Escolar Documentos
Profissional Documentos
Cultura Documentos
By
SCOTT MICHAEL LARWOOD
B.S. (California Polytechnic State University, San Luis Obispo) 1988
M.S. (Stanford University) 1993
DISSERTATION
Submitted in partial satisfaction of the requirements for the degree of
DOCTOR OF PHILOSOPHY
in
Mechanical and Aeronautical Engineering
in the
OFFICE OF GRADUATE STUDIES
of the
UNIVERSITY OF CALIFORNIA
DAVIS
Approved:
Committee in charge
2009
i
Donald Margolis
Fidelis Eke
Case van Dam
Dynamic Analysis Tool Development for Advanced Geometry Wind Turbine Blades
Copyright 2009
by
Scott Michael Larwood
Acknowledgments
The California Energy Commission partially funded this work through the PIEREnergy
Innovations Small Grant Program, grant number 54905A/06-16.
I would like to thank the men and women of the wind energy industry, especially my
former colleagues at Kenetech Windpower, who believed that wind could become a com-
petitive power source. I would also like to thank Case van Dam, my PhD adviser, for
accepting me into his research group (friends Eddie Mayda, Jon Baker, Henry Shiu, and
Ray Chow) at UC Davis. His wise and gentle guidance made my transition into academia
(from industry) very smooth. Kevin Jackson at Dynamic Design planted the seeds for this
dissertation, and encouraged me to apply for external funding. Mike Zuteck, one of the
most talented engineers I know, pushed the concept of the swept blade into reality, along
with leadership from Gary Kanaby at Knight and Carver and Tom Ashwill at Sandia. The
wind turbine code developers at NREL, Marshall Buhl and Jason Jonkman, laid excellent
groundwork for my code development, and tolerated my pesky questions.
Finally, I would like to acknowledge my family. My parents whole-heartily supported
my shocking decision to leave a good job in my middle age to go back for my doctorate.
It was the best decision of my life. At Davis I met my wife Brenna, and we now have a
wonderful son Paul. Thank you Brenna and Paul, for giving me the time to nish this work.
Our journey together has just begun.
ii
Contents
List of Figures vi
List of Tables viii
Nomenclature ix
Abstract 1
1 Introduction 2
1.1 Current Status of Wind Energy . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Advanced Wind Turbine Rotors . . . . . . . . . . . . . . . . . . . . . . . 3
1.3 Wind Turbine Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.4 Research Objectives and Motivation . . . . . . . . . . . . . . . . . . . . . 8
2 Blade Finite Element Modeling 10
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.2 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.1 Coordinate Systems and Transformations . . . . . . . . . . . . . . 11
2.2.2 Kinematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
2.2.3 Lagranges Equation . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.4 Steady State Equations and Axial Force . . . . . . . . . . . . . . . 25
2.2.5 Oscillation about Steady State . . . . . . . . . . . . . . . . . . . . 28
2.2.6 Beam Element Matrices . . . . . . . . . . . . . . . . . . . . . . . 29
2.2.7 Solution for Blade Frequencies and Mode Shapes . . . . . . . . . . 30
2.3 Finite Element Method Verication Results . . . . . . . . . . . . . . . . . 32
2.3.1 Curved Beam Deection Verication . . . . . . . . . . . . . . . . 32
2.3.2 Non-Rotating Curved Beam Modes Verication . . . . . . . . . . . 35
2.3.3 Rotating Tapered Beam Verication . . . . . . . . . . . . . . . . . 35
2.3.4 Rotating Beams from Leung and Fung 1988 Verication . . . . . . 38
2.3.5 Rotating Curved Beam Verication . . . . . . . . . . . . . . . . . 45
2.4 Conclusions and Recommendations . . . . . . . . . . . . . . . . . . . . . 47
iii
3 Dynamic Analysis of Curved Wind Turbine Blades 48
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.2 Coordinate Systems and Transformations . . . . . . . . . . . . . . . . . . 49
3.2.1 Local Undeected Axis . . . . . . . . . . . . . . . . . . . . . . . . 49
3.2.2 Local Deected Coordinate System . . . . . . . . . . . . . . . . . 52
3.2.3 Local Aerodynamics Coordinate System . . . . . . . . . . . . . . . 54
3.3 Kinematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
3.3.1 Positions and Displacements . . . . . . . . . . . . . . . . . . . . . 57
3.3.2 Velocities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
3.3.3 Accelerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
3.4 Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
3.4.1 Generalized Inertia Forces . . . . . . . . . . . . . . . . . . . . . . 69
3.4.2 Generalized Active Forces . . . . . . . . . . . . . . . . . . . . . . 74
3.5 Blade Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.5.1 Blade Root Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.5.2 Blade Gage Moment Outputs . . . . . . . . . . . . . . . . . . . . . 95
3.6 Verication with FAST and Adams . . . . . . . . . . . . . . . . . . . . . . 95
3.7 Validation with Field Test Data . . . . . . . . . . . . . . . . . . . . . . . . 106
3.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
3.8.1 Verication with Adams . . . . . . . . . . . . . . . . . . . . . . . 110
3.8.2 Validation with Field Test Data . . . . . . . . . . . . . . . . . . . . 111
3.9 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 112
4 Design Studies 113
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113
4.2 Baseline Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
4.2.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
4.2.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
4.2.3 Conclusions and Discussion . . . . . . . . . . . . . . . . . . . . . 121
4.3 Scaling to Larger rotors . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.3.1 Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
4.3.2 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
4.3.3 Conclusions and Recommendations . . . . . . . . . . . . . . . . . 128
Bibliography 129
A Blade Finite Element Matrices 134
A.1 Mass Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
A.2 Elastic Stiffness Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136
A.3 Gyroscopic Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
A.4 Spin Stiffness Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
A.5 Axial Force Stiffness Matrix . . . . . . . . . . . . . . . . . . . . . . . . . 142
A.6 Axial Reduction Matrix . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
B CurveFEM Program Structure 145
iv
C Summary of FAST Modications 147
v
List of Figures
1.1 Swept blade concept. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Analysis ow diagram. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1 First transformation of element . . . . . . . . . . . . . . . . . . . . . . . . 13
2.2 Second transformation of element with built-in structural twist . . . . . . . 14
2.3 Beam element coordinate system . . . . . . . . . . . . . . . . . . . . . . . 18
2.4 Beam element axial force (after Leung [1]) . . . . . . . . . . . . . . . . . 27
2.5 Nodal degrees of freedom . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.6 Curved beam for deection verication (after Kosmatka [2]) . . . . . . . . 32
2.7 Curve beam deection and slope for F
x
= 100 lb . . . . . . . . . . . . . . 33
2.8 Curve beam deection and slope for F
y
= 100 lb . . . . . . . . . . . . . . 34
2.9 Curved beam for natural response (after Kosmatka [2]) . . . . . . . . . . . 35
2.10 Rotating tapered beam (all dimensions in inches) . . . . . . . . . . . . . . 36
2.11 Tapered beam verication results . . . . . . . . . . . . . . . . . . . . . . . 37
2.12 Horizontal Cantilever (after Leung [1]) . . . . . . . . . . . . . . . . . . . . 38
2.13 Horizontal Cantilever Verication Results . . . . . . . . . . . . . . . . . . 39
2.14 Inclined Cantilever (after Leung [1]) . . . . . . . . . . . . . . . . . . . . . 40
2.15 Inclined Cantilever Verication Results . . . . . . . . . . . . . . . . . . . 41
2.16 L-beam (after Leung [1]) . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
2.17 L-beam with = 90
Verication Results . . . . . . . . . . . . . . . . . . 43
2.18 L-beam with = 30
Verication Results . . . . . . . . . . . . . . . . . . 44
2.19 Rotating curved beam verication results . . . . . . . . . . . . . . . . . . 46
3.1 First transformation of element . . . . . . . . . . . . . . . . . . . . . . . . 50
3.2 Second transformation of element with built-in structural twist . . . . . . . 51
3.3 Undeected blade element offsets . . . . . . . . . . . . . . . . . . . . . . 57
3.4 Sweep angle for swept blade . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.5 Blade tip twist verication . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.6 Blade tip deection verication . . . . . . . . . . . . . . . . . . . . . . . . 99
3.7 Generator power verication . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.8 Flap bending moment verication . . . . . . . . . . . . . . . . . . . . . . 101
3.9 Edge bending moment verication . . . . . . . . . . . . . . . . . . . . . . 102
3.10 Out-of-plane deection for STAR7d rst ap bending mode . . . . . . . . 104
3.11 Torsional deection for STAR7d rst ap bending mode . . . . . . . . . . 105
vi
3.12 Generator power validation . . . . . . . . . . . . . . . . . . . . . . . . . . 107
3.13 Blade pitch validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.14 Edge bending validation . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
3.15 Flap bending validation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
4.1 Parametric study of damage equivalent loads . . . . . . . . . . . . . . . . . 118
4.2 Parametric study of annual energy production . . . . . . . . . . . . . . . . 119
4.3 Parametric study of blade tip deection . . . . . . . . . . . . . . . . . . . 120
4.4 Flap bending comparison for scaled design . . . . . . . . . . . . . . . . . . 127
vii
List of Tables
2.1 XYZ Component Values . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Curved beam modes verication . . . . . . . . . . . . . . . . . . . . . . . 35
3.1 CurveFAST Blade 1 Degrees of Freedom . . . . . . . . . . . . . . . . . . 53
3.2 CurveFAST Degrees of Freedom . . . . . . . . . . . . . . . . . . . . . . . 56
3.3 Turbulent wind le parameters . . . . . . . . . . . . . . . . . . . . . . . . 96
3.4 Maximum percentage differences for the Adams/CurveFAST verication . 97
3.5 Extreme Load Verication, Normalized Maximum Values . . . . . . . . . . 103
4.1 Parametric Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
4.2 WindPACT model wind turbine parameters . . . . . . . . . . . . . . . . . 122
4.3 WP1500 blade properties . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
4.4 WP3000 blade properties . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
4.5 Scaled model results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
viii
Nomenclature
Roman
a acceleration vector
A sectional area
C matrix for Kanes method
C
D
drag coefcient
C
D
drag coefcient modied for sweep
C
L
lift coefcient
C
L
lift coefcient modied for sweep
C
M
moment coefcient
D damping matrix
EA extensional stiffness
EI bending stiffness
f vector in equation of motion for Kanes method
F element axial force
F force vector
ix
F
r
generalized active forces
F
r
generalized inertia forces
g gravitational acceleration
g
I inertia dyadic
j pitched system unit vectors
J
r
eigenvalue
transformation matrix
angle between relative velocity and chord line
lineal density
non-dimensional position along element
mass density
interpolation function
eigenvector element
matrix of eigenvector elements
angular velocity vector
rotational speed
spinning matrix
Subscripts
0 value at inboard element node
a applied, aerodynamic
Aero aerodynamic
ave average
AxRed axial reduction
xiii
B1 blade 1
Damp damping
e element
Elastic elastic
eq Equivalent
E1 rst edge mode
F1 rst ap mode
F2 second ap mode
Grav gravitational
l value at outboard element node
r partial velocity derivative
s structural, steady state
T1 rst torsion mode
Teet teeter degree of freedom
u ultimate
v oscillatory component
Superscripts
B1Tip blade 1 tip node
E inertial frame
H hub body
xiv
M1 blade 1 element body
S1 blade 1 element center of mass
Other Notation
( )
_
i
B1
1
i
B1
2
i
B1
3
_
_
= [T
CH
]
_
_
g
B1
1
g
B1
2
g
B1
3
_
_
(2.1)
with the transformation matrix:
[T
CH
] =
_
_
cos[PreCone(1)] 0 sin[PreCone(1)]
0 1 0
sin[PreCone(1)] 0 cos[PreCone(1)]
_
_
where PreCone(1) is the hub pre-coning angle for blade 1, which is positive downwind.
Currently there are no provisions for coning in CurveFEM; therefore, the T
CH
matrix is
equal to the identity matrix [I]. Coning could be added later to CurveFEM. The transfor-
mation to the blade pitched coordinate system (j
B1
) is:
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
= [T
PC
]
_
_
i
B1
1
i
B1
2
i
B1
3
_
_
(2.2)
with the transformation matrix:
[T
PC
] =
_
_
cos[BlPitch(1)] sin[BlPitch(1)] 0
sin[BlPitch(1)] cos[BlPitch(1)] 0
0 0 1
_
_
where BlPitch(1) is the blade 1 pitch. The pitch angle is relative to the chord line at zero
aerodynamic twist and is positive toward feather (leading edge upwind). Currently there
is no provision for pitch (can be added later) in CurveFEM; therefore the transformation
13
matrix T
PC
is equal to the identity matrix [I].
The next coordinate systemis the blade element systemaligned with the local undeected-
axis. It has the unit vectors Lj
B1
i
(i = 1, 2, 3). The transformation from the pitch system
is:
_
_
Lj
B1
1
(r)
Lj
B1
2
(r)
Lj
B1
3
(r)
_
_
= [T
EP
(r)]
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
(2.3)
The transformation development follows a similar method to Rao [28] p. 328, for nite
elements, but with a coordinate system change to match the FAST elements. The rst
stage involves a transformation matrix [
1
] between the pitched coordinates XY Z and the
coordinates x y z by assuming the y axis to be parallel to the Y Z plane (Fig. 2.1):
j
3
j
1
j
2
X
Z
Y
3
j
2
j
1
j
x
z
y
l
i
j
Figure 2.1: First transformation of element
_
_
x
y
z
_
_
= [
1
]
_
_
X
Y
Z
_
_
The next step is between the local coordinates xyz (principal axes) and the coordinates x y z
14
as (Fig. 2.2):
_
_
x
y
z
_
_
= [
2
]
_
_
x
y
z
_
_
Lj
2
Lj
1
x
y
2
j
1
j
x
y
_
x
y
z
_
_
= [T
EP
(r)]
_
_
X
Y
Z
_
_
15
From Fig. 2.1, the direction cosines of the longitudinal axis ( z or z) are:
l
o z
= l
oz
=
X
j
X
i
l
m
o z
= m
oz
=
Y
j
Y
i
l
n
o z
= n
oz
=
Z
j
Z
i
l
where the length of the element is:
l = {(X
j
X
i
)
2
+ (Y
j
Y
i
)
2
+ (Z
j
Z
i
)
2
}
1/2
As in the FAST2ADAMS.f90 preprocessor, the interior element nodes coincide with
the analysis nodes, which are interpolated from the properties in the blade input le. The
last node uses the values at the tip. Table 2.1 shows values of the XY Z components and
their FAST variable names.
Table 2.1: XYZ Component Values
Component Inner Elements Tip Value
X PrecrvRef(i) PrecrvRef(TipNode)
Y PreswpRef(i) PreswpRef(TipNode)
Z BlFract(i) BldFlexL +
BldFlexL + HubRad
HubRad
Because the unit vector
j
2
(parallel to the y axis) is normal to both the unit vectors j
1
(parallel to the X axis) and
j
3
(parallel to the z axis) the following vector analysis relation
holds:
j
2
=
j
3
j
1
_
_
j
3
j
1
_
_
=
1
d
j
1
j
2
j
3
l
oz
m
oz
n
oz
1 0 0
=
1
d
(j
2
n
oz
j
3
m
oz
)
where:
d = (m
2
oz
+ n
2
oz
)
1/2
16
Therefore, the direction cosines of the y axis with respect to the XY Z system are:
l
o y
= 0, m
o y
=
n
oz
d
, n
o y
=
m
oz
d
Because the x axis (unit vector
j
1
) is orthonormal to the y axis (
j
2
) and the z axis (
j
3
),
j
1
is:
j
1
=
j
2
j
3
=
j
1
j
2
j
3
l
o y
m
o y
n
o y
l
oz
m
oz
n
oz
j
1
j
2
j
3
0
noz
d
moz
d
l
oz
m
oz
n
oz
=
1
d
_
j
1
(m
2
oz
+ n
2
oz
) +j
2
(l
oz
m
oz
) +j
3
(l
oz
n
oz
)
_
l
o x
m
o x
n
o x
l
o y
m
o y
n
o y
l
o z
m
o z
n
o z
_
_
=
_
_
(m
2
oz
+ n
2
oz
)/d (l
oz
m
oz
)/d (l
oz
n
oz
)/d
0 n
oz
/d m
oz
/d
l
oz
m
oz
n
oz
_
_
The principal axes of the element (xyz axes) makes an angle
s
(built-in structural twist)
17
about the negative z axis. This direction is such that a positive
s
results in a lower angle
of attack. The tranformation between the systems is:
_
_
x
y
z
_
_
=
_
_
cos
s
sin
s
0
sin
s
cos
s
0
0 0 1
_
_
_
_
x
y
z
_
_
= [
2
]
_
_
x
y
z
_
_
so that:
[
2
] =
_
_
cos
s
sin
s
0
sin
s
cos
s
0
0 0 1
_
_
Note that this procedure breaks down when the z (or z) axis is aligned to the X axis. In
this case, m
ox
= n
ox
= 0 and therefore d = 0. However, this situation is not realizable
for a horizontal-axis wind turbine because it implies that two stations are at the same radial
position.
