Hi,
I am working with the Edinburgh (EEPA) model for my simulations, but I have encountered a problem having to do with parallel computing. My material bed is first begin consolidated by a flat plate up to a certain pressure and then a blade is moving horizontally through it. However, once my blade hits the material bed, something weird is happening at the borders between the CPU domains. The particles are spewed upwards, and it seems like the particles 'forget' that they were plastically deformed and return to their relaxed state. This is emphasized by a drop in horizontal force on my blade which goes down to the unconsolidated level.
Has anyone encountered such a problem before?
Although I guess this only happens for this contact model, which is not widely used by LIGGGHTS users...
I am able to fix it by using a single core for the simulation, but as this quadruples my computational time it would be great if there is another solution. Lowering my time step did not have an effect on this problem (base simulation is ran at 10% Rayleigh time, and lowered it to 5%)
| Attachment | Size |
|---|---|
 particles colored by velocity, showing particles moving upwards at borders | 98.18 KB |
