Controle adaptativo robusto combinado com regulador linear quadr atico aplicado a condicionadores de energia

Abstract

This work proposes an algorithm that combines two control laws with di erent compensation characteristics in order to obtain a robust and high performance control: A Linear Quadratic Regulator aided by a Kalman lter or LQRKF and a Robust Model Reference Adaptive Controller or RMRAC. The design of each used controller is performed independently and the combined algorithm is composed by the sum of the two separated control laws. The main motivation for the union of these two strategies is to obtain an algorithm that carries the properties of each controller involved. The RMRAC scheme is designed to ensure robustness against parametric variations and unmodeled dynamics, while the LQRKF regulator is designed to reduce the distortion e ects caused by the voltage at the Point of Common Coupling or PCC. Moreover, the LQRKF controller presents a fast transient response. The combined control structure ensures a proper tracking performance, considering di erent types of adversities, commonly found in real systems, such as high line impedance, noise measurement, frequency variation of network, unbalance of currents, among others. Moreover, it is developed the stability proofs of the proposed control system, from a candidate Lyapunov function. It is also shown the robustness of the control system, since regarded the design constraints. Finally, in order to verify the consistency of the proposal, extensive tests, simulation and experimental results are presented, by implementing the control system using a Digital Signal Controller or DSC for the control of a power conditioner comprising a shunt three-phase PWM inverter connected to the network through an inductive coupling lter.