2.2.2 Kinematics
This section draws on the work of Leung and Fung [1], mentioned by the author in
the introduction. They develop the element matrices for rotating, linear, constant property,
space-frame elements with six DOF at each node. The author changed their coordinate
systems to match the alignment in FAST, in addition to adding tapered element properties.
The beam element is located in space with three sets of orthogonal axes:
1. xyz along the local principal axes of the beam (with unit vectors Lj
i
)
2. XY Z that is the pitched blade system for no pre-cone or pitch (with unit vectors j
i
)
with X parallel to the spinning axis
3.
X
Y
Z, an inertial frame with
X parallel to X
18
x
Y
y
z
{r
h
}
{r
0
}
{r
g
}
{s}
X
Z
Figure 2.3: Beam element coordinate system
Figure 2.3 shows the rst two coordinate systems.
The undeformed element is at rest relative to frame XY Z and the position vector for
any point on the element is:
{r
0
} = {r
g
} + ({r
h
} {r
g
}), 0 1
or:
{r
0
} = {r
g
} + s{s}, 0 s l (2.4)
where {r
0
}, {r
g
}, and {r
h
} are respectively the position vectors of the point and the ends
of the beam with respect to frame XY Z, {s} is the unit vector along the beam axis with
{s} = {r
gh
}/ |r
gh
| and {r
gh
} = {r
h
} {r
g
}.
A spinning matrix [] used in the ensuing analysis for the element is:
[] =
_
_
0 0 0
0 0 1
0 1 0
_
_
(2.5)
19
The absolute position vector for any point on the beam is:
{r} = [X, Y, Z]
T
The absolute velocity vector with respect to the inertial frame is:
{v} = { r} + []{r} (2.6)
The absolute acceleration vector with respect to the inertial frame is:
{a} = {r} + 2[]{ r} + [][]{r}
The displacement vector {u} in the principal beam axis with reference to the local coordi-
nates xyz is:
{u} = [u, v, w]
T
where u, v, w are displacements along the xyz axes respectively. The displacement vector
{ u} in the principal beam axis with reference to the moving coordinates XY Z is:
{ u} = [R]
T
{u}
where [R] is the transformation between the global coordinates XY Z and the local coordi-
nates xyz. The elements in [R] are:
[R] =
_
11
12
13
21
22
23
31
32
33
_
_
(2.7)
For the case of zero coning and zero pitch, [R] is equal to [T
EP
], as in Eq. 2.3. Also, the
20
unit vector {s} is:
{s} = [
31
,
32
,
33
]
T
(2.8)
The position vector of a point on the deformed beam with reference to frame XY Z is:
{r} = {r
0
} + { u} = {r
0
} + [R]
T
{u} (2.9)
and:
{ r} = [R]
T
{ u}, because { r
0
} = 0 (2.10)
2.2.3 Lagranges Equation
Lagranges equation in vector form is:
d
dt
_
T
{ u}
_
T
{u}
+
U
{u}
= {F} (2.11)
where T and U are the kinetic and strain energies, respectively, and {F} is the generalized
force vector. For the free vibration problem, {F} = 0. The kinetic energy of the beam is:
T =
1
2
_
A{v}
T
{v}ds (2.12)
where is the mass density and A is the sectional area. The strain energy of the beam is:
U =
1
2
_
EA
_
w
s
_
2
ds +
1
2
_
EI
x
_
2
v
s
2
_
2
ds +
1
2
_
EI
y
_
2
u
s
2
_
2
ds +
1
2
_
F(s)
_
u
s
_
2
ds + (2.13)
1
2
_
F(s)
_
v
s
_
2
ds
At this point the analysis does not include torsion of the element, which is assumed in Le-
ung and Fung [1] to be uncoupled from rotation. The rst three terms in the strain energy
21
equation 2.14 are the standard beam extension and bending terms, which are found in nite
element texts such as Raos [28]. The last two terms account for the work done by the
element axial force F(s), which arises from rotation. They account for the beams resis-
tance to bending due to centrifugal forces, and are sometimes called axial reduction factors.
Their derivation, described in Section 3.3.1, are from the theory of vibrating strings, which
Meirovitch [41] describes in the context of rotating beams.
Substituting Eqs. 2.9 and 2.10 into Eq. 2.6 the absolute velocity vector is:
{v} = [R]
T
{ u} + []({r
0
} + [R]
T
{u}) (2.14)
The term in the kinetic energy equation 2.12 is therefore:
{v}
T
{v} = { u}
T
[R][R]
T
{ u} + {u}
T
[R][]
T
[][R]
T
{u} +
{r
0
}
T
[]
T
[]{r
0
} + 2{ u}
T
[R][]{r
0
} + (2.15)
2{ u}
T
[R][][R]
T
{u} + 2{r
0
}
T
[]
T
[]{u}
From Eqs. 2.7 and 2.5:
[R][R]
T
= [I] and []
T
[] = [
2
] =
2
_
_
0 0 0
0 1 0
0 0 1
_
_
(2.16)
For the nite element method, the element displacement vector {u
e
} is interpolated
from the nodal coordinate vector {q
e
} as in:
{u
e
} = [N]{q
e
} and { u
e
} = [N]{ q
e
} (2.17)
22
where N is the shape function matrix:
[N] =
_
_
N
1
N
2
N
3
_
_
(2.18)
Section 2.2.6 describes the development of these N terms. The nodal coordinate vector
{q
e
} is:
{q
e
} = [{q
1
}
T
, {q
2
}
T
, {q
3
}
T
]
T
(2.19)
where:
{q
1
}
T
= [u
1
,
y1
, u
2
,
y2
] (2.20)
are the bending nodal displacements in the xz plane, and:
{q
2
}
T
= [v
1
,
x1
, v
2
,
x2
] (2.21)
are the bending nodal displacements in the yz plane, and:
{q
3
}
T
= [w
1
, w
2
] (2.22)
are the axial nodal displacements.
From Eqs. 2.12 and 2.16 the kinetic energy for an element is:
T
e
=
1
2
{ q
e
}
T
[M
e
]{ q
e
} +
1
2
{q
e
}
T
[K
e
]{q
e
} + T
0e
+
{ q
e
}
T
{f
e
} + { q
e
}
T
[G
e
]{q
e
} + {F
e
}
T
{q
e
} (2.23)
23
where (following Leung and Fung [1]):
[M
e
] =
_
l
0
A[m]ds, [K
e
] =
_
l
0
A[k
]ds,
[G
e
] =
_
l
0
A[g]ds, {f
e
} =
_
l
0
A[N]
T
[R][]{r
0
}ds, (2.24)
{F
e
} =
_
l
0
A[N]
T
[f
x
, f
y
, f
z
]
T
ds,
T
0e
=
1
2
_
l
0
A{r
0
}
T
[
2
]{r
0
}ds
where M
e
is the element mass matrix, K
e
is the element spin-stiffness matrix, and G
e
is
the element gyroscopic matrix. The vectors {f
e
} and {F
e
} and the scalar T
0e
drop out of
the ensuing analysis. The terms in the element matrices are:
[m] =
_
_
N
T
1
N
1
N
T
2
N
2
N
T
3
N
3
_
_
[g] =
_
_
0 b
1
N
T
1
N
2
b
2
N
T
1
N
3
0 b
3
N
T
2
N
3
skew symmetric 0
_
_
(2.25)
[k
] =
2
_
_
a
11
N
T
1
N
1
a
12
N
T
1
N
2
a
13
N
T
1
N
3
a
22
N
T
2
N
2
a
23
N
T
2
N
3
symmetric a
33
N
T
3
N
3
_
_
where:
a
11
=
2
12
+
2
13
, a
12
=
12
22
+
13
23
,
a
13
=
12
32
+
13
33
, a
22
=
2
22
+
2
23
, (2.26)
a
23
=
22
32
+
23
33
, a
22
=
2
22
+
2
23
24
and:
b
1
=
13
22
12
23
,
b
2
=
13
32
12
33
, (2.27)
b
3
=
23
32
22
33
The total kinetic energy comes from assembly of the nite element matrices in typical
fashion, as described in Rao [28], with the matrices transformed to the XY Z coordinate
systems and the common nodes added for the compatibility relation. The total kinetic
expression is:
T =
e
T
e
=
1
2
{ q}
T
[M]{ q} +
1
2
{q}
T
[K
]{q} + T
0
+
{ q}
T
{f } + { q}
T
[G]{q} + {F
}
T
{q}, (2.28)
where {q} is the global nodal displacement vector in the rotating reference frame XY Z.
Similar to the kinetic energy, the total strain energy for an element is:
U
e
=
1
2
{q
e
}
T
([K
ee
] + [K
ge
]){q
e
}, (2.29)
where:
[K
ee
] =
_
_
EI
x
[K
a
]
EI
y
[K
b
]
EA[K
c
]
_
_
[K
ge
] =
_
_
[K
ga
]
[K
gb
]
0
_
_
, (2.30)
25
where K
ee
is the element elastic stiffness matrix and K
ge
is the element axial force stiffness
matrix. The terms in the element matrices are:
[K
a
] =
_
l
0
[N
1
]
T
[N
1
]ds, [K
b
] =
_
l
0
[N
2
]
T
[N
2
]ds,
[K
c
] =
_
l
0
[N
3
]
T
[N
3
]ds, (2.31)
[K
ga
] =
_
l
0
F(s)[N
1
]
T
[N
1
]ds, [K
gb
] =
_
l
0
F(s)[N
2
]
T
[N
2
]ds,
where the prime denotes differentiation with respect to s.
The total strain energy is:
U =
e
U
e
=
1
2
{q}
T
([K
e
] + [K
g
]){q}. (2.32)
2.2.4 Steady State Equations and Axial Force
The nodal displacement vector {q} is divided into the steady state displacement {q
s
}
and the oscillation about steady state, {q
v
}, so that:
{q} = {q
s
} + {q
v
} (2.33)
For steady state the oscillatory component and the derivatives are zero, as in:
{q
v
}, { q
v
}, { q
s
}, and{ q} = 0 (2.34)
Inserting Eqs. 2.33 and 2.34 into Eqs. 2.28 and 2.32 the steady state kinetic energy be-
comes:
T =
1
2
{q
s
}
T
[K
]{q
s
} + T
0
+ {F
}
T
{q
s
}, (2.35)
26
and the steady state strain energy becomes:
U =
1
2
{q
s
}
T
([K
e
] + [K
g
]){q
s
}. (2.36)
The Lagrangian equation 2.11 then reduces to:
T
{q
s
}
+
U
{q
s
}
= 0, (2.37)
and with Eqs. 2.35 and 2.36:
([K
e
] + [K
g
] [K
]){q
s
} = {F
}. (2.38)
The blade is assumed to be a cantilever beam and therefore the internal forces are
predetermined. The axial force stiffness matrix [K
g
] is therefore independent of the steady
state deformation. The derivation of the axial force F(s) assumes that the centrifugal force
does not change with deection. From Leung and Fung [1] the centrifugal force per unit
length along the beam is:
= A{s}
T
[][]{r
0
}. (2.39)
From Eqs. 2.4, 2.5, and 2.8, the above equation becomes:
= A(a + bs) (2.40)
where:
a =
2
(
32
Y
0
+
33
Z
0
) (2.41)
and:
b =
2
(
2
32
+
2
33
) (2.42)
and Y
0
and Z
0
are the coordinates of the element inboard endpoint {r
g
}. The axial force at
27
a point along the element, F(s) (Figure 2.4) is:
F(s) = F
0
_
s
0
(s)ds (2.43)
where F
0
is the axial force at the inboard node of the element.
F
0
F(s)
s
(s)
Figure 2.4: Beam element axial force (after Leung [1])
Substituting Eq. 2.40 into Eq. 2.43:
F(s) = F
0
_
s
0
A(a + bs)ds
= F
0
_
s
0
(s)(a + bs)ds (2.44)
where (s) is the lineal density (kg/m) along the element. The model assumes linearly
tapered properties, so the lineal density is:
(s) =
_
0
+
l
0
l
s
_
(2.45)
where the zero subscript represents the property at the inboard node, and the l subscript
represents the property at the outboard node. Substituting Eq. 2.45 into Eq. 2.44 and
integrating:
F(s) = F
0
cs
ds
2
2
es
3
3
(2.46)
28
where:
c =
0
a
d =
(
l
0
)a +
0
bl
l
(2.47)
e =
(
l
0
)b
l
2.2.5 Oscillation about Steady State
For a small oscillation about steady state, the nodal velocity is:
{ q} = { q
v
}, (2.48)
and the kinetic and strain energies are:
T =
1
2
{ q
v
}
T
[M]{ q
v
} +
1
2
{q
s
+q
v
}
T
[K
]{q
s
+q
v
} + T
0
+
{ q
v
}
T
{f } + { q
v
}
T
[G]{q
s
+q
v
}{F
}
T
{q
s
+q
v
}, (2.49)
and:
U =
1
2
{q
s
+q
v
}
T
([K
e
] + [K
g
]){q
s
+q
v
}. (2.50)
Substituting these into Eq. 2.11, the Lagrangian becomes:
[M]{ q
v
} + 2[G]{ q
v
} + ([K
e
] + [K
g
] [K
]){q
v
} = {0}. (2.51)
29
2.2.6 Beam Element Matrices
The beam element matrices are assembled using Eqs. 2.24 and 2.30. The inertia and
stiffness properties vary linearly along the element as in Eq. 2.45 for the mass and:
EA(s) =
_
EA
0
+
EA
l
EA
0
l
s
_
, (2.52)
for the extensional stiffness, and:
EI(s) =
_
EI
0
+
EI
l
EI
0
l
s
_
, (2.53)
for the bending stiffnesses, and:
GJ(s) =
_
GJ
0
+
GJ
l
GJ
0
l
s
_
, (2.54)
for the torsional stiffness, and:
J
(s) =
_
J
0
+
J
l
J
0
l
s
_
, (2.55)
for the polar moment of inertia about the centroidal axis.
The shape function matrices are:
[N
1
] =
_
(2s
3
3ls
2
+ l
3
)/l
3
, (s
3
2ls
2
+ l
2
s)/l
2
, (2.56)
(2s
3
3ls
2
)/l
3
, (s
3
ls
2
)/l
2
and:
[N
2
] =
_
(2s
3
3ls
2
+ l
3
)/l
3
, (s
3
2ls
2
+ l
2
s)/l
2
, (2.57)
(2s
3
3ls
2
)/l
3
, (s
3
ls
2
)/l
2
30
and:
[N
3
] = [(1 s/l), s/l] (2.58)
where l as the length of the element. These are standard shape (or interpolation) functions
for a space frame element, which can be found in Raos nite element text [28]. The nodal
degrees of freedom are ordered as in Figure 2.5
Lj
3
q
1
Lj
1
q
3
q
2
Lj
2
q
4
q
5
q
6
q
7
q
10
q
12
q
9
q
11
q
8
Figure 2.5: Nodal degrees of freedom
Appendix A list the entries for the beam element matrices.
2.2.7 Solution for Blade Frequencies and Mode Shapes
The model frequencies and mode shapes are from an eigenvalue solution of Equa-
tion 2.51. This equation cannot be solved with standard techniques because of the skew-
symmetric gyroscopic matrix. Meirovitch [42] outlines a solution that avoids complex
arithmetic, which combines the matrices into symmetric ones. Given the stiffness matrix:
K = [K
e
] + [K
g
] [K
], (2.59)
31
Meirovitch makes a modied stiffness matrix:
K
=
_
_
KM
1
K KM
1
K
G
T
M
1
K K+G
T
M
1
G
_
_
, (2.60)
and a modied mass matrix:
M
=
_
_
K 0
0 M
_
_
, (2.61)
which assembles into a standard eigenvalue problem with 2n entries:
K
{x
r
} =
r
M
{x
r
}, (2.62)
where n is the number of degrees of freedom, {x
r
} is an eigenvector (r = 1, 2, . . . , 2n),
and
r
is an eigenvalue. The solutions consist of n pairs of repeated eigenvalues and n
pairs of associated eigenvectors y
r
and z
r
, where y
r
is the real part and z
r
is the imaginary
part of eigenvector x
r
.
The author used Meirovitchs method for verication with Leungs results, which is
described in the Section 2.3.4. However, to use the modes from Meirovitchs method for
solving the response, a dynamics code would require twice the number of degrees of free-
dom; one mode for the real part and one mode for the complex part. Therefore the nal
method in CurveFEM assumed that the gyroscopic terms for the wind turbine blade are
considered small. This is standard practice in the industry. The blade motions perpendicu-
lar to the axis of rotation are small in comparison to motions parallel to the axis, as in blade
ap motion. Therefore, the analysis neglects the gyroscopic matrix and the eigensolution
for Equation 2.51 is:
K{x
r
} =
r
M{x
r
}, (2.63)
with K from Eq. 2.59. The solutions consist of n eigenvalues and n real eigenvectors. The
author veried this method for a rotating, tapered beam with published results from Baner-
32
jee, Su, and Jackson [43], which Section 2.3.3 describes. Other verications described
below are the deection and vibration of a non-rotating curved beam, and the vibration of
a rotating curved beam. Appendix B describes the nite element program structure.
