A Tutorial on the Dynamics and

Control of Wind Turbines and

Wind Farms
Lucy Pao
Richard and Joy Dorf Professor
Electrical, Computer, & Energy Engineering Department
University of Colorado at Boulder

Scientific Director
Center for Research and Education in Wind

Annual Joint IEEE RRVS, RRVS NIU Student Branch, & IEEE CSS Chapter Meeting IEEE-RRVS, November 2009
 Grateful Acknowledgement to:
Kathryn Johnson

Clare Boothe Luce Assistant Professor

Division of Engineering
Colorado School of Mines

Colorado School of Mines Site Director

Center for Research and Education in Wind

This talk is based on our joint paper:

L. Y. Pao and K. E. Johnson, “A Tutorial on the Dynamics and Control of Wind Turbines
and Wind Farms,” Proc. American Control Conf., St. Louis, MO, pp. 2076-2089, June 2009.

Available at http://ecee.colorado.edu/~pao/ACC09WindTutorialSessionPresentations
Lucy Y. Pao Acknowledgment IEEE-RRVS, November 2009
 Wind Energy
World wind energy, total installed capacity
 Fastest growing energy source
in the world
 Current global installed capacity
exceeds 120,000 MW, with a
projected growth of more than
20% per year for the next five
years
 Wind farms today produce
electrical power at a Cost-of-
Energy of approximately
$0.03/kWh, comparable to that
of coal and natural gas based ’08
power plants [data from www.wwindea.org]
Lucy Y. Pao Introduction 2 IEEE-RRVS, November 2009
 Increasing Turbine Size

 Typical size of utility-scale

wind turbines has grown
dramatically
 Large flexible structures
operating in uncertain
environments [video]
 Advanced controllers can
help increase energy
capture efficiency and
reduce structural loading
Decrease the cost
[diagrams and schematic from www.renewableenergy.no,
of wind energy www.aerospaceweb.org/aircraft/jetliner/b747, and
en.wikipedia.org/wiki/American_football]

Lucy Y. Pao Introduction 3 IEEE-RRVS, November 2009

 Outline
 Motivation and Wind Turbine Basics
 Wind Turbine Control Loops
 Issues in Turbine Control
 Advanced Turbine Control
 Wind Farms
 Offshore Wind
 Conclusions

Lucy Y. Pao Outline 4 IEEE-RRVS, November 2009

 Vertical vs. Horizontal Axis Wind Turbines

 Vertical-axis wind turbines

(VAWTs) more common among
smaller turbines
 HAWTs are the most commonly
produced utility-scale wind [photo from www.symscape.com]
turbines
 Advantages of horizontal-axis
wind turbines (HAWTs)
 Improved power capture capabilities
 Pitchable blades
 Improved structural performance
[photo courtesy of L. J. Fingersh, NREL]
Lucy Y. Pao Introduction 5 IEEE-RRVS, November 2009
 Vertical vs. Horizontal Axis Wind Turbines

Horizontal Axis Wind Turbine

 Vertical-axis wind turbines
(VAWTs) more common among
smaller turbines rotor
nacelle
 HAWTs are the most commonly
produced utility-scale wind
turbines
 Advantages of horizontal-axis
tower
wind turbines (HAWTs)
 Improved power capture capabilities
 Pitchable blades
 Improved structural performance NREL’s Control Advanced Research
Turbine (CART2)
Lucy Y. Pao Introduction 5 IEEE-RRVS, November 2009
 Wind Turbine Components

 Wind encounters rotor, Upwind HAWT

causing it to spin
 Low-speed shaft transfers
energy to the gear box
 Steps up speed
 Spins high-speed shaft
 High-speed shaft causes
generator to spin,
producing electricity
 Yaw system turns nacelle
so that rotor faces into
[figure courtesy of US Dept. of Energy]
the wind
Lucy Y. Pao Introduction 6 IEEE-RRVS, November 2009
 Wind Turbine Design Considerations

