search

IEAWindTask37 / IEA-10.0-198-RWT

Apache License 2.0
39 stars 25 forks source link

OpenFAST wind ramp simulation: pitch oscillations and rated wind speed not matching design specs #38

Open Guiiem opened 6 days ago

Guiiem commented 6 days ago

Hello!

I have tried a wind ramp simulation by setting WindType to 2 in the InflowFile and using the following time-series:

! Time     Wind    Wind    Vertical    Horiz.      Pwr.Law     Lin.Vert.   Gust     Upflow
!          Speed   Dir     Speed       Shear       Vert.Shr    Shear       Speed    Angle 
! (sec)    (m/s)   (Deg)   (m/s)                                            (m/s)   (deg)
0.00 3 0 0 0 0.2 0 0 8
1149.95 25 0 0 0 0.2 0 0 8

All the rest is left as it is in the last commit of master. I am using OpenFAST 3.5.0 and ROSCO 2.7.0.

After rated rotor speed is reached, there are some pitch oscillations. Even though they are small, the influence in power is quite large. Additionally, rated power is achieved at around 12 m/s, which is more than the 10.75 m/s documented in the Table 16 of the design report. I confirmed this in another steady simulation at 11 m/s with no upflow or shear, where I only got a generator power of around 9 MW.

windramp

Am I missing something?

Thank you! Guillem

ptrbortolotti commented 4 days ago

Hello @Guiiem ,

I'm not too sure what's going on. I've just used the latest release of https://github.com/WISDEM/weis to retune ROSCO v2.9.0. I attach the DISCON. The gen power is 10MW, let me know if this seems to work

Archive.zip

Guiiem commented 3 days ago

Hi @ptrbortolotti ,

Thanks for the reply and all your work with these models! I have tried the new files with the wind ramp and the oscillations are still there: windramp_retuned

If pitch saturation is removed (PS_Mode=0) the rated speed matches better the expected power curve, but the oscillations are still there. If I test it in a wind step simulation (still no saturation), it looks quite stable, even though the behavior at the 11m/s jump is a bit strange: windstep

By the way, the wind-ramp simulation of the IEA3.4MW cousin looks smooth and without oscillations, this difference is what made me curious what was going on here.

  1. What do you think about the pitch saturation? I can see that it changed in commit 059caf9. Before that, the pitch saturation used to start after 10 m/s, and now it starts around 9 m/s, below rated.
  2. What do think about the pitch oscillations? The controller looks stable in the wind-step and in single steady wind simulations, but I guess it is still not ideal that these oscillations appear in a wind ramp?

Thank you! Guillem