Research On The Fault Classification Of Gear Drive System Based On Torsional Vibration Analysis Method

With the rapid development of industrial automation in our country,mechanical equipment is developing in the direction of high efficiency,automation and intelligentize,rotating machinery has been widely used in the fields of aeronautics,astronautics,metallurgy,petrochemical engineering and construction machinery.Thus,high speed,large load and other mal-conditions can lead to its damage probability.Gearbox is an important transmission component in mechanical equipment,which is prone to failure due to its complex structure and bad working environment,and the gearbox faults can cause equipment health daterioration and even breakdown.Any rotating machinery gear can also face the same problems.Thus,it is great significance to improve the fault diagnosis ability and the level of fault identification of the rotating machinery system.Fault diagnosis based on torsional vibration analysis of gear transmission system can directly reflect the normal status and abnormal status.This paper studies the fessibility of fault identification of gear transmission component system based on the torsional vibration analysis method.The details are as follows:Firstly,the dynamic analysis of gear transmission system is carried out,which is based on three kinds of gear fault model,including the vibration of the gear system and the mathematical model analysis of the torsional vibration on shaft.Besises,describes some different types of typical gear failures,and summarizes the signal characteristics of typical gear fault.Secondly,this paper presents a new method of gear transmission sysytem fault detection based on the permanent magnetic acceleration sensor,and then the working principle of permanent magnetic rotary acceleration sensor and experiment scheme are described in detail.In addition,respectively set up a “magnetic levitation” experiment platform,two axis and three axis gear transmission system,and measuring the torsional signal on the shaft and transverse vibration signal on the experimental platform of the gear transmission system in different running state.Thirdly,the spectral analysis is performed between the torsional signal of two shaft and the transverse vibration signal,and then compare the results of two sets of different signals.The two groups of signals are decomposed by the wavelet packet multilayer,and the energy of each frequency node is extracted as the feature vector.And the two groups of signals are classified by Support Vector Machine(SVM)respectively,comparing the final fault recognition results.Finally,the experimental results shows that the gear fault identification method based on the torsional vibration signal on the shaft is superior than the traditional gear fault diagnosis method based on the gear box vibration noise signal,and it verifies that the method of torsional vibration signal detection based on rotation(angular)acceleration sensor is verified;the detection effect of torsional vibration signal on fault shaft is better than that effect of torsional vibration signal on not fault shaft.The experiment of three axis gear system verifies that the detection effect on the output axis is better than that on the input axis.