2.3 Finite Element Method Verication Results
2.3.1 Curved Beam Deection Verication
In following Kosmatkas work [2], the author veried the nite element stiffness matri-
ces (K
e
) with an example curved beam under deection. Figure 2.6 shows the cantilever
beam. The beam had six elements with applied forces at the end in the x and y directions.
1 in.
2 in.
y
z
Material: Aluminum 6061-T6
20 in.
0 90
y
x
Finite element representation
(6 elements)
Figure 2.6: Curved beam for deection verication (after Kosmatka [2])
The nite element results compare well with analytical results from Roarks [44] for beams
33
of varying curvature, shown in Figure 2.7 for the x-force direction and Fig. 2.8 for the
y-force direction.
0.00E+00
2.00E-03
4.00E-03
6.00E-03
8.00E-03
1.00E-02
1.20E-02
1.40E-02
1.60E-02
1.80E-02
0 10 20 30 40 50 60 70 80 90
Beam Curvature, degrees
D
i
s
p
l
a
c
e
m
e
n
t
,
i
n
.
0.00E+00
2.00E-04
4.00E-04
6.00E-04
8.00E-04
1.00E-03
1.20E-03
1.40E-03
1.60E-03
S
l
o
p
e
(
n
o
t
e
n
e
g
a
t
i
v
e
)
Finite Element
Roark's
y-displacment
x-displacment
slope
F
x
= 100 lbs
Figure 2.7: Curve beam deection and slope for F
x
= 100 lb
34
0.00E+00
5.00E-03
1.00E-02
1.50E-02
2.00E-02
2.50E-02
3.00E-02
3.50E-02
4.00E-02
4.50E-02
0 10 20 30 40 50 60 70 80 90
Beam Curvature, degrees
D
i
s
p
l
a
c
e
m
e
n
t
,
i
n
.
0.00E+00
5.00E-04
1.00E-03
1.50E-03
2.00E-03
2.50E-03
3.00E-03
3.50E-03
S
l
o
p
e
(
n
o
t
e
n
e
g
a
t
i
v
e
)
Finite Element
Roark's
y-displacment
x-displacment
slope
F
y
= 100 lbs
Figure 2.8: Curve beam deection and slope for F
y
= 100 lb
35
2.3.2 Non-Rotating Curved Beam Modes Verication
Again following Kosmatkas work [2], the author veried the nite element stiffness
and mass matrices (K
e
and M) with an example of curved beam free-response. Figure
2.9 shows the curved beam with xed-xed boundary conditions. Table 2.2 shows the
1 in.
2 in.
y
z
Material: aluminum 6061-T6
Six finite elements
20 in.
90
x
y
Figure 2.9: Curved beam for natural response (after Kosmatka [2])
nite element results compared to Kosmatka [2] and Blevins [45]. The results are in good
agreement with previous results for curved beams.
Table 2.2: Curved beam modes verication
Description Frequency from Frequency from Finite
Description Blevins [45] (Hz) Kosmatka [2] element
First out-of-plane 473.2 472.2 475
Second out-of-plane - 1343.8 1355
First in-plane 2277.5 2237.3 2231
2.3.3 Rotating Tapered Beam Verication
The author then conducted simulations with a tapered rotating beam to compare to
work by Banerjee, Su, and Jackson [43]. Banerjee et al uses the dynamic stiffness method
36
for solution, neglects the Coriolis terms and only considers the apping (or out-of-plane)
motion. The authors verication also neglected the Coriolis matrix, and did not used
Meirovitchs method for solution. Instead, the model used the standard eigenvalue solution
as in Eq. 2.63.
The verication used the beam with dimensions and properties shown in Figure 2.10.
y
x
z
E = 1.00 10
7
psi
= 0.00305 slugs/in
3
20
2
4
2
1
Figure 2.10: Rotating tapered beam (all dimensions in inches)
The author also compared with results for a non-spinning tapered beam by Mabie and
Rogers [46] and with an equivalent model with the NREL MODES program [29], described
in Section 2.1. The verication was conducted at rotational speeds representing the range
of Banerjees et al analysis. Figure 2.11 shows the verication results for the rst three
modes, which show good agreement amongst the methods.
37
0
1000
2000
3000
4000
5000
0 100 200 300 400 500 600 700 800 900
Rotational RPM
N
a
t
u
r
a
l
f
r
e
q
u
e
n
c
y
,
r
a
d
/
s
Banerjee (2006)
CurveFEM
Modes Program
Mabie (1972)
Mode 1
Mode 2
Mode 3
Figure 2.11: Tapered beam verication results
38
2.3.4 Rotating Beams from Leung and Fung 1988 Verication
The author compared the nite element program against results from Leung [1] for
several examples. The rst example was for a rotating horizontal cantilever (Figure 2.12)
of four elements.
Y
Z
X
Figure 2.12: Horizontal Cantilever (after Leung [1])
The verication runs were at zero rotational speed and a non-dimensional spinning
rate of 20 shown in Figure 2.13, with the lines digitized from Leung [1] and the triangles
representing the authors results.
The results show good agreement at zero rotational speed, which is comparable to re-
sults available in Roarks [44]. For the maximum rotational speed, the rst two bending
modes match well, but for CurveFEM the third and fourth modes have slightly lower natu-
ral frequency.
39
0
10
20
30
40
50
60
0 2 4 6 8 10 12 14 16 18 20
Spinning rate (Al
4
/EI )
1/2
N
a
t
u
r
a
l
f
r
e
q
u
e
n
c
y
A
l
4
/
E
I
)
1
/
2
Lines from Leung and Fung
Symbols are CurveFEM
Figure 2.13: Horizontal Cantilever Verication Results
40
The next example is a rotating beam inclined 45
A
l
4
/
E
I
)
1
/
2
Lines from Leung and Fung
Symbols are CurveFEM
Figure 2.15: Inclined Cantilever Verication Results
42
The next two examples are for an L- beam shown in Figure 2.16.
A
B
C
L
1
L
2
Figure 2.16: L-beam (after Leung [1])
The results show good agreement at zero rotational speed. For the 90
beam (Figure
2.17), the second and third modes at the highest rotational speed show lower natural fre-
quency for CurveFEM.
For the 30
beam (Figure 2.18), the results showgood agreement for the rst and second
modes at the highest rotational speed, and slightly lower frequency for the third mode for
CurveFEM.
43
0
20
40
60
80
100
0 4 8 12 16 20 24 28
Spinning rate (rad/s)
N
a
t
u
r
a
l
f
r
e
q
u
e
n
c
y
(
r
a
d
/
s
)
Lines from Leung and Fung
Symbols are CurveFEM
Figure 2.17: L-beam with = 90
Verication Results
44
0
20
40
60
80
100
0 4 8 12 16 20 24 28
Spinning rate (rad/s)
N
a
t
u
r
a
l
f
r
e
q
u
e
n
c
y
(
r
a
d
/
s
)
Lines from Leung and Fung
Symbols are CurveFEM
Figure 2.18: L-beam with = 30
Verication Results
45
2.3.5 Rotating Curved Beam Verication
The author veried CurveFEMwith rotating curved beamresults fromWang and Mahren-
holtz [47]. This is the only reference the author found for modes of rotating curved beams.
Wang and Mahrenholtz developed a model for bending of the minor axis of a rotating
curved beam with the Bernoulli-Euler approximation and with Coriolis terms neglected.
The authors used a Galerkin method for solution. The author compared results for straight
beams and beams with inplane curvature of 0.6 (arclength/radius) and zero hub radius. The
beam properties were according to the following relation:
EI
L
4
= 1,
where L is the total beam length. The author used 20 elements of equal length for the
CurveFEM model. The major axis bending, torsional, and extensional stiffnesses were
1000 times the value of the inplane bending stiffness. Figure 2.19 shows the comparison.
As in Wang and Mahrenholtz [47], the CurveFEM results show little difference in fre-
quency between straight and curved beams. The results match well for the rst mode. For
the second mode, both analysis predict lower frequencies for the curved blade; however,
CurveFEM predicts lower frequencies overall.
46
0
5
10
15
20
25
30
35
0 1 2 3 4 5 6 7 8 9 10
Rotational rate, rad/s
N
a
t
u
r
a
l
f
r
e
q
u
e
n
c
y
,
r
a
d
/
s
CurveFEM straight
CurveFEM curved
Wang straight
Wang curved
Second Mode
First Mode
Figure 2.19: Rotating curved beam verication results
47
2.4 Conclusions and Recommendations
The verication results show excellent agreement between the authors nite element
analysis and previously published results for curved beam deection and natural frequen-
cies. The results also show excellent agreement with previous work on rotating tapered
beams. For rotating beams that are not straight, the results show good agreement for
the fundamental mode; however, CurveFEM underpredicts the second bending mode fre-
quency. This may be due to differences in the solution method. Wang and Mahrenholtz
[47] do not mention their numerical procedure, and Leung and Fung [1] use a Newtonian
procedure. Without experimental results, it is not clear which method is most correct.
For future work, the author recommends the following:
Add pitch angle to the analysis
Add coning angle to the analysis
Add center of mass and elastic axis offsets to the analysis
Verify natural frequencies and mode shapes with experimental data on rotating curved
beams
The following chapter describes how the author used the CurveFEM-derived mode
shapes and frequencies in a modied wind-turbine dynamics code.
48
Chapter 3
Dynamic Analysis of Curved Wind
Turbine Blades
3.1 Introduction
This chapter presents the development of the equations of motion for the curved blade
to be used in the wind turbine analysis code FAST. The new program is called CurveFAST
to distinguish it from the existing version of FAST. This development is separate from the
nite element modeling of the blade that is used to solve for the blade mode shapes, as
explained in Chapter 2. These mode shapes enter in the equations of motion, shown below.
This development follows the theoretical development of FAST. Currently there is no
FAST theory manual, however, the author obtained from the developer, Jason Jonkman
of NREL, several documents outlining the equations of motion for FAST. Much of the
background is also documented in Jonkmans masters thesis [26], which worked from the
original code developed at Oregon State University [23]. The basis of the method is known
as Kanes Method, which can be found in Kane and Levinsons text [25].
FAST uses the aerodynamic subroutines of AeroDyn [17] for calculation of the aero-
dynamic forces. Moriarity and Hansen [48] cover the theory for AeroDyn. AeroDyn uses
49
blade element momentum (BEM) theory but has several improvements to account for un-
steadiness and asymmetric loading in wake. For CurveFAST, the author modied version
6.10a of FAST that used version 12.58 of AeroDyn.
The following development begins with dening coordinate systems, then moves sys-
tematically to kinematics and then kinetics.
3.2 Coordinate Systems and Transformations
The section describes the coordinate systems unique to CurveFAST. They are similar to
those from Section 2.2.1; however, the transformations are different in order to match those
used by the FAST2ADAMS.f90 pre-processor. The author retained the FAST2ADAMS.f90
transformations because CurveFAST would be veried with Adams.
3.2.1 Local Undeected Axis
The rst coordinate system unique to CurveFAST is the blade element system aligned
with the local undeected-axis with unit vectors Lj
B1
i
(i = 1, 2, 3). The superscript B1
refers to blade number one, and there are similar coordinate systems for the other blades.
This coordinate systemis transformed fromthe pitched coordinate system, with unit vectors
j
B1
1
, j
B1
2
, and j
B1
3
. The transformation is:
_
_
Lj
B1
1
(r)
Lj
B1
2
(r)
Lj
B1
3
(r)
_
_
= [T
EP
(r)]
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
(3.1)
The rst stage involves a transformation matrix [
1
] between the pitched coordinates
XY Z and the coordinates x y z (Fig. 3.1):
50
j
3
j
1
j
2
X
Z
Y
3
j
2
j
1
j
x
z
y
l
i
j
Figure 3.1: First transformation of element
_
_
x
y
z
_
_
= [
1
]
_
_
X
Y
Z
_
_
The next step is between the local coordinates xyz (principal axes) and the coordinates x y z
as (Fig. 3.2):
_
_
x
y
z
_
_
= [
2
]
_
_
x
y
z
_
_
The desired transformation between the xyz and the XY Z system is therefore:
[T
EP
(r)] = [
2
][
1
]
51
Lj
2
Lj
1
x
y
2
j
1
j
x
y
_
x
y
z
_
_
= [T
EP
(r)]
_
_
X
Y
Z
_
_
The rst transformation,
1
assumes small rotations about the X and Y axis, with:
1
=
_
TransMat(
1
=
B1
X
(r),
2
=
B1
Y
(r), 0)
where TransMat is the orthonormal transformation matrix subroutine used in FAST for
small rotations about the three axes. The rotations at a particular analysis node i are:
B1
X
i
=
X
i+1
X
i1
Z
i+1
Z
i1
B1
Y
i
=
Y
i+1
Y
i1
Z
i+1
Z
i1
where X is the blade precurve (RefAxisxb), Y is the blade presweep (RefAxisyb) and
Z is the length along the j
3
axis RNodes.
52
After the rst transformation, the principal axes of the element (xyz axes) make an
angle
s
(built-in structural twist) about the negative z axis. This direction is such that a
positive
s
results in a lower angle of attack. The transformation between the systems is:
_
_
x
y
z
_
_
=
_
_
cos
s
sin
s
0
sin
s
cos
s
0
0 0 1
_
_
_
_
x
y
z
_
_
= [
2
]
_
_
x
y
z
_
_
so that:
[
2
] =
_
_
cos
s
sin
s
0
sin
s
cos
s
0
0 0 1
_
_
3.2.2 Local Deected Coordinate System
The next transformation is from the undeected element to the deected coordinate
system, which is:
_
_
n
B1
1
(r)
n
B1
2
(r)
n
B1
3
(r)
_
_
= [T
DE
(r)]
_
_
Lj
B1
1
(r)
Lj
B1
2
(r)
Lj
B1
3
(r)
_
_
with
[T
DE
(r)] =
_
TransMat(
1
=
B1
Lx
(r),
2
=
B1
Ly
(r)
3
=
B1
Lz
(r)
B1
Lx
(r)
B1
Ly
(r)
B1
Lz
(r)
_
_
= [T
EP
(r)]
_
B1
jx
(r)
B1
jy
(r)
B1
jz
(r)
_
_
(3.2)
53
The rotations in the j system (
jx
,
jy
,
jz
) are:
B1
jx
(r) =
B1
4 1
(r)
4
(r) q
B1F1
+
B1
4 2
(r)
4
(r) q
B1F2
+
B1
4 3
(r)
4
(r) q
B1E1
+
B1
4 4
(r)
4
(r) q
B1T1
B1
jy
(r) =
B1
5 1
(r)
5
(r) q
B1F1
+
B1
5 2
(r)
5
(r) q
B1F2
(3.3)
+
B1
5 3
(r)
5
(r) q
B1E1
+
B1
5 4
(r)
5
(r) q
B1T1
B1
jz
(r) =
B1
6 1
(r)
6
(r) q
B1F1
+
B1
6 2
(r)
6
(r) q
B1F2
+
B1
6 3
(r)
6
(r) q
B1E1
+
B1
6 4
(r)
6
(r) q
B1T1
In these rotation equations, as an example, the symbol
B1
4 1
(r) is the eigenvector com-
ponent (subscript-prex 4) in the pitched (j) system for the x-component of slope (sweep
deection) for the rst blade ap mode (subscript-sufx 1) at a particular radial station for
blade 1. The symbol
4
(r) represents the interpolation function for the x-component of
slope. The CurveFAST analysis nodes coincide with the nite element nodes and the
interpolation functions are equal to 1. In the current version of the code the variable was
not be implemented. The symbol q
B1E1
is the rst ap mode degree of freedom for blade 1.
For reference, Fig. 2.5 shows the numbering of the element degrees of freedom. Table 3.1
shows CurveFASTs degrees of freedom for blade 1.
Table 3.1: CurveFAST Blade 1 Degrees of Freedom
Name Description
q
B1F1
Blade 1, First Flap Mode
q
B1F2
Blade 1, Second Flap Mode
q
B1E1
Blade 1, First Edge Mode
q
B1T1
Blade 1, First Torsion Mode
To reduce the number of multiplications during the simulation, the program uses Eq.
54
3.2 in the form:
_
B1
Lx
(r)
B1
Ly
(r)
B1
Lz
(r)
_
_
= [T
EP
(r)] []
_
_
q
B1F1
q
B1F2
q
B1E1
q
B1T1
_
_
with the matrix [] as:
[] =
_
B1
4 1
(r)
B1
4 2
(r)
B1
4 3
(r)
B1
4 4
(r)
B1
5 1
(r)
B1
5 2
(r)
B1
5 3
(r)
B1
5 4
(r)
B1
6 1
(r)
B1
6 2
(r)
B1
6 3
(r)
B1
6 4
(r)
_
_
The subroutine CalcT EP in the le FAST IO SML.f90 pre-multiplies the matrices
[T
EP
(r)] and [] at initialization.
