Pore structure characteristics of sandstone and applicability of fractal method in the Linxing Area, Ordos Basin, China

Abstract This study focuses on tight sandstone reservoirs in the Linxing area, Ordos Basin, China, and systematically compares the applicability of the box-counting and real pore-size distribution models for calculating fractal dimensions based on high-pressure mercury intrusion (HPMI) and nuclear magnetic resonance (NMR). In the HPMI experiment, the real pore-size distribution model showed better fitting accuracy in large and small pore segments, with average R2 values of 0.924 and 0.977, respectively, and fractal dimensions ranging from 2.208 to 2.397. Although the box-counting model achieved an R2 value of 0.981 in the mesopore segment, its calculated fractal dimensions exceeded the theoretical upper limit of three-dimensional Euclidean space, which lacks physical meaning. In the NMR experiment, when the bound water content exceeded 71.388%, the ability of the box-counting model to differentiate pore types was significantly weakened, whereas the real pore-size distribution model effectively distinguished between bound and movable water pores. In the dual-experiment comparison, the fractal dimensions of micropores dominated by bound water differed significantly between the HPMI (average of 2.446) and NMR (average of 1.665), indicating that NMR was more sensitive to micropore structures. This study is the first to adopt a “dual-experiment–dual-model” perspective to construct a multi-dimensional applicability evaluation framework for fractal dimension models, clarify flow response differences across pore sizes, and provide a theoretical basis for characterizing pore structure heterogeneity and optimizing model selection in the tight sandstone reservoirs of Linxing.