ZHU Zhaohui, LIU Jian, LI Xin, SHANG Ceng, WANG Guochuan

Abstract: 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 project. It is fourd that,compared with the conventional statistical model,the robust statistcal model can effectively 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.

Keywords: deformation monitoring; robust estimation; statistical model; gross error

(Journal of China Institute of Water Resources and Hydropower Research, Vol. 16, No. 2, 2018)