Recognition of Complex PQ Disturbances Based on SVM Cascaded Decision Tree

doi:10.11985/2023.02.014

Abstract

Abstract:

Aiming at the problem of features overlapping and difficult recognition of complex power quality(PQ) disturbance, a PQ disturbance recognition method is proposed based on support vector machine(SVM) cascaded decision tree. Firstly, based on the initial characteristics of the disturbance extracted by the S-transform, 7 two-class SVMs are constructed for 7 single PQ disturbances such as voltage sag, harmonics, and voltage interruption. Then, in order to reduce the feature overlap of the complex disturbance, two-class statistical machine learning of the SVMs are used to reconstruct the features of disturbance. Finally, the constructed features are used to realize multi-class recognition of the complex PQ disturbance by the SVM cascade decision tree, combining with the two discrete feature processing of the CART decision tree. Simulation results show that the proposed method can effectively reduce the degree of feature overlap for composite PQ disturbances with strong noise immunity, and accurately identify 21 types of PQ disturbances including double and multiple complex disturbances.

Key words: Power quality, complex disturbances, cascade classifier, support vector machine, decision tree

CLC Number:

TM714

CHEN Nuo, LÜ Ganyun, YE Jiaxing. Recognition of Complex PQ Disturbances Based on SVM Cascaded Decision Tree[J]. Journal of Electrical Engineering, 2023, 18(2): 149-156.

Figures/Tables 10

类型号	信号类型	表达式	参数
T1	电压暂降	$v (t) = 1 − a u t 1 − u t 2 cos (ω t)$	$0.1 ≤ a ≤ 0.9$ , $0.1 ≤ a ≤ 0.9$
T2	电压暂升	$v (t) = 1 + a u t 1 − u t 2 cos (ω t)$	$0.1 ≤ a ≤ 0.9$ , $0.1 ≤ a ≤ 0.9$
T3	谐波	$v (t) = cos (ω t) + α 3 cos (3 ω t) + α 5 cos (5 ω t) + α 7 cos (7 ω t)$	$f = 50 Hz$ , $f = 50 Hz$
T4	电压中断	$v (t) = 1 − a u t 1 − u t 2 cos (ω t)$	$0.9 ≤ a ≤ 1$ , $0.9 ≤ a ≤ 1$
T5	闪变	$v (t) = 1 + α cos (β ω t) cos (ω t)$	$0.1 ≤ α ≤ 0.2$ , $0.1 ≤ α ≤ 0.2$
T6	振荡暂态	$v (t) = cos (ω t) + a exp − (t − t 1) / τ ⋅ cos ω n (t − t 1) ⋅ u (t 1) − u (t 2)$	$0.1 ≤ a ≤ 0.8$ , $0.1 ≤ a ≤ 0.8$ , $0.1 ≤ a ≤ 0.8$ , $0.1 ≤ a ≤ 0.8$
T7	脉冲暂态	$v (t) = cos (ω t) + a u (t 1) − u (t 2)$	$1 ≤ a ≤ 10$ , $1 ≤ a ≤ 10$

