基于因果启发的图深度学习电力系统暂态稳定评估方法

赵晨浩; 焦在滨; 张林博; 顾瀚文

doi:10.7652/xjtuxb202611001

您当前的位置：

首页 >

文章列表页 >

基于因果启发的图深度学习电力系统暂态稳定评估方法

更新时间：2026-11-09

- 基于因果启发的图深度学习电力系统暂态稳定评估方法
- Transient Stability Assessment of Power Systems Based on Causal-Inspired Graph Deep Learning
- 西安交通大学学报 2026年60卷第11期页码：1-13
- 作者机构：
  
  西安交通大学电气工程学院，710049，西安
- 作者简介：
  
  赵晨浩（1995—），男，博士生；
  焦在滨（通信作者），男，教授，博士生导师。
- 基金信息：
  
  智能电网重大专项（2030）资助项目(2025ZD0806000)
- DOI：10.7652/xjtuxb202611001
  中图分类号： TM732
- 收稿：2026-01-20，
  
  网络首发：2026-06-01，
  
  纸质出版：2026-11-10
- 稿件说明：
移动端阅览
赵晨浩, 焦在滨, 张林博, 等. 基于因果启发的图深度学习电力系统暂态稳定评估方法[J/OL]. 西安交通大学学报,2026,60 (11):1-13. https://doi.org/10.7652/xjtuxb202611001.

ZHAO Chenhao, JIAO Zaibin, ZHANG Linbo, et al. Transient Stability Assessment of Power Systems Based on Causal-Inspired Graph Deep Learning[J/OL]. Journal of Xi'an Jiaotong University,2026,60 (11):1-13. https://doi.org/10.7652/xjtuxb202611001.
赵晨浩, 焦在滨, 张林博, 等. 基于因果启发的图深度学习电力系统暂态稳定评估方法[J/OL]. 西安交通大学学报,2026,60 (11):1-13. https://doi.org/10.7652/xjtuxb202611001. DOI：

ZHAO Chenhao, JIAO Zaibin, ZHANG Linbo, et al. Transient Stability Assessment of Power Systems Based on Causal-Inspired Graph Deep Learning[J/OL]. Journal of Xi'an Jiaotong University,2026,60 (11):1-13. https://doi.org/10.7652/xjtuxb202611001. DOI：

摘要

为了识别影响暂态稳定评估的关键拓扑结构和特征，提升因果关系的建模能力和模型决策的可解释性，提出一种基于因果启发的边图注意力网络电力系统暂态稳定评估方法。首先，搭建边图注意力网络编码器，联合提取节点与边特征，利用节点和边的注意力掩码，将电网图数据显式解耦为因果子图与非因果子图；其次，在因果子图上并行构建失稳机组识别任务，通过施加隐式因果约束，确保因果子图包含失稳关键节点与连接边；然后，引入因果干预与对比学习策略，消除非因果特征的虚假关联，增强模型对因果特征的鲁棒学习；最后，联合优化各部分损失项，利用因果子图实现准确的暂态稳定评估和关键失稳发电机的识别。在CEPRI-TAS-173算例上的仿真和测试结果表明：所提方法的暂态稳定评估准确率达99.01%，失稳机组识别准确率达98.07%，各因果模块的添加使得F1分数提升了3.69个百分点，验证了该方法具有更好的评估准确性、鲁棒性与可解释性。研究结果可为电网安全预警和紧急控制提供可靠决策依据。

Abstract

To identify the critical topological structures and features influencing transient stability and to enhance the causal modeling capability and the interpretability of model decisions

this paper proposes a causal-inspired edge-featured graph attention network for transient stability assessment of power systems.First

an edge-featured graph attention network encoder is constructed to j ointly extract node and edge features

and the grid graph data is explicitly decoupled into a causal subgraph and a non-causal subgraph using node and edge attention masks. Next

an unstable generator identification task is built in parallel on the causal subgraph to impose implicit causal constraints

ensuring that the causal subgraph contains the critical nodes and connecting edges associated with instability.Then

causal intervention and contrastive learning strategies are introduced to eliminate spurious correlations caused by non-causal features and to enhance the robust learning of causal features.Finally

the loss terms of each part are jointly optimized

and the causal subgraph is utilized to achieve accurate transient stability assessment and identification of critical unstable generators.Simulation and test results on the CEPRI-TAS-173 test case show that the proposed method achieves a transient stability assessment accuracy of 99.01 % and an unstable generator identification accuracy of 98.07%. Furthermore

the incorporation of causal modules improves the F1 score by 3.69 percentage points

demonstrating the superior assessment accuracy

robustness

and interpretability of the proposed method.The research findings can provide a reliable basis for decision-making in early warning and emergency control of power systems.

关键词

Keywords

references

赵兵,吴萍,蒋彦翃,等.响应驱动的新型电力系统暂态稳定判别方法综述[J].电网技术,2025,49(7):2633-2648.

