Feature extraction of diesel engine vibration signal based on waveletpacket transform and singularity value decomposition
Guobin Li, Guobin Li, Guobin Li · Zhendong yu chongji · 2011
In order to recognize the operating conditions of diesel engine,a novel approach for extracting features of vibration signals was proposed using waveletpacket transform and singularity value decomposition.The waveletpacket transform and the singularity value decomposition algorithms were introduced.The characteristic parameters were defined based on the singular value characteristic vector.The intrinsic relationship between the characteristic parameters and the operating condition of diesel engine were discussed.It is shown that the characteristic parameters proposed can sensitively reflect the working performance of diesel engine.When the working performance of diesel engine turns to be worse and the intensity of vibration increases,these parameters appear larger.Therefore the characteristic parameters proposed could be considered as parameters to inspect the condition and to diagnose the fault of engine.