季节性冻土区路基高边坡变形多因素时变预测模型

Multifactor time-varying model for the prediction of subgrade high-slope deformation in areas with seasonally frozen soil

为了研究季节性冻土区路基在各种影响因素作用下的运营质量,本文对路基高边坡进行沉降变形预测研究。鉴于当前路基变形各种预测模型均有其适用范围,总体预测波动性较大,精度较低,提出了一种基于支持向量机的季节性冻土区路基高边坡变形多因素时变预测模型。依托季节性冻土区路基观测系统,在分析相关监测数据的基础上,对季节性冻土区路基高边坡变形特征及其热稳定性影响因素进行分析,并总结主要因素对高边坡路基变形的影响规律。借助该周期性变化规律采用支持向量机理论构建了变形多因素时变预测模型。实验结果表明:本文模型在预测精度上较当前预测模型在预测精度上有所提高,可为岩土工程建设提供重要的参考依据。

This work predicts the settlement and deformation of high-slope roadbeds to investigate the running quality of subgrades under the influence of various factors in soil areas undergoing seasonal freezing.Different models for subgrade deformation prediction have large overall prediction volatility and low precision given their different specific application scopes.Thus,in this work,a multifactor time- varying model for the prediction of subgrade high-slope deformation in areas with seasonally frozen soil is proposed.The model is based on the support vector machine (SVM).Relevant monitoring data are analyzed on the basis of the observation system of subgrades in areas with seasonally frozen soil.The factors that influence the deformation characteristics and thermal stability of high-slope subgrades in areas with seasonally frozen soil areas are studied,and the influence of the main factors on the deformation of high-slope subgrades is summarized.A multifactor time-varying deformation prediction model is constructed by using SVM theory with the aid of this periodic change law.Experimental results show that the proposed model has better prediction accuracy than the current prediction model and can provide important reference for geotechnical engineering construction.