Professional Documents
Culture Documents
The Technology of Artificial Lift Methods Volume 4 Kermit E Brown PDF
The Technology of Artificial Lift Methods Volume 4 Kermit E Brown PDF
nag HPs Gane +7 a= JP.—Py) From the constant productivity index, Peinsey €an be calculated as follows: =P,-# Petes P= 4 (2.62) By substituting Equation 2.61 and Equation 2.62 Inflow Performance 31 into Equation 2.60, the flowing bottom-hole pres- sure at the total flow rate is: rae (Pe= + F0.125)P, J x[at yao [= 7] 263) Somes (3) The interval between the maximum oil fow rate and the maximum total Bow rate, dense << Giner In this interval, the composite TPR curve would have a constant slope, since the curve is mostly affected by water production. So, tan B must be determined (refer to Figure. 2.35) for calculating the fowing bottomn-hole pressure at a total ow rate as follows: (@) Take a total flow rate that is very close to the maximum oil flow rate, ie: 4 = 0.999 deras () Since the difference between 4, and Gonag is very small, we can assume that a; = a, and Bs ‘By and the tangent of these angles can be calcu- lated geometrically in the shaded triangle. (©) From the shaded triangle, tan B= CD/CG tan a, =CG/CD CD is the difference between the flowing bottom- hole pressure at point D, Pyrg, and the flowing bottom-hole pressure at point C, Pyic; that is: CD = Paty ~ Pate (2.65) Point D lies on the composite IPR curve, so: Prty = FulPwigt) + Po(Prtnte) Or, from Figure 2.3 Prty = Fy X Paty + Fe X Prete 0.1250Ps) 2.64) TH x[a+ 54 — 50[ 92? Soom Geaux a 9 demas J ‘Therefore: Poy =F.{p,— 222s) p= Fa Pe + FY(0.126)P, (2.66) From Figure 2.35, Pye = Pwig, where G also lies on the composite TPR curve for @. = Goxae Pyig = Fo(Prtgy) + Pa(Patuaier) AEG: = Qeaaes Parga = 0; therefore: Prag = FaPwteued = Pe (P rartanh (eS)32 Technology of Artificial Lift Methods Substituting Equations 2.66 and 2.67 into Equation 2.65 yields: OD =Prig— Pete 0.001 domes oF. (=) + Fc120r,[-1 7 3 =e - + os — a0 [ee }) 2.68) Gemas ~ CG is the difference between q, and Goma therefore: 0G = Gomar 0.999 Gemax= 0-001 Gomer (2:69) Hence, from Equations 2.68 and 2.69, tan a; or tan’; can be calculated. ‘The flowing bottom-hole pressure at flow rates between the maximum oil flow rate and the maximum total flow rave can be calculated by using the following equation: Pac Fe (P,—S2#) —(q,— 284) tan a) (2.69) ‘The maximum total flow rate (for the composite IPR curve) can be calculated by using the follow- ing equation: Ginaz = onus + Pong (tan @) or max = Goss + Fe ( fe ous) (tana) (2.70) 2.2272 CALCULATION OF THE TOTAL FLOW RATE AT CERTAIN FLOWING BOTTOM-HOLE PRESSURES FOR THE COMPOSITE IPR CURVE Refer to Figure 2.36, which shows that the composite IPR curve can be divided into three intervals, and in Pr Putte) i) Yb Yomax* Figure 2.36 Composte IPF Curves SGtmax Swmax every interval, the total flow rate at certain flowing bottom-hole pressures can be calculated as follows: (Y) For pressures between reservoir pressure and the bubble-point pressure, Py < PayPa), the variables that are needed to construct the composite IPR curves can be calculated by using the following equations: Suet a (2.78) =I P,— Py) (2.79) Gogae = +E (2.80) Gham = Gene +Fe (Pe) an a) (281) where tan a can be determined by applying Equa- tion 2.64. (2) The flowing bottom-hole pressure of the test is less than the bubble-point pressure, Pwin < Ps. For determining qu, Gomar) 224 digas the productivity Eindex J must be calculated. The following section shows the derivation of the equation to determine Jat this condition. Refer to Figure 2.36. = Fo (Go) + Fw (Gol (2.82) toe a{i-02(282) Gon = Te + (Gomax ~ A) Py =~ ey) og = JP, P+ {or® Ja any = JP, = Py + (FE) 8 (2.83) Gey =F [Pe = Patel (2.84) ‘Substituting Equation 2.83 and Equation 2.84 into Equation 2.82 yields: Pea PR Ped) eae Por a=, (3(2- | +Fae.- Pree] 2.86) PLA) BP p+ Es i + PAP Parad If J can be determined from Equation 2.86, and Ging, Can be calculated by applying Eq 2.79, 2.80, and 2.81, respectively. 22274 THE COMPOSITE IPR CURVES FOR THE RESERVOIR PRESSURE BELOW THE BUBBLE- POINT PRESSURE At this condition, the previous equations for con- structing the composite [PR curves for the flowing pres- sure below the bubble-point pressure can be used by changing P, to P,, setting the value of q equal to 0, and using the same procedure. EXAMPLE PROBLEM #21 (reservoir pressure greater than the bubble-point pres- sure) Given data: reservoir pressure = 2,550 psi bubble-point pressure = 2,100 psi Test data: Flowing bottom-hole pressure = Flow rate (total), q, = 800 b/d 300 psi Inflow Performance 33 Calewlate: Determine the composite IPR curves for Fy 0.25, 0.50, 0.75, and 1.0 Solution procedure: (2) (Example for the preliminary calculations for Fy = 0.50) Since Pris > Pet 500 J = 5335 75a 720 b/A esi (2,012,550 — 2,100) = 800 b/d (2,012,100) omae = 900 + FE 00, % = 3,233.83 b/d By applying Equation 2.68, CD can be calculated as follows: 0.001(8,288.33) 50S + 0.500.12612.100) PITTI DO) «far i-@P Saas |) = 7.89 cD= By using Equation 2.69, CG can be calculated as follows: By using Equation 2.70, the maximum total fow rate is: Anas = 3,298.83 + 0.50 (2850 - 424.68 b/d (2) The calculation of Pye at certain total fow rates is as follows: For q: = 600 b/d, a > ox: 10 — (600/2.0) 50 psi ) + 051012812,100 300-900, 13,289.33 — 900, X=1+ 81-80. 1,789.08 psi 300 b/4, domes <4 < Gina! 50 (2,850 ~ S288 38) =3,300 20 ~ 3,233.39(2.44) = 308.99 psi (3) The calculations of the total fow rates at certain flowing bottom-hole pressures are as follows: (a) By using Equation 2.67, calculate the flowing, bottom-hole pressure at di = donee = 3,233.33 b/d—that is: ee Prag * 0.50 (2880 aS For q = Poe ) = 488.67 psi ce34 Technology of Artificial Litt Methods (b) For Pac = 2,400 psi, Par > Py: 9 =2.0(2,550 ~ 2,400) = 300 b/d For Pr = 1,500 psi, Prig < Pwr < Pet 1,800 + 0.12640.60%2,100)~ 0,502,550) - 0.125(0.5012,1001 7 0.001805 ap 004461 ~si= 1085281 Bx 0, 80 = 0061 + LOOT = OORT . BO.001S05 = 2018.18 b/d For Par = 350 psi, Pee < Pug 466.67 + 8,289.39(2.44) ot 2.44 (4) The results of the preliminary calculations for other values of F. are shown in the following 350 3,261.15 b/d table: F. 00 025 080 as 3.00 a 20 20 20 20 20 % 000 $000 © 800.0000 8000 Gin 9.28808 9.20809 920898 3.23339 9.28089 wre zs 028 at 068 20 ane 458 at 2at 147080 Ginn S238S3 9901.73 3ans52 9708.98 5,100.00 Poe 0.00 2329 466.67 700.0 ~ (©) The results of the calculations of the total fow rates at the Sowing bottom-hole pressures for ev- ery value of Fy are shown in the following table: 3.302 (6) The plot of the composite IPR curves is shown in Figure 2.37, EXAMPLE PROBLEM #22 (Pyiee ig less than the bubble-point pressure] ‘aos 3360 PO Figure 257 Composite 1PR Curves for Example Problem #21 Given data: reservoir pressure bubble-point pressure test data: Pat cn 800 psi 2.400 ps 1.200 psi 1,480 b/d Calculate: Construct the composite IPR curve for Fy = 0.50. Solution procedure: (1) Preliminary calculation: 10~0.25,200/2.¢00r= o4.2n4/2,200¢ =070 J+ 50 sno 6300 ~ 2222) a sags o 02.600 ~ 2.4001 403.68 b/d 964 ONE 004 0.21 484 Using Equation 2.64, tan a tan 8 1,749.09" Ghanx = 1,749.09 + 0.80 (2.200 = Jo. 21) 101 = 1,859.18 b/d (2) The flowing bottom-hole pressure vs total flow rate relationships are as follows: eRaa Pa 2600 ° 2800 202 2400 404 2000793 1500 1,140 120011498 ‘300 1.862 4004777 459 (8) The plot of the composite IPR curve is shown in Figure 2.38, ay 8Po O58 — 900 Figure 238 Composite IPR Curve for Example Problem #22 EXAMPLE PROBLEM #23 (reservoir pressure is below the bubble-point pressure) Given data: reservoir pressure = 2,260 psi bubble-point pressure = 3,000 psi test data: Pres = 1,800 poi Gig, = 900 b/d Calculate: Plot the composite IPR curves for Fy = 0.0, 0.25, 0.50, 0.75, and 1.00. Solution procedure: (1) The preliminary calculations for every value of F,, are made by using the same procedure as in example problem #21 except the values of Ps Inflow Performance 35 = P, and qy = 0 must be set. The results are shown in the following table: Fe 20 025 05078 1.00 4 220 214209208 en 274090287857 2610.28 2ssaee2 = — ‘ana 025” ot 043088200 ten 8 soz az 235, 148 oso Ginn 2743.90 275655 282927 307587 4,500.00 (2) The relationship of the Sow rate vs the fowing bottom-hole pressure for every value of Fy is tabulated as follows: a@ Fe Pa 0 oa oso (O78 2.250 ° ° ° ° 2000-522 51S 5188 400 = 900g? woo 455315714508 1.838 000 2088202178 za 00 2ari_ 2323 ast? soo 25772618. 2es) 2700 0 arse ars7 3.078 (S) The plot of the composite IPR curves for every value of Fy is shown in Figure 2.39, Py Pt 200] 00] Figure 2.29 4 -9P0 Composite IPA Curves for Examele Problem #2936 Technology of Artificial Lift Methods 2.23. PREPARATION OF FUTURE IPR CURVES. ‘The preparation of future IPR curves becomes very important in determining future production rates for both fowing and artificial lift wells, These IPR curves ‘also represent a means of determining when a well is to be placed on artificial lift. They are necessary in looking at rate acceleration projects and comparing artificial lift methods. Figure 2.40 represents a typical set of these curves, PRESSURE > 3 7 RATE Figure 2:40 IPR Gurves for Prosent and Futuro Reservoir Pressures 2.281 FETKOVICH PROCEDURE In early years, Muskat noted that the productivity index on a well at time 1 could be related to the pro- ductivity of a well at time 2 by the following relation- “ata Fetkovich noted that, in material balance calculations for solution gas drive reservoirs, ky. Was approximately linear with pressure.** He then wrote the following equation, which approximates the change in oil perme- ability with pressure depletion: (2.87) (2.88) or: (2.89) where ky, is taken with respect to ki and is defined at a vanishing AP (zero drawdown). He then con- structed Figure 2.41, which defines a locus of values Telia sattd-ourt® Figure 241 Fotkovich Flot of Locus of Values (Kn! sB4) v8 Pressure of (Bee, vs pressure at zero drawdown. Using his two equations: oy = Joy Pry? = Pave (2.90 which defines rate vs drawdown, and: Fy ke P= 5. 2.91) which corrects for depletion, the final equation pro- posed by Fetkovich is obtained as follows: ., (&) ®, Ifa three- or four-point flow test is being conducted at the present time, it is possible to predict IPR curves at other static reservoir pressures by use of Equation 2.99. EXAMPLE PROBLEM #24 Given data (From example #16) ‘The equation describing this test w: {2,500} ~ Pa] 012 an 3.906 [SSPE Pee] Calculate: (Gongs When P, lowers to 1,800 psi 2) aor Pay = 600 psia when P, = 1,800 psia Solution procedure a Pr , (Be) mPa 1800 psi, 0) .s00r op cere (33te) 1,000 =819 b/d or 1,800 oe qo = 0.03062 (a) [2,800)2 — op: stg 5 mm (2) Find qo for Put = 800 psia when Py 1,800" 1,800 psia: a» = 0.03062 (EPG) (.8007 — Goran =102 b/d 2,500, By assuming other values of Py, a complete IPR curve can be prepared for P, = 1,800 psi. Also, other IPR curves can be prepared for other static pressures such as 1,500 psia, 1,000 psia, ete, CLASS PROBLEM #24a Using the data of class problem #16a, prepare future IPR curves for static pressures of 2,000 psia and 1,000 sia. CLASS PROBLEM #240 Using the data of class problem #16b, prepare future IPR curves for static pressures of 3,800, 9,000, and 2,400 psia. 2292 COMBINATION FETKOVICH AND VOGEL PROCEDURE FOR THE PREPARATION OF FUTURE IPR CURVES Eckmier noted that if we take the equations of Fetko- vich for static pressure at time 1 and divide by the inflow equation for static pressure at time 2 we arrive at an equation for determining dina, at time 2 after which we can use Vogel's equation directly for prepara- tion of the [PR curve."7 ey = Jo, (Pry? — Pa &) Say (Br (2.100) Pee (2.10) By solving fOr Goya (Setting n = 1.0 and Pur = 0) and dividing Equation 2.100 by Equation 2.101, we have: Goran! Gomexg = Pry! Peal? (2.102) ‘Therefore, we need only one test at the present time from which we can obtain Gomar,. By knowing Gems, and P,,, we can solve for Goma, at any other desired static pressure (P,,), and in turn, by using Vogel’s equa- tion, a complete IPR curve at time 2 (P,,) can be con- structed. EXAMPLE PROBLEM #25 Given data: P, = 9,000 psia (Py = 3,000 psia) ‘A test shows q = 200 b/d at a flowing pressure of 2,500 psia, Caleulate: (1) Gogue When the static pressure drops to 2,000 psia (2) ge for Pe = 2,000 psia and Pye = 1,000 psia Solution procedure: (2) Solve for Gongs at Pr = 3,000 psia: Inflow Performance 37 o/nas = {i -02 &) -08 ] (2) Find qa: at P, = 2,000 psia, Use the relationship: eos &) = (oy io omaxz \Pra/— \2,000, 720 Gennes = 3.97 = 218 b/4 (3) Use Vogel's equation to find qs for Py psia and P, = 2,000 psia: Oe, ise) 0] 219 [1-02 (352) -o8 (ie) |= 149 ra CLASS PROBLEM #252 Given data: 2,400 psia_ (P, > 2.400) 10 180 b/d for Pac Calewlate: 1,700 psia (1 Gagge When Pe drops to 1,800 psia (2) qn'Por Pur = 1,500 psia, 1,000 psia, and 700 psia when the static pressure is 1,900 psia GLASS PROBLEM #25b Given data: 1,790 psia_ (Py > 1,790) 95 b/d for Pq = 1,500 psia 10 Calculate: (D the maximum flow rate when P, drops to 1,600 psia (2) the flow rate for Par = 1,400 psia, 1,200 psia, and 1,000 psia when the reservoir pressure is 1,600 psia CLASS PROBLEM #25¢ Given data: P,= 2,850 psia qq = 685 b/d for Per = 2,100 psia FE=LO Calculate: (1) the maximum flow rate when P, drops to 2,000 sia (2) the flow rate for Pwe= 1,800 psia, 1,500 psia, and 1,200 psia when P, = 2,000 psia38 (8) the flow rate for Prt = 1,500 psia, 1,200 psia, and 1,000 psia when P, = 1,800 psia Technology of Artificial Lift Methods 2.233 STANDING'S PROCEDURE FOR PREDICTING FUTURE IPR CURVES Standing used Vogel's equation in combination with ‘Muskat’s relationship concerning present and future productivity indices in presenting a method for predict- ing future inflow curves." The procedure requires that ‘we use material balance calculations in order to deter- mine future saturations from which we can obtain k, values, ‘The following relationships are used: (Kew er ) 3 (Be) | (Fs 2.114 GS/s), exe In order to use Equation 2.114, we need present and future values of kro, #e and By. Therefore, we need to apply material balance calculations in order to estab- lish oil saturations vs P. since kre is a function of oil saturations—which in turn is a function of static pres- sure, Standing suggested using the Tarner"” gas deple- tion calculations for determining oil saturation. If kre values are not available, Standing also suggested that the work of AT. Corey could be used to obtain values for kro! i 2.115) Sh (Be iC. (2.103) a7)! GS, where = -o2(2)-08 (Ex) S.=S.+Sq wewefi-ens)-os] amo “SB (2.305) oF e116) where n is approximately 4 for consolidated sandstonee and nonvugular limestones. The value of n is less for (2.106) unconsolidated sandstones (2207) EXAMPLE PROBLEM #26 Vogel’s equation can be rearranged: Given data: =f: Pp 208 (Pe ae tome (1B) [24 08 ()] (2.108) i By substituting q = J (P, — Pai) into Equation 2.108 a and solving for J, we have reservoir pressure flowing pressure a (2.109) towrate i fw etfiioney to By dividing Equation 2.109 by Equation 2.107, we have: (, ee tat (1408 @u1) 2862 Jeb (ie : cs 4.2862 1s te 1,800 Substituting Equation 2.105 into 2.104, we have: | acl IP, Put) 9g (Pet\ o-EEfr-o2(B)-00(%)] enn Caleulate where J° for any future P, is determined from: He ( Kero ) ( kro, ° \neBo | HeBo/» It is necessary to test the well at the present time so that P, and the present value of J can be determined. (@) Calculate the present value (J3) from Equation 2,107, or J can be determined from Darcy's Law— that is: vt (2.112) 1.08 ¥ 10°? kh . 3, BB, (tm rete$ 8) (2) Calculate the future value of J from the relation- ship: 2.113) the flow rate for Per = 1,500 psia and 1,000 psia when reservoir pressure drops to 2,000 peia Solution procedure: (@ Using Equation 2.108, calculate J: 2.70 2.024 (2) Using Equation 2.107, calculate J3, as follows: 2.814.325) _ 2,500 = 4,528 b/d J; 3.1(8) Using Equation 2.112, calculate Jt, as follows: 0.18 (G829)0 2562) aaa EE (0.542111 3190) (4) Using Equation 2.111, calculate the flow rate for specific fowing pressures: For Pwr = 1,500 psia @ P; = 2,000 psia: (2.67)(2,000) 1,500" 0 0 -00( “teh For Pwr = 1,000 psia @ P, = 2,000 psia: 2, (2.67%2,000) fp aos (22) 18 ~04 (G52) | 2.234 COUTO'S PROCEDURE Couto's procedure, by using Equation 2.41 or 2.42, can also be used to predict future IPR curves if the relative permeability, the oil viscosity, and the forma- tion volume factor at certain reservoir pressures in the future are known, dg =(8.10) 2,000, 186 b/d 2,00 075 bid EXAMPLE PROBLEM #27 Given data: Use data from example problem #23. Calculate: the flow rate for Pwr = 1,500 psia and 1,000 psia when reservoir pressure drops to 2,000 psia Solution procedure: (D Using the data from example problem +26, Equation 2.41 can be simplified as follows: (62.5160) a= 0.400012 (Sea soe) 0.80 x 26000 (sappy aig; OKLO BD x {18 —0.8(1.0X1.0-R)) =2,964.25(1.0 ~ RKL.8 —0.80.0~R)] (2) For Pu = 1,500 psia and P, = 2,000 psia: R= 1,500/2,000 =0.75 go = 2.964.250 — 0.751(.8 — 0.811 —0.75)) =1E6 b/d (8) For Pee = 1,000 psi and P, = 2,000 R= 1,000/2,000=0.5 qo = 2,964.25(1 — 0.5)[1.8 — 0.81 —0.5)] 2,075 b/d Inflow Performance 39 CLASS PROBLEM #278 Given data: Present ture condition conaltion revervol preseure 2210 1980 flowing pressure ‘\g75 - flow rata 11932 - flow sttciancy 1 1 be oa 0.25 ene 6 Fa = 0.9429 ogres 5 1.3528, 3.2909 " 1,983 419853 te ‘033, 033 B 10 70 Cateulate: ‘Using Standing’s and Couto’s procedures, calculate Alow rates for Pwe= 1,600 psia, 1,400 psia, and 1,000 psia when the reservoir static pressure = 1,850 psia. ‘CLASS PROBLEM #27 Given data: Present Fare contin coneltion ‘rasanvoirprossuro (i) 2.95 2375 owing prossure (ps) 2581 - row rate (0/6) ‘582 = flow etficancy 10 10 ke 0.2 025, ow ci 6s a 0.2970 03298 8 1.4738, 1.3808: raft 1.590 1,590 mit 033, 033 nit 0 o Colewlate: (2) Caleulate flow rates for Pwr = 2.000 psi, 1,800 psi, and 1,500 psi when P, = 2,975 psi by using Couto's procedure. (2) By assuming FE = 1.0, calculate flow rates for Py = 2,000 psi, 1,800 psi, and 1,500 psi when P, = 2,375 psi by using Standing’s procedure. CLASS PROBLEM #27¢ Rework problem #27b, section (1) only, by assum- ing FE = 2.0. 