Novel prosperous computational estimations for greenhouse gas adsorptive control by zeolites using machine learning methods

To predict CO2 adsorptive capture, as a vital environmental issue, using different zeolites including 5A, 13X, T-Type, SSZ-13, and SAPO-34, different models have been developed by implementing artificial intelligence algorithms. Hybrid adaptive neuro-fuzzy inference system (Hybrid-ANFIS), particle swarm optimization-adaptive neuro-fuzzy inference system (PSO-ANFIS) and the least-squares support vector machine (LSSVM) modeling optimized with the coupled simulated annealing (CSA) optimization have been employed for the models. The developed models, validated by utilizing various graphical and statistical methods exhibited that the HybridANFIS model estimations for the gas adsorption on 5A, T-Type, SSZ-13, and SAPO-34 zeolites with average absolute relative deviation (AARD) % of 8.21, 1.92, 4.99 and 2.26, and PSO ANFIS model estimations for the gas adsorption on zeolite 13X with an AARD of 4.85% were in good agreement with corresponding experimental data. It could be deduced that the proposed models were more prosperous and efficient in favor of the design and analysis of adsorption processes than previous ones.

Automated summary

Different models have been developed using various zeolites to predict CO2 adsorptive capture. The developed models employ artificial intelligence algorithms such as Hybrid-ANFIS, PSO-ANFIS, and LSSVM with CSA optimization. These models provide more accurate and efficient predictions for gas adsorption on different zeolites compared to previous models.