while simulating melting of paraffin (PCM) using ANSYS Fluent the solver says Divergence detected in Temperature after the liquid fraction is 0.3740 (37.40%). Until this point the solver runs smoothly. What could be the problem? How can we solve it?
Description:
Physics: Melting/Solidification
2D model,
BC: one wall has constant temperature; other walls insulated.
Convergence: Energy =10e-6, others 10e-3
This kind of issue occurs when continuity diverges and generally it is linked to large time step. Initially, the buoyancy effects are weak therefore the simulations manage to remain stable with relatively large time steps but as the melting rate increases the solution starts to oscillate due to large natural convection effects and eventually leads to divergence. In PCM simulation you have to be very careful with the time step,mesh quality and numerical schemes. Specially, due to the limit on the time step, PCM simulations require large computational times.
You can find the appropriate numerical setup for PCM simulation in the following attached paper. The simulation step works well even in high convection zones. Also have a look at the section 2.3. discretization schemes.
You may also want to reduce continuity criteria to at least 10-4 , this way you won't get asymmetric solution.
It might be because of inappropriate time-step and grid selection. The good agreement can be calculated using Courant Number. Please refer to the attached. Or you can refine your mesh and lower the time step on a trial basis.
Hi
The melting problem is highly non-linear and the solution diverges for convergence criteria less than 10−3. So, set the convergence criterion as 10−3
- Badges
- Science topic