Study On Strength Reduction Technique Considering Rheological Property Of Rock And Soil Medium

For the slope, rheological property of rock and soil medium is one of the important reasons for the large deformation and loss of stability. It's necessary to take the rheological property into consideration when we evaluate the stability of slope, forecast deformation, early warn landslide disaster depending on monitoring deformation data. Based on strength reduction technique, this paper carries on an investigation in the effect of rheology on slope's deformation and stability, and rheological parameters are acquired by intelligent displacement back analysis method which includes Artificial Neural Network and Genetic Algorithm. Furthermore, this paper presents a new kind of prewarning method which involves judgment of macroscopical deformation phenomena, analysis of monitoring deformation data and numerical calculation results. Accordingly a set of early warning standards belonging to different deformation stage are put forward. In the end, this method is successfully applied to slope which lies to the west of Chong Qing Zhangfei Temple. The result proves that this method is reliable and supplies early warning of landslide with a new approach. The main study works and results are listed as follows:(1)Strength reduction technique (SRT) is applicable in evaluating the stability of slope in most case, because it doesn't need to assume the location and shape of failure surface. But SRT considers only the elasto-plasitic property of rock and soil medium, neglecting the rheological property at present. Rheology is basic mechanical property of rock and soil medium. Having some relations to rheology, many slopes yield large deformation or even lose stability in fact. In this paper, based on SRT the effect of rheology on slope's deformation and stability is analyzed. The final research shows that the rheological property of rock and soil medium makes deformation of slope larger, furthermore it makes slope more dangerous.(2)When the creep constitutive model is chosen in simulation, the FLAC code can always accomplish the calculation up to the time we set, regardless whether the slope can attain static equilibrium. So divergence of numerical calculation can never be a standard of loss of stability. Since the creep constitutive model in the paper is generalized Kelvin model, the deformation of slope will finally not change after undergoing some time if the slope is stable. If the strength of rock and soil is reduced near to critical state, the slope will yield great plastic shear strain on slide zone which is greater than creep shear strain. Displacement of slope above slide surface will increase all the time under this situation, so we can log displacements of some points to judge whether the slope arrive the critical state. (3)Based on the in site measuring displacement information, this paper back analyzes the mechanical parameters using Artificial Neural Network and Genetic Algorithm. Firstly the visco-elastic-plastic computing model of slope is built, and training samples for Artificial Neural Network can be obtained through a lot of calculation. Secondly let Artificial Neural Network study and identify these samples since it has the function of study and memory. To get the global optimization results, the initial weights are optimized by Genetic Algorithm which can search a global optimization results. By now the complex nonlinear mapping relation between displacement and mechanical parameters is built by Artificial Neural Network. If the in site measuring data is input to Artificial Neural Network, we can get a group of mechanical parameters. At last parameters obtained through back analysis method is verified by posterior variance examination, it's proved that the results are reasonable and can be used in subsequent processes.(4)To overcome the shortcoming of prewarning method which only employs purely mathematical model, this paper presents a new idea of predicting deformation and failure process in landslide. The whole procedure is divided into long-term forecasting stage, short-term forecasting stage, and temporary slip forecasting stage. Accordingly this paper presents new prewarning standards on each stage. These standards include slope macroscopical phenomena of landslide observed at the scene, feature of measuring data of some key points in site(such as displacement, velocity etc), numerical results and so on. Comparing the numerical results to measuring data, we can generally judge which stage the slope goes into at present and how much is the safety factor. Finally the entire process of deformation and failure of landslide is forecasted by this method.(5)At last, the early warning method is successfully applied to K108 slope which lies on Jinzhu highway and slope which lies to the west of Chong Qing Zhangfei Temple. It's proved that the early warning method this paper presents and prewarning standards are feasible and supply early warning of landslide with a new approach.