Stress field-driven fracture prediction in ultra-deep carbonate reservoirs: A coupled numerical-analytical framework based on Griffith and Mohr–Coulomb failure criteria

Ultra-deep carbonate reservoirs are currently a key focus in global oil and gas exploration. However, owing to complex tectonic evolution, intense diagenetic modification, and strong heterogeneity, the effective prediction of fracture distribution faces major challenges. The Ordovician Yijianfang and Yingshan formations in the Fuman Oilfield exhibit complex fracture networks controlled by strike-slip fault zones. This study proposes a coupled numerical–analytical framework that integrates Griffith and Mohr–Coulomb failure criteria with in situ stress field modeling to predict fracture development intensity. Results show that structural fractures are predominantly high-angle shear fractures, with linear densities of 1.06–2.38 fractures/m, of which 59.8% are unfilled microfractures. The composite fracture development index (FDI), calibrated against core and image log data, achieves a 73.3% correlation with observed fracture densities. It delineates the planar heterogeneity of fracture development and demonstrates that FDI high-value zones are strictly controlled by first- and second-order strike-slip fault zones. This multiparameter coupled approach offers a mechanistic alternative to conventional curvature-based or stochastic methods and provides a practical basis for well placement in ultra-deep carbonate fractured-vuggy reservoirs.