References 23

[1]	汪颖, 罗代军, 肖先勇, 等. IEEE ICHQP2018与电能质量发展方向[J]. 电力自动化设备, 2019, 39(4):197-203.
	WANG Ying, LUO Daijun, XIAO Xianyong, et al. IEEE ICHQP2018 and development directions of power quality[J]. Electric Power Automation Equipment, 2019, 39(4):197-203.
[2]	蒋洁, 周福林, 钟凯, 等. 基于组合赋权和改进灰色关联分析的电能质量综合评价方法[J]. 电力系统及其自动化学报, 2020, 32(8):47-53.
	JIANG Jie, ZHOU Fulin, ZHONG Kai, et al. Comprehensive evaluation method for power quality based on combinatorial empowerment and improved grey correlation analysis[J]. Proceedings of the CSU-EPSA, 2020, 32(8):47-53.
[3]	张杨, 刘志刚. 一种基于时频域多特征量的电能质量混合扰动分类新方法[J]. 中国电机工程学报, 2012, 32(34):83-90.
	ZHANG Yang, LIU Zhigang. A new method for power quality mixed disturbance classification based on time-frequency domain multiple features[J]. Proceedings of the CSEE, 2012, 32(34):83-90.
[4]	汪飞, 全晓庆, 任林涛. 电能质量扰动检测与识别方法研究综述[J]. 中国电机工程学报, 2021, 41(12):4104-4121.
	WANG Fei, QUAN Xiaoqing, REN Lintao. Review of power quality disturbance detection and identification methods[J]. Proceedings of the CSEE, 2021, 41(12):4104-4121.
[5]	占靳果, 朱清智, 孟阳, 等. 基于多层极限学习机的电能质量扰动多标签分类算法[J]. 电力系统保护与控制, 2020, 48(8):96-105.
	ZHAN Jinguo, ZHU Qingzhi, MENG Yang, et al. Power quality disturbances multi-label classification algorithm based on a multi-layer extremelearning machine[J]. Power System Protection and Control, 2020, 48(8):96-105.
[6]	王燕. 电能质量扰动检测的研究综述[J]. 电力系统保护与控制, 2021, 49(13):174-186.
	WANG Yan. Review of research development in power quality disturbance detection[J]. Power System Protection and Control, 2021, 49(13):174-186.
[7]	黄建明, 瞿合祚, 李晓明. 基于短时傅里叶变换及其谱峭度的电能质量混合扰动分类[J]. 电网技术, 2016, 40(10):3184-3191.
	HUANG Jianming, QU Hezuo, LI Xiaoming. Classification for hybrid power quality disturbance based on STFT and its spectral kurtosis[J]. Power System Technology, 2016, 40(10):3184-3191.
[8]	杨晓梅, 郭林明, 肖先勇. 基于可调品质因子小波变换和随机森林特征选择算法的电能质量复合扰动分类[J]. 电网技术, 2020, 44(8):3014-3020.
	YANG Xiaomei, GUO Linming, XIAO Xianyong. Classification of multiple power quality disturbances based on TQWT and random forest feature selection algorithm[J]. Power System Technology, 2020, 44(8):3014-3020.
[9]	易吉良, 彭建春, 谭会生. S变换在电能质量扰动分析中的应用综述[J]. 电力系统保护与控制, 2011, 39(3):141-147.
	YI Jiliang, PENG Jianchun, TAN Huisheng. A summary of S-transform applied to power quality disturbances analysis[J]. Power System Protection and Control, 2011, 39(3):141-147.
[10]	徐志超, 杨玲君, 李晓明. 基于聚类改进S变换与直接支持向量机的电能质量扰动识别[J]. 电力自动化设备, 2015, 35(7):50-58,73.
	XU Zhichao, YANG Lingjun, LI Xiaoming. Power quality disturbance identification based on clustering-modified S-transform and direct support vector machine[J]. Electric Power Automation Equipment, 2015, 35(7):50-58,73.
[11]	郭俊文, 李开成, 何顺帆. 基于改进不完全S变换与决策树的实时电能质量扰动分类[J]. 电力系统保护与控制, 2013, 41(22):103-110.
	GUO Junwen, LI Kaicheng, HE Shunfan. A real time power quality disturbance classification based on improved incomplete S-transform and decision tree[J]. Power System Protection and Control, 2013, 41(22):103-110.
[12]	李建闽, 林海军, 梁成斌, 等. 基于双分辨率S变换和学习向量量化神经网络的电能质量扰动检测方法[J]. 电工技术学报, 2019, 34(16):3453-3463.
	LI Jianmin, LIN Haijun, LIANG Chengbin, et al. Detection method of power quality disturbances based on double resolution S-transform and learning vector quantization neural network[J]. Transactions of China Electrotechnical Society, 2019, 34(16):3453-3463.
[13]	黄南天, 徐殿国, 刘晓胜. 基于S变换与SVM的电能质量复合扰动识别[J]. 电工技术学报, 2011, 26(10):23-30.
	HUANG Nantian, XU Dianguo, LIU Xiaosheng. Identification of power quality complex disturbances based on S-transform and SVM[J]. Transactions of China Electrotechnical Society, 2011, 26(10):23-30.
[14]	韩刚, 张建文, 禇鑫, 等. 多特征组合及优化SVM的电能质量扰动识别[J]. 电力系统及其自动化学报, 2015, 27(8):71-76,81.
	HAN Gang, ZHANG Jianwen, CHU Xin, et al. Power quality disturbance classification based on multi-features combination and optimizing parameters of SVM[J]. Proceedings of the CSU-EPSA, 2015, 27(8):71-76,81.
[15]	黄南天, 彭华, 蔡国伟, 等. 电能质量复合扰动特征选择与最优决策树构建[J]. 中国电机工程学报, 2017, 37(3):776-786.
	HUANG Nantian, PENG Hua, CAI Guowei, et al. Feature selection and optimal decision tree construction of complex power quality disturbances[J]. Proceedings of the CSEE, 2017, 37(3):776-786.
[16]	李祖明, 吕干云, 陈诺, 等. 基于混沌集成决策树的电能质量复合扰动识别[J]. 电力系统保护与控制, 2021, 49(21):18-27.
	LI Zuming, LÜ Ganyun, CHEN Nuo, et al. Recognition of hybrid PQ disturbances based on a chaos ensemble decision tree[J]. Power System Protection and Control, 2021, 49(21):18-27.
[17]	陈伟, 何家欢, 裴喜平. 深度信念网络在电能质量复合扰动识别中的应用[J]. 电力系统及其自动化学报, 2018, 30(9):75-82.
	CHEN Wei, HE Jiahuan, PEI Xiping. Application of deep belief network in power quality compound disturbance identification[J]. Proceedings of the CSU-EPSA, 2018, 30(9):75-82.
[18]	刘志刚, 张巧革, 张杨. 电能质量复合扰动分类的研究进展[J]. 电力系统保护与控制, 2013, 41(13):146-153.
	LIU Zhigang, ZHANG Qiaoge, ZHANG Yang. Review of power quality mixed disturbances identification[J]. Power System Protection and Control, 2013, 41(13):146-153.
[19]	张淑清, 李盼, 师荣艳, 等. 基于改进S变换的电能质量扰动分类新方法[J]. 仪器仪表学报, 2015, 36(4):927-934.
	ZHANG Shuqing, LI Pan, SHI Rongyan, et al. New method for power quality disturbance classification based on modified S transform[J]. Chinese Journal of Scientific Instrument, 2015, 36(4):927-934.
[20]	IEEE 1159-2009. Recommended practice for monitoring electric power quality[S]. Institute of Electrical and Electronic Engineers, 2009.
[21]	陈华丰, 张葛祥. 基于决策树和支持向量机的电能质量扰动识别[J]. 电网技术, 2013, 37(5):1272-1278.
	CHEN Huafeng, ZHANG Gexiang. Power quality disturbance identification using decision tree and support vector machine[J]. Power System Technology, 2013, 37(5):1272-1278.
[22]	张巧革, 刘志刚, 朱玲, 等. 基于多标签Rank-WSVM的复合电能质量扰动分类[J]. 中国电机工程学报, 2013, 33(28):114-120.
	ZHANG Qiaoge, LIU Zhigang, ZHU Ling, et al. Recognition of multiple power quality disturbances using multi-label wavelet support vector machine[J]. Proceedings of the CSEE, 2013, 33(28):114-120.
[23]	瞿合祚, 刘恒, 李晓明, 等. 基于多标签随机森林的电能质量复合扰动分类方法[J]. 电力系统保护与控制, 2017, 45(11):1-7.
	QU Hezuo, LIU Heng, LI Xiaoming, et al. Recognition of multiple power quality disturbances using multi-label random forest[J]. Power System Protection and Control, 2017, 45(11):1-7.

