基于KJHH模型的基坑开挖概率反分析方法

Probabilistic inverse analysis for foundation excavation based on KJHH model

基坑开挖工程中,最大地表沉降和最大墙体侧移是非常重要的两个变形量。然而由于土体存在变异性,基坑开挖变形难以准确预测。提出一种基于KJHH模型的基坑开挖概率反分析方法,可以同时预测最大地表沉降和最大墙体侧移。在贝叶斯更新框架下,动态融合各开挖阶段观测数据,利用多重数据同化集合平滑器(Ensemble smoother with multiple data assimilation, ES-MDA)更新土体参数,提高变形预测准确性。以台北TNEC基坑工程为例,验证了所提方法的有效性。实验结果表明:随着融合更多不同开挖阶段的观测数据,预测均值和实测值趋于一致;对于TNEC工程,假设先验分布服从对数正态分布得到的开挖变形预测结果略大于采用均匀分布时的预测结果;变形预测的准确性随着迭代次数和样本量的增加而提高。

The maximum surface settlement and the maximum wall lateral movement are two very important deformations in foundation excavation projects.However,the accurate deformation prediction in foundation pit excavation is a challenging due to the variability of the soil body.In this paper,a probabilistic inverse analysis method is proposed for foundation pits based on the KJHH model,which can predict the maximum surface settlement and lateral movement.Under the Bayesian updating framework,the observed data from each stage are dynamically fused and the soil parameters are updated using the ensemble smoother with multiple data assimilation(ES-MDA)to improve the prediction accuracy of deformation.The validity of the proposed method is verified through the TNEC foundation excavation project in Taipei.The results show that the predicted mean values and the measured values are in agreement with the integration of more observation data from different excavation stages.For the TNEC project,the excavation deformations obtained by the lognormal distribution assumption are slightly larger than those obtained by the uniform distribution assumption.In addition,the prediction accuracy of deformation increases with the number of iterations and the sample size.