Quantitative Analysis And Application Of Ballast Paticle Morphology Based On Computer Vision

The ballasted track,whose track bed is composed of ballast particles of different sizes and shapes,has the advantages of low construction cost and convenient maintenance.What's more,the compacted particles could form a stable interlocking structure so that the track bed has good bearing capacity and drainage characteristics.Hence,the ballasted track is widely used in high-speed railway.Under the dynamic load generated by the high-speed trains,there are more and more edge abrasion and particle breakage appearing,and the particles are prone to misalignment and rearrangement due to the loss of effective occlusion between the ballast particles,which leads to accelerating development of track bed and even operational safety issues.So the particle morphology and the interlocking structure between the particles play a key role in the long-term service of the ballasted track bed,and the evolution of the particle morphology and the bearing capacity characteristics during service are worth further study.This study developed a ballast particle image analysis device for capturing with three orthogonal images of an aggregate particle,and more importantly,a software application compatible with the device,combing to the algorithm of locally weighted regression smoothing and k-fold cross-validation,was programmed to obtain the mean particle surface and the key points describing the boundary of particle,conduct the rapid analyses of particle morphological property parameters,and model the three-dimensional particle with high accuracy.Its accuracy and reliability are verified by the comparison between the manually measured results and the calculated results.And this device realizes quantitative analysis of particle morphology(angularity index,sphericity,flat and elongated ratio,convexity).The Los Angeles abrasion test and the full-scale model test of ballasted railway were carried out to quantitatively analyze the morphological characteristics of ballast particles during the degradation.By adopting this method,the change trend and evolution mechanism of morphological properties of ballast particles during this process are explored,which provides a basis for evaluating the evolution of particle morphology and bearing capacity of the ballasted track bed.In order to further demonstrate the possible application of quantitative analysis of ballast particles,this study also carried out the static triaxial loading tests of ballast under different confining pressures,analyzed the variation rules of the deviator stress,circumferential strain,volumetric deformation,and explored the influence of confining pressure on the shearing strength of ballast samples.Furthermore,a three-dimensional discrete element model of the static triaxial test based on the true shape of ballast particles was built,and the effect of actual particle morphology and interlocking on the macroscopic shearing strength properties of ballasts at the particle level are revealed based on this DEM model.