The Application Of Combined Prediction Model In Building Settlement Monitoring

Building settlement monitoring is mainly through regular and periodic observation on the prearranged monitoring points in this settlement, use the comparative analysis between each period data and the first period initial data to obtain points plane coordinate and the elevation variation. The occurrence possibility of disaster accident is judged by the precision measuring settlement body variation. Finally settlement monitoring can monitor settlement process of the settlement body, and provide reliable guarantee for the project safety and related information.In this paper, through three kinds of prediction models introduction in Wuhu City Yijiang area settlement building 11# floor settlement monitoring in prediction work application, the first chapter mainly introduces the concept, significance, development status and main prediction model of the settlement monitoring, and briefly describe writing style arranged layout and related technical route.The second chapter make specific introduction and relevant technical specification etc mainly for monitoring project of this paper, involve:types and layout requirements of settlement monitoring points,observation cycle confirmation,observation network design, observation instrument selection etc.; the third chapter is mainly about the actual construction project actual measured data analysis technology, from data prediction processing start of settlement monitoring, mainly for the singular value test and least squares fitting in the monitoring data prediction processing finding the gross errors to eliminate; the fourth chapter mainly introduces three kinds of prediction models: linear fitting prediction, gray time prediction and time series prediction respectively, etc. By processing the same raw data (5-10 period settlement monitoring raw data) to predict 11,12 period settlement data, and comparing with the actual measured values of 11 and 12 settlement data, analysing accuracy and shortcomings of several prediction model; the fifth chapter mainly introduces the combination weighting methods,classic weighting method,and optimal weighting method. And finally use the classic weighting method combine with gray-time prediction, gray prediction and time prediction to make precision analysis, and ultimately selected the most appropriate engineering prediction model, but also for the future similar settlement monitoring prediction project to provide reference data, which generally is a good scientific research and economic benefits.