SPE 24057 Steam Injection Well Falloff Analysis for Monarch Zone, MOCO Steamdrive, California A.M. AI-Khatib, Mobil Oil Corp.; H.J. Ramey Jr., Stanford U.; and J.P. Busby, Mobil Oil Corp. SPE Members Copyright 1992, Society of Petroleum Engineers Inc. This paper was prepared for presentation at the Western Regional Meeting held in Bakersfield. California, March 30-April 1, 1992. This paper was selected for presentation by an SPE Program Committee following review of information contained in an abstract submitted by the author(s). Contents of the paper, as presented, have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material, as presented: does not necessanly reflect any position of the Society of Petroleum EngiAeers, its officers, or members. Papers presented at SPE meetmgs are subject to publication review by Committees of the Society of Petroleum Engineers. Permission to copy is restricted to an abstract of not more than 300 words. lIIustraltons may not be copied. The abstract should contain conspicuous acknowledgment of where and by whom the paper is presented. Write Librarian Manager, SPE, P.O. Box 833836, Richardson, TX 75083-3836. Telex, 730989 SPEDAL. ABSTRACT The effects of the unswept zone mobility on steam injectivity were investigated. Six falloff tests were run in a steamflood near Maricopa, California. The tests were analyzed using the rad i a 1 compos i te model wi th the aid of the Barua-Horne and the Kappa Saphir well test analysis software packages. Computer aided matching of the radial composite model with steam injection well falloff data provided remarkable agreement between field data and model simulation. It was possible to find skin, wellbore storage, steam zone mobility and compres- sibility, and the mobility and compressibility of the zone ahead of the steam front. low mobil ity ahead of the steam zone was found to 1 imit steam injec- tivity in several field cases. INTRODUCTION In 1979 Eggenschweiler et al' found that the radial composite model had an unexpected property which made it suitable for thermal injection well fallOff pressure transient analysis. Wellbore storage and skin effects lasted for short time (seconds) then there was a semi-log radial flow period for about one hour which contained the swept zone properties, followed immediately by a pseudo-steadystate Cartesian straightline which contained the swept zone volume for gas injection with gravity override. The pseudo-steady period appeared to last for about five hours. Thus, the properties of this pressure transient test were ideal. The test could be completed in a short time. Inspection of many sets of field test data indicated that the procedure was appropriate for both insitu combustion and steam injection. However, the swept volume usually appeared too large for both kinds of tests. Reference. iiid j[(uatrlltf_ It iiid of peper. 293 Onyekonwu 2 found it was necessary to reduce the apparent combustion zone volume by half. Numerical s i mulat i on showed the pressure test measured both the combustion zone and a high mobility zone volum, in the steam plateau in combustion. Walsh et al showed that the steam zone volume was controlled by a very large adiabatic compressibility due to steam condensation. Messner and Williams 4 applied the pseudo-steadystate method to steam injection falloff interpretation. A notable result of their study was that it was sometimes difficult to identify the correct pseudo- steady line. Eggenschweil er et a l' observed that there was a decades-long transi t ion zone between the i ni t hl semi-log straightline for the swept zone and the second semi-log straightline for the mobility of the region ahead of the swept zone. As a result, they concluded that it was likely that producing well interference would preclude measuring the mobility of the region ahead of the swept zone. However, Barua and