基于参数优化MCKD算法的交流接触器异常噪声信号特征提取

Hum Fault Feature Extraction of AC Contactor Based on Parameter Optimization MCKD Algorithm

交流接触器的不良品在线圈通电后发出异常电磁性噪声,超出规定噪声值。针对交流接触器声音信号的特点,提出最大相关峭度解卷积(MCKD)与包络谱分析结合的方法,提取故障频率,实现交流接触器的异常振动噪声故障诊断。首先对采集到的声信号带通滤波,减少噪声干扰,再通过粒子群算法搜寻MCKD算法最优参数,对信号解卷积,增强信号中周期性脉冲,MCKD算法处理后的信号通过Hilbert变换求包络谱,以故障频率幅值因子为判据判断交流接触器是否故障。通过对交流接触器实际测试验证,结果表明所提方法能够有效诊断交流接触器异常噪声。

The defective products of AC contactor emit abnormal electro-magnetic noise after the coil is energized,which exceeds the specified noise value.According to the characteristics of AC contactor sound signal,a method combining maximum correlation kurtosis deconvolution(MCKD)algorithm and envelope spectrum analysis is proposed to extract the fault frequency and realize the abnormal vibration noise fault diagnosis of AC contactor.Firstly,the collected original signal is filtered by a band-pass filter to reduce noise interference.Then,the optimal MCKD parameters are searched by the particle swarm optimization algorithm.The signal is deconvolved to enhance the periodic pulse in the signal.The signal processed by the MCKD algorithm is used to obtain the envelope spectrum through Hilbert transform,and the fault frequency amplitude factor is used as the criterion to determine whether the AC contactors is faulty.The actual test of AC contactor shows that the method can effectively diagnose the abnormal noise of AC contactor.