Discrepancy with OpenFAST blade loads

[AlexGuntner]

Dear @jjonkman,

I am conducting LES simulations using AMR-Wind coupled with OpenFAST, employing the IEA 15 MW turbine. Currently, I am calibrating the projection smearing function (epsilon) and using the recommended values from the AMR-Wind documentation.

For easier comparison with other studies, I aim to use a simplified wind turbine configuration—excluding the tower, shaft tilt, and pre-coning, with all DOFs disabled except GenDOF = True. I have set these values accordingly in the AeroDyn and ElastoDyn modules.

However, I am encountering discrepancies between results obtained from standalone OpenFAST and those from OpenFAST integrated with AMR-Wind, where velocity is sampled from the CFD simulation.

As shown in the figure, OpenFAST alone produces minimal differences in axial and tangential curves across configurations. In contrast, the AMR-Wind coupled results exhibit unusual oscillations, particularly in the tangential curve.

When I use the default configuration, with all DOF=True, the curves match pretty well:

Instead with everything but GenDOF set to false:

Could you share your thoughts on whether this behaviour indicates a potential bug in the AMR-Wind coupling or another underlying factor? Does OpenFAST make specific assumptions regarding pre-coning, shaft tilt, or similar parameters that might explain this?

Kind regards

[jjonkman]

Dear @AlexGuntner,

Given that your question is more about AMR-Wind than OpenFAST, I would suggest posting your question in the AMR-Wind repository: https://github.com/Exawind/amr-wind.

Best regards,

[AlexGuntner]

Dear @jjonkman,

I have raised this issue in the AMR-Wind repository (see here), but I have not received an answer so far.

Regarding the OpenFAST side of the question, I was expecting to see larger differences in load curves between wind turbine configurations, as presented in my first figure. I would like to kindly ask you if OpenFAST makes any specific simplifications when calculating aerodynamic forces with pre-coning and shaft tilt set to True.

Kind regards

[jjonkman]

Dear @AlexGuntner,

Which version of OpenFAST are you using? The influence of precone and tilt on the solution will be affected by the aerodynamic options you've selected in AeroDyn; which settings have you selected? Certainly though, looking at the mean over the azimuth will smear out the main effect of tilt, which is the variation with azimuth angle.

Best regards,

[AlexGuntner]

Dear @jjonkman,

I’m using the latest version of OpenFAST (v4.0.2)

For the standalone runs I used these aerodyn settings:

