Joint Internal Bearing Failure Based On Transient Characteristics Diagnostic Method Research

In rotary machinery,about one-third of the faults come from rolling bearings.Therefore,it is necessary to study the fault diagnosis method of rolling bearings.It can find out the early faults in advance,and then repair the mechanical equipment to avoid sudden failure.There is serious damage to mechanical equipment and danger to workers.At present,many scholars at home and abroad have studied many high-speed rotating bearings for a whole week,and have proposed a large number of methods for fault diagnosis research.However,there are few studies on fault diagnosis of internal bearing joints.The object studied in this paper is the internal rolling bearing of the joint in the connecting rod machinery,which can not rotate at high speed and continuously.The fault diagnosis method based on transient characteristics is studied in this paper.The main contents are as follows.(1)The three-dimensional model of the mechanical arm was designed and the physical object was machined.The experimental system of the internal fault bearing of the joint was established,including two modules of software and hardware.The software is the control motor program.The hardware is the structure of the mechanical arm and the motor.Prepare to collect the internal bearing fault signal of the joint.(2)The dynamic model of internal fault bearing of the joint is designed.The differential equations of motion of different bearing faults are established and solved.The vibration acceleration signals of different faults are obtained.The difference between the two is obtained and the difference is found.In addition,the spectrum of a single shock signal is figured out.(3)A method for extracting transient shock characteristics of joint internal bearing faults is proposed.The GT filter bank,spectral kurtosis and Hilbert transform method are adopted.The outer ring,inner ring and rolling element are respectively replaced by multiple fault bearings.The fault impact signal is extracted,and the characteristic frequency is used to verify that the impact is generated when the rolling element passes through the bearing fault location.Fourier transform is performed on the single impact obtained by intelligent extraction,and the impact spectrum structure of different fault bearings is found.There is a certain difference,and the single impact spectrum structure of the same fault is similar.(4)In the simulation and experiment,12 characteristic parameters in the single impact spectrum of different internal faults of the joint are extracted.The different fault types are classified by PCA and SVM,and the effect is better,but the classification effect of the method on the experimental data samples.It is not ideal.(5)The method based on deep belief network is studied,which realizes the classification and classification of experimental data of internal fault bearing types of joints,and makes up for the insufficiency of classification of experimental data based on principal component analysis and support vector machine.