轮轨撞击振源参数的虚拟反演

Virtual inversion of vibration source parameters of wheel-rail impact

摘要:

为了论证利用台阵观测反演城市轨道交通中由轮轨撞击产生的振源参数的可行性,采用了虚拟反演的方法,首先预设车轮与轨道接头撞击振源参数,利用已有精确解计算得到了当列车以一定速度通过弹性半空间表面时,在振源附近布设的一密集台阵测得的竖向振动位移时程曲线,并对该位移时程曲线进行谱分析.利用功率谱作为目标函数,采用微遗传算法,反演得到了脉冲振源的强度及列车运行速度.计算结果表明,利用两个点的功率谱作目标进行振源参数反演的结果误差较小.遗传算法在振源特性研究领域的应用是可靠的,为进一步采用遗传算法反演随机振源函数的研究提供了理论支持.

Abstract:

To demonstrate the feasibility of inverting wheel-rail impact source based on observation data,the method of virtual inversion is adopted.Firstly,a setting function was given to describe the source,and the corresponding vertical displacement time histories of the observation array were obtained by means of an existing analytical solution.Spectrum analysis was performed on the displacement time histories.Then taking the power spectrum as the objective function,the intensity of the source function and the velocity of the train were inversed by means of the micro-genetic algorithm.Test results show that the error is very small by using the power spectrum of two points as the objective function to inverse the vibration source parameters.Genetic algorithm is reliable in the research of vibration source characteristic,which provides a theoretical foundation for further application of genetic algorithm in the research of the random vibration source function.