一种准确度高的土工测试数据拟合新方法

A New Best Fit Approximation Method of Soil Test Data

针对目前岩土工程土工测试数据处理方法的不足,给出了一种能够减少端点导数误差的三次样条拟合方法。三次样条函数拟合是曲线拟合的一个公认的较好方法,它具有很好的分段光滑性。但当端点的导数值无法精确得到时,会给样条拟合的准确性带来较大的影响。为了尽量减少这种误差的出现,本文提出先用多项式函数对前、后各一小部分测量数据点进行拟合后,再对函数求导,得到所需端点导数带入三次样条函数,最终得到更加准确的拟合结果。采用文中建立的拟合方法,对地基沉降实测数据进行分析,拟合结果表明本文方法的工程适用性强、拟合准确度高。

In order to optimize the methods of current soil test data processing in the study of geotechnical engineering, this paper presents a cubic spline function with small error and precise results to analyze the test data, Cubic spline function fitting is a good method in the curve fitting because it is characterized by very good subsection smoothness. But it is not easy to gain precise results of cubic spline fitting at both ends since their first derivatives are not exact. In order to minimize the error, this paper proposes that the eventually accurate results could be gained first by fitting small amount of data in the preceding and the back part of the measured data respectively using multinomial function. Secondly, the derivatives of this function at both ends are taken, which are introduced to the cubic spline function. The test data of ground settlement were analyzed by this fitting method. Results show that this fitting method presents strong engineering applicability and high fitting accuracy.