Research Of Genetic Mechanism And Prediction Of Typical Debris Landslide In The Three Gorges Reservoir

Three Gorges area is one of the high incidence areas of geological disasters in China. With the operation of the Three Gorges Reservoir, influenced by the changing of water level and rainfall, many of the old landslides are newly deformed, and also and also some new landslide disasters are produced. Debris landslides widely spread in the Three Gorges. Taking The third phase project of geological disaster prevention emergency rescue and emergency implementation for Three Gorges Reservoir as an example, debris landslides accounted for80%of the geological disasters. Due to the special material composition and structural characteristics, the deformation of debris landslides is more easily affected by the changing water level and rainfall. The deformation and stability of shore debris landslides will directly control the sustainable development of economy, and also affect life and property security of people in the Three Gorges Reservoir. In result, this is one of the most important environmental and geological problems in the Three Gorges Reservoir. So the genetic, failure mechanism as well as the prediction research of debris landslides are extremely important and far-reaching significant to the geological disaster prevention and mitigation of people’s lives and property protection.In this thesis,43typical debris landslides in Zigui-Wanzhou section in Three Gorges Reservoir are selected as the study sample. Based on the statistic analysis of development law and deformation and failure characteristic, the dynamic origin, deformation and failure mode of the debris landslides are classified and summarized. Taking Baishuihe Landslide in Zigui County as research object, based on the study of Regional geological background and engineering geological conditions, the formation time of Baishuihe landslide, causes and developing process of the debris material are proposed. Combined with the landslide geological background and environmental factors, the genetic mechanism is analyzed from the internal factors and external factors. On the basis of the detailed investigation of the macro-geological surveys and landslide deformation monitoring data analysis, the deformation characteristics of Baishuihe landslide are researched. Sensitive factors of Baishuihe landslide are quantitatively evaluated by the method of grey correlation analysis.According to the above all geological analysis of Baishuihe landslide, simplified geological model is established. Combined with the situation of the Three Gorges Reservoir water level and the climatic characteristics of Zigui County, four typical conditions are selected for the Numerical Simulation. The author simulate the seepage field, the displacement field, strain field and plastic zones of the landslide on the combined effects of the reservoir water level and rainfall changes, using the SEEP/W modules of software GEO-Studio, finite element software ANSYS and finite difference software FLAC3D. The accuracy of geological analysis result is verified by analyzing the displacement equivalent cloud, shear strain increment equivalent cloud as well as the plastic zone maps. Finally, based on the landslide developmental stages, which is determined by analyzing the feature of fracture development and the displacement vector angle, author makes displacement prediction of Baishuihe landslide by the method of time series and neural network. According to the research content in this paper, some preliminary results are obtained as follows:(1)According to the statistic analysis of43study sample, we draw the following conclusions:The majority of debris landslides are large-scale landslide, The primary slope shapes are "convex type "and "up-convex down-concave type". In the formation of T2b and J2s debris landslides develop greatly. The elevation of the toe of debris landslides is always in the range of fluctuation of reservoir level. In the thick to middle-thick debris landslides, there is always more than one sliding belts exist. The mainly two deformation failure modes are push type and traction type. According to the difference of deformation stages and deformation position, the failure modes may be changed.(2)Formation time:Based on the comparing of the elevation of river terraces and the elevation of the shear opening of Baishuihe landslide, we conclude that the formation time of the whole landslide is later than the first river terrace, i.e. less than6500years from now. Cause of debris material:The clay with gravel soil and mass rock-soil is formed by slope wash and eluvium, and colluvial sourced the partial existence mass rock-soil with mud. The formation of the debris material is divided into many times. Sequence of the debris material is different with position varies. Genetic mechanism:the terrain feature, slope shape and the consequent structure provide advantage conditions for the formation of landslide. The block crack rockmass is one of the most important reason for the development of the two sliding belts. Reservoir level and rainfall factors affected Baishuihe landslide’deformation in two main ways:reducing the slope material mechanical strength and changing slope structure.(3) Based on the geological analysis and numerical simulation results, author analyzes the deformation features under the effect of the water level changes and rainfall. The deformation features under the action of reservoir level mainly manifested as:the reservoir level’s first time substantially change will caused strong deformation; in the low reservoir level condition the front of the landslide is deformed obviously. High reservoir level is in favor of the front part stability, but it’s not conducive to the middle part stability. The deformation features under the action of rainfall mainly manifested as:The infiltration of rainfall caused changes of groundwater seepage field in landslide, thus softening the rock and soil, reducing the shear strength of the rock mass, thereby reducing the stability of the landslide, finally caused the deformation of the landslide. Especially during the drawdown of reservoir water level, the heavy rain will be more unfavorable to the slope stability. Through the grey correlation analysis of sensitive factors, we can get the sensitive order of the influencing factors as follow:internal frictional angleφ)>rainfall(r)>cohesion (c)> water level (h).The warning region’s deformation is different, because of the differences of topographic. The eastern part of the front landslide is more sensitive to the influencing factors; which is the high-dangerous area of the warning region. The deformation mode of Baishuihe landslide is "Slip-Crack", and the failure mode is traction type. The deformation position at the present stage is located in the debris layer.(4) According to the feature of fracture development and the displacement vector angle, we determined that Baishuihe landslide have not entered the accelerated deformation stage. Author dissociate the displacement using the time series model, and predict the periodic term displacement through the BP neural network which selecting the appropriate factors. The superposition of the trend and periodic terms displacement is the prediction value of the overall landslide. These methods comprehensively consider the predisposing factors and dynamic changes of displacement. Based on the time series model and taking the water level changes and rainfall, which caused the periodic displacement in the landslide deformation, into consideration, the BP neural network can predict the relevant periodic deformation precisely. This model combines these advantages, which have better prediction accuracy.