完整桩极限承载力的偏最小二乘回归预测法

Prediction of the ultimate bearing capacity of the integrate pile by using the partial least-squares regression method

关于完整桩轴向极限载力Qu与桩长L、桩径d、桩的阻尼自振基频f1、桩的应力波波速c以及桩的单位动刚度Kd等参数,已知Qu的大小与上述5个参数存在一定的关系,现采用偏最小二乘回归方法对其进行描述。偏最小二乘回归方法是近年来产生和发展的一个具有广泛适用性的多元统计分析方法。其特有的选择因子方式与传统方法迥然不同,而其计算量比传统方法都小。它意义明确,计算简单,建模效果好,解释性强,日益成为工程技术人员和经济管理工作者能够熟练掌握的实用工具。

The ultimate bearing capacity of the integrate pile has relations with the pile's length,diameter,damped self-excited fundamental frenguency,velocity of the stress wave and unit dynamic stiffness. In this paper,the partial least-squares regression method is adopted to establish the prediction model of the ultimate bearing capacity of the integrate pile. The partial least-squares regression method,which has been generated and developed to fit the practical need,is a method for the multivariate statistic analysis and it can be used widely. This method is quite different from the traditional methods in the mode of selecting factors,and the calculating work is less. The introduced method is definite in meaning,simple in calculating,wonderful in the modeling effect and good in explanation,so it is becoming a practical tool in the fields of engineering and economic management.