AI-based Hydropower Plant Operation Planning System Developed by ExaWizards and Chugoku Electric Power Co. to Go into Full-Scale Operation. -System takes supply and demand into account to create operating plans comparable to those of skilled operatorsEffective use of water resources to promote a decarbonized society

NEWS

ExaWizards Inc. (Headquarters: Minato-ku, Tokyo; Representative Director & President: Makoto Haruta; hereafter, “ExaWizards”) announce that its operation planning system for hydroelectric power plants will begin full-scale operation in FY2023, in cooperation with The Chugoku Electric Power Company. This is the result of joint development work that began in fiscal 2021. The use of AI in the operation planning system for hydroelectric power plants enables advanced control similar to that of skilled operators. The AI also takes into account the supply and demand of electricity and market prices to optimize conditions. The system is scheduled to move into actual operation at the two hydroelectric power plants at the two dams where trial operations were conducted.

☑ Objectives of the initiative
The reuse of renewable energy is essential for achieving a decarbonized society, and advancements in this area are a major point of focus. Particularly, at hydroelectric power plants, which have long been a mainstay in providing a stable supply of electricty, highly efficient use of water resources is needed.

The Chugoku Electric Power Company has developed an extensive base of knowledge on the technologies and restrictions in the hydroelectric power plant operation through years of its involvement. Staff in the field have drawn on the know-how to make operational adjustments based on weather forecasts and by predicting the intake of water in the dam.

In recent years, they also need to tailor operations based on fluctuations in the supply and demand for electricity. However, due to the complex array of conditions affecting the the supply and demand for electricity, elaborate planning for it has become a challenge. To address this, an AI and data-driven optimization solution was considered from the perspective of transfering technology and establishing stable and efficient business operations.

☑ Implemented solution
At dams and other hydroelectric power plants with a water reservoir, the allocation of stored water for power generation is planned out on a weekly range, in addition to the power generation plan for the next day . Previously, this planning process relied on a simple model to predict the level of water intake in the dam. Then, highly skilled technicians fomulated these plans based on empirical methods, and by choosing from several different power generation patterns.

In this joint development project, deep learning was combined with optimization techniques to develop an AI model capable of performing the following power generation planning tasks.

1. Predict water intake in the dam over a weekly range (around 10 days) using rainfall forecasts for the area in which the dam is located
2. Optimize the allocation of reservoir-stored water resources used for power generation over a weekly range using forecasts of the water intake in the dam, market supply and demand, and market rates over an extended period of time
3. Formulate optimized next-day power generation plans (in 30 minute increments) using forecasts of market supply and demand, and market rates

This development project involved AI consultants and machine learning engineers from ExaWizards, and dam engineers from The Chugoku Electric Power Company working together to substantiate a broad range of tacit operational know-how and to be incorporated into the AI model.

☑ Expected benefits
We trialed the operation of the AI model developed for formulating power generation plans in FY2021 and FY2022 at the Sasanamigawa Dam (Sasanamigawa Power Station) in Yamaguchi Prefecture and the Sufugawa Dam (Sufugawa-Daiichi Power Station) in Shimane Prefecture, operated by The Chugoku Electric Power Company.

As a result, these trials showed that the AI model was reliable enough to be used in the actual operation of power plants, and that it could come up with plans to concentrate power generation on periods of high demand than ever before. The use of this AI model is expected to further optimize the use of water resources at dams and other water reservoir-based hydroelectric power plants, achieving a pronounced decarbonization effect.

Looking ahead, the two companies will continue joint development efforts in an aim to roll-out this model at other dams, while working to address apparent issues identified during trial operations.

Further, in an effort to propel decarbonization initiatives, ExaWizards is considering rolling this system out at hydroelectric power plant dam owned by municipalities and other bodies. The company is also looking to combine this AI model with its other AI technologies in future development and social implementation projects with a view to expand it across a broad range of renewable energy fields.

ExaWizards aims to optimize energy and resource use through the use of AI to achieve decarbonization and energy conservation, and to realize GX (Green Transformation) and solve social issues.

[ExaWizards Corporate Profile]
Company name: ExaWizards Inc.
Location: 21F, Shiodome Sumitomo Building, 1-9-2 Higashi-Shinbashi, Minato-ku, Tokyo
Representative: Makoto Haruta, President and Representative Director
Business: Industrial innovation and resolution of social issues through the development of services utilizing AI
URL: https://exawizards.com/en/

<Contact for public relations>
E-mail address of the Public Relations Division of ExaWizards Inc.: publicrelations@exwzd.com