基于 SVD 的 Preisach 磁滞模型辨识及其在压电驱动器中的复合控制

SVD-based Preisach hysteresis identification and composite control of piezo actuators.

机构信息

NUS Graduate School for Integrative Sciences and Engineering, Department of Electrical and Computer Engineering, National University of Singapore, 4 Engineering, 117576, Singapore.

出版信息

ISA Trans. 2012 May;51(3):430-8. doi: 10.1016/j.isatra.2012.01.002. Epub 2012 Jan 26.

DOI:10.1016/j.isatra.2012.01.002

PMID:22284264

Abstract

In this paper, the singular value decomposition (SVD) based identification and compensation of the hysteretic phenomenon in piezo actuators are addressed using a Preisach model. First, this paper presents an SVD-based least squares algorithm and a revision approach of the identification through updating the SVD. With the identified parameters and a log of the memory curve, a Preisach-based inversion compensator is constructed which is complemented with a feedback controller to address the inevitable and residual modeling errors. Experimental results are furnished for both the identification and compensation approaches. The Preisach-based feedforward controller significantly improves the tracking performance and reduces the root-mean-square (RMS) tracking error of a PID controller by 76.7% and 89% at 1 Hz and 25 Hz, respectively. With the proposed composite controller, the percent-RMS errors at 1 Hz and 25 Hz are reduced to 0.035% and 0.31%, respectively.

摘要

本文使用 Preiach 模型基于奇异值分解（SVD）对压电执行器中的迟滞现象进行识别和补偿。首先，本文提出了一种基于 SVD 的最小二乘算法和通过更新 SVD 进行识别的修正方法。利用所识别的参数和记忆曲线的日志，构建了基于 Preisach 的逆补偿器，并补充了反馈控制器以解决不可避免和剩余的建模误差。为识别和补偿方法都提供了实验结果。基于 Preisach 的前馈控制器显著提高了跟踪性能，将 PID 控制器在 1 Hz 和 25 Hz 时的均方根（RMS）跟踪误差分别降低了 76.7%和 89%。使用所提出的复合控制器，在 1 Hz 和 25 Hz 时的百分比 RMS 误差分别降低到 0.035%和 0.31%。