Horne considered computer regression matching of thermal falloff data with the radial composite model and concluded that both the mobility and porosity-compressibil ity product in the swept zone and ahead of the swept zone could be found as well as swept zone radius, wellbore storage and skin effect. This procedure was better for finding swept volume than the Cartesian straightline method. The results of Barua and Horne 5 were impressive because experience had shown it was often difficult to identify the pseudo-steady line correctly. They also showed comparisons between field data and simulated data which were in remarkable agreement. Thus, computer aided analysis by nonlinear re- gression showed it was possible to find the mobility ahead of the swept zone even though no semi -log straightline was evident. This observation was of critical importance to the present study. The main objective was to determine 2 Steam Injection Well Falloff Analysis for Monarch Zone, MOCO Steamdrive SPE 24057 why steam injectivity was so poor in parts of the Monarch zone steam injection project in the Midway Sunset Field. FIELD TESTS The steam drive project in question totals 60 acres of five-acre five-spot patterns in the Monarch zone, Midway Sunset Field. The formation is a shallow anticline at depths from 800 to 1,000 ft. The project was developed across the crest and southern flank of the Moco 35 anticline, which is a southeast trending structure with dips ranging from fi ve degrees on the crest to 40 degrees on the flanks. The oil gravity of the Monarch crude is 13 API degrees whil e the oil vi scosity ranged from 5,000 to 40000 cp at original reservoir temperatures (viscosity variations and effects will be discussed in a later section of this paper). Prior to the startup of steam injection, oil production averaged three to five BOPD/well. For 16 months of continuous steam injection, no significant response was observed from the patterns across the crestal portion of the steam drive. However, the south flank patterns were performing well. The poor crestal response was coupled with high injection pressure and low injection volumes. Cresta 1 injectors took 200 to 500 BSD feedwater rates at 700 to 800 psi, while the south flank injectors took 500 to 1,000 BSD feedwater rates at 200 to 400 psi. Note: Permission to inject at the higher pressures had been received from the D.O.G. Downhole temperature observation data offered little information on steam zone size or causes of low crestal injectivity. To determine causes of the low cresta 1 inject i vity trans i ent pressure testing was performed across the steamflood project. The main object i ve of the falloff tests was to determine the cause of poor steam injectivity. Six falloff tests were run in this study. Two different types of response were observed. Both were surprising and will be shown in the following. The types of tests were observed repeatedly and thus are not uni que occurrences. An example of each type result will be discussed. EQUIPMENT AND TEST PROCEDURE The pressure surveys were run through the use of a capillary tubing string filled with helium. The tubing was lowered into the well and placed at the middle of the perforated interval. A steam/liquid separator was used to measure a stable steam injection rate and quality. Upon achieving a stable injection rate and pressure, the well was shut-in. The bottomhole pressure through a surface read-out transducer were conducted. When sat i sfactory semi -log and 1 og-l og plots were achieved, the test was terminated. The shut-in time for the six falloff tests ranged from 20 to 60 hours, with two to three hours of injection rate measurement prior to shut-in. 294 WELL 'A' TEST Well 'A' was a poor injector located on the crest of the anticline. The reservoir and test data for this well are presented in Table 1. Figures 1-3 present a