Hysteresis Modeling of Magnetic Shape Memory Alloy using a NARMAX Structure Model

As a kind of smart material-based product, the magnetic shape memory alloy based actuator (MSMABA) is widely used in nano-positioning applications. However, the serious hysteresis in the MSMABA severely impacts its application in the field of micro/nanotechnology. To model the hysteresis of the MSMABA which possesses the multi-valued mapping characteristics and the rate-dependent behavior, a nonlinear auto-regressive moving average with exogenous inputs structure (NARMAXS) model is developed. In this method, the Bouc-Wen (BW) model as an exogenous variable function (EVF) is used to construct the NARMAXS model, and the proposed model can be applied to describe the behavior of hysteresis with multi-valued mapping. Based on the powerful function approaching capability, a wavelet neural network (WNN) is used to replace the nonlinear function of the NARAMXS model. To verify the effectiveness of the proposed method, the NARAMX model based on the radial basis function neural network (RBFNN) is used for comparison. The experiment results illustrate that the proposed NARMAXS model presents excellent modeling ability for hysteresis of the MSMABA.