空间直线度误差评定的逼近最小包容圆柱法

Evaluating spatial straightness error by approaching minimum enclosure cylinder

针对目前空间直线度误差评定中结果误差过大或者因采用进化算法耗时太长的问题,提出一种定向旋转包容圆柱轴线的方法.通过将测量点投影至最小二乘中线的中垂面,在中垂面内求出满足国标要求的2种情况的最小包容圆.针对2点在包容圆上的情况,做2次坐标变换,然后确定搜索方向,定向旋转圆柱体轴线,找到更加接近最小包容圆柱体的轴线,从而得到更小的空间直线度误差评定值.本方法主要计算过程中的搜索方向明确,无反复迭代,鲁棒性好.数据实验表明:本方法得到的误差评定结果比其他几种方法的都小,结果更接近真实值,适合于直线度误差评定精度要求高的场合.

An evaluating algorithm of rotating the axis of minimum enclosure cylinder in determined-direction was presented.The two kinds of minimum enclosure circles that meet the requirement for GB were computed,when the measured points were projected to the median plane of the least squares mean line.In the situation of the two points on the minimum enclosure circle,the coordinate transformation was done twice,then the rotating direction was determined,and the axis of cylinder was rotated to get a new axis that more close to the theoretical minimum enclosure cylinder.Therefore a very little straightness error was gotten.This algorithm is very robust for its determined searching direction and no iteration.Numerical experiment shows that the result of this algorithm is less than other′s,which indicates that the result is more close to the true value.This algorithm is more appropriate to evaluate spatial straightness error.