Research on vibration response optimization of the urban rail vehicle traction gear transmission system based on combined surrogate model
YU Wennian
DAN Jiaguo
CHEN Kai
LIU Yueqiu
WU Jing
WEI Zhicheng
CHEN Xiaohui
Abstract：[Objective]Aiming at the excessive vibration of the traction gear transmission system of urban rail vehicles and the problems of long operation time and slow convergence speed in optimization design,a multi-objective optimization method based on a combined surrogate model was proposed to achieve precise improvement of system vibration characteristics and enhancement of optimization efficiency.[Methods]Firstly,a lumped parameter method was used to establish the electromechanical coupling dynamic model of the system,and vibration response data were obtained to provide a foundation for subsequent optimization.Secondly,with structural parameters as inputs and vibration responses as outputs,a combined surrogate model integrating Kriging,radial basis function method(RBF)and polynomial response surface was constructed to quickly map the relation between parameters and vibration.Finally,taking bearing support stiffness and damping as optimization variables,the Pareto solution set was solved based on the non-dominated sorting genetic algorithm Ⅱ(NSGA-Ⅱ)to minimize vibration acceleration.[Results]After optimization,the root mean square value of the system's vibration acceleration is reduced by an average of 24%,with the root mean square value of the vertical vibration acceleration of the large gear decreasing by 38.4%;the optimization efficiency is increased by 97%,and the calculation time is shortened from 16 days to 10 min.The vibration suppression effect was stable at different rotational speeds,and the gear bearing capacity is not affected,providing a reference for the optimal design of the traction transmission system of urban rail vehicles.
Keywords：Urban rail vehicleElectromechanical couplingVibration responseSurrogate modelMulti-objective optimization
Publication Date：2026-03-15
Online Publishing Date：2026-09-12(First online date of this platform, not the publication date of the document)
Pages：10( 70-79 )