I seem to be having difficulties with this model. Even using a relatively simply mesh, it will not generate. However when I have managed to make it mesh, the solver quits unexpectedly just prior to start solving the analysis.
Any tips or help with this would be well received.
Here is a link to the share file:
Regarding the solver error that you are getting when it does mesh, move your files to a local drive and that should fix the issues.
Let me look at the model and see if I can provide some meshing guidelines.
Regarding the meshing, I would probably take a different approach with the geometry that you have.
Intead of including the air volumes on both end of the server racks I would remove them and neglect the surface resistances. Since we have already assigned a fixed flow rate through each rack by using the internal fan component, the need for the resistance is minimal.
The typical approach that the geometry will look like when it is imported into CFD looks like the following picture where you don't see the air volumes on the inlet side of the internal fans:
Removing the various volumes and surface materials will make it drastically easier to mesh your model.
My last comment, you are going to want to suppressed the mesh for the steel enclosures. This will also simplify your model.
Have a good weekend!
Thanks for your replies. regarding the file location - I am running them locally, just on a secondary hard drive.
I will give the geometry change a try now.
So I changed the geometry, and am trying a very coarse mesh, just trying to see if it will start to solve. So we are between 200 and 300 thousand elements. But it has hung on creating volume mesh for probably about an hour now? Normally it only takes this long for something in the millions of elements.
I reset the geometry settings (there were some edges I had hidden) and ran it with no changes to the model, and refined the mesh a little and it seems to be solving now.
I will post a share file shortly with my mesh in it, I would like some thoughts as to its accuracy?
Thanks for all your help!
Here is a link to my share file, complete with mesh and results set. I would appreciate some input on this, especially since this model had so many issues getting it to solve!
Now I am very confused - I made a change to the geometry to create an alternate set up (this is simply to go from a single equipment block inside each cabinet, to three blocks seprated by an equal sized air gap. This is to simulate an unblanked situation, and allow air to recirculate within the cabinet.
Then it created meshing errors, with only that change.
So I have been through and overhauled the cabinet frame design, making the vertical posts and top/bottom panels much thicker to ensure the minimum element length is quite long. I then removed all the side panels also, so there are even less geometry elements.
The automatic mesh for this is about 13,000 elements, and will generate the mesh with no problems. If I refine it, if crashes. It even suggests that I refine the mesh...?
So basically, a piece of geometry that is very similar to the original will not mesh. After I coarsen up the geometry substantially it generates and solves a coarse mesh, but crashed when I refine it?
Try out this support file. You had a lot of left over surface resistance materials and you didn't apply all your solid components between your server racks which resulted in a somewhat undesireable geometry situation.
I also switched your exit pressure to be applied to the CRAC exhaust so that we don't have a pressure boundary condition touching a flow rate boundary. This will cause the model to have issues. I would also consider adding the underfloor geometry so that you could have a closed loop system. For the CRAC you would use the heat exchanger material.