传感技术学报2017，Vol.30Issue(12)：1884-1889,6.DOI:10.3969/j.issn.1004-1699.2017.12.017
压电陶瓷迟滞神经网络建模与线性补偿控制
Neural Network Modeling and Linear Compensation Control for Hysteresis of PZT
摘要
Abstract
In order to suppressthe hysteresis nonlinearity of PZT,firstly the hysteresis data of PZT is obtained by ex-perimental measurements. Through the analysis of experimental data,relationship between input voltage and output displacement of PZT is linearized by introducing the linear equation and hysteresis model of PZT is established by combining the polynomial fitting with neural network.A feedforward controller is designed according to the hysteresis model. Compensation and controlling for hysteresis of piezoelectric ceramic is experimented by feedforward open-loop method and feedforward combined with PID method. Experimental results demonstrate that the primary ring hysteresis of PZT displacement is reduced 91.84% and the minor ring hysteresis is reduced 85.67% andthe mean relative error of the displacement track is 2.97% with feedforward open-loop method,and the primary ring hysteresis of PZT is reduced 96.42% and the minor ring hysteresis is reduced 88.42% andthe mean relative error of the dis-placement track is 2.04% with feedforward combined with PID method. This indicates that the control method can-suppress the hysteresis nonlinearity of PZT effectively.关键词
迟滞非线性/神经网络/前馈/控制Key words
piezoelectric ceramics/hysteresis nonlinearity/neural network/feedforward/control分类
信息技术与安全科学引用本文复制引用
许素安,金玮,梁宇恩,张锋..压电陶瓷迟滞神经网络建模与线性补偿控制[J].传感技术学报,2017,30(12):1884-1889,6.基金项目
国家自然科学基金项目(51105348) （51105348）
