Rotor Structural Loads Analysis Using Coupled Computational Fluid

Dynamics/Computational Structural Dynamics

Hyeonsoo Yeo Mark Potsdam

US Army Aviation Development Directorate − AFDD
Aviation & Missile Research, Development & Engineering Center
Research, Development, and Engineering Command
Ames Research Center, Moffett Field, California

ABSTRACT
Coupled CFD/CSD (RCAS/Helios and CAMRAD II/Helios) analyses are performed and the calculated rotor struc-
tural loads are compared with the flight test data obtained from the NASA/Army UH-60A Airloads Program. Three
challenging level flight conditions are investigated: 1) high speed with advancing blade negative lift, 2) low speed
with blade-wake interaction, and 3) high thrust with dynamic stall. The predicted flap bending and torsion moments,
pitch link and lag damper loads, in general, show reasonably good correlation with the test data. A nonlinear lag
damper model is essential for the accurate prediction of root chord bending moment and lag damper load. Both
analyses, however, significantly underpredict the chord bending moments, especially the 4/rev harmonic amplitude.
Parametric study shows that blade stiffness variations have only a small influence on the loads calculations. However,
modal damping in the first flap mode has a significant influence on the flap bending moments. Inclusion of a simple
one degree-of-freedom drivetrain model shows the potential importance of high frequency drivetrain modes for the
accurate prediction of the 4/rev chord bending moments and a need to develop a realistic drivetrain model.

NOTATION estimated fatigue design loads (oscillatory rotor loads). Of-

ten, these design loads are scaled from previous flight test
A rotor disk area, π R2 databases, as calculations are not trustworthy. The rotor loads
CT rotor thrust coefficient, T /ρ (ΩR)2 A that are transmitted to the airframe via the rotor hub and
CW weight coefficient, GW /ρ (ΩR)2 A swashplate are the dominant source of helicopter vibration.
GW gross weight Vibration reduction devices, active and passive, are used to
Lf fuselage lift meet very stringent vibration requirements. However, their
Ls stabilator lift cost and weight penalty have been excessive in part because of
LT R tail rotor lift inadequate vibration prediction capability. Accurate predic-
Mx hub roll moment (positive right down) tion capability of rotor loads and vibration at an early design
My hub pitch moment (positive nose up) stage has the potential to significantly reduce costly modifica-
R rotor radius tions, additional testing, weight penalties, and overall perfor-
TMR main rotor thrust mance degradation.
V∞ free-stream velocity
Rotor loads and vibration analysis is a challenging multi-
αs shaft angle (positive for rearward tilt) disciplinary problem due to coupling of the complex structural
µ advance ratio, V∞ /ΩR
deformations of rotor blades with the three-dimensional and
ρ free-stream density highly unsteady aerodynamic environment. In recent years,
σ solidity
there has been significant progress in rotorcraft aeromechan-
Ω rotor angular rotation rate
ics prediction capability using coupled computational fluid
dynamics (CFD) / rotorcraft computational structural dynam-
INTRODUCTION ics (CSD) analyses (Refs. 1–5). The CFD methods, which
use a high fidelity, Navier-Stokes, overset grid methodology
Accurate prediction of rotor blade loads and vibration is es-
with first principles-based wake capturing, overcame the lim-
sential for the successful design of rotorcraft. In the de-
itations of the conventional lifting line aerodynamics used in
sign of rotor dynamic components, fatigue life assessment
rotorcraft comprehensive codes. The CSD methods (through
of the components depends heavily on the accuracies of the
a comprehensive code) performs the sophisticated structural
Presented at the AHS 70th Annual Forum, Montréal, Québec, dynamics modeling and also carries out trim calculations.
Canada, May 20–22, 2014. This is a work of the U.S. Gov- Coupling a CFD code to a comprehensive code marries the
ernment and is not subject to copyright protection in the U.S. strengths of the two approaches and produces the highest fi-
1
delity solution currently possible. (11.3% radius) and then evenly distributed along the blade
The present authors demonstrated improved airloads pre- at 10% increments of the rotor radius (20% − 90%). Flap
diction capability using a loosely coupled CFD/CSD method bending moments were measured at all nine radial locations.
in three challenging level flight conditions of the NASA/Army Chord bending moments were not measured at 90%R and tor-
UH-60A Airloads Program (Ref. 1): 1) high speed, µ = 0.37, sion moments were measured only at 30%R, 50%R, 70%R,
CT /σ = 0.081 (C8534), with advancing blade negative lift, and 90%R. Additional instrumentation included strain gauges
2) low speed, µ = 0.15, CT /σ = 0.076 (C8513), with blade- on the pitch links, dampers, and rotor shaft.
wake interaction, and 3) high thrust, µ = 0.24, CT /σ = 0.129 All pressure signals were filtered using 550 Hz low-pass
(C9017), with dynamic stall. Blade section normal force and 6- pole Butterworth filters and digitized at a rate of 2142 sam-
pitching moment magnitudes and phases are accurately cap- ples/sec/channel. The non-pressure signals were filtered us-
tured in the coupled solutions, overcoming the inaccuracies of ing 110 Hz low-pass 6-pole Butterworth filters and digitized
airloads prediction using comprehensive analysis alone. Pre- at a rate of 357 samples/sec/channel. For a typical level-flight
diction of rotor airloads for these three flight conditions have test condition, a 5 second time slice (approximately 19 revo-
been performed by many researchers using various combina- lutions) was stored in the database. These data are stored in
tions of CFD/CSD tools (Refs. 6–9). All the coupled analyses, the master TRENDS database at the NASA Ames Research
in general, show satisfactory airloads correlation with the test Center.
data. For comparison with analyses, the raw data in the
Rotor structural loads also have been investigated for these TRENDS database needed to be post-processed. The first step
conditions by several researchers (Refs. 10–12). In general, was a zero azimuth reference correction. The zero azimuth
the prediction of the structural loads did not show the same reference for the TRENDS database aligned the center of the
level of correlation with the flight test data as observed for the rotor hub, the center of the elastomeric bearing, points on the
predicted airloads. blade quarter-chord, and the rotating beacon light on the tail.
In the present paper, both coupled In order to define the zero azimuth reference to be parallel to
RCAS (Ref. 13)/Helios (Ref. 14) and the pitch axis, a negative 7-deg blade azimuth shift was ap-
CAMRAD II (Ref. 15)/Helios analyses are performed plied to all azimuthally dependent data (Ref. 18).
and rotor structural loads results are compared with data from The second step was a correction for the signal delay
the same three level flight conditions of the NASA/Army caused by the antialiasing filters. Assuming an ideal 110 Hz
UH-60A Airloads Program. Calculated blade flap and chord low-pass 6- pole Butterworth filter, the nominal group delay
bending moments, torsion moments, pitch link loads, and lag (time shift) of the signal as it goes through the filter is 0.00558
damper loads are compared with the flight test data. Detailed sec (group delays are actually slightly non-linear for this filter
time history and harmonic responses are examined to assess type). With the standard flight rotor speed of 258 RPM, this
the high fidelity analysis codes’ accuracy in the calculation of results in an azimuthal delay of approximately 8.6 deg. This
rotor structural loads. Effects of a nonlinear damper model, azimuthal correction was applied to all the non-pressure chan-
blade stiffness, structural damping, and drive train dynamics nels from the flight test, including the shaft bending gauge.
on the rotor structural loads are also investigated. For the 550 Hz filters used for the pressure channels, the
equivalent delay was 0.00112 sec, resulting in a correction
of 1.7 deg.
DESCRIPTION OF THE TEST
The first correction has been applied to the test data since
The test data used in the present study were obtained during January 2005 (before 2005, the correction was applied in the
the NASA/Army UH-60A Airloads Program conducted from wrong direction). The second correction was first applied in
August 1993 to February 1994 (Ref. 16). The database pro- 2012 for the comparison between the full-scale UH-60A Air-
vides aerodynamic pressures, structural loads, control posi- loads wind tunnel and flight test data (Ref. 19). For this study,
tions, and rotor forces and moments, allowing for the valida- both corrections were applied to the test data.
tion of both aerodynamic and structural models. The steady
level flight test matrix contains a range of advance ratios
and gross weight coefficients as shown in Fig. 1. The three
DESCRIPTION OF ANALYTICAL METHOD
test points investigated here and an aerodynamic rotor lift The analytical results were obtained using coupled
boundary obtained in wind tunnel testing of a model rotor by RCAS/Helios and CAMRAD II/Helios. This section
McHugh (Ref. 17) are also plotted. describes each method and how they are coupled to produce
Two of the blades were heavily instrumented: one with a higher fidelity solution.
subminiature pipette-type pressure transducers and one with
a mix of strain-gauges and accelerometers. Absolute pres- RCAS and CAMRAD II comprehensive analyses
sures were measured at nine radial locations. Blade flap bend-
ing, chord bending, and torsion moments were measured with Rotorcraft Comprehensive Analysis Systems (RCAS) is a
two- or four-leg strain-gauge bridges bonded to the second in- comprehensive multidisciplinary, computer software system
strumented blade. The gauges were located at the blade root for predicting rotorcraft aerodynamics, performance, stability
2
and control, aeroelastic stability, loads, and vibration. RCAS program sponsored by the DoD High Performance Comput-
is capable of modeling a wide range of complex rotorcraft ing Modernization Office (Ref. 14). Helios uses an innova-
configurations operating in hover, forward flight, and maneu- tive dual-mesh paradigm that employs unstructured meshes
vering conditions. The RCAS structural model employs a in the “near-body” close to the surface to capture the wall-
hierarchical, finite element, multibody dynamics formulation bounded viscous effects and structured Cartesian grids in
for coupled rotor-body systems. It includes a library of prim- the “off-body” to resolve the wake through a combina-
itive elements including nonlinear beams, rigid body mass, tion of higher-order algorithms and adaptive mesh refine-
rigid bar, spring, damper, hinges and slides to build arbitrarily ment (AMR). An overset procedure facilitates the data ex-
complex models. RCAS has been used recently for perfor- change and also enables the relative motion between the two
mance and loads correlation of various rotors including the meshes. The parallel domain connectivity solver PUNDIT
UH-60A (Refs. 20–22). automatically handles the data exchange between the two
meshes. CFD is loosely coupled with CSD solvers (RCAS or
Comprehensive Analytical Model of Rotorcraft Aero-
CAMRAD II) to solve the rotorcraft structural dynamics and
dynamics and Dynamics (CAMRAD) II is an aerome-
trim. A lightweight Python-based software integration frame-
chanics analysis of rotorcraft that incorporates a combina-
work handles the data exchange between the modules.
tion of advanced technologies including multibody dynam-
ics, nonlinear finite elements, and rotorcraft aerodynamics. The near-body unstructured solver NSU3D is a node-
CAMRAD II has been used extensively for correlation of per- centered, finite-volume-based unsteady Reynolds-Averaged
formance and loads measurements of the UH-60A in various Navier-Stokes (URANS) solver that is spatially second-order
flight conditions (Refs. 23–26). accurate and is capable of handling mixed elements. Time-
accurate computations utilize a 2nd-order backwards-Euler
Both analyses include multiple aerodynamic options for time stepping scheme along with dual-time stepping for con-
airloads, wake-induced flow fields, and aerodynamic interfer- vergence of the nonlinear problem at each physical time step.
ence. Airloads models include two-dimensional (2D) airfoil The Spalart-Allmaras turbulence model is used. The struc-
and lifting-line models for rotor blade, wings, and 3D airloads tured solver SAMARC is used for the Cartesian off-body grid
for bodies. system. SAMARC solves the inviscid Euler equations using a
5th-order spatial discretization scheme and 3rd-order explicit
Structural modeling Runge-Kutta time integration scheme.
The computational grids model the standard UH-60A
The UH-60A Black Hawk is modeled in comprehensive anal- blade geometry. The unstructured blade mesh was generated
yses as well as CFD as an isolated rotor, not as a complete air- using GRIDGEN and AFLR3 software. The 4 rotor blade
craft. The structural model of the rotor employed in this study grids have 15.4 million nodes and 36.5 million cells. The
is developed from a common UH-60A master input database, finest off-body spacing is 5% chord with a fixed refinement
which was developed by Yeo and has been used extensively region surrounding the rotor plane; AMR is not used here.
for studies of the UH-60A (Refs. 7, 10, 23, 27). Detailed rotor The off-body grid contains 146 million unblanked grid points
pitch control system linkage geometry, stiffness, and nonlin- on 8 levels. Details of the flow solvers, input parameters, and
ear lag damper are also incorporated. To assess the structural grids are available in Refs. 4 and 14.
dynamic modeling, Ho et al. specified measured airloads from
