Integrating Spatial Data Fusion and Temporal Data in B-ANN for Enhanced Lithofacies Classification and Reservoir Characterization

In subsurface exploration, lithofacies classification is crucial for hydrocarbon reservoir characterization. In traditional methods, well logging data is often interpreted by hand, which can lead to mistakes and wasted time. The goal of this study is to make lithofacies classification more accurate and reliable by adding more complicated geological background knowledge to the Bayesian-Artificial Neural Network (B-ANN) model. Adaptive Markov models are used to record complex geological changes, and different geological data sources are combined in a way that doesn't affect the quality of the data. The suggested method is clearly better than existing ones, as shown by the fact that it achieved a classification accuracy of 97.6%, which is a lot higher than other methods.