The digital transformation of the domestic electric power industry is in full swing.

On the 10th, the Electronics and Telecommunications Research Institute (ETRI) and Korea Southern Power Co., Ltd. signed an MOU for the development of artificial intelligence (AI) technology for power facility prediction and control to make power plant operations smarter.

The two organizations plan to carry out this collaboration centered on practical technology verification, including △facility status diagnosis and operation prediction, △robot technology to replace dangerous tasks, and △development of multimodal-based language models.

ETRI Honam Research Center is currently promoting the development project of the 'Electricity Field Specialized Foundation Model' with support from the Ministry of Science and ICT and the Institute of Information and Communications Technology Planning and Evaluation (IITP).

This model is designed to detect abnormal signs in power plant equipment at an early stage and aims for an accuracy of prediction error of less than 4.0%.

The robot-controlled AI that replaces high-risk tasks is expected to have a task success rate of over 98% and the ability to perform over 20 types of tasks.

The research team plans to develop a large-scale language model specialized for the power industry domain and implement an AI agent that responds to user queries at an expert level.

The target performance is an average score of 85 points or higher on the qualification exam and a situational judgment accuracy of 94% or higher.

This is a figure close to the level of technology that actual power plant operators can trust.

ETRI Headquarters Director Kang Hyeon-seo said, “We will respond to the needs of the field.“We will realize smart and safe power plant operation using English-Korean AI technology,” he said.

Ryu Ho-yong, director of Southern Power, also expressed his expectations, saying, “This cooperation will be an opportunity to increase the competitiveness of the energy industry.”

The two organizations are expected to actively promote sharing of future technological achievements and exchange of overseas cases, contributing to the expansion of the electric power AI ecosystem.