[mohammad]

Hi everybody
I am doing a CFD simulation of a 2 bladed wind turbine under wind shear profile condition when the wind speed changes according to a logarithmic wind profile. The wind turbine is standstill ( not rotating) and I want to model a single blade at the zero azimuth angle ( = blades are supposed to be vertical).

The question is:
How should I simulate the B.C on the geometrical plane of symmetry?
What is the B.C? From the first moment it is clear that "Symmetric" boundary condition is not true because the wind speed distribution on the two sides of this plate are not the same.

I was thinking to use a domain and put a single blade somewhere in the domain. in this new model the domain is stretched out of both sides of the blade ( form the root and the tip).Therefore I can apply the wind profile to the domain and this wind profile contains some part of the wind profile just before the blade's root. ( SEE THE ATTACHMENT )

This new idea can simulate some effects of wind shear profile near the blade's root; however, because the other blade does not exist here the wind simulation is still not exact in the sense that the wind before the blade's root is not affected by the other blade.


Can anybody please help me to clarify which B.C should I use?

Regards,

[ghorrocks]

How should I simulate the B.C on the geometrical plane of symmetry?

It is not clear what you are modelling. If this is modelling the blade when stationary then apply a tangential velocity equal to the velocity at your boundary.

[mohammad]

Dear Glenn

Originally Posted by ghorrocks

It is not clear what you are modelling.

because I am going to simulate half model( one blade only) , I am asking about the B.C. of the plane of geometrical symmetry which passes through the hub and divides the wind turbine to two geometrically similar blades.

Originally Posted by ghorrocks

If this is modelling the blade when stationary then apply a tangential velocity equal to the velocity at your boundary.

According to your idea, if i apply this condition, still i am removing the effects( shear effects) of the removed part of model ( the other blade) which is not simulated in my simulation.

my question is that is there any way to simulate the effects of the wind shear ( friction and shear forces,....) on that plane?
Regards,

[ghorrocks]

Why not model this using the rotating frames of reference? Then you can do a model with the difficulty of a steady state simulation, but include the effects of blade rotation. Much more accurate and easier. Also you do not have the weird bottom boundary to worry about.

[mohammad]

Originally Posted by ghorrocks

Why not model this using the rotating frames of reference?

I am doing the standstill analysis....How can I use the rotary frame for a stationary case!!!???

[mohammad]

Hi everybody!!!!

Any help???!!!

[ghorrocks]

is there any way to simulate the effects of the wind shear ( friction and shear forces,....) on that plane?

Yes, as I said before just apply a tangential velocity to the plane equal to the boundary layer velocity at that height.

[mohammad]

Originally Posted by mohammad

Hi everybody
I am doing a CFD simulation of a 2 bladed wind turbine under wind shear profile condition when the wind speed changes according to a logarithmic wind profile. The wind turbine is standstill ( not rotating) ...........................................

Regards,

Dear all,
I have made the model and applied the velocity profile ,as mentioned by Glenn Horrocks.

However I cannot get a stable results.
I have tried many differnt timesteps.....also I switched to the Transient simulation . But, all of my attemps failed.

Do you have any idea? how should I get the results....

Thanks,

[ghorrocks]

http://www.cfd-online.com/Wiki/Ansys...gence_criteria

[s.q]

Originally Posted by mohammad

Dear all,
I have made the model and applied the velocity profile ,as mentioned by Glenn Horrocks.

However I cannot get a stable results.
I have tried many differnt timesteps.....also I switched to the Transient simulation . But, all of my attemps failed.

Do you have any idea? how should I get the results....

Thanks,

hi mohammad
i had simulated a wind turbine in fluent such as you.
and now i want to calculate the output power using torque(report-forces-moment)
but this output power is 10 times bigger than real amount.


would you plz guide me????

[mohammad]

Originally Posted by s.q

hi mohammad
output power is 10 times bigger than real amount.

Hi,
unfortunately the commercial CFD softwares are always erroneous. First make sure about your Turbu. model.
Then, check your grid resolution for a) adequate nodes in different areas including chordwise, spanwise and near wake zone.
Later check your y+ and # nodes inside the BL.
at last check the convergence.

good luck

[ghorrocks]

unfortunately the commercial CFD softwares are always erroneous.

Rubbish. If they are always wrong then why do people buy them? Used properly it is accurate for many cases.

The issue is almost certainly a problem in the way the simulation was set up, not the software itself.

Anyway, this vague question is an FAQ: http://www.cfd-online.com/Wiki/Ansys..._inaccurate.3F

[mohammad]

Originally Posted by ghorrocks

Rubbish !!!!!! .

First of all ghorrocks, I am sure that you do not need to become angry like this. This is a public forum, so please watch your words and be formal.... Knowlege is the best thing if comes with politness. I AM SURE politness BRINGS YOU MORE CREDIT.

[/QUOTE]

No doubt about ghorrocks' experience and knowledge, however, as all turbulence models are , to some extent, based on some theoritical and even hypothetical formulations, how could you stand on the CFD softwares' side and support them like this? If they are as exact as you say, why do scholars try for the improvement?????

[ghorrocks]

I apologise if my frankness offends you but I tend to get straight to the point By the way, I am not angry.

Yes, turbulence models are based on theoretical, empirical and even guessed models. But if they give you results which are accurate enough to improve your design that is all which matters. The key phrase is "accurate enough". I wish the scholars all best in their quest to make CFD more accurate and able to handle a wider range of applications, but software available now is adequate for most purposes.

[jiaojiao]

Originally Posted by ghorrocks

Yes, as I said before just apply a tangential velocity to the plane equal to the boundary layer velocity at that height.

as this said to use the tangential velocity,but I dont know which patchfield I should use.could anyone tell me?
I forget where I have ever seen the set that
[ inlet
{ type shearVelocity;
n (0 0 1); direction
y (0 1 0); vertical aixs
value 7; the velocity value
}
]but it is wrong when I try it,it shows that unkown patchfield type shearVelocity.I am waiting for your help!thanks so much!