energy harvesting and applications

1/53 © D J Inman

Energy Harvesting and Applications

D. J. Inman A. Erturk, M.A. Karami, C. DeMarqui, S. Anton, B. Joyce, J. Hobeck and Y. Wang

Center for Intelligent Material Systems and StructuresNSF I/UCRC Center for Energy Harvesting Materials and Systems

Virginia TechBlacksburg, VA 24061, USA

[email protected] www.cimss.vt.eduand

Institute for Smart TechnologiesUniversity of BristolBristol, BS* 1TR [email protected]

2/53 © D J Inman

OutlineIntroduction to the basics in energy harvestingSelf charging structuresEnhancements to vibration based energy harvesting using nonlinear dynamics (Erturk)

Piezomagnetic harvesting from a bistable beamPiezoelectric harvesting from a bistable plate

Piezoaeroelastic energy harvesting (Erturk and DeMarque)HarvestingSimultaneous harvesting and flutter suppression

Minimum energy controllers to work with harvesters (Wang)Application to Bridge Monitoring

3/53 © D J Inman

Energy Harvesting as used here refers to: Capturing low levels of ambient waste energy to convert to useable electrical energyThe goals are:

To increase battery life or to replace batteriesTo provide wireless sensor solutions to numerous problems

Most of the effort reported here focuses on harvesting mechanical vibration using the piezoelectric effectOther important harvesting mechanisms are

Use of the Seebeck effect using thermoelectric materials to capture thermal gradientsUse of the photovoltaic effect to capture solar energy

Other effects useful in harvesting mechanical energy are the electromagnetic effect and the electrostatic effect

4/53 © D J Inman

Some Sample Numbers

Mechanical Harvesting from ambient vibration produces values from milliwatts to microwatts. (eg 0.5 g at 8 Hz produces 10 mW)Small thermal electrics and solar can produce up to a few wattsSmall solar arrays can also produce of the order of a watt

5/53 © D J Inman

Piezoelectric based Energy Harvesting Involves 4 Components

Structural Dynamics

PiezoelectricConstitutive laws

Electric Circuits

wtt (x,t)+ Lw(x,t) = f(x,t)Bw =0

€

S=sET+dED=dT+εTE

v + iR =0

Storageand/or direct use

Linear Cantileverswith tip masses well Studied and modeled

Room here for advancesin material science

Need adaptive and nonlinear circuits to address issues of conditioning and optimalityMost studies focus on just optimal resistance

Batteries Li IonSuper capsMechanical not much investigated

6/53 © D J Inman

The piezoelectric effect couples mechanical strain to and electric field allowing for both sensing and actuation functions

y

x

xixi+1

Host structure

Piezoceramic

Neutral axisb

t

y1 y2

The constitutive equations are

S =sET + dED =dT + εTE

S is the strain, T is the stress, D is the electric displacement, E is the electric field, sE is the compliance measured at a constant electric field, eT is the permittivity measured at constant stress and d is the charge coefficient.

7/53 © D J Inman

Harvesting vibration using the piezoelectric effect can be configured in a number of ways

1. Layered onto a structure2. Cantilevered off of a vibrating mounting point (with tip mass)3. Stacked between two moving surfaces4. Flapping in the wind5. ????

Cantilevered with tunable tip mass: Tip mass

PZT

Moving base

Layered into a structural member:

PZT

Proof mass

Bimorph vs. unimorph

8/53 © D J Inman

Storage and Duty Cycles can extend the usefulness of harvested energy

Not all applications need continuous powerStorage devices can be used to enhance the usefulness of a harvesting systemSuppose telemetry of a sensor signal is required 1 sec out of every hour (following example illustrates)

Storage can be through batteries or capacitorsBatteries: limited cycles/hold charge wellCapacitors and super capacitors: large numbers of cycles, drain quickly

9/53 © D J Inman

Most harvesting of vibration using the piezoelectric effect is based on the idea of resonance using a cantilever

Cantilevered with tunable tip mass:

