Identification of fractional order circuits from frequency response data using seeker optimization algorithm

Abstract

Linear circuits and systems are generally described by traditional differential equations and integer order transfer functions based on the assumption that the dynamics are lumped and time invariant. However, as compared to the conventional integer order calculus, many dynamical systems are better represented by fractional calculus with interaction among the variables modelled by fractional integration and/or fractional differentiation. The present work proposes a generalized approach for the identification of fractional order systems in frequency domain using experimental data. To achieve the same, the system identification task has been framed as an optimization problem and solved using seeker optimization algorithm. The algorithm seeks to attain a set of system parameters for which the deviation between the simulated response of the identified system and experimental data is minimized. The proposed approach has been validated on a set of electrical circuits with varying configuration. The simulation and experimental results reveals that all of the test circuits are better represented by fractional order model, over a wide range of frequency.