Research On The Landslide Forecasting And Early Warning System Based On Xiakou Landslide

Due to the dangerous and complexity of landslides, the landslide forecasting has been an international concern for a long time. Taking the Xiakou landslide in Ya’an in Sichuan province for an example, this thesis analyzed its long-term monitoring data, and mainly studied the methods of monitoring data pre-processing, landslide forecasting and early warning. Then a landslide forecasting and early warning platform was developed. In this respect, the main results are as follows:(1) The deformation trend analysis and stability evaluation of Xiakou landslide. The geological environment conditions of Xiakou landslide was analyzed comprehensively. Based on the analysis of incline displacement, groundwater level and rainfall monitoring data from 2006 to 2010, the deformation characteristics and trend of Xiakou landslide before and after Wenchuan earthquake in 2008 were studied. The limit equilibrium method was used to calculate the stability of the landslide, it was showed that the safety factor was significantly affected by the groundwater levels in the forepart and central part of the new landslide, and its change represented the critical state of the landslide for a certain period of time. Based on the FLAC3D software, the numerical simulation analysis was carried out separately to study the landslide state with low groundwater level and high groundwater level, and then compared with the actual monitoring displacement.(2) The study on the model test of Xiakou landslide. The scale model of Xiakou landslide was built based on the similarity theory, and different storm conditions were adopted to study of the deformation and failure process of the landslide model under rainfall conditions. And they were compared with the actual deformation characteristics of Xiakou landslide, which both verified the sliding-stop-sliding intermittent deformation mode of the Xiakou landslide. Though analyzing the characteristics of the displacement, earth pressure and pore water pressure with changes in rainfall process, it can be concluded that the methods of landslide forecasting and early warning based on deformation monitoring is more reliable.(3) The design and development of a landslide forecasting and early warning platform. Based on a comprehensive study of the methods and models of landslide monitoring data preprocessing, long-term forecasting, medium-term forecasting and short-term forecasting, a landslide forecasting and early warning platform was developed by using the object-oriented software Delphi. The developed platform is feature-rich:it consists of data preprocessing module, long-term forecasting module, medium-term forecasting module, short-term forecasting module and automatic alarm module. And it provides landslide analysis, forecasting and early warning with different time scales.(4) The test and verification of the developed landslide forecasting and early warning platform. The reliability and calculation accuracy of the long-term forecasting module was verified through an international standard slope example. Based on the monitoring displacement in the model test, the automatic alarm module was used to capture the signs before the landslide deformation accelerated, and the short-term module was applied to predict the failure time of the landslide. After this test, it was proved that both the automatic alarm module and the short-term module have strong applicability and reliability.