Research On Spur Gear Wear Detection Method On Vibration Signal

Gear fault detection methods based on vibration signal analysis can effectively identify and diagnose gear faults,but there are still problems with the identification and diagnosis of early gear wear.Faults of gears usually produce shocks,with the result that transient excitations can be observed in the vibration signal.However,in the early wear stage,disturbed by high-speed meshing vibration and strong environmental noise,localized wear on the tooth surface can lead to weak transients in the vibration signal,making feature extraction difficult.Aiming at this problem,this paper takes spur gears as the research object,and studies the weak fault detection method of gear wear under non-stationary conditions.This paper first analyzes the common gear failure modes and fault characteristics by establishing a gear dynamic model of gears,and analyzes the spectral components of wear signals by establishing the mathematical model of vibration signals.Because the collected signal is a multi-component modulated signal,and the accuracy of the collection is greatly affected by the location of the measuring point and background noise,it causes the problem of difficulty in detecting weak wear faults.To solve this problem,a combination of Variational Modal Decomposition（VMD）based on spectral correlation analysis and Hilbert Transform（HT）is introduced into gear wear detection.For weak gear vibration signals that can characterize early wear,the modal numbers are initialized by an approximate complete reconstruction criterion.Meanwhile,the frequency corresponding to the maximum value of the signal power spectral density is used to initialize the center frequency of the VMD algorithm.Then,the Hilbert envelope spectrum is used to locate the Intrinsic Mode Function（IMF）containing the gear meshing frequency,and the wear feature vector is constructed using feature extraction.Finally,Support Vector Machine（SVM）optimized by Adaptive Particle Swarm Optimization（APSO）is used for diagnosis.In order to verify the effectiveness of the proposed method in monitoring the early health status of gears,this paper designs a gearbox experimental platform to collect experimental data of gears with six different health states for validation.The experimental results show that the detection accuracy reaches 94.4% at the feature number of 2520,which can provide a solution for the detection of weak faults arising from early gear wear.