Floating Feedback (and choice of local axis system)

Wind & Water Structural Analysis
1

Dear NREL,

My basic question is whether the tower top acceleration terms supplied as inputs to ServoDyn should be provided in the Tower-Top/Base-Plate Coordinate System or the Nacelle/Yaw Coordinate System. Further context follows.

I am using the IEA 15 MW RWT as a test case. I have stripped the model back a bit so that it just contains a tower and turbine, there is no monopile or semisub etc. In order to investigate the floating feedback term, I am forcing the base of the tower to oscillate back and forth using a displacement-controlled sinusoidal boundary condition. Note I am using my own structural solver which is coupled to OpenFAST for aerodynamics and control, rather than using OpenFAST itself directly, as I find it easier to manipulate external boundary conditions. A steady wind of 10m/s is applied to the model.

Comparing Fl_Mode = 0 (no nacelle velocity feedback) and Fl_Mode = 1 (translational velocity feedback), the blade pitch returned from ServoDyn is as shown below. When Fl_Mode = 0, the blade pitch fluctuates a bit because the wind speed seen by the turbine is affected by the applied back-and-forth motion. When Fl_Mode = 1, ServoDyn attempts to counteract the applied motions to help prevent against negative aerodynamic damping. Pitch keeps increasing or decreasing steadily because the translational acceleration is imposed and therefore unaffected by the blade pitch (this is an academic/artificial test case). This is all fine and makes sense, just trying to give some context to the question.

image
image1134×553 67.6 KB

Now consider 3 different models:

  1. Model set up at 0 degrees. 10m/s steady wind applied at 0 degrees.
  2. Tower rotated by 90 degrees (e.g. simulation of a semi-sub yaw scenario). Note that the turbine is also rotated with tower. 10m/s steady wind applied at 90 degrees.
  3. Turbine rotated by 90 degrees (e.g. simulation of a turbine yaw scenario). Note that the tower itself does not rotate. 10m/s steady wind applied at 90 degrees.

The key inputs being passed to ServoDyn from the structural solver are as follows:

!> \* Record 53: Tower top fore-aft     acceleration (m/s^2) \[SrvD input\]

u_SrvD%YawBrTAxp

!> \* Record 54: Tower top side-to-side acceleration (m/s^2) \[SrvD input\]

u_SrvD%YawBrTAyp

The axis systems used in OpenFAST are discussed in the following document: https://ebranlard-openfast.readthedocs.io/en/f-ff-doc/_downloads/d8bd014121d6505cb25cf49bee5eaa80/Old_FAST6_UsersGuide.pdf. The document dates from 2005 but I am assuming it is still valid.

  1. Tower-Top/Base-Plate Coordinate System: Fixed to the top of the tower. It translates and rotates as the platform moves and the tower bends, but it does not yaw with the nacelle.
  2. Nacelle/Yaw Coordinate System: Translates and rotates with the top of the tower, plus it yaws with the nacelle

I would assume that when Fl_Mode = 1, the results from Models 1, 2 and 3 should be very similar if not identical.

  1. If the tower top acceleration terms are supplied in the Tower-Top/Base-Plate Coordinate System, results from Models 1 and 2 are in agreement, but not Model 3 (which looks like the Fl_Mode=0 case).
  2. If the tower top acceleration terms are supplied in the Nacelle/Yaw Coordinate System, results from Models 1, 2 and 3 are all in agreement.

So my basic question is whether the tower top acceleration terms supplied as inputs to ServoDyn should be provided in the Tower-Top/Base-Plate Coordinate System or the Nacelle/Yaw Coordinate System. Instinctively one might expect the former, but results would suggest the latter.

Any advice appreciated.

Kind regards,

Aengus.

2

Dear @Aengus.Connolly,

I’ll just say that OpenFAST passes the tower-top translational accelerations to the DISCON controller through avrSWAP(53-54) in the tower-top/base-plate coordinate system, which does not yaw with the nacelle. The nacelle-yaw angle is itself passed to the DISCON controller through avrSWAP(37), which you could use to convert the translational accelerations to a coordinate system that yaws with the nacelle within the DISCON controller if so desired.

Best regards,

3

Thank you Jason, that makes sense :+1: