Heat simulation mesh lvl3 cylinder shape issue

Numerical methods and mathematical models of Elmer
Post Reply
laileng
Posts: 23
Joined: 15 Mar 2016, 08:01
Antispam: Yes

Heat simulation mesh lvl3 cylinder shape issue

Post by laileng »

Hi.
A box defined as heat source was placed on a cylinder shape heat sink
Only heat equation is applied, the material and boundary condition were defined properly
The simulation result was frustrated when increases the Mesh level to 3!
the box is placed at the center of the cylinder
the box is placed at the center of the cylinder
Capture.PNG (61.46 KiB) Viewed 5495 times
Is this the solver problem? because this does not happen to a square shape heat sink for mesh lvl3
If not the solver problem, What is the best way to explain this?
Thank you.

Regards
laileng
dmitry
Posts: 53
Joined: 09 Dec 2014, 16:52
Antispam: Yes
Location: Russia

Re: Heat simulation mesh lvl3 cylinder shape issue

Post by dmitry »

Hi!
I guess, in order to help you, we need the SIF file and the mesh, you used.
laileng
Posts: 23
Joined: 15 Mar 2016, 08:01
Antispam: Yes

Re: Heat simulation mesh lvl3 cylinder shape issue

Post by laileng »

Hi dmitry,

I wish to upload the attachment here, but the mesh file size has exceeds the maximum allowed upload size.
Is there any ways for me to send you the file?
Thank you.

Regards,
laileng
raback
Site Admin
Posts: 4812
Joined: 22 Aug 2009, 11:57
Antispam: Yes
Location: Espoo, Finland
Contact:

Re: Heat simulation mesh lvl3 cylinder shape issue

Post by raback »

Hi

Could you be more specific. Simulation being "frustrated" does not ring any bell to me.

Note that for mesh levels 3 in 3D you have 64 times more elements. So you can easily hit a wall in memory consumption.

-Peter
dmitry
Posts: 53
Joined: 09 Dec 2014, 16:52
Antispam: Yes
Location: Russia

Re: Heat simulation mesh lvl3 cylinder shape issue

Post by dmitry »

laileng wrote:Hi dmitry,

I wish to upload the attachment here, but the mesh file size has exceeds the maximum allowed upload size.
Is there any ways for me to send you the file?
Thank you.

Regards,
laileng
You can simply apload your case to some cloud like Dropbox or something else and share the URL/
laileng
Posts: 23
Joined: 15 Mar 2016, 08:01
Antispam: Yes

Re: Heat simulation mesh lvl3 cylinder shape issue

Post by laileng »

Hi raback,

Good day.
in the simulation, all the setting and parameters are the same, the only diff in mesh lvl.
cylinder- the highest temperature observed from lvl 1 and 2 are 69.3 and 69.4 respectively. But the for lvl3 is 67.4, the change is 2 degree diff from lvl2
a box on a square heat sink also studied, and the result is shown in figure below. The volume and thickness of the cylinder and square are the same, just diff in shape.
square- let consider the highest temperature, only slightly change between lvl1 and lvl2, where lvl3 is same with lvl2
therefore, i would expect the highest temperature of lvl3 same as lvl2 or does not change much for the cylinder case.
thank you.

Regards,
laileng
the box is located at the center of the square
the box is located at the center of the square
Capture.PNG (106.7 KiB) Viewed 5480 times
laileng
Posts: 23
Joined: 15 Mar 2016, 08:01
Antispam: Yes

Re: Heat simulation mesh lvl3 cylinder shape issue

Post by laileng »

Hi dmitry,

https://drive.google.com/file/d/0Bx1W27 ... sp=sharing
Thank you.

Regards,
laileng
dmitry
Posts: 53
Joined: 09 Dec 2014, 16:52
Antispam: Yes
Location: Russia

Re: Heat simulation mesh lvl3 cylinder shape issue

Post by dmitry »

Hi!
Unfortunately I was not able to solve with Mesh Levels = 3 on my computer due to lack of memory (as Peter said).
But I suppose,it's the mesh issue. IMHO: In the thin layers of 0.04 and 0.07 mm thickness the elements has wary sharp angles in Temp. gradient direction. After 3 splitting operations the angles become sharper that increase the errors.
In my opinion, there are two ways here:
The first one is to prepare fine mesh in the mesher on the thin layers of the domain.
The second one is to neglect that thin layers, instead you can calculate effective heat transfer coef. of the layers and use Heat gap boundary condition between heater and layer of 2 mm thickness.

That's my suggestion. Please correct me someone, if I'm wrong.
Post Reply