Page 246 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

• Extend forecast duration from medium-short-term (within 10 days) to cross-month and
cross-quarter predictions to guide long-term power trading and grid planning.
• Develop standards for AI-based meteorological forecasting in the energy sector.
• Develop a standardized knowledge base for AI-based meteorological forecasting in
collaboration with industry partners and regulatory bodies. This knowledge base will
include standardized data formats metadata schemas, and forecasting protocols to
ensure interoperability and scalability across renewable energy operators. The goal is
to establish an industry benchmark for AI-driven meteorological forecasting, promoting
adoption and consistency in the energy sector.
• Enhanced datasets for precipitation, temperature, and market-related data from partners.
• Develop advanced model for long-term forecasting.
• Deepen partnerships with meteorological firms to enhance data quality and modeling
support.

3 Use Case Requirements

REQ-01: It is required to have access to large-scale historical and real-time meteorological
data from partners like Beijing Jiutian Meteorological Technology Co., Ltd. for training and
fine-tuning the AI Large Model.

REQ-02: It is required to have cloud-edge collaboration infrastructure to support the three-
layer architecture (group, regional, and site-level processing).

REQ-03: It is required to have high-performance AI computing ability for training the
meteorological model efficiently.

REQ-04: It is required to have integration with power station systems and electricity trading
platforms to deliver real-time forecast data for grid dispatching, reporting, and trading.

REQ-05: It is required to comply with regional energy regulations (e.g., Jiangsu Province’s
forecasting standards) and electricity market requirements to avoid penalties and optimize
trading.

210