log-log pressure and pressure derivative graph, a semi-log graph and a Cartesian graph for the Table 1 injection falloff data. Both the 1 og-l og type curve and semi -log graphs indicate that skin effect is not a reason for poor injectivity. The semi-log and the Cartesian graphs indicate an apparent straightline which should contain the swept zone volume. However, the swept zone size is of incidental interest here. If there is no skin effect on the injector, and mobility is high in the steam zone as shown by the small slope of the semi-log graph straightline, then the mobility ahead of the steam zone must limit injec- tivity. The graphical straightline procedures of Eggenschweiller et all are not useful here as there is no second semi-log straightline evident. thus it is necessary to use computer aided i nterpretat ion with the radial composite model. Both the Barua-Horne 5 and Kappa Saphir 6 well test programs were used and gave almost identical results. The test analysis procedure was verified through semi-log, log-log and Cartesian plot computer matching. Results for Well 'A' are presented in Table 2, and a graphical comparison of field data and the program model are shown on Figure 3 and Figure 4. The simulated pressures were within two psi of the measured fi e 1 d data. The semi -log graph was also simulated and had an excellent fit to the measured data. WELL '8' TEST Well B was a good injector located on the southern flank of the anticline. The reservoir and test data for this well are presented in Table 3. Figures 5, 6 and 7 present a 1 og-l og pressure and pressure derivative graph, a semi-log graph, and a Cartesian graph of the Table 3 data. Both log-log and semi-log graphs indicate negligible skin effect which could enhance or reduce steam injectivHy. There is a surprising result in the long-time results after one hour shut in. The Well 'B' pressure derivative log-log type curve continues to decrease in Figure 5, while Well 'A' derivative increases beyond one hour in Figure 1. The semi-log graph, Figure 6, for Well 'B' shows no evidence of a Pseudo-steadystate period following the semi-log straightline (0.1 to 1 hours on Figure 6). The Cartesian graph of Well 'B' data on Figure 7 appears to curve smooth 1 y with no apparent pseudo- steadystate straightline. Compare the pseudo- steadystate rapid drop in pressure after the semi- log straightline on Figure 2 after one hour with the slight rise in pressure after one hour on Figure 6. Al so, the pressure deri vat i ve increases after one hour on Figure 1 for Case A, but decreases after one hour on Figure 5 for Case B. The Well 'B' test results presents a Case not ant i ci pated by Eggenschwiler et a1 1 The mobility of fluids moving ahead of the steam front is higher than the steam mobility in the swept zone. Ambastha 7 made an exhaustive study of properties of the radial-composite model leading to many SPE 24057 Steam In.iection Well Falloff Analvsis for Monarch Zone MOCO Steamdrive publications. Although he considered most apparent combinations of mobility and storativity ratios, we did not expect thermal oil recovery to 1 ead to mobility ratios greater than unity across the steam- oil interface. The Saphir program 6 users manual published by Kappa engineering, Paris, France (section 7) presents an example of the effect of mobility ratio and diffusivity ratio on the pressure derivative plot. Figure 9 shows tne behavior of the pressure derivative plots for mobility ratios of 0.1 to 10. The Well 'B' pressure and pressure derivative graph on Figure 5 resembles the low mobility cases on Figure 9. This suggests several thoughts. The pseudo-steadystate method proposed by Eggenschwiler et all can not be used to find the swept volume for such cases, but computer-aided interpretation with regression and the composite model may permit determination of all parameters. This