A rae ee | ae ok ae are ie eT) Loe tee ee re TWO-PHASE BAD WP oleae th EditionTWO-PHASE FLOW IN PIPES by Dr. James P.. Brill Dr. H. Dale Beggs Sixth Edition ‘Third Printing January, 1991 Copyright© 1978 by Dr. James P. Brill and Dr. H. Dale Beggs‘The notes included herein were prepared by Dr. James P. Brill and Dr. H. Dale Beggs for use in short courses taught by either or both of them and as a textbook in a graduate level course at The University of Tulsa. Publication of this material in whole or in part is prohibited without approval from the author: ‘The notes should not be considered to be a thorough review of all Published works on the subject of two-phase flow in pipes, The authors have attempted to compile, in a concise and uniform format, a collection of information necessary to perform pertinent engineering calculations ‘frequently necessary in designing two-phase flow facilities, ‘Mastering material contained in these notes will provide an individual ‘with a firm foundation in the important concepts which differentiate single from two-phase flow. ‘This material is not designed for students to merely read sbout a correlation and then use off-the-shelf computer programs to solve difficult design problems. Competent computer programmers could Program correlations covered after mastering concepts, or modify existing in-house programs to méet specific objectives,TABLE OF CONTENTS 1-PRINCIPLES OF TWO-PHASE FLOW References. 1.3. Evaluation of Friction Losses - The Friction Factor Concept en ‘Laminar Single-Phase Flow... Turbulent Single-Phase FLOW cereus ‘Smooth Wall. 1.4 Single-Phase Flow. Elevation Change Component... Friction Loss Component.. 1.5 Definitions of Variables Used in Two-Phase Flow. Surface Tension snnnennnnne1.6 17 18 19 TABLE OF CONTENTS (Cont) 1-PRINCIPLES OF TWO-PHASE FLOW Modification of the Pressure Gradient Equation foe Two-Phase Flow. ‘Two-Phase Friction... Acceleration Component. Flow Patterns... Calculation of Pressure Traverses.. Procedure for Iterating on Length Increment ; Procedure for Iterating on Pressure Increment... Procedure When Temperature Distribution is Unknown see Effect of Slippage on Composition Coupling Algorithm. References .. 1.10 Problem: 7TABLE ‘OF CONTENTS 2-FLUID PROPERTY CORRELATIONS Solubility of Natural Gas in Water. Water Formation Volume Factor Water Compressiblity.....TABLE OF CONTENTS (Cont) 2-FLUID PROPERTY CORRELATIONS Effeet of Non-hydrocarbon Impurities on 7. Calculation Algorithms for 2-£8CtOF menwinnnmnrnemn Z-Factor Comparison Study ano 2.13 Liquid Viscosity. Oil Viscosity... New Correlations for Water Viscosity, Emulsion Viscosity... 2.14 Gas Viscosity. Car etal. Correlation... Lee et al. Correlation... 2.15 Surface Tension. Gas-0il .. Model... Gas-Water. 216 References annum 2.17 Problems: Page 2-40 2-44 2-450 2-48 2-48 2-53a 2-53 2-55 2-58 2-58 2-60 2-64 2-64 2-65a 2-66 2-68 2-7131 3.2 3.3 3.4 3.5 3.6 3.7. Evaluation of Pressure Loss 3.8 Flow in ANNU snnonnnn 3.9 3.10 3.11 3.12 3.13 Coupling Wellbore Hydraulics to Reservoir 3.14 Temperature Prediction mem BAIS References semen 3.16 Problems4a 4.2 43 4.4 45 4.6 47 48 49 4.10 alt TABLE OF CONTENTS 4- HORIZONTAL FLOW YntrodUCtON nenncnnnsininnmrnenenn Horizontal Pressure Loss Prediction Method: Eaton et al. Correlation... Beggs and Brill Correlation. Dukler et al. Correlation (Case I - Constant Slip)... Guzhov et al. Correlation emninseinnnnnnenns Lockhart and Martinelli Corelation. Oliemans Correlation. Other Liquid Holdup Correlation Hughmark and Pressburg Correlationecmmun Hughmark Correlation.. Taitel-Dukler Dimensionless Momentum Balance Equations.esjsenoan Taitel-Dukler Flow Regime Determination.. Use of Panhandle Equations in Two-Phase Flow. Miscellaneous Pipeline Design Problems.. Design of Slug Catchers.... 4-77 Flow in Riser Pipes. rnmaninie 4281 Condensate Movement in Pipeline Networks... Use of Gradient Curves... References vreinnen ProblemsTABLE OF CONTENTS 5-INCLINED FLOW Sa 2 5.2 “2 5.3 23 ‘Use of Vertical Correlations. Beggs and Brill Correlation 34 Guzhov, Mamyev and Odishariya.Corelation, Beggs and Brill Correlation... Gregory, Mandhane and Aziz Method oun. 5.5 Downhill Flow. wn 5-32 Bonnecaze, Erskine and Greskovich Correlati 5-32 Gallyamov and Goldzberg Correlation.. 5-33 Greskovich Method. 5-33 Other Methods.... 5-34 5.6 Evaluation of Inclined Flow Correlation’ mewnmomene 5°34 Hilly Terrain Pipelines... 5.7 References Problems 5.861 6.2 69 64 65 66 67 6. FLOW THROUGH RESTRICTIONS Introduction. 1 Description of Restrictions - o2 Critical vs Suberitical Flow «+. 66 ‘Single-Phase Suberitical Flow - es Liquids»... 8 Critical Flow Correlations .. Suberitical Flow Correlations. Evaluation of Correlations ...+.seeeeuee0 Prediction of Two-Phase Critical Plow .. References... Problems... seesee 6919 sesee GAT b-49AL AB Ad AS Ag AT AS AD A.10 An APPENDIX FORTRAN Lletings of Pressure Gradient Subroutines~ - ~~ ~~ FORTRAN Listings of Fuld Physical Property and ‘Related Subroutines = = = = = = == = Properties of Saturated Water: temperature table = Properties of Saturated Water: pressure table = ‘Properties of Water: superheated vapor table Properties of Water: compressed Liquid table ~ Enthalpy-Log Pressure Diagram for Water and Steam = Ant ‘Vertical Flow Gradient Curves: 1.995 In, .D, = = sae Ants ‘Vertical Flow Gradient Curves; 4.494 In, LD, == =~ = ~~ === A-62 Horizontal Flow Gradient Curves: 2.0m, LD, - Horizontal Flow Gradient Curves: 4.0 tn, 1.D. Ae? Ant9 As A-st = As3T Antsva oe area formation volume factor compressibility factor specific heat taneter depth friction factor fraction acceleration of gravity gravitational constant mass flux rate igas of Liquid holdup *atlo of epecitic heats, C,/C, axtal Length molecular wetghe moles Reynolds number Froude number Liquid veloctey mumber gas velocity number Atameter oumber Mquid viscosity numbera volumetric flow rate-imesity conditions volumetric flow rate-standard conditions es constant in situ gas-liquid ratio Producing ges-ol1 ratio solution gas-oil ratio time temperature velocity volume specific volume nase flov rate lorizontal Length ‘Presnure ratlo across choke, py/p, mole fraction as compressibility factor vertical Length SREEK LETTERS roughnes specific gravity atfference angle no-slip gas or Liquid holdup viscosity mnsity ~ standard conditions density + ineeieu conditions Anterfactal tension DINENsIONS ye re let wile att a mHwee a ngs rg ate acceleration bubble bubble potne base conditions bean or choke erteteal ataneter dissolved elevation formation (rock) free frtetion Froude eas s28 velocity sntetat Kinetic energy Liguta quid velocity nixture slip ot jeudo-reluced Paeudo-erttteal Reynolds iiiee a y solution usp, sonic standard conditions suparfictal Hguid superfietal gas vater hortzontal direction vertical direction 1,2,3, ete, time or Location Je1= upstream, 2 = downstream sathCHAPTER 1 PRINCIPLES OF TWO-PHASE FLOW12 1 = PRINCIPLES OF TWO-PHASE FLOW 1.1 Introduction Tho prediction of pressure gradients, Liquid holdup and flow patterns occurring during the simultaneous flow of gas and liquid in pipes {s necessary for design in the petroleum and cheaical industries. Petroleum engineers encounter two-phase flow most frequently in well tubing and in flovlines, The flow may be vertical, inclined ox horizontal and methods must be avail- able for predicting pressure drop in pipes at any inclination angle. Off shore producing has necessitated transporting both gas and Liquid ph over long distances before separation. Besides being able to size these Lines from a prassure loss standpoint, the engineer must be able to calcu- Late Liquid content in the pipeline at various flow conditions in order to design separation and slug catching facilities. ‘Tworphase flow occurs frequently in the chemical processing the indusery a Feactors and process piping requires methods to predict pressure érop, Liquid holdup, and for heat transfer purposes, flow pattern, ‘Te purpose of this manual {s to present the basic problens Anvolved in two-phase design, outline the correlations currently available for design purposes, and to demonstrate the application of the various correlations. lesign of such facilities as condensers, has exchangers, 1.2 The General Energy Equation Tne theoretical basis for many flutd flow equations is the general ‘energy equation, an expression for the balance or conservation of energy between two points in a systes. The energy equation is developed first, and using thermodynamic principles, {s aodified to a pressure gradient equation. The energy balance sinply states that the energy of a fluid entering a control volume, plus any shaft work ie on or by the fluid, Plus any heat energy added to or taken from the fluid, plus any change of energy with time {n the control volume must equal the energy Leaving the control volume. Figure 1.1 may be used to illustrate this prineipl13 Figure 1.1 Flow System Control Volume Considering a steady state eysten, the anergy written as lance may be ‘ wey YAM ATE tah tatu tte yt Te se vhere: U'= internal energy, pV = energy of expansion or conpression, ay? = Kinetic energy, 38, af = potential energy, & heat anergy added to flutd,st Mi + vork done on the fluid, and 2 = elevation above reference datum Dividing this equation by m to obtain an anergy per unit mass balance and weiting in differential form gives: wea CB) Bh tat agtay no. ‘Tis form of che energy balance equation ts difficult to apply because of the internal anergy tera, 40 it t¢ usually converted to 4 mechanical ‘energy balance using well known thermodynamte relations, From thermo- dynanics: GO cccececeee ata ana where = epecitic enthalpy, 5 = entropy, and T > cenperature. Subatteuting fq, 1.3 into £4. 1.2 gives ete a2heag Au HEE ACE eek PEGE OL 1a For an irreversible process, the Clausis inequality states that 482 8, opa5 Tasedatdh,» where d 1 = losses due to irreverstbtlitten, such as friction. Using this relationship and 1.4 becomes ming no work 1 dove on or by the fluid, 2 rd yt ata = aesectenanasneceereesssiber Pe fo ars at = 0... see 1.5 TE ve consider a pipe inclined at sone angle 9 co we nortzontal, as tn Pig. 1.2, since az = dl ein 6 ae SAY YR at ind + at = is 7 $a oinb 4 a4, = 0 wuleiptyiog the equation by “Ee glves Be BEbe ey aub es oGE oO. a6 Fig. 1.2. Flow Geometry1.2 Supplemental Material Conservation of mass Conservation of mass simply means that for a given control volume such as a segment, of pipe, the mass in minus mass out must equal the ‘mass accumulation’, For a constant area duct, BF r= 0 For steady state flow, pv = constant, and no mass accumulation can occur. Equation S1 then becomes : $1 Bev) = 0 eesnmnnannnannnnennnnnennnnie SE Conservation of Momentum ‘Application of Newton's First Law to fluid flow in pipes requires that the rate of momentum out minus the rate of momentum in plus the rate of momentum accumulation in a given pipe segment must equal the sum of all forces on the fluids’, Figures 1.1 through 1.3 define the control volume and pertinent variables. Conservation of linear momentum can be expressed as Zonder However, the rate of momentum efflux term can be expanded as follows Hor =F oneorg .. s4 ‘Combining Eqs $2, $3 and $4 and assuming steady state flow [(pv)/at = 0] to eliminate the rate of accumulation of linear momentum gives 85 v3 @M a15b Solving for the pressure gradient, one obtains 4 Td 7 dv $Rane ld pgsno-prSt antnernarees $6 which is also frequently called the mechanical energy balance equation and is identical to Eq. 1.7. Thus, the steady state pressure gradient ‘equation is actually a result of applying the principles of conservation of mass and linear momentum. Eq, S6 clearly shows that the steady state pressure gradient is made up of three components, The first component is due to friction or shear stress at the pipe wall and will be described in greater detail below. Friction losses normally represent 5-20 % of the total pressure drop in wells and nearly all of the pressure drop’ in near-horizontal pipes... The second term in Eq. $6 is the pressure gradient due to elevation change (often called hydrostatic head or elevation component). It is normally the predominant term in wells and contributes from 80-95% of the pressure gradient. The elevation component is much smaller for pipelines, but is often comparable in magnitude to the friction component in hilly terrain pipelines. The final component is due to change in velocity (often called acceleration or kinetic energy component). It is normally negligible and can become significant only if a compressible phase exists at relatively low Pressures. For upward flow, pressure always drops in the direction of flow. It is common to adopt a sign convention that pressure drop is positive in the direction of flow. Eq. $6 must then be multiplied by -i to yield a positive pressure gradient. Evaluation of the wall shear stress or friction losses can be accomplished by defining a dimensionless friction factor that is the ratio of the wall shear stress to the kinetic energy of the fluid per unit volume'. Thus, te a pv where f is a Fanning friction factor. Throughout this monograph we will adopt a Darcy-Weisbach or Moody? friction factor, which is 4-times larger than the Fanning friction factor. Eq. S7 can then be solved for shear stresswe 88 Substituting Eq. S8 into Eq. $6, the frictional component of the pressure gradient equation becomes 2)-(224)-2 which is often called the Darcy-Weisbach? equation, Calculation of frictional pressure gradients requires determining values for friction factors. The procedure first requires evaluating whether the flow is laminar or turbulent. Laminar flow is considered 10 exist if the Reynolds number is less than approximately 2,000. The Reynolds number is defined as For laminar flow, an analytical expression can be derived for the fiction factor. The velocity profile for laminar flow can be integrated to yield the pressure gradient. This was done for flow in horizontal capillary tubes to give Eq. S11 which is the Poiscuillet equation, iC) - Since the pressure gradient in Eq, $11 is due only to wall shear stresses or friction, ‘the pressure gradient is identical to Eq. $9. Combining these equations gives seers SUL . 812 ur ability to predict flow behavior under turbulent flow conditions is a direct result of extensive experimental studies of velocity profiles and Pressure gradients. These studies have shown that both yelocity profile and pressure gradient are very sensitive to characteristics of the pipe wall, A logical approach to defining friction factors for turbulent flow i to begin 15615d with the simplest ‘case, ie. the smooth wall pipe, proceed to the partially rough wall and finally to the fully rough wall. ‘Smooth pipe is seldom encountered in the production of oil and gas. However, laboratory studies conducted with glass or plastic pipes often ‘approach smooth pipe. Numerous empirical equations have been proposed for predicting friction factors for smooth pipe. The equations most often used and their suggested ranges of applicability are: Drew, Koo and McAdams; f= 0.0056 + 05 G2" .. $13 3000 3x 10° BG ENE sane <1 Although the Blasius equation is considered less accurate, it receives greater use due to its similarity to the laminar friction factor equation. Both can be expressed in the form t=cMh, ‘The inside wall of a pipe is not normally smooth In turbulent flow the roughness can have a significant effect on the friction factor, and thus the pressure gradient. Wall roughness is a function of the pipe material, the method of manufacture, the age of the pipe, and the environment to which it is exposed. From a microscopic sense, wall roughness is not uniform. Individual protrusions, indentations, etc. vary in height, width, length, shape and distribution, The absolute roughness of a pipe, e, is the mean protruding height of relatively uniformly distributed and sized, tightly packed sand grains that would give the same pressure gradient behavior as the actual pipe.Dimensional analysis suggests that the effect of roughness is not due to its absolute dimensions, but rather to its dimensions relative to the inside diameter of the pipe, e/d. In turbulent flow, the effect of wall roughness has been found to be dependent on both the relative roughness and on the Reynolds number. If the laminar sublayer that exists within the boundary layer is sufficiently thick, then the behavior is similar to 2 smooth ‘pipe. The sublayer thickness is directly related to the Reynolds number. Nikuradse's? famous sand grain experiments formed the basis for friction factor data from rough pipes. His correlation for fully rough wall Pipe is given as Eq. S15 and is still the best one available, dpa 2tee(22) The region in which f varies both with Reynolds number and relative roughness is called the transition region or partially rough wall. An empirical equation describing the variation of f in the transition region that is the basis for modern friction factor charts was proposed by Colebrook? sis 1 ae, 187 primeater(F+ so) $16 Note that Eq. S16 degenerates to Eq. 2.15 for large Reynolds numbers corresponding to fully developed turbulent or rough pipe flow. Solving Eq. S16 for £ requires a trial and error process. Eq. S16 can be expressed as? ten[ite-ateg "E+ 182. ee si7 Nel Values of f are assumed (fg) and then calculated (fe) until they agree to within an acceptable tolerance. A direct substitution process of using the calculated value as the next assumed value results in convergence in only 2 or 3 iterations... The initial guess can be obtained from one of the explicit smooth pipe equations, or from explicit approximations to the Colebrook equation. Lse1st Numerous explicit approximations’ to the Colebrook equation have been proposed. One of the most accurate of these that is also relatively simple to use was given by Zigrang and Sylvester’, gota f-Hial 2] ~ In most cases Eq. S18 can be used in Hien of Eq. S16, wes $18 The variations of friction factors with Reynolds number and relative roughness based on Eqs. $12 and S16 are shown graphically in Fig. 1.4. It is important to emphasize that © is not a property which is physically measured. Rather, it is the sand grain roughness which would result in the same friction factor. The only way this can be done is by comparing the behavior of a normal pipe with one that is sand-roughened. Moody® has done this and his results, given in Fig. 1.5 are still accepted values. These values should mot be considered inviolate and could change significantly by such things as paraffin deposition, hydrates, erosion or corrosion. Thus, if measured pressure gradients are available, a friction factor vs Reynolds number relationship can be established and an effective relative roughness obtained from the Fig. 1.4. This value of e/d should then be used for future predictions until updated again, Initial values of roughness are often needed for design calculations. ‘The recommended value for new tubing is e = 0.00005 ft. A common value used to generate pressure gradient curves has been 0.00015 ft. For tubing that has been exposed to an environment that causes significant changes in Toughness, “very dirty” pipe can have roughness values of 0.00075 ft'!. For most wells, the friction component of the pressure gradient is small compared to the potential energy component. Consequently, approximate values for absolute roughness are normally sufficient. Calculate the pressure change in a water injection well. The following data are known.* D = 8,000 ft w= 20,000 bbl/D © =-90° py = 624 Ibm it?15¢ d =5.0 in, py = 1.0 ep .00006 ft The average velocity in the pipe is From Eq. S10 the Reynolds number is 299.5325) ran oseincat GR) , Since Nre > 2000 the flow is turbulent. The relative roughness for the pipe is i- 2.00006 _ 0.000144 oO) From Eq. S16, $18 or Fig. 1.4, f = 0.0155. Calculate the pressure gradient from Eqs. $6 and $9, neglecting acceleration effects. 