项目名称: 复杂叶片机器人砂带磨抛在线测量与余量优化
项目编号: No.51475187
项目类型: 面上项目
立项/批准年度: 2015
项目学科: 机械、仪表工业
项目作者: 李文龙
作者单位: 华中科技大学
项目金额: 86万元
中文摘要: 视觉引导的机器人砂带磨抛是复杂叶片精密加工的前沿技术,在线测量数据规模大导致长时间停机计算以及匹配算法陷入局部最优值加剧磨削颤振,是制约其应用的关键问题。本项目研究无干涉激光扫描方法,生成四元数表示的运动轨迹,操作机器人臂在线采集叶片型面点云;研究三维数学形态学、点云密度聚类和点-面距离函数,建立大规模点云精简的数学模型和快速计算方法,提出点云匹配全局最优解(或接近最优解的可行解)的定量判别方法,解决传统方法受在线测量缺陷影响或易陷入局部最优值的难题;根据叶盆叶背磨削深度工艺需求,重构基于不同权重的点云匹配目标函数,通过坐标系的微分平移和旋转实现凹凸面加工余量优化;在以上研究基础上,开发机器人砂带磨抛专用的在线测量软件,完成多款叶片的砂带磨抛实验,验证叶片加工精度和表面质量。项目研究成果将为中小型汽轮机叶片、航空叶片、螺旋桨桨叶的机器人砂带磨抛提供基础理论和核心技术。
中文关键词: 复杂曲面;砂带磨抛;在线测量;点云匹配;余量优化
英文摘要: Robot belt grinding based on vision method is an advanced technology of machining complex blades. However, its application is restricted by outage waiting and grainding chatter. 1) The visual cone theory is applied to research no-interference laser scanning method, generate quaternions motion curve, and control the robot manipulator for on-line scanning blade surface; 2) Mathematical morphology, density clustering and distance function is applied to research mathematical modeling and computing method of large point-cloud data processing, including measuring defect filtering, curvature-adaptive simplification and data-design model matching, and a quantitative analysis method is proposed to determine the optimal matching solution or its feasible solution. The aim is to overcome the problem of measuring defect disturbing or trapping into a local solution; 3) According to the grinding depth requirement, a new objective function of point-cloud matching is reconstructed, from which differential translation and rotation is perfomed to obtain allowance optimization of blade's concavo and convex surfaces; 4) A special on-line measuring software for robot belt grinding is developed to carry out the machining experiment of complex blades with different types, and test the machining accuracy and surface quality of blade. The project results will provide theory basis and key technolies for the robot belt grinding of small turbine blades, aviation blades and propeller blades.
英文关键词: Complex Surface;Belt Grinding;On-line Measurement;Point-cloud Matching;Allowance Optimization