Вы находитесь на странице: 1из 23
PIPE HYDRAULICS AND SIZING DRASMOHARIR DRPPARANIAPE WHY PIPE SIZING IS IMPORTANT ron-ptimsl Also, what is optimal today According, to 1979 American may not be optimam over along period (due survey, as much a8 309% ofthe total cost 10 fouling, change in relative cos, change cof a Nypical chemical paces plant goes operating schedule which affects the for piping. piping clements and valves. A tlizaton time of the pipeline, ete) significant “amount of operating cost (cnerey is also used up in forcing flow Pipe sizing is thus alot of experience, through piping including its components. engineering foresight and judgment than just Asighfcan amount ofthe mainenance theory. This paper attempts to review the Gost also fr the ping and associated pipe sizing procedures, the pressure drop ‘ings. ‘aleulation procedures Which are integral Proper sing, optimal in some pipe sizing procedurc, the pitfalls in these sense, i therefore very necesser. calculations, the confidence limits in calculated values and the factors of safety WHY IS IT DIFFICULT AND which must be incorporated in view of ATTIMES MEANINGLESS ‘known limitations of correlations, Different Piping must be sued before the N6eDAS are then cemented through plant i laid. ott. Layout, must be TeteSentative examples during the lecture in the cetiene "Couse én Pi compete (ee. eqipment must be ining Toc pipe racks esis, layout of Etgieting conducted by ping Cll at agit in rss else, gout ofr inte of ecg, Mane celcling realise presure Gop aed doing pipe sizing foreach pipe segment. PIPE SIZING PROCEDURES ‘This “chicken and egg’ scenario means Pipe sizing is generally done using that decisions vegadg pipe ing none ofthe following ete plant layout must be ime in most Se Tht ermal tt pre 2 Yelsiy onsertons xcept in few very Inge engineering lable presi drop consdersins fmanizatns which ean affrd ie 3) Economie considerations (Faving 10 cay oUt ppe sizing at 0 ‘premature stage invariably means that ‘The degre offical increases s one the recommended pipe size may not £85 ffom (1) © G). While pressure drop treet process requirement or may notbe cael ian integral pat of 2) and (3), the mest economic, et) it would need to be calculated in ease (I) ‘Normally a layout ie assumed als0 to quantly enenty requirement, sizing drawing on past pracicesand experience PRESSUE providing equipmeat sich as and pipes are sized. No iecond iteration Pumps! compressors, ele. It is, therefore, is carried. out. Actual layout which, YY. important to be conversant with femerges later mayb: significantly’ PreSSure drop calculation procedures. for Giffeent than what vas assumed during Yaiety of flow types tal are encountered sizing. The sizes thus may tum out to be the dst. ‘This paper assumes that the readers are conversant with presture drop ‘calculation procedures and concepts underlining them, at least for the single phase flow. The paper attempts ‘to Duid on this background. ‘The paper reviews the following ‘TYPES OF FLOW ‘Single phase, Two phe, Muiti= phase Horizont, nctned ‘Through straight rapes, through ‘complex routings Isothermal, non-isothermal Incompressible, compressible Laminar, Turbulent BERNOULLI’S EQUATION SINGLE PHASE PRESSURE DROP CALCULATIONS Horizontal, straight, constant cross- section segment Inclined, straight, constant cross- section segment Fitings and valves Equivalent length in actual terms Equivalent length in diameter terms TWO PHASE PRESSURE DROP CALCULATIONS, Flow regimes ani their idenuiications (Baker Parameters) Pressure drop caleulations (Lockhart Martinelli, Baker) Confidence levels in. caleuated pressure drops| ‘¢ Effect of inclination ‘+ Scientific approach MULTI -PHASE FLOW PRESSURE. DROP CALCULATIONS + A possible approach PIPE SIZING 4 Velocity considerations 4 Pressure drop considerations 4+ Beonomie considerations ‘TYPE OF FLOW Although the flow canbe ‘categorized on several basis the classification based on number of phases fnvolved isthe most commonly used. When the flowing medium bs unifoum piysicl properties across the low cross-section, the flow is « single-phase flow. Flow of pure single liquids, soltions of slid in liquids, ‘mixtures of ‘completely miscible tiguids, mixtures of gases and/or vapors come inthis ‘All other flows are multiphase flows. The two phase flow would involve {70 distinct phases such as Liquid with is vapor, a liquid with an incondensible gas, et Tiguid or gas!vapor stream with suspended solid partcles is also a two phase flow. However, ato phase flow would normally refer to two uid phases. When to immiscible liquids are involved with thie vvapor andor another inert gs, it is a three phase flow and s0 on, Eucigy seyuiied Wo sustain sult flows in ppestubes is avery important information which has tobe generated through calculations of pressure drop that