Prediction of sulfur solubility in supercritical sour gases using grey wolf optimizer-based support vector machine

Accurate knowledge of the solubility of sulfur in supercritical sour gas is essential in highly effective development of sour gas reservoirs. However, it is well-acknowledged that experimental measurements are expensive, time-consuming and cumbersome especially due to the highly nocuous H2S. As a direct consequence, a new meta-heuristic technique namely grey wolf optimizer-based support vector machine (GWO-SVM) was proposed to accurate prediction of the sulfur solubility in supercritical sour gases. The proposed GWO-SVM model considered the reservoir temperature, pressure and the mole fraction of methane, hydrogen sulfide and carbon dioxide as input parameters and the sulfur solubility as target parameter on the basis of grey correlation analysis. The accuracy and reliability of the presented model were evaluated through 170 data sets accessible to the literature and compared with three empirical correlations (Guo-Wang, Hu et al., Chrastil correlation) reported in previous literature. The results showed that the proposed model provides the closest agreement with experimental data with the global average absolute relative deviation of 4.65% and significantly outperforms all the existing methods considered in this work. Additionally, the outlier diagnostics was also operated for detection of the probable doubtful sulfur solubility data and identification of the validity and applicable range of all models considered in this work. (C) 2018 Elsevier B.V. All rights reserved.