2.235 PIVOT POINT METHOD Uhri and Blount derived the pivot point method to predict future inflow performance relationship (IPR) curves by considering two IPR curves of a well, which represents the different stage of depletion ofa solution- gasdrive reservoir.” This method can be applied40 Technology of Artificial Lift Methods graphically or numerically by calculating the relation- ship between the productivity index, (—dao/dPwd) at Py = P,, and reservoir pressures, "The productivity index is determined by differentiat- ing Vogel's equation, that is ge _ dows Pa) 9 908 (_pp—1.6Pe! . SB (02-10 con For the maximum flow rate, P, ‘0.0, and the equation becomes (2.118) and for Pur = Pe yg doe Pa 18 Pr, (2.119) Based on Equations 2.118 and 2.119, the productivity index at Py; = P, is equal to nine times the productivity index at Pur = 0.0, or: <{@ fess ane ePwch oat To predict future inflow performance relationships, this method needs two flow tests at different times One test can be in the past and one test at present, or both-can be at different times in the past. Using these two flow tests and applying Equations 2.118, 2.119, and 2.120, the relationship between (dq,/dPw) vs P, can be obtained. ‘The following procedure is used to predict future IPR curves by using the pivot point method graphically (refer to Figure 2.42) @) Calculate the maximum flow rate from each flow test using Vogel's equation. (2) For each flow test, calculate dqo/dPy« for Per = 0 and Pyr= Py using Equations 2.118 and 2.118. (3) Draw the X-¥ axis with dq./dPy as the vertical axis and Pwr as the horizontal axis. (4) Plot the value of dqo/dPww V5 Py a5 shown in Fig- ure 2.42 from the flow tests. Point P=dq,/dPwc at Pyy= 0 from first test Q=dqy/dPur at Par = P, from first test R= dq./dPy at Py, =0 from second test S=dqo/dPy; at Pr =P, from second test (5) Draw a line between P-Q and R-S and extend these ines until both lines intersect. This intersection is called the pivot point—PP. (6) If 0 is the point of origin, divide line OP (on the vertical axis) into several parts, for example into four parts—that is, P,, P2, and P:—and draw lines from the pivot point (PP) to P,, Ps, and Py—that is, PP-P;, PP-P,, and PP-P;—and then extend the Tines far enough to incorporate the envelope. (7) On the vertical or dq./dPy axis, determine points qu Ge, and qs, where the distances Og, Oq., and Og) are determined, based on the value of OP,, OP, and OP;, respectively, by applying Equation 2.120; that is: (2.120) OP, x9 OP, x9 Og: = OP x9 49g /8P yy OR 1.8CQqugy /P,) FOO 80510051265 RESERVOIR PRESSURE, PS! E56 Figure 242 Wustration to Construct Reservoir Pressure Envelope us- ing the Pivot Point Method (8) From qx, gs, and q3, draw horizontal lines until they intersect the extension of PP-P,, PP-P,, and PP-P,, respectively, and obtain points R,, Ra, and Ry. (9) Connect points 8, Rs, Rs, Ri, Q. and O. This curve is called the reservoir pressure envelope. (10) Ifthe horizontal axis is considered to be the reser- voir pressure P,, the vertical axis is 18 (Quex/P:) (11) Using the reservoir pressure envelope from step 9 and for certain values of reservoir pressure in the future, the value of 1.8 (quux/P,) can be deter- mined; then the maximum flow rate can be caleu- lated. (22) An IPR curve into the future can be predicted by applying Vogel's equation based on the future reservoir pressure and the maximum flow rate from step 11. Uhri and Blount also derived equations to predict IPR curves numerically, which can be obtained in two ways, that is: (D) numerical solution using the pivot point method (2) numerical solution using the P,-envelope equation For the numerical solution approach using the pivot point, the following equations were derived: ¥6 (Gaus), Pas? X Pes ~ (Gna) Pes? X Pr] (Gaels Pra? = (Greed Pr? Pic 12 aeesi (2122) Cat) “TreeyPp) 29 )@ Pero 22g — wary Ps; and (~dq/dP.1)* are the coordinates of the pivot, point, and subscripts 1 and 2 refer to well tests 1 and 2, respectively; subscript f indicates a future curve’s condition. ‘The following procedure may be followed to predict fature IPR curves: () By using Vogel’s equation, calculate the maximum flow rate for both tests—that is, (nas ad (Qnea) (2) By using Equation 2.121, calculate Ps, (3) Calculate (— da,/dPy)* by using Equation 2.122. (4) Fora particular future reservoir pressure, calculate (Ss), @ Peo, by using Squation 2.128. (5) Based on the result in step 4, the maximum flow rate at a particular future reservoir pressure can be obtained by using Equation 2.124, (6) ‘The future IPR curve can be obtained by using Vo- gel’s equation based on the values of Py and (Gras) For the numerical solution approach using the Py envelope equation, Uhri and Blount derived an analyti- cal expression for the P.-envelope. The equation is: Gus AXP, Peta where A and n are constants for a particular well. ‘These constants can be determined by using the values of P, and the maximum flow rate from the tests as follows: (2.125) P,P, asa (2.126) Pat Geet Gm n=P,, [A Pr,/(@nah — 1] (2.127) ‘The procedure to predict future IPR curves by using the P,envelope equation is as follows: (1) By using Vogel's equation, determine the maxi- mum flow cates for both tests. (2) Calculate A and n by using Equations 2.126 and 2.127, respectively. (3) Calculate the maximum flow rate for a particular future reservoir pressure by using Equation 2.125, (# By using Vogel's equation, obtain the IPR curve for a particular future reservoir pressure. EXAMPLE PROBLEM #28 Given date: two well tests: Tet PP 11,900,660 511.60 2 t7so 187528082 Inflow Performance 41 Calculate: (Q) the maximum flow rate when the reservoir pres- sure drops to 1,600 psia (2) the flow rate for Py * 1,200 psia when P. = 1,600 psia Solution procedure: (@) Calculate the maximum flow rate from each test by using Vogel's equation; then calculate dq/dPue for Pyy =O and Py = P, by using Equations 2.118 and 2.119, respectively. g/day Test # aoe @ PaO @Pu=Pr + 22028 Oza 2708 Ee 0g 70g OR 1 Blagman Pp) 360 3866 a Figure 243. P, Envelope for Example Problem #28 (2) Plot dq/dP,« vs P, and determine the pivot point as shown in Figure 2.43. (8) Assume several values of dq/dPv; @ Puy = 0 and calculate dq/dPur @ Pye = P, by using Equation 2.120 as follows:42 — Technology of Artificial Lift Methods aI Pa @ Pano 9/dPaz (assumed) @Pu=P, 045 135 0.10 080 0.05 04s (4) By using steps 6, 7, 8, and 9 of the procedure, draw the Prenvelope as shown in Figure 2.43. (8) Using the Penvelope, for P; = 1,600 psia, the value of 1.8(@ms)/P, = 140. The maximum flow rate is: nes = (1.40(1,600)/1.8 = 1,244 bo/d (6) By using Vogel's equation, the flow rate for Pa: = 1,200 psia ist wna fio-00 (3) 00) = 498 bo/d EXAMPLE PROBLEM #29 Given data (same as for example problem #28). Calculate: (using the numerical solution proposed by Uhri and Blount) Solution procedure: (1) Numerical solution using the pivot point method: (a) Calculate the maximum flow rates for both tests: Test # P, Se 228428 1681.40 - () Calculate Pz, by using Equation 2.121 as fol- lows: pew tlh. 40 st. 15)—2294 2541750500) * (@.284.26n3,150°~ 1,691 4081, 900 =-0181 (©) Calculate (~ $8)" by using Equation 2.122, ( #) (2,294.26) @ Caleulate tion 2.128: dae 0.1265 (a), -reacsanria 78 (e) Calculate the future maximum flow rate by using Equation 2.124: 0.1563 = 0.2 (@max)/1,600 0.153 x 1,600 (ea = So 250.14 b/d. (£) The flow rate @ Pax = 1,200 psi can be caleu- lated by using Vogel's equation as follows: qo= 1,250.14 [2-02 500 b/a (2) Numerical solution by using P,-envelope equa- tion: (@) The maximum flow rates for both tests had been calculated previously. (b) Calculate constant A, using Equation 2.126: 1,900 000k _ 1 TBOF 2,204.28 7,601.46. (6) Calculate constant n, using Equation 1.127 0.63270, 22) Sate (Pome 2,294.26 )- (@) The Ppenvelope equation is: naz 0.6827 P, P, B= 2895.57 (e} For Pr, = 1,600 psi: 1,600(~-0.63271,600) 1,600 — 2,895.57 (f) The flow rate at Py; = 1,200 psi is 600 b/d (by using Vogel's equation). A= 895.57 (Gna) 1,250.19 b/d CLASS PROBLEM #260 Given data: two well tests: Tete PF Pa 1 2328 12s (6s 2 2150 1,780 407 Calculate: (Q) the maximum Bow rates when the reservoir pres- sures drop to 2,000 psia, 1,600 psia, and 1,200 psia (2) the flow rates for flowing pressure = 900 psia when the reservoir pressures drop to 2,000 psia, 1,600 psia, and 1,200 psia CLASS PROBLEM #280 Given data: two tests were taken on a well: Te# Pa 1 19801250 Ga 2 18801300218a] ae FS RE Calewlate: (2) the maximum flow rates when the reservoir pres- sures drop to 1,500 psia, 1,200 psia, and 1,100 psia (2) the fow rates for fowing pressure = 950 psia when the reservoir pressures drop to 1,500 psia, 1,200 psia, and 1,100 psia CLASS PROBLEM #286 Given data: two well tests: Test # Pe Pye ' 2@80 1800501 2 20101700 27 Caleulate: (1) the maximum flow rates when the reservoir pres- sures drop to 1,900 psia, 1,750 psia, and 1,500 psia (2) the dow rates for Py;~ 1,000 psia when the reser- voir pressures = 1,900 psia, 1,750 psia, and 1,500 psia CLASS PROBLEM #264 Given data: two well tests: Tost # Pa + 1800274 2 3200271 Calculate: (D the maximum flow rates when the reservoir pres- sures drop to 1,400 psia, 1,200 psia, and 1,000 psia (2) the flow rates for Pay = 1,000 psia when the reser- voir pressures = 1,400 psia and 1,200 psia 2298 OTHER METHODS OF CONSTRUCTING FUTURE IPR CURVES Refer to pages 25 through 82 of Volume 1 of this series for work by Eckmier" and others in establishing fature IPR curves. 224 TRANSIENT IPR CURVES FOR OIL FLOW ‘There are many wells with low permeability that require a long time period to reach pseudo-steady-state Sow. It therefore becomes important to construct [PR curves at different times in order to predict the flow rate at these times. 2241 STATES OF FLOW IN THE RESERVOIR In a finite reservoir system, there are three states of flow (How regimes): Inflow Performance 43 (2) transient flow (@) late transient flow (8) pseudo-steady-state fow ‘These three states of flow are shown schematically by plotting flow rate vs time for a certain constant flowing pressure as shown in Figure 2.60. Transient fow occurs for a certain time, and its pres- sure behavior is essentially the same as that of an. infinite reservoir. As the time becomes larger, the pres- sure decline becomes a linear function of time; at this, condition, pseude-steady-state flow occurs. The flow pe- riod between the transient and pseudosteady-state is, called the late transient period in which, at the begin- ning of late transient period (or at the end of the tran- sient period), the pressure drop caused by production, has been felt at the drainage boundary of the well. At this condition, the IPR curves decrease as time in- creases. At the end of the late transient period or when the pseudo-steady state is reached, the transient IPR curve Is close to pseudosteady-state IPR curves. The duration of the transient period may be esti- ‘mated from the following equation: Be Clrel® 002687 where ty = duration of transient period, hours. ‘The late transient period may be estimated by the following equation: tee (2.128) the Cr lee? ti (2.129) 0.00088 ke Equation 2.129 was derived for a circularshaped reser- voir with an error percent of 0.0L. To talke care of some average errors in the shape factor, Ertle suggested us- ing the following equation to estimate the late tran- sient period yg = Seba Clee? 0.001085 ke For t
Cw #108, 760 205 B20 2zol H 26 6080 100 120 140 160 TEMPERATURE, °F Figure 2.490 Average Compressiblty of 100,000-ppm NaCiin Distiled ‘ater (attr Long and Chir) Teo 200 220 280; 260) 2.403685 aL = 220) eee fest shane ere 17 F 200) 40 60 60 100 120 a0 160 TEMPERATURE, OF 186 200 220 Figure 2.502 Average Compressibilty of 20,000-ppm NaClin Distited Water (after Long and Chieria)* (2.147) For determining formation compressibility, correla- tions of Hall” and van der Knaap® have been used extensively, but according to Newman,¥ those correla- tions do not apply to a very wide range of reservoir rocks. Figures 2.56 through 2.59 are the plots between initial porosity, di, vs pore volume (rock) compressibil- ity at 75% lithostatic pressure based on the large nurn- ber of samples studied by Newman, The lithostatic aus ona axel [ WS Feo SECENOV'S COEFFICIENT, eo 80105 eee ed Teo TEMPERATURE, OF Figure 251 Secenov's Coetfcient for Mathane (alter Long and iia T L 109 120 160 160 TEMPERATURE, oF Figure 2500 Average Gompressibity of 300,000-pom NaC in Listed Water (after Long anal Chien 10 ‘30 6080 190 205 220 Re is datermined by using Figure 254 B, is calculated by applying Equation 2.141 By-is determined by using Figure 2.55 (@) Formation or rock compressibility. Formation compressibility is defined by: SOLUTION GAS-WATER RATIO, Fay, CU FT/BBL. oli . wa a. ‘O14 eae TEMPERATURE, *F aye, CORRECTION FOR BRINE saLiniry ze sb Pea “HE de o 6 20 30 © TOTAL SOLIDS IN BRINE, ppm x10"3 Figure 252 Solubitty of Natura! Gas in Wator (ater Dodson and Standing)a — : 2 ls - Gs anne eq at 78 © if § asl 1. 1. 1 : Sa TEMPERATURE, *F Ee 3 SEP comederin mor eut gousron a eae t ae | ie = ge GAS-WATER RATIO, CU FT/BBL ‘Figure 2.53 Erect of Dissolved Gas on Water Comprassibity (alter Dodson and Standing) Tacs waste 3 “avgpice ron ae reurenarones e ir oie Te tee 5 B ccod | AVERAGE ERROR FOR 150% TO B908F 1S 9.39%, MAAN ERROR 118% ol i eet L 0 1000. -a000 S000 4000” 8600 PRESSURE, PSIA CORRE NON AACTOR Fa SALINITY ie” oo 2” as B° a} ee ‘TOTAL SOLIDS IN BRINE, PPM x10? Figure 2.54 Change of Natural Gas in Sotuton in Formation Water ih Prassura vs Prossure. Multiply Rx! ®p) by the Cor- ‘action Factor to Got Result for Bre (ater Ramey, data (of Standing!) inflow Performance 47 eure waren Aaa waren Puls naTuea oss e WATER FORMATION VOLUME FACTOR, By, RB/STB 8 101 i 100} 22) oo ase! 0 1600 #0058608 4000-8500 PRESSURE, p, PSA Figure 255 Formation Youme Factor of Puro Water and a Miura Of Natwal Gas and Water (data of Dodson and Standing FRUBLE SANSTONES Seas FORE VOLUME COMPRESSIBILITY, cf 4108, PSI“AT 75% LITHOSTATIC PRESSURE 3 rr a INITIAL POROSITY AT ZERO NET PRESSURE, ¢ 2.58. PorVoume Compressoy at 75% Lithstats Fresco \s Iniial Samote Perosity for Friable Sandstones (aftor Newnan) 7i 48 Technology of Artificial Lift Methods TREONSOLIDATED SANDSTONES mac's: Connevarion PORE VOLUME COMPRESSIBILITY, cf x 10°, PSIAT 75% LITHOSTATIC PRESSURE. rr ) INITIAL POROSITY AT ZERO NET PRESSURE, i Figue 257 Pore-Volume Compress at 75% Lihostte Pressure vs Initial Serpte Porosity for Unconsoidated Sandsiones (ater Newmany> pressure is reservoir depth multiplied by 1 psi/f. From these plots, it can be concluded that Hall's and van der Knaap’s correlations would not provide a good esti- mation: To obtain the best formation compressibility, it should be measured from a sample of reservoir rocks. Knowing Co, Cy Cw, and C; and also the gas, oil, and water sattiration, the total compressibility can be calculated by using Equation 2.132 or 2.183, EXAMPLE PROBLEM #29 (to estimate the decrement of the flow rate vs time at constant flowing pressure) Given data: 3250 psi O14 0.65 0.8807 cp 1,500 ft 708 2.0 X 10° pein Calculate: () The transient period by using Equations 2.128, 2,129, and 2.130. (2) Plot fiow rate vs time for constant flowing pres- sure at 2,000 psi. CONSCLIDATED SANDSTONES PORE VOLUME COMPRESSIBILITY, cj x108, PSIY AT 75% LITHOSTATIC PRESSURE g ‘ WALLS CORRELATION ie y Sethe ‘ose 8 os INITIAL POROSITY AT ZERO NET PRESSURE, $1 Figure 2.58. Pore-Volume Gomoresstiliy at 75% Liosiatc Pressure vs Inbal Sample Porosity for Consakdated Sandstones (ater Newnan) Solution procedure: (Q) The transient period calculation is figured by ap- plying: (a) Equation 2.128 as follows: 14 0.880702.0 x 10-82,50¢ Gi 0.002637 (2) = 21041 br (b) Equation 2.129 as follows: (0.14¥0.880742.0 x 10°61,500)? aa ‘0.00088 @) = 630.50 hr (e) Equation 2.130 as follows: # (@.14X0.8807)2.0 x 10-*1,500 ‘0.001055 (1) = 525.92 hr (2) Plot flow rate vs time (hours) for a constant flow: ing pressure of 2000 psi as follows: 20) = 1S2Sigo10.580700.25851 1.950 — 2.000 soso *[Gumamntimem) -sm9] ‘Simplifying the previous equation yields:© { PSI AT 78% LITHOSTATIC PRESSURE. a Te ‘Baise n= 8548 8 Tog + 434 By varying t from 1 to 700 hours, the plot of flow rate vs time can be made as shown in Figure 2.60. 