节点文献

激光飞行标刻系统及激光加工机器人控制与仿真研究

Research on Control and Simulation for Laser Flying Marking System and Laser Processing Robot

【作者】 蒋明

【导师】 曾晓雁;

【作者基本信息】 华中科技大学 , 物理电子学, 2011, 博士

【摘要】 激光加工机器人以高能激光束为加工工具,以多自由度机器人运动机构实现激光束运动传输,聚焦激光束在工件上按控制系统预先编程设计的2D二维或三维3D加工轨迹精确运动,实现各种材料的点、线、面多种形式多种工艺的控型控性加工,是当今最为活跃的先进制造技术之一,在机械制造、微电子加工、产品包装等各领域得到广泛应用。本文进行了2D激光飞行标刻系统和3D激光加工机器人仿真的研究。基于PC平台研制了激光飞行标刻控制卡和控制软件,研究进行了激光标刻标刻路径智能优化、激光飞行标刻控制算法以及最大飞行速度性能标刻速度影响因素等问题的研究。。研究了3D激光加工机器人运动建模仿真系统及3D激光加工路径的生成方法。在研制的激光飞行标刻软件平台上,开展了激光加工路径(Laser Processing Path, LPP)智能优化算法的研究。分析了点群加工、线条笔划加工、闭合轮廓加工三类LPP优化问题的数学模型,针对线条笔划路径优化问题,提出了一种采用正负整数进行染色体基因编码的新方案。结合λ? opt启发式算法和免疫算法思想设计实现了基于优生策略的免疫遗传算法。根据三角形三边关系定理设计了消除染色体自相交的缺陷的平滑免疫算子;提出了“最优染色体内部任意积木块必定是最优路径”的完美积木块准则,并据此设计了3段局部优化免疫算子。两类免疫算子的引入使遗传算法获得了均匀的进化过程、更快的收敛速度和优良的局域优化能力。应用该算法进行路径优化测试结果良好显示:,激光标刻跳跃路径经过优化后长度缩短了48.96%,振镜标刻整个图形的加工时间减少了5.34%~~8.57%。该优化算法可广泛应用于激光切割、焊接、打孔等激光加工控制系统以提高系统工作效率,。研究了激光飞行标刻扫描幅面约束特性和、激光飞行标刻控制算法对以及激光飞行标刻速度性能的影响影响因素等问题。首先分析了生产线恒速和变速运动状态下激光飞行标刻控制的的位移跟踪模型特点,设计实现了开环计算和闭环检测两类运动位移补偿算法并进行了实验,研究结果表明:开环计算跟踪算法在不需要增加额外硬件的情况下能完成对恒速运动工件的飞行标刻;闭环位移检测跟踪算法适合恒速和变速各种情况,具有速度自适应的优点。进一步研究了激光飞行标刻扫描幅面约束特性((Scanning Area Limitation , SAL))、最大标刻位移(Maximum MarkableOffset, MMO)、最大标刻时间(Maximum Markable Time, MMT)等问题,推导出激光飞行标刻最大飞行速度((Maximum Flying Velocity, MFV))与沿工件运动方向的最大标刻位移成正比,与标刻时间成反比,实验结果与MFV公式计算值结果吻合良好。还研究了标刻图形对象的不同扫描路径顺序对MFV的影响,分析比较研究了了完全进入后扫描标刻AEM((All Entered Marking, AEM))、逐行先进入先扫描标刻RFEFM((Rowed First Entering First Marking, RFEFM))、先进入先扫描标刻FEFM((First Entering First Marking, RFEFM))算法对MFV的影响。研究结果表明采用FEFM算法可以获得对于多种标刻图形的最佳的激光飞行标刻速度性能。MFV公式以及FEFM扫描算法的提出对于工业领域激光飞行标刻系统的性能评价和改进具有指导意义。进行了3D激光加工机器人仿真系统的研究。从激光加工应用的角度分析了5自由度激光加工机器人系统结构,推导了3P2R框架式机器人和5R旋转关节机器人正、逆向运动学计算公式,并分析了两种机器人运动机构的特点及定位误差。在自行开发的激光飞行标刻控制软件基础上,引入OpenGL图形引擎,自主开发实现了圆柱关节臂和STL模型关节臂的3D激光加工机器人建模装配及正、逆向运动仿真功能。研究了3D空间平面加工图形和示教3D示教插值曲线加工路径的编辑方法,并实现了机器人加工仿真。基于Open CASCADE造型内核开发了更具真实感的3D激光加工机器人仿真系统,设计实现了由IGES文件导入3D工件模型,通过对3D模型拓扑图元的分析,提取模型边数据点的位置信息和法线信息,经过融合处理后生成3D激光加工路径,实现了复杂3D零件复杂的激光加工路径的生成。3D激光加工机器人仿真的研究为研制开放式激光加工机器人控制系统以及激光加工离线编程系统奠定了基础。

