Evolution of Water Distribution and Connectivity in Coal during Sequential Imbibition Drying: Implications for Water Injection Effectiveness

Summary

Coal seam water injection is widely recognized as an effective mitigation measure for dynamic disasters in deep mining, yet its field performance remains highly variable, as the injected water continuously redistributes and dries before extraction. In this study, we aim to quantify the evolution of pore-scale water distribution and hydraulic connectivity during sequential imbibition and drying, and to demonstrate how this quantification can inform predictions for the injection-to-production window. To achieve this, we systematically characterized pore water evolution using nuclear magnetic resonance (NMR) imaging (NMRI) and T2 spectra, quantified its dynamic heterogeneity and hydraulic connectivity through multifractal analysis, and developed a coal-specific fractal model for predicting spontaneous imbibition (SI) dynamics. Specifically, NMRI shows imbibition advances longitudinally with transverse diffusion, while drying proceeds inward from the surface, leaving residual water concentrated in the sample interior. T2-based partitioning indicates that adsorption pores dominate early imbibition, whereas seepage pores and fractures lose water rapidly during drying, ultimately constituting less than 5% of the final residual water content. The proposed fractal model accurately captures the linear relationship between imbibed water weight and the square root of time (R2 > 0.97). Multifractal descriptors provide process-scale indicators: The generalized dimension spectrum width (ΔD) correlates linearly with total water content, quantifying distribution heterogeneity, and the Hurst exponent (H) tracks the evolution of water distribution correlation across pore sizes, serving as a statistical proxy for hydraulic connectivity changes during imbibition-drying. These metrics offer a quantitative basis for understanding and potentially optimizing coal seam injection design and operational timing.