Estudo do comportamento e controle de atuadores de liga de memória de forma utilizando redes neurais

• Autor(a) principal: Sarmento, Nathan Luan Dutra
• Data de Publicação: 2019
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFPB
• Texto Completo: https://repositorio.ufpb.br/jspui/handle/123456789/16873
• topic: Ligas de memória de forma; Filtro de Kalman; Realimentação de resistência elétrica; Redes neurais artificiais; Controle preditivo; Shape memory alloys; Kalman filter; Electric resistance feedback; Artificial neural networks; Predictive control; Engenharia elétrica; Resistência elétrica; CNPQ::ENGENHARIAS::ENGENHARIA ELETRICA

Shape Memory Alloys are intelligent materials and can be used as actuators in many situations. The control of Shape Memory Alloys actuators may be challenging due to the non-linear and hysteretic behavior of this type of material. This way, the use of Artificial Neural Networks is a interesting tool to work as sensor and to implement the control of Shape Memory Alloys actuators. Studies with Shape Memory Alloys springs have been conducted to understand the phase transformation phenomenon in order to implement a control strategy. Electric resistance feedback has been used as sensor element, and an Artificial Neural Network has been developed to learn the relationship between the electric resistance and the force applied by the spring. Besides, an actuation circuit has been developed and an Extended Kalman Filter has been designed to estimate the state variables of the system. Finally, a neural network based predictive control strategy has been proposed and its results have been shown.The use of Artificial Neural Networks were efficient both as sensor and as model into the predictive controller. In addition, the force applied by the Shape Memory Alloy springs has satisfactorily converged to the reference values of force in the closed loop controlled system.