How to identify OC3-Hywind through control inputs & outputs?

Lets consider OC3-Hywind (In FAST) as a black box with 2 inputs(blade pitch angle and constant gen torque) and several outputs(Blade and tower motions, …).
I want to identify this black box through numerical methods. To do this identification process well, I need to alter the blade pitch angle regularly or maybe irregularly and record the outputs. In this way I will be able to model the black box through its inputs and outputs.
Are there any opinions about how to do this process well?
Suppose we have no disturbances in system(constant wind and still water). In this way the blade pitch alteration must be good as possible to incite the all modes of motion of OC3.
How can I acquire the best Inputs alteration (fluctuations) to incite the all modes of motion of OC3?

Any other comments will be so helpful.


Hi Mehdi,

I think a good reference on these techniques could be found in some published papers:

[1] G. J. van der Veen, J.-W. van Wingerden, P. A. Fleming, A. Scholbrock, and M. Verhaegen, “Global data-driven modeling of wind turbines in the presence of turbulence,” Control Engineering Practice, vol. 21, no. 4, pp. 441–454, Apr. 2013.

[1] P. M. O. Gebraad, J. Van Wingerden, P. A. Fleming, and A. D. Wright, “LPV Identification of Wind Turbine Rotor Vibrational Dynamics Using Periodic Disturbance Basis Functions,” IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY, vol. 21, no. 4, pp. 1183–1190, 2013.

[1] P. M. O. Gebraad, J.-W. van Wingerden, P. A. Fleming, and A. D. Wright, “LPV subspace identification of the edgewise vibrational dynamics of a wind turbine rotor,” in Control Applications (CCA), 2011 IEEE International Conference on, 2011, no. i, pp. 37–42.

Iribas, Mikel, and I. D. Landau. “Closed loop identification of wind turbines models for pitch control.” Proceedings of the 17th Mediteranean Conference on Control and Automation. 2009.

Iribas-Latour, M., and I. D. Landau. “Closed loop identification of wind turbines models for pitch control.” Control and Automation, 2009. MED’09. 17th Mediterranean Conference on. IEEE, 2009.

Iribas‐Latour, M., and I‐D. Landau. “Identification in closed‐loop operation of models for collective pitch robust controller design.” Wind Energy (2012).

Hope these are helpful!


Thanx Paul,

Your mentioned papers are for terrestrial wind turbine, Are 'nt there similar works on floating wind turbine?


Hi Mehdi,

There could be but I’m not personally aware of them. At any rate these algorithms should be applicable.


Dear Paul,
You mean the hydrodynamic part of the system will be regarded as a static part of the system(like aerodynamic part)?
Could you please explain a little more how this method can be generalized on hydrodynamic part of the floating wind turbine ?
Any comments to help in this case will be appreciated.

Hi Mehdi,

I’m sorry, I’m not sure I understand your question. The papers I had sent documented methods for performing system identification of a wind turbine. They identify system models by exciting the system in some way (pitch or torque excitation) during closed-loop operation (it’s dangerous to run a turbine open-loop, that’s why the closed loop). They don’t, in my mind at least, have a dependence on what platform the turbine is mounted on to (land, fixed-bottom, floating) although this will effect the identification.

In terms of hydrodamics, it needn’t be static, these DOFs can be on. Although, perhaps there will be a difficulty in identifying models related to the platform as the available actuation is all up-tower.

Hope this helps!!


Hi Mehdi,

I think the references provided by Paul give quite a good picture of the complexities that play a role in system identification of wind turbines.

Maybe one more reference that might be interesting in the context of floating wind turbines is:

Control of floating wind turbines
G.J. van der Veen, LJ. Couchman, and R.O. Bowyer
2012 American Control Conference

This paper goes into the added complexity of the dynamics of floating turbines, showing that these dynamics are varying with the wind speed for example.

Pieter Gebraad
PhD student
Delft Center for Systems and Control
Delft University of Technology