Hello Everybody,
i'm pretty new to CFD. But i have created a DEM-CFD coupled model, which run pretty good until i changed the boundary conditions. The geometry is quite simple and looks like a big hopper with a block of particles on top. After a initialisation time a wall is deleted and all particles can fall to the ground. The hopper has an inlet/outlet on the top and a second on the bottom. The side walls are defined as walls. The mesh is created with blockMesh.
If the bottom the replaced by a wall boundary condition, i can't reach convergence. The last result file shows big differences in the pressure field. (What i want to see is a higher pressure on the bottom and a fluid stream through the particles)
The following output is written:
No Iterations 3
time step continuity errors : sum local = 1.99655e+54, global = 6.12444e+42, cumulative = 6.12444e+42
suppressing ddt(voidfraction)
DICPCG: Solving for p, Initial residual = 0.953441, Final residual = 5.42579e-07, No Iterations 29
time step continuity errors : sum local = 8.62479e+55, global = -3.94424e+47, cumulative = -3.94418e+47
ExecutionTime = 182.36 s ClockTime = 201 s
I tried different parameters for:
1. time step of the coupling (timestep DEM 1e5 / CFD 1e-3 )
2. solver tolerance (from 1e-5 to 1e-8)
3. net size (de- and increasing)
4. no turbulence model (in the moment i'm using "laminar")
The only change i could see was a bit bigger simulation time with bigger net elements. I'm using the cfdemSolverPiso.
Has anybody a idea?
Greetings,
Hendrik