Jtam.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 49, 3, pp. 757-764, Warsaw 2011 ENERGY HARVESTING IN A MAGNETOPIEZOELASTIC SYSTEM DRIVEN BY RANDOM EXCITATIONS WITH UNIFORM AND GAUSSIAN DISTRIBUTIONS Grzegorz Litak, Marek Borowiec Technical University of Lublin, Department of Applied Mechanics, Lublin, Poland e-mail: g.litak@pollub.pl Michael I. Friswell, Sondipon Adhikari Swansea University, School of Engineering, Swansea, United Kingdom A simple magneto-piezoelectric system excited by random forces mo- delled with a double well potential is considered. System responses for different realizations of noise with uniform and Gaussian distributions are compared. The results show negligible differences in the regions of small and high noise intensity. A more noticeable difference can be se- en in the intermediate region of noise just below the transition from isolated single well oscillations to coupled double wells oscillations. Va- riations in themechanical displacement in this transition region indicate that the transition between these types of behaviour is broader for a uni- form noise excitation. Consequently, the system excited with Gaussian noise tends more clearly to one of the different solutions (i.e. motion in a single well or in bothwells) while the uniform noise case demonstrates intermittency with multiple solutions. Key words: energy harvesting, piezoelectric transducer, random exci- tation 1. Introduction In the energy harvesting process, energy is derived from external sources (e.g., solar power, thermal energy, wind or hydro energy, salinity gradients, and also kinetic energy). This energy is captured and stored by autonomous devices, such as those used in wearable electronics and wireless sensor networks. In recent years, energy harvesting has attracted great attention as the energy generated can be used directly or used to recharge batteries or other storage devices, which enhances battery life (Anton and Sodano, 2007). 758 G. Litak et al. Many of the proposed devices use the piezoelectric effect as the trans- duction method (Arnold, 2007; Beeby et al., 2007). These devices are usually implemented as patches on cantilever beams and designed to operate at reso- nance conditions. The design of an energy harvesting device must be tailored to the ambient energy available. For a single frequency excitation the resonant harvesting device is optimum, provided it is tuned to the excitation frequency (Erturk et al., 2009; Litak et al., 2010; Stanton et al., 2010). One should also note that individual small devices may be combined in arrays to produce a larger andmore powerful device. 2. The magneto-piezoelectric harvester The system of our present investigation consists of a ferromagnetic cantilever beam that is excited at the support (Fig.1). Two permanent magnets are located symmetrically on the base near the free end, and the static system can have five, three or one equilibrium positions depending on geometry of the system (Erturk et al., 2009; Litak et al., 2010) and, in particular, the distance between the beam and the magnets. Fig. 1. Schematic of the piezomagnetoelastic device (Litak et al., 2010) In thepresentwork,weare interested in the casewhen the systemhas three equilibrium positions, two of which are stable, and the mechanical system is characterized by the classical double well potential. The non-dimensional equations of motion for this system (Erturk et al., 2009) are ẍ+2ζẋ− 1 2 x(1−x2)−χv=F(t) (2.1) and v̇+λv+κẋ=0 (2.2) Energy harvesting in a magnetopiezoelastic system... 759 where x is the dimensionless transverse displacement of the beam tip, v is the dimensionless voltage across the load resistor, χ is the dimensionless pie- zoelectric coupling term in the mechanical equation, κ is the dimensionless piezoelectric coupling term in the electrical equation, λ ∝ 1/RlCp is the re- ciprocal of the dimensionless time constant of the electrical circuit, Rl is the load resistance, and Cp is the capacitance of the piezoelectric material. The non-dimensional excitation F(t) is proportional to the base acceleration on the device, and is assumed to be uniform or Gaussian white noise, with zero mean and specified variance. 3. The harvester response to random excitation The system parameters are taken as (Erturk et al., 2009; Litak et al., 2010): ζ = 0.01, χ = 0.05, and κ = 0.5, while λ was 0.01. The excitation F(t) is stationary uniform or Gaussian white noise with standard deviation σF . Equations (2.1) and (2.2) are integrated using the fourth order Runge-Kutta- Maruyama algorithm (Naess andMoe, 2000; Litak et al., 2010). The standard deviations of the displacement x and the voltage v are calculated for a range of excitation noise amplitudes σF for both the uniform and Gaussian noise distributions. Figure 2 shows the signal to noise ratio σx/σF as a function of noise inten- sity σF for theGaussian (Fig.2a) and uniform (Fig.2b) distributions, respec- tively. For each value of σF depicted in this figure, five different realizations of noise were used. The simulated results for the different noise distributions are similar, and only small differences appear in the regions of small and high noise intensity. However, there is a noticeable difference in the intermediate region of noise just below the transition from isolated single well oscillations (for small σF) to coupled double wells oscillations (for large σF). The beam displacement in this region indicates that the transition between these types of behaviour is broader in the case of a uniform noise excitation. The form of the displacement response can be determined from the mean value of displacement (Fig.3), which shows slightly increased concentration close to the unstable equilibrium point x = 0 in the case of a uniform noise distribution (Fig.3b) above the transition region. The explanation is that the systemexcitedby theuniformnoisedistributionprefersmore frequenthopping between the potential wells. For further clarification, Fig.4 shows the number of hops (motion from one potential well to the other) between the potential wells for two types of noise. Clearly, the number of hops is zero for lower 760 G. Litak et al. Fig. 2. The displacement signal to noise ratio σx/σF versus noise intensity σF , where σx is the standard deviation of the beam displacement and σF is the standard deviation of the noise excitation for different noise distributions (a) Gaussian and (b) uniform noise levels. At a certain critical level, the number of hops starts to increase approximately linearlywith the noise intensity. The larger the number of hops, the higher the hopping frequency in the response spectrum. Theultimate aimof the harvester is to generate energy. Figure 5 shows the variance of voltage σ2v for the two noise distributions. Assuming the voltage has zeromean, this will approximate the energy generated. It is clear that the variance of the voltage is not sensitive to different noise distributions. 