Research On Fault Diagnosis Method Of Planetary Gear Transmission System Based On Sparse Decomposition

Planetary gear transmission system is widely used in industrial machinery because of its large transmission ratio and compact structure.As an important part of mechanical transmission system,planetary gear structure may cause major accidents if it fails.Therefore,the condition monitoring and fault diagnosis of planetary gear transmission system is of great engineering significance.The sparse decomposition algorithm proposed in recent years has achieved good results in fault signal extraction.In this paper,based on the sparse decomposition theory,fault signal analysis and fault feature extraction of the two core parts of gear and bearing in planetary gear transmission system are studied.The main contents are as follows:(1)This paper takes the planetary gear transmission system as the research object,and analyses the vibration signal forms of bearing and gear under different working conditions and different faults from the angle of signal processing.The fault signal modulation forms of bearings and gears under different faults are also studied.According to the actual working conditions of planetary gear system,a planetary gear simulation test-bed is established,and different working conditions and loads are set up.The fault signals are collected by the acquisition system.(2)By studying the sparse decomposition algorithm,this paper improves the low efficiency of the orthogonal matching pursuit algorithm,introduces the concept of residual ratio threshold,determines the optimal sparseness by observing the variation of residual ratio threshold,and improves the accuracy of fault signal extraction using sparse decomposition algorithm.Then,spectrum analysis and envelope analysis are combined with fault characteristic frequency to realize fault classification.The validity of the method is verified by simulation and experimental analysis.(3)In the case of unknown fault characteristic frequency,the above method can not be used to judge the fault category,so sparse classification algorithm is introduced to judge the fault category in the case of unknown fault frequency.In order to reduce the effect of noise on sparse classification,this paper preprocesses the fault signal as training signal by empirical mode decomposition,highlights the fault feature components in training signal,and makes the training signal have better identification.In this paper,the above methods are simulated and experimentally analyzed by using a variety of fault signals,which proves the effectiveness of the above methods.