Process-Materials Integrated System Modeling & Co-optimization

Advanced energy systems consist of components across a wide range of scales (see above). For example, the process-materials integrated systems are commonly seen in energy technologies, where the process is empowered by novel materials and the functioning of materials relies on optimal process design. The overall performance depends on both materials and process decisions, which are integrated closely. And to optimize the integrated systems needs simultaneous decision-making at both scales, which is rarely discussed in the literature.

We aim to bridge this gap and co-optimize process-materials systems. Our proposed approach is to learn and incorporate materials surrogate models into equation-oriented process models, thus enabling direct numerical optimization. From a methodological point of view, our approach will push the materials-process decision resolution Pareto front (see below) to the blue region with equation-oriented process model and descriptor-level materials decision variables.

Specifically, our project focuses on pressure swing adsorption (PSA) processes and microporous materials such as metal-organic frameworks (MOFs) and zeolites, as they represent a promising technological pathway for next-generation energy-efficient, emission-free gas separation processes. We are also keeping our methodologies generic enough to be adapted to other process-materials co-optimization applications.

Learn optimizable microporous materials structure-function relationships

• Developed a standardized and automated surrogate model learning pipeline
• Introduce predictive ML model to improve surrogate modeling efficiency and quality
• Accuracy-complexity trade-off via selections of descriptors, isotherm models, learning objectives etc.

Develop and implement fidelity-tunable PSA process models

• Developed 1D packed-bed adsorption column models at 36 different fidelity levels
• Implemented the column model following IDAES-PSE standards as a custom unit model
• Developing numerical simulation capabilities of generic PSA cycles

On-going: Integrate and co-optimize microporous materials surrogate model and PSA process model