9, ~10.0155)(62.4(9.531)" _(62.4)32.2)(sin -90") a Ee a 72. ) =-3:274 + 62400 = 59.126 BE = 00227 + 04333 = 0.4106 88 ‘The pressure change is then ‘Ap = (-0.0227 +.0.4333) (8000) = 71819434664 = 32845 psi Note that the pressure change consists of a loss due to friction of -181.9 psi and a gain due to elevation change of +3,466.4 psi.15h * In the examples in this text, the gravitational conversion constant, g¢ = 322 lbp filly s2 appears when problems are solved using customary wits, ‘Non-Newtonian Fluids Fluids encountered in the petroleum industry often act as non- Newtonian fluids. These include: many of the drilling muds; fluids such as cements, frac fluids and spacers used during well completion activities; and, many of the oils and oil-water mixtures that are produced. The material presented previously in this section is valid only for Newtonian fluids. The design of non-Newtonian piping systems becomes complicated since the use of conventional friction factor correlations is not directly applicable. These non-Newtonian fluids may assume any type of theological behavior depending on such factors as shear rate, temperature, and fluid composition. Two methods are commonly used to design piping systems transporting oil-water mixtures. The first method treats the mixture as a Newtonian fluid with an apparent viscosity that can vary with water fraction. This method is covered later and involves combining the viscosities of each phase to obtain a mixture viscosity. The second method treats the mixture as a non-Newtonian fluid and is based on the following assumptions 2: 1, The mixture is homogeneous. 2. Slippage between phases is neglected. Thus, in-situ holdups are the same as their respective input volume fractions. 3. The rheological behavior of the oil-water dispersion system is suitably described by the power-law (Ostwald-de Waele) model’, Power Law Model ‘The power-law stress-strain relationship can be expressed as YT een sacnsenansnnsens $19)If n’ is unity, Eq. $19 will describe Newtonian behavior, and K’ will be equal to the constant viscosity, ». For a typical oil-water mixture, n° will usually be less than unity, and Eq. S19 will describe pseudoplastic (shear thinning) behavior. It is also possible for an oil-water mixture to have n' greater than unity, resulting in dilatant (shear thickening) behavior, The type of behavior that a fluid system will follow is normally unknown and can be determined from laboratory experiments using an appropriate viscometer. These tests must be conducted for a specified set of the following operating conditions: input water fraction, temperature, and droplet size of the dispersed phase (or, indirectly, mixing speed). Once the fluid system is correctly characterized, the prediction of frictional pressure gredients for pipe flow of this particular oil-water dispersed system can be readily determined. Generalized Reynolds Number Metzner and Reed" introduced the concept of a “generalized Reynolds number” into the field of non-Newtonian flow. The usual Reynolds number definition is f ‘$12 which can also be written as pvd vn (HE) sto Where tw is the laminar wall shear sess, while the quantity (84) is also the true shear rate at the pipe wall or (- rd \y for a Newtonian 4 uid. Thus, the ratio ( F7) corresponds to the Newtonian viscosity, 1. For non-Newtonian flow, the relationship between the true shear rate and the apparent shear rate can be expressed as 1511-5} va) (3041) fee) ($5]-e82(¢9) a where n’ is the slope of a logarithmic plot of ty vs S22. From the same plot, the equation of the tangential line to the laminar flow curve can be written as where NRey.p is the generalized Reynolds number, and 1m is the apparent viscosity of the mixture. Eqs.S23 and $24 clearly reduce to the normal Reynolds number and the Newtonian viscosity for a Newtonian fluid, when n‘= 1 and K’= yp, Non-Newtonian Friction. Factor Laminar_Flow: Following Metzner and Reed, the Fanning friction factor for non- Newtonian laminar flow can be written as: roth rnnnanimeninan S2S15k Dodge and Metzner'* proposed the following implicit friction factor ‘equation, 4.0 (-a'R) vi rors 8 ENaen S26 Govier and Aziz'® suggested the following friction factor for power- Jaw pseudoplastic fluids flowing in rough pipes : =tua(i) , sennctnnnnnannnnsenne SET where f'-R = friction factor caleulated from Eq. $26, and f; = Fanning friction factor for rough pipe, calculated at the same generalized Reynolds number, fs = Fanning friction factor for smooth pipe, calculated at the same generalized Reynolds number. recommended the following; 1 _ 4.07 d Feta 2) s60- Eq. S28, for Newtonian fluids, essentially reduces to the von Karman equation, .06 top| (4) 3.36151 Erictional_ Pressure Gradient Once the Fanning friction factor, f, is determined for the particular dispersion system, the frictional pressure gradient can be calculated from, . $30 (2 228 Pave dL Jp, da As can be seen from the above equations, K’and n° are the two t parameters which will affect the frictional. pressure gradient ‘import calculation for a dispersion system. These two parameters can be experimentally determined with an appropriate viscometer. Example_S2_Non-Newtonian_Qil-Water Mixture Pressure Drop A horizontal pipe discharges an oil and water mixture at atmospheric pressure. Determine the pressure at the inlet of the pipe required to maintain a constant volumetric flow rate of the dispersion for ‘an input water fraction of 20%. The following are also given: q =30 f3/s 4d = 20 in, e 0.001 ft L = 10,000 ft Po = 53.74 Ibmift3 py = 62.23 Ibm/ft3 w = 08589 K = 71,475 Ibtst £2 a. Calculate the mixture density, using Pm = Po fo + Pw fy = 55.4 Ibm/ft3 b. For the given flow rate and water fraction, calculate the generalized Reynolds number, raat Rey. = BOD = 997K IOS‘The criterion for turbulent flow considered to be NReyy py > 1500. For this example the flow regime is turbulent, ¢. Compute the friction factor for turbulent flow of pscudoplastic fluids in rough pipes, using f, f= fur (#) 0.007 (1.006) = 0.00704 where, |. * fy.n is given by Eq. S26, f, is given by Eq. S14, fy is given by Eq. S16. 4. Calculate the pipe inlet pressure, py ; L2 : PL =P2 +2f Pm GF Ym = 205.9 psia References 1 2. Knudsen, J. G. and Katz, D. L. : Fluid Dynamics and Heat Transfer. McGraw Hill Book Co,, New York (1958). Moody, L. F.: "Friction Factors for Pipe Flow," Trans, ASME (Nov. 1944) 66, No. 8, 671-684, Allen Jr., T. and Ditsworth, R. L. : Fluid Mechanics. McGraw-Hill Book Co., New York (1975). - Poisouille, J. L.: Compte Rendus, 11, 961 and 1041 (1840): 12, 112 (1841). Drew, T. B., Koo, E. C. and McAdams, W. H. : Trans, Am, Inst. Chem, Engrs., 28, 56 (1930). 1.5m15a een e 10. Ml, 12, 13. 14, 15, 16. . Blasius, HL : Z. Math. Phys., 56, 1-37 (1908) Nikoradse, J.: Forschungsheft, p. 301, (1933). + Colebrook, C. F.: J. Inst, Civil Engrs, (London),11:133 (1938). Brill, J. P. & Beggs, H. D.: Two-Phase Flow in Pipes, The University of Tulsa (1978), Zigrang, D. J. and Sylvester, N. D.: "A Review of Explicit Friction Factor Equations,” Trans. ASME, J. Energy Res. Tech. (June 1985) 107, 280- 283, Theory and Practice of the Testing of Gas Wells, 3rd Ed. Energy Resources Conservation Board, Calgary, Alberta (1975). Martinez, A. E., Arirachakaran, S., Shoham, O. and Brill, J. P, : "Prediction of Dispersion Viscosity of Oil-Water Mixture’ Flow in Horizontal Pipes" SPE 18221, presented at SPE Annual Technical Conference and Exhibition, Houston, TX, Oct. 2-5, 1988, Bird, R. B., Stewart, W. E. and Lightfoot, E. N. : Transport Phenomena. John Wiley & Sons, New York (1960). ‘Metzner, A. B. and Reed, J. C. : "Flow of Non-Newtonian Fluids - Correlation of the Laminar, Transition, and Turbulent-Flow Regions,” AICHE J., ( Dec. 1955) 1, no.4, 434-440. Dodge, D. W. and Metzner, A. B. : "Turbulent Flow of Non-Newtonian Systems,” AICHE J. ( June 1959), 5 , No. 2, 189-203 Govier, G. W. and Aziz, K.; The Flow of Complex Mixtures in Pipes, ‘Van Nostrand Reinhold Co, New York (1972).16 ‘This equation can be solved for pressure gradient, and if ve consider a pressure drop ds being positive in the direction of lov de em path + ove + (ee a 8, ar (Care vhere 44, (aE “ts the pressure gradient due to viscous shear oe Tin ts the aredient 4 ane or friction losses. 1,3 Evaluation of Friction Losses - The Friction Factor concent In horizontal pipe flow the energy losses or pressure drop ie caused by change in kinetic energy and friction losses only, Since nost Of the viscous shear occurs at the pipe wall, the ratio of val shear stress (7,) to kinette energy per untt volume (pv"/2 g,) reflects the relative inportance of wall shear stress to the total losses. This group ard defines « friction factor. oerete To evaluate the wall shear stress, a force balance between Pressure forces and vall shear stress can be formed. Referring to Fig. 1.3,sere LY hE) cece + be Substituting Eq. 1.9 into Eq. 1.8 and solving for the prassure gradient due to friction gives 1.10 which 49 the well known Fanning equation. In terms of a Darcy-Weisbach x Moody triton fastor, fy bf) and 2 ae ig tree dL art z 2 Laminar Single-Thase Floy ‘The friction factor for laninar flow can be determined anelyti- ‘cally by combining Bq. 1.11 vith the Hagen-Poteeutlle equation for laminar flow or seeeeseeLel Bquating the expressions for frictional pressure gradient gives tovita ‘Example 1.2 Amo wich ha vison of 80 op (0.08 tetany of S01, 1 (01.8 kg/m) flown in 1 In. (25.4 mm) daniter pipe at th rate of 0.07 f° /eee (0.00198 m"/eec), Determine the pressure drop which wil occar over 8 100 ft (30,48 m) sectiod of the pipe. Solution: ve q/A = 0.01/[0/(4 x 127)] = 12. 84 ft/s00 igo 7 PAn = GONA2. £4)(0, 0899)/50/1408 = 1588 Flow la temainar, since Np, <2000 ‘Therefore f= GA/Ny, = 04/1593 = 0.04 patos? ab/ag, 4 = (0, 04)80)(12, 847(100)/(2(82.2(0, 0883) = erm thi/n? = 42,95 pat (295.45 k Pa)‘The equivatene expression for the Fanning friction factor 1s veld ‘The Darey-Weisbach or Moody* frtetion factor will be used Chroughout the remainder of thie text, 40 the subscript, a, will be dropped. Turbulent Stngle-Phase Floy Our ability to predict flow behavior under curbulent flow co ditions is a direct result of extensive experimental studies of velocity Profiles and pressure gradients. These studies have shown that both velo- city profile and pressure gradient are very sensitive to characteristics of the pipe wall, A logical approach to defining friction factors is to degin with the simplest case, i.e. the smooth wall pipe (Si), proceed to the partially rough wall (PRH) and finally co the fully rough wall (FRI). Only the most accurate enpirical equations available for friction factors are presented here, Although studies of velocity profiles are important, their use is purely academic. Saooth Walt Yor snooth wall pipes, several equations have been developed, each valid over different ranges of Reynolds nunbers. The equation which ts now used most comonly since it ts explicit in £ and also covers a range of Reynolds numbers 3000 ys an - Prreerereeeeeeeees teteeeen ee LIS, a ene Using the above definitions for velocity, an alternate equation for no-sltp holdup ts s0e1.36 Yiscostey ‘The viscosity of the flowing fluid 2 used to determining « Reynolds wunber as vell as other dinensionless numbers used as correlating Parameters. The concept of a tvo-phase viscosity ts rather nebulous and As defined dttterently by various investigators. Sas ‘The viscosity of various gases may be estinated from enpirical correlations described in chapter 2, on Enpirical corzalations for both gas free and gas saturated crude ‘olls are given in chapter 2. ‘Ligute ‘The viscosity of an ott vater mixture 4s usually calculated using the water-oil ratio as a weighting factor, The equation ts HT tH fy 1.37 Teochase Te following equations have been used to calculate a two-phy viscosity.18 seeeeeeeee D3: By TAL AL ty Nyeeseeeeseseeeeeeseeaee wy ey Surface Tension Correlations for the interfacial tension betveen vater and natural gts at various pressures and tesperatures are given in Chapter 2, The Anterfactal tension between natural gas and ervde ofl depends on ofl gravity, temperature and dissolved gas, anong other variables. Expirical correla- elona for eatination of this fluid proparty are given tn Chapter 2. When the Liquid phase contains both vater and ofl, the same = water surface t tom, aod f,, £, are oil and water fractions respectively. 1.6 Modification of the Pressure Gradient Equation for Two-Phase Flow The pressure gradient equation which 1¢ applicable to any fluid floving in 4 pipe inclined at an angie § from horizontal was given previously ‘The equation is usually adapted for tvo-phase flow by assuming that the gas- Liquid mixture can be considered homogeneous over « finite volume of the pipe.1-20 where p, 42 the density of the gas-liquid alxture in the pipe elenent. Considering a pipe element vhich contains Liquid and gas, the density of the mixture can be calculated from seceeeel oT BT ALB toy By aoe where H, is the Liquid holdup and 1s defined as volume At volume of pipe elenent The friction loss component becomes. fey Pe Vg Pea (2. (hy, where £,, and o; ere defined differently by different investigator: ‘The concept of two-phase friction factor is discussed in the following section. FeorBhase Friction Freviously 1© wan shown that the tem {2 sepresents che * Pressure losses due to friction vhen gas and liquid flov staultsneously in Pipes. This term is not analytically predictable excepe for the case of laminar single-phase flow, Therefore it must be determined by experimental means or by analogies to single phase flov. ‘The method which has received by far the most attention ts the fone resulting {n tvo phase friction factors, Auong the most common defini~ tons are the folloving:send Im general, the two phase frfetion factor methods differ only in the way the friction factor is deternined and to 2 Large extent on the flow pattern. For example, in the mist flew pattern Equation 1.44, based ‘on gas is normally used, whereas in the bubble flow regime Equation 1.43, based on liquid ts frequently used, The definition of pg can differ widely depending on the investigator as discussed previously. Most correlations attenpt to correlate friction factors with sone form of a Reysolds mmber, Recall that the single phase Reynolds number is defined as ae seveeeee LAO "he one consistent set of unite frequently used for calculating Ng, is: p= density sm Tba/ét? vm velocity tn ft/sec = pipe dteneter in fe be viscosity te tba/ ft Stnce viscosity ts nore comonly used in cp, the Reynolds sunber vith # in ep. ts 1.47 pa OBE eescceneneennenneeess122 ‘Another frequently found definition of Reynolds nuaber vhich agsin nerely Involves a staple untt conversion ts eevee eh 2 where Wd cece teceeeneee reese “Re a where we mass flov rate in Ibm/day and Ae tpe area, fe? w=ep ‘The various Reynolds numbers used to evaluate friction factors are defined when the friction factor correlations are discussed for the individual correlations in future chapters. One variation vhich deserves mention is that several correlations for predicting vertical floving pressure losses make use of only the numerator of the Reynolds number. These include the works of Poettmann and Carpenter, Baxendell, and Fancher and Frown. Acceleration Consonent ‘The acceleration component {9 completely (gnored by sone investi ‘estore and ignored in sone flow regines by others. When it is considered, various assumptions are nade regarding the relative magnitudes of paraneters involved to arrive a€ soue simplified procedure to determine the pressure drop due to kinetic energy change. From the discussion of che various components contributing to the Sotal pressure gradient {t follows that the princtpal considerations for developing pressure gradient equations are developing methods for predicting Liquid holdup and eworphase friction factor. This is the approach followed by almost all researchers in the study of evo-phase floving pressure gradients, 1.7 Flow Patterns Whenever tvo fluids vith different physical properties flow simultaneously in a pipe, there is a wide range of possible flow regimes. By flov pattern, we refer to the distribution of each phase in the pipe relative €o the other phase. Many investigators have attempted to predict the flow pattern that vill exist for various sets of conditions, and many Ai €ferent nanes have been given to che various patterns. Of even more stentficance, some of the uote reliable pressure lose correlations rely on123 a knowledge of the existing flov pattern. Also, since the increase of tvo- phase Lines from offshore platforms co on-shore facilities, increased con- gern has grown regarding the prediction of not only flow pattern, but pected Liquid slug sizes. Prediction of flow patterns for horizontal flow is a more diffi- cult problem than for vertical flow. For horizontal flow, the phases tend to separate due to differences in density, causing a form of stratified flow to be very common. Govier’has presented a series of flow pattern de~ scriptions for horizontal aix-water flow and vertical air-vater flow. ‘These are shown to illustrate the various patterns that can result and also to show that all depend to some extent on the relative magnitudes of vy and vg When flow occurs in a pipe inclined at sone angle other than vertical or horizontal the flow patterns take other forms, For in- clined upward flow, the pattern is almost alvays slug or mist. The effect of gravity on the Liquid precludes stratification, For inclined downward flow the pattern is usually stratified or annular. 1.8 Caleulation of Pressure Traverses ‘The calculation of: a two-pha: use of an iterative or erial flowing pressure traverse involves ind-error procedure if tenperatures or pipe tn- elimation change with location or distanc: flow conduit {s divided into « number of pressure or length increments and the flutd properties and pressure gradient are evaluated at average condi- tions of pressure, ceaperature and pipe inclination in the incresent, The accuracy of the pressure craverse calculation increases as the aumber of {1 crements increases, but 40 does the number of calculations performed, This presents no problen if a computer is available, but the time involved may be significant for hand calculations. The increments should be smaller at lover pressures where the preseure changes rapidiy with distanct puter calculations the pressure increments should be no Larger than one- Eenth of the average pres In calculating a traverse, the wee in the increment. Algorithns are given both for incrementing on pressure and incrementing on length. In both eases It Ie assumed that the tomporature gradient is known,‘vest DISPERSED ¥, Vou, ft/sec 8 of omecton = Ae Superficial Woter Velocity | Superticie Gos Velocity, Veg, ‘h /see HORIZONTAL FLOW PATTERNS I TWO-PRASE FLOW After Govier and AzizVERTICAL FLOW PATTERNS IN TWO-PHASE FLOW After govier and Aziz?1-26 lure for I Langth In 1, Starting with the known pressure value, p, , at location L select a pressure increaent, Ap. 2, Starting at the location vhere the pressure 18 known, estinate ‘a length inerenent, AL, corresponding to the pressure increment, Ai 3. Calculate the average pressure and, for non-{sothermal cases, the average temperature in the increnent. Temperature nay be a function of Location, 4, From Laboratory data or enpirical correlations, detemine the necessary fluid and PVT properties at conditions tenperature. 5. Galoulate the pressure gradient, average conditions of pressure, tenpe: the appropriate pressure gradtent correlation. 6. Calculate the length increnent corresponding to the selected Pressure increment, AL = Ap/(4p/AL). 7. Conpare the estimated and calculated values of AL obtained in eteps 2 and 6. If they are not sufficiently close, estinate @ new Length increment and go to step 3. Repeat steps 3 through 7 until the cotinated and calculated values are suffictently close. 