考虑数据缺失的图注意力网络暂态稳定评估

doi:10.11930/j.issn.1004-9649.202307019

摘要/Abstract

摘要：

基于人工智能的暂态稳定评估模型的性能高度依赖于系统的可观测性，而通信延迟和相量测量单元（phasor measurement units，PMU）故障等因素易导致数据缺失，使模型的评估性能下降。针对该问题，提出了一种基于图注意力网络（graph attention network，GAT）的暂态稳定评估模型。首先，根据原始网络拓扑及PMU配置方案获得表征系统可观测性的掩码矩阵，在任意PMU缺失的条件下，利用掩码矩阵训练模型；其次，通过GAT网络的多头注意力机制提取输入节点的时空信息，利用不同的权重聚合目标节点的邻域特征，实现对可观测数据的充分利用；最后，利用焦点损失函数加强模型对失稳样本的学习能力。仿真结果表明，所提方法可以最大限度地利用可观测数据，具有高精度和强鲁棒性，并且不受网络拓扑的限制，易于迁移。

关键词: 暂态稳定评估, 数据缺失, 图注意力网络, 掩码矩阵, PMU故障

Abstract:

The performance of the transient stability assessment model based on artificial intelligence is highly dependent on the observability of the system, while the factors such as communication delays and PMU faults can easily lead to data missing, which degrades the model's assessment performance. To address this problem, this paper proposes a transient stability assessment model based on graph attention network (GAT). Firstly, a mask matrix representing the system observability is obtained based on the original network topology and PMU configuration scheme, and the mask matrix is used to train the model under the condition of any PMU missing. Secondly, the spatio-temporal information of the input node is extracted through the multi-head attention mechanism of the GAT network, and different weights are used to aggregate the neighborhood characteristics of the target node to make full use of observable data. Finally, the focus loss function is used to enhance the model's learning ability for unstable samples. The simulation results show that the proposed method can maximize the use of observable data with high precision and strong robustness, and is not limited by the network topology and easy to migrate.

Key words: transient stability assessment, data missing, graph attention network, mask matrix, PMU fault

周生存, 罗毅, 易煊承, 吴亚宁, 李丁, 熊逸. 考虑数据缺失的图注意力网络暂态稳定评估[J]. 中国电力, 2024, 57(5): 157-167.

Shengcun ZHOU, Yi LUO, Xuancheng YI, Yaning WU, Ding LI, Yi XIONG. Transient Stability Assessment of Graph Attention Networks Considering Data Missing[J]. Electric Power, 2024, 57(5): 157-167.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://www.electricpower.com.cn/CN/10.11930/j.issn.1004-9649.202307019

