[DrDan]

Hello dear Foamers,

I have a problem while trying to learn to use the HeatTransfer solvers of OpenFOAM. Before trying what I intend to do when I understand everything, I started with a simple test case.
I build the model and then went to have a look at the tutorials. I've chosen the chtMultiRegionFoam solver since I'm interested in the behaviour over time, instead of a steady state solution. Then I used the snappyMultiRegionHeater tutorial as a base for my case.
I'm now already trying for several weeks to get it to give me plausible results. I don't have problems getting it to run, but the results I'm getting are far from what I expect.

I want to start with a cold structure (everything initialized to 300K) and have a hot gas flow (inlet T set to fixed 800K) through the domain and heat everything up. But that's not what I'm getting. I guess I'm having problems with my BCs.

If you want, you can have a look at the case here:
https://webdrive.gkd.de/?ShareToken=...1B2F225E09814A
(link valid till the end of September 2014)

There you can find the complete case, a picture of the mesh and the checkMesh log files for the two regions I used in the case.
(The case will finish in about 20 minutes overall on my 6 core Xeon)

The problem I'm having looks like the temperature field isn't coupled to the flow field. With my model being 6 mm long and a flow speed of 0.1 m/s, the whole fluid domain should be filled in 0.06 seconds. And I expect to see the steep gradient of the temperature rushing from the inlet to the outlet in that time. But even when I calculate to 0.12 seconds, the temperature field looks like a distribution where there is no flow at all.

I'm slowly going crazy with this. I searched google over and over again, even checked the OpenFOAM bugtracker to see if anything is being reported, but to no avail.

I know I'm asking a lot, but could anyone please have a look and guide me into the right direction?
I'd really appreciate some help.
Thank you very much

Daniel

[laurentD]

Hello Daniel,
i have exactly the same problem than you. In my case, a heat flow is introduced at the inlet but the heat flow doesn't spread into the fluid. It is making me crazy.
Have you solved it ?
Or did you find a clue to help me ?
Thank you in advance for your answer.
Have a good day.
Laurent (France)

[DrDan]

Hello Laurent,

acutally, yes, I figured out what went wrong with my simulation.

It turned out that OpenFOAM was calculating correctly from the beginning. It was my expectation of the result that was off.

Due to the things we work with, my models are usually very small. In this particular case, the whole model was about 6 mm long.
I set a rather slow inlet velocity, which would theoretically then need 60 ms to cross the whole model. And that's what I expected.

Since the result didn't match my expectations, no matter what I tried, I turned to cfd-online. Unfortunately no one could help me. So I played around some more. When I scaled the model up by a factor of 1000 and also the inlet velocity, I saw a result that I could accept. Then I slowly scaled it down again in steps of 10. It was an eye-opener when I noticed what was really going on.

At one point I started to see some strange circulation of the flow I didn't expect. Then it hit me. What was going on, is buoyancy.

By the way I set up the case, with the gravity perpendicular to the flow, the hot inlet was flowing to the top of the model, where all the temperature was dissipated to the wall. This was all happening within the first cells closest to the inlet. So I couldn't actually see what was going on and it looked like the temperature field wasn't moving.

There were a few ways I could now "fix" my calculation. Either turn the gravity into another direction. Or lower the gravitational constant (be careful if you set it to zero. In my case this lead to numerical instabilities. Rather try some very small number, if you want to turn gravity off). Or increase the speed of the flow at the inlet.

Regarding the BCs and the case setup itself, I kept close to the chtMultiRegionFoam tutorials. They work pretty well and weren't the cause of my problem at all.

I hope this little insight to my learning process is helpful.

Daniel

[laurentD]

Hello Daniel,
Thank you for your answer. What a good new to see you have solved your problem.
In fact i have solved the mine too, but the reason was more stupid.
In fact i used a mesher which works with milimeters. And i didn't realize that OpenFOAM use meters.
So my heat flux was spreading with a honnest velocity, but i couldn't see it because my model was very huge.
Now i know it...
Have a good day.
Laurent