A systematic method for permeability prediction, a Petro-Facies approach

In this study, using a relatively large and complete data set of a complex carbonate reservoir, it is proven that among the numerous methods proposed for the prediction of permeability, the porosity-facies based models are the best choice from a theoretical and practical point of view. Based on petrographic examinations and petrophysical interpretations, a systematic approach is proposed for permeability prediction. Porosity and pore type have been identified as the main influential attributes and Petro-Facies is the preferred way of permeability estimation in the un-cored wells. The Fuzzy C-Means (FCM) clustering method has been applied for the subdivision of the data space into 12 representative Petro-Facies and the corresponding relationships between porosity and permeability for each facies has been determined. After identification of themain responsive well log suite, based on the rank correlation, a classification tree approach was used for the population of Petro-Facies in the un-cored wells. Then, the relevant porosity–permeability relation was applied for permeability calculation. This study shows that by using a systematic approach for the identification of the controlling parameters of permeability and determining the proper permeability model, it is possible to achieve a reliable permeability prediction