基于CDE的空间直线度误差评定

Spatial Straightness Error Evaluation Based on CDE

针对空间直线度误差评定中求解精度不高等问题,提出基于改进差分进化算法的空间直线度误差评定方法。使用产品技术几何规范(GPS)公差标准,通过最小二乘法计算得到符合最小区域法的空间直线度数学模型;将差分进化算法(DE)加入Cubic混沌映射使种群初始化,并对算法中变异因子和交叉概率进行改进,经测试函数仿真对比证实,该算法在收敛速度、精度上均有一定提高。对两个评定实例进行误差评定,研究结果表明:相比于HTMLBO,PSO,DE,ABC算法,CDE算法在计算精度、速度、稳定性上更具有优势,在同一零件的内径上提取截面圆圆心坐标时,规定最佳提取点数可以有效降低算法误差。

Aiming at the problem of low accuracy in spatial straightness error evaluation,an improved differential evolution algorithm based on spatial straightness error evaluation is proposed.Based on the product technical geometry specification(GPS)tolerance standard,the spatial straightness mathematical model is established by least square method.The differential evolution algorithm(DE)is added into Cubic chaotic mapping to initialize the population,and the variation factor and cross probability in the algorithm are improved.The simulation results show that the algorithm has improved convergence speed and accuracy to some extent.The error evaluation of two evaluation examples is studied.The results show that compared with HTMLBO,PSO,DE and ABC algorithms,CDE algorithm has more advantages in calculation accuracy,speed and stability.At the same time,when extracting the center coordinates of the section circle from the inner diameter of the same part,the optimal extraction points can effectively reduce the algorithm error.