Modeling of fracture intensity by integrating geostatistical and neural network methods: The asmari reservoir, Rag-e-Sefid Oilfield, SW Iran

Modeling fractured reservoirs is crucial for hydrocarbon exploration, enhanced production, and effective reservoir management. To model the fracture network in the Asmari reservoir of the Rag-e-Sefid Oilfield (SW Iran), this research characterizes fracture intensity using borehole image logs, fracture driver data (e.g., seismic attributes, strain analysis, distance to fault), and fracture controller data (e.g., tectonic setting, bed thickness, lithology, and diagenetic processes). Fracture intensity is modeled using geostatistical methods, including both deterministic and stochastic approaches. Outputs from these methods are compared and calibrated based on the structural setting of the study area and fracture indicator data (e.g., mud loss and production data analyses). Finally, the selected fracture intensity data are integrated into a single model using an artificial neural network, creating a comprehensive fracture intensity model for the reservoir. The results show that fracture intensity increases near the Rag-e-Sefid and Nourooz-Hendijan-Izeh (NHI) Faults and in the folded parts of the Asmari horizon. The highest values of fracture intensity, and consequently porosity/permeability in the reservoir, are observed at the intersection of structures with the N-S Arabian trend and the NW-SE Zagros trend, where the fold axis has rotated. High fracture intensity in the NW part of the Rag-e-Sefid Oilfield is related to inversion tectonics, multi-stage reactivation along pre-existing basement structures, and an older deformation history in this area compared to its SE part.