[1]王 辉,陈佳宁,金 雪,等.基于ReliefF的时频联合特征及随机森林的配电网电缆故障识别方法[J].南京师范大学学报(工程技术版),2020,20(02):044-51.[doi:10.3969/j.issn.1672-1292.2020.02.007]
　Wang Hui,Chen Jianing,Jin Xue,et al.Fault Identification Method of Distribution Cable Based on Time-FrequencyDomain Features Extracted by ReliefF and Random Forest Algorithm[J].Journal of Nanjing Normal University(Engineering and Technology),2020,20(02):044-51.[doi:10.3969/j.issn.1672-1292.2020.02.007]
点击复制

基于ReliefF的时频联合特征及随机森林的配电网电缆故障识别方法
分享到：

南京师范大学学报（工程技术版）[ISSN:1006-6977/CN:61-1281/TN]

卷:
20卷
期数:
2020年02期
页码:
044-51
栏目:
电气工程
出版日期:
2020-05-15

文章信息/Info

Title:
Fault Identification Method of Distribution Cable Based on Time-FrequencyDomain Features Extracted by ReliefF and Random Forest Algorithm
文章编号:
1672-1292(2020)02-0044-08
作者:
王 辉1; 陈佳宁2; 金 雪1; 冯 双2
(1.国电南瑞南京控制系统有限公司,江苏 南京 211106)(2.东南大学电气工程学院,江苏 南京 210096)
Author(s):
Wang Hui1; Chen Jianing2; Jin Xue1; Feng Shuang2
(1.Nari Technology Co.,Ltd.,Nanjing 211106,China)(2.School of Electrical Engineering,Southeast University,Nanjing 210096,China)
关键词:
配电网电缆; 短路故障识别; 时频联合特征; ReliefF算法; 随机森林
Keywords:
cable of distribution network; short-circuit fault identification; time-frequency domain feature; ReliefF; random forest
分类号:
TM726
DOI:
10.3969/j.issn.1672-1292.2020.02.007
文献标志码:
A
摘要:
提出一种基于ReliefF算法的时频联合特征及随机森林的配电网电缆故障识别方法. 针对零序电压,从时域和频域构造23个故障敏感特征,采用ReliefF算法进行特征选择,得到最具分类能力的特征子集. 将特征子集作为基于随机森林的输入进行训练,得到最终的分类模型,实现了电缆故障类型识别. 所提方法与基于单一特征的方法相比,能够更加充分地挖据数据潜力,同时由于采用ReliefF算法筛除了无关特征,提高了算法效率. 最后采用Matlab软件进行仿真,并与决策树、KNN、SVM等算法进行比较,仿真结果验证了所提方法的可行性和高准确性.
Abstract:
This paper proposes a fault identification method of distribution cable based on time-frequency domain features extracted by ReliefF and random forest algorithm. 23 fault-sensitive features in time and frequency domain are constructed with bases of the zero sequence voltage,which are then selected with RefiefF method to obtain the subset of features with strong ability to identify the types of fault. The subset of features serves as the input of random forest to train the model identifying the type of short circuit fault. Compared with the method utilizing single feature,the proposed method exploits the potential of data and is more robust. Besides,because of the elimination of irrelative features,the efficiency of the method is improved. The simulation is carried out in Matlab and compared with decision tree,KNN and SVM algorithms. The result demonstrates the feasibility and high accuracy of the proposed method.

参考文献/References:

[1] 王成山,李鹏,于浩. 智能配电网的新形态及其灵活性特征分析与应用[J]. 电力系统自动化,2018,42(10):13-21.
[2]姜楠,王琦,王恩荣,等. 分布式光伏电源接入对配电网可靠性的影响研究[J]. 南京师范大学学报(工程技术版),2016,16(2):1-9.
[3]方毅,薛永端,宋华茂,等. 谐振接地系统高阻接地故障暂态能量分析与选线[J]. 中国电机工程学报,2018,38(19):5636-5645,5921.
[4]张姝. 配电网弱故障接地保护与定位方法研究[D]. 成都:西南交通大学,2018.
[5]吴京. 基于模分量小波能量谱的电缆故障识别方法研究[D]. 西安:西安科技大学,2014.
[6]DAS B. Fuzzy logic-based fault-type identification in unbalanced radial power distribution system[J]. IEEE Transactions on Power Delivery,2006,21(1):278-285.
[7]李小薇. 基于IPSO-SVM的电缆故障识别[D]. 西安:西安科技大学,2014.
[8]杨春宇. 电力电缆故障分析与诊断技术的研究[D]. 大连:大连理工大学,2013.
[9]孙萌. 基于小波奇异熵的配电网电缆接地故障研究[D]. 西安:西安理工大学,2017.
[10]苏立. 基于HHT变换和FOA_LSSVM的电缆故障诊断[J]. 计算机与现代化,2017(9):96-101,105.
[11]何清,李宁,罗文娟,等. 大数据下的机器学习算法综述[J]. 模式识别与人工智能,2014,27(4):327-336.
[12]JENKE R,PEER A,BUSS M. Feature extraction and selection for emotion recognition from EEG[J]. IEEE Transactions on Affective Computing,2014,5(3):327-339.
[13]何涛,胡洁,夏鹏,等. 基于ReliefF算法与遗传算法的肌电信号特征选择[J]. 上海交通大学学报,2016,50(2):204-208.
[14]SABAHI F,AHMAD M O,SWAMY M N S. Perceptual image hashing using random forest for content-based image retrieval[C]//2018 16th IEEE International New Circuits and Systems Conference(NEWCAS). Montreal,Canada:IEEE,2018:348-351.
[15]彭徵. 基于随机森林的文本分类并行化研究[D]. 湘潭:湘潭大学,2018.

备注/Memo

备注/Memo:
收稿日期:2019-03-07.
通讯作者:冯双,博士,讲师,研究方向:电力系统运行与控制、电力电子化电力系统. E-mail:sfeng@seu.edu.cn
更新日期/Last Update: 2020-05-15