Tip mass

PZT

Moving base

xp (t) =f0

(w n2 −w 2)2 + (2ζw n)

2sinwt

F sinwt

wn =km

Response:

Natural frequency of structure:

10/53 © D J Inman

Internal (strain rate) damping

External (air) damping Electrical term

Inertial excitation External damping excitation

Distributed Parameter Model (Erturk & Inman)

Internal capacitance of the piezoceramic Circuit

excitation term

Coupled electrical circuit equation:

11/53 © D J Inman

Steady state voltage response:

Steady state vibration response:

Closed form solution for both the mechanical and electrical response reveals backwards coupling

Backward piezoelectric coupling in the beam response (clearly not an electrically induced viscous damping term)

12/53 © D J Inman

Cantilever based energy harvesting in the linear region results in a single frequency device

Power out versus frequency of disturbance for a linear harvester

Issues: ambient energy is often broad band in the low frequency range, linear harvesting is narrowband, and frequency increases as the length of the cantilever decreases

13/53 © D J Inman

Here we exploit nonlinear effects to increase the band width of the energy harvester

Nonlinear behavior is purposefully introduced using added magnets to encourage harvesting over a broader frequency range

Once completed the power generated by the linear and nonlinear system are compare for exactly the same ambient input signal

14/53 © D J Inman

Limit Cycle Oscillations for Broad Band HarvestingA magnetic field causes the equation of motion of the harvesting piezoelectric cantilever to be nonlinear

Spacing of the magnets results in:5 equilibrium (3 stable)3 equilibrium (2 stable)1 equilibrium (1 sable)

Limit cycle oscillation is the possible producing large amplitude periodic response over a range of input frequencies

&&x + 2ζ &x−12x 1−x2( )−χv=fχosΩt

&v+ λv+κ &x =0

15/53 © D J Inman

NONLINEAR EFFECTS: Bistable Beam Induces More Power Over Wider Frequency Range

16/53 © D J Inman

Illustration of Limit Cycle Harvesting for low and high amplitude accelerations

8.5 mW at 0.35g an order of magnitude better then w/o magnets!

17/53 © D J Inman

Comparison of voltage vs velocity vs time of linear and nonlinear harvesters for increasing frequency

18/53 © D J Inman

Power Output Comparison of Linear vs Nonlinear

Excitation Frequency

5 Hz 6 Hz 7 Hz 8 Hz

Piezo-Magneto-

Elastic

1.57 mW 2.33 mW 3.54 mW 8.54 mW

Piezo-elastic 0.10 mW 0.31 mW 8.23 mW 0.46 mW

Linear Resonance

Note that at linear resonance the linear system will always win, however it is narrow band and falls off quickly away from resonance and that the nonlinear has higher values overall

19/53 © D J Inman

Such nonlinearities can also be induced by using a bistable plate

Bistable carbon-fiber plate with piezoceramic patches

The plate is clamped to a seismic shaker from its center point.

Arrieta, A.F., Erturk, A., Hagedorn, P., and Inman, D.J., 2010, A Piezoelectric Bi-stable Plate for Nonlinear Broadband Energy Harvesting, Applied Physics Letters (in press).

20/53 © D J Inman

Various nonlinear phenomena can be observed in the bistable plate, enhancing harvesting.

Chaos (12.5 Hz)

High-energy LCO (8.6 Hz)Voltage history samples

Intermittency (9.8 Hz)


21/53 © D J Inman

Large-amplitude oscillations generate very high power output over a range of frequencies.

Average power vs. Frequency

Average power vs. Load resistance

(98.5 kohm)


22/53 © D J Inman

Application of energy harvesting to Structural Health monitoring

The goal is to provide power for remote monitoring systems so that battery life can be extended and or batteries can be removedThree applications are under way

One is monitoring in flowsOne is the monitoring of a bridgeThe last is the monitoring of a wind turbine blade

23/53 © D J Inman

Piezoelectric Grass for Harvesting Energy from a Flow for Running Remote Sensors

In line configuration Staggered configuration

24/53 © D J Inman

3-span steel girder bridge (08/18/09 - Roanoke)

Approximation as a persistent single harmonic (0.05g at 7.7 Hz)

Acceleration signal measured on the bridge

Acceleration data of the bridge has been simplified to a harmonic function for simulations in the lab.

