Research on the Relationship Between Micropore Structure and Gas‐Bearing Capacity in Tight Sandstone Using NMR and Deep Learning

ABSTRACT

In tight sandstone reservoirs, some formations share similar lithological, logging, and physical characteristics, yet exhibit significant differences in gas‐bearing capacity. Therefore, how to accurately characterize the relationship between micropore structure and gas‐bearing capacity remains a challenging task that needs to be resolved. To address this, a new impact ranking analysis process based on nuclear magnetic resonance (NMR) experiments and deep learning has been proposed to better understand the impact of micropore structure on gas‐bearing capacity. This process primarily includes two parameters: bound water saturation and effective pore volume ratio obtained from NMR experiments. The adjacent point weights of these two parameters are the highest. The Kruskal–Wallis value and the adjacent point weight of the pore fractal dimension are also relatively high, indicating a secondary influence. Combined with the NMR T ₂ spectrum of reservoirs with different gas‐bearing capacities, this process effectively reflects the relationship between microscopic pore structure and gas‐bearing capacity. Blind well test results further confirm this relationship. Overall, the proposed process can effectively guide the efficient development of these reservoirs and holds broad application prospects for the oil and gas industry.