CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Why y+=1 is required for Heat Transfer predictions? (http://www.cfd-online.com/Forums/fluent/125723-why-y-1-required-heat-transfer-predictions.html)

 motorbean October 29, 2013 20:06

Why y+=1 is required for Heat Transfer predictions?

Dear All!

I'm using k-omega-SST turbulence model in Fluent to simulate conjugate heat transfer on a blade surface. And I want to AVOID the wall-functions approach

I've read quite a few posts in which it is said that y+<=1 is required for heat transfer predictions.

My questions is, since the upper bound of the laminar sublayer of a turbulent boundary layer is y+=5, (and in this laminar sublayer, the velocity profile varies linearly in the wall normal direction, am I right?), then why not simply go for y+<=5, instead of requiring y+<=1?

I haven't found any detail explanation about why y+<=1 is required for Heat Transfer Predictions. Can someone tell me, or give me any links/documents for detail explanations, please?

Thanks, so much!!

 Bionico October 30, 2013 03:40

Hi Motorbean,
k-omega SST required y+=1 in the boundary layer. I know that it's difficult to reach this target, so I think that y+<5 is anyway a good situation.
k-omega SST uses Enhanced wall function, so you should carefully avoid y+>10.

In conclusion the more you are near y=1, the more accurate the simulation will be.

Sorry for my bad english ;)

Regards

 motorbean October 30, 2013 05:26

Hi, Bionico, could you tell me why k-omega-SST requires y+=1? or do you some references about this?

I often see some people stressed that for heat transfer y+ should be <1, I would very much like to know the explanation behind this.

 Bionico October 30, 2013 06:13

Fluent R14.5 - Theory Guide (Chapter 4.4.4 - Wall Boundary Condition for k-omega):

"The wall boundary conditions for the k equation in the k-omega models are treated in the same way as the k equation is treated when enhanced wall treatments are used with the k-epsilon models"

So, considering that Enhanced wall treatment needs a value for y+ next to 1, you should refine your mesh as much as possible.
This requirement is important even for turbulence, not just heat transfer (anyway there's an interaction between them...)

Best regards

 motorbean October 30, 2013 06:43

Thanks Bionico. But do you know why Enhance Wall Treatment requires y+=1, rather than just y+<=5 ? Because I think,.. for y+<=5, the velocity variation is linear in wall normal direction, so what's the difference between y+=1 and y+=5?

Regards,
Motorbean

 Bionico October 31, 2013 04:29

Because Enhanced Wall Treatment equations resolve the viscous sublayer of the Boundary Layer, where the non-dimensional distance (y+) is <1.
y+<5 is anyway a good mesh but not optimal!
If you want to use a Standard wall function you can handle mesh with 30<y+<300.
Be careful to avoid y+ = 10-15 because there's the transition between the linear and logarithmic behaviour of the boundary layer!

See Fluent Theory Guide for more detailed (and graphical) information. :)

Regards

 motorbean November 5, 2013 18:33

Thanks. The last thing to mention: it seems that the range for viscous sublayer is normally considered as y+<5, rather than y+<1, right?

Regards

 Bionico November 6, 2013 03:06

Yes, you're right: I've read back the theory and the range for viscous sublayer is y+<5. ;)

Anyway if you use EWT, remember to go near 1 as much as possible.

Regards

 All times are GMT -4. The time now is 07:16.