Shahin Parchekhari - Department of Petroleum Engineering, Kish International Campus, University of Tehran, Iran\Department of Petroleum Engineering, University of Tehran, Tehran, Iran
Ali Nakhaee - Department of Petroleum Engineering, Kish International Campus, University of Tehran, Iran\Department of Petroleum Engineering, University of Tehran, Tehran, Iran
Ali Kadkhodaie - Earth Sciences Department, Faculty of Natural Sciences, University of Tabriz, Tabriz, Iran

In particular, quantitative laboratory measurements are challenging to perform due to their costs and time consumption. So, the need to explore other available data interconnectivity to permeability is of great importance. One of these data is NMR (Nuclear Magnetic Resonance) log data which have been used frequently in recent studies. It is considered to segregate different groups, which can be obtained through cluster analysis. Using reliable parameters in the cluster analysis helps to segregate different rock units which can be used in the permeability models. To select reliable parameters, cross plots of the permeability versus extracted features from the NMR T۲ distribution curve were plotted. Results indicate that TCMR, peak reading amplitudes, and T۲Lm are the best permeability indicators, respectively. Cluster analysis was performed on T۲LM, TCMR and Peak reading amplitudes, which showed the highest consistency with core derived data compared with other parameters. The crucial step is to determine the best estimate of the number of clusters. It is usually taken as a prior in most clustering algorithms. In this research, Davies-Bouldin criterion values versus the number of clusters were used to obtain the optimal number of clusters. The knee method, which finds the “knee” in many clusters vs. clustering evaluation graph, was used. A clustering model with the number of clusters from ۲-۱۰۰ was created. It showed the five is an optimal number of clusters. Subsequently, the Schlumberger-Doll-Research (SDR) coefficients for each cluster were modified using a curve fitting tool in the Matlab software. Results indicated that calculated permeability using cluster analysis showed a higher correlation by core derived permeability than the original SDR permeability model. Since this is the core part of the group attempt to use extracted T۲ distribution features in permeability estimation in carbonate reservoirs, more investigation is required to attempt satisfactory results to standardize the value of the coefficient of the permeability models in carbonate rocks with different petrophysical properties.

NMR, Peak Analysis, Clustering, Timur-Coates, SDR