3.2.3 Local Aerodynamics Coordinate System
The coordinate system for calculating and returning the aerodynamic loads is:
_
_
m
B1
1
(r)
m
B1
2
(r)
m
B1
3
(r)
_
_
=
_
_
cos
_
B1
p
+
B1
s
(r)
sin
_
B1
p
+
B1
s
(r)
0
sin
_
B1
p
+
B1
s
(r)
cos
_
B1
p
+
B1
s
(r)
0
0 0 1
_
_
n
B1
1
(r)
n
B1
2
(r)
n
B1
3
(r)
_
_
55
This is the same transformation as in FAST, with
p
representing the blade pitch and
s
rep-
resenting the built-in structural twist (both positive to lower angle of attack). The transfor-
mation does not include the
Lz
(elastic twist) term. The transformation to the trailing-edge
coordinate system, used in aeroacoustic calculations, is also the same as FAST, with:
_
_
te
B1
1
(r)
te
B1
2
(r)
te
B1
3
(r)
_
_
=
_
_
cos
_
B1
p
+
B1
a
(r)
sin
_
B1
p
+
B1
a
(r)
0
sin
_
B1
p
+
B1
a
(r)
cos
_
B1
p
+
B1
a
(r)
0
0 0 1
_
_
m
B1
1
(r)
m
B1
2
(r)
m
B1
3
(r)
_
_
where
a
is the local aerodynamic twist, a user input, which is usually equal to the structural
twist.
3.3 Kinematics
There are several references to the model degrees of freedom in CurveFAST in this
section; for example, the indices in the partial velocity derivatives refer to the degrees of
freedom. For reference, Table 3.2 shows these degrees of freedom for 2- and 3-bladed wind
turbines.
56
Table 3.2: CurveFAST Degrees of Freedom
No. 2 Bladed 3 Bladed Description
1 Sg Sg Platform Surge
2 Sw Sw Platform Sway
3 Hv Hv Platform Heave
4 R R Platform Roll
5 P P Platform Pitch
6 Y Y Platform Yaw
7 TFA1 TFA1 First Tower Fore-Aft
8 TSS1 TSS1 First Tower Fore-Aft
9 TFA2 TFA2 Second Tower Fore-Aft
10 TSS2 TSS2 Second Tower Fore-Aft
11 Yaw Yaw Yaw
12 RFrl RFrl Rotor Furl
13 GeAz GeAz Generator Azimuth
14 DrTr DrTr Drivetrain Flexibility
15 TFrl TFrl Tail Furl
16 B1F1 B1F1 Blade 1 First Flap Mode
17 B1E1 B1E1 Blade 1 First Edge Mode
18 B1F2 B1F2 Blade 1 Second Flap Mode
19 B1T1 B1T1 Blade 1 First Torsion Mode
20 B2F1 B2F1 Blade 2 First Flap Mode
21 B2E1 B2E1 Blade 2 First Edge Mode
22 B2F2 B2F2 Blade 2 Second Flap Mode
23 B2T1 B2T1 Blade 2 First Torsion Mode
24 Teet B3F1 Teeter or Blade 3 First Flap Mode
25 B3E1 Blade 3 First Edge Mode
26 B3F2 Blade 3 Second Flap Mode
27 B3T1 Blade 3 First Torsion Mode
57
3.3.1 Positions and Displacements
The position vector from the apex of the coned system (point Q) to the blade 1 node
(S1) is:
r
QS1
(r) =
_
h
B1
1
(r), h
B1
2
(r), hub radius + h
B1
3
(r)
_
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
+
_
u
B1
(r), v
B1
(r), w
B1
(r) + w
B1
AxRed
(r)
_
_
_
j
B1
1
(r)
j
B1
2
(r)
j
B1
3
(r)
_
_
(3.4)
where h
1
, h
2
, and h
3
are the undeected blade element offsets shown in Fig. 3.3. In the
code, the offsets are given by RefAxisxb(i), RefAxisyb(i), and RNodesNorm(i)
BldFlexL respectively. Also in the equation, u, v, and w are the deections in the j
system. The additional axial term w
AxRed
is the axial reduction term from blade bending.
j
3
h
1
j
1
h
3
h
2
j
2
Figure 3.3: Undeected blade element offsets
Similar to the angular deections (Eq. 3.3 and Fig. 2.5), the above deections in the j
58
system are:
u
B1
(r) =
B1
1 1
(r)
1
(r) q
B1F1
+
B1
1 2
(r)
1
(r) q
B1F2
+
B1
1 3
(r)
1
(r) q
B1E1
+
B1
1 4
(r)
1
(r) q
B1T1
v
B1
(r) =
B1
2 1
(r)
2
(r) q
B1F1
+
B1
2 2
(r)
2
(r) q
B1F2
(3.5)
+
B1
2 3
(r)
2
(r) q
B1E1
+
B1
2 4
(r)
2
(r) q
B1T1
w
B1
(r) =
B1
3 1
(r)
3
(r) q
B1F1
+
B1
3 2
(r)
3
(r) q
B1F2
+
B1
3 3
(r)
3
(r) q
B1E1
+
B1
3 4
(r)
3
(r) q
B1T1
with the values of the interpolation functions () again equal to one.
The axial reduction term is more complex. The innitesimal change in element length
(see Fig. 2.3) is:
dw
AxRed
= ds dl
AxRed
, (3.6)
where dl
AxRed
is the displaced length due to axial reduction, which is:
dl
AxRed
=
_
(ds)
2
+
_
u
s
ds
_
2
+
_
v
s
ds
_
2
_
1/2
(3.7)
The binomial expansion of Eq. 3.7 with the rst two terms is:
dl
AxRed
ds
_
1 +
1
2
_
u
s
_
2
+
1
2
_
v
s
_
2
_
(3.8)
From Eq. 3.6, the innitesimal change in element length is therefore:
dw
AxRed
ds
_
1
2
_
u
s
_
2
+
1
2
_
v
s
_
2
_
(3.9)
59
The total change in element length is:
w
AxRed
=
_
l
0
dw
AxRed
=
_
l
0
_
1
2
_
u
s
_
2
+
1
2
_
v
s
_
2
_
ds (3.10)
This equations resembles the terms in the strain energy relationship (Eq. 2.14), without the
axial force term. In the same manner as Eqs. 2.30 and 2.31:
[K
AxRede
] =
_
_
[K
AxReda
]
[K
AxRedb
]
0
_
_
, (3.11)
and:
[K
AxReda
] =
_
l
0
[N
1
]
T
[N
1
]ds, [K
AxRedb
] =
_
l
0
[N
2
]
T
[N
2
]ds. (3.12)
Appendix A lists the components of K
AxRede
. With the eigenvector matrix:
[] = [{
1
}, {
2
}, {
3
}, {
4
}]
and the vector of degree of freedom displacements:
{q} =
_
_
q
B1F1
q
B1F2
q
B1E1
q
B1T1
_
_
the deection due to axial reduction from Eq. 3.10 is:
w
AxRed
=
1
2
[[]{q}]
T
[K
AxRed
][[]{q}]
60
where [K
AxRed
] is the global axial reduction matrix. For a particular nodes axial reduction,
the eigenvectors and the axial reduction matrix include the element degrees of freedom at
the node in addition the degrees of freedom of all the inboard nodes. Using the transpose
property:
[[A][B]]
T
= [B]
T
[A]
T
the above relation becomes
w
AxRed
=
1
2
{q}
T
[]
T
[K
AxRed
][[]{q}]
Expanding:
w
AxRed
=
1
2
{{
1
}q
B1F1
+ {
2
}q
B1F2
+ {
3
}q
B1E1
+ {
4
}q
B1T1
}
T
[K
AxRed
]{{
1
}q
B1F1
+ {
2
}q
B1F2
+ {
3
}q
B1E1
+ {
4
}q
B1T1
}
Multiplying the terms through:
w
AxRed
=
1
2
_
{
1
}
T
[K
AxRed
]{
1
}q
2
B1F1
+ {
2
}
T
[K
AxRed
]{
2
}q
2
B1F2
+
{
3
}
T
[K
AxRed
]{
3
}q
2
B1E1
+ {
4
}
T
[K
AxRed
]{
4
}q
2
B1T1
+
2{
1
}
T
[K
AxRed
]{
2
}q
B1F1
q
B1F2
+ 2{
1
}
T
[K
AxRed
]{
3
}q
B1F1
q
B1E1
+
2{
1
}
T
[K
AxRed
]{
4
}q
B1F1
q
B1T1
+ 2{
2
}
T
[K
AxRed
]{
3
}q
B1F2
q
B1E1
+
2{
2
}
T
[K
AxRed
]{
4
}q
B1F2
q
B1T1
+ 2{
3
}
T
[K
AxRed
]{
4
}q
B1E1
q
B1T1
_
61
or:
w
AxRed
=
1
2
_
S
11
q
2
B1F1
+ S
22
q
2
B1F2
+ S
33
q
2
B1E1
+ S
44
q
2
B1T1
+
2S
12
q
B1F1
q
B1F2
+ 2S
13
q
B1F1
q
B1E1
+ 2S
14
q
B1F1
q
B1T1
+ (3.13)
2S
23
q
B1F2
q
B1E1
+ 2S
24
q
B1F2
q
B1T1
+ 2S
34
q
B1E1
q
B1T1
_
where:
S
ij
= {
i
}
T
[K
AxRed
]{
j
}; i, j = 1, 2, 3, 4
3.3.2 Velocities
The angular velocity of the blade 1 element body (M1) in the inertial frame is:
E
M1
(r) =
E
H
+
H
M1
(r) (3.14)
The analysis uses this angular velocity with the aerodynamic moments in the development
of the generalized active forces and for the blade twist in the generalized inertia force. The
simulation does not calculate the ap and edge rotation generalized inertia forces, which is
consistent with the Bernoulli hypothesis.
The angular velocity
E
H
is the angular velocity of the hub in the inertial frame and
the angular velocity of the element body in the hub frame is:
H
M1
(r) =
_
B1
jx
(r),
B1
jy
(r),
B1
jz
(r)
_
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
(3.15)
The time derivatives of the element angular deections come from taking the time deriva-
62
tive of Eq. 3.3:
B1
jx
(r) =
B1
4 1
(r)
4
(r) q
B1F1
+
B1
4 2
(r)
4
(r) q
B1F2
+
B1
4 3
(r)
4
(r) q
B1E1
+
B1
4 4
(r)
4
(r) q
B1T1
B1
jy
(r) =
B1
5 1
(r)
5
(r) q
B1F1
+
B1
5 2
(r)
5
(r) q
B1F2
+
B1
5 3
(r)
5
(r) q
B1E1
+
B1
5 4
(r)
5
(r) q
B1T1
B1
jz
(r) =
B1
6 1
(r)
6
(r) q
B1F1
+
B1
6 2
(r)
6
(r) q
B1F2
+
B1
6 3
(r)
6
(r) q
B1E1
+
B1
6 4
(r)
6
(r) q
B1T1
The velocity of a blade element node in the inertial frame is:
E
v
S1
(r) =
E
v
Q
+
H
v
S1
(r) +
E
H
r
QS1
(r) (3.16)
The velocity in the hub frame comes from the time derivative of Eq. 3.4:
H
v
S1
(r) =
H
dr
QS1
(r)
dt
=
_
u
B1
(r), v
B1
(r), w
B1
(r) + w
B1
AxRed
(r)
_
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
(3.17)
Pitch dynamics are not included in FAST, therefore the derivative of the j unit vectors are
zero. The time derivatives of the element deections come from the time derivatives of Eq.
63
3.5:
u
B1
(r) =
B1
1 1
(r)
1
(r) q
B1F1
+
B1
1 2
(r)
1
(r) q
B1F2
+
B1
1 3
(r)
1
(r) q
B1E1
+
B1
1 4
(r)
1
(r) q
B1T1
v
B1
(r) =
B1
2 1
(r)
2
(r) q
B1F1
+
B1
2 2
(r)
2
(r) q
B1F2
+
B1
2 3
(r)
2
(r) q
B1E1
+
B1
2 4
(r)
2
(r) q
B1T1
w
B1
(r) =
B1
3 1
(r)
3
(r) q
B1F1
+
B1
3 2
(r)
3
(r) q
B1F2
+
B1
3 3
(r)
3
(r) q
B1E1
+
B1
3 4
(r)
3
(r) q
B1T1
and the time derivative of Eq. 3.13:
w
AxRed
=
_
S
11
q
B1F1
q
B1F1
+ S
22
q
B1F2
q
B1F2
+ S
33
q
B1E1
q
B1E1
+ S
44
q
B1T1
q
B1T1
+
S
12
( q
B1F1
q
B1F2
+ q
B1F1
q
B1F2
) + S
13
( q
B1F1
q
B1E1
+ q
B1F1
q
B1E1
) +
S
14
( q
B1F1
q
B1T1
+ q
B1F1
q
B1T1
) + S
23
( q
B1F2
q
B1E1
+ q
B1F2
q
B1E1
) +
S
24
( q
B1F2
q
B1T1
+ q
B1F2
q
B1T1
) + S
34
( q
B1E1
q
B1T1
+ q
B1E1
q
B1T1
)
_
The partial angular velocity, used for Kanes method [25], comes from the partial ve-
locity derivative of Eq. 3.14:
E
M1
r
(r) =
E
H
r
+
H
M1
r
(r) (3.18)
with:
H
M1
r
(r) =
_
B1
jx
(r),
B1
jy
(r),
B1
jz
(r)
_
r
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
64
where the r subscript represents the partial velocity derivative. Expanding Eq. 3.18:
E
M1
r
(r) =
E
H
r
+
_
_
H
M1
B1F1
(r) for r = B1F1
H
M1
B1F2
(r) for r = B1F2
H
M1
B1E1
(r) for r = B1E1
H
M1
B1T1
(r) for r = B1T1
0 otherwise
(3.19)
with the additional partial velocites coming from the partial velocity derivative of Eq. 3.15:
H
M1
B1F1
(r) =
_
B1
4 1
(r)
4
(r),
B1
5 1
(r)
5
(r),
B1
6 1
(r)
6
(r)
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
H
M1
B1F2
(r) =
_
B1
4 2
(r)
4
(r),
B1
5 2
(r)
5
(r),
B1
6 2
(r)
6
(r)
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
H
M1
B1E1
(r) =
_
B1
4 3
(r)
4
(r),
B1
5 3
(r)
5
(r),
B1
6 3
(r)
6
(r)
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
H
M1
B1T1
(r) =
_
B1
4 4
(r)
4
(r),
B1
5 4
(r)
5
(r),
B1
6 4
(r)
6
(r)
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
65
The partial linear velocity comes from the partial velocity derivative of Eq. 3.16:
E
v
S1
r
(r) =
E
v
Q
r
+
H
v
S1
r
(r) +
E
H
r
r
QS1
(r) (3.20)
with:
H
v
S1
r
(r) =
_
u
B1
(r), v
B1
(r), w
B1
(r)
_
r
_
_
j
B1
1
j
B1
2
j
B1
3
_
_
Expanding Eq. 3.20:
E
v
S1
r
(r) =
E
v
Q
r
+
_
_
E
H
r
r
QS1
(r) for r = 4,5, . . . ,14
H
v
S1
B1F1
(r) for r = B1F1
H
v
S1
B1F2
(r) for r = B1F2
H
v
S1
B1E1
(r) for r = B1E1
H
v
S1
B1T1
(r) for r = B1T1
E
H
Teet
r
QS1
(r) for r = Teet
0 otherwise
(3.21)
66
with the additional partial velocites coming from the partial velocity derivative of Eq. 3.17:
H
v
S1
B1F1
(r) =
B1
1 1
(r)
1
(r)j
B1
1
+
B1
2 1
(r)
2
(r)j
B1
2
+
(
B1
3 1
(r)
3
(r) {S
11
q
B1F1
+ S
12
q
B1F2
+
S
13
q
B1E1
+ S
14
q
B1T1
})j
B1
3
H
v
S1
B1F2
(r) =
B1
1 2
(r)
1
(r)j
B1
1
+
B1
2 2
(r)
2
(r)j
B1
2
+
(
B1
3 2
(r)
3
(r) {S
22
q
B1F2
+ S
12
q
B1F1
+
S
23
q
B1E1
+ S
24
q
B1T1
})j
B1
3
H
v
S1
B1E1
(r) =
B1
1 3
(r)
1
(r)j
B1
1
+
B1
2 3
(r)
2
(r)j
B1
2
+
(
B1
3 3
(r)
3
(r) {S
33
q
B1E1
+ S
13
q
B1F1
+
S
23
q
B1F2
+ S
34
q
B1T1
})j
B1
3
H
v
S1
B1T1
(r) =
B1
1 4
(r)
1
(r)j
B1
1
+
B1
2 4
(r)
2
(r)j
B1
2
+
(
B1
3 4
(r)
3
(r) {S
44
q
B1T1
+ S
14
q
B1F1
+
S
24
q
B1F2
+ S
34
q
B1E1
})j
B1
3
3.3.3 Accelerations
In Kanes method [25], angular accelerations are:
E
N
i
( q, q, q, t) =
NDOF
r=1
E
N
i
r
(q, t) q
r
+
NDOF
r=1
d
dt
_
E
N
i
r
(q, t)
_
q
r
+
d
dt
_
E
N
i
t
(q, t)
_
for each body N
i
in the system.
