An innovative approach to discrete facture network modeling driven by geomechanics and multiple factors

Abstract

The Keshen 8 gas field, with reservoirs exceeding 6000 m in depth, relies on structural fractures to enhance permeability due to poor matrix properties resulting from intense compaction and cementation. This study aims to conduct a new approach to establish the discrete fracture model considering geomechanical characteristics. Through a combination of field observations, core samples, imaging logging, and thin section analysis, fracture attributes, distribution, and development were analyzed. Structural influences and reservoir characteristics were evaluated to identify the key factors controlling fracture formation and development. A discrete fracture network (DFN) model was constructed based on multi-scale and multi-constraint principles, with model reliability verified through comparison with field data. The results showed that: (1) Sear fractures dominated the fracture system, with high-angle fractures being the most prevalent. The fractures are mostly partially filled, with gypsum and calcite as the main filling materials. Fracture density primarily ranges from 0.41 to 0.76 fractures/m, with dominant orientations in the NNW-SSE and NEE-SWW directions. (2) Fractures are more developed near faults, especially near major faults, with layer thickness influencing fracture density, particularly in mudstone and sandstone layers. (3) Three major tectonic events have influenced fracture development, with the most significant fractures occurring during the third event in the Miocene-Pliocene. (4) A geomechanical-driven fracture strength model was constructed by combining various geological factors, including stress fields, fault proximity, and lithology, resulting in a more accurate representation of the fracture network in the reservoir. The findings of this research provide valuable insights into fracture characterization and modeling for gas reservoir development, contributing to more accurate reservoir simulation and management.