https://www.electricpower.com.cn/CN/Y2024/V57/I5/157

图/表 12

参考文献 25

1	汪泽州, 张明明, 钱峰强, 等. 含光伏接入的中压配电网集中调控优化策略[J]. 中国电力, 2023, 56 (2): 15- 22. DOI
	WANG Zezhou, ZHANG Mingming, QIAN Fengqiang, et al. Centralized regulation and optimization strategy for MV distribution network with PV integration[J]. Electric Power, 2023, 56 (2): 15- 22. DOI
2	马兆兴, 秦昌民, 朱文杰, 等. 考虑多时间尺度特征的电力系统稳定性界分析[J]. 电力系统保护与控制, 2021, 41 (2): 461- 475. DOI
	MA Zhaoxing, QIN Changmin, ZHU Wenjie, et al. Analysis of the stability bound of a power system considering multi-time scale characteristics[J]. Power System Protection and Control, 2021, 41 (2): 461- 475. DOI
3	康重庆, 杜尔顺, 郭鸿业, 等. 新型电力系统的六要素分析[J]. 电网技术, 2023, 47 (5): 1741- 1750.
	KANG Chongqing, DU Ershun, GUO Hongye, et al. Primary exploration of six essential factors in new power system[J]. Power System Technology, 2023, 47 (5): 1741- 1750.
4	蔡欢, 袁旭峰, 熊炜, 等. 柔性互联配电网运行调度研究综述[J]. 智慧电力, 2022, 50 (6): 92- 99, 106.
	CAI Huan, YUAN Xufeng, XIONG Wei, et al. Review on operation scheduling of flexible interconnected distribution network[J]. Smart Power, 2022, 50 (6): 92- 99, 106.
5	GURRALA G, DINESHA D L, DIMITROVSKI A, et al. Large multi-machine power system simulations using multi-stage adomian decomposition[J]. IEEE Transactions on Power Systems, 2017, 32 (5): 3594- 3606. DOI
6	VU T L, TURITSYN K. Lyapunov functions family approach to transient stability assessment[J]. IEEE Transactions on Power Systems, 2016, 31 (2): 1269- 1277. DOI
7	单瑞卿, 盛阳, 苏盛, 等. 考虑攻击方身份的电力监控系统网络安全风险分析[J]. 电力科学与技术学报, 2022, 37 (5): 3- 16. DOI
	SHAN Ruiqing, SHENG Yang, SU Sheng, et al. Risk analysis of power system cyber security considering identity of malicious adversaries[J]. Journal of Electric Power Science and Technology, 2022, 37 (5): 3- 16. DOI
8	汤涌, 姚伟, 王宏志, 等. 电网仿真分析与决策的人工智能方法[J]. 中国电机工程学报, 2022, 42 (15): 5384- 5406. DOI
	TANG Yong, YAO Wei, WANG Hongzhi, et al. Artificial intelligence techniques for power grid simulation analysis and decision making[J]. Proceedings of the CSEE, 2022, 42 (15): 5384- 5406. DOI
9	周艳真, 查显煜, 兰健, 等. 基于数据增强和深度残差网络的电力系统暂态稳定预测[J]. 中国电力, 2020, 53 (1): 22- 31.
	ZHOU Yanzhen, ZHA Xianyu, LAN Jian, et al. Transient stability prediction of power systems based on deep residual network and data augmentation[J]. Electric Power, 2020, 53 (1): 22- 31.
10	ZHANG Y C, XU Y, ZHANG R, et al. A missing-data tolerant method for data-driven short-term voltage stability assessment of power systems[J]. IEEE Transactions on Smart Grid, 2019, 10 (5): 5663- 5674. DOI
11	HE M, VITTAL V, ZHANG J S. Online dynamic security assessment with missing pmu measurements: a data mining approach[J]. IEEE Transactions on Power Systems, 2013, 28 (2): 1969- 1977. DOI
12	ZHANG Y C, XU Y, DONG Z Y. Robust classification model for PMU-based on-line power system DSA with missing data[J]. IET Generation, Transmission & Distribution, 2017, 11 (18): 4484- 4491.
13	ZHANG Y C, XU Y, DONG Z Y. Robust ensemble data analytics for incomplete PMU measurements-based power system stability assessment[J]. IEEE Transactions on Power Systems, 2018, 33 (1): 1124- 1126. DOI
14	谭本东, 杨军, 刘源, 等. 考虑数据缺失的电力系统暂态稳定自适应集成评估方法[J]. 电力系统自动化, 2021, 45 (23): 68- 75. DOI
	TAN Bendong, YANG Jun, LIU Yuan, et al. Adaptive integrated assessment method for transient stability of power system considering PMU data missing[J]. Automation of Electric Power Systems, 2021, 45 (23): 68- 75. DOI
15	REN C, XU Y, DAI B J, et al. An integrated transfer learning method for power system dynamic security assessment of unlearned faults with missing data[J]. IEEE Transactions on Power Systems, 2021, 36 (5): 4856- 4859. DOI
16	YU J J Q, LAM A Y S, HILL D J, et al. Delay aware power system synchrophasor recovery and prediction framework[J]. IEEE Transactions on Smart Grid, 2019, 10 (4): 3732- 3742. DOI
17	GAO P Z, WANG M, GHIOCEL S G, et al. Missing data recovery by exploiting low-dimensionality in power system synchrophasor measurements[J]. IEEE Transactions on Power Systems, 2016, 31 (2): 1006- 1013. DOI
18	REN C, XU Y. A fully data-driven method based on generative adversarial networks for power system dynamic security assessment with missing data[J]. IEEE Transactions on Power Systems, 2019, 34 (6): 5044- 5052. DOI
19	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.
20	曹鹏, 刘敏. 基于改进的整数规划法结合零注入节点的PMU优化配置方法[J]. 电力系统保护与控制, 2021, 49 (16): 143- 150. DOI
	CAO Peng, LIU Min. PMU placement method based on improved integer programming method combined with zero injection buses[J]. Power System Protection and Control, 2021, 49 (16): 143- 150. DOI
21	李佳玮, 王小君, 和敬涵, 等. 基于图注意力网络的配电网故障定位方法[J]. 电网技术, 2021, 45 (6): 2113- 2121.
	LI Jiawei, WANG Xiaojun, HE Jinghan, et al. Distribution network fault location based on graph attention network[J]. Power System Technology, 2021, 45 (6): 2113- 2121.
22	朱思婷, 管霖, 黄济宇, 等. 面向稳控策略校核的并联图注意力网络稳定评估模型[J]. 电网技术, 2023, 47 (9): 3836- 3846.
	ZHU Siting, GUAN Lin, HUANG Jiyu, et al. Parallel graph attention network based transient stability assessment model for stability control strategy checking[J]. Power System Technology, 2023, 47 (9): 3836- 3846.
23	李楠, 朱嫄, 崔莹. 考虑代价敏感的AC-LSTM暂态稳定评估[J]. 电力系统保护与控制, 2022, 50 (22): 160- 169.
	LI Nan, ZHU Yuan, CUI Ying. AC-LSTM transient stability assessment considering cost-sensitivity[J]. Power System Protection and Control, 2022, 50 (22): 160- 169.
24	李保罗, 孙华东, 张恒旭, 等. 基于两阶段迁移学习的电力系统暂态稳定评估框架[J]. 电力系统自动化, 2022, 46 (17): 176- 185.
	LI Baoluo, SUN Huadong, ZHANG Hengxu, et al. Transient stability assessment framework of power system based on two-stage transfer learning[J]. Automation of Electric Power Systems, 2022, 46 (17): 176- 185.
25	田书欣, 李昆鹏, 魏书荣, 等. 基于同步相量测量装置的配电网安全态势感知方法[J]. 中国电机工程学报, 2021, 41 (2): 617- 632. DOI
	TIAN Shuxin, LI Kunpeng, WEI Shurong, et al. Security situation awareness approach for distribution network based on synchronous phasor measurement unit[J]. Proceedings of the CSEE, 2021, 41 (2): 617- 632. DOI

