Research On Blind-separation Technology Of Rotor Vibration Signals With Coupling Faults

In pursuit of high performance and high thrust-weight ratio, the structure of the modern aero- engines is becoming more and more complex, which makes aircraft engines running at high speed will inevitably lead to large vibration of various parts,even causing faults,especially the rotor system. However, because of the large and complex structure, the signal generated by the engine's vibration is often not a single signal, but contains more than one f vibration signal. Thus, it is the key that correct extraction of the useful signal, which premises condition monitoring and fault diagnosis of an aero-engine in a real time.In this paper, according to the vibration characteristics of coupling faults of rotor system, a new model of rotor system with coupling faults was established, then, the nonlinear characteristics of this coupling fault system was analyzed, finally, the blind source separation was applied to the rotor fault diagnosis. The research work of this paper is carried out as follows:(1)The main context about blind source separation (BSS) basic theory and typical algorithms-FastICA algorithm is introducted systematicly, highlighting its performance in mechanical malfunction diagnosis;(2) A new model of rotor system with coupling faults was established. In this new model, the rotor shaft was considered as Equisection Euler Free Beam model, considering the imbalance, rubbing, cracking and loosening fault. With the modal truncation method, the system's responses are obtained using the numerical integral method. The ZT-3 type multi-function rotor test rig is used to verify the new model. Finally, the bifurcation diagrams are used to analysis the nonlinear characteristics of this coupling fault system;(3) The blind source separation (BSS) theory is used in machinery fault diagnosis.Firstly, the unbalance, rubbing,crack and loosening faults are simulated using that new model and ZT-3 type multi-function rotor test rig.And the vibration response of rotor system are obtained respectively by simulation and measurement sensors.Last, the simulated and measured rotor vibration data are separated successfully.