Development the models of oil Formation volume factor using artificial neural networks and fuzzy logic models

The oil formation volume factor with high accuracy method is a key role in the petroleum industry due to the wide use of it in the petroleum industry. It is readily obtained from laboratory PVT measurements or may be calculated from correlations such as Vasquez. Nevertheless, these measurements are either not available, or very costly to require. Thus, there is an essential need for a reliable method for obtaining the oil formation volume factor. The aim of this paper is predicting the oil formation volume factor using Artificial Neural Networks (ANN) and Fuzzy Logic (FL) tools. It is worth noticing that a data set consisting 800 of laboratory measurements on oil formation volume factor was gathered from different published resources. The paper also will use the current available models presented in the literature for predicting the oil formation volume factor and compare the average percent error of these models with the new base models. The results obtained depicted that new models were able to find the oil formation volume factor with higher accuracy than the current models for predicting oil formation volume factor. It is conspicuous results that the Artificial Neural Networks (ANN) model with coefficient 0.994 and Fuzzy Logic (FL) with coefficient 0.9993 provide the oil formation volume factor. The new developed models from the ANN and Fl models outperformed the prior models for the oil formation volume factor. It is obviously observed that the new models can be used to predict the oil formation volume factor with a high accuracy as compared with the other models used to be calculated from correlations such as Vasquez and Beggs.