Predictive Method For Landslide Dynamic Deformation Affected Sensitively By Rainfall And Research On The Effect Of Drainage Tunnel

Landslide is a common geologic hazard. The construction would lose a lot and the safety of resident would be threatened by landslide. Therefore, the research for landslide dynamic deformation prediction and treatment technology have important realistic significance.Focusing on the character that landslide is affected sensitively by rainfall, theoretical analysis and numerical simulation were used based on the field investigation and monitoring. The influence patterns of environmental factors for the landslide deformation were analyzed. The prediction methods of landslide creep deformation, landslide influenced by groundwater and groundwater variation supplied by rainfall were analyzed. The influence of backwater amount existing in unsaturated zone in rainfall supplying groundwater was analyzed. The principle about optimal location for drainage tunnel in landslide was analyzed. Finally, the main results and new ideas in this study are as follows:(1) By the use of autoregressive model, the displacements were converted into acceleration to predict landslide deformation based on the time series monitoring data about landslide deformation. The negative effect of nonlinear displacement curve to statistical model was eliminated.(2) The autoregressive model which only treats with single-factor time series was developed into vector autoregressive model which could treat with multi-factor time series. The relations about groundwater-landslide displacement, rainfall-groundwater and rainfall-groundwater-drainage discharge were analyzed by dimensional transformation. The several data types could have correlativity. And then, the analytical-statistical model for landslide was built;(3) The dimensional transformation based on soil analytical model about groundwater variation and landslide deformation made the dimension consistent according to the acceleration. Both quantities transformed were taken into account in vector autoregressive model. The landslide displacement could be predicted by monitoring data of groundwater level.(4) The variation of groundwater level could be divided into two parts, lifting by rainfall and natural retreat. The correlation analysis and vector autoregressive model analysis were used to analyze the relation between the amount lifting by rainfall and rainfall intensity. The landslide early warning could base on the study on lag characteristic and rainfall threshold, by the use of rainfall information.(5) The backwater amount in unsaturated zone in landslide was estimated. The water amount was set as an index of dimensional transformation for rainfall, drainage discharge from tunnel and groundwater. The groundwater variation was converted into release water amount, and then the vector autoregressive model was fit to simulate the rainfall in the condition of high initial groundwater level. The model could compute the backwater amount in unsaturated zone in the condition of low initial groundwater level. The rainfall threshold about rainfall supplying groundwater was based on the factor of the backwater influence.(6) The using conditions of drainage tunnel in landslide prevention were analyzed. The applied effect were discussed. And then the principle about the optimum location of tunnel was got. The roles of drainage tunnel in drainage in the condition of constant head and rainfall were analyzed.