Research On Fault Diagnosis Of Planet Bearings Based On Analysis Of Vibration And Motor Current

Planetary gearboxes have several parallel paths formed by planet gears to share the torque load.They are widely used in the transmission systems of helicopters,ships,wind turbines and vehicles,as a result of their compact structure and large transmission ratio.Once faults occur in the planetary gearboxes in mechanical systems,the faults can possibly decrease the machine efficiency,increase the industrial operating costs.Thus,it is valuable economically to monitor the health condition and diagnose the faults of planetary gearboxes.Planet bearings undertake the transmission load affected by the load distribution of planet gears,meanwhile they are responsible for supporting the planet gears rotating around the planet shafts.Hence,they are more prone to faults.Moreover,the motion of planet bearings is complex,comprising of the revolution around the sun gear center with the planet carrier and the spinning around their own shafts.1)The phenomenon of the unequal load sharing among planets is ubiquitous in various degrees of severity.In order to support the fault diagnosis of planet bearings,it is necessary to investigate the mechanism of unequal load sharing and its vibrational characteristics.This dissertation mainly takes account for different factors,including the position error of the planet shaft center,the input load,the planet number,the natural frequency of the system and the time-varying transfer path,and proposes the vibration signal model of the unequal load sharing among planets.The revealed characteristics of the gear vibration offer a reference for the following analysis of the spectral structure of planet bearing signals.2)In order to overcome the limitation of the traditional spectral kurtosis in extracting periodic impulses,we propose the infogram based on spectral negentropy to determine the frequency range of periodic impulses caused by planet bearing faults,then design the optimal bandpass filter parameters and diagnose faults by combining envelope demodulation and frequency demodulation methods;3)We establish a vibration signal model considering the slippage of rolling elements,and improve the Multi-point Optimization Minimized Entropy Deconvolution Adjusted(MOMEDA)algorithm,by using spectral negentropy to estimate the slippage parameter of rolling elements.Thus,the problem of the blur in vibration signal spectra and the weak fault signature is successfully resolved.4)A vibration signal model under time-varying speed conditions is established.Through combining the Multi-Order Probabilistic Approach(MOPA)and Adaptive Iterative Generalized Demodulation(AIGD),we identify the instantaneous fault characteristic frequencies via the high-precision time-frequency representation based on the accurate extraction of the instantaneous angular speed with large and rapid fluctuation,and finally realize the fault diagnosis of planet bearings under nonstationary conditions;5)In the wind power generation system driven by planetary gears,considering the influence of planet bearing faults on the generator current,we establish a signal model of the generator current and develop the analysis procedure of amplitude and frequency demodulation.Because the amplitude of the current signal is not modulated by the rotation of the planet carrier,the complexity of the signal is reduced.The current signal analysis avoids the problem that the measuring locations of the vibration sensors cannot be easily arranged.The above signal models and theoretical methods are substantiated by the experimental data analyses.