Discussion Forum

Hi

YES you are right, COMSOL is alwaye working in 3D, when you select 2D or 2D axi, or even 1D COMSOL assumes a default depth (in 2D its ussually 1[m] it's defined on the subdomain/main physics page, ususally named "thickss", it is to be understood as the "Z" direction as 2D is by defult x-y.

In 2D axi you need to multiply surfaces by 2*pi*r to get to volumes (for integration), its the "loop length" and must often be used for BC definition too, to get the units compliant.

In 1D its a area that is assumed (in Y and Z).

That is why one should always check the units (but some functions do not correctl transfer units, such as integration coupling variables so these might "flag red" even if they are correct. this applies to 3.5, its better in V4, but there are still a few holes in the current V4 too. It's quite a job to get all hese cases correct, but its very useful for us to check what we are doing, no ?

Have fun Comsoling
Ivar

Hi Ivar,

Thanks for the help. I am using 2D, but not axisymmetric. I suppose the thickness is then 1m, but it is not a variable in the subdomain physics section. It is a setting in some structural mechanics models, but not heat transfer or fluid flow. The equations are being solved in 2D, but it still confuses me why the boundary condition for heat flux seems to be 3D, ie - W/m^2 instead of W/m. Once my solution runs, the total heat input does not seem to correlate with a 1m depth though.

Jeff

Hi

you are right I'm s used to structural that's what comes up first ;)

Which version are you running ?

ivar

Version 3.5. Thanks.

Hi again

then if you go into the "Physics - Equations - Boundary Settings - Variable tab", at the bottom of the list there are a few constants, I believe it's these, but I'm not 100% sure, you should check with the doc and COMSOL support. But it seems natural to use 1[m] depth as default

You must have a 2D surface defines, and selected the appropriate physics on a boundary, otherwis COMSOL is not defining any variables, if not needed

U_htgh, L htgh, dVolbnd_htgh = 1

Infact in V4 you cannot even see these variables, neither any "depth info", this could be more explicit, I agree, could be worth to send a suggestion to COMSOL

Have fun Comsoling
Ivar

Hi Ivar,

Unfortunately I do not see any constants there. I will check the documentation a little more then. Thanks for the help.


Jeff

Hi
Do not forget to select the correct physics and settigns then APPLY to have comsol updates the required variables
I have updated my previous post
Ivar

Hi Ivar,

"htgh" is the general heat transfer and it looks like L_htgh refers to the convection length scale. Plus, I am using the non-isothermal module and those are not available.

I am experimenting with simpler models to find an appropriate comparison. However, you are right, it seems that it could be made easier.

Jeff

Hi

v3.5 see the htug.pdf p63, for V4 its p23 the depths re there, and in some cases its by default 1cm and not 1m !

Ivar

I am running COMSOL 4.1

I am trying to change the default cross-sectional area in a 1D time-dependent heat transfer problem. The cross-sectional area drastically changes the time scale. I have a cross-sectional area that is not a neat unit square of any standard unit.

I have looked at all these posts and searched the manuals, and I cannot figure out how to change this (or the depth assumed in a 2D simulation). How can I change these values?

Thanks!

Sorry, I'll post this as a new thread instead of tacking it on the end of this one.

Actually, VERY sorry, I was wrong, and I figured that out. Great forum!