Research On Wheel Set Wear Prediction And Lathing Strategy Optimization Of CRH6A EMU

Wheel pairs of EMU（Electrical Multiple Unit）are the key to smooth operation of EMU.When high-speed EMU runs on rails,wheel pair wear is accelerated due to continuous impact and friction between wheel pair and rails.Wheel pair wear will directly affect the wheel-rail contact relationship between wheel-pair tread and rail,thus affecting driving safety.Railway departments must formulate reasonable wheel-to-wheel repair strategies to carry out wheel-to-wheel repair so as to ensure a good wheel-rail contact relationship.At present,the railway department adopts the planned maintenance mode for EMU wheel pairs,which will make the maintenance cycle unreasonable.When the overrepair or breakdown occurs,resources will be wasted and there will be some potential safety hazards,causing the maintenance cost to be too high.By forecasting the wear of the wheel set,the maintenance personnel can timely repair the wheel set when the geometric parameters of the wheel set are near the limit of wear,so as to ensure the geometric dimensions of the wheel set are in a safe range of use.In addition,based on the predicted wear results and actual measured parameters,a reasonable repair plan is formulated to ensure the safe operation of EMU,improve the comfort of operation and save maintenance costs.In this paper,the wheel pair wear and haircut schemes of CRH6 A electric vehicle are studied and discussed.The main work is as follows:（1）Extensively consult the relevant literature of EMU wheel set wear,maintenance scheme and lathing strategy,understand the research background and development status,collect the geometric dimension data,equal taper,wheel set polygon and lathing data of crh6 a vehicle on a special line,and analyze the change law of wheel set wear,equal taper and polygon.（2）For the wear prediction of wheel pairs,a prediction model based on Gradient Lifting Decision Tree（GBDT）is proposed,which combines decision tree with Boosting algorithm of machine learning to solve the complexity of decision tree and avoid over-fitting.Through pre-processing of wheel pair data,2500 sets of sample data are obtained,characteristic quantities of wheel pair diameter and rim thickness are constructed,and basic parameters of GBDT prediction model are obtained through network training.Compared with the experimental results,the predictive model can forecast the wheel pair wear more accurately and can predict the change of wheel pair wear for up to one year,providing guidance for the transition from planned maintenance to condition-based maintenance.（3）CRH6A EMU is upholstery scheme is studied from the aspects of wheel pair wear,equivalent taper,polygon effect,etc.A wheel-to-wheel grouping upholstery scheme is put forward in the upholstery cycle of wheel pair,i.e.150,000 km upholstery cycle for 2 head trucks and 300,000 km upholstery cycle for 6 mid-trucks.Compared with the traditional225,000 km upholstery cycle,it guarantees the safe operation of motor vehicle groups.The service life of the wheel pair of EMU is extended.At the same time,it is proposed that the equivalent taper of wheel pair is safer between 0.035-0.06 when scheduled maintenance and case-by-case maintenance are carried out.The optimization of upholstery cycle and the parameters of equivalent taper have been applied in practical work with good results.