In this study, a novel rock physics-based data-driven algorithm was presented to identify three pore types using data related to three wells drilled in a carbonate reservoir located in Iran. The results of this study indicated that.

Generating the velocity-deviation log (VDL) the main pore types of reservoir under study can be categorized into three main groups.

Using the frame flexibility factor (γ), the γ values between −1 and 1 are the best class for Moldic/vuggy porosities. Interparticle pores are classified by γ<−1, and cracks and micro-fractures are classified by γ>1.

Defined pores make specific clusters to gathers in cross-plot of velocity–porosity as well as impedance–porosity

Using the sected features, defined pore types were identified through Parzen classifier with average accuracy of 76.7%.