Modelling, simulation and optimisation of a piezoelectric energy harvester

The power generation efficiency of piezoelectric energy harvesters is dependent on the coupling of their resonant frequency with that of the source vibration. The mechanical design of the energy harvester plays an important role in defining the resonant frequency characteristics of the system and therefore in order to maximize power density it is important for a designer to be able to model, simulate and optimise designs to match new target applications. This paper investigates a strategy for the application of soft computing techniques from the field of evolutionary computation towards the design optimisation of piezoelectric energy harvesters that exhibit the targeted resonant frequency response chosen by the designer. The advantages of such evolutionary techniques are their ability to overcome challenges such as multi-modal and discontinuous search spaces which afflict more traditional gradient-based methods. A single case study is demonstrated in this paper, with the coupling of a multi-objective evolutionary algorithm NSGA-II to a multiphysics simulator COMSOL. Experimental results show successful implementation of the schema with all 5 experimental tests producing optimal piezoelectric energy harvester designs that matched the desired frequency response of 250 Hz.