Stochastic Simulation of Undiscovered Petroleum Accumulations

Abstract

Geographic characteristics of undiscovered petroleum accumulations are important both to better natural resource management and improved exploration efficiency. Stochastic simulation is a useful tool that reveals uncertainties in petroleum exploration and exploitation applications. The lack of information regarding the locations of undiscovered petroleum accumulations presents a major difficulty to the application of this technique to petroleum resource assessment. In order to facilitate the locations of undiscovered petroleum accumulations, we propose a model-enhanced simulation approach that uses a geological model, in either the form of geological favorability or probability of petroleum occurrence derived from available geological and geophysical observations. The proposed approach employs a Fourier transform algorithm in the conditional simulation because it permits the spatial correlation-specific and location-specific features from different data sources to be studied separately and integrated in the frequency domain subsequently. This approach is illustrated by the analysis of the Rainbow petroleum play in the Western Canada Sedimentary Basin. The proposed approach produces a resource map showing the possible size of undiscovered petroleum accumulations with geographic locations. A comparison with the results from a traditional conditional simulation indicates that the proposed approach produces maps with improved features and predictions validated by the test data set.