Investigation Of Control Methodology For Magnetically Controlled Shape Memory Alloy Actuators

Magnetically controlled shape memory alloy is a kind of functioning material that has a short history of the development, but it possesses both the rapid frequency response and high control precision. At the same time, it is able to overtime the shortcoming of low deformation rate. Moreover, magnetically controlled shape memory alloy is a magnetic control material. It is easier to control than temperature shape memory alloy. So, a research of magnetically controlled shape memory alloy attracts intensive concern. For the time being, the research for the mechanism of deformation, the factors of deformation and the applications of magnetically controlled shape memory alloy actuator are far from enough, especially the method of magnetically controlled shape memory alloy actuator's control. In this article, we'll try to research in control method in magnetically controlled shape memory alloy actuator. In order to engage in the future related to lay a foundation.In this article, we'll introduce the mechanism of deformation, the static characteristic and the dynamical characteristic of magnetically controlled shape memory alloy actuator firstly. By analysing the character of magnetically controlled shape memory alloy actuator, we realize the relations between the inputs and the outputs which have hysteresis nonlinear relation. The method of making models is propsed for compensating the hysteresis nonlinear of magnetically controlled shape memory alloy actuator.Because of the superiority of the networks for modeling, a neural network is proposed for hysteresis nonlinear model of magnetically controlled shape memory alloy actuators. In order to solve the problem of one to many mapping, two displacement Sigmoid function use as the activation function of the neuron in the first hidden layer. The model and inverse model of the actuator is established with the neural network. The inverse model is used for open-loop control and feedforward plus feedback control. The experimental results demonstrate that the method is useful to control the magnetically controlled shape memory alloy actuators, and could get a higher accuracy.A Prandtl-Ishlinshii model for hysteresis modeling of the magnetically controlled shape memory alloys actuator is proposed. The model is composed of a number of simple linear operators called linear-play operators. Then, an inverse model of the actuator is established with a number of linear-stop operators. The inverse model is used for open-loop control and feedforward plus feedback control. The experimental results demonstrate that the method is useful to control the magnetically controlled shape memory alloy actuators, and could get a higher accuracy.Finally, a fuzzy neural network controller which based in reinforcement learning algorithm is designed. The simulation results illustrate the designed controller has well ability for controlling magnetically controlled shape memory alloy actuator.