8. SEL =, +7 AL and pep, +, ap. f average pressure and /aL in the increment at cure, and pipe inclination, using 4.38 ah fu ten than che tote conte Iengh, rsuen to Sween the last two values of L to obtain pressure at the end of the conduit. ‘The above procedure is iterative {£ either temperature or pipe inclination are functions of length or location. For isothermal, constant pipe inclination cases, steps 2 and 7 may be omitted. ‘This procedure must be used vith caution for calculating two-phase pressure eraverses for dovn- vard flov. In this im the dtrection of flov vhile the friction loss component causes a pressure decrease in the direction of flow. This can create conditions of either # negative or sero presvure gradient which will cause problems in step 6 of the procedure. ‘The result would be either a negative or infinite length increment. This Possibility te eliminated if the following procedure is used. je the hydrostatic pressure inerea12T bu, vores aT tet Hinterpolat Fine Lieeviee_| Flow GHART YOR CALCULATING & PRESSURE TRAVERSE (QTERATING ON LENGTH1-28 wee for Iterating op Pressure Increment 1, Starting with the know pressure, p,, at location Ly select @ Length oceans, Bt. 2 Evins pronouns Soeranst, Sp entspendig t Lng tacrnent, at se Cleaate the average pretuce an for so-tethenel cose, tes erage tnperttare to doce Fron Latortary data or mpicicl gerreatioty deters the owceency fd ol 7T properties at eodicione of average Pesre end crerentete 5, Calculate the pressure gradient, dp/dL, in the increment at J cenisios of prsvere, tmperttre sad pipe tnlintion, woe the apyropciate pressure qadlee cortlaton ¢. etetee te presareinzennecezepeding fo the length increment, Ap = AL (4p/dL), 7 ompate She otntat nt caesate vltes of bead tn saps 24nd 6H toy ace wee sufleleniy clove, eatones «ten postr tserenent and rar Eo step 3. Bepect steps 3 earmaph 7 act the eacinaed ot calelate values are autigtnty cose hsetet+tat ad pen tar: jected 9. tf Son ts tess chan the total conduit length, retern co etep 2. ing this procedure the length increnents can be selected s0 that thetr oum fe exactly equal to the total conduit length snd snterpola- fon is not required in the last step. Also, the calculation of « negative or zer: tem in step 6, This method 1s always fterative even {f temperature and inclination angle are constant, since fluid properties in an increnent depend on the unknown Pressure. Also, it 1s not possible to select Ap as some fraction of Pressure im the pipe, If the length increments are set equal, Large pressure reps vill result in the lov pressure, high velocity segnents. Pressure gradient in dovmvard flow presents no pi1-29 Read Data Toittalize = €(1, P) Gale, ap/au 4p, = AL(@p/ai) mn t0r Flow GHART FOR CALCULATING A PRESSURE TRAVERSE © fees 780 Print Bro IMeseage (LTERATING ON. PRESSURE)1-30 Procedure When Temperature Distribution is Unknown ‘When & more rigorous method of determining temperature distribution ie de~ sired, it is necessary to account for heat transfer to or from the flowing fluids, Heat transfer calculations for two-phase flow can be very Important when calculating prosoure gradients in geothermal wells, steam injection wells, wet gas pipelines in offshore locations or in cold climates, flow of high pour point erudes, ete. to general, heat transfer calculations are always preferable to assuming known tem- perature distribution, but for multicomponent systems, they require that elther the inlet or outlet temperature be known, ‘To perform heat transfer calculations, It Is first necessary to change the energy balance equation to a beat balance equation, Combining Eqs. (1.2) and (2.24) and aseuming no work is done on or by the fluid (4 W,= 0): viv yg as EE dd QO ccs eeceseeseee le a &, ; a If specific enthalpy and the heat added term are expressed as heat per unit mase, then the mechanical energy equivalent of heat constant, J, must be introduced. sare t82 ,£ azssaqno te see eeeses (1:80) Jn English unite, J has a value of 778 ft Ib/BTU, Expressing elevation in terms of pipe length and angle and solving for specific enthalpy gradient ccc eseeeeeees 5D) dh dog at #34 aug, ‘The heat added to the system per unit length, dq/dL is negative since heat is Jost to the surroundings when the fluld temperature ia greater than the surrounding temperature, ‘The heat loss gradient can be expressed as windy Fe seeeeesees 152)131 where = average fluid temperature over dL, 7 Storage proud or surrounding temperature over dT U = overall heat transfer coeffictent total mass flow rate, Heat Transfer Coetficlent ‘The overall heat transfer coefilclent can be a combination of several covifictents which depend on the method of heat transfer and the pipe configuration, For unburied pipelines, there will be convective heat losses between the flowing fluids and the pipe wall, conductive losses through the wall and through any ingula- ‘tion or coating material and convective losses to the enviroament, There can slo ‘bo significant thermal radiation transfer. A complex mixture of beat losses can exist in wellbores due to the variety of materials through which the heat must flow. For example, 2 well may be cased and the amulus can be cemented, Liquid fitted or gas filed, In general, the overall heat transfer coefficient Is the reciprocal of the ‘sums of the individual resistances to heat transfer®, Consider a buried pipeline, ‘Then ++ 2.89) — RyRy TR where R, = Fesletmee to conductive heat transfer from the plpe to the ground = resistance to conductive heat transfer through the pipe wall and coatings Ry © resletance to convective heat transfer between the flowing fluids ‘and the pipe wall Bp For steady state heat transfer, Rem be expressed as 7182 InaD,/s) "Ta cee teeeeeeeeeeeseeneeaeeees (Le 5A) where Do = depth from surface to centerline of pipe. (Frequently taken aa depth of fill on top of pipe). k_ = thermal conductivity of earth, For unsteady state heat transfer, Ry can be expressed ag 7 t By Tey Fe lasdoeheh! ++ (1-55) where f(t) 16 a time dependent function, ‘The conductive resistance due to the pipe wall Le normally expressed as the steady state, radial, one-dimensional conduction equation” mel) mr Lageaeerorars eee) where Tyy%, outer and inner radii of the pipe k= thermal conductivity of the pipe Ry, can require modificat ions to account for scale or paraftin buildup on walls or to account for insulation or corrosion protection coatings and wraps. ‘The convection term, Ry» is extremely complex in two-phase flow and depends ‘on flow pattern in addition to normally accepted single-phase flow parameters, Heat transfer due to thermal radiation is frequently combined with the convection term, ‘The convective resistance is inversely proportional to the Nusselt factor, Nyy and the thermal conductivity of the wetting fm on the pipe wall, N,, in turn depends on the flow characteristic. laminar, transition or turbulent and is normally correlated with three common dimensionless groups: Reynolds number, Grashof1-83 number and Prandtl number. In turbuleat and transition flow, convective heat transfer can be sigaficent ao that temperature differences between the ud and the wall are ignored, in almost all two-phase flow problems flow is turbulent, A ‘commonly used approximation of tho Nusselt factor” is ay us 0. = 0,027 Npe’® Npr tala Nags ses 57) 4 KNaw = thermal conductivity of fds see (1,58) ‘where Npy * Prandtl number = ¢,/k oy = specific heat at constant pressure For flow ina wellbore, heat transfer 1s 4 combinstion of convection in the ‘casing-tubing annulus and conduction into the earth. ‘The time dependent function in Eq. (1-55) for long times of a week or more can be expressed as ° x, tein {—2-) - 0.200... () where 7,7 outer radius of casing ‘== thermal diffualvity of the earth t= time since well began flowing. ‘Typleal values of f( range from 0.5 to 9.0. Eq. (1.59) 18 totally inadequate for short time hest transfer applications or ‘when non-steady flow ocouts, ‘This is the case for transient well tests, cyclic1:34 steam injection, etc, {In most two-phase design calculations involving buried plpeline or wellbores, R, 1s by far the largest of the posaltle resistances to heat tranefer. The result will be U=1/R,, Ifa pipeline 1s insulated, however, R can become the largest torm, resulting essentially In U=1/Ry, Seldom 1s convective heat transfer com= parable to the conductive terms. However, for unburied pipelines, the thermal radtation losses can make this term very important, Magnitudes of overall heat transfer coefficients can vary widely. Insulated, buried pipelines can have U values as low as 0,1 BTU/(hr°F ft). Uninsulated, un- buried Lines can have values higher than 100, Example Probl Flow in a wellbore is controlled by unsteady heat conduction to the ground. Calculate the overall heat transfer coeffictent after 30 daye (720 hours) If 4i=0.5 tt ko 14 BrU/@r °F) 0.04 0 ae, wos (ct) -om 9.5. ome 2{(0. 09 720) Solution: =__2.8 Ba. 576.051485 = 0,990 he f? °p/BrU v=a/a, =1/0,900 = 1.01 BrU/er #0 OF E rroblem Eetimate the value of R, for hot water injection fato a well f 20.4 BTU/be ft °F 4,-2.0 8 ip 7 200,000 Ib, ae p= 0.9 op = 0,726 Ib, /ft he T= 300°F gp 2 0-5 BTU/ID,, 200,000), _4 @.726) #0) at y 3 Nyy" OO Ne! Nee .08 18 0.02.81 x 105) 0.906)188 = ways 9.50 x10°° 2? rere Mixture Enthelpy ‘The total mixture enthalpy of a two-phase gas-liquid mixture at a given pressure and temperature can be calculated from a knowledge of the in-situ mass ‘or mole fraction of each phase and the enthalpy of each phase, If x 1s the no-slip ‘e38 mass fraction (quality), then the no-slip mixture enthalpy ia Demy HBL veesesreceesnereeesee + 60) where hy yb = specie enthalpy ofthe mixture, quid and gus, respectively, For several years {t has been assumed that geothermal welle normally flow ‘m the mist dow pattern Go that allppage 1s negligible. Recent unpublished work by Duler suggests that most geothermal wells flow in the alug flow pattern, slippage of gas past Lquid was shown by Gould” to have a stgnifieant effect on mixture enthalpy and thus the determination of temperature and heat balance. Although his ‘work pertained to compositional modeling of two-phase flow in pipelines, it probably 4s also important in geothermal welle, Gould showed that alippage has an effect on both enthalpy and in-situ quality. ‘Slippage results in accumulation of liguid and a corresponding decrease in quality. ‘The in-situ quality, x', can be related to Liquid holdup by a-x9/ (1.61) Solving for (2-62) ‘The mixture enthalpy ia then calculated from,1-31 henge thy ax) and accounts for the additional mass of Liquld present. Eifect of slippage on Composition “Gould” presented a development which suggested that slippage causes the composition and the feed rate of the flowing mixture to ditfer from the In-situ mix~ ture, Of eritical Importance is his conclusion that the composition of the quid present due to slippage in identical to that present without elippage, Although flash calculationa are not covered in thia text, the following is presented for those interested In Gould's development. ‘Under equillbriam conditions, a mole balance can be written for each com- ‘ponent at any point in the pipe. ZpFom Lata Ly ty ceseees (64) where F = toveltu total motes of foed 1," moles of liquid without slippage 1, = moles of liquid due to slippage V = moles of vapor 2, = mole fraction of component 1 in lula ‘= mole fraction of component j in vapor = mole fraction of component | in feed, ‘Te ita fod F Ls ls gor than the Lat fud to to ply by the amount of hugld present eto alippage, IL, i removed, Gould tates tht by deialtion te compotion ofeach pha will not change, However, the total molee hae changed to FF wlth comportion 21. A mote balance then yield,1-38 seseteeeeeees (1-65) By Pea Pog hy nota 2 ab compultn of he apt owing mire sn hat or ead sat Aw, 2, and are constant, ‘he nabs somprlen acolo lp 0 (1.66 ‘The composition of each phase, x, andy, are available from flash calculations of the input flowing mixture, F'. Because of slippage, the total in situ composition In the pipe, Zp will vary from pont to polnt according tothe last equation. es ition ‘Two levels of sophletioation can be employed when coupling the heat balance and mechanteal energy balance equations to calculate pressure and temperature changes simultaneously, In the simpler case, Eq. (2. 60) 1a used to calculate mixture enthalpy. In the more complex case, Eq. (1.65) 1a used, Gould” called these cases the constant composition and variable composition cases, respectively, with beat balance. Use of Eq. 1.60 results in ignoring slippage affects in the heat balance, although slippage effects are still accounted for in the pressure gradient ‘equation, A simplified flow chart of a computing algorithm for these cases Ia given on the next page. ‘The flow chart differs slgaificantly from those on pgs. 1-27 and 1-29 and $9 discussed below. Convergence on both pressure and temperature in a given pipe length Incre= ‘ment requires a double iterative procedure, Most efficient results are obtained ‘when converging on the least sensitive variable first. Normally, mixture enthalpy 15 more sensitive to temperature than to pressure, ‘This ie shown on a following page im which h (BTU/100-mole) is plotted vs‘ for several isobars and for a specific feed composition’®, Similar curves can be developed for any feed composition using roaulte from vapor-liquid oquilibrium flach calculations and empirical eathelpy correlations. The outer "DO loop” of the flow chart ix thus the temperature loop1-39 FLOW CHART FOR COUPLING MECHANICAL ENERGY AND. HEAT BALANCE EQUATIONS SUAS TF, Tb ats ied grearer Gunes Mw Ti femee he | (eg. + 60" Create OATVariation of Mixture Enthalpy with Preomure and Temperatireith ‘and the inner "DO loop" Is the pressure loop. ‘Cases have been found in which convergence can only be obtained using temperature as the inner loop, it is probable that this region corresponds to 3 ‘temperature range in which 2, 1s essentially a constant du to the simultaneous effects of pressure and temperature reduction, Zemert'* reports that convergence oa temperature can be accelerated by taking advantage of the linear trend of DH (see box with asterisk) with T. ‘Two values of ,; can be extrapolated to an Im- Drove value of, correrpnding to DEO, ‘Only Gould” has published results based on coupling the heat balance and mechanical energy balance equations, Others are mown to have duplicated his rresul:s for the constant composition caso with heat transfer, Gould reported a single computing case for the variable composition case, but a physically impo temperature discontinulty was reported at the pipeline inlet, The discontimity probably remulted from an incorrect representation of hy and several inveatigatora are currently studying this problem.ase References 1, Colebrook, C.F. J. Inst, Civil Engrs. (London), 11:139 1999): 2, Drew, T. B., Koo, E. C,, and McAdains, W. H.: Trang, Am. Inst. Chem, Engrs. 28, 56 c980). 3. Govier, G. W., and Aziz, K. "The Flow of Complex Mixtures in Pipes", (Book) Van Norstrand Reinhold Co,, New York, 1972, 4. Moody, L. F.1 "Prietion Factors for Pipe Flow", Trane, ASME, V.66, Ps 671, 194, 5, Nikuradse, J.: Forsehungsheft, p, 801, 1993, 6, Adams, J. A. and Rogers, D. F,: "Computer-Aided Heat Transfer Analysis", (Book) McGraw-Hill Book Co,, New York, 1973, 7, Sailas, A. P. + "Production and Transport of Oll and G: ‘Elsevier Sclentific Publishing Co., New York, 1975, '» (Book 8. Gould, 7. L.: "Vertical Two-Phase Steam-Water Flow in Geothermal Wells", Trans, AIME, 1974, 833, 9. Gould, T, L.: "Compositional Two-Phase Flow in Pipelines", Preprint No. ‘5685, Presented at SPE Annual Fall Meeting, Dallas, Texas, Sopt. 1075. 10. Goncalves, C. A. de C.: M.S, Theals, U. of Tulsa, in preparation, 11. Zeunert, D.1 Private Communication, Oil Services Co. of Iran,is 7 Problems 1, 1 When 0,009 m°/see of water flows in 2 76 rm pipeline, is the flow laminar fo turbulent? Water viscosity is 0,001 kg/m-sec. 1. 2 Dorive an expression for Reynolds mmber {a torms of q, d and v. ws ‘When ol! of kinematic viscosity 0.001 7/eee and specific gravity of 0,92 flows af a velocity of 5 ft/sec through a 2 in, plpeline, the head loas a 16 ft in 100 {8 of pipe. What willbe the bead loss when tbe velocity is increased to 10 fe/aee? 1. 4 ‘The mean veloctty in a 300 mm pipeline 1s 9 m/sec. ‘The relative zoughness of the pipe is 0,002 and the kinematic yiocosity of water is 9x10" m/sec. Determine the pressure drop in kN/m~ in 300 m of pipe. 1. 5 Ol having a viscosity of 3 ep and density of 60 1b, /f flows in a 6 in, smooth plpe at a velocity of 25 ft/sec. Calculate the fricHlon factor using the Drew, Xoo and MeAdams, the Blaslus, the Colebrook and the Jain equations, 1. 6 A smooth brass pipeline 200 ft long Isto carry & flow rate of 0.3 f° /see between two water tanks whose diference in ourtace elegation is 5 ft. If the water viscosities and denalties are 1 op snd 62.4 1b, /ft respectively, what pipe diameter is required? 7 1. 7 When frosh water flows through 6 m of $0 mm smooth pipe the head lost is 0.3m, Calculate the flow rate, 1, 8 A liquid of epecific grayjty 0.85 flows in a4 in, diameter commercial stool pipe at a rate of 0.15 ft°/eec. If the pressure drop over 2 200 fi length of horizontal pipe Je 0.25 pal, determine the viscosity of the Liquid. . 1. 9 Fresh water (p= 62.4gb, /ft° , 4 = 1.0 ¢p) flows tn a smooth pipeline at a eyootdo manor of 10% “Aer soverd youre of eit covered that al the origizal flow rate produces the same presmuze drop as the original ow ate, Entimate the relative rouginers of the deteriorated pipe. 1.10 ‘There Is a leak in horizontal 0.3 m pipeline having constant friction factor of 0,025. Upstream of the leak two gauges 600 m apart show a pressure drop of 188k Pa. Downstream from the leak two gauges 600 m spart show a Preanire Gop of 104 era, At what rate 1 water beg low rom the pie 2 1.12 Liquid and gas flow in a 2 in, pipe at In situ flow rates of 0.0277 and 0.0377 ft°/sec respectively, Calculate the superficial Liquid and gas velocities and the no-slip Liguid holdup.qe 1.18 1.16 att ‘The liquid holdup was measured fp be 0.95 in an Gin, pipe in which liquid and gas are flowing at rates of 0.5 ft/sec and 3,25 ft/sec respectively. Calcu- fate the superficial liquid and gas velocities, the actual liquid and gas, velocities, the slip velocity and the no~elip Liquid holdup. Discuss how slippage In two-phase pipe flow causes H, and }, to have dilfer~ ‘eat values. ff and gas having densities of 50 1b, /ft° and 4 1b, /ft respectively flow in a pipe at superficial velocities of 8 ft/Bec and 20 tt/Bbe respectively. ‘The liquid holdup is 0.47. Calculate P, 9, #04 A, ‘An oll well Ja producing a 0.75 specific gravity oil at the rate of 100 bbi/day. ‘The well also produces 60 bbI/day of 1.05 specific gravity salt water. Cal- ulate the liquid density, the WOR and the fraction of water flowing. Gas and liquid are flowing simultaneously in a 6 in, pipe Inclined at an angle of -2.27 degrees from horizontal, The following data are obtained at « given location in the pipe: a7 1-903 1°/e0c a, = 0,982 10/a0e y= 500 & 3 2 7 9 = 90 b, /2 9,710, / a Hy = 0.458 fyp = 0.042 Using Eq. (1.45) for the friction component; calculate the pressure gradient {in the pipe at this location, using a slip density.4s at as Given the folloving flov paraneters and fluid phyetcal properties: 4, % 1000 sTBO/D R, = 450 sof/stB0 By * 1000 sef /s780 = 1.27 bbL/STBO 4 = 1.995 tn, 20.85 Y, = 0.70 = constant . API = 30° (y, = 0.88) P = 2500 peta 1 = 20°F ©) Galevlate vey 5 Yeg and ae Dy W caleutate 2, and 9, at py Te 2 Oe tot sate 12°85 oa cetculace 22 | gn pai/ft if flow ts vertical. a ® Calculate v, and v, , the actus! Liquid and gas velocities at P, 7. Elevation (fe) ¢ floving only vater with the folloving terrain1-46 1.18 (continued) Aleo given: p, = 1000 psi ay = 5000 stBW/D 3, = 1.0 bOI/STEW Ay = 64 Toayee? thy © 1.0 ep d= 6.0 ta, £ = 0,018 Calculate che dowserean pressures p, and py . 1.19 Given the following: AB byte ast ep Pg = 4 Ib/fe Uy = 0.01 ep Ye = 6 ft/sec. @ = 2.0%, Mgt 9 ft/sec. © = 0.0006 fe. 4. Calculate the no-s1ip two-phase pressure gradient in psi/ft for vertical flow. Friction factors can be obtained frow the Jain equation 2.0 te [-« -2 Lona * ual] ae ke Ay the slip velocity 1s 3.0 ft/sec, caleulate the 1quid holdup, ileCHAPTER 2 FLUID PROPERTY CORRELATIONS22 2, FLUID FROFERTY CORRELATIONS 2.1 Introduction Solution of the pressure gradiont equation requires that one de able to calculate values of fluid density, velocity, and viscosity. Also, many of the two-phase flow pressure loss prediction methods require values for surface tension to calcalate correlating parameters. Pluids encountered in the production, injection, and transportation phases of the petroleum and natural gas industries are normally water, and hydrocarbons in the gaseous and/or liquid state, Although little emphasis is placed on fluld physical properties in this course, this section is included ‘since prediction of these properties is an integral part of pressure loss calculations. Whenever measured laboratory data are available on fluid proper- es, they should always be used. Even if available at values of pressure and ‘temperature not encountered in the wells or pipelines they can be used to improve predicted values, Empirical correlations are available for predicting ‘every fluid property of interest in multiphase flow operations. However, care must be taken to assure that correlations are not extended beyond thelr ranges of validity. Since required data are frequently not available at pressures and temperatures encountered at flowing conditions, they must be predicted with empirical correlations. The following sections summarize the best correlations available for predicting fluid physical properties. Listings of several FOATAAN subroutines for calcelating various properties are given in the Appendix. 