
c laayer. In ordeer to avoid propagatingg the frac
ttoward the carbonate lay yer, very smaall frac desig
gn should bee used (~ 40,000 Lb of pproppant). TThis small
ffrac size will not increasse the produ
uctivity from
m well as reqquired and thhe productioon rate is exppected to
ddrop rapidly after the fraacture. Propp
pant flow paack additive should be uused to avoidd proppant flflow back
aafter the treaatment and av
void damagiing the suckeer rod pumpp.
33. LGR Co
oncept and Design
D
TThe LGR is the process of dividing one or seveeral grids in the reservoiir model intoo smaller sizzed grids
aallowing en nhanced griid definition n, which is i essential for modelling hydrauulic fracturees using
ppermeabilityy multipliers.. Because thhe fracture do oesn’t have a perfect recctangular shaape, in fact iit usually
ttakes an ellip
ptical shape; it is requireed to give diffferent ratioss in the refinnement of eaach layer whhich can’t
bbe achieved using only oneo LGR. Th his limitationn leaves youu with the chhoice of eithher to use ammalgamed
LLGRs or to choose simp plicity and sacrifice mod deling the acctual geomeetry; instead you’ll needd to use a
rrectangular shaped
s fractu
ure which do oesn’t accuraately mimic the actual frracture geom metry.
IIn this modeel several LGGRs were neeeded in ord der to accuraately describbe the fractuure geometryy.5 LGRs
wwere used foor each well, one for eveery layer. Eaach layer waas divided sseparately with differentt ratios to
aallow sensitivity on the fracture
f leng
gth and heigh
ht.
TThis LGR reesulted in inncrease of th
he grid cellss number too be 77,076 cells, whichh is greater than the
ooriginal num
mber of cells by 4,836 cells (6.7% %). This incrrease in the cell numbeer resulted iin longer
ccomputing time
t by aboout 30% sim mulated on the same m machine whhich is actuaally an achievement
ccompared too the normall cartesian LGR;
L a norm
mal LGR w would have increased thee number off cells to
1116,856 cellls (61.8%) increasing the
t computiing time byy up to morre than fifthh times thee original
ccomputing tiime, and thee model requuired a moree powerful wworkstation to be used so as to simmulate the
rresults.
TThe actual fiield productiion history was
w received d after the im
mplementatiion of the opptimum frac scenario
aand was commpared to thee simulated results
r to sh
how a match with great aaccuracy witth the prediccted rates
aand pressurees. This mattch validatess that the teechnique useed in modelling this fraac actually is able of
mmimicking thhe real reserv
voir performmance.

Fig. 4: The dyna

amic without LGR
L Fig. 5: T
The dynamic w
with LGR

T
The keyword
ds used in th
his process were:
w
LLGR: It is thhe first keyw
word used in
n the LGR creation,
c fouund in the R
Runspec secttion, this keeyword is
uused to introduce the nummber of LGRRs present an
nd their mainn specificatiions.
8 SPE 148864

CARFIN: It is presented in the grid section, and it defines the cells included in the LGR and the number
of their subdivisions.
HXFIN &HYFIN: It is the keyword responsible for defining the ratios by which the cell should be
divided; if not included the cell will be divided equally. This is keyword should be Included in the
CARFIN keyword (i.e. Before ENDFIN)
AMALGAM: It is the keyword responsible for combining the separate LGRs in one group. Without this
keyword LGR can’t be introduced in two adjacent cells.
WELSPECL: Used instead of WELSPEC keyword for the wells located in the LGR, it used to introduce
the wells that are present in the LGR cells.
COMPDATL: Used instead of COMPDAT keyword for the wells located in the LGR, it used to
introduce the information about the completion for the wells included in the LGR cells.