flight test as prescribed external loads and then compared the Coupling procedure and trim
resulting response with the measured response (Ref. 28). The
close agreement between RCAS and CAMRAD II predictions The CFD/CSD coupling procedure uses the standard loose
and between the calculations and test data provides significant or “delta” coupling approach (e.g. Refs. 1 and 2). At
confidence in the structural dynamics modeling and analysis each coupling iteration the aerodynamic loads calculated by
methodology of the two codes. CFD are passed to CA (Comprehensive Analysis − RCAS
or CAMRAD II). After trimming with the CFD airloads, CA
Figure 2 compares the blade in-vacuo natural frequen- computes the blade deflections relative to the blade frame of
cies calculated by RCAS and CAMRAD II. The frequencies reference and passes them back to CFD. This sequence is re-
shown here are for a nominal zero collective pitch with very peated until the airloads, deflections, and control angles con-
small structural damping. The frequency predictions by the verge, typically in about 3-4 revolutions, depending on flight
two comprehensive codes show good agreement. There are conditions. The exchange of information between CFD and
strong couplings between modes for the fourth to sixth modes CA codes are handled by the Rotor Fluid Structure Interac-
and participation by each of the contrasting motions depends tion (RFSI) module.
on the rotor speed for the coupled modes.
For those cases that looked at the effects of the drivetrain,
a modified coupling procedure was used. In particular, it was
Helios necessary to return blade deflections from CA in the non-
rotating hub frame of reference, so that the drivetrain-induced
Helicopter Overset Simulations (Helios) is the rotary-wing oscillations would be included in the blade motions. Coordi-
product of the US Army and CREATE-AV (Air Vehicles) nate system transformations were then applied within Helios
3
to convert to the blade frame of reference translations and ro- Figure 4 compares the calculated and measured structural
tations. loads. Steady values were removed from both test data and
An azimuthal step size of 0.1 deg (3600 steps per rotor analyses. Figures 4(a), 4(d), and 4(g) show the oscillatory
revolution) was used in the Helios calculations. A 15-deg az- flap bending moments at 30%R, 50%R, and 70%R, respec-
imuthal step size, which is standard for the aerodynamic and tively. The two coupled analyses agree well with each other
structural dynamic calculations in CAMRAD II, was used for and show reasonably good correlation with the flight test data.
the high speed and low speed calculations. A 5-deg azimuthal The peak-to-peak amplitude is well predicted but there is an
step size was used for the high thrust condition. For the RCAS approximate 10-deg phase difference between the data and
calculations, a 5-deg azimuthal step size was used for all three analyses.
conditions. Figures 4(b), 4(e), and 4(h) show the oscillatory chord
The comprehensive analysis trim solution for the UH-60A bending moments at 11.3%R and 50%R and lag damper
flight-test data solves for the collective and cyclic controls re- load, respectively. Included in these figures are RCAS/Helios
quired to obtain the specified rotor thrust and shaft pitch and predictions using either a linear or nonlinear model of the
roll moments with fixed rotor shaft angle. Because there was hydraulic lag damper (the CAMRAD II/Helios predictions
no direct measure of rotor thrust, it was estimated from known use the nonlinear model). The same damper models have
quantities. The procedure is explained below. been used to study the structural dynamics based on the pre-
The vertical force balance equation is scribed airloads, and the damper characteristics are available
in Ref. 28. When a nonlinear lag damper model is used, both
TMR cos(αs ) = GW − L f − Ls − LT R (1) chord bending moment at 11.3%R and lag damper load are
reasonably well predicted. Both models, however, signifi-
where TMR is the main rotor thrust, GW is the measured gross cantly underpredict the magnitude of chord bending moment
weight, L f is the fuselage lift, Ls is the stabilator lift, and LT R at 50%R. The effects of the linear lag damper model on the
is the lift from the canted tail rotor. The fuselage pitch atti- flap bending and torsion moments are small and thus not in-
tude and stabilator angle were measured on the flight vehicle cluded in these figures.
and the fuselage and stabilator lift curve slopes were obtained
from wind tunnel test. The tail rotor lift was estimated from Figures 4(c), 4(f), and 4(i) show the oscillatory torsion mo-
the measured main rotor torque, distance between the main ments at 30%R and 70%R and pitch link load, respectively.
and tail rotors, sideslip angle, and cant angle. This method Large aerodynamic pitching moment on the advancing side
has been used for the comparison between the small-scale generates high torsion moments on the blade and pitch link
wind tunnel and full-scale flight test data for the UH-60A ro- loads on the advancing side. The torsion moment at 30%R
tor (Ref. 29). Aircraft moment trim can be specified using (most inboard location) and pitch link loads show very sim-
the first harmonics of measured hub moments. The rotor hub ilar waveforms. The analyses show accurate correlation of
moment is obtained from the shaft bending moment measured the torsion moments at 30%R, and thus pitch link loads, in
from the strain-gauge bridge. the first and second quadrants. However, the correlation is
not satisfactory in the third and fourth quadrants. In general,
The trim targets for the three level flight conditions inves- the torsion moment correlation at 70%R is worse than that at
tigated are in Table 1. The coordinate system used is a right 30%R.
hand coordinate system where the x-axis points forward, y
right, and z up. Figure 5 compares the half peak-to-peak amplitude of flap
bending, chord bending, and torsion moments along the blade
RESULTS AND DISCUSSION span. The highest flap bending moment occurs at 60%R, with
another peak near the blade root (Fig. 5(a)). This peak load
In this section, selected data from the three flight test counters is the greatest among the three flight conditions investigated.
are compared with predictions from the coupled RCAS/Helios Although the analyses are able to capture the trends very well,
and CAMRAD II/Helios analysis methods. These test data the peak load is underpredicted by about 14%. Chord bending
include flap and chord bending moments, torsion moments, moments are strongly affected by the damper loads, particu-
pitch link loads, and lag damper loads. The analytical results larly inboard on the blade. The peak load occurs at 30%R and
presented in this section are generated using a nonlinear lag is the greatest among the three flight conditions investigated.
damper model, unless otherwise specified. The analyses with the nonlinear lag damper model increase
the magnitudes and significantly improve the correlation com-
High speed condition, µ = 0.37, CT /σ = 0.081 (C8534) pared to that with the linear model. However, the effect of the
nonlinear damper model diminishes around mid span. Unlike
This condition is the maximum-speed counter for the steady the test data, the calculated chord bending moments continu-
level flight test conditions shown in Fig. 1. The rotor blade ally decrease from the blade root to the tip. For this case, the
aerodynamic environment at high speed is characterized by maximum peak loads are underpredicted by about 35%. The
compressibility, negative lift and large aerodynamic pitch- half peak-to-peak torsion moments increase from the blade
ing moment on the advancing side, with unsteady three- tip to root. The loads show similar magnitude to the high
dimensional flows at the tip (Fig. 3). thrust condition (shown later). The analyses show reasonably
4
good correlation with the test data and slightly underpredict phase difference between the test data and analyses is consis-
the peak magnitude at 30%R by 6%. tent with the high speed condition.
Figure 6 compares the half peak-to-peak and harmonic am- Figures 9(b), 9(e), and 9(h) show the oscillatory chord
plitude of pitch link and lag damper loads. In general, the bending moments at 11.3%R and 50%R and lag damper load,
magnitude decreases as the order of the harmonic increases. respectively. Only results with the nonlinear lag damper are
The half peak-to-peak pitch link correlation is excellent be- shown. Both chord bending moment at 11.3% and lag damper
cause the overprediction of the 1/rev harmonic is compen- load are overpredicted and the correlation is worse than for
sated by the underprediction of 2 and 3/rev harmonic compo- the high speed condition. As will be shown later for the high
nents. The 4/rev harmonic correlation is very good. The lag thrust condition, this is the only case that the current analy-
damper load correlation is reasonably good with the nonlin- ses show unsatisfactory root chord bending moment and lag
ear lag damper model included. The linear lag damper model damper load correlation. Because of the overall overpredic-
significantly underpredicts 1, 2, and 3/rev magnitudes. The tion, the chord bending moment correlation at 50%R appears
overprediction of the half peak-to-peak magnitude (about 5%) better compared to the high speed condition.
is caused by the overprediction of the 3/rev harmonic compo-
Figures 9(c), 9(f), and 9(i) show the oscillatory torsion mo-
nent.
ments at 30%R and 70%R and pitch link load, respectively.
Figure 7 compares the harmonic magnitude of flap bend- As mentioned earlier, the airloads measured from the flight
ing, chord bending, and torsion moments along the blade test show a sharp “down-up” impulse on the advancing blade
span. In general, lower harmonic components have larger and an opposite impulse on the retreating blade near the blade
magnitude. The maximum peak locations and shapes differ tip. This impulse is a significant source of negative peak of
for each harmonic. The flap bending moment correlation is, in torsion moments and pitch link loads around 270-deg. The
general, very good except for the 3/rev harmonic component. analyses are not able to capture the negative peak on the re-
The underprediction of the 3/rev harmonic component is the treating side.
cause of the underprediction of the peak-to-peak magnitude
shown in Fig. 5(a). The measured chord bending moments Figure 10 compares the half peak-to-peak amplitude of
show a sudden increase of 1/rev harmonic component and a flap bending, chord bending, and torsion moments along the
sudden drop of 3/rev harmonic component, although smaller blade span. In general, these loads are the smallest among
than the 1/rev harmonic, at 40%R and 50%R. The reason for the three flight conditions investigated. The measured flap
the sudden changes is not known at present. The analyses bending moments show much smaller magnitude at the root
with the nonlinear lag damper model show much better agree- than for the high speed condition. The magnitude at 70%R
ment with the measured data for the 1, 2, and 3/rev harmon- increased by about 60% compared to that at 60%R. The anal-
ics. Lag damper modeling has an important influence near the yses show reasonably good correlation with the measured half
root of the blade and its effect diminishes around mid span. peak-to-peak flap bending moments at all the radial locations,
Even with the nonlinear lag damper model, both analyses sig- except at 70%R. The analyses underpredict the magnitude at
nificantly underpredict the chord bending moment; the 4/rev 60%R by about 14%, similar to the high speed condition. The
harmonic magnitude is particularly poor. The torsion moment measured chord bending moments show almost constant mag-
correlation is, in general, very good. The measured peak mag- nitude along the blade span, which is quite a different distri-
nitude of the 3/rev harmonic component occurs at 70%R, un- bution than that at high speed and high thrust (shown later)
like the rest of the harmonics where the peak magnitude oc- conditions. The calculated chord bending moments continu-
curs at 30%R. The coupled analyses capture the trends, but ally decrease from the blade root to the tip, overpredicting the
significantly underpredict the 3/rev magnitude. half peak-to-peak chord bending moments near the root of the
