[p1nkman]

Hi everyone, Im trying to simulate a solar load model. Currently using a k-epsilon Realizable model and the max skewness of the geometry is 0.84.

However, I have been getting divergence in k before starting the iterations. Even if I set all the URFs to 0.1 it is still giving me divergence in k before even the iteration starts.

Anyone has any idea? Could this be a computer problem? Really need to fix this. Any help is highly appreciated. Thanks!!

[Светлана]

Switch off the solar load and check whether the simplified version converges. You can simplify it even further, disable turbulence, simplify fluid properties, decrease the time step or set it to adaptive, change to steady flow (sometimes this gets solved even if the flow is transient).

Please tell whether any of these simplifications, particularly when applied simultaneously, makes the solution converge.

[p1nkman]

Thank you for the quick response, Светлана.

While I checked the skewness forgot about orthogonal quality. Turns the orthogonal quality of the mesh was really poor, improving that region has solved the problem.

However I have been using steady time step until now with reduced URFs which has been giving me a poor continuity residual.

Do you reckon I should set the URFs to default and try with transient with varying time steps until I get better residuals? I have set the URF's to the following:

Pressure - 0.3 (Default)
Density - 0.3 (Default)
Body Forces - 1 (Default)
Momentum - 0.7
Turbulent Kinetic Energy - 0.5
Turbulent Dissipation Rate - 0.5
Turbulent Viscosity - 0.7
Energy - 0.7

This is what the residuals look like right now: