A New Approach in Reservoir Characterization Using Artificial Intelligence

Abstract The basic characteristics of a reservoir system are normally distributed spatially. Since there is no immediate measurement for the lithological parameters, they are to be computed from other geophysical logs or seismic attributes. In this study, oil field data from the exploration wells including logs and seismic data were used and studied using AI Artificial Intelligence to select the best locations for drilling according to porosity distribution. Moreover, the objective of this project is the study of reservoir characteristics and distribution using Artificial Intelligence (AI). The studied area is the Asma area in India where all seismic and well-logging data were collected. The process is first aligned in finding the porosity of the well-logging interpretation using density log and sonic log and then shed the prediction of porosity using seismic data using the reflected waves of the structure to estimate the acoustic impedance. The neural network was used to predict porosity. There were three outputs for each of well-logging, logging-seismic and seismic. All the outputs were compared where the accuracy was good with a ratio of 95% and a small error percentage of 5%. A map of the area was created where the porosity distribution of each of the outputs was determined. The best areas were identified for drilling the future production wells depending on the porosity and the type is sandstone.