基于动态PREISACH算子的压电陶瓷动态迟滞智能建模  

Intelligent Modeling of Rate-dependent Hysteresis in Piezoelectric Actuators Based on Dynamic Preisach Operator

作　　者：张新良[1] 谭永红[2] 

机构地区：[1]上海交通大学电子信息与电气工程学院自动化系,上海200240 [2]准林电子科技大学计算机系智能系统与工业控制实验室,桂林541004

出　　处：《系统仿真学报》2009年第9期2682-2686,共5页Journal of System Simulation

基　　金：上海师范大学重点学科项目(DZL811);上海市教委科技创新重点项目(09ZZ141);上海师范大学前瞻性科研项目(DYL200809);国家自然科学基金(60572055)

摘　　要：针对经典Preisach模型存在的不能反映迟滞依赖于输入变化率的动态特性且其权重函数难以确定的缺点,给出了Preisach模型中积分项的一种实现形式,动态Preisach算子,其参数是输入变化率的函数,该Preisach算子反映了迟滞依赖输入变化率而体现的动态特性的基本特征。通过引入动态Preisach算子,将存在于迟滞非线性中的多值映射转化为连续的一一映射,从而将神经网络等智能辨识工具应用到迟滞非线性的建模中。该方法简化了迟滞非线性的建模过程,结构简单,且能实现在线更新,具有实际的应用价值,实验结果验证了所得模型的有效性。The classic Preisach model is hard to be used to describe the dynamics of the hysteresis depending on the input change-rate. Moreover, the weighting function of the Preisach model is rather difficult to be determined so that it is impossible to adapt the change of the environmental conditions. A modified Preisach model, namely the dynamic Preisach operator, was proposed, wherein, the parameters are dependent on the change-rate of the input to describe the basic feathers of the dynamics of the hysteresis caused by the input change-rate. With the introduction of the dynamic Preisach operator into the input space, the multi-valued mapping of the hysteresis is transformed into the one-to-one mapping so the neural networks can be utilized for the identification of the rate-dependent hysteresis. The proposed model has a simple architecture that simplifies identification procedure for hysteresis and can adapt to different operating conditions. Finally, experimental results are illustrated to show the effectiveness of the proposed approach.

关 键 词：PREISACH模型 动态迟滞 神经网络 动态Preisach算子 一一映射 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]