Preliminary Study On Prediction Model And Early Warning Criterion Of Landslide Based On GNSS Monitoring

Landslide monitoring is one of the important means of landslide disaster prevention,prediction and early warning.Through monitoring,we can understand the deformation development process of landslide,which provides an important basis for the development trend prediction and occurrence time early warning of landslide.At present,the GNSS(Global Navigation Satellite System)technology with the characteristics of automatic real-time monitoring is gradually applied in the field of landslide monitoring.The monitoring of landslide is in the transition stage from non automation to automation.The prediction and early warning of landslide by using GNSS monitoring technology has become a new direction of landslide disaster prevention and control,so it is necessary to carry out the research of GNSS monitoring technology in landslide prediction and early warning.Based on the above reasons,on the basis of comparing GNSS technology with traditional monitoring technology,combined with the advantages of GNSS monitoring,this paper proposes that the landslide prediction and early warning of GNSS monitoring should be divided into two parts: medium and long-term displacement prediction and critical sliding time early warning according to the deformation stage of landslide.Through consulting relevant literature and combining with landslide engineering cases,the application of GNSS technology in landslide prediction and early warning is opened The preliminary research has been carried out and the following research results have been achieved:(1)Combined with the three-stage theory of landslide deformation stage and the measured S-T curve of landslide engineering case,the S-T curve is divided into four types from the shape of the S-T curve as the starting point,and the corresponding landslide deformation stage of each type is determined;the S-T curve with different characteristics is predicted by using the commonly used grey and BP neural network models in the landslide prediction model,and the S-T curve is obtained by different models According to the prediction effect of various models and the deformation stage corresponding to the S-T curve,the selection of medium and long-term prediction and critical sliding early warning models is determined.(2)For the medium and long term prediction of landslide,the S-T curve characteristics of different deformation areas of the landslide are determined by using GNSS monitoring data,and the deformation and damage stages and risk degrees of different deformation areas are obtained;the failure mode of the landslide is determined by combining rainfall and groundwater with deep displacement data and multi-stage profile changes;the failure mode of the landslide is determined by combining the data of deep displacement and multi-stage profile changes In this paper,the grey GM(1,1)-BP neural network displacement prediction model of trend term cycle term is established,and the deformation trend of landslide is discussed;the medium and longterm early-warning index based on monthly rainfall is determined by using the BP neural network model of cycle term.(3)Taking Zhouzhi G108 section as an example,according to the characteristics of S-T curve of monitoring point and the criteria of deformation rate,acceleration and tangent angle,the deformation stage of G108 section is divided,and the determination index of deformation stage based on displacement rate ratio is proposed;the instantaneous deformation rate determination index is proposed and combined with the characteristics of GNSS real-time monitoring Combined with the instantaneous tangent angle,the dynamic quantitative discrimination of the deformation after entering the critical sliding stage is carried out,and finally the landslide is successfully forewarned.