融合关键节点信息的数据驱动暂态电压稳定性评估模型优化方法

doi:10.19595/j.cnki.1000-6753.tces.251286

摘要
图/表
参考文献
相关文章 (7)

全文: PDF (1055 KB)
输出: BibTeX | EndNote (RIS)

摘要

深度学习模型为电力系统暂态电压稳定评估（transient voltage stability assessment,TVSA）问题提供了新的解决方案,但其评估结果可解释性与决策过程可控性面临严峻挑战。为应对这一挑战,本文提出了一种融合关键节点信息的TVSA模型优化方法。首先本文构建了基于 Transformer 的 TVSA 模型,通过分析模型对特征量的注意力权重分布,分析模型的评估规则,为模型决策逻辑优化提供理论基础。然后,利用故障延迟电压恢复指数标记各训练集样本的关键节点,实现失稳模式关键节点的捕捉。在此基础上,为了实现模型评估规则的可控性,从模型损失函数角度出发,通过构造注意力引导项,引导模型增强对样本中关键节点的注意力权重,强化对失稳模式关键节点信息的挖掘。同时,为防止引导优化对其余样本的评估规则的影响,在损失函数中构造注意力保持项,降低指定失稳模式的引导过程对其余失稳模式评估规则的干扰。在IEEE 39节点系统和西北某省电网系统上的仿真结果均验证了该方法的有效性。

	服务

	把本文推荐给朋友
	加入我的书架
	加入引用管理器
	E-mail Alert
	RSS
	作者相关文章
	缪业丰
	黄栋
	王怀远

关键词 ： Transformer, 暂态电压稳定性评估, 可解释性, 注意机制, 模型优化

Abstract：

A new approach for power system transient voltage stability assessment (TVSA) is offered by deep learning models, yet challenges persist in two key aspects: the interpretability of assessment results (e.g., how critical node characteristics are identified by the model) and the controllability of the decision-making process (e.g., how assessment logic for specific instability modes is adjusted). The mechanism by which the model focuses on critical nodes struggles to be revealed by traditional methods; moreover, when assessment rules for specific instability modes are optimized, the assessment logic of other modes tends to be disrupted by them, resulting in insufficient model reliability in complex power grid scenarios.
To improve the interpretability and controllability of TVSA models, an optimized training method integrating critical node information is proposed in this paper, which consists of three core steps: First, a TVSA model based on a Transformer encoder is constructed. Spatiotemporal features of bus node voltage phase angle sequences are learned by this model via a self-attention mechanism, and the model’s critical node assessment rules are visualized using attention weights. Second, the fault-induced delayed voltage recovery index (FDVRI) is introduced. Based on the dynamic recovery characteristics of node voltages within 0-4 cycles after fault clearance, critical node sets for each training sample are labeled to clarify the core feature carrier of the instability mode. Third, a loss function incorporating an attention guidance term and an attention retention term is designed. Critical nodes of the target instability mode are forced to be focused on by the guidance term through weight constraints, thereby correcting the model’s assessment rules; the stability of attention distribution for non-target modes is maintained by the retention term, avoiding interference of the optimization process on global assessment logic.
Validations are conducted on the IEEE 39-node system and the Northwest China Power Grid. Results show that the model’s attention mechanism toward critical nodes is clearly revealed by attention visualization, and the influence of critical nodes on assessment decisions is directly reflected by their attention weights. Dominant instability features are accurately mined by the model during training through the loss function integrating critical node information: while the assessment ability for specific instability modes is optimized, the physical consistency of the model’s overall assessment logic is ensured. After a specific instability mode is optimized using the proposed method, the accuracy of samples for this mode is increased by 8.72%, with no impact on the assessment of other instability samples.
The interpretability and controllability of TVSA models are enhanced by this optimized method through the introduction of critical node labeling and a dual-constraint loss function, providing an intuitive basis for operators to understand assessment logic. It is confirmed by validations on both standard systems and actual power grids that instability mode features are effectively captured by the method, offering an interpretable and controllable new path for power system TVSA.