blade and significantly underpredicting along the rest of the
Low speed condition, µ = 0.15, CT /σ = 0.076 (C8513) blade span. The measured half peak-to-peak torsion moments
do not decrease from the blade root to tip compared to the high
At low speed, the airloads are mainly determined by the in- speed and high thrust conditions. The maximum peak occurs
teraction between the blades and the vortices trailed from the at 70%R, unlike the 30%R for the high speed condition. The
preceding blades. The airloads measured from the flight test torsion moment correlation is good at 30%R. However, the
show a sharp “down-up” impulse on the advancing blade and analyses significantly underpredict at 70%R and 90%R.
an opposite impulse on the retreating blade near the blade tip Figure 11 compares the half peak-to-peak and harmonic
(Fig. 8). amplitude of pitch link and lag damper loads. The analy-
Figure 9 compares the calculated and measured structural ses overpredict the half peak-to-peak pitch link load by about
loads. Again, steady values were removed from both test data 12%. The 1/rev and 4/rev harmonics are accurately predicted.
and analyses. Figures 9(a), 9(d), and 9(g) show the oscillatory However, the 2 and 5/rev harmonics are overpredicted and
flap bending moments at 30%R, 50%R, and 70%R, respec- the 3/rev harmonic is underpredicted. As shown in the time
tively. Again, the two coupled analyses agree well with each history comparison, the lag damper loads are overpredicted
other. The analyses show reasonably good correlation with at all harmonics, except the 3/rev harmonic component. The
the flight test data at 30%R and 50%R. However, the peak-to- analyses overpredict the half peak-to-peak lag damper load by
peak amplitude is substantially underpredicted at 70%R. The about 35%.
5
 Figure 12 compares the harmonic magnitude of flap bend- used in the solution procedure was increased from the default
ing, chord bending, and torsion moments along the blade value of 10 to 16 for convergence and accuracy. The analyses
span. The flap bending moment correlation is very good ex- show reasonably good correlation with the flight test data at
cept the 3/rev harmonic, where the analyses significantly un- 30%R. However, the high frequency waveforms and magni-
derpredict. This is consistent with the results at the high speed tudes are not well captured at 60%R and 70%R.
condition. The underprediction of the 3/rev harmonic compo- Figures 14(b), 14(e), and 14(h) show the oscillatory chord
nent caused slight underprediction of the peak-to-peak mag- bending moments at 11.3%R and 60%R and lag damper load,
nitude shown in Fig. 10(a). respectively. The measured chord bending moments at 60%R
A recent study using a hybrid method, coupled viscous also appear to be band-edged. Only results with the nonlinear
vortex particle method (VVPM) with CFD/CSD, showed very lag damper are shown. Both chord bending moment at 11.3%
good correlation of the 3/rev flap bending moment for this and lag damper load are well predicted, consistent with the
condition (Ref. 12). In this method, the CFD solves only the results at the high speed condition. The test data at 60%R
flow field near the blade surface and the VVPM resolves the show a strong negative peak at 225-deg azimuth. However,
detailed rotor wake away from the blade surface. Improved the analyses are not able to capture it.
wake modeling from the VVPM might be the reason for the Figures 14(c), 14(f), and 14(i) show the oscillatory torsion
accurate prediction of the 3/rev harmonic component. How- moments at 30%R and 70%R and pitch link load, respectively.
ever, the same improvement was not obtained for the high At high thrust, large pitching moments due to two dynamic
speed and high thrust conditions. Even with the VVPM, the stall cycles on the retreating side have an important influence
3/rev flap bending moments were significantly underpredicted on the blade torsion responses. The measured torsion mo-
for those conditions. ments show stronger high frequency content compared to the
The measured chord bending moments show that the 2/rev high speed condition. For example, the torsion moment at
harmonic component continues to increase from the blade root 30%R has a large 6/rev component. The analyses show, in
to the tip, and both 1 and 3/rev harmonics also increase at general, good correlation but are not able to capture the de-
60%R and 70%R. The analyses are not able to predict this tailed waveforms.
trend. The chord bending moment correlation is in general
Figure 15 compares the half peak-to-peak amplitude of
better than the high speed condition. However, the analy-
flap bending, chord bending, and torsion moments along the
ses again significantly underpredict the 4/rev magnitude. The
blade span. The test data show that the highest flap bending
measured torsion moments show that the peak magnitudes of
and chord bending moments occur at the blade root, which is
both 2 and 3/rev harmonic components occur at 70%R. This
quite different from the high and low speed conditions where
explains the maximum half peak-to-peak torsion moment at
peak loads occur just outboard of the blade mid span. The
70%R shown in Fig. 10(c). The analyses capture this trend
analyses capture this trend well. For the flap bending moment,
for the 2/rev harmonic, but significantly underpredict the 3/rev
the peak magnitude is underpredicted by about 16%. For the
harmonic because the analyses are not able to capture the
chord bending moment, the peak magnitude is overpredicted
negative peak on the retreating side as shown in Figs. 9(c)
by about 10%. The half peak-to-peak torsion moments in-
and 9(f).
crease from the blade tip to root. The analyses show reason-
ably good correlation with the test data and slightly overpre-
High thrust condition, µ = 0.24, CT /σ = 0.129 (C9017)
dict the peak magnitude at 30%R by about 3%.
This condition is the maximum thrust counter for the steady Figure 16 compares the half peak-to-peak and harmonic
level flight test conditions. At this high thrust, aerodynamics amplitude of pitch link and lag damper loads. The pitch link
is dominated by dynamic stall and large negative pitching mo- load in this condition is the greatest among the three flight
ment. The measured data show two stall events in the fourth conditions investigated. The 1 and 5/rev harmonics in this
quadrant (Fig. 13). flight condition increased compared to the high speed condi-
Figure 14 compares the calculated and measured oscilla- tion. The analyses underpredict the half peak-to-peak, and 1,
tory structural loads. Structural loads at some radial locations 2, and 3/rev harmonic pitch link loads. The analyses over-
(e.g. flap and chord bending moments at 50%R) are not avail- predict the half peak-to-peak lag damper load by about 9%
able for this flight condition due to strain gauge malfunction. mainly because of the overprediction of 2 and 5/rev harmon-
Figures 14(a), 14(d), and 14(g) show the oscillatory flap bend- ics.
ing moments at 30%R, 60%R, and 70%R, respectively. The Figure 17 compares the harmonic magnitude of flap bend-
measured flap bending moments at 60%R appear to be band- ing, chord bending, and torsion moments along the blade
edged or clipped on the retreating side. The two coupled anal- span. The flap bending moment correlation is very good ex-
yses agree well with each other. As mentioned earlier, a 5- cept the 3/rev harmonic, where the analyses significantly un-
deg azimuthal step size was used in the CAMRAD II calcu- derpredict. The underprediction of the 3/rev flap bending mo-
lations for this flight condition. The 15-deg azimuthal step ment is observed for all three flight conditions investigated.
size, which was enough for the high and low speed condi- The 1 and 2/rev chord bending moment correlation is reason-
tions, was not sufficient for the high-frequency response cal- ably good; much better than the high and low speed correla-
culations (mostly 5/rev and above). The number of harmonics tion. The torsion moment correlation is worse than the high
6
and low speed conditions, where only 3/rev harmonic compo- 50%R. Chord stiffness values are uniformly increased along
nent is underpredicted. The analyses underpredict 1/rev har- the blade span by 20% and 50%, respectively, from the base-
monic and significantly overpredict 4 and 5/rev harmonics. line property. Even with the 50% increase, the chord bending
moment does not change much. Again, structural damping
Parametric studies of high speed condition tends to smooth out the waveform. Modal damping on the
chord modes is not critical because the hydraulic lag damper
The predicted flap bending moments, in general, show rea- provides large damping. Although not shown here, the damp-
sonably good correlation with the test data. However, there is ing value at the pitch bearing has a strong influence on the
the approximate 10-deg phase difference for both high and pitch link loads (but not torsion moments).
low speed conditions and the underprediction of 3/rev har-
The effects of drivetrain dynamics on the chord bending
monic for all three flight conditions investigated. Although
moments are examined next. A simple single-shaft drivetrain
the nonlinear lag damper model improves the root chord bend-
model is included in the CAMRAD II analysis (baseline anal-
ing moment, the coupled analyses significantly underpredict
yses have no drivetrain dynamics). This model is comprised
the peak-to-peak magnitudes, especially the 4/rev harmonic
of polar mass moment of inertia, spring stiffness, and damp-
component. To better understand these deficiencies, the ef-
ing, and represents one degree-of-freedom torsion dynamics.
fects of blade stiffness, structural damping, and drivetrain dy-
The effective drivetrain inertia can be considered as the sum
namics are evaluated at the high speed condition. In these cal-
of main rotor driveshaft inertia and engine inertia. The ef-
culations, comprehensive analysis is performed with the final
fective drivetrain stiffness can be considered as the sum of
(converged) delta airloads from CFD in order to save com-
main rotor driveshaft stiffness and engine driveshaft stiffness.
putation time. Fully coupled CFD/CSD analyses are carried
The coupled rotor/drivetrain model captures the effects of the
out selectively and almost identical results are obtained with
drivetrain on the collective lag modes. The analysis should
those with comprehensive analyses using the final delta air-
solve all four blades as the blades are coupled with the driv-
loads, confirming that these structural variations do not sig-
etrain. As explained earlier, CAMRAD II provides the blade
nificantly change airloads.
deflections in the non-rotating hub frame and Helios converts
Figure 18 examines the effects of increasing flap stiffness them to the deflections in the rotating blade frame. The data
on the flap bending moments. Flap stiffness values are uni- passed from Helios to CAMRAD II are airloads data for all
formly increased along the blade span by 10% and 20%, re- four blades.
spectively, from the baseline values and the flap bending mo-
ment results are compared with the flight test and the base- Various moment of inertia (50, 100, 1000 slug-ft2 ) and
line results with RCAS/Helios. The flap stiffness variations stiffness values (200000, 300000, 400000 ft-lb/rad) are used
changed the second flap frequency to 2.86/rev and 2.89/rev, to examine their effects on structural loads. Figure 21 shows
respectively, from the baseline value of 2.84/rev. At 50%R, frequency contents of various dynamic modes that commonly
these variations decrease the minimum peak around 135-deg occur in rotorcraft (Ref. 30). Large inertia values used in
azimuth and increase the maximum peak around 240-deg az- the parametric study approximate the first drivetrain torsion
imuth. The harmonic comparison shows that although there frequency and small inertial values approximate the second
is an increase in the 3/rev harmonic, the correlation of 1 and drivetrain torsion frequency. Figure 22 shows the effects of
2/rev harmonics deteriorates. stiffness values on the flap and chord bending moments with
a fixed moment of inertia value (1000 slug-ft2 ). The results
Figure 19 examines the effects of structural damping on
show that the drivetrain changes the phase of both flap and
flap bending moments. The baseline analytical model has
