The application of committee fuzzy inference system to improve results of fuzzy model in flow units prediction of petroleum reservoirs

An accurate model of reservoir rock helps in the prediction the production performance and recognition the flow units. Flow units are specific volume of reservoir, which control the flow of fluids with in petroleum reservoir. Each flow unit is characterized by a specific rang of flow zone indicator (FZI). This study presents, an intelligent model based on concept of committee fuzzy inference system (CFIS) to enhance the prediction of FZI. The proposed methodology comprises to steps. Firstly, the FZIs were divided in three low, medium and high populations. Three fuzzy models were constructed based on three classes of FZI values and the test data were introduced to fuzzy models, independently. Secondly, a CFIS is constructed. The CFIS is used a classifier neural network to combine the output of simple fuzzy models. For this purpose, petrographic data conventional logs and porosity-permeability core plug measurements from two wells of South Pars gas field is used. The results of this study demonstrate that the CFIS improved the accuracy of fuzzy model. This methodology may be used in modeling flow units at well locations for which cores may not be available or well preserved.