Research On Loose Integtation Of High-rate GNSS And Strong Motion Records For Broadband Coseismic Displacements

As one of the countries with frequent earthquakes and extremely high earthquake disaster risks,China has always regarded earthquake warning and monitoring as a very important technical means for seismic disaster reduction.Accurate co-seismic displacement is an important indicator for obtaining earthquake magnitude,source depth and source location,which can significantly enhance the precision and promptness of improve the accuracy and timeliness of earthquake warning.In traditional earthquake warning and monitoring systems,broadband seismometers are prone to saturation in the near-field of large earthquakes,while strong motion instruments have the advantage of unlimited range in strong ground motion,but are prone to baseline drift due to surface rotation and tilt,which can lead to underestimation of the earthquake magnitude.In recent years,high-rate GNSS has been widely used in the monitoring of crustal movement and surface displacement.This technology is not affected by range and surface movement,and can directly obtain co-seismic displacement information.However,due to the large observation noise and low sensitivity of high-rate GNSS,it is difficult to effectively capture subtle surface displacement changes.Therefore,integrating the characteristics of co-seismic displacement monitoring with strong motions and high-rate GNSS,this study focuses on constructing a joint observation mode of high-rate GNSS displacement records and strong motion acceleration records,complementing each other’s advantages,to achieve the goal of obtaining high-precision wideband co-seismic displacement.Centered on this core goal,the research contents of this paper mainly include:(1)This paper explains a systematic exposition of the basic positioning principles and coordinate time series model of GNSS,and provides a detailed introduction to the principle of seismic monitoring with strong motions and the data processing procedure.Based on this,the basic principle of the loosely coupled system of high-rate GNSS and strong motions based on the Kalman filter algorithm are analyzed and studied.Furthermore,the impact of noise parameters in the combined system on filtering results is analyzed according to the filtering formula.(2)The traditional Kalman filter algorithm assumes that the noise is Gaussian white noise,while the noise in actual GNSS data processing often has a certain correlation.In order to address the problem of the impact of colored noise on the performance of the GNSS and strong motion loose integration system,this paper constructs a Kalman filter-based system for highrate GNSS and strong motion loose integration system under colored observation noise.Firstly,based on the white noise and scintillation noise model,the noise parameters of the high-rate GNSS time series before the earthquake are estimated using the variance component estimation method.Then,the colored observation noise is added to the Kalman filter to construct a highrate GNSS and strong motion loose integration system with colored observation noise.Based on the experimental results,it has been demonstrated that the proposed algorithm has the capability to enhance the precision of the displacement results of the integration system by 43%compared to only considering Gaussian white noise,and can effectively reduce the impact of colored noise in high-rate GNSS on the performance of the loose integration system.(3)The strong ground motion during an earthquake can cause abrupt changes in the acceleration data recorded by a strong motion,resulting in non-linear and time-varying noise variance in the GNSS and strong motion integration system,which in turn reduces the accuracy of the integrated displacement.To address these issues,this paper presented a variancecompensated adaptive Kalman filtering algorithm to construct a high-rate GNSS and strong motion integration system.This algorithm utilizes the predicted residual covariance matrix to estimate and update the system noise covariance in real-time,which leads to more accurate state vector estimation.In addition,both simulation experiments and the August 8,2017 Ms 7.0Jiuzhaigou earthquake are used to validate the superiority and robustness of this method.Based on the experimental results,it has been demonstrated that compared to using a fixed system noise variance prior,the proposed algorithm improves the accuracy of the integrated displacement by 46%.The algorithm not only eliminates the baseline bias produced by the strong-motion instrument but also effectively suppresses the impact of the time-varying system noise variance,resulting in more accurate co-seismic displacement information.