METHOD FORECASTS COALBED METHANE PRODUCTION
Peet M. Soot
Northwest Fuel Development Inc.
Portland, Ore.
A method has been developed to predict the cumulative production from coalbed methane wells in the Black Warrior and San Juan basins.
The method is not precise for individual wells, but is statistically significant for groups of wells.
The technique involves ranking individual wells against a series of cumulative production curves. These generic curves are based on computer modeling of wells in the respective basin. The cumulative production curves provide the basis for projecting future production.
METHODOLOGY
Computer models for individual coalbed methane wells have been developed by a number of entities. Northwest Fuel Development Inc's method is based on Gas Research Institute/ICF Resources Inc.'s Comet model.1 The base-case parameters for the respective basins are as shown in Table 1. It is argumentative whether these parameters are precisely representative of the respective basins, but they are adequate to provide the generic production curves for the NW Fuel method.
The cumulative production curves generated from the Comet model (using the aforementioned parameters) for the Black Warrior and San Juan basins are presented in Fig. 1, respectively.1
The base-case curves from the Comet model were used to generate a series of cumulative production curves. The Black Warrior curves are presented in Fig. 2a.
One should note that for convenience, the rank value of an individual curve corresponds closely to the expected cumulative production in 10 years, e.g., Rank 1 = 106 MMcf, Rank 5 = 530 MMcf.
For the San Juan case, the 10-year production has been adjusted to be exactly 100 MMcf for Rank 1 wells and proportionately higher for wells in other ranks Fig. 2b.
Some of the high-production wells in the San Juan basin are not expected to follow the Comet model curves. These wells are in overpressured zones that do not behave like typical coalbed methane wells.
The ranks for these high-production cases were modified. Any well with a rank greater than 30 was cut in half to account for the different behavior. This modification should be studied more but for this analysis the estimate is sufficient.
Polynomial equations were fit to the curves in Fig. 2 so that individual wells could be assigned ranks by computer analysis rather than visual correlation.
Due to the double flexure in the generic curves, several polynomial equations were needed to describe each curve. As an example, three different equations were used for the San Juan curves.
The first equation covered production through the first 2 years. The second was for years 3 through 6. The last 4 years had a third equation.
A computer program assigned a rank to all producing wells in both the Black Warrior and San Juan basins using the appropriate polynomial equation.
One can also determine the rank for a given well graphically by using Fig. 2. By knowing the well's cumulative production and the time on-line, one can determine the well rank from the curves in these figures.
A similar ranking process was applied to the water production from wells in the respective basins. Fig. 3 shows the water production well ranks for the Black Warrior and San Juan basins.
These water production ranks are also based on Comet model simulations and comparisons with published water production data. Computer rankings were calculated with polynomial equations of these well rank curves. Visual rankings could be made from the graphs presented in Fig. 3.
RESULTS
A total of 641 conventional coalbed methane wells were in production in the Black Warrior basin at the end of 1989. By November 1991, there were 1,364 vertical producing wells. These totals exclude the entire Brookwood field where most of the wells are in-mine horizontal or "gob" gas wells. Also excluded are the Oak Grove field horizontal and gob wells. These were not included because the Comet model does not simulate these types of wells.
Horizontal well production could be simulated with models found in the literature, but that effort was beyond the scope of the present study.
Gob gas wells produce gas from the collapsed overburden at underground coal mines that use the longwall mining technique. Computer models simulating production from these wells are impossible to use without detailed information regarding the local geology. Hence, these wells were deleted from consideration in the present study.
The results from the 641 wells on-line before 1990 indicate that over 95% of the wells are Rank 4 or below, i.e., they will not produce more than 450 MMcf in 10 years.
Only 11 wells (less than 2% of the total) exceed Rank 7. The average rank for the Black Warrior basin was 1.8, which indicates 10-year production of less than 200 Mmcf.
The consistency of the NW Fuel method was evaluated by comparing ranking results over a 4-month period from Oct. 1, 1989, through Jan. 31, 1990. It was determined that the Black Warrior wells, which are currently producing gas, follow closely the series of curves in Fig. 2a.
A sampling of 56 wells from three different fields in the basin demonstrates that the rank for a given well can vary month to month, but the average rank of the entire sample hardly changes.
There are negative changes to offset most positive moves. The newer wells are somewhat less predictable. Forty-four percent of the wells on-line less than 1 year at the beginning of the test period changed rank. Rank changes of wells older than 1 year wells were 34%.
As noted in the calculation of averages, many of the individual rank changes offset each other.
Subsequent analysis of data through October 1991 corroborates these previous conclusions for the Black Warrior basin. The basin-wide average rank had increased to over 2, but most of this improvement came from a single field within the basin.
