Professional Documents
Culture Documents
CU Scholar
Electrical, Computer & Energy Engineering
Electrical, Computer & Energy Engineering
Graduate Theses & Dissertations
Spring 4-1-2016
Recommended Citation
Dunne, Fiona, "Optimizing Blade Pitch Control of Wind Turbines with Preview Measurements of the Wind" (2016). Electrical,
Computer & Energy Engineering Graduate Theses & Dissertations. 125.
https://scholar.colorado.edu/ecen_gradetds/125
This Dissertation is brought to you for free and open access by Electrical, Computer & Energy Engineering at CU Scholar. It has been accepted for
inclusion in Electrical, Computer & Energy Engineering Graduate Theses & Dissertations by an authorized administrator of CU Scholar. For more
information, please contact cuscholaradmin@colorado.edu.
Optimizing Blade Pitch Control of Wind Turbines
with Preview Measurements of the Wind
by
F. Dunne
Doctor of Philosophy
2016
This thesis entitled:
Optimizing Blade Pitch Control of Wind Turbines
with Preview Measurements of the Wind
written by F. Dunne
has been approved for the Department of Electrical, Computer, and Energy Engineering
Date
The final copy of this thesis has been examined by the signatories, and we find that both the
content and the form meet acceptable presentation standards of scholarly work in the above
mentioned discipline.
Dunne, F. (Ph.D., Electrical Engineering)
In above-rated wind speeds, the goal of a wind turbine blade pitch controller is to regulate
rotor speed while minimizing structural loads and pitch actuation. This controller is typically
feedback-only, relying on a generator speed measurement, and sometimes strain gauges and ac-
celerometers. A preview measurement of the incoming wind speed (from a turbine-mounted lidar,
for example) allows the addition of feedforward control, which enables improved performance com-
pared to feedback-only control. The performance improvement depends both on the amount of
preview time available in the wind speed measurement as well as the coherence (correlation as a
function of frequency) between the wind measurement and the wind that is experienced by the
turbine.
This thesis shows how to design an optimal collective-pitch controller that takes both preview
time and measurement coherence into account. Simulation results show significantly reduced pitch
actuation, improved generator speed regulation, and reduced structural loads compared to several
different baseline cases. In addition, linear-model-based results show how the benefit of preview
Effective lidar-based control also requires knowledge of the expected arrival time of the mea-
sured wind. Arrival time is the time it takes for the wind to travel from the measurement focus
location to the turbine rotor. Arrival time is often assumed to be equal to the distance traveled
divided by the average wind speed. This thesis, using field test data, studies deviations from this
assumption.
Control implementation across the full range of above-rated wind speeds can be achieved
through gain scheduling. The effect of gain scheduling implementation on effective feedforward
iv
and feedback gains is not straightforward. This thesis provides a detailed explanation of the ef-
fective gains resulting from two different gain-scheduling implementations. It includes an analysis
of a simplified version of a nonlinear gain scheduling feedback loop as well as verification through
Additional topics covered in this thesis include a model-inverse-based analysis of the condi-
tions under which lidar is beneficial, a breakdown of the maximum useful amount of preview time
by its different uses, and a comparison of two lidar-based individual pitch controllers.
Dedication
Acknowledgements
I would like to thank my advisor Prof. Lucy Pao for her tireless help, encouragement, and
kindness and her thorough, insightful feedback; Dr. Alan Wright for his crucial advising and sup-
port; the rest of my dissertation committee Prof. Behrouz Touri, Prof. Dale Lawrence, Prof. Eric
Frew, and Prof. Jason Marden for their generous time and feedback; Dr. David Schlipf, Dr. Eric
Simley, and Dr. Jason Laks for their especially significant suggestions and discussions in guiding
the direction of this research; the many anonymous peer reviewers who gave very helpful feedback
on parts of this work; a fantastic group of labmates and visiting researchers including Dr. Hua
Zhong, Marian Chaffe, Dr. Jeff Butterworth, Dr. Jason Laks, Dr. Eric Simley, Dr. Shervin Shajiee,
Jake Aho, Andrew Buckspan, Dan Zalkind, Arnold Braker, Dr. David Schlipf, Floris Teeuwisse,
and Bart Doekemeijer for both their friendship and their help and feedback on this research;
Prof. Kathryn Johnson, Dr. Na Wang, Andy Scholbrock, Dr. Pieter Gebraad, Bonnie Jonkman,
Neil Kelley, Dr. Rod Frehlich, Dr. Jason Jonkman, Dr. Paul Fleming, Jen Annoni, and many others
for valuable advice and discussions at NREL meetings; all the friends, Ultimate players, and hiking
and climbing partners who were a part of great fun and adventures outside of lab; and finally my
husband Daniel Garrett for always being there to support and encourage me.
vii
Contents
Chapter
1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Contributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
1.5 Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2 Model-Inverse-Based Analysis 19
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
viii
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.4 Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
4.1.1 Actuator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
4.1.3 Turbine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.2.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
ix
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86
Bibliography 114
Appendix
A Proof of Effect of Filters K, L, and H on Modeled Coherence and Power Spectra 119
B Analysis of Gain-Scheduling Implementation for the NREL 5-MW Turbine Blade Pitch
Controller 123
D Gain Scheduling Implementation Details of the NREL 5-MW and Other Controllers 141
xi
Tables
Table
4.3 Mean values of td (t) and Tv : either uncorrected or corrected for induction zone . . . 79
Figures
Figure
1.1 Wind coordinate system and wind turbine actuator axes of rotation . . . . . . . . . 3
1.4 Coherence bandwidth vs. number of lidar measurement samples per circle . . . . . . 11
2.1 Block diagram showing linearized models of turbine, feedback, and feedforward . . . 20
3.1 PSDs without and with ideal feedforward to eliminate generator speed error . . . . . 28
3.10 Overall frequency response of PI plus H2 feedback controllers, along with PI alone. . 44
3.14 Steady-state pitch angle versus steady-state wind speed, and linearization. . . . . . . 50
3.18 Linear-model expectations vs. simulation results: gen speed error and pitch rate PSDs 54
3.19 Linear-model expectations vs. simulation results for RMS generator speed error . . . 55
3.20 Linear-model expectations vs. simulation results for RMS pitch rate . . . . . . . . . 56
3.21 RMS gen speed error for a range of preview times and coherence bandwidths . . . . 59
3.22 RMS pitch rate for a range of preview times and coherence bandwidths . . . . . . . 59
5.5 PSDs for method BBB under its original simulation conditions. . . . . . . . . . . . . 96
5.6 Simulated turbine loads using individual pitch controllers, 13 m/s Great Plains wind 97
5.7 Simulated turbine loads using individual pitch controllers, 14 m/s Class A wind . . . 98
5.8 Bode plots of feedforward controllers A and B, excluding scheduled gains . . . . . . 100
B.2 Nonlinear gain-scheduling loop. A simplified part of Fig. B.1, where d = 0.109997/2. 125
B.3 Steady-state pitch angle βo as a function of steady-state wind speed. . . . . . . . . 125
B.5 Effective additional gain factor caused by the Integrate-First implementation . . . . 127
B.8 Blade pitch time series, simulation under different implementations . . . . . . . . . . 131
B.9 Generator speed time series, simulation under different implementations . . . . . . . 132
B.12 ‘Effective closed-loop gain’ with new f (u) divided by ‘desired closed-loop gain’ . . . 135
B.13 Blade pitch PSD, simulation with new f (u) implementation included . . . . . . . . . 135
B.14 Generator speed PSD, simulation with new f (u) implementation included . . . . . . 136
Chapter 1
Introduction
generator speed sensor, and sometimes strain gauges and accelerometers. Recently, however, it
has become feasible to obtain measurements of the speed of the wind approaching a turbine by
using a turbine-mounted Doppler lidar or other technology that remotely measures wind speed.
This provides an estimate of the wind speed that will arrive at the turbine with a few seconds of
preview. The topic of this thesis is incorporating this feedforward signal into the wind turbine’s
blade pitch control system in order to better counteract the effects of wind disturbances. The
specific focus is improved performance in the above-rated wind speed operating region, where the
control goals are rotor speed regulation, structural load reduction, and minimal pitch actuation.
1.1 Motivation
The impact of this research depends on the percent reduction in the cost of wind energy
due to improved control performance using lidar, the cost of a lidar relative to the cost of a wind
turbine, and the size of the wind energy industry. This section provides some brief background on
these factors.
A turbine-mounted lidar can scan the incoming wind field and provide preview wind speed
measurements a few seconds in advance [1]. This allows anticipatory blade pitch adjustments
which can reduce structural loads compared to feedback alone. Structural load reduction leads
to reduced cost of energy through either reduced material costs or increased turbine lifetime. A
2
small reduction in fatigue load often leads to a large increase in material lifetime. This is especially
true for composite blade materials, where with an S-N (Wöhler) curve exponent of 10 [2], a 7%
decrease in fatigue load leads to a doubling of number of cycles to failure. Therefore a relatively
small reduction in structural load can have a relatively large impact on cost of energy.
Coherent Doppler lidar has seen recent large improvements in cost, compactness, and reliabil-
ity because of fiber-optic technology and related components developed for use in the telecommu-
nications industry [3]. In addition, as wind turbines are becoming larger, a lidar becomes a smaller
percentage of the cost of a turbine. In 2013, the average newly installed US wind turbine was rated
at 1.87 MW, with a capacity-weighted average installed cost of $1,630/kW [4]. A Doppler lidar, at
a cost of roughly $170,000, is then currently about 6% of the installed cost of today’s average $3
million turbine.
Finally, the potential impact of this work scales with the growth of the wind energy industry.
The wind energy industry is growing rapidly, with wind power being the largest source of new
United States electricity in 2014 [5]. Wind power currently generates 4% of U.S. electricity, and
the U.S. Department of Energy has presented a goal of meeting 20% of U.S. electricity needs
with wind by 2030. Globally, wind power is projected to deliver 5.3% of the electricity consumed
1.2 Background
This section contains some background information necessary to understanding this thesis.
Figure 1.1: Blue lines show the perpendicular directions of the u (downwind), v (transverse), and
w (vertical) components of the wind when the turbine is yawed into the mean wind 4/43
direction. Red
lines show the axes of rotation for the three different turbine actuators.
1.2.1.1 Actuators
A modern multi-megawatt wind turbine relies on active control of the nacelle yaw, the gen-
erator torque, and the blade pitch angles. Fig. 1.1 shows the axis of rotation for each of these
Nacelle Yaw Nacelle yaw control keeps the wind turbine rotor directed into the wind.
It operates slowly in order to minimize gyroscopic forces and avoids making continuous small
corrections in order to avoid shock load cycles due to gear backlash [7]. A wind vane mounted on
the turbine measures yaw error, and the turbine is moved in the correct direction only when the
yaw error has been above some value for some period of time, often on the order of 5 degrees error
for the past 10 minutes. A yaw brake is used whenever the yaw drive is not in motion.
Generator Torque Generator torque actuation is achieved using the power electronics
that connect the wind turbine generator to the utility grid. Generator torque response to torque
commands is fast enough to be considered instantaneous relative to blade pitch control. The
τa − τg = I ω̇ (1.1)
4
where τa is the aerodynamic torque applied to the rotor by the wind, τg is the torque applied by
the generator, I is the moment of inertia of the rotor, generator, and drivetrain, and ω̇ is the rotor
P = τg ω
where P is the power produced by the generator and ω is the rotor speed. For simplicity, these two
equations assume a gearbox ratio of 1 and omit drivetrain and electrical losses.
Blade Pitch Blade pitch actuation is the rotation of an entire blade about its long axis.
This changes the angle of attack between the airfoil and the wind it experiences, and therefore
changes the lift and drag coefficients which determine the aerodynamic torque τa as well as the
thrust on the rotor. A pitch angle of 0◦ is often defined as the blade pitch angle that captures
maximum power, with increasing values towards 90◦ increasingly shedding power, typically done in
the pitch-to-feather direction for a more smooth and continuous reduction in power capture than
the pitch-to-stall direction. For a 5-MW turbine, pitch actuation is possible up to a bandwidth of
about 1 Hz, and this bandwidth generally decreases with increasing turbine size. Pitch actuation
is typically achieved using one pitch motor for each blade, which can provide redundant braking
capability [8]. This also enables the possibility of individual, rather than collective, blade pitch
control, which may be used to compensate for non-uniform wind over the rotor plane, at the
expense of increased pitch actuation. This thesis focuses mainly on collective pitch control, with
Wind turbine control can be divided into the four control regions shown in Fig. 1.2. In
Region 1, below the cut-in wind speed, the power production is zero because the wind speed
is too low for operation to be worthwhile. In Region 2, the turbine is controlled to maximize
power capture, typically achieved by holding the pitch angle constant at its optimal value of 0◦ ,
and controlling generator torque based on generator speed feedback as τg = kωg2 , where k is a
5
140
Power (%)
120 Rotor Speed (rpm)
Blade Pitch (deg)
100
80
1 2 3 4
60
40
20
0
0 5 10 15 20 25 30
Wind Speed (m/s)
Figure 1.2: A sample steady-state power curve, with corresponding steady-state rotor speeds and
blade pitch angles. Wind speeds are below cut-in in Region 1, below-rated in Region 2, above-rated
in Region 3, and above cut-out in Region 4. Transition Regions 1.5 and 2.5 are omitted. This thesis
focuses on blade pitch control in Region 3 (highlighted).
constant that depends on the turbine and the air density, and ωg is the generator speed, which is
approximately equal to the rotor speed ω times the gearbox ratio. This nonlinear torque feedback
law regulates the rotor speed to the aerodynamic optimal value for the current wind speed. In
Region 3, above the rated wind speed, the turbine is controlled to regulate rotor speed and power
to their rated values primarily by using blade pitch control. Pitch control in Region 3 is often gain-
scheduled proportional-integral (PI) control of generator speed error. Region 3 generator torque
is either held constant at its rated value or set to be inversely proportional to generator speed,
maintaining constant power. In Region 4, above the cut-out wind speed, the turbine is shut down
because the wind speed is too high. Not shown in Fig. 1.2 are Regions 1.5 and 2.5, which allow
The National Renewable Energy Laboratory (NREL) 5-MW turbine model [9] is a publicly-
available and very commonly used turbine model for control systems and other turbine design
research. This thesis mainly uses the NREL 5-MW turbine model. In addition, this thesis dis-
cusses lidar field test data taken from the NREL two-bladed Controls Advanced Research Turbine
6
(CART2), which is a smaller 600-kW wind turbine used for controls field testing at NREL.
Coherent Doppler lidar remotely measures wind speed by emitting a laser, often infrared,
which is reflected by aerosols, and by measuring the Doppler shift of the return signal to determine
wind speed. The resulting wind speed measurement is a spatially-averaged line-of-sight wind speed
centered at some chosen focus distance. The line-of-sight wind speed components are spatially
averaged along the line of sight, weighted by a Gaussian (for pulsed lidar) or Lorentzian (for
continuous-wave lidar) range-weighting function with peak centered approximately at the focus
The turbine is most affected by changes in the u-component of the wind: the component of
the wind vector in line with the mean wind direction, which is pointing directly at the turbine rotor,
assuming no yaw error. Fig. 1.1 shows the directions of the perpendicular wind components u, v,
and w in blue. The effect of the u component in comparison to that of the v and w components
is shown in Fig. 1.3, where for a linear model of the NREL 5-MW turbine, the u-component of
the wind, at most frequencies, has at least ten times the effect on the turbine as the v (transverse)
reconstructed from the lidar line-of-sight wind speed, which is often not directly in line with the
u-component. The lidar is mounted on top of a turbine nacelle or in the turbine hub, and one or
multiple beams are typically directed to sample the incoming wind field at distributed locations
representative of the wind approaching the entire rotor area. For example, the lidar might scan a
circle with a half-cone angle of 15 or 30 degrees. The u-component can be reconstructed based on
the line-of-sight measurement angle and the assumption that the v and w wind components are
zero. The true v and w components contribute to u-component-estimate error, which increases as
the lidar measurement angle increases. This is just one of several factors that affect the coherence
between the rotor-effective wind speed predicted using lidar measurements and the rotor-effective
Bode Diagram
150
To: Gen speed
From: u (m/s)
error (rpm)
0
10000
To: RootMycC
(kN·m)
5000
Magnitude (abs)
0
1000
To: RootMycH To: RootMycV
(kN·m)
500
0
600
(kN·m)
400
200
0
0 0.2 0.4 0.6 0.8 1
Frequency (Hz)
Figure 1.3: Magnitudes of transfer functions from uniform u (downwind), v (transverse), and w
(vertical) wind components, as deviations from the operating point, to generator speed error and
the out-of-plane root bending moments of the three blades, as deviations from the operating point,
presented in collective (RootMycC), vertical (RootMycV), and horizontal (RootMycH) multi-blade
coordinate transformation (MBC) components, for the NREL 5-MW turbine model with base-
line feedback controller (Multiply-First Implementation, to be discussed further in Appendix B),
linearized in 13 m/s wind.
8
The rotor-effective wind speed is the single uniform wind speed that has the same effect on
the turbine as the full 3D vector field, in terms of a particular turbine parameter: usually either
The rotor-effective wind speed experienced by the turbine can be determined by using a wind
speed estimator, essentially using the entire turbine as an anemometer [12]. One relatively simple
rotor-effective wind speed estimation method is the torque-balance method, in which the first step
is to solve for the aerodynamic torque according to equation (1.1), given generator torque and rotor
acceleration. Then a lookup table based on the turbine model is used to determine the current wind
speed given the current aerodynamic torque, blade pitch angle, and rotor speed. Many improved
versions of wind speed estimators have also been used, for example, a Kalman filter that accounts
In simulation, rotor-effective wind speed can also be estimated by using a weighted average
of the u-components of the wind speeds over the rotor disc. The specific shape of the weighting
function depends on the mean wind speed and whether aerodynamic torque or thrust is of interest,
but even a simple average over the rotor disc is sometimes used as a sufficient estimate.
The conversion from individual lidar samples of line-of-sight wind speeds to a prediction of
rotor-effective wind speed depends on the scan pattern. For a simple circular scan, the rotor-
effective wind speed prediction can be a running average over the past full circle of measurements,
effective wind speed predicted using lidar measurements and the rotor-effective wind speed experi-
enced by the turbine. More formally, we define ‘measurement coherence’ as the magnitude squared
9
2 ) as a function of frequency (f ):
coherence (γam
2 |Sam (f )|2
γam (f ) = (1.2)
Saa (f )Smm (f )
where Sam (f ) is the cross power spectral density between the rotor-effective wind speed predicted
using lidar measurements, wm (measured wind), and the rotor-effective wind speed experienced by
the turbine, wa (actual wind). The wind experienced by the turbine is estimated as described in
Section 1.2.3. Smm (f ) and Saa (f ) are, respectively, the individual power spectral densities of each
signal.
Magnitude squared coherence can range from zero to one, where zero means no correlation
between the measured and actual wind, and one means perfect correlation. In other words, a
coherence of one means there is some linear time-invariant (LTI) transfer function between the
two signals, and a coherence of less than one means some noise, disturbances, or nonlinearities are
present between the two signals. This coherence function between the actual and measured wind
is affected by wind evolution (which increases with focus distance), u-component estimation error
(Section 1.2.2) due to the line-of-sight angle (which decreases with focus distance, given a fixed scan
radius), the turbine induction zone (the wind in front of the turbine is slowed and redirected because
the turbine is capturing power), and the use of limited lidar measurement locations to represent
the effective wind speed over the entire rotor plane (lidar spatial averaging is usually beneficial).
The overall effect of these error sources is generally strongest at high frequencies, resulting in a
measurement coherence function that drops from close to one at low frequencies down to zero at
high frequencies, with the exact shape, DC value, and bandwidth depending on the measurement
bandwidth’, which gives an approximate frequency up to which preview measurements can be gen-
erally trusted to represent the effective wind speed experienced by the rotor. Coherence bandwidth
is sometimes defined as the frequency at which the magnitude squared coherence drops below 0.5.
In other cases, it is defined as the cutoff frequency of the lowpass filter whose magnitude squared
10
best fits the magnitude squared coherence. Coherence bandwidth may be expressed in temporal
frequency (Hz or rad/s) or spatial frequency (rad/m), where the average wind speed is the conver-
sion factor between the two notations. As average wind speed changes, coherence bandwidth often
remains more constant when expressed as a spatial frequency than when expressed as a temporal
frequency.
For a circular scan pattern, one factor affecting the measurement coherence is the number of
measurement samples per circle. Fig. 1.4 shows that coherence bandwidth generally increases as
number of samples per circle increases. It also shows that using a non-integer number of samples
per circle results in increased coherence bandwidth compared to using an integer number of samples
per circle, assuming a fixed scan rate of one circle per second. This is because sampling at varying
azimuth angles better distributes the measurement points throughout the structures in the wind.
(Note that this data is based on lidar simulations where the lidar model instantaneously takes each
measurement. Instead, with realistic non-instantaneous measurements, if the lidar were to con-
tinue scanning without pausing to take each measurement, a smearing effect would be introduced,
1.3 Contributions
The focus of this work is in improving wind turbine blade pitch control performance in above-
rated wind speeds using lidar-based wind speed measurements, with the control goals of rotor speed
regulation, structural load reduction, and minimal pitch actuation. The specific research questions
addressed are:
• How can an optimal controller be designed while accounting for measurement coherence
0.065
Coherence Bandwidth (rad/m)
0.06
0.055
0.05
0.045
0.04
0.035
0.03
4 5 6 7 8 9 10 11 12
# of Measurement Points Per Circle
Figure 1.4: Coherence bandwidth vs. number of lidar measurement samples per circle, for a
simulated lidar scanning a circle 80 m upwind of the NREL 5-MW turbine in 18 m/s wind. The
coherence bandwidth plotted is the pole location of the first-order lowpass filter, with DC gain 1,
whose magnitude squared best fits the magnitude squared measurement coherence resulting from
the simulation. Lidar scan rate: 1 second per circle. Scan radius: 73% span. (Data points are
limited to measurement sample rates that correspond to multiples of the 0.0125 s simulation time
step.)
