The Optimization Of Suspension Parameters Of Urban Rail Vehicles Based On Surrogate Model

Vehicle dynamic performance design is a key part of urban rail vehicle design,which affects the stability,safety and comfort of vehicle operation.There are a wide variety of design parameters related to the dynamic performance of the vehicle,and the range is uncertain.In the process of iterative optimization,a large number of human resources and computing resources will be occupied.Therefore,it is very necessary to combine design analysis with optimization to establish an efficient design optimization process.Based on the SIMPACK platform,a dynamic model of a metro vehicle is established,and the interface between Isight and SIMPACK is established.By analyzing and comparing the method of vehicle nonlinear critical speed acquisition based on Isight platform,it is found that dichotomy is the most economical way to collect critical speed.The series method is used to establish the process of collecting critical speed,stability index,derailment coefficient,wheel load reduction rate,lateral force of wheel axle and overturning coefficient.It is found that this process greatly saves the cost.Using the optimized design method of Latin hypercube test,the design parameters related to dynamics,including vehicle quality parameters,rotational inertia parameters,geometric parameters and suspension parameters,are tested in the design space,and the main effect analysis is carried out on the test data.It is found that the vehicle suspension system is the core system to optimize the dynamic performance index of the vehicle.The parameters of the quality design and the inertia of the body and the position of the center of gravity have great influence on the dynamic performance.The effect of primary and secondary suspension parameter on dynamic performance indexes is studied by parameter test method in a larger design space.Giving consideration to robustness and optimization performance index,the optimal solution range of primary and secondary suspension stiffness is reduced to 2~3MN/m,and the range of optimal solution of damping coefficient is reduced to 15 kN·s /m~40kN·s /m.The applicability of different surrogate model technology to the vehicle dynamics system is compared.The relationship between the suspension parameters and the dynamic response index is simulated by using the radial basis function model,and the genetic algorithm is used for multi-objective optimization to get the non inferior solution set.The non inferior solution set is brought into the dynamic performance index data acquisition process,and it is recalculated and compared with the original solution.It is found that 4 sets of solutions are better than the original ones.The 4 sets of solutions are calculated under 3 typical working conditions,and it is found that the dynamic performance is also significantly optimized.