Integrated deterministic and predictive discrete fracture network modeling for an Eocene carbonate reservoir, Bengal Basin, India

Abstract

Prediction of fracture porosity and permeability remains a challenge for fractured carbonate reservoirs. As natural fractures are heterogeneous and subseismic in scale, core data and image logs only provide partially sampled data, leading to sparse information on fracture length, height, orientation, spacing, and aperture. In the present study, an integrated discrete fracture network was generated that is capable of predicting fracture porosity in Eocene carbonates of the Bengal Basin in northeastern India. The predictive fracture modeling method used 3D kinematic and geomechanical restoration of interpreted seismic horizons to estimate infinitesimal stress and strain values and to characterize associated fracture sets. Seismic attribute analysis was used to extract faults and fractures from an ant-track cube, which provided sharper definition of discontinuities seen in conventional curvature attribute data. An integrated discrete fracture model was created using information from seismic attributes, seismic inversion, and strain distribution to determine fracture intensity. Faults and fractures that are seismically resolved were statistically analyzed, which indicated that spatial distribution of fracture length follows a power law. Based on theoretical concepts of fracture mechanics, linear aperture-to-length scaling was used to characterize aperture population. A present-day geomechanical earth model was used to identify open fracture sets. This model shows that northeast–southwest-oriented fracture sets are critically stressed and will contribute to porosity and permeability. Criticality of fractures to shear failure was analyzed by computing geomechanical parameters — slip stability and dilation tendency, based on the direction and magnitude of far-field stresses. Fractures with slip and dilation tendencies greater than 0.6 in the modeled discrete fracture network were taken as inputs for porosity and permeability estimation.