The Korea Electronics Technology Institute (KETI) has secured artificial intelligence technology capable of precisely diagnosing the battery status of Energy Storage Systems (ESS). This achievement was derived through joint research with overseas research institutions and is significant in that it is a technology designed with applicability in real-world operating environments in mind.

KETI announced on April 28 that it has developed AI technology capable of predicting the remaining lifespan and signs of abnormalities in ESS batteries in real time through international joint research with the University of Warwick in the UK. This research was conducted with support from the Korea Institute of Energy Technology Evaluation and Planning.

This technology differs from existing approaches in its data processing method and computational structure. First, it applies a distributed federated learning structure designed to enable training on each device rather than collecting data on a central server. This allows for maintaining or improving prediction accuracy while minimizing data transfer, and provides a structure that is advantageous in terms of personal information protection.

Another feature is the reduction in computational burden through model lightweighting. The research team improved the algorithm by reducing prediction errors and increasing processing speed compared to the existing approach. As a result, efficiency has been secured to the extent that AI models can be directly run even in battery management system (BMS) environments, rather than high-performance equipment.

The research results were published in 'Energy and AI,' an international academic journal in the fields of energy and artificial intelligence. Based on this, Dr. Ahn Jung-hoon, who led the joint research, was appointed as an honorary associate professor at the University of Warwick Manufacturing Group (WMG).

KETI plans to conduct technology verification centered on domestically produced ESS batteries and expand the scope of demonstration for related diagnostic technologies. At the same time, it intends to broaden the scope of technology application by participating in international joint research programs and standardization activities.