关键词: 超磁致伸缩, 智能构件, 神经网络, 前馈补偿, 迟滞非线性

Abstract: Because the problem of machining non-circle pin-hole, a giant magnetostrictive smart component is proposed. The intrinsic hysteresis observed in giant magnetostrictive material has impaired the motion accuracy. The GMM（Giant Magnetostrictive Material） smart component hysteresis has the nature that its relationship between output and input of smart component is one-to-two mapping and its output varies from the different frequency input. A new kind of architecture of neural network is proposed to approximate the smart components dynamic hysteretic characteristics. The smart component precision control is realized by combining the neural network created inverse model and a Proportional Derivative（PD） feedback controller. Simulation shows that this control strategy can eliminate the hysteretic nonlinear impact and achieve the precision control of the smart component.

Key words: giant magnetostrictive, smart component, neural network, feedforward compensation, hysteresis nonlinear