编号	安装PMU的母线位置	总数
1	3, 8, 10, 16, 20, 23, 25, 29	8
2	3, 8, 12, 16, 20, 23, 25, 29	8
3	2, 6, 9, 10, 13, 14, 17, 19, 20, 22, 23, 25, 29	13
4	2, 6, 9, 10, 11, 14, 17, 19, 20, 22, 23, 25, 29	13

编号	安装PMU的母线位置	总数
1	3, 8, 10, 16, 20, 23, 25, 29	8
2	3, 8, 12, 16, 20, 23, 25, 29	8
3	2, 6, 9, 10, 13, 14, 17, 19, 20, 22, 23, 25, 29	13
4	2, 6, 9, 10, 11, 14, 17, 19, 20, 22, 23, 25, 29	13

模型	布局方案误差率/%
模型	方案1	方案2	方案3	方案4
GAT	0.37	1.06	0.25	0.41
RVFL	0.42	1.22	0.49	0.50

模型	布局方案误差率/%
模型	方案1	方案2	方案3	方案4
GAT	0.37	1.06	0.25	0.41
RVFL	0.42	1.22	0.49	0.50

PMU缺失数量	GAT模型			RVFL模型
PMU缺失数量	R	P	${G_{\rm mean}}$	R	P	${G_{\rm mean}}$
0	98.90	99.23	99.06	98.73	98.97	98.85
1	98.58	98.95	98.76	98.60	98.37	98.48
2	98.46	98.23	98.34	98.40	97.85	98.12
3	98.18	98.00	98.09	97.98	97.57	97.79
4	97.87	97.72	97.80	97.53	97.25	97.39