Breaking New Ground: Novel Lithology Estimation Machine Learning Model for Pulsed Neutron Logging Technology

Abstract

The rapid advancement of machine learning techniques has opened up new possibilities for improving lithology estimation for outstanding reservoir characterization. This study introduces a groundbreaking machine learning model specifically developed for lithology estimation utilizing measurement data from Multi-Detector Pulsed Neutron (MDPN) logging. Traditional methods of lithology estimation have relied on limited data inputs, and non-automated interpretation leading to significant uncertainties and inconsistencies. On the other side, the proposed machine learning model is considered as an automated approach that is trained over well-log big data combined with high-computation advanced artificial neural networks algorithm. The model undergoes extensive data preprocessing and analytics, feature engineering, training and optimizing the model architecture/parameters, and model evaluation workflow. By leveraging the power of artificial intelligence, the proposed model is qualified to learn/capture the data pattern interrelationships between pulsed neutron logging measurements and lithology variations that enhance the accuracy and efficiency of lithology estimation in diverse geological formations. The developed model was evaluated with statistical metrics to assess the prediction accuracy for the model outputs versus log data measurements. The evaluation of the model's performance demonstrates that superior artificial neural networks technique has outstanding capability to accurately estimate lithology with an accuracy higher than 97% with low loss error evaluation metrics. The successful application of the model in a case study conducted in a Middle East region further validates its effectiveness and robustness. The integration of this innovative machine learning model with pulsed neutron logging technology offers a transformative solution to automate lithology estimation workflows in the oil and gas industry. This research paves the way for enhanced lithology characterization, leading to improved decision-making in reservoir evaluation and exploration activities.