10 ‘GONSOLIGATED LIMESTONES YAN OER eyaae's Comaeiation 3 1s ao 2s 30a INITIAL, POROSITY AT ZERO NET PRESSURE, $i Figure 259 Pore-Volume Compressbilly at 75% Lithostatc Pressure vs inital Sample Porosty for Limastonas (after PORE VOLUME COMPRESSIBILITY, cy «i Nowand® 1 sy 0}- Boo ue yare mays . ee ol g THE TRANSIENT PERI H 5 yg Orang ESTE | apse =e eGae. & | edb Se 3 T i | Hee col I rot t i Eee eae eae wats 5 PRODUCING TIME, HOURS ‘Figure 2.60. Flow Rata vs Producing Time at Constant Flowing Pros- sure (Example Prosiem 23) Inflow Performance 49 EXAMPLE PROBLEM #30 Given data: Reservoir pressure = 2,000 psia h=50 2 10° psi-t Fw = 0.25 ft aq=0 Catewlate: (2) The time needed to reach the pseude-steady-state period. (2) Prepare the IPR curves for t= 1 day, 2 days, 4 doys, and at t = ts. (@) Prepare the IPR curve at the pseudosteady-state period by using Equation 2.2. Solution procedure: (1) By using Equation 2.128, t,. can be calculated as follows: (0.18)0.42.0 x 10-544,300 0.008657 (5) = 153.81 hours = 6.41 days (2) For t ty che IPR curve can be calculated using Equation 2.2. The results are as follows: Flow rate @ t, days Flowing pressure 2 + he Euaton 22 2,000) o o 0 oO ° 4.800 2s aoe 95189 189 1.200 4290 408390 a7 78 ‘300 83 BI 85 SBT 587 400 ase at? 79 758, 758 0072 gatas 94s 345 ‘The IPR curves are shown in Figure 2.61. CLASS PROBLEM #30a Given data P= 4,790 psia yy =0.4984 cp B,=1.4997 1e=20 md 6 =0.20 n= 80 8 C.=05 x 10° psi-t 11500 fry = 0.33 ft S=0 aq Calculate: (2) The duration of the transient period () Plot the IPR curves for t= 2 days, 4 days, 6 days, and at t= (3) Plot the IPR curve for t = 20 days. CLASS PROBLEM #200 Giver data: P,=2,500 psia $=0.20 Ho = 0.4 ep h=100% — BL=1.25 stb/bbl50 Technology of Artificial Lift Methods 2000 109) 1000] 800] FLOWING PRESSURE, PS! 151 doy 600 1 2doys 400| 200 00405600609 10001200 OWL FLOW RATE, 8/0 Figure 261 IPR Curves at Ditferont Times during Transiont Period re=038f% Gy re=1L50% & 9.5 X 10°F psi! .0 md s eq Cateuiate: (2) The time needed to reach the pseudo-steady-state period (2) Prepare the IPR curves for t = 10 days, 30 days, 50 days, and t = te. (3) Prepare the IPR curve for t = 100 days. CLASS PROBLEM #30 Given data: P, = 3,570 psia k,=0.9 md h=70f By = 1.3567 1760 ft s =0 Calculate: WD ter (2) Prepare the IPR curves for t 80'days, and at t = tin (8) Prepare the IPR curve for t = 150 days. 0 days, 40 days, CLASS PROBLEM #304 Given data: Py = 9,225 psia k,=0.5 md a7 h=90f G.= 2.0 x 10°* ps B.=1.3209 4, = 0.6615 ep re=1860% He =O5 a aq Calculate: D toy (2) Prepare the IPR curves for t= 50 days, 100 days, 200 days, and t = tay (8) Prepare the IPR curve for t = 275 days. 2.9 INFLOW RELATIONSHIPS FOR GAS WELLS Al-Hussainy and Ramey derived the real gas flow equation by solving the partial differential equation of real gas flow, which comes from the combination of the continuity equation and Darcy's Law for radiai fiow.>? ‘The solution of the partial differential equation. after applying the proper boundary conditions and b: taking 147 psi and 60°F as pressure and temperature at standard conditions, respectively, is: m(P,) ~ mPa) = 2887 eT fice (cams) —8.23+0878+a | eras ges flow rate at standard condition, Mscf/éa3 reservoir temperature, °R as permeability of the formation, md hh=thickness of the gas zone, ft roducing time, hours porosity, fraction (ugh = initial viscosity, ep (Co) = initial total compressibility, psim! Tq = wellbore radius, ft ‘S=skin effect, dimensionless a'q=non-Darcy flow term 2.222 x 10"! B yeK bate m(P) = pseudo-pressure function, psi?/op Equation 2.148 can be applied to calculate pressures in a gas well that produces at a constant rate from an infinite reservoir or to calculate the flow rate at constant flowing pressure, including skin effect and the non-Darcy flow term for short” producing times. For long flowing (producing) times, Equation 2.148 becomes: 703 X 10-“ch){m(P,) — Pwo) T Untr/t4)— 0.75 + 5+ a) ‘The pseudo-pressure function, m(P), is defined as fol- lows: (2.149) Ge Pap P)= mm Yad (2.150)a kana ase pressure, psi ressure, psi gas viscosity, ep as compressibility factor, dimensionless A typical plot of (42) vs pressure is shown in Figure 2.62 (the plot is made for a gas gravity = 0.65 and a temperature from 150°F to 225°F) and in Figure 2.63 for a constant temperature (200°F) and gas gravities from 0.60 to 0.75. From those plots, it can be concluded that the values of (1,2) will be constant for pressures between 0 to approximately 1,000 psi. Therefore, in this case, (z,2) can be taken out of the integral, and the pseudo-pressure function can be defined by: nm) aafP® By solving the integral of Equation 2.151, the pseudo- pressure function becomes: usp m(P) P? ~ Py?) (2.152) wher Hi, =the average viscosity, op Z= the average gas compressibility Figura 262 Plot of uzZ vs Pressure (7, = 0.65) Figure 269 Plot of uz vs Pressure for Constant Temperature Inflow Performance 51 Gand Z are evaluated at an average pressure, P, where P is the root mean square pressure defined by: ESET" By substituting Equation 2.152 into Equation 2.149, wwe obtain: (2.183) 703 x 10-® kh (P, Paz (2.134) Ge THe B(In—075+S +049) From Figures 2.62 and 2.63, itis obvious that Bqua- tion 2.154 is valid for pressures below 1,000 psi, but several authors suggest that Equation 2.154 is valid for pressures below 2,500 psi, This is reasonable, since the plot between 2P#/y_Z vs pressure deviates from m{P) vs pressure at pressures greater than 2,500 psi, as shown in Figure 2.64 by the dashed line, yioter cl eae Figure 264 Plot of mP) ve Pressure At high pressure, Figures 2.62 and 2.63 show that the slope of the curves, 4 (y42)/AP, are constant, so: pam onnatant = (2.155) ‘The pseudo-pressure function can be defined by: miP)= Cf aP (2.156) s mP)=CP-P) gasp where C = (2PV/uZh Substituting Equation 2.157 into Equation 2.149 yields: 708 x 10-* kh C1P,— Pas) Tnie/r)- 0.75 +S aq) Equation 2.158 is valid for pressures above 5,000 psi. ‘The plot of 2P/u2Z vs pressure for a gas gravity of 0.70 and a temperature of 200°F is shown in Figure 2.65. The shaded area is the difference of two values of the pseudo-pressure function—that is, m(5,000) — m(4,000)—and it can be calculated by integration. An analytical procedure such as the trapezoidal rule or the composite Simpson's rule for certain values of AP (2.158) cs ‘a52 Technology of Artificial Lift Methods, 1 | -130001-mt40001 - | « WS appa, Por or apy2 ve P Yq + 070, TEMP. 200°F Figure 265 Plot of 2P/u2Z vs Pressure can be used to solve the integration and, of course, is easily handled by computer. By using 0 as the base pressure and epplying the composite Simpson's rule, the value of the pseudo-pres- sure function for some pressures can be calculated and plotted as shown in Figure 2.64 Using a plot such as Figure 2.64, the value of m(P) at certain values of pressure can be obtained, and by applying Equation 2.149, the inflow performance rela- tionship of a gas well can be calculated. ‘A quick estimate of AOFP for most medium-depth gas wells with low permeabilities less than about 7 md can be made with the following equation: hore = 77 & 10-7 kh (PF where: The following assumptions for this approximation have been made in Darcy's flow equation for gas: 703 x 10-* kh (P,2~ 0} Fn ilie—% FS+ aq) Fig= 0.02 op "= 200F ~ e50R Un t/t % +84 a9) = 7.08 ca 210810" ODI oy 20-1 eps (0.02)(660)7.08) For permeabilities of 100 divide by 2 For permeabilities of 600 divide by 4 ‘These are for estimates only and should in no way replace accurate calculations. Equation 2.154 can be rearranged into the following form: Pi Pw? = ba + ag? (2.159 where: px 2424 X 102m sTZ Untey/eu)~ 0.75 + S).y 159 kh 3.186 x 10°28 y,ZT hits 2.161 In the paper by Jones, Blount, and Glaze, the fow rate is given in sofa instead of MSCF/D and hence: 1.424 uZT (in to/te —% +8) kh 8.16 x 10-8 af RHER IO AyeDT 263 Bete Also, we may choose to write the radial Gow equatior as follows: 162 703 x 10-t kh (P?— Pes?) eT Zin tr. —% +S Faq) for q in Mefél ‘The constant is 703 X 10°? for sefé A brief discussion of each term and how it may be obtained follows: (2) k—permeability to gas can be obtained from cor: analysis from either conventional or sidewal h—thickness of producing zone is obtained from well logs. P, = 23.69)— that is, greater than 3.0—so the restrictions,a \, 1 It i i 1 1 { t 1x 10-4 1 1 i 1 1 1 [ I i I A0FP-20.98) rr a 7 ge WMSCE/O Figure 2.76 Plot of (Pt — Pat Vide v5 94 (Examole Problem #29) caused by turbulence effects are occurring in the well. The turbulence effect is normally caused by a lack of sufficient perforations in the well. (©) To improve the productivity of the well, addi- tional perforations are recommended. CLASS PROBLEM #998 Giver Data: four-point gas-well tests: "900 pia Fowg Gas tow 8 Pia ‘ot vectra “08 1 dase ae 3007 2 a0 ae Calculate; Recommend ways to improve this well based on the tests, g i CLASS PROBLEM #330 Given data: four-point gas-well tests: 4,140 psia Flowing Gas tow pressure, psia__ rate, MMsct/¢ 3957 157 3794 23 i 273 386 Inflow Performance 61 Calculate: Recommend ways to improve the productivity ofthis CLASS PROBLEM #90 Given data: four-point gas-well tests: Reson 4470 psia Flowing | Gas fow procsura, psia rate, MMsct?d 4228 1.887 4n0 2.261 4.005 2805, 39st 3297 Calculate: Recommend a way to improve the productivity of the gas well. 2.82 TRANSIENT IPR CURVES FOR GAS WELLS ‘The previous discussion mentioned that Equation 2.148 is the solution of the partial differential equation of real gas fow for “short” periods of fowing time. Equation 2.148 can be used to find the gas flow rate vs pressure relationship during the transient Row pe- ri For pressures below 2,500 psi, where u¢Z can be as- sumed constant, Equation 2.148 becomes: kt ied (sais aa + 0.87/S + a’q)} (2.171) ‘The duration of the transient period can be calcu- lated by applying Equations 2.128, 2.129, and 2.190, as noted in the discussion of the transient IPR curves for oil flow. The following example problem will show how to construct the IPR curve in a gas well during the transient period. EXAMPLE PROBLEM #94 Given data: P,= 5250 psi k=O0.01md He = 023 ep 30 & 0.45 2.0 X 10° psi! 0 reservoir temperature = 200°F Catewiate: (2) The ducation of the transient period and of the late transiant period. (2) Draw the IPR curves at t = 10 days, 48 days, 100 days, and at t = tks = 118 days Solution procedure: (2) Calculate the transient period by using Equation 2.128 and caleulate the late transient periods by using Equations 2.129 and 2.130.62 Technology of Artificial Lift Methods At pressure = 5,250 psi, (jidh = 0.028 ep Using Equation '2.128, the transient period can be'valculated as follows: 12X0.02842.0 x 10-*X2,100)" (0.00263740.00) 1,141.87 hr = 47.56 days Using Equation 2.129, the late transient period ee (03 20.028412.0 x 10°8N2,1008 ft (0.00088)(0.01) = 8,420.21 hr = 142.51 days Based on Ertle’s Equation (Equation 2.190)" the late transient period is: (0.12K0.0284X2.0 x 10-#42,100 (0.001055)0.00) 2,852.88 hr = 118.87 days. (2) The pressure vs flow rate relationship at t = 10 days, 48 days, 100 days, and at the late transient periods can be calculated by using Equation 2.171, and the results are shown in the following, table! Gas Flow Rate, Msct/é tea 00 tae tae Pressure days days Eu 2ng Eg. 2190 250 ° ° 0 ° ° 5,000 432 368 © 370365362 4900021371819 832 w1a.17.89 3000 «9728 «945 ated = 31.60 2000 «S047 4599327421 aa.8e 1900 «58355330 508B 5034 ea79 Ce ee The plot of the IPR curves is shown in Figure 2.77. CLASS PROBLEM #34a Given data: P= 6050 psi 01 md Calculate: (1) The duration of the transient p (2) Plot the IPR curves for t = 2 Gaye 6 days, and at pseudo-steady-state conditions. CLASS PROBLEM #340 Given data: P,= 4/957 psia =: Z= 0.9203 k= 0.25 md 0.0258 h=20 ft 210°F 280, 000 2 4510 days Z 1288 co z 2 $2100 gore 5 zoool g th i 18 gore 1000} \ eee eee °5 10 20° 30 30 30, 60 7 GAS FLOWRATE, MSCF/D Figure 277 Transiont IPR Curves tor Example Problem #34 1,750 ft 1650.72 4479 $= 0.20 4 C= 8.0 10° ps; Calculate: () The duration of the transient period. (2) Plot the IPR curves for t = 10 days, 20 days, and for pseudo-steady-state conditions. CLASS PROBLEM #34¢ Given data: Calculate: (2) The duration of the transient period. (2) Plot the IPR curves for t = 20 days, 40 days, and at pseudo-steady-state conditions. CLASS PROBLEM saad Given data: P,= 4,870 psia py = 0.0260 ¥=004md = Z= 0.9698 T=220°F e068 $= 014 = 2.8 x 10-8 psi-t aCaleulate: (2) The duration of the transient period. (® Plot the IPR curves for t = 30 days, 60 days, and at pseudosteady-state conditions. 24. IPR CURVES FOR FRACTURED WELLS," BY DR. JIM LEA 2.41 DETERMINING THE PERFORMANCE OF TIGHT GAS WELLS Most low-permeability gas wells (k= 0.1, 0.01, 0.001, and 0,0006 md) are more difficult to analyze because the long testing times required make it practically i possible to obtain stabilized test date. Because of this, multipoint well test procedures such as the fow-after- flow or isochronal four-point tests do not give useful information. Since most low-permeability wells are fractured to obtain commercial production rates, the analysis of the tight-gas wells, of necessity, is broken into a pre- frac and a post-frac well testing period. For most low- permeability wells, testing the well before fracturing is necessary to obtain data on the interwell permeabil- ity, skin, and initial reservoir pressure. “These values are routinely found from semi-log anal- ysis or by matching the buildup curve to type curves when the well is exhibiting “radial” dow or is in the transition to radial flow. However, once the well is fractured, the effective wellbore area increases tremen- dously, and it can become impossible to obtain radial flow data, The post-frac testing can be used to deter- mine the increase in effective wellbore ares caused by fracturing if the pre-frae data (primarily k and P,) was obtained. Then, if both pre-ftac data (lk, Pr, and S) and post-frac data (X; and S) are obtained, the user can input a reservoir model with the appropriate data land predict reservoir performance into the future. ‘Type curves may also be used. However, it is neces- sary to have a bottom-hole pressure to make predic- tions with a reservoir model. Many times, this pressure is assumed to be a constant for reservoir studies. ‘The thrust of this section is to determine how to predict gas-well production by considering liquids that accumulate during low-volume gas production. This means that the bottom-hole pressure (BHP) will change with time. The reservoir model (input with the pre- and post-frac data) or a type curve (reservoir model generated) becomes the tool for predicting when liquid loading will become a problem. ‘The use of the reservoir model or type curve is intro- duced later, but the concept is to generate IPR curves by plotting BP vs rate on linear scales and then plot a tubing performance curve on the same plot. The in- tersection of the IPR or reservoir curve with the tubing performance curve indicates where the well will dow. IPR curves for future conditions intersected with tub- ing performance curves will prediet when the well will ‘experience liquid loading problems. (1) Prefrae testing. Pre-frac testing must be done to obtain values for k, S, and P,. The primary reason for testing before fracturing is to obtain data while * Adapted with permission from the Ol! and Gas Journal. Inflow Performance 63 the well is exhibiting radial dow. This will allow the user to perform a semi-log analysis of the data col- lected, which is the only way that c (permeability) of the reservoir can be calculated from transient well test- ing (with the exception of typecurve analysis) while the data is changing into radial fow. The general anal- ysis is as follows: (a) General Procedure for Pre-Frac Testing, (2) Estimate time required for pressure buildup well test. (2) Obtain test data. (8) Make a type curve log AP, log At) to identify domi- nant fow mechanisms during different time periods of the test. (@) If good radial flow data is present, use a semi-log. plot to calculate S, kh, and P*; then calculate P,. © If data are just changing from storage or linear Bow to radial fow, use the appropriate type curve to calculate Kh and S. (©) Dimensionless Type Curves, An alternative to the conventional methods of making a log-log plot is to use a dimensionless type curve for an unfractured well as shown in Figure 2.78 By plotting data on log-log paper of the same sized grid, it is possible to determine when storage ends and also when radial flow begins, Also, it is possible to obtain skin S if an approximate match of the data to the curves is made, The preferred method is to try to calculate the storage first and then make a match to obtain the appropriate skin and the value to (dimensionless time), which corre- sponds toa match point. From the match point, a Ih can be found from the dimensionless pressure. How- ever, since there may be several places on the curve where a match to data can be made, Figure 2.78 is primarily used by some analysts only to determine ‘where radial flow begins; then semi-log analysis is used. 2.42 TYPE CURVE MATCHING® (D) Select the theoretical type curve to be matched based on known well conditions (i., fractured well, unfractured well, constant pressure, constant pro- duction, etc). (2) Using piece of transparent paper, overlay the soale of the theoretical type curve and trace the grid. (3) Blot the data type curve on the overlay using the Figure 2.78 Py vs ty lor Well wih Storage and Skin Effect (after ‘Agenval, Atussairy, and Famey)* (Worcing Copy on 298) 364 — Technology of Artificial Lift Methods underlying grid for sealing purposes. Temporarily ignore the curves plotted on the theoretical type curve: (4) Keeping the horizontal and vertical grid lines on the data type curve parallel to the corresponding grid on the theoretical type curve, move the data type curve until a best match is obtained between the overlying data points and the underlying curvels) (5) Place a match-point mark on the data type curve at any known point (Pp, tn) of the theoretical type curve, Though any point can be used, eycle intersec- tions are convenient, and Pp = tp = 1 is usually the most convenient. (6) Replace the overlay to the position in which the data points were plotted ‘Type curve: Pp to (8) Compute using Figure 2.78 as an example: 1,424 qusZT Pi 1 0,000264kt, toute? Overlay: Ap? = Ap, t kh= (Pp) 6G, where! Ap = (p# — py), 6 = porosity Note: These symbols are standard SPE nomencla- ture; see reference 35 for detailed definitions. 