Numerical instability when simulating the IEA 22MW turbine in FAST.Farm

[Bartdoekemeijer]

Hi OpenFAST community,

Looking for some help, as I've been breaking my head on this...

Context
For the European SUDOCO project, I am preparing a simple three-turbine simulation in FAST.Farm with the new IEA 22MW wind turbine model. In the end, I'd like to showcase the ZeroMQ functionality in ROSCO that can be used to couple advanced wind farm control algorithms directly to the turbine controller.

The problem
I took the NREL 5MW demonstration case for FAST.Farm and replaced the turbines with the IEA 22MW wind turbine definitions. I then updated the grid and wake settings to what I have seen Kelsey Shaler and others use for the IEA 22MW turbine. Here are my simulation files: fastfarm_iea22mw_monopile.zip. However, the simulation turns unstable instantly.

**************************************************************************************************
 FAST.Farm

 Copyright (C) 2024 National Renewable Energy Laboratory
 Copyright (C) 2024 Envision Energy USA LTD

 This program is licensed under Apache License Version 2.0 and comes with ABSOLUTELY NO WARRANTY.
 See the "LICENSE" file distributed with this software for details.
 **************************************************************************************************

 FAST.Farm-v3.5.3-dirty
 Compile Info:
  - Compiler: GCC version 12.3.0
  - Architecture: 64 bit
  - Precision: single
  - OpenMP: Yes, number of threads: 16/16
  - Date: Apr 26 2024
  - Time: 16:40:44
 Execution Info:
  - Date: 05/29/2024
  - Time: 11:28:26+0200

  Heading of the FAST.Farm input file:
    Sample FAST.Farm input file
 Running AWAE.
 Running InflowWind.
 Running WakeDynamics.
 Running FASTWrapper.
 Running OpenFAST.
 OpenFAST input file heading:
     Generated with AeroElasticSE FAST driver

 Running ElastoDyn.
 Nodal outputs section of ElastoDyn input file not found or improperly formatted.
 Running AeroDyn.
 AD15 Nodal Outputs: Nodal output section of AeroDyn input file not found or improperly formatted.
 Skipping nodal outputs.
    WARNING: WakeMod=11 is a temporary hack. Using new projection method with WakeMod=1.
 >>> ReadBladeInputs: Computing cant angle (BlCrvAng), AeroProjMod=2
 >>> ReadBladeInputs: Computing cant angle (BlCrvAng), AeroProjMod=2
 >>> ReadBladeInputs: Computing cant angle (BlCrvAng), AeroProjMod=2
 Running InflowWind.
 Running ServoDyn.
 Running ServoDyn Interface for Bladed Controllers (using GNU Fortran for Linux).
 Using legacy Bladed DLL interface.
 Running OpenFAST.
 OpenFAST input file heading:
     Generated with AeroElasticSE FAST driver

 Running ElastoDyn.
 Nodal outputs section of ElastoDyn input file not found or improperly formatted.
 Running AeroDyn.
 AD15 Nodal Outputs: Nodal output section of AeroDyn input file not found or improperly formatted.
 Skipping nodal outputs.
    WARNING: WakeMod=11 is a temporary hack. Using new projection method with WakeMod=1.
 >>> ReadBladeInputs: Computing cant angle (BlCrvAng), AeroProjMod=2
 >>> ReadBladeInputs: Computing cant angle (BlCrvAng), AeroProjMod=2
 >>> ReadBladeInputs: Computing cant angle (BlCrvAng), AeroProjMod=2
 Running InflowWind.
 Running ServoDyn.
 Running ServoDyn Interface for Bladed Controllers (using GNU Fortran for Linux).
 Using legacy Bladed DLL interface.

 Farm_Initialize:Farm_ValidateInput:OutFmt produces a column width of 10 instead of 20 characters.
 Farm_Initialize:For efficiency, NumPlanes has been reduced to the number of time steps (41).

  Time: 0 of 60 seconds.

