Experiment-based hysteresis identification of a shape memory alloy–embedded morphing mechanism via stretched particle swarm optimization algorithm

Abstract

Prandtl–Ishlinskii model is a phenomenological model for complex nonlinear hysteretic behaviors in mechanisms integrated with smart material actuators such as piezoelectric, magnetostrictive and shape memory alloy. In this article, experimental data obtained from a bio-inspired morphing wing mechanism actuated by a shape memory alloy wire are used to derive the generalized Prandtl–Ishlinskii model. The unknown parameters of the generalized Prandtl–Ishlinskii model are identified using a modified type of particle swarm optimization algorithm, that is, stretched algorithm. Accuracy of the trained model is evaluated by two different input signals. In addition, for each input signal, statistical prediction error analysis is implemented to test the model validity and accuracy. Results confirm that the presented model with identified parameters properly predicts hysteresis behavior of the mechanism for different input signals and the model yields a small estimation error.