SPE 84848 Forty Years of Steam Injection in California - The Evolution of Heat Management

E.J. Hanzlik and D.S. Mims, Members SPE, ChevronTexaco

Copyright 2003, Society of Petroleum Engineers Inc. This paper was prepared for presentation at the SPE International Improved Oil Recovery Conference in Asia Pacific held in Kuala Lumpur, Malaysia, 2021 October 2003. 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 necessarily reflect any position of the Society of Petroleum Engineers, its officers, or members. Papers presented at SPE meetings are subject to publication review by Editorial Committees of the Society of Petroleum Engineers. Electronic reproduction, distribution, or storage of any part of this paper for commercial purposes without the written consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may not be copied. The abstract must contain conspicuous acknowledgment of where and by whom the paper was presented. Write Librarian, SPE, P.O. Box 833836, Richardson, TX 75083-3836 U.S.A., fax 01-972-952-9435.

Abstract Steam injection EOR began in California forty years ago, and has been highly successful. As a consequence, Californias thermal recovery operations represent a leading source of EOR production in the world. The understanding of most early thermal recovery operators was limited to the concept of heat reduces heavy oil viscosity, and reduced viscosity means more production. Steam injection was attempted in almost any reservoir having viscous oil with little appreciation of other recovery process considerations. Although several early pilot projects were steamfloods, most early applications were cyclic stimulation. During the late 1970s, steamflooding became predominant, and many people considered steamflooding to be a displacement process (hence the term steam drive). With this paradigm and high oil prices, there was little impetus to understand efficient use of heat. The predominant philosophy was If you want more oil, inject more steam. With the later collapse of oil prices, operators returned to review process fundamentals and to determine how to more efficiently operate steam projects. This paper discusses the shift to an override, or gravity drainage, model concept. This helped lead to reduced steam injection and improved thermal efficiency through the use of heat management. This paper discusses the shift to the concept of steam override and gravity drainage as steamflood recovery mechanisms and the subsequent use of heat management practices that improved thermal recovery efficiency. Thermal Recovery in California Early Days: 1960-1966 Early efforts to improve heavy oil productivity in California used bottomhole heaters or in-situ combustion.

Bottomhole heaters did improve well productivity1, but their success was limited by both relatively small heat input rates and the fact that they depend upon conductive heat transfer from the wellbore to the formation. Although some success was obtained with in-situ combustion projects, the high fuel (coke) deposition from the heavy crudes resulted in high air requirements. This, coupled with the difficulty in operating and controlling the in-situ process, limited its application. The first official mention of steam injection in California was cyclic steam injection in the Yorba Linda Field in 1960.2 In 1961, cyclic injection began in the Kern River Field, and in 1962 two additional fields started cyclic steam (Coalinga and McKittrick), and steamflooding began in Kern River. Although companies were beginning to test steam injection, in-situ combustion was still popular, as five new combustion projects were begun in 1962.3 Secrecy surrounded the early steam tests as noted in this reference to testing in the Coalinga Field in 1962, In other portions of the field, recovery of oil by various means of production stimulation appears to be on an increasing trend. The degree of success by these methods is a secret closely guarded by the operators, and little information has been made available to the public.2 However, neighboring operators were observant, and steam generators were difficult to hide. A dramatic expansion of steam injection projects was soon to follow. Cyclic Injection In the three years 1960-1962, cyclic steam injection was tested in four fields. In 1963, cyclic injection was tested in five additional fields, more than doubling the number of fields. However as evidence of success became apparent, there was an amazing increase in the number of applications. Cyclic steam injection was tested in thirty-two additional fields in 1964, and in eighteen more in 1965. Thus, in only six years cyclic steam injection went from a single field test to testing or application in 59 separate fields. Although the earliest tests were generally carefully selected (seven of the first nine fields tested went on to long-lived highly successful steam operations), many of the tests in 1964-1967 were the result of enthusiasm to test the process. Much of this enthusiasm came from the potential to obtain a rapid increase in cash flow and the ability to conduct a test with minimal permitting requirements and the use of rental steam generation equipment. The effects of high temperatures on casing and cement were poorly understood, as were the range of desirable reservoir properties. Thus, a number of tests were short-lived and unsuccessful. However, in only seven years, cyclic steam injection was essentially a full-grown process in California. In

