Application of Non-Parametric Regression Techniques to Estimate the Reservoir Permeability of Bahariya Formation

Abstract

The objective of this work is to introduce a systemic workflow for a regional understanding of Bahariya reservoir characteristics, identification of rock units, and reservoir permeability. Specifically, the Alternating Conditional Expectation (ACE) algorithm and the Artificial Neural Networks (ANN) were applied on well log data from about 100 cores covering the different geological and depositional features. This approach was applied to different testing wells addressing different geological and sedimentary features with variable log characteristics from the convention high-resistivity to low-contrast (LRLC) behaviors. The established permeability profiles exhibit high correlation coefficients for training and testing datasets. Additionally, it shows high accuracy that matches the field experience even with LRLC characteristics.