Questions about inflow wind

Wind & Water Controls
1

Deal all,

I have some questions about the wind speed.

  1. I noticed that there are 9 wind points in total for wind speed measurement, how could I figure the exact positions of them, could I change the position by myself?

    image
    image799×177 1.58 KB

  2. For a closed loop control scheme, if the Uop depends on a lookup table based on steady wind speed, which wind speed should be referred to? Should I pick WindVx in several points for the mean value or just the Point 1?

  3. How could I generate Turbulent wind with varying mean wind speed?

  4. What are the following outputs used for?

    image
    image759×124 1.96 KB

Thank you very much for your kind volunteering and precious time, looking forward to your reply.

Best regards,
Yarong

2

Dear @Yarong.Zhou,

Here are my answers:

  1. The coordinates of up to 9 point where wind data may be output from InflowWind are specified by the user via inputs WindVxiList, WindVyiList, WindVziList.

  2. It is not common for a control system to be based on a tabulated look up of steady wind speed, which would not be measurable. For those control systems that do rely on wind speed, typically a wind speed estimator is used (based on rotor speed and pitch angle). A nacelle-mounted anemometer is often used as well as an input to a turbine control system.

  3. TurbSim can be used to develop full-field turbulent wind inflow data that be can used by InflowWind.

  4. I’m actually not too familiar with these InflowWind outputs, but my understanding is that they are tied to Lidar functionality built within the Simulink interface (documented a bit here: Linking measured wind data from LIDAR to Turbsim). Please note that a new and more powerful Lidar module is being developed and interface to OpenFAST to replace this feature.

Best regards,

3

Dear Jason,

Thanks for your detailed reply. Here are my feedbacks.

  1. I found the descriptions of Wind Point in InflowWind input file, but I didn’t find WindList. How can I get the information about WindList?

    image
    image739×184 6.63 KB

  2. Many controllers are implemented with set points due to the perturbation linear model. (Q1: Generally, how do they choose the OP? according to where they get the linear model?) When they are validated, turbulent wind with constant mean wind speed are usually adopted. (Q2: In cases like this, the OP will not be changed, right? ) But in natural environment, how can we define a mean wind speed of inflow wind for a wind turbine? How to divide time and zone? Or can I take the view that if the mean wind speed doesn’t change, there is no need to change the OP?
    By the way, another confusion also occurred after the discussion with my colleagues working on aerodynamic. When we linearize the fast model, we used steady wind speed, and then we also get OP here. But the “wind speed” which affects the force should be a weighting sum related to blade, in turbulent cases, things may not like that in steady case. Is previous OP still effective enough?

  3. I have ever tried TurbSim, but what it generated is with constant mean wind speed.

    image
    image943×330 17.2 KB

    image
    If I want turbulent wind with varying mean wind speed (whatever step or ramp) , is there any change that I can make in TurbSim.inp file?

Thanks for your kind volunteering. I’m sorry the second part is a bit disordered, if you cannot get the point, I will make it clearer next time. Looking forward to your reply

BR,
Yarong

4

Dear @Yarong.Zhou,

Regarding (1), InflowWind module input parameters WindVxiList, WindVyiList, WindVziList are specified by the user in lines 9-11 of the InflowWind input file (as of OpenFAST v.3.2). See the InflowWind User’s Guide for more information: 4.2.9. InflowWind Users Guide and Theory Manual — OpenFAST v3.3.0 documentation.

Regarding (2), it is common for a wind turbine controller to be based on linear models, but typically multiple linear models are generated at different operating points (OP) associated with different mean wind speeds and gain-scheduling (often based on pitch angle) is used to transition across the range of conditions. Time-domain simulations of wind turbines typically involve running distinct turbulent wind simulations at different mean wind speeds, with each simulation being run for around 10 minutes and multiple turbulence seeds used at each mean wind speed. For example, covering the full operational range of a wind turbine with a cut-in wind speed of 3 m/s and a cut-out wind speed of 25 m/s could involve 66 10-minute simulations, i.e., splitting the wind speed range into 11 bins and running 6 unique simulations (6 unique seeds) within each wind speed bin (at mean wind speeds of 4 m/s, 6 m/s, …, 24 m/s). I’m not sure I understand your last question in (2), but certainly the linearized model does not include the nonlinear dynamics and full-field turbulent excitation captured by the full nonlinear model.

Regarding (3), again, wind turbine simulations with turbulence are typically run about a given mean wind speed and separate simulations are run at different mean wind speeds. Related questions have been discussed previously on this forum, e.g., see: Stitch Together TurbSim Timeseries.

Best regards,