Identificação e controle ativo de vibrações aplicado em estruturas flexíveis

Abstract

The present work proposes the active control of vibrations on a stainless-steel beam using electromagnetic actuators, with the advantage of controlling the structure without any mechanical contact, with the ease and high repeatability of exciting the structure from them. Using an LQR controller associated with a state observer and an inverse model of electromagnetic actuators, it was possible to train and use an Adaptive NeuroFuzzy Inference System (ANFIS) to significantly simplify the control loop. At first, only the LQR controller was replaced by the ANFIS controller, designed from its inputs and outputs. Then a new ANFIS was designed that replaced the LQR controller and the Inverse Actuator Model. The proposed methodologies were validated through numerical simulations and then experimental tests were performed. Another comparison was made by analyzing the energy consumption of the controllers for each of these configurations. From the results, one can conclude the validity of the proposed methodology.