An insight into the modeling of sulfur content of sour gases in supercritical region

Abstract

Sour gas reservoirs are one of the well-known energy re-sources in the world so investigation of their problems especially deposition of sulfur is of interest for chemical and petroleum engineers. Due to this fact, prediction of sulfur solubility in supercritical sour gases, which is known as the most important factor in the deposition of sulfur, is considered as the main aim of the current investigation. In this work, least-squares support vector machine (LSSVM) and adaptive neuro-fuzzy inference system (ANFIS) were coupled with particle swarm optimization (PSO) to estimate sulfur solubility in terms of reservoir pressure, temperature, and composition of sour gas. The ability and potential of proposed algorithms were investigated through 170 experimental data, which are available in literature. Also, the proposed algorithms results were compared with four published methods results to show the great capability of LSSVM-PSO and ANFIS-PSO in the estimation of sulfur solubility. On the other hand, the effectiveness of input parameters was evaluated and it was acclaimed that pressure has the most impact on sulfur solubility in supercritical sour gases.