2.2 Solution Gas-Ol Ratio Under conditions of increasing pressure, a crude oil will absorb available gas into solution. Conversely, gas will evolve from 2 saturated ‘crude ofl under conditions of decreasing pressure. ‘The composition of the Mquid and gas phases is a vapor-liquid equilibrium problem, For light oils, ‘Such 8 condensate fluids, the amount and composition of each phase should be predicted by flash vaporization calculations, However, for normal gravity ‘erudes, both of asphalt, paraffin and mixed bases, empirical correlations are23 available, ‘The two most geneial correlations were developed by Lasater™! and by standing 29 ‘The Lasater correlation {8 froqpently considered to he the more accurate and was developed from data on black ofl systems produced in canada, Western and Mid-Continental U, 6., and South America, The ‘Standing correlation is based only on gas-crude systems from California. Both correlations are represented by the following functional relationship. Py Hy Yer THY) Chierici, et al§ suggest that the Lasater correlation be used for °API>15 and ‘the Standing correlation for °API<18. Lasater Correlation ‘The basis for the Lasster correlation ie the following relationship: Ye o%, T whore: y, = mol fraction of gus = fy» M.) » 1M, = effective molecular weight of tank ofl, RY/379.8 B50, ye PT Pais, T=R, R,» sef/stno. M,, wae correlated with tank oll API gravity as in Figure 2.324 ‘eae ‘ee weg gare Sega eae My Fig. 2.1. Effective molecular weight related to tank-oll gravity (From Lasater™ ‘The empirical correlation was represented by Figure Fig. 2.2, Lasater's correlation of ‘bubble-point-pressure factor ‘with gas-mole fraction, (From Lasater #1) ‘Fricklprovides two equations to deseribe the line in Figure 2.2 but they have been found to create a discontinuity where the two coincide, For performing hand calculations which require a value ofp, oF Rgy Figure 2.8 can be used and combines Lasater’s work into a convenient nomogram,25 Chart for determining bubble-point pressure or solution gas-oll ratio by Lasster's correlation, (From Leseter®? ) Fig. 2, ‘The procedure for using Figures 2.1 and 2,2 to obtain a value of R, {8 a8 follows, 5, 1, Caleulate_° fs.c. where p, ts any “saturated” pressure F and vs Se the gravity of the total free gas at standard conditions, 2, Obtain y, from Figure 2.2, 3. Obtain M, for the known API gravity from Figure 2,1,246 4, Calculate ‘hn upper Iii for R equal tothe total prodvetng gualigid ratio, Fy must be established, The oosficient C was proposed by Cheri et l.° ant is used to adjust R, to ft available date st reservotr conditions, 1 no measured R, data are available, C 1s assured to be 1,0, ‘Standing Correlation Direct caleulation of solution gas-oil ratio is pesstble for the Standing correlation using the following equation, 1/0, 88 nye, [Be « a00085029 | Sec. [Te “novo | where py Ye» Care the same as for Lasater and T = °F, ‘a. ‘When performing hand calculations, it is frequently easier to uuse the nomograph shown in Figure 2.4 to obtain a value of R, from the ‘Standing correlation,a7 Fig. 2.4, Chart for determining bubble-point pressure or solution gas-oil atio by Standing's correlation, (California Research Corpors~ ton, copyright 1947) 2.8 Oil Pormation Volume Factor The oll for..ation volume factor, B,, is merely a factor used to predict the change in volume of ofl as it undergoes a pressure and temperature change, The change in volume is 2 combination of compressibility ‘effecte,thermal expansion and mass transfer, which is represented by gas evoly- ing from or gotng into solution. This B, is the volume of oll at p and T occupted by one stock tank barrel of oll and the gae in solution, ‘The only empirical correlation frequently used to predict B, was developed by Standing. 29 For a saturated crude, B, = 0.972 + 0,000147 1-275 4 ¢tee ren (Gh) +1257 © se where: B, = bbl/STBO Tr R, + scl/STBO C= additive factor similar to that for R,, Hf no measured data are available, C is assumed to be 0.0. When performing hand calculations, it is frequently easier to use the nomograph of the above equations which is shown in Figure 2.5, Pig. 2.5. Chart for determining ofl-formation volume factor by Standing's correlation. (California Research Corporation, copyright 1947)20 Lasater R, fall available gas 1s n solution, oo that an increase in pressure results tn an undersaturated crude, then Standing's equations cannot be used. For this case, the compressibility ofthe oil must be estimsted, and B, values can be approximated from: B, = By Exp [-c,-n) | pis calculated from the Lasster or Standing solution gas correlations using R, =R,,amd B,, is obtained from the Standing correlation for this R Although ¢, 18 not a constant, it can be considered so for most pipe flow problems and can be estimated from the correlation by Trube ®}which is Gesczthed tn the next section, Example 2,1 Estimate the solution gas-oll ratio and of! formation volume factor at 2 pressure of 2,625 psia and temperature of 200°F for a gas-crude system having a tank oil gravity of 30°API and a gas specific gravity of 0.8, ‘Use both the Lasater and the Standing correlations for estimating R, and the Standing correlation for B, Po Ye = _2,625)(0.6) = 3,282 T (460+ 200) From Fig. 2.2, = 0,59 From Fig. 2.1, M,, = 380 Yoo MLS __= 0.876 TS. $+ APT29a New Correlations for R,, B, and c, Vasquez A., M. E.: "Correlations for Fluid Physical Property Prediction, " ‘M. 5. ‘Thesis, Tulsa University ‘A. Correlation to predict gas gravity at 114.7 pala, Ys 44.7 ', », a + Peep son teed 1.0 srt tol: et) 10 where: They” temperature, ed APL= API gravity Pgep * Prosmurey pala B, Correlation for solution gas-oll ratio, R, (acl/STBO) 1) API «90° : 1.0887 pean” ager A 5 7.68 2) API>30° 1.187 “gay. QP ygl0.d2 A 36,06 where A=—APL_ F + 460 C. Correlation for oll formation volume factor below the bubble point pressure, B, (b81/STBO) 1) API ¢ 30° By #1.0+ 4,677 R, x10" + 0.1751 x20 D ~2.8108% 08RD2) API> 30° By +10+467R, x10 + 0.11210 D+ 0.1337 x10" R, D where De a) ‘g(L14. 7), D. Caleulstion of oll formation volume factor above the bubble point pressure, 2, @HI/STBO), using new correlation for oll compressbity, ©, fal”) oP B= Bye 1) Use Bi or Bt above to determine p, for R=, 2) Use C1 oF C2 above to determine B.., 3) Catealate c, from:2-10 (879, 3)(850) v, y, a lls 3 c = Love. gan. ] [Lose ] a. = 507.1201/s1B0, Check: From Fig, 2.3, R, = 500 sof/STBO sanding Bs ayeyge[ Be « 00025 Arm J 0.88 o's TS y+ 00091 (F) 1o-012900)_]3/0.88 = 0.90.0 2526) . 20° TO To" g(0- 00055200) =552,.9 scf/STBO Check: From Fig. 2.4, R, = 590 sof/STBO Standing B, : (Using 620 for R) = 756.5 B, = 0.972 + 0.000147 F175 = 0,972 + 0, onon4merse..9)?"27%B11 = 1,327 bbl/STBO Check: From Fig. 2.5, B= 1,996 bbl/STBO 2.4 Condensate Systeme For condensate eystems, R, and B, laboratory data are seldom avallable, However, liquid and gas mass fractions and densities are froquently avatlable from flash vaporization data, It is possible to construct , and B data as a function of available ash data, The user should be aware that ‘many times these values take on seemingly impossible values, It 18 possible for @ well to produce single phase gas at the sand- face and at the surface, yet go through a two-phase flow section in the well ue to retrograde condensation, When using a Mblack oil" model based on de- scribing mass transfer between phases with values of R, and B., a finite value of af must be assumed at stock tank conditions to simulate a ligutd phase in the tubing. Single phase gas can be simulated using B, valves of 0,0 and B, should merely be considered a8 a "device" to assure that the product q’ B, {is a measure of the in-situ vohumetrte low rate, Thus values of B, lees than 1,0 should not concern the user. Jn a similar way, retrograde condensation can result in a very strange behavior of the R, data, Here again, the user should realize that 4% Rp Tepresents the total produced gas, even though q may be a physically ‘pon existent variable, Under retrograde condensation, a decrease in Pressure causes condensation rather than vaporization, Since q/ R, represents (he gas in solution, we are faced with an increasing R, with decreasing pressure. ‘An alternate way to express this is thatthe free gas, q/ @, = R,) decreases ‘5 pressure declines in the retrograde regio. For retrograde systems, itis feasible for R, to take on negative valu To predict pressure losses when condensate fluids flow through Pipes, it is recommended that a compositional model be used rather than a "black oil" model. Given the total mass flow rate of a well and the composition (of the total feed, one can perform flach vaporization calculations at selected212 values of pressure and temperature, These calealations will yield values of 1. Vy py and o, ae functions of pressure and temperature, tn situ volu~ metric flow rates can then be determined from: my =LM, mV, ee = Tete/ mt my L oe /imytm)) °e 1. = motes liquid/mole feed V = moles vepor/mole feed rm, = mass of liquid/mole tata feed, gz mass of gas/mole total food, Supertictal velocities can thea be calculated direct, which in turn will yield values of liquid holdup. 2.5 Ol Compresssbaity Tete necessary to estimate value of oll compressibility (or coetficient of seothermal compressibty to predict oll formation volume factors of undersaturated crude oils, Normally it is suflicient to estimaie '& value at the bubble point pressure and an average flowing temperatare and esume that tie a constant. According to the Theory of Corresponding States, a reduced oettletent of tsothermal compressibility can be defied for pure liquids as22a 2.4 Supplemental Material COMPOSITIONAL MODEL A complex mixture of hydrocarbon compounds or components ean exist as a single- phase liquid, a single-phase gas, or as a two- phase mixture, depending on the pressure, temperature and composition of the mixture. Unlike a single component or compound, a multicomponent mixture will exhibit an envelope rather than a single line on a pressure-temperature diagram when two phases exist simultaneously. A typical phase diagram for a multicomponent system is given in Fig. 1. Fig.1 permits a qualitative classification of the types of reservoirs that are encountered in oil and gas systems. Typical oil Teservoirs have temperatures below the critical temperature of the hydrocarbon mixture. Volatile oil and condensate reservoirs normally have temperatures between the critical temperature and the cricondentherm for the hydrocarbon mixture. Dry gas reservoirs have temperatures above the cricondentherm. Many condensate fluids exhibit retrograde condensation, a Phenomena in which condensation occurs during pressure reduction. This abnormal or retrograde behavior occurs in the shaded region within the two-phase envelope of Fig. 1. When volatile oil and condensate fluids flow through pipes, their phase behavior is best predicted with a variable composition model. The procedure is known as vapor-liquid equilibrium (VLE) or “flash calculations. YLE__Calculations Given the following definitions: F = number of moles of feed or mixture L = number of moles of liquid Y = number of moles of vapor % = mole fraction of component i in feed Xi = mole fraction of component i in liquid Yi= mole fraction of component i in vapor2-12 FIGURE 1 TYPICAL PHASE DIAGRAM FOR MULTICOMPONENT HYDROCARBON SYSTEM ‘vise “sunssaud 8 3. alonuas3i ry g oy 100180 200850 30 * RESERVOIR TEMPERATURE,226 ‘A simple mass balance on the feed and on each component yields: F=L+V ZFexul+yV Equation 1 can be modified to give L v Brive ‘An equilibrium constant for component i can be defined as: yw x ‘Liquid _Phase K Solving Bq. 2 for xi gives ar-yv Ze -aKy F ye BEY | OTE F ine Simplifying gives wl-Heang= 2 alefig] y¥ x FOw=1)41) =z ‘Therefore,212d Yanor—Phase Solving Eq. 2 for yi gives vi Bk Bea APT Simplifying gives WV Ang = BF wetngEe% of s+ (t-#)] 2% yfuetonen] = Therefore, Pe ZK 14 K-21) Solution__Procedure Solving for yi and xi requires a trial and error procedure. Numerous solution algorithms have been proposed to accelerate convergence, especially near the critical point where convergence is difficult and slow. Following is a simple algorithm that is acceptable for most cases and demonstrates the procedure required to perform a VLE calculation.2ei26 ‘The sum of the mole fractions in the liquid and gas phases must = 1.0. Therefore, Yas Done 10 ‘Therefore, combining Eqs. 6 and 8 yiclds Equation 9 is often called the Rachford and Rice equation. A second ‘order Newton convergence scheme can be used to converge the Rachford and Rice equation to nearly zero (0.000001 is sufficient). The procedure requires that the following be known: ‘the number of components, n the mole fraction of each component in the mixture, Zi the equilibrium constant for cach component, Kj 4 first guess for the mole ratio, (V/F}j Given a first guess, such as 0.5 for (V/F)j, an improved value of (V/F) can be estimated from w @) ®,-@)- ie where the derivative, f(V/F) is obtained by differentiating Eq. 9 with respect to V/F. t(%) 2 7 ® Xeon) 10Convergence is obtained when (Bp) l<20810% ns Once convergence is obtained for V/F, the composition of each phase can be determined from Eqs. 6 and 8. ‘The above procedure requires values for Kj at the pressure and temperature of interest. Several methods exist for determining K values, including use of charts such as given in Fig. 2. The most accepted way to determine K values is from equations of state (EOSs) and will be covered in a later section, Bubbie_and_Dew Point Curves For temperatures less than the critical temperature, the bubble point curve separates single-phase liquid from the two-phase region. Thus a bubble point pressure is the maximum pressure for which gas will be present with the liquid. At a bubble point pressure, V=0, and Zo Therefore, Eq. 8 simplifies to aK = TOCK=1 or neAK ZK and Enedames. fet B 222el2g FIGURE 2 EQUILIBRIUM RATIOS FOR PROPANE Koy Kel PRESSURE. PSIA~—= PROPANE = convirnts So rata2-12h The dew point curve in Fig. 1 separates the single-phase gas and two-phase regions. At a given temperature, two dew point Pressures can exist. At a dew point pressure, L=0, V=F, and Z; = 9; ‘Therefore, Eq. 6 simplifies to «A 4" THIKHD ~ and 4 i ft Bubble and dew point pressures can be calculated with a trial and error procedure in which various pressures are assumed and a pressure is found that produces K values that satisfy Eqs. 13 or 14. The following suggestions will help in determining if a given Pressure is in a single-phase or the two-phase region. a eS 2m at > 1, one pe bes pad ial 1 ot se SaK ! wer/Pes 2 a = [148,01 TET 6 Ca, Cy and Sj in Eqs, 22-26 are dependent on the specific EOS used and are defined in the next section.212k ‘Thus, K values are functions of the compositions of each phase, and the compositions in turn are a function of the K values. In addition, the compressibility factor for each phase, Z and Zy in the fugacity coefficient equations, are also functions of the compositions of each phase as described in the next section. This leads to an additional ial and error procedure to determine K values from an equation of state. Equations 22 and 23 have been found to work well for relatively simple mixtures. When the number of components increases, the mixing rules in these cquations must be modified to account for the more complex molecular forces present. The resulting equations are 233 AM Tafa and weed Dv yas wnennnnnnnunnennnee B am where y= (1-85) Saf In Eq. 29, 31; are the binary interaction coefficients between two components. They are obtained by fitting predicted binary bubble point pressures to experimental data for a given EOS. The Parameters ajor aj are defined byEquations of State The two most accepted EOSs in use today are the Redlich- Kwong equation as modified by Soave (SRK) and the Peng-Robinson equation (PR). Both are two constant cubic EOSs that have been modified for mixtures such as volatile oils and gas condensate systems. Soave-Redlich-Kwong (SRK) Cubic equations for cach phase based on the SRK EOS are given in Eqs 30 and 31. Bh BAZ (AL~By-BE)— ALBL = 0 cemisoninnsnens 30 Zy~ Zh + 2Zy (dy ~By -BY) = AyBy = 0 Where the coefficients A and B for each phase are given by Eqs. 22- 23 or 27-28 and Eqs. 24-25. For the SRK EOS, C, = 0.42748 and Cy = 0.08664. In addition, 31 S\= 0.480 + 1.574 @,~ 0.176 oF 32 ‘The Pitzer acentric factors, w, are constants for each component and are available in the literature. For hypothetical or pseudo- components such as those used to characterize the heavy ends in 2 multicomponent hydrocarbon mixture, the acentric factors can be estimated from 2121212m Peng-Robinson (PR) Cubic equations for each phase based on the PR EOS are given in Eqs 34 and 35. i ~ UBL + (hy 382-282, — (ALB. —BE-BE)= 0 vane 34, By ~ (-By)Zh + (Ay 3B} - 2By)Zy ~ (AyBy -BF-B) = 0... 35 where the coefficients A and B for each phase are. given by Eqs. 22- 23 or 27-28 and Eqs. 24-25. For the PR EOS, Cy= 0.45724 and Cy 0.0780. In addition, S,= 0.37464 + 154226 ,- 026992 a? 36 Roots Equations 30-31 and 34-35 all have three roots, only one of which is significant. The largest roots of Eqs. 31 and 35 represent the correct compressibility factors for the vapor phase. The smallest Positive roots of Eqs. 30 and 34 represent the correct compressibility factors of the liquid phase. The same second order Newton convergence scheme that was described earlier can be used to find the desired root if a reasonable guess is provided to initiate the trial and error direct substitution method. A good initial guess for the liquid root is based on an assumed specific gravity of 0.8. Thus PAM = 1.8 Py! Equation 30 can be written as Fis) = Zhe Zit Ze (AL~Bp-BE)— AB,2128 ‘An improved value for Zt.4+1 can be obtained from ase Yay ~ See where Fed) = pee = az = 24 + BBE) Convergence is obtained when | Zyer- tel < 10 A similar approach can be used with Eq. 34 for the PR EOS. ‘The above approach is also used to find the appropriate roots for the vapor phase from Eqs. 31 and 35, However, an initial guess of Zy = 1.0 should be used. Densities Once the compressibilities of each phase is determined, the gas and liquid densities can be determined from PM Pee Fee and i PAM, oe er a It is generally felt that liquid densities calculated from Eq. 43 are not sufficiently accurate. A better approach would be to use the Correlation proposed by Hankinson and Thomson.Computer _Aigorithm A computer algorithm is presented in Fig, 3 that first performs a VLE calculation based on estimated K values, The resulting compositions for each phase are then used to determine the roots of the gas and liquid cubic equations based on either the SRK or PR EOSs. The appropriate roots are then used to calculate fugacity coefficients and improved K values. Convergence on all K values is required, otherwise a direct substitution of improved K values for estimated K values is performed and the process is repeated. The initial estimate of K values is obtained form Eq. 44.2-12p FIGURE 3 PROCEDURE FOR PERFORMING VLE CALCULATIONS 18248 14s asmmed that a similar equation is valid for hydrocarbon mixtures and ‘Trube” has correlated c, with P_ and T, with the result shown in Figure 2.6, Fig, 2.6, Varation of pseudoreduced compressibility with pscudoreduced pressure for values of pseudo- reduced temperature, (After Trube 51.) ‘The ol! compressibility is calculated from oP, ey" ete P244 ‘Tins, for « given pressure and temperature, one must estimate values of P, and T, to obtain ¢. and then ¢,, Values of c, and P, are not normally avaiable for quid However, the reduced temperature can be estimated from Figure 2,7 after making appropriate adjustments to the bubble point pressure and the specific ravity of the Liquid, 38 8 88888 | i ‘Fig. 2.7, Variation of pseudocritical temperature of reservoir oils with specific gravity and bubble-point of liquid correlated to 60°F (After Trube®218, ‘Either the Standing or Lasater correlations em be used to ad- Just the bubble point pressure to 60°F, A more detailed procedure is necessary to estimate the specific gravity of the flowing liquid at 60°F and ‘babble point pressure, ‘Normally in two-phase flow systems the API gravity of the stock tank oil, the producing gas gravity, and the producing gas-oll ratio ‘are mown, Also, at the bubble point pressure, the solution gas-oll ratio and producing gas-oil ratio are equal, The concept of apparent liquid den- sities of natural gases was proposed by Katz 17 and the resulting work appears. ‘fm Figure 2.8, i u u r Fig, 2.8, Apparent ligud densities of natural gases (After Katz7 ‘To obtain the specific gravity of the flowing guid at 60°F and ‘babble point pressure, the following procedure is used: 1, From Fig, 2,6 determine the apparent liquid deneity of R, at 60°F and 14.7 pela, 6, 5 -216 2, Caloulate the mass of R,/STBO, m,. mR, 5 1B,/8TBO ene eet a 3, Calculate the mass of gas-free crude/STBO, m,. 