[krisipanikou]

Hello,

I am simulating natural convection heat transfer in a vertical can witch contains liquid, but I can't see the bernard-cells as I expected.. Is this major problem??

[yashganatra]

It depends on the Rayleigh number. What is your rayleigh number? Aspect ratio of the cavity? Post some more details. The critical Rayleigh number is 1708.

[sbaffini]

It also depends on the Turbulence model you are using. I have no experience on this specific problem but i guess it might be difficult, if not impossible, to see them in a steady k_eps simulation.

[yashganatra]

I had done the simulation ; it's not easy to capture ; Could you explain how does it depend on the turbulent model?

I had done an unsteady state but first I had done the steady state and then proceeded to unsteady.

Thanks,
Yash

[sbaffini]

Ok, i don't know the exact details of the phenomenon but, of course it depends on the viscosity and, of course, at some point, if you increase the temperature jump you're gonna get convection. So, what i was saing is that a given turbulent model may largely affect the viscosity and the overall system to the point that the instability is for much higher Rayleigh numbers, if any.

Of course there are people doing it (unsteady of course):

http://www.sciencedirect.com/science...42727X99000077
http://www.iket.fzk.de/cube/lib/file...0f60e4ecfe.pdf

But they don't seem completely happy. Also, if you set up Ra sufficiently high of course you're gonna get the phenomenon. But are you getting it right?

[yashganatra]

Thank you for your reply. Previously i thought that one could get these cells only till a critical number of 1708.
It depends on a lot of factors you are right

[krisipanikou]

Thank you all for the replies. It's not easy to capture, that's true!