high reliability monitoring and control of wind...

AEROSPACE ENGINEERING AND MECHANICS

High Reliability Monitoring and Control of Wind Turbines

Peter Seiler Department of Aerospace Engineering & Mechanics

University of Minnesota


2

Turbine Components

Figure from the US DOE

Eolos Field Station


3

1. Maximize captured power

2. Minimize structural loads

3. Reduce operational downtime

Performance Objectives

pCAvP 3

21

Power in Wind Power Coefficient: Function of turbine

design, wind conditions, and control


Outline

4

1. Overview of UMN

/ Eolos Research

2. Redundancy Management

in Commercial Aviation

3. Blade health monitoring

using energy harvesting

4. Conclusions…


Eolos Consortium

5

Triaxial DC Accels

Fiber-optic Strain

LP HP/LP HP/LP/LE/TE LP/HP

Established via US DOE Grant

http://www.eolos.umn.edu/

Wind Field Station

2.5MW / 96M Clipper Liberty

(Commissioned on 10/25/2012)

UMN

Field

Station

~25mi

http://www.eolos.umn.edu/


Collaboration with Mesabi Range CTC

6

~190mi

UMN

MRCTC

96m

80m

27m

31m

36.6m

40m

Liberty C96

(Eolos) 225kW Vestas V27

(Mesabi Range)

CART3

(NREL)

Mesabi Range CTC Wind Energy Technology Program

offers A.A.S. degree for maintenance

of utility scale wind turbines.

(V27 shipped from Antwerp on 9/28/2010)


Overview of Research Projects

7

1.31 1.32 1.33 1.34 1.35 1.36 1.37 1.38

x 105

6

6.5

7

7.5

810 min Averaged Wind Speed

Time (s)

10

min

Ave

rag

ed

Win

d S

pe

ed

(m

/s)

8.2 8.4 8.6 8.8 9 9.2 9.4 9.6

0.2

0.25

0.3Clipper C96 Cp Curve for beta=1.25

Tip Speed Ratio

Cp

V27 Control

(Thorson,

Janisch)

Blade Health

Monitoring

(Lim, Mantell,

Yang)

Distributed

Estimation

(Showers)

Wind Farm

Control

(Annoni, Yang,

Sotiropolous,

Bitar)

Active Power

Control

(Wang)

Multivariable

Design Tools

(Ozdemir, Escobar

Sanabria, Balas)


V27 Control Design

8

Accomplishments:

• Mesabi Range rewired turbine, removed stock

controller and installed Master/Slave CRIOs

• UMN designed turbine state logic and rotor

speed tracking.

Future: Fixed speed power generation

References:

• Vestas V27 Test, Petersen, 90

• CART Commissioning, Fingersh/Johnson 02, 04


Wind Farm Modeling and Control

Objectives: • Develop control-oriented models

• Design control laws for increased power capture and load mitigation (Bitar, Seiler, ‘13 ACC)

Simulators: • Saint Anthony Falls Virtual Wind Simulator (Yang, Kang,

Sotiropoulos 2012; Chamorro, Porte-Agel 2011)

• NREL SOWFA (Churchfield, Lee, Michalakes, Moriarty, 2012)

Selected References: • Jensen, ‘83 Risø Report

• Steinbuch, de Boer, Bosgra, Peters, Ploeg, ‘88 JWEIA

• Johnson, Thomas, ‘09 ACC

• Pao, Johnson, ‘09 ACC

• Brand, Soleimanzadeh, 11 EWEA

• Marden, Ruben, Pao, ‘12 ASM

• Wagenaar, Machielse, Schepers, 12 EWEA

• Fleming, Gebraad, van Wingerden, Lee, Churchfield, Scholbrock,

Michalakes, Johnson, Moriarty, ‘13 EWEA 9

SAFL Wind Tunnel Tests

(Chamorro, Porte-Agel)


CFD Results

10

Decreasing Lead Turbine Induction Factor

Park Model (Jensen, ‘83):

22

2 where)1(kxD

Davvvv

Simulation: Turbine Located at x=0.5

Park model fit shown with k=0.01

Summary: Opportunity to optimize

total power output but validated

control-oriented models are needed.

