Characterization of Enhanced Coalbed Methane Recovery Injection Wells

Abstract An integral part of enhanced coalbed methane recovery (ECBM) is injection of flood gas into a coal deposit via injection wells. Understanding of these wells is critical to maximizing coal gas recovery. Although ECBM is a relatively new and rapidly evolving technology, existing gas injection well technology can be readily adapted to ECBM injectors. This paper discusses characterization of two nitrogen injectors using spinner surveys, Hall plots, reservoir simulation, and pressure transient analysis. Using actual field data, this paper addresses nitrogen injectivity, conformance of injected gas, reservoir sweep, and implications for coal gas recovery. Introduction Over the last decade, coalbed methane has evolved into a significant domestic gas resource. As of year end 1995, US domestic coalbed methane production was 973 bcf, approximately 5% of US gas consumption while coal gas reserves are 13 tcf, roughly 6% of proved US domestic gas reserves. To recover a large portion of this resource, reservoir pressures in the coal deposit must be low. Because coals store gas primarily by sorption, reducing reservoir pressure to one fourth ifs original value will release only about half the gas sorbed to the surface of the matrix. Injection of inert gases, such as nitrogen, reduce the methane partial pressure and strip additional methane from the coal. This process is called enhanced coalbed methane (ECBM) recovery and, when properly done, results in increased gas rates and recoveries. Characterization of injection wells in ECBM field projects is crucial to understanding field response and managing the reservoir for maximum value. The purpose of this paper is to document the use of conventional gas injection well technologies to help understand two ECBM injectors. Following a general description of the wells, step rate tests, Hall plots, spinner results, pressure falloff tests, and simulation studies are discussed. General Well Description This ECBM project involved several producers and two centrally located injectors, as shown in Figure 1. The coal deposit was approximately 3300 ft deep and comprised of three layers with nominally 75 ft of net coal in a 105 ft gross interval. Both wells had been completed in all zones as case through, perf, and frac producers for a number of years before being selected for conversion to nitrogen injectors. Figure 2 shows Well A configured for injection. Over the course of the seven month project, a total of 115 MMCF of nitrogen was injected into Well A at an average rate of 535 MCFD while total nitrogen injection into Well B was 196 MMCF at an average rate of 1142 MCFD. Step Rate Tests Step rate tests were conducted in both injectors to determine fracture parting pressure. Beginning with a nitrogen injection rate of 400 MCFD, the rate was stepped up by 400 MCFD every two hours up to a maximum injection rate. Maximum injection rate in Well A was 3.2 MMCFD at a calculated bottomhole pressure of 1,660 psia. Throughout both tests, operational constraints made it difficult to maintain constant injection rates, reducing confidence in the data. The high rate data is especially suspect. One of the key learnings from these tests, which can be carried forward to future step rate tests of enhanced cbm recovery injectors, is their sensitivity to rate variations and, consequently, every effort should be made to maintain stable rates within each given step. Interpretation of both these tests was difficult due to their unorthodox behavior. Final rates and pressures for Well B are plotted in Figure 3. Plots of rate and bottomhole pressure from a step rate test typically show a linear trend at low rates and pressures followed by a breakover to a second trend with a shallower slope at high rates and pressures. The intersection of the two lines is interpreted as formation parting pressure. As seen in Figure 3, Well B shows the opposite behavior. That is, the slope of the high rate and pressure data, the second trend, is steeper, not shallower, than that of the low rate and pressure data. P. 171^