Title

Implementing radial basis function neural networks for prediction of saturation pressure of crude oils

Document Type

Article

Publication details

Tatar, A, Najafi-Marghmaleki, A, Barati-Harooni,A, Gholami, A, Ansari, HR, Bahadori, M, Kashiwao, T, Lee, M. & Bahadori, A 2016, 'Implementing radial basis function neural networks for prediction of saturation pressure of crude oils', Petroleum Science and Technology, vol. 34, no. 5, pp. 454-463.

Published version available from:

http://dx.doi.org/10.1080/10916466.2016.1141217

Peer Reviewed

Peer-Reviewed

Abstract

This study highlights the application of radial basis function (RBF) neural networks for perdition of saturation pressure of gas condensates and oils. The experimental data were collected from literature and cover a vast geographic distribution. Genetic algorithm (GA) was used to determine the optimum values of spread and maximum number of neurons for developed RBF model. The input parameters of the model were the C1 through C7+fraction of gas condensates, crude oil, nonhydrocarbon fraction of crude oil (nitrogen [N2], carbon dioxide [CO2], and hydrogen sulfide [H2S]), specific gravity and molecular weight of C7+ (SGC7+, MWC7+) and temperature. The output of model was the saturation pressure of crude oil. Different statistical and graphical methods were utilized to examine the accuracy of implemented GA-RBF model. Results of modeling study showed that the GA-RBF model is effective and robust in reproducing the whole data points with an acceptable accuracy.