Dear CFD experts:

I am working on a conjugate heat transfer case:

Flow through a heat sink with several vertical fins. Heat source is attached to the bottom of the heat sink. Cool air flows through the fins to take heat away. Flow speed is pretty constant at around 1 m/s. The minimum cell size is about 0.03 meter. Hence, I will assume that the delta t should be around 0.03 seconds, correct? Unfortunately, I need to run a 2 hour simulation to see if the heat sink (fin) design is good. This means, long calculation time.

Now, since the flow itself is pretty steady, can I run a steady state cold flow case first, then, use it as the starting flow for the conjugate heat transfer case? But, in this case, does CFL criteria still applies? Am I still limited to 0.01 seconds delta t? Is there any way to speed up the transient calculation?

phsieh2005

Thought I would chime in since no one responded...

Yes I agree. The time step for the complete P,u,v,w (K,E) would be small, perhaps proceeding at about CFL ~ 0.5 to 1.0. This is generally due to the pressure coupling which by nature is elliptic, but is advanced, by varying solution methods, as a parabolized transport equation.

After solution of the velocity state, preserve your velocity field solution S = [u,v,w], then begin your energy equation solution (assuming non-bouyancy effects). Depending on the discretization methods, e.g. implicit advection, you should be able to advance this solution at a much higher CFL number, perhaps 5 to 10.