Study On Seepage Monitoring Data Analysis Of Earth-rockfill Dam

Seepage exists widely in earth-rockfill dams in water conservancy projects. Seepage might cause leakage, piping effect, landslide and even endanger the whole dam. So far by now great achievements have been made in dealing with data of in-situ observation, but there are still some problems in the methods of safety supervisory materials: 1. gross error; 2. time effect; 3. dependent variable correlation of observation; 4. multi-measuring-point data fusion. Combined with the new methods of seepage monitoring data analysis in earth-rockfill dam projects, some researches have been made as follows:1. Introduce the method of principal component analysis and robust regression analysis to solve the problems of dependent variable multiple correlation and gross error, to reduce man's intervene and to raise the reliability of analysis. Build dynamic regression model, use the method of moving stations to analyze the observed data and deal with the dam's time effect with dynamic parameters. By analyzing real projects, it is drawn that dynamic statistic model is coming near real data better than static model.2. Build FEM or FDTD models, obtain in reverse the seepage coefficient of different areas by using part of the data of different observation stations. By generalized inversion, taking all the observation stations into account, comparing the changing relation of seepage coefficient of every year, dynamic features of inner dams are obtained and thus providing scientific basis for dams security. In the process of inversion, it is proposed that limited advantages be transformed to limitless advantages and good effects have been achieved. By analyzing inversion data of Qingshan reservoir dam, the conclusion is the same as the data obtained by previous basic material and data.3. In order to invert effectively seepage coefficient of dams in different areas and know something about the working methods of inner dams along with its economy and reliability, models of best arrangements of observation stations have been built. The number of observation stations is considered on the basis of dendrogram analysis and economy .By introducing generalized inversion and sensitivity analysis and genetic algorithm, observation stations are to be arranged best. Take Qingshan reservoir for instance, researches have been made to learn whether the arrangements are reasonable or not.