A Bayesian approach to identify clay minerals from petrophysical logs in Gonbadly Gas field, northeastern Iran

More often clay matrix is the major factor to reduce the porosity and permeability in sandstone facies. Consequently determination of clay minerals is of prime importance in reservoir quality assessment. The present study aims to identify four different types of clay mineral namely kaolinite, illite/cholorite, halloysite, and montmorilonite from Petrophysical Logs (PLs) using Cation Exchange Capacity (CEC) parameter. In this regard, PLs related to two wells of Shurijeh Formation (Early cretaceous) in Gonbadly gas field, Northeast of Iran, were used. Utilizing measured CEC data and proper PLs, the CEC log were generated by employing MLP neural network. Relying on this fact that clay minerals can be classified based on their CEC value, the formation under study were divided into five zones by implementing four cut offs on CEC log. Finally, Bayesian classifier was applied on PLs to identify the desired zones. According to the obtained results, the method proposed in this study is able to identify desired clay types with average accuracy of 68.5% in single well analysis step and 65.75% for generalization step.