the flow would cause in & conduit of given cross-section, and extent. This information is then wsed in locating equipments, sizing pipes, deciding their routes, rating Pressure generating equipments ts. Temperature of the flowing ‘medium affects physical properties such as density and viscosity which in tum Ihave a bearing on the pressure drop. When the temperature is constant over ‘the pipe segment under consideration, or the temperature change along the flow path is not significant enough so a to ‘cause appreciable change in the physical properties, ii treated as an isothermal flow. When the temperature change is signifieant, it is nonsothermal flow. ‘When the density ofthe lowing medium isnot stongly corelated with the pressure, the medium is termed as incompressible and the flow as, incompressible’ flow. Liquid flow (Gingle, or multiphase) would come in ‘his category naturally. However when gases/vapors. which xe compressible (Ghats thei density is strong funtion fof pressure) ate involved, but the pressure drop slong the flow path is not ficant enough to affect the medium density, their flow may also be treated as incompressible flow. Otherwise, the flow of gesapos i compesle In some flow situations, expecially two and muhiphase flaws, the inclination of the flow conduit from horizon! is of great significance. Also ‘whether the flow in the inclined conduit is upward or downvard is also an important consideration. Inthe case of single phase flow, the inclination is important inthe sense that i affects the ‘overall energy balance given forthe low Suation Uy the fanuus Bexuulli's fcqustion, But the Tow wpe and Inydrauic pressure drop are not affected by the pipe inetination, CY BERNOULLI'S EQUATION In its original. form, Bemouli's ‘equation is merely statement of conversation ‘of energy for flowing medium. Consider a Segment of an inlined conduit of variable cross-section a8 shown in Fig and fi flowing trough it. The enerey ofthe Hid at any locetion may be expressed inter of @ vertical column of the flowing fluid itself. ‘The height al any point along the conduit is then seen as comprising of three components, the pressure (Pip), velocity head (v?2g) and elevation head @, Bemoulli’s theorem states thatthe sum of these three componcats constant everywhere along the flow pa, This is true if there are no external inputs or withdrawals fiom the conduit Applied a the two points 1 and 2 ofthe inclined pipe shown (Fig), the Bemoulli’S equation can te ‘writes os follows: P,lptv;?2gtZ,=P,/ptv, */2e+Z2 SEE ‘When the pipe is horizontal (Z,= Z,) and the conduit cross-section is uniform (vi=¥5), the pressures at the two points, 1 and 2, should be equal, Tis isnot the case ‘cause the flow is confined ky the pipe and ther isa resistance to flow caused by fiction between the fluid and the ‘wal, ction between diferent layers of ‘uid flowing a different velocities and the small of big swisls creatod in the liquid ave to flow turbulence. Flow ‘against these resistances causes ‘generation of heat rang the temperature of the Mid a it Bows. This Aemperstare rise is not enough todo any ‘work as this energy transfarmed: into thermal energy is as good as lst energy. ‘This expressed in pressure units or ‘expressed in terms of an equivalent ‘column of the flowing Did is called ‘ctional pressure drop or head lose Tneorporting this fact into the ‘Bernoulli's equation yields the following form which is generally used in calculating fictional pressure drop in flow: Pip+v)2g+Z, =Py/p+v,*/2e+ Za APIp SINGLE PHASE PRESSURE ‘DROP CALCULATIONS Single phase flow is assifid as LAMINAR, TRANSIENT OR TURBULENT. ‘The deciding factor is the REYNOLD'S NUMBER defined as allows R.=2e : H It is a Dimensionless number if ‘the quantics aré in consist units. For Reynolds number valves up 40 2000, the flow is termed laminar and for values above 4000, it i a turbulent flow. The range 2000-4000" is termed as the tratsition region, D in the definition of the Reynold’s number is the actual diameter if the flow cross-section is crear suchas in ete), D is defined in terms of the Hydraulic rdis( R98 follows: D=4x Hydeaulic radius ‘The HYDRAULIC RADIUS is defined a8 rato of flow erose-sectioal area to the weted perimeter For example, it the ‘ise of roctanglar cross-section with sides Sand b, the flow cross-section i ab while the weted perimeter is 222), Silay, for fn annular region a5 shown (Fig2), the Iydratic radius es shown: R, = AEA DIN r=) "ae, + DM ‘With D defined inthis general sense in the definition of Reynold's number, the limiting values of the number for laminar, ransient and turbulent flows remain the same as given earlier. The linear velocity used inthe definition of Reynold's number is obined by dividing the volametic flow ty cies ate for fiw ‘Ateraaive bit quvlet foms of cnn of Ress manber wich fe