Global optimisation approach for designing high-efficiency piezoelectric beam-based energy harvesting devices

The paper proposes a novel methodology for developing high-power energy harvesting gravity-based devices using an array of piezoelectric beams for wind energy applications. The methodology incorporates a global multidimensional constrained optimisation algorithm, which accounts for the physical size of the device, the physical, geometrical and electrical properties of the piezoelectric beams, and the power management circuit to increase the device's efficiency. As the beams are plucked sequentially, they vibrate out-of-phase, which consequently leads to charge cancellation issues. The paper proposes and incorporates an electrical circuit design to avoid such problems, being able to further increase the efficiency of the device by 35% when compared against the output from the standard energy harvesting (SEH) circuit with independent rectifiers. The proposed optimisation methodology is applied to the devices utilising flexible polyvinylidene fluoride beams. The developed dynamic numerical model of the beams’ vibration is validated using the experimental results and the results of a Finite Element Analysis. To study the electro-mechanical coupling of the beams, an electric circuit and the power management circuit are created and modelled in Matlab/Simulink software. The optimised device delivers 6 to 17 times higher energy output compared to the unoptimised device. The performance of this device was also compared to that of the device with much stiffer LiNbO₃ beams (Clementi et al., 2021, [1]) to demonstrate the direct applicability of such devices to power sensors and transmitter units for structural monitoring of wind turbine blades. It has been demonstrated that the LiNbO₃-beam device yields an energy output with one order of magnitude higher. The applied optimisation methodology enabled a 0.057 × 0.017 × 2 m³ dimension device to produce a power output in the range from 0.5 to 1 W depending on the blades’ speed, resulting in 1.06 mW/cm³ power density of the device.