[benjamin_]

Hi everybody

I'm currently experiencing difficulties to achieve a 3D non-structured tetra simulation .

My aim is to simulate the ejection of 3 different solid parts from a cavity in a wall.

I'm using our in-house 6DOF model for each element, which are moving free in a tetrahedral domain (the 6DOF model is OK).

The begining of the simulation is quite satisfying, but after a few thousands time steps, some high and low pressure values appear in the domain , far from the bodies (so far from the expected physic changes).

After 2000 time steps theese high pressure zones starts to appear on the walls, and amplify by the time. I stopped my simulations at 5000 times steps (~0.3s) because the physic is not good anymore.


Theese "drops" are likely to appear near from the fixed walls.

I don't undestand what's happening, but it looks like something incremental for me.

Here I resume the characteristics of the simulation :

Fluent 13 3d
Pressure-based , unsteady
Tetrahedral mesh
Variable time-step, but an average very low time-step (~0.00005 s !!)
Fluid : water
Boudary conditions :
velocity inlet on front face and flow inlet
pressure outlet on flow outlet
k-epsilon turb. model

Dynamic mesh smoothing, remeshing / Local Cell
Size remeshing interval : 1
Max cell skewness : 0.9

Each body : UDF 6DOF Rigid Body
Fixed walls (wall & cavity) : Dyn. Mesh Stationary

P-V coupling : SIMPLE

Spat. Discretization :
Gradient : Green-Gauss Cell Based
P : 2nd order
Momentum : 2nd order
k : 2nd order
epsilon : 2nd order

Time : 1st order implicit

Under-Relax. :
P : 0.25
Density : 1
Body force, Mom., k , eps, turb. visc : 0.5

I don't see any other relevant characteristics...


Did some of you has ever experienced similar issues ?

Thanks by advance

Benjamin