Research On Detection Of Abnormal Gearbox Based On Optical Fiber Sensor

In the process of modern industry,the machinery and equipment are developing towards the direction of intelligence.Gearbox which is an important transmission device in mechanical equipment,will cause serious loss in the fault event,so it is essential to detect faults of the gearbox.Fiber Bragg grating(FBG)sensors have many excellent features and have become a research focus in the structural health detection.The application of FBG sensors in the gearbox faults detection,has higher accuracy than the conventional acceleration sensors.Therefore,the design of a gearbox fault detection system based on FBG senesors has important research significance and application prospects.To solve the problem of noise in the gearbox vibration signal extracted by FBG sensors,an improved FBG signal denoising method based on orthogonal matching pursuit algorithm is proposed.The sparsity of FBG sensing signals is determined by sparse transformation.The threshold and the elimination strategy are designed to reduce the time complexity of the improved algorithm.Termination condition is modified to improve the signal-to-noise ratio.The simulation results show that the proposed method can effectively remove the noise in the FBG signal.To solve the problem that the accuracy of the gear fault identification and fault location is poor,and the traditional Empirical Mode Decomposition(EMD)algorithm is easily affected by mode mixing,a gearbox fault diagnosis method based on improved EMD algorithm is proposed.Firstly,the FBG sensing signal is preprocessed,and the signal spectrum is homogenized by adaptively compensating the Gaussian white noise,so that the intrinsic mode function(IMF)is free from the effect of the mode mixing.A comprehensive evaluation index for selecting effective IMF components is proposed,and four characteristics of effective components are extracted.The multi-class support vector machine is used to identify six gear faults.Finally,the Hilbert demodulation method is employed to obtain the envelope spectrum of the IMF components,and the peak frequency and the multiplication frequency can be obtained.The position of the fault gears can be located by the frequency comparison.The gearbox fault experiment platform is established to verify the effectiveness of the proposed method.The experimental results show that the proposed method can effectively identify fault types such as normal,mild wear,severe wear,pitting,cracks and broken teeth,and the identification accuracy is above 90%.At the same time,the proposed method can precisely locate the position of the faulty gear.