Study On Dynamic Evalution Of Dam Bedrock By Safety Monitoring Based On SVM And Grey Theory

The research and application on monitoring data analysis of dam has madeconsiderable progress, which played a huge role in ensuring safe operation of dam.However, there are still many problems and deficiencies in data analysis. Dam bedrockis the main part of the dam. The study on comprehensive evalution on dam bedrock isstill less. Most studies have focused on single index of dam bedrock and gainedevalution results through these single index. In view of the problems and insufficiencyof current analysis methods and models, indicators weight model of dam bedrock hasbeen establishend and safety comprehensive evalution on the bedrock operational statushas been made combining with the grey theory. In the meantime, SVM model has beenintroduced to analysis and forecast of monitoring data. Dynamic evalution system hasbeen formed combining grey theory with SVM model. The system achieved moreeffective and rational evalution on current and future status of dam bedrock， whichmeeted the actual need for engineering applications.Because of many influent factors of the safety of dam bedrock, the current technologylevel is still difficult to obtain comprehensive information of dam bedrock runningstatus. Just because the limited awareness and disrupted by pseudo information, themonitoring information of dam bedrock is not complete, that’s mean, gray information.Gray clustering comprehensive evalution model was used to estimate the safety statusof dam bedrock in this paper in order to avoid the in insufficiency of evalution by singleindex. For the importance of weight in a comprehensive evalution system and the basisof current weight model, two aspects of tradition projection pursuit (PP) model can beimproved, resulting in a constrained ME-PP weight model which fully taped theinformation of dam monitoring data and considered characteristics of the dam rookfactors.The weight model was used to estimate the status of the dam bedrock.Since the dam operation managers not only want to know the current status of thedam bedrock but also want to know the development of the status, SVM model wasintroduced to dam bedrock displacement predicition.The article establied a SVM modelbetween the displacemet of dam bedrock and the environmental variables to predict thedisplacement of the dam bedrock. Based on small samples of the monitoring data， twosuitable small sample mathematical model, grey theory and SVM model, werecombined—using dynamic SVM model to predict the displacement of the dam bed rock and using grey theory to dynamicly evaluate the status of the dam bedrock in whichconstrained ME-PP weight model was applied.