Research On Monitoring And Early Warning Model Of Railway Landslide Disaster Based On Big Data Analysis And Mining

China,as a country with frequent landslides.The area along Yunnan-GuizhouSichuan Railway,which is the focus of this research,is prone to landslide.Once landslide occurs,it will cause incalculable huge losses.In order to ensure the safety of railway during construction and operation,It is necessary to study the early-warning model of landslide along railway.Due to the complex causes of landslide disasters,people’s early warning technology is far from mature.Most of the current research on landslide monitoring and early-warning is based on the analysis and early warning of sensor monitoring data,and its research results have some limitations.In order to solve this problem,this thesis will carry out the research work on railway landslide disaster monitoring and early warning model based on big data analysis and mining.It provides a new solution for landslide disaster monitoring and early warning.The main work of this thesis is as follows:(1)Write python program to collect the data of landslide disaster sites and their influencing factors nationwide in recent ten years,and use big data analysis technology to research the causes of landslide disasters.(2)Proposed an improved and optimized logistic regression method to mine the data pattern of landslide disaster,and use this data pattern to build a landslide risk prediction model.The experimental comparison proves that the prediction model built by the improved logistic regression method is higher than the traditional logistic regression method in detection rate,false negative rate and accuracy rate.(3)Study the application of BP neural network deep learning method in the field of slope surface deformation prediction,and improve it based on it.This thesis proposes a time series prediction method of ground deformation based on moving average method(MV-BP).The experimental results show that the accuracy of the improved algorithm in predicting the time series of surface deformation is higher than other algorithms,and it can predict the surface deformation of the monitored slope more accurately,and then predict the occurrence time of landslide disaster.(4)Proposed a spatial region division method based on grid element,which can divide different slopes in the monitoring area.By using this method,not only can the precision of landslide hazard risk prediction be improved,but also the spatial geographical range of landslide hazard can be predicted.(5)Propose a landslide risk monitoring and early warning model that combines the probability of landslide occurrence,damage area and socio-economic loss according to the standards of the disaster monitoring and early warning industry and the actual needs of monitoring and early warning along the railway.The experimental results show that the early warning model proposed in this thesis is more reasonable and effective in practical application.(6)Develop the geological disaster early warning platform along the railway and present the real-time monitoring and early warning results to users in a visual way.Let users know the results of real-time monitoring and early warning more intuitively and clearly.