Utilization of RBF-ANN as a novel approach for estimation of asphaltene inhibition efficiency. (18th August 2018)

• Record Type: Journal Article
• Authors: Tashayo, Behnam; Zarei, Fariba; Zarrabi, Houman; Lariche, Milad Janghorban; Baghban, Alireza
• Is Part Of: Petroleum science and technology. Volume 36:Number 16(2018)
• Page Start: 1216
• Page End: 1221
• Publication Date: 2018-08-18
• Subjects: asphaltene -- inhibition -- precipitation -- predicting model -- RBF-ANN; Liquid fuels -- Periodicals; Petroleum -- Periodicals; 665.505
• DOI: 10.1080/10916466.2018.1463260
• Languages: English
• ISSNs: 1091-6466

Abstract

ABSTRACT: One of the problematic concerns in petroleum industries is the deposition of heavy fractions of crude oil such as asphaltene fraction during production and transportation. The utilization of inhibitors is known as a relative low cost and effective method for asphaltene inhibition. In this study, Radial basis function artificial neural network (RBF-ANN) was applied to predict asphaltene precipitation reduction in terms of structure and concentration of inhibitor and oil properties. In order to training and testing of RBF-ANN the required data are extracted from reliable sources. The predicted asphaltene precipitation reduction values were compared with the actual data statistically and graphically. The coefficients of determination for training and testing phases of RBF-ANN were determined as 0.995906 and 0.994853 respectively. These evaluations showed that the RBF-ANN as a predictive tool has great capacity to estimate effect of asphaltene inhibitors on reduction of asphaltene precipitation.