Analysis of Logging Identification and Methane Diffusion Properties of Coals with Various Deformation Structures

Abstract

Methane diffusion in coals with different deformations is important for gas resource evaluation. In this study, the adsorption/desorption and diffusion behavior in different coal structures were investigated, and the multi-logging parameter model of back-propagation (BP) neural network was employed to identify coal structures, as well as the effects of diffusion changes on gas contents in the study area, were revealed. The results indicate that with the increase of coal deformation intensity, the Langmuir VL increases, and both specific surface area and pore volume increase as well. Diffusion coefficients in coals decrease gradually with the decrease of pressures, and the development of meso/macropores and micropores in tectonic coals promotes gas migration, which increase the diffusion efficiency, especially the fast and reduction diffusion stages (S1 and S2). Comparisons between intact and tectonically deformed coals in terms of gas diffusion behaviors modeled by the unipore and bidisperse numerical methods show that the analytical solution of the bidisperse model is more appropriate to describe diffusion behaviors in tectonic coals due to the enhanced complexity of mesopores and micropores in coals. Different coal structures exhibiting similar values in well logs lead to multiple interpretations of coal structures. The proposed BP model is an effective way to identify the nonlinear relationships between well logs and coal structures by extracting key information from logging datasets. The modeled results show that the degree of heterogeneity of the gas content in different coal structures is high and the higher diffusion rate in granulated coal zone somewhat reduced the measured gas content.