【作者】 李欣欣；

【导师】 陈子辰； 王文；

【作者基本信息】 浙江大学 ， 机械制造及其自动化， 2008， 博士

【摘要】 本论文以研究大行程、高精度定位平台为主要目的,结合浙江省科技计划(国际合作)项目“纳米级超精密定位平台的研制”(No.2004C24003),研制了以平面电机及超磁致伸缩致动器为驱动单元,以基于平面电容传感器的二维位移直接解耦测量系统为测量单元的大行程高精度二维定位平台,并进行了动力学分析、优化、控制和相应的技术基础研究与实验研究。第一章阐述了超精密定位平台的研究背景与意义,介绍了其中的驱动与测量技术,综合评述了当前国内外超精密定位平台大行程驱动方式、两级驱动方式以及位移测量技术的研究现状及其发展趋势,重点分析了实现大行程高精度二维定位的若干基础技术。最后给出了论文的主要研究内容,阐述了各章节结构之间的关系。第二章研究了超磁致伸缩致动器的动态磁滞模型。在致动器(准)静态磁滞模型的基础上,结合涡流阻抗的估算方法,建立了致动器在不同频率下的动态模型;研究了动态模型的数值计算方法,并对仿真与实验结果进行了对比分析。第三章根据超磁滞伸缩致动器(GMA)微动平台的磁、机特性,推导了GMA微动平台的动力学模型;建立了基于一阶有限差分-强跟踪滤波的扩展卡尔曼-布斯滤波方法,对GMA微动平台的位移测量信号进行滤波:并对所提滤波方法进行了仿真分析与实验验证。第四章研究了基于磁滞逆补偿的超磁致伸缩致动器自适应控制方法。采用二分法数值求解超磁滞伸缩致动器(GMA)(准)静态与动态磁滞全逆模型,并将逆模型应用于GMA精密定位控制系统的前馈补偿环节。对模糊PID控制、广义最小方差—模糊PID控制以及广义预测PID控制等自适应控制算法的控制效率、控制精度进行了仿真与实验研究。利用磁滞前馈逆补偿结合多模广义预测PID控制算法,提高GMA微动平台的定位精度和起动阶段的平稳性。第五章针对二维位移的测量,提出了基于平面电容传感器原理、具有直接解耦效果的二维位移测量方法。对比了无限大平行板电容器电容量计算公式、Hereen模型、MAXWELL静电场理论以及有限元等四种电容量计算方法的计算结果,在此基础上计算了传感器的灵敏度。最后,结合灵敏度指标对传感器的结构参数进行了设计。第六章研制了基于平面电容传感器的二维位移直接解耦测量系统。研究了基于平面电容传感器的二维位移直接解耦测量系统研制所采用的主要技术:采用电荷放大器检测电路对传感器输出微弱信号进行放大检测,以消除PCS分布电容的影响;采用锁相检测原理对检出的信号进行相敏解调,以消除由检测电路寄生电容造成的相移;利用研制的传感器所产生的相位若为90度的两路信号进行运动辨向、1/4周期计数、相位细分以及位移量整合计算;最后,叙述了测量系统所采用的抗干扰技术。第七章进行了宏/微两级驱动的大行程高精度二维定位系统实验研究。构建了宏/微驱动二维大行程精密定位实验平台,以及基于平面电容传感器的测量系统实验平台,对其进行了实验研究,分析了影响平台性能的主要因素。更多还原

【Abstract】 Ultra-precision positioners with large range, have been widely applied in many fields. Most of these positioner are driven in coarse-fine mode. For 2-dimension positioning, the type, with one dimension positioner added on to another dimension one, is commonly used. Accordingly, one set of measuring system for one DOF’s measurement, two set are applied for two DOF（X-Y） displacement measurements, which leads to Abbe error. Here, a new coarse-fine positioning method is put forward, with planar motor as the coarse driving part and giant magnetostrictive actuator （GMA）, a smart material actuator, as the fine one, which avoids added-on installation for 2-DOF positioning as mentioned.For coarse-fine mode, fine positioning is decisive for the whole positioning accuracy, though coarse positioning also plays an important role. Therefore, inherent nonlinearity and hysteresis in GMA, greatly affecting the accuracy, has to be handled for the goal of ultra-precision positioning.Hysteresis exhibited by giant magnetostrictive actuator is rate-independent when the input frequency is low and can be modeled by static Jiles-Atherton model. It becomes rate-dependent due to the eddy current effect and the magnetoelastic dynamics when the input frequency gets high. In this case, the eddy current loss represented by eddy current impedance and the estimation of the GMA’s eddy current impedance through effective permeability are studied.Before computing the error between the feedback displacement signal and the desired position, the feedback signal should be filtered to minimize the influence of noise and time-variant model parameters, caused by environment or by estimation deviation. To this end, the extended Kalman filter combined with 1^st-order devided difference and strong tracking filtering is established.To attenuate the nonlinear and hysteresis behaviors, the inverse of the Jiles-Atherton hysteresis model for hysteresis compensation, in other words, linearizing the GMA model, is employed. After the inverse hysteresis developed and the feedback displacement signal filtered, feedback control laws are designed to meet the performance specifications. Several adaptive control algorithms, respectively, combined with common PID controller are considered, including Fuzzy-PID control, generalized minimum variant Fuzzy-PID control, multi-mode generalized predictive control PID control. The performance of these control methods are illustrated through simulations and experiment studies.On the other hand, to erase Abbe error for 2-DOF measurement, the principle of the planar capacitive sensor-based 2-D direct-decoupling measurement is proposed. Four methods for the sensor’s capacitance computing—equation for infinite parallel capacitor, equation by Hereen’s model, equations derived from Maxwell static electric field theory and analysis by finite elements method-are provided, and their results are compared with. Based on capacitance computing, the sensitivity of the designed sensor is figured out. To balance the dimension and the sensitivity of the sensor, structural design is conducted to minimize the sensor area with the given sensitivity. Moreover, the techniques for signal detection and conditioning, increasing displacement computation and anti-interference implementation are discussed.At last, the coarse-fine positioner is established, the planar capacitive sensor-based 2-D direct-decoupling measuring system is constructed, and the according experiments have been conducted to verify above theoretical studies.更多还原