Neural Network Control And Dynamic Model As Well As Experiment For Giant Magnetostrictive Actuator

Giant magnetostrictive actuators, characterized by large strain, high force, fast response and nanometer solution and so on, have a wide range of potential applications in super-precision positioning, robotics, and vibration control, etc. However, the relation between the input magnetic field and output displacement for giant magnetostrictive actuators exhibits the hysteresis nonlinearity, which presents a challenge in the applications of actuators. To control and use giant magnetostrictive actuators, the neural network control and dynamic model as well as experiment for giant magnetostrictive actuators are selected as the subject of dissertation for Ph.D. The main research work is following:1. RBF neural network, DRNN and Fuzzy RBF neural network are used separately as the identification for the actuator. Simulation results of the three neural network identifications show that the identification ability of fuzzy RBF neural network is stronger, and there is better identification performance than simply neural network. And there are lesser parameters in the fuzzy RBF neural network with 5 parameters.2. There are different hysteresis nonlinearity for the actuator driven with the signal of different frequency and different amplitude. It is necessary to establish hysteresis compensation control for the actuator system. Neural network supervisory control using feedback error is used to compensate hysteresis nonlinearity of the actuator. DRNN, RBF and fuzzy RBF are used separately as controller in the control system. Simulation results show that the control performance of DRNN controller is better than the other two neural networks controller.3. Neural network adaptive PID control with nonlinear prediction model is designed. In the control system, there is a fuzzy RBF neural network identification to forecast the hysteresis nonlinearity of the actuator, there is a BP neural network to optimize PID parameters, and PID with excellent parameters feed back to improve control performance. Simulation results show that the control strategy can on-line obtain inverse model of hysteresis nonlinearity for the actuator and eliminate the impact of hysteresis nonlinearity. The control strategy can track the dynamic characteristic real time, and adapt the change of the reference input without knowing the mathematics model of the giant magnetostrictive actuator.4. A dynamic linearity model of the actuator has been founded, based on the electro-magnetic theory and mechanical vibration principle. The model quantifies the relation between output displacement and input current by analyzing mechanical impedance of the Terfenol-D rod, spring and output shaft for the actuator. The output displacement for the actuator has been calculated. The calculating results of the dynamic linearity model can describe the output displacement characteristic for the actuator with bias magnetic field and without bias magnetic field when the driving magnetic field frequency from 0Hz to 200Hz. when the driving magnetic field frequency from 200Hz to 2200Hz, the dynamic linearity model can describe the output displacement characteristic for the actuator without bias magnetic field. The dynamic linearity model can provide the groundwork for the design and optimum of actuators.5. An experimental study on the displacement output characteristics for the designed actuator is carried out with bias magnetic field and without bias magnetic field. Experimental data collected from the test device show that below the driving magnetic field frequency 200Hz, at the same drive current value the displacement output decreases with input current frequency increasing. There are no double frequency phenomena between the output displacement and input current with bias magnetic field, and double frequency phenomena without bias magnetic field. Over the driving magnetic field frequency 800Hz, there are both double frequency phenomena with bias magnetic field or without bias magnetic field. Through experiments, frequency-dependent behaviors of the actuator were investigated. The experiment results provide groundwork for the optimum and application of actuators.