------------------------------------------------------------------------------
Running ROSCO-v2.9.0
A wind turbine controller framework for public use in the scientific field
Developed in collaboration: National Renewable Energy Laboratory
                            Delft University of Technology, The Netherlands
------------------------------------------------------------------------------
Generator speed:    3.0 RPM, Pitch angle:   0.0 deg, Power:     0.0 kW, Est. wind Speed:  11.0 m/s

------------------------------------------------------------------------------
Running ROSCO-v2.9.0
A wind turbine controller framework for public use in the scientific field
Developed in collaboration: National Renewable Energy Laboratory
                            Delft University of Technology, The Netherlands
------------------------------------------------------------------------------
Generator speed:    3.0 RPM, Pitch angle:   0.0 deg, Power:     0.0 kW, Est. wind Speed:  11.0 m/s

 T1:FARM_InitialCO:FWrap_t0:FAST_Solution0:CalcOutputs_And_SolveForInputs:SolveOption2:AD_CalcOutpu
 t:RotCalcOutput:BEMT_CalcOutput(node 3, blade 1):UA_CalcOutput:UA_BlendSteady:Temporarily turning
 off UA due to high angle of attack or low relative velocity. This warning will not be repeated
 though the condition may persist.
 T2:FARM_InitialCO:FWrap_t0:FAST_Solution0:CalcOutputs_And_SolveForInputs:SolveOption2:AD_CalcOutpu
 t:RotCalcOutput:BEMT_CalcOutput(node 3, blade 1):UA_CalcOutput:UA_BlendSteady:Temporarily turning
 off UA due to high angle of attack or low relative velocity. This warning will not be repeated
 though the condition may persist.

 Warning: SkewedWakeCorrection encountered a large value of chi (90.007 deg), so the yaw
 correction will be limited. This warning will not be repeated though the condition may persist.
 See the AD15 chi output channels, and consider turning off the Pitt/Peters skew model (set
 SkewMod=1) if this condition persists.

 Warning: SkewedWakeCorrection encountered a large value of chi (90.007 deg), so the yaw
 correction will be limited. This warning will not be repeated though the condition may persist.
 See the AD15 chi output channels, and consider turning off the Pitt/Peters skew model (set
 SkewMod=1) if this condition persists.


 T1:FARM_UpdateStates:FWrap_Increment:FAST_Solution:FAST_AdvanceStates:AD_UpdateStates:BEMT_UpdateS
 tates:UpdatePhi(node 44, blade 2):BEMT_UnCoupledSolve:There is no valid value of phi for these
 operating conditions: Vx = 3.0007, Vy = 38.691, rlocal = 106.21, theta = -6.28773E-02, geometric
 phi = 7.70969E-02. This warning will not be repeated though the condition may persist. (See
 GeomPhi output channel.)
 T2:FARM_UpdateStates:FWrap_Increment:FAST_Solution:FAST_AdvanceStates:AD_UpdateStates:BEMT_UpdateS
 tates:UpdatePhi(node 44, blade 2):BEMT_UnCoupledSolve:There is no valid value of phi for these
 operating conditions: Vx = 3.0007, Vy = 38.691, rlocal = 106.21, theta = -6.28773E-02, geometric
 phi = 7.70969E-02. This warning will not be repeated though the condition may persist. (See
 GeomPhi output channel.)

 Time: 5 of 60 seconds.  Estimated final completion at 11:58:54.
 T2:FARM_UpdateStates:FWrap_Increment:FAST_Solution:FAST_AdvanceStates:ED_ABM4:ED_CalcContStateDeri
 v:SetCoordSy:Small angle assumption violated in SUBROUTINE SmllRotTrans() due to a large blade
 deflection (ElastoDyn SetCoordSy). The solution may be inaccurate. Simulation continuing, but
 future warnings from SmllRotTrans() will be suppressed.
  Additional debugging message from SUBROUTINE SmllRotTrans(): 4.9834 s

As I didn't get that to work, a colleague shared their existing simulation set-up for the IEA 22MW semi-sub. In that simulation, I use three identical semi-sub IEA 22MW turbines with the same turbine controller and ElastoDyn file, only differing in the ROSCO library file that they call. Here are the set-up files: fastfarm-iea22mw-semi.zip.