SPE 84848

1966, the first year for which data is available in California DOG annual Reports, cyclic injection totaled 262,000 BSPD (barrels of steam per day, cold water equivalent) injected into 6621 wells in 38 fields (Note: all injection and well count data were taken from California Division of Oil and Gas annual reports, 1966-2001). This volume of steam is very close to the average rate of 279,000 BSPD injected over the last 35 years. Steamflood For the first several years of steam experimentation, interest in steamflooding nearly matched that in cyclic stimulation. By the end of 1963, there were steamflood tests in six fields compared to the nine fields with cyclic tests. However, greater capital requirements and slower production response led to a much slower rate of expansion. By 1966, there were steamflood projects in seventeen fields, injecting 43,000 BSPD using 135 injectors. Steamfloods accounted for only 14% of total steam injection, a very low value compared to the 80-85% values common in later years. Growth Years: 1967-1986 Cyclic Injection As noted above, and shown in Figure 1, the volume of steam used in cyclic projects had reached 262,000 BSPD by 1966. This rate would remain nearly constant until 1981. The number of fields in which cyclic steam was tested would continue to increase at a slow rate until 1981, but the number of fields with active projects would actually decrease until the first oil price shock of 1973 (Figure 2). This decrease was a natural consequence of the exuberant experimentation of 1964-1966. Operators determined the reservoir conditions under which cyclic steam succeeded or failed. As oil prices increased after 1973, the number of active cyclic projects showed a generally increasing trend until 1983. However, it is interesting to note that there was relatively little change in the amount of steam used, or in the number of wells receiving cyclic stimulations (Figures 1 and 3). Some of the larger stimulation projects were now being converted to steamfloods, and the new projects tended to be smaller. Although some of these projects were in fields which had not previously had steam injection, others were in fields which had seen prior testing. Steamflood The volume of steamflood injection began to rapidly increase from 1967 to 1973, but the number of fields with active steamfloods remained essentially constant at 10-14 (Figures 1 and 2). The injection rates increased by a factor of 8.5 from 48,000 BSPD to 411,000 BSPD - in this six-year period. The success of early pilot projects was being followed by project expansions and a steady, but slow rate of experimentation in additional fields. Following the price increases of 1973, there was a rapid expansion of the number of fields with steamflood projects. A peak of 24 fields with active steamfloods was reached in 1977, and the number would range between 19 and 23 for the next ten years. However, the number of steam injectors and the volume of steam being injected would continue to grow dramatically from 1973 through 1985. Steamflood injection volumes

increased by another factor of 4.4, to nearly 1.9 million BSPD, and the number of steam injectors increased from 1136 to 5001 (Figures 1 and 3). Total Steam Injection During these growth years, oil prices increased greatly. There was a small price drop in 1982, but prices were stable in 1983-1985. This high price environment encouraged rapid expansion. Steam injection grew at a rate of nearly 11% per year during this period, essentially all in the form of increased steamflood injection. The percentage of steam used in steamflood projects grew from 15% to 85% during this period, and steamflooding continues to account for 80% of steam injection (Figure 4). California reached its peak steam injection rate in 1985. Combined cyclic and steamflood injection that year was 2.07 million BSPD into 13,798 wells. There were 8794 cyclic stimulation wells and 5001 steamflood injectors. Achieving Maturity In 1986 California heavy oil prices collapsed from a value of $20/barrel to less than $9/barrel. Steam injection rates began a general downward trend which has continued to the present. There was not an immediate drop in thermal oil production, as recently expanded steamflood areas were in the early stages of production response. Maximum thermal recovery oil production was reached in 1988 at 511,000 BOPD, after which it began a steady decline (Figure 5). This combination of stable or declining production and lower prices forced operators to more closely examine their operating procedures. Although Californias average steam/oil ratio (SOR) had dropped somewhat from its peak due to production response from immature steamfloods, it was still nearly 4.5 when the 1986 price collapse began. Since generating steam is the largest operating expense item in a steamflood, efficient use of steam began to receive much more attention. This emphasis has continued to the present, and has become even more critical as larger numbers of projects pass their peak production period. Heat Management Optimization of heat use in a steamflood requires an understanding of the physics which govern the process and having methods in place to adequately monitor and analyze those factors. The early analytical tools available to steamflood designers for estimating heat losses and the growth of the steamchest were based upon frontal advance theory, e.g., Marx and Langenheim4, and Myhill and Stegemeier5. With the frontal advance concept, higher steam injection rates both reduce ultimate heat losses to the underburden and overburden, and result in a more rapid recovery of oil. Thermal observation wells were drilled in early projects to provide verification of the rate of advance of the steam as predicted by the frontal advance models. However, post-flood cores taken to evaluate early pilots showed that gravity override was a significant factor6. Thus, there was significant inconsistency between the analytical models and the physics of the process (Figure 6). However, many engineers still

