大坝安全监测统计模型的比较与选择

Research and Opting on Statistical Models of Dam Safety Monitoring

统计模型在大坝安全监测资料分析中占有重要地位,是建立混合模型的基础。从统计模型选择的角度出发,对多元线性回归、逐步回归以及偏最小二乘回归的应用特点和效果进行了探讨,其中偏最小二乘回归是一种旨在消除多重相关性的新型回归技术。重点结合某水电站大坝安全监测资料,从模型解释能力、统计参数的检验、拟合和预测效果、效应量的分离等方面对上述3种回归方法进行了比较性研究,得出了一些有益的结论。

Statistical models of dam safety monitoring are the most useful method in the analysis of dam safety monitoring data, which are necessary for us to get deterministic models and hybrid models. The essential theory of three kinds of statistical models, which are Multiple Linear Regression, Stepwise Regression and Partial Least - Squares Regression （PLSR） , are presented and their applied characteristics are discussed in this paper. PLSR can effectively overcome the harms of Muhicollinearity among variables. Combined with the dam safety monitoring data of WanAn hydropower station, the comparison of model validity analysis shows that the statistical models based on PLSR are reasonable on fitting accuracy, forecasting capability, test for statistic parameter and dividing of time effect.