chord bending moments. A smaller stiffness value (softer
very small modal damping (0.02% of critical damping) to rep-
shaft) generates a slightly larger phase shift. The 10-deg phase
resent structural damping of the UH-60A blade. The modal
difference of flap bending moment is eliminated in the third
damping value is increased to 0.5%, 1.0%, and 2.0%, respec-
and the beginning of the fourth quadrants.
tively. Although the same modal damping was applied to all
the modes for the results shown here, almost identical results Figure 23 shows the effects of moment of inertia values
were obtained by applying the modal damping only to the first on the flap and chord bending moments with a fixed stiffness
flap mode. The first flap mode frequency of 1.04/rev is very value (300000 ft-lb/rad). The changes in moment of inertia
close to resonance frequency and thus modal damping of this do not affect the flap bending moment as shown in Fig. 23(a).
mode strong sensitivity. The structural damping has a signifi- However, they have a significant effect on the chord bending
cant influence on both magnitude and phase. At 30%R, struc- moment. The chord bending moment at 50%R in Fig. 23(b)
tural damping tends to smooth out the waveform, eliminating shows that the analyses start to capture the peaks and valleys
a small hump around 120-deg azimuth. Harmonic compar- within the test data and noticeable differences in the wave-
ison shows that structural damping significantly increases 1 forms between the test data and the baseline results start to
and 2/rev harmonic amplitudes and has a small influence on diminish. Figures 23(c) and 23(d) show that there are sub-
3/rev harmonic. Structural damping has more influence on the stantial increases in the half peak-to-peak magnitude and 4/rev
inboard blade section than the outboard section. harmonic component around the mid span and thus the overall
Figure 20 examines the effects of chord stiffness and struc- correlation significantly improves.
tural damping variations on the chord bending moments at The results show the potential importance of high fre-
7
quency drivetrain modes. However, the limitations of the chord bending moment magnitude and the 4/rev harmonic
drivetrain model used in the parametric study should be component around the mid span and thus the overall chord
noted. The present single degree-of-freedom model can rep- bending moment correlation significantly improves. How-
resent only the fundamental mode of the rotor coupled with ever, due to the limitations of the present simple model, a
a low frequency drive system. Accurate simulation of a more complex drive train model, which can capture higher
high frequency drive system should match both frequency and frequency modes, should be developed and incorporated into
impedance. Thus, a more complex drive train model, which analyses.
can capture higher frequency modes, should be developed.
ACKNOWLEDGMENT
CONCLUSIONS
The authors gratefully acknowledge Dr. Wayne Johnson at
Structural loads from the NASA/Army UH-60A Airloads Pro- NASA Ames Research Center for many discussions and ad-
gram are compared with calculations obtained using coupled vice. The authors also acknowledge Dr. Vaidyanathan Anand
RCAS/Helios and CAMRAD II/Helios analyses at three level at the Boeing Company for the valuable discussions on the
flight conditions: 1) high speed, µ = 0.37, CT /σ = 0.081 drivetrain dynamics modeling.
(C8534), 2) low speed, µ = 0.15, CT /σ = 0.076 (C8513),
and 3) high thrust, µ = 0.24, CT /σ = 0.129 (C9017). Cal- REFERENCES
culated blade flap and chord bending moments, torsion mo- 1
ments, pitch link loads, and lag damper loads are compared Potsdam, M., Yeo, H., and Johnson, W., “Rotor Airloads
with the flight test data. From this study the following conclu- Prediction Using Loose Aerodynamic/Structural Coupling,”
sions are obtained: Journal of Aircraft, Vol. 43, (3), May-June 2006, pp. 732-742.
2 Datta, A., Nixon, M., and Chopra, I., “Review of Rotor
1) The two coupled analyses agree well with each other
for all three conditions investigated. A 15-deg azimuthal step Loads Prediction with the Emergence of Rotorcraft CFD,”
size in the comprehensive analysis is sufficient for the high Journal of the American Helicopter Society, Vol. 52, (4), Oc-
and low speed conditions. However, an azimuthal step size tober 2007, pp. 287-317.
of 5 deg is required for the high thrust condition, where high 3 Bhagwat, M. J., Ormiston, R. A, Saberi, H. A., and
frequency responses are important.
Xin, H., “Application of Computational Fluid Dynam-
2) The predicted flap bending moments, in general, show ics/Computational Structural Dynamics Coupling for Analy-
reasonably good correlation with the test data. However, there sis of Rotorcraft Airloads and Blade Loads in Maneuvering
is an approximate 10-deg phase difference for both high and Flight,” Journal of the American Helicopter Society, Vol. 57,
low speed conditions and underprediction of 3/rev harmonic (3), April 2012.
for all three flight conditions investigated.
4 Sankaran, V., Potsdam, M., Wissink, A., Datta, A., Jayara-
3) When a nonlinear lag damper model is used, both root
man, B., and Sitaraman, J., “Rotor Loads Prediction in Level
chord bending moment and lag damper load are reasonably
and Maneuvering Flight Using Unstructured-Adaptive Carte-
well predicted. However, the effects of the nonlinear damper
sian CFD,” American Helicopter Society 67th Annual Forum,
model diminishes around mid span. Even with the nonlinear
Virginia Beach, VA, May 3-5, 2011.
lag damper model, both analyses significantly underpredict
the peak-to-peak magnitude of chord bending moments, espe- 5
Yeo, H., Potsdam, M., and Ormiston, R. A., “Rotor Aeroe-
cially the 4/rev harmonic component. lastic Stability Analysis Using Coupled Computational Fluid
4) Torsion moment correlation is good except for the un- Dynamics/Computation Structural Dynamics,” Journal of the
derprediction of the 3/rev harmonic magnitude at both high American Helicopter Society, Vol. 56, (4), October 2011.
and low speed conditions. However, torsion moment correla- 6 Sitaraman,J., Datta, A., Baeder, J., and Chopra, I., “Cou-
tion is worse at the high thrust condition. The analyses signif-
pled CFD/CSD prediction of Rotor Aerodynamic and Struc-
icantly overpredict both 4 and 5/rev harmonics. tural Dynamic Loads for Three Critical Flight Conditions,”
5) Blade stiffness variations have only a small influence 31st European Rotorcraft Forum, Florence, Italy, September
on the prediction of flap and chord bending moments. How- 13-15, 2005.
ever, modal damping in the first flap mode has a significant 7 Biedron,
influence on the flap bending moments. Large damping val- R. T., and Lee-Rausch, E. M., “Computation of
ues deteriorate the correlation, especially at the inboard blade UH-60A Airloads Using CFD/CSD Coupling On Unstruc-
section. tured Meshes,” American Helicopter Society 67th Annual Fo-
rum, Virginia Beach, VA, May 3-5, 2011.
6) Inclusion of a simple single-shaft drivetrain model
shows the potential importance of high frequency drivetrain 8 Rajmohan, N., Marpu, R. P., Sankar, L. N., Baeder, J. D.,
modes on load predictions. Simulation of the first drive- and Egolf, T. A., “Improved Prediction of Rotor Maneuvering
train torsion frequency changes the phase of the structural Loads using a Hybrid Methodology,” American Helicopter
responses. Simulation of the second drivetrain torsion fre- Society 67th Annual Forum, Virginia Beach, VA, May 3-5,
quency shows a substantial increase in the half peak-to-peak 2011.
8
 9 Zhao, J., and He., C., “Coupled CSD/CFD and Viscous 21 Wachspress, D. A., Yu, K., Saberi, H., Hasbun, M. J., Ho,
Vortex Particle Method for Rotorcraft Comprehensive Anal- J. C., and Yeo, H., “Comprehensive Analysis Predictions of
ysis,” American Helicopter Society 68th Annual Forum, Fort Helicopter Rotor Airloads using a Fast Loose-Coupled Aero-
Worth, TX, May 1-3, 2012. dynamics Solution,” American Helicopter Society 68th An-
10 Datta,
nual Forum, Fort Worth, TX, May 1-3, 2012.
A., and Chopra, I., “Prediction of the UH-60A Main
22 Yu, K., Wachspress, D. A., Saberi, H., Hasbun, M. J.,
Rotor Structural Loads Using Computational Fluid Dynam-
ics/Comprehensive Analysis Coupling,” Journal of the Amer- Ho, J. C., and Yeo, H., “Comprehensive Analysis Predic-
ican Helicopter Society, Vol. 49, (4), October 2008, pp. 351- tions of Helicopter Rotor Structural Loads using a Fast Loose-
365. Coupled Aerodynamics Solution,” American Helicopter Soci-
11 Opoku,
ety 68th Annual Forum, Fort Worth, TX, May 1-3, 2012.
D., Makinen, S. M., and Wake, B. E., “Investiga-
23 Yeo, H., Bousman, W. G., and Johnson, W., “Performance
tion of Loose and Tight Aeroelastic Coupling for High Speed
and High Thrust Flight Conditions,” American Helicopter So- Analysis of a Utility Helicopter with Standard and Advanced
ciety 67th Annual Forum, Virginia Beach, VA, May 3-5, 2011. Rotor,” Journal of the American Helicopter Society, Vol. 53,
12 Zhao, J.,
(3), July 2004, pp. 250-270.
and He., C., “Rotor Blade Structural Loads Anal-
24 Shinoda, P. M., Yeo. H., and Norman, T. R., “Rotor Perfor-
ysis Using Coupled CSD/CFD/VVPM,” American Helicopter
Society 69th Annual Forum, Phoenix, AZ, May 21-23, 2013. mance of a UH-60 Rotor System in the NASA Ames 80- by
120-Foot Wind Tunnel,” Journal of the American Helicopter
13 Saberi, H. A., Khoshlahjeh, M., Ormiston, R. A., and Society, Vol. 49, (4), October 2004, pp. 401-413.
Rutkowski, M. J., “RCAS Overview and Application to Ad-
25 Yeo,H., and Johnson, W., “Assessment of Comprehen-
vanced Rotorcraft Problems,” AHS Fourth Decennial Special-
ists’ Conference on Aeromechanics, San Francisco, CA, Jan- sive Analysis Calculation of Airloads on Helicopter Rotors,”
uary 21-23, 2004. Journal of Aircraft, Vol. 42, (5), September-October 2005, pp.
1218-1228.
14 Sankaran, V., Wissink, A., Datta, A., Sitaraman, J., Ja-
26 Yeo,H., and Johnson, W., “Prediction of Rotor Structural
yaraman, B., Potsdam, M., Kamkar, S., Katz, A., Mavriplis,
D., Roget, B., Saberi, H., Chen, W.-B., Johnson, W., and Loads with Comprehensive Analysis,” Journal of the Ameri-
Strawn, R., “Overview of the Helios Version 2.0 Compu- can Helicopter Society, Vol. 53, (2), April 2008, pp. 193-209.
tational Platform for Rotorcraft Simulations,” 48th AIAA 27 Ormiston,
Aerospace Sciences Meeting Including the New Horizons Fo- R. A., “An Investigation of the Mechanical Air-
rum and Aerospace Exposition, Orlando, FL, January 4-7, loads Problem for Evaluating Rotor Blade Structural Dynam-
2010. ics Analysis,” American Helicopter Society 4th Decennial
Specialists’ Conference on Aeromechanics, San Francisco,
15 Johnson, W., “Technology Drivers in the Development of CA, January 21-23, 2004.
CAMRAD II,” American Helicopter Society Aeromechanics 28 Ho,J. C., Yeo, H., and Ormiston, R. A., “Investigation
Specialist Meeting, San Francisco, CA, January 19-21, 1994.
of Rotor Blade Structural Dynamics and Modeling Based
16 Bousman, W. G., and Kufeld, R. M., “UH-60A Airloads on Measured Airloads,” Journal of Aircraft, Vol. 45, No. 5,
Catalog,” NASA TM-2005-212827, August 2005. September-October 2008, pp. 1631-1642.
17 McHugh, 29 Tung,C., Bousman, W. G., and Low, S., “A Comparison
F. J., Clark, R., and Solomon, M., “Wind Tun-
nel Investigation of Rotor Lift and Propulsive Force at High of Airload Data Between Model-Scale Rotor and Full-Scale
Speed - Data Analysis,” NASA CR 145217-1, October 1977. Flight Test,” American Helicopter Society Aeromechanics
and Product Design, Bridgeport, CT, October 11-13, 1995.
18 Kufeld, R. M., and Bousman, W. G., “UH-60A Air-
30 Mihaloew, J. R., Ballin, M. G., and Ruttledge, D. C. G.,
loads Program Azimuth Reference Correction,” Journal of the
American Helicopter Society, Vol. 50, (2), April 2005, pp. “Rotorcraft Flight-Propulsion Control Integration: An Eclec-
211-213. tic Design Concept,” NASA TP 2815, April 1988.
19 Norman, T., R., Peterson, R. L., Maier, T., H., and Yeo,
H., “Evaluation of Wind Tunnel and Scaling Effects with the
UH-60A Airloads Rotor,” American Helicopter Society 68th
Annual Forum, Fort Worth, TX, May 1-3, 2012.
20 Jain,R., Yeo, H., Bhagwat, M., and Ho, J. C., “An Assess-
ment of RCAS Performance Prediction for Conventional and
Advanced Rotor Configurations,” American Helicopter Soci-
ety 70th Annual Forum, Montréal, Québec, Canada, May 20-
22, 2013.
9
 Table 1. Rotor trim conditions

