Hi all,
This is an issue I have had before & thought I had solved it, but have encountered similar problems with another model.
I am attempting to simulate particles being thrown outwards by an impeller, and have previously had issues with particles going straight through the impeller mesh without contacts being detected. On earlier models, changing the pair-style to Hertz (no cohesion) instead of Hooke, and also using neigh_modify every 1 delay 0 seemed to do the trick with the particles interacting with the impeller. Interactions with outer particles after leaving the impeller also seemed to work well, even though solution time is quite long.
Particle velocity after leaving the impeller can be around 80m/s.
However, I am now trying to model collisions with a second (fixed) mesh, and the particles just go straight through - especially if the faces of the mesh are relatively straight. If I add some curvature to the faces, the result is slightly improved but more often than not the particles will go straight through the front face, interact with internal faces inside the mesh and then re-emerge out of one of the front faces at a much slower velocity. The node spacing on the mesh is less than or equal to the particle size.
I have tried reducing neighbor skin distance to 0.0001 metres and also the timestep down to 1e-8, but these don't seem to lead to any improvement.
Am I missing something here?