12
• How much preview time is useful to the feedforward controller, and how much is provided
by the lidar?
(1) An optimal control design process that incorporates lidar measurements and
optimal control, we mean a control design that minimizes some cost function, allowing
tradeoffs between different outputs. Simulation results show significantly improved Region
3 performance, with tradeoffs between performance measures being easily tunable through
the choice of weights. For example, one implementation on the NREL 5-MW turbine in 18
m/s wind with 17% turbulence intensity results in reductions of 48.3% in RMS generator
speed error, 43.2% in pitch rate standard deviation, 7.7% in blade root out-of-plane damage
equivalent load, 18.4% in tower base fore-aft damage equivalent load, and 3.3% in shaft
torque damage equivalent load, compared to the baseline PI feedback only case.
surement coherence and preview time. Increases in both measurement coherence and
(3) An improved understanding of the amount of preview time that is useful for
blade pitch control, broken down by different sources of delay. The major factors
that determine the maximum amount of useful preview time, including pitch actuator delay
given a controller that expects a constant preview time. This analysis is based on
13
lidar and turbine data provided by NREL after field experiments carried out on a 600-kW
wind turbine known as the two-bladed Controls Advanced Research Turbine (CART2) at
NREL.
shows that a change in the order of multiplication with a gain-scheduling factor significantly
changes the effective gains of the nonlinear closed loop. In addition, after correcting for
these effective gain changes, one PI feedback-only control implementation has significant
method to turn lidar-based control on and off smoothly and an integrator anti-windup im-
signal.
individual pitch control designs created by two different researchers are studied, with the
differences broken down by feedforward controller, feedback controller, and lidar configu-
ration.
are usually proprietary. However, the most common commercial Region 3 blade pitch control
implement individual-pitch gain-scheduled PI control. Both of these strategies have been field
tested [14].
14
1.4.2.1 Feedback-Only
In the research community, a wide variety of feedback-only Region 3 pitch control meth-
ods have been studied. These include model predictive control (MPC) [15, 16, 17], `1 optimal
control [18], state-space methods including Linear-Quadratic Regulator (LQR) and disturbance-
accommodating control (DAC) [19, 20], Linear-Quadratic-Gaussian (LQG) control [21], periodic
control [22], adaptive control [23, 24], H∞ control [25], a classical/LQG/H∞ comparison [26],
sliding-mode control design via H∞ theory [27], feedback linearization [28], and LPV H2 /H∞
control [29].
Control Design Recently and often simultaneously with this dissertation research, many
Region 3 pitch control methods that include feedforward of a wind speed preview have also been
studied. The majority assume perfect measurements in the original design process, although many
are later simulated with imperfect measurements and re-tuned based on simulation, or designed
assuming perfect measurements and then augmented with a prefilter to filter out the uncorre-
lated high frequency portion of the wind measurements. In the following list, the designs where
measurement coherence is instead directly considered in the design process are marked with a *.
• The first known lidar-based control study [3] replaces state estimates of vertical wind shear
in a DAC controller with measured values from a simulated lidar, resulting in reduced blade
root loads.
• *Successful field testing [30, 31, 32] of lidar-based collective pitch control has been imple-
mented.
trol to achieve tower fore-aft oscillation reduction. This strategy accounts for measurement
15
• Two different methods of fatigue load reduction are discussed in [34]. One is an adaptive
• In [35], model-inverse control and preview designed using Tomizuka’s H2 Preview Control
method [36] are studied. Both of these design processes are based on an assumption of
• Lidar for extreme event control, rather than fatigue load reduction, is discussed in [37].
• In [38], a lidar-based H2 controller is tested with simulated wind evolution. This H2 control
design is optimal assuming a white noise wind spectrum and no measurement error. The
thesis [39] contains further discussion of H2 preview design and considers additive white
• Reference [39] also discusses MPC using a wind speed measurement and a DAC approach.
MPC with wind measurement is also discussed in [40], [41], [42], and [43].
• Reference [44] uses spinner-mounted lidar for individual pitch control to reduce once-per-
Wind and Lidar Modeling and Scanning In addition to control design research, re-
search related to Region 3 blade pitch control with a wind speed preview also includes wind and
• Reference [10] discusses lidar measurements for wind turbine control, showing how lidar
configuration, including measurement range, angular offset, and measurement noise affect
• Reference [46] discusses a coherence model for wind evolution, based on data from large
eddy simulations.
• Reference [47] describes the design of a minimum mean-square error prefilter based on lidar
measurement and rotor-effective wind speed statistics. This prefilter is ideal for augmenting
• Reference [48] discusses the effects of the induction zone on lidar measurements.
• Reference [49] discusses a model for the correlation between lidar measurement and rotor-
effective wind speed, based on the Kaimal wind spectrum and accounting for measurement
• Reference [32] discusses wind field reconstruction, lidar modeling, and experimental imple-
mentations.
• Reference [50] discusses a variety of scan patterns, provides rules of thumb for obtaining
lidar measurements representative of the entire rotor plane, and discusses both collective
Most of the controllers above are designed assuming perfect measurements, although many
are simulated and sometimes tuned in imperfect conditions. Instead, this work directly accounts
for measurement coherence, in addition to available preview time, in an optimal control design
process. Two other control designs mentioned above also directly include measurement coherence.
17
The first is the collective pitch feedforward control design discussed in [30, 31, 32]. We use
a very similar control design as a baseline for comparison in Chapter 3, where it is referred to as
lowpass filter feedforward (LPF FF). In contrast to LPF FF control, this dissertation research uses
an optimal control strategy that allows tradeoffs between different goals, includes a model of wind
turbine dynamics, and allows optimal design of both the feedback and feedforward parts of the
controller.
The second control design that directly includes measurement coherence is [33]. Using a
different approach to [33], this dissertation research addresses the feedback design with the same
optimal control method as the combined feedforward/feedback design and hence allows a consistent
evaluation of the benefit of feedforward. In addition, the results in this thesis are for discrete-time
rather than continuous-time design, control blade pitch angle directly rather than controlling a
power reference, and allow modeling of coherence functions with non-unity DC values.
with a minimum-mean-square-error prefilter as in [47]. However, this prefilter is designed given some
available preview time, where the amount available to the prefilter depends on the amount used
by the feedforward controller. Instead, the current dissertation treats the prefilter and controller
as one whole system, essentially solving the two problems together for optimal sharing of preview
time.
Many studies that look at the amount of preview time that is useful to a feedforward controller
quote values in the fraction of a second range [39, 45]. This dissertation research instead finds that
several seconds of preview are useful and comes to this conclusion by accounting for the delay
introduced by lowpass filtering to remove the uncorrelated high frequency portion of the wind
A method to predict arrival time of lidar measurements is also introduced in [42]. The work
in this thesis, however, provides additional insights through the use of real experimental data.
18
1.5 Organization
bance feedforward control. This single-output linear analysis provides initial insight into the ques-
tion of when lidar allows improved performance, while accounting for measurement uncertainty.
Chapter 3 covers an H2 optimal control design process that includes measurement coherence and
preview time, along with simulation results, as described in Contribution 1. It also includes linear-
described in Contribution 5. Chapter 4 provides a preview time analysis. The first section provides
a breakdown of how preview time is used, as described in Contribution 3, and the second section
uses CART2 field experiment data to provide an analysis of available preview time and an expla-
chapter contains its own detailed conclusions, and Chapter 6 contains general conclusions and
future recommendations. Appendix A provides a proof addressing the modeled coherence and
power spectra. Appendix B explains new results in gain scheduling implementation, as described
Chapter 3 has partially appeared in [51]. Section 4.1 has partially appeared in [52]. Section 4.2
has partially appeared in [53]. Chapter 5 has partially appeared in [52]. Parts of Chapter 3 and
Model-Inverse-Based Analysis
2.1 Introduction
One possible method for lidar-based wind turbine blade pitch control is to use model-inverse
feedforward control [54]. This method involves inverting the linear plant model for some chosen
output. Except when additional prefiltering is used, model-inverse control assumes that a perfect
trol because the reference is usually perfectly known. However, blade pitch control using wind
not perfectly known. In this chapter, we analyze the results of a model-inverse blade pitch controller
that is designed under the assumption of perfect measurements but is then subjected to error in
We study the linearized system in Figure 2.1, where we have added feedforward control in an
attempt to reduce the effect of the wind disturbance on some output y. In this section, we assume
−1
perfect feedforward control: F = −Tyβ Tyw , where Tyβ is the closed-loop transfer function from βF
to y, and Tyw is the closed-loop transfer function from wa to y. We call this choice of F perfect
feedforward control because it results in y=0 when combined with perfect wind measurements
Figure 2.1: Block diagram showing linearized models of the wind turbine P , the feedback controller
C, and the feedforward controller F . The labeled signals are the differences from the linearization
operating point: measured wind speed wm , actual wind speed wa , feedforward pitch command
βF , generator speed Ω, and some output of interest y. For example, y may be equal to Ω, to a
blade bending moment, or to a tower bending moment. Each of these signals, excluding Ω, may
be a vector of three values when considering individual pitch control, or they may each be a single
value when considering collective pitch control. T is the closed-loop model containing the feedback
controller and the turbine.
y = Tyw wa + Tyβ F wm
−1
y = Tyw wa + Tyβ (−Tyβ Tyw )wm
y = Tyw (wa − wm )
Therefore, at any given frequency, perfect feedforward is better than no feedforward at reduc-
ing y(f ) whenever ||wa (f ) − wm (f )|| < ||wa (f )||. That is, perfect feedforward is helpful whenever
the norm of the error in measurement is less than the norm of the actual wind. There is an ex-
ception at DC, where an integrator in the feedback loop makes Tyw (0) = 0, and therefore y(0) = 0
−1
Perfect feedforward (F = −Tyβ Tyw ) will never be possible because of errors in plant mod-
eling and because Tyβ often can only be inverted approximately, either because y is a vector of
outputs, or, even if y is a single output, because of non-minimum-phase zeros in the plant model.
When non-minimum-phase zeros are present, approximate inversion methods include the non-causal
21
series expansion, nonminimum-phase zeros ignore, zero-phase-error tracking controller, and zero-
magnitude-error tracking controller methods [54]. In this section, we study Figure 2.1 with imper-
−1
fect feedforward control: F = −Tyβ Tyw + . Then solving for y:
y = Tyw wa + Tyβ F wm
−1
y = Tyw wa + Tyβ (−Tyβ Tyw + )wm
||Tyw (f )(wa (f ) − wm (f )) + Tyβ (f )(f )wm (f )|| < ||Tyw (f )wa (f )||
Now the error in the output y is the sum of two terms: one due to measurement error and
another due to the error in F . It is possible but unlikely that these two sources of error would
cancel each other out. To reduce y, we need a balance between accurate measurements and “small”
measurements, where the tradeoff depends on the “sizes” of , Tyw , and Tyβ , and “small” and “sizes”
refer to the magnitudes or norms of the signals and transfer functions. Because errors increase with
frequency, this suggests that the value used for wm should be a lowpass-filtered version of what was
actually measured.
2.4 Conclusion
formance at a given frequency only when the norm of the measurement error, combined with the
error due to an imperfect model-inverse, is less than the norm of the wind disturbance. Otherwise
feedforward is harmful compared to feedback alone. However, the wind measurements can be low-
pass filtered before being input into an ideal model-inverse controller. Simley [47] has extended
the work in this chapter to the design of an optimal prefilter for the wind measurements; using
this prefilter, feedforward is never harmful. Combining a model-inverse controller with a prefilter
has limitations: only a single output can be effectively minimized using model-inverse control, and
22
when the turbine model is non-minimum phase, either infinite preview time or an approximation
3.1 Introduction
H2 optimal control allows the designer to make tradeoffs between multiple objectives. In
above-rated wind speeds, the wind turbine control goals are rotor speed regulation, structural load
minimization, and minimal pitch actuation. Improvements in meeting these goals result in increased
turbine lifetime and therefore reduced cost of energy. The benefit of a lidar sensor in achieving
these control goals is a function both of preview time, which is approximately determined by the
lidar focus distance divided by wind speed, and of measurement coherence: the correlation as a
function of frequency between the wind measurement and the wind that is actually experienced by
the turbine.
This chapter explains an H2 optimal control design method that accounts for measurement
coherence and preview time in the design process. It also includes simulation results in addition to
linear-model-based results, and the linear-model-based results are validated through simulation.
proportional-integral (PI) feedback controller, rather than directly applied to the turbine. This
means that there is an H2 feedback part that is added in parallel to the standard PI feedback,
and there is also an added H2 feedforward part. By wrapping the combined feedforward/feedback
H2 controller around the standard PI feedback controller, we facilitate smoothly turning off the
lidar-based part of the control, returning to PI-control when necessary. Keeping the PI controller
included also means that the augmented turbine model used in the H2 control design process is
24
less uncertain because it contains feedback. In addition, adding H2 feedback allows greater control
freedom than using PI feedback alone; the differences in operation between these two feedback
We use Hazell’s method of solving for the optimal controller [55], taking advantage of the
chain of delays in the problem structure to reduce the size of the Riccati equations we need to solve.
This makes the solution faster, more numerically accurate, and easier to implement than standard
H2 optimal control.
For a range of different preview times, coherence bandwidths, and cost function weights, we
solve for an optimal H2 combined feedback/feedforward wind turbine blade pitch controller whose
output is to be added to that of the baseline PI feedback controller. We present simulation results
for a small number of cases and linear model-based results for a larger number of cases.
This chapter is organized as follows. In Section 3.2, we describe our design methods, including
the turbine models, the cost function, the augmented plant used for H2 synthesis, and details on
solving the H2 problem. In Section 3.3, we describe the baseline PI feedback controller for the 5-
MW turbine and also discuss a lowpass filter feedforward controller used for comparing results. In
Section 3.4, we explain the simulation environment and implementation considerations and present
simulation results. We show linear-model-based results in Section 3.5, and finally we summarize
Simulations are performed using NREL’s FAST code [56], an aeroelastic wind turbine sim-
ulation code, using the land-based version of the NREL 5-MW turbine model [9]. Table 3.1 lists
properties of the turbine. All 16 available degrees of freedom are turned on in simulation. The
standard turbine model does not include a pitch-actuator model, so we have added appropriate
25
Table 3.1: Properties of the NREL 5-MW turbine.
dynamics to represent a pitch actuator (a second-order lowpass filter [57] with a cutoff frequency of
1 Hz and a damping ratio of 0.7) because without a pitch actuator model, there is no delay between
pitch command and pitch actuation, and the benefit of preview is not as apparent. In simulation,
dynamic inflow is turned on, whereas in linearization, dynamic inflow is not an option.
The turbine model is linearized using FAST, with reduced degrees of freedom: only the
generator, first tower fore-aft mode, and first flapwise blade mode are turned on for linearization
because these are the modes that have a noticeable effect on the value of the turbine transfer
functions at the frequencies (below about 0.4 Hz) where collective pitch actuation is to be used.
Using a limited set of degrees of freedom reduces the complexity of solving for the optimal controller.
The turbine is linearized about a wind speed operating point of interest (18 m/s) within the range
of above-rated wind speed operation, which occurs from 11.4 m/s to 25 m/s. After linearization,
the unobservable generator degree of freedom is removed from the linear turbine model (the plant),
and the plant is converted to discrete time using the sample-and-hold method with a sample rate
of 10 Hz. The plant is also augmented with a linearization of the turbine model’s standard Region
3 constant-power torque feedback controller [9] and the actuator model described above. The
τ = (−36.716 Nm/rpm) · Ω
26
where τ is generator torque in Nm (as a deviation from the operating point) and Ω is generator
The linearization and augmentation result in an overall linear turbine model P . P has two
outputs: y and Ω, and two inputs: w and β. The vector y is the cost to be minimized and is
described in Section 3.2.2. The other 3 signals are scalars: Ω is the generator speed error in rpm
and is used for feedback control, w is the rotor-effective wind speed in m/s (as a deviation from the
operating point), and β is the collective blade pitch command in degrees (as a deviation from the
operating point). For example, at the 18 m/s operating point, the turbine model transfer function
from w to Ω is
or the sum of the variances of the components of the augmented linear turbine model’s cost output
The cost function includes the generator speed error Ω, the overall pitch rate β̇, and the H2
pitch acceleration β̈H2 ; along with scalar weights αpitchrate and αH2 pitchaccel , which allow tuning of
Generator speed error is penalized because improved generator speed regulation is itself a
control goal, and because improved generator speed regulation generally also indirectly leads to
27
structural load reduction, as shown in Fig. 3.1. In Fig. 3.1, when the ideal feedforward controller
is used to eliminate generator speed error, it also significantly reduces blade root moment, tower
base fore-aft moment, and pitch rate at low frequencies. However, it significantly increases tower
bending and pitch rate at the 0.33 Hz tower resonant frequency. For simplicity, direct penalization
of structural loads is not included in our cost output. Our design framework does allow for the
possibility of direct penalization of structural loads, but leaving out direct penalization of structural
The overall pitch rate is penalized in order to preserve the life of the pitch actuators. This
also indirectly leads to reduced tower base fore-aft moment at the resonant frequency, compared to
The H2 pitch acceleration output is the second derivative of the pitch command output of the
H2 part of the controller. In our design, this pitch command output is added to the pitch command
output of the baseline PI feedback controller, and together the two commands sum to form an overall
pitch command β. The H2 pitch acceleration is penalized to minimize high frequency actuation
from the H2 part of the controller because of model uncertainty and unmodeled dynamics at high
frequencies.
The weight αpitchrate allows tuning of the tradeoff between generator speed error and overall
pitch rate. For reference, we found that the αpitchrate value for which the baseline PI controller
is optimal is 2693.6 in 18 m/s wind. In Section 3.4.3, Simulation Results, we call this value of
The penalty on the second derivative of the H2 controller output (H2 pitch acceleration) is
automatically tuned through an iterative simulation process that updates αH2 pitchaccel in order to
approximately minimize the first two terms in the cost function in simulation:
N
X 2
2 2
|Ωi | + αpitchrate β̇i (3.2)
i=0
After αH2 pitchaccel is updated, the iterative process then designs a new controller using this new
penalty, simulates the turbine with the new controller, and calculates the performance metric (3.2)
28
Linear−Model Prediction of
Power Spectral Densities (Power/Hz)
30000
no F
Gen speed
error (rpm)
10000
0
6e+007
TwrBsMyt (kNm) RootMyc (kNm)
4e+007
2e+007
0
3e+010
2e+010
1e+010
0
Pitch rate (deg/s)
40
20
0
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
Frequency (Hz)
Figure 3.1: Linear-model predictions of power spectral densities of generator speed error, blade
root out-of-plane moment (RootMyc), tower base fore-aft moment (TwrBsMyt), and pitch rate for
the NREL 5-MW turbine, with all 16 degrees of freedom, linearized at 18 m/s, assuming a rotor-
averaged Kaimal wind spectrum with class A turbulence. Blue represents feedback-only control
(no F ), and green represents feedback control augmented with an ideal feedforward control (F )
designed to perfectly eliminate generator speed error.
29
for the current iteration. The process is repeated until the performance metric (3.2) improvement
has been less than 0.5% for each of the past three iterations. This automatic tuning is done
individually for each controller. Ideally, in order to show the benefit of lidar, we would like to
use the same cost function in designing both the combined feedforward/feedback controller and
the feedback-only controller to which it is compared, but in practice, αH2 pitchaccel will differ for
each individual controller even when αpitchrate is the same. However, the two controllers can still
be designed with the same overall objective: minimize (3.2) in simulation over the set of optimal
As an alternative to H2 control, our control design framework also allows for implementation
of H∞ control, which instead minimizes the largest possible amplification over all frequencies of
a sinusoidal input, and allows for robustness guarantees. Through use of a simple switch, our
design framework can solve either type of optimal controller for our linear augmented plant model.
However, we have not yet fine-tuned the cost function, uncertainty-modeling, and implementation
consideration changes necessary for good simulation results in the H∞ control case, so further detail
The standard H2 problem configuration is shown in Fig. 3.2. Given the augmented plant
PH2 , the controller KH2 is solved to minimize the H2 norm of the closed-loop transfer function
from u1 to y1 : ||Ty1 u1 ||2 . This is equivalent to minimizing the 2-norm of the output vector y1 for a
Fig. 3.3 shows the augmented plant that models our turbine with imperfect preview wind
measurements, with blocks grouped to show how they fit into the standard problem configuration.
In Fig. 3.3, everything outlined in red corresponds to PH2 in Fig. 3.2, and the remaining feedfor-
ward controller H2 FF and feedback controller H2 FB in Fig. 3.3 correspond to KH2 in Fig. 3.2.
30
u1 y1
PH2
u2 y2
KH2
Corresponding input and output signals are highlighted in matching colors and defined below.
n1 wm
u1 = ; u2 = βH2 ; y1 = y ; y2 =
n2 Ω
The linear turbine model, P , is described in Section 3.2.1.2; its output y contains the weighted
outputs to be minimized as described in Section 3.2.2, and its output Ω is generator speed error, used
for feedback. PI FB is the linearized model of the baseline gain-scheduled PI feedback controller.
