[mwe]

Hello,

I posted here several weeks ago, regarding this problem and have tried several ideas to fix the problem since, none of which solved the issue. Maybe someone can help me, with more details of my problem available:

Basically I have a pipe with multiple layers:

Cooling Fluid -> Heat Transfer Pipe -> Phase Change Material -> Plastics tube

The Plastics tube is supposed to keep the PCM attached tightly to the Pipe. Since the PCM changes density on a relevant scale during phase change, the plastics tube expands a little when the PCM melts (therefore the Tube and the PCM need to be morphed). And that's what I'm trying to simulate (since movement within the PCM due to changes in volume/density is relevant for the actual heat transfer).

I have the folling initial conditions: Everything is at 20°C and then heat is transfered into the pipe.

However: When I run the simulation it seems to converge for about 10 to 15 iterations and then the stress solver diverges and the force residuals go up until some cell gets a negative volume and the simulation is stopped.

So this is what I tried up to this point:
- solving stress first (for some reason star cmm+ would not let me do that, eg. I freezed all other solvers)
- solving everything but stress first (it still diverged after I switched back on the stress solver & morpher)
- replacing the PCM with an ideal gas to see if my EOS for a compressible fluid was faulty (didn't work)
- morphing at inner iterations (didn't work)

I am using the following physics models:

PCM: VOF/Eulerian Multiphase/Segregated Temperature/Segregated Flow
Cooling Fluid: I removed that for speed and set a heat source as boundary condition
Solids: Solid Stress/Solid Energy/Constant Density

This is the Error Message I get:

Warning: 1 cell in region copper_tube has zero or negative volume. This may be due to an invalid imported mesh or to errors initializing an interface.
Repair the mesh before proceeding.

However, this doesn't make any sense, since this is also displayed when the morph-solver is frozen and is not displayed when I use the same mesh in a model without any stress simulation.

I'd be glad for any ideas/advice on this topic, since I have been working on this for several days and my attempts up til now did not achieve any success. Thank you!

[me3840]

Why do the forces go up?

[ping]

i would also be simplifying the physics in all materials to the most basic ie no phase change no expansion eg with water as fluids and not motion or morpher and get that going then step by step add required properties and features

also check your time step since it might need to be much smaller

[mwe]

Thanks for your replies. Ping, I will try your approach with future problems. Sounds quite reasonable

Oh I did not mean the forces going up but the force residuals.

It seems creating a higher quality thin layer mesh with twice the amount of cells did solve the problem for now.