混合粒子群算法计算数控加工逼近误差研究

Research on Hybrid Particle Swarm Optimization Algorithm Calculating Numerical Control Machining Step Error

数控精加工中,逼近误差是刀具进给方向上相邻刀位点之间的加工误差,它的高效、精确计算是生成高质量数控加工刀轨的前提。为了提高逼近误差的计算效率,提出一种结合遗传算法的混合粒子群优化算法。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和适应度计算模型,以适应度值的最大值作为逼近误差;运用Tent映射进行粒子种群初始化,提出基于sigmoid函数的惯性权重系数和基于迭代次数的学习因子两种非线性控制方法;引入遗传算法的交叉和变异策略提高粒子全局搜索能力,以此构建出混合粒子群优化算法并进行编程实现。以典型自由曲面为例计算逼近误差并生成等误差刀轨,使用本算法的刀轨生成时间小于几何迭代算法和标准粒子群算法,验证了算法的可行性和有效性。

The step error is the machining error between adjacent tool positions in the feed direction,its ef-ficient and accurate calculation is a prerequisite for generating high-quality NC machining tool paths.In or-der to improve the computational efficiency of the step error,a hybrid particle swarm optimization method combining genetic algorithm was proposed.The mapping relationship between the cutter contact interval in step error calculation and the particle search interval and the fitness calculation model were established.The maximum fitness value was taken as the step error.The particle population was initialized using the Tent mapping.Two nonlinear control methods based on the sigmoid function for the inertia weight coefficient and based on the number of iterations for the learning factor were proposed.The crossover and mutation strategies of genetic algorithm were introduced to improve the global search ability of particles.Then a hy-brid particle swarm optimization method was constructed.The functionality of the aforementioned parallel hybrid particle swarm optimization method was implemented,some typical free-form surfaces were taken as examples to calculate the step errors and equal error tool path is generated.The calculation results show that the tool path generation time of the proposed algorithm is lower than that of the geometric iterative algo-rithm and the standard particle swarm optimization algorithm,which verifies the feasibility and effectiveness of the proposed scheme.