New Oil Formation Volume Factor Correlation for Nigerian Crude Oils

• Main Authors: Atthi, A.J., Sulaimon, A.A., Akinsete, O.K.
• Format: ["eprint_typename_conference\_item" not defined]
• Published: Society of Petroleum Engineers 2022
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85136831724&doi=10.2118%2f211968-MS&partnerID=40&md5=591cb225ea28b03fdcac23394e7e41ca; http://eprints.utp.edu.my/33728/
• Institution: Universiti Teknologi Petronas

A comprehensive description of reservoir fluid properties is critical in developing solutions and resolving reservoir engineering issues. The oil formation volume factor, βo, is an indispensable reservoir fluid property in reservoir engineering calculations. In this study, we used a total of 11040 data points from 1840 oil samples to develop new βo correlations for the Nigerian crude oils specifically, and another set of correlations for the other regions herein referred to as the global crude oils. Linear regression (LR), multiple linear regression (MLR), multiple non-linear regression (MNLR), neural network (NN), support vector machine (SVM), and the group method of data handling (GMDH) techniques were used to develop several correlations. Results show that the GMDH method yielded the best correlation while the MNLR is the least accurate. The root means square error (RMSE) for the Nigerian, and Global correlations are 0.0033, and 0.0256 respectively. The two correlations are reliably better in terms of accuracy than the existing correlations. The new correlations would facilitate a more accurate reservoir characterization, and reliable design of surface equipment. © 2022, Society of Petroleum Engineers.