基于单神经元PID控制器的压电陶瓷控制方法  

Control Method for Piezoelectric Actuator Based on Single Neuron PID Controller

作　　者：梁宇恩[1] 钱飞[2] 许素安[2] 

机构地区：[1]浙江机电职业技术学院,杭州310053 [2]中国计量学院,杭州310018

出　　处：《计算机测量与控制》2015年第11期3657-3660,共4页Computer Measurement &Control

基　　金：国家自然科学基金(51105348)

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,文章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。In order to reduce the impact of hysteresis nonlinear characteristics of piezoelectric ceramic actuator on micro-positioning system,hysteresis curve of PZT is obtained under different conditions through experimental tests,then the characteristics of hysteresis is further analyzed in this paper.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,feedforward control and neuron self-adaptive PID algorithm is applied,and the hysteresis error of piezoelectric actuator has been compensated.Simulation results show that the mean relative error of the displacement track is reduced to 1.34%.So the output displacement of piezoelectric actuator can be controlled precisely.Thus the hysteresis effect of piezoelectric on micro-positioning system is diminished,and the positioning accuracy of micro/nano-positioning system has been improved.

关 键 词：压电陶瓷致动器 迟滞非线性 神经网络 单神经元 

分 类 号：TP273.21[自动化与计算机技术—检测技术与自动化装置]