The actual wind wa is represented by white noise n2 multiplied by the wind spectrum K. The
signal wa arrives at the turbine after a delay equal to the amount of preview time, delay = z −τP /Ts ,
where τP is the preview time and Ts = 1/10 s is the sample time. In this study, we look at a range
of preview times from 0 to 10 seconds in linear modeling, and in simulation focus on a preview time
of 3.944 seconds, determined by the wind speed and choice of lidar scan pattern. The measured
wind, wm , is the signal that is measured by the lidar and is input to the feedforward controller. In
this model, it is represented by wm = Lwa +Hn1 where L is typically a lowpass filter, H is typically
a highpass filter, and n1 is white noise. L and H are designed to make the coherence between wa
and wm equal to a desired coherence, and to make the power spectrums of the two signals match:
Saa = Smm .
On a real turbine, the coherence between the actual and measured wind can be measured,
using an observer to determine the actual wind. We can then set up our model’s wa and wm to match
this desired coherence. The power spectrum of both signals can also be measured on a real turbine.
The actual and measured wind tend to have similar power spectrums, and we set them equal for
simplicity in our model. This is reasonable because the difference in power spectrums tends to affect
31
n1 n2
H
K
+ L wa
wm
KH2 H2 delay
FF
w y
H2 βH2 β
+ + P
FB
PI
FB
Ω PH2
Figure 3.3: Augmented plant with feedback and feedforward controllers. Color coding corresponds
with Fig. 3.2.
32
n1 n2
K H
L +
wm wa
delay
Figure 3.4: Intermediate step in rearranging top part of Fig. 3.3 into Fig. 3.5. Color coding
corresponds with Fig. 3.2.
only the resulting controller, not the cost. For example, say Smm = 4Saa , and we correctly model
the coherence, but we incorrectly assume Smm = Saa while solving for an H2 optimal controller
and cost. The resulting cost is still correct, and the resulting feedforward controller simply needs
r r
Saa 1 Saa
to be multiplied by = to be corrected. The caveat is that if varies as a function
Smm 2 Smm
of frequency, it will have a nonzero phase delay, and therefore some additional preview time may
The augmented plant as shown in Fig. 3.3 does not yet meet the requirements for H2 syn-
• Each measurement output (each component of y2 in Fig. 3.2) should have direct feedthrough
n1
from a disturbance input (a component of u1 ). n2
y1 ). L delay +
wm
Direct feedthrough means that an output’s value at the current sample time is affected (at
wa(delayed)
least partially) by an input’s value at the current sample time. The second condition is met
because the pitch rate is included in our cost, because there is direct feedthrough from the H2
33
1 2
H2 2
2 H2
H2
Figure 3.5: ‘Equivalent’ augmented plant to Fig. 3.3, rearranged to a format more suitable for H2
synthesis. Color coding corresponds with Fig. 3.2.
34
pitch command βH2 to the pitch command β and because, through careful choice of discrete-time
implementation of the pitch actuator model, there is direct feedthrough from pitch command β to
pitch rate. The position of the delay block in Fig. 3.3 is the main roadblock in meeting the first
condition because it prevents any direct feedthrough from a white noise disturbance n1 or n2 to
the generator speed measurement Ω. We therefore rearrange the augmented plant in two steps as
follows.
First, note that magnitude squared coherence is commutative, and the H, L, and K filters
have been chosen to give a desired coherence between wa and wm and make Saa = Smm . Therefore,
from Fig. 3.3, we can flip the top half of the block diagram as shown in Fig. 3.4 so that the signal
that once reached wa now reaches wm , and vice versa. By swapping the inputs to wa and wm ,
we have changed neither the coherence between the two signals nor the fact that Saa = Smm .
Therefore this new arrangement in Fig. 3.4 is ‘equivalent’ for our purposes. (Note that while we
use Saa = Smm for simplicity, it is not a requirement. The details for modeling a non-unity ratio
Starting with Fig. 3.4, the second step is to move the delay block, which is now at the output
of a summing junction, to each input side of the summing junction. This is simply using the
distributive property of multiplication over addition. This results in a delay block after the L filter
and another delay block after the H filter. Any delay on the H side of the summing junction can
Fig. 3.5 is ‘equivalent’ for our purposes to Fig. 3.3, and the delay block is no longer in the
path from n2 to the generator speed output Ω. The two augmented plants are ‘equivalent’ but not
exactly equal because they match in magnitude, and in coherence between wa and wm , but the
phase from the white noise inputs to the augmented plant outputs differs between the two. This
phase difference does not matter because the goal is only to minimize the closed-loop magnitude,
In addition to fitting into the standard H2 problem configuration, our rearranged augmented
plant also now fits into the framework of the “generalized regulator problem with both previewable
35
and non-previewable disturbances” solved by Hazell [55]. Hazell’s method recognizes that when a
chain of delays is a specific part of the augmented plant structure, then the order of the delay block
(the number of samples of delay) does not need to be included in the dimension of the two Discrete
DARE, the same result can be achieved using a reduced-dimension DARE, along with a discrete
Lyapunov equation and a Stein equation. Instead of the high-dimension output-feedback DARE,
the same result can be achieved using a reduced-dimension DARE and by recognizing that the states
of the delay chain can be perfectly reconstructed by the controller. Fig. 3.6 contains essentially the
same rearranged augmented plant as shown in Fig. 3.5, but with four outlined blocks corresponding
to the four blocks of Hazell’s framework (Figure 1.5 within reference [55]). The dimension of each
Fig. 3.7 shows the magnitude of our filter K. It models magnitude of the rotor-averaged
wind speed from simulation with an IEC Kaimal spectrum [59], assuming Class A turbulence (high
turbulence) and the normal turbulence model (NTM), with a mean wind speed of 18 m/s. The
19 P36 !1
X j=1 vu (i, j)3 3
wa∗ = wCP (i)fA (i)
36
i=1
where wa∗ is the rotor-averaged wind speed, wa = wa∗ − 18 m/s is the actual wind experienced
by the turbine (as a deviation from the operating point) for use in coherence calculations, i is a
radius index, j is an azimuth index, wCP (i) is a weight representing the relative contribution to
power coefficient at that radius, fA (i) is the fraction of the swept area represented by that radius
point, and vu (i, j) is the u-component of the wind speed at point i,j. Rotor-averaging gives a
more accurate representation of the frequency content of the rotor-effective wind disturbance than
directly using the Kaimal spectrum, which is representative of the wind at a single point in space.
The Kaimal spectrum is one of the simplest of many spectral models of the wind. All models have
36
1 2
2 H2
Figure 3.6: ‘Equivalent’ augmented plant to Fig. 3.3, rearranged to a format more suitable for
H2 synthesis, as in Fig. 3.5, but with blocks outlined to correspond with those in Hazell’s frame-
work [55]. (The delay and L blocks are also swapped compared to Fig. 3.5.)
2
10
1
10
0
Magnitude
10
−1
10
−2
10
−3
10 −4 −3 −2 −1 0
10 10 10 10 10
Frequency (Hz)
Figure 3.7: Magnitude of the filter K at 18 m/s operating point, obtained using rotor-average wind
speed from simulation with Class A normal turbulence model (NTM) [59].
37
lowpass filter characteristics, with more energy in the wind at lower frequencies. The phase of
our filter K does not matter because the input to K is white noise. To determine this low-order
filter K with magnitude matching the magnitude obtained from simulation, we use log-Chebychev
magnitude design to fit a minimum-phase state-space model (Matlab [60] function fitmagfrd()),
achieving a close match with two poles and two zeros. We require the same number of zeros as
After obtaining this state-space model, we also artificially increased the value of the direct
feedthrough term (the ‘D’ matrix) in order to further increase the likelihood of finding a valid solu-
tion to the Riccati equations. This artificial increase in the direct feedthrough term is responsible
for the rise in the magnitude of K with frequency at frequencies above 1 Hz, as seen in Fig. 3.7,
but it does not significantly affect the magnitude of K at the lower frequencies (those below about
p
|L| = coh(f ) (3.3)
∠L = 0 (3.4)
p
|H| = |K|2 (1 − |L|2 ) (3.5)
∠H = irrelevant (3.6)
where coh(f ) is the desired measurement coherence as defined in (1.2). This choice of magnitudes for
L and H ensures that the coherence between wa and wm equals the desired measurement coherence
(coh(f )) and that Saa = Smm as proven in Appendix A. (For Saa /Smm 6= 1, in order to leave
coherence unchanged while correctly modeling the ratio of Saa /Smm , H should remain unchanged
p
from above based on the original K and L, the L above should be multiplied by Saa /Smm , and
p
K should be multiplied by Smm /Saa .) The choice of phase for L in (3.4) ensures that the average
phase between the two signals is zero, which is typically true for real measurements when we have
In this study, our desired magnitude squared coherence is the magnitude squared of a scaled
38
Butterworth lowpass filter (our L). This filter is determined by three design parameters:
(1) We choose a filter order of 1 because this best fits the magnitude squared coherence found
in simulation.
(2) We find solutions for a range of different coherence bandwidths, which is our name for the
cutoff frequency of the filter (the frequency where the magnitude response of the filter is
p
1/2 of its DC value). For various lidar trajectories studied by Schlipf for the NREL 5-
MW turbine, achievable coherence bandwidths ranged from 0.03 rad/m to 0.07 rad/m [32].
Our best-fit coherence bandwidth for our simulation with an 18 m/s IEC wind field and
a modeled lidar, as described in Section 3.4, is 0.2186 Hz, or 0.076 rad/m, approximately
equal to that of Schlipf’s optimal lidar trajectory (2π · 0.2186 Hz/(18 m/s) = 0.076 rad/m).
This is the highest coherence bandwidth we would expect to see with current experimental
lidar configurations, and it is at the top of the achievable range in part because wind
evolution and the induction zone are not modeled in our simulations.
(3) A scaling factor χ ≤ 1 is included because otherwise the control design process expects al-
most perfect coherence at low frequencies, and therefore almost eliminates feedback control
because it expects that feedforward can almost completely take over. However, feedback
work well.
(0.95)(0.064357)(z + 1)
L= ; Ts = 1/10 s. (3.7)
z − 0.8713
p
This choice of L satisfies |L| = coh(f ) but does not satisfy ∠L = 0. It is impossible for a causal
filter to have both this desired magnitude and a phase of zero. To approximately correct for the
φL (f )(in degrees) 1
tL (f ) = −
360 degrees f
39
where tL (f ) is the time delay of L and φL (f ) is the phase of L. The time delay of L is relatively
constant as a function of frequency at all frequencies below the bandwidth, and drops as frequency
increases above the bandwidth. We make the approximation tL (f ) = tL (flow ), where flow is a
very low frequency relative to the bandwidth. We can then subtract this constant value of time
delay, tL (flow ), from the delay block that precedes the L filter in the augmented plant (Fig. 3.6).
Then although the phase of L does not satisfy (3.4), the phase across both L and the delay block
together is approximately correct. At higher frequencies, the amount we subtract from the delay
block is an overapproximation, so our model shows slightly less high-frequency delay, and therefore
less high-frequency preview available than there really is. However, we do correctly model the low
coherence values at higher frequencies, so the resulting controller would be unable to make much
use of additional high-frequency preview time. This small error in modeling available preview time
is especially not a concern for preview times above about 6 seconds, because as we will see in
Section 3.5, the cost quickly flattens out as a function of preview time.
A filter H that meets the magnitude requirement of equation (3.5) is H = K · HnoK , where
HnoK is a filter which satisfies |L|2 + |HnoK |2 = 1. In the case where L has a scaling factor of
χ = 1, HnoK is simply the highpass Butterworth filter with the same order and cutoff frequency
as L. With a non-unity scaling factor χ, an HnoK with a very good approximation to the required
p
HnoK = 1 − χ2 Lω2 /Lω1
where Lω1 and Lω2 are lowpass Butterworth filters of order n where n is the order of L, with DC
gains of 1, and with cutoff frequencies ω1 and ω2 respectively, where ω1 < ω2 . The cutoff frequency
ω2 is set to match the cutoff frequency of L, so Lω2 = L/χ. The cutoff frequency ω1 can then be
found by noting that the ratio between ω2 and ω1 determines the amount by which |HnoK | rises
1
as frequency goes from DC to infinity. |HnoK | must rise overall by a factor of p because
1 − χ 2
p
its DC magnitude is 1 − χ2 , and it must have unity magnitude as frequency approaches infinity
because |L| = 0 as frequency approaches infinity. The overall increase of |HnoK | will be 20n dB per
40
1
ω2
0.8
0.6 ω1
0.4
|L|
0.2 |HnoK|
|L|2+|HnoK|2
0 −3 −2 −1 0 1
10 10 10 10 10
Frequency (Hz)
Figure 3.8: Magnitude of example L from equation (3.7), along with corresponding |HnoK | and
|L|2 + |HnoK |2 . Cutoff frequency ω1 is labeled on |HnoK |, and cutoff frequency ω2 is labeled on
both |L| and |HnoK |.
ω2
decade between ω1 and ω2 , where the number of decades between ω1 and ω2 is log10 . An
ω1
increase of 20n dB is equivalent to being multiplied by a factor of 10n . In equation form, we then
have
ω2
!
1 log10
p = (10n ) ω1
1 − χ2
This equation can be rearranged and simplified to solve for the cutoff frequency ω1 :
1
!
p
ω1 = ω2 1 − χ2 n
This completes the design of HnoK . Fig. 3.8 shows |L| and |HnoK | for the example L from equa-
tion (3.7). Their magnitudes squared sum to almost exactly one, as required.
The rearranged augmented plant described in Section 3.2.3 and shown in Fig. 3.6 still does
not quite meet the requirements of H2 synthesis because the turbine model P does not have direct
feedthrough from its disturbance input (wind) to its measurement output (generator speed). Since
our turbine model transfer function has one fewer zero than pole, we simply multiply its disturbance
41
input by z to add one more zero, and then add one more sample of delay (z −1 ) to the delay block
In the special case of infinite coherence bandwidth with a scaling factor of χ = 1, the filter
L = 1 and H = 0, and therefore, once again we do not meet a requirement of H2 synthesis because
there is no direct feedthrough from a disturbance input to each measurement output: as shown in
Fig. 3.6, the measurement outputs are wm and Ω, and when H = 0, this requirement is not met
for the measurement output Ω. Our solution to this problem, in this infinite coherence bandwidth
case only, is to simply delete the measurement output Ω, leaving only the measurement output wm .
This means we are using no feedback control, only feedforward control. Fortunately, in this case,
the measurement Ω is redundant, and anything we could have done with feedback control can also
be done with feedforward control because the assumption in this perfect feedforward case is that
all inputs to the plant model and the plant model itself are perfectly known, therefore the outputs
We use Hazell’s method [55] to solve for the minimum-H2 -cost controller. Both standard
H2 synthesis and Hazell’s method require the solution of two discrete algebraic Riccati equations
(DAREs). Hazell’s method is relevant for H2 (or H∞ ) synthesis in problems where the augmented
plant PH2 contains a significant amount of delay, which often appears in modeling a preview of a
reference or disturbance signal. The method is an improvement upon standard H2 synthesis for
these problems because it allows the size of the DAREs to be reduced to the order of the plant and
Two additional techniques, in addition to using Hazell’s method, are used to minimize the
occurrence of numerical problems that can prevent accurate solution of the two DAREs. The first is
to use a balanced realization (Matlab [60] function balreal()) on the augmented plant model before
doing the H2 control synthesis. The second is multiplying the output vector y by a scalar. Some-
times the algorithm cannot solve for the controller that minimizes y, but can solve the equivalent
problem of minimizing 10y, for example. We automatically loop through increasing factors until a
The frequency and impulse responses of two sample H2 controllers are shown in Figs. 3.9,
3.10, 3.11, and 3.12. The controllers, as labeled in the legends, are summarized in Table 3.2, which
lists the parameters and explains the abbreviated names of the controllers discussed throughout
Chapter 3.
In Fig. 3.9, Feedback Parts, the feedback parts of the two H2 controllers have approximately
the same magnitude at low frequencies as the PI feedback controller to which they are added in
parallel, but they have the opposite phase, meaning they are canceling much of the low frequency
action of the PI feedback control. The peak at about 0.32 Hz corresponds to the first tower mode.
This low-frequency magnitude reduction compared to PI alone can be seen in Fig. 3.10, Feed-
back Overall. In addition, H2 FF/FB, medPP control has a low-frequency magnitude reduction in
its overall feedback compared to H2 FB, medPP. This is expected because when the feedforward
controller is compensating for the wind disturbances at low frequencies, it frees the feedback con-
troller to shift towards disturbance rejection at higher frequencies, since the feedback disturbance
rejection across all frequencies is limited by the Bode sensitivity integral [61]. This low-frequency
magnitude reduction for H2 FF/FB, medPP feedback is also consistent with other work [30, 50]
that has shown reductions in PI gains to be beneficial when using lidar-based feedforward control.
controller, shown in Fig. 3.11, has a similar bandwidth to the coherence bandwidth used in the
design, but also contains peaks at the tower frequency and its harmonics.
The impulse response of this same feedforward part is shown in Fig. 3.12. This is the output
of the linear feedforward block when its input is an impulse at time zero. The blue line indicates the
3.944 s preview time, and the first 3.944 s of this response represents preview actuation. In addition,
the negative initial response indicates that the feedforward part of the controller is nonminimum
phase.
43
Table 3.2: A list of names and properties of controllers discussed throughout Chapter 3
H2 FF/FB, medPP
−40 PI FB
−60
−80
−100
135
90
Phase (deg)
45
0
−45
−90
−4 −3 −2 −1 0
10 10 10 10 10
Frequency (Hz)
Figure 3.9: Frequency response of H2 and PI feedback controllers, as individual parts (before
addition in parallel). (For H2 FF/FB, only its feedback portion is shown here.) Input: generator
speed error (Ω) (rpm). Output: pitch (β) (rad).
44
Magnitude (dB)
−20 H FF/FB, medPP
2
no H2, PI FB only
−40
−60
90
45
Phase (deg)
−45
−90
−4 −3 −2 −1 0
10 10 10 10 10
Frequency (Hz)
Figure 3.10: Overall frequency response of PI plus H2 feedback controllers, along with PI alone.
Input: generator speed error (Ω) (rpm). Output: pitch (β) (rad).
−40
Magnitude (dB)
−60
−80
H2 FF/FB, medPP
−100
1080
900
Phase (deg)
720
540
360
180
0
−4 −3 −2 −1 0
10 10 10 10 10
Frequency (Hz)
6
Amplitude
−2
0 2 4 6 8 10 12
Time (sec)
Figure 3.13: Block diagram of control implementation in simulation. The baseline gain-scheduled
proportional-integral feedback (PI FB) controller is represented by everything not colored blue.
Blue blocks and lines represent the H2 control additions, so that the overall block diagram represents
the combined H2 feedforward/feedback controller.
Our baseline controller is the standard 5-MW turbine collective-pitch feedback controller, a
gain-scheduled PI controller [9], with three changes from the original referenced controller. This
controller is represented in Fig. 3.13 by all parts of the block diagram that are not colored blue.
Other aspects of this figure will be explained later in this section, and in the upcoming Section 3.4.2.
The first change from the original referenced controller [9] is that because the gain-scheduling factor
is based on blade pitch angle, our addition of an actuator model introduces the design choice of
whether to use the signal from before or after the actuator model. We chose to use the signal after
the actuator model so that the gain scheduling is based on the actual pitch angle rather than the
pitch command, but the two options appear to produce very similar simulation results.
Second, the gain-scheduling multiplication happens before the proportional-integral (PI) con-
trol blocks. The original gain-scheduling order, with multiplication after the PI blocks, has un-
intended consequences, effectively reducing control gains from their design values. Our change in
multiplication order preserves the originally-designed control gains and also allows easier implemen-
47
is contained in Appendix B.
even when feedforward is added. The original anti-windup method limits the integrator state to
be within minimum and maximum values, with the lower limit being zero. With feedback alone,
and a lower pitch angle saturation limit of zero, this correctly works to prevent the integrator from
accumulating a negative value in below-rated wind speeds where the blade pitch is saturated at
zero and the generator speed is below the setpoint. However, if a feedforward signal with a positive
average offset value is contributing to the blade pitch command, the average value of the integrator
is reduced. The integrator will often need to accumulate a negative value in order to perform its
normal function. The original anti-windup method prevents this because of the lower limit of zero
Our new anti-windup method does not limit the integrator state. Instead, we use the purple
switch block shown in the lower-left of Fig. 3.13. This switch holds the generator speed error
input to the integrator at zero whenever the pitch command saturation block is saturated. This
replacement of the input with zero prevents windup. Because this method does not place any limits
on the integrator state, the integrator can now operate at negative values when necessary.
We compare our H2 controller results not only to the baseline PI controller, but also to a
lowpass filter feedforward (LPF FF) controller which is similar to a design that has been successfully
lowpass filter Second, the delayed measurement is lowpass filtered using a cutoff frequency equal
48
to:
• the coherence bandwidth, in linear-model results, where we assume for simplicity that
• the transfer function estimate bandwidth, in simulation, where for improved perfor-
mance we do not assume equal power spectrums. This refers to the transfer function
estimate between measured and actual wind. In this case we will use the name “LPF
FF*.”
lookup table Finally, the delayed-and-filtered measurement is input to a lookup table which maps
The output of LPF FF is the resulting feedforward pitch command βH2 , which is added to that of
3.4 Simulations
Simulations are performed using NREL’s FAST code [56], an aeroelastic wind turbine sim-
ulation code, augmented with a continuous-wave Doppler lidar model [10]. We use a simulation
sample rate of 80 Hz, where the controller is resampled from 10 Hz to 80 Hz using the Tustin
approximation. We use a lidar sample rate of 40 Hz, and a lidar scan rate of 1 second per circle.