Counter µ CT /σ αs Mx , ft-lb My , ft-lb

C8534 0.37 0.081 −7.31◦ −5350 −5025
C8513 0.15 0.076 0.75◦ −129 −5552
C9017 0.24 0.129 −0.15◦ −354 −193

0.18
McHugh's lift boundary
0.16
high thrust, C9017

0.14

0.12
C /σ
W

0.1

0.08

0.06 low speed, C8513 high speed, C8534

0.04
0 0.1 0.2 0.3 0.4 0.5
Advance ratio

Fig. 1. UH-60A Airloads Program level flight test matrix.

6
RCAS
CAMRAD II
5
Normalized frequency, /rev

0
0 0.2 0.4 0.6 0.8 1 1.2
Normalized rotor speed

Fig. 2. Rotor blade natural frequency comparison.

10
 0.3 0.3

0.2 0.2

0.1 0.1

n
n
Mc

Mc
2

2
0 0

Flight test (C8534)

-0.1 CAMRAD II/Helios -0.1
RCAS/Helios

-0.2 -0.2
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg
(a) Normal force @ 67.5%R (b) Normal force @ 92.0%R

0.02 0.01

0.01
0.005

0
n

m
Mc

Mc

0
2

-0.01

-0.005
-0.02

-0.03 -0.01
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg
(c) Pitching moment @ 67.5%R (mean removed) (d) Pitching moment @ 92.0%R (mean removed)

Fig. 3. Blade section normal force and pitching moment, µ = 0.37, CT /σ = 0.081 (C8534).

11
 1400 5000 600
Flight test (C8534)
CAMRAD II/Helios

Oscillatory CBM @ 11.3%R, ft-lb

RCAS/Helios
Oscillatory FBM @ 30%R, ft-lb

Oscillatory TM @ 30%R, ft-lb

RCAS/Helios
(linear lag damper)
700 2500 300

0 0 0

-700 -2500 -300

-1400 -5000 -600

0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg Azimuth, deg
(a) Flap bending moment @ 30%R (b) Chord bending moment @ (c) Torsion moment @ 30%R
11.3%R
1400 5000 600
Oscillatory CBM @ 50%R, ft-lb
Oscillatory FBM @ 50%R, ft-lb

Oscillatory TM @ 70%R, ft-lb

700 2500 300

0 0 0

-700 -2500 -300

-1400 -5000 -600

0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg Azimuth, deg
(d) Flap bending moment @ 50%R (e) Chord bending moment @ 50%R (f) Torsion moment @ 70%R

1400 6000 1600

RCAS/Helios
Oscillatory lag damper load, lb
Oscillatory FBM @ 70%R, ft-lb

Oscillatory pitch link load, lb

(linear lag damper)

700 3000 800

0 0 0

-700 -3000 -800

-1400 -6000 -1600

0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg Azimuth, deg
(g) Flap bending moment @ 70%R (h) Lag damper load (i) Pitch link load

Fig. 4. Comparison of time history of calculated and measured structural loads, µ = 0.37, CT /σ = 0.081 (C8534).

