ISSN Approved Journal || eISSN: 2582-8185 || CODEN: IJSRO2 || Impact Factor 8.2 || Google Scholar and CrossRef Indexed
Fast Publication within 48 hours || Low Article Processing Charges || Peer Reviewed and Referred Journal || Free Certificate
Machine learning-based prediction of well logs in the Niger Delta for improved hydrocarbon exploration: Comparison of models for density log predictions
Department of Earth Sciences, Federal University of Petroleum Resources, Effurun 330102, Delta State, Nigeria.
Research Article
International Journal of Science and Research Archive, 2024, 13(02), 2677–2689.
Article DOI: 10.30574/ijsra.2024.13.2.2379
Publication history:
Received on 29 October 2024; revised on 07 December 2024; accepted on 09 December 2024
Abstract:
This study explores the usefulness of machine learning methods to predict well-log data in the Niger Delta, a geologically complex region that poses challenges for traditional prediction methods. The research specifically compares the effectiveness of various machine learning models, including Linear Regression (LR), k-Nearest Neighbors (KNN), Random Forest (RF), Gradient Boosting Regressor (GBR), and the traditional Gardner's equation. The models were evaluated based on their performance metrics, including Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), and R-squared (R²) scores. The results indicate that machine learning models, particularly the tree-based models (R² scores above 0.90, MAE and RMSE scores below 0.05), outperform Gardner's equation (R² scores below 0.30, MAE and RMSE scores above 0.1) in predicting density logs, signifying higher accuracy and robustness. The study also underlines the importance of data quality, the selection of appropriate models, and the need for hyperparameter optimization to improve model performance. The results indicate that the incorporation of sophisticated machine learning models into exploration workflows can significantly improve subsurface predictions, thereby boosting exploration effectiveness and success rates within the Niger Delta. This study gives valuable insights into the potential of machine learning in geophysical applications and offers practical recommendations for the adoption and implementation of these techniques in the oil and gas industry. The research emphasizes the need for ongoing efforts and partnerships to enhance the use of machine learning techniques in predicting well-log data, to optimize hydrocarbon exploration in various geological contexts.
Keywords:
Machine Learning; Regression; Well logs; Gardner’s Equation; Density
Full text article in PDF:
Copyright information:
Copyright © 2024 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0
