Fault Diagnosis Of Mechanical Equipment With PCA Method

In this paper mechanical equipment fault diagnosis classfication method wasstudied, gear and bearing were the objects. Feature extraction: The feature value of timedomain includes peak value;The feature value of frequency domain is MSF;the characteristicof AR model and GREEN function were extracted.The PCA method was used for detecting the rolling bearing fault of fan in powerplant. From the vibration signals of reference and fault stations the feature of time domainand the feature of frequent domain , the characteristic of AR model and GREEN functionwere extracted and PCA was calculated. The results showed that the reference and faultstations of fan can be distinguished clearly in the PCA diagram.A vibration signal of gearbox was analysed by PCA and KPCA methoddifferencely ,Normal and fault state of gear was distinguished.The result showed that they allcan do the pattern recognition,but the change course of the gear can better to showed by theKPCA, a distance function was used to identify the stations of the gearbox.The SVM methodwas used to find the support vector from the PC and KPC.The database of bear was empoldered by Delphi 7,the parameter of bear can bequeried and the characteristic of bear can be calculated.