2421 POST-FRAC TESTING Post-frac testing involves testing a well with a tre- mendous increase in apparent wellbore area (Ay) caused bythe large fracture face surfaces: Awa 4Xch where X; = % the tip-to-tip fracture length. For a fractured well, the test may cover times that include linear flow, transition, and radial fow or may last only through transition or possibly may exhibit only linear fiow. If there is no radial flow, a type-curve ‘match (less accurate than semi-log analysis) is required to obtain the kh of the well. To illustrate why a fractured well may exhibit linear flow for a very long time, consider the criterion that tp must be greater than 20 to get to radial fow. Then, for an unfractured well: to dusCirn* asthe) ="0 o0026ak EXAMPLE TO DETERMINE At IN HRS: Given data: t= 20 i= 1.20%, psi (total system compressibility) 03 OL md Solution procedure: then: (20X0.05X0.02K0.5/5,0000.8)%8,600) (0.000264)0.) .0068 hr. for a fractured well, let r. = 500 ft (effective radius after fracturing). Then: Atthr) = 18,96 hrs =2.1 yrs. ‘The above example shows how it may be impossible to get to any radial flow analysis and demonstrates, the necessity of obtaining the permability before the well is fractured, Atthr) = 24.54 sec = 2.43 GENERAL PROCEDURE When a well is fractured, there may or may not be a significant pressure drop along the fracture. A signifi- cant drop will give a low fracture flow capacity, Feu ‘This is determined by type-curve matching the date against Figure 2.73. This curve is for constant well bore pressure, which is near the way most tight gas wells are produced, If there is little pressure drop, the dimensionless fracture flow capacity Fcp will be large, on the order of 500, Dimensionless Fracture flow capacity =k where: = fracture permeability w= fracture width k= formation permeability = fracture % length Ifthe data indicates that there is significant pressure drop in the fracture, the analysis must proceed using this curve to obtain'the fracture length from the toa, Figure 2.79 Log-tog Type Curves lor Fite Capacity Vertical Fractures (Constant Webore Pressure) (after Agarwal, Carte, and Polieck)* (Working Copy on p. 395) Soa & (dimensionless time on the abscissa) and the approxi- mate kh from the 1/qo, which is the dimensionless reciprocal rate on the ordinate. Ifthe match on Figure 2.78 indicates approximate infinite fracture flow capacity, it is possible to use the type curve shown in Figure 2.80 for a fracture of inf- nite capacity. Also, conventional plots suck as the square root plot (Figure 2.81) and the tandem square root plot (Figure 2.82) can be used for infinite capacity fracture data. The tandem square root plot has the advantage that it can be extrapolated to a P*, which will be the initial reservoir pressure P, if the well has always flowed in linear dow. Also, the tandem square root plot will account for short flow times prior to shutin, whereas the square root plot requires an ex- tended stabilized flow. Note that the square root plots allow the user to obtain only the product of aw/k. Inflow Performance 65 (ay = wellbore producing area) and does not uniquely identify the permeability. 2.431 INFLOW CURVES FOR TIGHT-FORMATION GAS WELLS: Because of low-permeability considerations, it is a very different problem to obtain a deliverability expres- sion for a fractured low-permeability well than for a well in which stabilized data can be determined easily. ‘To review briefly, the normal expression for the de- liverability of a gas well is the familiar back-pressure equation: Q= CPP — Pac ® where: rate of fow, Mefd C= anumerical coefiicient characteristic of the par- ticular well P, = shutin reservoir pressure, psia numerical constant characteristic of the well Ifa well has properties such that stabilized data can be collected within a short period of time, the procedure is to plot the data on a log-log plot of (P.? ~ Pec) vs 4. The n value is determined from the slope, and the C value is determined from the horizontal ‘displace- ment of test data on the graph. However, if an excessive amount of time is required to obtain stabilized data, the isochronal or modified isochronal method can be used to test the well and obtain a deliverability expression. The isochronal or modified isochronal well tests involve plotting nonsta- bilized data to obtain the slope (or n) and require at ‘TIS PLOT cooD ONLY FOR INFINITE caPaCtTY FaacTunes- USE TYPE cuRVE TO See IF FRACTURE CAPACITY AEAR Leh BEFORE USING, ale - - | aya 68S amor TeveR ie oe vag guatear vi Vasc 4 (reser tees wel f 2 ai} 2 (CANNOT GET EFFECTIVE WELLBORE AREA Joe Te pnion KNOWLEDGE OF 8) l- ey) F L L L 4 9 Y 2 2 3 $ Figure 281 Typical Linear Flow Plot66 Technology of Artificial Lift Methods CONEAR Foe Tis 108 OF avacrsis 20 (ga tere ne aORE GEA (FHAC-ConaThy SSUES InPLuITEFRacTuae cows TMU FRACTURE FREE SKIN [thy aio Be Fame. faacrune Tree cunie etna To See. act Fon cAPREITY Lo FORE tae one Times e035 #4 INITIAL RESenvOIR PRessune) AUAYS is'1F Fuoweo OT Aa (WuST HAVE A FROW PREAFRACT ANALYSIS) 2 VTTS-Vit, VAR 163.5 @ (MCFD) Ten (*R) S60) VO oa by = slope Figure 282 Example Tandem Square Root Builéup Plot least one stabilized test value to evaluate the C con- stant. For still lower permeability wells, nonstabilized data ccan be collected, and analytical methods can be used to predict a “stabilized” point. This method has limited uuse because the expression obtained is valid only when the stabilization is predicted to occur. However, an MHF (massive hydraulic fracture) low- permeability gas well will exhibit some type of flow because of the fracture (linear or linear with low-frac- ture conductivity) early in the well life and may not reach stabilization for years. ‘To illustrate how to obtain IPR curves for the MHF well, the followifig example is presented using @ type curve to determine fracture conductivity of a given well. Then, as an extension of this example, the type curve will be used to generate a present and future IPR expression. The type curve that is used (Fig. 2.79) in the example is a constant pressure solution for a fractured well as a function of fracture flow capacity.** EXAMPLE PROBLEM #95 (TO PREPARE IPR CURVES): Given date (for type curve analysis and IPR construe tion initial reservoir pressure, pi, psi = 2,555 reservoir temperature, T, *R = 720 formation thickness, h, ft ~ 82 formation permeability, k, md (permeability determined from semi-log analysis of the well before fracturing) = 0.0081 formation porosity, ¢, (fraction of PV)=0.107 total system compressibility, c,, psi-? = 2.34 x 10-* initial gas viscosity, 1. cp= 0.0195 difference between the initial and flowing pressures squared (P,*— Pye), psit, = 6.28 x 10° AFTER.FRAGT: aiticrs, 20 625 38 478 50 408 245» 109 300 208 3.28% 10 150 250 4.00% 10" 250 208 are i 300 172 24% 10" PVT Data Reservoir temperature = 720°R 2, ps! viscosity, P| 2 Factor 147 ore: 4,000 600 0.0146 0972 1.200 0.0383 osse1 4800 0.016 09413 2000 oor7e 9402 2855 0.0195 oes where: fracture flow capacity = kew/kxe racture width, ft fracture permeability, md formation permeability, md fracture half length, ft = tea oteke Foo = iy = dimensionless fracture flow capacit; 1X, Pee 2 khap? Qo TAzquZT where: ‘eciprocal dimensionless rate (ordinate of Figure 2.79) gas viscosity, cp ‘zal gas deviation factor ‘T= reservoir temperature, °R AP? =P? — Po, psia? P,= initial reservoir pressure, psia jottom-hole flowing pressure, psia be Zz ‘The type-curve abscissa is: E 2.634 x 10~*kt (AO x? where = formation porosity, fraction (uC). viscosity-compressibility product at initial condition, ep /psi t= flowing time, hr Solution procedure: ‘The following steps for type-curve matching are re- peated to determine the fracture flow capacity. UW) 1/qys time data are plotted on tracing paper using the log-log scale of the type curves. Main X and ¥ axes also are drawn on the tracing paper. This plot is shown in Figure 2.83. (2) Since formation flow capacity kh is known from a prefracturing buildup test and from semi-log analy- sis for a value of 1/q~0.001, the corresponding value of 1/q5 = 0.11 is computed: 1___khaP? (0.0081\32P.? — Pa?) go 1424quyZT (1.424)(1,000X0.0157)(0.95X720) 910" =2EE I ona = 160 10-42,585:~ 5009=0.11 2 2.2% 10 ww? ae MATCH =] POINT 1 3 wo} +4 0 +_DAYS. 4 r 0 100 wo Figure 289 Type Curve Match (Example Problem) inflow Performance 67 Note that in this example, the 2Z product is evalu- ated at (P, + Pu)/2. To be more correct concerning the variations in 12 caused by steep pressure gra- dients in the reservoir, the m(p) (pseudo-presstre) form of the above equation could be used as opposed to the AP?/j.Z form used here. The pressure form is used here since it will be easier to illustrate the following IPR calculations without a relationship be tween m(P) and P as required when using pseudo- pressure, The position of 1/q = 10°? on the ¥ axis of the tracing paper is fixed in the relationship to 1/q = O.11 on the Y axis of the type-curve graph paper. (8) ‘The tracing paper is moved horizontally along the X axis until a match is obtained. The “match” is shown in Figure 2.84 (® A mateh point is chosen where: t= 100 days = 2,400 hrs toxy= 2.2 X 10? @ Feo = 50 (5) Compute the fracture half length: pu O84 10-9Kt mCi tox (2.694 X 10°°0,0081N2.400)_ © GL07N0.0195)2.34 x 10-9122 x 10° 76,710 ft? X= 690 ft or 2X; (fracture length) = 1,380 ft (6) Compute flow capacity of fracture: ew = Feplkex) = 50 (0.00811,690) 2.44 IPR COMPUTATIONS Now that the fracture length and dow capacity have been defined, it is possible to calculate future IPR ex- pressions. The times at which it is desirable to caleu- late IPR curves are tabulated; then a corresponding ‘Tox, is calculated. From Toxp enter Figure 2.79 and at Boo = 50, and find a corresponding 1/qo: Tox, = 10007 Flowing time (days) 1o3.381%h0 gp thom Fig. 2.79) 20 34 x10 0.108) 1925 402 « 10° 0s 385 8.04 x 10-2 062 730 181 x 10-1 oa 1.095 Bat x to oa 1.480 3:22 x 10" 101 11328 a0 aa With the above constants established, an expression for a deliverability curve at each 1/qp (time) can be found: 1.85 x 107+ Wap ‘The coefficient in the above expression (1.65 X 10-4) contains 2, which isa function of pressure. An approx- imate average value of this quantity is used to obtain (Pe-P.2) ada 68 — Technology of Artificial Lift Methods 10 (EXAMPLE PROBLEM) 22x 10% [| —— a va apa 1 DAY: oars I. La iii we 107 10? 0 10" 1 tg = 2.834 410 tht uc COvxe Figure 2.84 Type Curves for Finite Capacity Fractures: the constant for facilitating the example calculations so that it does not vary slightly as pressures change After substituting the values of 1/qy corresponding to given flowing times, rate vs pressure curves (PR curves) can be found as shown in Figure 2.85 beginning at 20 days and extending to 60 months into the life of the reservoir. Once these IPR curves are calculated, tubing perfor~ mance or'J curves can be plotted on the same graph ane es panto, who ‘Figure 285 Interaction of Reservoir Inflow and Tubing Performance ‘tor Low-Permeatilty Fractured Gas Well to indicate when the well may begin to load with liquid: and cease to produce. The tubing performance curves shown in Figure 2.85 were calculated by using the Gray correlation’? for condensate ges wells with 20 bol MMef input as liquid condensate production. ‘Note that the IPR curves for the example tight gaz well are very steep. This indicates that reducing the flowing bottom-hole pressure at any particular time would not increase the fow rate very much, However, because of the sharp decline in production over the early life of the well, it is obvious that the well can experience severe loading conditions, leadinz to a total reduction in flow rate fairly early if, for in- stance, the tubing is too large. ‘Thus, the transient nature of the IPR curves is more significant than a slight change in bottom-hole pres- sure and the resulting PI (productivity index) effects. ‘These particular results show that a current reduc- tion in flowing bottom-hole pressure (removal of some liquids in the wellbore) will not produce much of a gain in production. If deliquefication methods are not initiated at the appropriate time, the production can drop into a tubing performance area where liquids can continually accumulate and shut off production. Figure 2.86 shows the production that might be ex- pected over the life of the well (data from the example for preparing IPR curves) at a constant bottom-hole pressure (500 psi). But, if the actual case of liquid load- ing is considered, the production is calculated to cease in 2 years with the use of 1.995-in. ID tubing. This is, indicated in Figure 2.86 by observing that the 1.995- in, tubing performance curve, if extended upward and to the left, would barely intersect, or be tangent to, the 24-month IPR curve. Further depletion beyond 24Mow Figure 2.86 Projected Low-Parmgabiliy Fractured Gas-Wall Produc: {ion win Tubing Periormancs Considered months would result in no intersection of the tubing curve with the IPR curve, and no further production, would occur. However, use of the smaller tubing extends the life of the well to beyond 60 months. ‘This shows the importance of considering tubing per- formance when making reservoir projections. A considerable amount of flowing well data is re- quired to first make the required type-curve match and then project reservoir performance. Also, knowledge of the liquid/gas production is required. [f the reservoir Auuids and conditions do not match the conditions under which the type curve (Figure 2.79) was generated, a reservoir simulator can be used to generate future IPR These should be calculated by finding rates at se- lected, fixed bottom-hole pressures as a function of time and plotting curves (Figure 2.85 along values with the same production times. Further, a tubing performance relationship that matches Seld performance must be found to generate tubing performance curves. Smaller tubing may be only a temporary solution, and consideration must be given to other methods of deliquefication such as plunger lift, which might re- quire a larger diameter tubing string. Note that a constant-pressure type curve was used with a tubing curve that was not completely constant pressure with time. In fact, near the loading condition, the tubing curve turns up sharply. [tis felt that, since the tubing curve is fairly flat (constant pressure) before loading, this approach should give a good approxima- tion to reality. However, if the tubing curve changes, sharply with time, computer-generated curves, where each point of fow has the reservoir and tubing solution in agreement, may be necessary to obtain satisfactory results 2.45 CHARACTER OF TUBING CURVES For wells flowing gas and liquid, the tubing perfor- mance curve has a minimum flowing bottom-hole pres- Inflow Performance 69 tioare cas Inst uve aonucTioN Rare = Figure 2.87 Liguid Loading Identified with Tubing and IPA Curves sure as a function of rate, and as such, can cross an IPR curve in two locations such as point A and B (Fig- ure 2.87) Point A can be shown to be a stable flow point be- cause, with slight fuctuations in fow rate, the Gow point ‘tends to return to point A. This is because, at Ax the IPR curve has less available pressure than re- quired by the tubing curve, so flow increases. By a similar argument, the tubing-IPR intersection at B is shown to be unstable. In fact, dry-tubing-IPR intersec- tion to the left of the minimum in the J curve or tubing curve can be shown to be unstabl However, for the steep tight-gas inflow curve, argu- ments similar to the above show that the tight-gas in- flow-tubing curve intersection is stable even though it is to the left of the minimum in the J curve. Point C would have to be reached for unstable conditions with the tight-gas inflow curve. However, based on lim- ited data, it seems that even with steep tight-gas inflow curves dificulties arise if the well is fowed to the lef of the J curve. This may be caused by the slug flow pattern precipitating a more pronounced heading type of flow. ‘SUMMARY ‘The purpose of this chapter has been to show the various procedures used in the construction of IPR. curves for both oil and gas wells. Our objective is to prepare these IPR curves using the best data and proce- dures available. ‘We have not attempted to discuss the more sophisti- cated reservoir simulation techniques. However, if res- ervoir models are readily available, they may be used in place of a less rigorous solution procedure. IPR curves for gas condensate reservoirs and many wells producing from secondary and tertiary recovery projects are good examples of times when more sophis- ticated techniques are needed.70 24 2 26 2. 