Estimating electrical resistivity from logging data for oil wells using machine learning

Abstract Formation resistivity is crucial for calculating water saturation, which, in turn, used to estimate the stock-tank oil initially place. However, obtaining a complete log can be challenging due high costs, equipment failure, or data loss. To overcome this issue, study introduces novel machine learning models that predict electrical of wells, using conventional well logs. The analysis utilized gamma-ray (GR), delta time compressional logs (DTC), sonic shear (DSTM), neutron porosity, and bulk density. dataset 3529 logging points from horizontal carbonate wells which were develop different random forest (RF) decision tree (DT) algorithms. obtained results showed both with accuracy, over 0.94 training testing data. Comparing based on accuracy consistency revealed RF model had slight advantage DT model. Based analysis, it was found formation more significantly impacted by GR compared DTC These new ML offer low-cost practical alternative providing valuable information geophysical geological interpretation.