PATRIOTA, A. H. C.; http://lattes.cnpq.br/7861803918085082; CAMPOS, Andrei Henrique Patriota.
Résumé:
Shaped memory alloy based actuators have been employed in a wide variety of applicati-
ons over the past decades. In fact, due to its main characteristics, as a good force-by-mass
ratio, in addition to its shape recovery capacity, even after being deformed by the applica-
tion of high mechanical stresses, this type of material has been used in applications ranging
from biomedical engineering to robotic systems, always being associated with quiet and
smooth displacements. Moreover, when it comes to linear displacements, the characteris-
tics of this class of actuators place them as an option for conventional electromechanical
actuators, such as electric machines and solenoids. However, a major drawback arises
when dealing with these alloys: position control based on mechanical systems model with
shape memory alloys is not always possible, due basically to the difficulties of precisely
modeling the behavior of the material, whose parameters vary with the temperature, and
whose phase change characteristic, responsible for the shape memory effect, presents a
hysteretic behavior. In this context, the present work aims at proposing control techniques
that attempt to circumvent the need to obtain an analytical model of a plant composed
of a nickel-titanium wire actuator, at the same time as it seeks to obtain a mathematical
representation for it, both by means of the mathematical equations that govern its submo-
dels, as well as using a system identification tool. Some classical controllers are glimpsed,
and a hybrid Fuzzy-PID controller is implemented in order to provide a smart mechanism
for tunning the PID controller. Finally, the performance indices of these controllers are
raised considering pre-established operating conditions for comparison purposes.
