Optimization and Analysis of CO2 Capture in RPB using Cognitive Computing and Evolutionary Algorithm

Due to industrialization, deforestation and many other anthropogenic activities, carbon emission is increasing at a rate of approximately at a rate of 1% in past few years. Now, reduction of atmospheric carbon dioxide (CO2) has become a significant concern and challenge for every country across the globe. This paper is a sincere effort to study, analyse and further optimize, amine based post-combustion carbon (PCC) capture. Monoethanolamine (MEA) in rotating packed beds (RPB) has been extensively studied for CO2 chemical absorption. Enhancing CO2 capturer efficiency necessitates a thorough comprehension of the complex interrelationships within the key parameters. This study focuses on modelling and optimisation of CO2 absorption efficiency in MEA by artificial intelligence and genetic algorithms (GA). Machine learning (ML) and Artificial Neural Networks (ANN) are versatile instruments employed to model and forecast diverse complex and highly non-linear phenomena. The established process models have been established by published steady-state experimental data. Subsequently, SHAP analysis has been applied that reveals the input factors such as solvent concentration, flow rate, and rotational speed are the primary determinants of CO2 absorption in RPB. To assess the model's performance, the acquired results have been examined using statistical measures, including MSE, RMSE, and R2 value. The modelling results have been utilised to optimise CO2 absorption, employing GA under various operating conditions to ascertain the optimal values for the input variables that correlate to maximized CO2 capture.