With OpenFOAM CFD, we are studying katabatic/anabatic flows in a simple mountain slope-valley geometry as shown in the attached image. We would appreciate very much any help in explaining why we are obtaining such a strange flow behaviour (and in principle a bit unexpected) in case number 2, while cases 1 and 3 behave as expected.
Please, see attached images, indicating the obtained results as well as the imposed boundary conditions in all 3 studied cases; particularly, the flow speed is set to 0 everywhere, and the only differences from case to case are:
case 1: initial temperature in the fluid field/domain is constant + temperature in the mountain slope wall decreases as we move up to the peak
case 2: initial temperature in the fluid field/domain decreases with altitude + temperature in the mountain slope wall decreases as we move up to the peak
case 3: initial temperature in the fluid field/domain decreases with altitude + temperature in the mountain slope wall is constant
Thanks!
You should provide more details about your formulation and problem setting.
Assuming you are showing 3 steady solutions, case 1 and 2 seems not to be different as the solution at the steady state should forget the initial condition and should depend on the BCs. But from your description case 1 and 2 have the same BCs. and the solutions should be consequently the same.
In conclusion, if I understand well your cases, some difference should be found in the BCs. setting.
Hello Jose Ignacio Rojas Gregorio ,
if you say initial temperature at which time or iteration have you stopped the simulations? If you are doing a stationary simulation and if I correctly understand your case, you normally should get the same results, assuming your solutions is converged for case 1 and 2.
Best regards
Christian
Dear Christian and Filippo, Thanks for your comments!
Sure the key point is in the "steady" analysis, we agree with that: the initial conditions have an influence on how convergence develops, but not in the final result (in a steady simulation).
We are doing first a steady analysis, and then we do transient (soon I'll tell you until when we do steady and when or under which criterion we move on to transient)
Maybe we could try doing steady only, but very well converged for all the studied cases...
And ultimately, even if we do transient, since the problem is steady, it is a fake transient, only used for the resolution method... but given that the problem is steady, it should not be affected by the initial conditions...
I'll tell you more details soon. And thanks again for the help!
Jose
- Badges
- Science topic