采用振动信号零相位滤波时频熵的高压断路器机械故障诊断

Mechanical Fault Diagnosis of High Voltage Circuit Breakers Utilizing Zero-phase Filter Time-frequency Entropy of Vibration Signal

为了以较少的计算量从高压断路器振动信号中获取机械状态信息,提出一种零相位滤波时频熵方法。通过小波包方法分析高压断路器振动信号的时频分布,并利用边际谱和瞬时能量密度计算时频平面的划分间隔。依据频带间隔,采用零相位数字滤波器提取频带分量;依据时间间隔,将各频带分量划分为多个子时间段。用子时段幅值包络的积分构造各频带分量的信息熵,并共同组成时频熵向量,对振动信号的时频特性进行量化表达。在真空断路器上进行模拟试验,获得了正常状态、铁心卡涩、软连接松动和绝缘拉杆故障等模式的样本数据。引入支持向量机分类器,对不同状态的样本数据进行诊断,准确率达90%。验证了通过零相位滤波时频熵方法诊断高压断路器机械故障的可行性。

To obtain the mechanical condition information contained in vibration signals of high voltage circuit breakers with small calculation cost,a zero-phase filter time-frequency entropy method was proposed for condition feature extraction.The time-frequency distributions of vibration signals were analyzed via wavelet packet decomposition method.The frequency intervals and the time intervals of the time–frequency plane were calculated based on the marginal spectrum and its instantaneous energy density level.According to the frequency intervals,different frequency components were separated from vibration signals using zero-phase digital filters.According to the time intervals,frequency components were segmented into time subsections.Entropy of a frequency component was constructed adopting the integral of the amplitude envelope of each time subsection,and all entropies composed a time-frequency entropy vector.Through these measures,the characteristics of a time–frequency plane were quantified by the time–frequency entropy vector.Simulative experiments on a vacuum circuit breaker were conducted and vibration signals of normal condition,friction in solenoid,loose flexible connection and defective insulating tie were collected.Support vector machine classifiers were introduced for fault diagnosis of different conditions and an accuracy of 90% was gain.The results verify the feasibility of the mechanical fault diagnosis approach for high voltage circuit breakers employing the zero-phase filter time–frequency entropy scheme.