Seismic shaker

Accelerometer

Piezoelectric and electromagnetic generators

Acceleration measured on the shaker

Experimental setup

25/53 © D J Inman

Piezoelectric and electromagnetic power outputs have been measured for an acceleration input of 0.05g (RMS: 0.035g) at 7.7 Hz.

Piezoceramic patches

Accelerometer

Seismic shaker

Rare earth magnets

Combined piezoelectric-electromagnetic generator configuration

Coil

Electromagnetic part : 0.22 V for 82 ohms = 0.6 mW (per coil)

Piezoelectric part : 11.2 V for 470 kohms = 0.3 mW

Power output of a single generator (for 0.05g) = 0.9 mW

26/53 © D J Inman

Increased base acceleration amplitude results in a larger power output. (0.1g, RMS: 0.07g at 7.7 Hz yields 2.7 mW).

Electromagnetic part : 0.42 V for 100ohms = 1.8 mW (from a single coil)Piezoelectric part : 21 V for 470 kohms = 0.94 mW

Power output of a single generator (for 0.1g) = 2.7 mW

[click on the movie]

27/53 © D J Inman

Compact Contact-less prototype has been fabricated and is waiting to be tested

Magnet rotor needs to be matched with blade rotor output profile

Multiple blade geometries have been printed

Piezoelectric Low Speed Wind Harvester

Hope to gather wind speeds at a few kmh

28/53 © D J Inman

Example of how Energy harvesting enables other technologies: Monitoring of wind turbines

Impact Impact or cracking detected via Acoustic Emission (AE) Signals activated periodically

and/or by impact.

NO Sleep

YES

Impedance activated via AE and periodically

to see if significant damage exists

NO Sleep

YESBroadcast damage state

Fatigue

SensorData

Blade Energy

Energy harvesting from blade vibration and centripetal force

29/53 © D J Inman

Uses the interplay between gravity and centripetal force to harvest energy

-10 -5 0 5 10-10

-8

-6

-4

-2

0

2

4

6

8

10

Path of magnet

30/53 © D J Inman

Sample of energy harvesting capability for wind turbine blades

0 1 2 3 4 5 6 7 8 9 100

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

1.8

2

Time t, s

Load

Voltage

, Volts

Model w/o SpringData

Voltage versus time

31/53 © D J Inman

Prospects of using harvested energy to perform Control

1. The first example is harvesting energy from low induced wing vibrations in an aircraft wing• This action of harvesting automatically induces a

shunting effect which acts a a vibration suppression system

• The result is an increase in flutter speed whilst simultaneously harvesting energy

2. The second is a look at performing active control using only harvested energy to provide vibration suppression.• As a first step, we examine which control laws for

vibration suppression will use the least amount of energy

32/53 © D J Inman

Flow Induced Energy HarvestingTypical section with piezoceramics

Experimental setup

33/53 © D J Inman

Flow Induced Energy Harvesting ResultsPiezoaeroelastic equations

Model validation – piezoaeroelastic response

at the flutter boundary

34/53 © D J Inman

Total damping vs. Airflow speed

Tip displacement

Electrical power

De Marqui, Jr., C., Erturk, A., and Inman, D.J., 2010, Piezoaeroelastic Modeling and Analysis of a Generator Wing with Continuous and Segmented Electrodes, Journal of Intelligent Material Systems and Structures, 21 (in press) doi: 10.1177/1045389 X10372261.

The time-domain piezoaeroelastic solution can predict the electromechanical response for airflow speeds below the linear flutter speed.

