路基沉降不确定性预测研究

Uncertainties in the Embankment Settlement Predictions

路基沉降计算和预测是工程建设中一个重要的课题,但是很多情况下理论计算的沉降量与实际沉降量有较大的出入。所以必须对路基沉降进行监测,根据实测数据控制填土速率,保证路堤在施工中的安全与稳定;根据实测沉降曲线预测工后沉降,使工后沉降控制在设计允许范围之内。但沉降预测模型的参数是不确定的,用概率和概率分布去描述更加合适,将模型参数视为随机变量,基于贝叶斯理论和MCMC方法,借助WinBUGS软件,建立了贝叶斯沉降时间序列不确定性预测模型。实例分析表明,该方法所得结果合理可信,将其应用于路基沉降预测是可行和值得研究的。

The analysis and prediction of settlement is one of the key issues in the foundation design practice. In many cases thesettlement values from the theoretical analysis and the monitoring results are often not even close. Therefore, it is important tomonitor the embankment settlement during the construction, and the embankment filling rate should be controlled by the settlement rate so that the safety of the embankment can be assured during the construction. The post construction settlement magnitude can also be predicted and mitigated based on the measured settlement during the construction. However, there are uncertainties in the settlement model parameters, it is more appropriate to describe the settlement pattern with probability and proba-bility distribution methods. The parameters are described as random variables and the model is established based on the BayesJan theory using WinBUGS software in the time domain. The prediction results are compared with the observed values. It isconcluded that this method is a reliable method.