基于稳健估计的大坝变形监测统计模型分析

Statistical model analysis of dam deformation monitoring based on robust estimation

回归分析是大坝安全监测统计模型的核心,测值粗差的存在使回归参数变异,造成统计模型一定程度上失真,所以在回归分析中必须对其进行处理。本文采用稳健估计的思想,根据残差大小对监测值进行赋权迭代计算,使常规统计模型具有抗差性。将稳健估计的思想应用于某工程实例分析,发现与常规统计模型相比,稳健统计模型能够有效抵抗粗差对模型参数求解结果的影响,并能较好反映水压、温度和时效等效应量的影响。

Regression analysis is the core of the dam safety monitoring statistical model. The variations of the regression parameters are caused by the gross errors, resultion in the distortion of to some extent- Therefore, it must be dealt with in the regression analysis. In this paper, the idea of robust estimation is used to calculate the weighted iteration of the monitoring data according to the size of residual,so that the conventional statistical model is robust. The idea of robust estimation is applied to the analysis of the proj-ect. It is fourd that,compared with the conventional statistical model,the robust statistcal model can effec-tively resist the influence of the gross error on the solution of the model parameters,and can better reflcet the effects of water pressure,temperature and aging.