清江隔河岩库区茅坪滑坡变形分析

Deformation analysis on Maoping landslide in Geheyan reservoir area

茅坪滑坡是清江隔河岩库岸规模最大的滑坡，具有稳定性差、崩滑方量大等特点。在地表位移监测资料分析的基础上，选用逐步回归分析法建立了滑坡体位移模型。逐步回归法是将ｍ 个自变量先做为候选变量，计算出各自的贡献（ 偏回归平方和） 与方差比Ｆ，再与给定等级检验临界值Ｆα作比较，若Ｆ≥Ｆα，自变量被引入；若Ｆ＜ Ｆα，自变量将被剔除。如此计算、比较，使模型中的自变量减少到最合理的程度。此法与经典多元线性回归方法相比，不仅简化了计算过程，消除了计算误差，且能对所建模型进行修正。模型计算成果表明，茅坪滑坡体主要受库水位涨落与大气降水作用的影响，具有中后部推移和前部牵引下滑的变形特征。

Maoping landslide is the largest landslide in Geheyan reservoir bank, and has such features as low stability, large amount of slide mass. On the basis of analysis on monitored data of the surface displacement, the displacement model of landslide mass is established with progressive regression analysis method. The method is carried out as follows: takes m independent variables for selection and computes contribution of the variables (partial regression square sum) and variance ratio F one by one, and compares the results with given grade check critical value F\-α, ifF≥F\-α,introduce the independent variable, if F<F\-α, delete the independent variable. The computation and comparison are carried out repeatedly in this way until the independent variables are reduced to the most reasonable number. Compared with classical multi variable linear regressive method, this method not only can simplify the computation process and eliminate computation error, but also can revise the established model. The results of model computation demonstrate that Maoping landslide mass is mainly influenced by water level fluctuation in the reservoir and precipitation in the area and has the deformation characteristics of pushing in the middle and back side of the mass and pulling and sliding in the front part, with great potential danger. Subject words Systematization of observation Slide monitoring Regression model Regression analysis Geheyan Hydropower StationFurthermore, the gain margin and phase margin for the uncertain closed-loop system with the desgn of robust quadratic guaranteed stable control.