We scan a circle using a half-cone angle of 30◦ , a radius of 46.2 m (73.3% span), and a distance
of 80 m ahead of the turbine rotor. The measured wind signal wm used as input to the feedfor-
∗ − 18 m/s, where w ∗ is the running average of the past full circle of
ward controller is wm = wm m
measurements corrected for line-of-sight angle as in [10] and 18 m/s is the operating point. The
wind field used in simulations is an IEC wind field of 1 hour duration, with 18 m/s average wind
speed and 17% turbulence intensity. The first 100 seconds of data are discarded before calculation
In this turbulent 18 m/s wind field, we use H2 control designs based upon a linear turbine
model, linearized at 18 m/s. These linear designs require gain scheduling in order to operate
effectively throughout the large range of wind speeds encountered in the wind field, which has an
instantaneous hub-height minimum of 7.4 m/s and maximum of 29.1 m/s. We found the best
results by separately scheduling the feedforward and feedback parts of the H2 controller, as shown
in Fig. 3.13, because this allows the preview (feedforward) part to be scheduled based on the
previewed wind speed while the feedback part maintains the same scheduling as the PI feedback,
gives the actual steady-state pitch angles for a given steady-state wind speed, each plotted versus
wind speed in Fig. 3.14, wSS is a vector of steady-state wind speeds, and 0.2606 is the steady-state
The H2 feedback scheduling is accomplished because its input is the generator speed error
that has already been multiplied by the gain-scheduling output. It is important that the same
gain-scheduling signal is used for both the PI and the H2 feedback parts because the H2 feedback
design partially cancels out the PI feedback, and the gain scheduling needs to match for this to
happen accurately. We also multiply the H2 input signal by 1/0.297 = 3.37 to correct the scaling
to 1 at the 18 m/s operating point, where the gain-scheduling factor has value 0.297, because the
H2 feedback block was designed based on an unscaled generator speed error input.
50
Actual
Linearized at 18 m/s
0.4
0.3
Operating Point
0.2
0.1
0
15 20 25 30
Steady−State Wind Speed (m/s)
Figure 3.14: Steady-state pitch angle versus steady-state wind speed, along with the curve’s lin-
earization at the 18 m/s operating point.
Additionally, the anti-windup implementation for the two feedback parts matches. The signal
that is the input to the PI feedback integrator is the same signal used as input to the H2 feedback,
as shown in Fig. 3.13, so whenever the scheduled generator speed error is switched to zero because
of pitch saturation, it affects both feedback controllers in the same way. This should prevent
integrator windup in the H2 feedback in addition to the PI feedback, allowing a smooth transition
Finally, lidar may not always be available, and the lidar-based part of the control may need
to be turned off. Both feedforward and feedback parts of the H2 controller should be switched off
together because the feedback part is designed to operate together with the feedforward part and is
not optimal operating alone. Smooth and instantaneous switching between PI-only and PI-plus-H2
control operation should be achievable as follows. Let a(k) be the value of the signal at point ‘a’ in
Fig. 3.13 at time sample k, b(k) be the value of the signal at point ‘b’ in Fig. 3.13 at time sample
k, and ξ(k) be the internal state of the integrator at time sample k (assuming a forward-Euler
51
integrator implementation where the state equals the output). When turning off the H2 part of the
controller at time sample k, set a(k) = a(k + 1) = a(k + 2) · · · = 0, and ξ(k) = ξ(k)∗ + a(k − 1)/KI ,
where ξ(k)∗ is the value the integrator would normally have under standard operation without this
addition, and KI is the integral gain shown in Fig. 3.13. This results in no step change in the total
control command a + b because a(k) + b(k) = b(k) = b(k)∗ + a(k − 1), where b(k)∗ is the value b(k)
would normally have without the addition to the integrator state. When turning on the H2 part
of the controller at time k, the H2 control states should be initialized based on their steady-state
values for the current wind speed, and the initial H2 control output value, a(k), should be scaled
and subtracted from the internal state of the integrator at time k: ξ(k) = ξ(k)∗ − a(k)/KI . Again,
by following these steps, there should be no step change in the total control command a + b.
Fig. 3.15 shows a summary of simulation results for simulation conditions described in Sec-
tion 3.4.1, relative to the baseline PI control. All controllers shown are an overall improvement
compared to the baseline PI control, with the exception of the H2 FF/FB, hiPP controller increas-
ing generator speed error. The medium-pitch-penalty H2 feedback-only controller (H2 FB, medPP)
results in only a small performance improvement over PI FB, but for controller with feedforward
included, the performance improvements are substantial. For example, the medium-pitch-penalty
in RMS generator speed error, 39.0% in pitch rate standard deviation, 7.5% in blade root out-of-
plane damage equivalent load, 16.9% in tower base fore-aft damage equivalent load, and 4.3% in
shaft torque damage equivalent load, compared to the baseline PI feedback only case. In addition,
the tradeoff between pitch rate and generator speed error can be easily controlled by tuning the
pitch rate penalty αpitchrate , as shown by the differences between the ‘lowPP’, ‘medPP’, and ‘hiPP’
results. Blade root and tower base loads tend to decrease with pitch rate standard deviation, while
Damage equivalent loads were calculated using rainflow counting, assuming an S-N-curve
52
slope m = 10 for blade material and m = 3 for tower and shaft materials [2]. When m = 10, a 7%
decrease in DEL corresponds to a doubling (1/0.9310 = 2.1) in lifetime, and when m = 3, a 21%
Fig. 3.16 shows the same simulation results as Fig. 3.15 but now shown relative to the LPF
FF* control. The H2 FF/FB controller can be tuned such that performance is improved in all
five metrics compared to LPF FF*. For example H2 FF/FB, medPP results in reductions of 0.9%
in RMS generator speed error, 27.3% in pitch rate standard deviation, 1.2% in blade root out-of-
plane damage equivalent load, 4.7% in tower base fore-aft damage equivalent load, and 0.9% in
shaft torque damage equivalent load, compared to the LPF FF* case.
Fig. 3.17 also shows the same simulation results as Fig. 3.15 but now shown relative to
the H2 FB, medPP control. The H2 FF/FB, medPP results in reductions of 50.6% in RMS
generator speed error, 31.4% in pitch rate standard deviation, 5.1% in blade root out-of-plane
damage equivalent load, 16.1% in tower base fore-aft damage equivalent load, and 3.2% in shaft
Fig. 3.18 shows power spectral densities of generator speed error and pitch rate, both as-
expected based on the linear model used for control design and as-recorded from simulation. General
trends in the linear case based on reduced-order linear modeling carry over to the simulation case
generator speed error and RMS pitch rate that should be observed in simulation. The square
root of the area under each curve in Fig. 3.18 is equal to the standard deviation of that signal,
which also equals the RMS value when the mean is zero. Linear-model-based predictions of RMS
generator speed error, along with the corresponding simulation results of RMS generator speed
error, are shown in Fig. 3.19, and linear-model-based predictions of RMS pitch rate, along with
the corresponding simulation results of RMS pitch rate, are shown in Fig. 3.20. The results are
53
1.6
RMS(gen speed error)
std(pitch rate)
1.4
mean blade root OOP DEL
tower base FA DEL
1.2 shaft torque DEL
normalized to PI FBonly
1
0.8
0.6
0.4
0.2
0
LPF FF* H FF/FB, H FF/FB, H FF/FB, H FB,
2 2 2 2
lowPP medPP hiPP medPP
Figure 3.15: Performance metrics (OOP: out-of-plane, DEL: damage equivalent load, FA: fore-aft),
normalized so all baseline PI feedback bars would equal 1 if displayed, for five controllers listed in
Table 3.2.
3
RMS(gen speed error)
std(pitch rate)
mean blade root OOP DEL
2.5
tower base FA DEL
shaft torque DEL
normalized to LPF FF*
1.5
0.5
0
H2 FF/FB, H2 FF/FB, H2 FF/FB, H2 FB, PI FB
lowPP medPP hiPP medPP
Figure 3.16: Performance metrics (OOP: out-of-plane, DEL: damage equivalent load, FA: fore-aft),
normalized so all LPF FF* bars would equal 1 if displayed, for five controllers listed in Table 3.2.
54
1.8
RMS(gen speed error)
1.6 std(pitch rate)
mean blade root OOP DEL
tower base FA DEL
1.4
1.2
0.8
0.6
0.4
0.2
0
H2 FF/FB, H2 FF/FB, H2 FF/FB, PI FB LPF FF*
lowPP medPP hiPP
Figure 3.17: Performance metrics (OOP: out-of-plane, DEL: damage equivalent load, FA: fore-aft),
normalized so all H2 FB, medPP bars would equal 1 if displayed, for five controllers listed in
Table 3.2.
PI FB
LPF FF*
6000 6000 H2 FB, medPP
H2 FF/FB, medPP
Power Spectral Density (Power/Hz)
4000 4000
2000 2000
0 0
5 5
To: pitch rate (deg/s)
4 4
3 3
2 2
1 1
0 0
0 0.1 0.2 0.3 0.4 0 0.1 0.2 0.3 0.4
Frequency (Hz) Frequency (Hz)
Figure 3.18: Linear-model expectations and simulation results of generator speed error and pitch
rate power spectral densities, for four controllers listed in Table 3.2.
55
RMS Generator Speed Error, in % of Baseline PI
140
LPF FF*
120 H2 FB, medPP
H2 FF/FB, medPP
100
H FF/FB, lowPP
2
Simulation
80 H FF/FB, hiPP
2
60
40
20
0
0 50 100 150
Linear Expectation
Figure 3.19: Linear-model expectations vs. simulation results for RMS generator speed error, as
percentages of baseline PI feedback-only (PI FB), for controllers listed in Table 3.2. Best-fit line
equation: y = 0.93x + 3.4
The trends of Figs. 3.19 and 3.20 can be used to map linear-model-based expectations to
a better prediction of actual simulation results. For example, if the linear model predicts an H2
feedforward/feedback controller to have 40% of the RMS pitch rate of the baseline PI feedback
controller, we can use the line fit through the points in Fig. 3.20, and determine that the simulation
result is likely to show this H2 feedforward/feedback controller to actually have 50% of the RMS
simulation, allowing quick study over a wide range of design parameters. The predictions in this
section should be referenced to Section 3.4.3.1 to determine how these predictions generally translate
to actual simulation results, and the results should be considered most trustworthy when the
preview time and coherence bandwidth are close to the simulation conditions of 0.2186 Hz coherence
bandwidth and 3.944 seconds of preview. Limited simulations with varied coherence bandwidths
56
100
80
Simulation
60
LPF FF*
H2 FB, medPP
40
H2 FF/FB, medPP
20 H2 FF/FB, lowPP
H2 FF/FB, hiPP
0
0 20 40 60 80 100 120
Linear Expectation
Figure 3.20: Linear-model expectations vs. simulation results for RMS pitch rate, as percentages
of baseline PI feedback-only (PI FB), for controllers listed in Table 3.2. Best-fit line equation:
y = 0.89x + 14.2
57
and preview times have shown similar trends to these linear-model-based predictions.
Fig. 3.21 shows linear-model-based expectations of RMS generator speed error, and Fig. 3.22
shows linear-model-based expectations of RMS pitch rate, for a range of coherence bandwidths
and preview times. LPF FF control and H2 control have relatively similar generator speed error
predictions, but H2 control has much lower pitch rate predictions. LPF FF results improve with
Fig. 3.21 and Fig. 3.22 quantify the H2 control improvements possible as preview time in-
creases from 0 to 10 seconds, and as coherence bandwidth increases from 0 to infinity, where 0
represents no feedforward and infinity represents perfect measurements. Both increasing coherence
bandwidth and increasing preview time reach a point of diminishing returns. The maximum useful
preview time is the preview time above which RMS generator speed error and pitch rate remain flat.
The maximum useful preview time increases as coherence bandwidth decreases because at lower
coherence bandwidths, more lowpass filtering is required, and this introduces more delay, which
can be compensated with preview time [52]. However, even with infinite coherence bandwidth,
which does not require lowpass filtering to minimize measurement error, up to 3 seconds of preview
time is useful because of our pitch rate and H2 pitch acceleration penalties. The zero coherence
bandwidth result for the LPF FF controller is the same as PI FB. The zero coherence bandwidth
result for the H2 controller shows only a small improvement in generator speed error compared to
For coherence bandwidths other than zero and infinity, there are no results below some
minimum preview time. This is because our design process involves subtracting the delay (tL (flow ))
of the lowpass filter (L) from the delay block as explained in Section 3.2.3.2. Before this subtraction,
the amount of delay in the delay block is equal to the amount of preview time. If tL (flow ) is
greater than the amount of preview time, the subtraction results in a negative delay, which is not
implementable.
The overall cost must theoretically be monotonically decreasing with increasing preview time
and with increasing coherence bandwidth, because there is only increased information. However,
58
the generator speed error and pitch rate, being individual components of the cost function, can
and do individually violate this monotonicity, as long as an increased cost in one of these output
components is balanced by a corresponding decreased cost in another, including β̈H2 , which is not
displayed here.
H2 optimal control design, within the limits of our modeling assumptions and cost func-
tions, allows a comparison of the best performance possible with lidar versus the best performance
possible without lidar, while using a realistic model of wind measurements in the design process
by including a known coherence model and preview time. In simulation, H2 optimal combined
feedforward/feedback control significantly reduces structural loads, pitch rate, and generator speed
error compared to three other types of control: baseline PI feedback only, H2 optimal feedback
only, and lowpass filter feedforward control. For example, the medium-pitch-penalty H2 combined
feedforward/feedback controller, on the NREL 5-MW turbine in 18 m/s wind with 17% turbu-
lence intensity, results in reductions of 51.3% in RMS generator speed error, 39.0% in pitch rate
standard deviation, 7.5% in blade root out-of-plane damage equivalent load, 16.9% in tower base
fore-aft damage equivalent load, and 4.3% in shaft torque damage equivalent load, compared to
the baseline PI feedback only case; and reductions of 0.9% in RMS generator speed error, 27.3%
in pitch rate standard deviation, 1.2% in blade root out-of-plane damage equivalent load, 4.7%
in tower base fore-aft damage equivalent load, and 0.9% in shaft torque damage equivalent load,
compared to the LPF FF* case. In addition, tradeoffs between outputs can be easily tuned through
the choice of weights. Linear-model-based results are also presented alongside simulation results for
comparison and validation, and additional linear-model-based results are presented to show how
the benefit of lidar is expected to vary across a range of coherence bandwidths and preview times.
Suggested future work includes improving gain scheduling by solving for linear H2 controllers
at multiple operating points and transitioning between them, field testing, and using the same
control design techniques to study the benefit of lidar for individual pitch control. This chapter
59
18 m/s Linear−Model Prediction: RMS Generator Speed Error
1
Coherence
Bandwidth (Hz)
normalized to PI FBonly
0.8
0
0.025
0.6 0.05
0.1
0.2186
0.4 0.4
Inf
0.2
0
0 2 4 6 8 10
Preview Time (s)
Figure 3.21: 18 m/s linear-model expectations of RMS generator speed error, normalized so that the
expected RMS generator speed error of baseline PI feedback-only equals 1, for a range of preview
times and coherence bandwidths, for lowpass filter feedforward (LPF FF) control in dashed lines
and H2 feedforward/feedback control with medium pitch penalty (H2 FF/FB, medPP) in solid
lines.
1
Coherence
Bandwidth (Hz)
normalized to PI FBonly
0.8
0
0.025
0.6 0.05
0.1
0.2186
0.4 0.4
Inf
0.2
0
0 2 4 6 8 10
Preview Time (s)
Figure 3.22: 18 m/s linear-model expectations of RMS pitch rate, normalized so that the ex-
pected RMS pitch rate of baseline PI feedback-only equals 1, for a range of preview times and
coherence bandwidths, for lowpass filter feedforward (LPF FF) control in dashed lines and H2
feedforward/feedback control with medium pitch penalty (H2 FF/FB, medPP) in solid lines.
60
has discussed collective pitch control, where each blade receives the same pitch command based on
rotor-averaged wind speed measurements and generator speed error. Lidar-based individual pitch
control could be implemented using additional cyclic pitch commands, allowing the turbine to react
to previews of changing wind shear or yaw error. Chapter 5 discusses individual pitch control, but
The maximum preview time that is useful for feedforward control depends on the particular
choice of cost function or output to be minimized. For the simplest case where minimizing only
generator speed error is of interest, the maximum useful preview time can be broken down into a
sum of three values. The first is determined by the actuator. The second is determined by the
lowpass filtering which is necessary because the measured wind does not exactly match the actual
wind experienced by the turbine, especially at high frequencies. The third is determined by the
turbine itself. Figure 4.1 shows estimates for these three values for the NREL 5-MW turbine,
along with the approximate expected preview time available from the lidar, assuming a fixed focus
distance 80 m upwind, a 0.5 s delay due to the scan pattern, a coherence bandwidth fixed at
0.06 rad/m in spatial frequency regardless of wind speed, and using Taylor’s frozen turbulence
hypothesis [62, 63], which assumes that the frozen wind field is marched toward to turbine at a
constant average wind speed. The preview time estimates shown for the actuator, lowpass filter,
4.1.1 Actuator
Our actuator is modeled by a second-order lowpass filter with a natural frequency of 1 Hz and
a damping ratio of 0.7. Figures 4.2 and 4.3 show how phase and time delay vary with frequency for
this model. At frequencies below 0.2 Hz, the time delay remains at about 0.225 seconds. Therefore
62
7
Available from Lidar
6 Turbine + Actuator + LPF
Required Preview Time (s) For:
Turbine + Actuator
Turbine
5
0
12 14 16 18 20 22 24
Wind Speed (m/s)
Figure 4.1: Cumulative required preview times to compensate the delay introduced by the turbine,
actuator, and lowpass filter. Available preview time from lidar also shown, with 0.5 s scan pattern
delay included. Assumptions include a fixed 80 m upwind focus distance, fixed 0.06 rad/m coherence
bandwidth, Taylors frozen turbulence hypothesis, and generator speed error minimization only.
63
0
−45
phase φ (deg)
−90
−135
−180 −2 −1 0 1 2
10 10 10 10 10
frequency f (Hz)
Figure 4.2: Phase φ of second-order actuator model (natural frequency of 1 Hz and damping ratio
of 0.7).
0.225 seconds of preview time is required to compensate for this actuator delay.
At high frequencies, lidar-based wind preview measurements are not an accurate representa-
tion of the rotor-effective wind speed experienced by the turbine. Therefore some lowpass filtering
must be included in any feedforward controller that relies on these measurements. The ideal pre-
filter is Sam (f )/Smm (f ), where Sam (f ) is the cross power spectral density between the measured
wind wm and the actual wind wa experienced by the turbine, and Smm (f ) is the power spectral
density of wm . This ideal prefilter has been referred to both as the transfer function estimate
from wm to wa [32] and as the minimum mean squared error Wiener smoothing filter for esti-
mating wa given wm [47]. The magnitude of this ideal prefilter is equal to the square root of the
magnitude squared coherence, normalized by the ratio between the power spectra of wa and wm :
r
2
Saa
|Sam (f )/Smm (f )| = γam . If Saa = Smm , the magnitude squared of the prefilter is simply
Smm
equal to the measurement coherence. Therefore the coherence bandwidth is a rough estimate of
The more that is filtered out by this lowpass prefilter, the greater the delay it introduces, and
therefore the greater the maximum useful preview time: Doubling the order of the lowpass filter
64
0.3
0.25
time delay τφ(s)
0.2
0.15
0.1
0.05
0 −2 −1 0 1 2
10 10 10 10 10
frequency f (Hz)
Figure 4.3: Time delay τφ of second-order actuator model (natural frequency of 1 Hz damping ratio
of 0.7). τφ = −φ/360◦ /f
65
(LPF) doubles the amount of useful preview time corresponding to this filter because it doubles the
phase delay. Doubling the cutoff frequency (the location of the LPF poles) cuts this useful preview
time in half because the majority of the phase delay shifts to higher frequencies. For this example
we assume a measurement coherence that is best fit by a first-order LPF with a cutoff of 0.06
rad/m in spatial frequency, and we assume that this spatial frequency will remain approximately
constant as the wind speed increases. Multiplication with wind speed gives the cutoff frequency in
Hz, and because that cutoff frequency increases with wind speed, the delay introduced by the filter
decreases with wind speed. The resulting delay is plotted in Figure 4.1 as the LPF component.
4.1.3 Turbine
The third value affecting maximum useful preview time appears because for some output
y, there is often more phase delay in the Blade-Pitch-to-y transfer function than there is in the
Wind-Speed-to-y transfer function of the wind turbine. This difference in phase delay is equal to
−1
the phase advance of the ideal model-inverse feedforward controller F = −Tyβ Tyw , where Tyβ is
the closed-loop (feedback included) transfer function from blade pitch to output, and Tyw is the
For ease of calculations in this section, we let y be the generator speed output Ω, and we use
−1
collective pitch feedforward control. Then the equation above simplifies to F = −Pyβ Pyw , where P
is the open-loop plant, and the feedback controllers do not matter. This is because the torque and
collective pitch feedback controllers C both take their input from generator speed, and since the
feedforward is designed to reduce generator speed error to zero, these torque and collective pitch
−1 −1
−Tyβ Tyw = −Pyβ Pyw
Any individual pitch feedback control also does not matter because it does not affect the average
66
Figure 4.4: Preview time required for F with all available degrees of freedom (DOFs) turned on in
linearization. Above 0.5 Hz, time values begin to vary much more widely, but we do not intend to
use frequencies any higher for feedforward control.
blade pitch, which is what we are adding to with this collective pitch feedforward control.
To find the preview time required to implement F in the above configuration, we first linearize
the open loop turbine model at 7 different wind speeds from 12 m/s to 24 m/s and, at each wind
−1
speed, solve for F = −Pyβ Pyw . For each F , we then take the phase advance in degrees (φ) at
−τφ = φ/360◦ /f
Figure 4.4 shows the results when all available degrees of freedom are turned on in lin-
earization. The preview time required varies with wind speed, but stays fairly constant over the
frequencies of interest.