The term required for the angular acceleration comes from the time derivative of the
partial angular velocity (Eq. 3.19):
67
d
dt
_
E
M1
r
(r)
=
d
dt
_
E
H
r
_
+
_
_
E
M1
B1F1
(r) for r = B1F1
E
M1
B1F2
(r) for r = B1F2
E
M1
B1E1
(r) for r = B1E1
E
M1
B1T1
(r) for r = B1T1
0 otherwise
Linear accelerations are:
E
a
X
i
( q, q, q, t) =
NDOF
r=1
E
v
X
i
r
(q, t) q
r
+
NDOF
r=1
d
dt
_
E
v
X
i
r
(q, t)
_
q
r
+
d
dt
_
E
v
X
i
t
(q, t)
_
(3.22)
for each point X
i
in the system.
The term required for the linear acceleration is given by the time derivative of the partial
68
velocity (Eq. 3.21):
d
dt
_
E
v
S1
r
(r)
=
d
dt
_
E
v
Q
r
_
+
_
_
E
H
r
_
H
v
S1
(r) +
E
H
r
QS1
(r)
for r = 4,5,6
d
dt
_
E
H
r
_
r
QS1
(r)+
E
H
r
_
H
v
S1
(r) +
E
H
r
QS1
(r)
H
v
S1
B1F1
(r) for r = B1F1
[S
22
q
B1F2
+ S
12
q
B1F1
+ S
23
q
B1E1
S
24
q
B1T1
]j
B1
3
+
E
H
v
S1
B1F2
(r) for r = B1F2
[S
33
q
B1E1
+ S
13
q
B1F1
+ S
23
q
B1F2
S
34
q
B1T1
]j
B1
3
+
E
H
v
S1
B1E1
(r) for r = B1E1
[S
44
q
B1T1
+ S
14
q
B1F1
+ S
24
q
B1F2
S
34
q
B1E1
]j
B1
3
+
E
H
v
S1
B1T1
(r) for r = B1T1
d
dt
_
E
H
Teet
_
r
QS1
(r)
+
E
H
Teet
_
H
v
S1
(r) +
E
H
r
QS1
(r)
for r = Teet
0 otherwise
3.4 Kinetics
Kanes equations of motion [25] for a simple holonomic system are:
F
r
+ F
r
= 0 (r = 1, 2, . . . , NDOF)
69
where the generalized active forces are:
F
r
=
w
i=1
E
v
X
i
r
F
X
i
+
E
N
i
r
M
N
i
and the generalized inertia forces are:
F
r
=
w
i=1
E
v
X
i
r
(m
N
i
E
a
X
i
) +
E
N
i
r
(
E
H
N
i
)
where the forces (F
X
i
) are applied at the center of mass point (X
i
) for each rigid body
(N
i
). The time derivative of the angular momentum of rigid body N
i
about its center of
mass X
i
in the inertial reference frame (E) is:
E
H
N
i
=
I
N
i
N
i
+
E
N
i
I
N
i
N
i
(3.23)
where
I
N
i
is the inertia dyadic of the body.
FAST uses Kanes equation of motion in matrix form, as in:
[C(q, t)]{ q} + {f( q, q, t)} = {0}
or:
[C(q, t)]{ q} = {f( q, q, t)}
3.4.1 Generalized Inertia Forces
For the generalized inertia forces, the program determines the contribution from each
blade element and the tip-node body. In the ensuing analysis, the author has only developed
the equations for blade 1; there are similar equations for the other blades.
70
The generalized inertia force for a blade 1 element that includes torsion is:
F
r
|
M1
(r) = NodeMass(r)
E
v
S1
r
(r)
E
a
S1
(r)dr (3.24)
+
E
M1
r
(r)
_
E
H
M1
(r)
_
where NodeMass is:
NodeMass(r) = AdjBlMs
B1
BMassDen
B1
(r) DRNodes
B1
(r)
where AdjBlMs is the blade mass adjustment factor in the blade le, BMassDen is the
blade mass density per unit length for the analysis node, and DRNodes is the element
length. Note that, as in FAST2ADAMS.f90 pre-processor, the node mass was not adjusted
for additional length from blade sweep and curvature. Therefore, there is an error that
grows with sweep angle. This can be countered by multiplying the blade mass density
with:
BMassDen
new
= BMassDen
old
Actual element length
DRNodes
With the addition of twist, the model includes the generalized inertia force for the twist
motion. For the twist motion the blade element dyadic is:
I
M1
(r) = ((InerBFLp
B1
(r) +InerBEdg
B1
(r))
DRNodes(r) +SmllNmbr)n
B1
3
(r)n
B1
3
(r)
The variable InerBFlp is the ap inertia per unit length. Similarly, InerBEdg is the
edge inertia per unit length. DRNodes is the element length, and SmllNmbr is a small
number (9.999E-4) so that an innite acceleration is not computed. This inertia is similar
to the one computed in FAST2ADAMS.f90
Using the Eq. 3.24 and previously derived terms, the total generalized inertia force for
71
blade 1 is:
F
r
|
B1
=
_
BldNodes
j=1
NodeMass(j)
E
v
S1
r
(j)
__
14
i=1
E
v
S1
i
(j) q
i
_
+
_
19
i=16
E
v
S1
i
(j) q
i
_
+
E
v
S1
Teet
(j) q
Teet
+
_
14
i=4
d
dt
_
E
v
S1
i
(j)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(j)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(j)
_
q
Teet
_
+
E
M1
r
(j)
_
I
M1
(j)
__
14
i=4
E
M1
i
(j) q
i
_
+
_
19
i=16
E
M1
i
(j) q
i
_
+
E
M1
Teet
(j) q
Teet
+
_
14
i=7
d
dt
_
E
M1
i
(j)
_
q
i
_
+
_
19
i=16
d
dt
_
E
M1
i
(j)
_
q
i
_
+
d
dt
_
E
M1
Teet
(j)
_
q
Teet
_
+
E
M1
(j)
I
M1
(j)
E
M1
(j)
__
+ . . .
Continuing with the blade 1 generalized inertia force:
72
F
r
|
B1
= . . . +
m
B1Tip E
v
S1
r
(BldFlexL)
__
14
i=1
E
v
S1
i
(BldFlexL) q
i
_
+
_
19
i=16
E
v
S1
i
(BldFlexL) q
i
_
+
E
v
S1
Teet
(BldFlexL) q
Teet
+
_
14
i=4
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(BldFlexL)
_
q
Teet
_
M1
r
(BldFlexL)
_
I
B1Tip
__
14
i=4
E
M1
i
(BldFlexL) q
i
_
+
_
19
i=16
E
M1
i
(BldFlexL) q
i
_
+
E
M1
Teet
(BldFlexL) q
Teet
+
_
14
i=7
d
dt
_
E
M1
i
(BldFlexL)
_
q
i
_
+
_
19
i=16
d
dt
_
E
M1
i
(BldFlexL)
_
q
i
_
+
d
dt
_
E
M1
Teet
(BldFlexL)
_
q
Teet
_
+
E
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
_
Rearranging the blade 1 generalized inertia forces (for a two-bladed turbine) into matrix
73
form:
[C(q, t)]|
B1
(Row, Col) =
_
BldNodes
j=1
NodeMass(j)
E
v
S1
Row
(j)
E
v
S1
Col
(j)+
E
M1
Row
(j)
I
M1
(j)
E
M1
Col
(j)
_
+
m
B1Tip E
v
S1
Row
(BldFlexL)
E
v
S1
Col
(BldFlexL) +
E
M1
Row
(BldFlexL)
I
B1Tip
M1
Col
(BldFlexL)
(Row, Col = 1, 2, . . . , 14; 16, 17, 18, 19; 24)
74
and:
{f( q, q, t)}|
B1
(Row) =
_
BldNodes
j=1
NodeMass(j)
E
v
S1
Row
(j)
__
14
i=4
d
dt
_
E
v
S1
i
(j)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(j)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(j)
_
q
Teet
_
+
E
M1
Row
(j)
_
I
M1
(j)
__
14
i=7
d
dt
_
E
M1
i
(j)
_
q
i
_
+
_
19
i=16
d
dt
_
E
M1
i
(j)
_
q
i
_
+
d
dt
_
E
M1
Teet
(j)
_
q
Teet
_
+
E
M1
(j)
I
M1
(j)
E
M1
(j)
__
m
B1Tip E
v
S1
Row
(BldFlexL)
__
14
i=4
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(BldFlexL)
_
q
Teet
_
M1
Row
(BldFlexL)
_
I
B1Tip
__
14
i=7
d
dt
_
E
M1
i
(BldFlexL)
_
q
i
_
+
_
19
i=16
d
dt
_
E
M1
i
(BldFlexL)
_
q
i
_
+
d
dt
_
E
M1
Teet
(BldFlexL)
_
q
Teet
_
+
E
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
_
(Row = 1, 2, . . . , 14; 16, 17, 18, 19; 24)
3.4.2 Generalized Active Forces
The blade 1 generalized active forces include the gravitational forces, the aerodynamic
forces, the elastic forces, and the damping forces, as in:
F
r
= F
r
|
GravB1
+ F
r
|
AeroB1
+ F
r
|
ElasticB1
+ F
r
|
DampB1
75
Blade Gravity Forces
The generalized forces due to gravity are in the same form as the current version of
FAST, and is:
F
r
|
GravB1
=
_
BldFlexL
0
B1
(r)g
E
v
B1
r
(r) z
2
dr
m
B1Tip
g
E
v
B1
r
(BldFlexL) z
2
(subscript r = 3, 4, . . . , 14; 16, 17, 18, 19; 24)
where g is the gravitational acceleration and z
2
is the tower base unit vector that points
vertical. In matrix form:
[C(q, t)]
GravB1
= 0
and:
{f( q, q.t)}|
GravB1
(Row) =
_
BldFlexL
0
B1
(r)g
E
v
B1
Row
(r) z
2
dr
m
B1Tip
g
E
v
B1
Row
(BldFlexL) z
2
(Row = 3, 4, . . . , 14; 16, 17, 18, 19; 24)
Blade Aerodynamic Forces
The application of the aerodynamic forces is the same as in FAST, with:
F
r
|
AeroB1
=
_
BldFlexL
0
_
E
v
S1
r
(r) F
S1
AeroB1
(r)+
E
M1
r
(r) M
M1
AeroB1
(r)
dr +
E
v
S1
r
(BldFlexL) F
S1
TipDragB1
(BldFlexL)
(subscript r = 1, 2, . . . , 14; 16, 17, 18, 19; 24)
76
where F
S1
AeroB1
(r) and M
M1
AeroB1
(r) are aerodynamic forces and pitching moments applied to
point S1 on blade 1 expressed per unit span. The pitching moment can include moments
due to airfoil moment coefcients (C
M
) and moments due to offsets of the aerodynamic
center to the element center of mass.
In matrix form the generalized active aerodynamic forces are:
[C(q, t)]
AeroB1
= 0
and:
{f( q, q.t)}|
AeroB1
(Row) =
_
BldFlexL
0
_
E
v
S1
Row
(r) F
S1
AeroB1
(r)+
E
M1
Row
(r) M
M1
AeroB1
(r)
dr +
E
v
S1
Row
(BldFlexL) F
S1
TipDragB1
(BldFlexL)
(Row = 1, 2, . . . , 14; 16, 17, 18, 19; 24)
As stated in Chapter 1, FAST uses the AeroDyn subroutines to compute the aerody-
namic forces. However, there is no accounting for blade sweep in AeroDyn. The author
therefore made corrections to the two-dimensional airfoil tables that are used to compute
the forces.
The aerodynamic correction for sweep takes into account that the pressure forces on
the blade are dependent on the relative velocity normal to the rotor leading edge. This is
the same analogy used for xed wing aircraft (see Hoerner [49]). This assumption is not
used by researchers in the rotorcraft community; however, researchers with the STAR pro-
gram studied the use of this assumption. They used a lifting-surface wake code originally
developed for helicopter rotors and modied for wind turbines [50] to study the effect of
sweep on a model wind turbine. It was found that the spanwise loading closely matched
the typical modication to the lift coefcient for sweep:
77
C
L
= C
L
cos
2
(3.25)
where is the angle between the relative velocity and the chord line of the blade section.
The drag coefcient was not altered for sweep. The results showed the assumption was
acceptable to about 20
(3.26)
Figure 3.4 shows the angle for the swept blade, assuming no out-of-plane curvature
(RefAxisxb). The sweep angle is the angle between the r velocity vector and the
perpendicular to the reference axis. The formula for is:
=
where is the angle between the pitch axis and radial position vector, and is:
=
where is the pre-sweep which is machined at the root face the balance the root pitching
moment and is the blade sweep angle. The formula is the arctangent of the deriva-
tive of the formula for the STAR-blade local sweep displacement of the reference axis
(RefAxisyb):
RefAxisyb = Tip Sweep (distance)
_
Blade Station Start of sweep
Blade Length Start of sweep
_
Sweep Exponent
(3.27)
for blade stations outboard of the sweep onset station.
78
R
N
o
d
e
s
Radial position vector
(no curvature,
RefAxisxb = 0)
)
Swept
reference axis
RefAxisyb
Local
Tangent
Pitch or
j
3
axis
Figure 3.4: Sweep angle for swept blade
79
In addition to the lifting surface studies, van Dam and Saephan built a computation
uid dynamics (CFD) model of the STAR [51] rotor, which showed good agreement to the
BEM corrected for sweep as above. The researchers therefore had further condence in
using this assumption for sweep.
The CurveFAST modication duplicates the current method used in the Adams model,
by supplying the relative velocities to AeroDyn in the plane perpendicular to the leading
edge, and the drag coefcients in the two-dimensional airfoil tables are adjusted according
to Eq. 3.26.
Blade Elastic Forces
For the blade elastic forces, the model uses the results from the blade nite element
analysis. From the FAST kinetics development [23] the Blade 1 generalized elastic force
is:
F
r
|
ElasticB1
=
V
B1
q
r
(3.28)
The model approximates the blade elastic potential energy by using four mode shapes, as
in:
V =
1
2
q
2
B1F1
{
1
}
T
{K
e
} {
1
} +
1
2
q
2
B1F2
{
2
}
T
{K
e
} {
2
} + (3.29)
1
2
q
2
B1E1
{
3
}
T
{K
e
} {
3
} +
1
2
q
2
B1T1
{
4
}
T
{K
e
} {
4
}
where for example {
1
} is the rst mode eigenvector and {K
e
} is the elastic stiffness
matrix for the non-rotating blade. The eigenvectors and stiffness matrix are referenced to
the pitched j system. CurveFAST internally calculates the elastic stiffness matrix to reduce
the truncation error in the matrix multiplications. The derivation of this matrix is in Chapter
80
2. The eigenvectors are included at the end of blade data input le.
Substituting Eq. 3.29 into Eq. 3.28:
F
r
|
ElasticB1
=
_
_
q
B1F1
{
1
}
T
{K
e
} {
1
} for r = B1F1
q
B1F2
{
2
}
T
{K
e
} {
2
} for r = B1F2
q
B1E1
{
3
}
T
{K
e
} {
3
} for r = B1E1
q
B1T1
{
4
}
T
{K
e
} {
4
} for r = B1T1
0 otherwise
The matrix multiplications of and K
e
are computed at initialization in the subroutine
coeff in the le FAST SML.f90.
