A Machine Learning-Driven Pore-Scale Network Model Coupling Reaction Kinetics and Interparticle Transport for...

Area of Science:

• Chemical Engineering
• Catalysis Science
• Computational Chemistry

Background:

• Catalytic processes in porous systems involve complex interactions between microkinetics and transport phenomena.
• Accurate modeling requires bridging disparate spatial and temporal scales, posing significant computational challenges.
• Existing methods often struggle to capture the intricate interplay governing apparent catalytic performance.

Purpose of the Study:

• To develop an efficient multiscale modeling framework for reaction-transport coupled catalytic processes.
• To integrate machine learning to accelerate microkinetic modeling within a dual-network approach.
• To provide mechanistic insights into catalyst arrangement and transport limitations for reactor optimization.

Main Methods:

• Development of a pore-scale dual-network model with kinetics (DNMK).
• Integration of machine learning (ML)-based surrogates to accelerate the microkinetic module.
• Application and validation of the DNMK framework for sorption-enhanced CO2 hydrogenation to methanol.

Main Results:

• Achieved up to a 750-fold computational speed-up compared to traditional methods.
• Preserved full physical and chemical fidelity in the multiscale modeling.
• Identified optimal catalyst-sorbent configurations for enhanced apparent activity and reactor performance.

Conclusions:

• DNMK offers a high-resolution, ML-driven platform for digital catalytic experimentation.
• The framework enables predictive, in silico optimization of catalyst scaling, utilization, and process intensification.
• DNMK reduces reliance on experimental trials, paving the way for data-driven reactor design.