基于变权重组合模型的路基沉降预测方法

Prediction Method of Subgrade Subsidence based on Variable Weight Composite Model

通过监测数据的沉降预测分析，是解决当前高速铁路路基沉降计算精度不足的主要途径。由于采用常见的反S形成长曲线模型在路基沉降预测时，不能完全预测出沉降曲线的整个发展过程，本文提出了一种基于多种曲线变权重组合的沉降预测模型。该方法结合多种反S形成长曲线模型，赋予一个未知的权重系数，通过最小方差方法对权重系数进行求解，从而得到沉降预测公式。工程实例的计算结果表明：该方法具有较好的适应性和较高的精度，可以应用于高速铁路路基最终沉降量的预测。

Settlement prediction based on the data monitoring is the main solution for computing precision of the high-speed railway subgrade settlement. Inverse sigmoid curve is a new method for subgrade settlement prediction, in which the settlement development process can＇t be predicted properly. This paper presents a new model by combining these inverse sigmoid curves,in which each curve has an unknown weight that is solved by the least variance method and the final settlement can be calculated. The engineering results showed that this method has good adaptability and high precision, which can be used to predict the final subgrade settlement of the high speed railway.