CFD Online Discussion Forums

CFD Online Discussion Forums (
-   CFX (
-   -   Heat transfer coefficient from CFX (

Sulfi September 14, 2007 01:51

Heat transfer coefficient from CFX
Hi all,

I am trying to find out the htc distribution in a 3d aerfoil used for a axial turbine. The usual trend of htc is maximum at the leading edge and then reduces down as we go away from LE to the TE. But i am getting a reverse trend. I am prescribing aerfoil wall temp as 1000F and the inflow velocity is 1126F. Any suggestions why CFX is giving such an inverse trend(The qflux is also max at the trailing edge)?. Do i need to use the tbulk for proper prediction?

Thanks, Sulfi

Glenn September 14, 2007 10:19

Re: Heat transfer coefficient from CFX
The HTC profole characteristic from LE to TE of an airfoil will depend on the shape of the airfoil. It is not uncommon for the HTC magnitude to rise as one traverses aft from the stagnation point (line). For example, on the suction side of an airfoi, the flow may be rapidly accelerating immediately aft of the leading edge and could easily result in an HTC which rises over this region above the stagnation HTC value. Look at your flow field around the airfoil shape and see if this may be the case. Are you sure the HTC should be monotonically reducing as you progress aft of the LE? Why?

Glenn Horrocks September 14, 2007 19:16

Re: Heat transfer coefficient from CFX

Also, as Sulfi mentioned the wall heat transfer coefficient is based on a local estimate of the fluid bulk temperature which often varies by quite a bit over the domain. Activating the expert parameter mentioned will fix the bulk temp to a defined temperature and this is how most wall heat transfer coefficients are reported in the literature.

Glenn Horrocks

Sulfi September 15, 2007 08:02

Re: Heat transfer coefficient from CFX
Glenn, Thanks for the reply! From journals, As Glenn pointed i have come across couple of cases where the htc at suction side is more than the the LE htc value. I would like to validate my case for building confidence. The difficulty here is that i am using steam properties. Could you suggest some open literature where i can get some data for 1st stage nozzle validation for steam turbine.

Thanks Sulfi

Wooster September 18, 2007 19:18

Re: Heat transfer coefficient from CFX
Regarding the Bulk temperature setting, what would be a good approximation? I'm looking for a HTC in one of my models as well and I've been wondering how the bulk temp effects the thermal problem. Thanks!


sulfi September 19, 2007 01:04

Re: Heat transfer coefficient from CFX
You need to set the Tbulk as inlet temperature.


Glenn Horrocks September 19, 2007 18:17

Re: Heat transfer coefficient from CFX

Wooster - the Tbulk temperature does not affect the simulation, only the way the HTC is reported in the results file. In the simulation heat transfer at the wall is calculated directly as a heat flux with no direct use of HTC's. The HTC is calculated from the wall heat flux.

Sulfi - You can set the Tbulk to be anything you like. Inlet temperature would be a good guess for most applications.

Glenn Horrocks

valypoor September 20, 2007 13:52

Re: Heat transfer coefficient from CFX *NM*

wooster September 24, 2007 12:51

Re: Heat transfer coefficient from CFX
Thanks guys.

I'm looking for an easy way to just select the heated surface and figure out a heat transfer coefficient for a numerical model. Looks like I'll have to setup an equation in Pre for later use in Post. I'll play around with the bulk temp and see how that works for me.


sulfi September 25, 2007 06:31

Re: Heat transfer coefficient from CFX

Anybody aware of validation database(or journal where we can find the standard airfoil information) for htc around Gas turbine airfoil.

Thanks in advance, sulfi

Pete September 29, 2007 21:41

Re: Heat transfer coefficient from CFX
The initial guess for tbulk does effect the reporting of the HTC in POST. Could this make a difference if you are using CFX convection for your thermal analysis?

Karkoura March 23, 2011 15:10


I am having the same problem. Could you please let me know where you can modify the Tbulk in CFX?


karolrecko March 23, 2011 17:36

areaInt(Wall Heat Transfer Coefficient) problem
Hi everyone!

I have a problem. Why if I set function in CFX Post:

areaInt(Wall Heat Transfer Coefficient)@wall

I get a bad unit:

[kg m^2 s^-3 K^-1]=[W/K]? Where is a m^2?

How is Ansys calculated this?

best regards


singer1812 March 24, 2011 10:36


[W m^-2 K^-1] = [J s^-1 m^-2 K^-1] = [N m s^-1 m^-2 K^-1] = [kg m s^-2 m s^-1 m^-2 K^-1] = [kg s^-3 K^-1]

Integrated over an area

[kg s^-3 K^-1]*[m^2] = [kg m^2 s^-3 K^-1]

which is what you got.

Sure you didnt want areaAve(h)@wall or areaInt(h)@wall/area of wall?

karolrecko March 24, 2011 16:02

Thanks for your reply!!

I calculated flow over a cylinder. Cylinder has a 294K and fluid has 293K.
I cut the cylinder in mean hight by plane and take polyline by boundary intersection- wall of cylinder and plane (green line on the picture)

Difference temperature between cylinder and fluid is 1K, so


why if I take this chart:
it isn't the same like this:

singer1812 March 24, 2011 16:14

Did you specify Tbulk as 293K or are you using first node temperature for your h?

If you just specified 293K as the fluid inlet, and asked for a wall h, without specifying the Tbulk, CFX will use the first node temp for its h calcs, which probably wont be 293K, it will be something higher.

That would drive your delta T down, in which you would expect your charts to show a higher value for h, which it does.

karolrecko March 24, 2011 16:21

I set the Initialisation domain temperature - 293K and fluid inlet 293K. that's wrong?

Where I might specify the Tbulk temperature?

singer1812 March 24, 2011 16:25

Your initializing just fine.

But if you want a fixed Tbulk for any h calcs you need to set that Tbulk. It doesnt have to be fluid temperature, you can set it to whatever you want.

Go into the expert parameter section under Discretisation Tab>>Miscellaneous>>tbulk for htc to set the Tbulk to a fixed value.

karolrecko March 24, 2011 16:47

Thanks a lot!! Problem is sloved.Singer1812, I owe you one.

All times are GMT -4. The time now is 17:22.