Toward Large-Scale Simulation of Railroad Dynamics: Coupled Train–Track–Discrete Element Method Model

Abstract

The development of a large-scale high-fidelity model of train, rail, crosstie, and ballast offers a virtual laboratory for studying train–track dynamics. Currently, Train–Track (TT) models integrate the whole train and track system together, but lack explicit representation of ballast particles and simplify them as one-degree-of-freedom mass blocks only moving vertically, whereas models based on Discrete Element Method (DEM) for detailed ballast granular mechanics rarely include detailed representations of the rail and train because these multi-body systems are difficult to model within a DEM framework. To overcome these shortcomings, a large-scale TT-DEM coupled model with more than 480,000 polyhedron ballast particles was established to simulate track dynamic responses. To make this size model feasible with available computing resources, the TT and DEM models were coupled with a proportional–integral–derivative (PID) algorithm to eliminate the need for iteration within each time step. Additionally, the DEM time step was increased, cross-software communication was streamlined, and DEM data extraction was improved. Collectively, these improvements resulted in a model speed-up of about 200 times. The proposed TT-DEM model was validated by comparing predicted and field measured crosstie displacements. These comparisons showed that the TT-DEM model more closely represents the nonlinear system behavior than the conventional TT model and offers the advantage of studying the ballast at the particle level. A study of the thirty-crosstie TT-DEM ballast particle response to train track loading identified significant horizontal ballast forces that are not included in the TT model or single-crosstie TT-DEM models.