Weak Fault Diagnosis Of Planetary Gearbox Using Rigid-flexible Coupling Model And Image Recognition

The planetary gearbox is used extensively in agriculture,aerospace,and wind power because of the merits of small volume as well as high transmission ratio.However,due to its working conditions that have been high load,it is often possible to have a wheel tooth wear,breakage,tooth surface,and rotation,and wheeled cracks.Generation of these faults will have a serious impact on mechanical operation,and even a major mechanical accident.Therefore,it is especially important for the accurate and efficient fault identification of the planet gearbox.The paper regard planetary gearbox as object of study,rely on rigid-flexible coupling model and the physical HFXZ-I platform,make use of SIFT algorithm to extract the weak fault features,and make use of HIKSVM to diagnose the weak faults.The research contents of the paper is as follows:The rigid-flexible coupling dynamic model is founded by Solid Works and ADAMS.With the planetary gearbox HFXZ-I physical fault diagnosis experiment platform and in accordance with the failure generation mechanism of the planetary gearbox gear and the analysis results in the ANSYS Workbench.Solid Works is used to set up the model of normal and fault conditions in the planetary gearbox to make sure that there is no intervene.Using the ANSYS APDL software,the components that need to be arranged vibration acceleration sensor in the planetary gearbox are converted into a flexible body;With the method of adding materials,restraints and contacts to the model in ADAMS,the rigid-flexible coupling dynamic model is built.Using rigid-flexible coupling dynamic model,the measuring point is optimizated.This part set up 7 working conditions include normal,planetary wheel wear,planetary wheel crack,planetary wheel break,sun gear wear,sun gear cracks and sun gear break.Besides,with the combination between the speed analysis and spectrum analysis,the correctness of this model is proved.Considering the development of "smart components",3 test points are arranged outside and inside of the planetary gearbox housing.Then,get the vibration acceleration signal and eigenvalue at each measurement by dynamics simulation,and determine the order of measurement point.The results show that the highest priority is the center above the ring gear and above the planet carrier output shaft bearing.Finally,the paper designs an optimized structure,and its correctness is proved through simulation study.Making a research in regard to weak fault diagnosis of planetary gearbox based on image recognition.First,the original vibration acceleration signal which has been collected is transferred into a 2D color image using signal image conversion method;Besides,extract the image feature points by SIFT;Moreover,cluster the feature vector and building a Bo W model.Finally,classify it by utilizing the HIKSVM for the purpose of accomplish the planetary gearbox problem recognition.The signals of the physical experiment as well as the simulation experiment are gathered.The simulation experiment as well as the physical experiment verifies the method.In the simulation experiment,set normal working conditions and three weak failure conditions of planetary gear 1 tooth wear,planetary gear 2 tooth wear,and sun gear 1 tooth wear.In the physical experiment,set normal working conditions and four weak failure conditions of planetary wheel corrosion,planetary gear 1 tooth wear,2 tooth wear,and sun gear 1 tooth wear.In the simulation experiments,the accuracy of the method is as high as96.7%,and in the physical experiments,the accuracy is as high as 97.3%.It can be concluded that taking this way can extract plenty of signal features and does not require noise decrease on the original signal,which enhances the correctness of fault mode recognition.