Controle adaptativo robusto de tempo discreto por modelo de referência e controle repetitivo

Abstract

This work is motivated by the recent advances in the electrical-electronics and in process control field, that extensively use microprocessing systems and require apropriated good performance control techniques with the ability of automatic tuning to deal with unmodeled dynamics and disturbances, using only input/ouput measurement information. Firstly, a literature review is presented concerning some control techniques and related with goal of this work. In the sequence, a robust model reference adaptive control (RMRAC) algorithm is presented and the stability analysis of two adaptation algorithms is carried out. To allow the RMRAC controller implementation in microprocessed based systems a discrete-time robust model reference adaptive controller (DRMRAC) and a discretized modified recursive least squares adaptation algorithm are proposed. Besides, to minimize steady state tracking errors due to periodic disturbances, a plug-in repetitive controller (RPC) is added to the DRMRAC controller. Moreover, an adaptation algorithm is proposed to tune automatically the repetitive controller gain, resulting in an adaptive repetitive control algorithm (ARPC). The stability analysis of proposed control algorithms is carried out. Results of their implementation to control the output voltage of an uninterruptible power supply (UPS) are presented to show the good performance ones. A descentralized approach to multivariable systems (MIMO) using the DRMRAC-RPC controller, a dq transformation and considering the subsystems interconections as unmodeled dynamics is presented. The performance verification of the descentralized MIMO controller is carried out througth the implementation of the same one to control the output voltages of a three-phase uninterruptible power supply.