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

Area of Science:

• Environmental Science
• Materials Science
• Chemical Engineering

Background:

• Carbon dioxide (CO2) adsorptive capture is crucial for mitigating environmental issues.
• Zeolites are widely studied materials for CO2 adsorption.
• Developing accurate predictive models for CO2 capture is essential for process design and optimization.

Purpose of the Study:

• To develop and validate artificial intelligence (AI) models for predicting CO2 adsorptive capture using various zeolites.
• To compare the performance of different AI algorithms in modeling CO2 adsorption.
• To identify efficient AI models for the design and analysis of adsorption processes.

Main Methods:

• Implementation of Hybrid adaptive neuro-fuzzy inference system (Hybrid-ANFIS).
• Application of particle swarm optimization-adaptive neuro-fuzzy inference system (PSO-ANFIS).
• Utilizing least-squares support vector machine (LSSVM) optimized with coupled simulated annealing (CSA).
• Validation using graphical and statistical methods, including Average Absolute Relative Deviation (AARD).

Main Results:

• Hybrid-ANFIS demonstrated high accuracy for CO2 adsorption on 5A, T-Type, SSZ-13, and SAPO-34 zeolites (AARDs: 8.21%, 1.92%, 4.99%, 2.26%).
• PSO-ANFIS showed good performance for CO2 adsorption on zeolite 13X (AARD: 4.85%).
• All developed AI models showed good agreement with experimental data.

Conclusions:

• The developed AI models, particularly Hybrid-ANFIS and PSO-ANFIS, are effective and efficient for predicting CO2 adsorptive capture.
• These models offer a prosperous approach for the design and analysis of zeolite-based adsorption processes.
• AI-driven modeling provides a valuable tool for advancing CO2 capture technologies.