Research On The Vibration Signal Separation And Fault Diagnosis Of Planetary Gears

Planetary gear transmission is widely used in the mechanical device, theoperating state of the planetary gear plays an important role on the entire device, sothe operating state of the planetary gear-line monitoring and fault diagnosis isparticularly important. Fault diagnosis based on the vibration signal is a commonmethod for gear fault diagnosis. So, a method based on time-domain synchronousaveraging combined with window function and improved independent componentanalysis method for extraction and separation of the planetary gear vibration signal ispresented. The main contents are as follows:1. The paper calculates the parameters of the planetary gear box and finds out thelaw of the transmission, According to the feature of a particular planetary gear backturned to the minimum number of turns to the starting point, the photoelectric encodercan acquires the whole cycle signal, Research on window function leakagemechanism and propose time-domain synchronous averaging combining withwindowing function can get complete planetary gear signal;2. In order to isolate useful fault signal, use the improved FastICA algorithmfault, the paper researches the principle of fast independent component analysisseparation, and carries out an in-depth analysis of a nonlinear function of the objectivefunction of the algorithm, finding three non-linear functions of the original algorithmare so complex that the software calculates slow and affect the separation time andeffect, because of the multiplication and addition characteristics of Huber-M function,it is ease and simple to operate, through replacing the nonlinear function of FastICAalgorithm and get the new HM-FastICA algorithm, and finally, through experimentalanalysis to verify the effectiveness of the separation;3. This paper optimizes the planetary vibration signal acquisition system. Theexperiment analysis the entire life cycle of the planetary gear vibration signal, in orderto get damaged signal faster, this paper consideres the man-made damage, uses threesenors to ensure the accuracy of the signal, respectively, using the time-domainsynchronous averaging method to de-noise the vibration signal acquisition,interception and failure characteristics of the separation. Experimental results show that the method can not only effectively36planetary gear signal, but also clearlyreflect the current planetary gear vibration, indicating the effectiveness of the method;4. The use of HM-FastICA separately separated the time domain and frequencydomain signals, the results show that HM-FastICA algorithm allows higher precisionsignal separation effect, the algorithm is more convenient and the useful signalobtained can be strengthened, the meshing frequency and its harmonics appear in thevicinity of the modulated signal can be used as indicators of the fault monitoring, theseparation results can judge the work condition of the planetary gears, finallydimensionless parameters analysis the results, it can be seen that kurtosis can reflectthe current distribution of the vibration energy of the planet gear.