System identification of a NiTi-based SMA actuator using a modified Preisach model and adaptive control

We present an experimental method for the modeling and system identification of shape memory alloy-based (SMA) wire actuators, employing minimum-variance adaptive tuning and control to find the parameters of a proposed modified Preisach model. The material compositions of SMAs have inherent properties that induce phase transformations between austenitic and detwinned martensitic crystal structures, with the former occurring at high temperatures and low stresses and the latter at low temperatures and high stresses. Thermally- and mechanically-induced phase transformations, known as the shape memory effect (SME) and superelasticity (or pseudoelasticity), respectively, allow the recovery of plastic deformations and enable SMA wires to behave as actuators. For both phenomena, forward and backward phase transformations display hysteretic nonlinearities. Here, we classify the SME phase transformation hysteresis and modify the classical Preisach model to account for superelastic characteristics observed experimentally. By simultaneously controlling the temperature and stress during static and dynamic experiments, we can identify the nonlinear mapping between temperature, stress, and strain, which combined with the thermodynamic model of actuation cycles describe the complete dynamics of the SMA actuator. The actuator model is validated with experiments in which the SMA wire is statically and dynamically loaded.