35/53 © D J Inman

Predictions for wing based harvesting and passive control

The analytical model predicts that the piezoaeroelastic system will harvest 10.7 mW at an air speed of 9.32 m/s

The shunting effect of the energy harvester simultaneously adds damping to the system and predicts an increase in flutter speed of 5.5% (that is a reduction in vibration)

36/53 © D J Inman

Examination of Low Power Control Laws

The goal is to find the feedback control law for vibration suppression the uses the least amount of energy

Fix the performance by fixing the settling time and overshoot and then computing several different control laws to obtain the desired response and then comparing the energy required for each

Hu(t) x(t)

Controller

G

_r(t) Plant

37/53 © D J Inman

Proposed Hybrid Control Laws

Use several common vibration suppression controllers,then use a switching algorithm over the top:

38/53 © D J Inman

The response of four controllers and their hybrid implementation

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2-5

0

5

Time(s)--(a)

Dis

p.( m

m)

0 0.5 1 1.5 2-5

0

5

Time(s)--(b)

Dis

p.( m

m)

0 0.5 1 1.5-5

0

5

Time(s)--(c)

Dis

p.(m

m)

0 0.5 1 1.5 2-5

0

5

Time(s)--(d)

Dis

p.( m

m)

0 0.5 1 1.5 2-5

0

5

Time(s)--(e)

Dis

p.(m

m)

Open-loop

PPFBang-bang-PPF

PIDBang-bang-PID

NonlinearBang-bang-nonlinear

LQRBang-bang-LQR

39/53 © D J Inman

Power consumption for each of the 8 controllers

0 0.2 0.4 0.6 0.80

50

100

150

200

Time(s)(a)

Inst

.Pow

er(m

W)

0 0.2 0.4 0.6 0.80

50

100

150

200

Time(s)(b)

Inst

.Pow

er( m

W)

0 0.2 0.4 0.6 0.80

50

100

150

200

Time(s)(c)

Inst

.Pow

er(m

W)

0 0.2 0.4 0.6 0.80

50

100

150

200

Time(s)(d)

Inst

.Pow

er( m

W)

PPFBang-bang-PPF

PIDBang-bang-PID

NonlinearBang-bang-nonlinear

LQRBang-bang-LQR

40/53 © D J Inman

Summary of the average power used by each controller (experimental):

The best choice of controller for use with harvested energy

41/53 © D J Inman

SELF CHARGING STRUCTURES AND COMBINED EFFECTS: piezoelectric, solar and thermal energy harvesting with flexible thin-film batteries.

Aluminum substructure components

Flexible thin-film battery

Piezoceramic patch

Flexible solar panel

Heat sink

Thermoelectric generator

Heater

New generation self-charging structures with flexible piezoceramic, solar panel and battery layers


Thermal power

Vibration power

Solar power

Thin-film battery

Regulator circuit for

impedance matching

Efficient circuit design for battery charging

42/53 © D J Inman

Extending the concept of the self charging structure to solar and thermal energy harvesting


Piezoelectric transducer

Flexible solar panels

Thin film Battery

70 mW

30 mW/g2

10 mW

43/53 © D J Inman

3.E/M CHARACTERIZATION: of the multifunctional system for different vibration, solar and thermal energy levels.

Different stages of fabrication

Aluminum (innermost) Flexible solar panel

(outermost)

Thin-film battery

Base excitation using a shaker

Solar power (outdoor) Thermal power Vibration power at resonance

Load resistance [ohms]

Pow

er [m

W]

Tem

pera

ture

[o C]

Load resistance [ohms]

Power [m

W]

hot side

cold side

temperature difference

power output

Piezo layer

Assembly: 4” x 1” x 0.06”

