Innovative lithology identification enhancement via the recurrent transformer model with well logging data

In the field of oil and gas exploration and development, accurately predicting lithology is crucial for strategic decision-making. Identifying subsurface rock types helps determine the distribution, quantity, and recoverability of oil and gas reserves, guiding exploratory efforts and increasing the likelihood of successful resource extraction. Our research focuses on enhancing lithology prediction accuracy through the development of the Recurrent Transformer model. This innovative model combines the Transformer architecture with recurrent elements, optimizing the processing of well logging data, inherently comprising time-series information. The recurrent structure effectively learns local contextual information, improving responsiveness to geological variations impacting lithology. Additionally, we integrated the Recurrent Scale-wise Attention (RSA) mechanism into our model, uniquely suited to well logging data. RSA captures and interprets multi-scale information by implementing attention mechanisms at varying scales, enhancing the model's understanding of sequence data. The recursive element within RSA further strengthens the model's ability to process contextual nuances. Key well log curves, including Density (DEN), Acoustic (AC), Gamma-ray (GR), and Compensated Neutron Log (CNL), serve as primary data sources for extracting geological features. These features are input into the Recurrent Transformer, establishing correlations between lithological characteristics and well logging parameters. Comparative analysis with leading-edge models using an experimental dataset revealed the Recurrent Transformer's high prediction accuracy and remarkable generalization capabilities across diverse lithological prediction tasks and geological conditions. This model represents a significant advancement in machine learning for well logging lithology prediction, providing geologists and engineers with a precise and efficient tool, enhancing resource exploration and development quality and reliability. The Recurrent Transformer model showcases the potential of integrating advanced machine learning in geosciences, opening new horizons in oil and gas resource exploration and utilization.