粒子群优化算法计算数控加工逼近误差研究

Research on optimized particle swarm algorithm calculating numericalcontrol machining step error

逼近误差的高效、精确计算是生成高质量数控加工刀轨的前提,为了提高给定精度下逼近误差的计算效率,提出逼近误差和粒子群计算原理紧密融合的粒子群优化算法。建立逼近误差计算的刀触点区间与粒子搜索区间之间的映射关系,根据逼近误差几何原理和粒子适应度特点,提出粒子群算法下的逼近误差适应度函数。为了实现最优解迭代搜索前期范围大、后期效率高,提出基于余弦函数的惯性权重系数非线性控制方法和基于迭代次数的学习因子控制方法。最后以上述方法完成粒子群算法的优化和开发,并用于自由曲面的逼近误差计算,实验结果显示计算时间比常用的几何迭代算法更短。

Calculating step error efficiently and accurately is the premise to generate high-quality CNC machining tool path.In order to improve the calculation efficiency of step error with the given precision,a particle swarm optimization algorithm is proposed which closely integrates principles of step error and particle swarm calculation.The mapping relationship between intervals of cutter contact and particle search was established.According to the geometric principle of step error and the characteristics of particle fitness,the fitness function of step error in particle swarm optimization algorithm was proposed.In order to make the iterative search of optimal solution large range in the early stage and high efficiency in the late stage,the nonlinear control method of inertia weight coefficient based on cosine function and the learning factor control method based on iteration number are proposed.Finally,the optimization and development of particle swarm optimization algorithm is completed with the above methods,and used to calculate step error of free-form surface.Experimental results show that the calculation time is shorter than that of commongeometric iterative algorithm.