commonly wed rs follows R28 “ie Where Gis the linear mass velocity of lid ooo \ yee | R631 \" vt \ eo Where W is the mass flow rate Whe, D is pipe 1D in inches and pis density in Ibift® ‘The fiitional pressure drop is ‘calculated using Dar’s equation as follows. fp is termed as the Darcy's fiction factor and is related to the Reynold's number and pipe roughness. ‘The applicable and widely used graphs are given in several text s00ks. For turbulent region, the fiction factor value should be read fiom an appropriate curve for a pipe of roughness & by caleulating its ratio with pipe iameter (@/D). ‘The log-log plo is difficult 10 ‘ea an the reading is enor prone due to nomlincarty of ssale. Several correlations are therfore proposed by various authors so that he fition factor ‘can be calculated from the Reynolds imumber, Some of the frequently used correlations are given lat. In the case of implicit correlations, an iterative approach is necessary to get the value ofthe fiction factor for given value of Reynold’s number, Newton-Rhspson method may be sed forgetting the value in fewer iterations. FFanning’s equation is also used in place of Darcy's equation as follows: he") (ep) ‘Comparison should show that the Darcy's fiction factor is obviously four times the Fanning’ ction facto, fy While asing. any fieon fetor vs Reynolds number graph to Teed ffiten factor and then while using, it inthe formula. to calculate the pres drop, eae must be taken to choose the compatible raph and compatible comelaion. This i oflen 8 source of err. ‘Anober fiction factor is also attbuted to Churchill which is all of Fanning’s icon ctor). The conesponding formula far pressure. drop faleulaion thus has a factor 8 in. the numerator instead of 4 in Paoning’s ‘equation. So, one needs 10 be realy very areal in handling this prevailing muliple ‘efnitons scenano. Generally, chemical ‘ginesing literature uses Faningsfetion factor and Process industy fellows the Davey ction ator. fone uses the f vs R, plot, i necessary to note whether itis for Fanning, Darcy of Chueh icton aco. There fsa simple way to do it which any engineer hl Know If you do ney ponder overs Tie and you would petit. Several simplified coreltions are available to ealeuat fiction factors from Reynol's number under diferent conditions of flow. Some ofthe commonly tse ones ae given Below wih reference the Darcy's definition of fiction factor. Suitable" multiplying factors must be wed t0 convert these comeatons for othe friction factors LAMINAR REGION. 64k, TURBULENT REGION Rough commercial pipes, R, less than 50000: fs6.8x10R,* ‘Smooth Pipe, less than 34€0000 ‘foster, Avot Blau ution fooi6R, Smooth or rough pie, Ress than 3400000, developing. turbulent flow: Most f vs R, plots would mark transition between developing, turbulent flows by a broken line. Most fow situations {in process in industry would fal inthe fly developed turbulent region and Blazius ‘equation (especially the one with R, with exponent ~02) given above is widely used. ‘The roughness factor eis dependent fon the pipe material and method of fabrication and some representative values are given in the Table 1. Note the wide ‘arason in perceptions of the roughness by different authors, In most plots, Mocdy"s roughness values are used. Because of the variation in friction factor definition’ and roughness values, tis advisable to stick to ‘one plot with Full knowledge ofthe fietion factor it pertains to andthe roughness values itrefest. “The frictional pressure drop caleulated by any of the above methods should be multiplied by the effotive length ‘ofthe pipe segment to get the net fitional ‘drop across the segment. This is then used in the Bernoulli's equation to obtain the actual pressure drop between pipe origin and ‘destination, The effective length is the ‘actual pipe length if the pipe line is straight and long enough so tha pressure drop due to extra turbulence created at the entrance when fluid enters the pipe from an equipment or atthe ext when the pipe feeds into another equipment are relatively insignificant as compared to overall {tional pressure drop. In ease the pie has fitings such as elbows, tees, valves, expanders, reducers, ete, an hypothetical straight pipe length of same diameter asthe fin pipe on which the Sings exit ia added inplace of each ofthe fitngs. The effective length is the sum of the stsight-run pipe Tength plus. the total equivalent for all fitings. Entrance and exit of fuid in and from the pipe segment also adds 10 turbulence and to extra pressure drop. This effect is also incorported by adding pressure. However, if itis more than this ‘equivalent length of these. The actual engineering tolerance, above approach of ‘equivalent lengths for important fitings segmenting the pipeline may be adopted. are given in real terms (ie. length of ‘A. good. practice would be to pipe to be added) in Tables 2-5. (The _ealeulate pressure: drop over the pipe run tables are taken from te famous paper assuming fluid properties a inlet or average ‘on practical pressure drop calculations" temperature/ressure conditions to. begin by Robert Kern) with, If the pressure drop so calculated is In another approach, equivalent within 10% or less of the actual pressure length of fitings are mentioned in terms __Jevels at which the fui is flowing, one may ‘of diameters of the pipe. This number ignore the effect of temperature/pressure should then be multiplied by the pipe change IF the pressure drop exceeds 10% af size to get the equivaert length of pipe flow pressure, the above approach of {o be added. The equivalent lengths for _segmeriing may be restored to, valves and fiings in temns of diameters ‘ce reported in several books and are not. ©» TWO PHASE PRESSURE DROP. een he Anges of te tel GAECUEAMNG Spuleghns of ifeet ~< Scceanaiataeee mse (oar dp in eo of 0 “et in phase flow is dependent on the flow regime. ‘a operas Ung a pe ae tae spore. Using acl spe Fue hu fr cnldon 7 opines fre posible as shown in Fig. Flow rei nia tifiaton i done by folowing Baker's “hove procedures applicable to‘ etifeaton is done by following Bak ‘aids, ic. Liquids and pases. Proce. In eases the temperature varies ‘Two Baker parameters By and By scross the pipe seymeat, the physical are ealeulated as follows: properties vary. Also if the flid is ‘2s/vapor, is volumetis low rate may vary duc to pressure changes arising out Bx"531(W./W,)} of temperature change as well as due to pressure drop. To account for these ‘effects, i may be a good practice 10 ‘divide the whole line into segments over ‘each of which, the temperature change is ot 80 significant ab to change the Inthe above definitions, following unis are ope scaly. The ropeies we wot suitably updated to incorporate WY Vapor flow rt, tbe ‘amber and pseu changes & ON Liquid low rate, vit traverses these hypothatical segments. " Calculation overall the segments thus p,~ Vapor density, tot” ee etgs prams acon te (Tt ey, a! pipe may be of hiporunce incase or ~Intemal cros-sctional ara, f? compressible Quids. Itmay be ignored if Hy. Viseosty of liquid, eP it is less than 10% of the total fluid, ~ Surface tension of ligu dymefom Ne at oh te et Phy Fy penonte So Sein is Pec enrems IS SDE paKeR’S METHOD me etaet Tlie postin ae Depg oe i ate fim anh eel ate jes cae, apse enn ct Toning pnitonotic, Sauk ga ere ogy it DANE CEL; lanes dp kn Pein Sonn ge ps pep “Efe dop lain Festtwebiripaton fen pone pat in Seliger Sues sn clowns dws GAP y ‘These comeations were derived by LOCKHART MARTINELLI the respective authors by _ extensive METHOD experimentation on aiewatet flow, but ‘Assuming that only te Higuid mostly on smaller diameter pipes. Thee ows in the pipe line, eabulste the applicability for lager dimension industal pressure drop that it would cause over pipes i suspect, However, these remain the Imost used corelations, Beiter approaches 9 voit pipe length, (AP), Sina, : ‘ ‘to phase flow pressure dop estimation are considering tha only vapocgs lows in" hase flow pressure drop ‘the pip, alclate the presse dropper Snes GAGE Tn two phase. flow calculations, eee eee confidence levels are low. Aso, itis not safe orelations are to be used in geting tp overdsign here asthe flow regime may these two pressure drops. Lockhart change and ene may get an undesirable low Martinelli Modulus, X, is thendefined 8s Sogime uch ae’ sug’ flow, Exteme an precaution is therefore necessary at : tngincering stage in designing pipes for to X*(AP) MBP) phase flow and the engineer must be ready {o handle problems that may surfece at the For this value of modulus, a commissioning stage. mutipier Ys or Yy the ca fom The Bale map is apple only if tne plot in Figs and it is appropriately low line al on has & used in oe of te fllovinglatons owt fet om flow pater andthe low ct the "two phase pee drop, YekiMe: may change for same. vapor fe har liquid flows in same size pipe line i the CPDex ee unit feng. Multiplying ipetinatons are diferent. Also, in inclined Pedra eroetve engin (atter pipes, it matters whether the low is upward including, equivalet lengtis of the or downward. Extensive work has. been fitings) of the pipe, one ges the tial PRportad "on these. aspects but. nds ‘so pase ftiona drop. prctices ted to ignore this fact BP) ¥, (AP), MULTIPHASE —_ PRESSURE. DROP CALCULATIONS Two immiscible or partly miscible liquid phases end a gas phase ‘comprising. of vapors of these liquids andlor other gases give rise to three phase flow situations. ‘There are no reported reliable presture drop calculation approaches for three phase flow. What is proposed bere is 2 possible extension of the Locthart Martinelt approach which was reasonably