Machine Learning Predictions of Best Optimal Solutions in Evolutionary Multi-Objective Optimization of Industrial Gas Sweetening Process

Natural gas (NG) is one of the most important energy sources to fulfil the growing energy demand of the world. It contains impurities like carbon dioxide (CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf>) and hydrogen sulfide (H2S), the process of removing impurities from the NG is known as gas sweetening. The industrial gas sweetening process is optimized considering two objective functions (Hydrocarbon recovery & CO<inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</inf> removal), five decision variables, and one imposed constraint. A Promax-based model is designed and a non-dominated sorting genetic algorithm (NSGA-II) is used. The machine learning approach is used to find relative error using the decision tree algorithm and Pareto ranking is carried out by Vikor method. Machine learning outcomes show the potential for predicting the output from a complex and nonlinear industrial chemical engineering process.