Figure 4.5 shows results when only 5 degrees of freedom (DOFs) are turned on in linearization:
first flapwise blade mode (×3 blades), generator, and drivetrain rotational-flexibility. Because
Figures 4.4 and 4.5 are very similar, we can conclude that the required preview time depends
mainly on some or all of only these 5 DOFs. A third set of linearizations was also created with
only two DOFs: generator and drivetrain rotational-flexibility. The resulting required preview
time was almost exactly zero at all frequencies and at all wind speeds! This shows that the first
67
Figure 4.5: Preview time required for F with 5 DOFs turned on in linearization: first flapwise
blade mode (×3 blades), generator, and drivetrain rotational-flexibility. Above 0.5 Hz, time values
begin to vary much more widely, but we do not intend to use frequencies any higher for feedforward
control.
flapwise blade mode is the main reason why the phase delay from Blade-Pitch-to-Generator-Speed
is different than the phase delay from Wind-Speed-to-Generator-Speed. In other words, the amount
It is also interesting to note that when blade flap DOF is turned on, a nonminimum phase
zero appears in the Blade-Pitch-to-Generator-Speed transfer function, but not in the Wind-Speed-
phase zero and the difference in phase delay appear only under the same circumstances. Preview
time can also be used for model-inverse-based control when the plant has non-minimum-phase
(unstable) zeros. Non-minimum-phase zeros often appear in plants with non-collocated sensors
and actuators, and they appear in the linearized 5-MW model when blade flexibility degrees of
freedom are turned on. In this case, the ideal model-inverse feedforward controller is unstable, but
including the non-causal series expansion [64, 65]. Unlike model-inverse-based control, H2 control
synthesis does not address non-mimimum-phase zeros directly, but their presence in the plant does
affect the resulting H2 optimal controller and the performance it can achieve.
A smaller amount of preview time than a plotted ‘required’ value is still useful because it
68
allows a better approximation to the ideal F than would be possible without any preview time.
In addition, required turbine preview time may increase greatly from the less than 0.3 sec shown
here if we focus on other outputs besides generator speed. A controller that is designed only
to regulate generator speed is not ideal for reducing loads, although it should help reduce them
somewhat. Previous work done using preview control [35] suggests a turbine required preview time
of 2 seconds when only penalizing pitch rate and generator speed error, and 20 seconds when tower
sway is also penalized. However, other work [66] shows that the preview time required to minimize
loads depends on wind speed and is at most 1 second. Future work should further investigate the
required turbine preview time when designing for minimizing pitch actuation and blade root and
So far in this thesis we have assumed a fixed, known amount of preview time, according to
Taylor’s frozen turbulence hypothesis [62, 63], and used the FAST simulation environment which
does not model wind evolution, but instead models a frozen 3D wind field marching toward the
turbine at a constant average wind speed. However, in real-world operation, the wind field evolves,
and, given a fixed lidar focus distance, the preview time is always changing and not precisely known.
Fig. 4.6 shows an example block diagram of combined feedforward/feedback wind turbine control
using a lidar, now with additional blocks included to analyze and manage the real-world variation
We use vu (t) to refer to the upstream estimate of the approaching rotor-effective wind speed.
Throughout this section, the rotor-effective wind speed is estimated based on line-of-sight lidar
measurements distributed over a vertical plane upstream of and parallel to the turbine rotor plane,
by using the running average of the past full circle of measurements corrected for line-of-sight
angle as in [10], while assuming that the turbine is perfectly aligned with the wind. We use vr (t)
to refer to the wind experienced at the turbine rotor, as determined by a wind speed estimator,
which essentially uses the turbine as an anemometer. The relationship between vu (t) and vr (t)
can be characterized by their coherence (correlation as a function of frequency) and the time delay
between them. This section focuses on the time delay between the two signals using data from field
tests on a 600-kW wind turbine known as the two-bladed Controls Advanced Research Turbine
(CART2) [67], a research wind turbine that is used to test advanced control algorithms at NREL.
Arrival time ta (t) is the time it takes for the wind to travel from the measurement location
to the turbine rotor. Preview time tp (t) and time delay td (t) are closely related to ta (t), as shown
visually in Fig. 4.7. Given one of these three timing signals, the others can be easily determined.
Compared to the wind at the measurement location, the upstream vu (t) is delayed by one-half
the time it takes to complete one scan pattern because vu (t) always depends on the most recent
complete scan, which includes measurements taken one-half scan pattern ago on average. The most
70
Figure 4.6: Block diagram of wind turbine control using a turbine-mounted lidar under real-world
conditions with varying preview time. Physical devices and processes are shown in blue, and control
system components are shown in green. It is likely that a different low-pass filter (sometimes
called a prefilter) is also contained in the feedforward part of the Combined Feedforward/Feedback
Controller.
71
Table 4.1: Properties of the NREL CART2 turbine.
recent complete scan is used to provide a representation of the wind across the full rotor disk.
There are mixed conclusions in the available literature on the importance of accurate timing
predictions. One study involving a collective pitch feedforward controller [57] shows (in its Figure
9) that errors of more than 5 s are needed before the combined feedforward/feedback controller
performance is worse than that of feedback alone. Another study [42] states that “lead or lag errors
in the wind speed measurement, which is fed to the controller, severely reduce the performance of
the controller.” The second study uses peak time of cross-covariance between vu (t) and vr (t) to
predict timing, but it does not use real-world data for these signals.
In this section, we analyze timing by using real-world data. We show how quickly timing
can vary, and we show how helpful it would be to correctly predict it if that were possible, rather
than assuming it stays constant for a constant average wind speed. In Section 4.2.1, we describe
the lidar measurements and how we obtain vu (t) and low-pass filter it into LPF(vu (t)). We also
describe the wind turbine and the wind speed estimator used to obtain vr (t). In Section 4.2.2, we
describe taking windowed cross-covariances between LPF(vu (t)) and vr (t) to find td (t), the time
delay between them. In Section 4.2.3, we pass the signal vu (t) through the varying time delay
td (t), and show that this provides an improved preview measurement for use in a time-invariant
Lidar measurements were obtained from CART2 field tests [30] at NREL. Table 4.1 contains
We use the four data sets shown in Table 4.2, downsampled to a rate of 40 Hz. These are
72
Wind at
measurement
location ds, Delay due to scan time
vu (t)
dLPF, Delay due to low-pass
Arrival filtering of vu(t)
time LPF(vu (t))
ta(t)
Preview
time
tp(t)
Time
delay
td (t)
Wind at
rotor dr, Delay due to rotor effective
wind speed estimation
vr (t)
Figure 4.7: Visual representation of possible introduced time delays and the naming conventions
used in Section 4.2. Delays in blue are constant in time, and red indicates functions of time that
vary with the evolving wind field. In Section 4.2, dLP F = dr = 0, and ds = 0.67s. We use the word
“timing” to refer in general to all three signals ta (t), tp (t), and td (t).
73
Table 4.2: Data sets from 2012 CART2 field tests
the four longest continuous data sets available from the 2012 CART2 field tests, excluding data
taken before torque calibration, before hard target problems were solved, when lidar measurements
were not available, and when the turbine was not operating in a region suitable for the wind speed
estimator to produce accurate results. Set 1 had several periods of up to 4 s long of missing lidar
measurements; we linearly interpolated to replace missing data. This did not cause any noticeable
problems, and although we often plot data from Set 1 alone, we also provide a summary of results
The lidar used in the field tests is a modified Windcube [68], a pulsed lidar that scans a
circular pattern at five focus distances, as shown in Fig. 4.8. Although the lidar can achieve other
scan patterns, the circular pattern was chosen because it was expected to provide measurements
well correlated to the wind experienced by the turbine. We chose to use data from the first focus
distance of one rotor diameter (42.7 m) upwind of the lidar, because these measurements were best
correlated to the turbine. The lidar was mounted on the wind turbine nacelle at a distance of 1.66 m
downwind of the hub (rotor center). Thus the measurement location is 42.7 m − 1.66 m = 41.0 m
ahead of the rotor. Each circle of six measurement points was scanned in 1.33 s, yielding an average
scan time delay (Fig. 4.7) of ds = 1.33/2 s = 0.67 s. The lidar measures the wind component along
its line of sight. Each measurement is adjusted using the lidar line-of-sight angle to find the wind
speed in the x-direction (directly downwind toward the turbine), under the assumption that the y
and z components of the wind are zero. A running average over the past six adjusted measurements,
from the first measurement plane only, was provided as an estimate of the x-component of the rotor-
74
50
40
z (m)
30
20
10
−20
0 0
0 −20 20
−40 −60 −80 y (m)
x (m)
Figure 4.8: Lidar measurement configuration. The lidar is shown as a white square on the nacelle of
the turbine. The black dots represent the locations of wind speed measurements used in Section 4.2.
The white dots represent locations where wind speed measurements were also recorded.
average wind speed. This is our vu (t), which is shown in Fig. 4.9.
Fig. 4.9 also contains the low-pass filtered version of vu (t), LPF(vu (t)). This includes low-pass
filtering with a cutoff frequency of 0.1432 Hz and notches at once-per-revolution (1P), twice-per-
revolution (2P), and the drivetrain torsion frequency (0.695 Hz, 1.39 Hz, and 3.36 Hz, respectively).
These particular low-pass and notch filters do not introduce any delays because zero-phase filtering
(filtfilt() in MATLAB [60]) is used, which is possible because the filtering is not done in real time.
Essentially, the filtering is done both forward in time and backward in time so that any effects
of the filter on phase are canceled out. Together these filters minimize the magnitude of vu (t) at
frequencies at which it may be correlated to vr (t) because of tower and lidar motion rather than
The wind speed estimator uses the torque balance method, which assumes
τr − τg = J Ω̇ (4.1)
where τr is the torque applied by the wind to the rotor, τg is the torque applied at the generator
(multiplied by the gearbox ratio), J is the drivetrain moment of inertia (as observed from the rotor
side of the gearbox), and Ω̇ is the rotor acceleration. Wind speed estimators that neglect rotor
acceleration may introduce a time delay because of the inertial response of the rotor.
After τr is solved for using the above equation, a lookup table is used to determine the rotor-
75
15 vu(t)
LPF(vu(t))
14 vr(t)
Wind speed (m/s)
13
12
11
10
8
300 350 400 450 500 550 600
Time (s)
Figure 4.9: Upstream measurement (vu (t)), low-pass filtered vu (t), and rotor estimate (vr (t)).
Five-minute portion from Set 1 data.
76
effective wind speed for the given τr , blade pitch angle β, and rotor speed Ω, with an adjustment
where R is the rotor radius, CP is the power coefficient (fraction of wind power captured), and λ
is the tip speed ratio (ΩR/vr (t)). Pitch, torque, and rotor speed signals are low-pass filtered and
notch filtered (at 1P, 2P, and the drivetrain torsion frequency) using zero-phase filtering before
being used in the estimator. The result of the lookup table is the rotor-effective wind speed vr (t)
The time delay td (t) between LPF(vu (t)) and vr (t) can be estimated from the times of peak
windowed cross-covariances between LPF(vu (t)) and vr (t). However, simply shifting a window
across the data and recording each time of peak cross-covariance results in a td (t) that contains
sudden jumps to unrealistic values. Three main steps can reduce these occurrences. First, we choose
a Gaussian window with a large enough window size. Second, we detrend (subtract the best-fit line
from) the data within each window before applying the cross-covariance function. Third, we enforce
a rate limit on td (t), along with a lower limit of 0 s. These steps yield the blue td (t) curves shown
in Fig. 4.10. We show results for three window sizes for comparison, with standard deviations of
σ = 30 s, 15 s, and 60 s, and total window sizes of 180 s, 90 s, and 360 s, respectively. Throughout
the remainder of Section 4.2, we use only the td (t) resulting from the σ = 30 s window. This choice
of window size is a trade-off because smaller sizes give better time resolution and tend to lead to
the greatest increases in coherence bandwidth (when td (t) is used as described in Section 4.2.3);
however smaller sizes are also more likely to produce non-physical results (which can be impossible
Based on Taylor’s frozen turbulence hypothesis [62, 63], ta (t) is typically assumed to be equal
77
25
Time of peak cross covariance (td(t)) σ=30 s
Time of peak cross covariance (td(t)) σ=15 s
Time of peak cross covariance (td(t)) σ=60 s
20
Expected td(t) (TLPF(v (t)))
u
Expected t (t) (T )
d v (t)
r
15
Time (s)
10
0
0 500 1000 1500 2000 2500
Window center time (s)
Figure 4.10: Time (td (t)) of peaks of windowed cross-covariances of LPF(vu (t)) with vr (t). σ
describes Gaussian window size; td (t) with σ = 30s is used throughout this chapter and can be
assumed when not specified; plots using the other two sizes are presented here for comparison. This
plot also includes two versions of Tv as in (4.3). Set 1 data used.
78
to D/v, where D is the distance traveled from the measurement location to the rotor and v is the
average wind speed. Fig. 4.10, in addition to showing td (t), also shows the expected td (t) computed
as
Tv = D/v − ds (4.3)
where D = 41 m, ds = 0.67 s, T stands for Taylor, and the subscript v varies depending on what
choice of v we use in the equation. We use either LPF(vu (t)) or vr (t) for v in Fig. 4.10.
At approximately 2,000 s, there is a large spike in td (t). This corresponds to a data set
portion when the lidar measurements are relatively inaccurate because of low wind speeds. In this
portion, the wind speed as determined by the turbine-based estimator is below 6 m/s; whereas the
lower limit of a possible lidar measurement is 6 m/s as a result of the technique used to eliminate
In the remaining data (before 1,900 s) in Fig. 4.10, td (t) on average matches its expected
value Tv relatively well, as shown in Table 4.3. It is likely that induction zone effects (slowing of
wind as it approaches the rotor) are responsible for td (t) being longer than expected on average.
U∞ −1
= 1 − a[1 + ξ(1 + ξ 2 )−1/2 ] (4.4)
U
where U∞ is the undisturbed velocity and ξ = x/R, where x is the distance from the rotor (negative
upwind) and R is the rotor radius [69]. The axial induction factor a is approximately the optimal
value 1/3 in below-rated wind speeds because the turbine is extracting as much power as possible
from the wind. This is when we expect the induction-zone slowdown effect to be strongest. As
wind speed increases above rated, the axial induction factor decreases, and the slowdown effect
diminishes. The CART2 has a rated wind speed of approximately 13 m/s, and we have little data
Integrating (4.4) with respect to ξ and then multiplying by U/U∞ /ξ gives the arrival time
multipliers M shown in Fig. 4.11. To calculate a theoretical arrival time ta (t) that accounts for the
79
Table 4.3: Mean values of td (t) and Tv : either uncorrected or corrected for induction zone
where D is the measurement distance and v is the wind speed at the measurement location. Our
1.12. This means ta (t) should be 12% greater than the value we calculated without accounting for
the induction zone. Table 4.3 shows the uncorrected and corrected versions of TLPF(vu (t)) , where
the corrected version is (D/v)M −ds . It also shows the uncorrected and corrected versions of Tvr (t) ,
where the corrected version is (D/v)M ∗ − ds , where M ∗ results from integrating (4.4) with respect
to ξ and then multiplying by 1/ξ. The multiplication factor U/U∞ is not included in M ∗ because
vr (t) is assumed to be a delayed estimate of U∞ instead of U . In both cases, the induction zone
On shorter time scales, td (t) often differs from its expected value Tv by several seconds, as
shown in Fig. 4.10. Timing does not stay constant for a constant average wind speed. This may
be in part because of the evolution of the wind field, and also the 3D nature of the wind field in
which the wind speed at a given location does not always match the speed of travel of the turbulent
airflow structure carrying that wind speed. The timing may also be affected by wind shear and
wind components in the y and z directions, which exist although we assume them to be zero. In
this section, our goal is not to improve the modeling to predict timing, but to use the time lag
found in post-processing to estimate how much room for improvement exists in the current model.
80
1.25
a=1/6
a=1/3
1.2 a=1/2
1.1
1.05
1
−8 −6 −4 −2 0
x/R
Figure 4.11: Arrival time multipliers M for use in (4.5). x is the measurement distance from the
rotor (positive downwind), R is the rotor radius, and a is the axial induction factor.
A controller incorporating lidar measurements, for example the controller used in the field
testing [30], can be designed for some given preview time and some given correlation or transfer
function estimate. The controller may be adapted as arrival time and correlation change, but so far
this has been done on a timescale of a few minutes. In Fig. 4.10 td (t) (σ = 30 s) jumps from 3 s to 6
s in only 68 s, for example. One way to address varying arrival time is to delay the measured signal
vu (t) by varying amounts before sending it to the controller. This essentially uses interpolation to
stretch and shrink various parts of vu (t) in time. The amount of varying delay that should be used
on vu (t) is equal to the preview time tp (t) minus the constant amount of preview expected by the
controller.
For this method to be successful, two requirements must be met by the preview time tp (t).
First, tp (t) must always be greater than or equal to the amount of preview required by the controller.
This is required for real-time operation because it is impossible to use negative delay. Second, let
∆ represent taking the differences between adjacent elements in a time series, where ∆t = 1/40 s
because of the 40 Hz sample rate. Then the slope ∆tp (t)/∆t (and thus ∆ta (t)/∆t and ∆td (t)/∆t
as well) must always be greater than −1. If ∆tp (t)/∆t = −1, two different wind speeds would
be expected to arrive at the same time, and below a slope of −1, wind speeds would be expected
81
to arrive in a swapped order than when they were measured. In situations when arrival time has
a slope ≤ −1, a possible solution is to overwrite the older measurement with the more recently
With a small amount of zooming in, td (t) in Fig. 4.10 appears to meet the requirement of
∆td (t)/∆t > −1. However, upon zooming in enough to see individual time samples, we see that the
time of peak cross-covariance is discretized at the sample rate (40 Hz). This results in steep steps,
many with slope equal to −1. To solve this problem, we filter this data with a boxcar filter of 11
samples in length, which, for this data, provides just enough smoothing to keep ∆td (t)/∆t > −1.
The boxcar filter is centered so that no time delay is introduced. After smoothing the data, Fig. 4.12
was created by stretching and shrinking LPF(vu (t)) from Fig. 4.9 according to td (t) from Fig. 4.10.
Because of this, it sometimes lags and sometimes leads the original LPF(vu (t)), and this results in
a better correlation at low frequencies to vr (t). We are most interested in improving low-frequency
correlation because the low frequencies contain the most power in the wind and have the most
Fig. 4.13 shows coherence (correlation as a function of frequency) using three methods of
variable time delay. TLP F0.003 (vu (t)) means that the wind speed v used in (4.3) is the lidar mea-
surement filtered with a first-order low-pass filter with a cutoff of 0.003 Hz. This is most similar
to the method used in the field tests: the time delay is updated slowly, on the order of minutes,
based on a filtered version of the measured wind speed. Using Tvu (t) , the time delay is updated
instantaneously, with no filter on the measured wind speed. Surprisingly, this improves coherence
compared to TLP F0.003 (vu (t)) , even though by using a time delay with such high frequency variations,
we are subjecting the signal to a swapped order of data points due to slopes being below −1 as
previously described. Fig. 4.14 shows that there is a general trend toward improved coherence
bandwidth as the low-pass filtering cutoff frequency is increased. However, the minimum avail-
able preview time also decreases as the low-pass filtering cutoff frequency is increased, therefore
low-pass filtering should not be completely neglected; instead some filtering should be employed
to ensure availability of a minimum preview time for control. Using td (t) for the time delay gives
82
16
LPF(vu(t))
14
Wind speed (m/s)
13
12
11
10
8
300 350 400 450 500 550 600
Time (s)
Figure 4.12: Low-pass filtered upstream measurement LPF(vu (t)), LPF(vu (t)) stretched and shrunk
by filtering using the variable time delay td (t) plotted in Fig. 4.10, and rotor estimate (vr (t)). Five-
minute portion from Set 1 data.
mscohere( vr(t) , vu(t) w/ variable time delay according to: )
83
1
0.9
0.8
0.7
0.6
Magnitude
0.5
0.4
0.3
TLPF
0.2 (v (t))
0.003 u
Tv (t)
0.1 u
td(t)
0
0 0.05 0.1 0.15 0.2
Frequency (Hz)
Figure 4.13: Magnitude squared coherence between turbine estimate vr (t) and lidar measurement
vu (t) with each of the three different methods of variable time delay listed in the legend. Set 1
data.
the best coherence bandwidth. Although creating the signal td (t) in real time is not possible, this
method estimates an upper limit of coherence bandwidth improvement that is possible to achieve
through improved knowledge of arrival time. Table 4.4 shows the results from these three methods
on the four data sets in terms of coherence bandwidth. We define coherence bandwidth as the pole
location of the first-order low-pass filter whose magnitude squared best fits the magnitude squared
coherence.
On average across all four data sets, our results show a 26% increase in coherence bandwidth
when going from using TLP F0.003 (vu (t)) to td (t), and a 13% increase in coherence bandwidth when
going from using Tvu (t) to td (t). These increases in coherence bandwidth, for example, would allow
a pitch controller to provide an improvement in its combined goal of generator speed regulation
and minimal pitch actuation as quantified in Fig. 3.21 and Fig. 3.22.
4.2.4 Conclusions
Using data from field tests on NREL’s CART2 wind turbine, we have shown how the arrival
time of the lidar measurements varies, we have filtered the measured wind speed signal using a
variable time delay, and we have found an upper limit on the improvement that can be obtained
84
0.09
0.08
0.06
0.05
0.04
Set 1 data
0.03 Set 2 data
Set 3 data
Set 4 data
0.02 −5 −4 −3 −2 −1 0 1
10 10 10 10 10 10 10
LPF cutoff frequency (Hz)
Figure 4.14: Coherence bandwidth between turbine estimate vr (t) and lidar measurement vu (t) with
variable time delay using TLP Fx-axis (vu (t)) (the wind speed v used in (4.3) is the lidar measurement
filtered with a first-order low-pass filter with cutoff frequency shown on the x-axis). The dashed
lines represent no filtering (infinite cutoff frequency).