Conclusion
• The modeling of hydraulic fractures could be achieved using amalgam LGR as previously shown
with small effect on the computing times and will yield reliable and accurate results as concluded
from the comparison of the postfrac production to the simulated rates.
• Hydraulic fracturing stimulation is expected to increase the production rate from A-1 well by 2.5 to
3 times.
• Long half length (above 600 ft) and high fracture conductivity (above 10,000 ft MD) are required
and can be achieved in order to maximize the production rate and ultimate recovery from the
reservoir.
• Massive hydraulic fracturing treatment will be required in order to achieve the required objective.
• The key risks associated with the fracture treatment are the propagation of the fracture toward the
carbonate interval below M.AR/G reservoir. Bending on the permeability of this carbonate layer,
water production may be seen after the frac treatment.
SPE 148864 9

Nomenclature
LGR (Local Grid Refinement): it's a widely used expression for the process of dividing one or several
grids in the reservoir model into smaller sized grids allowing enhanced grid definition, which is useful
for modeling wells or hydraulic fractures and other complex reservoir structures.
DOE: Department of Energy; governmental department whose mission is to advance energy technology
and promote related innovation in the United States.
Tailored pulse fracturing: employed to control the extent and direction of the produced fractures by
the ignition of precise quantities of solid rocket fuel-like proppants in the wellbore to create pressure
‘pulse’ which creates fractures in a more predictable pattern.
Foam fracturing: using foam under high pressure in gas reservoirs. It has the advantage over high-
pressure water injection because it does not create as much damage to the formation, and well cleanup
operations are less costly.
CO2, sand fracturing: increases production by eliminating much of the inhibiting effects of pumped
fluids such as plugging by solids, water retention, and chemical interactions.
: The minimum horizontal stress (in-situ stress)
: Poisson’s’ ratio
: Overburden stress
: Biot’s constant
: Reservoir fluid pressure or pore pressure
: Tectonic stress
: Average Compressional Velocity
: Shear Velocity
: Total vertical stress (overburden)
10 SPE 148864

References
1. Salz, L.B.: “Relationship between Fracture Propagation Pressure and Pore Pressure”, paper SPE
6870 presented at the 1977 SPE Annual Technical Conference and Exhibition, Denver, Oct. 7-12. 9.
2. Veatch, R. W. Jr. and Moschchovidis, Z. A.: “An Overview of Recent Advances in Hydraulic
Fracturing Technology”, paper SPE 14085 presented at the 1986 International Meeting on Petroleum
Engineering, Beijing, March 17-20.
3. Gidley et al.: Recent Advances in Hydraulic Fracturing, SPE Monograph 12, Richardson, Texas,
(1989), 62-63.
4. Applied Petroleum Reservoir Engineering, B.C. Craft & M.F. Hawkins
5. SCHLUMBERGER Simulation Software Manuals 2009.1
6. “Petroleum Resources Classification and Definitions”, based on the SPE, WPC, AAPG, and SPEE
guidelines.
S
SPE 148864 11

A
Appendix-A
A: The differrent stress profile
p scena
arios

Fig. A-1:
A Stress Pro
ofile Scenario # 1 (Base casee)

Fig. A-2: Streess Profile Sceenario # 2

Fig. A-3: Streess Profile Sceenario # 3

12 S
SPE 148864

Fig. A-4: You

ung’s Modulu
us Profile

A
Appendix-B
B:

Fig. B-1: Fra

ac design optimization

Fig. B-2: Fra

acture Profile Assuming Rocck Mechanic Model
M Scenarrio # 1 (Base caase) and mediium proppantt size (16-
20)
SPE 148864 13

Fig. B-3: Fracturee Profile Assum

ming Rock Meechanic Modeel Scenario # 2
and medium proppant sizee (16-20)

Fig. B-4: Fracturee Profile Assum

ming Rock Meechanic Modeel Scenario # 3
and medium proppant sizee (16-20)

Fig. B--5: A-1 Fractu

ure Profile Asssuming Rock M
Mechanic Moodel Scenario # 3
and medium proppant sizee (12-18)