Study On The Prediction And Early Warning Technology Of Xinpu Landslide In Three Gorges Under The Action Of Reservoir Water Level And Rainfall

Landslides are the most frequent geological disasters in my country.Since the construction of the Three Gorges Water Conservancy Project has changed the original geological environment balance of the reservoir area,the Three Gorges reservoir area has become one of the areas with the most serious landslide disasters in my country.Carrying out prediction and early warning research is an important means of disaster prevention and mitigation.This paper takes the Xinpu landslide in the Three Gorges Reservoir area as the research object.In previous studies,there was a lack of consideration of the impact of rainfall on the landslide over time,and there were few quantitative studies on the time-lag parameters between rainfall and reservoir water level.In addition,the existing early warning models lack specific The early warning criteria and early warning standards of step landslides do not take into account the characteristics of landslides,so it is difficult to carry out effective prediction and early warning work for step landslides on the reservoir bank.Based on the above questions,the main research contents and conclusions of this paper are as follows:(1)This paper first aims at the problem that the impact of rainfall on landslides decays with time,and introduces the pre-precipitation index(API)with an attenuation factor as a rainfall feature.Through comparative analysis with the traditional cumulative rainfall,it is found that API can weaken the cumulative characteristics brought about by the accumulation while considering the previous rainfall,which is more in line with the physical process of the impact of rainfall on landslides.(2)Considering that there is a time-lag effect in the response of landslide deformation to reservoir water level and rainfall,the lag correlation between the displacement rate of each monitoring point of Xinpu landslide and rainfall and reservoir water level was quantitatively analyzed using Pearson time-delay correlation analysis method.The results show that: with the increase of rainfall days,the correlation between cumulative rainfall and landslide displacement rate increases within a certain number of days,and the time-lag parameter gradually decreases to a negative value;API is significantly correlated with landslide rate,but because API is not completely eliminated Accumulative characteristics,the time lag parameter is between-3～0.(3)A PSO-BP neural network landslide displacement prediction model considering time lag and rainfall attenuation effects was constructed,and based on the research results of landslide deformation response law,a comparative experimental analysis was carried out on the use of different rainfall characteristics and whether to consider the cause hysteresis.The experimental results show that: compared with the cumulative rainfall,API prediction results have higher early warning accuracy;taking into account the time lag parameters can significantly improve the prediction effect of the prediction model,compared with the prediction model without taking the time lag effect into account,using 1 day,The accumulative rainfall of 5 days,10 days,30 days,and 60 days and the prediction accuracy of API as a rainfall feature increased by 25%,28%,41%,29%,-12%,and 44%,respectively.With a large negative impact,the prediction accuracy of the 60-day cumulative rainfall is reduced.(4)Finally,for the early warning of Xinpu step-type landslides,based on the consensus that the reservoir water level and rainfall are the main causes of the periodic deformation of the landslide,the reservoir water level and rainfall are used to weaken the periodicity of the landslide displacement rate and make the landslide The deformation conforms to the three-stage creep characteristics again,and the trend displacement rate ratio and the trend displacement tangent angle are calculated.At the same time,considering the relationship between the deformation rate of the landslide in the uniform velocity stage and the tangent angle of the imminent sliding,and comprehensively considering the improvement of the tangent angle and the displacement rate ratio early warning model standard,the Xinpu landslide early warning criterion and early warning model based on the displacement trend rate ratio were constructed.This study can provide a reference for other wading landslide early warning work.