A Method for Fault Detection of Micro-Motors Based on Wavelet Packet Energy Spectra

The research proposes a fault detection method for micro-motors based on wavelet packet energy spectra. The vibration signals of micro-motors under different fault types were tested. Afterward, the decomposition and reconstruction algorithm of wavelet packet transform is used to calculate the energy spectra of each sub-band which are then normalized. The energy changes of signals in each frequency band contain abundant information on the operation state of micro-motors. The energy of signals in different frequency bands shows significant differences under normal and fault conditions. With the eigenvectors built based on frequency-band energy spectra, the mapping relation from the energy change to fault can be constructed. The preliminary experimental results demonstrate that wavelet packet energy spectra can be applied to fault detection of micro-motors.