要約

Capillary pressure is one of the main parameters which is widely used to characterize and describe reservoir rock properties. Although some methods have been proposed to determine capillary pressure in reservoir rock, these methods may not be able to determine the capillary pressure accurately. In this study, an artificial neural network (ANN) algorithm was developed to estimate the capillary pressure in a hydrocarbon reservoir in the Middle East. A complete data set of several core samples includes porosity (Φ), normalized porosity (Φ_z), permeability (k), rock quality index (RQI),flow zone indicator (FZI); water saturation and drainage capillary pressure curves were applied to develop the ANN model. The ANN model which was designed in this study contains two separate parts. The first part categorized the reservoir rock into discrete groups with similar ranges of porosity and permeability, and the second part estimated the capillary pressure for each group. The results of this study revealed that ANN is an appropriate method to estimate the capillary pressure in reservoir rocks, particularly those which have heterogeneity in rock properties.