大电网调控人机混合增强智能:概念内涵、应用框架、关键技术以及系统验证OA北大核心CSTPCD

With the rapid development of artificial intelligence technologies such as deep learning,it has become an inevitable trend for the new power systems in the future to realize the informatization,digitization and intelligence of power grids through digitally empowered power.The application of artificial intelligence technology in power grid regulation and control is more extensive,but in the process of practical application,due to the uncertainty,openness and vulnerability of the power grid,it is faced with the problems of model credibility and interpretability,which makes the application of pure artificial intelligence technology difficult to meet the needs of power grid regulation with safety requirements.In order to solve the above problems,it is necessary to introduce human supervision and interaction,integrate the human intelligence and machine intelligence,and realize the hybrid enhancement of human-machine intelligence in power grid regulation.This paper first introduces the concept and connotation of hybrid-augmented intelligence(HAI)technology and its research status in different fields.Then,according to the characteristics of power grid regulation,from the two dimensions of artificial intelligence and regulation application,the application framework of human-machine hybrid-augmented intelligence(HHI)for power grid regulation is proposed,and the goal and basic principle of human-machine interaction coordination for power grid regulation are presented.On this basis,the key technologies involved in the HHI of power grid regulation are analyzed and discussed,and the human-machine hybrid emergency control decision support verification system for large power grid regulation is developed,which can be used for the application test and verification of various power grid regulation scenarios such as thermal stability adjustment and frequency control of provincial power grid.It provides reference for the application of HHI technology in the field of power grid regulation.