Learning Forum

[Jure Stošić]

Hello, I am a student trying to simulate heating and cooling of an object through natural convection (picture below). The material used for cladding is Copper (Cu), and the core is Silicone Dioxide (SiO2). Heat will be generated from the cladding through passing current to desired temperature (400C), the core will remain at room temperature (presumed 22C). I want to observe how fast my object reaches desired temperature, and how long it takes for it to cool to room temperature.
I cannot seem to find any topic related to my problem, and transient thermal simulations yielded impractical results. I saw that the duration of the simulation influences the rate at which my object will reach desired temperature.
I would greatly appriciate any help, even a small hint to point me on the right path, thank you in advance.


Core diameter (125 μm),
Cladding outer diameter (165 μm),
Length (1000 μm).

[Essence]

Hello,
What do you mean by "impractical results"? How is the mesh? Did you use adquate time step size? Did you setup the case properly?

[Jure Stošić]

Hello,

thank you for a fast reply, I am not really sure with which boundaries I could achieve my simulation. I went through Ansys learning for meshing, and the mesh connects fine like the course teaches. I used time stepping 'Program Controlled'. I am attaching mesh and Analysis Settings for further information. By 'impractical results' I mean that if I change 'Number of Steps' or 'Step End Time' the heating of fiber cladding changes per step so I am not really able to read how fast the object heats up, also my professor said the cooling through natural convection should be around 150 ms - 250 ms, which I am not able to achieve for the last week.

[Essence]

I have a couple of questions. Please help me understand your query better.

You mentioned SiO2 and copper. Hence, I assume SiO2 at the core and copper around it (as shown in the figure). Is that right? If it is, then I do not see any fluid domain around cylinder for natural convection to occur.

If my above assumptions are right, then it would be better to use Ansys Fluent instead of Ansys Mechanical. You need to create a fluid domain enclosing  the cylinder.

[Jure Stošić]

Yes, you are right, I have tried modelling air surrounding the model as a solid, but was not able to correctly set boundaries. To my understanding a heated face cannot act as a convection plane element?
So yes, the copper is heated by current passing through, I can model the air surrounding the geometry, but fluent also needs 'inlet'/'outlet' suggesting air flow, which is not present in my query.

[Essence]

You can provide pressure inlet and pressure outlet BCs. Then you can just leave the guage total pressure to 0 Pa at both the BCs. You can simulate this analysis in Fluent.

[Rob]

To add, make sure the density of the gas (air) is incompressible ideal gas or it won't rise. You also need to read up on operating density for natural convection problems.