Scientists receive $1M grant for AI-powered clean energy research

As the world races to combat environmental degradation and climate challenges, transitioning to renewable energy has become a top priority. However, the inconsistent nature of wind, solar, and other renewable sources poses a significant challenge to maintaining a stable energy supply, which has slowed the transition.

An interdisciplinary team of scientists is collaborating to find a workable solution by harnessing the power of artificial intelligence and microwave plasma, blending knowledge from chemistry, materials science, and engineering.

The National Science Foundation awarded a $1 million grant to this project, titled “Multidisciplinary High-Performance Computing and Artificial Intelligence Enabled Catalyst Design for Micro-Plasma Technologies in Clean Energy Transition.”

This project aims to leverage machine learning for catalyst discovery and develop new characterization methods for studying chemical reactions under extreme conditions such as plasma. The goal is to improve the efficiency of catalysts in hydrogen generation, carbon capture, and energy storage.

The University of Houston team includes Jiefu Chen, associate professor of electrical and computer engineering; Lars Grabow, professor of chemical and biomolecular engineering; Xiaonan Shan, associate professor of electrical and computing engineering; and Xuquing Wu, associate professor of information science technology. They are collaborating with Su Yan, an associate professor of electrical engineering and computer science at Howard University.

“By enhancing the efficiency of catalytic reactions in key areas such as hydrogen generation, carbon capture and energy storage, this research directly contributes to these global challenges,” said Chen, the principal investigator of the project. “This interdisciplinary effort ensures comprehensive and innovative solutions to complex problems.”

Discovering materials for new catalytic processes is a slow and challenging task requiring expertise in various disciplines such as robotics, AI, material science synthesis testing modeling. The researchers plan to assemble a robotic synthesis and testing facility while simultaneously programming the AI model for unsupervised learning.

Automating the experimental testing process with robotic facilities will make catalyst design more efficient by closely integrating theory and experiments through advanced machine learning techniques according to Shan and Wu.

The project has four major research thrusts:

1. Machine learning-driven catalyst discovery for plasma-assisted chemical reactions: The team will use a graph neural network model trained on the Open Catalyst Project dataset.

2. Multiscale multiphysics microwave-plasma simulation: New methods will be developed to model complex interactions involving electromagnetics plasma physics thermodynamics.

3. Design synthesize characterize catalyst support material architecture: Researchers will optimize supports for efficient microwave-assisted reactions.

4. Bench scale demonstration using micro-plasma catalyst system: A bench scale reactor will demonstrate efficiency.

Another important component is establishing a multidisciplinary research education program involving machine learning computational catalysis applied electromagnetics material synthesis advanced characterization. It aims to educate train next-generation STEM workforce.

“This project will help create a knowledgeable skilled workforce capable addressing critical challenges clean energy transition,” Grabow said. “Moreover interdisciplinary project transformative advances insights knowledge lead tangible economic impact future.”

He said the team is open partnering industry related projects further development during after project.