44/53 © D J Inman

Circuits and Implementation

Thermoelectric Harvester

PZT vin

+

-

iin

vS

+

-vo

io

iL

Ld Co Ro

Rin+

-

R1

R2

R3

RC1

RC2CC

vL

+

-

Cin

M

D

DC

vin

+

-

iin

vo

io

iL

Ld Co Ro

Rin+

-

R1

R2

R3

RC1

RC2CC

vL

+

-

Cin

M

D

DC

Solar/TEG

vin

+

-

iin

vo

io

iL

Ld Co Ro

Rin+

-

R1

R2

R3

RC1

RC2CC

vL

+

-

Cin

M

D

DC

Solar/TEG

Piezoelectric Harvester

Solar Harvester

Switching Impedance Matching Circuits

Thin-film Battery #1

Thin-film Battery #2

45/53 © D J Inman

UAV APPLICATION: embed piezoelectric and thin-film battery into the wing spar of a UAV to form a multifunctional spar capable of powering a local sensor

Multifunctionality Load bearing + Power generation + Energy storage

Self-charging structuresembedded in wing spar

Low-power sensor node being powered locally by harvesting system

• Incorporate energy harvesting devices and novel storage elements into UAVs.

• Provide local power source for low-power sensors in aircraft

• Flight endurance should remain unchanged with addition of harvesting

46/53 © D J Inman

Low Frequency MEMS Cantilever Harvesting

MEMS scale demands a cantilever of short lengthThe resonance frequency of a cantilever is inversely proportional to the square of its length: short beams mean high frequenciesAmbient energy in most systems are lowA potential solution is to use a zigzag arrangement of cantilevers

49/53 © D J Inman

Five popular piezoelectric ceramics have been considered (PZT-5A, PZT-5H and 3 single crystal types).

An order of magnitude difference for the d31 values of single crystals is not the case for their effective e31 values due to the effect of elastic compliance.

Compliance|d31|

50/53 © D J Inman

Comparison of these 5 Bimorph Cantilevers:For resonance excitation, the peak power does NOT differ by an order of magnitude as d31 does.

For different dynamic flexibilities at resonance, the peak power for resonance excitation does not depend much on d31 .

When the dynamic flexibilities of the bimorphs are artificially made identical, the maximum power outputs for resonance excitation become very similar. Larger power outputs of the single crystal bimorphs are due to their larger dynamic flexibilities (rather than their very large d31 constants).

Mechanical damping (hard to control due to clamping conditions and adhesive layers) can change the entire pictureModifications in the harvester’s geometry might be much more effective than the type of active material used.

Bimorph with larger d31 gave less power!

51/53 © D J Inman

Effect of the material properties of PZT for Harvesting Applications

It has been observed from the theory that d31 never appears solely and is always in multiplication with the elastic stiffness (importantly, larger d31 of single crystals comes with lower elastic stiffness – hence the effective e31 values have the same order of magnitude).

The dynamic flexibility and mechanical damping of a piezoelectric energy harvester can be much more important than the active material being used.

Dynamic flexibility of a piezoelectric energy harvester basically depends on the mechanical design. It is more difficult to control the mechanical damping due to clamping conditions and/or adhesive layers (and it should be noted that mechanical damping can change the entire picture!..).

These results might help those in the material side of piezoceramic development produce better materials for harvesting.

52/53 © D J Inman

Summary and ConclusionsIntroduced the concepts of harvesting ambient waste energy vibrationIllustrated how nonlinearity can be used to enhance energy harvestingPresented several examples of the usefulness of harvesting such low level energyTo review most applications are in charging batteries for sensor type applications (wireless sensing and monitoring)Its not safe the world renewable energy but rather “save the batteries” level and is an enabling technology for wireless systems

Thanks for your attention (sponsors follow)

53/53 © D J Inman

Acknowledgements

Questions/comments?

The support from the US Air Force Office of Scientific Research MURI under grant number F 9550-06-1-0326 “Energy Harvesting and Storage Systems for Future Air Force Vehicles” and “Simultaneous Vibration Suppression and Energy Harvesting” Grant number FA9550-09-1-0625, both monitored by Dr. B. L. Lee and the US National Institute of Standards and Testing is gratefully acknowledged.

The authors also gratefully acknowledge the support of CAPES (Brazil) and the GR Goodson Professorship Endowment.

energy harvesting and applications

Documents