基于最小投影的离散元道砟模型的粒径识别

摘要

The shape of ballast particles and ballast gradation have an important effect on the settlement and deformation characteristics of ballast.Therefore,when it comes to studying the settlement behavior of ballast by using discrete element method,a numerical model of ballast particle is required to be constructed,to accurately reflect the ballast particle shape and to reasonably evaluate the size of ballast particles.In this paper,using the theory of convex hull in computational geometry,a random convex polyhedron was generated to simulate the shape of ballast particles.Meanwhile,based on the bound rectangle of minimum projection,a method for evaluation of grain size was proposed to analyze the grain size of particle model of ballast.The flaky and elongated ballast particles in the simulation were also explored on the basis of bounding box of the model.The results showed that the shape of ballast particle can be well simulated by the ballast particle model of random convex polyhedron.Similarly,the assessment method based on bound rectangle of minimum projection can accurately reflect the polyhedral ballast grain size,which is more accurate than the evaluation method of equivalent grain size.The flaky and elongated ballast particles can be effectively identified on the basis of bounding box of the model to control the number of flaky and elongated particles.%道砟颗粒形状和粒径级配对道砟道床沉降和变形有重要影响,因此采用离散元法研究道砟道床沉降问题时,需要建立反映道砟颗粒形状的模型,并合理评估颗粒模型的大小.本文利用计算几何凸包理论,建立了随机凸多面体,以模拟道砟颗粒的形状;提出了基于最小投影矩形包围面的粒径评估方法,对道砟颗粒模型的粒径大小进行了分析,并基于道砟颗粒模型的包围长方体,分析了道砟颗粒模型的针状和片状情况.分析结果表明:随机多面体道砟颗粒模型能够较好地模拟道砟颗粒形状;采用最小投影矩形包围面的评价方法能够真实地反映多面体道砟颗粒模型的粒径,该评价方法比等效粒径评价方法更精确;利用道砟颗粒模型包围长方体的数值判别方法,能有效识别随机道砟颗粒模型的针片状情况,控制针状颗粒和片状颗粒的数量.