Shear Wave Sonic Log Estimation from Petrophysical Data Utilitizing Machine Learning Models, Lower Indus Basin Pakistan

Summary The research addresses a significant challenge in petroleum exploration by developing a machine learning approach to estimate shear wave velocity in the Sawan Gas Field, Pakistan. This innovation is particularly important because shear wave velocity data, while crucial for lithology discrimination, is often unavailable due to cost and technical constraints. The study implements multiple machine learning algorithms, including XGBoost, Random Forest, and Neural Networks, to predict shear wave velocities from existing well log data. Testing on multiple wells in the Sawan Gas Field demonstrated high accuracy, with R2 scores above 0.90. The methodology proves especially valuable for the heterogeneous sandstone intervals of the Lower Goru Formation, where traditional lithology discrimination methods face limitations. The developed approach offers a cost-effective solution that can be applied to similar depositional environments, enhancing reservoir characterization capabilities without requiring additional expensive logging operations.