In matrix form, the elastic forces for blade 1 are:
[[C(q, t)]|
ElasticB1
= 0
{f( q, q, t)}|
ElasticB1
(16) = q
B1F1
{
1
}
T
{K
e
} {
1
}
{f( q, q, t)}|
ElasticB1
(17) = q
B1E1
{
3
}
T
{K
e
} {
3
}
{f( q, q, t)}|
ElasticB1
(18) = q
B1F2
{
2
}
T
{K
e
} {
2
}
{f( q, q, t)}|
ElasticB1
(19) = q
B1T1
{
4
}
T
{K
e
} {
4
}
{f( q, q, t)}|
ElasticB1
(other rows) = 0
As in FAST, blade damping is assumed to be proportional to the stiffness, which is
a typical assumption in modal analysis. The damping, or dissipative function in the La-
grangian development from nite element analysis as in Rao [28], Eq. 12.20, is given
as:
R =
1
2
{ q}
T
[D]{ q} (3.30)
where Q is the vector of nodal displacments and {C} is the damping matrix. For a partic-
81
ular mode (e.g. rst blade ap):
{q}
B1F1
= q
B1F1
{
1
} (3.31)
Damping proportional to mass and stiffness is (Rao[28], Eq. 12.81):
[D] = a [M] + b [K
e
] (3.32)
where a and b are constants and [M] is the mass matrix. For a particular mode, the modal
damping constant is (Rao[28], Eq. 12.85):
i
=
a + b
2
i
2
i
where is the square root of the mode eigenvalue in rad/s. Setting a to zero and solving
for b:
b =
2
i
i
(3.33)
Substituting Eq. 3.33 into Eq. 3.32 for a particular mode:
[D]
i
=
2
i
i
[K
e
] (3.34)
Substituting Eqs. 3.31 and 3.34 into Eq. 3.30 the dissipation becomes:
R =
1
2
q
2
B1F1
2
1
1
{
1
}
T
{K
e
} {
1
} +
1
2
q
2
B1F2
2
2
2
{
2
}
T
{K
e
} {
2
} + (3.35)
1
2
q
2
B1E1
2
3
3
{
3
}
T
{K
e
} {
3
} +
1
2
q
2
B1T1
2
4
4
{
4
}
T
{K
e
} {
4
}
82
Similar to the elastic force the generalized damping force takes the following form:
F
r
|
DampB1
=
R
B1
q
r
(3.36)
Substituting Eq. 3.36 into Eq. 3.35:
F
r
|
DampB1
=
_
2
1
1
q
B1F1
{
1
}
T
{K
e
} {
1
} for r = B1F1
2
2
2
q
B1F2
{
2
}
T
{K
e
} {
2
} for r = B1F2
2
3
3
q
B1E1
{
3
}
T
{K
e
} {
3
} for r = B1E1
2
4
4
q
B1T1
{
4
}
T
{K
e
} {
4
} for r = B1T1
0 otherwise
In matrix form, the damping forces for blade 1 are:
[[C(q, t)]|
DampB1
= 0
{f( q, q, t)}|
DampB1
(16) =
2
1
1
q
B1F1
{
1
}
T
{K
e
} {
1
}
{f( q, q, t)}|
DampB1
(17) =
2
3
3
q
B1E1
{
3
}
T
{K
e
} {
3
}
{f( q, q, t)}|
DampB1
(18) =
2
2
2
q
B1F2
{
2
}
T
{K
e
} {
2
}
{f( q, q, t)}|
DampB1
(19) =
2
4
4
q
B1T1
{
4
}
T
{K
e
} {
4
}
{f( q, q, t)}|
DampB1
(other rows) = 0
83
3.5 Blade Loads
3.5.1 Blade Root Loads
FAST constructs the equations of motion with loads that are typical quantities of inter-
est. When the equations of motion are solved, these loads are readily available for output
and do not require further computations. For the blades, these loads are the blade root
forces and moments. Kanes method breaks the forces into the following (for a 2-bladed
turbine):
F
X
i
Source
( q, q, q, t) =
_
24
r=1
F
X
i
Sourcer
(q, t) q
r
_
+F
X
i
Sourcet
( q, q, t) (3.37)
where F
X
i
Sourcer
are the partial forces and F
X
i
Sourcet
is all the components of F
X
i
Source
not of
this form. For the moments:
M
N
i
@X
i
Source
( q, q, q, t) =
_
24
r=1
M
N
i
@X
i
Sourcer
(q, t) q
r
_
+M
N
i
@X
i
Sourcet
( q, q, t) (3.38)
where M
N
i
@X
i
Sourcer
are the partial moments and M
N
i
@X
i
Sourcet
are all the components of M
N
i
@X
i
Source
not of this form.
Blade 1s generalized active force in terms of the loads acting on the hub center of mass
(point C) is:
F
r
|
B1
=
E
v
C
r
F
C
B1
+
E
H
r
M
H
B1
(r = 1, 2, . . . , 24)
Because the hub is rigid, the forces on the hubs center of mass are related to the blade
root loads (F
S1
B1
(0) and M
H
B1
(0)) by:
F
S1
B1
(0) = F
C
B1
and:
M
H
B1
= M
H
B1
(0) +r
CS1
(0) F
S1
B1
(0)
84
or:
M
H
B1
= M
H
B1
(0) +
_
r
QS1
(0) r
QC
F
S1
B1
(0)
Because of the two-point velocity in Kane and Levinson [25]:
E
v
C
r
=
E
v
Q
r
+
E
H
r
r
QC
the generalized active force expands to:
F
r
|
B1
=
_
E
v
Q
r
+
E
H
r
r
QC
_
F
S1
B1
(0) +
E
H
r
_
M
H
B1
(0) +
_
r
QS1
(0) r
QC
F
S1
B1
(0)
_
(r = 1, 2, . . . , 24)
Applying the cyclic permutation law of the scalar triple product:
a (b c) = (a b) c
the generalized active force becomes:
F
r
|
B1
=
E
v
Q
r
F
S1
B1
(0) +
E
H
r
{r
QC
F
S1
B1
(0)} +
E
H
r
{M
H
B1
(0) +
_
r
QS1
(0) r
QC
F
S1
B1
(0)}
(r = 1, 2, . . . , 24)
85
which simplies to:
F
r
|
B1
=
E
v
Q
r
F
S1
B1
(0) +
E
H
r
_
M
H
B1
(0) +r
QS1
(0) F
S1
B1
(0)
(3.39)
(r = 1, 2, . . . , 24)
This generalized active force must produce the same effects as the generalized active and
inertia forces from blade 1. Therefore:
F
r
|
B1
= F
r
|
B1
+ F
r
|
AeroB1
+ F
r
|
GravB1
+ F
r
|
ElasticB1
+ F
r
|
DampB1
(r = 1, 2, . . . , 24)
Because
E
v
Q
r
and
E
H
r
are equal to zero unless r = 1, 2, . . . , 14; Teet, the blade elastic
and damping forces do not contribute to the root loads. Therefore:
F
r
|
B1
= F
r
|
B1
+ F
r
|
AeroB1
+ F
r
|
GravB1
(r = 1, 2, . . . , 14; Teet)
Expanding:
F
r
|
B1
=
_
BldFlexL
0
E
v
S1
r
(r)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +
_
BldFlexL
0
E
M1
r
(r)
_
M
M1
AeroB1
(r)
I
M1
(r)
E
M1
(r)
E
M1
(r)
I
M1
(r)
E
M1
(r)
_
dr +
E
v
S1
r
(BldFlexL)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
+
E
M1
r
(BladeFlexL)
_
I
B1Tip
M1
(BldFlexL)
E
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
(r = 1, 2, . . . , 14; Teet)
86
Because:
E
v
S1
r
(r) =
E
v
Q
r
+
H
v
S1
r
(r) +
E
H
r
r
QS1
(r)
the equation above becomes:
F
r
|
B1
=
_
BldFlexL
0
_
E
v
Q
r
+
H
v
S1
r
(r)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +
_
E
v
Q
r
+
H
v
S1
r
(BldFlexL)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
+
_
BldFlexL
0
_
E
H
r
r
QS1
(r)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +
_
E
H
r
r
QS1
(r)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
+
_
BldFlexL
0
E
M1
r
(r)
_
M
M1
AeroB1
(r)
I
M1
(r)
E
M1
(r)
E
M1
(r)
I
M1
(r)
E
M1
(r)
_
dr +
E
M1
r
(BladeFlexL)
_
I
B1Tip
M1
(BldFlexL)
E
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
(r = 1, 2, . . . , 14; Teet)
Because
H
v
S1
r
(r) is equal to zero (rigid hub) and
E
M1
r
(r) is equal to
E
H
r
with the con-
87
straint (r = 1, 2, . . . , 14; Teet), the preceeding equation simplies to:
F
r
|
B1
=
_
BldFlexL
0
E
v
Q
r
(r)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +
E
v
Q
r
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
+
_
BldFlexL
0
_
E
H
r
r
QS1
(r)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +
_
E
H
r
r
QS1
(r)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
+
_
BldFlexL
0
E
H
r
_
M
M1
AeroB1
(r)
I
M1
(r)
E
M1
(r)
E
M1
(r)
I
M1
(r)
E
M1
(r)
_
dr +
E
H
r
_
I
B1Tip
M1
(BldFlexL)
E
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
(r = 1, 2, . . . , 14; Teet)
With the cyclic permutation law of the scalar triple product, the preceeding equation be-
88
comes:
F
r
|
B1
=
_
BldFlexL
0
E
v
Q
r
(r)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +
E
v
Q
r
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
+
_
BldFlexL
0
E
H
r
_
r
QS1
(r)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
_
dr +
E
H
r
_
r
QS1
(BldFlexL)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
__
+
_
BldFlexL
0
E
H
r
_
M
M1
AeroB1
(r)
I
M1
(r)
E
M1
(r)
E
M1
(r)
I
M1
(r)
E
M1
(r)
_
dr +
E
H
r
_
I
B1Tip
M1
(BldFlexL)
E
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
(r = 1, 2, . . . , 14; Teet)
The root force and moment come from the comparison of the preceeding equation with Eq.
(3.39):
F
S1
B1
(0) =
_
BldFlexL
0
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
89
and:
M
H
B1
(0) +r
QS1
(0) F
S1
B1
(0) =
_
BldFlexL
0
_
M
M1
AeroB1
(r)
I
M1
(r)
E
M1
(r)
E
M1
(r)
I
M1
(r)
E
M1
(r)
dr +
_
BldFlexL
0
r
QS1
(r)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +r
QS1
(BldFlexL)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
+
_
I
B1Tip
M1
(BldFlexL)
E
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
or:
M
H
B1
(0) =
_
BldFlexL
0
_
M
M1
AeroB1
(r)
I
M1
(r)
E
M1
(r)
E
M1
(r)
I
M1
(r)
E
M1
(r)
dr +
_
BldFlexL
0
r
QS1
(r)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +r
QS1
(BldFlexL)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
+
_
I
B1Tip
M1
(BldFlexL)
E
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
_
r
QS1
(0)
__
BldFlexL
0
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
90
or:
M
H
B1
(0) =
_
BldFlexL
0
_
M
M1
AeroB1
(r)
I
M1
(r)
E
M1
(r)
E
M1
(r)
I
M1
(r)
E
M1
(r)
dr +
_
BldFlexL
0
_
r
QS1
(r) r
QS1
(0)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +
_
r
QS1
(BldFlexL) r
QS1
(0)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
+
_
I
B1Tip
M1
(BldFlexL)
E
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
Therefore:
F
S1
B1
(0) =
_
BldFlexL
0
_
F
S1
AeroB1
(r)
B1
(r)
_
gz
2
+
_
14
i=1
E
v
S1
i
(r) q
i
_
+
_
19
i=16
E
v
S1
i
(r) q
i
_
+
E
v
S1
Teet
(r) q
Teet
+
_
14
i=4
d
dt
_
E
v
S1
i
(r)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(r)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(r)
_
q
Teet
__
+F
S1
TipDragB1
m
B1Tip
_
gz
2
+
_
14
i=1
E
v
S1
i
(BldFlexL) q
i
_
+
_
19
i=16
E
v
S1
i
(BldFlexL) q
i
_
+
E
v
S1
Teet
(BldFlexL) q
Teet
+
_
14
i=4
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(BldFlexl)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(BldFlexL)
_
q
Teet
_
91
and:
M
H
B1
(0) =
_
BldFlexL
0
_
M
M1
AeroB1
(r)
I
M1
(r)
__
14
i=4
E
M1
i
(r) q
i
_
+
_
19
i=16
E
M1
i
(r) q
i
_
+
E
M1
Teet
(r) q
Teet
+
_
14
i=7
d
dt
_
E
M1
i
(r)
_
q
i
_
+
_
19
i=16
d
dt
_
E
M1
i
(r)
_
q
i
_
+
d
dt
_
E
M1
Teet
(r)
_
q
Teet
_
M1
(r)
I
M1
(r)
E
M1
(r)
_
dr +
_
BldFlexL
0
_
r
QS1
(r)r
QS1
(0)
_
F
S1
AeroB1
(r)
B1
(r)
_
gz
2
+
_
14
i=1
E
v
S1
i
(r) q
i
_
+
_
19
i=16
E
v
S1
i
(r) q
i
_
+
E
v
S1
Teet
(r) q
Teet
+
_
14
i=4
d
dt
_
E
v
S1
i
(r)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(r)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(r)
_
q
Teet
__
dr +
_
r
QS1
(BldFlexL) r
QS1
(0)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
_
14
i=1
E
v
S1
i
(BldFlexL) q
i
_
+ . . .
92
Continuing the previous equation:
M
H
B1
(0) = . . . +
_
19
i=16
E
v
S1
i
(BldFlexL) q
i
_
+
E
v
S1
Teet
(BldFlexL) q
Teet
+
_
14
i=4
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(r)
_
q
Teet
__
+
_
I
B1Tip
__
14
i=4
E
M1
i
(BldFlexL) q
i
_
+
_
19
i=16
E
M1
i
(BldFlexL) q
i
_
+
E
M1
Teet
(BldFlexL) q
Teet
+
_
14
i=7
d
dt
_
E
M1
i
(BldFlexL)
_
q
i
_
+
_
19
i=16
d
dt
_
E
M1
i
(BldFlexL)
_
q
i
_
+
d
dt
_
E
M1
Teet
(BldFlexL)
_
q
Teet
_
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
B1
(r)
E
v
S1
r
(r)dr m
B1TipE
v
S1
r
(BldFlexL)
(r = 1, 2, . . . , 14; 16, 17, 18, 19; Teet)
and:
93
F
S1
B1t
(0) =
_
BldFlexL
0
_
F
S1
AeroB1
(r)
B1
(r)
_
gz
2
+
_
14
i=4
d
dt
_
E
v
S1
i
(r)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
E
v
S1
Teet
(r) q
Teet
__
dr
m
B1Tip
_
gz
2
+
_
14
i=4
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(BldFlexl)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(BldFlexL)
_
q
Teet
_
+
F
S1
TipDragB1
with the moment as in Eq. 3.38:
M
H
B1r
(0) =
_
BldFlexL
0
_
I
M1
(r)
E
M1
r
(r) +
_
r
QS1
(r)r
QS1
(0)
B1
(r)
E
v
S1
r
(r)
_
dr m
B1Tip
_
r
QS1
(BldFlexL) r
QS1
(0)
E
v
S1
r
(BldFlexL)
I
B1Tip
M1
r
(BldFlexL)
(r = 1, 2, . . . , 14; 16, 17, 18, 19; Teet)
94
and:
M
H
B1t
(0) =
_
BldFlexL
0
_
I
M1
(r)
__
14
i=7
d
dt
_
E
M1
i
(r)
_
q
i
_
+
_
19
i=16
d
dt
_
E
M1
i
(r)
_
q
i
_
+
d
dt
_
E
M1
Teet
(r)
_
q
Teet
_
M1
(r)
I
M1
(r)
E
M1
(r) +
_
r
QS1
(r)r
QS1
(0)
_
F
S1
AeroB1
(r)
B1
(r) {gz
2
+
_
14
i=4
d
dt
_
E
v
S1
i
(r)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(r)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(r)
_
q
Teet
___
dr +
_
r
QS1
(BldFlexL) r
QS1
(0)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
_
14
i=4
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
_
19
i=16
d
dt
_
E
v
S1
i
(BldFlexL)
_
q
i
_
+
d
dt
_
E
v
S1
Teet
(r)
_
q
Teet
__
+
_
I
B1Tip
__
14
i=7
d
dt
_
E
M1
i
(BldFlexL)
_
q
i
_
+
_
19
i=16
d
dt
_
E
M1
i
(BldFlexL)
_
q
i
_
+
d
dt
_
E
M1
Teet
(BldFlexL)
_
q
Teet
_
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
95
3.5.2 Blade Gage Moment Outputs
The blade gage moments are computed in the subroutine CalcOuts. The equation for
the moment at the gage location R
Span i
is:
M
M1
B1
(R
Span i
) =
_
BldFlexL
R
Span i
_
M
M1
AeroB1
(r) +
I
M1
(r)
E
M1
(r)
E
M1
(r)
I
M1
(r)
E
M1
(r)
_
dr +
_
BldFlexL
R
Span i
_
r
QS1
(r) r
QS1
(R
Span i
)
_
F
S1
AeroB1
(r)
B1
(r)gz
2
B1
(r)
E
a
S1
(r)
dr +
_
I
B1Tip
M1
(BldFlexL)
E
M1
(BldFlexL)
I
B1Tip
M1
(BldFlexL)
+
_
r
QS1
(BldFlexL) r
QS1
(R
Span i
)
_
F
S1
TipDragB1
m
B1Tip
_
gz
2
+
E
a
S1
(BldFlexL)
_
At the time of this writing, the author had not modied the blade gage moment outputs
in CurveFAST.
3.6 Verication with FAST and Adams
For the initial CurveFAST development, the author used a straight blade with no ta-
per with uniform inow to ensure that aerodynamics and load were matching between
CurveFAST, FAST and ADAMS. For the swept-blade verication, the author used the same
STAR7d model that was used in the STAR program development [10]. The blade prop-
erties were proprietary and could not be included in this document. For the verication, the
author compared 10-minute turbulent simulation runs at average speeds of ranging from
8-13 m/s. The author generated the wind les with NRELs TurbSim program [18]; the le
parameters are below in Table 3.3.
96
Table 3.3: Turbulent wind le parameters
Parameter Value
Random Number Generator RANLUX
Vertical grid point dimension 13 m
Horizontal grid point dimension 13 m
Time step 0.05 s
Time series length 660 s
Usable length of time series 630 s
Hub height 61 m
Grid Height 65 m
Grid Width 65 m
Vertical mean ow angle 8
.
Each model run was a 10-minute turbulent simulation, which had a nominal 10-minute
average wind speed from 3 m/s to 25 m/s in 2 m/s steps. The NREL program TurbSim
[18] generated the wind les with the Kaimal spectrum for the IECA normal turbulence
model. Each wind speed step consisted of ve simulations with the random seed equal to
the system clock.