4. Conclusions This paper has extended the analysis in our previous paper (Litak et al., 2010) by comparing the effects of Gaussian and uniform noise distributions on the harvesting system. For the range of parameters investigated, the beam displacement results only differ in the region of the system response where the system transitions from single well vibrations to vibrations characterized Energy harvesting in a magnetopiezoelastic system... 761 Fig. 3. The mean values of displacements for different noise distributions (a) Gaussian and (b) uniform Fig. 4. The number of hops between potential wells for different noise distributions (during the investigated simulation interval) (a) Gaussian and (b) uniform 762 G. Litak et al. Fig. 5. The variance of the generated voltage, σv, for different noise distributions (a) Gaussian and (b) uniform by hopping between the potential wells. Note that in the uniform excitation case, the escape from the potential well is better defined because the noise is limited toagivenband. Incontrast, for theGaussian system, a large amplitude excitation may occur due to the distribution function tails. Furthermore, the lack of tails for the uniform excitation breaks the system ergodicity. In the short time scale, themost important effect is that the noisy force disturbances are usually larger in the case of the uniform noise distribution. Note that the differences between the investigated noise distributions may be larger for different values of λ, which defines the relaxation properties of the electrical part of the system. Finally, very similar responseswere obtained in termsof the voltage output (Fig.5); this implies that the broadband noise assumption in the previous paper (Litak et al., 2010) is a reliable approach to optimize the system design. Daqaq (2011) also studied theGaussianwhite noise excitation of a bistable inductive generator. He showed that in the limit of higher noise intensity, which corresponds to σF > 0.05 in our work, the shape of the double well potential is not important. High excitation levels lead to a large amplitude system responsewhere the potential barrier is regularly traversed. In contrast, Energy harvesting in a magnetopiezoelastic system... 763 our results consider the crossover betweenweakand fairly strong levels of noise intensity where the intermittency may play an important role. Acknowledgements The authors gratefully acknowledge the support of the Royal Society through International Joint Project No. HP090343. GL would like to thank Prof. Utz Von Wagner for useful discussions. 5. References 1. Anton S.R., Sodano H.A., 2007, A review of power harvesting using piezo- electric materials (2003-2006), Smart Materials and Structures, 16, R1-R21 2. Arnold D.P., 2007, Review of microscale magnetic power generation, IEEE Transactions on Magnetics, 43, 3940-3951 3. Beeby S.P., Torah R.N., Tudor M.J., Glynne-Jones P., O’Donnell T., SahaC.R.,RoyS., 2007,Amicro electromagnetic generator for vibration energy harvesting, Journal of Micromechanics andMicroengineering, 17, 1257- 1265 4. Daqaq M.F., 2011, Transduction of a bistable inductive generator driven by white and exponetially correlatedGuassiannoise,J. Sound andVibration,330, 2554-2564 5. Erturk A., Hoffmann J., Inman D.J., 2009, A piezomagnetoelastic struc- ture for broadband vibration energy harvesting,Appl. Phys. Lett., 94, 254102 6. Erturk A., Inman D.J., 2009, An experimentally validated bimorph can- tilever model for piezoelectric energy harvesting from base excitation, Smart Materials and Structures, 18, 025009 7. Erturk A., Inman D.J., 2008, A distributed parameter electromechanical model for cantileveredpiezoelectric energyharvesters,Journal of Vibration and Acoustics-Transactions of the ASME, 130, 041002 8. Litak G., Friswell M.I., Adhikari S., 2010, Magnetopiezoelastic energy harvesting driven by random excitations,Applied Physics Letters, 96, 214103 9. Naess A., Moe V., 2000, Efficient path integration methods for nonlinear dynamic systems,Probab. Eng. Mech., 15, 221-231 10. Stanton S.C., McGehee C.C., Mann B.P., 2010, Nonlinear dynamics for broadband energy harvesting: Investigation of a bistable piezoelectric inertial generator,Physica D, 239, 640-653 764 G. Litak et al. Pozyskiwanie energii w piezo-magnetycznym układzie sprężystym, pobudzanym siłą stochastyczną o rozkładzie jednorodnym i normalnym Streszczenie W pracy analizowany jest prosty układ piezo-magnetyczny, pobudzany losowo z potencjałem o dwóch studniach. Porównywane są odpowiedzi układu przy róż- nej realizacji szumu, o rozkładzie jednorodnym i normalnym (Gaussowskim).Wyniki przedstawiająnieznaczne różnicew obszarachniskiej i wysokiej intensywności szumu. Bardziej zauważalną różnicęmożnadostrzecwobszarzepośrednimszumu, tużponiżej przejścia z oscylacji w pojedynczej studni potencjału do oscylacji w dwóch sprzężo- nych studniach. Zmiany pracy układu w tym obszarze sygnalizują, że obszar przejść pomiędzy takimi typami rozwiązań jest szerszyprzypobudzaniu szumemo rozkładzie jednorodnym. Natomiast układ pobudzany szumem o rozkładzie normalnym wyraź- niej wykazuje tendencje do pracyw zakresie jednego z typów rozwiązań.W rezultacie przy szumieGaussowskimukładdążydo ruchuwobrębie tylko jednej lubdwóch stud- ni potencjału, podczas gdy w obecności szumu jednorodnego, w zachowaniu układu pojawia się zjawisko intermitencji w realizacji dwóch rozwiązań. Manuscript received February 28, 2011, accepted for print April 18, 2011