Unexpected relation between Nusselt and Reynolds (chtMultiRegionFoam)
 

Hello everybody, I have a question and I hope that someone can help me...
I have been running chtMultiRegionFoam to get the evolution of temperature from an object, similar to a semi-infinite cylinder, to the ambient. The cylinder tip is approximately in the center of the domain (say y=0), while the boundary condition on the other end (in the infinite direction, y=y_up) is zero derivative.

Around the cylinder is a stream of air, flowing in the direction from the infinite side to the tip, parallel to the cylinder axis in the y<0 direction. We are interested in the average temperature on the cylinder tip. I have been measuring it and following a simplified model (http://en.wikipedia.org/wiki/Lumped_capacitance_model) to find the heat transfer coefficient h, i.e. I fit my results with the exponential curve and then use tau = (m Cp) / (h A) to solve for h.

The problem is that the heat transfer coefficient (or Nu) decreases when the velocity (or Re) is increased, which is the opposite of what happens in all other configurations I have found :confused:. As far as I have seen, the advection towards the tip maintains a slight temperature gradient inside the wire, with the highest temperatures being near the tip.

I assume that the gas moves so fast that the heat does not have time to diffuse in the direction normal to the cylinder wall and it remains warm as it moves to the tip, resulting in a slower heat flux from the tip to the ambient. Other than that, the results do not seem surprising (or unrealistic).

Could anyone advise me? Do I need to change my model? Is my simulation wrong? Is this something that has been studied before in detail? Should I calculate h in another way?

No one can help me with this?

Hello there Mr startingWithCFD,

I have to say that the description of your case is a little foggy.
I think I understand how the geometry looks like (still am not sure...), but the initial conditions, and most importantly boundary conditions (BCs) are a complete mystery.

Add to the description the following, and maybe we can find some solution:
- what are the boundary conditions on the outer walls of the fluid for temperature?
- what are the inflow conditions for the fluid (U,p,T)?
- what are the outflow conditions?
- what drives the fluid flow (inlet? outlet? body force?)?
- are you running compressible version?
- are you using turbulence models?
- what BC are you using for fluid-solid coupling?
- very important- how do you calculate your Re and Nu?
- any other details that make your case specific.

Unfortunately we do not hold the same books and manuals. Though all of them say the same all over the world, they put it in different words.
Furthermore, OF allows for such a variety of case setups, that without a detailed specification of what you used, it is nearly impossible to point for the troublemaker.

Summing up- make a clear, brief and "eatable" description of the case and methodology, then state the problem, and maybe someone will be able to help ;)

Best,
Pawel

ps
it feels quite strange to address you Mr (Mrs?) startingWithCFD. Maybe think of getting a nick, or even a name :)