Improved Correlation of Oil Recovery Factor for Water Driven Reservoirs in the Niger Delta

• Published in: ABUAD Journal of Engineering Research and Development Vol. 8; no. 2; pp. 233 - 241
• Main Authors: Daniel, Ayodele, Olanipekun, Faisal Olasubomi, Isehunwa, Sunday Oloruntoba
• Format: Journal Article
• Language: English
• Published: College of Engineering of Afe Babalola University, Ado-Ekiti (ABUAD), Ekiti State, Nigeria 29.07.2025
• Subjects: Correlations; Machine Learning; Python; Recovery Factor
• ISSN: 2756-6811; 2645-2685; 2645-2685
• DOI: 10.53982/ajerd.2025.0802.23-j

Recovery factor is one of the most important variables for a reservoir engineer as it plays a major role in determining the economic viability of oil and gas projects and by implication what projects to mature. Over the years, many different approaches have been taken to estimating recovery factor of oil and gas reservoirs generally and these include simulations, volumetric method and correlations. All these methods have their high inherent cost except for correlations which are not only easy and quick to use but also low cost. Even though correlations have been developed in the past for the recovery factor of Niger Delta crude, none has employed the data analytics and machine learning techniques. Data from strong water driven crude oil reservoir in the Niger Delta was used in this study. After data cleaning and quality checking, cleaned data was used to train the machine learning model using multiple linear regression algorithms optimized with batch gradient descent method. This was implemented using Python code developed for this work. The model developed had an excellent performance on the training set as the coefficient was about 0.84. The mean absolute error is about 0.018. The results obtained showed better model performance and generalization than any previously existing model.