The author compared the NTMresults by calculating the damage equivalent load (DEL),
the annual energy production (AEP), and the maximum blade deection.
The damage equivalent load is the load (or moment in this case) for a given number of
cycles (one million in this case) that will produce the same damage for the lifetime load
spectra. Freebury and Musial [54] and Sutherland [55] discuss this method. The calculation
of the DEL involves the use of Miners rule, which is:
n
1
N
1
+
n
2
N
2
+ +
n
n
N
n
= Damage Fraction, (4.1)
where n is the number of cycles for a particular stress level, and N is the number of allow-
able cycles for a particular stress level. The damage fraction for failure is normally unity
(1.0). For the equivalent damage of one million cycles, Eq. 4.1 becomes:
116
n
1
N
1
+
n
2
N
2
+ +
n
n
N
n
=
1 10
6
N
eq
(4.2)
The number of allowable cycles, N, is from the W ohler (or S-N) curve:
N =
_
u
_
m
, (4.3)
where
a
is the applied stress,
u
is the ultimate stress, and m is the inverse of the W ohler
(or S-N) curve. The DEL analysis assumes a linear relation between stress and load (or
moment in this case). Equation 4.4 then becomes:
N =
_
M
a
M
u
_
m
, (4.4)
where M
a
is the applied moment and M
u
is the ultimate moment. Equation 4.4 is substi-
tuted into Eq. 4.2 to obtain:
n
1
_
M
a1
Mu
_
m
+
n
2
_
M
a2
Mu
_
m
+ +
n
n
_
Man
Mu
_
m
=
1 10
6
_
Meq
Mu
_
m
, (4.5)
where M
eq
is the damage equivalent moment for one million cycles. Solving for M
eq
:
M
eq
=
_
(M
ai
)
m
n
i
1 10
6
_
1/m
, (4.6)
The rainow cycle counting algorithm in NRELs Crunch program [56] determines the
number of cycles per simulation run for thirty moment levels. The author wrote an Excel
program to use the rainow counts to compute the DEL. The program adds up the cycles for
the ve 10-minute runs at each average wind speed step. It then multiplies these counts by
6/5 and by the annual number of hours for the wind speed assuming a Rayleigh distribution.
The Rayleigh probability distribution is from:
117
F(V ) = 1 exp
_
4
_
V
V
ave
_
2
_
, (4.7)
where F(V ) is the frequency that the velocity V or lower will occur, and V
ave
is the average
wind speed of the distribution. The author used an annual average wind speed of 8.5 m/s,
which corresponds to IEC class II. The number of hours at wind speed V
i
is from:
Hours at wind speed V
i
= 8760(F(V
i
) F(V
i1
)), (4.8)
with the calculation initiated at:
V
i1
= V
i
1 m/s (4.9)
The calculation for the annual energy production (AEP) averages the generator power
of the ve simulations at a particular average wind speed, and then multiplies by the number
of hours at that wind speed using Eq. 4.8. The total sum of these kilowatt-hours is the AEP.
The peak deection is simply the maximum deection from the model runs.
To determine the distribution in AEP and DEL for the baseline model, the author ran 5
separate 60-case runs. The standard deviation of AEP was 0.4%. The standard deviation of
ap and edge DEL (for W ohler slope of 10) were 1.4% and 0.15% respectively.
For the utter study, the author followed the work of Lobitz [57] who had found that
utter could be initiated using Adams. Lobitz had determined a utter speed for the STAR
rotor using the classical approach with a nite-element code and the Theodorsen function;
however, these results are considered proprietary. The author conducted the study for this
project using the STAR7d model in calm conditions. The rotor speed was varied from rated
to a point above rated using an induction-generator model with the rated speed set above
the nominal rated-speed. As in the Lobitz study, the simulation did not include the airfoil
moment coefcient, had blade damping at 0.05% critical, and used the Beddoes-Leishman
dynamic stall model, described the AeroDyn theory manual [48].
118
4.2.2 Results
Figures 4.1, 4.2 and 4.3 show the variation in DEL, AEP, and maximum blade 1 deec-
tion respectively for the parametric studies in Table 4.1.
-5.00%
-4.00%
-3.00%
-2.00%
-1.00%
0.00%
1.00%
2.00%
3.00%
4.00%
5.00%
70% 80% 90% 100% 110% 120% 130%
Percentage of baseline
P
e
r
c
e
n
t
a
g
e
o
f
n
o
r
m
a
l
i
z
e
d
f
l
a
p
D
E
L
f
o
r
m
=
1
0
Torsional stiffness
Tip sweep
Sweep exponent
Figure 4.1: Parametric study of damage equivalent loads
These results show an increase in ap DEL with torsional stiffness, and a decrease in
DEL with tip sweep. The sweep exponent has minimal effect on the ap DEL.
119
-5.0%
-4.0%
-3.0%
-2.0%
-1.0%
0.0%
1.0%
2.0%
3.0%
4.0%
5.0%
70% 80% 90% 100% 110% 120% 130%
Percentage from baseline
P
e
r
c
e
n
t
a
g
e
o
f
b
a
s
e
l
i
n
e
A
E
P
Sweep exponent
Torsional stiffness
Tip sweep
Figure 4.2: Parametric study of annual energy production
These results show less than 1% variation in AEP for the range of the parameters stud-
ied. The AEP increases with torsional stiffness and sweep exponent, but decreases with tip
sweep.
These results show an increase in blade deection with torsional stiffness. The results
for tip sweep and sweep exponent are not conclusive.
For the utter study, the CurveFAST model showed a slow exponential rise in blade
tip twist from the start of the simulation. The same model run in Adams did not exhibit
this behavior. The CurveFAST model without aerodynamic loads also did not exhibit this
behavior. With aerodynamic loads, the CurveFAST model exhibited stable behavior with
the 0.2% blade damping, which is still below typical values for blade structural damping
[6]. The models did not indicate the utter speed indicated by Lobitz.
120
-5%
0%
5%
10%
15%
20%
70% 80% 90% 100% 110% 120% 130%
Percentage from baseline
P
e
r
c
e
n
t
a
g
e
o
f
b
a
s
e
l
i
n
e
m
a
x
i
m
u
m
d
e
f
l
e
c
t
i
o
n
Sweep exponent
Torsional stiffness
Tip sweep
Figure 4.3: Parametric study of blade tip deection
121
4.2.3 Conclusions and Discussion
The parametric study results show the importance of low torsional stiffness to reduce
fatigue loads (DEL) and blade deection. Both Liebst [9] and Zuteck [8] discuss the im-
portance of lowering the torsional stiffness to achieve load relief.
Higher tip sweep reduces fatigue loads (Fig. 4.1); however, the results do not show
whether or not blade sweep decreases tip deection. For future study, the IEC extreme
load cases in addition to the turbulent operating cases should be run to study changes in
maximum blade deection with tip sweep.
Modifying the sweep curve exponent has less impact on loads. Note that the sweep
remains in the outboard blade portion, which Zuteck [8] notes as important. He shows that
a circular curve extending inward to the root is less effective than more outboard curvature.
For all of studies, the change in annual energy production is less than 1% (Fig. 4.2).
This value is less than the expected accuracy for power performance tests. Therefore,
changing the parameters within the range of this study would probably have no noticeable
effect on the annual energy production.
The utter study was not successful in showing a utter speed that was calculated by
a classical approach. Researchers should investigate the analytical approach to utter by
Lobitz [57], who used NASTRAN to calculate the utter speed; another approach would
be to adapt the authors CurveFEM code for these calculations.
4.3 Scaling to Larger rotors
4.3.1 Method
For the scaling studies, the author used the baseline WindPACT models [58] of 1.5
MW and 3.0 MW size. The 1.5 MW is included in the FAST code distribution [16]. Craig
Hansen, who was involved in the WindPACT studies, provided the 3.0 MW model data.
122
The airfoil tables for these models do not include the sectional moment coefcient, so
the author disabled the rst torsion mode degree-of-freedom for the study. The author
ran the baseline models in the same manner as the parametric studies to obtain annual
energy production and damage equivalent loads. Table 4.2 shows the WindPACT model
parameters.
Table 4.2: WindPACT model wind turbine parameters
Model WP1500 WP3000
Rated power, kW 1500 3000
Radius, m 35.5 49.5
Maximum chord, m 2.8 3.96
Rated generator speed, rpm 1800 1800
Rated rotor speed, rpm 20.463 14.469
Hub radius, m 1.65 2.325
Hub height, m 84 119
Gearbox efciency, % 100 100
Generator efciency, % 95 95
Gearbox ratio 87.965 124.4
123
Table 4.3 shows the blade properties for the WindPACT 1.5 MW turbine.
Table 4.3: WP1500 blade properties
Station Blade Twist Chord Pitch Aero Mass
Fraction (deg) (m) Axis Center (kg/m)
Ratio
1 0.0000 11.1 1.925 0.500 0.50 1447.61
2 0.0211 11.1 1.890 0.500 0.50 173.89
3 0.2105 11.1 2.800 0.340 0.25 204.04
4 0.4737 3.1 2.147 0.310 0.25 157.61
5 0.7368 0.6 1.494 0.280 0.25 72.66
6 1.0000 0 0.906 0.250 0.25 11.35
Station Flatwise Edgewise Torsional AE Airfoil
Stiffness Stiffness Stiffness Product Filename
(N m
2
) (N m
2
) (N m
2
) (N)
1 7.6815E+09 7.6815E+09 2.6552E+09 1.7153E+10 cylinder
2 1.1281E+09 1.1281E+09 3.9418E+08 2.5464E+09 cylinder
3 3.0477E+08 6.4782E+08 1.9215E+07 2.7043E+09 s818 2703.dat
4 8.5919E+07 2.7108E+08 8.4613E+06 2.0742E+09 s818 2703.dat
5 1.3668E+07 7.0329E+07 1.6868E+06 9.2581E+08 s825 2103.dat
6 2.3129E+05 7.8741E+06 1.7943E+05 1.1847E+08 s826 1603.dat
124
Table 4.4 shows the blade properties for the WindPACT 3.0 MW turbine.
Table 4.4: WP3000 blade properties
Station Blade Twist Chord Pitch Aero Mass
Fraction (deg) (m) Axis Center (kg/m)
Ratio
1 0.0000 11.1 2.673 0.500 0.50 2514.27
2 0.0211 11.1 2.673 0.500 0.50 342.34
3 0.2105 11.1 3.960 0.340 0.25 373.58
4 0.4737 3.1 3.036 0.310 0.25 302.91
5 0.7368 0.6 2.113 0.280 0.25 136.39
6 1.0000 0 1.281 0.250 0.25 16.68
Station Flatwise Edgewise Torsional AE Airfoil
Stiffness Stiffness Stiffness Product Filename
(N m
2
) (N m
2
) (N m
2
) (N)
1 2.5916E+10 2.5916E+10 8.9665E+09 2.8944E+10 cylinder
2 4.5189E+09 4.5189E+09 1.5791E+09 5.0974E+09 cylinder
3 1.3320E+09 3.4791E+09 6.2408E+07 4.9586E+09 s818 2703.dat
4 3.4479E+08 9.6503E+08 2.5625E+07 4.0027E+09 s818 2703.dat
5 5.4736E+07 2.4072E+08 5.1116E+06 1.7456E+09 s825 2103.dat
6 6.9195E+05 2.3690E+07 5.4377E+05 1.7586E+08 s826 1603.dat
125
The author scaled the baseline turbines for swept blades by increasing the rotor swept
area by 25% percent. This was an approximate method for obtaining an increase of 5%
AEP. The new models maximum chord was the same as the baseline, which was at the
25% radial position. To optimize the new blade planform, the author used the Betz method
outlined in Gasch and Twele [5]. For this method, the author scaled the WindPACT opti-
mal tip-speed ratio (7) by the ratio of the new rotor radius to the baseline rotor radius. The
new blade used the maximum chord of the baseline blade at 25% radial station, and the
Betz-optimized chord at the 75% radial station. The chord distribution was linearly tapered
from these two positions, with the tip chord not falling under 300 mm. The twist distribu-
tion was from the Betz-optimized twist distribution based on the maximum L/D value and
corresponding angle of attack for the 75% radial position airfoil. The maximum twist was
11.1 degrees, which was the WindPACT value.
The new blades stiffness properties were the same as the baseline at the same blade
station normalized by the blade length. This was approximately the same situation with the
STAR7d and its baseline blade. The only exception was the torsional stiffness, which the
author varied in the design process. The author adjusted the blade lineal density by a com-
mon factor so that the static moment was the same for the baseline and the extended blade.
The STAR program was able to accomplish this matching of the blade static moments.
The author ran simulations of the new rotors with the WT Perf performance code [59]
to determine the optimal pitch, tip speed ratio, and power coefcient. The author then used
these settings to determine the variable speed control parameters for CurveFAST. For the
pitch control, the author used WindPACT pitch control model and ran simulations with
stepped wind speeds above rated. The author adjusted the proportional gain as necessary
to ensure stable behavior.
The author ran normal turbulence model simulations as for the baseline machines, with
annual energy production and damage equivalent loads computed in post-processing. The
author compared the results to the baseline and adjusted turbine design parameters in order
126
to obtain a 5% AEP increase with no increase in ap bending DELs.
4.3.2 Results
Table 4.5 lists the resulting designs, performance, and loads for the two scaled rotor
models, Curve1500 and Curve3000.
Table 4.5: Scaled model results
Model Curve1500 Curve3000
Baseline WindPACT radius, m 35.5 49.5
Stretched radius, m 39.15 55.35
Maximum chord, m 2.8 3.96
Tip sweep, m 2.8 2.376
GJ modication 74% lower same as baseline
Percent AEP over baseline 5.0% 5.1%
Percent Flap DEL under 6.6% 10.2%
baseline (m = 10)
Percent Edge DEL over 3.7% 1.7%
baseline (m = 10)
The table shows that the AEP has increased 5% or higher for the swept-blade designs,
with a decrease in the ap-bending DEL. The larger turbine required less tip sweep and no
reduction in torsional stiffness to achieve the design goal. Both designs show an increase
in edge DEL.
127
Figure 4.4 below shows the ap DEL versus wind speed for the Curve1500 model and
a fatigue-curve slope m = 10. The author has normalized the DEL at an arbitrary value
slightly above the maximum value.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 5 10 15 20 25 30
10-minute average wind speed
N
o
r
m
a
l
i
z
e
d
b
l
a
d
e
1
f
l
a
p
b
e
n
d
i
n
g
D
E
L
WindPACT 1.5
Curve1500
Figure 4.4: Flap bending comparison for scaled design
The results in the gure show for the swept rotor that the ap DEL is higher below rated
wind speed but falls below the straight rotor above rated.
128
4.3.3 Conclusions and Recommendations
The new swept designs lower ap bending loads compared to the baseline straight
blades and provide increased energy capture. The swept-blade concept does not seem to
suffer at larger scales, with the required torsional exibility and tip sweep being lower in
comparison to smaller scale. However, the swept blade concept does not offer relief for the
edge bending loads, which will become critical for larger turbines.
Given the tentative success of this project and the successful operation of the test tur-
bine, the author recommends that this concept should be explored further for commercial
application. The author recommends continuing the design studies to a more detailed effort
which would include detailed section properties, and running of the full IEC extreme load
cases to ensure that baseline loads are not exceeded.
129
Bibliography
[1] Leung, A. Y. T. and Fung, T. C., Spinning Finite Elements, Journal of Sound and
Vibration, Vol. 125, No. 3, 1988, pp. 523537.
[2] Kosmatka, J. B., Structural Dynamic Modeling of Advanced Composite Propellers by
the Finite Element Method, Ph.D. thesis, University of California Los Angeles, 1986.
[3] Harrison, L., Prospects and Pundits, Windpower Monthly, April 2004, pp. 6.
[4] Anon., Net System Power: A Small Share of Californias Power Mix in 2005, CEC-
300-2006-009-F, California Energy Commission, April 2006.
[5] Gasch, R. and Twele, J., Wind Power Plants, Solarpraxis, Berlin, 2002.
[6] Burton, T., Sharpe, D., Jenkins, N., and Bossanyi, E. A., Wind Energy Handbook, J.
Wiley, Chichester, New York, 2001.
[7] Lobitz, D. W., Veers, P. S., and Migliore, P. G., Enhanced Performance of HAWTs
Using Adaptive Blades, Wind Energy 96, Energy Week Conference and Exhibition
Book VIII Volume 1, PennWell Conferences and Exhibitions, Houston, TX, 1996, pp.
4145.
[8] Zuteck, M., Adaptive Blade Concept Assessment: Curved Planform Induced Twist
Investigation, SAND2002-2996, Sandia National Laboratories, October 2002.
[9] Liebst, B. S., Wind Turbine Gust Load Alleviation Utilizing Curved Blades, Journal
of Propulsion, Vol. 2, No. 4, 1986, pp. 371377.