SPE 84848

perceived steam injection in terms of a displacement process in which higher injection rates equated to higher oil production rates. This intuitive belief in the benefits of high injection rates was reinforced by the belief that high injection rates would heat more oil faster with a consequent increase in production rate. Also during the field expansion of the process, steam would migrate to adjacent unheated vertical and horizontal areas of the field causing oil production there to increase, thus reinforcing this belief. Thermal reservoir simulation became widely available in the mid-1970s. These simulators included the required physics, and showed the effect of gravity override. However, they were relatively difficult and expensive to use. In addition, limited computer power required studies to use relatively small number of grid cells, which tends to smear fluid interfaces and results. Although the simulation results were generally directionally correct, many operating unit personnel had limited faith in the results. In the mid-1980s, two new analytical methods (Vogel7 and Neuman8) became available which are based upon the gravity override effects. Some key conceptual differences are that: (1) the dominant production mechanism for viscous oil steamfloods is gravity drainage of the heated oil, (2) oil rates are largely unaffected by steam rates higher than the minimum required to maintain the steamchest, (3) injection rates should be proportional to project area (not project volume), and (4) decreasing injection rates with time minimizes heat losses to produced fluids and casing blow. Figure 7 shows an example of the difference in heat requirements to maintain a steam chest at year one versus year five of a typical California steamflood pattern. The heat requirement in the fifth year is less than half the requirement in year one. The publication of these two analytical methods (1984 and 1985) was timely, given the price collapse which would occur in 1986. Early Monitoring and Management Examples Monitoring the use of injected heat in steam projects is not a new concept. The Deerfield steamflood pilot conducted by Carter Oil Company (now part of Exxon) in the 1950s included thermal observation wells. As cited in the SPE paper reporting on the pilot, Temperature data obtained from the thermal observation wells installed in the center five-spot enabled us: (1) to write energy balances on the reservoir and to determine the energy distribution, (2) to correlate energy requirements with steam front advance and (3) to form a picture of fluid movement through the oil sand.9 Their analysis included energy lost to produced fluids and fluids which migrated outside the pattern. Similarly, the concept of varying steam injection rates during the life of a steamflood was independently shown by both reservoir simulation and field testing in the early 1970s. Conclusions of reservoir simulation studies by Chu and Trimble10 included: (1) the optimal constant steam rate is proportional to pattern size rather than sand thickness, (2) economics of a project can be improved over the constant rate case by using higher steam rates in the initial stages and then decreasing the rate with time, (3) decreasing rate in a hyperbolic fashion appeared superior to a linear decrease. Bursell and Pittman6 reported that rate reductions late in the life of three pilots in the Kern River Field resulted in improved

