Uso de Aprendizado de Máquina na Estimação de Parâmetros do Controlador PID-2DOF Aplicado no Controle de Velocidade de um Motor de Indução Trifásico

Abstract

This study investigates the use of Machine Learning (ML)in estimating the parameters of a Proportional-Integral-Derivative 2 Degree of Freedom controller applied to a three-phase induction motor coupled to a cylindrical robotic arm control system. Using classical tuning methods for a PID-2DOF controller as a starting point, the results were checked in terms of the following specifications: rise time, settling time and overshoot. The results of the tests carried out show that the Machine Learning model provides results with adequate precision, being able to identify parameters that increase the speed and robustness of the system’s control. This work highlights the contribution of the Machine Learning algorithm for use in control systems, highlighting its application in certain test scenarios and providing a basis for future research and improvements, including the exploration of different data sets and Machine Learning models.