1m, = ¥, (5.616) (62.4), Ib, /STBO 4, Catonlate the mass of gus + ofl/STBO, mp ny=m,sm, 5, Determine the off volume V,, (t°/STBO) from 6, Calculate the off density from oer ‘ 7. Caloulate the specific gravity of the bubble point oll at 60°F trom Having corrected the bubble point pressure and apecific gravity of the quid to 60°F, the pseudocritical temperature, ‘T,, is obtained from Fig. 2.7, The peewhoeritical pressure, P., can be estimated from Figure 2.9 at the came abscissa value that yielde the value of T, above, This value of P, is considered more accurate than one obtained from Fig. 2.7 for a value of ¥,, calculated in step 7 above.247 ; Approximate variation of peeudocritieal pressure ‘and temperature of liquid corrected to 60°F, (After Trube 31) ‘Psoudoreduced pressure and temperature can be calculated ‘The paeudoreduced compressibility is then obtained from Fig. 2.6. Finally, ‘the oll compressibility can now be calculated from2.18 Example 2.2 A well is producing ofl with gas at a producing gas-oll ratio of 1580 scf/STBO. The specific gravity of the gus in 0.6 and the tank ofl gravity 18 S0°API, ‘The temperature at which the bubble point pressure cocurs ts ‘approximately 200°F, Estimate the oll compressibility at 200°F and the bubble potnt pressure, 1, Determine p, at T = 60°F, (Use Lasster correlation). From Fig. 2.3 at R, = §80 sef/STBO: > 2,350 pote 2, From Fig. 2.8, 9, = 26.9 Ib, /ft® Sms R, pp 40, =y, (0.0764) Mee, eee. = (880)(0. 80,0764) = 92.4 1b, /STBO 4, m= ¥,(6. 615)(62.4) = (0.876)(5, 628)(62, 4) = 808,9 1b, /sTBO 5 mem, +m, = 82,44 306.9 = 329.3 1b, /sTBO 6. vy= 5.615 +_My La = 5.615 + 92,4/26.3 =s.sar te 09 MT = 839,9/6, 847 = 49,85 fe = 49.55/62.4 = 0.79 7Bet9 9, From Fig, 2.7 at y,= 0.79 and p,= 2,150 pata, 6, T= 60 °R 10, From Fig. 2.9 at T,= 860°R, Po = 400 pela Wap Tt Po Te = 2150/4680 = 660/960 =4.48 = 0.767 12, From Fig, 2.6, ¢, = 0,01 13. gt cP = 0.04.49 =2,08x107 psi 7 21506 Solty of Natural Gas in Water ‘The solubility of hyérocarboo gue component i inversely proportional to thelr molecular weight, ‘Thus methane is more soluble than ethane, ethane than propane, ete, MoCett states thatthe solubility of each component fs two to three times greater than that of the next heavier paraffiato component, ‘The solubility of methane in water can be used to extinate the solubility of natural gas in pore water with an accuracy of about 5%, Figure 2,10 can be used to estimate this value, Fig. 2.10, Solubility of Methane in Pure Water (After Culberson and MeKetta 7 »a1 ‘To account for the effect of water salinity, the result obtained from Fig, 2,10 must be adjusted by the correction factor from Figure 2.11, AQ a Fig. 2.11. Effect of salinity on the amount of gas in solution when fully saturated with gas. (After Eichelberger 2°.) 2.1 Water Formation Volume Factor It is a common practice to neglect the compressibility of water, ‘the thermal expansion of water and the solubility of gas In water, Since these fare the three Items that contribute to expansion or shrinkage of ofl and water, ‘the not result ts to assume water formation volume factor, B, 181.0. A typical ‘graph of B,, ve p appears in Figure 2.12,222 Fig. 2.12, Typical graph, formation-volume factor of water vs pressure, (After McCain Note from Fig, 2.12 that B, always increases with decreasing pressure, Which means that, neglecting temperature changes, water will expand as Pressure drops, Thus the effect of gas evolving from solution only par- ‘tially offests water expansion with pressure reduction. ‘Normally tt {s assumed that the gas-oil and gas-water bubble point pressures are equal, Two methods are commonly used to estimate B., below the bubble potnt pressure, Method 2 ‘The formation volume factor for pure, dead (no gas in solution) water ata desired pressure and temperature can be estimated from Figure 2.13, Correction forthe effect of gas in solution inthe pure water in ob- {ted from Figure 2.14.2-23 “TTI 5 Be 5. z i a i 4 OOS Beessine Paik Fig, 2.18, Formation volume Fig. 2,14, Difference, 4B,» be- {actor of pure water (after ‘ween B,, of gas-saturated pure ‘Keenan and Keyes 20.) ‘water and of pure water (after Keenan and Keyes *°,) A final vale of B, a0 fnction of pressure and temperature, gus in soltion sod isnolved solids is calculated from BYTE, ty SB, where B= value from Fig, 2.13. y= value from Fig, 2.11 AB, = value from Fig, 2.14 ‘The following approximate relationship based on Figs. 2.12 and 2.14 and asguming y = 1.0 was given by Gould,” 6 4 6 2 6 5 By £10 41.210 Ty +1,02107° Te ~ 9,884 10° B where Ty = F-60 and T= °F.228 Method 2 eccatsnaggeta that te Kats atid for evtinaing 2, xn se ued timate By By TG +B Vg) O48 Vayg) whore A 8 the change in volume during presmure eduction to stmospherie prenmure and A Vy ts the change {9 volume due to reduction in temperature to. 60°F st atmospheric pressure, These values can be obtained from Figures 2,16 and 2.16. Fig. 2.15. A Vir a6 a function of temperature (after McCain *2)Fig. 2.16, 4 V,, a0 a fimction of temperature and pressure (after MeCain 28.) ‘MoCain states that Method 2 {a valid for oil-field waters with varying brine ‘concentrations, since effects of brine concentration on A Vi. AB AVyyn ATC offactting. ‘Water formation volume factors above the bubble point pressure are calculated the same way as for gas-oil systems, Thus, By 7 By, EP] -e, =P) ] +2-26 2.8 Water Compresefbility ‘The compresutbilty of water having natura! gas in solution can be estimated using Figure 2.27. The top chart yields o,, for pure water and ‘the bottom chart yields a multiplicative correction fsctor for dissolved gas. ‘The effect of dissolved solids 1s neglected, Fig. 2.17, Compnessibiity of ure water, sncluding effects fof gus in solution. (after Dodson and standing *),dar Example 2.8 ‘The well in example 2.1 (pg. 2-9) Is also producing water with 10% dissolved solids (100,000 ppm), Determine the gas in solution in the water at the bubble point pressure, 2626 psia, and 200°F, Also calculate 3 and ontimate o,, at the p, of 2628 pala and 200°, Py From Fig. 2.10: R,., (pure water) = 14 ecf/STEW From Fig. 2,11: Correction factor for effect of brine, y= 0.65 Ryy, 71 (0.65) = 9.1 scl/STBW ©) Matt ts Prom Pg, 2.38 BY gure water) = 1.080 8/Or5H rom Fig. 214 am, =.008 By =Bva = 1.029 + (0.68) (0.005) = 1,032 bbI/STBW Method 2; From Fig, 2.18: BV pq = 0-089 ri 2.18: = 0,006 ‘rom Fig, 2,1 Vp 7-4. B, = O+AV,) +a 0) = (0.996) @.039) = 1.035 bbI/STBW ° From Fig. 2.17: 0, (pure water) = 3.15 x10%psi7* Correction for R,,, = 1.08 ey = G18 x 10% a.08) =3.402 x 10° pai?2.9 Specific Gravity of Free and Dissolved Gas ‘As the pressure in a gas-oil system changes, the composition of ‘each phase also changes. When gas first evolves from a crude oil, the gas composition is primarily Methane, As pressure declines, higher and higher molecular weight hydrocarbons vaporize, tending to increase the gravity of both the free gas and the gravity of the gas in solution. Katz presented Figure 2,18 for predicting the dissolved ges gravity. Solubility - Cubic Fest Per Barrel of Residual Oil R, (sef/STBO) Fig, 2.18, Predtetion of gas gravity from ‘Solubility and erude-oil gravity (after Katz 12) ‘A practical limit is placed on the above value that ¥gq & Vg 0.56, where 0,56 is the specific gravity of Methane and y,, is the average specific ‘gravity of the total separator gas, all expressed at standard conditions, ‘The specific gravity of the free gas an be obtained from the following material2-29 batance calculation: wea" yd, (OT) og Mee BB) ETO Ye Mp Bp 060 yy weer "at fms, 4) (0.076) gy # B=) 4, (0.0760) Vg £8, 0.0705 ving by @ 0.0100, Ry Yea * BRD Yey “Rp et solving for ¥, 5 {A practical limit is aloo placod on ¥g, a8 follow By, 2-58 ‘ee Yet = Although most engineers tend to neglect the effect of changing 28 composition with pressure and temperature, the effect can be important. I tie procedting ealevlations are performed, the resulting valucs of Yee should be used to calculate gas physical properties such as density and viscosity.2-30 na similar way, values of ¥, ebould be used to caloulate the effect of ‘solution gas on the oll density covered in a Inter section. It is also common practice to neglect the effect of any solution gas in water on the composition of the tree gas, 2,10 Oil Density Oil specific gravity 1s defined as the ratio of the density of of! to the density of water, both taken at the same pressure and temperature, Yo Pe! Bar Sometimes specific gravity 1s given as 60°/60° which means both off and ‘water were measured at standard conditions of 60°F and atmospheric pressure, Its frequent practice to identify this with a subscript sc, The petroleum industry uses mother gravity term called API gravity which is defined by Yo 7241.5/091.5 + APD ee Prediction of ofl density at in-situ conditions is vitally tmportant {tn two-phase pressure loss predictions, Normally the gas composition and Maid composition are not Imown, but estimates are available for'y wy HR, one met for predicting», based on adjusting the gas free crude by the apparent ligu density of solution gas was described in section 2.8, This method, proposed by Katz "7, was stated by McCain” to be accurate to within 3% of experimental remus, ‘An alternate method is available for predicting», for sasirated comides which is similar to the Katz method, (62.4) + ¥, (0.0764) Ry oe Sat2-31 ‘Above the bubble point pressure, off density must be calculsted from 07 Pan PL, 0-H] ‘where pip 1s detersuined from one ofthe previous methods for R=, and B= Bay 2,11 Water Density Water density can be calculated using the same approach as for ‘ofl, except that the effects of gas in solution are normally neglected, Thus D Ny 24) ‘The density of the water phase at standard conditions as a function of total Aiesclved solkds om be obtained from Figure 2. Fig. 2.19, Density of brine as « function of total dissolved solids, (After McCain”),282 Example 2.4 Calculate the ofl and water densities at bubble point pressure and 200°P for the fluids of Examples 2.2 and 2,3, (pgs, 2-9, 2-27) at R, = 580, API= 30° From Fig. 2. Yea 0.79 (Bince Yq <0-8, use ; Xa) (62.4) + ¥_4(0.0764) R 247 Sat (0, 816)62.4) + 0, 0(0,0768)6630) - 3.614 BH = 40,95 + 4.32 45.27 b/s From Fig. 2.19: 9, = 67 1b, /a° ¢ For B, = 1.035, = 67/1, 085 = 64.18 th, /tt°2.12 Gas Density Natural gases are complex mixtures. of hydrocarbon compounds ‘vith various impurities euch as nitrogen, carbon dioxide and hydrogen sulfide, ‘Laws for ideal gases or mixtures of ideal gases are inadequate for predicting the behavior of natural gases. ‘The literature contains several hundred equations of state for gases depending upon the type of gas, the composition, ‘he conditions of pressure and temperature, and the degree of sophisticstion. ‘The procedures for calculating the density of natural gas are familiar to tiost engineers, Perhaps the most frequently used of the equations for predicting density is the "compressibility" or"'Engineering" equation of state, Engineering Equation of State Jn its most common form, the Engineering equation of state is expressed as pVeaaRT. ‘The gas constant, R, has a value of 10.72 if other variables have units of pei, f°, Ib, -moles, °R. ‘The mmber of moles ean be expressed a8 a 284, Cd Combining these two equations and solving for the density for m/V): ope = Piet w2.70 Pit. ty (10, 72)T aT ‘The free gas gravity ¥,,, 18 function of pressure and temperature since st ‘depends on the composition and amount of gas that has evolved from or gone {nto solution in the oil. ‘An alternate way to calculate gas density, and one that 1s con- ‘sistent with the procedure for handling ofl and water, is based on the gas formation volume factor, B,. B, can be derived as fotlows,204 ‘Assuming a constent number of moles, n: aha Y= (BU) const a (Gr Pog 2 T gL sso Voc Tec? ¥so = 0427 (620) p .0) =0,0263 27 P ‘The gas density at any pressure and temperature can be obtained from B, and where hg 7 Net 2064, ‘The gas compressibility factor accounts for all deviation of a real gas from ideal gas behavior, Thus z= _Viacuay . V cides!) ‘Aa exhaustive coverage of the z-factor is beyond the scope of this book, The ‘most generally accepted procedure for determining t is based on the well-known, ‘Theorem of Corresponding States which essentially states that real ges mixtures will have the same z-factor for the same values of pseudoreduced pressure, Pp» tnd poowdoradiced temporature, . whorePeeudocritical Pressure and Temperature For natural gases it is necessary to determine pseudocritical pressure and temperature before one can obtain a z-factor, If gas composition in known, the pseudocritical values can be estimated from a mole fraction weighted average of critical values of each component, This approach, known as Kays rule!®, is expressed as follows: Since the composition of the free gas phase in a gas-oll system is normally ‘not known, the previous approach can seldom be used, Fortunately, the measurement of gas gravity is much less tedious than measuring composi- tion and is usually obtained on natural gases. A correlation by Brown, ct at, > for estimating the pseudocritical properties of natural gases from gas gravity is shown in Figure 2.20,tee rtcl rer, ve Pent tenet, dyes Renin Fig, 2,20, Pseudocritical properties of natural gases (after Brown et al?) For computation purposes, equations for straight lines are frequently used for the miscellaneous gases lines of Pig, 2.20, The follow- {ng equations can be used, P= 708,78 - 57.5 ¥, pe st T1694 31d y pe ef ‘Compressibility factors for natural gases were correlated as functions of pseudoreduoed pressure and temperature by Brown? , His291 correlation covered P, values up to & but was found tobe inaccurate for Pip Values above § by Standing and Kats $9, Standing and Katz extended ‘Brown's correlation to pseudoreduced pressures of 15. Their correlation ‘appeare in Figure 2,21, Charts for determining 2=factors at low paoudoreduced pressures and for very low pseudoreduced pressures (near atmospheric presaure) sppear in Figures 2.22 and 2.23. Figures 2,21-2,28 are restricted tonstural gases containing small amounts (,e, less than 2) of carbon Aioxide and hydrogen sultide,2-38 Pstuve reouceo enessune Pe er Sonia eee aa i Rea UM Le i i ; ei Evaroaat esses © [HE 7 . ° a 2 “e PSCUDO REOUCED PRESSURE % Fig, 2.21. Compressibility factors for natural gases, (after Standing and Katz°°,)+ unos xeye) “oanssaad ojsoydsouye ( g waoxg 10172) “somssoad psonpazopnasd se0u odbeS pean 405 stojouy AaTIqesoLdwIoD “ge"Z “BA —_- Moy ye soeuB yeaNFEH soy BAOVORs AyTIGIESOIdWOD “7e"% “SA 190 rea (00 t00_ 0» 0 2-39240 Effect of Non-Hydrocarbon Impurities on 2 Carbon dioxide and hydrogen sulfide are the two main non- hydrocarbon tmpurities that cause z-factors to deviate from those predicted by Fig. 2.21. Nitrogen can be considered as a hydrocarbon component for purposes of predicting #-factors, Whea very precise values are destred for the density of gases with many impurities, it {s necessary to use equations of state developed specifically for those gases, ‘Two-phase pressure loss prediction methods do not justify more sophisticated equations of state than described here. Robinson et al.2° developed a method for estimating the effect of the acid gas constituents on z-factors for acid gas content (& H, $+% CO,) ‘upto 60%, ‘The method involves obtaining a correction ratio, and dividing the ‘ratio into the approximate 2-factor obtained from Figures 2,21-2.28, Thus 2 = _x (approximate) Correction Factor Correction factors for combined CO, and H, § mole percents appear in Figures 2,24-2.26, ‘This method ean be usod for P, from 0 0 7 and T,. from 1,1 t0 2.0, Wichert and Aziz ® suggested three possible methods for de- termining z-factors of sour gases, The first of these was a modifeation of the Pitzer ot al.” mothod end was found to be the most accurate, but was ‘cumbersome for computer use, ‘The most accurate method which is well ‘suited for computer application was found to be a modified Redlick and Kwong?5 equation of state, A computer listing of the modified R-K method was pub- shed by Wichert and Aziz. The final method, which is the simplest to use, uses the Standing-Katz z-factor chart but with peeudoreduced values based ‘on an adjustment to peeudocritical values, Ax adjustment factor ¢ was de- fined ao fomt tobe a tucton at ¥oq, 2 yg WA=%69 ty ge 'c0, *7H,80 WOL PERCENT METHANE PLUS NITROGEN 20 MO. PERCENT carRoW ow XIDE Pus HYOROGEN suLPHIDE | Fig. 2.24, Compressibility correction ratios for sour natural gases, (after Robinson etal. 28.)Fig. 2.25, Compresstbility corregiion ratios for sour natural gases, (after Robinson et al. 7.)Fig. 2.26. Compressibility corrgstion ratios for sour natural gases, (after Robingon ot al.26 ,)24 Pots = 120 a? a ‘The preudocritical temperature and pressure are adjusted as follows: Tea Teo f = +BQ-B ¢) Poo Pre Tye / Tye * BAB) €) where primed values are the corrected pseudocritical values and unprimed values are determined from Kay's rule or from Fig. 2.20, Caleulation Algorithms for z-factor ‘Soveral approaches for obtaining factors with a computer algorithm ere available in the literature, Perhaps the most acourate method 4s to read in a large array of values from Fig, 2.21 and employ some type of {interpolation technique. Computer storage limitations may limit this approach. Sarem?” presented a method which weed Legendre Polynomials ‘to fit the curves of Fig. 2.21. Other methods include the works of Hankinson et al,” and Satter and Campbell “ynich make use of the acentricity factor ‘originally proposed by Pitzer”, A simple unpublished Fortran subroutine ‘that hae beon found to be a fair representation of Fig. 2.21 is Listed in Appendix A as SUBROUTINE ZFACBB. A modification to this subroutine recommended by Standing Is listed in Appondlx A as SUBROUTINE ZFACST.2-45 ‘A recent algorithm for determining z-factors based on Fig. 2.21 was presented by Yarborough and Hall °5 , They developed a “hard-sphere” equation of state which resulted in a nonlinear equation that must be solved with 2 trial-and-error method sich as Newton-Raphson. A sample computer program was also published. Yarborough and Hall recommended use of the Wichert and Aziz method to correct pseudocritical temperamre and pressure for CO, and H,S. A subroutine of the Hall and ‘Yarborough method is listed in Appendix A as SUBROUTINE ZFACHY. Example 2,5 Caleulate the gas density of 2°0..80 specific gravity gas contain- tng 10% CO, and 10% Hy§ (mol %) at « pressure of 2000 psia and a tempera é. ‘ture of 150°R, 2, Determine z-factor: Method 1 From Fig. 2.20: P. Calculate P. 2 Pe . late Pp AAT, Pp = D/P, = 2000/863 = 3,02 . = 610/422 = 1,4 Typ T/T pg" 610/422 = 1.48 From Fig. 2.21: (approximate) = 0,745G. Takacs: Oll & Gas Jour., Dec. 20, 1976, pg. 64-66. Method Range Commente Gray-sime Tp 1: 05-8.0, Pp 0-28.0 20x 20 matrix with interpolation sarem ‘Tye 1-05-2.95, Py: 0114.9 Legendre polynomis Carifle-cillett Ty 12-8.0, Pps 0.0-15.0 polynomials up to degree 6 Papay Unimown ‘simple equation Hall-Yarborough Ty 1.29.0, P, ‘Starling-Carnahan State Ean, Brat Unknown simple equation Dranchuk et al, ‘Tai 1-05-8.0, Pp: 0.2-90.0 BWR state eqn, curve fit Dranchuk-Kassem TT, 1,0-8,0 for Py: 0,2-90,0 _Starling-Carnahan State Ean. 171.0 for Py <1.0246 From Fig, 2.24: Correction ratio = 0,942 Calculate 2 - factor: az Correction ratio = 0.745 = 0,191 942 fethod 2 (Wichert and Aziz *8 Determine A and B: A Ty,g Yoo, 7030 #0:10=0.20 Boy g "0-10 Myf Calculate e: enzo ah? ah§ + 15,005. 