In this research paper, we studied absorption of carbon dioxide in amino acid solutions theoretically via the artificial intelligence (AI) method. The CO2 loadings in different solutions were predicted as a function of input parameters including amino acid parameters and operational parameters such as pressure and temperature. For the training of the data, we used differential evolution, and for the decision and prediction sections, we used the fuzzy interface system. The model can be abbreviated as differential evolution fuzzy interface system (DEFIS). Moreover, various methods and tuning parameters in the DEFIS were used to increase the accuracy of the model. Cluster influence range (CIR) was used for tuning the model, and different population sizes were used, including 9, 18, 27, and 36, with the purpose of observing the effects of these parameters on the training. The results from AI overlapped measured results meaning that we can use AI for optimization of the separation process for CO2 sorption in chemical solvents. The neural methodology indicated a reliable and robust prediction of CO2 solubility in different chemical solvents.
Read full abstract