An Artificial Neural Network Approach to Formation Damage Prediction Due to Asphaltene Deposition

Abstract

Abstract Formation damage due to asphaltene deposition could have bad consequences on oil production through heavy oil recovery, miscible flooding, and even primary recovery. Many tests were carried out by researchers to find the amount of permeability decline, but the limits in which the asphaltene deposited in oil or at the pore surface was not determined. In this paper, neural network is used to predict rock-fluid interactions describing permeability decline due to asphaltene deposition. The second stage of asphaltene deposition (if we separate it into precipitation from liquid phase, and deposition on pore surface) is considered in which n-hexane is used to flocculate asphaltene particles in order to determine the effect of deposition on sandstone rock due to changing of pressure, temperature, and composition of reservoir oil. An artificial neural network is trained using data gathered by performing various dynamic experiments with pre-separated oil asphaltene content, moreover, some of the experimental data used to test the network and resulted a good agreement so it could predict the trend of permeability reduction due to deposition of asphaltene. This network can be used to predict the trend of formation damage due to asphaltene deposition under various pressures and asphaltene concentrations. Thus, this study guides us to a novel explanation of flow behavior in porous media.