Font Size: a A A

Research On Inspection Method Of Gearbox Assembly Quality Based On Binary Signal

Posted on:2022-03-02Degree:MasterType:Thesis
Country:ChinaCandidate:R P CaoFull Text:PDF
GTID:2492306335489914Subject:Master of Engineering (Mechanical Engineering Field)
Abstract/Summary:PDF Full Text Request
Gearbox as one of the main transmission parts of the car,its health state directly affects the overall running state of the car.Therefore,it is essential to inspect the assembly quality of the gearbox before it leaves the factory.In this paper,an M type gearbox is studied,and an on-line testing system of gearbox assembly quality is developed.The main research contents of this paper are as follows:First of all,describes the selected topic background and significance of this topic,to understand the current development situation of gearbox fault diagnosis technology,according to the research object,starting from the existing mode of M type gearbox test,chose to assembly quality better ability to characterize acoustic pressure signal and vibration signals for detection,design and build the corresponding binary signal acquisition platform,set up a database of different gearbox assembly quality.Secondly,for the binary signal of gearbox,this paper gives some signal processing methods.In order to prevent the influence of interference signals in binary signals on the detection of gearbox assembly quality,a denoising method combining EMD and moving average filter was designed by comparing the current common signal denoising methods,and the denoising effect of this denoising method was verified by using simulation signals.Then,the assembly quality detection model of M-type gearbox is established by deep learning.Taking the processed binary signals as input,the current popular convolutional neural network is trained,and the assembly quality detection model of M-type gearbox is established.By optimizing the depth and main parameters of the model,the model with the best recognition effect is obtained.Through comparative experiments,it is proved that the method of detecting binary signal is better than single signal.In order to enhance the generalization ability of the model and the traceability ability of assembly nonconformity,a feedback training model was adopted to update the assembly quality testing model of Mtype transmission.The best,designed and developed an M type gearbox assembly quality online testing system.According to the requirements of cooperative enterprises for gearbox assembly quality inspection,an online inspection system for M-type gearbox assembly quality was designed and developed using Lab VIEW.The feasibility of the system was verified through field test.
Keywords/Search Tags:Gearbox, Sound pressure, Vibration, Denoising, Convolutional neural network, Quality inspection
PDF Full Text Request
Related items