Key words： Transformer TVSA interpretability attention mechanism model update

收稿日期: 2025-07-21

PACS:

TM712

基金资助:

福建省自然科学基金资助项目（2022J01113）

通讯作者: 王怀远男,1987年生,副教授,硕士生导师,研究方向为电力系统运行与控制,电力系统暂态稳定分析与控制,电力系统频率稳定性分析与控制等。E-mail：79749544@qq.com

作者简介: 缪业丰男,1999年生,硕士研究生,研究方向为人工智能技术在电力系统中的应用。E-mail：915292584@qq.com

引用本文:

缪业丰, 黄栋, 王怀远. 融合关键节点信息的数据驱动暂态电压稳定性评估模型优化方法[J]. 电工技术学报, 0, (): 20251286-20251286. Miao Yefeng, Huang Dong, Wang Huaiyuan. An Optimization Method for the Data-Driven Transient Voltage Stability Assessment Model Integrating Critical Node Information. Transactions of China Electrotechnical Society, 0, (): 20251286-20251286.

链接本文:

https://dgjsxb.ces-transaction.com/CN/10.19595/j.cnki.1000-6753.tces.251286 https://dgjsxb.ces-transaction.com/CN/Y0/V/I/20251286

[1] Liu Y, Sun K, Yao R, et al.Power System Time Domain Simulation Using a Differential Transformation Method[J]. IEEE Transactions on Power Systems,2019, 34(5):3739-3748.
[2] 王长江,姜涛,陈厚合,等.基于相位校正李雅普诺夫指数的电力系统暂态电压稳定评估[J].电工技术学报,2021,36(15):3221-3236.
Wang Changjiang, Jiang Tao, Chen Houhe, et al.Transient voltage stability assessment of power systems based on phase correction maximum Lyapunov exponent[J].Transactions of China Electrotechnical Society,2021,36(15):3221-3236.
[3] 王长江,尹浩帆,姜涛,等.计及切换补偿的新能源电力系统暂态电压稳定评估方法 [J/OL].电工技术学报,2025,1-14 [2025-05-21].
Wang Changjiang, Yin Haofan, Jiang Tao, et al.Transient voltage stability assessment approach for renewable energy power system considering switching compensation[J/OL]. Transactions of China Electrotechnical Society,2025,1-14[2025-05-21].
[4] 陈厚合,张赫,王长江,等.基于卷积神经网络的直流送端系统暂态过电压估算方法[J].电网技术,2020,44(8):2987-2999.
Chen Houhe, Zhang He, Wang Changjiang, et al.A method estimating transient overvoltage of HVDC sending-end system based on convolutional neural network[J] .Power System Technology, 2020,44(8):2987-2999.
[5] Yang H, Zhang W, Chen J, et al.PMU-based voltage stability prediction using least square support vector machine with online learning[J].Electrical Power System Research,2018,160:234-242.
[6] 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(02):778-791.
Zhen Yongzan, Ruan Cheng.Hierarchical self-adaptive transient voltage stability assessment based on cost-sensitive SVM and multivariate decision tree[J].Power System Technology,2024,48(02):778-791.
[7] 朱利鹏,陆超,孙元章,等.基于数据挖掘的区域暂态电压稳定评估[J].电网技术,2015,39(4):1026-1032.
Zhu Lipeng, Lu Chao, Sun Yuanzhang, et al.Data mining based regional transient voltage stability assessment[J].Power System Technology,2015,39(4):1026-1032.
[8] Zhu L, Lu C, Sun Y.Time Series Shapelet Classification Based Online Short-Term Voltage Stability Assessment[J].IEEE Transactions on Power Systems,2016,31(2):1430-1439.
[9] 季佳伸,吴俊勇,王彦博,等.基于深度残差网络的电力系统暂态电压稳定评估[J].电网技术,2022,46(07):2500-2511.
Ji Jiashen, Wu Junyong, Wang Yanbo, et al.Power system transient voltage stability assessment based on deep residual network[J].Power System Technology,2022,46(07):2500-2511.
[10] 卢锦玲,郭鲁豫.基于改进深度残差收缩网络的电力系统暂态稳定评估[J].电工技术学报,2021,36(11):2233-2244.
