摘要
目前设备的机械故障诊断技术的研究多限于定性诊断,而故障诊断中故障程度的定量评估更能有效的指导设备维护。该文提出了一种低压万能式断路器分合闸故障程度定量评估的方法。首先对断路器工作模式进行识别,即利用局部均值分解(local mean decomposition,LMD)将采集到的分合闸振动信号自适应分解,求取主要乘积函数(product function,PF)的改进多尺度排列熵(multi-scale permutation entropy,MMPE)构成特征向量,再经过降维后,作为改进支持向量机(support vector machine,SVM)的输入量,实现断路器工作模式的识别;当断路器处于故障模式时,对采集的振动信号求取多尺度排列熵偏均值(partial mean of multi-scale permutation entropy,PMMPE),作为故障程度定量评估指标,并参照所求得的不同故障模式的故障程度特性曲线,可实现分合闸故障程度的定量评估。经实测数据验证表明,所提方法可以完成断路器工作模式的有效识别,且PMMPE指标相较于峭度、能量和多尺度排列熵平均值指标,能够更加有效的完成低压万能式断路器分合闸故障程度的定量评估。
At present the research of mechanical fault diagnosis technology for equipment is limited to qualitative diagnosis, but quantitative evaluation of fault degree in fault diagnosis can more effectively guide the equipment maintenance. This paper put forward a quantitative evaluation method of fault degree for low voltage conventional circuit breaker switching fault. Firstly, working mode recognition of conventional circuit breaker was carried, so the switching vibration signal was adaptively decomposed by local mean decomposition （LMD）, and then the modified multi-scale permutation entropy （MMPE） as feature vector of the main product function （PF） was calculated, and after dimension reduction, the feature vector was inputted to improved support vector machine（SVM）, to realize the recognition of working modes of conventional circuit breaker; Secondly, when the breaker was in fault mode, the partial mean of multi-scale permutation entropy （PMMPE） of the vibration signal was obtained as quantitative evaluation indicator of fault degree, and it could realize the quantitative evaluation of switching fault degree according to the fault degree characteristic curve in different fault modes. The experiment result shows that this method can identify the working modes, and the PMMPE can more effectively accomplish the quantitative evaluation of switching fault degree than indicators such as kurtosis, energy and the mean of multi-scale permutation entropy.
出处
《中国电机工程学报》
EI
CSCD
北大核心
2017年第18期5473-5482,共10页
Proceedings of the CSEE
基金
河北省教育厅资助科研项目(ZD2016108)
天津市科技特派员项目(16JCTPJC51700)~~
