Discussion Forum

Hi, anybody.
I met a problem. It is related with the inconstant result
between 3D PDE mode and 2D axial symmetry model.
Maybe I set some error parameters in PDE.

-------------------------------------------------------------------
My model is very simple half sphere with a hole in center.
Inter radius is 1.3mm, and outer radius is 40mm. The model is
heated with Joule heat and the heat power density is in inverse proportion
to the 4th power of position's distance to center -- r.
In the position r=1.3mm, the heat power density is 100000000 [W/m3]
I want to get the temperature distribution plot after 60 seconds.
Initial temperature is 293.15K

Please check attached sketch map
-------------------------------------------------------------

I build 2 models, one is 3D model and the other one is 2D axial symmetry
model. For each model, two methods is applied: one is standard comsol
heat transfer equation, and the other one is PDE equation.
In PDE equation, both 2D and 3D, parameter is same.
"c" item is set to thermal conductivity (k)
"Da" item is set to density*specified heat (rho*c)
"f" item :
in 3D model, f= 100000000*(0.0013^2/(x^2+y^2+z^2))^2
in 2d model, f= 100000000*(0.0013^2/(r^2+z^2))^2
For detail information, please check attached mph file.

The following is result,
*******************************************************************

3D Standard 3D PDE 2D Standard 2D PDE

386K 386K 386K 435K

*******************************************************************

Obviously, the 2D PDE case result is inconstant with the other 3 cases.
It has no relationship with model mesh size.

if I set the heat source power density to a constant value, the 4 results
are same.

So I suspect there are some problem for "f" item in 2D PDE mode.

Anyone can help me?

======================================================================