Research Progress on Heterogeneity Characterization Methods of Deep Coal-Rock Reservoirs

Abstract

Against the backdrop of global energy structure optimization and the carbon peaking and carbon neutrality goals, deep coalbed methane (CBM) exploitation has become a strategic direction for China's CBM industry, yet the extreme and cryptic heterogeneity of deep coal-rock reservoirs severely restricts its efficient development. Accurate characterization of this heterogeneity is the key to optimizing deep CBM development plans and improving exploitation efficiency. This paper systematically reviews the 2020–2025 research progress of deep coal-rock reservoir heterogeneity characterization methods, elaborating on the multi-scale connotation and characteristics of the heterogeneity at both micro and macro levels. It comprehensively introduces experimental characterization technologies covering micro-nano pore and macro fracture-tectonic characterization, as well as the multi-scale technology fusion workflow. The paper also expounds on quantitative evaluation methods including fractal theory, geophysical and logging data inversion, and numerical simulation, and summarizes the application of machine learning and deep learning in intelligent characterization, which has become the core driving force of this research field. Finally, the paper points out the current key challenges such as difficult deep in-situ characterization, ineffective multi-scale information fusion, the "black box" problem of intelligent algorithms and scarce high-quality data, and proposes future research directions including developing in-situ dynamic 4D characterization technologies, fusing physical mechanisms with data-driven models, constructing deep coal rock big data and knowledge graphs, and applying digital twin technology. This study provides important guidance for the accurate characterization of deep coal-rock reservoir heterogeneity and the efficient exploitation of China's deep CBM resources.