Yaw DOF in FAST 7

Dear FAST users,

I am trying to get more insight on the yaw system of the 5Mw reference onshore wind turbine.
In order to do that, i downloaded the latest FAST V7 version (Aerodyn 13 included).
In order to set the simulations, I made the following assumptions:

1.- Turbine is UpWind, and rotates clockwise when looking from upwind to downwind.
Although this is very common for UpWind turbines, it is not so clear to me when reading aerodyn and FAST manuals.
2.- Azimut angle increases when turbine rotates clockwise.
3.- Yaw reference coordinate system is defined by vectors d1, d2 and d3
d1 (x pointing downwind = IEC x)
d3 (y pointing left when looking from upwind to downwind = -IEC y)
d2 (z pointing upwards = IEC z)
4.- A positive yaw angle takes place when rotating FROM d1 TO d3.
That is: clockwise is positive when looking the turbine from above.
5.- Variable HorWndDir (FAST) is positive when the wind rotates FROM d1 TO d3.
That is: clockwise when looking the turbine from above.
6.- Variable YawBrMzn is the Action Torque over the yaw bearing according to d2 axis.
That is: Negative when looking the turbine from above.

Next figures tryes to point out these assumptions graphically and the simulations I have tried to set up
in FAST 7.02.

https://www.dropbox.com/s/zu4trmgd6enj361/YawBrMz_due_to_Tilt.png
https://www.dropbox.com/s/ow4lsm145tl3dfg/YawBrMz_due_to_HorWndDir.png
https://www.dropbox.com/s/veowcb3pk6vilxr/YawBrMz_due_to_YawError.png

YawBrMzn behaves as I expected in all cases when setting yawdof to false, meaning:

1.- YawBrMzn is negative due to tilt.
2.- YawBrMzn decreases when HorWndDir goes negative.
3.- YawBrMzn decreases when YawPos goes positive.

So far, all is ok (unless one of the former assumpions were wrong).

Now, when I switch on the yaw control logic, my aim is to evaluate the torque to be supplied by the actuator in order to perform a certain yaw manoeubre. Next figure show the results for variable YawBrMzn in the cases Yawdof = true and false in a yaw manoeuvre.

https://www.dropbox.com/s/sulcjfqy0m9e7n4/YawManoeuvre.png

The first thing I dont understand is that, when setting yawdof = on and YawSpr and YawDamp = 0, the turbine is stable. That is: it achieves yawerror = o for all initial HorWndDir I set as initial condition.
Could you please explain how the phisics work in this case?
The most probable is that I miss something …
Thank you in advance.
Kind regards,
Javier.

Dear Javier,

All of your assumptions are correct except (4) and (5). In FAST v7, positive yaw angle and positive wind direction are defined about the positive zn (d2) axis (counterclockwise when viewed from above). The complication here is that FAST’s module AeroDyn defines the wind direction differently–in AeroDyn, a positive wind direction is defined about the negative zn (d2) axis (clockwise when viewed from above). So, if you input a positive wind direction in AeroDyn’s wind file it will show up as a negative direction output from FAST. See the following forum topic for more information: http://forums.nrel.gov/t/is-positive-yaw-error-clockwise/582/1.

I’m not able to see any of your figures, so, I can’t comment on any specifics. I will just say, though, that free-yaw wind turbines with upwind rotors can remain stably upwind in some wind conditions; however, there are certainly wind conditions where the upwind rotor will yaw downwind. Likewise, it is known that free-yaw downwind rotors can become stuck upwind, potentially to its peril.

Best regards,

Dear Jason,

The free yaw orientation issue is not as simple as I thougth. I have started reading about this and it looks like it depends very much on tilt, coning, operational velocity, wind speed, pitch, implementation of the yawed wake influence, … I mean: you could obtain very different results depending on these items.
I tried to update my previous post with links to the figures I mentioned before, but anyway I think I understand your answer clearly. What I will try to do is mapping the behaviour of YawBrMzn for different misalignements and wind speeds in order to analyze the global behaviour of the system.
Thank you very much for the information.
Kind regards,
Javier.