Railway ballast bed stabilisation

DELA

Hello everyone,

I am trying to analyze the mechanical performance of a railway ballast bed under dynamic stabilization operation using EDEM. Can you provide me with information or the best materials on how to model the ballast bed, sleeper, and especially how I can apply the vertical downward pressure to the sleeper in order to stimulate the load acting on the sleeper?

RWood

Hi,

I'd recommend a look at the EDEM eLearning courses on Altair University:

https://learn.altair.com/course/view.php?id=147

There are a host of introductory videos showing how to set up many aspects of EDEM, including materials and applying forces to geometries. I would recommend starting there.

There are also many tutorials on the Altair community site you can go through, including one on material calibration, to help you get started in creating a suitable material:

https://community.altair.com/community?id=kb_article_view&sysparm_article=KB0037196

Richard

jpmorr

Railways embankments have quite a lot "stuff" happening and require a little bit of thinking on how best to set up. The first question you will want to be asking is about the ballast particles and how to capture the bulk behaviour of ballast, which is highly frictional with lots of interlocking. So you should decide on how to represent tis material to ensure you capture the correct frictional behaviour. Will spherical particles suffice? Or do you need to work with multi-spheres or polyhedra? And if so, what shape fidelity do you require? You should have some calibration test to compare against and lots of direct shears tests have been carried out experimentally on ballast and are a good indicator.

The sleeper is relatively easy to consider as it can essentially be considered rigid and a CAD geometry can be simply imported for that. Do note, that sleepers do actually deform slightly under the significant loads of trains and this slight deformation can lead to a slightly different stress distribution than the rigid body. There is plenty of geotechnical literature on the effect of flexible foundations if you want to go down that route eventually. The deformation could be modelled as a deformable body with EDEM using the coupling interface or some bonded particle in the shape of a sleeper. 

The question on loading depends on what you are trying to apply. A static, constant stress can easily be applied using the motion control in EDEM and should also work well for regular sine type waves for cyclic loading. If you want to apply a realistic force-time series from a real train, then that's a bit more complex and requires a proper PID type controller to ensure the dynamics are adequately applied. That would need to be done through the coupling interface.

When you put it all together you will have some system that should be able to replicate the real behaviour. Here's an example of a system we have used in the past under a real train load.