The Adaptive Kalman Filtering In The Application Of Deformation Monitoring Data Processing Research

With the progress of human society and the rapid development of economy, all kinds of projects are strongly under construction in each city. And the safety of engineering buildings is directly related to the production process of the national economy, therefore, the prompt and accurate for deformation monitoring and the analysis and prediction of monitoring data are particularly important.This paper introduces theoretical knowledge of classical adjustment, rank defect adjustment, and quasi-stable adjustment about the deformation monitoring network, these three adjustment methods are used respectively in calculating data of a settlement leveling network, for example, by calculating the results to show that different adjustments reference can get different deformations, when adjustment is made, the selection of the frame of reference should be according with the actual situation. Also the method of the stability analysis of point location about deformation monitoring network is expounded, the main research is on the average gap method, which is used for the overall test for the stability of monitoring network, as well as the t-test of the judgment about the significance of single point displacement, these two methods are applied to the stability analysis of point location about a settlement monitoring network.At present, there are many traditional processing methods of deformation monitoring data, and Kalman filter is also widely used as a kind of dynamic data processing method. A standard Kalman filter dynamic model is a combination of the parame-ter estimation and prediction, which can reflect the real time state of the moving target, but it also has some limitations, which spread easily in some conditions, that can cause shortcomings such as uncredible state estimation, the difficult selection of initial parameters and so on, so there are some deficiencies in the practical application. And adaptive Kalman filter is in the use of observed data to recursive filtering, and at the same time, parameters and the noise statistic characteristics of unknown or inexact model are properly estimated and revised in real time, which can abate the effect of error model and make filtering results more close to the reality. So, this paper systemic expounds the basic principle and the determination of initial value of Kalman filter, and then explores the use of variance compensation adaptive Kalman filter theory to establish the theory, method and model of monitoring data processing, analyses a tall building engineering, the deformation monitoring data are processed by using standard Kalman filter and variance compensation adaptive Kalman filter method respectively, and compared with the measured data, it is concluded that data processing result from variance compensation adaptive Kalman filter is better, which can meet the need of project.