Discussion Forum

Hi

you have a delete entity in 4.2, and in any case with the union the "hollow" domain can lways be excluded from the simulation by deleting in in the main physics node (also valid for 3.5)

--
Good luck
Ivar

Thanks for that tip, but the problem is that the material in the hollow is not showing up as a separate domain that can be deleted; it is part of the domain of the entire block. (That is, when in the physics mode, only two domains show up, the screw and the block; there is no separable portion in the hollow.) Is there a way I should have constructed the geometry in the first place so that more than two domains would be produced? My procedure originally was to import the screw from a CAD file, and then add a block. Thanks.

Hi

1 you can also in the geoemtry do boolean and difference operations, including splits and delete. If you have two overlapping domains you must be in "assembly mode (in which case you must clean your geoemtry manually), normally in "union mode" the domains are split up.

This applies to 3.5 too, the cmmands are just slightly different

--
Good luck
Ivar

I looked at your model in Comsol 4.2. Usually in a case like this, the simple geometry sequence you put (import screw from CAD, put surrounding block) should give the desired result without any ado. That is, you'd end up with three domains: the block outside the screw, the screw itself, and the hollow inside the screw.

However, in your case Comsol seems to have problems recognizing that the first and third of the described domains are in fact not connected. I think that has to do with the complicated surface model of the screw, which creates many overlapping volumes and even for a human it would be hard to tell what's what. (Click through the individual boundaries to see what I mean.)

A work-around that seems to work in your case is to convert the CAD object to a surface and then convert it back into a solid. Somehow, then, Comsol recognizes the domain relations correctly.

Best,

John

Thanks for that last set of tips by Ivar and John H. I found also that another work around is to create the block first, then import the screw, then create a Comsol object of the screw and block (not sure what this does, but...) then do a union of the two, and then a split. I have to re-trace those commands (in 3.5a, and then also in 4.2 etc.) just to double-check that this sequence was in fact what worked, but anyway, it ended up giving me the ability to isolate the domain inside the screw and delete it. So anyway, I think I'm good at this stage. I do agree with John H. that the complicated surfaces of the screw may be causing an issue. But anyway, thanks for the nice help. I appreciate it.