A novel computing paradigm for parameter identification of piezoelectric energy harvesting systems subjected to uncertain loads

In this paper, a recent innovative computing paradigm able to solve non linear equations systems and scalar or multi-objective optimization problems, is used for parameter identification of a second order dynamic system that describes the electromechanical response of energy harvesting devices subjected to uncertain loads. The identification problem is recast through a proper set of ordinary dynamic differential equations which equilibrium points correspond to the problem solutions. Exponential asymptotic convergence to equilibrium points is guaranteed from the Lyapunov theory. The main advantage with respect to standard iterative algorithms (Newton-Raphson etc.) is the possibility to deal with some uncertainty in the measured data. Two case studies are analyzed to assess the effectiveness of the proposed algorithm: identification of damping coefficients in second order electromechanical systems under harmonic base excitations and real time identification in presence of random base motion.