Set 1. 2. 3. %∆ %∆
TLP F0.003 (vu (t)) Tvu (t) td (t) 1→3 2→3
1 0.0814 0.0846 0.0996 22 18
2 0.0740 0.0993 0.1065 44 7
3 0.0757 0.0726 0.0838 11 15
4 0.0381 0.0431 0.0486 28 13
Mean 0.0673 0.0749 0.0846 26 13
Coherence bandwidths [Hz] between turbine estimate vr (t) and lidar measurement vu (t) with
three methods of variable time delay.
85
through better prediction of arrival time. The data show that we can improve the prediction of
average arrival time by using an induction zone correction when using Taylor’s frozen turbulence
hypothesis. This allows a good prediction of average arrival time, but arrival time can temporarily
A method for accurate real-time prediction of these quick variations in arrival time is outside
the scope of this work. Instead, we used post-processing to obtain these variations in arrival
time from CART2 field test data. Knowing these variations in advance, compared to a method
of predicting arrival time similar to that used in the field tests, would have increased coherence
bandwidth between measured and rotor-estimated wind by 26% on average, and therefore could
have improved control performance as quantified in Chapter 3. This work sets an upper limit of
possible performance improvement because real-time prediction at best will be no more accurate
5.1 Introduction
This chapter discusses two different methods [35, 66] for implementing lidar-based individual
pitch control that were previously designed under separate studies. Both of the two control designs
in this chapter are feedforward controllers that are intended to be added on to a standard feedback
controller as shown in Figure 5.1. Many design features differ between the two methods including
lidar scan patterns, and feedback and feedforward control strategies. The goal of this chapter is to
determine which features are most important in achieving good load reduction and which choices
In this chapter, Section 5.2 describes the turbine model, baseline controller, and simulation
conditions. In Section 5.3, we describe and compare the two controllers. Section 5.4 explores details
on individual pitch (IP) feedforward control. Finally, Section 5.5 outlines conclusions and future
work.
Simulations are performed using a full non-linear turbine model, the NREL 5-MW reference
turbine [9], in the FAST [56] software code. All 16 available degrees of freedom (DOFs) are turned
on in simulations. For all simulations, a 2nd-order pitch actuator model has been added to the
87
Wind
Wind Speed (ahead of turbine)
Preview
LIDAR
DELAY/
FEEDFORWARD
EVOLUTION
Pitch Commands
(3 blades) Outputs
TURBINE
FEEDBACK
5-MW turbine model, with a natural frequency of 1 Hz and a damping ratio of 0.7 [57].
vidual pitch (IP) feedback-only controller was also designed [8], and we will use this as our baseline
controller. In addition to generator speed, the controller inputs are the three out-of-plane blade
root bending moments, and the rotor azimuth, which is used for the multi-blade coordinate [70]
(MBC) (or d-q axis) transformation into horizontal and vertical blade root bending components due
to horizontal and vertical wind shear. The horizontal and vertical components are each controlled
with PI controllers. The collective component is controlled with the same PI feedback control as
the standard collective-pitch controller. An inverse MBC transformation transforms the outputs of
The NREL TurbSim [71, 72] stochastic full-field inflow simulator is used to provide realistic
wind fields for the turbine simulations. Most of the simulations described below are based on
extensive observations taken in the high-plains environment of Southeast Colorado that now has
a large operating wind farm. The Great Plains (GP-LLJ) spectral model available in TurbSim is
The boundary conditions for the TurbSim simulator are shown in Table 5.1. These values
88
Table 5.1: TurbSim Boundary Conditions for 90m Hub Height, 5-MW Turbine, Great Plains
(Lamar, Colorado) Inflow Simulations. uhub – hub-height mean wind speed, RiT L – vertical stability
parameter, αD – vertical power law shear exponent, u∗D – mean friction velocity (shearing stress)
over the rotor disk, coh struct – coherent structures, n.a. – jet instead of power law.
Ensemble ID uhub (m/s) RiT L αD u∗D (m/s) Coh Struct? Jet Height (m)
AR1 13 −0.10 0.077 0.514 N no jet - pwr law
AR2 13 0.02 0.139 0.422 N no jet - pwr law
AR3 13 0.20 0.363 0.135 N no jet - pwr law
AR4 13 0.02 n.a. 0.289 Y 90
AR5 13 0.20 n.a. 0.160 Y 90
are derived from the averages of subpopulations (e.g., AR1) of actual measured wind conditions
associated with 13 m/s (above-rated) hub-height mean wind speeds. The (y-z) grid encompassing
the turbine rotor disk contains 31×31 points of three orthogonal wind components with a sample
rate of 20/second and a total record length of 630 seconds. The first 99 seconds of data from each
simulation are discarded before calculating any performance measures to allow initial transients to
settle out. Thirty-one different realizations of each subpopulation were created, each 630 seconds
long.
In addition to the wind fields shown here, we also mention below the use of a 14 m/s average,
class A wind field. This field is also 630 seconds long and created using TurbSim, but has a greater
turbulence intensity (18%), and uses the IEC Kaimal Normal Turbulence Model (NTM) [59] spectral
feedforward controller A [35] and feedforward controller B [66]. Each feedforward controller was
designed for use in combination with its own individual pitch feedback controller. Both individual
pitch feedback controllers are very similar, both following the description above in Section 5.2.1, but
there are slight differences between the two, including differing gain-scheduling implementations.
(See Appendix B for more on gain-scheduling implementations.) We will refer to these as feedback
89
Table 5.2: Method notation. x, y, and z can each be either A, B, or (none).
Method xyz
x feedforward controller
y feedback controller
z lidar configuration
controllers A [35] and B [66]. The two feedforward controllers were also each developed with their
own lidar configuration, which we refer to as lidar configuration A [35, 73] and lidar configuration
B [66]. To refer to different combinations of feedforward controller, feedback controller, and lidar
configuration, we will use ‘method xyz’ as shown in Table 5.2. Only a few of all possible combina-
tions will be discussed. A summary of the differences between feedforward controllers is shown in
Table 5.3, and a summary of the differences between lidar configurations is shown in Table 5.4.
preview. Its FIR filter coefficients were originally chosen heuristically. They were then optimized by
using a genetic algorithm, trying thousands of variations, and converging on the set of coefficients
with the “best” performance. Performance was based on fatigue load reduction (blade and tower
damage equivalent loads and nacelle accelerations), RMS pitch rate, peak rotor speed, and average
power achieved in a simulation of the nonlinear turbine, with all 16 DOFs, in above rated wind
Figure 5.2.
measurements [73] are taken 65 m ahead of the turbine, one measurement ahead of each blade, at
about 75% span. We will refer to this wind measurement scheme as lidar configuration A. The three
blades’ feedforward signals are controlled separately, each using its own wind speed measurement.
Wind speed measurements are sampled at 80 Hz, matching the simulation rate and controller
update rate for convenience. However, the higher frequencies of the wind speed measurements are
not used, since the FIR filter acts like a lowpass filter, with a cutoff frequency of approximately
0.09 Hz. Each measurement is separately filtered by the FIR filter described above, which has a
DC gain of 1. In series with each FIR filter is a lookup table from steady-state wind speed to blade
pitch. Results of this design (method AAA) are shown in Figure 5.3. Results are also shown here
for a collective pitch variation, where the three wind speed measurements are averaged together,
and the same average is fed in to each blade’s identical feedforward control channel.
Feedforward controller B uses a pulsed lidar model, sweeping a circle in 2.4 seconds, with
12 points at each of five different distances, as shown in Figure 5.4. This trajectory has been
implemented with a lidar system developed and installed on the nacelle of a 5-MW turbine [68]. In
the simulation, effects such as obstruction of the laser beam by the blades, volume measurement,
and mechanical constraints of the scanner from real experimental data were considered to obtain
realistic measurements. For instance, the same loss of about 30% of points could be observed in
the simulation and in the measurements due to obstruction by the moving blades.
where v0 is the horizontal hub-height wind speed and δH and δV are the horizontal and vertical
shear, respectively. These components are found by using a least squares method on the past
12 measurements (the past full circle). We will refer to this wind measurement scheme as lidar
configuration B.
Lidar configuration B is more realistic than lidar configuration A, and it also appears to
have a higher bandwidth for providing accurate measurements: feedforward controller B, when
receiving inputs from lidar configuration B, works best in terms of simulation performance when
lowpass filtered with a cutoff frequency of 0.06 rad/m multiplied by wind speed in m/s. But when
92
Figure 5.3: Turbine loads using individual pitch baseline feedback control alone [8] (method A ),
and with added feedforward controller A (method AAA), in individual pitch and collective pitch
versions. Percents displayed are the average of the eight bars. Feedforward controller A reduces
overall loads by 4.9% in the individual pitch version. The results are averages from FAST simula-
tions across 155 wind files representing AR1 through AR5.
200
150
vertical [m]
100
50
0
100
0
150 200
50 100
−100 0
−50
horizontal [m] upwind [m]
Figure 5.4: Lidar configuration B pulsed lidar scanning pattern. Figure courtesy of D. Schlipf.
93
receiving inputs from lidar configuration A, the spatial cutoff frequency is reduced to 0.04 rad/m.
Lidar configuration B likely provides a more accurate wind speed preview because measurements
are taken at five different radii instead of just one. This was originally done with pulsed lidar, with
all 5 measurements taken at the same time. A CW lidar could produce equivalent measurements
by refocusing between the five points, one after the other, if the CW lidar’s sample rate was at
Lidar configuration A detects the blade effective wind speed needed by feedforward con-
troller A. On the contrary, lidar configuration B is trying to capture the rotor-effective wind char-
acteristics required by feedforward controller B. The correlation of one single measurement ahead
of each blade with each blade’s effective wind speed should be greater than the correlation of those
three point measurements with the effective wind characteristics of the whole rotor disk. Therefore
feedforward controller A combined with lidar configuration A can compensate loads above the the
once per revolution (1P) frequency, if a good correlation with the blade effective wind speed can be
obtained above the 1P frequency. To capture comparable correlation with the rotor-effective wind
characteristics for feedforward controller B, more measurement points are necessary. Lidar config-
uration B was designed to improve the correlation with the rotor-effective wind characteristics of
wind speed and horizontal and vertical shears. Nonetheless, using lidar configuration A also shows
Lidar configuration A has the advantage that it retains more information on how the wind
speed varies with azimuth. Blade loads can be caused by both wind speed differences over azimuth
and wind speed differences in the x-direction (up/downwind). For changes in the x-direction, we
estimate that our bandwidth for good measurements is around 0.1 or 0.2 Hz (at 13 m/s). For
changes over azimuth however, our bandwidth depends on the rate at which our lidar samples
and spins: the number of lidar measurements per lidar revolution. This could easily translate to
higher than 0.1 or 0.2 Hz as the blade sees it. For example, a low-level jet can cause a high wind
speed at hub height and lower wind speeds at both the top and bottom of the rotor plane. As
the blades spin through this, they see a 2P, or 0.4 Hz load. We will show below that when using
94
B takes enough measurements to capture this low level jet, but the information is lost when these
measurements are simplified into the 1P d0HV components [70] (average wind speed and vertical
Using only the 1P d0HV components of wind measurements as in lidar configuration B can be
called cyclic feedforward. The advantage of cyclic feedforward is that it can be used with different
lidar types and scan patterns, and is more independent of the preview time/scan distance. It should
work for all wind speeds and can easily be adapted in real applications. To combine the advantages
to 1P loads. This would involve transforming the ring of measurements into 2P components in
After capturing the d0HV wind measurements, feedforward controller B first delays them so
that the remaining preview time is just enough to compensate for the phase delay of the actuator,
lowpass filter, and the turbine. The measurements are then sent to the feedforward controller,
which is simply a lowpass filter followed by a set of static gains. The filter cutoff frequency is
determined based on the correlation between the lidar measurements and the turbine effective
wind speed. Then for v0 , there is a lookup table from steady-state wind speed to steady-state pitch
to get u0 , the average feedforward signal. The horizontal and vertical wind shears, δH and δV ,
are each multiplied by a scalar that is optimized to yield the best uH and uV blade pitch control
components that cancel the wind disturbance. The blade pitch components uH and uV are added to
the feedback controller in the multi-blade coordinate (MBC) domain. This simpler design has the
benefit of being easily tunable, which is very important when dealing with modeling uncertainty.
Originally, simulation for method BBB was done with a stochastic full-field wind (23 × 23
grid, ∆t = 0.25s) with a mean velocity of 16 m/s and a turbulence intensity of 18%. The results
presented in Figure 5.5 show greatly reduced tower and blade loads. When originally studied
separately, with different wind fields, different performance metrics, and slightly different baseline
feedback controllers, method BBB vs method B appeared to have significantly greater load
95
To specifically compare the two feedforward controllers, we simulated method AAA and
method BAA in the same wind fields, using the same performance metrics. Results are shown in
Figure 5.6 for a less turbulent 13 m/s wind field, and in Figure 5.7 for a more turbulent 14 m/s wind
field. Both feedforward controllers perform similarly, with feedforward controller A showing slightly
more load reduction in the 14 m/s wind, and feedforward controller B showing slightly more load
reduction in the 13 m/s wind. The more turbulent 14 m/s wind allows for more load reduction,
with both controllers averaging about 9% overall, versus 5% in the 13 m/s wind. Blade root and
tower base load reductions were the original performance metrics of feedforward controller B, and
looked at individually, these two measures consistently have greater load reduction than the average
of all 8 bars. Because performance for feedforward controller A was originally measured using this
8-bar average, this accounts for some of the originally perceived differences in the performance of
the two controllers. Both controllers improve rotor speed regulation compared to baseline as shown
by the lower peak rotor speeds, and, in the more turbulent 14 m/s wind, they also very slightly
improve power capture. Feedforward controller B has a lower RMS pitch rate than A but does not
regulate rotor speed quite as well as A. Feedforward controller A reduces tower top DEL more than
feedforward controller B. As will be shown in Figure 5.10 below, much of feedforward controller A’s
tower top DEL reduction is at 0.6 Hz, which translates to 0.4 Hz in the rotating (blade) coordinate
system [70]. Therefore the better tower top load reduction by feedforward controller A may be
due to the increase in its magnitude at 0.4 Hz, which is shown in Figure 5.8 and discussed further
below.
This same pitch-rate activity vs. rotor-speed error tradeoff that is appearing for feedforward
control also appears when comparing feedback controllers A and B. Feedback controller B was
designed to use reduced feedback gains when combined with a feedforward controller. This greatly
reduces RMS pitch rate and somewhat increases peak rotor speed. In some cases the reduced
feedback gains also lead to reduced loads. Overall this feedback gain reduction appears beneficial
because it reduces pitching action and allows the feedforward controller more control authority.
96
6
10
5
10
tower base pitching moment
9
10
[kNm2/Hz]
8
10
7
10
0
10
−2
10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
frequency [Hz]
Figure 5.5: Power spectral densities for method BBB under its original simulation conditions.
Figure courtesy of D. Schlipf.
97
Figure 5.6: Turbine loads using individual pitch baseline feedback control alone ( A ) [8], with added
individual pitch feedforward controller A, and with added individual pitch feedforward controller B.
The rated rotor speed is 12.1 rpm, and the rated power is 5000.0 kW. x-accel, y-accel, and z-accel
are respectively fore-aft, side-to-side, and up-down nacelle accelerations. The wind field used is
AR2 s13, from the Great Plains set, at 13 m/s average wind speed.
98
Figure 5.7: Turbine loads using individual pitch baseline feedback control alone ( A ) [8], with added
individual pitch feedforward controller A, and with added individual pitch feedforward controller B.
The rated rotor speed is 12.1 rpm, and the rated power is 5000.0 kW. x-accel, y-accel, and z-accel
are respectively fore-aft, side-to-side, and up-down nacelle accelerations. The wind field used is the
IEC Kaimal NTM spectral model with a Class A turbulence level, at 14 m/s average wind speed.
99
Figure 5.8 shows Bode plots of feedforward controllers A and B. Feedforward controller A
was designed for 13 m/s average wind speed. Feedforward controller B was designed for the full
range of region 3 wind speeds, and its cutoff frequency and preview time vary with wind speed.
Here the design for the 13 m/s average wind speed is shown. The two feedforward controllers, while
designed using very different methods, look strikingly similar. Both designs drop to 50% magnitude
at about 0.1 Hz. This has been roughly estimated to be the highest frequency at which we can
assume that the wind we measure with lidar matches the wind speed we see at the turbine, at an
average wind speed of 13 m/s [63, 74]. In addition to the filters shown in Figure 5.8, each controller
also contains a lookup table with gains that vary depending on wind speed. Both gain-scheduling
Figure 5.9 shows the controllers’ phase converted to time delay according to τφ = −φ/360◦ /f ,
where τφ is the time delay in seconds, φ is the phase in degrees, and f is the frequency in Hz. The
phase and time delay matter mostly below the cutoff frequency, where the magnitude is above
about 0.5, and here the controllers differ somewhat. Below 0.1 Hz, controller A has a time delay
of about 5 seconds. Controller A is also designed to receive its wind preview input 5 seconds in
advance, so all of the available preview is used by the controller. Controller B, on the other hand,
has a time delay of about 2.9 seconds at frequencies below 0.1 Hz. Controller B is designed to
receive its wind preview input 3.9 seconds in advance, so an extra 1.0 seconds is still available to
compensate for the additional delays caused by the actuator and turbine. This time value is also
tunable in B.
Here bi is the pitch feedforward command to blade i, where i =1, 2, or 3. A represents the transfer
function for controller A plotted in Figure 5.8, and wi is the wind speed at blade i. Controller B is
100
1
A
B
magnitude [−]
0.5
0
−3 −2 −1 0
10 10 10 10
frequency [Hz]
−90
phase [deg]
−180
−270
−360
−3 −2 −1 0
10 10 10 10
frequency [Hz]
Figure 5.8: Bode plot of feedforward controllers A and B (transfer functions A and B respectively),
excluding scheduled gains.
5
time delay [s]
3 A
B
0
−3 −2 −1 0
10 10 10 10
frequency [Hz]
Here bc is the average, bv is the vertical component, and bh is the horizontal component of the feed-
forward pitch command to the blades. B represents the transfer function for Controller B plotted in
Figure 5.8, and wc is the average, wv is the vertical component, and wh is the horizontal component
of the wind speed at the blades. These two methods (applying the controller in individual vs MBC
coordinates) are not equivalent for the vertical and horizontal components of the signal, because
frequencies change between the rotating and non-rotating frames. The two methods are equivalent
only for the collective component (average) of the signal. This is shown in the equations below,
if ~b = TM
−1 ~
BC ATM BC w
~ (controller applied in MBC coordinates)
In summary, feedforward controllers A and B, while designed using very different methods: A
with a genetic algorithm optimizing over repeated simulations and B based on correlations between
lidar measurements and rotor-effective wind, end up with very similar Bode plots (both lowpass
filters with cutoff frequencies matching within 15% of each other) and also relatively compara-
ble simulation results. Feedforward controller B is simpler to design and more tunable, which is
beneficial because there will always be modeling uncertainty. It is a good idea to explore using
feedforward at frequencies higher than 1P, as was attempted in controller A, to lead to even better
performance. This higher frequency operation may be the reason for controller A’s greater load
reduction in the tower top DEL. Reducing feedback gains when feedforward is present is a good
102
method to reduce pitch rate, but there is some tradeoff in the form of increasing rotor speed error.
Lidar configuration B is more realistic and provides more accurate wind speed estimates because it
takes more measurements, covering more of the rotor plane. A 1P MBC transformation is conve-
nient in converting a large set of lidar measurements to usable controller inputs, but it loses some
useful spatial information as in the example of the low-level jet. In more turbulent wind fields, the
addition of feedforward control provides a greater percentage load reduction than in less turbulent
wind. Finally, blade root and tower base load reduction is often greater than load reduction in
other areas of the turbine structure, so one should be careful to compare controllers under the same
performance metrics.
When wind speeds are constant in time but varying over the rotor plane, loads appear on
the blades only at the once per revolution frequency (1P)(0.2 Hz) and at its harmonics (2P, 3P,
etc.). It was therefore theorized that individual pitch (IP) control is only useful at these distinct
frequencies. To test this theory, we look at loads as a function of frequency for both collective pitch
Figure 5.3 shows that Controller A reduces the tower top DEL, blade root DEL, and RMS
configuration. These three loads are plotted as a function of frequency in Figures 5.10, 5.11, and
5.12. On the non-rotating parts of the structure, the 1P, 2P, and 3P blade loads described above
respectively transform to loads at 0P (DC), 3P, and none, through an MBC transformation [70].
Figure 5.10 shows that tower top fore-aft loads are reduced by the IP feedforward at below 0.1 Hz,
increased from 0.15 to 0.3 Hz, and reduced again between 0.4 and 0.7 Hz, when compared to CP
feedforward. This does fit with load reduction near 0P and 3P as expected. Figure 5.11 shows that
the IP action reduces blade root loads from 0.23 to 0.45 Hz. This includes 2P, but not 1P and 3P.
