Combination of genetic algorithm and CFD modelling to develop a new model for reliable prediction of normal shock wave in supersonic flows contributing to carbon capture

Abstract

Carbon dioxide separation and capture using green and efficient methods is an important issue in studies related to climate change. The supersonic separator is one of the efficient and reliable methods that can be used to separate impurities, including carbon dioxide, from gas streams. Reliable estimation of normal shock wave position plays a vital role in the proper design and simulation of supersonic separators. Many studies have used a one-dimensional theoretical (ideal) model of a normal shock wave for the estimation of the shock position and pressure recovery, but the accuracy of the ideal model of normal shock may be insufficient in some situations, as reported in the literature. A novel approach is presented in this paper to provide new equations for normal shock waves by the combination of computational fluid dynamics (CFD) and genetic algorithm. The comparison of the proposed model with several experimental data and the ideal model of a normal shock wave indicate that the present model provides more accurate predictions than the traditional model of a normal shock wave. The present model showed an average absolute relative deviation (AARD) of 1.80%, which is about six times less than AARD of the ideal model, indicating the robustness of the proposed model. Consequently, the present model can be employed as an accurate and efficient tool for the prediction of shock position and design of converging–diverging nozzles.