基于GMA的微致动结构建模理论和控制策略研究

In recent years, in order to satisfy the demand of advanced manufacturing equipments at many fields, such as national defence industry, ultraprecision machining etc. The application of GMA(Giant Magnetostrictive Actuator) in manufacturing equipments is paid more attention. The traditional GMA has some shortages in modeling theory of micro actuating structure and control strategy aspects, such as nonlinear model building imcompletely, complex parameters being handled ideally, lacking of status mornitoring method etc. These shortages directly cause actuating accuracy can't satisfy requirements of high-precise application. The GMA is a complex multi-physics field parameters coupling nonlinear system, the different physicial field parameters, such as stimulus, temperature, magnetic circuit, prestress and frequency etc, will couple each other and change in the whole actuation process. It can cause actuate error on one hand, and can reflect working status of GMA on the other hand. Therefore, in this study, the recognition of feedback feature information from GMA's multi-physics field status parameters which been collected real time by distributed fiber grating sensors is researched firstly. Second, the novel theory and methods of GMA's multi-physics field coupling on-line nonlinear modeling for micro actuating structure is researched based on the feedback feature information. Finally, the adaptive mechanical control strategy corresponding on-line nonlinear modeling is provided and implemented. When facing precision machining, manufacturing equipments can make GMA satisfy accuracy requirements of precision machining in high-end application, by effectively responding the change of various nonlinear parameters and environmental factors using model adaptive control.

近年来，为满足国防工业、超精密加工等领域对先进制造装备的需求，超磁致伸缩材料致动器GMA在制造装备中的应用越来越受到重视。传统GMA在微致动结构建模理论和控制策略方面，存在非线性模型构建不完备、复杂参数理想化处理、状态监测手段缺乏等不足，直接导致致动精度无法满足高精度要求。GMA是一个复杂多物理场耦合非线性系统，激励、温度、磁路、预应力、频率等不同物理场参数在致动过程中相互耦合不断变化，一方面导致致动误差，另一方面反映GMA工作状态。因此，本项目首先研究从分布式光纤光栅传感器实时采集的GMA多物理场耦合状态参数中辨识反馈特征信息，然后以此为依据研究GMA多物理场耦合的微致动结构在线非线性建模新原理和新方法，最后提出并实现依托非线性建模的自适应控制策略。使得制造装备在面对精密加工时，GMA能通过模型自适应控制策略有效应对致动过程中的各种非线性参量和环境因素变化，满足高端应用对精密加工的要求。

传统GMA在微致动结构建模理论和控制策略方面，存在非线性模型构建不完备、复杂参数理想化处理、状态监测手段缺乏等不足，直接导致致动精度无法满足高精度要求。针对以上不足，本项目历时3年，2015年搭建基于光栅传感技术的GMA在线动态建模的实验平台，采用光栅传感器对GMA输出位移、GMM棒位移、振动和温度进行在线测量；2016年基于实验平台，采集真实的实验数据，并采用神经网络方法对GMA进行动态建模；2017年基于实验平台，采集大量实验数据，依据数据驱动理论，采用LS-SVM方式进行GMA在线动态建模，并对模型的性能进行了评估: 0-1000Hz内，数据驱动模型能够实现-1.2%~1.1%范围内的控制精度；采用弱光栅传感技术，能够达到1uε的测量精度。项目进行期间，发表SCI期刊论文2篇，EI中文期刊论文1篇，EI国际会议论文3篇，申报发明专利3项。本项目验证了采用数据驱动理论能够实现对GMA的在线动态建模，能够实现对GMA的动态精确控制，能够克服传统模型存在的不能适应动态、高频、温度变化等各方面不足。