Research On The Application Of Partial Least Squares Regression In Analysis Of Dam Safety Monitoring

Dam prototype monitoring is not only an important measure that can grasp the operational status and guarantee safe operation of the dam but also effective mean that can test design results and monitor dam construction quality and understanding of the various physical changes in the law. Monitoring data analysis is identifying, analyzing, extracting and summarizing for the monitoring data and finding key issues from numerous monitoring data. It can extract contains information from the raw data and provide valuable information for the construction and management of the dam. Partial least squares regression method is used to analysis seepage and deformation of Heihe earth-rockfill dam because commonly statistical methods have shortcomings and defects. The study is not only of practical importance for hydraulic projects but also making useful exploration and supplementary for safety monitoring data analysis of dam and safety evaluation of a comprehensive and complementary and beneficial exploration.This paper focus on research of seepage and deformation of Heihe earth-rockfill dam, firstly based on analysis of the time and space property,the seepage and deformation rules of the dam are concluded and the safety status of the dam is given in qualitatively. Then choosing the closely related factors of seepage and deformation, the seepage and deformation statistical model are established. The seepage and deformation of partial least squares regression model are concluded from partial least squares regression method .The main conclusons are as follows:(1) Based n analysis of the time property and space property,the seepage rules of the dam are concluded and explained sciencely. On the basis of this reasonable choicing reasonable factors, The seepage of partial least squares regression model are established. Evaluate the stability of seepage field using the reasonably criterions.(2)Bootstrap Methods are used to test the parameters of establishment partial least squares seepage statistical model. Variables that have no explanation of the system are deleted so a more reasonable and effective set of variables are got.(3)The paper Analyzing vertical deformation observational data of Heihe earth-rockfill dam and made qualitative evaluation for the dam vertical deformation situation. Deformation is effected by filling high and aging, the dam deformation analysis model are established. Achieving the qualitative analysis of the deformation thruogh partial least squares regression. Evaluate the stability of deformation using the criteria of aging.