Advanced intelligent techniques for modeling oxygen storage in zeolite-based porous materials

Area of Science:

• Materials Science
• Chemical Engineering
• Computational Chemistry

Background:

• Oxygen and nitrogen separation from air is vital for industrial and medical applications.
• Zeolites offer promising porous structures for gas storage and separation.
• Predicting gas uptake in zeolites is complex, requiring advanced modeling.

Purpose of the Study:

• To forecast oxygen (O2) uptake capacity in zeolites using advanced machine learning (ML) methods.
• To develop robust predictive models for O2 storage based on experimental data.
• To identify key factors influencing O2 uptake in zeolites.

Main Methods:

• Utilized ML techniques: Generalized Regression Neural Network (GRNN), Cascade Forward Neural Network, and Multilayer Perceptron.
• Constructed a database of 750 experimental O2 uptake values.
• Input features included pressure, pore volume, temperature, and surface area.

Main Results:

• The GRNN model demonstrated superior performance with a root mean square error of 0.03 and R² of 0.9991.
• Models accurately captured O2 uptake trends under varying pressure and temperature conditions.
• Sensitivity analysis indicated pressure positively influences O2 storage, while temperature has the most significant effect.

Conclusions:

• ML techniques are effective for precisely forecasting gas storage in zeolites.
• GRNN shows exceptional predictive power for O2 uptake capacity.
• Findings provide insights for advancing air separation technologies.