逐步回归算法在边坡安全监测中的运用

考虑时效、降雨和温度对边坡变形的影响,深入分析了边坡变形的外在影响因素及各因子的表达形式,建立边坡变形统计回归模型,应用Matlab软件和逐步回归算法,实现边坡变形的回归分析与预测。东津水电站溢洪道左岸高边坡变形监测资料分析成果表明:边坡变形统计拟合值与实测值接近,复相关系数较大,剩余标准差较小。该模型能较好地反映边坡的变形规律和发展趋势,为分析边坡运行时的安全状态和预测变形趋势提供了指导思想,有较大的工程实用价值。

Application of Stepwise Regression Algorithm to Slope Safety Monitoring

Considering the effect of time, rainfall and temperature on slope deformation, we analyzed the deformation factors and their expressions in depth. We also established a statistical regression model of slope deformation through Matlab software and stepwise regression algorithm to carry out regression analysis and prediction of slope deformation. Analysis on the deformation monitoring data of high slope on the left bank of spillway of Dongjin hydropower station showed that the statistical regression values of slope deformation were in good agreement with measured values, with large multiple correlation coefficients and small residual standard deviations. This model can well reflect the slope deformation and its trend, and it provides guidelines for analyzing the safety and predicting the deformation trend.  