was the case for the Well 'B' test. Test results for Well 'B' are presented in Table 4. A graphical comparison of field data and simulated results are shown on Figure 8. The simulated pressures were within 2 psi of the measured data. The semi-log graph was also simulated and had a near perfect fit to the measured data. Some of the data presented in Tables 1 and 3 require discussion. The steam injection formation volume factor was computed using specific volumes of saturated 1 iquid and vapor and the injected steam qua 1 i ty. Saturation pressure was taken from the flat portion of the semi-log graphs, about 0.1 to 1 hours on Figures 2 and 6. The total adiabatic compress i bil ity was evaluated for the steam swept region by the method described by Walsh et all. The net thickness was taken as the perforated interval. DISCUSSION OF TEST RESULTS WELL 'A' As mentioned earlier, Well 'A' was a poor injector located on the crest of the anticline. A mobility ratio of 105 was calculated for this test. The mobility ratio is defined as the inner zone mobility divided by the outer zone mobility. This is the expected result because the viscosity of steam (see Tables 1 and 3) is very low compared to oil viscosity. This sort of example is typical of most steam injection falloff tests we've analyzed. The Eggenschwiler et all method can be used to find the inner swept zone volumes and properties. But in this case, we reached the conclusion that steam injectivity was limited by poor oil mobility ahead of the steam front. Thus, it was necessary to use computer-aided interpretation to obtain properties of the region ahead of the steam zone. This is not possible with the Eggenschwiler et all method. It was found that the mobil ity ahead of the steam zone was very low compared to good injectors located down fl ank. The obvious reason for poor mobil ity ahead of the steam-swept zone was high oil viscosity. This steam injection project was located in a zone which had a poor insitu combustion project in the 1960's. It was suspected that oxidation had increased in place oil viscosity on the crest. 295 Thus, oil samples were taken for viscosity measurement. Project-wide testing of the oil viscosity indicated a wide range of oil viscosities ranging from 5,000 cp to 300,000 cpo Two oil samples near crestal steam injectors yielded viscosities above 100,000 cpo Next to these steam injectors (on the crest of the anticline) an insitu combustion project had existed. The above normal Monarch oil viscosities resulted from low temperature oxidation of the crude due to a poor burn. The Well 'A' case is typical of most steam injection falloff tests we've seen. We present this example mainly to show that computer-aided interpretation wi th the compos ite radi a 1 model wi th storage and skin effect can provide a remarkable interpretation of steam pressure falloff data. In almost all cases, the difference between field pressures and computed pressures from the composite model and interpreted parameters was less than one psi for the entire duration of the falloff test. The low oil mobility ahead of the steam-swept region found for Well 'A' was an unusual finding. The only general result we wi sh to emphas i ze is that poor injectivity is not always a result of a large skin effect. WELL 'B' Well 'B' was a good south flank injector. The producers on the southern flank of the project were excellent and steam production was evident two years after the project was started. The Well 'B' example was found to be typical of several good injectors in the project. The mobility of the fluids ahead of the steam zone was five times that of the steam in the swept zone. This was a surprise. As a result, there was no obvious pseudo- stead1state data in the test and the Eggenschwiler et al method would lead one to suspect there was no steam-swept zone. The strange pressure