A new empirical correlation for estimating bubble point oil formation volume factor

To determine the bubble point oil formation volume factor (Bob), which is one of the most important PVT properties, several correlations have been proposed for different regions. None of the correlations could be applied as a universal correlation due to regional changes in crude oil compositions and properties. In this paper, a new correlation is proposed to predict the Bob for Middle East crudes. Genetic Algorithm (GA) was used as the dominant tool for development of the new correlation. A total of 429 data sets of different crude oils from Middle East reservoirs were used. These Data include Bob and conventional PVT properties. Among those, 286 data sets as training data and 143 data sets as test data were randomly selected for constructing the correlation and for correlation validation, respectively. The measured mean squared errors (MSEs) of predicted Bob from the correlation in the test data were 0.0029 and the correlation coefficient (R2) between predicted values from the model and experimental values in the test data was 0.9646. These results show a very good agreement with experimental data and are more accurate for Middle East crudes than those of all existing empirical correlations.