The great development of high speed rails all around the world has arisen a great interest in the optimization of the superstructure, which has to be adpated so to withstand a variety of extreme loads. The effect of these loads in the mid and long term is not well known.

Railway engineering is a special field inside civil engineering, because it is susceptible to optimization, due to the number of units of some of its elements (sleepers, rail) needed per kilometer of railway line. An improvement, even a small one, in one of these elements would mean a great advantage for the infrastructure as a whole.

Ballast response is one of the outstanding features of these structures. It is a granular, angular material, having very specific characteristics (granulometry, strength, etc.). Although this material has traditionally been used with good performance, the new loads derived from the high speed rails have caused some problems, such as ballast projection (the aerodynamics produced by the pass of the train at high speed provokes the ejection of some of the ballast particles, which may hit the train underbody and its elements, such as the brakes, the wheels, etc). This has open the discussion on the convenience of using slab track instead of ballasted track.

The possibility of simulating the behaviour of the ballast under the loads of the train, as well as the dynamic load of the latter, will constitute a great advance in this field, because it will allow to test innovative solutions before undertaking in situ tests and putting them into practice.

In order to get this goal, a numerical code based on the Dicrete Element Method is planned to be developed, taking into account that this method has showed its ability to model granular materials in dynamic problems. However, the modeling of the rail-sleeper-ballast as a whole requires a deep research on some of the issues involved, such as:

1.   Those which characterize the ballast: a) friction coefficients for the sleeper-ballast and ballast-base contacts, b) internal friction coefficient for the ballast and c) fatigue strength

2.    Those inherent to the model: a) mesh generation parameters and b) damping parameters

3.    Those related to the loads: magnitude, direction and duration of the loads.

Further issues which have showed to be important in the phenomenon and thus will be taken into account are: ballast consolidation, train velocity, temperature and traffic.

Figure 1: Numerical simulations of internal friction coefficient tests for granular material.

  
Figure 2: Numerical simulation of the behavior of the ballast under lateral loads.

Figure 3: Numerical simulations of ballast projection.


Figure 4: Numerical simulation of ballast behaviour under triaxial loading.