个人信息Personal Information
教授
博士生导师
硕士生导师
性别:男
毕业院校:中国科学院沈阳自动化研究所
学位:博士
所在单位:机械工程学院
学科:车辆工程. 机械制造及其自动化
办公地点:综合实验2号楼423
电子邮箱:xujt@dlut.edu.cn
Kinematics performance oriented smoothing method to plan tool orientations for 5-axis ball-end CNC machining
点击次数:
论文类型:期刊论文
发表时间:2019-07-01
发表刊物:INTERNATIONAL JOURNAL OF MECHANICAL SCIENCES
收录刊物:SCIE、EI
卷号:157
页面范围:293-303
ISSN号:0020-7403
关键字:5-axis machine; CNC machining; Tool orientation; Kinematics performance; Angular acceleration
摘要:When adopting 5-axis machine to mill the parts, it is desired to avoid the drastic change of tool orientation for improving the kinematics performance of 5-axis machining while ensuring no machining interferences. For this purpose, a kinematics performance oriented smoothing method is proposed to plan the tool orientations, which is focused specifically on minimizing the angular accelerations imposed on the rotary axes of 5-axis machine. In this method, with several specified representative tool orientations (RTOs), two B-spline curves, which represent the displacements of the rotary axes, are used to join smoothly the RTOs together and then to determine the tool orientations at other areas. The solutions for the two B-spline curves are achieved by solving a least-square objective function which minimizes the angular accelerations of the rotary axes. To restricting simultaneously the interpolated tool orientations in the geometric feasible domains (GFDs) of tool motion, a simple alternate strategy of first smoothing the tool orientation and then checking the machining interference is developed so that tool orientation planning and its geometric constraints are decoupled and the complicated constraint optimization process of tool orientation can be greatly simplified. Since the proposed method works in the machine coordinate system (MCS), it can not only ensure the smooth motions of the rotary axes without the machining interferences, but also can generate directly the rotary axis orders. Finally, the proposed method is validated by the experiments.