The mineral composition in coalbed methane (CBM) reservoirs significantly influences fracture morphology, making the description of reservoir heterogeneity challenging. This study develops a fracture propagation model for CBM reservoirs that incorporates the varying mineral properties within the reservoir’s lithology. Dynamic logging data are considered to characterize rock mechanical properties, which form the basis for in situ stress estimation. Using an adjusted critical circumferential stress calculation for coal rock, the model considers the impact of complex lithology on fracture propagation. A comprehensive fractal index is introduced to capture the influence of different minerals on fracture morphology and propagation randomness. Models representing clay, quartz, and pyrite with varied compositions were constructed to explore the effects of each mineral on fracture characteristics. In single-component models, clay-rich reservoirs exhibited the highest induced fracture density, with quartz and pyrite showing approximately 65% and 20% of the fracture density observed in clay, respectively. Fractures primarily propagated toward quartz-rich regions, while pyrite significantly inhibited fracture growth. In mixed-mineral models, increasing the quartz proportion by 40% resulted in a 20 m increase in fracture length and a 30% reduction in fracture density. Fractures predominantly propagated around pyrite boundaries, demonstrating pyrite’s resistance to fracture penetration. Clay and quartz promote fracture development, whereas pyrite hinders fracture formation. The fracture inversion model presented here effectively captures the influence of complex mineral distributions on fracture morphology, offering valuable insights for optimizing fracturing production strategies.

Hydraulic fracturing is a crucial technique for efficient development of coal reservoirs. Coal rocks typically contain a high density of natural fractures, which serve as conduits for fracturing fluid. Upon injection, the fluid infiltrates these natural fractures and leaks out, resulting in complex fracture morphology. The prediction of hydraulic fracture network propagation for coal reservoirs has important practical significance for evaluating hydraulic fracturing. This study proposes a novel inversion method for predicting fracture networks in coal reservoirs, explicitly considering the distribution of natural fractures. The method incorporates three distinct natural fracture opening modes and employs a fractal probability function to constrain fracture propagation morphology. Based on this method, the study compares hydraulic fracture network morphologies in coal reservoirs with and without the presence of natural fractures. The results show that while both reservoir types exhibit multi-branch fracture networks, reservoirs containing natural fractures demonstrate greater branching and a larger stimulated reservoir volume (SRV). Additionally, the study employs a fractal dimension calculation method to quantitatively describe the geometric distribution characteristics of fractures. The analysis reveals that the geometry and distribution of natural fractures, as well as reservoir geological parameters, significantly influence the fracture network morphology and fractal dimension. The contact angle between natural and hydraulic fractures affects propagation direction; specifically, when the contact angle is π/2, the fractal dimension of the hydraulic fracture network is maximized. Moreover, smaller lengths and spacings of natural fracture led to higher fractal dimensions, which can significantly increase the SRV. The proposed method offers an effective tool for evaluating the hydraulic fracturing of coal reservoirs.