Generating Machine-Optimal Numerical Control Paths

本研究的目的是生成最快的刀具轨迹来扫描三维表面。这将通过首先开发一个数学公式来实现，该公式捕获了加工的所有实际约束，包括电机速度限制，加速度限制，表面光洁度要求，机器运动学和路径不连续的影响。公式化的问题很可能不会产生封闭形式的解。将测试几种数值方法。一种“贪婪”的方法已经显示出初步的希望。在这种方法中，工具路径是使用局部最优性准则代替全局最优性准则生成的。第二种方法是使用参数化“基”曲线生成候选刀具轨迹族。将研究的第三种方法是全局搜索；表面将被分解成越来越小的单元，并且每个单元中的刀具路径方向将单独变化。第三种方法在计算上很昂贵，但也可能产生最好的结果。如果成功，该研究将为高速加工提供更快的刀具路径。随着数控机床主轴转速的不断提高，直线驱动系统日益成为制约机床加工性能的瓶颈。升级到更快的直线电机技术是非常昂贵的。通过这项研究实现的刀具路径将更好地利用当今机床结构与新主轴技术的能力，并在不显著增加成本的情况下实现更高的生产率。

The objective of this research is to generate the fastest tool paths to sweep a three-dimensional surface. This will be accomplished by first developing a mathematical formulation which captures all the practical constraints of machining, including motor speed limits, acceleration limits, surface finish requirements, machine kinematics and the effects of path discontinuities. It is likely that the formulated problem will not yield a closed-form solution. Several numerical approaches will be tested. A "greedy" approach has shown initial promise. In this approach, the tool paths are denerated using a local optimality criterion in place of the global optimality criterion. A second approach is to generate candidate tool path families are using parametric "basis" curves. A third approach which will be studied is a global search; the surface will be broken into smaller and smaller elemefltts, and tool paths directions will be varied individually in each element. The third approach is computationally expensive, but is also likely to yield the best results.If successful, this research will yield faster tools paths for high speed machining. As the spindle speeds available in CNC machine tools have increased, linear actuation systems have increasingly become the bottleneck in machining performance. Upgrading to faster linear motor technologies is very expensive. The tool paths enabled by this research will better utilize the capabilities of todays machine tool structures with new spindle technologies, and enable greater productivity without significantly greater costs.