Adaptive control of shape memory alloy actuators for underwater biomimetic applications
Academic Article
Overview
Research
Identity
Additional Document Info
Other
View All
Overview
abstract
In actuator technology active or smart materials have opened up new horizons in terms of actuation simplicity, compactness, and miniaturization potential. One such material is the nickel-titanium shape memory alloy (NiTi SMA), which is gaining widespread use in a variety of applications. The numerous advantages of SMA over traditional actuators are of particular interest in the area of underwater vehicle design, particularly the development of highly maneuverable vehicles of a design based on the swimming techniques and anatomic structure of fish. An SMA actuation cycle consists of heating/cooling half-cycles, currently imposing a limit on the frequency of actuation to well below 1 Hz in air because of slow cooling. The aquatic environment of underwater vehicles lends itself to cooling schemes that use the excellent heat-transfer properties of water, thus enabling much higher actuation frequencies. A controller for SMA actuators must account not only for large hysteretic nonlinearities between actuator output (strain or displacement) and input (temperature), but also the thermal control for resistive heating via an applied current. The control of SMA in water presents a problem not encountered when actuating in air: accurate temperature feedback for the SMA is very difficult in water. We overcome this problem by using a simplified thermal model to estimate the temperature of the wire in conjunction with an adaptive hysteresis model, which relates the actuator output to the estimated temperature. Experimental results are provided, showing that this method for control of an SMA wire works equally well both in air and in water, with only rough estimates (easily obtained) of the thermal parameters. Successful tracking of reference displacement signals with frequencies up to 2Hz and relatively large amplitudes have been demonstrated experimentally.
