Research on Aero-engine Vibration Fault Based on Statistical Feature and Artificial Intelligence Method

The vibration signals of the aero-engine compose of multi-frequency components as well as an abundant of noise. The signal was decomposed based on wavelet transform, the statistical feature of generalized roughness vector of the signal was formed by making use of the local energy distributions and the roughness of the sub-band signal. Then the desired parameters serve as the fault characteristic vectors to be input into two neural networks classifiers respectively and the fault patterns were identified by the output of the classifiers. These results were fused by applying the Dempster-Shafer (D-S) evidential theory of the information fusion. The analysis results from the aero-engine vibration signals show that the fault diagnosis method can classify working conditions and fault patterns effectively.