 Upwind vs. downwind Upwind HAWT

 Tower shadow
 Variable or fixed pitch
 Initial cost
 Ability to control loads
and change
aerodynamic torque
 Variable or fixed speed
 Aerodynamic efficiency
 Electrical power
processing
[figure courtesy of US Dept. of Energy]
 Number of blades
Lucy Y. Pao Introduction 7 IEEE-RRVS, November 2009
 Operating Regions
 Region 1: Low wind speed Example Power Curves for 2.5 MW Wind Turbine
(below 6 m/s)
 Wind turbines not run, because Variable
power available in wind is low Speed
compared to losses in turbine Turbine
system Power
Fixed Speed Turbine Power
 Region 2: Medium wind speeds
(6 m/s to 11.7 m/s) Wind
 Variable-speed turbine captures Power
more power
 Fixed-speed turbine optimized for
one wind speed (10 m/s)
– Max difference in example curves is
150 kW.
 For typical wind speed distributions,
in this example, variable-speed
Region 1
turbine captures 2.3% more energy
than constant-speed turbine
 Region 3: High wind speeds (above 11.7 m/s)
 Power is limited to avoid exceeding safe electrical and mechanical load limits
Lucy Y. Pao Introduction 8 IEEE-RRVS, November 2009
 Outline
 Motivation and Wind Turbine Basics
 Wind Turbine Control Loops
 Wind Inflow
 Sensors
 Actuators
 Torque Control
 Pitch Control
 Issues in Turbine Control
 Advanced Turbine Control
 Wind Farms
 Offshore Wind
 Conclusions

Lucy Y. Pao Outline 9 IEEE-RRVS, November 2009

 Wind Turbine Control Loops

turbine axis
Wind
instantaneous wind field Pitch Motor
Pitch
 e Pitch Angle

Desired
Rotor
 Controller

Speed Load
d Torque Torque
Controller

Power Converter
Speed
Sensor Rotor Speed

Lucy Y. Pao Walk Around the Loops 10 IEEE-RRVS, November 2009

 Wind Inflow
 Differential heating of atmosphere
is driving mechanism for earth’s
winds
 Numerous phenomena affect wind
inflow across a wind turbine’s rotor
plane
 Sea breezes
 Frontal passages
 Mountain and valley flows
 Nocturnal low-level jet
 Rotor plane of MW utility-scale
turbines span from 60m to 180m Hourly profiles of mean wind speed
above the ground after sunset on 15 Sept 2003
 Virtually impossible to obtain a
good measurement of the wind [Figure courtesy of R. Banta, Y. Pichugina,
speed encountering the entire span N. Kelley, B. Jonkman, and W. Brewer]
of blades

Lucy Y. Pao Walk Around the Loops 11 IEEE-RRVS, November 2009

 Characterizing the Wind
Example Weibull Distributions
 Average wind speed k = 1.5
 Spatial k=2

probability density
k = 2.5
 Temporal k=3

 Frequency distribution of
wind speeds
 Spatial
 Temporal
wind speed w (m/s)
 Prevailing wind direction w k 1   w k 
 Frequency of other wind f ( w )  k k exp    
c  c 
directions  
c : scale parameter k : shape parameter
 Capacity Factor
actual energy output over time period
CF =
energy output if turbine operated at max output over same time period
Lucy Y. Pao Walk Around the Loops 12 IEEE-RRVS, November 2009
 Wind Turbine Control Loops

turbine axis
Wind
instantaneous wind field Pitch Motor
Pitch
 e Pitch Angle

Desired
Rotor
 Controller

Speed Load
d Torque Torque
Controller

Power Converter
Speed
Sensor Rotor Speed

Lucy Y. Pao Walk Around the Loops 13 IEEE-RRVS, November 2009

 Sensors
 Rotor speed measured Upwind HAWT
on either high-speed
(generator) or low-speed
(rotor) shafts
 Gear box ratio known
 Anemometers used for
supervisory control
purposes
 Usually located on
nacelle behind rotor
plane
poor measurement
of wind
 Power measurement
devices
[figure courtesy of US Dept. of Energy]