Lu Jinling,Guo Luyu.Power system transient stability assessment based on improved deep residual shrinkage network[J].Transactions of China Electrotechnical Society,2021,36(11):2233-2244.
[11] Wang Y, Lu C, Zhang X.Convolution Neural Network-Based Load Model Parameter Selection Considering Short-Term Voltage Stability[J].CSEE Journal of Power and Energy Systems,2024,10(3):1064-1074.
[12] 朱林,张健,陈达,等.面向暂态电压稳定评估的卷积神经网络输入特征构建方法[J].电力系统自动化,2022,46(01):85-93.
Zhu Lin,Zhang Jian,Chen Da,et al.Construction method for input features of convolutional neural network for transient voltage stability assessment[J].Automation of Electric Power Systems,2022,46(01):85-93.
[13] Seyedi Y, Karimi H, Mahseredjian J.A Data-Driven Method for Prediction of Post-Fault Voltage Stability in Hybrid AC/DC Microgrids[J].IEEE Transactions on Power Systems,2022,37(5):3758-3768.
[14] 姜涛,董雨,王长江,等.基于图卷积和双向长短期记忆网络的受端电力系统暂态电压稳定评估[J].电网技术,2023,47(12):4937-4951.
Jiang Tao, Dong Yu, Wang Changjiang, et al.Transient voltage stability assessment for receiving-end power system based on graph convolution and bidirectional long short-term memory network[J].Power System Technology,2023,47(12):4937-4951.
[15] 百度. AI-人机交互趋势研究[R].北京:百度人工智能交互设计院(AIID),2019.
Baidu.AI-human-computer interaction trend research [R].Beijing:Baidu Artificial Intelligence Interaction Design Institute(AIID),2019.
[16] 孙宏斌,黄天恩,郭庆来,等.面向调度决策的智能机器调度员研制与应用[J].电网技术,2020,44(01):1-8.
Sun Hongbin, Huang Tianen, Guo Qinglai, et al.Automatic operator for decision-making in dispatch: research and applications[J].Power System Technology,2020,44(01):1-8.
[17] 蒲天骄,乔骥,赵紫璇,等.面向电力系统智能分析的机器学习可解释性方法研究 (一):基本概念与框架[J].中国电机工程学报,2023,43(18):7010-7030.
Pu Tianjiao, Qiao Ji, Zhao Zixuan, et al.Research on interpretability methods of machine learning applied in intelligent analysis of power system (Part I): Basic concepts and framework[J].Proceedings of the CSEE,2023,43(18):7010-7030.
[18] 胡润滋,马晓忱,孙博,等.基于特征选择的暂态安全评估方法及其可解释性研究[J].电网技术,2023,47(02):755-763.
Hu Runzi, Ma Xiaochen, Sun Bo, et al.Transient safety assessment and its interpretability based on feature selection[J].Power System Technology,2023,47(02):755-763.
[19] 甄永赞,阮程.基于强化学习的混合元启发式暂态电压稳定特征选择方法及可解释性研究[J].电网技术,2024,48(04):1519-1532.
Zhen Yongzan, Ruan Cheng.A hybrid element heuristic transient voltage stability feature selection method based on reinforcement learning and interpretability research[J].Power System Technology,2024,48(04):1519-1532.
[20] 周挺,杨军,詹祥澎,等.一种数据驱动的暂态电压稳定评估方法及其可解释性研究[J].电网技术,2021,45(11):4416-4425.
Zhou Ting, Yang Jun, Zhan Xiangpeng, et al.Data-driven method and interpretability analysis for transient voltage stability assessment[J].Power System Technology,2021,45(11):4416-4425.
[21] 朱利鹏,陆超,黄河,等.基于广域时序数据挖掘策略的暂态电压稳定评估[J].电网技术,2016,40(1):180-185.
Zhu Lipeng, Lu Chao, Huang He, et al.Wide-area time series data mining based transient voltage stability assessment[J].Power System Technology,2016,40(1):180-185.
[22] 朱利鹏,陆超,黄河,等.基于时序轨迹特征学习的暂态电压稳定评估[J].电网技术,2019,43(06):1922-1931.