The IP feedback controller is designed to work at 1P. This explains why 1P loads are lower than
2P loads. It also may explain why IP feedforward does not help at 1P: the IP feedback may have
103
done it so well that there is less remaining for IP feedforward to do. Finally, Figure 5.12 shows
that IP feedforward makes fore-aft nacelle acceleration worse from 0.2 to 0.3 Hz, and very slightly
better from 0.32 to 0.65 Hz. This follows a similar pattern to tower top loads, except for the part
near DC, since a DC fore-aft nacelle acceleration does not make much sense for a turbine that does
In addition to the frequencies of the loads, it is also useful to look at the frequencies of
actuator operation. Figure 5.13 shows that IP feedforward, when compared to CP feedforward, only
increases pitching action at 2P and 4P (0.4 Hz at 0.8 Hz). This does not imply that IP feedforward
is not doing anything at other frequencies; the blades can switch to pitching independently without
necessarily increasing the amount that they pitch. Figure 5.13 also shows that both CP and IP have
reduced pitching action below 0.1 Hz compared to feedback alone. Generator speed error (shown
indirectly in Fig. 5.14) and blade loads are also greatly reduced below 0.1 Hz, so it is encouraging
that in this frequency range, preview allows us to have less blade pitch action, less generator speed
Figure 5.14 shows generator torque versus frequency. Both CP and IP feedforward greatly
reduce generator torque action below 0.11 Hz (because they are reducing generator speed error
here). From 0.2 Hz up to 1 Hz, IP is slightly lower than both CP and feedback-only, except at
the spike at 0.32 Hz where both CP and IP are higher than feedback only. The plot of generator
speed (not shown) looks almost exactly the same as this generator torque plot, except in different
units. This is because the region 3 generator torque controller is simply set up so that its torque
command is inversely proportional to generator speed (after lowpass filtering with a 0.25 Hz cutoff).
However, upon zooming in on the generator speed plot, there is no difference between CP and IP
feedforward from 0.2 to 1 Hz, except that exactly at the 0.32 Hz spike, IP is lower than CP. The
spike shown in the generator speed plot at 0.32 Hz corresponds to the first tower fore-aft mode and
the first tower side-to-side mode. This spike appears in the plots of generator torque, generator
speed, drivetrain torsion, tower bending fore-aft and side-to-side, and nacelle motion in x, y, and
z directions. It is important because this one frequency makes up the overwhelming majority of
104
nacelle side-to-side motion, which is often a problem, showing increased loads when feedforward
control is added, because it is very lightly damped (1% damping ratio). In this set of data, IP does
slightly better than CP at reducing nacelle side-to-side velocity at 0.32 Hz, but both are worse than
feedback alone.
These plots over frequency have confirmed that IP feedforward load reductions all happen at
or near the 1Protating =0Pnonrotating or 2Protating =3Pnonrotating frequencies. Nonrotating loads are
reduced at both frequencies, and blade root loads are only reduced at 2P because IP feedback-only
already does a good job of reducing them at 1P. This implies that we might be able to add IP
feedback to the other loads to eliminate the need for individual pitch feedforward. We have also
gained some insight into the cause of the increased nacelle side-to-side motion that often occurs
when feedforward is added by noting the frequency at which it occurs and the many other measures
that also contain a spike at that frequency. Finally, we have seen that below 0.1 Hz, feedforward
control reduces blade pitch actuation, generator speed error, and blade loads, all at the same time.
We have compared two feedforward controllers and lidar configurations, and studied individ-
ual pitch feedforward load reductions as a function of frequency. Future work is needed to test
the idea of combining the best features of the two lidar configurations into a configuration that
is realistic with broad coverage of the rotor plane like lidar configuration B and also addresses
2P loads like lidar configuration A. Future work is also needed to determine how effective it is
to use additional feedback from additional turbine load sensors instead of adding individual pitch
feedforward control.
105
Figure 5.10: Tower top fore-aft pitching moment as a function of frequency. Average FFT mag-
nitude over AR1 s12 through AR1 s27. (This smaller subset of the wind fields was used due to
computer memory limitations.)
106
Figure 5.12: Fore-aft nacelle acceleration as a function of frequency. Average FFT magnitude over
AR1 s12 through AR1 s27.
108
Figure 5.13: Blade pitch angle as a function of frequency. Average FFT magnitude over AR1 s12
through AR1 s27.
109
Figure 5.14: Generator torque as a function of frequency. Plot is also representative of relative
magnitudes of generator speed as a function of frequency because generator torque is inversely
proportional to generator speed. Average FFT magnitude over AR1 s12 through AR1 s27.
Chapter 6
It has recently become feasible to use lidar to capture preview measurements of the wind
approaching a wind turbine, and this has opened up a whole new realm of possibilities for preview
control. One way that we can use these measurements is in a Region 3 blade pitch control system.
These preview measurements allow improved compensation for wind disturbances, resulting in
improved Region 3 blade pitch control performance. The major design challenges involved in
incorporating preview measurements into a blade pitch controller include imperfect correlation
between the preview measurement and the wind that arrives at the turbine rotor, a varying and
shows that if a control design assumes perfect measurement coherence, then the lidar mea-
surements allow improved performance at a given frequency only when the magnitude of
the measurement error is less than the magnitude of the wind disturbance.
• An H2 optimal control design process that accounts for the both the measurement coherence
and available preview time is explained. Simulation results show significantly reduced pitch
actuation, improved generator speed regulation, and reduced structural loads, compared
to several different baseline cases. One of these cases is a simple feedforward controller
(LPF FF) that has been field tested with good results. Another of these baseline cases is
111
a feedback-only controller that was created using the same optimal control design process,
only without lidar measurements available. This comparison allows a true evaluation of
• Linear-model-based predictions quantify how the benefit of lidar changes with measurement
• When generator speed regulation is the only control goal, at least one second of preview
time is useful for pitch control, with the majority used to cancel the delay introduced by
preview time increases with decreasing coherence bandwidth. When an optimal controller
penalizes pitch rate in addition to generator speed error, five or more seconds of preview
may be useful.
• An analysis of available preview time from CART2 field experiment data shows evidence
that the induction zone is responsible for the average preview time being greater than the
amount predicted by Taylor’s frozen turbulence hypothesis. The field data analysis also
shows that, in order to maximize measurement coherence, the amount of variable time delay
filtering the wind speed signal does increase the minimum preview time available, so the
choice of amount of filtering should be made considering the tradeoff between measurement
• A comparison of two different lidar configurations and individual pitch feedforward and
feedback controllers shows the benefits of each method, with feedforward controller A re-
sulting in better rotor speed regulation and feedforward controller B resulting in lower
RMS pitch rate, lidar configuration A retaining more information on how the wind speed
varies with azimuth and lidar configuration B capturing more samples over the rotor plane
to better represent the overall rotor-effective wind speed. A breakdown of load reductions
112
• Using the same H2 optimal control design method with vertical and horizontal
controller. We discussed H2 collective pitch control, where each blade receives the same
pitch command based on rotor-averaged wind speed measurements and generator speed
error. Lidar-based individual pitch control could be implemented using additional cyclic
pitch commands, allowing the turbine to react to previews of changing wind shear or yaw
error.
Gain scheduling could be improved by solving for linear H2 controllers at multiple operating
points, perhaps every one or two m/s throughout Region 3, and transitioning between them
as the average wind speed varies. Additionally, the stability of this gain scheduling strategy
• Field testing. When a turbine with a mounted lidar, possibly the 600-kW CART2 or
CART3 turbine at NREL, is available for field testing, the first step would be to determine
the measurement coherence and minimum available preview time from the current lidar
configuration. Then the same control design procedure that is explained in Chapter 3 for
the NREL 5-MW turbine could be followed for the new turbine model at several Region
3 operating points, for the given preview time and measurement coherence. The resulting
to be field tested.
• Direct penalization of structural loads. Feedback signals from blade and tower strain
gauges could be included along with the generator speed error Ω as inputs to the feedback
part of the H2 optimal controller. In addition, these signals could be included as compo-
113
nents of y, the performance output which defines the cost function to be minimized. This
should allow the control designer to directly trade off blade and tower loads in addition to
generator speed error and pitch actuation. However, the additional complexity makes it
more difficult to reliably find a valid solution to the optimization problem, and it may be
necessary to explore other methods of posing and/or solving the optimization problem.
augmented plant with a model of measurement coherence and preview time that is explained
in Chapter 3 can also be used in solving for an H∞ controller, using Hazell’s [55] solution
more direct penalization of the standard deviations of outputs, whereas H∞ control offers
improves performance. Appendix B explains the differences between two different gain
scheduling methods that were originally assumed to be equivalent: Integrate First and
Multiply First. The effective gains resulting from Multiply First are more straightforward,
and we recommend this method when feedforward is included in a control system. The
effective gains resulting from Integrate First are not straightforward but can be corrected
to result in desired effective gains. After correction, Integrate First consistently shows
improved performance over Multiply First that was not predicted by simplified models,
• Further analysis of maximum useful preview time for outputs of interest other
than generator speed error. Section 4.1 discusses how preview time is useful in canceling
the time delays introduced by the lowpass filter, actuator, and the turbine itself. The
amount of time delay introduced by the turbine itself was shown for the case where only
the generator speed error output is considered. We suggest a similar analysis for other
outputs of interest.
Bibliography
[2] International Electrotechnical Commission. IEC 61400-13 Ed.1: Wind turbine generator sys-
tems - Part 13: Measurement of mechanical loads. 2001.
[3] Harris M, Hand M, Wright A. Lidar for turbine control. National Renewable Energy Labora-
tory, Golden, CO, Tech. Rep., NREL/TP-500-39154, 2006.
[4] Wiser R, Bolinger M. 2013 wind technologies market report. U.S. Department of Energy,
Energy Efficiency and Renewable Energy, 2014.
[5] American Wind Energy Association. Wind was largest source of new electricity
in 2014, congress still must provide long-term policy certainty. March 2015. URL
http://www.awea.org/MediaCenter/pressrelease.aspx?ItemNumber=7294, [Online; ac-
cessed 8-May-2015].
[6] Broehl J, Labastida RR, Hamilton B. World wind energy market update 2015. Navigant Re-
search, 2015.
[7] Manwell JF, McGowan JG, Rogers AL. Wind Energy Explained: Theory, Design and
Application. 2 edn., Wiley, 2010.
[8] Bossanyi EA. Individual blade pitch control for load reduction. Wind Energy 2003; 6:119–128.
[9] Jonkman J, Butterfield S, Musial W, Scott G. Definition of a 5-MW reference wind turbine
for offshore system development. National Renewable Energy Laboratory, Golden, CO, Tech.
Rep., NREL/TP-500-38060, 2009.
[10] Simley E, Pao L, Frehlich R, Jonkman B, Kelley N. Analysis of light detection and ranging
wind speed measurements for wind turbine control. Wind Energy 2014; 17(3):413–433.
[11] Simley E. Wind speed preview measurement and estimation for feedforward control of wind
turbines. PhD Thesis, University of Colorado Boulder 2015.
[12] Østergaard K, Brath P, Stoustrup J. Estimation of effective wind speed. Proc. The Science of
Making Torque from Wind, 2007.
115
[13] Simley E, Pao LY. Evaluation of a wind speed estimator for effective hub-
height and shear components. Wind Energy 2014; doi:10.1002/we.1817. URL
http://dx.doi.org/10.1002/we.1817.
[14] Bossanyi EA, Fleming PA, Wright AD. Field test results with individual pitch control on the
NREL CART3 wind turbine. Proc. AIAA Aerospace Sciences Meeting, Nashville, TN, 2012.
[15] Bottasso CL, Croce A, Savini B, Sirchi W, Trainelli L. Aero-servo-elastic modeling and control
of wind turbines using finite-element multibody procedures. Multibody System Dynamics 2006;
16(3):291–308.
[16] Kumar AA, Stol K. Scheduled model predictive control of a wind turbine. Proc. European
Wind Energy Conference, 2010.
[17] Friis J, Nielsen E, Bonding J, Adegas F, Stoustrup J, Odgaard P. Repetitive model pre-
dictive approach to individual pitch control of wind turbines. Proc. IEEE Conference on
Decision and Control and European Control Conference (CDC-ECC), 2011; 3664–3670, doi:
10.1109/CDC.2011.6160948.
[18] Schuler S, Schlipf D, Cheng PW, Allgower F. `1 -optimal control of large wind turbines.
Control Systems Technology, IEEE Transactions on July 2013; 21(4):1079–1089, doi:
10.1109/TCST.2013.2261068.
[19] Wright A, Stol K, Fingersh L. Progress in implementing and testing state-space controls for
the controls advanced research turbine. Proc. 24th ASME Wind Energy Symposium, Reno,
NV, 2005; 88–100.
[20] Wright AD, Fingersh LJ. Advanced control design for wind turbines. National Renewable
Energy Laboratory, Golden, CO, Technical Report, NREL/TP-500-42437, 2008.
[21] Stol K, Balas M. Full-state feedback control of a variable-speed wind turbine: a comparison
of periodic and constant gains. Journal of solar energy engineering 2001; 123(4):319–326.
[22] Stol KA, Fingersh LJ. Wind turbine field testing of state-space control designs. National Re-
newable Energy Laboratory, Golden, CO, Subcontractor Report, NREL/SR-500-35061, 2004.
[23] Frost SA, Balas MJ, Wright AD. Adaptive control of a utility-scale wind turbine operating in
region 3. Proc. AIAA Aerospace Sciences Meeting, Orlando, FL, 2009.
[24] Magar KT, Balas MJ, Frost S. Direct adaptive control for individual blade
pitch control of wind turbines for load reduction. Journal of Intelligent
Material Systems and Structures 2015; doi:10.1177/1045389X14566527. URL
http://jim.sagepub.com/content/early/2015/01/20/1045389X14566527.abstract.
[25] Geyler M, Caselitz P. Individual blade pitch control design for load reduction on large wind
turbines. Proc. European Wind Energy Conference, Milano, Italy, 2007; 82–86.
[26] Grimble M. Horizontal axis wind turbine control: Comparison of classical, LQG and H∞
designs. Dynamics and Control 1996; 6(2):143–161.
[27] Long Y, Hanba S, Yamashita K, Miyagi H. Sliding mode controller design via H∞ theory
for windmill power systems. Proc. IEEE International Conference on Systems, Man, and
Cybernetics, vol. 1, IEEE, 1999; 56–61.
116
[28] Kumar A, Stol K. Simulating MIMO feedback linearization control of wind turbines using
FAST. Proc. AIAA/ASME Wind Energy Symposium, 2008.
[29] Lescher F, Zhao JY, Borne P. Switching LPV controllers for a variable speed pitch regulated
wind turbine. International Journal of Computers, Communications & Control 2006; 1(4):73–
84.
[30] Schlipf D, Fleming P, Haizmann F, Scholbrock AK, Hofsäß M, Wright A, Cheng PW. Field
testing of feedforward collective pitch control on the CART2 using a nacelle-based lidar scanner.
Proc. The Science of Making Torque from Wind, Oldenburg, 2012.
[31] Scholbrock AK, Fleming PA, Fingersh LJ, Wright AD, Schlipf D, Haizmann F, Belen F. Field
testing LIDAR based feed-forward controls on the NREL controls advanced research turbine.
Proc. AIAA Aerospace Sciences Meeting, Grapevine, TX, 2013.
[32] Schlipf D. Lidar-assisted control concepts for wind turbines. PhD Thesis, Universität Stuttgart
2015.
[33] Kristalny M, Madjidian D, Knudsen T. On using wind speed preview to reduce wind turbine
tower oscillations. Control Systems Technology, IEEE Transactions on 2013; 21(4):1191–1198.
[34] Wang N. Lidar-assisted feedforward and feedback control design for wind turbine tower load
mitigation and power capture enhancement. PhD Thesis, Colorado School of Mines 2013.
[35] Dunne F, Pao LY, Wright AD, Jonkman B, Kelley N, Simley E. Adding feedforward blade pitch
control for load mitigation in wind turbines: Non-causal series expansion, preview control, and
optimized FIR filter methods. Proc. AIAA Aerospace Sciences Meeting, Orlando, FL, 2011.
[36] Tomizuka M, Fung D. Design of digital feedforward/preview controllers for processes with
predetermined feedback controllers. ASME J. Dyn. Sys., Meas. & Ctrl. Dec 1980; 102:218–
225.
[37] Pace A, Johnson KE, Wright A. Lidar-based extreme event control to prevent wind turbine
overspeed. Proc. AIAA Aerospace Sciences Meeting, Nashville, TN, 2012.
[38] Laks J, Simley E, Pao L. A spectral model for evaluating the effect of wind evolution on wind
turbine preview control. Proc. American Control Conference, Washington, D.C., 2013.
[39] Laks JH. Preview scheduled model predictive control for horizontal axis wind turbines. PhD
Thesis, University of Colorado Boulder 2013.
[40] Körber A, King R. Model predictive control for wind turbines. Proc. European Wind Energy
Conference, 2010.
[41] Soliman M, Malik O, Westwick D. Multiple model MIMO predictive control for variable speed
variable pitch wind turbines. Proc. American Control Conference, IEEE, 2010; 2778–2784.
[42] Mirzaei M, Soltani M, Poulsen N, Niemann H. Model predictive control of wind turbines using
uncertain LIDAR measurements. Proc. American Control Conference, Washington, D.C., 2013.
[43] Schlipf D, Schlipf DJ, Kühn M. Nonlinear model predictive control of wind turbines using
lidar. Wind Energy 2013; 16(7):1107–1129.
117
[44] Kragh KA, Hansen MH. Individual pitch control based on local and upstream inflow measure-
ments. Proc. AIAA Aerospace Sciences Meeting, Nashville, TN, 2012.
[45] Ozdemir A, Seiler P, Balas G. Fundamental limitations of preview for wind turbine control.
Proc. AIAA Aerospace Sciences Meeting, Nashville, TN, 2012.
[46] Simley E, Pao LY. A longitudinal spatial coherence model for wind evolution based on large-
eddy simulation. Proc. American Control Conference, Chicago, IL, 2015.
[47] Simley E, Pao L. Reducing LIDAR wind speed measurement error with optimal filtering. Proc.
American Control Conference, Washington, D.C., 2013.
[48] Simley E, Pao LY, Gebraad P, Churchfield M. Investigation of the impact of the upstream
induction zone on LIDAR measurement accuracy for wind turbine control applications using
large-eddy simulation. Proc. The Science of Making Torque from Wind, Lyngby, Denmark,
2014.
[49] Schlipf D, Cheng PW, Mann J. Model of the correlation between lidar systems and wind
turbines for lidar-assisted control. Journal of Atmospheric and Oceanic Technology 2013;
30(10):2233–2240.
[50] Bossanyi EA, Kumar A, Hugues-Salas O. Wind turbine control applications of turbine-
mounted lidar. Proc. The Science of Making Torque from Wind, Oldenburg, 2012.
[51] Dunne F, Pao LY. Benefit of wind turbine preview control as a function of measurement
coherence and preview time. Proc. American Control Conf., Washington, D.C., 2013; 629–634.
[52] Dunne F, Schlipf D, Pao LY, Wright AD, Jonkman B, Kelley N, Simley E. Comparison of
two independent lidar-based pitch control designs. Proc. AIAA Aerospace Sciences Meeting,
Nashville, TN, 2012.
[53] Dunne F, Pao LY, Schlipf D, Scholbrock AK. Importance of lidar measurement timing accuracy
for wind turbine control. Proc. American Control Conf., Portland, OR, 2014; 3716–3721.
[54] Dunne F, Pao LY, Wright AD, Jonkman B, Kelley N. Combining standard feedback con-
trollers with feedforward blade pitch control for load mitigation in wind turbines. Proc. AIAA
Aerospace Sciences Meeting, Orlando, FL, 2010.
[55] Hazell A. Discrete-time optimal preview control. PhD Thesis, Imperial College University of
London February 2008.
[56] Jonkman J, Buhl ML. FAST user’s guide. NREL/EL-500-38230, NREL Report: National
Renewable Energy Laboratory, Golden, CO, 2005.
[57] Schlipf D, Kühn M. Prospects of a collective pitch control by means of predictive disturbance
compensation assisted by wind speed measurements. Proc. German Wind Energy Conference
DEWEK, Bremen, 2008.
[58] Green M, Limebeer DJ. Linear Robust Control. Prentice Hall: Englewood Cliffs, New Jersey,
1995.
[59] International Electrotechnical Commission. IEC 61400-1 Ed.3: Wind turbines - Part 1: Design
requirements. 2005.
118
[60] Matlab, Simulation and Model-Design Software Package. The MathWorks, Inc.: Natick, MA.
URL http://www.mathworks.com.
[61] Bode HW. Network analysis and feedback amplifier design. van Nostrand, 1945.
[62] Taylor GI. The spectrum of turbulence. Proc. Royal Society of London, Series A, Mathematical
and Physical Sciences, vol. 164, 1938; 476–490.
[64] Rigney BP, Pao LY, Lawrence DA. Nonminimum phase dynamic inversion for settle time
applications. IEEE Trans. Control Systems Technology Sept 2009; 17(5):989–1005.
[65] Widrow B, Walach E. Adaptive Inverse Control - a Signal Processing Approach. Wiley: Hobo-
ken, NJ, 2008.
[66] Schlipf D, Schuler S, Grau P, Allgöwer F, Kühn M. Look-ahead cyclic pitch control using lidar.
Proc. The Science of Making Torque from Wind, Heraklion, Greece, 2010.
[67] Bossanyi E, Wright A, Fleming P. Controller field tests on the NREL CART2 turbine. National
Renewable Energy Laboratory, Golden, CO, Tech. Rep., NREL/TP-5000-49085, 2010.
[68] Rettenmeier A, Bischoff O, Hofsäß M, Schlipf D, Trujillo JJ, Kühn M. Wind field analyses
using a nacelle-based lidar system. Proc. European Wind Energy Conference, Warsaw, Poland,
2010.
[69] Medici D, Ivanell S, Dahlberg JÅ, Alfredsson PH. The upstream flow of a wind turbine:
Blockage effect. Wind Energy 2011; 14(5):691–697, doi:10.1002/we.451.
[70] Bir G. Multi-blade coordinate transformation and its application to wind turbine analysis.
Proc. AIAA Aerospace Sciences Meeting, Reno, NV, 2008.
[71] Kelley ND, Jonkman BJ. Overview of the turbsim stochastic inflow turbulence simulator:
Version 1.21 (revised Feb. 1, 2007). National Renewable Energy Laboratory, Golden, CO,
Tech. Rep., NREL/TP-500-41137, 2007.
