I'm trying to simulate the solidification and melting process of a PCM in ANSYS Fluent, using solidification and melting module. To do so, since properties for solid and liquid phases are different, I'm trying to apply these difference for these two phases separately. I've already tried using piecewise function for tempratures over and below phase change temprature. But it causes divergence in my simulation.
I'm not intimately unfamiliar with that solver, but how are you handling properties at the transition? Are you using the piecewise definition to smooth over the phase transition or did you define a sharp step discontinuity? Often times defining a wider "mushy zone" between solid and liquid can improve convergence without significantly decreasing simulation accuracy. It should also stay energy consistent as long as you're latent heat is distributed over the temperature width of the mushy zone.
Thank you for your answer @Nicholas Robert Jankowski. I was using two piecewise polynomials. For the liquid phase it was variable with the temperature but for the solid phase it was constant. I have not payed attention to probable sharp step made by the change in the function. I have to check it. If the values at the boundary were not almost identical, I'll try using a linear function to make the transition smooth between these two phases in the mushy zone .
Again, thanks for your answer.
Hi dear Farhad, I suggest that you use COMSOL Multiphysics for simulation. I have already simulated the phase-change material in this software. This work can be did simply in this software! I can help you if you wish.
this is my problem too. if you found the solution please guide me too.
tnx
Aref Ghayedi I used a piecewise function with 3 zones. For temperatures over Liquidus temperature, i used the specific polynomial for liquid phase, and for temperatures below Solidus temperature, the specific polynomial for solid phase. Finally to smooth the transition between the last value for solid phase function (@T solidus) and first value of liquid phase function (@T liquidus) i've used a linear interpolation to connect these two values with a smooth transition. I don't know if i'm right or wrong but the problem with divergence has been solved.
Note that these two polynomials can be constant values. For example the density for solid phase can be constant.
Farhad Afsharpanah Thank you kindly
I didn't understand that you are using a temperature range or a constant temperature (for phase change temperature). which one do you use?
and
Is this the only solution for density change that you found?
Aref Ghayedi You're welcome. For the phase change i'm using two different values of T_solidus and T_ liquidus. So i have three zones of 1.solid, 2.liquid and 3.the transition zone (mushy zone). As Mr. Jankowski suggested, i simply used a linear function for this transition area to eliminate the divergence problem. If you have problem understanding me again, message me, we can exchange contacts.
For the density change , no, i'm still searching.
Check out this question that i've asked:
https://www.researchgate.net/post/How_to_use_a_temperature_dependent_density_function_just_for_the_buoyancy_term_in_the_momentum_equation_in_ANSYS_Fluent
I'm trying to find a way for using a temperature dependent density, just for buoyancy term in momentum equation. As you probably know, using a temperature dependent density (as a polynomial function in fluent) for the whole momentum equation can make the simulation harder.
Of course Ali Radwan
Firstly, the PCM domain is defined as a liquid, then the properties for the solid, transition and liquid phases are applied. The density variation is simulated using boussinesq approximation. Also, for the variation of the specific heat and thermal conductivity you can use a piece-wise polynomial function with three stages (3 temperature ranges for solid, transition and liquid phases) . For the divergence you need also to check the mesh quality and apply smaller time step, especially during initializing your solution.
the way i do it and i think its best is to use is as follows (check tutorials on solidification):
1- choose solidification as your model
2- define density from boussenisq approximation (need to add value for density and thermal diffusivity)
3- also define solidus and liquidus temperatures
I hope this will solve your problem
also check
http://www.afs.enea.it/project/neptunius/docs/fluent/html/tg/main_pre.htm
- Badges
- Science topic