A study of oil formation volume factor correlation for Malaysian crude using multi linear regression.

Abstract

The most trustworthy techniques for PVT investigation of reservoir fluid characteristics are laboratory measurements. However, real-time decision-making requires the use of quicker and less expensive procedures, and mathematical correlations are employed to forecast the properties. Due to either oversimplifying assumptions or using the wrong methodology, most of the known correlations do not provide precise estimations. In this study, a new correlation was created using the multi-linear regression technique to estimate the oil formation volume factor (Bo). It is a series of inductive techniques used to model multi-parametric data sets mathematically on computers. 93 PVT data sets from Malaysian crude oil analyses were used to create the correlation. The information includes the following: temperature (T), oil viscosity (µo), oil gravity (γo), gas gravity (γg), and solution gas-oil ratio (Rs). Results analysis reveals that the new correlation is more accurate than the current existing correlations from the literature. The Absolute Average Percent Relative Error (AAPRE) of 5.879%, with maximum and minimum average relative errors (max. AAPRE and min. AAPRE) of 21.598% and 0.366% respectively. The Standard Deviation (SD) is the lowest with only 4.429%. These indicate that the new correlation for the oil formation volume factor is superior to the other correlations for Malaysian crudes.