Evaluation of the efficiency of artificial intelligence techniques of permeability results of a lower Albian carbonate reservoir of Campos Basin using the Winland method and cluster analysis

Abstract

The selection of artificial intelligence (AI) techniques in the petroleum industry is very efficient because it will have a direct impact on the outcome of the study and the future use of the operation. In this study, we discuss the efficiency of the AI techniques by verification the results of permeability of a lower Albian carbonate reservoir of Campos Basin using the Winland method and cluster analysis. The permeability measured by AI techniques as fuzzy logic (FL), artificial neural network (ANN) and genetic algorithm (GA) were applied in three wells, being the first used for learning and the others as blind tests. Well logs are gamma ray, density, sonic, neutron porosity and SDR permeability logs. ANN obtained better performance compared to the FL, but the results have become better with GA. Employing artificial intelligence (AI) modern techniques together with a dataset composed by well logs, lithological information and sample laboratory measurements of permeability and porosity.