Fracture modeling of carbonate rocks via radial basis interpolation and discrete fracture network

Abstract

For carbonate reservoirs primarily composed of micrite particles, fractures serve as the main channel for fluid flow, determining the storage capacity and development efficiency of oil and gas. However, controlled by various formation factors, these fractures are typically of a small scale and have an irregular distribution, making them difficult to be accurately characterized and predicted by conventional methods. Therefore, this study proposed a discrete fracture network (DFN) modeling method based on radial basis interpolation, which enhanced the comprehensiveness of modeling restraint condition and significantly improved the accuracy of fracture distribution prediction. First, a model of the primary spatial distribution of fractures was derived from the difference calculation of statistical parameters of fractures based on radial basis function (RBF) interpolation. Next, an RBF-driven interpolation was programmed as an auxiliary tool for the DFN model by taking into account both stochasticity and uncertainty. The primary and secondary restraint maps used for DFN model were generated under the control of geological parameters and seismic attributes, respectively. Finally, the stochastic DFN model was sequentially modified based on different levels of constraint conditions, and its accuracy was validated using actual data. The results indicate that the DFN model with RBF is more reliable and highly consistent with the actual geological conditions and well production data.