Discontinuity of temperature at solid fluid interface

[dibyendu]

The temperature data along a line connecting points (0.5,-0.25,0.2) and (0.5,0.5,0.2) has been extracted from an existing tutorial on flow over a heated plate. Here are some data points for the solid and fluid regions:

For the solid region:

(-0.00125, 304.18)
(-0.001, 304.17)
(-0.00075, 304.17)
(-0.0005, 304.16)
(-0.00025, 304.15)
(0, 304.15)

For the fluid region:

(0, 304.12)
(0.0005, 304.12)
(0.001, 304.11)
(0.0015, 304.1)
(0.002, 304.1)
(0.0025, 304.09)

It’s evident that the temperature difference at the interface is larger than the temperature difference between neighboring cell points in the solid and fluid regions. This discrepancy indicates a discontinuity, contrary to the expectation of temperature continuity at the interface based on the boundary condition.

N.B. - OpenFOAM has been used to solve both fluid and solid regions

[Makis]

Hi @dibyendu,

sorry for the late response, I had no time to look into the forum for a while.

Are these results from the default tutorial? I assume with unmodified buoyantPimpleFoam and laplacianFoam? Using which OpenFOAM and adapter versions?

Are they the cell-center values, or the face-center values? Note that the cell sizes are also different in each region.

I would plot across the interface and check for any discontinuities. This does not seem to be the case:

In the OpenFOAM-preCICE paper, we had a detailed validation study on this. See the error we measured in Figure 8 (page 14).

If you want to check what preCICE itself does, then you can use watchpoints and ASTE.

[dibyendu]

Dear Makis,

Thank you for your

response. I’ve encountered an issue while attempting to plot data over a line from grouped datasets. My system consistently crashes and closes automatically when I try this in both Paraview 5.11.1 and Salome 9.7.0. However, I was able to successfully plot over a line when done separately, and I’ve attached the resulting image here.

To investigate the discrepancies further, I extracted data using Paraview and examined it for continuity. It appears that you are correct in pointing out that the interpolation from face center to cell center is causing temperature discrepancies at the interface.

[Makis]

I usually first group the two meshes and then apply a slice before plotting over a line, as this makes it easier to get a line that crosses locations valid for both domains.

If you have issues with ParaView, try getting a portable binary from the paraview website. I also had some issues with 5.11, and 5.12 worked. Note that OpenFOAM ships ParaView, and if you have multiple OpenFOAM installations, things might get a bit messed up.

[dibyendu]

Dear Makis,
paraview 5.12 works perfectly. Thank you for your kind help.