Research Of Gear Fault Enhancement Detection Based On Minimum Entropy Deconvolution

It is very important for predictive maintenance and healthy management of transmissions and drivelines to detect gear failure early and to estimate the remaining life,but the early failure characteristics of gears are easily submerged in other vibration signals or background noise.Therefore,it is necessary to carry out the enhancement detection technology of weak gear fault.At present,much research on gear fault enhancement detection at home and abroad focuses on reducing the background noise in the signal and on decomposing or reconstructing the signal which can characterize the fault from the vibration signal.On the one hand,Time Synchronous Average(TSA)is a typical technology of gear vibration signal noise reduction.However,the conventional realization of TSA generally requires additionally installing a phase-locked multiplier or speed sensor that can provide a time stamp signal,which brings a limitation to the engineering application of TSA technology.On the other hand,some of methods of signal decomposition and reconstruction lack of strict theoretical basis,and the physical meanings of obtained signal by these methods are not very clear,which makes it sometimes difficult to explain the physics.Focusing on the above problems,the enhancement detection technique of gear fault impact signal based on Fixed Minimum Entropy De-convolution(FMED)is studied and the application of enhanced signal to TSA algorithm in the absence of phase discrimination signal in this paper,and the main works are as follows:(1)The mechanisms and vibration signal characteristics of gear in several typical failure modes are analyzed,and the existing simulation model of vibration signal related to gear in these typical fault modes are improved in order to lay the foundation for the following research;(2)The theory and algorithm of Minimum Entropy De-convolution(MED)and Fixed Minimum Entropy De-convolution(FMED)are studied.Then,the enhancement detection method of gear fault based on FMED is proposed,and the validity of the proposed method is verified by simulation data.(3)The TSA theory based on angular re-sampling is studied.Then,a new TSA algorithm based on FMED is proposed,and the validity of the proposed method is verified by simulation data.(4)The gear pitting failure test was carried out on a two stage wind turbine gearbox,and the validity and practicability of the above methods are verified by test data.The results shows that: both the MED and FMED methods can realize the enhancement detection of gear fault by enhancing the fault impact of gear vibration signal and restraining the meshing vibration signal and noise;the FMED filter is more robust than the MED filter and it can enhance the detection of fault impulses that are disturbed by strong background noise;the proposed new TSA algorithm based on FMED technique can obtain an effective TSA signal in the absence of phase discrimination signal.