基于谱峭度解卷积的特高压直流输电换流阀空冷器故障诊断方法

Fault Diagnosis Method of Converter Valve Air Cooler Based on Spectral Kurtosis Deconvolution

针对特高压直流输电换流站换流阀的空冷器电机在实际运行过程中易出现的复合故障,该文提出了空冷器电机早期复合故障的诊断方法。基于频谱分析法和快速峭度图解决了空冷器电机潜在的,由于微弱故障累积而导致的换流阀乃至换流站运行不安全、不稳定的问题。首先,从振动信号入手,通过分析快速峭度图,得到复合故障模式下严重故障的谱峭度值,根据谱峭度值构建带通滤波器,进而得到严重故障的故障特征;然后,再构建带阻滤波器,通过多次滤波人为地不断衰减严重故障的故障特征,从而凸显微弱故障,经过多次滤波后使原本耦合的复合故障得到有效分离;最后,采用换流阀空冷器电机运行的实际信号,对所提算法进行了验证。

In view of the compound faults that are easy to occur in the actual operation of the air cooler motor of the converter valve in the UHVDC transmission station,this paper puts forward the early compound fault diagnosis method of the air cooler motor.Based on the spectrum analysis method and fast kurtosis diagram,the problem of unsafe and unstable operation of converter valve and converter station caused by potential weak fault accumulation of air cooler motor is solved.Firstly,starting from the vibration signal,the kurtosis value of the serious fault under the compound fault mode is obtained,and the band-pass filter is established according to the kurtosis value to get the fault characteristics of the serious fault.Secondly,we build the band stop filter,which continuously attenuates the fault characteristics of the serious fault through the multiple filtering,so as to highlight the weak fault.After multiple filtering,the original coupled compound fault is effectively separated.Finally,the application effect of the algorithm is verified by the actual signal of the air cooler motor.