Neural Network Model for Hysteresis Non linearity of Magnetic Shape Memory Alloy Actuator

作者：Zhou Miaolei^*; Wang Shoubin; Gao Wei

来源：Journal of Computational and Theoretical Nanoscience, 2013, 10(12): 2931-2935.

DOI：10.1166/jctn.2013.3304

• Magnetic Shape Memory Alloy
• Neural Network
• Hysteresis
• Activation Function

As a leading material in a new class of functional materials, magnetic shape memory alloy has received a great deal of attention. Hysteresis nonlinearity between input magnetic field and output displacement of magnetic shape memory alloy actuators is a big restriction for its high accuracy displacement control, and establishing hysteresis nonlinearity model is the most popular and effective method to solve this problem. A neural network model is proposed for the hysteresis nonlinearity of magnetic shape memory alloy actuators in this paper. The structure of this neural network model has four layers and a lateral connection. In order to solve the problem of one to many mapping which can not be realized by the neural network, a double sigmoid function is used as the activation function in the first hidden layer. The activation function in the second hidden layer is a sigmoid function. The learning method of the neural networks is momentum BP algorithm. The experimental results demonstrate that this method can build an effective hysteresis nonlinearity model of magnetic shape memory alloy actuators, and could obtain a higher accuracy; the maximum displacement error of this model is only 0.55%.

• 出版日期2013-12
• 单位吉林大学