Ptot = 0.3834 Ptot = 0.3888 Ptot = 0.3726


Active Power Control

Objectives: • Use gain-scheduling to track arbitrary power set-

point commands (Wang, Seiler, ‘13 Draft)

• Investigate feasibility for ancillary services

Selected References: • Kirby, Dyer, Martinez, Shoureshi, Guttromson,

‘02 Oak Ridge Report

• Keung, Li, Banakar, Ooi, ‘09 TPS

• Juankorena, Esandi, Lopez, Marroyo, ‘09 CPEEED

• Spudić, Jelavić, Baotić, Perić, ‘10 Torque

• Tarnowski, Kjaer, Dalsgaard, Nyborg, ‘10 PES

• Laks, Pao, Wright, ‘12 ACC

• Aho, Buckspan, Pao, Fleming, ‘13 ASM

• Jeong, Johnson, Fleming, ‘13 WE

11


Multivariable Control Design

12

Objective:

• Develop a framework to easily tune advanced (robust) control designs for wind turbines (Ozdemir, ‘13 PhD)

• Integrate advanced sensors (LIDAR) for preview control (Ozdemir, Seiler, Balas, ‘12 ASM, ‘12 ACC, ‘13 ASM, ‘13 TCST)

• Optimal Multi-Blade Coordinate Transformation (Seiler, Ozdemir, ‘13 ACC)

Selected (LIDAR) References: • Harris, Hand, and Wright, ’06 NREL Report

• Laks, Pao, Wright, ‘09 ASM

• Mikkelsen, Hansen, Angelou, Sjöholm, Harris, Hadley, Scullion, Ellis, Vives, ‘10 AWEA

• Schlipf, Schuler, Grau, Allgöwer, Kühn, ‘10 Torque

• Laks, Pao, Wright, Kelley, B. Jonkman, ‘10 ASM

• Laks, Pao, Simley, Wright, Kelley, ‘11 ASM

• Dunne, Pao, Wright, B. Jonkman, Kelley, Simley, ‘11 ASM

• Korber, King, ‘11 AWEA

Figure from Harris,

Hand, and Wright, ‘06


Distributed Estimation

13

Objectives: • Identify turbine model from real-time data

• Use measurements from upstream turbines to estimate wind for use as feedforward signal for downstream turbines.

Selected References: • Odgaard, Damgaard, Nielsen, ‘08 IFAC

• Knudsen, Bak, Soltani, ‘11 WE

• Van Wingerden, Houtzager, Felici, Verhaegen, 09 IJRNC

• Gebraad, van Wingerden, Fleming, Wright, 11 CCA

40 60 80 100 120 140 1607.9

8

8.1Wind Speed

Time (s)Win

d S

pe

ed

(m

/s)

40 60 80 100 120 140 1600.4

0.5

0.6Cp Comparison

Time (s)

Cp

Turbine Model

Table

6.7 6.75 6.8 6.85 6.9 6.950.4

0.5

0.6Cp Curve

Tip Speed Ratio

Cp

Turbine Model

Table

40 60 80 100 120 140 1605

10

15Wind Speed

Time (s)Win

d S

pe

ed

(m

/s)

40 60 80 100 120 140 1600

0.5

1Cp Comparison

Time (s)

Cp

Turbine Model

Table

5.5 6 6.5 7 7.5 80

0.5

1Cp Curve

Tip Speed Ratio

Cp

Turbine Model

Table

1.31 1.32 1.33 1.34 1.35 1.36 1.37 1.38

x 105

6

6.5

7

7.5


Time (s)

10

min

Ave

rag

ed

Win

d S

pe

ed

(m

/s)

8.2 8.4 8.6 8.8 9 9.2 9.4 9.6

0.2

0.25


Tip Speed Ratio

Cp

FAST

Simulations

Liberty Real-time Data


Overview of Research Projects

14

1.31 1.32 1.33 1.34 1.35 1.36 1.37 1.38

x 105

6

6.5

7

7.5


Time (s)