3h, #120 (0.29 0,288) +15 0.1% - 0,149 23.8 Calculate Pana, at o-e Tbe ‘Pe Poe” Poe Tyo (Tyg * BAB) «)at = s9y(998.2)/[ 422 + 0.1)0-0.1)08.8) ] = 022.8 Calculate P,, and 7, + Pos p= 2000/622.4= 3,21 pr iat Poe Tp," *610/990,2 91,89 Tye From Fig. 2.21: 20.788 2, Calealate B, : (Use z = 0,791) B, = 0.0263 2 T . > = _{0..0283) (0, 791) (610)_ = 0.00683 eo "@000) Tet 8, Caleulate p= Yel. 0768) e a = _@.8(0,0764) = 8,940, /tt? (0.00683) m248 2.19 Liquid Viscosity ‘The role of viscosity in pipe flow problems is not well understood, Tt can have an important effect on Reynolds number and thus determine if the flow mechaniom ie laminar, turbulent or transition, It can bave little meaning at all if a Suid exhibits non-Newtonian behavior. Its effect is highly un- Predictable tn gas-liquid, liquid-liquid, gas-solié, or Mquid-solid fiow problems, Emulstons may form of varying stability whose viscosity can neither be pre- dicted or even defined, Nevertheless, it is frequently important to be able to predict the viscosity of an individual phase in most fluid flow problems in Pipes, Viscosity is used in predicting friction factors and in determining guid boldup for gas-liquid flow, Fortunately, liquid viscosity measurement is dane routinely as Part of a PVT test, However, these results are either reported at 100°F and 210°F and atmospheric pressure, or at reservoir temperahire and a range of pressures For pipe flow problems, these measurements are essentially useless and em- irioal correlations are normally used to estimate viscosity. ‘Viscosity, lke other physical properties of iquide, is sensitive to changes in pressure and temperature, Increasing temperature always ‘causes a decrease in viscosity, Increasing pressire always increases viscosity ‘above the bubble point, However, below the bubble point, an increase in pressure ‘causes an increase in solution gas whch in turn decreases viscosity. Although the role of viscosity in pressure loss prediction is discussed in later chapter: ‘empirical correlations available to estimate values of dynamic viscosity are ‘given in the remainder of this section, G8 Viscosity. ‘The procedure for determining dynamic viscostty of of! at a fiven value of pressure and temperature is to determine yy, (T), the viscosity ‘atone atmosphere and the desired temperature, and then adjust 4o,,(T) for the effects of pressure and solution geo. ‘The most commonly used correlation for obtaining Hyp)(T) was Proposed by Beal , and is shown in Figure 2.27,gaset GbiE ae i } i t ; } i Cubes peer Aer to F ontanaan roe Fig, 2.27, Gas-free crude viscosity as a function of tempera ture and API gravity, (after Beal?) ‘The stock tank oil viscosity at the desired temperature can also be obtaizied by interpolating between two experimentally determined values wormally 100°F and 210°F) using the ASTM standard kinematic viscosity ve ‘temperature charts, An alternate interpolation scheme that can be used is: fon “Hop 00° [( a2]2-50 ‘The viscosity of gus free crude at the dosired temperature, tion) (€F) can be used to estimate in-situ live oll viscosity, 1, , with the Chew and Commally correlation 5 , Chew and Connally found that a plot of x, ¥8 Hyp (T) for a given value of R,, would yield a straight line on log-log papor, This shown in Figure 2.28, i 1 | | } ql i ! Fig. 2,28, Viscosity of gas-saturated crude oils at desired temperature and pressure. (after Chew and Connally °.)281 Chew and Connally simplified Fig, 2.28 by determining the slope of each R, curve and the corresponding 4, for a value of 4gp*1.0 6p, ‘These {nteroopt aod slope valuee were replotied v8 Ry and appear in Figure 2.29, EE Sie of Teel i Ze 1 aren wa oe nme am eme= See awaes Jet | ts. mrt SLE EET SOUT GAO RAT, 788 Pig. 2,29, A and factors for use in Chew and Comnally's viscosity correlation. (after Chew and Connally § .) After obtaining values of A and b from Pig, 2.29, live oil viscosity is ealeu- lated from: 154 og}? ‘Since Beal's data did not include R, values over 1600 scf/STBO, ‘the Chew and Connally correlation should be extrapolated beyond this level with caution, ‘When the live oll viscosity ts Inown at reservotr temperature and ‘bubble point pressure, improved predictions of live ofl viscosity variation with P and T can be obtained, ‘The procedure 1s to solve the Chew and Connally equation for Hoy(T). Values of op, obtained from Fig. 2.27 are plotted vs ‘T on ASTM paper and the curve is shifted in the y-direction until the curve ‘and the point for the calculated jiyp(T) coincide, The resulting curve 4s then2-82 used to determine 115,)(T) for use in the Chew and Connally equation, When oil is undersaturated, viscosity values obtained above must be corrected for the influence of pressure. ‘This can be done using Figure 2,30, ATE OF WeREASE OF VISCOSITY ABOVE ‘BueeLE Pow PressunE nvtire ocout o ead taut at ox av waace Pout chon Fig. 2.30, Hato of inereaso of viscosity above bubble-poinl pressure, (after Beal ? .)To assist in hand calculations, Beal also presented Figure 2,31 which was calculated directly from Figure 2.30, Fig, 2.51, Prediction of erude-oll viscosity above bubble-point pressure, (after Beal? ,) Water Viscosity Data on viscosity of ofl field waters are very scarce, Water viscosity will increase as pressure is increased, should increase with in- ‘creasing dissolved solids, and should decrease significantly with gas in solution, No data are available on the effects of gus in solution, owever,, the effet of temperature on the viscosity of water was reported by Van Wingen 2 and appears in Figure 2,52, 2083cai New Correlations for Oll Viscosity eggs, H. D. and Robinson, J. R.: “Estimating the Viscosity of Crude Oi! Systems", Jour. Pet. Tech., Sept. 1975, pg. 1140. Teeiyte edele tlalaaet belt Yee eto sop “28-4 ye 10% seas lace et where: T = °F sef/STBO B, Live oll Viscosity B By tation where: A = 10,715 (R, +200)" 545 Be 5.44 n, + 150) 998 I Vazquez A., M. E.: "Correlations for Fluid Physical Property Prediction", M.8. Thesis, Tulsa University. Viscosity of oll above the bubble point pressure, 1, (cP). lsh = [-0.039 p x10". 5.0] m= 26 ph x98 q where P,, is determined fi 24, forR = yl determined trom page Rye Ryy Ay determined from I above,Fig, 2.52, Viscosity of water at ofl-feld temperature and pressure. (after Van Wingen *,y ‘The curves in Fig. 2.92 suggest that effects of pressure and salinity are fhegligible, An expression which can be used to approximate an average curve for Fig: 2.92 te: ey © Exp (1.003 = 2,479 x10"? 7 + 1,982 x 207° 7) where T is in °F. 284When total dissolved solids are large, Frick’? provides a ‘correction ratio which can be used to estimate brine viscosity, ‘This ratio is shown in Fig. 2.33, Fig, 2.33, Ratio of brine viscosity ‘to pure-water viscosity vs salinity. {after Frick.) ‘The brine viscosity is obtained by multiplying the viscosity from Fig. 2.32 by the correction ratio of Fig, 2.33. For computer use, the following equation ccan be used which was developed for an average temperature of approximately 2, 20°F, (8:5) 209 +1.0 Ratios io Emulsion Viscosity ‘There fs little useful information in the Kterature on viscosities ‘of erude-oil emulsions. Woelflin™ presented data which can be used to estimate the viscosity of a brine-in-oll type of emulsfon from a known clean-ofl viscosity. ‘Most emulsions encountered in the production of ofl and water are of the brine-in-oil type where brine globules ive dispersed in oil. Emulsions ‘can be characterized as tight, medinm, or loose depending primarily upon the size of the brine globules, ‘Tight emulsions are common when a well Producing oi! and water flows with a high gas-oil ratio and is subjected to extreme conditions of agitation, A loose emulsion would commonly be found 4a a pumping well where agitation ts minimized, When an emulsion forms, the2-56 viscosity of the emulsion can be many times greater than the viscosity of elther the dispersed or dleperaing phases, Figure 2.94 shows how the ratio of emulsion viscosity to oll viscosity varies with increasing brine %. ire er pee, Fig, 2.34, Variation in viscosity of emulsions of crude oll and brine with brine content, (after Woelttin ,) Above a % brine of 60-70%, the emulsion inverts and brine becomes the con- ‘timous phase. The viscosity of the emulston then ca be approximated by the bbrine viscosity. However, near the inversion point, extremely high viscosities can be attained, ‘The liquid viscosity equation provided in section 1.5 cannot be ‘used whon emulsions occur. However, Fig, 2.24 can be used to approximate the viscosity of the Hquid-lqsid mixture,281 Example 2.6 Determine the oil and water viscosities for Examples 2.1, 2.3 and 2.4 at the p, of 2626 peta and 200°F, (see pye 2-9, 2-27, 2-45). From Fig. 2.29: A= 0,480 b = 0,675 Calculate 1,2 2, Determine Water Viscosity: From Fig. 2.92; iy, 90.92 ep From Fig. 2.53: Correction Ratio = 1,2 Caloulate Bay = 0.82)(1.2) = 0,39 op2-88 2.14 Gas Viscosity Laboratory data describing the viscosity of hytirocarbon gas mix- ‘ures at elevated pressures and temperatures are seldom available. Fortunately, the effect of gas viscosity on pressure drop and ob liquid boldup 1s normally considered negligible, The flow of compressible fluids ts slmost always ‘turbulent, and if the single phase flow anslogy were considered, friction losses ‘would be independent of gas viscosity. (One notable difference will be discussed Aster with the Baton et al° correlation). ‘When gas composition is known, gas mixture viscosities at low pressures (e.g, 1 atm.) can be calculated to within 2 or 3% accuracy using the Herning and Zipperer equation, a Tari where i, = viscosity of component 4 at 'T ‘This equation provides an excellent simple method to predict gas viscosities ‘at apy temperature and at Jow pressures, ‘The viscosity of gases under pressure can be estinated using ‘either the Catr et al.4 correlation or the more recent and simpler to use correlation by Lee et al. 22, Carr et al, Correlation The Carr otal correlation is based on first determining pat atmospheric pressure, 11, and the desired temperature, and then applying359 '# correction ratio to obtain, at the desired pressure, ‘The correlation for #1, {valid for temperature inthe range 49 <7 <400°R, Corrections can be made for the presence on Np, CO, or H,Sin the gas. Also note that the speci- ‘le gravity ofthe free gas, Ypr should be ueod. A plot of thetr correlation appears in Figure 2.35, Fig. 2.95. Viscosity of gases at 1 atmosphere pressure. (after Carr et al. *.) Using te viscosity at 1 atmostphere, 4, from either the preceding figure or from the Herning and Zipperer Equation, the remainder of the Carr et al correlation can be used, The viscosity ratio w/s, can be obtained from Figure 2.86 using paeudoreduced values for the free gas as already deter= mined for the gas compressibility factor,2-60 Fig. 2.36, Correlation of viscosity ratio with peoudoreduced pressure and temperature, (after Carr et al. *.) ‘The in-sita gas viscosity {s then obtained from 4" os hy The precoding Algure has a very restrictive limitation of 1,0< Pp <20,0, It has been found that there are numerous applications involving multiphase flow or single phase gas flow where conditions are outside the range of validity of the Carr et al Correlation, For this reason, the Lee et al. ‘correlation is considered to be more versatile, Lee et al, Correlation ‘The Lee et al. correlation is merely a series of equations as follows: =k. 104 % He. 104 ex OK oly261 where K=_@.4+ 0.02 Th 269+ 19M + T X=3,5+ 986 +0,01M T y= 24 = 02K In the above equations, T= °R, yg = ¢Ps M = molecular weight and 9 = gm/em®, ‘Tins, to use this correlation, the free gas ETavty, Yq Which Is algo the gus ensity expressed in gm/em at standard conditions, must be converted to an In-situ gravity or density. ‘This can be done with the following equation. Fg 7 Yq OOOH) G2 pe aan aT ar = 0,0438 YB ot B [A imple subroutine that can be used to calculate gas viscosity with the Lee et al correlation is Listed in Appendix A as SUBROUTINE GASVIS.262 Example 2.7 Determine gas viscosity for the gas in Example 2.5 using both the Carr et al, correlation and the Loe et al, correlation, (sce pg. 2-45). ‘Gare of al Correlation From Fig. 2.3 1, 70-0111 Correction for 10% H,S = + 0.0003 Correction for 10% CO, = + 0, 0006 4, = 0.011 + 0,0008 + 0, 0006 = 0,0120 op ‘From Fig, 2.36: P82 Using PL, = 8.2163 Eeo.ot al Correlation MH Ye Oye) = 0.8) (29) = 28.2 K=_.4+ 0,029 thd 209,0+19,0M2 7 = [9.4 + (0.0203. 2) I (610) 209.0% (19.0) (8.2) + 610 = 117.98, X=3.5+ 986 +0.01M T =9.5+ 986 + (0.00 (23.2) 0 = 5.95 Y= 24-02% +24 ~ (0.2)66.85) = 133 Pt Nop (0764) (52) __p, ain Zr a - (0,0764)(520)(2000) = 0,1496 gm/em? 4, N10. 7 1} ER. 4 werk. 10 exp ox 3 = 017.9007 oy [.25,0.2480)5%) = 0,0177 op2064 2.18 Surface Tension ‘The relative tmportance of surface tension on pressure loss when ‘tworphase flow occurs in a pipe is normally small. Values of gas-liquid surface tension are used to determine flow regimes and liquid holdup, and even here, the effect is usually minor. Nevertheless it is frequently necessary to predict values of gas-oil surface tension, ¢, and ges water surface tension, c,.. A ‘gas-liquid surface tension 1 then calculated as described in Section 1,5, Gas-on ‘The surface tension between a natural gas and hydrocarbon Lquids normally varies from approximately 85 dynes/em at low pressures and API gravities to 0 dynes/cm at the critical pressure when complete aiscfllity cosure, An estima of for stock ank ol 0.6, at emowpheric pressure) ca be obtained from the empaztca work of Baker ant Sera, A plot of thetr correlation appears in Figure 2.37, a 4) aN “SURFACE TENSION OF CRUDE OILS. ats et AT ATMOSPHERIC PRESSURE x 2 cr 28 6 ra 22 » op (ovnes ren ceNTIMETER) o 10 20~«8DksiSSCtSSCtCSSC«iSC«ta ‘API GRAVITY AT 60°F Pls, 2.37. Surface snsion of crude ols af atmnoepherie pressure, (after Baker and265 ‘Since the precise effect of temperature on oy, 1s not knows, ‘extrapolation beyond the temperature range of 100°F >T > 68°F is not re~ ‘commended. ‘The effect of pressure (or solution gas) is to decrease the surface tension, ‘The surface tension of crude ofl containing dissolved ‘gas expressed as a % of ¢,, can be obtained from Figure 2.38. The ‘graph can be extrapolated to higher pressures, provided the % does not be~ come negative, 106 TTT —— al ; id) Lag. SURFACE TENSION OF CRUDE OILS [7 EFFECT OF DISSOLVED GAS IN SOLUTION 70]---i SN L-|- AT VARIOUS SATURATION PRESSURES & . i ATMOSPHERIC TEMPERATURE “OF GAS FREE GIL AT ATMOSPHERIC PRESS. SURFACE TENSION OF CRUDE OIL CONTAMING O1SeQivED CAS AG PERCENT OF SUNEACE TENSION SATURATION PRESSURE, PSIA Fig, 2.38, Effect of golution gas on surface tension of crude oflv, (after Baker and swerdott.) ‘The lve oll nurfae tension is then obtatned trem on (785) ‘An equally valid estimate of gas-oil surface tension can be obtained from the work of Katz ot al, 1° as shown in Figure 2.39,2-658 wheres = surface tension, dynes/em P = parachor for given component X= mole fraction in liquid phase y= mole fraction In gas phase b= density, gm/ec M = molecular weight2-66 Sete tin tt Lom ao di ° eae een te Pig, 2.89, Bifect of saturation pressure on Katz et al. also state that "in general, one can be fairly sare that a hydrocarbon liquid, condensate, or crude ofl, saturated with natural gas in the reservoir at pressures of 3,000 psia or more, will have a surface tension of 2 dynes/em or less," Gas-Water ‘The murface tension of water-gus systems has been investigated bs Hocott!® and by Hough et al.7®, Thetr works have been combined into one dis scram by Katz et al, and is shown in Figure 2.40.261 we rater, ni Fig. 2.40, Eifect of pressure and temperature on surface tension of water, {etter Kats ot al,2®) Although these results suggest that all experiment have been valid, their use to predict ¢,, as functions of auch things as salinity of the water and gas composition ie highly questionable, Nevertheless, one can linearly inter polate between the curves of 14°F and 260°F from the Hough otal, work and obtain acceptable estimates of o,,.‘References: 1. Baker, 0. and Swerdloff, W.: "Finding Surface Tension of Hydrocarbon Liquids," Of and Gas Jour. (January 2, 1956)125. 2, Beal, C.: "The Viscosity of Air, Water, Natural Gas, Crude Ofl and Its Associated Gases at Oil Field Temperatures and Pressures," Trans. ADME (1946), 94, 3, Brown, G. G., et al: Natural Gasoline and the Volatile Hydrocarbons, N.G.ALA. (948). Carr, N, Ly, et al: "Viscosity of Hydrocarbon Gases under Pressore,"* ‘Traps, AIME (1954), 264. 5, Chew, J, and Connally, C. A. Jr.: "A Viscosity Correlation for Gas- Saturated Crude Olls," Trans. AIME (1958), 25. 6. Chlericl, G. L., et al: "Iwo-Phase Vertical Flow in Oil Wells-Prediction of Pressure Drop,” Trans, AIME (1974), 927, 7. Culbertson, 0. L, and McKetta, J. J.: "Solubility of Methane in Water at Pressures to 10,000 Pata," Trans, AIME (1951), 222. 8, Dodson, C. R, snd Standing, M, B.: "Pressure-Volume-Temperature ‘and Solubility Relations for Natural Gas Water Mixtures", Drill, and Prod, Prac., API 1944), 173. 9, Eaton, B. A., et al: "The Prediction of Flow Patterns, Liguid Holdup and ‘Pressure Losees Occurring During Contimous Two-Phase Flow in Hori~ zontal Pipelines," Trans. AIME (1967), 615. 10, Eichelberger, W. C, nd, and Engr. Chem., Vol. 47 (1955), 2223, 11, Frick, 7. C,: Petroleum Production Handbook, Millet the Printer, Ine. Dallas, Texas, Vol. 1 0962), Gould, T, L,: "Vertical Two-Phase Steam-Water Flow in Geothermal Wells," Trans, AIME (1974), 633. 13, Hankinson, R. W., ot al: "Predict Natural Gas Properties," Hydrocarbon Proc. (April 1969) 108. 14, Hering, F, and Zipperer, L.: "Calculation of the Viscosity of Technical Gas Mixtures from the Viscosity of the Individual Gases," Gas und ‘Wasserfoch (1986), 49.15. Hooott, C, R.:MInterfactal Tension Between Water and Oil Under ‘Reservoir Conditions," Trans, AIME (1989), 184. 16, Hough, E, W., et al:"Interfacial Tensions at Reservoir Pressures and ‘Temperatures; Apparatus and the Water-Methane System," Trans. AIME 951), 57. 17. Katz, D. L.+ "Prediction of the Shrinkage of Crude Oils," Drill. and Prod. Prac., API (1942), 187. 18, Katzy D, Lay ot alas Handbook of Natural Gas Engineering, McGraw il ‘Book Co,, ine, New York (0958). 49, Kay, W. B.: "Density of Hydrocarbon Gas and Vapors," Ind, and Eng. ‘Chem,, Vol, 28 (1936), 1014. 20, Keenan, J. H, and Keyes, F. G,: Thermodynamic Properties of Steam, John Wiley & Sons, Tne., New York (1936). 24, Lasater, J. A.: "Bubble Point Pressure Correlation, " Trans, AIME 958), 979, 22, Lee, A, L., et al: "The Viscosity of Natural Gases," Trans, AIME (1969), 997. 28, McCain, W, D, Jr. ; The Properties of Petroleum Fluids, Petroleum Publishing Co., Tulsa, Oklahoma (1973). 24, Pitzer, K. S, ot al: "The Volumetric and Thermodynamic Propertios of ‘Fluids, Part I, Compressibility Factor, Vapor Pressure and Entropy fof Vaporization," Jour. of A.C.5, (uly 1955), 2492, 25, Redlich, ©, and Kwong, J, N. §.: "The Thermodynamics of Solutions, V, An Equation of State, Fugacities of Gaseous Solutions," Chem, Rev. 1949), 233, 26. Robinson, R. L, Jr. and Jacoby, R. H.: "Better Compressibility Factors," Hydrocarbon Proc. (April 1965), 141. 27, Sarem, A. M,;'"Z-factor Equation Developéd for Use in Digital Com puters,"" Oil and Gas Jour, (September 16, 1961), 118, 28, Saiter, A., and Campbell, J. M,"Non-Ideal Behavior of Gases and ‘Their Mixtures," Trans. AIME (1969) Pt. 2, 393.2-70 29, Standing, M, B.:"A General Pressure-Volume-Temperatire Correlation = = for Mixtures of California Oils and Greases," Drill, and Prod, Prac. , API (947), 275, 30, Standing, M, B. and Katz, D. L,: "Density of Natural Gases," Trans. AIME 942), 140, SL, Trube, A. S.: "Compressibility of Undersaturated Hydrocarbon Reservoir Fluids," Trans, AIME (1957), 341, ‘82, Van Wingen: Secondary Recovery of Oil in the United States, APT (1950), 127. 83, Wichert, E. and Aziz, K.: "Calculate Z's for Sour Gases, " Hydrocarbon Proc. (May, 1972), 119. 34, Woelflin, W.: "The Viscosity of Crude-Oll Emulsions," Drill, and Prod, Prac., API 1942), 148, 'How to Solve Equation of State for Z-tactors," Off and Gas Jour, (February 18, 1974), 86,an Problems 2.1 Given the following information, determine the fluid propertion Indloatd at p= 765 psis and T = 197 °F. je Obat pe 147 pala T+ 609F; APE at", Rye Eve eterno 8) Solution gas-oll ratio uaing the Lasater, Standing, and Vazquez-Beggs correlations ») Oil formation volume factor using the Standing and Vazquer-Beges correlations 2.2 For the results of problem 2,1 for the Standing correlations, ‘calculate the ofl density. Also calculate the gas density, 2.3 For the results of problems 2.1 and 2.2 calculate the following Saturated ofl viscosity using the Beal-Chew and Connally and the Beggs- Robinson correlations ) Gas viscosity using the Carr et al and the Lee et al correlations ) Saturated oi! surface tension using the Baker & Swerdloff correlation, 2.4 Given the following information, determine the required fluid properties at 1700 peta and 160 °F using the designated empirical correlations. vgr O18 @ pa Ut pele, 7 =60°R API = 33 Ry = 1000 sel/STDO Fluid Property Corrélationts) Solution Gas-Oil Ratio, R, (ect/STBO) Lasater, Standing, Vazquez" Gil Formation Volume Factor, B,(bbI/STBO) Standing, Vazquez” Specie Gravity of Dissolved Gas, Yq Katz Speelfle Gravity of Free Gas, ¥g¢ Equation Ol Density, 9, Ib,,/1°) Equation Pseudocritical Temperature, T, (°R) Brown Peeudo Critical Pressure, Po(psia) Brown212 Eluid Property Corretatton(s) Gas compressibility actor, 2 Standing & Katz Gas Density, 4 (by) Equation Oi Viscosity, 1, (0) ‘Beal-Chew & Connally, Beggs & Robinson” Gas Viscosity, 1g (ce) Lee et al. Surface Tension g, (éynes/em) Baker & swerdlott ‘Use results from this correlation when determining a variable for which this variable is required. 2.5 Solve problem 2,4 by writing a computer program that calle the various subroutines in the Appendix. 