Robust Design of Piezoelectric Energy Harvesting Devices using Multiobjective Optimization Techniques
Palavras-chave:
energy harvesting, uncertainties, piezoelectric sensors, robust devices, optimizationResumo
For the robust design of piezoelectric energy harvesting devices, it is necessary to estimate the mean
and variance of the harvesting performance due to uncertainties in device parameters. This is done here using
Polynomial Chaos Expansions (PCE). Aiming at overcoming the total computational cost required for the robust
optimization, a discussion on the selection of the most relevant uncertain parameters and on the degree and con-
vergence of the PCE is performed. Then, with a satisfactory modeling choice, a multi-objective optimization using
Non-dominated Sorting Genetic Algorithm (NSGA-II) is performed to determine Pareto fronts and box-plots that
allow to choose the harvesting devices with better compromise between performance mean and dispersion.