10

min

Ave

rag

ed

Win

d S

pe

ed

(m

/s)

8.2 8.4 8.6 8.8 9 9.2 9.4 9.6

0.2

0.25


Tip Speed Ratio

Cp

V27 Control

(Thorson,

Janisch)

Blade Health

Monitoring

(Lim, Mantell,

Yang)

Distributed

Estimation

(Showers)

Wind Farm

Control

(Annoni, Yang,

Sotiropolous,

Bitar)

Active Power

Control

(Wang)

Multivariable

Design Tools

(Ozdemir, Escobar

Sanabria, Balas)


Motivation for Monitoring

15

Damaged Gearbox (Image courtesy of Mesabi Range

Community and Tech. College)

Failures Rates Table from: “Wind turbine downtime and its importance

for offshore deployment”, Faulstich, Hahn, Tavner,

Wind Energy, 2010.


Motivation for Monitoring

• Cost of wind energy dominated by capital (installation) + operations & maintenance

• Monitoring can be used to reduce O&M costs

• Preventative maintenance during low wind

• Continued operation after failures

• Large literature of wind turbine monitoring

• 2011 IFAC Competition (Benchmark from Odgaard, Stoustrup, and Kinnaert, 2009 SAFEPROCESS).

• Variety of methods including model-based, data-driven, physical redundancy

• Question: Can design techniques developed for aerospace systems be applied for turbines?

16


17

Commercial Fly-by-Wire

Boeing 787-8 Dreamliner • 210-250 seats

• Length=56.7m, Wingspan=60.0m

• Range < 15200km, Speed< M0.89

• First Composite Airliner

• Honeywell Flight Control Electronics

Boeing 777-200 • 301-440 seats

• Length=63.7m, Wingspan=60.9m

• Range < 17370km, Speed< M0.89

• Boeing’s 1st Fly-by-Wire Aircraft

• Ref: Y.C. Yeh, “Triple-triple redundant 777 primary flight computer,” 1996.


18

777 Primary Flight Control Surfaces [Yeh, 96]

• Advantages of fly-by-wire: • Increased performance (e.g. reduced drag with smaller rudder), increased

functionality (e.g. “soft” envelope protection), reduced weight, lower recurring costs, and possibility of sidesticks.

• Issues: Strict reliability requirements • <10-9 catastrophic failures/hr

• No single point of failure


19

Classical Feedback Diagram

Sensors

Primary

Flight

Computer

Pilot

Inputs Actuators

Reliable implementation of this classical

feedback loop adds many layers of complexity.


20

Triplex Control System Architecture

Sensors

Primary

Flight

Computer

Column

Actuator

Control

Electronics

Pilot

Inputs

Each PFC votes on redundant

sensor/pilot inputs

Each ACE votes on redundant

actuator commands

All data communicated

on redundant data buses

Actuators


21

777 Triple-Triple Architecture [Yeh, 96]

Sensors

x3

Databus

x3

Triple-Triple

Primary Flight

Computers

Actuator Electronics

x4


22

777 Triple-Triple Architecture [Yeh, 96]

Sensors

x3

Databus

x3 Actuator Electronics

x4

Left PFC

INTEL

AMD

MOTOROLA

Triple-Triple

Primary Flight

Computers


Ram Air Turbine

23

Ram air turbine: F-105 (Left) and Boeing 757 (Right)

http://en.wikipedia.org/wiki/Ram_air_turbine

http://en.wikipedia.org/wiki/Ram_air_turbine


24

Summary of Redundancy Management

• Main Design Requirements: • < 10-9 catastrophic failures per hour

• No single point of failure

• Must protect against random and common-mode failures

• Basic Design Techniques • Hardware redundancy to protect against random failures

