Metodologia integrada baseada em algoritmos genéticos para parametrização de motores de indução e sintonia de controladores vetoriais

Abstract

This work proposes and validates a sequential and automated methodology, based on Genetic Algorithms (GAs), for the parameter estimation of the three-phase induction motor and the subsequent tuning of PI controllers for the IFOC technique. The estimation is performed using experimental drive data, while the tuning optimizes a multi-objective cost function to ensure system robustness. In the bench validation, the GA successfully identified a set of effective parameters that compensate for the non-linearities of the real motor. The controller, tuned based on these parameters, demonstrated high robustness in stress tests, maintaining precise control even under disturbances more severe than those used in the optimization itself. Additionally, the analysis revealed that low-computational-cost GA configurations were sufficient to achieve maximum performance, demonstrating the high efficiency of the method. It is concluded that the proposed methodology is a viable and powerful tool for automating drive design, delivering a high-fidelity control system with experimentally proven robustness.