Zhao Bing,Wu Ping,Jiang Yanhong,et al.A survey on response-driven transient stability determination of new generation of power system[J].Power System Technology,2025,49(7):2633-2648.

胡博,谢开贵,邵常政,等.双碳目标下新型电力系统风险评述:特征、指标及评估方法[J].电力系统自动化,2023,47(5):1-15.

Hu Bo,Xie Kaigui,Shao Changzheng,et al.Commentary on risk of new power system under goals of carbon emission peak and carbon neutrality:characteristics,indices and assessment methods[J].Automation of Electric Power Systems,2023,47(5):1-15.

Xiao Tannan ,Wang Jianquan,Gao Yifan,et al.Improved sparsity techniques for solving network equations in transient stability simulations [J].IEEE Transactions on Power Systems,2018,33(5):4878-4888.

刘俊,孙惠文,吴柳,等.电力系统暂态稳定性评估综述[J].智慧电力,2019,47(12):44-53.

Liu Jun,Sun Huiwen,Wu Liu,et al.Overview of transient stability assessment of power systems[J]. Smart Power,2019,47(12):44-53.

杨博,陈义军,姚伟,等.基于新一代人工智能技术的电力系统稳定评估与决策综述[J].电力系统自动化,2022,46(22):200-223 .

Yang Bo,Chen Yijun,Yao Wei,et al.Review on stability assessment and decision for power systems based on new-generation artificial intelligence technology[J]. Automation of Electric Power Systems,2022,46 (22):200-223 .

闫炯程,李常刚,刘玉田.数据驱动的新型电力系统安全风险预警综述[J].电网技术,2024,48 (12):4989-5002 .

Yan Jiongcheng,Li Changgang,Liu Yutian.Review of data-driven security risk early warning research of the new-type power system[J].Power System Technology,2024,48(12):4989-5002.

甄永赞,阮程.基于代价敏感支持向量机和多变量决策树的分级自适应暂态电压稳定评估[J].电网技术, 2024,48(2):778-788.

Zhen Yongzan,Ruan Cheng.Hierarchical self-adaptation transient voltage stability assessment based on cost-sensitive SVM and multivariate decision tree[J]. Power System Technology,2024,48(2):778-788.

戴远航,陈磊,张玮灵,等.基于多支持向量机综合的电力系统暂态稳定评估[J].中国电机工程学报, 2016,36(5):1173-1180.

Dai Yuanhang,Chen Lei,Zhang Weiling,et al.Power system transient stability assessment based on multisupport vector machines[J].Proceedings of the CSEE, 2016,36(5):1173-1180.

张哲,秦博宇,高鑫,等.融合注意力机制和卷积神经网络的电网暂态电压稳定评估及可解释性分析[J].电网技术,2024,48(11):4648-4657.

Zhang Zhe,Qin Boyu,Gao Xin,et al.Transient voltage stability assessment and interpretability analysis of power grids based on the fusion of attention mechanism and convolutional neural network[J].Power System Technology,2024,48(11):4648-4657.

Kim J,Lee H,Kim S,et al.Real-time power system transient stability prediction using convolutional layer and long short-term memory [J].Journal of Electrical Engineering & Technology,2023,18(4):2723-2735.

杜一星,胡志坚,李犇,等.基于双向门控循环单元的电力系统暂态稳定评估[J].电力系统自动化, 2021,45(20):103-112.

Du Yixing,Hu Zhijian,Li Ben,et al.Transient stability assessment of power system based on bi-directional gated recurrent unit[J].Automation of Electric Power Systems,2021,45(20):103-112.

黄济宇,管霖,郭梦轩,等.图深度学习技术在智能暂态稳定评估中的应用及展望[J].电网技术,2023 , 47(4):1500-1511.

Huang Jiyu,Guan Lin,Guo Mengxuan,et al.Application and prospect of graph deep learning technique in intelligent transient stability assessment[J].Power System Technology,2023,47(4):1500-1511.

管霖,黄济宇,蔡锱涵,等.图深度学习技术在电力系统分析与决策领域的应用与展望[J].高电压技术, 2022,48(9):3405-3422.

Guan Lin,Huang Jiyu,Cai Zihan,et al.Application and prospect of graph deep learning technique in power system analysis and decision[J].High Voltage Engineering,2022,48(9):3405-3422.

王铮澄,周艳真,郭庆来,等.考虑电力系统拓扑变化的消息传递图神经网络暂态稳定评估[J].中国电机工程学报,2021 ,41(7):2341-2349.

Wang Zhengcheng,Zhou Yanzhen,Guo Qinglai,et al. Transient stability assessment of power system considering topological change:a message passing neural networkbased approach[J].Proceedings of the CSEE,2021, 41(7):2341-2349 .

刘雨晴,刘曌,王小君,等.融合同步知识和时空信息的电力系统暂态稳定评估框架[J].电网技术, 2025 ,49(6):2334-2346.

