High Efficiency, Nonlinear Vibration Energy Harvesting using Electrical Switching D. Mallick, A. Amann, S. Roy Tyndall National Institute Cork, Ireland Micro-energy 2017, Gubbio, Italy
Internet of Things (IoT) Wirelessly connecting Anything 2
Internet of Things (IoT) Wirelessly connecting Anything Trillion Sensor Visions How to POWER trillions of connected devices? Batteries Not suitable for fit-and-forget applications Solar Energy Harvesting Thermal RF Vibration Roadmap for the Trillion Sensor Universe, Internet of Things talk by Dr. Janusz Bryzek 3
Mechanical Energy Harvesting What releases mechanical energy Almost Anything Online database Real Vibrations - NiPS lab, University of Perugia, Italy. - EH Networks Database EPSRC Funded Network. Vibration data from more than 600 sources comprising Human Activities Machine Automobile/Aerospace Large Structures Neri et al., J. Intell. Mat. Syst. Struc., 23.18 (2012): 2095-2101. http://eh-network.org/data/ 4
Mechanical Energy Harvesting What releases mechanical energy Almost Anything 23% Sources single dominant, stationary frequency. 53% Sources contains single dominant, non-stationary frequency multiple dominant, stationary frequencies white/filtered noise - broadband R. Rantz & S. Roundy, SPIE SSMNEHM, International Society for Optics and Photonics, 2016. 5
Wideband Energy Harvesting Target: Wideband Operation Linear Nonlinear Equation of motion of simplest energy harvesting systems m x(t) + (c m + c e ) x(t) + F x = m y(t) Linear F x = kx Nonlinear F x = kx + k n x 3 Nonlinearity through stiffness Increase in Bandwidth No tuning needed; No array of devices needed Gain in output power P nonlinear 4 P linear π 6
Nonlinear Wideband Operation (Tyndall) Device Schematic Cross-section ICP- DRIE etched SOI spring structure Packaged Device Load Power vs Frequency Response 0.02g 100 μm 200 μm Electroplated Double Layer Copper Coil Bandwidth 82 Hz @ 0.5g with 2.8 µw D. Mallick et al., Journ. Microelectromech. Syst., 26(1): 273-282 (2017). 7 7
Low Frequency Operation (Meso-scale) Advantages of FR4 (Flame Retardant 4): Standard PCB material Low Young s Modulus (21 GPa) Useful for low frequency applications. Low cost!! Meso-scale Prototype Volume = 2.65 cm 3 Mass = 3 gm Bandwidth (BW) 10 Hz @ 0.5g Maximum Power ~ 0.5 mw @ 0.5g D. Mallick et al, Smart Mater. Struct., 24, 015013 (2015). 8
Nonlinear Hysteresis Frequency Domain Response Basin of Attraction Plots within Hysteresis a Up Sweep b c d Down Sweep Multiple steady state solutions - Hysteresis How to operate in the frequency varying environment? Blue Low Energy Red High Energy D. Mallick et al, Smart Mater. Struct., 24, 015013 (2015). 9
Surfing the High Energy Branch (I) Concept Implementation Switching voltage 2V During electrical energy injection Injected energy switches the state from actual to desired attractor Low Energy Branch High Energy Branch Maintains steady state without continuous energy input 0 10 20 30 40 50 Time (sec) D. Mallick et al, Phys. Rev. Lett., 119:197701 (2016). 10
Modelling of the Switching Scheme V A (t) = V OA sin 2πf A t for t i t t f 0 Otherwise V OA Amplitude f A frequency t i and t f - starting and ending times of the switching period The net current I C through the coil: I C x, t = I L I A = γ x R L R C R C V A (t)r C + γ xr A R C R A + R A R L + R L R C The coupled electromechanical equation of motion of the oscillator m x + 2mρω n x + kx + k n x 3 + γi C = m z 11
Surfing the High Energy Branch (II) Altering the well established frequency/amplitude scan responses Fixed Acceleration 0.5g Fixed Frequency 70 Hz 12
Surfing the High Energy Branch (II) Probabilistic Study on Switching Mechanism High Low D. Mallick et al, Phys. Rev. Lett., 119:197701 (2016). 13
Successful/Unsuccessful Switching Phase Space Diagram Initial state after switching stops 14
Successful/Unsuccessful Switching Yellow Successful switching Green Unsuccessful switching Fixed switching signal amplitudes (diamonds) - 5V (squares) - 15V (circles) - 25V Blue Low Energy Red High Energy 15
Surfing the High Energy Branch (II) Probabilistic Study on Switching Mechanism High Low D. Mallick et al, Phys. Rev. Lett., 119:197701 (2016). 16
Surfing the High Energy Branch (II) Evolution of Net Electrical Energy E 0 - Energy to apply switching signal once P S - Probability of successful switching in first attempt E T - Total energy spent to switch the state E T = P s E 0 k=1 k(1 P S ) k 1 = E 0 P S k - Number of attempts As P S ~ 0.8, E T - not very high D. Mallick et al, Phys. Rev. Lett., 119:197701 (2016). 17
Surfing the High Energy Branch (III) Response under randomly varying vibration Electrical Actuation Point R R : frequency/amplitude varying input excitation Increase in harvesting efficiency: 340 times D. Mallick et al, Phys. Rev. Lett., 119:197701 (2016). 18
Electrical Switching in MEMS EM VEH Switching voltage 8V Accl. - 0.5g Energy gap between High and Low energy branches is low Electromechanical coupling is very low as well Inefficient transfer of energy 19
Self-switching/Automatic Operation Proposed Scheme I Nonlinear Generator Power Conversion Circuit Control Circuit Energy Storage Extra Space for linear Generator What benefit in overall Power Density? Linear Generator 20
Self-switching/Automatic Operation (II) Proposed Scheme II Nonlinear Generator Power Conversion Circuit Switch Comparator V ref Power loss even when there is no multi-stability 21
Remarks Conclusions: Multi stability major challenge in nonlinear oscillator based applications Applications in Electronic Circuits, Photonics, Semiconductor Materials, Chemical Reactions, Cell Biology and many more Mechanical Energy Harvesting Huge improvement energy conversion efficiency in a real application environment. Method independent of device scale or transduction methods Future Work: Automatize the Mechanism: Self-controlled feedback loop development 22
Acknowledgement: SFI PI grant award (2012)- MEMS vibrational energy harvesting 11/PI/1201 Thank You Contact: dhiman.mallick@tyndall.ie 23