HI all,

I am doing a simple heat transfer in a box. i have applied a heat load to the floor.

I am using P1 radiation model for this problem.

The residuals dont seem to be converging. they are oscillating up and down after few iterations.

Please advice me.

Thanks Kind regards Elleana

hello,

try to modify the under-relaxation factor. if the residuals don't seem to be converging, adapt your mesh by checking it with y+ option ( adapt->Yplus).

Hi, I have played enough with the relaxation factor. i will work with the Y+.

Thanks Kind regards Elleana

try to change discretization methods. sometime unstable scheme can be due to that.

If you are sure about your set up, then the most important thing to do is check your mesh. If you could afford then have a finer mesh, specially in the places where you could have higher gradients.

If the residual oscillating and not conerging, mesh is the main culprit. (assuming you set the calculation correctly).

Now to check up set up, try to do a very small and simplified case. Where you could have a simple geometry. (like a box , tube etc but simple one). But it should have similar boundary conditions as original case. Once you get solution here you know that your set up is okey, only thing remaining is to run this set up on complicated grid. By modifiying this complicated grid, you should get solution.

Hi guys,

Thanks you very much for your earlier reply.

I tried to make a simple box and good mesh.

Applied a Heat flux on the floor.

I am using the incompressible ideal gas option for air in material properties.

I have used P1 radiation model.

Have activated the gravity in the appropriate direction.

Started the simulation by intialising the room at 23 degrees and i have all the control variables set with first order upwind.

While solving for the equations it is pretty good up till few hundred iterations after that it is not oscillating too much but is stabilizing with minor jumps. results in a straight line.

Dont know where iam going wrong. Iam not getting sensible results.

Please advice me.

Thanks Kind Regards Elleana

which version are you using. In one of the calculations we use a particular turbulence model. Now by using older version 6.2.16 i could get the solution converged, but using the versions 6.3. onwards, the same mesh same set up behaves exactly what you have written. The residuals falls to values of 0.01 and stays there. Just keeps oscillating around it. I am not sure how it is relevant to your case, but i thought i will share it.

Hi,

I will check this out and get back to you.

Thanks

Kind regards

Elleana

Yes, if possible do try to check with other versions of fluents. For us, we are still using 6.2.16 for few reasons, like calculation speed, accuracy and convergence.

though things should improve in latest versions, with fluent this doesn't seem like the thing. And it is very difficult to get Fluent people to admit that there is something wrong with their solver.

Hi, since the problem what u have specified is already simplified. instead of jumping on to radiation model directly. 1.get a converged solution of simple flow / or only energy eqaution . 2.solve radiation for the same without initializing the solution. As per fluents help and my experience ,P1 results and computational time it takes ,give it edge over other radiation models.