Convergence Problem using 2. order implicited time discretization
 

Hey guys,

I'm simulating melting behaviour in a 2D geometry, dominated by natural convection.

I already did a grid dependency study and calculated a time step, which should be small enough.

So far my residuals fulfilled the criteria easily (10-6/10-6/10-6 and 10-10 for energy) with an average of 6 Iterations.

I changed from First order to second order implicid (nothing else!) and now the residuals for continuity stop decreasing at around 10-3 and I have no idea why.

Any suggestions?

Thanks in advance

Tom

Screenshot of my residuals

Any ideas?