【作者】 褚彪；

【导师】 唐玉国； 竺长安；

【作者基本信息】 中国科学技术大学 ， 精密机械与仪器， 2014， 博士

【摘要】 大型光学系统及高端科学仪器对我国科学研究、国防安全、经济建设和社会民生等各领域的进步和发展具有极其重要的作用。大面积、高精度衍射光栅则广泛应用于光学天文望远镜、惯性约束激光核聚变(ICF)装置、大型光刻系统和高分辨率中阶梯光栅全谱仪等场合。大面积、高精度光栅具有刻划幅面宽、行程长、刻槽结构复杂、槽形精度要求苛刻等特点,决定了机械刻划的制造加工方式。光栅刻划机要求在几百毫米的行程范围内,完成每毫米内几千甚至上万道刻线的刻划,并且保证很高的加工精度。同时,光栅质量与光栅刻线形态密切相关,而光栅刻线是由毛坯位置和刻刀位置共同决定,这就需要一套性能良好的超精密进给系统来完成对光栅毛坯在分度方向上的定位。针对光栅刻划的特殊需求,确定了宏微驱动的进给方式。分度系统的位移进给包括宏微驱动结构、位移测量反馈系统和计算机控制系统三个方面,针对分度系统的二级驱动模式,分析了宏驱动的丝杠螺母传动系统和微致动的压电驱动结构力学特性对系统定位精度的影响,并从状态空间角度,分析了光栅刻划机二级传动控制模型的可控性和可观性。由于宏微驱动的长传动链造成的刚度不足和间隙等问题,分析了在系统运行过程中产生的爬行现象与影响爬行的刚度、质量、摩擦系统、阻尼之间的关系。光栅刻划机微致动平台承载光栅毛坯实现分度方向上的位移进给,其定位精度直接影响着光栅的加工质量。微致动系统是基于柔性铰链的柔性机构,论文中对不同形状柔性铰链进行了对比分析,对柔性簧片进行了理论建模,基于半梁模型对微致动系统的刚度进行了分析,研究了铰链参数以及光栅毛坯质量对平台刚度的影响；同时,对微致动平台进行动力学建模,分析系统主要参数对模态的影响,并在此基础上对微致动平台关键零件柔性铰链的尺寸进行了优化。高定位精度要求的微致动平台,需要相应的控制系统。为了得到更适于光栅刻划进给系统的控制策略,就需要对系统进行辨识,从而获得相关参数。根据辨识模型的复杂性,和微致动平台本身的非线性和时变性等因素,需要控制算法在运行过程中具备一定的自学习和自适应特性。因此文中分析了基于单神经元PID和基于BP神经网络的PID控制算法的微致动进给方式,并将其定位数据与定长PID进行对比。光栅刻划控制实验研究表明,结合BP-PID自适应控制算法,工作台可实现近5nm的控制定位精度。通过对测试数据的数据统计分析,包括刻线的误差分析,验证了光栅刻划机分度进给系统控制策略选择的有效性。更多还原

【Abstract】 Large optical systems and high-end science equipments play an important role in the development in the area of scientific research, national defense security, economic construction, people’s livelihood, etc. Large-area, high-precision diffraction gratings are widely applied to the astronomical optics telescopes, ICF instruments, large lithography systems and echelle grating spectrometers. Because of the characteristic such as large ruling area, long stroke, complex Groove shapes and high accuracy demands, the diffraction grating can only be manufactured by machines. The diffraction grating ruling engine must scratch thousands of grooves per millimeter within the stroke of Hundreds of millimeters, which means that a long travel ultra-precision feeding system with high-quality performance is definitely crucially need to position the grating blanks. Considering with the special demand for the grating manufacturing, the dual-stage feed drive strategy is selected. The large stroke nano positioning system consists of macro-micro feeding system, measuring feedback system and control system. Since the coarse-fine positioning system may have impact on location accuracy, the dynamic and static characteristics of the macro/micro-drive device are analyzed in this paper, including the screw-nut-driven mechanism and piezoelectric-driven system. Besides, theoretical models of the coarse positioning system and the fine positioning system have been developed, and controllability and observability of the system is discussed. Moreover, considering that the stick slip phenomenon may arise during the manufacturing process, this paper studies the relationship between stick slip behavior and its impact factors, such as the friction coefficient and damping coefficient, mass, stiffness, etc.The fine positioning subsystem is a flexible mechanism based on the leaf spring. The different type of flexible components is discussed in this paper. Theoretical modeling of the flexible hinge is accomplished. On the basis of the half beam model, stiffness of the micro feed system is studied, research is carried out on the size of the flexible hinge and the mass of the blank (pressure on the stage), which influence the stiffness of the designed device. Meanwhile, dynamic equation of the flexible mechanism is established to analyze the main factors which influence the modal frequency of the fine positioning stage, focusing on the material types and dimension parameters of the leaf-spring. The SQP is adopted to optimize the main parameters of the mechanism to acquire better characteristic. The finite element analysis is also employed. The transport of the grating blank at the nanometer scale is key to the diffraction grating fabrication technology. The blank is placed on the fine actuate stage while the positioning accuracy Determines the grating processing quality. To achieve nanometer grade, the control system must be adopted. In order to investigate suitable nano-precision control strategy to the feeding system, system identification of the positioning system is a practical method to reveal the mechanism characteristics. Thus, the Neural Network Control is chosen. In this paper, the constant PID controller, the single neuron network based PID controller and the BP neuron network based PID controller are analyzed contrastively while grating manufacturing tests are carried out. Experiment data shows that the positioning precision of the designed feeding system can reach nearly5nm, which verifies the effectiveness of the chosen control strategy.更多还原