The basin average, excluding the Cedar Cove field, still had a 1.8 rank. This leads one to the conclusion that the NW Fuel Method can provide an indication of future production for a group of wells, but it may deviate for any given well especially on a month-to-month basis.
The Oak Grove field (Black Warrior basin) was studied in detail because there is considerable published geologic information for that area.
Ranks for individual wells along with approximate locations for faults on the west side of the Oak Grove mine were plotted. The fault locations were estimated from a paper presented by Pashin, Chandler, et al.1
It was found that there is a definite correlation between well rank and location relative to the faults. The best wells appear to be on the upthrown side of the faults, although there were isolated exceptions to this general rule.
The proximity to the mine is also an important variable. It is hypothesized that the mine has a tendency to dewater the coal seam. Wells near the border of the mine are higher rank than those farther away.
Results for the San Juan basin are plotted on a map in Fig. 4. This map illustrates that a few townships have very high reserve values relative to the entire basin.
Townships 30n, Ranges 6w and 7w have average well ranks exceeding a value of 30. The basin as a whole has a rank average of less than 10. Most of the geographic extent of the basin has well ranks less than 5.
RECOVERY POTENTIAL
The NW Fuel method was used to estimate the percentage of coalbed methane originally in place that could be recovered from virgin coal. The Oak Grove field was used as an example because some wells in this field have been on-line for nearly 10 years.
This should be fairly representative of the Black Warrior basin. A 32 section area in Townships 18-19s, Range 6w was studied. Estimates for original gas-in-place (OGIP) were taken from the GRI/ICF report for the Black Warrior basin.
Part of this area is within the highest OGIP contour in the basin. The area exceeds 9 bcf/sq mile.
The average rank was calculated for all wells drilled within a section. Not every section had an equal number of wells.
To normalize the estimates, 16 wells were assumed to be drilled in each section, i.e., 40 acre spacing.
All 16 wells within the section had their recovery averaged by the wells already drilled and in production.
The average for all 32 sections studied is 43% recovery of the OGIP in 10 years. This average is raised by the sections where nearly all, or in one case greater than 100%, of the OGIP will be recovered. More than one third of the sections will produce less than 20% of the OGIP within 10 years. The median value is 27% recovery.
One can conclude that drilling in a random virgin area within the basin should be projected to produce less than one half of the OGIP. If proper well selections are made with knowledgeable coalbed methane geologists, then this recovery percentage can be improved.
PRODUCTION CORRELATIONS
A report published by GRI evaluated initial and peak production rates from coalbed methane wells in the Black Warrior basin.3 The Geological Survey of Alabama worked as a subcontractor to the University of Texas in reviewing variables which might affect coalbed methane production rates in the basin.
There were several surprising results. No variable provided a meaningful correlation with initial or peak production rates: not coal thickness, not well depth, not well spacing, not permit number. (The latter variable might have shown improvements in completion techniques with newer wells, i.e., higher permit numbers).
A weak correlation was shown between coalbed methane production and water production. This may be a function of higher permeability for gas and water within the coal in certain areas.
Completion types were also evaluated. None of the techniques (water, foam, gel, combination) were significantly better than the alternatives. The foam fracs may be judged somewhat less effective than the other options.
The researchers' conclusion was that geologic structure and proximity to coal mines were the primary variables affecting coalbed methane production. The northwest-oriented faults were nonproductive, but the northeast oriented lineaments provided significant conduits for coalbed methane flow.
Nearby coal mines would also enhance production by removing water from the coal seam. This conclusion was supported by the fact that a much larger percentage of the high production wells did not even require stimulation.
This is fully supportive of the conclusions drawn by the application of the NW Fuel method to the Black Warrior basin.
REFERENCES
- Sawyer, W.K. , Zuber, M.D., Kuuskraa, Y.A., and Horner, D.M., "Using Reservoir Simulation and Field Data to Define Mechanisms Controlling Coalbed Methane Production," Proceedings of the Coalbed Methane Symposium, Tuscaloosa, Ala., Nov. 16-19, 1987, pp. 295-307.
- Pashin, J.C., Chandler, R.V., and Mink, R.M., "Geologic Controls on Occurrence and Producibility of Coalbed Methane, Oak Grove Field, Black Warrior Basin, Alabama," Proceedings of the Coalbed Methane Symposium, Tuscaloosa, Ala., Apr. 17-20, 1989, pp. 203-09.
- Pashin, J.C., Ward, W.E. II, Winston, R.B., Chandler, R.Y., Bolin, D.E., Hamilton, R.P., and Mink, R.M., "Geologic Evaluation of Critical Production Parameters for Coalbed Methane Resources, Part 2-Black Warrior Basin," Gas Research Institute, Annual Report, GRI-90/0014.2, August 1988-July 1989.
Copyright 1991 Oil & Gas Journal. All Rights Reserved.