论文部分内容阅读
为了解决水轮机叶片坑内修焊空间加工作业需求,研制了适用于复杂曲面的两端吸附式爬壁机器人,该机器人由永磁间隙吸附式移动平台、多自由度机械臂(包括3个主动关节和3个被动关节)和永磁间隙吸附式末端作业单元组成。针对给定末端路径,这种结构的机器人需基于局部平面假设来完成主动关节的轨迹生成。但经仿真分析,在1.5m半径外球面上的简化造成的误差达到5mm以上,不满足修焊工艺要求精度。为此,提出在机械臂加工运动过程中,通过Jacobi矩阵将末端作业单元在Descartes坐标系下的误差转换为关节角修正量以完成动态修正的算法。仿真实验表明,该算法可有效降低运动路径误差至1mm以下。
In order to solve the need of repairing welding space in turbine blade pit, two-end adsorption wall climbing robot suitable for complex curved surface is developed. The robot consists of permanent magnet gap adsorption type mobile platform, multi-degree-of-freedom manipulator 3 passive joints) and permanent magnet gap adsorption end of the unit. For a given end path, a robot of this structure needs to complete the trajectory generation of the active joint based on the local plane assumption. However, the simulation analysis shows that the error caused by the simplification on the 1.5 m radius spherical surface reaches more than 5 mm, which does not meet the requirements of the repair welding process. For this reason, an algorithm is proposed to convert the error of the end-effector unit in Descartes coordinate system into the joint angle correction through the Jacobi matrix to complete the dynamic correction during the machining process of the manipulator. Simulation results show that this algorithm can effectively reduce the error of moving path to below 1mm.