Computational model for power optimization of piezoelectric vibration energy harvesters with material homogenization

•Optimization of a piezoelectric energy harvester using homogenization is presented.•Power output of unimorph and bimorph vibration energy harvesters are optimized.•Single crystals, polycrystals and piezocomposites are considered.•Effective material properties are computed using homogenization implemented in FEM.•Orientation and volume fraction of piezoelectric phase are the design variables. Piezoelectric vibration power harvesters are being studied in the literature since they have high energy conversion from mechanical vibrations. A computational model that optimizes piezoelectric vibration energy harvester output power using homogenization of piezoelectric material is presented in this work. This computational model allows piezoelectric material tailoring to create piezoelectric vibrational energy harvesters capable of producing higher electrical power. The materials considered in the study are single crystal and polycrystals of BaTiO3 and PZN-4.5%PT, and piezopolymer PVDF-TrFE and the piezocomposites of these materials. The computational model is used to optimize the harvester power output of the unimorph vibration harvester configuration. The harvesters are modelled using the finite element method which is validated comparing analytical results for four traditional harvester configurations, viz., unimorph, bimorph, longitudinal generator and transverse generator. Single crystals, polycrystals and piezocomposites made by piezoceramic and piezopolymer materials are considered in the optimization procedure. Polycrystalline and piezocomposite properties are computed through a computational model based in the homogenization theory, which is implemented using the finite element method. Electrical resistance is used as the surrogate for the electrical machine connected to the harvesters. The design variables considered are the crystal orientation for single crystal materials, microstructural orientation distribution of the grains for polycrystalline materials, the piezoceramic material volume fraction and piezopolymer orientation for piezocomposites and/or the circuit resistance. A simulated annealing algorithm based in Metropolis algorithm is used as the optimizer. Several examples are presented and discussed considering excitations near as well as far away from resonance frequency. Harvesters with material composites having optimal material configurations that deliver enhanced electrical power have been identified.