The Applied Research On Adaptive Kalman Filter In Data Processing Of Deformation Monitoring

With the progress of human society and the rapid development of economy,the construction process has been accelerated.The scale of engineering structures,modeling,the difficulty in an increasingly high demand.Large buildings or structures of safety directly related to the production process of national economy.Therefore,the deformation monitoring for the large-scale building or structure and analysis of monitoring data are become particularly important.At present,the common data-processing methods of the deformation monitoring are regression analysis,time series analysis,and the method of gray system analysis.Also application of Kalman filter is a broader way to deal with dynamic data.It introduced the concept of state space,with the state transition equation of the system,according to the previous value and the current observations,it estimated the new recursive state estimators. The adaptive Kalman filter has been on the dynamic information to amend the noise variance vector and calculates the closer actual status.If adopted inaccurate mathematical models and noise design,Kalman filtering may lead to state estimation distortion,even lead to filter divergence.And Adaptive Kalman filtering is use the observational data to recursive filtering, and at the same time,to properly estimate and amendment the unknown or inaccurate model parameters and the statistical characteristics of noise to weaken the impact of model error,and make the filter results closer to reality.Based on lots of the documents and materials,firstly this article describes the significance of the deformation monitoring and the monitoring technology and methods of data processing at home and abroad.Secondly,the research expound the basic principles of Kalman filtering systematically,and focus on research the use of adaptive Kalman filtering variance theory to set up the monitoring of data-processing theory,methods and models. Finally,based on the deformation monitoring data of the Sanmenxia Water Control Project, using the classical least squares,general Kalman filtering,adaptive Kalman filter variance compensation three methods to carry on processing,and with the Matlab software establishment data reduction program,to verify the results of data processing for compensates Adaptive Kalman filter.Through the comparative analysis,Adaptive Kalman filter variance compensation data processing better meet the needs of the project.