Page 1 of 1

Solid-liquid phase change

Posted: 06 Jan 2016, 15:22
by Martin2000
Hi Elmer users,

I wanted to simulate the phase change of water from liquid to solid with the enthalpy method. I am not sure, if I get the right input parameters because the time until phase change is happening is very short, so it would be nice if anybody who is practised in phase change simulations could take a look at the sif. First I had also Problems with the convergence tolerance. I solved this problem in decreasing the nonlinear relaxation factor to 0.25. My Problem is in 2D and the water is in a ractangular box with the size of 0.25*0.5 m.

I appreciate any hints.

Yours sincerely, Martin.

Re: Solid-liquid phase change

Posted: 06 Jan 2016, 21:33
by raback

For various phase change strategies the test case of Julien may be very usefull. See,


Re: Solid-liquid phase change

Posted: 08 Jan 2016, 19:32
by XisZero
Hi Martin,

As far as i can see in your SIF, the Enthalpy is way too small.
I dont know what values you are using, but try with multiplying by 1000 as the unit of enthalpy should be J/m^3 . (333 kJ / kg, density = 1000 kg/m^3 --> 333 MJ/m^3).
Your cp is in J/kg K i think?

In addition to Peters link, you could take a look at a phase change UDF i have written for some tests:

Hope it helps.



Re: Solid-liquid phase change

Posted: 29 Feb 2016, 21:18
by Martin2000
Hi XisZero,
thank you for your response.
Are you sure, that the unit of the enthalpy is J/m³?
After having simulated it with the new values (in J/m³), I think, the time until Phase Change is very long (about 13-16 h).
I appreciate any hints.

Yours sincerely, Martin.

Re: Solid-liquid phase change

Posted: 09 Dec 2019, 22:27
by sebastien ROUQUETTE
Hello Dear Elmer Users,

Actually i still have the same question .... i did 3 simulations (heat transfer / welding case) where i want to include the release of latent heat. But before doing this, i wanted to understand how to use the "phase change" option.

1st simulation: no latent heat and heat capacity is used

2nd: simulation: i used the phase change option with enthalpy (defined as integral of density * heat capacity) ...

but 2nd simulation did not give the same thermal results of the 1st????

3rd one: i did not multiply the heat capacity by the density.. just integrated it over the temperature interval.. and i got the same results as 1st simulation....

So the unit of "enthalpy" is really J/m3 ... because it is more J/kg ... i don't understand ...

If someone has an idea? if you need a sif file...i can attach it!

Regards, Sebastien

Re: Solid-liquid phase change

Posted: 10 Dec 2019, 09:18
by sebastien ROUQUETTE
well well...
after one night of thinking and few more simulations....

I may have done a mistake!!! i have always defined the "heat capacity" in my "material library" also with the definition of "enthalpy" ... so i guess that the resolution of my heat transfer employed "heat capacity" instead of "enthalpy" as it was given! (i did not check the source code to check this statement!).
If i removed the definition of "heat capacity", it worked well with the enthalpy given in J/m3 unit (and not the one with the one given in J/kg which led to a computation abort!!).

As a conclusion, if you want to work with the "enthalpy" instead of "heat capacity"... don't give the "heat capacity" in the material section ... or there is probably a keyword to force the use of "enthalpy"? if someone knows? cheers.

Have a good day

Re: Solid-liquid phase change

Posted: 22 Feb 2020, 17:13
by NJank
In previous discussion the transient phase change solver required the inclusion of both Enthalpy and Effective Heat Capacity due to an issue getting transient enthalpy in the first time step. See discussion here:

I have been doing this exclusively in all of my standard phase change models. The output behavior matches the analytical Neumann melting model. Maybe the issue is only with the Heat Capacity and not the Effective Heat Capacity keyword?