12
 Half peak-to-peak FBM, ft-lb 1200

1200
Flight Test (C8534)
CAMRAD II/Helios
600 RCAS/Helios

900

Pitch link load, lb

Flight Test (C8534)
CAMRAD II/Helios
RCAS/Helios 600
0
0 0.2 0.4 0.6 0.8 1
r/R
(a) Flap bending moment
300
4000
Half peak-to-peak CBM, ft-lb

0
half
0 1 2 3 4 5
peak-to-peak /rev
2000 (a) Pitch link load

4000
Flight Test (C8534)
CAMRAD II/Helios
RCAS/Helios RCAS/Helios
(linear lag damper) RCAS/Helios (linear lag damper)
0 3000
Lag damper load, lb

0 0.2 0.4 0.6 0.8 1

r/R
(b) Chord bending moment
600 2000
Half peak-to-peak TM, ft-lb

1000

300

0
half
0 1 2 3 4 5
peak-to-peak /rev
(b) Lag damper load

0 Fig. 6. Comparison of half peak-to-peak and harmonic

0 0.2 0.4 0.6 0.8 1
r/R magnitude of calculated and measured pitch link and
(c) Torsion moment damper loads, µ = 0.37, CT /σ = 0.081 (C8534).

Fig. 5. Comparison of half peak-to-peak magnitude of cal-

culated and measured blade structural loads, µ = 0.37,
CT /σ = 0.081 (C8534).

13
 1000 Flight Test (C8534) 2500 500
CAMRAD II/Helios
RCAS/Helios

1/rev CBM, ft-lb

1/rev FBM, ft-lb

1/rev TM, ft-lb

RCAS/Helios
(linear lag damper)
500 1250 250

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(a) 1/rev magnitude (b) 1/rev magnitude (c) 1/rev magnitude
1000 2500 500

2/rev CBM, ft-lb

2/rev FBM, ft-lb

2/rev TM, ft-lb

500 1250 250

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(d) 2/rev magnitude (e) 2/rev magnitude (f) 2/rev magnitude
1000 2500 500
3/rev CBM, ft-lb
3/rev FBM, ft-lb

3/rev TM, ft-lb

500 1250 250

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(g) 3/rev magnitude (h) 3/rev magnitude (i) 3/rev magnitude
1000 2500 500
4/rev CBM, ft-lb
4/rev FBM, ft-lb

4/rev TM, ft-lb

500 1250 250

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(j) 4/rev magnitude (k) 4/rev magnitude (l) 4/rev magnitude
1000 2500 500
5/rev CBM, ft-lb
5/rev FBM, ft-lb

5/rev TM, ft-lb

500 1250 250

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(m) 5/rev magnitude (n) 5/rev magnitude (o) 5/rev magnitude

Fig. 7. Comparison of harmonic magnitude of calculated and measured structural loads, µ = 0.37, CT /σ = 0.081 (C8534).

14
 0.2 0.2

0.1 0.1
n

n
Mc

Mc
2

2
0 0
Flight test (C8513)
CAMRAD II/Helios
RCAS/Helios

-0.1 -0.1
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg
(a) Normal force @ 67.5%R (b) Normal force @ 92.0%R

0.01 0.02

0.01
0.005

0
m

m
Mc

Mc

0
2

-0.01

-0.005
-0.02

-0.01 -0.03
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg
(c) Pitching moment @ 67.5%R (mean removed) (d) Pitching moment @ 92.0%R (mean removed)

Fig. 8. Blade section normal force and pitching moment, µ = 0.15, CT /σ = 0.076 (C8513).

15
 1000 2500 200
Flight test (C8513)
CAMRAD II/Helios

Oscillatory CBM @ 11.3%R, ft-lb

Oscillatory FBM @ 30%R, ft-lb

Oscillatory TM @ 30%R, ft-lb

RCAS/Helios
500 1250 100

0 0 0

-500 -1250 -100

-1000 -2500 -200

0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg Azimuth, deg
(a) Flap bending moment @ 30%R (b) Chord bending moment @ (c) Torsion moment @ 30%R
11.3%R
1000 2500 200
Oscillatory CBM @ 50%R, ft-lb
Oscillatory FBM @ 50%R, ft-lb

Oscillatory TM @ 70%R, ft-lb

500 1250 100

0 0 0

-500 -1250 -100

-1000 -2500 -200

0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg Azimuth, deg
(d) Flap bending moment @ 50%R (e) Chord bending moment @ 50%R (f) Torsion moment @ 70%R

1000 4000 600

Oscillatory lag damper load, lb
Oscillatory FBM @ 70%R, ft-lb

Oscillatory pitch link load, lb

500 2000 300

0 0 0

-500 -2000 -300

-1000 -4000 -600

0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg Azimuth, deg
(g) Flap bending moment @ 70%R (h) Lag damper load (i) Pitch link load

Fig. 9. Comparison of time history of calculated and measured structural loads, µ = 0.15, CT /σ = 0.076 (C8513).

16
 1000 Flight Test (C8513)
CAMRAD II/Helios
Half peak-to-peak FBM, ft-lb

RCAS/Helios

600
Flight Test (C8513)
CAMRAD II/Helios
500 RCAS/Helios

450

Pitch link load, lb

300
0
0 0.2 0.4 0.6 0.8 1
r/R
(a) Flap bending moment
150
2400
Half peak-to-peak CBM, ft-lb

0
half
0 1 2 3 4 5
peak-to-peak /rev
1200 (a) Pitch link load

4000

Flight Test (C8513)

CAMRAD II/Helios
RCAS/Helios
0 3000
Lag damper load, lb

0 0.2 0.4 0.6 0.8 1

r/R
(b) Chord bending moment
200 2000
Half peak-to-peak TM, ft-lb

1000

100

0
half
0 1 2 3 4 5
peak-to-peak /rev
(b) Lag damper load

0 Fig. 11. Comparison of half peak-to-peak and harmonic

0 0.2 0.4 0.6 0.8 1
r/R magnitude of calculated and measured pitch link and
(c) Torsion moment damper loads, µ = 0.15, CT /σ = 0.076 (C8513).

Fig. 10. Comparison of half peak-to-peak magnitude of

calculated and measured blade structural loads, µ = 0.15,
CT /σ = 0.076 (C8513).

17
 400 Flight Test (C8513) 1600 120
CAMRAD II/Helios
RCAS/Helios

1/rev CBM, ft-lb

1/rev FBM, ft-lb

1/rev TM, ft-lb

200 800 60

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(a) 1/rev magnitude (b) 1/rev magnitude (c) 1/rev magnitude
400 1600 120

2/rev CBM, ft-lb

2/rev FBM, ft-lb

2/rev TM, ft-lb

200 800 60

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(d) 2/rev magnitude (e) 2/rev magnitude (f) 2/rev magnitude
400 1600 120
3/rev CBM, ft-lb
3/rev FBM, ft-lb

3/rev TM, ft-lb

200 800 60

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(g) 3/rev magnitude (h) 3/rev magnitude (i) 3/rev magnitude
400 1600 120
4/rev CBM, ft-lb
4/rev FBM, ft-lb

4/rev TM, ft-lb

200 800 60

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(j) 4/rev magnitude (k) 4/rev magnitude (l) 4/rev magnitude
400 1600 120
5/rev CBM, ft-lb
5/rev FBM, ft-lb

5/rev TM, ft-lb

200 800 60

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(m) 5/rev magnitude (n) 5/rev magnitude (o) 5/rev magnitude

Fig. 12. Comparison of harmonic magnitude of calculated and measured structural loads, µ = 0.15, CT /σ = 0.076
(C8513).

18
 0.4 0.4

0.3 0.3

0.2 0.2
n

n
Mc

Mc
2

2
0.1 0.1

Flight test (C9017)

0 0
CAMRAD II/Helios
RCAS/Helios
-0.1 -0.1
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg
(a) Normal force @ 67.5%R (b) Normal force @ 92.0%R

0.02 0.02

0.01 0.01

0 0
m

m
Mc

Mc
2

-0.01 -0.01

-0.02 -0.02

-0.03 -0.03
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg
(c) Pitching moment @ 67.5%R (mean removed) (d) Pitching moment @ 92.0%R (mean removed)

Fig. 13. Blade section normal force and pitching moment, µ = 0.24, CT /σ = 0.129 (C9017).