--- WIND TURBINES ---
3             NumTurbines   - Number of wind turbines (-) [>=1] [last 6 columns below used only for Mod_AmbWind=2 or 3]
WT_X            WT_Y            WT_Z            WT_FASTInFile   X0_High         Y0_High         Z0_High         dX_High         dY_High         dZ_High        
(m)             (m)             (m)             (string)        (m)             (m)             (m)             (m)             (m)             (m)            
0	0	0	"./wind_turbines/IEA-22-280-RWT-Semi.T1.fst"	-174	-175	1	6	5	5
1400	0	0	"./wind_turbines/IEA-22-280-RWT-Semi.T2.fst"	1226	-175	1	6	5	5
2800	0	0	"./wind_turbines/IEA-22-280-RWT-Semi.T3.fst"	2626	-175	1	6	5	5

I continue to see numerical artefacts in the simulated flowfield on the first turbine in my simulation:

What I find particularly odd is that I use the same turbine definitions and controller set-up across the three turbines. What is even more interesting: if I remove the first turbine from the FAST.Farm set-up file, the problem occurs on the second turbine (then being the first turbine defined).

--- WIND TURBINES ---
2             NumTurbines   - Number of wind turbines (-) [>=1] [last 6 columns below used only for Mod_AmbWind=2 or 3]
WT_X            WT_Y            WT_Z            WT_FASTInFile   X0_High         Y0_High         Z0_High         dX_High         dY_High         dZ_High        
(m)             (m)             (m)             (string)        (m)             (m)             (m)             (m)             (m)             (m)            
1400	0	0	"./wind_turbines/IEA-22-280-RWT-Semi.T2.fst"	1226	-175	1	6	5	5
2800	0	0	"./wind_turbines/IEA-22-280-RWT-Semi.T3.fst"	2626	-175	1	6	5	5

When I swap the order of turbines, the first defined turbine being the one most downstream, the problem moves to the most downstream turbine. This confirms that the problem is really concerned with the turbine that is first defined in the FAST.Farm set-up file:

--- WIND TURBINES ---
3             NumTurbines   - Number of wind turbines (-) [>=1] [last 6 columns below used only for Mod_AmbWind=2 or 3]
WT_X            WT_Y            WT_Z            WT_FASTInFile   X0_High       Y0_High         Z0_High         dX_High         dY_High         dZ_High        
(m)             (m)             (m)             (string)        (m)             (m)             (m)             (m)             (m)             (m)            
2800	0	0	"./wind_turbines/IEA-22-280-RWT-Semi.T3.fst"	2626	-175	1	6	5	5
1400	0	0	"./wind_turbines/IEA-22-280-RWT-Semi.T2.fst"	1226	-175	1	6	5	5
0	0	0	"./wind_turbines/IEA-22-280-RWT-Semi.T1.fst"	-174	-175	1	6	5	5

Software
I use OpenFAST 3.5.3 and ROSCO 2.9.0.

Solutions I have already tried:

• Different turbine locations.
• Different initial conditions for the turbine rotor speed.
• Turbulent inflow. I see similar artefacts still, but mostly masked by the turbulence.
• Higher timestep in the turbine and in the flow field.

Things work just fine for the NREL 5MW simulations. What am I missing here?

[jjonkman]

Dear @Bartdoekemeijer,

I have no experience running the IEA Wind 22-MW RWT myself, but if you are having issues in FAST.Farm at the start of a simulation, I would first verify that you can run the simulation in standalone OpenFAST (uncoupled from FAST.Farm) and that this simulation produces reasonable results. (FAST.Farm will not produce reasonable results if the underlying OpenFAST model has issues.)

Regarding your monopile simulation, it appears that your OpenFAST model is numerically unstable Perhaps this is because of poor initial conditions of rotor speed and blade pitch, which we generally recommend to be initialized to their expected (mean) values conditioned on the mean hub-height wind speed. At the rated wind speed of 11 m/s for the IEA Wind 22-MW RWT, these would be around 7 rpm and 3 deg, respectively, based on my understanding of this model.

Regarding your semisubmersible simulations, I think the wakes of all turbines look a bit off. Again, I would ensure that the OpenFAST model works standalone. Regarding the differences you are seeing between the turbines based on their order, I would guess this is caused by differences in the waves at each turbine location. When Mod_WaveField = 2 (as you have used), then FAST.Farm will take the waves defined for the first wind turbine in the farm and propagate them to the other turbines in the wind farm based on each wind turbine's location.

Best regards,