基于多粒度NRS和改进Bi-LSTM的电力系统暂态稳定评估*

doi:10.11985/2023.03.025

摘要/Abstract

摘要：

当前基于数据驱动的深度学习算法在电力系统暂态分析领域被广泛应用，但是实际中电力系统测得的数据样本中暂态失稳的情况较少并且对失稳样本的重视不足，导致分类模型的抗干扰性、泛化能力较弱，影响整个模型的评估性能。针对此问题，提出一种基于多粒度邻域粗糙集和改进双向长短期记忆网络(IM-Bi-LSTM)的电力系统暂态稳定评估方法。首先采用邻域粗糙集在不同粒度级别下寻找最优的约简子集，再利用Bi-LSTM神经网络完成对特征子集进行时序信息的提取，并且在模型中引入注意力机制，对与失稳样本相关的特征增加更多的权重；通过焦点损失函数，引入权重系数调整模型训练的倾向性，解决失稳样本与暂态稳定样本间的不平衡问题，提高模型的评估性能。在IEEE10机39节点系统上的试验结果表明，相较于其他算法，所提方法的分类精度更好、结果更稳定。

关键词: 电力系统, 邻域粗糙集, 双向长短期记忆网络, 注意力机制, 焦点损失, 暂态稳定评估

Abstract:

At present, data-driven deep learning algorithm is widely used in the field of power system transient analysis, but in practice, there are few transient instability in the measured data samples of power system, and insufficient attention is paid to the instability samples, resulting in the weak anti-interference and generalization ability of the classification model, which affects the evaluation performance of the whole model. To solve this problem, a power system transient stability assessment method based on multi-granularity neighborhood rough set and improved bi-directional long-short-term memory network(IM-Bi-LSTM) is proposed. Firstly, the neighborhood rough set is used to find the optimal reduction subset at different granularity levels, and then the BiLSTM neural network is used to extract the timing information of the feature subset, and the attention model is introduced into the model to add more weights to the features related to the unstable samples. Through the focus loss function, the weight coefficient is introduced to adjust the tendency of model training, solve the imbalance between unstable samples and transient stability samples, and improve the evaluation performance of the model. The experimental results on the IEEE 10 machine 39 bus system show that compared with other algorithms, the proposed method has better classification accuracy and more stable results.

Key words: Power system, neighborhood rough sets, bi-directional long-short-term memory network, attention model, focus loss, transient stability assessment

中图分类号:

TM712

李志兵, 肖健梅, 王锡淮. 基于多粒度NRS和改进Bi-LSTM的电力系统暂态稳定评估^*[J]. 电气工程学报, 2023, 18(3): 232-241.

LI Zhibing, XIAO Jianmei, WANG Xihuai. Transient Stability Assessment of Power System Based on Multi-granularity Neighborhood Rough Set and Improved Bi-directional Long-short-term Memory Network[J]. Journal of Electrical Engineering, 2023, 18(3): 232-241.

