Application Of Parameter Adaptive Extended Kalman Filtering Theory On Tunnel Deformation Predication

With the rapid and all-sided develop of our country’s tunneling,the amount of the tunneling which is more large scale,demanding higher specification requirements and based on deteriorate geological environment is on the rise; Monitor Measuring is of great significance on the land of construction control and methods examination,when these engineering were designed and constructed based on New Austrian Tunneling Method (NATM). However monitor measuring don’t show its ability fully on design examination and construction guidance,cause of the imperfection of monitoring means and post date processing methods and the quality of feedback information. According to the mentioned problems above, the author propose the "research on Application of parameter adaptive extended kalman filtering theory on tunnel deformation predication" based on the existing research results had carried out further in-depth study through a series of on-site monitoring and geological data.Firstly, the method of monitoring data analyses for tunnel construction was summarized, and the regression analysis, time series analysis, gray forecast method and kalman filter method were studied. Furthermore, compare the advantages and shortcomings of deformation prediction apply in tunnel construction monitoring measurement data processing were pointed out.Secondly,establish the kalman filter method for the main research approach,based on the comparation.In consideration of the non-linear speciatlty of the development of serounding rock deformate which is impact by many unknown factors.further more kalman filter is just an effective method to solve linear problem,so associate translate non-linearity into linearity to make the method work,in other words,is extended kalman filter method.however,if want this method to work normally,precise model parameter and noise accountment must be known,but they are usually unknown in practical application,thus peoples carry out all kinds of self-adaptive methods.now most self-adaptive method research are focus on variance self-adaption,and less concern parameter adaption.in order to prevent filter form divergency.the paper propose a way to adust the model parameter by residual error’speciality factor which use to judge the convergency and divergency of the filter,to make model parameter adaption come true.Lastly,establishing crown settlement and perimeter deflection parameter adaptive extended kalman filter predict model to process tunnel deformate prediction base on a series of on-site monitoring data.Comparing the parameter adaptive extended kalman filter and standard extended kalman filter’s predict values to observed values at the same time,evaluating parameter adaptive extended kalman filter’s deformate predict ability base on the comparation,thus to discuss its feasibility.The research work in this paper is on the forefront of science. Using advanced mathematical calculation and processing methods, the Application of parameter adaptive extended kalman filter on tunnel deformation predication are studied in this paper. The results have high theoretical and application value which can provide theoretical basis for dynamic feedback analysis of tunnel construction.