[stefanke]

I have implemented the layer addition/removal feature into the existing engine solver (engineFoam/dieselEngineFoam). The layer addition/removal itself (pure mesh motion) works fine. I have to stress that my faceZone is a 4 Layers away from the piston (so I do not use the piston surface).

A strange thing happens in the expansion stroke:

Crank angle = 30 CA-deg
Mean u':1.67458

-> switching from deformation mode into layering mode

Crank angle = 30.0099 CA-deg
Mean u':5.36223

Crank angle = 30.0103 CA-deg
Mean u':11.4833

Crank angle = 30.0107 CA-deg
Mean u':32.7286
.
.
.
.

-> solver blows up

Why does u' increase? Are there any special things to know?

A helping hand is very appreciated

[stefanke]

some additional information:

the turbulence instensity increases unphysically (see my last posting) as soon as the first layer is added. So something goes wrong in the layer addition and that's causing my solver to blow up!

In the compression stroke (layer removal) all works well.

Any ideas?

[hjasak]

A comment (this is known stuff): if you are throwing away the layer next to the wall, where wall functions operate, you keep throwing away your boundary layer. Consider removing a cell layer away from the wall.

Hrv

[stefanke]

Thanks Hrv for you respond.

As I have said before I do not use a faceZone which corresponds to a wall boundary (e.g. piston surface). I have defined my faceZone on top of the static part (piston+3 layers) si the cell layer is away from a wall. Ok there are some cells in the layer touching liner but this should be ok. Any more ideas?