Energy microgenerators for mobile devices powering

1
Energy microgenerators for mobile devices
powering

• Francesco Cottone
• Virgo Thermal Noise Group
• Seminario di passaggio al III anno
• Dottorato in Fisica XX ciclo
• Università di Perugia
• 17-10-2006

2
Outline

• Introduction mobile powering problem
• Microgenerators vibration noise as energy source
• Finite Elements modelling of linear and
  non-linear oscillators
• Conclusions and forthcoming work

3
ICT trends and energy requirement

• Moores Law transistors etc doubling every 1-2
  years

• Bells Law New computing class every 10 years

pervasive and near-invisibility computing!!!
Faster, Smaller, Numerous
4
Mobile powering
Wireless Sensor Network
source
Streaming Data to/from the Physical World
destination
Multihop Networking
5
Applications exsamples of WSN

• Environmental Monitoring
• Habitat Monitoring
• Integrated Biology
• Structural Monitoring

• Interactive and Control
• RFID
• Medical biosensors
• Emergency medial response
• Pursuer-Evader
• Intrusion Detection
• Automation
• Interactive museum exhibits
• Military applications

6
Wireless Sensors development
Smart Dust Technology (U.C. Berkeley)
MEMS LASER SCANNER 6 azt - 3 elev degrees

• Laser mote
• 650nm laser pointer
• 2 day life full duty

Operative range From 20 m to 20 Km !!

• Atmel Microprocessor
• RF Monolithics transceiver
• 916MHz, 20m range, 4800 bps
• 1 week fully active, 2 yr _at_1

7
Goal requirements for mobile sensors powering

• Small (lt1cm3)
• Lightweight (lt100 gr)
• Low Power (lt100 µW)
• Long-lasting (2-10 yr)
• Inexpensive (lt1 )
• Low data rate
• wireless platforms
• Flexibility

Spec 6/2003 Mote on a chip INTEL
Present (cubic centimeter)
Future (cubic sub-millimeter sub-micrometer)
8
Present-day power requirements for computing
electronics
Current challenge (prototypes are below 100µW of
mean power requested)
Advances in IC (Integrated Circuit) manufacturing
and low power circuit design and networking
techniques have reduced the total power
requirements of a wireless sensor node to well
below 1 milliwatt. Chandrakasan et al, 1998,
Davis et al, 2001
9
Comparison onEnergy and Power sources
S. Roundy 2
10
Batteries Vs Renewable Energy Sources
S. Roundy 1
11
Batteries Vs Renewable Energy Sources

• If outdoor sunlight, or relatively intense indoor
  light it available, solar cells appear to be the
  best alternative.
• Solar cells are a mature technology and a mature
  research area.
• If projected lifetime is longer than 1 year,
  vibrations offer an attractive alternative for
  certain environments. It was therefore decided
  to pursue research into the conversion of
  vibrations to electricity.

12
Vibrations Noise as Energy source
Base of Milling Machine
Microwave Oven Casing
Displacement vs. Frequency
Displacement vs. Frequency
Acceleration vs. Frequency
Acceleration vs. Frequency
S. Roundy 1
13
Existing methods for vibration energy scavenging
Capacitive Change in capacitance causes either
voltage or charge increase.
Inductive Coil moves through magnetic field
causing current in wire.

• Piezoelectric
• Strain in piezoelectric
• material causes a charge
• separation (voltage across
• capacitor)

Amirtharajah et. al., 1998
14
Energy storage density comparison

• Electrostatic transducers are more readily
  implemented in standard micro-machining
  processes.
• Electrostatic trasducers require a separate
  voltage source (such as a battery) to begin the
  conversion cycle.
• Electromagnetic transducers typically output AC
  voltages well below 1 volt in magnitude (too low
  power density)

15
MEMS Capacitive Converters
A MEMS in-plane gap-closing test structure
fabricated in the microlab at UC Berkeley.
full size test device tungsten mass
4mm
Interdigitated fingers
16
Piezoelectric Mechanical-electrical conversion
Constitutive Equations
d strain s stress Y Youngs modulus d
piezoelectric coeff. D electrical
displacement e dielectric constant E electric
field
17
Piezoelectric Converters
Where k equivalent spring stiffness of beam
m attached proof mass bm damping
coefficient a1 geometric constant a2
geometric constant d31 piezoelectric
coefficient tc thickness of one
piezo-ceramic layer VR voltage across load
18
MEMS µPG PIEZO at micron scale
4 N. E. duToit et al. Cambridge MIT-Institute
2005
3 Y. Ammar et al. TIMA Lab Grenoble VIBES
project, october 2005
2
19
Driven mechanical oscillator dynamics
damping ratio
Natural frequency
Freq. Resp. function
Resonance frequency
20
Simple model of mechanical to electrical energy
conversion