• Dissimilar hardware / software to protect against common-mode failures

• Voting: To choose between redundant sensor/actuator signals

• Encryption: To prevent data corruption by failed components

• Monitoring: Software/Hardware monitoring testing to detect latent faults

• Operating Modes: Degraded modes to deal with failures

• Equalization to handle unstable / marginally unstable control laws

• Model-based design and implementation for software


Blade Structural Health Monitoring (SHM)

• Data/Power transportation to/from sensors

• Retrofit capability desirable (no cabling)

• Preventative maintenance

• Shortened down time

• Good for unpredictable

working conditions

SHM benefits

Challenges

SHM Example (Rumsey, Paquette, White, Werlock,

Beattie, Pitchford, van Dam, Structural health monitoring of wind turbine blades, 2008)

25


Proposed SHM System

Sensors (Strain/Acceleration)

Telemetry (Wireless transceiver)

Energy Harvester (Piezo-electric)

Issues:

1. Low power in blade vibration

2. Blade loading difficult to model / measure 26


Proposed SHM System

Energy Harvestor = Sensor

Telemetry (Wireless transceiver)

Solution:

1. Use harvested energy as the sensor

2. Rely on triple redundant measurements 27


Approach

• Estimate harvested energy

• Properties of energy harvester (size, efficiency, etc)

• Power available in blade vibrations

• Design low-rate health monitoring algorithm

• Assess feasibility of proposed SHM algorithm

28


Harvested Strain Energy

tfEE

EVwstrain 2

0

0

0 Harvested Strain Energy

29



EH Design Variables:

tfEE

EVwstrain 2

0

0

0

= KEH ,

EH Design Factor


η,

30




tfEE

EVwstrain 2

0

0

0

= Pavail ,

Available

Strain Power

= KEH ,

EH Design Factor


η,

31




tfEE

EVwstrain 2

0

0

0

= Pavail ,

Available

Strain Power

= KEH ,

EH Design Factor

Charging Time


η,

32


Experimental Set-up

33

Front Side P1 Type Back Side P2 Type Overall set-up

Energy

Harvester

Strain

Gages

SMART MATERIAL MFC P2 M2814 Energy Harvester

327.1434.3060.117004.0 mmKeh

Signal Conditioner

Transceiver tfEE

EVwstrain 2

0

0

0

92.4mJ (Transmission)

280mJ (Strain Meas.+Trans.)


Experimental Set-up

34

Front Side P1 Type Back Side P2 Type Overall set-up

Energy

Harvester

Strain

Gages

SMART MATERIAL MFC P2 M2814 Energy Harvester

327.1434.3060.117004.0 mmKeh

Signal Conditioner

Transceiver tfEE

EVwstrain 2

0

0

0

92.4mJ (Transmission)

280mJ (Strain Meas.+Trans.)

Need to estimate available

strain power in blade vibrations


Modeling Blade Strain

Flap DOF

Edge DOF

Initial conditions

Number of Blades

Blade Mode Shapes

Wind Conditions

. . .

Inputs Blade Displacement

Shear Force

Bending Moments

. . .

Outputs

Ref. Jonkman, J. M., Buhl Jr, M. L., “FAST user’s

guide,” NREL, Golden, Colorado, USA, 2005.

35


Modeling Blade Strain

Flap DOF

Edge DOF

Initial conditions

Number of Blades

Blade Mode Shapes

Wind Conditions

. . .

Inputs Blade Displacement

Shear Force

Bending Moments

. . .

Outputs

Ref. Jonkman, J. M., Buhl Jr, M. L., “FAST user’s

guide,” NREL, Golden, Colorado, USA, 2005.

