基于Sobol’法的轨道车辆平稳性的全局灵敏度分析

Global sensitivity analysis on ride quality of railway vehicle based on the Sobol' method

建立经典轨道车辆垂向多刚体模型,推导出模型的振动微分方程组,并采用虚拟激励法和三角级数反演理论求解轨道车辆Sperling平稳性指标。基于Sobol序列对模型11个参数进行随机采样,运用Sobol’法实现对Sperling指标的全局灵敏度分析,得到各个参数对车辆平稳性指标的一阶灵敏度及总灵敏度。最后选取车体质量和一系悬挂刚度2个参数进行优化分析。研究结果表明:平稳性指标随着速度的提高而增大,但影响平稳性的主要因素受速度改变的影响较小;车体质量、转向架质量、二系悬挂刚度和车辆定距之半对车辆平稳性影响较大;车体质量和二系悬挂阻尼与其他参数交互作用明显;提高车体质量和降低一系悬挂刚度能有效地提高平稳性。

The classical vertical multi-rigid-body model of railway vehicle was established,and the corresponding differential equations for the dynamic model were developed.The Sperling ride quality index of railway vehicle was determined by the pseudo excitation method and trigonometric series inversion theory.The eleven parameters involved in the dynamic model were sampled randomly based on the Sobol sequence.The global sensitivity analysis on the Sperling index was realized by the Sobol'method.The first-order sensitivity and the total sensitivity of each parameter to the ride quality index were obtained.Finally,the vehicle mass and the first suspension stiffness were selected to be optimized and analyzed.The results show that ride quality index increases with the increase of vehicle speed but the effect of change in speed on the main factors of the ride quality is not significant.The vehicle mess,the bogie mess,the second suspension stiffness and half of the vehicle distance play remarkable roles on the ride quality of railway vehicle.The vehicle mess and the second suspension damping have significant interactions with other parameters.Increasing the vehicle mass or decreasing the first suspension stiffness can improve the ride quality significantly.