svecessful in using single phase flow Correlations and predicing two phase flow pressure drop. The approach would ‘be something like this: StepI Consider only that the liquid phase including the two ligids is flowing through the pipe. Let thse liquids be T and L . Using Lockhart Marinelli method or other method (say Baker’), calculate the pressure drop per unit length that would be caused in this case. Letthisbe AP, ‘Step IT ‘Consider only gusivapor is lowing and calculate the pressure crop that would ‘occur per unit Jength using single phase pressure drop comelation. Let this be Abo ‘Step It Calculate the Lockhart Martine modulus a8 was done ia the two phase Flow situation a fallows KPa ARy Ae Step IV For this value of modus, ¢ multiplier Y, Gey Jor, (or¥) is then rad forthe plot in Fig, 5 and itis appropriately used in cone of the fllowing relations to get the thre phase pressure drop, (AP)ap per uit Jeng (afte including equivalent lengths of the fitings) of the pipe, one gets the ttl three pase coma pressure drop. (APY AR (AP)y oA 1 may be appreciated that this is nothing but using the Lockhart Martinelli approach on itsell In absence of any other ‘correlation with proven merit this is ikely (0 bea good engineering aproseh, PIPE SIZING “The earlier mentioned three pipe sizing approaches are discussed ere in brie. PIPE SIZING BASED ON VELOCITY CONSIDERATIONS This is the simplest of approaches Herein, recommended values of linear velocities for the flowing medium are used slong withthe design flow rates to baek out the pipe diameter. Recommendations for the linear velocities may arise due to process ‘considerations, material of construction considerations, corrosion considerations, ‘economic considerations. based on prior experience etc. or a combination of these. Consider the following exampies 8) In a steam carying pipe, if the linear steam velocity is beyond a certain value, the flowing steam may pick up the condensate, break it up into fragments ‘These entetined condensate droplets may impinge agnnst the pipe wall causing erosion and erosion corrosion, ¥) Too low a steam velocity in steam headers may mean a large diameter pipe for design requirement of steam, ‘This would inerese pipe cost, insulation cost, cic. thereby adversely affecting econonies. © A gaseous steam camrying Particulates (uch as pnewnatc solid transport lines) must flow above a ‘minimum velocity to eliminate solids setting down at pipe botom causing flow obstruction, increased pressure Arop ete. A gaseous steam —canying pariculates must not flow above a Certain linear velocity tc eliminate severe erosion of pipeline or elbows ©) Aline carrying two phase must be of suitable dimension so. that cersin ‘two phase flow regimes (aich as slug low) are avoided or acerain regime jis guaranteed (such as concentric flow). 4) Linear velocities in exhaust lines should be below eertsin level to keep ‘nose within acceptable levels These are just representative examples to help appreciate the origin of such restrictions on linear velocities of flowing medium, ‘Some of the more accepted linear velocities in a variety of design eases are ‘complied in Tables 6 and 7. PIPE SIZING BASED ON AVAILABLE PRESSURE DROP. This is a more involved method of pipe sizing and perhaps the most important. Pipes are sized here to meet ‘ertain process requirements. ‘These process requirements are translated into the maximum hydraulic pressure drop that ‘one can accept over the pipe segment of interest. A minimum pipe size which causes a pressure drop at the most equal to this maximum acceptable pressure drop is thus recommended. Any size more than this size would also be acceptable, but would be ‘uneconomical as it would involve higher capital cost. The procedure would be one of tial and emor. A commercial pipe size would be assumed in terms of NB. The pressure design ofthe pipe would decide the Schedule. From the appropriate tables, the ID of the pipe size would be obtained. Taking this asthe hydraulic diameter and for the design flow rates, hydraulic pressure drop over the proposed pipe route is calelated wsing appropriate pressure drop comelations. If this pressure drop is more than the acceptable eve, a highee pipe size is taken for next traf the pressure drop is ‘much smaller than ‘that acceptable, next lower pipe size can be tried. Minimum pipe size meeting the pressure drop requirement is recommended. Some important sitwatons where pipe sizing needs to be done using avaliable pressure drop consideration ae as follows: 1. Suetion Pipe Sizing fora pump: A. igi fs to be pumped from a storage tank to ‘an equipment. The storage tank pressure is fixed. On its way ffom the storage tank to the pamp suction, the liquid ‘would loose pressure dve (0 frictional pressure drop. If this pressure drop is excessive, the fluid. pressure as