基于图深度学习的电网调度系统人机交互模式  

Human computer interaction mode of power grid dispatching system based on graph deep learning

作　　者：辛阔[1] 单政博 陈兴望 程哲[1] XIN Kuo;SHAN Zheng-bo;CHEN Xing-wang;CHENG Zhe(China Southern Power Grid Power Dispatching Control Center,Guangzhou 510663,China)

机构地区：[1]中国南方电网电力调度控制中心,广州510663

出　　处：《信息技术》2022年第9期61-66,共6页Information Technology

基　　金：中国南方电网有限责任公司科技项目(ZDKJX-M20180074)。

摘　　要：针对传统电网调度系统存在数据检索时间长、传输质量差的问题,提出了基于图深度学习的电网调度系统人机交互模式优化设计。采用最小二乘法,计算交互平面参数,得到节点与人机交互平面的距离限差,完成电网调度系统人机交互数据的预处理,减少数据噪声点;利用人机交互模式的自动布局,建立了优化设计数学模式,结合图深度学习的运行方法,求解了人机交互模式的布局,实现电网调度系统人机交互模式的优化设计。实验结果表明,该方法不仅可以缩短人机交互数据的检索时间,还可以提高数据的传输质量。On the basis of the problems of long data retrieval time and poor transmission quality in traditional power grid dispatching system,an optimal design of human-computer interaction mode for power grid dispatching system based on graph deep learning is proposed.The least square method is used to calculate the parameters of the interaction plane,and the distance tolerance between the node and the human-computer interaction plane is obtained,so as to complete the data preprocessing of the human-computer interaction of the power grid dispatching system and reduce the data noise points.By using the automatic layout of the human-computer interaction mode,the optimal design of mathematical model is established,and the layout of the human-computer interaction mode is solved by combining with the operation method of graph deep learning,therefore,realizing the power grid dispatching The optimization design of man-machine interaction mode of network dispatching system.Experiment results show that this method can not only shorten the retrieval time of human-computer interaction data,but also improve the quality of data transmission.

关 键 词：图深度学习 电网调度系统 人机交互模式 自动布局模式 

分 类 号：TM734[电气工程—电力系统及自动化]