Torsional Vibration Signal Analysis And Fault Diagnosis Of Planetary Gear Box

Planetary gearboxes have some advantages,including big transmission ratio and high transmission efficiency,so the planetary gearboxes play a significant role in engineering fields.Because its structures are more complex than that of fixed-axis gears,the faults such as petting and fatigue crack occur easily under varying conditions.Therefore,condition monitoring and fault diagnosis for planetary gearboxes is more important to prevent sudden accident and reduce equipment maintenance costs.In the thesis,torsional vibration measurement is proposed as a new diagnosis method.The main contents are as follows:Combining the incremental encoder and high frequency clock pulse counting is presented to measure torsional vibration signals from planetary gearboxes,and then an acquisition system to sample the torsional vibration signal is established.According to the high frequency clock pulse counting principle,the high-resolution incremental encoder and multi-function data acquisition card are selected.By analyzing the Fourier spectra of stationary torsional vibration signals and Order spectra of non-stationary torsional vibration signals,and then comparing the features with these of transverse vibration signals,the results show that the measuring method is effective and available to diagnose planetary gearbox faults.And torsional vibration signals have simpler transmission paths,simpler frequency spectrum and high SNR,so it is more sensitive to the faults of planetary gearboxes.The effect of torsional vibration signals sampled at different sampling locations is studied.First the torsional vibration signals of the input and output shafts are separately measured without extra loads.Comparing the frequency spectra of the two signals,the facts indicate that the signals collected from the input shaft are more sensitive to detect the faults than that from the output shaft.On the contrary,the torsional vibration signals sampled from the output shaft are better than these from the input shaft for the fault detection under heavy load.Aiming at the fault feature extraction of planetary gearbox under heavy loads and strong background noise,a novel method based on minimum entropy deconvolution is presented.In this method,minimum entropy deconvolution is introduced to stress the impulsive signal,and the fault features can be extracted easily.The effectiveness of the method is validated by both simulated signal and the torsional vibration signals of fault planetary gearbox.Via comparing the effect over different loads,the results demonstrate that the proposed method is more efficient in the heavy loads.At the same time,the minimum entropy deconvolution technique can solve the problem that the fault feature can not be extracted in the torsional vibration signals acquired from the input shaft.