Lucy Y. Pao Walk Around the Loops 14 IEEE-RRVS, November 2009

 Sensors
Several types of sonic and propeller anemometers
 Rotor speed measured on a meteorological tower at NREL’s NWTC
on either high-speed
(generator) or low-speed
(rotor) shafts
 Gear box ratio known
 Anemometers used for
supervisory control
purposes
 Usually located on
nacelle behind rotor
plane
poor measurement
of wind
 Power measurement
devices

[Photo courtesy of L. J. Fingersh, NREL]

Lucy Y. Pao Walk Around the Loops 15 IEEE-RRVS, November 2009
 Additional Sensors
 Strain gauges Upwind HAWT
 Tower
 Blades
 Accelerometers
 Position encoders
 Drive shaft
 Blade pitch
actuation systems
 Torque
transducers
[figure courtesy of US Dept. of Energy]

Lucy Y. Pao Walk Around the Loops 16 IEEE-RRVS, November 2009

 Wind Turbine Control Loops

turbine axis
Wind
instantaneous wind field Pitch Motor
Pitch
 e Pitch Angle

Desired
Rotor
 Controller

Speed Load
d Torque Torque
Controller

Power Converter
Speed
Sensor Rotor Speed

Lucy Y. Pao Walk Around the Loops 17 IEEE-RRVS, November 2009

 Actuators
 Yaw motor Upwind HAWT

 Slow (usually < 1 deg/s)

 Generator
 Fast (time constant usually
> 10x that of rotor speed)

[figure courtesy of US Dept. of Energy]

Lucy Y. Pao Walk Around the Loops 18 IEEE-RRVS, November 2009

 Actuators
Inside the Nacelle of the 3-Bladed Controls
 Yaw motor Advanced Research Turbine (CART3) at NREL’s
National Wind Technology Center (NWTC)
 Slow (usually < 1 deg/s)
 Generator
 Fast (time constant usually
> 10x that of rotor speed)

[Photo courtesy of L. J. Fingersh, NREL]

CART3 is a 600 kW wind turbine with a

40 m rotor diameter that is used at NREL’s
NWTC as an experimental test bed
for advanced controllers.

Lucy Y. Pao Walk Around the Loops 18 IEEE-RRVS, November 2009

 Actuators
Three pitch motors on the CART3
 Yaw motor
 Slow (usually < 1 deg/s)
 Generator
 Fast (time constant usually
> 10x that of rotor speed)
 Blade pitch motor
 Fast [Photo courtesy of L. J. Fingersh, NREL]

– Up to 18 deg/s for 600 kW turbines CART3 is equipped with

– Up to 8 deg/s for 5 MW turbines independent blade pitch capability.

 Collective vs. Individual Pitch

Lucy Y. Pao Walk Around the Loops 18 IEEE-RRVS, November 2009

 More on Actuators
 Operational blade pitch angle data from CART2:
 CART2 is a 2-bladed, 600 kW
wind turbine with a 43 m diameter
rotor at NREL’s NWTC
 Data from a normal shut-down
event caused by the wind speed
decreasing into Region 1
• Pitch rate limited to approx 5 deg/s
• Lag between commanded and
actual pitch can be represented by a
1st-order filter
 Teetering hinge on 2-bladed turbines
 Allows rotor to respond to differential loads
when blades in vertical position
Lucy Y. Pao Walk Around the Loops 19 IEEE-RRVS, November 2009
 Torque Control

turbine axis
Wind
Pitch Motor
instantaneous wind field Pitch
 e Pitch Angle

Desired
Rotor
 Controller

Speed Load
d Torque Torque
Controller

Power Converter
Speed
Sensor Rotor Speed

Lucy Y. Pao Walk Around the Loops 20 IEEE-RRVS, November 2009

 “Standard” Torque Control

R  c  Kˆ 2
Tip-speed ratio:  C pmax
w 1
K  R 5

2 3
 c = generator (control)
torque
 ̂ = measured rotor speed
  = air density
 R = rotor radius
 CPmax = maximum power
coefficient
 * = optimum tip-speed ratio

