提高级配指标计算精度方法研究

级配曲线是颗粒分析试验的重要成果图件,用其计算的不均匀系数Cu与曲率系数Cc是路基填料划分的重要指标。传统作图法整理级配曲线时,工作量大且精度不高。在分析线性插值、多项式插值、Lagrange插值与分段抛物线插值算法的基础上,提出了修正样条函数拟合的方法,得到土的光滑颗粒级配曲线。针对非单调样条函数,采用最佳逼近算法求解,克服了曲线出现平台时病态函数问题。文中提供了求解的关键代码,实践表明该法可以提高计算级配指标的精度和自动化程度。

A Research to Enhance Computation Precision of Gradation Index

The gradation curve is an important achievement graphic document of the grading analytical investigation.Nonuniformity coefficient（Cu） and curvature coefficient（Cc） are important target to divide the road padding.The amount work is great and computation precision is low when using traditional graphing method.Based on the analysis of common algorithm such as linear interpolation,polynomial interpolation,Lagrange interpolation,sectoring parabola interpretation,the paper proposes the revision spline function fit method,and then obtain smooth gradation curve.In view of the non-monotonous spline function,the optimal approximation algorithm is used to solve and overcome the curve which has platform caused ill-conditioned curve.The method enhances the computation precision and automation of gradation index and key codes are provided.