At the 2025 AI for Good World Summit hosted by the International Telecommunication Union (ITU), the AI-based new energy meteorological power prediction solution, co-developed by China Huadian Corporation, Huawei, and Beijing Jiutian Weather, was selected and featured in the ITU’s AI for Good – Innovate for Impact Report 2025.
The solution also received the Innovate for Impact Use Case Award.
Leaders from China’s Ministry of Industry and Information Technology (MIIT) and the ITU Deputy Secretary-General, jointly unveiled a collection of case studies featuring practical AI solutions in areas like healthcare, climate resilience, education, and digital inclusion. Leaders from China’s MIIT and Deputy Director of the ITU Telecommunication Standardization Bureau, jointly presented the Innovate for Impact Use Case Winner Certificate to case representatives.
This solution stood out among 234 cases from 32 countries, indicating that the collaborative effort in the AI-driven new energy sector is highly recognized by international authorities and serves as an innovation model for global green energy development.
Launched by the ITU and over 40 United Nations agencies, the AI for Good Global Summit highlights groundbreaking AI applications in health, climate, and sustainable infrastructure. It stands as one of the world’s leading platforms for AI ethics and practices.
Zhou Qian, standards expert from Huawei’s Industry ICT Solution Development Management Department, shares the case
Technological innovation drives industry transformation
Based on Ascend AI platform, this solution creates a three-layer cloud-edge synergy architecture of “group meteorological large model – regional micro-meteorological engine – station edge power prediction.” This delivers intelligent new energy power forecasting.
The solution uses the world-leading Transformer-ViT architecture and integrates high-precision meteorological grid data with multi-modal time series data from stations, addressing the limitations of traditional models in handling complex weather conditions.
Leveraging the high-speed parallel computing capabilities of the Ascend AI NPU chip, a single server can generate high-precision meteorological predictions for the next 10 days within 10 seconds, more than ten thousand times faster than traditional CPUs.
In the pilot project at Jiangsu Huadian New Energy Company, the 15-minute weather forecast accuracy reached 97.24%, the 4-hour forecast accuracy was 91.72%, and the 24-hour forecast accuracy exceeded 90%.
Practical tests show that this solution can reduce assessment costs by 27% for wind power and 15% for solar power in the first half of 2024, slashing the operational costs of power stations.
The solution optimizes grid dispatch strategies through spatiotemporal dynamic analysis, effectively reducing wind and solar power curtailment.
It also optimizes electricity trading strategies by integrating tariff and load forecasting capabilities. At the device level, it supports AI-driven preventive maintenance, such as adjusting turbine positions in response to typhoon warnings to minimize equipment damage.
This use case offers a replicable technical approach for the efficient operations of new energy stations and the safe, stable running of power grids.
Leveraging digital, intelligent technologies for a zero-carbon future
Looking ahead, Huawei will continue to explore digital and intelligent technologies to build a zero-carbon future. By focusing on AI, IoT, and cloud-edge synergy, Huawei aims to develop more solutions to help global power customers create new power systems that are highly reliable, energy-efficient, and intelligent.
Together with our customers and partners, we will accelerate the clean, electric, and intelligent transformation of the global energy industry.
To learn more about Huawei intelligent new energy solutions, please visit: https://e.huawei.com/cn/industries/grid/power-generation/new-energy/intelligent-new-energy
Read more on Smart Energy International
