Complex Fault Diagnosis Of Gear Box Based On Improved Algorithm Of Morphological Component Analysis

Gearbox has the advantage of fixed transmission ratio,compact structure,high reliability,high transmission efficiency and strong bearing capacity.As the middle transmission link of the power system and the actuating mechanism,gearbox is widely applied in various mechanical.The normal and stable operation of the gearbox has an important influence on the transmission of motion and power.Once the failure occurs,it directly affects the normal work of the whole machine.The condition monitoring and fault diagnosis of the gear box have important value of scientific research and practical significance.There are many methods of gearbox fault diagnosis,but the analysis of vibration signal method is the most extensive,convenient and mature.Vibration signal analysis is the key to gearbox fault diagnosis.Through statistical analysis of the fault location of the gear box,the gear and the bearing are the most vulnerable parts of the gear box.The failure mode,failure performance and vibration mechanism of the gear and bearing are analyzed,and the mathematical model of the vibration of the gear and bearing is analyzed in the case of local failure.When the gear is locally malfunction,the amplitude and frequency components often appear in the vibration signal.When the bearing is partial fault,the periodic and transient impact component often appears in its vibration signal.LabVIEW is a graphical programming software which widely applied in the field of measurement and control system.On the basis of it,EEMD-HHT virtual instrument is implemented and applied in gearbox fault diagnosis.Hilbert-Huang transform is composed of two parts by EMD decomposition and Hilbert transform.The signal is decomposed into several intrinsic mode functions(IMFs)by EMD decomposition,and then each intrinsic mode component perform Hilbert transform to obtain the instantaneous spectrum of each one.Combination of the instantaneous spectrum shows time-frequency information of the signal.The vibration signal of the disabled gear from sensors is decomposed to the IMFs and its instantaneous frequency spectrum is obtained by HHT virtual instrument,and the fault of the gear box is diagnosed.The HHT method performed well in the diagnosis of a fault in the gear box.When the gearbox has compound faults,each fault signal will interfere with each other,and the MCA is used to decompose the fault signal.As a new signal decomposition method,the MCA method takes advantage of the morphological difference of the components of the signal and reconstructs the components of the signal through different overcomplete dictionaries to achieve the signal decomposition.In the MCA method,a semi-soft threshold wavelet denoising method and maximum median method is proposed to update the threshold to overcome the shortcoming of the linear decreasing of the threshold.After obtaining good results in simulation experiments,the MCA method is applied to get the fault signals of gear and bearing in compound fault signal of gear box,and the specific fault of gear and bearing is obtained through Hilbert marginal spectrum.