Additional linear hydrodynamic damping in OC3 Hywind

Hello,

I’m running a number of simulations making use of the OC3 Hywind model as provided on your server for some comparison efforts with alternate mooring models. Reading the “Definition of the Floating System for Phase IV of OC3” document, I noticed “additional linear damping” coefficients have been prescribed for the platform hydrodynamics. I’m having difficulty figuring out where these can be specified. Could anyone kindly explain how these additional damping terms can be applied to the FAST simulation?

Thanks,
Matt

Dear Matt,

Good Point. As you rightly noticed, it is not possible to specify the additional linear damping directly through the FAST or HydroDyn input files.

Instead, to model the OC3-Hywind system in FAST, we’ve compiled a custom version of FAST that includes this additional linear damping as a hardcoded feature. We’ve also included in this customized version of FAST the depth-varying diameter (tapered region) for the viscous drag calculation. We’ve included this customized compilation of FAST in the OC3-Hywind model archive available from here: wind.nrel.gov/public/jjonkman/NR … Hywind.zip.

If you need to recompile FAST yourself for some reason, please let me know and I can tell you the source code changes that are required to accomodate these two customizations.

I hope that helps.

Best regards,

Jason,

Thanks for your fast reply. In the meantime, I had gone ahead and coded in a means of inputting additional platform damping coefficients using a file of the same format as the WAMIT hydrostatic restoring (.hst) file. I’m glad to hear there wasn’t a pre-existing input for specifying the additional damping that I overlooked.

I’ve made a number of additions and modifications to the FAST source code in order to enable coupling with an external dynamic mooring model, so I need a source code solution. (I’ll have my work cut out for me trying to transfer over all my additions once a new version of FAST is released!) I’d be grateful for a description of the code changes required for the variable diameter viscous drag calculations, and to see how you implemented the additional platform damping.

Best regards,
Matt

Dear Matt,

Two model-specific changes (customizations) must be made to HydroDyn for the OC3-Hywind spar buoy. In order to have the viscous drag coefficient depend on the depth-varying platform diameter, the following logic must be added before variable ViscousForce is computed in SUBROUTINE HydroCalc.f90/FltngPtfmLd():

IF ( WaveKinzi0(JNode) > - 4.0 ) THEN
PtfmDiam = 6.5
ELSEIF ( WaveKinzi0(JNode) < -12.0 ) THEN
PtfmDiam = 9.4
ELSE
PtfmDiam = 9.4 + ( 6.5 - 9.4 )*( WaveKinzi0(JNode) + 12.0 )/( -4.0 + 12.0 )
ENDIF

In order to augment the platform damping with additional linear damping in surge, sway, heave, and yaw, the following logic must be added after variable PtfmFt is computed in SUBROUTINE HydroCalc.f90/FltngPtfmLd():

PtfmFt(1) = PtfmFt(1) - 100000.0XD(1)
PtfmFt(2) = PtfmFt(2) - 100000.0
XD(2)
PtfmFt(3) = PtfmFt(3) - 130000.0XD(3)
PtfmFt(6) = PtfmFt(6) - 13000000.0
XD(6)

Also, the torque and pitch control system for this turbine is implemented as a DLL in the style of GH Bladed. So, to run FAST with this controller, the BladedDLLInterface.f90 source file must be used according to the instructions documented in the FAST User’s Guide.

I hope that helps.

Best regards,

Thank you Jason. I did compile my version with the Bladed-style DLL interface. The code changes you provided will be very helpful.

Matt

Hello,

I apologize for posting in such an old post but I have a short question related to this topic.

I was wondering if there is any reason why it was recommended to add linear damping instead of quadratic damping. Is there any advantage in doing so?

BR,

Guido.

Dear Guido,

Just a few comments:

  • Linear damping tends to be more important for low-amplitude motion and quadratic damping tends to be more important for large-amplitude motion.
  • Linear damping is easier to work with, as many types of analyses assume linearity.
  • It is often advantageous to linearize a nonlinear model, which tends to work well for small perturbations about the linearization point.

Best regards,

Dear Jason,

Thank you very much for your reply.

BR,

Guido.