[mariam.sara]

Hello
I attempted to evaluate Nu-avg at the bottom of a square cavity filled with a nanofluid (illustrated in photo) I create a line at the bottom of the cavity then using function calculator I select lengthavg and location as line 1 and I was define a variable name it as Nu which I give it expression as:
Nu=Wall Heat Transfer Coefficient*L/kw
the problem CFX give me too high Nu-avg compared to Fluent case (i.e. in CFX Nuavg=130.45 while in Fluent Nuavg=10.26) is there a problem in my method of calcu in CFD post?

mariam

[ghorrocks]

What temperature is the wall HTC being normalised against? You probably want to make sure it is normalised against a specific value, rather than the CFX calculated reference value.

[mariam.sara]

Hello Glen
The temperature at the bottom wall (which I want to evaluate Nu-avg along it ) is fixed and I give it as given temp boundary in cfx-pre its same as Th in photo? Is it same as normolised temperature that you mean?

[ghorrocks]

That is not really what I meant.

Convection is defined as q" = h(Ts-Ta) where Ts is the surface temperature and Ta is an ambient temperature. CFX does not know what Ta should be so it has a guess based on local conditions. But if you are comparing to published results Ta is probably a known value. So the heat transfer coefficient (h) is referenced to Ta - for a given heat flux, h will be a function of Ta.

If you want to compare to other results you are going to have to make sure your h is referenced the correct Ta. If you don't your Nu calc will be wrong.

[mariam.sara]

Alright if I know Ta how I can set it in CFX-Post?

[ghorrocks]

In CFD-Post the variable "Wall Heat Transfer Coefficient" is defined using the Ta defined in CFX-Pre. If you did nothing it will be the default value (where CFX guesses based on local conditions). If you used the expert parameter which sets the reference temperature it will be that.

If you have completed the simulation then you will have to recalculate h from the Wall Heat Flux. If Wall Heat Flux = q", Ta = the reference temperature you want to use and T is the wall temperature then h = q"/(T-Ta).

[mariam.sara]

I cannot find exactly where Ta in CFX pre? may you send an illustrated photo for it?
About your second answer you mean I need to evaluate h as h=Wall Heat Flux/(Tw-Ta) then use this h to evaluate Nusselt as Nu=h*L/k?

[ghorrocks]

You define Ta. Whatever you are comparing against will have a reference temperature for h, that is what you use.

And yes, this h will then be able to compare to published values.

[mariam.sara]

Originally Posted by ghorrocks

You define Ta. Whatever you are comparing against will have a reference temperature for h, that is what you use.

And yes, this h will then be able to compare to published values.

Thank you but where I can set the value of reference temperature in cfx pre you not show that to me?

[ghorrocks]

I can't remember the exact name. Search the forum or check the expert parameters in the documentation.

[mariam.sara]

Originally Posted by ghorrocks

I can't remember the exact name. Search the forum or check the expert parameters in the documentation.

I found where to set bulk temp for htc in CFX-Pre it in insert>expert parameter but I need to know how much value I need to put there? do I chose it randomly?

[ghorrocks]

You set it to the reference temperature you want to use for the HTC.

[mariam.sara]

Hence it will be any temperature ?