基于改进边图卷积网络的电力系统小干扰稳定评估模型  

Small-signal Stability Assessment Model Based on Improved Edge Graph Convolutional Networks of Power System

作　　者：郭梦轩 管霖[1,2] 苏寅生 姚海成[3] 黄济宇[1] 朱思婷 钟智 GUO Mengxuan;GUAN Lin;SU Yinsheng;YAO Haicheng;HUANG Jiyu;ZHU Siting;ZHONG Zhi(School of Electric Power,South China University of Technology,Guangzhou 510641,Guangdong Province,China;Guangdong Provincial Key Laboratory of Intelligent Operation and Control for New Energy Power System(China Southern Power Grid Electric Power Research Institute),Guangzhou 510663,Guangdong Province,China;Dispatching&Communication Center,China Southern Power Grid,Guangzhou 510530,Guangdong Province,China)

机构地区：[1]华南理工大学电力学院,广东省广州市510641 [2]广东省新能源电力系统智能运行与控制企业重点实验室(南方电网科学研究院),广东省广州市510663 [3]中国南方电网电力调度控制中心,广东省广州市510530

出　　处：《电网技术》2022年第6期2095-2103,共9页Power System Technology

基　　金：国家自然科学基金项目(52077080);南方电网公司科技项目(ZDKJXM20180084)。

摘　　要：小干扰稳定问题对电力系统安全稳定的影响日益突出。采用样本学习的思路建立从稳态运行信息到关键振荡模式的映射模型,为大电网振荡特性的快速预测和评估提供了新的技术路径。采用图卷积网络,并引入边卷积的设计来考虑输电通道潮流分布的影响,建立了小干扰稳定评估的边图卷积网络模型(edge graph convolutional networks for small-signal stability assessment,EGCN-SSA)。采用卷积增强技术改善网络退化现象,并建立多任务学习框架,同时预测多模式的振荡频率和阻尼比。在IEEE10机39节点上的算例和模型对比验证了所提出模型的性能以及对拓扑变化的适应能力。The influence of small-signal stability on the security and stability of the power system is becoming more and more prominent.A mapping model from the steady-state operation information to the important oscillation modes is established by using the idea of sample learning,which provides a new method for the rapid prediction and evaluation of the large power system oscillation characteristics.In this paper,the graph convolution networks based on the edge convolution are introduced to consider the influence of the power flow distribution in the transmission lines,establishing the edge graph convolutional networks for the small-signal stability assessment(EGCN-SSA).The convolution enhancement technology is designed to improve the network degradation.Meanwhile,a multi-task learning framework is established to predict the oscillation frequency and damping ratio of the important modes.The performance of the proposed model and its adaptability to the topological changes are verified on the IEEE 39-bus system.

关 键 词：小干扰稳定评估 图深度学习 边图卷积 多任务学习 

分 类 号：TM721[电气工程—电力系统及自动化]