Research On Fault Diagnosis Of Planet Bearing Based On Signal Deconvolution

Planetary gearbox is widely used in mechanical equipment that needs to provide large transmission ratio and transmission stiffness.Its good working state is conducive to ensuring the safe and effective operation of the equipment.Because the planet bearing is a relatively easy fault component in the planetary gearbox,the fault diagnosis of the planet bearing is of great significance to further improve the safety and reliability of the planetary gearbox.In the planetary gearbox,the planet bearing has two movements of rotation and revolution at the same time,its fault signal will be subject to complex modulation,and its fault vibration signal will be coupled with the vibration signals of other components.Under the interference of gear signal and noise,the fault feature of planet bearing is easily filtered out as noise,which makes it difficult to extract local weak fault features effectively.Aiming at the problem of fault diagnosis of planet bearing inner race,a fault feature extraction method combining Fast Iterative Filter Decomposition and parameter adaptive Multipoint Optimal Minimum Entropy Deconvolution was proposed.The vibration signal was decomposed by Fast Iterative Filter Decomposition,and the components containing more bearing fault information in the decomposition result are selected by envelope entropy to reduce the influence of gear signal and noise signal.The Multipoint Optimal Minimum Entropy Deconvolution was used to enhance the fault features of the selected components.In order to improve the performance of the algorithm,the joint index based on kurtosis and square envelope spectral kurtosis was proposed as the fitness function of the Particle swarm optimization algorithm.The filter length of the Multipoint Optimal Minimum Entropy Deconvolution was adaptively selected to effectively extract the fault features of the inner race of the planet bearing.In the outer race fault signal,there is amplitude modulation caused by the rotation of the planetary gear,resulting in weaker fault impact than that of the inner race.Therefore,a method combining Discrete Random Separation and Maximum SecondOrder Cyclostationary Blind Deconvolution was proposed.Firstly,the influence of strictly periodic signals such as gear signals was suppressed by Discrete Random Separation,and then the fault impact characteristics were enhanced by using the Maximum Second-Order Cyclostationary Blind Deconvolution technology.Finally,the fault characteristics were intuitively reflected by the square envelope spectrum.Aiming at the parameter optimization of the Maximum Second-Order Cyclostationary Blind Deconvolution algorithm,a salp swarm optimization algorithm with the improved fault characteristic ratio index as the fitness function was proposed to optimize the cycle frequency and filter length.In the simulation study,through the analysis of the characteristics of the fault vibration signal of the bearing in the planetary gearbox,the numerical simulation signals of the inner race and outer race of the planet bearing were established.In the experimental research,taking NGW single-stage planetary gearbox as the research object,a comprehensive experimental platform of planetary gearbox was built.In the experiment,the artificial local fault was used to simulate the inner and outer race faults of planet bearing under actual working conditions,and the fault vibration signal and speed pulse signal are collected respectively.Simulation and experimental results verify the effectiveness of the proposed method.