Implementing a Cluster Analysis with the Aid of Self-Organized Map Artificial Neural Network for Assessment of the Asmari Reservoir Rock Type (Gachsaran Oil Field)

Identification of geological and reservoir changes in accordance with changing rock types is possible through log-facies studies and confirmation of those facies with core data. In the present study, first, in order to determination of microfacies and diagenesis processes, the Asmarri formation has been studied using thin sections of cores. Then well logging and core data of 5 wellbores from Asmari reservoir in Gachsaran field were clustered with the aid of Self-Organizing Map Neural Network algorithm, and then 5 facies models were selected among log facies. The log facies were acceptably confirmed through consideration and comparison of these models with tow cases: (1) core data (porosity, permeability and capillary pressure curve) and (2) petrography. Afterward, the log facies were presented as reservoir rock types. Regarding the good results in distinguishing reservoir horizons, the model was propagated in other wellbores in which only logging data were available, and the trend of changes in rock types were investigated through the entire field.