phase change comparison elmer/Comsol

[raback]

Hi Julien,

Very impressive results! And I'm very happy that Elmer beats Comsol hands-down - if this is the correct interpretation.

You seem to take very good use of Elmer. One challenge in coupling the many equations is achieving 2nd order accuracy with time. The cases involving mesh update tend easily to be just first order. Also coupling in natural convection may sometimes be difficult with the default time-stepping strategy. You could try to use "Nonlinear Timestepping". You can find some examples of it among tests. Basically it is a strategy to estimate the stiffness matrix and rhs vector accounting properly the previous timestep in some timestepping schemes. The standard scheme assumes that the stiffness matrix corresponds to just to current time.

-Peter

[julien givernaud]

Thank you very much Peter,

I will use the "Nonlinear Timestepping" in future works. Elmer is a very good tool!

Julien

[dmitry]

Hi!
After close look of this topic I founded in the code

Code: Select all

 
Material 1
 Phase Change Intervals(2,1) = 1680.2 1696

and

Code: Select all

Equation 1
  Check Latent Heat Release = True

1) Please, can someone explain, what this keywords responsible for? When I modeled crystallization of casting in mold using effective heat capacity(temporal) I have obtained results without this keywords! And it seems to be in agreement with expirement.
2) During visualization to see the phase change interfaces I created isosurfaces with temperature of solidus and liquidus. Is it right approach?

[mzenker]

Hi,

1) see this post and this one.

HTH,

Matthias

[sebastien ROUQUETTE]

Hi Julien

Excellent post! i study arc welding processes... but your case is very interesting and answers some of my questions!

Go on...

Cheers

[sebastien ROUQUETTE]

i just discovered your post! excellent work! another excellent contribution!

[annier]

Hi Sebastien,
discovered or rediscovered it? it seems you have already discovered it several months ago.

Yours Sincerely,
Anil Kunwar

[sebastien ROUQUETTE]

You're right Anil.... i so busy that i did not remember this post (funny thing is that is saved into a file a while ago).

One good thing is: even after few month, i'm still agree with my self

Have a nice day Anil

[annier]

Hi All,
I am trying to apply enthalpy based method for heat transfer and fluid flow analysis in phase change of Sn0.7Cu alloys.
The phase change interval has been taken as

Code: Select all

 Phase Change Intervals(2,1) = 500.0 503.0

The melting point of Sn0.7Cu is 500.0 K.
The example set by Julien is so wonderful. Thank you.

Yours Sincerely,
Anil Kunwar

[alexbrown]

I tried to replicate this phase change comparison case, when i run the "Compute ALE moving mesh with phase change solid liquid" case, the analysis was successful and vtu file is generated, but when i tried to open the vtu file in paraview, the below error occur:

Code: Select all

Cannot read point data array "mesh update" from PointData in piece 0.  The data array in the element may be too short.

I am using Elmer version 8.4. The generated vtu file is attached.