steam/oil ratios. All of these observations are consistent with our current understanding of the steamflood process. Recognition of Heat Management by California Operators As noted in the preceding section, by 1975 there were examples of how to conduct heat balances and evidence supporting the steam over-ride/gravity drainage model, including the principle of reducing injection rates in mature projects. However, rapidly increasing oil prices and a mental analogue of a displacement mechanism for the steamflood process led to a common belief that higher injection rates automatically resulted in increased oil production rates. Figure 8 shows the average injection rate per well and average SOR for steamflood operations in four of the largest California thermal fields (Coalinga, Kern River, Midway-Sunset, and South Belridge). It shows that in the late 1960s, and with current heat management concepts, average injection rates for steamfloods were about 225 BSPD/well. From Figure 7 we see that early life steam chest heat minimum requirements at 60 psig are about 70 MMBTU/day, or approximately 200 BSPD. Thus the rates used in the first steamflood projects were quite reasonable. However, we also see that average rates per well increased to more than 300 BSPD/well during the 1970s, and did not drop below 300 BSPD/well until 1992. It appears that most operators did not carefully review overall heat and mass balances for their steamflood projects during this period, and they generally appear to have had too high a rate of heat injection. During the 1990s an increasing amount of effort was applied to heat management11, which resulted in a general decrease in the amount of steam injected in each pattern. The significant consolidation which occurred in the thermal industry in the past seven years (from six major operators to two) has also resulted in more rigorous application of heat management. Besides the emphasis placed on heat management by the surviving operators, control of larger contiguous field areas has made coordination and optimization of injection operations easier. Despite the present maturity of many projects and the accompanying production declines, Figure 8 shows that it has been possible to maintain a nearly constant average SOR for the four large fields, despite significant production declines. In fact the SOR for the Kern River field was at an historic low in 2001. Heat Management Tools As shown in Figure 9, heat management is a basic energy balance. Note that only a few key parameters can be measured on the surface steam in, production and casing venting out. Some parameters can be directly measured subsurface using logs temperatures of the overburden, underburden and reservoir. Other information can only be inferred. There are three approaches for calculating energy requirements: analytical methods and reservoir simulation which require assumptions to estimate heat requirements, or direct methods which account only for energy components that can be measured on the surface or by sub-surface logging tools. In this paper we will not discuss reservoir simulation heat management tools.

SPE 84848

Analytical Methods The basic analytical methods are those of Vogel, and Neuman. Neuman is generally used prior to steam breakthrough. After the steamchest has expanded across the project area, heat requirements can be set based upon extensions of the basic concepts expounded by Vogel. The computations are a function of: Initial reservoir conditions/properties, Mass and heat content of injected and produced fluids, Steamchest temperature and pressure (from temperature observation wells). Time on injection, and Area of the project/control volume.

Unknown variables are influx and efflux from the control volume. However, simultaneous solution of the mass and energy equations may allow estimates of steam migration, water encroachment, heat scavenging, or steam chest collapse. Although the analytical methods use an overly simple conceptual model (do not account for Darcy flow, no geologic detail), and require high quality input data, they have the advantage of emphasizing energy control for operational decisions. Direct Methods Direct methods have been developed for use where there are significant numbers of temperature observation (TO) wells. They provide simple and fast interpretation to account for reservoir heat accumulation and losses to the overburden/underburden. The techniques also are easily exportable to 3D/4D visualization, which assists with analysis of large data sets and project areas. Conductive heat loss to the overburden/underburden can be easily calculated by taking the dT/dz value directly from a temperature log. The rate of heat loss is then quickly determined by the function: Qob = kA dT/dz, where k is the overburden heat conductivity, A is the project area, and dT/dz is the vertical temperature gradient. Similarly, the heat accumulation rate can be estimated by taking the difference in the average heated zone temperature between successive time-lapse temperature logs and calculating the rate of heat input required to heat the formation in the control volume: Qr = MrVrdTavg/dt, where Mr is volumetric heat capacity, Vr is the project volume, and dTavg is the change in the average project temperature over the time period dt. This simple formulation assumes no changes in volumetric heat capacity or project volume with time. Where there are sufficient TO wells, the direct method offers the advantage of simplicity and fewer assumptions when determining the required injection during routine monitoring after establishing a steamchest. However, the direct methods do not take into account heat losses that occur beyond the area being analyzed (e.g., production wellbore, surface lines, influx and efflux).

