Neural network model for hysteretic characteristic of shape memory alloy

Abstract

Shape memory alloy has obvious hysteretic characteristics, which is crucial for the utilization of shape memory alloy, so a rational model is needed to describe them. In the current paper, a model for shape memory alloy based on the neural network is established, in which the activation function is modified so that it can naturally describe the hysteretic curve of shape memory alloy. The proposed neural network model is used to simulate temperature-induced hysteresis and stress-induced hysteresis, and the results show that the proposed model has a good effect and has certain adaptability. The reason why the proposed neural network model can describe hysteresis is qualitatively explained, and it is further pointed out that the model can be regarded as a generalized model of the Preisach model. The model is further extended to accommodate hysteretic curves under different conditions. The proposed neural network model has a simple form and high accuracy, especially in comparison with traditional models, and can be used for the modelling of shape memory alloys or other materials with hysteretic characteristics.