how to output the temperature on the boundary?

[chengan.wang]

Hello everyone

I have a small question here:

Now I'm working on the coupled problem between natural convection and radiation. I want to use the temperature in the domain and on the boundary. Here, I try to ouput the temperature on the hot wall, on which the temperature is constant. But the result got by the following code is variable.(The cordinates are correct)

!***************************************
open(21,file="chaud.dat")
do ilelt = 1, nlelt

ifac = lstelt(ilelt)
iel = ifabor(ifac) ! associated boundary cell
write(21,*)cdgfbo(1,ifac),cdgfbo(3,ifac),rtpa(iel,ivar)
enddo
close(21)
!***************************************

I guess I couldn't use "rtpa(iel,ivar)" to get teh temperature on the boundary face. If somebody knows how to deal with it?

Thank you!

[Jacques Fontaine]

Hello,

There is an example (true if there is no wall function) in usproj (v2.2):

Code: Select all

! Compute reconstructed value in boundary cells

    compute grad T with grdcell -> gradT

    do ifac = 1, nfabor (loop on boundary faces)
      iel = ifabor(ifac)  (boundary faces -> cell connectivity)
      diipbx = diipb(1,ifac) (compute vector II', cf: theory guide) 
      diipby = diipb(2,ifac)
      diipbz = diipb(3,ifac)
      TempIp(ifac) =   rtp(iel,isca(iscalt))              &
                    + diipbx*gradT(iel,1)                   &
                    + diipby*gradT(iel,2)                   &
                    + diipbz*gradT(iel,3)
    enddo

So we have:


Next, you need to reconstruct the value at the face (cf: theory guide (condli)):

Code: Select all

iclvar = iclrtp(isca(iscalt), icoef) 

do ifac = 1, nfabor (loop on boundary faces)
  TempF(ifac) = coefa(ifac, iclvar) + coefb(ifac, iclvar)*TempIp(ifac)
enddo 

To output the result, it is quite simple in sequential mode, in parallel you have to be prudent and to use synchronization functions (in this case paragv).

Best regards,

[chengan.wang]

Thank you for your answer.
In fact, I use orthogonal mesh and I have two kinds of boundary conditions, one is Dirichlet, left and right wall, I can use "coefa" directly to get temperature; I apply the Neumann condition on the top and bottom walls, if I can also calculate the temperature the same way you show me?

Sincerely,

Chengan WANG

[Jacques Fontaine]

If you use orthogonal meshes you can use coefa and coefb directly. Same way for Neuman/Dirichlet BC.

In my previous post please note my edit in the 2nd "Code box" (iclvar added).

Best regards,

[chengan.wang]

Hello,
If there is wall function (when I use LES), what should I do for ouput the exact temperature?
Best regardes
Chengan