Research on fault feature extraction method of abnormal noise of diesel engine based on wavelet energy spectrum

A data acquisition system has been designed to capture the vibration signal for the engines. A test has been carried out for the vibration signal acquisition of piston knock and piston pin knock of 5.9L Cummins 6BT diesel engine. The test data has been processed by continuous wavelet transform. The scale-energy spectrum and the 3-dimentinal (3D) energy spectrum plot have been generated. The fault feature of abnormal noise of piston knock and piston pin knock of the engine under various technical condition and operating status has been extracted. The research result shows that, the setup of scale-energy spectrum and 3D energy spectrum plot by applying continuous wavelet transform is an efficient method for multi-source abnormal noise extraction for diesel engines, it can not only differentiate piston knock and piston pin knock visually and efficiently, but also can provide useful information to identity the fault level of them.