With a new $4 million grant from the U.S. Department of Energy (DOE), researchers at Case Western Reserve University will lead a project to improve the lifecycle of photovoltaic (PV) solar energy systems-from fabrication and electricity production through their end of service.
It's an effort to modernize and upgrade the nation's reliance on renewable energy using data science and artificial intelligence (AI). Research to understand and prevent common causes of early breakdowns and power loss, such as damage from harsh outdoor conditions and extreme weather, can extend solar-panel lifespans, the DOE said.
The Case Western Reserve-led project is headed by Roger French, the Kyocera Professor in Materials Science at Case School of Engineering and director of the university's Solar Durability and Lifetime Extension (SDLE) Research Center. The SDLE Research Center is leveraging its expertise in data science modeling and its Common Research Analytics and Data Lifecycle Environment, which provides high-performance computing to power these models.
"Taking advantage of our large repository of photovoltaic data and results, we can train a Neuro-Symbolic AI to learn from PV systems across their whole life," French said. "ArgoPV AI will allow the PV community to answer critical questions such as how to maximize PV power plant output or how to increase lifetime.
"ArgoPV will make the relationships among materials used, PV module design, how much energy is then generated and how this relates to end of life accessible for decision-making," he said. "This opens up possibilities for new materials and design choices in the future."
The grant is one of four the DOE committed for PV systems research and development, totaling $15.7 million. The goal: to reduce the cost and impact of solar energy technologies.
"The selected projects will maximize the environmental benefit of solar energy technologies by increasing system lifetime and work to facilitate materials recovery once the system is decommissioned," the DOE said. "Keeping solar panels in the field longer by making them more durable and easier to repair will slow and reduce the flow of solar panels into the waste stream."
Specifically, French and his team, which includes Yinghui Wu, Laura Bruckman and Jing Ma of CWRU, Kris Davis and Mengjie Li at the University of Central Florida, and Jen Braid (a former SDLE research associate professor) and Cliff Hansen of New Mexico-based Sandia National Laboratories, will focus on developing its ArgoPV technology.
"ArgoPV uses a holistic approach to the lifecycle of these power plants," French said. "By building an AI-empowered ecosystem, it allows researchers and PV owners to make decisions and optimize energy use, cost and environmental impact. ArgoPV will be the real-world research lab for PV systems and energy generation."
It is designed to quickly adapt, respond and model new hypothetical scenarios using generative AI. Traditional PV analytical systems receive and make decisions based on existing data-like how much energy a solar panel is producing at a certain time. ArgoPV includes the traditional method as one part of the lifecycle model.
"ArgoPV will help PV analysts and the PV community glean new insights into the lifecycle stages of these facilities," French said.
Collaborators providing data and industry input for the project include: Braskem US, Maxeon Solar, Silfab Solar Inc., SolarEdge Technologies, ESA Solar, Duke Energy Corp. and Univers.
