多点間統計を用いた侵食谷形状モデリング

Abstract

Geologic concepts are essential for building geologic models for the exploration and development of oil and gas reservoirs. It is, however, difficult to deliver geologic concepts into geologic models with traditional stochastic modeling when the geologic concepts featured have complex morphologies. Multiple-point geostatistics, regarded as one of the best ways to overcome this difficulty, are now being put into practical use. In multiple-point geostatistics, geologic concepts are captured as multiple-point patterns through training images. However, multiple-point geostatistics has not been used for building geologic models featuring complex morphology.Therefore, we applied multiple-point geostatistics to the reproduction of a complex-shaped reference model that is a close replica of dendritic, incised valleys. Pseudo-wells and three-dimensional seismic data based on the reference model were prepared, and training images were created by using process modeling to simulate the development of incised valley geomorphology.As a result, the geologic model built using multiple-point geostatistics properly reproduced the complex morphology of the reference model. Data of various scales and resolutions were able to be integrated with the model. Therefore, the study shows that a multiplepoint geostatistical approach is effective for modeling reservoirs composed of complex-shaped incised-valley-filling deposits.