------- AERODYN v15.03.* INPUT FILE ------------------------------------------------
IEA 15 MW Offshore Reference Turbine
====== General Options ============================================================================
False Echo - Echo the input to ".AD.ech"? (flag)
Default DTAero - Time interval for aerodynamic calculations {or "default"} (s)
1 Wake_Mod - Wake/induction model (switch) {0=none, 1=BEMT, 3=OLAF} [Wake_Mod cannot be 2 or 3 when linearizing]
1 TwrPotent - Type tower influence on wind based on potential flow around the tower (switch) {0=none, 1=baseline potential flow, 2=potential flow with Bak correction}
1 TwrShadow - Calculate tower influence on wind based on downstream tower shadow (switch) {0=none, 1=Powles model, 2=Eames model}
True TwrAero - Calculate tower aerodynamic loads? (flag)
False CavitCheck - Perform cavitation check? (flag) [UA_Mod must be 0 when CavitCheck=true]
False Buoyancy - Include buoyancy effects? (flag)
False NacelleDrag - Include Nacelle Drag effects? (flag)
False CompAA - Flag to compute AeroAcoustics calculation [used only when Wake_Mod = 1 or 2]
AeroAcousticsInput.dat AA_InputFile - AeroAcoustics input file [used only when CompAA=true]
====== Environmental Conditions ===================================================================
"default" AirDens - Air density (kg/m^3)
"default" KinVisc - Kinematic viscosity of working fluid (m^2/s)
"default" SpdSound - Speed of sound in working fluid (m/s)
"default" Patm - Atmospheric pressure (Pa) [used only when CavitCheck=True]
"default" Pvap - Vapour pressure of working fluid (Pa) [used only when CavitCheck=True]
====== Blade-Element/Momentum Theory Options ====================================================== [unused when Wake_Mod=0 or 3, except for BEM_Mod]
1 BEM_Mod - BEM model {1=legacy NoSweepPitchTwist, 2=polar} (switch) [used for all Wake_Mod to determine output coordinate system]
--- Skew correction
1 Skew_Mod - Skew model {0=No skew model, -1=Remove non-normal component for linearization, 1=skew model active}
False SkewMomCorr - Turn the skew momentum correction on or off [used only when Skew_Mod=1]
1 SkewRedistr_Mod - Type of skewed-wake correction model (switch) {0=no redistribution, 1=Glauert/Pitt/Peters, default=1} [used only when Skew_Mod=1]
default SkewRedistrFactor - Constant used in Pitt/Peters skewed wake model {or "default" is 15/32*pi} (-) [used only when Skew_Mod=1 and SkewRedistr_Mod=1]
--- BEM algorithm
True TipLoss - Use the Prandtl tip-loss model? (flag) [unused when Wake_Mod=0 or 3]
True HubLoss - Use the Prandtl hub-loss model? (flag) [unused when Wake_Mod=0 or 3]
True TanInd - Include tangential induction in BEMT calculations? (flag) [unused when Wake_Mod=0 or 3]
True AIDrag - Include the drag term in the axial-induction calculation? (flag) [unused when Wake_Mod=0 or 3]
True TIDrag - Include the drag term in the tangential-induction calculation? (flag) [unused when Wake_Mod=0,3 or TanInd=FALSE]
Default IndToler - Convergence tolerance for BEMT nonlinear solve residual equation {or "default"} (-) [unused when Wake_Mod=0 or 3]
500 MaxIter - Maximum number of iteration steps (-) [unused when Wake_Mod=0]
--- Shear correction
False SectAvg - Use sector averaging (flag)
1 SectAvgWeighting - Weighting function for sector average {1=Uniform, default=1} within a sector centered on the blade (switch) [used only when SectAvg=True]
default SectAvgNPoints - Number of points per sectors (-) {default=5} [used only when SectAvg=True]
default SectAvgPsiBwd - Backward azimuth relative to blade where the sector starts (<=0) {default=-60} (deg) [used only when SectAvg=True]
default SectAvgPsiFwd - Forward azimuth relative to blade where the sector ends (>=0) {default=60} (deg) [used only when SectAvg=True]
--- Dynamic wake/inflow
2 DBEMT_Mod - Type of dynamic BEMT (DBEMT) model {0=No Dynamic Wake, -1=Frozen Wake for linearization, 1:constant tau1, 2=time-dependent tau1, 3=constant tau1 with continuous formulation} (-)
29.03 tau1_const - Time constant for DBEMT (s) [used only when Wake_Mod=2 and DBEMT_Mod=1]
====== OLAF -- cOnvecting LAgrangian Filaments (Free Vortex Wake) Theory Options ================== [used only when Wake_Mod=3]
../IEA-15-240-RWT-OLAF/IEA-15-240-RWT_OLAF.dat OLAFInputFileName - Input file for OLAF [used only when Wake_Mod=3]
====== Unsteady Airfoil Aerodynamics Options ====================================================
True AoA34 - Sample the angle of attack (AoA) at the 3/4 chord or the AC point {default=True} [always used]
3 UA_Mod - Unsteady Aero Model Switch (switch) {0=Quasi-steady (no UA), 2=B-L Gonzalez, 3=B-L Minnema/Pierce, 4=B-L HGM 4-states, 5=B-L HGM+vortex 5 states, 6=Oye, 7=Boeing-Vertol}
True FLookup - Flag to indicate whether a lookup for f' will be calculated (TRUE) or whether best-fit exponential equations will be used (FALSE); if FALSE S1-S4 must be provided in airfoil input files (flag) [used only when UA_Mod>0]
3 IntegrationMethod - Switch to indicate which integration method UA uses (1=RK4, 2=AB4, 3=ABM4, 4=BDF2)
0.1 UAStartRad - Starting radius for dynamic stall (fraction of rotor radius [0.0,1.0]) [used only when UA_Mod>0; if line is missing UAStartRad=0]
1

Best regards

[jjonkman]

Dear @AlexGuntner,

For preconed rotors (or situations with blade large deflection), I recommend setting BEM_Mod = 2, which should provide improved accuracy compared to the legacy solution.

For tilted rotors (or other cases with skewed flow), I recommend setting Skew_Mod = 1 with SkewMomCorr = TRUE, which should provide improve accuracy compared to legacy solution.

Best regards,

[AlexGuntner]

Dear @jjonkman,

I’ve run a set of simulations using the settings you recommended for the IEA15MW wind turbine with the default configuration (DOF=True, ShaftTilt = -6.0, and PreCone = -4.0). As shown in the plot below, enabling both BEM_Mod = 2 and Skew_ModCorr = True results in a noticeable spike near the blade tips in both axial and tangential loading. This behavior is absent when these options are disabled (running the exact files uploaded from the IEA15MW git repository). It almost seems like the Prandtl tip-loss correction is not functioning properly when skew correction is active.

I observe a similar pattern when using OLAF in a simple configuration (no shaft tilt, no precone, no blade elasticity), despite configuring the wake length and other parameters according to the OpenFAST documentation. This behaviour can also be seen in the recent study by Trigaux et al, Figure 4 (1).

I'd appreciate your thoughts on this—also regarding the results of OLAF.

Kind regards

[AlexGuntner]

@jjonkman do you have any thoughts on this?