基于混合门控循环单元子层的多任务暂态稳定评估

doi:10.12204/j.issn.1000-7229.2022.02.008

摘要/Abstract

摘要：

在暂态功角稳定评估和暂态电压稳定评估的相关研究中,通常分别构建独立的评估模型,这阻碍了不同任务间的信息共享,浪费了计算和存储资源。考虑不同评估任务间往往存在相似性和差异性,为更好地实现二者同时评估,文章提出了一种基于混合门控循环单元(gated recurrent unit, GRU)子层的多任务暂态稳定评估模型。由于电力系统暂态过程具有明显的时序特性,模型采用GRU子层高效地提取量测数据中的时序特征;并在模型结构中引入门控机制以自动调节各个子层在构建不同评估任务特征表示时的占比,不仅促进了不同任务间的信息共享,还削弱了不同任务间差异性对模型训练的负面影响。在IEEE 39节点测试系统中的仿真实验结果表明,文章提出的模型具更好的评估性能和计算速度。

关键词: 暂态稳定评估, 深度学习, 多任务学习, 门控循环单元(GRU)

Abstract:

In the related research of transient power-angle stability assessment and transient voltage stability assessment, independent assessment models are usually constructed, respectively, which hinders the information sharing among different tasks and wastes computing and storage resources. Considering the similarities and differences among different assessment tasks, in order to better realize the two assessments at the same time, a multi-task transient stability assessment model based on sub-layer of hybrid gated recurrent unit (GRU) is proposed in this paper. Because the power system transient process has obvious time-series characteristics, the GRU sub-layer is used to extract the time-series characteristics of the measured data efficiently. The gating mechanism is introduced into the model structure to automatically adjust the proportion of each sub-layer in constructing the feature representation of different evaluation tasks, which not only promotes the information sharing among different tasks, but also weakens the negative impact of the differences among different tasks on model training. The experimental results in IEEE 39-bus test system show that the model proposed in this paper has better evaluation performance and computing speed.

Key words: transient stability assessment, deep learning, multi-task learning, gated recurrent unit (GRU)

中图分类号:

TM712

孙黎霞, 彭嘉杰, 田屹昀, 陈欣凌, 袁欢. 基于混合门控循环单元子层的多任务暂态稳定评估[J]. 电力建设, 2022, 43(2): 63-69.

SUN Lixia, PENG Jiajie, TIAN Yiyun, CHEN Xinling, YUAN Huan. Multi-task Transient Stability Assessment Based on Sub-layer of Hybrid Gated Recurrent Unit[J]. ELECTRIC POWER CONSTRUCTION, 2022, 43(2): 63-69.

图/表 10

图1 门控循环单元的结构

Fig.1 Structure of a gated recurrent unit

图2 硬参数共享多任务学习

Fig.2 Hard parameter shared multi-task learning

图3 混合子层多任务模型

Fig.3 Multi-task learning with multi-gate mixture-of-expert

图4 多任务模型评估流程

Fig.4 Process of multi-task model evaluation

表1 暂态稳定评估结果

Table 1 Transient stability assessment results

图5 特征提取效果

Fig.5 Feature extraction effect

图6 门控机制输出结果

Fig.6 Result of gating mechanism

表2 风机替换方式

Table 2 Replacement mode of wind power generators

表3 新能源系统下暂态稳定评估结果

Table 3 Results of transient stability assessment for new energy system

表4 考虑噪声干扰时的暂态稳定评估结果

Table 4 Transient-stability assessment results with considering noise

参考文献 24

[1]	KUNDUR P. Power system stability and control[M]. New York, USA: McGraw-Hill, 1994.
[2]	IEEE/CIGRE Joint Task Force on Stability Terms and Definitions. Definition and classification of power system stability[J]. IEEE Transactions on Power Systems, 2004, 19(3):1387-1401. doi: 10.1109/TPWRS.2004.825981 URL
[3]	谢小荣, 贺静波, 毛航银, 等. “双高”电力系统稳定性的新问题及分类探讨[J]. 中国电机工程学报, 2021, 41(2):461-475.
	XIE Xiaorong, HE Jingbo, MAO Hangyin, et al. New issues and classification of power system stability with high shares of renewables and power electronics[J]. Proceedings of the CSEE, 2021, 41(2):461-475.
[4]	周悦, 谭本东, 李淼, 等. 基于深度学习的电力系统暂态稳定评估方法[J]. 电力建设, 2018, 39(2):103-108.
	ZHOU Yue, TAN Bendong, LI Miao, et al. Transient stability assessment of power system based on deep learning technology[J]. Electric Power Construction, 2018, 39(2):103-108.
[5]	孙黎霞, 白景涛, 周照宇, 等. 基于双向长短期记忆网络的电力系统暂态稳定评估[J]. 电力系统自动化, 2020, 44(13):64-72.
	SUN Lixia, BAI Jingtao, ZHOU Zhaoyu, et al. Transient stability assessment of power system based on bi-directional long-short-term memory network[J]. Automation of Electric Power Systems, 2020, 44(13):64-72.
[6]	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.
[7]	田芳, 周孝信, 史东宇, 等. 基于卷积神经网络的电力系统暂态稳定预防控制方法[J]. 电力系统保护与控制, 2020, 48(18):1-8.
	TIAN Fang, ZHOU Xiaoxin, SHI Dongyu, et al. A preventive control method of power system transient stability based on a convolutional neural network[J]. Power System Protection and Control, 2020, 48(18):1-8.
[8]	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. A data-driven method for transient voltage stability assessment and its interpretability analysis[J]. Power System Technology, 2021, 45(11):4416-4425.
[9]	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(6):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(6):1922-1931.
[10]	梁修锐, 刘道伟, 杨红英, 等. 数据驱动的电力系统静态电压稳定态势评估[J]. 电力建设, 2020, 41(1):126-132.
	LIANG Xiurui, LIU Daowei, YANG Hongying, et al. Data-driven situation assessment of power system static voltage stability[J]. Electric Power Construction, 2020, 41(1):126-132.
[11]	汤涌. 电力系统电压稳定性分析[M]. 北京: 科学出版社, 2011.
	ZHANG Y, YANG Q. An overview of multi-task learning[J]. National Science Review, 2018, 5(1):30-43.
[13]	GUPTA A, GURRALA G, SASTRY P S. An online power system stability monitoring system using convolutional neural networks[J]. IEEE Transactions on Power Systems, 2019, 34(2):864-872. doi: 10.1109/TPWRS.2018.2872505 URL
[14]	陈庆超, 韩松, 毛钧毅. 采用多层次特征融合 SPP-net 的暂态稳定多任务预测[J/OL]. 控制与决策, 2021:1-10[2021-05-10]. https://doi.org/10.13195/j.kzyjc.2020.1568.
	CHEN Qingchao, HAN Song, MAO Junyi. Multi-task prediction for transient stability using multi-level feature fusion based SPP-net[J/OL]. Control and Decision, 2021:1-10[2021-05-10]. https://doi.org/10.13195/j.kzyjc.2020.1568.
[15]	MA J Q, ZHAO Z, YI X Y, et al. Modeling task relationships in multi-task learning with multi-gate mixture-of-experts[C]// Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. London United Kingdom. New York, NY, USA: ACM, 2018: 1930-1939.
[16]	CHO K, VAN MERRIENBOER B, GULCEHRE C, et al. Learning phrase representations using RNN encoder-decoder for statistical machine translation[J]. Computer Science, 2014: 1406. 1078.
[17]	刘取. 电力系统稳定性及发电机励磁控制[M]. 北京: 中国电力出版社, 2007.
	LIU Qu. Power system stability and generator excitation control[M]. Beijing: China Electric Power Press, 2007.
[18]	霍思敏, 王科, 陈震海, 等. 基于轨迹输入特征支持向量机的湖南电网暂态稳定在线识别[J]. 电力系统保护与控制, 2012, 40(18):19-23.
	HUO Simin, WANG Ke, CHEN Zhenhai, et al. Hunan power grid transient stability online detection based on support vector machine with trajectory input features[J]. Power System Protection and Control, 2012, 40(18):19-23.
[19]	力系统安全稳定计算技术规范: DL/T 1234—2013[S]. 北京: 中国电力出版社, 2013.
	Technique specification of power system security and stability calculation:DL/T 1234—2013[S]. Beijing: China Electric Power Press, 2013.
[20]	KINGMA D, BA J. Adam: A method for stochastic optimization[EB/OL].(2018-09-01) [2019-12-01]. http:www.oalib.com/paper4068193.
[21]	SRIVASTAVA N, HINTON G, KRIZHEVSKY A, et al. Dropout: A simple way to prevent neural networks from overfitting[J]. Journal of Machine Learning Research, 2014, 15(1):1929-1958.
[22]	PAIMA. Energy function analysis for power system stability[M]. Berlin,Germany:Springer Science & Business Media, 2012.
[23]	ABADI M, BARHAM P, CHEN J, et al. Tensorflow: A system for large-scale machine learning[C]// Proceedings of the 12th USENIX conference on operating systems design and implementation. Savannah, GA, USA, 2016:265-283.
[24]	CHOLLET F. Keras: Theano-based deep learning library[EB/OL]. (2018-05-09)[2021-05-10]. https://github.com/fchollet/keras.

