Problem in modeling a simple heat loss through convection

classic Classic list List threaded Threaded
2 messages Options
Reply | Threaded
Open this post in threaded view
|

Problem in modeling a simple heat loss through convection

hamid
Hi mates,
I'm trying to make a very simple model for a pin losing heat through free convection to an environment. To account to for the buoyancy force, I have used the term "alpha*g*rho_air*(T-Tamb)" (taken from a sample by FEATool).
For boundary conditions, I have used Temperature and Pressure for the vertical wall against warm plate.
For the top part I have used "Natural outflow" with "convective heat flux" and for the bottom part I have tried "convective heat flux" and "Insulation" with "Wall no slip" and "Natural out flow".
Neither of these conditions give me a stable answer. Anybody has any idea what's going wrong here? I have uploaded the file here as well. pin_2D.fea

Thanks for your help beforehand.

PS: In the samples of FEATool for convection inside a box, I replaced the above-mentioned term with Pr.Ra for the Fy and it totally changed the results and velocity profile, but at least the model could converge.
Reply | Threaded
Open this post in threaded view
|

Re: Problem in modeling a simple heat loss through convection

Precise Simulation
Administrator
Unless you are absolutely sure, the solution should (easily) converge to a stable solution, the answer usually is that it is a difficult non-linear problem, possibly also that no stable solution exists. The workflow for highly non-linear problems is usually:

- Increase the non-linear relaxation parameters and maximum number of non-linear steps in the solver settings
- Perform multiple restarted solver steps, while in each step restarting using the old/previous solution and increasing the parameters leading to the non-linearity little by little
- Switch to a time dependent solver