Finding an Optimal Gas Injection Scenario for Coal Bed Methane Recovery Enhancement
Ky Anh DONG, Martin MANSER, Abdul Rahman MOHD ALIAS - 2014
Australian School of Petroleum
University of Adelaide
Since 1980, scientists have conducted research about gas injection into coal seams target to improve methane (CH4) recovery. Findings show injecting Carbon Dioxide (CO2) induces CH4 desorption from coal matrix and also sequestrates greenhouse gases. Due to CO2's high adsorption affinity, CO2 causes matrix swelling and decreases gas injectivity. Nitrogen (N2) injection on the other hand does not have the swelling issue, but leads to an early breakthrough of N2 which deteriorates the quality of gas production. This is due to the low sorption affinity of coal on N2. Recently, several simulation-based studies have produced better findings with injecting a mixture of CO2 and N2. For most of the studies, optimum gas injection composition was found based on the following objective, maximizing CH4 recovery with low N2 production. In all of these studies, the composition of the injected mixture was constant throughout the injection period. In this research project, it is shown that the aforementioned method is not the best-obtained optimal injection strategy as CH4 production can be further increased. Therefore, an alternative approach is presented to increase the gas injection performance. This approach involves varying CO2 fraction over multiple injections to limit the severity of swelling.
The objective of this study is to find optimal gas injection scenario with varying gas compositions. This is achieved by using a methodology involving a series of sensitivity analyses using a compositional flow simulator (ECLIPSE-E300) as this study requires the simulation of multiple components. The sensitivity analyses consist of three parts. Part one investigates varying fractions of CO2 over multiple injections. This pattern is then compared with alternating injections of pure N2 and CO2. Furthermore, it is advantageous to consider the effect of adsorption time and CO2 sequestration which was investigated in parts two and three respectively. The described methodology is first evaluated using ECLIPSE-E300 which includes extended Langmuir isotherm and modified Palmer Mansoori model are applied. The simulation was then validated against a previous work of Shi and Durucan (2008) to ensure these models have been applied. The application of extended Langmuir isotherm allows for the multi-component adsorption/desorption process to be investigated through the whole simulation study. Additionally, the modified Palmer Mansoori model was used to simulate the permeability changes due to swelling and shrinkage. A semi-synthetic model is constructed to further evaluate the methodology.
Results confirmed that, a pattern involving sequential injections of CO2 produce better results compared to an alternating pattern. The sequential pattern draws the following advantages:1) Higher CH4 production total, 2) deferment in permeability reduction, 3)delays the N2 breakthrough and 4) larger CO2 sequestration.