Aplicação do controle preditivo generalizado em sistemas mecânicos

Abstract

The present contribution presents a theoretical, experimental and simulated study on the application of Generalized Predictive Control (GPC) in mechanical systems. The mechanical systems discussed in this work are a mass-spring-damp structure of one degree of freedom and a parallel planar mechanism of two degrees of freedom. The experimental structure consists of an aluminum platform supported by flexible beams of stainless steel. The structure was subjected to small excitations using a test hammer and coupled with an accelerometer to acquire data. From the measurements obtained it was possible to establish the Frequency Response Function to identify the parameters of the structure. With the structure parameters established, the active vibration control was performed using the GPC and compared to the LQR control. On the other hand, the complete dynamic model of the mechanism was obtained using the Lagrange principle. A position control for trajectory tracking was implemented in the mechanism using the Computed Torque Control and the results using GPC were compared to those obtained with PID control. The results demonstrate the efficiency of the GPC and its power advantages over the other controllers submitted to the comparison. The analysis and the results of the controllers performance were carried out computationally by using softwares Matlab® e Simulink.