Development of AI for Automated Management of Renewable Energy Power Plants Completed... Commercialization as Early as December

[Financial News] Development has been completed on an artificial intelligence (AI) agent capable of managing renewable energy power plants in real time, with commercialization planned as early as December. The technology is expected to reduce renewable energy generation losses, ease management burdens, and centralize research on renewable energy power plant data.
According to sources in the science community on the 4th, KISTI posted a request for proposals on the Korea Online e-Procurement System (KONEPS) on the 1st of this month for the development of an agent-based data preprocessing and AI model training pipeline to utilize operations and maintenance (O&M) AI models for renewable energy power plants. The notice marks the final stage in commercializing three types of AI agents for renewable energy, whose development KISTI has completed. The project will run through December 11 after the contract is signed with the selected company. The winning bidder will be responsible for the user interface (UI) and design work for the completed AI agents. Commercialization is planned as early as December.
Park Hyung-wook, head of KISTI's Honam branch and the project lead, stated, "These AI agent models were developed as part of the energy sector of the 'Region-Led AI Transformation Project,' and all three models have been completed. We expect that AI models in Honam, which accounts for more than 30% of renewable energy nationwide, will be unified, enabling standardized research."
In practice, the AI agent is a service that monitors power generation status based on field data collected through various sensors, including inverters and environmental sensors, at power plants. It can forecast short-term power generation and detect abnormalities in generation equipment.
A KISTI official from the Honam branch of the Data Analysis Research Division explained, "The system is configured to allow users to ask various questions through the large language model (LLM) agent screen arranged on the side. This LLM agent is equipped with a short-term power generation forecasting AI model developed by KISTI, a System Marginal Price (SMP)/Renewable Energy Certificate (REC) forecasting AI model, and an equipment anomaly detection AI model. It supports users in making decisions about the status of the power plants they own."
The agent is expected to become available around December through an open-platform portal to be established at Jeonnam Technopark (TP). Users will be able to access it by registering the Sitekey issued for each power plant. The service targets small-scale power generation operators, and free access is planned for small and medium-sized companies located in South Jeolla Province. A demonstration is currently underway with five companies.
According to KISTI, renewable energy power plants that have been in operation for more than 10 years are experiencing abnormalities and performance degradation, and poor management can result in cumulative generation losses of more than 30%. Although systematic O&M services are therefore needed, conditions vary from plant to plant, including the types of inverters used, and there is also a shortage of specialized service providers. Real-time monitoring of generation equipment through the AI agent is expected to enable faster responses to abnormalities and improve the profitability of renewable energy operators.
[email protected] Yeon Ji-an Reporter