• If acceleration magnitude is relatively constant
  with frequency, output power is inversely
  proportional to frequency.
• There is an optimal level of electrically induced
  damping that is designable.
• It is better to have too much electrical damping
  than too little.

z transducer displacement y magnitude of input
damping
Power assuming w wn. A acceleration
amplitude of input vibrations m proof mass
Wiliams and Yates 6
21
Power efficiency of piezoelectric generators
S. Roundy et. Al, Pervasive Computing, 2005

• Narrow frequency
• Actively and passive tuning resonance frequency
  of generator
• Wide bandwith designs
• Proof mass m
• Improve the strain from a given mass
• Coupling coefficient k31
• Thin-film piezoelectric-material properties

for
22
The Idea Non-linear oscillators
Bi-stable systems
barrier
w
Potential barrier height
23
Non-linear oscillators2D Finite Element Model
of inverted pendulum
f16.58Hz first resonance mode
Frequency response
Harmonic excitation force _at_ first mode frequency
20000 dof
24
2D FEM of an inverted pendulum
repulsive force to create double well potential
Repulsive magnetic force, harmonic excitation
25
2D FEM of an inverted pendulum
Harmonic Excitation at 6.58Hz
broadening of the resonance peak
W0.1
W0.001
26
2D FEM of an inverted pendulum
Gaussian Noise Excitation µ0 s0.5 - 1 Fs100Hz
27
2D FEM of an inverted pendulum
28
Energy Efficiency

• Noise Excitation
• Increment up to thirty times when the barrier
  height increase relative to no barrier case
• Harmonic Excitation
• not so much improvement relative to the situation
  of a tuned external force at resonance frequency
• almost one order of magnitude relative to out of
  resonance excitation

29
2D FEM of a bi-stable membrane
vertical compression
Differential Pressure

• Stainless steel membrane of 1mm thickness and
  20mm length
• contracting clamping of 0.1mm we induce bistable
  response on displacement when we applicate
  external force (for example harmonic)

30
2D FEM of a bi-stable membrane

• Possible applications
• Piezoelectric micro generator
• Using PZT material for membrane
• Pressure sensor
• Actual pressure sensors are based on position
  measurement of the membrane so they are affected
  by aging of material therefore a decalibration in
  time

31
Preliminary experimental tests
Experimental Setup of driven piezo inverted
pendulum
32
Preliminary experimental tests
digital simulation and preliminary exp. results
System response to simulated floor vibration
with different spectral properties.
33
Conclusions

• Piezoelectric converters appear to be the most
  attractive for meso-scale devices with a maximum
  demonstrated power density of approximately 200
  mW/cm3 vs. 100 mW/cm3 for capacitive MEMS
  devices.
• If external excitation is harmonic and its tuned
  to resonance frequency of the device we can
  obtain high mechanical energy transfer but it is
  possible to reach it even with a bi-stability
  dynamics
• For an external excitation with noise or out of
  resonance, FEA shows that a non-linear bistable
  beahviour amplify up to a factor 30 or more the
  vibrational energy extraction
• Most of vibrational energy sources live al
  relative low frequency (for example near hundred
  Hz) so its hard to tune the resonance mode of
  the device to external frequencies

34
Forthcoming works

• We must compare FEA data with further
  experimental measurements for macro and
  sub-millimetric piezo generators
• Scalability is an important issue for
  micropowering
• We want to implement a real sub-centimeter and
  sub-millimeter prototype and characterize its
  power conversion capability
• in order to optimize energy harvesting we need to
• investigate new piezo material with high quality
  factor and high piezoelectric couling constant
• characterize response at various vibrational
  sources

35
Bibliography

• S. Roundy, P.K. wright, and J. Rabaey, Energy
  Sacvenging for Wireless Sensor Networks with
  Special Focus on Vibrations, Kluwer Academic
  Press, 2003.
• S. Roundy et al., Improving Power Output for
  Vibration-Based Energy Scavengers, Pervasive
  Computing IEEE 2005.
• 3 Y. Ammar et al. TIMA Lab Grenoble EUSAI
  conference
• N. E. duToit et al. Cambridge MIT-Institute 2005
• S. Roundy et al. A study of low level vibration,
  Computer Communication 2003
• Wiliams, Yates, Analysis of a micro-electric
  generator for microsystems, proceeding of the
  Trasducers 95/Eurosensors (1995) 687-695
• www.intel.com
• http//www.ife.ee.ethz.ch/tvonbuer/energyharvesti
  ng_links.html
• http//www.darpa.mil/mto/mpg/index.html