【Abstract】 Laser processing robot is one of the advanced manufacturing techniques which integrates the technologies of lasers, robots, CIMS, CAD\CAM, material processing and so on. The laser beam with high power density was transferred through fiber or mirrors mounted on multi-joints mechanics and the focused laser was controlled to move along with the pre-programmed graphics accurately on the material surface to realize various 2D or 3D processing works of dots, lines or curves and surfaces.In this dissertation, the studies are focused on the control system of 2D laser flying marking and the simulation system of 3D laser processing robots. The control card based on PCI bus and the control software based on windowsWindows OS for laser flying marking were developed, the related algorithms were investigated, which includes motion-tracking model, graphics objects scanning path and laser processing path (LPP) optimization. Different modeling methods for 5-dof laser processing robots were analyzed, the kinematics formula and processing path generating methods were studied and a processing simulation system for 3D robot was realized.Laser processing path optimization problems were analyzed and an immune genetic algorithm was designed to optimize them. A novel path sequence chromosome model that consists of signed integer genes was presented, in which the absolute value of the gene is the stroke index, the position of gene in the chromosome indicates the processing order and the sign represents the processing direction. Two immune operators, smooth immune operator and 3-segs optimal immune operator based on heuristic information were brought forward to improve the GA search performance. Smooth immune operator is based on the common triangle inequality theorem, which was designed to eliminate the self-intersection faults of chromosome. 3-segs optimal immune operator is based on perfect building block theorem and it is a local optimal operator. The combination of both immune operators would improve the search speed and local-search performance of GA significantly. The experimental results showed that 48.96% of air path in a marking graphics was shortened and 8.57% of total laser marking time was reduced. This algorithm could be widely used for other applications such as laser welding, laser marking, laser show and so on.The control algorithms for laser flying marking system have been studied in this dissertation. Different motion-tracing methods including open-loop computing method and closed-loop offset-detecting method were analyzed and compared. For open-loop computing method, the moving offset is computed according to a motion tracing formula, which is only suitable for applications of constant speed. For closed-loop offset detecting method, moving offset is fed back directly from the rotary coder and it represents real displacement of work piece at any moment, hence it is suitable for both applications of constant or variable velocity. Secondly, the flying speed performance of laser flying marking system have been studied. The scanning area limitation (SAL) specialty and defined maximum markable time(MMT) and maximum marking offset(MMO) were proposed. A maximum flying velocity (MFV) formula of laser flying marking system has been deduced and the impacting factors, including marking time, the length of marking graphics in the moving direction and the marking order of graphics object, were studied. All objects entered marking(AEM), rowed first object entered first marking (RFEFM) and first object entering first marking(FEFM), were presented. Experimental results satisfied theoretical MFV calculations and it showed that there was a best MFV performance when using FEFM scanning path algorithm in transverse moving direction. Above work will benefit for the industrial field to improve the performance of laser flying marking system.3D laser processing robot simulation has been investigated in this dissertation. Firstly, the joints mechanics and laser transferring path of different dof robots were analysed. The forward and inverse kinematics equations of 3D laser processing robots (3P2R and 5R structure) have been derived and the location errors of both structure were analyzed. Based on the OpenGL environment we developed a simulation system for 3D laser processing robots, and realized the function of editing processing paths of plane graphics and intersection spline in 3D environment. Based on the Open CASCADE 3D engine, we developed another simulation system for 3D robot, in which the robot model is assembled with IGES files joints model and complex processing path of 3D accessory edges could be generated automatically from its model topology data. These work could be useful for the development of control system or off-line program system for laser processing robots in the future.

  • 【分类号】TN249;TP242
  • 【被引频次】3
  • 【下载频次】681
  • 攻读期成果
节点文献中: