基于深度学习模型融合的电压暂降源识别方法
|
摘要
电压暂降源的识别是制定电压暂降治理方案和明确事故责任的基础。电压暂降源可分为单一电压暂降源和复合电压暂降源,电网设备的复杂化和用电模式的区域化对基于物理特征的传统电压暂降源识别方法提出了新的挑战。该文提出一种基于模型融合的电压暂降源识别方法,通过深度学习算法中的卷积神经网络获取电压暂降信号的时序特征和空间特征,采用深度置信网络替换卷积神经网络中用于提纯高维特征和起分类器作用的全连接层,从而增强网络的多标签分类能力。利用仿真和加噪数据对网络进行迭代训练和反复测试,验证了融合模型的高识别精度和抗噪性能。对比传统的电压暂降源识别方法,生成的模型具有良好的泛化能力,能够有效应用于实际工程中。
The recognition of the voltage sag sources is the basis for formulating a voltage sag governance plan and clarifying the responsibility for the accident. The causes of voltage sag can be divided into single voltage sag sources and composite voltage sag sources, the complexity of grid equipment and the regionalization of power consumption patterns present new challenges to the traditional recognition method of voltage sag sources based on physical characteristics. This paper proposed a voltage sag sources' recognition method based on model fusion, using the convolutional neural network(CNN) in the deep learning algorithm to obtain the timing characteristics and spatial characteristics of voltage sag signals, the deep confidence network(DBN) was used to replace the full-connected layers in CNN for purifying high-dimensional features and acting as a classifier, in order to enhance the network's multi-tag classification capabilities. Utilize simulated and noise-added data to iteratively train and test the network repeatedly, high recognition accuracy and anti-noise performance of the fusion model have been verified. Compared with the traditional voltage sag sources' recognition method, the generated model has good generalization ability and can be effectively applied in practical projects.
The recognition of the voltage sag sources is the basis for formulating a voltage sag governance plan and clarifying the responsibility for the accident. The causes of voltage sag can be divided into single voltage sag sources and composite voltage sag sources, the complexity of grid equipment and the regionalization of power consumption patterns present new challenges to the traditional recognition method of voltage sag sources based on physical characteristics. This paper proposed a voltage sag sources' recognition method based on model fusion, using the convolutional neural network(CNN) in the deep learning algorithm to obtain the timing characteristics and spatial characteristics of voltage sag signals, the deep confidence network(DBN) was used to replace the full-connected layers in CNN for purifying high-dimensional features and acting as a classifier, in order to enhance the network's multi-tag classification capabilities. Utilize simulated and noise-added data to iteratively train and test the network repeatedly, high recognition accuracy and anti-noise performance of the fusion model have been verified. Compared with the traditional voltage sag sources' recognition method, the generated model has good generalization ability and can be effectively applied in practical projects.
引文
[1]IEEE.1564-2014 IEEE guide for voltage sag indices[S].New York:IEEE,2014.
[2]肖湘宁,韩民晓,徐永海,等.电能质量分析与控制[M].北京:中国电力出版社,2010.Xiao Xiangning,Han Minxiao,Xu Yonghai,et al.Power quality analysis and control[M].Beijing:Electric Power,2010(in Chinese).
[3]李夏林,刘雅娟,朱武.基于配电网的复合电压暂降源分类与识别新方法[J].电力系统保护与控制,2017,45(2):131-139.Li Xialin,Liu Yajuan,Zhu Wu.A new method to classify and identify composite voltage sag sources in distribution network[J].Power System Protection&Control,2017,45(2):131-139(in Chinese).
[4]王克星,宋政湘,陈德桂,等.基于小波变换的配电网电压暂降的干扰源辨识[J].中国电机工程学报,2003,23(6):29-34,54.Wang Kexing,Song Zhengxiang,Chen Degui,et al.Interference source identification of voltage sag in distribution system based on wavelet transform[J].Proceedings of the CSEE,2003,23(6):29-34,54(in Chinese).
[5]赵凤展,杨仁刚.基于短时傅里叶变换的电压暂降扰动检测[J].中国电机工程学报,2007,27(10):28-34,109.Zhao Fengzhan,Yang Rengang.Voltage sag disturbance detection based on short time fourier transform[J].Proceedings of the CSEE,2007,27(10):28-34,109(in Chinese).
[6]Manjula M,Sarma A V R S.Classification of voltage sag causes using probabilistic neural network and Hilbert-Huang transform[J].International Journal of Computer Applications,2010,1(20):22-29.
[7]陈丽,王硕,孔维功.基于改进S变换的复合电压暂降源识别特征分析[J].电力系统保护与控制,2014,42(4):27-33.Chen Li,Wang Shuo,Kong Weigong.Method to identify composite voltage sag disturbance sources based on generalized S-transform[J].Power System Protection&Control,2014,42(4):27-33(in Chinese).
[8]贾勇,何正友,赵静.基于小波熵和概率神经网络的配电网电压暂降源识别方法[J].电网技术,2009,33(16):63-69.Jia Yong,He Zhengyou,Zhao Jing.A method to identify voltage sag sources in distribution network based on wavelet entropy and probability neural network[J].Power System Technology,2009,33(16):63-69(in Chinese).
[9]吕干云,方奇品,蔡秀珊.基于多分类支持向量机的电压暂降源识别[J].电力系统保护与控制,2010,38(22):151-155.LüGanyun,Fang Qipin,Cai Xiushan.Identification of voltage sags source based on multi-class support vector machine[J].Power System Protection&Control,2010,38(22):151-155(in Chinese).
[10]赵莹,赵川,叶华,等.应用主成分分析约简电压暂降扰动源识别特征的方法[J].电力系统保护与控制,2015,43(13):105-110.Zhao Ying,Zhao Chuan,Ye Hua,et al.Method to reduce identification feature of different voltage sag disturbance source based on principal component analysis[J].Power System Protection&Control,2015,43(13):105-110(in Chinese).
[11]李晨懿,杨家莉,徐永海,等.模糊综合评价在电压暂降源识别中的应用[J].电网技术,2017,41(3):1022-1028.Li Chenyi,Yang Jiali,Xu Yonghai,et al.Application of comprehensive fuzzy evaluation method on recognition of voltage sag disturbance sources[J].Power System Technology,2017,41(3):1022-1028(in Chinese).
[12]Schmidhuber J.Deep learning in neural networks[J].Neural Networks,2015,61:85-117.
[13]Zhao Guangquan,Zhang Guohui,Ge Qiangqiang,et al.Research advances in fault diagnosis and prognostic based on deep learning[C]//Prognostics and System Health Management Conference.Chengdu,China:IEEE,2016:1-6.
[14]周飞燕,金林鹏,董军.卷积神经网络研究综述[J].计算机学报,2017,40(6):1229-1251.Zhou Feiyan,Jin Linpeng,Dong Jun.Review of convolutional neural network[J].Chinese Journal of Computers,2017,40(6):1229-1251(in Chinese).
[15]陈先昌.基于卷积神经网络的深度学习算法与应用研究[D].杭州:浙江工商大学,2014.Chen Xianchang.Research on algorithm and application of deep learning based on convolutional neural network[D].Hangzhou:Zhejiang Gongshang University,2014(in Chinese).
[16]朱乔木,党杰,陈金富,等.基于深度置信网络的电力系统暂态稳定评估方法[J].中国电机工程学报,2018,38(3):735-743.Zhu Qiaomu,Dang Jie,Chen Jinfu,et al.A method for power system transient stability assessment based on deep belief networks[J].Proceedings of the CSEE,2018,38(3):735-743(in Chinese).
[17]刘方园,王水花,张煜东.深度置信网络模型及应用研究综述[J].计算机工程与应用,2018,54(1):11-18,47.Liu Fangyuan,Wang Shuihua,Zhang Yudong.Survey on deep belief network model and its applications[J].Computer Engineering and Applications,2018,54(1):11-18,47(in Chinese).
[18]Hinton G.A practical guide to training restricted Boltzmann machines[J].Momentum,2010,9(1):926-947.
[19]Tan R H G,Ramachandaramurthy V K.Simulation of power quality events using Simulink model[C]//2013IEEE 7th International Power Engineering and Optimization Conference.Langkawi,Malaysia:IEEE,2013:277-281.
[20]丁宁,蔡维,锁娟,等.电压暂降源识别方法研究[J].电网技术,2008,32(S2):55-59.Ding Ning,Cai Wei,Suo Juan,et al.Research on voltage sag sources recognition method[J].Power System Technology,2008,32(S2):55-59(in Chinese).
[21]肖先勇,李逢,邓武军.雷击与短路故障的S变换特征量识别方法[J].高电压技术,2009,35(4):817-822.Xiao Xianyong,Li Feng,Deng Wujun.Identification method of lightning stroke and short-circuit faults using S-transformed characteristics[J].High Voltage Engineering,2009,35(4):817-822(in Chinese).
[22]祁博,邹金慧,范玉刚,等.基于Hilbert-Huang变换和小波包能量谱的电压暂降源识别[J].中国电力,2013,46(8):112-117.Qi Bo,Zou Jinhui,Fan Yugang,et al.Identification of voltage sags source based on Hilbert-Huang transform and wavelet packet energy spectrum[J].Electric Power,2013,46(8):112-117(in Chinese).
[23]储佳伟,袁晓冬,陈兵,等.结合小波分析和改进型DTW距离的配电网电压暂降源辨识方法[J].电网技术,2018,42(2):637-643.Chu Jiawei,Yuan Xiaodong,Chen Bing,et al.A method for distribution network voltage sag source identification combining wavelet analysis and modified DTWdistance[J].Power System Technology,2018,42(2):637-643(in Chinese).
[2]肖湘宁,韩民晓,徐永海,等.电能质量分析与控制[M].北京:中国电力出版社,2010.Xiao Xiangning,Han Minxiao,Xu Yonghai,et al.Power quality analysis and control[M].Beijing:Electric Power,2010(in Chinese).
[3]李夏林,刘雅娟,朱武.基于配电网的复合电压暂降源分类与识别新方法[J].电力系统保护与控制,2017,45(2):131-139.Li Xialin,Liu Yajuan,Zhu Wu.A new method to classify and identify composite voltage sag sources in distribution network[J].Power System Protection&Control,2017,45(2):131-139(in Chinese).
[4]王克星,宋政湘,陈德桂,等.基于小波变换的配电网电压暂降的干扰源辨识[J].中国电机工程学报,2003,23(6):29-34,54.Wang Kexing,Song Zhengxiang,Chen Degui,et al.Interference source identification of voltage sag in distribution system based on wavelet transform[J].Proceedings of the CSEE,2003,23(6):29-34,54(in Chinese).
[5]赵凤展,杨仁刚.基于短时傅里叶变换的电压暂降扰动检测[J].中国电机工程学报,2007,27(10):28-34,109.Zhao Fengzhan,Yang Rengang.Voltage sag disturbance detection based on short time fourier transform[J].Proceedings of the CSEE,2007,27(10):28-34,109(in Chinese).
[6]Manjula M,Sarma A V R S.Classification of voltage sag causes using probabilistic neural network and Hilbert-Huang transform[J].International Journal of Computer Applications,2010,1(20):22-29.
[7]陈丽,王硕,孔维功.基于改进S变换的复合电压暂降源识别特征分析[J].电力系统保护与控制,2014,42(4):27-33.Chen Li,Wang Shuo,Kong Weigong.Method to identify composite voltage sag disturbance sources based on generalized S-transform[J].Power System Protection&Control,2014,42(4):27-33(in Chinese).
[8]贾勇,何正友,赵静.基于小波熵和概率神经网络的配电网电压暂降源识别方法[J].电网技术,2009,33(16):63-69.Jia Yong,He Zhengyou,Zhao Jing.A method to identify voltage sag sources in distribution network based on wavelet entropy and probability neural network[J].Power System Technology,2009,33(16):63-69(in Chinese).
[9]吕干云,方奇品,蔡秀珊.基于多分类支持向量机的电压暂降源识别[J].电力系统保护与控制,2010,38(22):151-155.LüGanyun,Fang Qipin,Cai Xiushan.Identification of voltage sags source based on multi-class support vector machine[J].Power System Protection&Control,2010,38(22):151-155(in Chinese).
[10]赵莹,赵川,叶华,等.应用主成分分析约简电压暂降扰动源识别特征的方法[J].电力系统保护与控制,2015,43(13):105-110.Zhao Ying,Zhao Chuan,Ye Hua,et al.Method to reduce identification feature of different voltage sag disturbance source based on principal component analysis[J].Power System Protection&Control,2015,43(13):105-110(in Chinese).
[11]李晨懿,杨家莉,徐永海,等.模糊综合评价在电压暂降源识别中的应用[J].电网技术,2017,41(3):1022-1028.Li Chenyi,Yang Jiali,Xu Yonghai,et al.Application of comprehensive fuzzy evaluation method on recognition of voltage sag disturbance sources[J].Power System Technology,2017,41(3):1022-1028(in Chinese).
[12]Schmidhuber J.Deep learning in neural networks[J].Neural Networks,2015,61:85-117.
[13]Zhao Guangquan,Zhang Guohui,Ge Qiangqiang,et al.Research advances in fault diagnosis and prognostic based on deep learning[C]//Prognostics and System Health Management Conference.Chengdu,China:IEEE,2016:1-6.
[14]周飞燕,金林鹏,董军.卷积神经网络研究综述[J].计算机学报,2017,40(6):1229-1251.Zhou Feiyan,Jin Linpeng,Dong Jun.Review of convolutional neural network[J].Chinese Journal of Computers,2017,40(6):1229-1251(in Chinese).
[15]陈先昌.基于卷积神经网络的深度学习算法与应用研究[D].杭州:浙江工商大学,2014.Chen Xianchang.Research on algorithm and application of deep learning based on convolutional neural network[D].Hangzhou:Zhejiang Gongshang University,2014(in Chinese).
[16]朱乔木,党杰,陈金富,等.基于深度置信网络的电力系统暂态稳定评估方法[J].中国电机工程学报,2018,38(3):735-743.Zhu Qiaomu,Dang Jie,Chen Jinfu,et al.A method for power system transient stability assessment based on deep belief networks[J].Proceedings of the CSEE,2018,38(3):735-743(in Chinese).
[17]刘方园,王水花,张煜东.深度置信网络模型及应用研究综述[J].计算机工程与应用,2018,54(1):11-18,47.Liu Fangyuan,Wang Shuihua,Zhang Yudong.Survey on deep belief network model and its applications[J].Computer Engineering and Applications,2018,54(1):11-18,47(in Chinese).
[18]Hinton G.A practical guide to training restricted Boltzmann machines[J].Momentum,2010,9(1):926-947.
[19]Tan R H G,Ramachandaramurthy V K.Simulation of power quality events using Simulink model[C]//2013IEEE 7th International Power Engineering and Optimization Conference.Langkawi,Malaysia:IEEE,2013:277-281.
[20]丁宁,蔡维,锁娟,等.电压暂降源识别方法研究[J].电网技术,2008,32(S2):55-59.Ding Ning,Cai Wei,Suo Juan,et al.Research on voltage sag sources recognition method[J].Power System Technology,2008,32(S2):55-59(in Chinese).
[21]肖先勇,李逢,邓武军.雷击与短路故障的S变换特征量识别方法[J].高电压技术,2009,35(4):817-822.Xiao Xianyong,Li Feng,Deng Wujun.Identification method of lightning stroke and short-circuit faults using S-transformed characteristics[J].High Voltage Engineering,2009,35(4):817-822(in Chinese).
[22]祁博,邹金慧,范玉刚,等.基于Hilbert-Huang变换和小波包能量谱的电压暂降源识别[J].中国电力,2013,46(8):112-117.Qi Bo,Zou Jinhui,Fan Yugang,et al.Identification of voltage sags source based on Hilbert-Huang transform and wavelet packet energy spectrum[J].Electric Power,2013,46(8):112-117(in Chinese).
[23]储佳伟,袁晓冬,陈兵,等.结合小波分析和改进型DTW距离的配电网电压暂降源辨识方法[J].电网技术,2018,42(2):637-643.Chu Jiawei,Yuan Xiaodong,Chen Bing,et al.A method for distribution network voltage sag source identification combining wavelet analysis and modified DTWdistance[J].Power System Technology,2018,42(2):637-643(in Chinese).