基于模式识别技术的铁路道砟模型库建立及离散元模拟

Establishment of Railway Ballast Model Base and DEM Simulation Based on Pattern Recognition Technology

颗粒形状是影响铁路道砟力学特性的重要因素,考虑颗粒形状是进行真实铁路道砟力学特性数值模拟研究的首要问题。本文基于模式识别技术,开发了一种颗粒提取技术,进而能够从复杂的原始道砟图像中自动识别出道砟轮廓。该技术为道砟颗粒的形状评估奠定了基础。在对道砟颗粒进行统计形状分析的基础上,建立了道砟颗粒的几何形状库。最后,基于所提出的道砟颗粒形状库和离散元方法,对真实形状的道砟颗粒进行了双轴压缩试验,研究了道砟颗粒细长度对其力学特性的影响。

Particle shape is an important factor affecting the mechanical properties of railway ballast.Considering the particle shape is the primary problem in the numerical simulation of the mechanical properties of railway ballast.Based on pattern recognition technology,this paper developed a particle extraction technology,which can automatically identify the ballast contour from the complex original ballast image.The technology laid a foundation for the shape evaluation of ballast particles.Based on the statistical shape analysis of ballast particles,a geometric library of ballast particles was established.Finally,based on the proposed shape library of ballast particles and discrete element method,biaxial compression tests were carried out on real shaped ballast particles,and the influence of fine length of ballast particles on its mechanical properties was studied.