I am trying to simulate a turbulent flow over a PCM plate(2-D). I have used the k-epsilon model with enhanced wall function. A coupled algorithm is used. I have treated phase change material as a solid domain. When I compared my results with published data, what I got is the melting fraction is coming very low as compared to the reference paper(let's say at time(t)=3500 sec, the reference paper has a liquid fraction of 0.24 while I am getting only 0.17. I have used constant values for all properties of PCM (same as the reference paper). Can anyone point out where I am doing wrong? I have used the same boundary conditions as the reference paper. Which parameter is responsible for such a low value melting fraction?
It is my humble request to share some suggestions.
Can anyone share any basic model on PCM with the fluid flow?
Regards
Prakash S
Dear Prakash,
I think some factors could be affecting the validation of your code. Apart from the turbulent flow simulation, that I do not have experience, I recommend you to check:
· The grid size and time step used;
· The mush-zone constant (C or Amush);
Besides, since you are using another simulation results as a reference, make sure that it was validated with experimental results. I hope it could be helpful.
Regards,
Thiago Torres
Dear Rocha sir,
the grid size and time step I am using are the same as the validation paper.
The value of the mush-zone constant is 10^5.
Yes, sir, it was validated with the experimental result.
but still, I am not getting good results.
Dear Prakash,
If all the parameters are equal to those of your reference, I cannot think what could be the issue. As a last recommendation, make sure your solution is converging in the entire simulation. When I was performing a phase-change simulation, after several time steps (about 24000=2400s) the solution started not converging. After verifying many parameters, I figured out that the mush-zone constant was not appropriate. I hope you find a way.
Thiago Torres
Dear sir,
My solution is converging everywhere. That is very true, It happens when there is no any flow, just a temp/heat flux condition(natural convection). Is it possible to change the mush-zone constant if it is already mentioned?
Sir, how to overcome convergence issues, when simulating the melting of PCM with natural convection?
It is not easy to identify the cause of convergence problems. It could be related to the mesh quality, under-relaxation factors, the own mush-zone constant, and several other parameters. I could not give you a different reccomendation then to consider a trial-and-error approach. Maybe someone with more experience could help.
Best regards,
Thiago Torres
- Badges
- Science topic
- Similar topics
- Geoscience
- Asia
- India