Blade Deflections show High Frequency Response

Hi Jason/All,

I’m interested in examining blade structural responses. I created a simple test case, using the IEA-15-240-RWT-UMaineSemi files as a starting point. I made the following minor modifications:

  1. Fixed all platform motions, to allow me focus purely on the blades/tower (PtfmSgDOF etc. = False)
  2. Preventing the rotor from spinning, to keep things simple, no changes in relative wind velocity due to blade rotation (GenDOF=False, RotSpeed=0). Rotor facing wind (YawDOF=False).
  3. Blade pitch fixed at 30degs (BlPitch(1) etc. =30).
  4. Disabled ServoDyn (CompServo=0), to simulate steady conditions without any modifications stemming from control.
  5. Because rotor is not spinning, I disabled wake induction (WakeMod=0) and selected steady aerodynamics (AFAeroMod=1).

I am running OpenFAST V2.6.0.

I realise the case is somewhat artificial as in reality the rotor would be spinning for wind speeds above cut-in. Regardless:

  1. When a steady wind of 10m/s is applied, blade tip deflections quickly settle to steady state.
  2. For higher wind speeds (e.g. 20m/s), there seems to be some high frequency noise in the tip deflections. Suggests the blade is vibrating at high frequency.

Just wondering is this expected behaviour? I’ve seen a similar pattern in a more complicated model, this test case is designed to demonstrate the issue in a simple way. Perhaps I’m missing something, like structural damping. Do you have any suggestions?

PS: the plots are presented in an external tool but the data is taken from the OpenFAST out file.

Kind regards,

Dear @Aengus.Connolly,

By “vibrating at high frequency”, can you confirm which frequency this is? I would guess the frequency corresponds to the blade edgewise-bending natural frequency. This would suggest that the issue you are facing is the classic edgewise instability brought about by negative damping introduced in deep stall, e.g., as discussed in the following forum topic: Designing for yaw errors using FAST.

Best regards,

Hi Jason,

Thanks for your reply.

It is as you suspected, numerical instability in the edgewise direction, as outlined in the linked post. A small increase to damping in this degree of freedom (via BldEdDmp1 parameter) greatly reduces the effect.

My current model is somewhat of an artificial test case, but it’s great to have an understanding of the root cause for when I come across it again in real-world scenarios.

Thanks again.

Kind regards,

1 Like