Algoritmos de controle para melhoria da regulação de tensão de alternadores de rotação variada

Abstract

This work presents a study of control algorithms to improve the voltage regulation of varied rotation alternators. Many voltage control algorithms are presented in the literature, but most of them do not present enough results to guarantee their usage in locomotive applications. In addition, the generator speed change is not addressed on the mathematical models used in voltage regulation control algorithms in the literature. Therefore, a comparative study is necessary, following some pre-defined metrics, so that a technique could be proposed and then applied and evaluated in a digital Automatic Voltage Regulator (AVR) controlled by an embedded system in a real generator. This study evaluates two PID controller optimization algorithms with off-line tuning, PSO (Particle Swarm Optimization) and GA (Genetic Algorithm). The controller tuning is performed in a simulation environment, where the mathematical model of the alternator is obtained through tests. The algorithms are compared considering the rise time, settling time, steady-state error, overshoot and tuning time as comparison metrics. A better response from the PSO is obtained through simulations, and this response is validated in a digital AVR on a real generator.