图/表 12

表1

图1

图2

图3

表2

表3

表4

表5

图4

表6

图5

表7

参考文献 22

[1]	曾波, 吴昆霖. 大规模特高压交直流混联电网特性分析与运行控制[J]. 南方农机, 2018, 49(17):214.
	ZENG Bo, WU Kunlin. Characteristic analysis and operation control of large scale high voltage AC/DC hybrid power grid[J]. China Southern Agricultural Machinery, 2018, 49(17):214.
[2]	赵晋泉, 邓晖, 吴小辰, 等. 基于广域响应的电力系统暂态稳定控制技术评述[J]. 电力系统保护与控制, 2016, 44(5):1-9.
	ZHAO Jinquan, DENG Hui, WU Xiaochen, et al. Review on power system transient stability control technologies based on PMU/WAMS[J]. Power System Protection and Control, 2016, 44(5):1-9.
[3]	罗深增, 李银红, 石东源. 广域测量系统可观性概率评估及其在PMU优化配置中的应用[J]. 电工技术学报, 2018, 33(8):1844-1853.
	LUO Shenzeng, LI Yinhong, SHI Dongyuan. Wide area monitoring system observability probabilistic evaluation and it’s application in optimal PMU placement[J]. Transactions of China Electrotechnical Society, 2018, 33(8):1844-1853.
[4]	石访, 张林林, 胡熊伟, 等. 基于多属性决策树的电网暂态稳定规则提取方法[J]. 电工技术学报, 2019, 34(11):2364-2374.
	SHI Fang, ZHANG Linlin, HU Xiongwei, et al. Power system transient stability rules extraction based on multi-attribute decision tree[J]. Transactions of China Electrotechnical Society, 2019, 34(11):2364-2374.
[5]	RAHMATIAN M, CHEN Yu, PALIZBAN A, et al. Transient stability assessment via decision trees and multivariate adaptive regression splines[J]. Electric Power Systems Research, 2017, 142:320-328. doi: 10.1016/j.epsr.2016.09.030
[6]	SUN Zhifang, LI Yi. Optimization of decision tree machine learning strategy in data analysis[J]. Journal of Physics:Conference Series, 2020, 1693(1):012219. doi: 10.1088/1742-6596/1693/1/012219
[7]	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 multi-support vector machines[J]. Proceedings of the CSEE, 2016, 36(5):1173-1180.
[8]	姜涛, 王长江, 陈厚合, 等. 基于正则化投影孪生支持向量机的电力系统暂态稳定评估[J]. 电力系统自动化, 2019, 43(1):141-148.
	JIANG Tao, WANG Changjiang, CHEN Houhe, et al. Transient stability assessment of power system based on projection twin support vector machine with regularization[J]. Automation of Electric Power Systems, 2019, 43(1):141-148.
[9]	於万里, 王艳, 纪志成. 氨糖发酵过程建模与工艺参数优化研究[J]. 系统仿真学报, 2020, 32(10):1895-1902. doi: 10.16182/j.issn1004731x.joss.20-FZ0400
	YU Wanli, WANG Yan, JI Wangcheng. Study on modeling and optimization of process parameters for ammonia fermentation[J]. Journal of System Simulation, 2020, 32(10):1895-1902. doi: 10.16182/j.issn1004731x.joss.20-FZ0400
[10]	GOMEZ F R, RAJAPAKSE A D, ANNAKKAGE U D, et al. Support vector machine-based algorithm for post-fault transient stability status prediction using synchronized measurements[J]. IEEE Transactions on Power Systems, 2011, 26:1474-1483. doi: 10.1109/TPWRS.2010.2082575
[11]	HU W, LU Z, WU S, et al. Real-time transient stability assessment in power system based on improved SVM[J]. Journal of Modern Power Systems and Clean Energy, 2019, 7(1):26-37. doi: 10.1007/s40565-018-0453-x
[12]	张林林, 胡熊伟, 李鹏, 等. 基于极限学习机的电力系统暂态稳定评估方法[J]. 上海交通大学学报, 2019, 53(6):749-756.
	ZHANG Linlin, HU Xiongwei, LI Peng, et al. Power system transient stability assessment method based on extreme learning machine[J]. Journal of Shanghai Jiao Tong University, 2019, 53(6):749-756.
[13]	高昆仑, 杨帅, 刘思言, 等. 基于一维卷积神经网络的电力系统暂态稳定评估[J]. 电力系统自动化, 2019, 43(12):18-26.
	GAO Kunlun, YANG Shuai, LIU Siyan, et al. Transient stability assessment for power system based on one-dimensional convolutional neural network[J]. Automation of Electric Power Systems, 2019, 43(12):18-26.
[14]	邵美阳. 基于深度置信网络的电力系统暂态稳定评估[D]. 北京: 北京交通大学, 2020.
	SHAO Meiyang. Transient stability assessment of power system based on deep belief network[D]. Beijing: Beijing Jiaotong University, 2020.
[15]	SHUANG Wu, LE Zheng, WEI Hu, et al. Improved deep belief network and model interpretation method for power system transient stability assessment[J]. Journal of Modern Power Systems and Clean Energy, 2020, 8(1):27-37. doi: 10.35833/MPCE.2019.000058
[16]	朱乔木, 陈金富, 李弘毅, 等. 基于堆叠自动编码器的电力系统暂态稳定评估[J]. 中国电机工程学报, 2018, 38(10):2937-2946.
	ZHU Qiaomu, CHEN Jinfu, LI Hongyi, et al. Transient stability assessment based on stacked autoencoders[J]. Proceedings of the CSEE, 2018, 38(10):2937-2946.
[17]	LI X, YANG Z, GUO P, et al. An intelligent transient stability assessment framework with continual learning ability[J]. IEEE Transactions on Industrial Informatics, 2021, 17(12):8131-8141. doi: 10.1109/TII.2021.3064052
[18]	胡清华, 于达仁, 谢宗霞. 基于邻域粒化和粗糙逼近的数值属性约简[J]. 软件学报, 2008(3):640-649.
	HU Qinghua, YU Daren, XIE Zongxia. Numerical attribute reduction based on neighborhood granulation and rough approximation[J]. Journal of Software, 2008(3):640-649.
[19]	LIN T Y, GOYAL P, GIRSHICK R, et al. Focal loss for dense object detection[J]. IEEE Transactions on Pattern Analysis & Machine Intelligence, 2017(99):2999-3007.
[20]	张文修. 粗糙集理论与方法[M]. 北京: 科学出版社, 2001.
	ZHANG Wenxiu. Rough set theory and method[M]. Beijing: Science Press, 2001.
[21]	MI Jusheng, WU Weizhi, ZHANG Wenxiu. Approaches to knowledge reduction based on variable precision rough set model[J]. Information Sciences, 2003, 159(3):255-272. doi: 10.1016/j.ins.2003.07.004
[22]	ZHAO Suyun, TSANG E C C, CHEN Degang. The model of fuzzy variable precision rough sets[J]. IEEE Transactions on Fuzzy Systems, 2009, 17(2):451-467. doi: 10.1109/TFUZZ.2009.2013204

