Optimal fractional order adaptive controllers for AVR applications

Abstract

This work presents strategies for fractional order model reference adaptive control (FOMRAC) and fractional order proportional-integral-derivative control (FOPID) applied to an automatic voltage regulator (AVR). The paper focuses on tuning the gains and orders of the FOPID controller and the gains and orders adaptive laws of the FOMRAC controller, with the goal of minimizing non-linear and high dimensionality objective functions, using sequential quadratic programming (SQP), particle swarm optimization (PSO), and genetic algorithms (GA). Two models used for AVR have been studied and reported in the literature and are the bases of the three case studies reported in this paper. To analyze the advantages and disadvantages of the proposed MRAC, comparisons are made with the previous results, i.e. with the results obtained by a PID controller and an MRAC controller optimized by GA. We demonstrate through some performance criteria that fractional order controllers optimized by the PSO algorithm improve the behavior of the controlled system, specifically the robustness with respect to model uncertainties, and improvements with respect to the speed convergence of the signals.