28, Ea Technology of Artificial Litt Methods REFERENCES Standing, M.8. Vlumac and Phase Behavior of Oi Fels Hyaro- earoon Systems. New York New York: elnhold PUbEShng Comp 4852 Brown K. E. Technology of Aria! Lit Methods. Vol 1. Tus, ‘Oklahoma Penna PuDishing Co. Lasater 1. A "Bunbie Pint Pressure Covrataton ofthe IME (1888), 9.373. Geen, AS. "Pseuostead.State Flow Equation and Procuetity Index tora Wt with Noncrcula Drainage Area” Journal of Petro- toum Tachnolegy Mathews, ©. S. and D. G. Russel, Pressure Buildup and Flow Tests Was. Monogragn Senes. SPE of All Jones, Love G. EM Blount, and O. H, Glaze.” ‘Term Moltipe Rate Flow Tests to Precci Periormanes of Wells Having Turulenee” SPE 6722. SPE of AIME, 1076. Sukarno, Pasa. "Comparison of the Methods ir Preciting Inow Feriormance Rolatonenp Curves” MS. Thesis, Universy of Tusa, 1982, Vogel, J. V.“inlow Performance Relatonstips for Solution Gas Drve Wels." Jouma’ of Pevateum Technology (lsnvary 1968), p. 85-08. Wile, W. T. "Reserva Performance During Two-Phase Flow.” slournal of Potoloum Technology (February 1985), p. 210-246, Ercimoter, James A Personal Communicaton Neeoy, Buca, Unpublished Work. Hasan, M."0il Wot Partrmance in ths Presence ota Gas Phase.” ‘SPE 7762. SPE of ANE, 1979. Standing. M. 8. “Infow Performance Rolationships for Damaged ‘Wels Producing by Solution Gas Dave Reservors." Journel of Pe- ‘roleum Technology (Novarnba" 1970), p. 1399-1460, Hamison, Dave Personal Communication Diae-Catto, Lu Evane and Michaal Galan “Gonotal inflow Perfor. mance Relationship for Soluion Gas Drive Reservar." Journal of Petroloum Technolgy (February 1982), p. 285-268. Fetkovioh, Mu. "The Isechronal Tasting of Ol Wall. SPE 4529, SPE of AINE Eickmeier, James A "How to Accurately Prodiet Future Wol Pro- duciviios™ World Oi (hay 1968), p. 8. Standing i. B. "Concerning the Calculation of inflow Performance 6! Wells Producing trom Solution Gas Drve Reservovs." Journal 9f Petroleum Technology (September 1971), p.1141~1142. ‘Farner, "How Diferont Sze Ges Gap and Pressure Maintenance Progra Affect Amount of Racoverabie O.” OV Week, v. Ws, no. 2 Gare 12,1948), p. S24 Uw 8G. and EM: Blount. "Pivot Point Method Quickly Predicts ‘Well Performance.” Wold OW (May 1982}, p. 182-164 Ere, dm Personal Commuriction Transactions : Trube, Alber §, "Compressibilty of Undersaturated Hydrocarbon Fuads." Trensastions of the AIME, v 210 (1857), p. 941-344 Ramey, HJ." lapid Method of Estimating Reservoir Compressi- bil.” Joumal of Petroleum Technology (Apa 1864), p. 447—458, Trube, Alber S. "Compressiuity of Natura! Gases.” Transactions of the AIME, v. 210 (1957), p. 55-257. Cong, Giordana and Gianluigi Grier, "Salt Content Changes Com- pressiblity of Aasarvoir Sines." Petroleum Engineer (July 196%), p. 3-25 10 8.94 Bodson,C. A. and M. 8. Standing. “Pressure-Volume-Tempereture and Solubility Relatons for Natual-Gas-Water Mixtures.” API Dil tng and Production Practice (1944), p. 173-178. Hall, Howard N. "Compresstiliy of Reservotr Racks.” Transactions Of the AIME, v. 188 (1883), p. 805-311 van der Knap, W, "Noningar Behavior ol Blastc Porous Media” Transactions ofthe AME, v. 216 (1859). g. 175-167, Newman, GH. “Pore-Volume Compressibity of Consotdated, Fri 20, at 22, 33, 9 36, o, 38, 39, 40. 42, 46. 47, 43, 50, 5 52 53 able, and Unconsoldate Reser Rocks Urderyosttic Load ing! Joumal of Porwoum Technotgy Wreburany 1073), 12S. Se ‘AbHusssiny A. and HJ. Ramey, “Appkeaton of Real Ges Flow “Troory t Wel Tasing ane Dolversbty Forecasting” Journal of Pavoloun Technology (ay 1988), Garr Lot a "vecoay of Hyerocarvon Gases under Pressure.” Transactions of the aIMe 1984, p. 2a atin, Car. Petoteum Enginearng, Ding and Well Compioton Englewood Ciifs: Prentice-Hall, 1960, ie i Himmararke, Alok K. "Analysis of Producivty Raducton Ove to Nor Darey Fw and Tae Skin in Grave-Packed Well.” SAE 10084" SPE of NIE, 1801 ee, sim F "Avoid Promatire Liquid Loading in Tight Gas Wells by Using Prefer and Posten Test Osta” Di and Gas Joumal (Gostember 20,1882) p. 129. ‘Agarwal BG. A ArFuseany, nd HJ. Rarny J “An nvestga- {ian ol Welore Storage ané Skin Etec im Unatendy Uauid Flow IC Analyecal Tresiment” SPE douma (Setembe" 1670) 9. 279 230, ‘Agerwal, A. Gy FD, Garter. end C. 8, Polack. “Evaluation and Petomance Prediction of Low Permeabiy Gas Wale Stnulted by Massive Hyrule Freeing" Journal or Pol Technolgy (aren 1975)» 38 Kets, K Let at Handbook of Natural Gas Engineering. Now York, New Yor MeGraw si Book Go, fats 1988 Bar, G.S.and Aut. "The Anaya Of Modifiglsochonal Tests {o Predict the Siatized Oaivertiy of Gas Woe witout Usng ‘Stabtzed Flow Data" SPE 6794, SOE of AME. To76 Gray. FE "Vaca Flow Comrlatonn Ges Wels. in User Marual for API 48 Subsurface Contolad Salty Valve Sting Computer Proorem. Aap. B. dune 1874 Albert Eneray Resources Gansorietion Soard. Teor’ and Prac tee ofthe Testing of Ges Wels 94 od. 1875 Ae, Rand DL. lock "Unsteady State Gas Flon—Use of Draw dawn Data inthe Preicbon of Gas Well Behar” dour of Canadian Petoloum Tactnoley. v2 no. Beal, Carton, "The Vscosty 0! Ar, Wate, Natal Gas, Grice (ll ane its Assocated Gases at i Eau Temperatures ond Pros. Sores" Transactions of te AIME.v. 765 (1500) 9, 38 uonder, MH "The lsoctvona’ Method of Betortinng the Fow Sharacteistcs of Gas Wall.” Transacnons of ne Anis (1955) Bake, UP. Pundernentatof Resenor Engmvering Eeewet Seu ‘ific Publishing Co., 1978. ae Beet. D-N. "Detiminaton of Average Reser Prossur From Bullcua Surveys” Tansevtons of te AINE (1985) Earoughes, Rober . ul Advances In Wel Tost Analysis. 2nd ‘9. Monograph 8. Naw York, New Yor: SE of AE, 1077 Froozabadl A. anc 0. Kai "An Ana of ig Velocy Gas Fiow Trvough Porous Media” Jounal of Patoeum Tecrrogy (February 1975),p 21-216 Groene, W.F. “Anata the Petormance of Gas Watt.” 25th ‘Arrua! Seusrwosiom Paoisum Short Course Processing. is Sock, Texas, 1878. p. 125-186 Havelena, © and A'S. Oden "The Matoral Balance as an Eavation ofa Svaig tne” Transactions of he AME (136) Homer. 0”. "Pressure Suldip i Welt” ard Word Petroum Congress. Sect ip S02-8e Wailer. C. C~ A £ Byes, ane . A, utchnson. “The Estimation ot Permoablity ane Reserv Pressures om Eottorsole rot Sure Butasyp Onaractansicn” Transactone of the SIME (1050 IMuskat M. The Flow of Homogeneous Motte Trough Porous Mé- dia New York, New York: NckrawHl Book Cos 1307 van Evercinga Fan Hurst “The Applicaton of the LaPlaco ‘Transtornaton to Fiow Problems in Reserore=Tanwactons of the AME (1843), 8885 aa& Chapter Multiphase Flow in Pipes by Kermit €. Brown, Zelimir Schmidt, and Dale R. Doty 3.1 INTRODUCTION A brief discussion of multiphase flow in pipes will be given in this chapter. The subject has been thor- oughly covered in Volume 1 and in numerous other references.'## However, a brief review and a discussion of the most important correlations is presented here. The three principal components of the equation for predicting pressure loss in any fluid flow problem are (2) elevation or static component @) friction component (@) acceleration component For vertical and inclined flow, the elevation component is normally the most important. ‘The equation for the How of uids in pipes, which is good for any fluid (single phase or multiphase), and for any angle (upward direction) of dow, (refer to Figure 3.1) states: ‘Total Loss Loss Loss Pressure = Caused by + Caused by + Caused by (3.1) Loss Elevation Friction Acceleration If we take the pressure loss (AP) that occurs over a distance (2), we can write an equation in terms of the pressure gradient commonly used in units of psi/ft: 2h, ap eva Aarer” AZireion CZacenerton The elevation component is taken over the vertical distance only with friction, and acceleration is taken over the entire length. ‘A more specific equation for single-phase flow that represents all terms may be written as follows: ap. 32 where: density velocity pipe diameter = acceleration caused by gravity conversion factor f= friction factor aP/dZ-= pressure gradient ‘m= mixture properties ‘The elevation component for vertical or inclined flow is by far the most important of all three components. For vertical flow, it generally contributes greater than 80% of the total loss and may range from 70 to 98%. Its also the most difficult to properly evaluate because so many variables have an effect on it. [tis the principal component that causes wells to load up and die. One of the easiest ways to visualize its importance is in the loading of gas wells, Many low-flow-rate gas wells, if making some liquid, will eventually accumulate suf- ficient liquid to cause the well to die or flow at very low Bow rates. The prediction of when to expect this to occur in gas wells is quite difficult and complex and is discussed in detail in Chapter 6. On ae ng + £082, BveY o- gopsino + 2% + aes EPPS az «OS i: For multiphase flow, this may be written: eel dP pasin + foxPmV¥m™ + PmVnldVn) (a4) Aeon Be Qed gd) a Figure 3.1. Ganeral Fw Schematici i i 72 Technology of Artificial Lift Methods Gas Loup VALVE CLOSED Figure 9.2 Holdup mustration In order to determine the elevation component, it is necessary to determine the density of the mixture in a static condition. In order to do this, we must be able to determine that fractional part of the pipe oceu- pied by liquid and that fractional part of the pipe occu- pied by gas. Refer to Figure 3.2. That part of the pipe occupied by liquid is called holdup (Hy) and will vary from the top to the bottom of the well, as well as vary- ing with numerous factors such as liquid rate, gas rate, viscosity, etc., as noted in Figure 3.3. Figure 3.3 repre~ sents the Hagedorn and Brown correlation for predict- ing the holdup value.* Once H, is determined, the den- sity can be readily calculated: A RL el CORRELATION BASED ON: [waive sizes: r1m.— 21m VISCOSITIES + 0.86 e>~ 10 ep NoLouP FacTOR /¥ hi 1 ‘ 107 tor ot * [trotor)es~ (95r}e4] 7a Tarn a] Floure 3.9. Holeup Correlation (ater Hagedom and Brown) gm ES Pm = pH) + pel — Hd @s where: Pm = Tif? ‘The gradient in psi/ft is py/144 and represents the elevation gradient component at one set of conditions ‘The friction component can vary from the rather complex correlations of Ros* to the simplified correl=- tions of others such as Hagedorn. Refer to Figure 34 for Hagedorn’s correlation. Hagedorn used a standard Darcy-Weisbach plot but incorporated a two-phase Rey nolds number—that is: OE PT TT TT 09 ee ot 3 “os ER (Waele 4 or os S 05 ol Eos g z é = = 5 os é = 3 © oes 2 2 s Z 90), z Boe : ‘oe 3 z 00s = os SMOOTH PIPES 0002 ..001 i 900.95 008 ‘ooe bt 000,01 WT EE aEe gh EP ESE ge EP EEE ge 090,008, “000,001 Figure 3.4 Fricton-Factor Correlation (ater Hagedorn ane! Brown) TWO-PHASE REYNOLDS NUMBER (Na)xe = 22 x 10°? g5 | Gin mat ue Mixture properties are used to determine Nz.. The most significant one is viscosity. Refer to Volume 1, Technology of Artificial Lift Methods, for more de- tailed discussions on viscosity.* The viscosity of a mix- _ture of liquid and gas has no meaning because the ‘two phases are in reality separated. The most common “method for determining mixture viscosity is to take the average viscosity based on that portion of the pipe occupied by each phase—that is: pes = aol + all — Hd an = (88) and: B= Fop + Feet where: fraction, oil fraction, water L= viscosity, liquid be = viscosity, gas fin = viscosity, mixture holdup However, other methods for determining u,, are used such as the one by Hagedorn where: pe = ey, BO as) None of these methods give a true mixture viscosity. Emulsions are a complete and separate problem, and even after additional research, the correct prediction of pressure loss for emulsified ow leaves much to be desired. ‘A rather serious problem is that viscosity also affects the holdup term. Therefore, true holdup values as a function of viscosity may be questionable. Although the true friction loss and the true elevation loss may not be quite correct, the sum of the tworplus accelera- tion losses may give the correct total loss, which is what we are striving for. Frictional losses may comprise from 1 to 30% of the total pressure loss. Very low rates such as 100,000 scfd of gas in 1996-in. ID tubing have very little frictional loss, but 3,000 b/d of oil with a gas-oil ratio of 1,000 sef/bbl have very high frictional loss in 1,995-in. ID tubing, ‘The third component is the acceleration term, which sometimes is referred to as the kinetic energy term. It constitutes a velocity-squared term and is based on a changing velocity that must occur between various positions in the pipe—that is, vat, In about 98% of the actual field cases, this term approaches zero but can be significant in some instances, showing up to 10% of the total pressure loss. In those cases of low pressure and hence low densities and high gas volumes or high gasoil ratios, a rapid change in velocity is oc- curring and the acceleration component may become significant. It should always be included in any com- puter calculations. Only one final equation is shown here—that is, the one by Hagedorn and Brown. 9652 x 10"d>pq | P™ AZ a9 aP_g- Me peasant Multiphase Flow in Pipes 73 where: Po = pukhs + Pall —HD ap Sem psilte B/ge= 1.00 w= mass flow rate, Iba/day = pipe ID, Ba = average density, tby/ft? a= mixture velocity, ft/sec Refer to Volume 1, Technology of Artificial Lift Meth- ods, by K. E, Brown for complete details and example problems on this and other procedures for calculating pressure losses. Notice that the two principal unknowns in Equation 3.9 are density (pq) and friction factor (f). In order to calculate pm, we must know holdup (Hy), and this is obtained from Figure 3.3, The friction factor is obtained from Figure 3.4. Increments of length or pressure are then assumed and the average gradient determined from which the correct distance between any two pres- sure points is determined. 3.2. VERTICAL MULTIPHASE FLOW For vertical multiphase flow, the general equation reduces to: ap fatavat , pavatvn + faba; Patan 7 Ta P89 gdh NO For Hagedorn’s equation we ean write a(S) ap te SG) Mash = * 5e658x tonaes, Peay AD) Again, the principal unknowns are pm and f. The correlations that are most widely used at the present time are listed as follows: WD for vertical multiphase dow (oil wells) (a) Hagedorn and Brown! &) Duns and Rost (©) Ros and Gray? Modification of original work of Ros and Duns and Ros: (@) Orkiszowski® (e) Beggs and Brill? (0 Aaa ‘There are numerous other correlations, but the pravi- ous six are used more widely than all others. (2) for vertical flow (dry gas wells) (@) Cullender, Smith, and Poettman'? 18) for vertical flow (wet gas wells) (a) Ros and Gray’—available through the API; handles both water and condensate produc: tion (b) Beggs and Brillt—has been found to be good for gas wells with gas-liquid ratios less than 150,000 to 75,000 sef/bbl and does reasonably well for relatively dry gas wells By assuming various flow rates, a "tubing intake” curve can be prepared. Figure 3.5 shows a plot of pres- = e3 i 74 Technology of Artificial Lift Methods sures at the entry of the tubing string vs flow rates. This may also be called a “node outflow” pressure curve. Itis equally important for both oil and gas wells. Its shape will differ depending upon the correlation cased, with the Ros and Gray’ correlation being recom- mended for gas wells making liquid. Figure 3.5 is an extremely important plot because it shows the critical rates below which the flow in the tubing becomes unstable and a loading in the tubing string can be expected. Another manner in which to visualize this is that the gas is slipping past the liquid and therefore the liguid falls back. As the rate de creases, the pressure required at the entry of the tubing continues to increase. For a particular well, the tubing intake curve and inflow performance curve must be analyzed together. The following series of figures shows the significance of the tubing entry or node outflow curve. PRESSURE —= PRESSURE —> RATE —— Figure 8.6 Sabie Flow Figure 3.6 shows a condition of stable flow caused by the IPR and tubing entry curve crossing to the right of the critical point. Figure 8.7 shows a condition whereby the tubing en- try curve crosses the IPR curve at two positions. The ‘well will flow at position 2 to the right of the critical point and will exhibit stable flow at this position. The flow at position 1 will not occur on this well unless the operator purposely places it in this position by choking the well back, and when he does this, he is shifting the position of the tubing entry curve. Note Figure 9.8. Here, the IPR curve is not shifted for 2 particular set of conditions; therefore, we can only shift the tubing entry curve, and this will kill the well. Notice that as the choke size decreases, the wellhead pressure Py» increases and hence the tubing entry pres- STABLE RATE PRESSURE —> RATE —— Figure 3.7 Wail Flowing at Position 2 PRESSURE—= 0 ° RATE —= Figure 8.8 How Choking Shits the Intake Curve and Fiaaly Kits the Wel!