36

iE

iiE

ieEI

cM

,

,

,2

iF

iiF

ifEI

tM

,

,

,2

Result: Calculate blade edge/flap

strain using (FAST) simulated

nodal bending moments


Characterize the strain energy available for typical wind turbines:

Wind Conditions : 6 m/s, Rated Speed, 24 m/s + Low / High Turbulence

Wind Turbine Case Studies

CART3 WindPact Offshore

600 kW 1.5 MW 5.0 MW

37.1 rpm 20.5 rpm 12.1 rpm

Rated Power

Rated Speed

6, 14, 20 m/s 3, 12, 28 m/s 3, 11, 25 m/s Wind Speed

20m/1.8ton 35m/3.9ton 63m/17.7ton Length/Weight

34.9 m 84 m 87.6 m Hub Height

37


020 40 60

80

120

130

140

150-1500

-1000

-500

0

500

1000

1500

020

4060

80

120

130

140

150-1500

-1000

-500

0

500

1000

1500

Strain Simulation in Time & Span

Wind Conditions

24 m/s, Low Turbulence

FAST

Flapwise Strain Edgewise Strain

Stra

in (μ-ε

)

38


Strain Analysis

120 125 130 135 140 145 150-1500

-1000

-500

0

500

1000

1500

Time (sec)

Str

ain

(m

icro

)

Edgewise

Flapwise

0 0.2 0.4 0.6 0.8 10

200

400

600

Max Strain (Edgewise)

|Str

ain

| (u

-str

ain

)

Frequency (Hz)

0 0.2 0.4 0.6 0.8 10

200

400

600

Max Strain (Flapwise)|S

tra

in| (u

-str

ain

)

Frequency (Hz)

FFT

Pick one location

Dominant Freq.

39


5 MW WT model 24 m/s, Low Turbulence

Available Strain Power in Blade Span

0 10 20 30 40 50 600

10

20

30

40

50

60

fEPavail

2

0

Span direction (m)

Stra

in P

ow

er (

W/m

3)

Edgewise

Flapwise Nominal E0 = 1 GPa

FFT analysis for 9 locations

ε = Strain Amplitude (mean to peak)

f = Dominant Frequency

40


Available Strain Power

LWS: 6 m/s MWS: Rated HWS: 24 m/s

LT: Low HT: High

6 Wind Conditions:

3 Wind Turbines: 600 kW, 1.5 MW, 5.0 MW

Wind Speed Turbulence Intensity

600kW1.5MW

5.0MWLWS/HT

LWS/LT

MWS/HT

MWS/LT

HWS/HT

HWS/LT

0

10

20

30

40

50

60

Edgewise Strain Power

Pow

er

availa

ble

[W

/m3]

600kW1.5MW

5.0MWLWS/HT

LWS/LT

MWS/HT

MWS/LT

HWS/HT

HWS/LT

0

10

20

30

40

50

60

Flapwise Strain Power

Pow

er

availa

ble

[W

/m3]

41


Apply to EH/Telemetry Design

tPKw availEHstrain

Harvested Strain Energy (μJ)

42



tPKw availEHstrain


92.4 μJ, Transmission only

Available for 5MW WT Power Pavail = 60, 40, 13 W/m3

43


11 12 13 14 15 16 17 181

2

3

4

5

6

7

8

9

10

11

Design factor of EH, KEH

(mm3)

Charg

ing tim

e (

min

)


tPKw availEHstrain




44

13 W/m3

40

60


11 12 13 14 15 16 17 181

2

3

4

5

6

7

8

9

10

11


(mm3)

Charg

ing tim

e (

min

)

KEH = 14.27 mm3

MFC P2


tPKw availEHstrain




45

13 W/m3

40

60


11 12 13 14 15 16 17 181

2

3

4

5

6

7

8

9

10

11


(mm3)

Charg

ing tim

e (

min

)

KEH = 14.27 mm3

MFC P2


tPKw availEHstrain




46

13 W/m3

40

60

Summary: Power only sufficient for

very low transmission rates.

Question: Can blades be monitored

with low rate data?


