Solid-Fluid region interface mesh for chtMultiRegionFoam

ishan_ae · February 3, 2021, 03:12

I am trying to simulate a CHT analysis of high temperature flow through exhaust manifold geometry using chtMultiRegionFoam in transient PIMPLE solver mode until steady state is reached.

It has 3 inlets and one outlet.
On the solid outer boundary, I am using a convective BC with HTC and Ta supplied.

This is how the geometry looks like.

I am running upto 1 second of time with deltaT corresponding to a large limited Courant number of 50 and diffusion of 10.

Upto 0.8 seconds, all the parameters reach a steady state. The residuals are all good for solid and fluid regions.

I noticed that the results for fluid region are as expected, but for the solid region it is as if the conduction never happens from fluid to solid. I am monitoring the maximum temperature in the solid region and it reaches a steady state as well.

My question is, does the mesh transition at the interface matter when it comes to CHT simulations ?

Is this an issue with large Courant number ?

On the fluid side, yes, we need to have boundary layer resolution, but on the solid side we need to only solve for conduction. My understanding says that, from a physical point of view this transition wont be needed. But, from a numerical point of view a large mesh size jump will dissipate the results.

This is the zoomed out view of the mesh. Green is fluid and pink is solid.

This is the zoomed in view at the interface.

What is your experience and recommendation?

dlahaye · February 3, 2021, 07:02

Wonderful to meet you hear again!

Are you employ splitMeshRegion prior to running chtMultiRegionFoam?

Cheers, Domenico.

julieng · February 3, 2021, 07:26

Hello

I think that 0.8 sec is not enough to reach a stationnary solution in thermal process in solid region (of course depending of your initial conditions).

Julien

ishan_ae · February 3, 2021, 08:10

Hello Domenico,

Good to hear from you again.

I am not using splitMeshRegions. I am using a commercial pre-processor which outputs the fluid and solid regions. The common interface is output as mappedWall type on which I have applied different BC.

@Julieng: I am again running the simulation. I deleted my files from the HPC. I will show the monitor plots and additional contour plots hopefully tommorow.

keitaro7_14 · February 8, 2021, 13:09

I would verify the log of chtMultiRegionFoam. Is the solid region being resolved? Your tolerances for the solid region are small enough? Then, is the boundary for the fields well defined?

Then, as previously indicated, maybe the time is too low to reach the steady state.

February 3, 2021, 03:12	Solid-Fluid region interface mesh for chtMultiRegionFoam	#1
ishan_ae Member ishan Join Date: Oct 2017 Posts: 78 Rep Power: 9	I am trying to simulate a CHT analysis of high temperature flow through exhaust manifold geometry using chtMultiRegionFoam in transient PIMPLE solver mode until steady state is reached. It has 3 inlets and one outlet. On the solid outer boundary, I am using a convective BC with HTC and Ta supplied. This is how the geometry looks like. I am running upto 1 second of time with deltaT corresponding to a large limited Courant number of 50 and diffusion of 10. Upto 0.8 seconds, all the parameters reach a steady state. The residuals are all good for solid and fluid regions. I noticed that the results for fluid region are as expected, but for the solid region it is as if the conduction never happens from fluid to solid. I am monitoring the maximum temperature in the solid region and it reaches a steady state as well. My question is, does the mesh transition at the interface matter when it comes to CHT simulations ? Is this an issue with large Courant number ? On the fluid side, yes, we need to have boundary layer resolution, but on the solid side we need to only solve for conduction. My understanding says that, from a physical point of view this transition wont be needed. But, from a numerical point of view a large mesh size jump will dissipate the results. This is the zoomed out view of the mesh. Green is fluid and pink is solid. This is the zoomed in view at the interface. What is your experience and recommendation?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
How I can introduce my power heat (W) in chtMultiRegionFoam?	aminem	OpenFOAM Pre-Processing	32	August 29, 2019 03:23
Problem simulating the temperature rise in a composite material (chtMultiRegionFoam)	Adam_K	OpenFOAM Running, Solving & CFD	2	March 27, 2019 07:51
My radial inflow turbine	Abo Anas	CFX	27	May 11, 2018 02:44
[mesh manipulation] Importing Multiple Meshes	thomasnwalshiii	OpenFOAM Meshing & Mesh Conversion	18	December 19, 2015 19:57
[ANSYS Meshing] Combine solid mesh generated in workbench mesh and fluid mesh in fluent meshing ?	RPjack	ANSYS Meshing & Geometry	2	August 27, 2015 10:33

February 3, 2021, 07:02		#2
dlahaye Senior Member Domenico Lahaye Join Date: Dec 2013 Posts: 802 Blog Entries: 1 Rep Power: 19	Wonderful to meet you hear again! Are you employ splitMeshRegion prior to running chtMultiRegionFoam? Cheers, Domenico.

February 3, 2021, 07:26		#3
julieng Senior Member julien Join Date: Dec 2018 Posts: 107 Rep Power: 8	Hello I think that 0.8 sec is not enough to reach a stationnary solution in thermal process in solid region (of course depending of your initial conditions). Julien

February 3, 2021, 08:10		#4
ishan_ae Member ishan Join Date: Oct 2017 Posts: 78 Rep Power: 9	Hello Domenico, Good to hear from you again. I am not using splitMeshRegions. I am using a commercial pre-processor which outputs the fluid and solid regions. The common interface is output as mappedWall type on which I have applied different BC. @Julieng: I am again running the simulation. I deleted my files from the HPC. I will show the monitor plots and additional contour plots hopefully tommorow.

February 8, 2021, 13:09		#5
keitaro7_14 Member JuanMi Join Date: Nov 2017 Posts: 41 Rep Power: 9	I would verify the log of chtMultiRegionFoam. Is the solid region being resolved? Your tolerances for the solid region are small enough? Then, is the boundary for the fields well defined? Then, as previously indicated, maybe the time is too low to reach the steady state.