Safety Analysis And Design Of Stabilizing Pile Reinforced Slope Based On Reliability Theory

Landslides are widely distributed,complex,destructive and difficult to predict,which causing huge economic and life losses every year worldwide.As an important reinforcement method in geotechnical engineering,stabilizing piles are widely used to reinforce slopes.Recently,the analysis of pile-reinforced slope is still based on the deterministic analysis method which generalizes the unknown and uncertain factors into the factor of safety to assess the stability state of slopes.However,there are following limitations of the deterministic method.Firstly,it cannot effectively quantify the engineering uncertainty and risk;Secondly,the assessment results include the subjective experience judgment of engineers;In addition,this method ignores the parameter uncertainty and inherent spatial variability of geomaterials.Therefore,it is necessary to establish the stability analysis theory and optimal design method of pile-reinforced slope considering the uncertainty of geotechnical parameters,so as to accurately evaluate the slope safety and finally achieve the technically feasible and economically efficient design of stabilizing piles.In this thesis,the reliability analysis theory of pile-reinforced slope based on probabilistic analysis is developed,and the influence of uncertainty of geotechnical parameters on the reliability,quantitative risk and failure mode of pile-reinforced slope is studied.The main research contents and conclusions are as follows:(1)A new method for stability and reliability analysis of pilereinforced slopes considering the rotational anisotropy of soil parameters is proposed.The results show that the spatially variable soils with rotational anisotropy have significant effects on the statistical characteristics of factor of safety,failure modes and probability of failure of pile-reinforced slopes.Dip slopes tend to have lower stability and reliability with more potential failure modes,while the anti-dip slopes show the opposite results.Therefore,the difference in failure probability based on different strata rotation angles could be significant.For soil slopes with rotational anisotropy,reliability analyses based on the commonly used assumption of horizontal transverse anisotropy may underestimate or overestimate the failure probability of pile-reinforced slopes,then leading to completely different pile design results.(2)A probabilistic analysis method of pile-reinforced slopes based on investigation borehole data was developed and used to systematically investigate the effects of survey data on the statistical parameters of safety factor,failure probability,quantitative risk and structural response variability of anti-slip piles including for different borehole survey layout.The results show that the statistical values of the results based on the conventional unconditional random field analysis may have larger uncertainties than those obtained from the conditional random field analysis using real site information.In general,increasing the number of boreholes and sampling points can significantly reduce the uncertainty of geotechnical parameters,thus obtaining more accurate quantitative risk assessment and structural response analysis of the piles.The number,location and sampling distance of boreholes in the survey and design plan can significantly affect the results of probabilistic analysis.When the number of boreholes is limited,it is recommended to arrange the boreholes uniformly along the slope surface,which can improve the accuracy of the factor of safety and its credibility,and thus provide more accurate assessment results of the failure probability and failure risk.(3)The influence of the spatial variability of soil parameters on the response of piles and the failure mode of pile-reinforced slopes is systematically investigated.Firstly,based on the random field theory,the strain wedge model,which used to be only applicable to homogeneous soils in traditional methods,is extended to spatially variable soils.Then,a new method of slope stability analysis based on the improved strain wedge model is proposed;The proposed method can conveniently obtain the pilesoil pressure above the slip surface and the subgrade reaction below the slip surface to evaluate the stability of the reinforced slope.Finally,the proposed improved strain wedge model and Monte-Carlo simulation method are used to investigate the influence of soil spatial variability on the structural response of stabilizing pile and the failure mode of the pilereinforced slope.The results show that the soil spatial variability shows significant influence on the response of the pile and the slope failure mechanism,and reveal the reasons for the discrepancy between the engineers’ understanding of the potential failure mode of the anti-slip pilereinforced slope and the actual results based on the traditional qualitative analysis method.(4)A new method of adaptive and efficient Kriging metamodel for slope reliability analysis is proposed.The proposed method uses WhittleMatérn non-classical two-parameter autocorrelation function and determines the optimal autocorrelation structure of the data via global optimization of correlation coefficient and smoothness coefficient.The results show that the computational efficiency of the proposed method is significantly improved compared with the traditional direct Monte-Carlo simulation method.Meanwhile,the proposed method shows higher computational accuracy in fitting the response surface and calculating the failure probability of slope than the traditional Kriging model based on the single-parameter autocorrelation function.In summary,the proposed method overcomes the limitation of subjective selection of autocorrelation function in the traditional Kriging method,and improves the accuracy of the Kriging model.Finally,a multi-objective optimal design method for stabilizing piles is proposed by integrating the relationships among slope failure probability,construction cost and variation coefficient of the factor of safety,and the effectiveness of the method is verified by designing an optimal pile reinforcement scheme for a landslide in Jiang Yong County,Hunan Province as an example.