(I PRESSURE RATE — Figure 2.9 Dead We wure increases. Finally, upon reaching Pyny the well is dead and is very likely to be dead at Pang Figure 8.9 shows a condition whereby the well is dead because the IPR and tubing entry curve do not intersect. 3.21 THE USE OF VERTICAL MULTIPHASE FLOW GRADIENT CURVES: Refer to Volume 1, Technology of Artificial Lift Methods, pages 158-166, for numerous examples in the use of these curves.t Refer to Volumes 3a and 3b, Technology of Artificial Lift Methods, for numerous gradient curves of all types.'* Only two examples are shown in this section, and they are the same as those shown in Volume 3a. Refer to Figures 3.10 and 3.11 for further explana- tions. 3.3. HORIZONTAL MULTIPHASE FLOW 3.31 INTRODUCTION For horizontal flow, the general equation reduces to: AP _ fapaVn’ , pa¥advn 12) db gd gedL ‘The elevation component drops out since no fluids are to be lifted vertically. Mathematically, sin @ = 0 at zero angle in Equation 3.3. If the acceleration terms are neglected, we have the rather simple equatior @P _ faa aL Peed ‘The principal unknowns in this equation are the den- sity and the friction factor. It is then important to Multiphase Flow in Pipes 75 have a good holdup correlation for horizontal flow as well as a good friction factor correlation. One of the best horizontal multiphase flow correla- tions has been found to be the one of Beggs and Brill It covers the entire range of rates and pipe sizes and was originally intended to be good for any angle of flow as well as for vertical flow. Field experience has shown it to be one of the best for horizontal fow. ‘Some of the other correlations that are quite often used are those of Dukler,? Eaton et al,,"° Lockhart and Martinelli," and Baker. Baker's correlation was used almost exclusively in the past by numerous com- panies and is still used with some modifications. Vari- ous modifications of the Lockhart and Martinelli cor- relation are also used by several companies. ‘One of the best holdup correlations was the one pro- posed by Eaton et al." (Figure 3.12), Excellent results, have been obtained by using Eaton's holdup correlation, in conjunction with Dukler’s correlation—that is, using Dukler’s procedure but substituting Eaton’s holdup correlation for the one proposed by Dukler. 3.92 THE USE OF HORIZONTAL MULTIPHASE FLOW GRADIENT CURVES Refer to Volume 1, Technology of Artificial Lift Methods, pages 191-196, for numerous examples in the use of horizontal multiphase flow gradient curves.! Only one example is shown here, and it is the same one shown in Volume 3a. Refer to Figure 3.13 and note that, if the separation ssure is known, the wellhead pressure can be deter- mined and vice versa. 3.4 INCLINED FLOW 3.41 INTRODUCTION ‘There are numerous cases whereby the tubing string deviates from the vertical and/or where the flow line deviates from the horizontal. This causes a more difi- cult solution to the multiphase flow problem. There have been numerous attempts to solve the inclined flow problems, but there is still much work that needs, to be done in resolving this problem. The most recent, ‘work of Beggs and Brill? appears to offer the most prom- ise, Beggs and Brill conducted laboratory experimental, work in small pipes for various angles of flow (both upward and downward from the horizontal). Our expe- rience has shown that they offer the best solution for inclined flow lines. Their correlation, along with spe- cific examples, can be found in reference ‘The Flanigan correlation also is used quite often for inclined fow—in particular for larger lines and gas, transmission lines. Complete details, along with exam- ple problems, can be found in reference 1. 3.42 DIRECTIONALLY DRILLED WELLS More and more directional wells are being drilled from offshore platforms and even from numerous land wells in the Arctic and in other places such as man- made islands. Some companies are suggesting that we drill more wells from central locations, thereby reduc- ing overall costs, Angles greater that 60° from the ver- tical are fairly common, and angles greater than 70°76 — Technology of Artificial Lift Methods 160 PSI PRESSURE, 100 PSI 3,360 PSI 8 16 24 32 40 48 56 A a VERTICAL FLOWING PRESSURE GRADIENTS (OL PERCENT = 50) TUBING SIZE = 24¢7N. 1D PRODUCTION RATE = 3,000 8L/0 GAS SPECIFIC GRAVITY = 0.65, AVERAGE FLOWING TEMPERATURE = 150"F OO1L API GRAVITY = 35.0 API WATER SPECIFIC GRAVITY = 1.07 ses GIVEN DATA, ‘2s IN. OD TUBING (2.4474N. 1D) LIQUID FLOW RATE = 1.000 6/0 (50% WATER) DEPTH » 12.000 FT PAODUGING GOR = 800 ScF/aSL PRODUGING GLA = 800/2 = 400 SCF/BBL WELLHEAD PRESSURE = 160 Pst FIND THE FLOWING BOTTOM-HOLE PRESSURE ‘SOLUTION: “ 1. Fine the equivalent depth corresponding to 160 i wellhead pressure. To do this, proceed vertically downward from 60 pei at zero depth Lunt! intersecting the 400-sct/obI GLR line. This is a1 a depth of 1400 tt. Note the pressure scale Is in BO-psi inerements and tne Gopth ecale in 200-H incomonts [Add the equivalent depth of 1.400 ft to th of 12,000 f1 and obtain 13.400 rom 19.400 ft on the vertical ecale. proce: horizontally to the 400 scl/bbl line and read flowing aressure 0! 3.360 ps) 7 o 3 13 0001 “Hidag 8 Figure 3110 How to Find tne Flowing Botiom-Hole Pressurea Multiphase Flow in Pipes 77 3,800 PSI 320 PSI PRESSURE, 100 PSI 24 32 40 56 i i VERTICAL FLOWING PRESSURE GRADIENTS (OIL PERCENT = 50) TUBING SIZE = 2.4415N. ID PRODUCTION RATE = 1.000 BLO GAS SPECIFIC GRAVITY = 0,65 AVERAGE FLOWING TEMPERATURE = 150°F OWL API GRAVITY = 35.0 APL WATER SPEGIFIC GRAVITY = 1.07 2.600 FT GIVEN DATA: TUBING 1D = 2.41 QUiO FLOW Rar: DEPTH = 12.000 FT GOR = 800 SCFBaL GLA = 400 ScF/BBL STATIC PRESSURE = 3.000 PSI PRODUCTIVITY INDEX = 5 8/0 PSI [ASSUME CONSTANT) ‘OWING WELLHEAD PRESSURE. 1.000 8/0 (S0"» WATER) o FIND 7 SOLUTION: 1. Caloulate the flowing battom-note pressure or obtain it from an inflow performance curve. For a constant Pl Py =P, - & = 4,000 - 1.000 ~ 3.800 psi a 5 2, Locate 3.800 psi on the 400:set/DpI line and note a depth of 14.600 ft, 3, Subtract the woll depth of 12.000 ft ftom the death of 14,600 ft to obtain 2.800 ft, 4. From 2.600 ft on the graph, procaaa norizontally to the 400-5c1/b0I line and read 920 ps. This is tne permissible wellhead pressure, 44 000't ‘Hid3G Ss 8 50 14,600 FT| 400 200 i 300 : 400 i epee se 2 3 i Seas sssss ° eee Figure 9.11 How to Find the Flowing Wellhead Prassure| | 78 — Technology of Artificial Lift Methods WATER-GAS DATA 2" PIPE ~ 50-2500 BPD 4" PIPE - 50-5500 BPD 0-10 MM SCF/DAY LIQUID HOLDUP - FRACTION 001 O0r or 19 10 we (yt Leg (oft) Yor (Fe) oer (Tos owe 312 un Coreaton tr HonzontalFow ater Eaton or have been drilled. However, additional careful plan- ning should be done in extending these wells too far from the vertical. For example, some wells with an overall deviation of between 60-70° can be expected to produce 30-35% less than a vertical well to the same formation for the same tubing size. In addition, artif- cial lift problems are increased drastically, with some lift methods being ruled out completely. For good plan- ning, the production engineer, the reservoir engineer, and the drilling engineer should get together in estab- lishing drilling and completion programs. Highly devi- ated wells should be analyzed to see if the objective flow rate can be obtained with the planned casing and tubing sizes under consideration. 3.421 CORRELATIONS FOR DIRECTIONALLY DRILLED WELLS It is recommended that the multiphase flow correla- tion be used in the form of Equation 3.4: @P finPinVe® . PmV¥mdVm Tay e/edon sing + bate. Pavan (eatin) (tees) neato! where: Pm = pHa + pall — HD) 3) f= mixture energy loss factor scan be seen, the elevation component is converted to vertical distance only, whereas friction and accelera~ tion ere taken over the entire length of the piping system. The reliability of this method depends mostly upon the correct determination of Fi. Reasonable reliability has been obtained by using a vertical flow holdup correlation such as Figure 33 as proposed by Hagedorn. This can be expected to be reliable up to about 35-40° deviation from the vertical after which the holdup values for directional fow begir. to deviate from vertical flow values. Hagedorn’s eque: tion would then appear as noted in Equation 3.11, aP fee Bon TREO : 2(2) (an faa Other correlations such as those of Orkiseewski Ros, etz, could be handled in the same manner. The greatest error comes from assuming that the holdus Gy is the same for directional fow and vertical ow "The next best approach to the problem isto use the correlation of Beggs and Brill, which gives a holduy value for each angle of flow.* This should be the bes procedure to use, However, when using the publishec work of Beggs and Brill, © appears that most of the calculated pressure losses will be higher than the truc values. In other words, lower rates will be predictec than obtainable when using this correlation, butt! always consistent ‘The improved correlation of Beggs and Brill give: good results. Figure 8.14 shows the effect of the anzlc of deviation on the Sowing pressuresraverse. Figure 3.15 shows the effect of the angle of deviation on fox rate for a gas it well sa eae Epa sind + a eM 3.43 MULTIPHASE FLOW CORRELATIONS FOR INCLINED FLOW LINES The best correlation for inclined multiphase flow it flow lines appears to be the one by Beggs and Brill” It covers the complete range of angles of inclination. ~ including both upward and downward flow, It also ap. pears to handle fow in larger transmission lines ‘well as condensate flow when an adequate composi tional model is used. Adequate example problems can *, be found in Volume 1, Technology of Artificial Lift © Methods, along with 2 complete discussion of their correlation.” # The correlation that has been most widely used in past years has been the one of Flanigan,’ and it is still used extensively by several companies althougi: several modifications may be included. It is probably the simplest solution to the inclined flow problem and may be used in conjunction with any good horizontal flow correlation such as the one by Beggs and Brill? Dukler* ete. A holdup correction factor is obtained based on superficial gas velocity, which accounts for both upslope energy loss and downslope recovery in one factor, Ps 3.5 FLOW-THROUGH RESTRICTIONS 3.51 SURFACE CHOKES 3.511 MULTIPHASE FLOW THROUGH A CHOKE ‘The equation most commonly used for multiphase ¢ flow through surface chokes is the one by Gilbert, = which was based on field datasi? & 485 (GLRYs Pee ed (3)Multiphase Flow in Pipes 490 PSI 100 PSI PRESSURE, 100 PSI 12 16 HORIZONTAL FLOWING PRESSURE GRADIENTS (ALL OIL) FLOW-LINE SIZE = 2.5004N. 1D PRODUCTION AATE = 1,500 BL/s GAS SPECIFIC GRAVITY = 0.65, AVERAGE FLOWING TEMPERATURE = 120°F ‘OWL API GRAVITY = 35.0 API WATER SPECIFIC GRAVITY = 1.07 6,000 FT 6,600 FT 8 GIVEN DATA: 2H:N. 1D FLOW LINE LENGTH = 6.000 FT RATE = 1,500 B/D (100% O1L) GOR = 800 Scr/a6t. SEPARATOR PRESSURE = 100 PSI FIND THE REQUIRED WELLHEAD PRESSURE, SOLUTION: 1. Find the equivalent length corcasponding to 100 ps! separator prossure. Locate 100 pi at zera length and proceed vertically downward unt intersecting the 800-sct/bb} gas-il ratio lina ata length of 600 ft. Add the equivalent length of 800 ft to the fine length ‘6.000 tt ane obtain 6,500 ft From 6,800 ft proceed horizontally to the 800-sc!/bb! line and read a flowing wellhead pressure of 490 psi 14 000°! ‘HLON|T © gay &42eec5 o o 833 §888onq 3 § ‘ S88 Ss8888sS 8 3 3 sss 3 3 8 Fro 2.13 Henaonal sutphase Flow to Find Flowing Wolleed Pressure 7980 Technology of Artificial Lift Methods PARISON OF PROJECTED VERTICAL oe PRESSURE TRAVERSES 9" 10008/0 ° 7200) 26.00 PRESSURE, PSI 36.00 Figure 6.14 Comparison of Projected Vericel Pressure Travetses wellhead pressure (psia) gas-liquid ratio (Mef/bbl) flow rate (b/d) d=choke diameter (64ths of an inch) 3.512 GAS FLOW THROUGH A CHOKE Texas A & I University published research work in 1946 that it had conducted concerning gas fow through } the Thornhill-Craven standard 6-in-long positive flow bean.!* The flow coefficients were found to be: Tenia Sake ~~~ wage ton Average dosage ae Soadens¢ comico * es oreoa x ea anes iar onset : : aa oases | % Ses oanee we mse Sasrs * ao aera * se aise x eae asst Note that the discharge coefficients become essentially constant at 0.880 for sizes greater than %, in. ‘The above coefficients are used in the equations: cP. VyeT PRODUCTION RATES POSSIBLE FOR VARIOUS DEVIATION ANGLES 3 7 oeumnont 2-577 a PRESSURE fmunerads) of PS) Hopes Pbawance 7 = 7 © RATE thuncreds of 8/0) Figure 8.18 Production Rates Possible for Various Deviation Angles and: 605.97 Ce Py = DEST CIA Ps (3.141 Vint ; coefficient of discharge pressure upstream of choke, psia A= area of choke, in? specific gravity of gas emperature (absolute) ‘The basic formula for gas flow through an orifice is as follows: 155.50 A Py 2g = 7 RN RA Vet (3.251 where: g = acceleration of gravity K =C,/C, = ratio of specific heats at constant pres sure and constant volume Pp = downstream pressure (psia) R= ratio Pp/Py2R, R, = critical flow pressure ratio 2 \KiR-» =e) Using air as a basis: and: @.16) ‘This equation is valid as long as the downstream pressure is less than approximately one-half of the up- stream pressure, & RR3.8. TWO-PHASE FLOW IN PIPELINES, BY DA. ZELIMIR SCHMIDT AND DR. DALE R. DOTY 3.81 INTRODUCTION ‘There are numerous production systems whereby pressure losses as well as heading and slugging in pipe- lines become very important, Large production lines such as those in the Prudhoe Bay area or ocean-floor ~eompletion lines such as those in the North Sea offer specific examples of large lines for two-phase flow. ‘The pigging of these lines may also become impor- tant, and calculations for pigging become critical. ‘The effects of riser pipes on offshore platforms be- come critical in predicting slugging in horizontal lines. Calculations as to slug sizes and frequency are critical in designing separation facilities to handle the large separated slugs of gas and liquid. ‘This section presents a discussion and a manner of solution to this problem. 3.62 STEADY-STATE PRESSURE LOSS CALCULATIONS ‘The prediction of pressure gradients, liquid holdups, and flow patterns for pipelines that are transporting both a gas and liquid phase is a common problem in the petroleum industry. Such two-phase flow occurs frequently in well tubing and in Gow lines. The Bow can be vertical, inclined, or horizontal, and the result- ing interaction between phases can create a number of different possible dow patterns. Figure 3.16 displays the commonly accepted two-phase dow patterns. Flow pattern, along with pipe inclination, liquid holdup, Buid composition, flow rates, and duid physical proper- ties, determines the nature of the pressure loss along the length of the pipeline. Because the exact effect of each of these factors in determining the pressure loss is not theoretically known, physical correlations com- bined with the pressure-loss equation are required to solve practical problems. The results that follow are also valid for single-phase dow. (2) Steady-State Continuity Equation. The theoreti- cal basis for determining pressure loss results from combining the physical laws of conservation of mass and momentum. Under steady-state conditions, the conservation of mass for a pipeline segment of length AL (See Figure 3.17) can be expressed as: Rate of mass transport across position 1 = rate of mass transport across position 2, In terms of the two-phase density (p) and the two-phase velocity (v) at positions 1 and 2 the conservation of mass can be expressed as: PAY = pA ve in where it is assumed that p and v are constant across the cross-sectional area (A) of the pipe. Dividing this equation by AL and taking the limit as AL goes to zero yields the steady-state continuity equation: aipv) aL (2) Steady-State Pressure Loss Equation. Under steady-state conditions, the momentum in a pipe seg Ga) g Q Multiphase Flow in Pipes 81 SEGREGATED FLOW ANNULAR INTERMITTENT FLOW per SLUG) DISTRIBUTIVE FLOW Fours 3.16 ‘Scrematc Diagram of Possible Flow Patterns in Two- Phase Pipaines ard82 Technology of Artificial Lift Methods. Poiva+Po {3 Nay Foure 3.17 Flow in a Ppe Segment ment of length AL does not change; therefore, the con- servation of the linear momentum can be expressed as: Rate of momentum transport across position 1 ~ Rate of momentum transport across position 2 + the sum of the forces acting on the fluid = 0 In terms‘of the variables defined in Figure 3.17, this ‘equation can be expressed as: (@ wy ke Ata PLA ~EPsin 6A aL — (rd AL)=0 (3.19) where 7 is the two-phase wall shear stress. Dividing this equation by AL and A and taking the limit as AL goes to zero yields: -£ (C2)-Z-Losino Fao 20) ag) ae A Differentiating p v? and using the continuity equation Bives: