Learning Forum

[kiki1]

 Transient calculation can be calculated for a long time after the start, but it suddenly diverges in the middle (about 0.77s)

Does anyone know why? Thank you.

The geometric model, mesh, solution setting and results before divergence of the problem are as follows：

Calculation of water vapor condensation in 100mm * 20mm space by axisymmetry

The two condensing surfaces a and B have a temperature of 273k, the C surface is adiabatic, and the D surface is a symmetrical boundary

The volume fraction of water vapor in the initial state calculation space is 1, the temperature is 380K and the velocity is 0

The reference calculation example can complete the calculation (5S) under the same setting. When I calculate to about 0.77s, the continuity residual suddenly diverges, I conducted two calculations, and the results are the same,

The 0.7s residuals and results are as follows:

[Amine Ben Hadj Ali]

Where does it start condensing? Wall or bulk? Which model is used for phase change: yours or user defined?

[kiki1]

Condensation starts from the left and right cold wall surfaces at the same time;
The Lee model of the system is used to simulate condensation.

[Amine Ben Hadj Ali]

The Lee Model itself is not completely correct for this kind of investigation. Either one needs to refine and refine or to have a sub-model which deals with wall condensation. I hope you are doing auto-saving to have a recovery restart file from which you can start again and perhaps uses a smaller time step size and to check where it starts diverging.
In general it is not recommended to use Explicit VOF when having mass transfer.

[kiki1]

Thank you for your reply and tips. Maybe I should write a UDF file to limit condensation near the wall, but why do you think there is a sudden divergence in transient calculation