Dynamic Modeling and Control System Design for Shape Memory Alloy Actuators

Abstract

Shape memory alloy (SMA) is a type of smart material which remembers its original state. It is light weight and small, and known to provide high contraction force with low noise. Its application has wide range from robotics to medical science. One of its potential applications in space is a supporting system of membrane structure that can be used as synthetic aperture radar (SAR) antenna to achieve high flatness. It exhibits nonlinear phenomena called hysteresis when it's electrically heated. Hysteresis is a nonlinear phenomenon that refers to the dependence of a physical system on the environment. Hysteresis in SMA causes a major difficulty in control system design. Un-modeled or poorly modeled hysteresis introduces inaccuracy in tracking and the performance of the system. Experimental test bench is constructed for one set of SMA actuators that resembles the membrane structure's supporting system. Hysteresis is obtained by running open loop test with the test bench. Dynamic model of the SMA wires is developed using classical Preisach model and modified Maxwell model. Then the inverse model is implemented in feed-forward loop to compensate for nonlinear hysteresis. Simple feedback controllers are added to correct the modeling errors. Experimental results reveal that the error is significantly reduced when comparing feedback controller with hybrid feedback and feed-forward controller.