Lucy Y. Pao Walk Around the Loops 21 IEEE-RRVS, November 2009

 Torque Control: Perfect Measurements

 When measurements are perfect and turbine is

perfectly modeled, “standard” torque control
leads to optimal operation in the steady-state
   aero   c 
1
J
5 2 
1 C p C pmax
  R   3  3 
2J  * 

C pmax
Cp  3    0
*3
C pmax
Cp  3    0
*3
Lucy Y. Pao Walk Around the Loops 22 IEEE-RRVS, November 2009
 Torque Control Summary
 Data from CART2
 Key features of
standard torque
control
 Nonlinear
 Only required
measurement is
rotor speed
 Saturates at rotor
speeds near rated
 Speed regulation
achieved via pitch
control

Lucy Y. Pao Walk Around the Loops 23 IEEE-RRVS, November 2009

 Pitch Control

turbine axis
Wind
Pitch Motor
instantaneous wind field Pitch
 e Pitch Angle

Desired
Rotor
 Controller

Speed Load
d Torque Torque
Controller

Power Converter
Speed
Sensor Rotor Speed

Lucy Y. Pao Walk Around the Loops 24 IEEE-RRVS, November 2009

 PID Pitch Control
 Speed regulation at high
winds typically achieved KDs
using PID pitch control
d  e c
KP


 e = error in rotor speed KI

 d = desired rotor speed s
 c = control pitch angle
̂

 Pitch rate actuation limits

may be up to 8 deg/s
Lucy Y. Pao Walk Around the Loops 25 IEEE-RRVS, November 2009
 Pitch Control Variations
 Derivative term may be filtered e 
to reduce measurement noise KDs
s 
errors
 KP, KI, and KD may be gain 1.2

Gain Schedule Factor GS

1

scheduled due to system 0.8

nonlinearities 0.6

0.4

 Pitch control signal can be 0.2

given as angle or rate of 2 4 6 8 10 12 14

Blade Pitch (deg.)
16 18 20 22

change
 Pitch control may be collective
or independent
 MIMO control options available

Lucy Y. Pao Walk Around the Loops 26 IEEE-RRVS, November 2009

 Pitch and Torque Control
 Pitch control
saturated below
rated
 Saturation
value chosen
to optimize
energy capture
 Pitch and
torque control
loops
complement
each other

Lucy Y. Pao Walk Around the Loops 27 IEEE-RRVS, November 2009

 Outline
 Motivation and Wind Turbine Basics
 Wind Turbine Control Loops
 Issues in Turbine Control
 Size
 Multiple control loops
 Control while stopped
 Modeling inaccuracies
 Advanced Turbine Control
 Wind Farms
 Offshore Wind
 Conclusions
Lucy Y. Pao Outline 28 IEEE-RRVS, November 2009
 Increasing Turbine Size

 Increased flexibility may

lead to structural vibrations
 Tower motion (fore-aft and
side-to-side)
 Drive train torsion
 Blade bending and twisting
 Rotor is larger than some
“coherent” wind turbulence
structures
 Requires individual blade
pitch control [diagrams and schematic from www.renewableenergy.no,
www.aerospaceweb.org/aircraft/jetliner/b747, and
en.wikipedia.org/wiki/American_football]
Lucy Y. Pao Turbine Control Issues 29 IEEE-RRVS, November 2009
 Multiple Control Loops

 Transition between regions 2 and 3

sometimes leads to maximum turbine
loads
 Switching between
torque and pitch
control may
exacerbate problem
 CART2 field data
during a bad
transition
Lucy Y. Pao Turbine Control Issues 30 IEEE-RRVS, November 2009
 Control while Stopped
 Supervisory control may stop
turbines due to faults or high
winds
 Little active control usually
performed while stopped
 Yaw control may still be Most active
performed to point turbine into control
performed
the wind here

