【作者】 郭锐；

【导师】 赵万生； 王振龙；

【作者基本信息】 哈尔滨工业大学 ， 机械制造及其自动化， 2007， 博士

【摘要】 由于可以获得较高的精度和良好的表面质量,微细电火花加工已经成为微细工具和微细零部件制造的重要手段之一。微细电火花加工是在突破微细电极在线制造这一瓶颈后才获得飞速发展的,之后在学术界和产业界的共同努力下,微细电火花加工专用机床也逐步研制出来,为微细电火花加工技术的研究提供了强有力的平台支持。数控系统是数控机床的核心,微细电火花加工数控系统与加工工艺密切相关,因此本文开展了微细电火花加工数控系统的研究工作。本文首先分析了微细电火花加工系统的设计要点,进行了数控系统的总体设计。提出了直线电机+直线光栅尺的全闭环运动控制方案,取消了从电机到工作台的一切中间环节,提高了系统的运动精度和灵敏度,同时也分析了在微细电火花加工中应用直线电机所需注意的问题。开放源码和基于Internet的协作开发模式带给Linux更强的稳定性和健壮性,而且软件资源丰富,成为数控系统理想的软件开发平台。RTAI和RTLinux是目前仅有的硬实时Linux,本文分析了RTAI和RTLinux的特点,在实时任务和实时内核之间设计了实时抽象层,封装了RTAI和RTLinux的编程接口,为实时任务提供了一致的API,基于实时抽象层开发的实时任务在无需修改源码的情况下,可以在RTAI或RTLinux下编译运行。PMAC运动控制器是一个实时多任务计算机系统,可以实现复杂的实时运动控制任务。本文基于PMAC运动控制器实现了上下位机体系结构的微细电火花加工数控系统;针对直线电机提出了PID+速度前馈+加速度前馈的运动控制算法,提高了系统的响应速度和控制精度,并不失稳定性;研究了定位误差补偿技术,构建了直线运动定位误差补偿表,获得了较高的定位精度。通过运动轨迹缓冲和分段实现了电火花加工特有的回退运动,基于Linux操作系统开发了译码、界面和工艺数据库,系统功能稳定,使用方便,满足了微细电火花加工的要求。为了获得较高的加工精度,微细电火花加工必须在线测量电极的尺寸。这是因为如果离线测量然后二次安装加工,那么安装误差将影响加工特征的最终精度。为了使数控系统具有在线测量电极的功能,本文在对显微成像系统进行光学分析的基础上,基于机器视觉技术构建了微细电极在线检测系统,系统由卤素光源、变焦显微镜头、CCD摄像机和6自由度支架组成,具有1.61μm的分辨率。在Linux操作系统下,基于V4L2 API开发了图像采集程序,使用mmap()内存映射方法获取图像数据。实现了IplImage数据结构和QImage类的转换,使图像既可以基于OpenCV进行处理,又可以基于Qt进行显示,通过Canny边缘检测算法提取了微细电极的边缘轮廓。后续实验表明系统在线测量值与扫描电镜离线测量值的相对误差在±5%以内,满足了微细电极在线检测的要求。为了验证数控系统的各项性能指标,本文在自行研制的微细电火花加工机床上进行了块电极电火花磨削实验和微三维结构电火花铣削实验,利用数控系统的视觉功能进行微细电极的在线测量和补偿,实验结果证明了本文开发的微细电火花加工数控系统的可行性、稳定性和可靠性,为我国研制出拥有自主知识产权的微细电火花加工数控系统奠定了坚实的基础。更多还原

【Abstract】 Owing to the high precision and good surface quality that it can offer, micro Electrical Discharge Machining (EDM) is potentially an important process for the fabrication of micro-tools, micro-components and parts with micro-features. Computer Numerical Control (CNC) system is a key component of the machine tool, and micro-EDM CNC system has a close relationship with the machine arts. Therefore, a micro-EDM CNC system is intensively studied in this dissertation.By analyzing the characteristics of a micro-EDM system, an overall design of the CNC system is taken. By using linear motor and linear encoder, a full loop motion control policy is proposed. It cancels all intermediate transport segments and greatly improves the motion accuracy and sensitivity of the machine tool. On the other hand, it also adds some difficulties when being used in a micro-EDM machine.Apart from the proprietary operating system, Linux offers abundant functionalities, better stability and robustness. So it is an ideal software development platform for a CNC system. By analyzing the characteristics of RTAI and RTLinux, a real-time abstraction layer is designed to provide a uniform API for RTAI and RTLinux. The developed micro-EDM CNC system allows real-time programs to use either RTAI or RTLinux without any code changes.PMAC motion controller is a multi-task real-time system, and it can do sophisticated motion planning calculations. Therefore, a micro-EDM CNC system is developed, in which PMAC is employed as the subordinate computer, and a PC-based platform is applied as the principal computer running on the Linux operating system. A PID + speed feed-forward + acceleration feed-forward motion control policy is proposed, which improves the response speed and accuracy, and without losing stability. The techniques of motion trajectory buffer and motion trajectory segmentation are used to implement backward control of motion trajectory. Under the Linux operating system, the left CNC programs are developed, such as a CNC decoder, a GUI and a database.On-the-machine measurement of electrode dimensions is necessary to achieve good accuracy in micro-EDM. This is because after a part that has been taken out of the machine for measurement is reinstated for more machining, the resetting error will drastically affect the final accuracy of the machined features. Therefore, a micro-EDM machine vision system is developed based on the analysis of the optical system, which is composed of a halogen type light source, a zoom lens, a CCD camera and a bracket which has 6-degree of freedom. It has a resolution of 1.61μm. Based on the Linux operating system, an image capturing program is developed with the V4L2 API, in which the image data is acquired with the method of mmap() memory map. The conversion between the data structure of IplImage and the class of QImage is implemented. Therefore the image can be processed by using OpenCV, as well as showed with Qt. The contour of micro electrodes can be extracted by means of the Canny edge detector. The experiment shows that the relative error between the CNC system and the scanning electron microscope (SEM) is less than±5%.Based on the self-developed machine, experiments of block electrical discharge grinding and micro-EDM milling have been carried out. By using the machine vision function of the CNC system, micro electrodes can be measured and compensated on-line. The experimental results prove its feasibility, stability and reliability.更多还原