Research On Extraction Model Of GNSS And Accelerometer Fusion Structural Vibration Information

Dynamic deformation monitoring is one of the contents of vibration monitoring of large civil structures and a necessary guarantee for safe operation of large projects.As the main technology of dynamic deformation monitoring,Global Navigation Satellite System global navigation satellite system(GNSS)plays an important role in the field of structural health monitoring with its unique advantages.This paper presents the application of GNSS and accelerometer in vibration monitoring of large civil engineering structures.Key techniques such the multipath error correction model,an RTS smoothing model based on multi-rate Kalman filtering,and the Hilbert Huang transformation model for CEEMDAN decomposition.The main contents are as follows:(1)Review and summarize the principle of GNSS measurement and positioning,analyze the main error sources of GNSS in the field of structural health monitoring,and improve the accuracy of monitoring data by means of difference method and hardware improvement.The weakening method of multipath effect is studied.(2)Aiming at the low-frequency error problem in GNSS dynamic deformation monitoring data,empirical mode decomposition and standardized modulus cumulative mean are introduced.By decomposing the monitoring data and using the cumulative mean value of the standardized modulus to extract the slowly changing trend items in the GNSS monitoring data,the purpose of weakening the multipath error is achieved.(3)Aiming at the high-frequency error problem in the GNSS dynamic deformation monitoring data,a multi-rate Kalman filter method is proposed,which increases the sampling rate of the GNSS monitoring data and removes the GNSS monitoring data by fusing the GNSS and acceleration sensor data and RTS smoothing.the presence of high frequency noise.(4)Aiming at the problem of frequency feature extraction of large structures,a method combining CEEMDAN and Hilbert transform is proposed.Compared with the traditional fast Fourier transform method,the proposed method can extract and display vibration events in multiple dimensions of time-frequency-energy,thereby improving the accuracy of monitoring data and increasing data integrity.(5)Taking Changchun Hairong Plaza Building as an example,the dynamic monitoring data is processed and analyzed through multi-rate Kalman filtering.And for the first time,the energy difference formula is applied to super high-rise buildings to determine the number of components of variational mode decomposition,and then the frequency spectrum of the components is analyzed to extract the main mode vibration frequency of super high-rise buildings.