19
 800 3000 600
Flight test (C9017)
CAMRAD II/Helios

Oscillatory CBM @ 11.3%R, ft-lb

Oscillatory FBM @ 30%R, ft-lb

Oscillatory TM @ 30%R, ft-lb

RCAS/Helios
400 1500 300

0 0 0

-400 -1500 -300

-800 -3000 -600

0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg Azimuth, deg
(a) Flap bending moment @ 30%R (b) Chord bending moment @ (c) Torsion moment @ 30%R
11.3%R
800 3000 600
Oscillatory CBM @ 60%R, ft-lb
Oscillatory FBM @ 60%R, ft-lb

Oscillatory TM @ 70%R, ft-lb

400 1500 300

0 0 0

-400 -1500 -300

-800 -3000 -600

0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg Azimuth, deg
(d) Flap bending moment @ 60%R (e) Chord bending moment @ 60%R (f) Torsion moment @ 70%R

800 6000 1600

Oscillatory lag damper load, lb
Oscillatory FBM @ 70%R, ft-lb

Oscillatory pitch link load, lb

400 3000 800

0 0 0

-400 -3000 -800

-800 -6000 -1600

0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg Azimuth, deg
(g) Flap bending moment @ 70%R (h) Lag damper load (i) Pitch link load

Fig. 14. Comparison of time history of calculated and measured structural loads, µ = 0.24, CT /σ = 0.129 (C9017).

20
 1000 Flight Test (C9017)
CAMRAD II/Helios
Half peak-to-peak FBM, ft-lb

RCAS/Helios

1200
Flight Test (C9017)
CAMRAD II/Helios
500 RCAS/Helios

900

Pitch link load, lb

600
0
0 0.2 0.4 0.6 0.8 1
r/R
(a) Flap bending moment
300
3000
Half peak-to-peak CBM, ft-lb

0
half
0 1 2 3 4 5
peak-to-peak /rev
1500 (a) Pitch link load

4000
Flight Test (C9017)
CAMRAD II/Helios
RCAS/Helios
0 3000
Lag damper load, lb

0 0.2 0.4 0.6 0.8 1

r/R
(b) Chord bending moment
600 2000
Half peak-to-peak TM, ft-lb

1000

300

0
half
0 1 2 3 4 5
peak-to-peak /rev
(b) Lag damper load

0 Fig. 16. Comparison of half peak-to-peak and harmonic

0 0.2 0.4 0.6 0.8 1
r/R magnitude of calculated and measured pitch link and
(c) Torsion moment damper loads, µ = 0.24, CT /σ = 0.129 (C9017).

Fig. 15. Comparison of half peak-to-peak magnitude of

calculated and measured blade structural loads, µ = 0.24,
CT /σ = 0.129 (C9017).

21
 600 Flight Test (C9017) 2000 300
CAMRAD II/Helios
RCAS/Helios

1/rev CBM, ft-lb

1/rev FBM, ft-lb

1/rev TM, ft-lb

300 1000 150

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(a) 1/rev magnitude (b) 1/rev magnitude (c) 1/rev magnitude
600 2000 300

2/rev CBM, ft-lb

2/rev FBM, ft-lb

2/rev TM, ft-lb

300 1000 150

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(d) 2/rev magnitude (e) 2/rev magnitude (f) 2/rev magnitude
600 2000 300
3/rev CBM, ft-lb
3/rev FBM, ft-lb

3/rev TM, ft-lb

300 1000 150

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(g) 3/rev magnitude (h) 3/rev magnitude (i) 3/rev magnitude
600 2000 300
4/rev CBM, ft-lb
4/rev FBM, ft-lb

4/rev TM, ft-lb

300 1000 150

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(j) 4/rev magnitude (k) 4/rev magnitude (l) 4/rev magnitude
600 2000 300
5/rev CBM, ft-lb
5/rev FBM, ft-lb

5/rev TM, ft-lb

300 1000 150

0 0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R r/R
(m) 5/rev magnitude (n) 5/rev magnitude (o) 5/rev magnitude

Fig. 17. Comparison of harmonic magnitude of calculated and measured structural loads, µ = 0.24, CT /σ = 0.129
(C9017).

22
 Flight test (C8534)
1400 1000 Flight Test (C8534)
Baseline (RCAS/Helios)
110% EI_flap Baseline (RCAS/Helios)
Oscillatory FBM @ 50%R, ft-lb

120% EI_flap 110% EI_flap

120% EI_flap
700

1/rev FBM, ft-lb

0 500

-700

-1400 0
0 45 90 135 180 225 270 315 360 0 0.2 0.4 0.6 0.8 1
Azimuth, deg r/R
(a) Flap bending moment @ 50%R (b) 1/rev magnitude

1000 1000
2/rev FBM, ft-lb

3/rev FBM, ft-lb

500 500

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R
(c) 2/rev magnitude (d) 3/rev magnitude

Fig. 18. Effects of flap stiffness on flap bending moments, µ = 0.37, CT /σ = 0.081 (C8534).

23
 1400 Flight test (C8534) 1000
Baseline (RCAS/Helios, 0.02% damping)
0.5% damping
Oscillatory FBM @ 30%R, ft-lb

1.0% damping
700 2.0% damping

1/rev FBM, ft-lb

0 500

-700

-1400 0
0 45 90 135 180 225 270 315 360 0 0.2 0.4 0.6 0.8 1
Azimuth, deg r/R
(a) Flap bending moment @ 30%R (b) 1/rev magnitude

1000 1000
Flight Test (C8534)
Baseline (RCAS/Helios, 0.02% damping)
0.5% damping
1.0% damping
2.0% damping
2/rev FBM, ft-lb

3/rev FBM, ft-lb

500 500

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R
(c) 2/rev magnitude (d) 3/rev magnitude

Fig. 19. Effects of structural damping on flap bending moments, µ = 0.37, CT /σ = 0.081 (C8534).

24
 6000 6000

Oscillatory CBM @ 50%R, ft-lb

Oscillatory CBM @ 50%R, ft-lb

3000 3000

0 0

-3000 -3000 Flight test (C8534)

Flight test (C8534) Baseline (RCAS/Helios, 0.02% damping)
Baseline (RCAS/Helios) 0.5% damping
120% EI_lag 1.0% damping
150% EI_lag 2.0% damping
-6000 -6000
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg
(a) Effect of chord stiffness (b) Effect of structural damping

Fig. 20. Effects of stiffness and damping variations on chord bending moments @ 50%R, µ = 0.37, CT /σ = 0.081 (C8534).

Fig. 21. Modal frequencies of rotorcraft dynamics system (Ref. 30).

1400 Flight test (C8534) 6000

Baseline (CAMRAD II/Helios)
K = 400000 ft-lb/rad
Oscillatory CBM @ 50%R, ft-lb
Oscillatory FBM @ 50%R, ft-lb

K = 300000 ft-lb/rad
700 K = 200000 ft-lb/rad 3000

0 0

-700 -3000 Flight test (C8534)

Baseline (CAMRAD II/Helios)
K = 400000 ft-lb/rad
K = 300000 ft-lb/rad
K = 200000 ft-lb/rad
-1400 -6000
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg
(a) Flap bending moment @ 50%R (b) Chord bending moment @ 50%R

Fig. 22. Effects of drivetrain stiffness on flap and chord bending moments, µ = 0.37, CT /σ = 0.081 (C8534).

25
 1400 Flight test (C8534) 6000
Baseline (CAMRAD II/Helios)
I = 1000 slug-ft^2

Oscillatory CBM @ 50%R, ft-lb

Oscillatory FBM @ 50%R, ft-lb

I = 100 slug-ft^2
700 I = 50 slug-ft^2 3000

0 0

-700 -3000 Flight test (C8534)

Baseline (CAMRAD II/Helios)
I = 1000 slug-ft^2
I = 100 slug-ft^2
I = 50 slug-ft^2
-1400 -6000
0 45 90 135 180 225 270 315 360 0 45 90 135 180 225 270 315 360
Azimuth, deg Azimuth, deg
(a) Flap bending moment @ 50%R (b) Chord bending moment @ 50%R

4000 2500
Flight Test (C8534)
CAMRAD II/Helios
Half peak-to-peak CBM, ft-lb

I = 1000 slug-ft^2
I = 100 slug-ft^2
3000 I = 50 slug-ft^2
4/rev CBM, ft-lb

2000 1250

1000

0 0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
r/R r/R
(c) Half peak-to-peak chord bending moment (d) 4/rev chord bending moment

Fig. 23. Effects of drivetrain inertia on flap and chord bending moments, µ = 0.37, CT /σ = 0.081 (C8534).

26