Modeling the reservoir fluid behavior of black oil systems using a RBF network

This paper presents the application of radial basis function neural network (RBFNM) to model the behavior of black oil systems. The RBFNM is trained u sing PVT analysis of numerous black-oil samples collected from various Kuwa iti oil fields. The model is tested using properties of other samples that have not been used during the training process. The accuracy of the model i n predicting the PVT properties has been compared for training and testing samples to several PVT correlations. The comparison indicated that the RBFN M is much more accurate than published correlations in predicting the prope rties of the crude oils under study. The behavior of the model in capturing the physical trend of the PVT data has also been checked against experimen tally measured PVT properties of the test samples.