Investigation of Form Effect on Ballast Mechanical Behavior Based on Discrete Element Modeling

As an essential component of the ballast track, the ballast layer provides functionali-ties such as drainage, load distribution, as well as strength and stability for the rail-way track. The mechanical behaviors of ballast track such as its permanent settle-ment, breakage, force propagation and void ratio are in a great extent influenced by the form distribution of ballast stones. Its reasonable design will greatly improve the mechanical behaviors, and thus prolong the maintenance cycle of ballast track, or reduce the number of ballast stones needed for construction.

This dissertation focuses on proposing optimized ballast stones in the ballast aggre-gate in regard to their geometrical forms. As the first step, a ballast random form gen-erator, which is designed to generate ballast form databases with different form dis-tributions, is proposed. 15 databases are created for further usage. Afterwards, Dis-crete Element Method (DEM) based simulations are performed to investigate the mechanical behaviors of ballast aggregates. The simulation model is established based on a box test, whose result is presented firstly. Establishment and calibration process of the model are expatiated afterwards. A parameter study regarding to cru-cial modeling parameters is also performed. Using the validated parameters and the 15 generated form databases, DEM simulation models with different form distribu-tions of ballast stones in the ballast aggregate are proposed. Simulative methods to quantify the mechanical behaviors are elaborated. Based on the obtained results, the interrelation between mechanical behaviors of ballast aggregate and the form effect on mechanical behavior of the ballast aggregates are studied. The optimized ballast aggregate is proposed based on the findings stated above.

The proposed optimized ballast aggregate is expected to be a reference for construc-tion of ballast track in real world. The modeling technic and the calibrated modeling parameters can be used fo