[sa har]

hi,i simulate solar air heater with natural convection in fluent (3D and solar radiation simulate with source term in absorber) .
i use k-e model for turbulence model and boussinesq for natural convection.
after running, report:volume integral:turbulent:wall y+:maximum and minimum for air channel is 18-0 or in contour of y+ is also.
but i read that y+ must be between 30-300 for k-e model. how can increase y+?where is my problem?
please explain about y+.
thanks alot

[davidwilcox]

the standard k epsilon assumes that the flow is fully turbulent. This means that it follows the log-law distribution near the boundary. If you look at the log-law in your normal (u^+) vs (y^+) plots, you notice that the straight line (log region) does not equal 0 when u^+ is 0. This of course violates the no-slip condition. Which is why we have wall functions for the standard k-epsilon.

The log law is based on the balance between production of turbulence kinetic energy and its dissipation rate ( convective and diffusive terms for k are negligible).

Because you have a wall function, it takes care of the viscous sub-region. This means that you do not have to refine to the death near the wall regions when you are using the k-epsilon. Typically the smallest y+ should be around 11 ( the overlap region in the (u^+) vs (y^+) plots).


As to what is yplus. Think of it as a form of Reynolds number. In this case the length scale is y where y is the normal distance from the wall. So, to increase the y plus, you need to increase y.

Hope this helps.

[sa har]

Originally Posted by davidwilcox

the standard k epsilon assumes that the flow is fully turbulent. This means that it follows the log-law distribution near the boundary. If you look at the log-law in your normal (u^+) vs (y^+) plots, you notice that the straight line (log region) does not equal 0 when u^+ is 0. This of course violates the no-slip condition. Which is why we have wall functions for the standard k-epsilon.

The log law is based on the balance between production of turbulence kinetic energy and its dissipation rate ( convective and diffusive terms for k are negligible).

Because you have a wall function, it takes care of the viscous sub-region. This means that you do not have to refine to the death near the wall regions when you are using the k-epsilon. Typically the smallest y+ should be around 11 ( the overlap region in the (u^+) vs (y^+) plots).


As to what is yplus. Think of it as a form of Reynolds number. In this case the length scale is y where y is the normal distance from the wall. So, to increase the y plus, you need to increase y.

Hope this helps.

david thanks alot. your helps is very useful to understand about y plus. how can calculate y plus in fluent?
when i see countur of y plus in fluent my minimum of y plus is 0 !!! how can i increase it?when i use bigger mesh,my maximum y plus increase but minimum is fixed and 0 !!

[davidwilcox]

y plus will always be 0 in the fluid region. Make sure your yplus contours are plotted at your walls.

[sa har]

Originally Posted by davidwilcox

y plus will always be 0 in the fluid region. Make sure your yplus contours are plotted at your walls.

david thaks alot
i use realizable k-epsilon model for turbulence with enhanced wall treatment.
now my y+ is between 11 to 27, is enough for solve this problem? but i read that y+ must be between 30 to 300.
regards

[fabiobodnar]

Natural convection with turbulence model? I use laminar model, with the solution methods recommended for Boussinesq, than have on tutorial.

[sa har]

Originally Posted by fabiobodnar

Natural convection with turbulence model? I use laminar model, with the solution methods recommended for Boussinesq, than have on tutorial.

hi,i used boussinesq model but my grashof number is 10^8 when i use hydraulic diameter.how was your rayleigh number or grashof number?

[fabiobodnar]

Originally Posted by sa har

hi,i used boussinesq model but my grashof number is 10^8 when i use hydraulic diameter.how was your rayleigh number or grashof number?

Hi, my number of Rayleigh is high, for my tank. I use laminar because the velocity of fluid is low in natural convection... so it works.