Application of Heat Management Figure 10 shows the heat injection history of a project along with a comparison of the heat calculated by several analytical methods and the TO well direct method. One can see that there is general agreement between the two analytical methods which include heat losses from production, and the data from the direct TO well method. The Vogel method without consideration of production losses under-predicted heat requirements. It is also evident that there was a period in 1994 during which this project was over-injected. This was followed by a period of slight under-injection, and then balanced injection. Figure 11 demonstrates how decreasing the injection rate does not necessarily translate to a loss in oil production rate. This project was being greatly over-injected in 1992-1993. Heat injection was reduced by approximately 60% in two steps, yet we see that oil production was not adversely affected. Similarly, Figure 12 shows another project in which reduction from 15,000 BSPD to 4,000 BSPD in two steps over a period of two years resulted in no change in the wellestablished production decline behavior of the reservoir. Finally, Figure 13 shows a case where a project which had adequate injection and well density showed no response to increased steam injection and increased well count. If a project is receiving sufficient heat and has sufficient well density, the rate is fixed by the rate of gravity drainage increased heat and well density will not accelerate oil recovery. Heat management is an interactive reservoir management process. It should involve production monitoring, temperature surveillance, an integrated reservoir model, and surface steam measurement/control. Although this paper has concentrated on the subsurface aspects, measurement and control of steam distribution (including liquid/vapor phase splitting) in surface facilities and the wellbore is also very important. Conclusions Thermal recovery in California is at mature stage in its life cycle. Widespread adoption of demonstrated technical principles (e.g., heat management) may not occur until forced by commercial circumstances. Increased efficiency has come from both and improved understanding of the steamflood process and consolidation of the industry. Modern heat management techniques are improving process performance and extending field life.

Summary of Operations, California Oil Fields, Vol. 46, No. 2, California Division of Oil and Gas, San Francisco, 1961. 2 Summary of Operations, California Oil Fields, Vol. 52, No. 2, Part 1, California Division of Oil and Gas, Sacramento, 1967. 3 Summary of Operations, California Oil Fields, Vol. 48, No.
Taken from slide presented by M. R. Guzman, ChevronTexaco, personal communication.

SPE 84848

2, California Division of Oil and Gas, San Francisco, 1963. Marx, J. W. and Langenheim, R. H.: Reservoir Heating by Hot Fluid Injection, Trans AIME (1959), 216, 312-314. 5 Myhill, N. A. and Stegemeier, G. L.: Steam Drive Correlation and Prediction, SPE 5572 presented at SPE 50th Annual Fall Meeting, Dallas, Sept. 28- Oct. 1, 1975; J. Pet. Tech. (Feb. 1978), 173-182. 6 Bursell, C. G. and Pittman, G. M.; Performance of Steam Displacement in the Kern River Field, J. Pet. Tech., (August 1975), 997-1004. 7 Vogel, J. V.: Simplified Heat Calculations for Steamfloods, J. Pet. Tech., (July 1984), 1127-1136. 8 Neuman, C. H.: A Gravity Override Model of Steamdrive, J. Pet. Tech., (January 1985), 163-169. 9 Valleroy, V. V., Willman, B. T., Campbell, J. B., and Powers, L. W.: Deerfield Pilot Test of Recovery by Steam Drive, J. Pet. Tech., (July 1967), 956-964. 10 Chu, C., and Trimble, A. E.: Numerical Simulation of Steam Displacement Field Performance Applications, J. Pet. Tech., (June 1975), 765-776. 11 Ziegler, V. M., Crookston, R. B., and Sanford, S. J.: Recommended Practices for Heat Management of Steamflood Projects, SPE 25808, presented at International Thermal Operations Symposium, Bakersfield, Feb. 8-10, 1993. __________________________________________________
4

