Integrated computational and experimental framework for the assessment of railroad ballast life-cycle behavior

ballast;discrete element modeling;triaxial test;filed test;strength;permanent deformation;fouling;degradation

Qian, Yu

关键词: ballast;    discrete element modeling;    triaxial test;    filed test;    strength;    permanent deformation;    fouling;    degradation;   

Others  :  https://www.ideals.illinois.edu/bitstream/handle/2142/72751/Yu_Qian.pdf?sequence=1&isAllowed=y

美国|英语

来源: The Illinois Digital Environment for Access to Learning and Scholarship

【 摘 要 】

Ballast, typically comprising large sized aggregate particles with uniform gradation, is an essential layer in the railroad track substructure. Functions of ballast include facilitating load distribution and drainage, maintaining track geometry and track stability, and providing track resilience and noise absorption. Throughout its service life, ballast goes through changes in gradation and aggregate particle shape properties. These changes affect the particulate nature of ballast layer behavior, which has not been thoroughly understood from laboratory experiments and field performance. Moreover, numerical simulations that treat ballast as a continuum layer may often fail to address individual particle interactions at the micro-scale, which is a key perspective in modeling more realistically the behavior of any aggregate layer assembly. The main focus in this research effort has been to develop an integrated computational-experimental framework to evaluate ballast life-cycle behavior. Laboratory experiments and numerical simulations based on the Discrete Element Method (DEM) were performed to better understand micro-mechanical interactions affecting complex ballast assembly behavior and to provide quantitative evaluations of field response and performance under dynamic train loading. Starting from aggregate particle scale, grain size distributions and particle morphological properties of clean and fouled ballast materials were analyzed in the laboratory in order to provide input properties for the DEM numerical simulations. Moving to the laboratory scale, large-scale triaxial tests were performed in the laboratory under controlled monotonic and repeated loading conditions for shear strength and permanent deformation characteristics, respectively. The experimental studies indicated that the shearing rate had no significant influence on the results of the ballast triaxial tests. Compared to clean ballast, fouled ballast samples in dry condition had higher permanent deformation under repeated loading due to particle degradation but not necessarily lower strength than clean ballast. The laboratory experiments were simulated using the DEM platform, which integrated aggregate image analysis technology into numerical modeling with the ability to generate particles in the DEM simulations for the same grain size distributions and morphological properties of ballast particles used in the laboratory tests. The DEM simulations of shear strength and permanent deformation tests could capture complex ballast behavior reasonably accurately. With the use of a newly introduced “incremental displacement” shearing method, numerical simulations could be completed with less iteration or computational expense. Using the integrated computational-experimental framework, effects of geogrid reinforcement and presence of moisture on ballast at different stages of its life-cycle have also been investigated. Furthermore, full validation of the framework approach was accomplished through field testing efforts; ballast settlement trends computed from the full-scale track numerical simulations were compared with the actual ballast settlements measured in the Transportation Technology Center (TTC) High Tonnage Loop (HTL) field tests. An application of this integrated computational-experimental framework was the design of a field “settlement ramp” to reduce differential settlement in a track transition zone. The measured settlement trends of the engineered bridge approach at TTC HTL were reasonably close to those predicted by the DEM model.