Hierarchical Approach to Modeling Karst and Fractures in Carbonate Karst Reservoirs in the Tarim Basin

Abstract

Karst reservoirs in the Tarim Basin, northwestern China, were formed by subaerial exposure and karstification from the Ordovician formation and represent the main plays. Predicting the storage capacity and quantifying permeability heterogeneities are challenging while important for field development planning. In this paper we present a hierarchical approach to modeling karst and fractures with geoscience and engineering data for selecting locations of new wells and for the reservoir simulation. Karst and fractures at multiple scales contribute significantly to reservoir volumes in place and well productivity. Fracture-karst units in wells were determined using log-based electrofacies validated against core data, image logs and drilling data to quantify different karst features and fracture patterns hosted in units. A 3-D architecture model of karst system was constructed with extracted karst features at the seismic-scale based on multi-attribute seismic facies analysis. The karst network model was generated with karst-fracture units at wells, inverted seismic impedance volume, and 3-D karst architecture model. Porosity estimates of the karst system were conditioned with log data, mud loss data, seismic impedance volume and karst network model. Karst horizontal and vertical conduits were modeled and their permeabilities were empirically derived. Based on fracture length relative to the seismic resolution, fractures were modeled at two scales. Diffuse fractures at a small scale were modeled stochastically conditioned with image log data and the karst fracture unit model. A discrete fracture network (DFN) model at a large scale was deterministically built by meshing fracture lineaments automatically tracked from the curvature enhanced attribute. The DFN model was embedded into a geocellular grid model in which geometries of the large fractures were maintained explicitly. The calculation of effective horizontal and vertical permeabilities of the fracture system was scale dependent and decoupled. Fracture geometric parameters and permeabilities were calibrated against well test data. Streamline simulation was performed in the static model to calibrate spatial fracture densities. After two-step conditioning, fracture models were updated and then upscaled. Flow properties of karst and fractures from the wellbore to the seismic scales were combined based on their impacts on fluid flow. Integration of karst network model and history match of water cut and bottom hole pressure using streamline simulation helped the oil/water contact (OWC) assessment and allowed the identification of dynamic compartments. Combing karst networks, dynamic compartments and modeled geological scenarios allowed targeting potential highly productive zones where new well locations could be selected. The case study demonstrated that the hierarchical approach to karst and fracture modeling and calibration allowed building a realistic reservoir model and better understanding of the reservoir complexity.