Merging physical domain knowledge with AI improves prediction accuracy of battery capacity

Electric vehicles (EVs) have recently become commonplace, appearing in everything from passenger cars to buses and taxis. Electric vehicles offer the advantages of being environmentally friendly and having cheap maintenance expenses, but their owners must be cautious of deadly accidents if the battery runs out or approaches the end of its life. As a result, accurate capacity and lifespan projections for lithium-ion batteries, which are widely utilized in electric vehicles, are critical.

Professor Seungchul Lee of POSTECH and Ph.D. candidate Sung Wook Kim of the Department of Mechanical Engineering collaborated with Hanyang University's Professor Ki-Yong Oh to develop a novel artificial intelligence (AI) technology that can accurately predict the capacity and lifespan of lithium-ion batteries. This breakthrough in study, which significantly enhanced prediction accuracy by combining physical domain knowledge with AI, was just published in Applied Energy, an international academic publication in the subject of energy.

A physics-based model, which simplifies the sophisticated internal structure of batteries, and an AI model, which leverages the electrical and mechanical reactions of batteries, are the two ways for forecasting battery capacity. The traditional AI model, on the other hand, needs a lot of data to train. Furthermore, when applied to untrained data, its prediction accuracy was poor, necessitating the development of next-generation AI technology.

The researchers coupled a feature extraction technique that varies from traditional methods with physical domain knowledge-based neural networks to efficiently forecast battery capacity with less training data. As a consequence, battery prediction accuracy improved by up to 20% while testing batteries of varying capacity and lifespan distributions. The consistency of the results was confirmed to assure their dependability. These findings are expected to pave the way for the application of highly trustworthy physical domain knowledge-based AI to a variety of sectors.

POSTECH's Professor Lee stated: "The limitations of data-based AI have been overcome using physics knowledge. The difficulty of building big data has also been alleviated thanks to the development of the differentiated feature extraction technique."

"Our research is significant in that it will contribute in propagating EVs to the public by enabling accurate predictions of remaining lifespan of batteries in next-generational EVs," said Professor Oh of Hanyang University.