a __pvav_g (ae 2 oa eeaine (5) 20) It is traditional to view this equation as a pressure loss equation. Therefore, the total pressure loss is con- sidered to be composed of acceleration, elevation, and friction components: (iE) n” (tn CE) mtr * (Ea ‘This equation is more difficult to apply than it would appear from a first examination. The difficulty arises from the fact that. both the two-phase velocity and den- sity are not truly constant across the pipe cross section. ‘Also, the two-phase velocity and density are compli- cated by the fact that, for most flow patterns, the gas and the liguid phases are not moving with the same actual velocity. This “slippage” between phases implies that, in addition to the gas and liquid mass fow rates, a knowledge of the liquid and gas holdups is required to determine the two-phase velocity and density. This, along with additional information, must be supplied by external correlations, (3) Two-phase correlations. Use of the pressure loss equation requires a determination of two-phase veloc- ity, density, and wall shear stress, In order to simplify the wall shear stress term 7, it is traditional to intro- duce a dimensionless Moody friction factor (f) defined as the ratio of four times the two-phase wall shear to the two-phase kinetic energy per unit volume: pa 22) eran & In terms of the two-phase friction factor, the pressure Joss equation can be expressed asi (tere aie The problem of calculating the two-phase pressure loss has been reduced to integrating the pressure loss along the length of the pipeline. This can be accom- plished only when proper values can be assigned to p. v, and f. As observed earlier, the definition of 9, v, and f for each portion of the pipeline depends upon the steady-state gas and liquid flow rates, the nature of the flow pattern, liquid holdup, and the fluid physica} properties for each phase. It is the purpose of two-phase flow correlations to define the relationship between these variables ‘The two-phase flow correlations commonly used are the Dukler-Eaton” and the Beggs and Brill? correla- tions. The Dukler-Eaton correlation was developed for horizontal flow in pipelines and gives reasonably gooé results for pressure loss and liquid holdup predictions However, the Dukler-Eaton correlation does not 4. rectly take the flow pattern into consideration. ‘The Boggs and Brill correlation removes this shor'- coming by including the flow pattern in the pressure drop and liquid holdup calculations. Moreover, the cor- relation was developed for any pipeline inclination. When compared to the Dukler-Eaton correlation, the ‘Beggs and Brill correlation tends to slightly underpre- dict, pressure losses, and the Dukler-Eaton tends to slightly overpredict pressure losses. It is very important to accurately define the ow pattern, since the pressure less calculation is flow- regime dependent. The different flow patterns arise from different physical distributions of gas and liquid in the pipeline. A schematic diagram of possible flow patterns for pipelines is showm in Figure 8.16. Various, “maps” are available for predicting the flow pattern from the gas and liquid flow rates, the pipeline inclina- tion, and the fluid physical properties, and each pres- sure loss correlation uses its own map. The Mandhane et al!” map shown in Figure 3.18 is probably the most, widely used map for predicting fow pattern because flow pattern can be predicted from a knowledge of the gas and liquid flow rates. The pipeline outlet flow pat- tern under steady-state flow is of interest, since slug flow can require special liquid storage facilities. The flow-pattern maps commonly used are the Mandhane map, the Beggs map,’ and the Mukherjee map The use of more than one map serves as a method of cross- referencing. : zi 4 Waa oobe L n n Figure 3.18 Flow Patter Map for Two-Phase Ppalines (altar Mand nana at a)® 9.83. LIQUID FLOW BEHAVIOR PREDICTIONS. Calculations have been performed for two separate cases that can result in liquid slugs fowing out of the pipeline, ‘These include normal slug flow and pigging. ‘The calculations for each case are different and are given in the following sections. (D) Normal stug flow. When stead: ina pipeline, time-averaged gas and liquid mass flow rates are constant at the inlet and outlet of the pipe- line. If slug flow occurs, instantaneous flow rates and pressures are not constant, and liquid slugs and follow- ing gas bubbles vary in velocity and length in a random way. Proper design of separators must account for ran- dom slug flow and be based on the flow characteristics of slugs. Figure 3.19 shows a schematic diagram of a slug ow model in a horizontal pipe. The model is com- posed ofa gas bubble and a liquid slug containing many small bubbles. The hydrodynamic model was developed using the following assumptions: (1) Small gas bubbles and liquid in the liquid slug travel with the same velocity. 12) Liquid film does not contain any gas bubbles. (3) Negligible liquid droplets occur in the gas bubble. To develop a relationship among variables, an overall mass balance for the liquid in the slug unit of length. tL, + Ly in Figure 3.19 was used. The mass of liquid that flows out of the pipe during one period T is the sum of the mass of the liquid in the liquid slug m, and in the liquid film mt. However, as the slug flows out of the pipe with a velocity vq, the gas bubble follow- ing the liquid slug overruns or bypasses part of the liquid slug. The portion of liquid that is overrun will not flow out with the slug. The mass balance can be expressed as: my = my, + My ~ ML, 20 where: Multiphase Flow in Pipes 83 RIOD LIQUID SLUG RESIDENCE TIME GAS BUBBLE AESIDENCE TIME Lj = Liguio SLUG LENGTH = GAS UBBLE LENGTH V, = Gas veLociTY Vi = UQUIO Fit VELOcITY Gas SUBBLE VELOCITY MIXTURE VELOCITY HL, = UQUID HOLOUP IN LIQUID SLUG. Hu = LIQUID HOLOUP IN GAS BUBBLE Figure 3.19 Schomate ofan idealized Liqud Sug and Gas Subbie— Siug Unie. my, =vaT Ap mu voiTs A Hs Pe me = YorTy A Hes pr muy = (voc Ya) TA Hh pu ‘Substituting the mass equations into Equation 3.24 and solving for vs, gives: ty van Vor Hoa + Vor ap Has ~ (er Ya) Hs (6:25) ‘A similar development for mass conservation of gas phase results in: Mg = Mey + Ms — Mee (3.26) where: My = Ver TA pe mp = vu Te AC — Hes) pe mp = ve T,A(L— Hu) pe imige = (ac Va) TAC = Hla Pp ‘Substituting the mass equations into Equation 3.26 and solving for vs, gives: toe — Hy) + vu = HD ml in @2n = (ve) (1 ~ Hes) Consider a coordinate system that travels in the pipe with the velocity of the nose of the gas bubble, vir. Continuity considerations on the gas phase across the liquid slug and the gas bubble result in: fa) A (1 — Hs Pe = (Yor Ve) ACL Hi) pe (3.28) Solving for v, gives: vee Hs Ba Ve= Ver (oe Ve) (3.29) 56 i i 84 — Technology of Artificial Lift Methods A similar development for continuity consideration on the liguid phase results in: (Wor — Va) A Bus p= (Vor ~ Vo A Has pr (3.30) Solving for the liquid film velocity gives: V1 Wn Yad BE + vp sp Ho The time average liquid holdup, Hy, can be expressed in terms of the liquid holdup and the period of the liquid slug and gas bubble, Thus: HL T=Hu T+ Hy Te (3.32) Combining Equations 3.92 and 3.25 and solving for ver gives: Yan Ve Hue wn (3.33) Variables vp, ve, and vor are uniquely defined with Equations 3.28, 8.31, and 3.33. Combining Equations 8.25 and 8,32 results in an equation that does not con- tain Ty or T,, Thus, a new equation must be developed for either Ty or Ts. The average liquid slug length is related to the liquid slug residence time or period by: L Ts (3.34) Equations 3.27, 8.29, 3.31, 3.93, and 3.34 must be solved to obtain values of vs, Ve, Vig, Ts, and Ty, To solve these equations, it is necessary to have empirical correlations for Hy, Hi, Hy and L,, Experimental data were availa- ble from two Tulsa University test facilities. The tests, were conducted in 2-in. pipe with the Piping Compo- nents Facility and in 1'4-in. pipe with the Inclined Flow Facility. Kerosene and air were used for fluids, and liquid holdup was measured with @ capacitance sensor and associated electronics. Data collected in- cluded one hundred tests for H, and approximately fifty tests each for Hi. and His, The following empirical correlations were developed using Va and Vs, as inde- pendent variables: Hu = LO — 0.01 exp [a + bin vig + ett ‘} (3.35) where: a= 4.47108 ~ 0.13691 va. b= 0.05831 + 0.08070 va. 0.02124 — 0.01169 va, Hy=10—O0lexpla+binviete(in ve] (8.36) where: a= —0.52728 +0.49899 vp, b= 2.01451 ~ 0.17878 vs, = 0.20271 + 0.01819 va. H.=10-O0lexplatbinva+c(inys?] (8.87) where: a= 4.27148 ~ 0.26172 va, b= 0.08495 + 0.12992 va. = 0.00406 ~ 0.01826 vat An empirical correlation for the average slug length ‘was developed using data from the Prudhoe Bay field of Alaska The data were collected on 16- and 12-in, ‘two-phase pipelines. The slug length was found to be Jog-normally distributed with a homogeneous variance of 0.3. A linear regression analysis of experimental average liquid slug lengths resulted in the following correlation. Le exp [2.668 + 5.441 (in d)** +0.059 in vq) (9.38) During production of a liquid slug from a pipeline. the liquid slug and film are continuously overrun by the faster moving gas bubble. Thus, not all of the slug Jength in the pipeline is produced. The total liquid vol- ‘ume produced from the pipeline in time T is defined by: VuawwTA (3.39) Distribution of this liquid is not uniform in the pipe, and part comes from the liquid slug and part from the liquid film beneath the gas bubble. The liquid pro- duced from the liquid slug during time T is: Vir = va Ts AH (3.40) ‘The liquid produced from the liquid film during time ‘T, is the difference between Vip and Viz, and can also be calculated from: Vie =v Te A Bas a1 ‘The volume of the liquid slug in the pipeline, which is not totally produced because of gas-bubble overran, is defined by: Vu= ve Ts A Hy (3.42) The difference between Vis and Vier will not be pro- duced with the slug. The total liguid in a slug unit Giguid slug and ges bubble) shown in Figure 5.39 is given as: Vir ve TyA Hist va TsAHi: (6.45 ‘This volume is greater than Viy because of the overrun- ning phenomenon. (2) Pigging, When ges and liquid flow into a pipe- line, density differences tend to separate the phases. ‘The gas flows more rapidly, leaving the liquid to Bow along the bottom of the pipe at a much lower velocity, ‘The Liquid is moved because of the viscous drag of the gas on the free liquid surface Liquid holdup in a pipeline is primarily e function of liquid and gas volumetric flow rates and pipeline configuration. The liquid holdup contained in a pipe- line, for @ constant liquid flow rate, decreases rapidly with increasing gas flow rates, At higher gas flow rates when annular or mist flow prevails, liquid holdup ap- proaches asymptotically a value of the no-siip holdup. ‘The liquid holdup always increases with an increase in liquid flow rate. When gas and liquid flow in hilly terrain pipelines, liquid holdup increases in the uphill sections and decreases in the downhill sections. As a result of this, liquid tends to accumulate in the low spots or valleys of the pipeline. ‘The pressure drop in a pipeline increases as liquid holdup increases, The pipeline efficiency is therefore inversely proportional to the liquid holdup. In order to increase the efficiency of a pipeline, liquid holdup must be reduced. This can be accomplished by pigging the line at regular intervals using rubber spheres in- fiated to the desired diameter. The major variable dur-ing pigging is the volume of liquid that will be removed by a sphere. Liquid separation or storage volume at the sphere receiving end must be capable of handling the maximum size of slug. The velocity with which the liquid slug exits the pipe is assumed to be that of the sphere. Assuming 100% removal efficiency of a sphere and that all liquid removed by the sphere is in the form of a continuous slug, the volume can, ~be estimated by integrating the difference between the “Higuid holdup and the no-slip holdup in the pipeline, ‘This additional liquid volume to be handled by slug catchers can be estimated from: vena fa val) EX 4a) =AS (Bax ‘Thus, the pipeline is divided into n calculation length inerements, and average liquid holdup and noslip holdup are determined for each increment. Surnming the liquid volumes in all calculation increments using Equation 3.44 yields an estimate of the total liquid volume swept by a sphere. ‘A more rigorous estimate of the liquid slug volume can be obtained by developing an expression that per. forms a volumetric balance at the time the slug formed by the sphere first reaches the outlet end of the pipe- line. The volume of liquid swept from the inlet to the sphere location, X, must be equal to the void volume ahead of the sphere that would exist, assuming steady. state flow without spheres. Thus: Vu! Fn AX) Afign-rvax=afa-Hyax 48) Equation 9.45 would require a numerical solution. However, some simplifications can be made. Compari- son of Hi and Az values from inlet to outlet in steady- state flow shows thet very little change occurs in either variable. Equation 3.45 then becomes: AH. - WX =AGd-HYL-X) Algebraic simplification yields an expression for slug lengeh: Hyak =u-x=1 [BSS] Coraparison with Equation 3.44 suggests that « more accurate estimate can be obtained from: 5614 V. Adhd (3.46) where: ‘Vi. = liquid volume, bbl Av is based on outlet conditions ‘The time required to produce the liquid slug can be calculated from: T= & gan ‘The time required for the liquid slug to reach the outlet after introducing a sphere at the inlet is: Multiphase Flow in Pipes 85 Lol, The assumption of 100% removal efficiency is not possible in a pipeline. As the sphere travels, some amount of liquid will be bypassed. The amount of liquid slipping past the sphere will affect the velocity of the sphere, the outlet liquid velocity, and the slug length. T (3.48) ‘SUMMARY ‘There are numerous areas whereby additional study is needed in the area of multiphase flow. Ernulsions are a very difficult problem, and the pre- diction of when emulsions can be expected to form is questionable. Once an emulsion exists, the prediction of pressure loss becomes more complex and question- able. Present-day correlations also give trouble in han- dling viscous crudes. The principal correlations availa- ble fail to predict accurate results and give widely vary- ing answers. ‘The prediction of well loading also leaves much to be desired, with the Ros and Gray correlation appear- ing reasonably good for condensate wells that may also produce some water. The principal multiphase flow cor- relations give widely varying results. REFERENCES 1. Brown, Kermit E. and H. 0. Boggs. Technology of Arial Lit Methods, Vol. Tulsa, Oklahoma: PennWell Books, 1980. 2, Aziz, K, G, W. Govier and M. Fogaras. "Pressure Drop in Walls Producing Ol ae Gas." Journal of Canaan Petrofaum Technol. (gy (luly- September 1972), pa. 348 23, Bi JP. and HD. Baggs, Two-Phase Flow in Pipes, Tulsa: Unk- versity of Tulsa, 1978, 4 Hagedom, Aion A. and Kermit €. Srown. “Experimental Study of Prassuta Gragints Oczurripg During Continuous Two-Phase Fw in Small DamaterVerleal Conduits” Joumaf of Patou Tectno!- 29y (Apa 1965), p. 475, 5, Ros, NG J."Simutanaous Flow of Gas and Liquid as Encountered in Woll Tusing” Journal of Potoieum Technology, 13 (October 1961), p. 1097. 6, Dune H Jp and WN. G, J. Ros, “Verical Flow of Gas and Ligut Mitures in Wells,” bin Wert Pavoioum Congress, Franktun, Got many. 7, Gray, H.€,Vanica Flow Correlations in Gas Wels.” User Manual ‘or API 143 Subsurface Controlled Satety Valve Sizing Comouter Program. App. B. June, 1974, 8, Onkisgewshi J. "Predicting Two-Phase Pressure Orops in Vertical Pipa.” Joumal of Patoloum Tachnology (une 1967) 8. Baggs, H.0. and J.P. Ball, "A Study of Two Phase Fw in Inctined Pines." Joumal of Pevoloum Technology (May 1975), p. 607 10, Katz, Donald L, etal. Handbook of Naniral Gas Engineering. New York, New Yoric NeGraw Hill Bock Co, 1958, pp. 305-012. 11, Brown, KE, Technology of Artifical Lift Mattods. Vale. 3a and 3b, Tuisa, Oklahoma: PanaWell Zooks, 1980. 42, Due, AE, ot al. "Gas-Liquid Flow in ipalines." Aascarch Ro- sufs, Vol. American Gas Associaton, American Potroloum int: tuto, May 1889. 19, Eaten, 8. A oa “The Prediction of Flow Pattors, Lculd Holdup land Pressure Losaes Qcouring Dung Continuous Two-Phase Flow in Horzortal Pipainas." Transactons of the AIME (1968). 14, Locher A. We and RC. Martine, "Proposed Correlation of Data forisothormal Two-Phase, Two.Component Flow in Pipes." Chem tal Engineering Progress, v. 45, N-1 (January 1949), pp. 39-48, 15, Baker, Ovid, “Goeign of Pipalinas for the Simuiteneous Flow of Ol and Gan” Oi-and Gas Journal, 52 (1954), pp. 185-195, 46, Flanigan, Osn, “Effect of Uphill ow on Pressure Drop in Dasign ‘of Two-Phase Gatnonng Systoms.” Ci and Gas Journal (Maren 10, 1958), p. 192, seine86 17. Gilber, W. E. "Flowing and Gas-Lit Well Performance." Dring and Production Practice, API, 1954, p. 143. 18, Texas A & | University. Report on the Calliration of Positive Flow Beans. Kingsville, Texas, 1948, 18, Mandhane, J. M. el al." Flow Pattern Map for Gas-Liquid Flow in Horizontal Pipes." intemational Journal of Multiphase Flow. + (1974), pp. 537-553. 