Zhu Lipeng, Lu Chao, Huang He, et al.Transient voltage stability assessment based on sequential trajectory feature learning[J].Power System Technology,2019,43(06):1922-1931.
[23] 房佳姝,刘崇茹,苏晨博,等.基于自注意力 Transformer 编码器的多阶段电力系统暂态稳定评估方法[J].中国电机工程学报,2023,43(15):5745-5759.
Fang Jiashu, Liu Chongru, Su Chenbo, et al.Multi-stage transient stability assessment of power system based on self-attention transformer encoder[J].Proceedings of the CSEE,2023,43(15):5745-5759.
[24] Q. Chen, N. Lin, S. Bu, H. Wang, B. Zhang.Interpretable Time-Adaptive Transient Stability Assessment Based on Dual-Stage Attention Mechanism[J]. IEEE Transactions on Power Systems, 2023, 38(3): 2776-2790.
[25] 高发骏,王怀远,党然.基于Transformer 的暂态稳定评估模型的可解释性分析与模型优化研究[J].电力系统保护与控制,2023,51(17):15-25.
Gao Fajun, Wang Huaiyuan, Dang Ran.Interpretability analysis and model update research of a transient stability assessment model based on Transformer[J].Power System Protection and Control,2023,51(17):15-25.
[26] 任顺鑫,王怀远,李剑,等.主导模式引导的电力系统暂态稳定数据驱动评估方法[J].中国电机工程学报,2025,45(14):5589-5601.
Ren Shunxin, Wang Huaiyuan, Li Jian, et al.Data-driven method for transient stability assessment of power system guided by dominant pattern[J].Proceedings of the CSEE, ,2025,45(14):5589-5601.
[27] Li Q, Xu Y, Ren C.A hierarchical data-driven method for event-based load shedding against fault-induced delayed voltage recovery in power systems[J].IEEE Transactions on Industrial Informatics,2021,17(1):699-709.
[28] Dong Y, Xie X, Zhou B, et al.An Integrated High Side Var-Voltage Control Strategy to Improve Short-Term Voltage Stability of Receiving-End Power Systems[J/OL].IEEE Transactions on Power Systems,2016,31(3):2105-2115.DOI:10.1109/TPWRS.2015.2464695.
[29] 杨金洲,李业成,熊鸿韬,等.基于阻抗的新能源接入的受端电网暂态电压失稳高风险故障快速筛选方法 [J/OL].电工技术学报,2025,1-13 [2025-05-21].
Yang Jinzhou, Li Yecheng, Xiong Hongtao, et al.A fast screening method for the high-risk faults with transient voltage instability in receiving-end power grids interconnected with new energy based on impedance[J/OL].Transactions of China Electrotechnical Society,2025,1-13[2025-05-21].
[30] 杨金洲,李业成,熊鸿韬,等.新能源接入的受端电网暂态电压失稳高风险故障快速筛选[J].电工技术学报,2024,39(21):6746-6758.
Yang Jinzhou, Li Yecheng, Xiong Hongtao, et al.A fast screening method for the high-risk faults with transient voltage instability in receiving-end power grids interconnected with new energy[J].Transactions of China Electrotechnical Society,2024,39(21):6746-6758.
[31] 全国电网运行与控制标准化技术委员会(SAC/TC 446).电力系统安全稳定计算规范:GB/T 40581-2021[S].中国标准出版社,2021.
National Standardization Technical Committee for Power Grid Operation and Control (SAC/TC 446).Specifications for Power System Security and Stability Calculation:GB/T 40581-2021[S]. China Standards Press,2021.
[32] Zhu L, Hill D J, Lu C.Intelligent short-term voltage stability assessment via spatial attention rectified RNN learning[J].IEEE Transactions on Industrial Informatics,2021,17(10):7005-7016.
[33] Wang H, Wang Q.Adaptive cost-sensitive assignment method for power system transient stability assessment[J].International Journal of Electrical Power & Energy Systems,2022,135:107574.