Who's the strongest in "AI+Energy"? CSG reveals the results of the power prediction competition

 

Artificial intelligence (AI) is not just a "power-hungry monster," but also a core tool for energy transformation and efficiency improvement. Energy and AI have formed a mutually empowering and symbiotic relationship.

 

In the two-way rush of "Energy for AI" and "AI for Energy," some companies have formed a winning combination of "AI+Energy".

 

Recently, China Southern Power Grid (CSG) released the "Notice on the Annual Ranking of AI Load Forecasting Technology Competition Results for the 2024-2025 Season." The notice shows that Envision Smart won the championship with the highest comprehensive score.

 

The notice states that in order to fully implement the national "Artificial Intelligence+" action deployment, implement the digital development strategy requirements of CSG, and continuously promote the digital and intelligent transformation of the dispatching domain, CSG's dispatching center launched a year-long AI load forecasting technology competition in July 2024, using large-scale power grid dispatching load forecasting as a scenario. This attracted a large number of high-level talent teams to engage in fierce competition for the next-day and 10-day load forecasting (including unified dispatching caliber and system caliber) in the southern region (five provinces and regions + Shenzhen), achieving significant results.

 

Other teams that achieved excellent results along with Envision Smart include Beijing TsingSoft Innovation Company and China Southern Power Grid Artificial Intelligence Company, which won second and third place respectively.

 

In the notice, the first-ranked "fusion model" was obtained by CSG through a certain algorithm dynamically fusing the prediction results of teams with excellent performance, and is not an actual participating manufacturer.

 

 

 

This competition has high gold content, and the participating teams are very strong, including energy and power companies, universities and research institutes, AI and software companies, and even CSG's own AI company.

 

For example, TsingSoft Innovation Technology Co., Ltd., which won second place, was founded in 2005. Its founders and core management and technical personnel team come from well-known domestic universities such as Tsinghua University, Peking University, and Xi'an Jiaotong University. It is a high-tech enterprise specializing in power system load forecasting and decision-making software and related consulting services.

 

The company mainly provides software products and consulting services to customers in the fields of power market, grid dispatching, power planning, and power marketing load analysis and forecasting decision-making.

 

China Southern Power Grid Artificial Intelligence Technology Co., Ltd., which won third place, is wholly owned by CSG Digital Power Grid Research Institute Co., Ltd., and focuses on the research and development and application of artificial intelligence technology in the power field. The company undertakes the supply of core equipment for power artificial intelligence technology, basic infrastructure empowerment services, and the development of integrated solutions for artificial intelligence and energy. Its business covers core scenarios such as intelligent power system dispatching, load forecasting optimization, and intelligent equipment maintenance.

 

In September 2023, the company released an independent and controllable power large model innovation platform, building an AI application system covering seven scenarios including intelligent creation, equipment inspection, and power dispatching. In 2024, it carried out research on brain-like architecture equipment procurement and humanoid robot key technologies, and simultaneously implemented computing power server transportation, installation, debugging, and maintenance service projects in Guiyang, Guangzhou, and other places.

 

Envision Smart, which won first place, also has a significant background. It belongs to Envision Group, a world-leading green technology company. Its EnOS, the world's largest intelligent IoT operating system, has connected 788 GW of energy assets, equivalent to one-quarter of the global renewable energy installed capacity.

 

The author learned from the information that Envision Smart has won the championship many times in similar competitions.

 

According to the "Letter on Announcing the Results of the 2023 Renewable Energy Power Prediction Value Ecosystem Cultivation Plan Competition" released by CSG in 2023, among the participating teams in the renewable energy power generation prediction ecosystem cultivation plan, Envision Smart ranked first with a cumulative network-wide average score of 95.3437 points and won the competition's special prize.

 

CSG stated that the competition invited domestic and foreign renewable energy prediction algorithm teams to continuously conduct fair and just on-site technical competitions, striving to select "top-level" renewable energy prediction players suitable for the southern region from a global perspective and build a "flexible and open" renewable energy power prediction value ecosystem.

 

Image caption: In the 2024 load power prediction ecosystem cultivation plan, Envision Smart's comprehensive score was among the top.

 

Envision Smart is the only company in these competitions to win the championship in both power generation power prediction and load power prediction. It is also the only company that uses AI technology to empower the power generation side, load side, and power market simultaneously.

 

Envision's AI-enabled new power system has covered an innovative product system covering the entire chain of power generation, power grid, load, energy storage, and power plant management.

 

The series of AI load and power prediction technology competitions organized by CSG have attracted widespread attention in the industry because the competition lasts for a relatively long time, covering a full year. For AI prediction, long-term cycle prediction is very important. Predictions of ten days or one or two months are not as meaningful because they cannot cover various situations and curves of load changes. Only by extending it to one year and encountering various peak and valley cycles of the load can the performance of the AI strategy be fully reflected.

 

In addition, we know that CSG covers Guangdong, Guangxi, Hainan, Yunnan, and Guizhou provinces and is responsible for ensuring the power supply of Hong Kong and Macau. The climate, environment, and load in these regions vary greatly.

 

Load forecasting for these regions is a great test of AI technology level and understanding of renewable energy output. In addition to cities with high electricity demand such as the Pearl River Delta, it also faces weather forecasts brought about by different climatic environments in these provinces.

 

Among these provinces, there are coastal provinces such as Guangdong, Hainan, and Guangxi, which often face typhoon attacks, which have a huge impact and fluctuation on wind power.

 

There are also plateau areas such as Yunnan and Guizhou, which face various extreme weather conditions and rainfall, as well as extreme high and low temperatures. The Lingnan region sometimes experiences snow, and this special cold wave will bring additional load demand.

 

In such a complex environment, the prediction and matching of renewable energy power and load is a true test of the "AI+Energy" combination.

 

And this is not a superficial imitation of reality, but a real-world battle.

 

According to CSG, they are actually already using the competition prediction results to support actual daily operations.

 

After each team submits its prediction results, CSG will integrate these prediction results with better performance and directly push them to the provincial power grids' central dispatching centers for reference in their daily dispatching plans. Sometimes, they are also used for actual dispatching operations, including the arrangement of unit start-up and shutdown plans.

 

Renewable energy and AI are moving towards integration, which will bring fundamental changes to energy management.