Smart Energy Co., Ltd.
The “2nd Solar Power Generation Prediction AI Competition” will be held to compete for ideas to solve the problem of solar power generation, where “power generation fluctuates depending on the weather” [Total prize money: 1 million yen]
Held as an industry-academia collaboration between Smart Energy, the University of Tokyo, and Weathernews. Entries start on Friday, June 21st.
……
Smart Energy Co., Ltd. (Headquarters: Minato-ku, Tokyo; Representative Director: Takuya Ogushi; hereinafter referred to as Smart Energy) is co-hosting the “2nd Solar Power Generation Prediction AI
Competition” in collaboration with ClimCORE, the University of Tokyo’s Institute for Advanced Science and Technology, and in cooperation with Weathernews Co., Ltd. ”
(https://www.solar-forecast.jp/) will be held, and entries will start from June 21st (Friday).
This competition aims to improve AI prediction technology and promote the utilization of advanced weather data, as well as to recognize unique technologies and ideas that solve the problems of solar power generation, which is said to be a naturally variable power source. A total of 84 teams entered. This year’s second event will differ from the previous one in terms of difficulty settings, etc., and we hope that more people will participate as an opportunity to feel the importance of prediction technology in solar power generation.
[Image: https://prtimes.jp/i/6241/21/resize/d6241-21-50711a3e672fc14e626a-0.png&s3=6241-21-fd030a10d90df029b8d0dacb2ac5ec2b-720×480.png] Background
Traditionally, solar power plants were mainly mega solar power plants installed on vast tracts of land, but due to a lack of suitable land, the trend is for low-voltage power generation equipment that can be installed on the roofs of buildings to increase in the future. As the number of low-voltage power generation facilities distributed across the country increases, it will become difficult to estimate and predict the amount of power generated, and the stable supply of solar power generation will become more complicated. In order to respond to the coming future, Smart Energy, which operates and maintains solar power plants, is working with Makoto Iida, a specially appointed associate professor at the University of Tokyo’s Institute for Advanced Science and Technology, and Weathernews Inc., to develop technology for predicting solar power generation using AI. We held this competition for the first time last year with the aim of further development.
Last year, more teams than expected participated, and we realized the high level of attention this field is receiving.This year, we decided to hold the “2nd Solar Power Generation Prediction AI Competition” with even more advanced content. did.
About the solar power generation prediction AI competition
The first competition last year was held to compete on the accuracy of AI models created to predict the amount of power generated at power plants owned by Smart Energy based on power generation data and weather data from the past two years. . A total of 84 teams
participated, and their model performance was evaluated using Smart Energy’s solar power generation prediction platform. The team with the highest accuracy was presented with the “Overall Forecast Award,” and the team with the highest prediction accuracy by power plant was presented with the “Short-term Forecast Award by Power Plant.” -Contents of the 2nd competition-
This year, we have made the following changes from last time. ■Change in event period
Last time, the competition period was for one month from July 17th (Marine Day) to August 11th (Mountain Day), but this year the competition period will be for two weeks from September 2nd to 13th. September tends to be more difficult to predict than midsummer due to the effects of typhoons, autumn rain fronts, etc., making predictions more difficult.
■Holding multiple departments
Based on the idea that verification and reproducibility of AI models are important in order to award prediction technologies that can be put into practical use, we will hold a competition in the following two categories.
・Open category
This is the department where you submit predicted values and source code. In addition to prediction accuracy, validation of the AI model will also be scored.
・Closing section
This is a category where grades are based only on predicted values. This applies to those who cannot submit source code.
(Prize settings are different from the open category)
Overview of “2nd Solar Power Generation Prediction AI Competition”
[Table 2: https://prtimes.jp/data/corp/6241/table/21_1_d7551ef97711627ec3498a44e728c63d.jpg] Details/Contact information
・“2nd Solar Power Generation Prediction AI Competition” Official website: https://www.solar-forecast.jp/
・Contact: Solar power generation forecast AI competition office: support@solar-forecast.jp
*Entries will be accepted from the official website after June 21st (Friday). Smart Energy Co., Ltd. Company Profile
Representative Director: Takuya Ogushi
Established: 2007
Location: 1-18-17 Nishi-Shinbashi, Minato-ku, Tokyo
Business details: Sales of renewable energy electricity such as solar power and hydropower, planning, design, construction, management, maintenance and sales of power generation facilities, equipment and systems, solicitation of investment in environmental funds, operation and countermeasures, and emissions. Quantity system
research/consulting and business support
Solar O&M results: 4,357MW (as of February 2024, group total) HP: https://www.smart-energy.jp/
Instagram: https://www.instagram.com/chikyunomikata_smartenergy/ Founded in 2006. We are a group of professionals who aim to realize a decarbonized society through business, with the mission of “Connecting a Beautiful Earth.” We provide solutions to solve global environmental problems through renewable energy operation and maintenance (O&M), new power business support, asset management business, renewable energy citizen fund formation business, environmental consulting business, etc.
More details about this release:
https://prtimes.jp/main/html/rd/p/000000021.000006241.html