In order to improve the accuracy of vibration monitoring and fault diagnosis, a blind source separation frame which is combining mathematical morphology preprocessing and Fast ICA arithmetic was brought out based on the characteristics of dual rotors aero engine case vibration signal. It is that the original signal was preprocessed by open—close and close—open combinatorial mathematical morphology. The signal noise can be well filtered, but its fundamental form can be reserved. The Fast ICA arithmetic was used to get back the high and low rotor-s vibration waveform. The frame was used to analyse the real testing data. The result shows that the method can well respectively resume the high and low rotor signal frequency domain structure information. It is very helpful to improve the accuracy of fault location.
Hu Aijun��Tang Guiji��An Liansuo.De-noising technique for vibration signals of rotating machinery based on mathematical morphology filter[J].Chinese Journal of Mechanical Engineering��2006��42(4)��127-130(in Chinese)