Forum

Please or Register to create posts and topics.

Wake field simulation and output

Page 1 of 2Next

Hi Pro.Marten

I tried to calculate the wake field of 5MW wind turbine. When I intercepted the downstream x=100m and z=90m and checked the distribution of y axis, I found that only half of the velocity field distribution was uniform, while the other half was chaotic, as shown in Figure 1. I wanted to know why. Figure 2,3,4 shows the parameters I set.

The blue color in Figure 1 is calculated by OPENFAST software, and the orange color is calculated by Qblade. I want to get the same and similar results as the blue color. May I ask where is the problem with my setting?

Uploaded files:
  • You need to login to have access to uploads.

Hi,

your wake settings seem to be ok – but because your simulation is quite short (only 1000 steps) you should maybe increase the wake relaxation factor to something like 0.8. If you want to obtain good velicity distributions you need to make sure that the wind turbine wake is fully developed and has reached a stationary state. This should take around 20 full rotor revolutions. With your settings (6° azimuthal discretization) and a wake relaxation factor of 0.8 you would need to simulate for about 1500 timesteps.

To get smooth velocity distributions you should also sample the wake velocities for several timesteps and calculate an average because the tip vortices can sometimes be visible in the vlocity fields, depending on the time you take a snapshot.

BR,

David

Hi David

Hello, I would also like to ask you, what is the difference between the fixed bound vortex core radius and the initial vortex core radius? I see the manual where it’s 0.6 times the spanwise length. I understand that one can be taken in the calculation. In addition, for a 5MW wind turbine, how many of these two values do you recommend?

In addition, for the value of turbulent vortex viscosity, I know that 1000 is recommended in the manual. I would like to know if it is OK for this value to exceed the stipulated 800.

Do you have any more recommended papers on vortex core radius and turbulent vortex viscosity?

Thank you so much !

Hi Ron,

the bound vortices are the ones located on the blade. The wake vortices are the free vortices that are convected into the wake.

Our implementation is slightly different from OpenFAST, that is the bound vortex core radii are defined as a fraction of the local panel width whereas the wake vortex core radii are defined as a fraction of chord (or panel length). The default values should work for most cases, but there is no clear consensus as to which would be the ideal value as this also depends on a lot of other settings – so dont be afraid of experimenting with these values.

The same applied to the turbulent vortex viscosity, there is no clear consensus as to which is the “best” value. The value that is proposed in QBlade’s GUI is already tuned for the turbine size, so it should be fine, however if you want to “smoothen” your wake (as also mentioned in the OpenFAST manual) dont be afraid to increase this value. You can try increasing by a factor of 10, or even 100 to see its influence – which is not overly dramatic…

BR,

David

robytian has reacted to this post.
robytian

Hi David

I want to get smooth curves, so I turn off Wake Rollup, I turn off Vortex Stretching, and the vortex viscosity is set to 100000, and it works out like the figure below, and I think that might be the reason for the output, because there might be tip vortices somewhere. I would like to know how I should set the wake plane to avoid this.

As you can see in the picture, I want to get the velocity distribution down 1D (126m), and I want it to be smooth.

Best regards,

Tian

 

Uploaded files:
  • You need to login to have access to uploads.

Hi Tian,

to get a smooth velocity distribution you need to average the wake velocities over several timesteps.

To do this you can store the replay of the simulation. This stores the wake vortices at every timestep. After the simulation is finished you can then create velocity distributions at every timestep and average them (click the “+ all next” button in the velocity plane options).

I would strongly suggest to turn the wake roll-up on as otherwise you will get a very unrealistic near wake velocity distribution…

BR,

David

Hi  David

Thanks for your advice, I will open roll up and store replay as you said in the subsequent simulation.

In addition, I would like to discuss some issues with you.

There’s something else I’d like to discuss with you,

First, is it correct to set the wake zones as I understand it? If I set the first one to a large size, it means that near-wake induction is adopted in the whole domain. And the wake zone factors are all set to 1, so I don’t know if that makes sense.(Refer to the figure from the previous question)

Second, in your wake model, roll up is wake self-induction, and Vortex strecting is the stretch term in the vortex equation, including the vortex core growth model, which you have considered. However, regarding the turbulence term in the equation of vortex dynamics, I would like to ask how you consider this problem. The turbulent term I mentioned is not referring to turbulent winds.

Thirdly, this figure is the result of simulation with my own code, which takes into account turbulence, vortex core growth and vortex stretching, etc. The simulation results using Qblade code in the figure did not consider vortex roll up and vortex stretch, but the simulation results obtained by opening these two items are still similar to the overall trend. I do not know why this is. What I want to know is, this is 1D, why is the velocity at the root of the blade so big, and how do I set it so that it doesn’t spread out too much at the blade root?

BR,

roby

Uploaded files:
  • You need to login to have access to uploads.

Hi Roby,

the turbulence is considered by the “turbulent vortex viscosity” term – which is affecting how fast the vortex core is growing over time.

The wake zoning is a technique to reduce the computational requirements of the self-induction step by reducing the number of free vortex elements. The idea is that further away from the rotor the wake can be modeled coarser without loosing too much accuracy. The default settings should be aedequate in most cases – but depending on the focus of your simulations you can play around with these parameters, e.g. inrease the length of the near wake if your main interest are wake dynamics.

I dont really understand your last question, but it looks like in your results the vortex core might be extending into the wake “center” so that it affects the velocity there…

BR,

David

Dear David

Thank you for your answer. What I meant was, why would the velocity at the center of the wake be so high? Does that make sense? If you want the velocity of this position (about blade root) to decrease, do you know how to set it, just like the other result in the figure (The result of my own code in the last problem).

BR,

roby

Hi Roby,

the velocity at the wake center is so large because there is no “rotor” causing induction at the wake center and the nacelle influence is not modeled.

In fact, at the wake center the velocity is even slightly larger than the freestream velocity, caused by the induction of the wake vortices. This is also true due to the continuity equation. In the attached image red regions have higher velocity than the freestream and blue regions have lower velocity.

BR,

David

Uploaded files:
  • You need to login to have access to uploads.
Page 1 of 2Next