Research On Fault Diagnosis Of Gearbox Based On Oil Monitoring And Grey Theory

Gearbox,as the main power transmission component of engineering machinery equipment,is prone to wear and tear because of its complex structure and poor working conditions.Severe wear of gearbox may lead to the failure of mechanical equipment,resulting in serious safety accidents and loss of shutdown maintenance.Statistics show that approximately 80%of the scrap parts in gearbox,35%of the operation failure and 38.5%of the gear failure were caused by severe wear.Therefore,the research on gearbox wear fault diagnosis has great application prospects and socio-economic significance.This study focused on the bulldozer gearbox,the wear condition of gearbox was monitored based on oil monitoring technology.Combined with grey theory,the wear failure evaluation model and location analysis model were established to realize quantitative analysis and location analysis of gearbox wear condition.The main research contents were as follows:(1)In order to effectively identify the wear faults of the gearbox,four common wear types of the bulldozer gearbox and their corresponding wear debris types were analyzed.The wear condition of bulldozer gearbox was tracked and monitored by ferrography analysis and laser particle size analysis,and the reliability of oil monitoring technology in identifying the wear fault of bulldozer gearbox was verified by oil sample analysis.It also provided a theoretical basis for the feasibility of quantitative analysis of gearbox wear faults based on fault assessment model.(2)In order to realize the quantitative analysis of gearbox wear fault,the two-factor parameters of ferrography analysis and spectral analysis of gearbox lubricants were extracted as fault characteristic indexes.Based on the validity of grey theory for fault diagnosis of small sample data,a grey target model for gearbox wear fault assessment was established.Considering that the resolution coefficients in traditional grey target models are usually determined by human experience,there are some shortcomings such as subjectivity and low universality,which will directly affect the resolution of the model.This study improved particle swarm optimization(PSO)algorithm in a non-linear way to adaptively optimize the resolution coefficients of the model and obtain the optimized grey target model.Then,the reliability of the optimization method for gearbox wear fault assessment was verified by comparing with the traditional grey target model.(3)In order to realize the wear location analysis of the gearbox,the element characteristics of wear debris in gearbox lubricating oil were analyzed by spectral analysis technology.This work took the element wear rate as the location analysis index.Based on the reliability of grey GM(0,N)model for describing the multivariable parameter relationship of small data and the multi-objective optimization function of genetic algorithm,a model of MOGA-GM(0,N)on the basis of multi-objective genetic algorithm(MOGA)and grey GM(0,N)model was proposed.With the minimum posterior difference ratio and maximum small error probability of the model as optimization objectives,study the multivariate linear relationship between target elements and related elements.According to the contribution degree of each related element to the target elements,the main wear sources of wear elements were judged.And the reliability of the location analysis method was verified by an example.