SUGAR BEET CLAMP CFD MODELLING: IMPLEMENTING A STRATEGY FOR DEALING WITH PRECISION ISSUES
Feature image: A clamp cover blowing in the wind.
OpenFoam version: 9
Solver: pimpleFoam and clampPimpleFoam v5
Case: clamp_14
THE PROBLEM
In the last post in this series, it was noted that there was an issue with dt being too small for this case. The proposed solution was to run the simulations in two steps: one step to solve for p and U, then a second step that kept p and U constant and solves for T. So that’s what we’ve done…
THE SOLUTION
The system can basically be described as:
- run pimpleFoam for 90 second to get a stable p and U
- pass p and U to clampPimpleFoam (with the pimple loop turned off) and solve for T over 900 seconds (15 minutes).
- repeat.
This was achieved by having two subfolders in the case folder, and running it all from an Allrun file.
The programming in the all Allrun file is not elegant, but it works. Given there this was my first Allrun file, a couple of points:
- the first three lines have to be there.
- The first line tells the shell that the file is talking “shell”. If it had “/bash” instead of “/sh”, it would have been speaking “bash”.
- The second line sets the working directory
- The third line tells shell that whatever comes after runApplication is an OpenFoam command.
Regarding making the two subcases align, there was an issue with the Tf field that was a result of the phi field not copying with the mapFields function. In the end, this file just needed to be copied in it’s own line of command.
THE RESULT
Looked fine, but I didn’t bother recording the animation – I was too excited to run an actual test of the model using actual real data and this version of the model that might actually be functional!. To see an animation of the first test of the model with real data, see the next post [LINK].