Research on nonlinear modeling and precision displacement control of giant magnetostrictive actuator

To address the nonlinear issues in the precision displacement control of the giant magnetostrictive actuator(GMA), a nonlinear mathematical model was developed based on the J-A hysteresis model. This model takes into account the dynamic loss characteristics of giant magnetostrictive materials (GMM) as well as the structural nonlinearity of the actuator. A long short-term memory neural network (LSTM) feedforward compensation PID control strategy is proposed to achieve precise control of GMA by adjusting controller parameters in real-time. The control model is constructed using simulation software, and a GMA system test platform is developed. Simulation analysis and experimental verification are conducted to evaluate the performance of the GMA system. The results show that the LSTM feedforward compensation PID control has good position tracking and anti-interference capabilities, and the dynamic tracking error is kept within 1μm, which greatly improves the control accuracy.