derivative behavior sU9gested the low mobility ratio considered by Ambastha. An important result is that computer- aided interpretation yielded swept zone size and composite system parameters even in this case. A remaining question is how can the mobility ratio be less than unity for steam injection? One potential cause is the thickness of the steam zone compared to the thickness of the oil zone ahead of the steam zone. Because of gravity override, the steam zone is probably less than 20% of the oil zone thickness. Another factor may be rel ative permeabil ity. Messner and Wi 11 i ams 4 remarked that lab data indicated that the relative permeability to steam was only 10% in the swept zone. Although both thickness and permeability effects are in the direction of increased mobility ahead of the steam zone, the low steam viscosity compared to any liquid viscosity makes steam injection appear a high mobility process. We can only say that we have seen several examples like Well 'B' for other good steam injectors. 3 4 Steam Injection Well Falloff Analvsis for Monarch Zone MOCO Steamdrive CONCLUSIONS 1. Computer-aided pressure falloff testing of the steam injection wells in Monarch reservoir yielded good insight to the reservoir properties. 2. It was possible to determine the unswept zone mobility, radius and storativity with computer-aided pressure analysis. 3. Poor mobil i ty of the unswept zone was determi ned to be the rna in reason for poor steam injectivity in the crestal patterns of the Monarch steam injection project. 4. Increased mobility ahead of the steam zone was observed for several good injectors. It was possible to interpret this kind of falloff data wi th several computer programs for the radial composite case, although the Eggenschwiler et all pseudo-steadystate method was not helpful. ACKNOWLEDGEMENTS The authors wish to thank the management of Mobil Exploration & Producing U.S. Inc. for permission to publish this work. REFERENCES 1. Eggenschwiler, M., Satman, A., Ramey, H. J., Jr., and Cinco-Ley, H.: "Interpretation of Injection Well Pressure Transient Data in Thermal Oil Recovery," SPE 8908, presented at the 50th Annual California Regional Meeting, SPE of AIME, Los Angeles, California, Apr. 9- 11, 1908. 2. Onyekonwu, M. 0.: "Interpretation of Insitu Combustion Thermal Recovery Falloff Tests," Ph.D Thesis, Stanford Univ., California, 1985. 3. Walsh, J. W., Jr., Ramey, H. J., Jr., and Brigham, W. E.: "Thermal Injection Well Falloff Testing," SPE 10227, presented at the 56th Annua 1 Techn i ca 1 Conference and Exhibition, SPE of AIME, San Antonio, Texas, October 5-7, 1981. 4. Messner, G. L., and Will iams, R. L.: "Application of Pressure Transient Analysis in Steam Injection Wells," SPE 10781, California Regional Meeting, San Francisco, California, Mar. 24-26, 1982. 5. Barau, J., and Horne, R. N.: "Computerized Analysis of Thermal Recovery Well Test Data," SPE 12745, California Regional Meeting, Long Beach, California, Apr. 11-13, 1984. 6. Saphir User's Manual, Section 7, Kappa Engineering, Paris, France, 1986. 7. Ambastha, A. K.: "Pressure Transient Analysis for Compos ite Systems," Ph. D. Thes is, Stanford Univ., California, 1988, page 45. 296 '\PF ?40"i7 TABLE 1 WELL At RESERVOIR AND TEST DATA STEAM FORMATION VOLUME FACTOR, TOTAL ADIABATIC COMPRESSIBIUTY, STEAM VISCOSITY NET THICKNESS INITIAL PRESSURE POROSITY INJECTION RATE FEEDWATER INJECTION TIME INJECTION aUAUTY RBBUBBLFEEDWATER PSI-1 CP FT PSIA BBUD HOURS WT.