Fractal Fault Diagnosis of Rotor System Based on Morphological De-nosing

As an experimental tool for analyzing movement of chaos, we intend to introduce the correlation dimension for analyzing vibration signals of rotor system in this paper. For noise corruption existing in field-measured vibration signal, morphological filter is also introduced for de-noising the vibration signal. The analysis and comparison between the correlation dimensions of the noised and de-noised vibration signals of rotor system at different states are done. The result shows that reflecting the system´s characteristic with the correlation dimension of noised signal is largely unreliable, so the field-measured signal must be de-noised prior to the correlation dimension calculation. The favorable de-noising effect by morphological filter is also showed. The correlation dimension of de-noised vibration signals can reflect and distinguish the actual state of rotor effectively. As a character to classify the fault type and state of rotor in fault diagnosis, the correlation dimension after de-noising is feasible.