Feature Extraction And Diagnosis Of Early Crack Fault In Gear-Rotor Drive System

The gear-rotor system is one of the most common mechanical systems,has become the most important link in industrial production.The gears that are the most important part of the system,its health may directly affect whether the whole system can operate.As the most common form of gear failure,the early cracks are difficult to identify and the impact signal is weak.If the early crack phenomenon in the gear-rotor system can be effectively identified in the monitoring link of the equipment,the failure of the gear system can be prevented in advance,this paper mainly studiesThe response of gear pair system with root crack is analyzed as follows:Modeling analysis of involute gear,the meshing stiffness of gears and the meshing stiffness of cracked gears are analyzed based on the energy method;the transmission system of the healthy gear was modeled and the vibration characteristic response of the cracked gear system was analyzed.Root crack fault diagnosis based on EMD and BP neural network:According to the fundamental theory of dynamics,for the root crack of 4mm,gear simulation processing and EMD decomposition under the operating state of 1000r/min rotation speed.In terms of experiment,the signals of different crack degrees were collected.And it is processed by combining EMD with BP neural network.The difference of gear signal under different crack degree is analyzedIdentification of root crack damage degree based on nuclear principal component analysis:Dynamic simulation was carried out for three conditions:healthy gear,1mm root crack,and 3 mm root crack,PC A and GRA were used to calculate the correlation degree between the sample to be tested and multiple comparison samples,and the correlation degree was combined with the experiment.Fault diagnosis of multi-crack gear pair system based on EMD and choi-williams distribution:The nonlinear dynamic model of gear-rotor-bearing drive system with 16 degrees of freedom is established.Based on the energy method in chapter 2,six kinds of time varying meshing stiffness of the gear,such as continuous double cracks,double cracks with one gear tooth spaced,double cracks with two gear teeth spaced,double cracks with five gear teeth spaced,and double cracks with eight gear teeth spaced,were calculated and compared.Based on the EMD theory in chapter 3,the empirical mode decomposition of the vibration displacement signal in the direction of torsion is carried out,the IMF component is expressed by CWD,and the spectrum diagram of STFT is compared with that of CWD.