[Alicia C.]

I am currently doing a 2D simulationg of a curtain of hot falling particles. I am running a steady-state simulation using DPM with two-way coupling. The turbulence model is realizable k-epsilon. The particles fall (surface injection type) from a hopper and hit a horizontal surface (where they disappear, trap condition). Due to the high temperature of the particles, the hot air should rise by natural convection; however, the upward air velocity becomes very high, even when the air is already cold, which doesn't make sense. I think that for some reason it's converging, but the results don't make any sense. The particles have a diameter of 350 μm.

I have performed the simulation with different enclosure sizes, large enough to avoid any suction effect, but this issue still persists. I have also run the same simulation in 3D, and it converges perfectly and gives very reasonable results. I’ve tried different mesh types, relaxation factors... I don't know what else to try at this point. Any input is welcome. Thank you.