KEH = 0.78 mm3

(ZnO Nanowire EH) η=6.8%, E=30GPa, V=0.38mm3(10x20cm2, 20 layers)


tPKw availEHstrain


280 μJ, Single data packet measurement/transmission


Ref. G. Zhu, et al. Flexible High-Output Nanogenerator Based on Lateral ZnO Nanowire Array, 2010

0.5 0.6 0.7 0.8 0.9 10

2

4

6

8

10

12


(mm3)

Charg

ing t

ime (

hour)

13 W/m3

40

60

47


Proposed SHM Algorithm

48

Turbine

Blade Strain

BL1

BL2 BL3

Harvested Energy

760 770 780 790 800-400

-200

0

200

Time (sec)

B1-L

P-2

5-S

Str

ain

(m

icro

)

760 770 780 790 800-400

-200

0

200

Time (sec)

B2-L

P-2

5-S

Str

ain

(m

icro

)

760 770 780 790 800400

600

800

1000

Time (sec)

B3-L

P-2

5-S

Str

ain

(m

icro

)

760 770 780 790 8000

0.05

0.1

0.15

0.2

0.25

Time (sec)

Str

ain

Energ

y (

uJ)

760 770 780 790 8000

0.05

0.1

0.15

0.2

0.25

Time (sec)

Str

ain

Energ

y (

uJ)

760 770 780 790 8000

0.05

0.1

0.15

0.2

0.25

Time (sec)

Str

ain

Energ

y (

uJ)

2000 4000 60000

200

400

600

t(sec)

t(sec)

2000 4000 60000

200

400

600

t(sec)

t(sec)

2000 4000 60000

200

400

600

t(sec)

t(sec)

Key idea: Transmit single pulse when

harvested energy exceeds threshold

(Harvested energy is correlated with damage)

Pulses

SHM Algorithm

Detect blade

damage based

on pulse timing


Problem Set-up

49

FAST / EOLOS

Wind Turbine

Piezo-electric

Energy Harvester Blade Strain

BL1

BL2 BL3

Rosemount, MN

Wind Data Fail

Simplified Damage Model

Strain Energy

Accumulation

760 770 780 790 800-400

-200

0

200

Time (sec)

B1-L

P-2

5-S

Str

ain

(m

icro

)

760 770 780 790 800-400

-200

0

200

Time (sec)

B2-L

P-2

5-S

Str

ain

(m

icro

)

760 770 780 790 800400

600

800

1000

Time (sec)

B3-L

P-2

5-S

Str

ain

(m

icro

)

760 770 780 790 8000

0.05

0.1

0.15

0.2

0.25

Time (sec)

Str

ain

Energ

y (

uJ)

760 770 780 790 8000

0.05

0.1

0.15

0.2

0.25

Time (sec)

Str

ain

Energ

y (

uJ)

760 770 780 790 8000

0.05

0.1

0.15

0.2

0.25

Time (sec)

Str

ain

Energ

y (

uJ)

760 770 780 790 8001

1.05

1.1

1.15

1.2

Synthesizing Blade Damage

(Paquette, et al. 46th AIAA ASM, ‘08)


50 60 70-1000

0

1000

Time (sec)

Edge|S

train

| (u

-str

ain

)

FAST Simulation Result (OS5MW WT)

50

FAST Simulator

(OS5MW WT)

Transceiver

Signal Cond. Strain Energy

EH1

EH2 EH3

TurbSim

Wind Data

Signal

Transmission

Damaged Blade (Progressive ) 1000 2000 30005

10

15

20

Time (sec)

WIn

d V

elo

city (

m/s

)

40 50 60 700

5

10

15

EH

Energ

y (m

J)

t(sec)

40 50 60 700

5

10

15

EH

Energ

y (m

J)

t(sec)

40 50 60 700

5

10

15

EH

Energ

y (m

J)

t(sec)

2000 4000 60000

200

400

600

t(sec)

t(sec)

2000 4000 60000

200

400

600

t(sec)

t(sec)

2000 4000 60000

200

400

600

t(sec)

t(sec)

1000 2000 3000 4000 5000 6000 70000.8

1

1.2

1.4

t(sec)

Ration (

-)

Threshold

1-2

2-3

3-1

2000 4000 6000

300

350

400

450

500

Time (sec)

t(sec)


Clipper Raw Data Result (Healthy)

51

Strain Energy

Hub Height Wind Speed

Signal Transmission

0 500 1000 1500 2000 2500 3000 35003

3.5

4

4.5

5

5.5

Time (sec)

