This study details characterization and Discrete Fracture Network (DFN) modeling of a naturally fractured carbonate reservoir with natural fractures and karst, or “non-matrix,” that are essential to production. Characterization of non-matrix follows a workflow where non-matrix is interpreted from image logs and core. Prediction of non-matrix intensity trends away from well control is performed using regression to identify covariance with static grid properties and modeled using Sequential Gaussian Simulation (SGS). DFNs are built to model fractures and karst with defined statistical distributions of orientation, size, aperture, and permeability. Several DFN realizations are created to capture uncertainty. A hybrid Embedded Discrete Fracture Model plus Dual Porosity Dual Permeability (DPDK) approach is employed for simulation. Each DFN is split into 3 different size bins, including “small” fractures upscaled and added to matrix; “medium” fractures upscaled and assigned to the DPDK grid; and “large” fractures and karst that are modeled as EDFM. The resulting EDFM+DPDK models are used in practice for forecasting and signposting primary production and IOR outcomes. This work highlights a workflow to incorporate fracture characterization with fracture prediction to build DFNs that are geologically robust, consistent with dynamic well test data, and application to EDFM+DPDK models for forecasting.

This content is only available via PDF.
You can access this article if you purchase or spend a download.