Author Topic: The calculated temperatures is lower than specified temperature  (Read 3096 times)

Offline pitney1

  • Jr. Member
  • **
  • Posts: 65
  • Reputation: +0/-0
  • Searching for solution
    • View Profile
Advertisement
Hi,

I am running a simulation with energy equation activated. The temperature at the outlet is lower than the specified temperature at the inlet. I have tetrahedral mesh.

Please help.

Thanks.

Offline william

  • Full Member
  • ***
  • Posts: 159
  • Reputation: +15/-0
  • Know it, share it.
    • View Profile
Re: The calculated temperatures is lower than specified temperature
« Reply #1 on: February 13, 2012, 09:05:03 PM »
When using a tet-mesh and second-order scheme the calculated temperatures will be higher or lower than specified temperatures on inlets. This seems to only happen when using tetrahedral mesh, a hex-mesh will not give this problem.

The solution to the problem seems to be to use node-based solution scheme instead of the default cell-based scheme. This was verified with a simple test case as well as with a large (1.8M) cell case and in both cases this helped. Other suggestions I have collected are:

1) W-cycle for energy equation in the Multi-grid options usually gives
good convergence in conduction dominated simulations.

2) In many cases, the solution may not be converged if with a energy URF
less than 0.95. It is advisable to converge the case finally with 0.98
or more.

3) When the mesh has tetrahedral elements and temperature gets limited
(eithermaximum or minimum), we can try switching the secondary-gradient
calculations off using (rpsetvar 'temperature/secondary-gradient? #f).
Note that this is only to give a stable solution in the beginning and to
get the final solution, we need to switch that on eventually.