【摘要】 从圆柱度误差的测量要求和圆柱度误差评定搜索算法两个方面研究了提高测量精度的方法。为了减小圆柱度误差测量中的工件倾斜误差,设计并分析了两点垂直布局的调平方法;根据工件轴线的方向余弦,计算得到了两点调整的高度值,克服了手动调整存在的问题,实现了工件快速精确调平并提高了工件圆柱度的测量精度。由于圆柱度误差评定是对满足最小条件的圆柱轴线的搜索,文中针对Nelder-Mead单纯形法的收敛精度依赖于初始解和收敛速度较慢,提出了拟牛顿法和Nelder-Mead单纯形法相结合的联合算法来实现全局最优解的快速准确搜索。对经典测试函数的Matlab仿真及实际测量数据的应用表明,该联合算法能有效地提高收敛速度和收敛精度,其收敛速度提高了50%,收敛精度提高了1倍,从而提高了工件圆柱度误差的测量精度。更多还原

【Abstract】 How to improve the measuring accuracy of a roundness measuring instrument was researched in consideration of the cylindricity error requirement and cylindricity error assessment searching algorithm.To reduce inclination error of a workpiece in measuring cylindricity error,a leveling method using a dual-point vertical layout was proposed and analyzed.According to the direction cosine of the axis of workpiece,the amount of leveling was defined and calculated to overcome the problem brought by manual adjustment technology and the accurate leveling and high precise measurement were completed.As the assessment of cylindricity error is to search for a cylinder axis that statisfies the minimum condition and the Nelder-Mead simplex method has lower convergence precision and convergence rate because it depends on initial solution,so that a combined algorithm with quasi-Newton method and simplex algorithm was proposed to achieve the fast and accurate search for global optimums.The Matlab simulation andexperimental testing on classical functions indicate that the combined algorithm improves the convergence rate and convergence precision effectively and the convergence rate has been increased over 50% and convergence precision increases over once as compared with that of Nelder-Mead simplex method,which ensures and improves the measuring precision of workpieces.更多还原