Abstract: Aiming at the special condition of the port crane gearbox and the difficulty of tracking the degradation trend, the symbol sequence entropy algorithm, statistical properties of vibration energy spectrum and Weibull fitting theory were studied. The effective value of the normal energy spectrum was introduced as the threshold to reduce the random fluctuation of the original symbol sequence entropy algorithm. A degeneration feature extraction method based on improved symbol sequence entropy and sliding window Weibull fitting was proposed. The lifetime data of the hoisting gearbox monitored online by a terminal in Shanghai Port was tested. The results show that the improved symbol sequence entropy is able to describe the complexity of the vibration energy spectrum and the proposed degradation character parameter can stably reflect the performance degradation trend. The deeper of the degradation degree, the larger of the character parameter, and this method has the smooth and accurate advantages.
Key words: port crane; degradation feature extraction; symbol sequence entropy; Weibull distribution; gearbox