Underground coal seam gas displacement by injecting nitrogen: Field test and effect prediction

Abstract

The existing ground well gas injection displacement engineering shows that the technology of displacement by injecting gas is a coal gas stimulation technology with good application effect. Based on the field test of underground nitrogen injection to promote the discharge of coal seam gas in Liuzhuang coal mine, Anhui Province, China, the variation laws of the flow rate of mixed gas, the composition of mixed gas, the pure flow rate of methane in gas producing borehole are analyzed. Meanwhile, the laws of “discharging while injecting” and conventional emission are compared and analyzed. The results show that the injected nitrogen with a higher pressure than the coal seam gas can effectively increase the pressure gradient between the fracture and the gas-producing boundary, improving the mixed flow dynamics inside the coal seam. Compared with the mixed gas flow rate at the end of conventional emissions, the average mixed gas flow rates in each gas producing borehole after injecting nitrogen increase by 2.20 ~ 5.83 times, and the average pure flow rates of CH4 increase by 1.42 ~ 2.66 times. The technology of displacement by injecting nitrogen can significantly promote gas emission from coal seam. Compared to the prediction values of conventional emission in the corresponding gas producing boreholes, the total gas emission increase by 64.9%, 222.5%, 132.3% and 149.1% under the condition of displacement intervention, respectively. The output process of underground coal seam gas displacement by injecting nitrogen can be divided into two stages approximately. In stage I, driven by the pressure gradient, CH4 in the original free state is the main product. While in stage II, the desorbed CH4 from adsorbed state is the main product driven by the concentration gradient. In addition, a back-propagation artificial neural network is applied to the effect prediction of coal seam gas displacement by injecting nitrogen. The root mean square error of gas pure flow rate in each gas producing boreholes is 0.0075, 0.0266, 0.0492 and 0.0379 respectively, which is in good agreement with the actual measured values. It shows that the application of artificial neural network to the prediction of nitrogen injection displacement effect is feasible.