Study On The Fluid-solid Coupling Characteristic Of Soil And Rock Blending Landslide And Its Prediction And Forecast

Associated with the scientific research projects of ministry ofcommunications "The study of monitoring-prediction technology in westhighway geology disaster (2003-318-802-01)", the fluid-solid couplingcharacteristic of soil and rock blending landslide and its prediction andforecast were studied comprehensively, with the combined researchmethod of laboratory tests, field tests, theory analysis, numericalsimulation. The main contents include followings:1. The permeability coefficient of soil and rock blending at differentgravel content were measured by artificial constant head permeameter,which show there exists index relationship between them. A empiricalformula for calculating the composite permeability coefficient of soil androck blending is put forward based on power averaging method and theformula is proved by the experimental results, have offered a concise anduseful computational tool for the theory calculation of coarse-grained soilpermeability coefficient.2. Soil and rock blending is a typical porous medium, and itsseepage is closely related to the particle-size, porosity ratio and particleshape. With self-made constant head permeameter, the influence of gravelcontent, porosity ratio and particle shape on the permeability coefficientof soil and rock blending was studied by lab orthogonal test. Orthogonaltest confirmed the influence orders of three factors to permeabilitycoefficient of soil and rock blending and the significance levels ofdifferent factors.3. Through artificial rainfall simulation tests and field syntheticmonitored on a soil and rock blending slope, some results were obtained:the deformed zone of soil and rock blending landslides caused by rainfallinfiltration lies within the top 4m soil layer, where the deformation valueof slope surface is the biggest, which gradually reducing from dome to the deep part of slope. The average percentage of infiltration during thefirst 2 hours is 86%, it reduce gradually with time later because of theincrease of the surface runoff. The average percentage of infiltration dropto a relatively stable value (50%) 6 hours later. Rainfall infiltration leadspore-water pressure to increase, which may result in a reduction of shearstrength due to a decrease in effective stress and wetting-inducedsoftening, the double effects of rainfall infiltration are the main reason ofrainfall infiltration induced landslides in soil and rock blending slope.4. By the situ-monitoring of excavation tests on a well-instrumentedsoil and rock blending slope, these results show: Most of soil and rockblending landslides caused by excavation are shallow draught slopefailures, and deformed zone lies within the top 4m soil layer, anddeformation value of slope surface is the biggest, which reduce fromdome to the deep part of slope gradually; The alert height of the free faceis 3m in such soil and rock blending slope, when the free face exceed 3mheight proper defend measures are necessary; Under the influence ofstrong rainfall, the uncovered soil and rock blending slope afterexcavation easily lead to collapse accident.5. In consideration of the deformation of the soil and rock blendingand the compressibility of the fluid, the dynamic models of porosity andpermeability are acquired; Quantitative analysis the effect of stress fieldon the porosity and permeability and the effect of seepage force on thestress field, fluid-solid coupling mathematical model in soil and rockblending are acquired; By Galerkin theory, differential coefficientequations are dispersed on space and time domains, then the couplingfinite element equations of the seepage field and stress field areestablished.6. Under the fluctuation of different water levels caused by rainfallinfiltration, the seepage-stress fields, the deformation trends and thefailure process of soil and rock blending landslide are analyzed withFLAC(Version 5.0). In addition, the change of landslide stability is discussed after water levels fluctuation in this dissertation.7. On the basis of Verhulst inverse-function mechanism predictionand forecast model, and in combination with Back Propagation NeuralNetwork, a new error fitting of Verhulst inverse-function mechanismprediction and forecast model is built. The establishment of the predictionand forecast model and the process of application are discussed in detail.In engineering practice, the application research is compared with actuallandslides case nationally and internationally. The results are similar toactual data and proved the correction of model, which provides a newway for landslide prediction and forecast.