Magneto-mechanical Optimization and Analysis of a Magnetostrictive Cantilever Beam for Energy Harvesting

The requirement of self-powered stand-alone sensors or wireless sensor networks (WSN) with a “fit and forget” paradigm, that is, without the need of batteries, seems to have attracted the interest of technological research in last years. Activity in development of devices, Known as Energy Harvesters, able to transform ambient energy into a kind of usable energy were boosted. Among this quite large set of solutions, those able to transform mechanical vibrations into electric power and referred to as Kinetic Energy Harvesters (KEH) are among the most known and studied, [1]. Such special set of self powered systems found a noteworthy interest in the health monitoring of civil structures or in automotive or railway applications, where, in the latter, the KEH potentialities have been identified for low cost self power sensors in freight railway wagons, [2]. Among them, one solution is represented by the cantilever configuration exploiting magnetostrictive materials, [3], based on Fe-Ga (Galfenol) alloys, showing interesting mechanical and magnetostrictive properties, [4]. Vibrations energy spectrums are complex and usually distributed over a quite large frequency range, while, conversely, cantilever KEH work in a relatively narrow band and this requires both modelling and experimental effort in order to provide an “optimal” system tuning. In this paper as a first step, different configurations employing one or more Fe-Ga strips over an Al substrate have been developed and tested through the analysis of a converted RMS power vs. frequency in order to detect the geometric and physical parameters affecting resonance frequency, bandwidth and converted power by Villari effect, [2]. The selected cantilever is implemented by bonding two single Galfenol strips 120X15X0,3 mm to a 120X15X2 mm Aluminium foil. The latter guarantees a better mechanical resistance to cyclic loads and provides a larger stiffness yielding to a consequent increase of the resonance frequency. The device has been equipped by a 300 turns coil, as it can be observed in Fig. 1. Experimental tests were performed with an electrodynamic shaker (Fig. 2a) in the 50–200 Hz frequency interval with sinusoidal acceleration in the ${1-4 g}$ amplitude range [5]. The excitation level has been detected by a reference accelerometer, while the beam bending is measured by a strain gauge based system. The effect of a resistive load on the RMS converted power is shown in Fig. 2b) where the power response with different resistive loads and 4g acceleration is shown. The best result of 305mW is achivied with a 9,9 W. The low energy conversion required the design of a suitable magnets set in order to supply a sufficient magnetic bias to the active material with the aim to increase the converted power [6]. In order to increase the flux captured by the coil, two magnets were attached to the structure [7]. The field generated by two magnets was measured spanning from 43 kA/m on the upper strip down to 3 kA/m on the lower strip. In Fig. 2c) the effect of the magnetic bias on the RMS converted power is shown. However, it should be outlined that the applied bias is the result of an optimization procedure since, as known, high magnetic biases pushes the material towards saturation and energy conversion is no longer observed [6]. Considering the importance of the bias, tests were carried out by varying the number of magnets (Fig. 2d). In this case it is important to evaluate the position of the magnet on the beam as it influences the resonance frequency and the magnetoelatic coupling [8]. These results seem quite encouraging, due to the relatively high converted energy (37mW) and the simple device structure. However, a thorough and accurate modelling analysis in connection to the experimental data, in order to quantitatively show the dependence of the converted power on geometric and physical parameters will be discussed in the full paper.