摘要
为提高断路器机械振动信号的时—频特征分类能力、减小噪声干扰和提高断路器状态识别的准确性,提出一种基于S变换与优化随机森林算法的高压断路器机械故障诊断方法。首先,对高压断路器原始振动信号进行S变换;然后对S变换得到的时—频矩阵进行局部奇异值分解,以每个子矩阵的最大奇异值为特征向量;之后,将特征向量输入到随机森林中,以泛化误差与诊断准确率为综合指标对树的棵数进行寻优,构建最优随机森林分类器,最终实现对高压断路器机械故障状态的准确判别。对断路器实测振动数据开展对比实验,结果表明,新方法的特征类可分性好,整体故障识别准确性高。
In order to improve the time⁃frequency feature classification ability of circuit breaker mechanical vibra⁃tion signals,reduce noise interference and improve the accuracy of state recognition of circuit breakers,a mechani⁃cal fault diagnosis method for high voltage circuit breaker based on S⁃transform and optimal random forest algorithm is proposed.Firstly,the vibration signal of high voltage circuit breaker is processed by S⁃transform;after that the time⁃frequency matrix obtained by S⁃transform is decomposed by local singular value decomposition,and the maxi⁃mum singular value of each sub⁃matrix is the feature vector;Then,the feature vector is input into the random for⁃est,and the number of trees is synthesized by generalization error and diagnostic accuracy.The optimal random for⁃est classifier is constructed to realize the accurate identification of mechanical fault state of high voltage circuit break⁃er.Comparative experiment is carried out on the measured vibration data of circuit breaker,and the experimental re⁃sults show that the new method has good classification of features and high accuracy of fault identification.
作者
张欣
张静
高旭
任新卓
顾承天
朱玮翔
ZHANG Xin;ZHANG Jing;GAO Xu;REN Xinzhuo;GU Chengtian;ZHU Weixiang(Hangzhou Power Supply Company of State Grid,Hangzhou 310009,China;Department of Electrical Engineering,Northeast Electric Power University,Jilin 132012,China)
出处
《高压电器》
CAS
CSCD
北大核心
2020年第6期225-231,共7页
High Voltage Apparatus
基金
国家自然科学基金资助项目(51307020)。
关键词
高压断路器
机械故障诊断
S变换
局部奇异值分解
随机森林
high voltage circuit breaker
mechanical fault diagnosis
S⁃transform
local singular value decomposition
random forest