20. Mukherjee, H. “An Experimental Study of Incined Two-Phase Flow.” Diss, University of Tusa, 1979. 21. Schmidt, 2. “Experimental Study of Two-Phase Slug Flowin a Pips- {ine-Fiser Pipe Syston.” Dss., University of Tusa. 1977. 22. Bil, J.P. etal. “Analysis of Two-Phase Tests in Large Diameter Prudhoe Bay Fiald Flowlines." SPE 8305. SPE Annwal Fall Meating, September, 1878. Technology of Artificial Lift Methods 2 2, ES 2. Baker, O. et al. “GasLiquid Flow in Pipelines.” Dosign Manual Yol 2. AGA-AP! Projact NX-26, Octabor 1970. Cortvile, T. 9a. "Two-Phase Flow Key to Ottshore Line Design.” i and Gas Jourral (August 10, 1861), pp. 71-75. unite, RS. "Pradicion of Condansata Flow Rates in Largo Diam- ‘ter High Pressure Wat Gas Pipolinas." 4PEA Joumal (1976), p. a7 kel, F. Principles of Heat Transter. New York, New York: intext Education Publishers, 1873 Yona, IR. et al. "Comparison of Liquid Holdup and Friction Factor Correlations for Gas-Liquid Flow in Honzonta Pipes.” vournal of Petroleum Technology (May 1875), pp. 564-567 ag ay Ce Chapter Nodal systems analysis by Kermit E. Brown, Dale A. Doty, Carl Granger, Lewis Lediow, Joe Mach, Eduardo Proafio, Zelimir Schmidt, and A. Paul Szilas 4.1 INTRODUCTION ‘The objective of nodal systems analysis is to combine the various components of the oil or gas well in order to predict flow rates and to optimize the various compo- nents in the system. This approach was discussed by Mach, Proafio, and Brown and is given here essentially the same as in the original paper (copyright SPE of AIMB:! ‘An approach is presented for applying systems anal- ysis to the complete well system, from the outer bound- ary of the reservoir to the sand face, across the perfora- tions and completion section to the tubing intake, and up the tubing string, including any restrictions and downhole safety valves, the surface choke, the flow line and separator. Figure 4.1 shows a schematic of a simple producing system. This system consists of three sections or mod: ules: ( flow through porous medium (2) flow through vertical or directional conduit (3) flow through horizontal pipe or inclined Bow line Figure 4.2 shows the various pressure losses that can occur in the more complex system from the reservoir to the separator. Beginning from the reservoir, these are noted as: resstre loss in porous medium resstire loss across completion ressure loss across regulator, choke, or tubing nipple ressure loss actoss safety valve ressure loss across surface choke ressure Loss in surface flow line total pressure loss in tubing string, which includes AP, and AP, ‘44 = Pun ~ Pose = total loss in surface flow line, including surface choke 87 ‘The various well configurations may vary from the very simple system of Figure 4.1 to the more complex system of Figure 4.2 or any combination thereof, and present-day completions more realistically include the various configurations of Figure 4.2 (especially off shore) ‘This chapter will discuss the manner in which to interrelate the various pressure losses. In particular, the ability of the well to produce fiuids will be inter- faced with the ability of the piping system to handle these duids. The manner in which to treat the effect of the various components will be shown by nodal concept. In order to solve the total-producing-system prob- Jems, nodes are placed to segment the portion defined by different equations or correlations. Figure 4.3 has been prepared showing locations of the various nodes. This figure is the same as Figure 4.2 except only the node positions are shown. A node is classified as functional when a pressure differential exists across it and the pressure or low-rate response can be represented by some mathematical or physical function. More realistically, we will refer to a node at the bottom of the well, at the top of the well, etc., as a solution node. These two solution positions were discussed by Brown and Begg: Node 1 represents the separator pressure, which is, usually regulated at a constant value; however, some separator pressures do change with rate and should be properly accounted for. There are two positions whereby the pressures are not functions of flow rates. These are P, at node 8 and Peep at node 1. For this reason, any trial-and-error solution to the total system problem must be started at node 1 (Psey), node 8 (P,) or both node 1 and 8 if'an intermediate node such as Bor 6 is selected as the solution node. Once the solution node is selected, the pressure drops or gains from the starting point are added until the solution node is reached. Example problems are worked to show the nodal system approach. For example, the flow rate pos- sible can be determined by utilizing node 8 (P,), node88 — Technology of Artificial Lift Methods Figure 4.1 Complete Producing System 6 (Pad, node 3 Px9), and node 1, (Paes) or other positions. ‘The node selected depends upon which component we want to isolate for evaluation. The effect of tapered strings, surface chokes, and safety valves can also be evaluated in this manner. Figure 43 Location of Various Nodes (after Mach, Prost, and Brown.' © SPE of AIME) In summary, a nodal approach is presented in order to effectively evaluate a complete producing system. All of the components in the well, starting from the static pressure (P,) and ending at the separator, are considered. This includes flow through the porous mo : Tene => SALES “APs = (Py Peep) >} as UNE fa rye 7 P, SEPARATOR Pose an ‘SURFACE CHOKE | LIQUID. STOCK et AP 5* (Putosc) Fe "TANK WELD 6, TEE EEE Le Zi oa MME TELE sv"Pose)_|_N AL ffosy Le Leer AP, = Pr -Pywis = LOSS IN POROUS MEDIUM ae; a7 AP, = Pyfs~Pwt = LOSS ACROSS COMPLETION Put~Puh APs = Pup-Por =" "RESTRICTION BOTTOM HOLE! BP, = Pygy"Posy =” “SAFETY VALVE seen AP, © Pub Posc =” "SURFACE CHOKE, ap, = 7 Gee) APG = Pose-Psep = IN. FLOWLINE hee AP, = Pyf-Pwh = TOTAL LOSS IN TUBING AP = Pyn-Poep + “" FLOWLINE Ys Figure 4.2. Pressure Losses in Complete System (after Mach, Proaio, and Brown.* © SPE of AIME) ee eudium, flow across the perforations and completion, flow up the tubing string with passage through a possible downhole restriction and safety valve, and flow in the horizontal flow line with passage through a surface choke and on to the separator. Various positions are selected for solution nodes, and the pressure losses are converged on that point from both directions. Nodes can be effectively selected to __show better the effect of certain variables such as inflow ‘ability, perforations, restrictions, safety valves, surface chokes, tubing strings, flow lines, and separator pres- sures The appropriate multiphase flow correlations and equations for restrictions, chokes, etc., must be incorpo- rated in the solution. An effective means of analyzing an existing well, making recommended changes. or planning properly for a new well can be accomplished by the nodal system analysis, This procedure offers a means to more eco- rnomically optimize producing walls. 4.2. SOLUTION PROCEDURE FOR OIL WELLS 4.21 INTRODUCTION In order to best illustrate the solution procedure, the following example was presented by Mach, Proaiio, and Brown? and will be worked by taking the solution node at several different positions. EXAMPLE PROBLEM (OIL WELL) Given Data (flowing oil wel: separator pressuce: 100 psi fow line: 2-in., 3,000 ft long WOR: depth: 5,000 f mid perf. GOR: 400/sef/bbI productivity index = 10 A simple system such as shown in Figure 4.4 is as- sumed. For purpose of illustration only let us assume that a constant J of 1.0 exists for all flowing pressures tubing size, 2%-in OD ‘P,: 2200 psi Figura 4.4 Nodes tor Simple Producing System (after Mach, Froaio. fang Brown, @ SPE of AIME) Nodal Systems Analysis 89 940 —Pa) vi cuRVE FOR EXAMPLE nessun UwnOREDS OF P51) ATE (THOUSANDS OF 20) Figure 4.5 (PR Curve for Example Problem (Constant J = 1.0) for this well. In reality, wo know that two-phase flow will occur below the bubble-point pressure of 1,800 psi found from Figure 2.14. However, for the flow rates obtainable with 2%-in. OD tubing, the rates will differ very little from @ straight-line J plot as compared to a Vogel solution (see Figure 4.5). In order to apply the constant J plus Vogel solution, we will assume constant J of 1.0 from 2,200 psi, to 1,800 psia (bubble point) and a Vogel curve behavior from 1,800 to zero pressure, (2.02.00) Lg 7 -02,200~ 1,800 + = = 400+ 1,000 = 1,400 b/d or for the constant J case, dmax = 1.0(2,200 — 0) = 2,200 b/d. Other pressures are assumed, and the IPR eurve is constructed as noted in Figure 4.6. For purposes of illustration, we will show the con- stant J solution for simplification in working the prob- lem in most cases. Gnas = 99> eaten ssranco apo AHS RATE (HNOREDS OF 8/0) Figure 4.8. IER Curve for Example Problem (Voge! Solution) 4.22 SOLUTION AT BOTTOM OF WELL (NODE § FROM FIGURE 4.3) Probably the most common solution position is at the bottom of the well—that is, at the canter of the You might also like
Sucker-Rod Pumping Handbook: Production Engineering Fundamentals and Long-Stroke Rod PumpingFrom EverandSucker-Rod Pumping Handbook: Production Engineering Fundamentals and Long-Stroke Rod PumpingRating: 5 out of 5 stars5/5 (9) Hydraulic Fracturing in Unconventional Reservoirs: Theories, Operations, and Economic AnalysisFrom EverandHydraulic Fracturing in Unconventional Reservoirs: Theories, Operations, and Economic AnalysisNo ratings yet Well Integrity for Workovers and RecompletionsFrom EverandWell Integrity for Workovers and RecompletionsRating: 5 out of 5 stars5/5 (3) Enhanced Oil Recovery, II: Processes and OperationsFrom EverandEnhanced Oil Recovery, II: Processes and OperationsRating: 5 out of 5 stars5/5 (4) Fundamentals of Gas Lift Engineering: Well Design and TroubleshootingFrom EverandFundamentals of Gas Lift Engineering: Well Design and TroubleshootingRating: 5 out of 5 stars5/5 (2) PVT Property Correlations: Selection and EstimationFrom EverandPVT Property Correlations: Selection and EstimationRating: 5 out of 5 stars5/5 (1) Caso de Analisis Capital HumanoDocument1 pageCaso de Analisis Capital HumanoSaúl Oswaldo HurtadoNo ratings yet Advanced Production Decline Analysis and ApplicationFrom EverandAdvanced Production Decline Analysis and ApplicationRating: 3.5 out of 5 stars3.5/5 (4) Surface Operations in Petroleum Production, IIFrom EverandSurface Operations in Petroleum Production, IIRating: 3.5 out of 5 stars3.5/5 (2) Modern Chemical Enhanced Oil Recovery: Theory and PracticeFrom EverandModern Chemical Enhanced Oil Recovery: Theory and PracticeRating: 4 out of 5 stars4/5 (2) Well Test Analysis for Fractured Reservoir EvaluationFrom EverandWell Test Analysis for Fractured Reservoir EvaluationRating: 5 out of 5 stars5/5 (1) The Practice of Reservoir Engineering (Revised Edition)From EverandThe Practice of Reservoir Engineering (Revised Edition)Rating: 5 out of 5 stars5/5 (3) Fundamentals of Numerical Reservoir SimulationFrom EverandFundamentals of Numerical Reservoir SimulationRating: 3 out of 5 stars3/5 (2) Hydraulic Fracturing Explained: Evaluation, Implementation, and ChallengesFrom EverandHydraulic Fracturing Explained: Evaluation, Implementation, and ChallengesNo ratings yet Practical Reservoir Engineering and CharacterizationFrom EverandPractical Reservoir Engineering and CharacterizationRating: 4.5 out of 5 stars4.5/5 (3) Low Salinity and Engineered Water Injection for Sandstone and Carbonate ReservoirsFrom EverandLow Salinity and Engineered Water Injection for Sandstone and Carbonate ReservoirsNo ratings yet Gas Well Deliquification: Solutions to Gas Well Liquid Loading ProblemsFrom EverandGas Well Deliquification: Solutions to Gas Well Liquid Loading ProblemsNo ratings yet Petroleum Production Engineering, A Computer-Assisted ApproachFrom EverandPetroleum Production Engineering, A Computer-Assisted ApproachRating: 4.5 out of 5 stars4.5/5 (11) A Practical Companion to Reservoir StimulationFrom EverandA Practical Companion to Reservoir StimulationRating: 5 out of 5 stars5/5 (1) Advanced Water Injection for Low Permeability Reservoirs: Theory and PracticeFrom EverandAdvanced Water Injection for Low Permeability Reservoirs: Theory and PracticeRating: 4 out of 5 stars4/5 (2) Special and Different: The Autistic Traveler: Judgment, Redemption, & VictoryFrom EverandSpecial and Different: The Autistic Traveler: Judgment, Redemption, & VictoryNo ratings yet What Squirt Teaches Me about Jesus: Kids Learning about Jesus while Playing with FidoFrom EverandWhat Squirt Teaches Me about Jesus: Kids Learning about Jesus while Playing with FidoNo ratings yet Strangers' Voices In My Head: A Journey Through What Made Me Who I Am from My MindFrom EverandStrangers' Voices In My Head: A Journey Through What Made Me Who I Am from My MindNo ratings yet Best Friends (Lyrics) (SECURED)Document4 pagesBest Friends (Lyrics) (SECURED)Simón Huanca GallardoNo ratings yet 2 3 Control-Cadena-FrioDocument8 pages2 3 Control-Cadena-FrioSimón Huanca GallardoNo ratings yet AntibióticoDocument26 pagesAntibióticoSimón Huanca GallardoNo ratings yet Coordinador Scania PDFDocument31 pagesCoordinador Scania PDFSimón Huanca Gallardo100% (1) SDP 1000032 03Document14 pagesSDP 1000032 03Simón Huanca Gallardo100% (1) 03aislador Campana 10 KV PDFDocument2 pages03aislador Campana 10 KV PDFSimón Huanca GallardoNo ratings yet Manual Senior G6 PDFDocument142 pagesManual Senior G6 PDFSimón Huanca GallardoNo ratings yet Tablas de Carga AntiguaDocument4 pagesTablas de Carga AntiguaSimón Huanca GallardoNo ratings yet Manual SeniorDocument168 pagesManual SeniorSimón Huanca GallardoNo ratings yet Funcionamiento y Reglaje de Los PreostatosDocument1 pageFuncionamiento y Reglaje de Los PreostatosSimón Huanca GallardoNo ratings yet AtlasCopco XAS 48 Series 8 Brochure Pre-Launch PDFDocument8 pagesAtlasCopco XAS 48 Series 8 Brochure Pre-Launch PDFChoirulNo ratings yet Formato D02.07.f01-Infraestructura - Tecnologica - 2020 NSDocument7 pagesFormato D02.07.f01-Infraestructura - Tecnologica - 2020 NSJonnyJaimesSilvaNo ratings yet La DomotiqueDocument39 pagesLa Domotiquemat konanNo ratings yet Practice Quiz AnswersDocument40 pagesPractice Quiz AnswersReddy Freddy75% (4) Unidad 5. Maquinas Eléctricas. Pedro Isaac Alvarez HDZDocument31 pagesUnidad 5. Maquinas Eléctricas. Pedro Isaac Alvarez HDZisaacNo ratings yet Manual Nobreak XNB 600 720 1200 1440 1800 VA 01-21 SiteDocument16 pagesManual Nobreak XNB 600 720 1200 1440 1800 VA 01-21 SiteJoao Cesar Sousa SilvaNo ratings yet Grupo 7 - Viernes 18-Noviembre-2021Document36 pagesGrupo 7 - Viernes 18-Noviembre-2021Austin OrtegaNo ratings yet Conceptos: Top Sirloin RNDocument1 pageConceptos: Top Sirloin RNjuan manuel arellano pinedaNo ratings yet Unidad 4 Algoritmos Repetitivos - Parte 1Document34 pagesUnidad 4 Algoritmos Repetitivos - Parte 1Cristhian Lucio Quispe NinaNo ratings yet Uncertainty of Mineral Resource Estimates: From Confidence Intervals To Resource ClassificationDocument8 pagesUncertainty of Mineral Resource Estimates: From Confidence Intervals To Resource ClassificationEXPL RI-6 SINGRAULINo ratings yet Apple Marketing ChallengeDocument5 pagesApple Marketing ChallengeMazlenashahNo ratings yet Entrenamiento Afex para T - Cnicos WestfireDocument128 pagesEntrenamiento Afex para T - Cnicos WestfireCesar MuñozNo ratings yet 3.2 Procedimiento Constructivo 22222Document36 pages3.2 Procedimiento Constructivo 22222Alexis Polanco DelgadoNo ratings yet Auditor Lider ISO 9001-2015 Registro IRCADocument3 pagesAuditor Lider ISO 9001-2015 Registro IRCAJorge PachecoNo ratings yet Ultromat - at - Atf - 96 - v5 - Manual - EnglishDocument51 pagesUltromat - at - Atf - 96 - v5 - Manual - EnglishMADOURI Sid ahmedNo ratings yet Funcionamiento Del Driver CNC DIY para Motores Paso A PasoDocument21 pagesFuncionamiento Del Driver CNC DIY para Motores Paso A PasoJuNo ratings yet Avksvmc Butterfly Digital2020Document15 pagesAvksvmc Butterfly Digital2020Nam Nguyễn ĐứcNo ratings yet Protocolo - Seccionador de PotenciaDocument1 pageProtocolo - Seccionador de PotenciajoseNo ratings yet Costeo Basado en ActividadesDocument8 pagesCosteo Basado en ActividadesIleani CruzNo ratings yet 01 Kernsmethod 150909121926 Lva1 App6892 PDFDocument45 pages01 Kernsmethod 150909121926 Lva1 App6892 PDFPenjual AirNo ratings yet Diagrama de Un Transmisor de AMDocument1 pageDiagrama de Un Transmisor de AMlph0240No ratings yet Huamanchay PCDocument151 pagesHuamanchay PCServicio FarmaciaNo ratings yet Acceptance of IoT Based and Sustainability Oriented SM - 2022 - Sustainable CitiDocument13 pagesAcceptance of IoT Based and Sustainability Oriented SM - 2022 - Sustainable CiticomputerwolfNo ratings yet Pmkvy Guidelines Junior Software DeveloperDocument4 pagesPmkvy Guidelines Junior Software DeveloperJitendra KumarNo ratings yet 06 - Chapter 2Document29 pages06 - Chapter 2Jessa Joy Alano LopezNo ratings yet Audi 2Document2 pagesAudi 2Gina Vanessa Palmera SalcedoNo ratings yet Club Inwi Compagne 21Document6 pagesClub Inwi Compagne 21Omnia OmniaNo ratings yet Adriana Carballo GuerreroDocument4 pagesAdriana Carballo GuerreroJonathAn CalapiñaNo ratings yet p3 Osciloscopio PDF para WordDocument4 pagesp3 Osciloscopio PDF para WordBRHAYAN JEANPIERRE RAMIREZ GOMEZNo ratings yet