2.6 Gae, oll and water are flowing through a pipe, The following fluid properties, flow rates and other data are known. Ny = 1-000 APT gravity © 30°API T= 200% = 1000 peta € = 0,0006 ft, a, = 2,441 inches Pp 44 = 500 STBO/D ay, = 100 STBW/D B= £00 set/STBO Yep TOP ‘Write a computer program that calculates the following at the given T and p using subroutines given: in the Appendix.2.9 2.10 aan 2-13 =u Lth a) ty Yor My OAL Hp) (7 Hf * He A 2 ss correlations for Rand 1, by Ry Bey Yyg oye, (Ute Vanquer and Boggs corral 7 1 Bes gr Pr Py D Npen’ t 8 oo 1) %sg> Yon, I the quality ofa saturated stesm-water mixture 16 0.01, calculate the following 311 atm, (gee Appendix for steam tables) 1 Novslip mixture density 2 No-slip liquid holdup @ Mixture specific enthalpy tthe mixture apecitic enthalpy of « saturated steam-waler system at 100 psia 16 250 BTU/Ib,, celolate the steam quality. ‘Water at 100 °F and 2500 pala undergoes an isenthalpie pressure reduction to 2000 pata. Calculate the final temperature, Superheated steam at 600 °F and 200 psia undergoes an isenthalpic pressure reduction to 100 psia, Caloulate the final temperature, Dry steam Is injected into a well with 3 in, dlameter tubing ata rate of 100,000, 1b,,/D- Wellhead (surface) conditions are 600 °F and 100 psia, Useful relation ships are: babys oh, (x) ve Lp-= specific volume at) i 00 + vg %214 2.12 Values of specific volume in ft"/Ib,, and enthalpy in BTU/Ib,, are given 1m the Appendix for superheated vapor (all steam) and for saturation conditions (two-phase); ‘a, ‘The heat loss in the well for this rate ie 70 BTU/b,,. If the bottomhole pressure Is 120 pele, estimate the bottomhole tomperatare, 'b, If the heat loss were 400 BTU/Ib,, and the bottomhole pressure ‘were 200 pela, determine the botiombole temperature, quality, ‘po-slip liquid holdup, and superficial gas and liquid velocities. ‘A Ywet gas Is flowing ina pipeline with a diameter of 40,0 in, at a mass ffow rate of 10°1b,,/D, A ash caloulation is performed at 1000 pela and 118 °F regulting in the following: (see pg. 2-12) ‘L= 0,02 moles liquid/mole feed V = 0,98 moles vapor/mote feod = 75.0 1b, = mole My = 23,016 /Ib_ ~ mole = 28.0 1B, Mb, ‘From the vapor and liquid compositions, gas and liquid denaitites were ccaleulated to be 45.010, /t 3 9g = 4.01, /t8 ‘4, Caleulate the gas mass fraction (quality), , Caloulate the no~slip Liquid holdup, ©. Caleulate the superficial gas and liquid velocities,CHAPTER 3 VERTICAL FLOW32 3. VERTICAL FLOW 3.1 Introduction ‘The general pressure gradient equation was derived previously and given fo, ep ov, ey, Tet ky ain $2 = Ho, sin 8 + where st Re and the definitton for 9 and the dentty corm used in the acceleration component vary with different investigators, For vertical flow, 0 = 90°, sin 8 = 1, db = az, and the equation can be written as 8-8), + The pressure drop caused by elevation change depends on the density of the two-phase mixture and is urvally calculated using « Liquid holdup value. Except for conditions of high velocity, most of the pressure drop in vertical flow is caused by this component. The pressure drop caused by friction losses requires evaluation of a evo-phase friction factor. The pressure drop caused by accelerating the fluids 1s sonetines consider negligible and Le usually calculated only for cases of high flow veloctries. Many correlations have been developed for predicting two-phase flowing pressure gradients which differ in the manner used to calculate these chree components of the total pressure gredient. Some tnvestigetors chose to assume that che gas and Liquid pheser cravel at the same velocity (no slip- page between phases) for evaluating the mixture density and evaluate only ¢ friecion factor enpirically, Others developed mathods for calculating both Liquid holdup and friction factor and some chose to divide the flov conditions Anto patterns or regines and develop separate correletions for each flow regime. The correlations discussed tn this section are firet closettied according to thetr complexity and the methods use ach of che At components are givens to evaluate three pressure gr:33 3.2 Classification of Correlations ‘The vertical flow correlations discussed in this section cen be placed tn one of tally three categories. These categories i ‘4, No olip, no flow regise consideration. In correlations whieh fie Ante this category, the mixture density 48 calculated based on the input gas-liquid ratio, Thet is, the ges and Liquid ere assumed ro be traveling at the sene velocity in the pipe, The oaly cor= relation required ts for two-phase friction factor. No dtstinetion is made for different flov regines. b, Slip considered, no flow regine consideration. Methods tn this ‘category require correlations for both liquid holdup and friction factor, Since it 1s considered that the Liquid end g: travel at different velocities, # method must be provided for predicting the portion of the pipe occupied ty Liquid at any location, The sue correlations for Liquid holdup and friction fector # for all flow regines. c+ Slip considered, flow regine « required to predict Liquid holdup and friction factor, but setho: co predict vhich defined flow regine exists ore nece: the correct flow regine 1s established, the appropriate holdup and feiecion factor correlations ere determined, which are usually different for each flov regine. The method for calculating the accoleretton pressure gredieat also depends on flo regine. ‘The vertical flow correlations discussed and the category in vhich they belong ere tsted below. ed sidered. Not only are correlations ry. Once Varstcel Flow Correlation Sategory Foertmenn and Carpenser't . Baxendell and Thomas” . Fancher snd Brova® . Wagedora and Brows » Duns and Ros? e Orktszevses'? Aste, Govier and Foy e Chiertet, Ciueet,and Selocens * e Bagge and Beii13 ea4 3.3 Flow Regines Flow patterns or regimes frequently encountered in vertical two-phase flov are shown in Fig. 3.1 Most tnvestigators who consider flow regines define four regines which nay occur in a vertical pipe, Although different anes are given these regines, most of the methods discussed in this sec- tion use essentially the sane descriptions for these four flov regines. ‘The names and descriptions given by Orkiszewskt vill be used in this dts- cussion. A brief description of the manner in which che fluids are di tributed in the pipe for each flow regine {8 given. RUBBLE FLOW ‘The pipe ts slaost completely filled vith Liqutd and the free gas phase 4¢ present in snall bubbles, The bubbies move at different velocities and except for their density, have Little effect on the pressure gradient. ‘The wall of the pipe is alvays contacted by the Liquid phase. SLUG FLOW ‘The gas phase {s nore pronounced. Although the Liquid phase 1s sti11 continuous, the gas bubbles coalesce and form plugs or slugs vhich aluoet £411 the pipe cross section. Tha gat bubble velocity f+ greater then that of the Liquid. The Liquid in the film around the bubble may nove dovovard ‘at low velocities. Both the gas and Liquid have atgntficant effects on the Pressure gradient. TRANSITION FLW ‘The change from a continuous 1iguid phase to a continuous gas phase cccurs. The gas bubbles my Join and Liquid aay be entrained tn the bubbles. Alehough the Liquid effects are significant, che gas predoainant. Mist Lae The gas phase is continuous and the bulk of the liquid is entrained 448 droplets tn the gas phase, The pipe vall ta coated with @ Ligutd fila, but the gas phase predominantly controls the pressure gradient. Boundaries for the various flow regines aay be defined differently for ateferent investigator shown in Tig. 3.2. +A typical vertical tworphase flow regime map is35 hy Fig.3.2- Flow Regime Map6 3.4 Category g Correlations The three correlations considered in this category axe based on the sane approach and differ only in the correlation presented for friction factor. The baste equation for calculating « pressure gradient at given conditions of pressure and teuperature i Bee ae, for any constatent set of units. Poettmann and Carpenter preferred to base the equation on « total mass flow rate. The equation and units given belov are identical to their original equation except for the constant in the denominator, vhich has been modified ro a Dazcy-Weisbach or Moody equation rather then « Fanning equation for friction less. edie as ed ], cess Sh ae" Ta fs, 2.9652 x10" 5, © | M2» pressure gradient, pot/te. = no-elip density, 1ba/ ft? = LD, of tubing, fe % w= total mass flow race, Ibe/day a f = tworphase frfetion factor, dimensionless Jn each method the friction factor was correlated empirically with the mmerator of the Reynolds number. The friction factor correlations for ‘the methods sf Foettsann and Carpenter, Baxendell and Thomas and Fancher 0d Brown are shown tn Figures,3,3, 3.4 and 3.5. Since the munerator of the Reynolds oumber {s not dimensionless, unite must be apectfied for the absctusa in the graphs. For the graphs presented in Figures 3.3, 3.4 and 3.5, the units for che abscissa are Iba/ft-sec, Tn the Fancher and Bromn correlation the three curves for gas-liquid ratio ean be considered to represent gas-liquid ratios for 1500, 2250 and 3000 for interpolation purpesta wv Fig.3.3-Poettmann Carpenter Friction Factor Correlation th /teraeeMy 3.4-Baxendell & Thomas riction Factor Correlationcher & Brown, p ne! . Friction Factor Correlation 10 Fig.3.5~Far3.10 Exenple Problea by Foettmenn ~ Carpenter Method Give Vag "409 fe/aee 01,7 56.6 tbo/en fe vy 2665 fe/see 9, + 2.0 Ibe/eu ft a= 269 te wp 7 7687 Th Pa 720 peta Calculate the flowing pressure gradient at chese conditions. 1. Determine no-slip density «2S 5 + 4.05” G27” 0399 + ody = (56.6) (.399) + (2.84)(.607) = 23.97 tba/eu Fe fy 2. Determine friction factor vd = (23.97) (6.74) (.249) = 40.23 Prom Fig. 3.3, £9 .021 Determine total pressure gradient 2 2 fag ¢ fap g7 + Catan» [07.879 084003] "'2,9652x 10! p a 2.9652% 1043.97 (.269)> se 2» 23,97 + 1.03 + 25.40 BE» 0.176 Ps3.5. Category b Correlation ‘The only correlation discussed in this category 14 that of Hagedorn ond Brova, This correlation vas developed from date obtained fron # 1500-f2 experinentel vell, ‘The Liquid holdup wes not nessured but was calculated £0 Satisfy the measured pressure gradient after the pressure gradience due to friction ond acceleration vere accounted for. The Liquid holdup ts not @ true measure of the portion of the pipe occupied by iquid bur ts merely « correlating persneters Two-Phase Densiey. Jn order to calculate the pressure gradient component due to elevation change » value of Liquid holdup must be determined. Calculation of holdup Anvelves three correlations, tvo of vhich are correction factors for liquid Viscosity. Several dimensionless correlating perazeters vhich must be used are given below. These equations include constants Which make them dinension- ‘ally consistent when common “oil f1eld" units are used. 4 Ry they, fe Liquid Velocity Number 4 ty "18%,5 Pe, Gas Velocity Munber iy 08728 fii Te, Pipe Diseter Muaber M, = 0.1s726 6, five, Liquid Viscostey number where: gy 7 ft/sec Mgg * fate py = tay te? = dyses/em ase antewhen the Liquid stream includes both of! and vater, the Liquid properties are weighted in the following manner, PL mf BR, 17% fot % > HL Hof ty fy > shore oe eo rey th Thete dimensionless ousbers for any sat of conatatent unite vere defined by Ros as: wa * nS ‘The correlation for Liquid holdup divided by a secondary correction factor 4s shown tn Fig. 3.6 Gt, = "L 4). the correlating finetion requires @ 7 scondary Value of Gh, which fs correlated with H, tn Pig. 3.7. me correction factor correlation {4 show fn Tig. 3.8, Once a value for Liquid holdup has been deteratned from Figures 3.6, 3.7 and 3.8, the pressure grade tent caused by elevation change te ealeslated free yey : (),, HTL go] Exiction Eactor ‘Te pressure gradient component due to friction is given by318 "NOLLVISNYOD YOLIVI~dNGIOH-9°E' 913 "N dN i i | i > Ott ~ 49 98°0 : ga1Liso9SIA wi| +83z1s ONranL NO,d3SV8 NOLLW13uY0O wh / 4OLOVS dNQTOHseu 9 1N31914309 YIGNAN ALISOISIA UOI NOLLVISUYOI-L'E 914 W ol or 100" 10" my TTT TIT; TTT Trt 100° T Ta5 YOLIVI NOMLITYYOD AUVGNOITS NOI NOILV19YUN0I-8'E 914 wre "NZ (oe ™N “ND or_60 80 20° 90 sO #0 <0 2010"16 where ‘This equation can be written in tems of mass flov rate by multiplying and dividing by the square of the pipe ares. This gives seeereeeeee DT, #) soe. a oe), 25,40 6, ‘The expreseion given by Hagedorn and Brown vas Gp) > — £ 2,9682 x lp a? where we nase flov rate, Ib/dey, = denaity based o Liquid holdup, toe/ft°, d= pipe faeide diameter, ft, and f= tworph friction factor ‘the two-phase friction factor 1s correlated vith a tvorphase Reynolds number using a standard Moody diagran, Fig. 3.9, The Raynolds number 1s calculated from sees dD vhereseat reso000: par “sequiny spjoudey solsnas 5 « (tk) [owes s » ¢ Cpe ywacss + € (wie gieney sy COIR Wweos0 se0000 toe ‘000 383s 3 3Acceleration Tera ‘The pressure gradient due to acceleration is given by 0, de) (40 Ce ace 2 ae where 2 eT) Yr Cye%) TE we define 2 as ms Z () 0, 4 2) sae $@) -HG aa the totst pressure gradtent can be calesleted fren (a2 + @ ao. @at@e 312 $e Example Problem by Hagedora ~ Brown Nethod ctvea: Wag 7 4109 tt/eee p= 720 psta Van 7 2065 ft/sec T= 128° F d= 0,249 fr = 3,00 in, B= 18 ep Nyy * 6:02 bg 018 op Nyy = 9629 94, = 9646 Iba/eu ft = 0.08 9, = 2.86 Ibm/eu ft Ngo 41.36 €/8 = .0006 Neglecting scceleration, calculate the floving pressure gradient ot chase conditions,3-19 Determine liquid holdup and evorphase density. Deteraine CH, from Fig. 3.7 for H, = 0.08 to de 0.0055 To determine Hy ealeulate / ¥, Wen * (3) (3) ) ao +(e) ts) (2B eon c.er9 ena = 0.00032 sen ne 6 he Oat oe . 8 zo daceraine 4, coeutee MEAL « OILS? 5 cay nee aa From Fig. 3.8, 451 a= Gi t= 0.520 a = 0.520 og = yh, +P, GL - Hy) = (56.59(-520) + 2.04 CL ~ .520) » 29.49 + 1.36 pg 7 30:79 the/eu fe (2 = Kp, = 30.79 ibt/eu fe 2, Determine friettonsl tera. 1688 povd 1 1 EEE eat ah Catan Oetag) "TH (56.9 2.65942.) 09) bt tfeate , wc hen, F107? x sore = o65 ep et as 29 ie + = 92135, Tem Mss 39, {© 0601 for af8 © 086 2 fg) fa%e*s? , cocares epee + @),- “a SESE AGES = auumeien 3. Calculate total pressure gradient age) + (2) ~ + - bt. ek BiB) + (B),c9e9 + ve ous BE roa3-0 3.6 Category ¢ Correlations All of the methods in this category essentially used the three flow regimes discussed in section 3.3, except for the Bagge and Brill method. Sone of the studies involved only # change in calculation procedure {n one oF more flow regines fron previously published methods. For example the Orkiszevskt method presents new correlations for the slug flov regine only. ‘The correlations used in the other flow regimes were taken from previously Published work. Each method is discussed separately and the correlations for Liquid holdup or mixture density, friction factor and acceler are given for each flow regine. The Limits for determining which flow regime exists are given also, Buns_and_Ros ‘The Duns and Ros correlation is « result of an extensive laboratory study n vhich liquid holdup and pressure gradiente vere seasured. Correla- tions were developed for elip velocity (from which holdup ean be calculated) and friction factor for each of three flow regines (Fig. 3.1), The flow Fepines are defined as fonctions of the dimensionless quantities K, Ky, Lys Uys Ly typ and M, where Pe ee ren settee eeneeeBelD 275 16 75 +86 M, sence Deb 1, and L, are functions of Ny as show in Fig. 3.11, Duns and Ros presented correlations for « dimensionless elip velocity from which actual slip velocity ‘and Liquié holdup can be calculated using the folloving equations: seer BAS where seveeeeedeI6san saoquiny ewiBey moj -LE bly aySolution of Eq. 3.16 for Liquid holdup yields 2 a2 verry + | (y, thyy, 7 Z a . sae ‘s ‘The procedure for calculating the pressure gradient due to elevation change a: 1, Calculate the dimenstontess slip velocity §, using the appro- priate correlation, The correlation for $ te different for ich flow regine. 2. Solve Bq. 3-15 for the slip velocity, ¥,. 3. Caleulace che Liquid holdup, H, using Zq. 3.17. 4, Calculate the mixture density, 9, =, Hy + eg(1-B)- 5. Caloulate the pressure gradient due to elevation change, (ae) ta te rere SUBBLE FLOW luntts: 0H, SL, +h Hy, Two-Phase Density 3.19 ; F, and F, are given in Fig. 3.12. They are functions of the liquid viscosity number N,. F5 ean be obtained from: 7% seeeeneescsceeess 320 where Fy and F,.are also obtained fron Fig. 3.12. Exiction Factor ZigBee seseeeeeeee 32SuoqUINN Aloojan dys |1qqNg-zrE B143-24 uol}9e1109 UOIIDIA4 aiqqng-el'e’ Bly A Py ny tfFrom experimental data Duns and fos arrived at the following qe 13.22 fy By lf ceseeees £ A obtained free a Moody dtagran (Fig. 2.9) a9 function of the Reynolds number of the Liquid: us Yn! 1 ‘The factor £, 48 a correction for the in-situ gas Ifquid ratio, and ts given 2/3 tn Fig. 3.13 as a function of the group £ (,,/¥,,) ¥y/%. The factor fy te considered by Dune and Ros as an additional correction factor for both Liquid viscosity and in-eitu gas-Liqatd ratio and becomes important for |. Te te calculated from socseeesaeeseeeneneeD26 Ascolenatton Tere ‘The acceleration term vas considered to be negligible in the bubble flow eegine. ae How Latte: by #2, My SUgy Shy ~ Phase Fy» %q and F, are found in Fig. 3.16 as a function of the Liquid viscoriey umber, Hy» and Hi, = 0.029 Wy + Fy, sees 113.263-26 suaquiny Ausojeq dig Bnig-rre-614 4327 Extetion Factor The pressure gradient due to friction in this flow repine ts calcu ated using the procedure given for bubble flow. Acceleration Term ‘The acceleration term was consti flow regize. 1d to be negligible in the slug Mist FLow Limtes: N,, >I, ty he ‘uorPhase Denaizy Duns and Ros assumed that with the high gas flow rates in the mist flow region the slip velocity was zero, Therefore, the mizture density 1s the no-slip density and is calculated from nokta le Friction Factor In this region, the friction term te basi ys Thus: on the gas phase 2 @, = fe og 3.28 2p Age 7 Since there 4s no el{p, the friction factor is that given in a Moody Asagtam, but as a funccion of a Reynolds muaber of the gas 13.29 Duns and Zos noted that the wall roughness for mist flov is affected by the Lim of Liquid on the wall of the pipe. The ripples of the wall {Im cause a drag on the gas. This process ts governed by forn of the Weber ounber: 2613.303-28 and 1s also affected by Liquid viscosity. This influence was accounted for by making H,, a function of « dimensionless number containing Liquid viscosity, 2 1a & ‘The functional relationship is shown in Fig. 3.15 vhere the coordinates are Wve Nae 7 Sue Ma The value of roughness may be very enall but e/d never becomes smaller than the value for the pipe eself. At the transition sone to slug flow, £/€ may approach 0.5, Between these Limits, ¢/4 can be obtained from the following equations which were developed from Fig. 3.15. we ‘a 3.31 0749 0, Nye Sas .005 = Z tated dlesbal 13.32 matt 0.302 vans oy O, 4) yg np noose EP a 7 Poy a Where, + gucllgutd tncerfecta tension, dynt/en, Pg * Gas density, 1be/t23, = eoprtictal gas velocity, fe/eec, end G = pipe dteneter, fe Valoes of £ forthe nls flow regan may be found for 14> 0.05 from £ o|———+___ 4 0.067 ray] xa eee eens (4 to, 9(0.27¢/8)]? ‘As the vave height on che pipe walle increase, the actual ares through which the gas can flow is decreased, since the GLameter op flow of gas ts d+. Dune and Tos suggested that the prediction of frtceton Jose could be refined Ly substitution of dee for d and “and? for ¥, @o? shroughout the calculation of friction gradient. In this case the datermin- tion of roughness, €, 1s teerative,3-29 SSOUYIYL WIIy MO}y ISIW)-S'e'B1g NN3-30 Acceleration Term ewe dation ge eT Sa csc seeneceseeeessnereennennee BS a then the total pressure gradient can be calculated from a et az 1s at BANSITI Limits: L, <¥,, Increased accuracy vas clained Lf the gas density used in the mist flow pressure gradient calculation vas modified to aie where 9, 14 the gas density calculated a¢ the given conditions of pressure and teaperature. This modification accounts for pome of the Liquid entrained in the gus,S31 Example Problem by Duns-Ros Method gg 08 f/see @,, = 80 dynes/em Ygpt 265 ft/s00 p= 720 peta S a= 0.09 8% Teans’y 7 p= 1b ye Oo N,y= 9.28 Hg 0018 op 4, = 56.6 Ibm/ou tt fe a7 0.0006 tg 72.04 Ten/on Ny 2413 Nz 0.08 ‘Neglecting acceleration, calculate the flowing pressure gradioat st these con ttions. 1, Caleulate flow regime Limits and determine flow regime, From Fig, 3,21, Ly #153, L, = 0.68 Ly +E, Np, =1.59 + 0.88 02) «6.63 1,7 50+96N,, = 50+96 (6,02) = 266,72 0.78 Tag 75+ AN, VO = 397.8 Since L, +1 Ny

