[hamednoordoost@yahoo.com]

hi
i try to solve temperature around an sphere that inner temperature of sphere is known and the surface temperature need to calculated. there is free convection(in air) around sphere. after running, the convection are true but conduction part (in thickness of steel sphere) not. all of shallow sphere become same temperature with inner surface of sphere.
what is the problem? (and i have the wall-shallow in interface of solid and fluid)
thanks in advance
Hamed

[hamednoordoost@yahoo.com]

a picture on my solution are attached
the problem are on inner circle that are same temp as inner wall!!!

[flotus1]

Not sure if I got your problem right. Maybe you could add a sketch where you explain what you mean with "inner circle" and "inner wall".
So you have a hollow sphere with a temperature boundary condition at the inner wall and you are simulating the free convection at the outer wall. right?

Your results look quite plausible in this case.

[hamednoordoost@yahoo.com]

thanks flotus1, i have a new problem on this case...scale...
the scale of this problem is -1 to +1 m and the above solution is on -1 to +1 cm!!!
when i change scale to -1 to +1 m solution became(attached image) not like the free convection that should be around sphere!!
does the scale are very important at free convection? what's the solution?
(i solve on 10,000 iterate and it's not what should be)
thanks again

[flotus1]

Yes, the scale is important.

But even more important are the boundary conditions.
Could you explain (add labels to the image above) what types of BC you assigned to which wall. I am especially curious about the BC at the far field, because it looks like a wall in the pictures you posted.
Also why are you doing a transient calculation? Are you expecting turbulence from buoyancy?

[hamednoordoost@yahoo.com]

all of the BC's are wall... the far field at 30 degree for simulating environment air.... the smallest circle(inner surface of sphere) are wall at 328 degree and intersection between solid and circle are coupled wall.
i solve for steady first and answer wasn't right then, as suggest by a friend ,change it to transient (the answer wasn't right either)

[flotus1]

If you use a wall at the outer boundary, you are not simulating free convection.
It is basically the convection around a sphere inside another sphere.
I think pressure boundaries would be a more approptiate choice for the flow field.

In your second simulation, it looks like there is no gravitation at all. Check the setup again. Apart from that, there is just not enough information to guess what went wrong in the second simulation.
The first one looks pretty good though now that we know the outer boundary is a wall.

[hamednoordoost@yahoo.com]

you mean to use pressure-far-field at outer boundary?(some friend suggest that but then what to do with bussinesq?)
then i have to use ideal gas instead of bussinesq! and it give me floating point error!!
or should i use another pressure BC?
thanks for your patience

[flotus1]

Did you already read the section about natural convection in the Fluent manual?
14.2.4 "Natural convection and Buoyancy-Driven Flows"

It provides usefull hints for the problems you are facing.

[sajeesh]

hi,

I am a beginner in fluent.i want to simulate the sphere moving in a fluid..the sphere is moving by the drag force and pressure force by the fluid initially.and then the motion of sphere modify the flow field ..finally the sphere reaches steady movement ...which method i should follow ...in ansys...plas help me ..or refer any tutorial to start with..somebody is telling VOF method.or particle transport method ..which i should follow pls help me.

[flotus1]

Start your own thread and stop hijacking other threads that have nothing to do with your topic.

[sajeesh]

sorry sir...i wont repeat

[hamednoordoost@yahoo.com]

hi again
i solved my problem. for natural convection it is appropriate to stimulate the environment by using pressure far field as your environment BC.
*pressure far field are more better than wall with fixed temp.
*my problem with scale was solved by adding more mesh at near the sphere...( i think by adding scale you need to ad mesh to correct the size... maybe)

thanks flotus1