[10] Larwood, S. and Zuteck, M., Swept Wind Turbine Blade Aeroelastic Modeling for
Loads and Dynamic Behavior, AWEA Windpower 2006, American Wind Energy As-
sociation, Pittsburgh, PA, 2006, pp. 117.
[11] Manwell, J. F., McGowan, J. G., and Rogers, A. L., Wind energy explained : theory,
design and application, Wiley, Chichester ; New York, 2002.
[12] Molenaar, D.-P., Cost-effective design and operation of variable speed wind turbines :
closing the gap between the control engineering and the wind engineering community,
Ph.D. thesis, Delft University Press, 2003.
[13] Anon., Wind turbines - Part 1: Design Requirements (Final Draft), IEC 61400-1
Ed. 3 (88/228/FDIS), International Electrotechnical Commission, 2005.
130
[14] Anon., Regulations for the Certication of Wind Energy Conversion Systems, Supple-
ment No. 2 to the 1993 Edition, Germanischer Lloyd, Hamburg, 1993rd ed., 1998.
[15] Buhl Jr, M. L., The NWTC Design-Codes Suite: An Overview, A Collection of the
2005 ASME Wind Energy Symposium Technical Papers at the 43rd AIAA Aerospace
Sciences Meeting and Exhibit, ASME/AIAA, Reno, NV, United States, 2005, pp.
156166.
[16] Jonkman, J. M., NWTC Design Codes (FAST), Available at:
http://wind.nrel.gov/designcodes/simulators/FAST/, Last update 5 July 2005.
[17] Laino, D. J., NWTC Design Codes (AeroDyn), Available at:
http://wind.nrel.gov/designcodes/simulators/aerodyn/, Last update 5 July 2005.
[18] Kelley, N. and Jonkman, B., NWTC Design Codes (TurbSim), Available at:
http://wind.nrel.gov/designcodes/preprocessors/turbsim/, Last update 13 April 2006.
[19] Laino, D. J., NWTC Design Codes (IECWind), Available at:
http://wind.nrel.gov/designcodes/preprocessors/iecwind/, Last update 01 September
2005.
[20] Wright, A. D. and Elliott, A. S., Development of a Wind Turbine System Dynamics
Model Using the Automatic Dynamic Analysis of Mechanical Systems (ADAMS)
Software, Windpower 92, American Wind Energy Association, Seattle, WA, 1992,
pp. 329338.
[21] Hansen, A., ADAMS Overview, Wind Turbine Design Workshop, National Renew-
able Energy Laboratory, Boulder, Colorado, 2005, pp. 110.
[22] Anon., Design Codes FAST and ADAMS for Load Calculations of Onshore Wind
Turbines, Tech. rep., Germanischer Lloyd WindEnergie GmbH, 26 May 2005.
[23] Wilson, R., Freeman, L., Walker, S., and Harman, C., FAST Advanced Dynamics
Code, OSU/NREL 96-01, Oregon State University, Department of Mechanical En-
gineering, March 1996.
[24] Jonkman, J. M. and Buhl Jr, M. L., New developments for the NWTCs fast aeroe-
lastic HAWT simulator, 2004 ASME Wind Energy Symposium, AIAA/ASME, Reno,
NV, 2004, pp. 181191.
[25] Kane, T. R. and Levinson, D. A., Dynamics: Theory and Applications, McGraw-Hill,
New York, 1985.
[26] Jonkman, J. M., Modeling of the UAE Wind Turbine for Renement of FAST AD,
NREL/TP-500-34755, National Renewable Energy Laboratory, December 2003.
[27] Lobitz, D. W., Ramications of aeroelastic analysis approximations as blade designs
approach stability boundaries, 2004 ASME Wind Energy Symposium, American So-
ciety of Mechanical Engineers, Reno, NV, 2004, pp. 203210.
131
[28] Rao, S. S., The Finite Element Method in Engineering, Elsevier Butterworth-
Heinemann, Burlington, MA, 4th ed., 2005.
[29] Buhl Jr, M. L., NWTC Design Codes (Modes), 26 May 2005.
[30] Bielawa, R. L., Johnson, S. A., Chi, R. M., and Gangwani, S. T., Aeroelastic Analysis
for Propellers, NASA CR-3729, National Aeronautics and Space Administration,
December 1983.
[31] Kosmatka, J. B. and Friedmann, P., Vibration Analysis of Composite Turbopro-
pellers Using a Nonlinear Beam-Type Finite-Element Approach, AIAA Journal,
Vol. 27, No. 11, 1989, pp. 16061614.
[32] Bauchau, O. A. and Hong, C. H., Large Displacement Analysis of Naturally Curved
and Twisted Composite Beams, AIAA Journal, Vol. 25, No. 11, 1987, pp. 14691475.
[33] Bir, G. and Chopra, I., Aeromechanical Stability of Rotorcraft with Advanced Geom-
etry Blades, Mathmatical Computer Modeling, Vol. 19, No. 3/4, 1994, pp. 159191.
[34] Celi, R. and Friedmann, P. P., Aeroelastic Modeling of Swept Tip Rotor Blades Using
Finite Elements, Journal of the American Helicopter Society, Vol. 33, No. 2, 1988,
pp. 4352.
[35] Hodges, D. H., Shang, X., and Cesnik, C. E. S., Finite Element Solution of Non-
linear Intrinsic Equations for Curved Composite Beams, Journal of the American
Helicopter Society, Vol. 41, No. 4, 1996, pp. 313321.
[36] Hodges, D. H., Geometrically Exact, Intrinsic Theory for Dynamics of Curved and
Twisted Anisotropic Beams, AIAA Journal, Vol. 41, No. 6, 2003, pp. 11311137.
[37] Rosen, A. and Rand, O., Static Aeroelastic Behavior of Curved Helicopter Blades in
Hovering and Axial Flight, Vertica, Vol. 7, No. 3, 1983, pp. 241257.
[38] Rosen, A. and Rand, O., Numerical Model of the Nonlinear Behavior of Curved
Rods, Computers & Structures, Vol. 22, No. 5, 1986, pp. 785799.
[39] Rosen, A. and Rand, O., A Model of a Curved Helicopter Blade in Forward Flight,
Vertica, Vol. 11, No. 1/2, 1987, pp. 211231.
[40] Tarzanin Jr., F. J. and Vlaminck, R. R., Investigation of the effect of blade sweep on
rotor vibratory loads, NASA CR-166526, October 1983.
[41] Meirovitch, L., Principles and Techniques of Vibrations, Prentice Hall, Upper Saddle
River, NJ, 1997.
[42] Meirovitch, L., A New Method of Solution of the Eigenvalue Problem for Gyro-
scopic Systems, AIAA Journal, Vol. 12, No. 10, 1974, pp. 13371342.
132
[43] Banerjee, J. R., Su, H., and Jackson, D. R., Free vibration of rotating tapered beams
using the dynamic stiffness method, Journal of Sound and Vibration, Vol. 298, 2006,
pp. 10341054.
[44] Young, W. C. and Budynas, R. G., Roarks Formulas for Stress and Strain, McGraw-
Hill, New York, seventh ed., 2002.
[45] Blevins, R. D., Formulas for Natural Frequency and Mode Shape, Van Nostrand Rein-
hold Company, New York, 1979.
[46] Mabie, H. H. and Rogers, C. B., Transverse Vibrations of Double-Tapered Cantilever
Beams, The Journal of the Acoustical Society of America, Vol. 51, No. 5 (Part 2),
1972, pp. 17711774.
[47] Wang, J. and Mahrenholtz, O., Bending Frequencies of a Rotating Curved Beam,
Ingeniur-Archiv, Vol. 44, No. 6, 1975, pp. 399407.
[48] Moriarty, P. J. and Hansen, A. C., AeroDyn Theory Manual, NREL/EL-500-36881,
National Renewable Energy Laboratory, December 2005.
[49] Hoerner, S. F., Fluid-Dynamic Lift, Published by Liselotte A. Hoerner, 1985.
[50] Kocurek, J. D., Lifting Surface Performance Analysis for Horizontal Axis Wind Tur-
bines, NREL Report No. STR-217-3163, National Renewable Energy Laboratory,
1987.
[51] van Dam, C. P., Saephan, S., and Chao, D., Aerodynamic Design and Analysis of
Innovative Wind Turbine Blades, AWEA Windpower 2006, American Wind Energy
Association, Pittsburgh, PA, 2006.
[52] Anonymous, Wind turbines generator systems- Part 13: Measurement of mechanical
loads, IEC TS 61400-13 First edition, International Electrotechnical Commission,
2001.
[53] Buhl Jr, M. L., Wright, A. D., and Pierce, K., FAST AD Code Verication: A Com-
parison to ADAMS, 2001 ASME Wind Energy Symposium, American Institute of
Aeronautics and Astronautics, Inc. and American Society of Mechanical Engineers,
Reno, NV, 2001, pp. 368373.
[54] Freebury, G. and Musial, W., Determining Equivalent Damage Loading for Full-
Scale Wind Turbine Blade Fatigue Tests, 38th AIAA Aerospace Sciences Meeting and
Exhibit, American Institute of Aeronautics and Astronautics, Reno, Nevada, 2000.
[55] Sutherland, H. J., On the Fatigue Analysis of Wind Turbines, SAND99-0089, San-
dia National Laboratories, June 1999.
[56] Buhl Jr, M. L., NWTC Design Codes (Crunch), 01 April 2008.
[57] Lobitz, D. W., Aeroelastic Stability Predictions for a MW-sized Blade, Wind En-
ergy, Vol. 7, 2004, pp. 211224.
133
[58] Grifn, D. A., WindPACT Turbine Design Scaling Studies Technical Area 1.Com-
posite Blades for 80- to 120-Meter Rotor, NREL/SR-500-29492, National Renew-
able Energy Laboratory, April 2001.
[59] Buhl Jr, M. L., NWTC Design Codes (WT Perf), 26 May 2005.
134
Appendix A
Blade Finite Element Matrices
This appendix lists the equations for the elements of the blade nite element matrices.
The elements are from the upper triangle. Elements from the upper triangle that are not
listed assume the value zero. The matrices are symmetric or skew-symmetric as noted.
A.1 Mass Matrix
The element mass matrix [M] is symmetric, with the elements:
M
1,1
= l/35(10
0
+ 3
l
)
M
1,5
= l
2
/420(15
0
+ 7
l
)
M
1,7
= 9l/140(
0
+
l
)
M
1,11
= l
2
/420(7
0
+ 6
l
)
M
2,2
= l/35(10
0
+ 3
l
)
M
2,4
= l
2
/420(15
0
+ 7
l
)
M
2,8
= 9l/140(
0
+
l
)
135
M
2,10
= l
2
/420(7
0
+ 6
l
)
M
3,3
= l/12(3
0
+
l
)
M
3,9
= l/12(
0
+
l
)
M
4,4
= l
3
/840(5
0
+ 3
l
)
M
4,8
= l
2
/420(6
0
+ 7
l
)
M
4,10
= l
3
/280(
0
+
l
)
M
5,5
= l
3
/840(5
0
+ 3
l
)
M
5,7
= l
2
/420(6
0
+ 7
l
)
M
5,11
= l
3
/280(
0
+
l
)
M
6,6
= l/12(3J
0
+ J
l
)
M
6,12
= l/12(J
0
+ J
l
)
M
7,7
= l/35(3
0
+ 10
l
)
M
7,11
= l
2
/420(7
0
+ 15
l
)
M
8,8
= l/35(3
0
+ 10
l
)
M
8,10
= l
2
/420(7
0
+ 15
l
)
M
9,9
= l/4(
0
/3 +
l
)
M
10,10
= l
3
/840(3
0
+ 5
l
)
M
11,11
= l
3
/840(3
0
+ 5
l
)
136
A.2 Elastic Stiffness Matrix
The element elastic stiffness matrix [K
e
] is symmetric, with the elements:
K
e1,1
= 6/l
3
(EI
y0
+ EI
yl
)
K
e1,5
= 2/l
2
(2EI
y0
+ EI
yl
)
K
e1,7
= 6/l
3
(EI
y0
+ EI
yl
)
K
e1,11
= 2/l
2
(EI
y0
+ 2EI
yl
)
K
e2,2
= 6/l
3
(EI
x0
+ EI
xl
)
K
e2,4
= 2/l
2
(2EI
x0
+ EI
xl
)
K
e2,8
= 6/l
3
(EI
x0
+ EI
xl
)
K
e2,10
= 2/l
2
(EI
x0
+ 2EI
xl
)
K
e3,3
= (EA
0
+ EA
l
)/2l
K
e3,9
= (EA
0
+ EA
l
)/2l
K
e4,4
= 1/l(3EI
x0
+ EI
xl
)
K
e4,8
= 2/l
2
(2EI
x0
+ EI
xl
)
K
e4,10
= 1/l(EI
x0
+ EI
xl
)
K
e5,5
= 1/l(3EI
y0
+ EI
yl
)
K
e5,7
= 2/l
2
(2EI
y0
+ EI
yl
)
K
e5,11
= 1/l(EI
y0
+ EI
yl
)
137
K
e6,6
= (GJ
0
+ GJ
l
)/2l
K
e6,12
= (GJ
0
+ GJ
l
)/2l
K
e7,7
= 6/l
3
(EI
y0
+ EI
yl
)
K
e7,11
= 2/l
2
(EI
y0
+ 2EI
yl
)
K
e8,8
= 6/l
3
(EI
x0
+ EI
xl
)
K
e8,10
= 2/l
2
(EI
x0
+ 2EI
xl
)
K
e9,9
= (EA
0
+ EA
l
)/2l
K
e10,10
= 1/l(EI
x0
+ 3EI
xl
)
K
e11,11
= 1/l(EI
y0
+ 3EI
yl
)
K
e12,12
= (GJ
0
+ GJ
l
)/2l
A.3 Gyroscopic Matrix
The element gyroscopic matrix [G] is skew-symmetric, with the elements:
G
1,2
= (156b
1
)
0
l/420 + (72b
1
)(
l
0
)l/840
G
1,3
= (147b
2
)
0
l/420 + (70b
2
)(
l
0
)l/840
G
1,4
= (22lb
1
)
0
l/420 + (14lb
1
)(
l
0
)l/840
G
1,8
= (54b
1
)
0
l/420 + (54b
1
)(
l
0
)l/840
G
1,9
= (63b
2
)
0
l/420 + (56b
2
)(
l
0
)l/840
G
1,10
= (13lb
1
)
0
l/420 + (12lb
1
)(
l
0
)l/840
138
G
2,3
= (147b
3
)
0
l/420 + (70b
3
)(
l
0
)l/840
G
2,5
= (22lb
1
)
0
l/420 + (14lb
1
)(
l
0
)l/840
G
2,7
= (54b
1
)
0
l/420 + (54b
1
)(
l
0
)l/840
G
2,9
= (63b
3
)
0
l/420 + (56b
3
)(
l
0
)l/840
G
2,11
= (13lb
1
)
0
l/420 + (12lb
1
)(
l
0
)l/840
G
3,4
= (21lb
3
)
0
l/420 + (14lb
3
)(
l
0
)l/840
G
3,5
= (21lb
2
)
0
l/420 + (14lb
2
)(
l
0
)l/840
G
3,7
= (63b
2
)
0
l/420 + (70b
2
)(
l
0
)l/840
G
3,8
= (63b
3
)
0
l/420 + (70b
3
)(
l
0
)l/840
G
3,10
= (14lb
3
)
0
l/420 + (14lb
3
)(
l
0
)l/840
G
3,11
= (14lb
2
)
0
l/420 + (14lb
2
)(
l
0
)l/840
G
4,5
= (4l
2
b
1
)
0
l/420 + (3l
2
b
1
)(
l
0
)l/840
G
4,7
= (13lb
1
)
0
l/420 + (14lb
1
)(
l
0
)l/840
G
4,9
= (14lb
3
)
0
l/420 + (14lb
3
)(
l
0
)l/840
G
4,11
= (3l
2
b
1
)
0
l/420 + (3l
2
b
1
)(
l
0
)l/840
G
5,8
= (13lb
1
)
0
l/420 + (14lb
1
)(
l
0
)l/840
G
5,9
= (14lb
2
)
0
l/420 + (14lb
2
)(
l
0
)l/840
G
5,10
= (3l
2
b
1
)
0
l/420 + (3l
2
b
1
)(
l
0
)l/840
G
7,8
= (156b
1
)
0
l/420 + (240b
1
)(
l
0
)l/840
139
G
7,9
= (147b
2
)
0
l/420 + (224b
2
)(
l
0
)l/840
G
7,10
= (22lb
1
)
0
l/420 + (30lb
1
)(
l
0
)l/840
G
8,9
= (147b
3
)
0
l/420 + (224b
3
)(
l
0
)l/840
G
8,11
= (22lb
1
)
0
l/420 + (30lb
1
)(
l
0
)l/840
G
9,10
= (21lb
3
)
0
l/420 + (22lb
3
)(
l
0
)l/840
G
9,11
= (21lb
2
)
0
l/420 + (22lb
2
)(
l
0
)l/840
G
10,11
= (4l
2
b
1
)
0
l/420 + (5l
2
b
1
)(
l
0
)l/840
A.4 Spin Stiffness Matrix
The element spin-stiffness matrix [K