[72] Jonkman BJ. Turbsim user’s guide. National Renewable Energy Laboratory, Golden, CO,
Tech. Rep., NREL/TP-500-46198, 2009.
[73] Simley E, Pao LY, Frehlich R, Jonkman B, Kelley N. Analysis of wind speed measurements us-
ing coherent lidar for wind preview control. Proc. AIAA Aerospace Sciences Meeting, Orlando,
FL, 2011.
[74] Simley E, Pao LY, Kelley N, Jonkman B, Frehlich R. Lidar wind speed measurements of
evolving wind fields. Proc. AIAA Aerospace Sciences Meeting, Nashville, TN, 2012.
[75] Hansen MH, Henriksen LC. Basic DTU wind energy controller. Department of Wind Energy,
Report Number: DTU Wind Energy E-0018, 2013.
Appendix A
wa = Kn2 (A.2)
p
|H| = |K|2 (1 − |L|2 ) (A.3)
p
|L| = coh(f ) (A.4)
where equations are represented in the frequency domain; wm and wa are signals (in this thesis
they represent the measured and actual wind); H, L, and K are linear systems; n1 and n2 are
2
γam = coh(f ) (A.5)
2
where γam is the magnitude squared coherence between wa and wm ; Smm is the power spectral
Lemma 1: If
wm = Hn1 + Lwa
120
Lemma 2: If
p
|H| = |K|2 (1 − |L|2 )
wm = Hn1 + Lwa
wa = Kn2
where n1 and n2 are uncorrelated zero-mean white noise with equal power |n1 |2 = |n2 |2 , then
Smm = Saa
Combining the two lemmas, we see that given (A.1) through (A.3), it is true that Smm = Saa
p
2 = |L|. If additionally given (A.4), we see that γ 2 = coh(f ), and our proof is complete.
and γam am
121
p
2
γam
s
|Sam |2
= definition of coherence
Saa Smm
wa wm ∗
=√
S S
aa mm
∗
w
a (Hn 1 + Lw )
a
= √ plug in wm
Saa Smm
wa n H ∗ + wa wa∗ L∗
∗
1
= √ distribute
Saa Smm
wa wa∗ L∗
=√ phase of n1 is random, adds 0 on average
Saa Smm
wa wa∗ L∗
=r
Smm
Saa Saa
S
aa
wa wa∗ L∗
= r
Smm
Saa
Saa
r
wa wa∗ L∗ Saa
=
wa wa∗ Smm
r
wa wa∗ L∗ Saa
= ∗
wa wa Smm
r
Saa
= |L∗ | L does not change between realizations
S
r mm
Saa
= |L|
Smm
122
Smm
∗
= wm wm
= |H|2 |n1 |2 + |L|2 |wa |2 n1 and n∗1 in cross terms, random phase, add 0 on average
= Saa
Saa = wa wa∗ = Kn2 n∗2 K ∗ = KK ∗ |n2 |2 = |K|2 |n2 |2 = |K|2 |n1 |2 (A.7)
Appendix B
B.1 Introduction
added. In the NREL 5-MW baseline proportional-integral (PI) blade pitch controller [9], because
of the type of gain-scheduling implementation that is used, the effective values of the feedback
gains are not obvious, even when only the baseline feedback controller alone is used. The addition
of feedforward can further change the effective feedback gains, and the effective feedforward gains
can also be affected by the gain-scheduling implementation. This appendix explains the effects of
The original implementation of the NREL 5-MW baseline proportional-integral (PI) blade
pitch controller [9] is shown in Fig. B.1. The generator speed error is integrated and multiplied by
proportional and integral (P and I) gains KP and KI , and then the sum of the resulting signals is
multiplied by the output of the gain-scheduling block (f (u) = 1/(1 + u/0.109997)). An optional
feedforward signal addition is shown in blue, but this is zero for the original controller. Finally, the
signal is saturated and rate limited, and the output is the collective blade pitch command to the
turbine.
124
feedforward
control signal
(= 0 in baseline
generator operation)
speed blade pitch
(rpm) + KP + command
_ X +
∫ KI saturation rate
1173.7 0° to 90° limit
rpm 8 °/s 1-sample
delay
(desired
speed) f(u)=1/(1+u/d) 1/z
Figure B.1: Block diagram of the baseline gain-scheduled proportional-integral feedback blade pitch
controller for the NREL 5-MW turbine, with the original (Integrate First) control implementation.
In the original implementation, the integral is saturated with an upper and lower limit to
avoid integrator windup when the pitch command is saturated. Integrator saturation is omitted
here because this analysis focuses on operation within saturation limits and covers cases both with
and without feedforward. When feedforward with a positive average value is included, a different
anti-windup method is required to prevent integrator windup while also allowing the integrator to
Fig. B.2 shows the gain-scheduling feedback loop part of Fig. B.1, except that it omits for
simplicity the 1-sample delay and saturation and rate limits that are contained within the loop in
actual implementation. In Fig. B.2, β1 is the signal from the PI blocks, β2 represents the optional
feedforward signal which is zero for the original controller, βo is the blade pitch command, and
Let
β1 = β1 + ∆β1
β2 = β2 + ∆β2
βo = βo + ∆βo
where β1 represents the average value of β1 , and ∆β1 represents a deviation from this average
value, which can be thought of as the amplitude of a sinusoid or other small signal. Similarly, β2
and βo are the average values of their respective signals, and ∆β2 and ∆βo are the amplitudes
125
2
1 X + o
f(u)=1/(1+u/(2d))
Figure B.2: Nonlinear gain-scheduling loop. A simplified part of Fig. B.1, where d = 0.109997/2.
of their respective signals. The value βo is determined by the average wind speed according to
Fig. B.3, which shows the steady-state pitch angle that will make generator speed error equal zero
for each constant wind speed. The following analysis of this loop treats βo as an independent
equations below.
0.5
Pitch Angle
0.4
0.3
Steady-State
0.2
¡ ¢
(rad) βo
0.1
0
10 15 20 25
Steady-State Wind Speed (m/s)
Figure B.3: Steady-state pitch angle βo as a function of steady-state wind speed.
126
∆βo y β1 + ∆β1 , β2 − βo
=
∆β1 ∆β1
βo − β2
β1 =
f (βo )
βo = β2 + β1 f (βo )
Note that for small ∆β1 and for β2 = 0, the following approximation can be made
∆βo
≈ 1/(1 + 2βo /(2d))
∆β1
as derived in Appendix C.
∆βo
Fig. B.4 shows the resulting effective gain of the Fig. B.2 loop when β2 = 0, for
∆β1
different values of ∆β1 . The black line shows that the average output value βo is simply equal to
β1 f (βo ). However, the colored lines show that the effective gain for deviations about the operating
point is significantly lower than f (βo ). As wind speed (and βo ) increases, the discrepancy grows
between the effective closed-loop gain and f (βo ). The ratio of the two is shown in Fig. B.5, which
plots, for ∆β1 = 0.001 rad, the blue line from Fig. B.4 divided by the black line from Fig. B.4
∆βo
/f (βo ) . This can be considered an effective additional gain, in addition to f (βo ), which
∆β1
was the gain originally intended in the design [9].
When β2 is nonzero, the behavior of the loop changes. The effective gain of the loop, assuming
an input amplitude of 0.001 rad for either ∆β1 or ∆β2 , is shown in Fig. B.6 for different values
of β2 . As β2 rises, β1 decreases, and the effective gain of the loop rises. When β2 = βo , then
127
∆β1 (rad)
0.8
0.001
Effective Gain
0.05
0.6
(∆βo /∆β1 )
0.1
0.2
0.4
0.2
0
0 0.1 0.2 0.3 0.4
βo (rad)
0.9
0.8
0.7
( 0.2606 , 0.59 )
0.6
0.5
0 0.1 0.2 0.3 0.4
βo (rad)
∆βo
Figure B.5: Plot of the blue line from Fig. B.4 divided by the black line from Fig. B.4 /f (βo ) .
∆β1
This represents the effective additional gain factor caused by the Integrate First implementation.
The point marked at 59% is the effective additional gain for the 18 m/s wind speed operating point
(0.2606 rad).
128
1.4
1.2 β2 /βo
1 0
Effective Gain
0.333
0.8 0.667
0.6 1
1.33
0.4
0.2
0
0 0.1 0.2 0.3 0.4
βo (rad)
Figure B.6: Effective closed-loop gain of the gain-scheduling loop shown in Fig. B.2, from β1 to βo in
solid lines and from β2 to βo in dotted lines. The solid black line, in addition to plotting effective gain
from β1 when the β2 = βo , is also a plot of the gain-scheduling function f (u) = (1/(1+u/0.109997)).
∆βo ∆βo
β1 = 0. The result is that = f (βo ), and = 1. These effective closed-loop gains depend
∆β1 ∆β2
on β1 because, assuming there are variations in βo , then β1 is multiplied by an inversely-varying
gain-scheduling signal (inversely because of the negative slope of the gain-scheduling function), and
Numerical simulations of the Fig. B.2 loop, with 1-sample delay included, produced very
similar results to these analytical results, when tested at various operating points using sinusoidal
inputs. These results are an approximation because inputting a sinusoid to the closed loop results
Instead of the original Integrate First implementation, Fig. B.7 shows a control implemen-
tation where the gain-scheduling multiplication happens before the integrator and the other PI
blocks. It makes no difference whether a constant gain is implemented before or after the gain-
scheduling multiplication, but the position of the integrator does make a difference. Both Integrate
First and Multiply First methods have appeared in published controllers, with known examples
listed in Appendix D.
129
feedforward
control signal
(= 0 in baseline
generator operation)
speed blade pitch
(rpm) + command
_ X KP + +
∫ KI saturation rate
1173.7 0° to 90° limit
rpm 8 °/s 1-sample
delay
(desired
speed) f(u)=1/(1+u/d) 1/z
Figure B.7: Block diagram of the baseline gain-scheduled proportional-integral feedback blade pitch
controller for the NREL 5-MW turbine, with the Multiply First control implementation.
In this Multiply First configuration, because the integrator is inside the loop, and the in-
tegrator accumulates some average value, the average value of the integrator works identically to
adding in some average value through the feedforward control signal (β2 in Fig. B.2). Further, the
average value that the integrator accumulates is such that it, together with feedforward if included,
maintains the correct blade pitch command for the current operating point. Therefore, regardless of
the average value of the feedforward control signal, the effective closed-loop gains for this Multiply
First implementation should match the black lines (labeled ‘1’) in Fig. B.6.
Figs. B.8, B.9, B.10, and B.11 show blade pitch and generator speed simulation results in both
the time and frequency domains for the NREL 5-MW turbine, augmented with a pitch actuator
model inside the gain-scheduling loop, under four different PI blade pitch control implementations:
Integrate First as in Fig. B.1, Multiply First as in Fig. B.7, Multiply First with an additional gain
factor of 0.59 applied to both P and I gains, and Integrate First with an additional gain factor of
1/0.59 applied to both P and I gains. Multiply First results are significantly different from Integrate
First results until either an additional gain factor of 0.59 is included in the Multiply First imple-
mentation or an additional gain factor of 1/0.59 is included in the Integrate First implementation.
This is consistent with the analytical results for the simplified loop in Fig. B.2, which predicted
Integrate First to result in an additional effective gain factor of 0.59, as labeled in Fig. B.5.
130
Figs. B.10 and B.11 also show that in both the higher effective gain and the lower effective gain
cases, Integrate First works better than Multiply First at reducing both pitch rate and generator
speed error. Further investigation is needed to explain why this is the case.
In addition to the 18 m/s Class A turbulence (17% turbulence intensity) wind field results
shown in Figs. B.8, B.9, B.10, and B.11, the four implementation cases were also simulated in Class
A, B, and C turbulence conditions for average wind speeds of 12, 14, 16, 18 and 20 m/s. For each
average wind speed, the corresponding additional effective gain factor from Fig. B.5 was used when
needed. The conclusions drawn from each of these simulations are the same as those for the 18 m/s
Class A simulation. Multiply First and Integrate First results are significantly different from each
other until correction according to Fig. B.5, and after correction, in both the higher-effective-gain
and the lower-effective-gain cases, Integrate First performs better than Multiply First at reducing
Two of the four implementation strategies shown in the simulation results in Figs. B.8, B.9,
B.10, and B.11 use an additional gain factor that depends on the constant average wind speed of the
simulation. Because they assume a constant average wind speed and rely on knowledge of what that
average is, these two implementations are not ready for real-world use. Real-world implementation
would be complicated because it would require an estimate of the average wind speed, and if this
estimate depended on blade pitch, then this would create another nonlinear feedback loop in need of
correction. This section instead presents a simpler way to correct Integrate First so that it behaves
similarly to Multiply First and can be easily implemented in real-world situations with changing
Instead of multiplying the P and I gains by a correction factor that depends on the average
wind speed, effective gains can be changed by changing the gain-scheduling function f (u). The goal
∆βo
is to find a new f (u) such that, for a small-amplitude sinusoid, equals 1/(1 + (βo )/0.109997),
∆β1
when β2 = 0. In Fig. B.4, this desired effective closed-loop gain is shown in black, and the actual
131
25
Integrate First
Multiply First
Multiply First, Gains X 0.59
Blade Pitch Angle (deg)
15
10
5
300 320 340 360 380 400
Time (s)
Figure B.8: Blade pitch time series from simulation of the NREL 5-MW turbine, augmented with
a pitch actuator model, in an 18 m/s wind field with 17% turbulence intensity, under four different
PI blade pitch control implementations. Results from a 100-second portion of a 1-hour wind field
simulation.
132
1350
Integrate First
Multiply First
1300
Multiply First, Gains X 0.59
Generator Speed (rpm)
1200
1150
1100
1050
300 320 340 360 380 400
Time (s)
Figure B.9: Generator speed time series from simulation of the NREL 5-MW turbine, augmented
with a pitch actuator model, in an 18 m/s wind field with 17% turbulence intensity, under four
different PI blade pitch control implementations. Results from a 100-second portion of a 1-hour
wind field simulation.
133
1.5
0.5
0
0 0.1 0.2 0.3 0.4
Frequency (Hz)
Figure B.10: Blade pitch rate power spectral density from simulation of the NREL 5-MW turbine,
augmented with a pitch actuator model, in a 1-hour 18 m/s wind field with 17% turbulence intensity,
under four different PI blade pitch control implementations.
4
x 10 Generator Speed Error (rpm)
2.5
Power Spectral Density (Power/Hz)
Integrate First
Multiply First
2 Multiply First, Gains X 0.59
Integrate First, Gains X 1/0.59
1.5
0.5
0
0 0.1 0.2 0.3 0.4
Frequency (Hz)
Figure B.11: Generator speed error power spectral density from simulation of the NREL 5-MW
turbine, augmented with a pitch actuator model, in a 1-hour 18 m/s wind field with 17% turbulence
intensity, under four different PI blade pitch control implementations.
134
effective closed-loop gain, using the original f (u), is shown in dark blue.
The new f (u) that, for small-amplitude inputs, results in the desired effective closed-loop
gain is f (u) = 1/(1 + u/(2·0.109997)) where now d = 0.109997 instead of d = 0.109997/2. Using
this new f (u), Fig. B.12 shows the new effective gain divided by the desired gain for different input
amplitudes. As input amplitude approaches zero, this ratio approaches 1 as desired. Appendix C
provides a derivation confirming that using f (u) = 1/(1 + u/(2·d)) results in an effective gain
approximately equal to 1/(1 + (βo )/d) when the input amplitude is small and when there is no
feedforward signal.
Results of simulations for the Integrate First method with the original P and I gains and
the new f (u) are shown in Figs. B.13 and B.14, along with two of the implementations previously
presented in Figs. B.10 and B.11. Results are very similar to the Integrate First method with the
B.6 Conclusions
The Multiply First implementation maintains the proportional and integral gains at their
unscheduled values times f (βo ). The Integrate First implementation effectively further reduces the
proportional and integral gains, as plotted for a simplified loop model and confirmed in simulation,
except when augmented with a feedforward control signal whose average value is βo . This feedfor-
ward control augmentation serves the same function as moving the integrator into the loop does in
the Multiply First method. When gains are corrected to make Integrate First and Multiply First
have the same effective gains, the Integrate First method has better performance. Gain correction
as implemented in simulation is not ready for real world implementation, but a simple change to
the gain-scheduling function f (u) has almost the same effect in simulation as gain correction, and
Our recommendations are that when feedforward control is included, the Multiply First
implementation should be used for simplicity. When feedback control alone is used, the use of
the Integrate First implementation, with the simple change to the gain-scheduling function f (u),
135
0.9
∆β1 (rad)
0.001
0.85 0.05
0.1
0.2
0.8
0 0.1 0.2 0.3 0.4
βo (rad)
Figure B.12: ‘Effective closed-loop gain’ divided by ‘desired closed-loop gain’ for the implementation
∆βo
with new f (u). ‘Effective closed-loop gain’ is when β2 = 0, in a modified version of the gain-
∆β1
scheduling loop shown in Fig. B.2 where now d = 0.109997 instead of d = 0.109997/2, and ‘desired
closed-loop gain’ is 1/(1 + βo /0.109997).
Multiply First
3 Integrate First,
Gains X 1/0.59
2.5 Integrate First,
New f(u)
2
1.5
0.5
0
0 0.1 0.2 0.3 0.4
Frequency (Hz)
Figure B.13: Blade pitch rate power spectral density from simulation of the NREL 5-MW turbine,
augmented with a pitch actuator model, in a 1-hour 18 m/s wind field with 17% turbulence inten-
sity, for the new f (u) implementation and two of the previously displayed implementations from
Fig. B.10.
136
Multiply First
6000 Integrate First,
Gains X 1/0.59
5000 Integrate First,
New f(u)
4000
3000
2000
1000
0
0 0.1 0.2 0.3 0.4
Frequency (Hz)
Figure B.14: Generator speed error power spectral density from simulation of the NREL 5-MW
turbine, augmented with a pitch actuator model, in a 1-hour 18 m/s wind field with 17% turbulence
intensity, for the new f (u) implementation and two of the previously displayed implementations
from Fig. B.11.
137
should be considered because of its improved performance. The reason for this improvement is
not yet understood, and further study is suggested as future work. One possible next step is to
use the rotor-effective wind speed as a scheduling variable instead of using the average blade pitch
angle, and to then compare the resulting performance outputs to the Multiply First and corrected
Integrate First methods. Another possible next step is to incrementally decrease the turbulence in
the modeled wind field in order to study whether the difference in performance between the two
βo = β1 · f (βo ) (C.1)
1
f (u) = u (C.2)
1+
2d
2d∆β1 (d + βo )2 (C.3)
then
y(β1 + ∆β1 ) − βo 1
≈ (C.4)
∆β1 βo
1+
d
where
p
y(u) = −d + d2 + 2du (C.5)
βo
β1 = (C.6)
f (βo )
The meaning of these equations is as follows. Equation (C.1) can be read directly from
β1 = β1 + ∆β1
βo = βo + ∆βo
with β1 representing the average value of β1 , and ∆β1 representing a deviation from this average
value. Similarly, βo represents the average value of βo , and ∆βo represents a deviation from this
average value.
139
In equation (C.2), let d = 0.109997/2 to represent the original f (u), or let d = 0.109997 to
represent the proposed new f (u) from Section B.5. This derivation does not rely on a particular
choice of d. Inequality (C.3) is true when ∆β1 is small. The left side of equation (C.4) represents
the effective closed-loop gain: the deviation from the mean output due to a deviation from the mean
∆βo
input, all divided by the amplitude of the deviation from the mean input . This effective
∆β1
closed-loop gain is approximately equal to the gain shown on the right side of equation (C.4). When
d = 0.109997, the right side is the desired gain from Section B.5. Equation (C.5) shows how to
calculate y(u), where y(β1 ) = βo when β2 = 0. Equation (C.5) can be derived by combining (C.1)
y(β1 + a) − βo
a
q
−d + d2 + 2d β1 + a − βo
= by plugging in (C.5)
a
s
βo
−d + d2 + 2d + a − βo
f (βo )
= by plugging in (C.6)
a
s
βo
−d + d2 + 2d βo 1 + + a − βo
2d
= using 1/f (u) from (C.2)
a
q
2
−(d + βo ) + d2 + 2dβo + βo + 2da
=
a
q 2
−(d + βo ) + d + βo + 2da
=
a
s
2da
−(d + βo ) + d + βo 1+ 2
d + βo
=
a
!
da
−(d + βo ) + d + βo 1 + 2
d + βo √ x
≈ binomial approximation & (C.3) 1+x≈1+ for small x
a 2
!
da
(d + βo ) 2
d + βo
=
a
d
=
d + βo
1
=
βo
1+
d
Appendix D
The NREL 5-MW baseline proportional-integral (PI) blade pitch controller implementation
GK is the gain-scheduling correction factor, PC KI is the integral gain at rated wind speed, and
IntSpdErr is the integral of the generator speed error. The integration occurs before the gain-
scheduling correction factor is multiplied, indicating that this controller uses the Integrate First
implementation.
The NREL CART3 baseline blade pitch controller Simulink model (CART3 Baseline V3.mdl)
shows its gain correction factor multiplication happening before the discrete-time integrator. This
The Basic DTU Wind Energy controller [75] also uses a Multiply First implementation, as
shown in the ‘Integral Term’ block equation in the Figure 1.3 within reference [75], where the gain
factor ηk multiplies only the error terms, and not the θI,k−1 term which holds the value of the
integral state from the previous time step. This means the multiplication is essentially happening
to the input of the integral, not the output. However, this blade pitch control implementation has
additional complications because the gain factor depends on both pitch angle and generator speed
error, and there is PID control of power error in addition to generator speed error.