[1]	肖发龙, 吴岳忠, 沈雪豪, 何震凯, 秦烨. 基于深度学习和知识图谱的变电站设备故障智能诊断[J]. 电力建设, 2022, 43(3): 66-74.
[2]	朱庆, 郑红娟, 唐子逸, 韦思雅, 邹子骁, 吴熙. 基于生成对抗网络的综合能源负荷场景生成方法[J]. 电力建设, 2021, 42(12): 1-8.
[3]	唐子卓, 刘洋, 许立雄, 郭久亿. 基于负荷数据频域特征和LSTM网络的类别不平衡负荷典型用电模式提取方法[J]. 电力建设, 2020, 41(8): 17-24.
[4]	庄家懿, 杨国华, 郑豪丰, 王煜东, 胡瑞琨, 丁旭. 并行多模型融合的混合神经网络超短期负荷预测[J]. 电力建设, 2020, 41(10): 1-8.
[5]	陈倩，齐林海，王红. 基于LSTM网络的谐波多标签分类[J]. 电力建设, 2019, 40(7): 101-106.
[6]	肖泽青，华昊辰，曹军威. 人工智能在能源互联网中的应用综述[J]. 电力建设, 2019, 40(5): 63-70.
[7]	杨懿男，齐林海，王红，苏林萍. 基于生成对抗网络的小样本数据生成技术研究[J]. 电力建设, 2019, 40(5): 71-77.
[8]	梁荣，杨波，马润泽，吴健，吴奎华，林振智，文福拴. 利用多源信息和深度置信神经网络的配电系统空间负荷预测[J]. 电力建设, 2018, 39(10): 12-19.
[9]	杨甲甲, 刘国龙, 赵俊华, 文福拴, 董朝阳. 采用长短期记忆深度学习模型的工业负荷短期预测方法[J]. 电力建设, 2018, 39(10): 20-27.
[10]	史佳琪，张建华. 基于深度学习的超短期光伏精细化预测模型研究 [J]. 电力建设, 2017, 38(6): 28-.
[11]	王晓辉，朱永利，郭丰娟. 基于同步监测和深度学习的电容器介损角辨识[J]. 电力建设, 2017, 38(5): 76-.
[12]	石鑫，朱永利. 深度学习神经网络在电力变压器故障诊断中的应用[J]. 电力建设, 2015, 36(12): 116-122.
[13]	方怡,王建,王晓茹 . 电力系统暂态稳定评估的关联规则挖掘[J]. 电力建设, 2008, 29(2): 83-86.