IEEE P3732

New IEEE Standard - Active - Draft.
The cloud-edge collaborative intelligent computing architecture in the power industry is specified. Construction specifications and functional boundaries for power network cloud-level intelligent agent (PNCIA) and power network edge-level intelligent agent (PNEIA) are defined. Electrical-computing collaborative regulation principles, resource scheduling mechanisms, and dynamic task offloading specifications are established. User priority classification standards are defined. Information security requirements covering system security, identity and access control,
communication security, and data privacy protection are established.

This standard focuses on the construction and technical specifications of cloud-edge collaborative intelligent agents in the power industry, while also addressing the intelligent computing requirements for complex scenarios in new-type power systems, including such complex computing tasks as equipment state perception, intelligent scheduling, rapid response, and fault recovery.
The standard comprises two major technical subsystems:
1. Design of cloud-edge collaborative intelligent computing architecture and technical requirements for intelligent agent construction:
1.1 Terms and definitions: Define basic terminology, system composition, and hierarchical structure of cloud-edge collaborative intelligent agents in the power industry.
1.2 An overall architecture design: Establishes an integrated cloud-edge-end intelligent agent system, clarifying the functional boundaries and interaction mechanisms between edge autonomy and cloud coordination.
1.3 Cloud-layer intelligent agent construction requirements: Determine the construction rules and capacity configuration of the cloud-layer agent.
1.4 Edge-layer intelligent agent construction requirements: Determine the construction rules and capacity configuration of the edge-layer agent.
2. Specifications for cloud-edge collaborative computing of intelligent agents in the power industry
2.1 Cloud-layer optimization and global learning mechanisms: Standardize a strategy formulation, global optimization, and multi-agent training processes for cloud-layer intelligent agents.
2.2 Edge-layer autonomous regulation and rapid response mechanisms: Define specifications for regional detection, autonomous optimization, and decision output for edge-layer intelligent agents.
2.3 Power-computing collaboration and regulation mechanisms: Establish mechanisms to ensure efficient, economical, and reliable operation of power and computing resources.
2.4 User priority identification and scheduling standards: Provide rules for classifying user load priorities and mapping them into regulation and optimization processes.
The purpose of this standard is to establish a unified technical framework for cloud-edge collaborative intelligent computing in the power industry. In refining computing mechanisms under varying communication, energy, and computational resource conditions, the standard enhances the intelligence level, operational efficiency, response resilience, and computing capabilities of the power sector.