100

80
Cum. Cyclic Cum. Steamflood Active Cyclic Act. Steamflood

60

Fields
40 20 0 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Year
Figure 2 Number of Fields With Steam Projects

2500
Stimulation Total

14000

1977 Data Missing

12000

2000

Number of Injected Wells

Steam Injection (MBSPD)

10000

1500

8000

1000

6000

4000

500
2000

Stimulation Steamflood

0 1965 1970 1975 1980 1985 1990 1995 2000 2005

0 1965 1970 1975 1980 1985 1990 1995 2000 2005

Year
Figure 1 California Annual Steam Injection

Year
Figure 3 Wells on Injection

SPE 84848

100 90 80

Frontal Displacement

Gravity Drainage

Percent of Steam Injected

Steam
70

Steam
60 50 40 30 20 10 0 1965

Oil

Oil

Oil Rate Most Dependent On Injection Rate

Oil Rate Most Dependent On Reservoir Parameters

1970

1975

1980

1985

1990

1995

2000

2005

Figure 6 Mechanistic Concepts

Year
Figure 4 - Growth of Steamflood Injection
90 80
600 6

70 60 50 40 30 20 10
t= 1 Year t= 5 Years
Assumptions: Static aquifer 2.5 Acres Overburden & reservoir conduction and convection Ti = 90F

500

400

MBOPD

300

200

100

MBOPD SOR

SOR

MMBTU/Day

0
0 2005

20

40

60

80

100

120

Steamchest Pressure (PSIG)

0 1960 1965 1970 1975 1980 1985 1990 1995 2000

Year

Figure 7 Heat Required to Maintain a Steamchest

Figure 5 Thermal Production and Steam/Oil Ratio

SPE 84848

2500

600
Note: Data not available to calculate SOR prior to 1975. No wellcount for 1977.

6
2000

500

5
MMBTU/Day
1500

Calc. w/TO Wells Vogel (no prod losses) Vogel (w/prod losses) ChevronTexaco Template Injection

Bbls Steam/Day/Injector

Steam/Oil Ratio

400

1000

300

500

200

100

BSPD/Injector Steam/Oil Ratio

0 4/1/1993

4/1/1994

4/1/1995

4/1/1996

4/1/1997

4/1/1998

4/1/1999

Date

0 0 1965 1970 1975 1980 1985 1990 1995 2000 2005

Figure 10 Comparison of Analysis Methods

Year
Figure 8 Largest Four Fields Performance
40 BOPD/Active Well Project BSPD Heat Requirement 20000

35

17500

30

15000

25

12500

Input - Output = Accumulation

Injected Steam m s qs Produced Liquids & Gases m o qf m w qw m cb qcb

20

10000

15

7500

Overburden Losses, o Steam

10

5000

Fluid Influx m i qi

Liquid Region

2500

Fluid Efflux m e qe

0 0 1/1/92 7/1/92 1/1/93 7/1/93 1/1/94 7/1/94 1/1/95 7/1/95 1/1/96 7/1/96 1/1/97

mInjection + mWater Influx = m(o,w)Prod. + mCB + mEfflux + mZ(Accumulation) qS Injection + qi Water Influx = q(o,w) Prod. + qCB + qe Efflux + qob + qz (Process)

Date

Figure 11 Example Reduction in Excessive Steam Rate

Figure 9 Heat and Mass Balance

Project Total BSPD

BOPD/Active Well

SPE 84848

16000 8000 First Reduction: 15,000 to 8,000 BSPD

14000

12000 6000 10000

BOPD

8000 4000 6000

2000 Second Reduction: 8,000 to 4,000 BSPD

4000

2000

0 1/1/99

7/1/99

1/1/00

7/1/00

1/1/01

0 7/1/01

Date
Figure 12 Impact of Heat Management

500

400

Bbl/Day

10000

300

200

BOPD BSPD Active Prod. 1000 1975

100

1980

1985

1990

1995

0 2000

Year

Figure 13 No Response to Increased Steam and Well Count

Active Producers

BSPD