Research on Fault Diagnosis Algorithm for Vibration Signals of Electromechanical Equipment
Fault signals operated in a complex condition is easy to be lost in the noises,however,the traditional methods for feature exaction can not get good results,reducing the accuracy of fault diagnosis.To solve this problem,a new algorithm for fault diagnosis is purposed based on band variance of wavelet packet decomposition and parameters optimized support vector machine.First,the vibration signal wavelet packet decomposition coefficients were calculated for each sub-band variance and used as feature vectors.Then the feature vectors acted as SVM input to optimize its parameters and achieve fault "multi-to-multi" classification.The simulation results show that the algorithm can improve the fault diagnostic accuracy of electromechanical equipment vibration signal.
NETL
NSTL
万方数据
维普
