Development of AI-based Hydropower Planning Support System and Patent Application

Chubu Electric Power Co., Inc. (President & Director: Kingo Hayashi; "Chubu Electric Power") and TSUNAGU Community Analytics Co., Inc. (President: Osamu Kuribayashi; "TCA") have developed a system that uses AI to support optimal power generation planning at hydroelectric power plants and yesterday jointly filed a patent application for the AIs incorporated in the system.

Chubu Electric Power is working to increase power generation from existing power sources as well as developing new power sources in order to realize a decarbonized society. For hydroelectric power generation, Chubu Electric Power aims to maximize power generation by improving the accuracy of forecasting water inflow into dams through the promotion of DX and by operating with an optimal power generation plan.
In addition, power generation plans for hydroelectric power plants have been currently prepared by skilled employees, who spend about half a day based on their experience and know-how, causing issues to improve operational efficiency and train younger employees.

The system enables to simultaneously optimize the power generation operations of multiple hydroelectric power plants where water flows in and out of different rivers, which consists of the three AIs listed below.

(1) Inflow Forecasting AI: Predicts the amount of water inflow into the dam based on the next day's forecast rainfall.
(2) Historical Search AI: Searches for past power generation plans under similar conditions based on conditions such as dam water levels and available water volume on the following day.
(3) Optimization Calculation AI: Creates power generation plans based on conditions such as predicted inflow at each dam, dam water levels, and power generation operation constraints to meet objectives such as maximum power generation and maximum power sales.

Chubu Electric Power and TCA have together conducted the verification at hydroelectric power plants under Chubu Electric Power's service area in the Hida River system, which has a particularly complex configuration, from April 2021. As a result, it is confirmed an annual power increase of approximately 2% (up to approximately 30 million kWh/year for the entire water system, equivalent to the annual power consumption of approximately 10,000 standard households) can be expected through operation based on the power generation plan formulated using the system.

It is also possible to reduce the time required to develop power generation plans to less than a quarter of the conventional time by utilizing the system, significantly improving efficiency. Furthermore, the system will enable the training of young employees, which until now has mainly involved on-the-job training, to acquire skills in a short period of time through simulations using the system.

Starting with full-scale operation in the Hida River system in FY2024, the two companies will continue to introduce the system into other water systems, thereby contributing to the realization of a decarbonized society by continuously working to increase the amount of electricity generated by hydropower without emitting CO₂.

Attachment

• Overview of AI-based hydropower planning support system