Win

d (

m/s

)

50 60 70 80 90

-600

-400

-200

0

200

400

Time (sec)

Str

ain

(m

icro

)

B1-LE-RT-S

B2-LE-RT-S

B3-LE-RT-S

Strain @ 3 Blades

0 1000 2000 30000

100

200

300

400

500

600

700

800

900Signal firing timing of EHs

Tim

e b

etw

een t

ransm

issio

ns (

sec)

Time (sec)

EH1(Blade1)

EH2(Blade2)

EH3(Blade3)

2100 2150 2200 2250 2300

50

60

70

80

EH

Energ

y (m

J)

t(sec)

2100 2150 2200 2250 2300

50

60

70

80

EH

Energ

y (m

J)

t(sec)

2100 2150 2200 2250 2300

50

60

70

80

EH

Energ

y (m

J)

t(sec)

EH1

EH2

EH3


Clipper Data (same data) with Synthetic Fault

52

Damage Model Signal Transmission

(Paquette, J., et al. 46th AIAA ASM, NV 2008)

Fail

Interpretation

500 1000 1500 2000 2500 3000 35000.95

1

1.05

1.1

1.15

1.2

1.25

Time (sec)

Tim

e d

iffe

rence b

etw

een s

ignal tr

ansm

issio

ns (

sec)

Threshold

1-2

2-3

3-1

Blade 3 Strain Output

0 1000 2000 30001

1.05

1.1

1.15

1.2

Time(sec)

Str

ain

Weig

th (

-)

0 1000 2000 30000

100

200

300

400

500

600

700

800


Tim

e b

etw

een t

ransm

issio

ns (

sec)

Time (sec)

EH1(Healthy)

EH2(Healthy)

EH3(Damaged)


53

Conclusions

• Advanced monitoring and control techniques can continue to reduce the costs of wind energy.

• Energy harvesting can be used to power sensors • Max. strain: ~20 to 33% of the blade length

• Max. available strain power for harvesting: ~60 W/m3

• Long charging time is required given current EH technology

• Total harvested energy can be used to monitor blade • Harvested energy is correlated with damage

• Transmit single pulse when harvested energy exceeds threshold

• Rely on triple redundant measurements


Future Work

1. Experimental validation of proposed SHM algorithm • Build test beam specimens with variety of damage types

• Design a power conditioner/booster to maximize EH performance (matched resistance).

• Vibrate test specimen to mimic realistic operating conditions

• Evaluate ability of SHM algorithm to detect damage

2. EH development: ZnO Nanowire array • Ref: Zhu, Yang, Wang, Wang, Flexible High-Output Nanogenerator

Based on Lateral ZnO Nanowire Array, ’10 Nano Letters

0 5 10 15-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

Time (sec)

Norm

aliz

ed B

lade S

train

(unitl

ess)

Blade 1

Blade 2

Blade3

EOLOS Wind Turbine Pulses

Blade Strain Lab-scale Set-up

Energy Harvester 0 1000 2000 3000

0

100

200

300

400

500

600

700

800


Tim

e b

etw

een t

ransm

issio

ns (

sec)

Time (sec)

EH1(Healthy)

EH2(Healthy)

EH3(Damaged)

54


55

Acknowledgments

• Institute for Renewable Energy and the Environment • Grant No. RL-0010-12: “Design Tools for Multivariable Control of Large

Wind Turbines.”

• Grant No. RS-0039-09: “Improved Energy Production for Large Wind Turbines.”

• Grant No. RS-0029-12: “Development of self-powered wireless sensor for structural health monitoring in wind turbine blades”

• US Department of Energy • Grant No. DE-EE0002980: “An Industry/Academe Consortium for Achieving

20% wind by 2030 through Cutting-Edge Research and Workforce Training”

• Eolos Wind Energy Consortium: Provided Liberty data

• US National Science Foundation • Grant No. NSF-CMMI-1254129: “CAREER: Probabilistic Tools for High

Reliability Monitoring and Control of Wind Farms”

high reliability monitoring and control of wind...

Documents