 Extreme loads may occur

during “parked” conditions,
usually in high winds
 Fault detection and health
monitoring are also of interest
Lucy Y. Pao Turbine Control Issues 31 IEEE-RRVS, November 2009
 Modeling Inaccuracies
 Torque control assumes perfect knowledge of the
turbine’s Cp surface
 Errors can be costly
 Effect of a 5% modeling error in the optimal tip speed
ratio
 Energy loss of around 1% - 3% in Region 2
 Assume we reach DOE’s 20% Wind Energy by 2030 goal
– requires ~300 GW of installed
capacity
 Assume the cost of energy is
$0.03 per kilowatt-hour (kWh)
 Thus, a 1% loss of energy is
equivalent to a loss of $630
million per year

Lucy Y. Pao Turbine Control Issues 32 IEEE-RRVS, November 2009

 Advanced Control Strategies

 Adaptive control
 Feedforward control
 Using wind speed estimates
 Using wind speed measurements
FEEDFORWARD
CONTROLLER

 
FEEDBACK
TURBINE
CONTROLLER

 Lots of others under development

Lucy Y. Pao Advanced Control 33 IEEE-RRVS, November 2009
 Advanced Blades
 New configurations and
actuators under
development
 Multiple pitch actuators per
blade
– Will allow different pitch angles
at different radial positions
 Microtabs Multiple pitch actuators per blade
– Will change aerodynamic forces
 Air valves
– Will change aerodynamic forces
 Advanced blade concepts
Microtabs (not to scale)
will likely require new
control systems

Lucy Y. Pao Advanced Control 34 IEEE-RRVS, November 2009

 Outline
 Motivation and Wind Turbine Basics
 Wind Turbine Control Loops
 Issues in Turbine Control
 Advanced Turbine Control
 Wind Farms
 Offshore Wind
 Conclusions

Lucy Y. Pao Outline 35 IEEE-RRVS, November 2009

 Wind Farm Considerations
 Wind farms can take advantage of economies of
scale
 May differ from individual turbines in
 noise
 safety
 visual
 environmental effects

Lucy Y. Pao Wind Farms 36 IEEE-RRVS, November 2009

 Wind Farm Layouts and Control
 Control focuses
 Electricity
 Aerodynamics
 Control goal is to maximize “array efficiency”
given existing configuration

Lucy Y. Pao Wind Farms 37 IEEE-RRVS, November 2009

 Offshore Wind Motivation
 Advantages to offshore wind
 Wind resource typically higher and more
consistent
 Turbine size is not
limited by transportation
constraints
 Visual and noise effects
can be avoided more
easily
 More area available,
especially near
population centers
Lucy Y. Pao Offshore Wind 38 IEEE-RRVS, November 2009
 Offshore Wind Turbine Configurations

U.S. has
relatively
more deep
water near
the shoreline
than Europe,
so more
floating
turbines are
likely in the
U.S.
[Image courtesy of NREL]
Lucy Y. Pao Offshore Wind 39 IEEE-RRVS, November 2009
 Floating Platforms
 Floating
platform
configurations
have been
borrowed from
offshore oil rig
technologies

[Image courtesy of NREL]

Lucy Y. Pao Offshore Wind 40 IEEE-RRVS, November 2009
 Offshore Wind Challenges
 Waves can excite structural modes
for both floating and fixed offshore
turbines
 Deep water anchors are expensive
 U.S. has more deep water near
population centers than Europe
 What is the best way to control a
floating inverted pendulum with a
large spinning mass at its top?
 What actuators are necessary?
 How will control affect the energy
capture?
Graphic courtesy of NREL
Lucy Y. Pao Offshore Wind 41 IEEE-RRVS, November 2009
 Conclusions
 Large, flexible turbines lend
themselves to control solutions, and
turbines are getting larger and more
flexible
 Existing turbine controllers tend not
to take advantage of the wealth of
available control theory
 Industry has been slow to adopt
advanced control strategies for both
individual turbines and wind farms
 Offshore wind turbine control is a big
prospective area for research
Lucy Y. Pao Conclusions 42 IEEE-RRVS, November 2009
 Thank You

Lucy Y. Pao Thanks IEEE-RRVS, November 2009