You might also like

• The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  

  From Everand

  The Subtle Art of Not Giving a F*ck: A Counterintuitive Approach to Living a Good Life

  Mark Manson

  Rating: 4 out of 5 stars

  4/5 (5814)
• The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  

  From Everand

  The Gifts of Imperfection: Let Go of Who You Think You're Supposed to Be and Embrace Who You Are

  Brene Brown

  Rating: 4 out of 5 stars

  4/5 (1092)
• Never Split the Difference: Negotiating As If Your Life Depended On It

  

  From Everand

  Never Split the Difference: Negotiating As If Your Life Depended On It

  Chris Voss

  Rating: 4.5 out of 5 stars

  4.5/5 (844)
• • Magazines
  • Podcasts
  • Sheet music
• Principles: Life and Work

  

  From Everand

  Principles: Life and Work

  Ray Dalio

  Rating: 4 out of 5 stars

  4/5 (609)
• The Glass Castle: A Memoir

  

  From Everand

  The Glass Castle: A Memoir

  Jeannette Walls

  Rating: 4.5 out of 5 stars

  4.5/5 (1717)
• Grit: The Power of Passion and Perseverance

  

  From Everand

  Grit: The Power of Passion and Perseverance

  Angela Duckworth

  Rating: 4 out of 5 stars

  4/5 (590)
• Sing, Unburied, Sing: A Novel

  

  From Everand

  Sing, Unburied, Sing: A Novel

  Jesmyn Ward

  Rating: 4 out of 5 stars

  4/5 (1104)
• Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  

  From Everand

  Hidden Figures: The American Dream and the Untold Story of the Black Women Mathematicians Who Helped Win the Space Race

  Margot Lee Shetterly

  Rating: 4 out of 5 stars

  4/5 (897)
• Shoe Dog: A Memoir by the Creator of Nike

  

  From Everand

  Shoe Dog: A Memoir by the Creator of Nike

  Phil Knight

  Rating: 4.5 out of 5 stars

  4.5/5 (540)
• The Perks of Being a Wallflower

  

  From Everand

  The Perks of Being a Wallflower

  Stephen Chbosky

  Rating: 4.5 out of 5 stars

  4.5/5 (2104)
• The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  

  From Everand

  The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers

  Ben Horowitz

  Rating: 4.5 out of 5 stars

  4.5/5 (348)
• Bad Feminist: Essays

  

  From Everand

  Bad Feminist: Essays

  Roxane Gay

  Rating: 4 out of 5 stars

  4/5 (1018)
• Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  

  From Everand

  Elon Musk: Tesla, SpaceX, and the Quest for a Fantastic Future

  Ashlee Vance

  Rating: 4.5 out of 5 stars

  4.5/5 (474)
• Her Body and Other Parties: Stories

  

  From Everand

  Her Body and Other Parties: Stories

  Carmen Maria Machado

  Rating: 4 out of 5 stars

  4/5 (822)
• The Outsider: A Novel

  

  From Everand

  The Outsider: A Novel

  Stephen King

  Rating: 4 out of 5 stars

  4/5 (1866)
• The Emperor of All Maladies: A Biography of Cancer

  

  From Everand

  The Emperor of All Maladies: A Biography of Cancer

  Siddhartha Mukherjee

  Rating: 4.5 out of 5 stars

  4.5/5 (271)
• The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  

  From Everand

  The Sympathizer: A Novel (Pulitzer Prize for Fiction)

  Viet Thanh Nguyen

  Rating: 4.5 out of 5 stars

  4.5/5 (122)
• Angela's Ashes: A Memoir

  

  From Everand

  Angela's Ashes: A Memoir

  Frank McCourt

  Rating: 4.5 out of 5 stars

  4.5/5 (441)
• Brooklyn: A Novel

  

  From Everand

  Brooklyn: A Novel

  Colm Toibin

  Rating: 3.5 out of 5 stars

  3.5/5 (1947)
• The Little Book of Hygge: Danish Secrets to Happy Living

  

  From Everand

  The Little Book of Hygge: Danish Secrets to Happy Living

  Meik Wiking

  Rating: 3.5 out of 5 stars

  3.5/5 (401)
• A Man Called Ove: A Novel

  

  From Everand

  A Man Called Ove: A Novel

  Fredrik Backman

  Rating: 4.5 out of 5 stars

  4.5/5 (4770)
• Steve Jobs

  

  From Everand

  Steve Jobs

  Walter Isaacson

  Rating: 4.5 out of 5 stars

  4.5/5 (807)
• The World Is Flat 3.0: A Brief History of the Twenty-first Century

  

  From Everand

  The World Is Flat 3.0: A Brief History of the Twenty-first Century

  Thomas L. Friedman

  Rating: 3.5 out of 5 stars

  3.5/5 (2259)
• The Art of Racing in the Rain: A Novel

  

  From Everand

  The Art of Racing in the Rain: A Novel

  Garth Stein

  Rating: 4 out of 5 stars

  4/5 (4203)
• Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  

  From Everand

  Devil in the Grove: Thurgood Marshall, the Groveland Boys, and the Dawn of a New America

  Gilbert King

  Rating: 4.5 out of 5 stars

  4.5/5 (266)
• The Yellow House: A Memoir (2019 National Book Award Winner)

  

  From Everand

  The Yellow House: A Memoir (2019 National Book Award Winner)

  Sarah M. Broom

  Rating: 4 out of 5 stars

  4/5 (98)
• A Tree Grows in Brooklyn

  

  From Everand

  A Tree Grows in Brooklyn

  Betty Smith

  Rating: 4.5 out of 5 stars

  4.5/5 (1929)
• A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  

  From Everand

  A Heartbreaking Work Of Staggering Genius: A Memoir Based on a True Story

  Dave Eggers

  Rating: 3.5 out of 5 stars

  3.5/5 (231)
• Yes Please

  

  From Everand

  Yes Please

  Amy Poehler

  Rating: 4 out of 5 stars

  4/5 (1898)
• Team of Rivals: The Political Genius of Abraham Lincoln

  

  From Everand

  Team of Rivals: The Political Genius of Abraham Lincoln

  Doris Kearns Goodwin

  Rating: 4.5 out of 5 stars

  4.5/5 (234)
• The Woman in Cabin 10

  

  From Everand

  The Woman in Cabin 10

  Ruth Ware

  Rating: 3.5 out of 5 stars

  3.5/5 (2522)
• Fear: Trump in the White House

  

  From Everand

  Fear: Trump in the White House

  Bob Woodward

  Rating: 3.5 out of 5 stars

  3.5/5 (738)
• Wolf Hall: A Novel

  

  From Everand

  Wolf Hall: A Novel

  Hilary Mantel

  Rating: 4 out of 5 stars

  4/5 (3811)
• John Adams

  

  From Everand

  John Adams

  David McCullough

  Rating: 4.5 out of 5 stars

  4.5/5 (2409)
• On Fire: The (Burning) Case for a Green New Deal

  

  From Everand

  On Fire: The (Burning) Case for a Green New Deal

  Naomi Klein

  Rating: 4 out of 5 stars

  4/5 (74)
• The Light Between Oceans: A Novel

  

  From Everand

  The Light Between Oceans: A Novel

  M.L. Stedman

  Rating: 4.5 out of 5 stars

  4.5/5 (789)
• Manhattan Beach: A Novel

  

  From Everand

  Manhattan Beach: A Novel

  Jennifer Egan

  Rating: 3.5 out of 5 stars

  3.5/5 (792)
• The Constant Gardener: A Novel

  

  From Everand

  The Constant Gardener: A Novel

  John le Carré

  Rating: 3.5 out of 5 stars

  3.5/5 (104)
• The Unwinding: An Inner History of the New America

  

  From Everand

  The Unwinding: An Inner History of the New America

  George Packer

  Rating: 4 out of 5 stars

  4/5 (45)
• Rise of ISIS: A Threat We Can't Ignore

  

  From Everand

  Rise of ISIS: A Threat We Can't Ignore

  Jay Sekulow

  Rating: 3.5 out of 5 stars

  3.5/5 (137)
• Little Women

  

  From Everand

  Little Women

  Louisa May Alcott

  Rating: 4 out of 5 stars

  4/5 (104)