Research On Analysis Method Of Slope Stability Based On Inverse Reliability Theory

Slope engineering is one of the classic research field of geotechnical engineering. Its object is the rock soil mass that experienced a long period of geologic history, whose occurrence environment, composition, physical and mechanical characteristics have different degrees of uncertain factors. In the analysis and design of slope engineering, the safety coefficient and the reliability index are often used to depict uncertainty degree of these factors. But the safety coefficient mainly reveal relative quantitative relationship between average resistance and average load, the reliability indexes uncover probability of the failure risk that needs to bear passively under given conditions. For projects that exist uncertainty factors, it is impossible to completely eliminate the failure risk, in order to control the risk within acceptable range, the emphasis is keeping a clear mind on the influence degree of related factors to the risk and making a specified direction for the sound engineering treatment measures based on appropriate design and analysis. However, the current emphasis on correlation analysis basically emphatically reflected one of the aspects, not to complete all functions at the same time. Therefore, this article established a new method of slope stability analysis based on active control theory of risk.Firstly, according to two kinds of slope failure modes:sliding along the regular sliding surface and arbitrary sliding surface, Bishop pattern and Janbu pattern are chosen as representatives. Based on the concept of risk control, once inverse reliability theory was used to deduce the slope stability of inverse reliability algorithm and to derive related formulas of the two types of sliding mode. Then, the slope inverse reliability analysis framework that contain double control indexes of reliability and safety factors was formed, which can complete slope stability calculation in simple cases. Based on an acceptable risk of failure in advance, this method made a one-to-one correspondence between the reliability index and stability coefficients and achieve active control of slope failure risk.Subsequently, uncertain description of geotechnical parameter and function expression of the state of slope stability determine the practicality of inverse reliability design analysis methods and the validity of the calculation results. Uncertain characteristics of the numerous parameters that affecting the slope stability often embodied in diverse phenomenon such as randomicity, interval, gray, etc, and also showed the non-uniformity of affecting degree of various parameters on slope stability. Therefore, based on the establishment of a full space, multi-level and robustness of uniform design for the extraction of test sample, the gray relational theory was embedded into the interval analysis model for solving interval correlation between various factors and response. And by combining the theory of comparative advantage to sort the interval correlation and to build up the slope stability factors significantly integrated assessment method. This method not only provides a screening tool when establish functions for mechanical state of the slope stability, but also prepares appropriate means for further improving the practicability of risk control analysis methods.Combining with the characteristics of the actual project whose slope parameters have different distribution forms and coupling relations, in-depth research and discussion have been launched from the construction conditions of response surface function model itself. First, by introducing the Nataf transformation to make the slope parameters independently normalized. Then, using independent normalized parameters to fit the response surface function model, so that the constructed response surface model can be applied to the actual slope engineering whose parameter distributions have diversity and relativity. Finally, the aforementioned fitting response surface models and once reverse reliability theory are combined to construct a more reasonable and effective active control of slope failure risk analysis method.Thirdly, in order to further enrich and perfect the research content of active control of slope failure risk analysis method and to improve the applicability of response surface modeling tools in processing slope engineering implicit function functions, a comprehensive consideration for function expression forms of response surface model and sampling way have been taken into based on the complexity and highly nonlinear characteristics of slope limit state function. In selecting the response surface model, to conquer limitations of the application scope for polynomial model, the Kriging surrogate model which can handle complex nonlinear problems more effectively was adopted. For sampling method of test sample points, the more efficient and adaptive Latin Hypercube experimental design methods and the active learning method were combined to search for the best training samples. Integration of the above response surface model and sampling way will establish a more perfect slope stability inverse reliability response surface technology and broaden the application scope of failure risk analysis method for implicit function of complex slope. Eventually, using the significant identification methods of slope stability influence factors and the active control analysis methods of instability risk constructed in this article to guide the large numbers of slope stability analysis and design for highway side slope of TongPing Highway that connect northern Hunan and eastern Hubei. In the sixth chapter, based on the above method, two sections of the slope stability analysis and design process and main conclusions were shown. Work products provided a practical and effective treatment for slopes along TongPing Highway, at the same time, they verified that the active control of slope instability risk analysis methods established in this paper were rational and applicable.