Research On Suspension Parameters And Dynamic Performance Optimization Of Straddle Monorail Vehicle

With the development of urban rail transit,straddle monorail vehicles are selected by more and more urban rail transit operators.In the case of a large number of straddle monorail vehicles put into operation,exploring the relationship between vehicle design parameters and dynamic performance can help the design unit to optimize the dynamic performance and improve the operation quality of urban rail transit.In this paper,the modeling method and basic process of multi-body system dynamics are analyzed from the general calculation process of multi-body system dynamics,and the commonly used computational solvers involved in the research are introduced.Then,based on the simplified model,using the original design parameters such as the inertia and mass of the vehicle structural components,the UM simulation software and the Ansys finite element software were used to establish the straddle monorail vehicle model and the track beam finite element model.For the evaluation of vehicle dynamics performance,the evaluation method of railway vehicle dynamics and the dynamic evaluation method of straddle monorail vehicle summarized by experts and scholars are used for reference.The safety index is judged from the revised overturning coefficient,the radial force of running wheel,the acceleration of car body and the radial force of guide wheel,while the stability index is analyzed from the aspects of Sperling stability,comfort and operation quality.The rationality of the model is verified by comparing the measured data of Southwest Jiaotong University from the aspects of track beam vibration and car body vibration.Then,using the UM-Isight simulation joint platform,the dynamic performance of the original design suspension parameters under two different load conditions of AW1 and AW3 is compared and analyzed.It is found that the AW3 condition has a relatively large influence on the dynamic performance,which needs to be emphatically analyzed.Then,the variable parameter test of the primary and secondary suspension parameters of the vehicle under AW3 condition is carried out,and the sensitivity of the secondary and primary suspension parameters to the dynamic response is studied.It is found that the comfort index,the lateral stability index of the front measuring point,the vertical stability index of the rear measuring point,the overturning coefficient of the first stable wheel and the vertical / lateral vibration acceleration of the vehicle body have more obvious influence on the change of the suspension parameters,which needs to be optimized.The lateral stability index of the rear end measuring point,the vertical stability index of the front end measuring point,the wheel overturning coefficient and other evaluation values change more slowly,and the sensitivity to the influence of the change of the suspension parameters is small.According to the conclusion of sensitivity analysis,six indexes which have great influence on dynamic performance are selected as optimization objectives.Six suspension parameters are used as design variables.Robust optimization method and multiobjective optimization method based on surrogate model are used to optimize the design parameters and dynamic response of vehicle suspension.According to the theoretical knowledge of robust optimization design,the larger the signal-to-noise ratio is,the more robust the design is.Therefore,the robust optimization needs to focus on evaluating the relationship between the signal factor and the noise factor.In this paper,the robust optimization of the dynamic performance of the straddle monorail is carried out.Three uncontrollable factors that may occur in the operation of 5 levels and 125 groups are selected as noise factors,and 20 sets of suspension parameters are selected as controllable factors.The robust optimization method is used to form the orthogonal table of experimental design for optimization.Then,2500 calculations are carried out for 6 optimization objectives,and 6 sets of suspension parameters with the largest signal-to-noise ratio are obtained.The results of each parameter combination are verified respectively,and the optimization rate of the robust optimization solution is obtained.Among them,the optimization rates of vertical stability,vertical vibration acceleration of vehicle center of mass and comfort are about72.33 %,78.41 % and 53.22 %.The maximum optimization rates of the lateral vibration acceleration of the vehicle center of mass and the overturning coefficient of the stabilizing wheel are about 0.98 % and 7.06 % respectively.The maximum optimization rate of lateral stability is about 0.98 %.Finally,six groups of optimization solutions are verified,and the optimization effect is : combination five = combination six > combination three >combination one > combination two > original parameters > combination four.However,since the robustness optimization is calculated by the single-objective optimization method,the consistency of the optimization results is poor.It is necessary to verify all the optimization solution combinations in order to eliminate the solutions that violate the design constraints.Optimization based on surrogate model can obtain the approximate relationship between variables and responses by constructing fitting formulas.When the accuracy of the approximate model meets the analysis requirements,the constructed model approximation can be used to replace the simulation program for multi-objective optimization analysis,so as to obtain a compromise Pareto design solution set that meets the requirements.In this paper,based on the multi-objective optimization of the surrogate model,the ± 50 % design value of the original suspension parameters is selected as the variable,By using the optimal Latin hypercube method,2500 groups of training sample points of suspension parameters-dynamic performance response values are generated.Based on the training sample points,the orthogonal polynomial surrogate model and the response surface surrogate model are obtained by fitting,and the model error verification is carried out.It is found that the error index of the response surface model is about 0.60890,and the error index of the orthogonal polynomial model is about 0.68028.Two surrogate models can be used to replace the simulation program for optimization analysis.Finally,the NSGA-II algorithm is used to optimize the two surrogate models.Based on the optimization of orthogonal polynomial surrogate model,a total of 1449 sets of solutions are obtained.The convergence optimization solution of the optimization algorithm is checked,and 10 sets of optimal Pareto solutions are obtained.Among the optimization rates of the 10 sets of solutions,the optimization rates of the vertical vibration acceleration,vertical stability and comfort of the vehicle center of mass are larger,and the maximum is about 80.97 %,72.19 % and 57.15 %,respectively.The maximum optimization rates of the lateral vibration acceleration and lateral stability of the vehicle center of mass are about 9.04 % and 0.78 % respectively.The optimization rate of the stability wheel overturning coefficient is negative,and the optimization rate is about-9.60 %,but all meet the requirements of the evaluation value.Based on the optimization of the response surface surrogate model,a total of 1959 sets of solutions that meet the requirements are obtained.The convergence optimization solutions of the optimization algorithm are checked,and 6 sets of optimal Pareto solutions are obtained.Among the optimization rates of the 6 sets of solutions,the optimization rates of the vertical vibration acceleration,vertical stability and comfort of the vehicle center of mass are also large,and the maximum values are about 93.30 %,73.90 % and 57.59 %,respectively.The optimization rate is higher than that of the orthogonal polynomial surrogate model.The maximum optimization rates of the lateral vibration acceleration and lateral stability of the vehicle center of mass are about 8.08 % and 0.90 %respectively.The optimization effect of the overturning coefficient of the stabilizing wheel is the same as that of the orthogonal polynomial surrogate model.The optimization rate is about-11.63 %,and the optimization result is slightly worse.Finally,the optimal solutions obtained by robust optimization and multi-objective optimization are compared and analyzed.It is found that except for the combination of four solutions that violate the constraints of robust optimization,the evaluation response value intervals obtained by multi-objective optimization and robust optimization are not much different,and the optimization effect is equivalent.The response surface model optimization effect is better than the orthogonal polynomial model,and the multi-objective optimization method is slightly better than the robust optimization method.Robust optimization is a combination of optimization solution parameters based on each single-objective optimization problem,and the optimization results between different optimization solutions are slightly discrete.All the optimized solutions need to be further verified separately to eliminate the solutions that do not meet the requirements.The research cost of multi-objective optimization is slightly higher.The optimization rate obtained by the multi-objective optimization of suspension parameters based on the surrogate model is slightly higher than that of the robust optimization.The optimization results of the optimization solution are balanced and consistent.In the study of practical problems,the multi-objective optimization method can be preferred according to the needs.