Data Fusion And Stability Analysis Of Landslide Monitoring Based On Multi-source Asynchronous Sensors

Landslide is a kind of geological disaster with wide distribution and frequent occurrence.It has the characteristics of outburst,seasonal and creep.In the comprehensive information monitoring of landslide,there are many kinds of sensors and various positions.The types of sensors can be divided into: landslide displacement monitoring sensors,environmental information monitoring sensors,it is necessary to optimize the layout of various monitoring sensors to maximize the role of the monitoring sensors.In the massive monitoring data,the monitoring information of sensors with different frequency,position,depth and precision is fused to reflect the comprehensive deformation characteristics of landslide.Through the displacement time series analysis of landslide monitoring,the stability of the landslide is evaluated to judge whether the landslide is safe or not.The paper relies on the actual engineering project and conducts research according to the main line of “high-precision model construction – sensor optimization design – monitoring information fusion – stability analysis”.The main research contents and innovations are as follows:(1)Construction of high-precision model of landslide in accumulation layer and optimal design of sensorsIn this paper,a high-precision 3d geological model of landslide is constructed by using high-precision 3d laser scanner and geological exploration data,and corresponding mechanical parameters are given to each stratum.In view of the shortcomings of traditional landslide deformation monitoring,which usually uses empirical method to select monitoring points,lines and regions,this paper adopts numerical simulation to calculate the deformation under different rainfall conditions,and determines the key monitoring area of landslide body Optimization Design of landslide monitoring sensor placement,landslide monitoring point,Monitoring Line sensor optimization design to achieve theoretical innovation.Through the simulation of monitoring point and the analysis of Factor of safety,the early warning threshold of landslide is determined,which provides the criterion condition for the monitoring and early warning of landslide body.(2)Multi-sensor data fusion for landslide monitoringIn order to overcome the shortcoming that the Kalman filter state estimation needs the known noise information accurately,the attenuation Memory factor is introduced,and the Adaptive Fading Memory Kalman filter(AFMKF)is used to reduce the noise effect.Aiming at the uncertainty of attenuation Memory factor,this paper presents an Adaptive attenuation Memory Kalman filter algorithm(AFMKF with HPOC,Adaptive Fading Memory Kalman Filtering with High-Precision Observation Constraints),Periodic high-precision observations are used to update the innovation sequence,and the updating of the attenuation memory factor is constrained to improve the filtering ability of the Algorithm.Simulation results show that the root mean square error of the conditional AFMKF algorithm is 41.31% lower than that of the traditional Kalman filter algorithm.The data fusion algorithm based on AFMKF with HPOC is designed according to the CKF(Centralized Kalman Filter Filter)data fusion algorithm,and the monitoring data of landslide is fused to reflect the deformation characteristics of landslide as a whole.(3)Stability analysis of multi-scale landslideIn view of the fact that the traditional displacement time series analysis method cannot consider the relationship between the micro-scale displacement variation and the stability,the fractal theory and rescaled Range Analysis(R/S analysis)are studied in this paper to judge the stage stability and the whole stability of landslide.The results show that the multi-fractal dimension spectrum of the displacement time series is consistent with the spectrum of the acceleration deformation process,and the stability of the landslide is poor at this stage the analysis results of long-scale displacement time series show that the stability situation of the landslide will continue and have long-range persistence.