Discussion Groups

Simulation CFD

Reply
Member
Jaloboch
Posts: 5
Registered: ‎10-22-2013

Weird Natural Convection

166 Views, 3 Replies
10-23-2013 01:03 AM

Hi!


I'm new in CFD and am having some problems trying to simulate the natural convection of air due to four hot walls with known identical temperature.

 

Basically, I assigned a temperature BC of 40C to the 4 sides of a chimney, pressure gage of 0 to the two openings (10cm in diameter) at the top and bottom, and a temperature BC of 35C to the inlet at the bottom (fresh air entering at 35C).
I also set the air to be variable and the proper gravity as well (0,0,-1 in my case).
Everything's going fine until the ~100th iteration, where the air direction and the magnitude seems to go haywire as in the attachment.

 

May I know what mistakes have I done that caused this or what proper BCs to place?

 

Your help is very much appreciated.
Thanks!
Jason

Please use plain text.
Product Support
wildej
Posts: 854
Registered: ‎08-25-2011

Re: Weird Natural Convection

10-23-2013 05:59 AM in reply to: Jaloboch

Hi,

 

I am not quite sure what is expected here - why is the goal of running this model?

 

Are the 4 walls really all fixed at 40C or is this just the ambient? I would assume that what is happening is that all the air is simply reaching 40C and then nothing much is going to happen.

 

Kind regards,

Jon



Jon Wilde
Please use plain text.
Member
Jaloboch
Posts: 5
Registered: ‎10-22-2013

Re: Weird Natural Convection

10-24-2013 04:22 AM in reply to: wildej

This is the cfz file.

Please use plain text.
Product Support
wildej
Posts: 854
Registered: ‎08-25-2011

Re: Weird Natural Convection

10-24-2013 11:36 AM in reply to: Jaloboch

As a public update. Jason and I chatted about this and it turns out that this is a glass chimney, ambient temp at bottom and top but the air is heated through solar radiation.

 

It was decided that simply applying 40C to the walls was not sufficient to heat the internal air (it all just ends up at 40C with nowhere sensible to go :smileyhappy: ) and it would be far smarter to model the larger building within a solar dome and have solar radiation active.

 

This is the next step, which should work far better.

 

Thanks,

Jon



Jon Wilde
Please use plain text.