Stability Analysis for Subgrade Settlement Prediction by Curve Fitting Methods

2nd International Symposium on Resource Exploration and Environmental Science

生态环境科学

Yin, Lihua^1 ; Sun, Xudong^2 ; Yang, Ping^3 ; Li, Meixia^3

CCCC First Highway Consultants Co. Ltd., Xi'an, Shaanxi

710075, China^1

China Railway Group Limited, Beijing

100000, China^2

China Railway Pearl River Delta Investment Development Co. Ltd., Foshan, Guangdong

528300, China^3

关键词: Curve fitting methods;    Error amplification;    Evaluation indicators;    Fitting parameters;    Prediction methods;    Regression formulas;    Stability analysis;    Subgrade settlement;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/170/3/032050/pdf
DOI  :  10.1088/1755-1315/170/3/032050

学科分类：环境科学（综合）

来源: IOP

【 摘 要 】

The curve fitting methods are widely used in predicting the settlement of subgrade because of few and easily determined parameters, especially for hyperbola method, Xingye method and Asaoka method. In applying this kind of prediction methods to predict the subgrade settlement, the afterward settlement can be obtained by the regression analysis of advance monitoring data, and the stability analysis for subgrade settlement prediction by curve fitting methods has become a noteworthy and valuable issue in engineering. In this paper, the evaluation indicators of fitting parameters by hyperbola method, Xingye method and Asaoka method are discussed, and the stability of predicted results by these three methods is analyzed through some engineering examples. The results show that there is a closely relationship between the stability of predicted results by curve fitting methods and fitting parameters in the regression formula. In predicting the subgrade settlement based on hyperbola method and Xingye method, the error amplification coefficient of fitting parameters is small and the predicted results show good stability when setting different initial time. However, as for Asaoka method, the error amplification coefficient of fitting parameters is large and the predicted results show instability when setting different initial time and time spacing. In addition, the predicted results obtained from Asaoka method are greatly affected by randomness of sample points.