% TIME - PRESSURE DATA HOURS 0.0000 0.0011 0.0022 0.0033 0.0044 0.0056 0.0078 0.0100 0.0144 0.0189 0.0222 0.0256 0.0289 0.0330 0.0378 0.0433 0.0489 0.0556 0.0711 0.0900 TABLE 2 PSI 639.8100 639.2900 638.4600 637.3000 635.9800 634.9000 632.9700 631.9000 630.3000 629.4800 629.0300 628.8100 628.4300 628.2600 627.9500 627.3300 626.9600 626.9300 626.3200 625.8500 WELL A, TEST RESULTS HOURS 0.1144 0.1456 0.1833 0.2067 0.2611 0.3300 0.4167 0.4678 0.5256 0.5900 0.6622 0.7433 0.8344 0.9367 1.05 1.18 1.32 1.48 1.66 1.87 PERMEABIUTY TO STEAM FIRST ZONE, md SKlNFACTOR MOBIUTY RATIO, ZONE1 TO ZONE 2 DIFFUSIVITY RATIO, ZONE 1 TO ZONE 2 SWEPT ZONE RADIUS, Ri FT. PSI 625.2800 624.5200 623.6000 623.7100 623.0900 622.6000 621.9000 621.3500 620.6300 620.5300 620.0900 619.9300 619.7700 619.0100 618.7100 617.8800 617.6200 616.6900 616.6000 615.7800 HOURS 2.10 2.35 2.64 2.97 3.74 4.19 4.71 5.28 5.93 6.65 7.47 8.38 9.41 10.56 11.86 13.31 14.93 16.76 18.81 21.15 30.2 0.0326 0.020 50 639.81 0.28 660 2000 65 PSI 615.1900 614.2900 613.7000 612.8100 611.1500 609.8400 608.9200 607.8500 606.8400 605.4300 603.9500 601.9600 600.1300 598.2000 596.1900 594.0300 591.7000 589.1500 586.5100 583.6900 273 -0.6 105 1000 29 TABLE 3 WELL B, RESERVOIR AND TEST DATA STEAM FORMATION VOLUME FACTOR, TOTAL ADIABATIC COMPRESSIBIUTY, STEAM VISCOSITY NET THICKNESS INITIAL PRESSURE POROSITY INJECTION RATE, FEEDWATER INJECTION TIME INJECTION aUAUTY RBBUBBLFEEDWATER PSI-1 CP FT PSIA SBUD HOURS WT% TIME - PRESSURE DATA HOURS 0.0000 0.0014 0.0028 0.0042 0.0056 0.0069 0.0083 0.0097 0.0111 0.0125 0.0153 0.0181 0.0208 0.0236 0.0278 0.0319 0.0361 0.0417 0.0472 0.0542 0.0611 0.0694 0.0792 0.0889 0.1000 0.1125 0.1431 0.1611 TABLE 4 PSI 204.16 203.87 203.45 202.96 202.43 201.91 201.40 200.91 200.46 200.03 199.24 198.57 197.99 197.46 196.82 196.26 195.82 195.31 194.90 194.49 194.10 193.79 193.37 193.03 192.70 192.33 191.77 191.56 WELL at TEST RESULTS PERMEABIUTY TO STEAM FIRST ZONE, md SKlNFACTOR MOBIUTY RATIO, ZONE1 TO ZONE 2 DIFFUSMTY RATIO, ZONE 1 TO ZONE 2 SWEPT ZONE RADIUS, Ri FT. HOURS 0.2081 0.2358 0.2664 0.2997 0.3386 0.3803 0.4811 0.5422 0.6089 0.6856 0.7772 1.1106 1.2522 1.5856 1.7856 2.0106 2.2606 2.8606 3.2106 4.5772 5.1439 5.7772 6.5106 7.3106 9.2772 10.4439 13.2772 16.7772 21.1939 SPE 2405 7. 106.0 0.0502 0.019 50 224 0.28 340 2000 70 PSI 190.86 190.59 190.26 190.10 189.81 189.54 189.08 188.91 188.58 188.38 188.09 187.35 187.14 186.84 186.67 186.51 186.36 186.10 185.97 185.44 185.30 185.12 184.94 184.64 183.87 183.50 183.35 183.17 182.81 1421 -.9 0.4 1.72 967 10 2 640 . . . 10 1 Q. "0 ~ D. "0 10 0 *+++ HfI-++ + *+++-f!-+ -I'I++++++t#+ *++ ~ ...... #+++++++* ,f-+++t+ ... ++#+ . +++ .. -.- + .. "" ........ . .. . . . - +- . - .. . .- . . .. -. .. .- - -. -- +- .. -. .- - . . + 630 620 PWS 610 psia 600 590 580 . 10 -1 10 -3 10 -2 10 -1 10 0 10 1 10 2 dt, hours Figure 1 - (Well A) Pressure difference and derivative log-log type curves. 640 630 620 psla 610 600 590 0 10 t, hours Figure 3 - (Well A) Cartesian graph of falloff - . -... -.. 0_010 0.100 1.000 dt, hours SPE 2405 '1 --.- -- -. -.. --. 10.00 . . Figure 2 - (Well A) Semilog Graph of Test Data 20 10 2 10 2 .L U. UJ....L -. .. #-H 1+-+f'I--fITTT 10 1 10 1 ~ - I - ~ ~ *+"1" ++ ~ t..- ... -+'" . ~ .. 1-.". .... _____ .-.... at- .......... - . lit- -.
. "-
"- GIS ~ Q. o/J ~ Q. ~ 10 0 10 0 + . 10 1 10 -3 10 2 10 1 10 1 10 1 dt, hours cit, hours Figure 4 - (Well A) Pressure Difference and Pressure Derivative Type Curve Figure 5 - (Well B) Pressure Difference and Pressure Derivative Type Curve 210 205 . . 200 ... . . . . . . PWS 195 .. .. ... psla 190 185 .-. 180 0.010 0.100 1.000 10.00 dt, hours Figure 6 - (Well B) Semilog Graph of Test Data w o o C\I "0 a. 200 190 o 10 2 10 1 Q. 'g ~ a.. 'g 10 0 .. 10 20 10 -1 10 -3 10 -2 t, hours dt. hours Figure 7 - (Wen B) Cartesian Graph of Field Data Figure 8 - (Well B) Comparison of field data and simulated results from computer-aided interpretation. 0 C. "C "C c: til 0 C. "C 10 2 ,--------,-------,--------,-______ ,-______ -, ______ --, 10 1 10 0 10- 1 1 0 ~ L-______ -L ______ ________ L-______ -L ______ ______ 10 -1 10 0 10 1 10 2 10 3 10 4 10 5 Figure 9 - Effect of Mobility Ratio on Shape of Pressure, S = 0, CD = 1000, Mobility = Diffusivity Ratio = 1 D. 2, .5, .1.