Abstract

Prediction on permeability is an essential task in reservoir engineering as it has great influences on oil and gas production, while porous media grain size, sorting, cementing, porosity, specific surface area, direction and location of grain and irreduction water saturation have effects on permeability. In this project we studied the effect of porosity, specific surface area and irreduction water saturation as main parameters on permeability distribution in the reservoir; the main goal of this research was permeability prediction in carbonat reservoir using neural network approach. Our studies showed a good agreement between our neural network model prediction and lab data or core analysis. This approach can be a useful tool for prediction permeability when core tests are not available.

Key words: Permeability prediction, neural network, specific surface area, irreducible water saturation, porosity.