基于BP神经网络的超磁致作动器建模与控制

Modeling and Control of Giant Magnetostrictive Actuators Based on BP Neural Network

超磁致作动器的迟滞非线性是其应用的一大障碍,对此基于BP神经网络建立超磁致作动器GMA的Hammerstein模型,在此模型基础上,设计前馈逆补偿与PID反馈控制相结合的复合控制策略。针对建模范围内的所有单频和复合频率的输入信号,控制器都能保证跟踪控制效果。通过实验实时跟踪的结果进一步验证建模和控制的效果。

The hysteresis nonlinearity of the GMA is a major obstacle in the application of such material in actuators. Establishes a Hammerstein model based on the BP neural network to model the hysteresis nonlinearity. Based on the model, designs a composite control strategy combining feed-forward compensation and PID feedback control. The controller ensures tracking control for all single and complex frequency input signals within the modeling range. Finally, the results of the real-time tracking experiment on GMA further validate the modeling and control results.