基于图卷积神经网络的电力系统暂态稳定评估
摘要:随着新能源高渗透、特高压输电互联及电网数字化智能化等新兴技术的高速发展，电力系统结构变得更加复杂，因此对电力系统的安全稳定运行工作提出了更高要求。基于此，给出一种以图卷积神经网络模型为基础的暂态稳定分析方法，将短时模拟与神经网络预测相结合，减少暂态稳定分析所需周期，该方法适用于多种仿真分析场景。同时，将图卷积神经网络代理模型与遗传算法相结合，快速稳定地完成针对预想故障集的暂态稳定中预防控制优化策略的求解。与传统暂态稳定分析方法相比，提出的方法所需运行时间短、效率高。基于IEEE-30和IEEE-39节点系统的测试结果验证了所提方法的适用性、高效性和优越性。
Abstract：With the rapid development of emerging technologies such as high penetration of new energy, ultra-high voltage transmission interconnection and power grid digital intelligence, the structure of power system has become more complex and diverse, which puts forward various requirements for the safe operation and stability of the power system. Based on this, a transient stability analysis based on graph convolutional network model is given, which combines short-time simulation and neural network prediction to reduce the cycle time required for transient stability analysis. The method is applicable to a wide range of simulation analysis scenarios. Meanwhile, the agent model of graph convolutional network is combined with genetic algorithm to complete the solution of preventive control optimization strategy quickly and stably in transient stability for the contingency set. Compared with the traditional transient stability analysis, the method proposed in this paper requires shorter running time and is highly efficient. The test results based on the IEEE-30 and IEEE-39 node systems verify the applicability, efficiency and superiority of the proposed method.
标题：基于图卷积神经网络的电力系统暂态稳定评估
title:Transient Stability Assessment of Power Systems Based on Graph Convolutional Neural Networks
作者：薛栋烽,姚方,文福拴,张新宇,岳文全
authors:XUE Dongfeng,YAO Fang,WEN Fushuan,ZHANG Xinyu,YUE Wenquan
关键词：新能源,图卷积神经网络,预想故障集,暂态稳定,
keywords:new energy,graph convolutional network,contingency set,transient stability,
发表日期：2023-11-17
- 文件大小:
- 1.71 MB
- 下载次数:
- 60
-
高速下载
|
