Hello,
I am currently trying to simulate a fluidized bed for 10 s. Unfortunately, the simulation keeps aborting with error messages after 5.8 s. Among them is e.g. that the time step falls below the selected minimum time step. I have already tried various things, e.g. increasing the time step factor or the spring stiffness, which seems to have a great influence in this case. I have also changed the number of cores on which I calculate, but this has not helped either.
I would be very grateful if someone could take a look at the case and maybe have an idea what settings I could change.
Thanks in advance and best regards
Fritzi
BFB_Large_Fluidizedl200.mfx (12.8 KB)
calc_collision_wall_mod.f (42.4 KB)
calc_force_dem.f (15.4 KB)
particle_input.dat (31.4 MB)
usr1.f (2.5 KB)
I briefly looked into your file. I’m mostly accustomed to the TFM model, did you try that also? It might be easier to get to work at first. Or at least start off with something simple, that also works.
One thing though - your initial volume of particles takes up almost all the fluid volume. I would feel better to reduce this to a third of the volume perhaps. Also - did you check that your number of initial particles is actually possible and does not violate what’s packing? I’ve had problems with this in DEM before, when initial conditions was packed too tight, where it will run for a few seconds, and then fail.
Thank you very much for your speedy response. As my work is actually about Coarse Graining, switching to TFM is not an option. The initial settling simulation with pure granular flow worked in order to obtain the input file for the fluidized simulation. So that should not be the problem. And the particles only take up a third of the whole cylinder. One can see this after starting the simulation. But still thanks for your ideas.
Have you tried setting your IC upwards velocity to something other than zero, perhaps closer to what you have on your BC? This would perhaps be a good idea for the fluid at least, the particles less so. When starting from zero as you do now you might maybe experience a pressure build-up as the particles settles/falls down altogether as a “piston”.
To help - do you have an image of the simulation after just a few timesteps, and what it looks like almost as it crashes?
I attached two images: one at the start and one before it crashes (at around 5.8 s). As I’m using a user defined function for the velocity, there is a start velocity of 0.1 m/s which then ramps up to a value of 0.6 m/s.
I can’t really see what’s wrong here unfortunately, but I’m doing similar troubleshooting fairly often. Have you achieved the simulation to work when using simpler inputs, say, not ramp the velocity or not using any other user functions?
My general recommendation would be to let the simulation “settle” a little longer before ramping up, and get a simpler simulation to work first. Then it is easier to see what setting that is causing the errors.
Also, are you expecting the particles to exit out the top? Why not have the cylinder longer? Does the internal pressure change a lot from start to this point?
