The Rock Physical Phase Partition Method of Multilayer Sandstone Reservoir Based on Fuzzy Inference Networks

摘要

A kind of discrimination method based on weighted fuzzy inference networks is put forward in this paper according to the rock physical phase division problem of ' heterogeneous multilayer sandstone reservoir. In this method, the fuzziness of rock physical phase classification and meeting character of characteristic parameters are fully considered, and the cumulated expert experience in research are combined with the quantitative indexes which reflect the rock physical phase change. The rock physical phase classification types and deviding indexes are defined by the core analysis data of coring well and explaining results of expert, then the classification standard pattern library of rock physical phase is established. The rock physical appearance discrimination model of heterogeneous multilayer sandstone reservoir based on weighted fuzzy inference networks is established by adopting the adaptive learning mechanism. The experimental results demonstrate that the method has good application effect by processing the actual data from Daqing oilfield.