I’m unable to get my velocities to reach realistic values, I’ve had results on the order of 1e+23. I’ve reviewed the boundary conditions multiple times so I know that they are good. I’ve tried compressible and incompressible simulations. I was able to get realistic velocities and somewhat realistic pressures in the case of a single room but the h value was all over the place in the convergence plot. Expanding this setup to the half ward case was too much computationally and I only got to ~20s, with a 1 s timestep, after 13 core hours. Does anyone have any insight or tips for me?
I saw that you made some improvements already!
A good advice is to start with steady state first and move on to transient once you have established a proper setup.
General recommendations from my colleague:
- Extend inlets/outlets
- He is using flow conditions at inlet and outlet, the open face can then be bit bigger if possible
Let us know if that helped!
Jousef
Hi Jousef,
I ended up extending the domain for the pressure inlet and this did help me.
My issue with h being all over the place got fixed as well. I had been selecting eConst in the material section but was failing to save that change. Now on my convergence plot it shows e instead of h and the residual decays as I would expect.
Just commenting this in case anyone comes by looking for a fix to a similar issue, The goal was to decrease the pressure in the room by a significant amount so I had to use Compressible flow, incompressible was not giving realistic velocities for my set up. I was also having issues where two rooms separated by the air under a door were at about 20 Pa different from each other but removing the door fixed this issue. In future simulations if I need the door I will be increasing the width of the space under the door to attempt to fix that.