Smoothing for Tsource in cfdemSolverPisoScalar.C

Submitted by yesaswi92 on Fri, 01/11/2019 - 01:20

Hello,

I'm trying to solve the heat transfer problem for a static particle bed with interstitial air. The particles are bigger than the mesh used for the CFD solution. From the source code, I see that the heat exchange field (EuField in cfdemSolverPisoScalar.C or the Tsource in cfdemSolverPisoScalar.C) is just being added to the CFD cells that coincide with particle centroids. But Tsource in all the other cells that also occupy this particle is being taken as zero.

I am wondering why not use a smoothing model to distribute this Tsource?

Is there a specific reason for not implementing this?

Thank You
Yesaswi

Forums:

mardussa | Wed, 10/30/2019 - 23:42

I am looking at a very similar problem with a denser liquid instead of air. I have not found the distribution of Tsource (or any of the other particle properties as seen by the continuous phase) to be an issue.

Which voidfractionModel are you using? I am using the divided model which places additional tracking points inside of the particles for mapping of the void fraction, and I assumed that this would also distribute the other quantities being passed to the continuous phase.

Otherwise I would suggest you increase the size of your CFD mesh to be closer to that of your particles.
Rob