Research On Fault Diagnosis Of Gear Box Based On Hilbert-huang Transform

The gear box is a key transmission component in the mechanical equipment.But due to its relatively complex structure and the harsh working environment in general and together with the complex machining process of the main components inside the gear box,causing the gearbox to easily malfunction during operation.In addition,the vibration signals in the event of a failure are all non-stationary signals.The commonly used time-frequency analysis methods cannot meet the requirements for processing non-stationary signals.Therefore,appropriate methods are required to deal with the vibration signals of the gearbox.Aiming at the common fault problems and vibration characteristics of gearboxes,this paper will use the Hilbert-Huang transform method to analyze the fault signals of the gearbox,adaptively decompose the original signals by EMD method,and then perform Hilbert transform to obtain Hilbert spectrum and Hilbert margin.Spectrum,by analyzing the Hilbert spectrum and Hilbert’s marginal spectrum,to determine if a failure occurred and what kind of failure occurred.This paper addresses two important issues in the Hilbert-Huang transform.Firstly,for the noise reduction problem of the signal,the wavelet threshold denoising method based on EMD method is adopted in this paper.The wavelet threshold denoising is performed for the IMF component after EMD decomposition,and the disadvantages and deficiencies of soft threshold function and hard threshold function are proposed.A multi-level threshold function is used to set the threshold.The simulation shows that the wavelet threshold denoising method based on EMD method can well solve the problem of signal denoising;secondly,it proposes an improvement to the endpoint effect that occurs in EMD decomposition.The endpoint matching method of waveform matching finds the position inside the signal that matches the slope of the signal at the endpoint and copies a signal in front of the position to the endpoint.The result shows that this method can handle the endpoint effect better..The vibration signals of the concentrated faults of gears and bearings in the gearbox were collected through experiments.Based on the previous theoretical methods,the MATLAB software was used to process the gear tooth surface wear and gear tooth breakage faults and the vibration signals of the rolling bearing inner ring faults,using Hilbert.The spectrum and the characteristics of Hilbert’s marginal spectrum are used to determine the type of failure.The results are consistent with the actual failure types,which verifies the feasibility and accuracy of the Hilbert-Huang transform method.Based on the theory of Hilbert-Huang transform,a gearbox fault diagnosis system was developed using MATLAB GUI.The system was used to process a vibration signal.The results showed that the system could correctly analyze the signal data and display the corresponding graphs.