Liu Yuqing,Liu Zhao,Wang Xiaojun,et al.Power system transient stability assessment framework based on fusion of synchronization knowledge and spatialtemporal information[J].Power System Technology, 2025 ,49(6):2334-2346.

钟智,管霖,苏寅生,等.基于图注意力深度网络的电力系统暂态稳定评估[J].电网技术,2021,45(6):2122-2130.

Zhong Zhi,Guan Lin,Su Yinsheng,et al.Power system transient stability assessment based on graph attention deep network[J].Power System Technology,2021 , 45(6):2122-2130.

Xia Shiwei,Zhang Chenhui,Li Yahan,et al.GCNLSTM based transient angle stability assessment method for future power systems considering spatialtemporal disturbance response characteristics [J].Protection and Control of Modern Power Systems, 2024,9(6):108-121.

傅太国屹,杜友田,吕昊,等.基于多尺度图注意力网络的电力系统暂态稳定评估[J].电力系统自动化, 2025 ,49(3):60-70.

Fu Taiguoyi,Du Youtian,Lü Hao,et al.Transient stability assessment of power systems based on multiscale graph attention network[J].Automation of Electric Power Systems,2025,49(3):60-70.

王小君,窦嘉铭,刘曌,等.可解释人工智能在电力系统中的应用综述与展望[J].电力系统自动化, 2024,48(4):169-191.

Wang Xiaojun,Dou Jiaming,Liu Zhao,et al.Review and prospect of explainable artificial intelligence and its application in power systems[J].Automation of Electric Power Systems,2024,48(4):169-191.

刘晓菲,尹硕,张艺涵,等.基于可解释图注意力网络的新型电力系统无功优化[J].电网技术,2026, 50(3):1008-1018.

Liu Xiaofei,Yin Shuo,Zhang Yihan,et al.Reactive power optimization of new power systems based on interpretable graph attention network[J]. Power System Technology,2026,50(3):1008-1018.

Wang Ziming, Chen Jun, Chen Haopeng. EGAT:edge-featured graph attention network [C]//Artificial Neural Networks and Machine Learning-ICANN 2021 . Cham,Germany: Springer International Publishing, 2021:253-264.

文载道,王佳蕊,王小旭,等.解耦表征学习综述[J].自动化学报,2022,48(2):351-374.

Wen Zaidao,Wang Jiarui,Wang Xiaoxu,et al.A review of disentangled representation learning[J].Acta Automatica Sinica,2022,48(2):351-374.

赵梦阳,张旭,曲存全.GraphCCL:自注意力增强的因果图对比学习[J].指挥与控制学报,2025,11(5):569-576.

Zhao Mengyang,Zhang Xu,Qu Cunquan.GraphCCL:causal graph contrastive learning with self-attention augmentation[J].Journal of Command and Control, 2025,11(5):569-576.

肖智伟,王童威,王怀远.基于样本空间分布特性抽样的电力系统暂态稳定评估[J].电网技术,2025, 49(10):4156-4163.

Xiao Zhiwei,Wang Tongwei,Wang Huaiyuan.Power system transient stability assessment based on sampling of sample spatial distribution characteristics[J]. Power System Technology,2025,49(10):4156-4163.

张宇驰,蔡国伟,刘铖,等.基于网络能量轨迹特征学习的新能源电力系统暂态稳定评估[J].电力系统自动化,2025 ,49(10):20-28.

Zhang Yuchi,Cai Guowei,Liu Cheng,et al.Transient stability assessment of power system with renewable energy based on network energy trajectory feature learning[J].Automation of Electric Power Systems, 2025 ,49(10):20-28.

Guo Tingyan,Milanovic′J V.Online identification of power system dynamic signature using PMU measurements and data mining [J]. IEEE Transactions on Power Systems,2016,31(3):1760-1768.

IEEE Std C37.118.2-2011 IEEE standard for synchrophasor data transfer for power systems [S].

Chan C S,Kong Huanqi,Liang Guanqing.A comparative study of faithfulness metrics for model interpretability methods [C]//Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics. Stroudsburg,PA,USA:ACL,2022:5029-5038.

赵晨浩,焦在滨,李程昊,等.基于双塔Transformer的电力系统暂态稳定评估[J].全球能源互联网, 2024,7(5):521-529.

Zhao Chenhao,Jiao Zaibin,Li Chenghao,et al.Power system transient stability assessment based on twotower Transformer model[J].Journal of Global Energy Interconnection,2024,7(5):521-529.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

一维卷积神经网络在机械故障特征提取中的可解释性研究

Sinc卷积结合变分频带注意力机制的变压器异响智能诊断模型研究

考虑易用性和可解释性的自组织映射-即时学习风险预测框架

采用多传感器融合的空间全方位旋转角度测量方法

多形态狭缝型槽道结构的叶片前缘气膜冷却特性研究