基于单神经元PID控制器的压电陶瓷控制方法

The Control Method for Piezoelectric Actuator Based on Single Neuron PID Controller

摘要:

为了减小压电陶瓷致动器迟滞非线性固有特征对微定位系统的影响,本文通过实验测试得到了压电驱动器在不同条件下具有的迟滞曲线，并进一步分析了该迟滞现象所呈现出来的特征和规律。根据以上的分析和测量获得的数据，建立了基于多项式拟合算法的神经网络迟滞模型，并能精确地反映该迟滞现象。在这种模型的基础上，采用前馈控制和神经元自适应PID算法，对压电致动器迟滞误差进行了补偿。系统仿真结果表明，位移跟踪的平均相对误差降低至1.34%，实现了精确控制压电致动器的输出位移,从而降低了压电陶瓷迟滞现象对系统定位精确度的影响，提高了微/纳米定位系统的定位精度

Abstract:

Hysteresis nonlinearity is the intrinsic characteristic of piezoelectric actuator. In this paper, hysteresis curve of PZT is obtained under different conditions through experimental tests, then the characteristics of hysteresis is further analyzed. According to the above analysis and the data obtained by the measurement, the neural network hysteresis model based on polynomial fitting algorithm is established, and the model can reflect the hysteresis of PZT accurately. Based on this Model, feed-forward control and neuron self-adaptive PID algorithm is applied. The hysteresis error of piezoelectric actuator has been compensated and the output displacement of piezoelectric actuator can be controlled precisely. Thus the hysteresis effect of piezoelectric on micro-positioning system is diminished. Simulation results show that the mean relative error is reduced to 1.34% . This indicates that positioning accuracy of micro/nano-positioning system has been improved , and the system itself has certain self-adaptability and anti-interference ability.