[mrswordf1sh]

Hi guys,

I am using ANSYS FLUENT to simulate the flow and calculate the power output of a horizontal axis wind turbine. I am using mostly k-e Realizable turbulence model and I use kw-SST sometimes. As for the wall treatment, I select either Enhanced Wall Treatment or Menther-Lechner together with k=e model. My y+ is below 5 (I can make it lower if necessary).

Someone said that, to use k-e Realizable together with Enhanced wall treatment, I have to make my y+ lower then 1. Is this true about the y+ and Enhanced wall treatment? Will it be problem if my y+ is less then five.

The other question is, I know that menter-lechner wall treatment independent from y+ and my my y+ is below 5 (y+<5). I am calculating the power output of a wind turbine using in fluent using k-e Realizable together with menther-lechner wall treatment method. Will people criticise me about using Menter-Lechner wall treatment for Wind turbine simulations? What`s your advice?

Many thanks.

[mrswordf1sh]

Originally Posted by mrswordf1sh

Hi guys,

I am using ANSYS FLUENT to simulate the flow and calculate the power output of a horizontal axis wind turbine. I am using mostly k-e Realizable turbulence model and I use kw-SST sometimes. As for the wall treatment, I select either Enhanced Wall Treatment or Menther-Lechner together with k=e model. My y+ is below 5 (I can make it lower if necessary).

Someone said that, to use k-e Realizable together with Enhanced wall treatment, I have to make my y+ lower then 1. Is this true about the y+ and Enhanced wall treatment? Will it be problem if my y+ is less then five.

The other question is, I know that menter-lechner wall treatment independent from y+ and my my y+ is below 5 (y+<5). I am calculating the power output of a wind turbine using in fluent using k-e Realizable together with menther-lechner wall treatment method. Will people criticise me about using Menter-Lechner wall treatment for Wind turbine simulations? What`s your advice?

Many thanks.

no response?

[halowine]

I don't know in detail the wall-treatment algorithm but the goal of a wall-model is to 'model' the boundary layer... In fact the wall-models work better if your 1st point lies in the log-law layer (30<y+<200) or even in the outer layer (y+>200). The condition of y+<5 is for resolved simulation, i.e. without wall-model/wall-treatment.

In conclusion,
if you want to resolve the boundary layer, go below y+<5 (laminar layer) and desactivate the wall-treatment in fluent, with y+<5 it means that the rest of your flow should be resolved so you should also desactivate the turbulence model and do a LES/DNS... (note that you also need some specifications on x- and z-direction... LES: x+~50 y+~1 z+~20 or DNS: x+~20 y+~1 z+~8)

if you want to model the BL, go over y+>30 and the wall-model should take care of finding back the good law-of-the-wall.

Are you sure that your first cell is at y+=5 ? At the scale of a windturbine your mesh is going to be huge !

cheers

[mrswordf1sh]

Originally Posted by halowine

I don't know in detail the wall-treatment algorithm but the goal of a wall-model is to 'model' the boundary layer... In fact the wall-models work better if your 1st point lies in the log-law layer (30<y+<200) or even in the outer layer (y+>200). The condition of y+<5 is for resolved simulation, i.e. without wall-model/wall-treatment.

In conclusion,
if you want to resolve the boundary layer, go below y+<5 (laminar layer) and desactivate the wall-treatment in fluent, with y+<5 it means that the rest of your flow should be resolved so you should also desactivate the turbulence model and do a LES/DNS... (note that you also need some specifications on x- and z-direction... LES: x+~50 y+~1 z+~20 or DNS: x+~20 y+~1 z+~8)

if you want to model the BL, go over y+>30 and the wall-model should take care of finding back the good law-of-the-wall.

Are you sure that your first cell is at y+=5 ? At the scale of a windturbine your mesh is going to be huge !

cheers

Thanks. Yes but it is a model wind turbine which has a rotor diameter of 1 meter. Mesh consist of nearly 6 million elements.

[bpulat]

I suggest you should take a look at the following paper.
There are valuable information regarding to your question.
http://www.southampton.ac.uk/~nwb/le...om_77_2006.pdf