ISSN: 2578-4846
Authors: Hamblin S, Darmaki NA, Zarooni MA, Nantongo H, Boukadi F, Obeng P, Edusah E* and Osumanu J
Mathematical models and machine learning applications such as Artificial Neural Networks (ANN) have been adopted in hydrocarbon exploration, drilling, production, and reservoir engineering. Thanks to ample data sets and computing power, statistical analysis, analytics, and model prediction replace time-consuming and expensive laboratory measurements. This study used ANN to create two models for predicting water (krw) and oil (kro) relative permeability profiles for predominantly North American water-wet sandstone reservoirs. The developed model was compared to the Modified Corey and Ibrahim and Koederitz’s equations. The coefficient of multiple determination (R2) and Root-Mean-Square Error (RMSE) were used to evaluate the accuracy of the new model. The developed model showed a superior fit and can, therefore, be utilized to generate krw and kro profiles for North American water-wet sandstone reservoirs.
Keywords: Relative Permeability; North American; Water-Wet; Sandstone; ANN; Empirical Models