基于调控云的人工智能资源多级共享框架与关键技术  

Multi Level Sharing Framework and Key Technologies for Artificial Intelligence Resources Based on Dispatching and Control Cloud

作　　者：陶洪铸 王佳琪 陈建[3] 高尚 徐楷 王丁子睿 TAO Hongzhu;WANG Jiaqi;CHEN Jian;GAO Shang;XU Kai;WANG Dingzirui(National Power Dispatching and Control Center,State Grid Corporation of China,Beijing 100031,China;China Electric Power Research Institute,Beijing 100192,China;State Grid Tianjin Electric Power Company,Tianjin 300010,China;State Grid Electric Power Research Institute,Nanjing 211100,China;State Grid Beijing Electric Power Company,Beijing 100032,China;Faculty of Computing,Harbin Institute of Technology,Harbin 150006,China)

机构地区：[1]国家电力调度中心(国家电网),北京100031 [2]中国电力科学研究院有限公司,北京100192 [3]国网天津市电力公司,天津300010 [4]国网电力科学研究院有限公司,南京211100 [5]国网北京市电力公司,北京100032 [6]哈尔滨工业大学计算学部,哈尔滨150006

出　　处：《新型电力系统》2024年第4期371-385,共15页NEW TYPE POWER SYSTEMS

基　　金：国家电网有限公司总部科技项目(5108-202212052A-1-1-ZN)。

摘　　要：近年来人工智能技术在电网调控领域中的应用不断发展,随着调控云平台建设,相关业务对人工智能资源的多级共享提出了强需求。为此提出了一种基于调控云架构的人工智能资源多级共享框架,实现了数据样本集、算法模型、强化学习环境仿真器等资源在多级调控云上集成、存储、调用的全过程管理,同时阐述了电网连续运行数据样本生成,知识图谱分级共享,模块化仿真器构建等关键技术,并对该框架在调控业务场景中的应用进行了分析。该框架具备人工智能资源的全过程管理能力,能够为人工智能技术在调控领域中的深化应用提供有力的技术支撑。In recent years,the application of artificial intelligence technology in the field of power grid dispatch and control has continued to develop,and with the construction of dispatching and control cloud,related businesses have put forward a strong demand for the sharing of artificial intelligence resources on the two-level architecture.This paper proposes a multi-level sharing framework for artificial intelligence resources based on dispatching and control cloud architecture,which realizes the full-process management of integrating,storing,and calling artificial intelligence resources such as data sample set,algorithm model,and reinforcement learning environment simulator on the multi-level dispatching and control cloud.Meanwhile,it elaborates on key technologies such as the generation of continuous operation data sample for power grid,hierarchical sharing of knowledge graph,and modularized simulator construction,and analyzes the application of this framework in the field of dispatching and control business scenarios.This framework has the ability to manage the entire process of integrating,storing,and calling artificial intelligence resources.This framework possesses the capability of full-process management of artificial intelligence resources,and can provides strong technical support for the deepening application of artificial intelligence technology in the field of power grid dispatching and control.

关 键 词：电网调控 人工智能 资源共享 调控云 

分 类 号：TM7[电气工程—电力系统及自动化]