A technology has been developed that can diagnose the health of used batteries and easily determine whether they can be recycled. This technology is expected to have a significant impact on battery health management, as it can diagnose devices regardless of their type.

UNIST (President Yong-Hoon Lee) announced on the 8th that a team led by Professors Dong-Hyeok Kim and Yun-Seok Choi of the Department of Energy and Chemical Engineering and Professor Han-Kwon Lim of the Carbon Neutral Graduate School developed DeepSUGAR, a system that can diagnose the health status of battery components based on deep learning, which trains computers independently.

It combines the generative artificial intelligence technology 'Generative Adversarial Network (GAN)', which creates new creations through learning, and the 'Convolutional Neural Network (CNN)', which can effectively process images.

DeepSUGAR converts voltage, current, and capacity data obtained during charging and discharging of lithium batteries into the three primary color values of light and visualizes them. Based on this, a deep learning model is used to predict the battery's health. It can be applied regardless of battery configuration, such as modules and packs, and is differentiated from existing battery diagnosis methods.

Professor Kim Dong-hyeok explained, “By leveraging DeepSUGAR’s ability to visualize charge and discharge data, we have built a verification system that can determine whether a used battery can be recycled without disassembling it.”

The system developed by the research team utilizes generative AI to extract charge and discharge data from battery modules based on the battery's health status. This allows the team to determine whether the internal modules are recyclable without disassembling the battery or conducting actual charge and discharge tests.

First author Seo-Jeong Park, a researcher in the integrated master's and doctoral program in the Department of Energy and Chemical Engineering, explained, "The developed system was able to simplify the recycling process by using generative AI to check whether each internal component module can be reused just by charging and discharging the pack." She added, "It is expected to make a great contribution to the field of battery recycling because it can be applied universally without being limited to the type of device."

In addition, co-first author Dongjun Lim, a researcher in the integrated master's and doctoral program in the Department of Energy and Chemical Engineering, said, "Not only in the field of battery recycling, but also in real life, we can predict the health status of internal modules through battery pack diagnosis." He added, "We expect that this will help realize green energy in various fields in the future because only modules with degraded performance can be replaced."

This research was supported by the UNIST Carbon Neutrality Demonstration Center, the Ministry of Science and ICT's Bio-Medical Technology Development Project, and the Korea Industrial Technology Evaluation and Planning Institute with funding from the Ministry of Trade, Industry and Energy. It was published online on October 17th in the international academic journal 'Journal of Materials Chemistry A' and was selected as the cover paper for the November issue.