类型号	信号类型	类型号	信号类型
T1	电压暂降	T12	电压暂升+振荡暂态
T2	电压暂升	T13	闪变+振荡暂态
T3	谐波	T14	谐波+振荡暂态
T4	电压中断	T15	电压暂降+脉冲暂态
T5	闪变	T16	电压暂升+脉冲暂态
T6	振荡暂态	T17	谐波+脉冲暂态
T7	脉冲暂态	T18	电压暂降+振荡暂态+谐波
T8	电压暂降+谐波	T19	电压暂升+振荡暂态+谐波
T9	电压中断+谐波	T20	闪变+脉冲暂态+谐波
T10	闪变+谐波	T21	电压暂降+振荡暂态+ 脉冲暂态+谐波
T11	电压暂降+振荡暂态

扰动类型		分类准确率(%)
扰动类型		20 dB	30 dB	40 dB
T1	电压暂降	98.2	100.0	100.0
T2	电压暂升	95.7	97.3	99.5
T3	谐波	96.3	99.3	100.0
T4	电压中断	96.6	99.2	100.0
T5	闪变	96.3	98.6	100.0
T6	暂态振荡	96.8	99.8	100.0
T7	脉冲暂态	96.0	97.5	98.7
T8	电压暂降+谐波	95.1	97.7	98.5
T9	电压中断+谐波	95.6	97.8	99.2
T10	闪变+谐波	95.8	99.4	100.0
T11	电压暂降+振荡暂态	96.4	97.6	98.3
T12	电压暂升+脉冲暂态	97.0	97.9	98.2
T13	闪变+振荡暂态	96.7	97.0	100.0
T14	谐波+振荡暂态	95.8	98.3	100.0
T15	电压暂降+脉冲暂态	96.8	98.6	99.6
T16	电压暂升+脉冲暂态	95.1	97.6	98.3
T17	谐波+脉冲暂态	96.1	97.7	98.7
T18	电压暂降+振荡暂态+谐波	97.0	97.4	97.9
T19	电压暂升+振荡暂态+谐波	95.5	97.2	97.5
T20	闪变+脉冲暂态+谐波	96.1	96.8	97.2
T21	电压暂降+振荡暂态 +脉冲暂态+谐波	94.2	96.3	96.7
	平均准确率(%)	96.2	98.1	99.0

分类方法	不同噪声强度下的分类准确率(%)
分类方法	20 dB	30 dB	40 dB
SVM	94.6	95.2	95.7
决策树	94.2	94.5	94.8
量化神经网络^[12]	95.9	96.9	98.0
SVM级联决策树	96.2	98.1	99.0

分类方法	训练样本	测试样本	分类准确率(%)
决策树	840	210	89.2
	1 680	420	93.5
	3 360	840	94.5
SVM级联决策树	840	210	95.3
	1 680	420	97.9
	3 360	840	98.1