特征	特征集描述
1	发电机机械功率平均值
2	发电机加速功率的平均值
3	发电机加速度方差
4	发电机所受冲击最大值
5	发电机所受冲击最小值
6	转子加速度平均值
7	转子相对加速度最大值
8	转子角度最大值与最小值之差
9	转子角速度最大值与最小值之差
10	转子动能最大值与最小值之差
11	转子动能最大值
12	转子动能平均值
13	发电机所受冲击最大值
14	发电机所受冲击最小值
15	转子角度最大值与最小值之差
16	转子角速度最大值与最小值之差
17	转子动能最大值与最小值之差
18	转子动能最大值
19	转子动能平均值
20	有功功率总和
21	功角极值
22	发电机加速功率方差
23	系统的总能量变化

评估指标	真实标签
评估指标	稳定	失稳
稳定	TP	FP
失稳	FN	TN

指标	表达式
ACC	$\frac{T P+T N}{T P+T N+F N+F P}$
MA	$\frac{F N}{T P+F N} $
FA	$\frac{F P}{F P+T N} $
G-mean	$\sqrt{(1-M A)(1-F A)}$

粒度级别	约简特征子集
R1	23,15,21,6,22,4,20,13,5,3,14,19,2
R2	23,15,21,22,10,4,20,13,5,3
R3	23,15,17,5,21,22,20,3,6,13,4,14,2,7,19
R4	18,23,15,5,21,10,22,3,14,20,4,13,2

粒度级别	ACC(%)	MA(%)	FA(%)	G-mean(%)
R1	98.30	1.31	3.14	97.77
R2	96.28	2.36	5.35	96.13
R3	95.14	4.20	5.66	95.07
R4	93.99	1.05	11.95	93.34
原始数据集	95.57	5.51	3.15	95.66