Reliability Evaluation And Design Optimization For A Slope Reinforced With Stabilizing Piles

The stabilizing piles are commonly used countermeasures to treat large-scale landslides,of which the design scheme directly influences the safety as well as the cost of stabilizing engineering.In this thesis,the methods such as literature survey,theoretical analysis and numerical simulation were used to study the stability evaluation and design optimization of a slope reinforced with stabilizing piles.The main contents of the thesis are as follows:(1)The stability evaluation methods for a slope reinforced with stabilizing piles were reviewed.The state of the art of the stability evaluation and design optimization for a slope reinforced with stabilizing piles were summarized.(2)A numerical model that considered the pile-soil interaction was established for a slope reinforced with stabilizing piles.The shear strength reduction method that based on the large slope surface displacement criterion was employed to compute the factor of safety(Fs)of the piles-reinforced slope.The failure modes of the piles-reinforced slope were investigated.The influence of design parameters of piles on the computed Fs and the failure modes of the piles-reinforced slope were studied(3)A regression model that quantitatively depicted the relationship between the design parameters of the pile-slope system and its mechanical response was constructed using the support vector machine(SVM)model that trained with the sampling points prepared via the uniform design(UD).The pile-slope system mechanical response included the Fs of the piles-reinforced slope and the bending moment of the stabilizing piles.Based on the surrogate model of SVM,the optimal design scheme of the stabilizing piles that simultaneously satisfied required safety and minimum cost was obtained by means of solving a non-linear optimization problem(4)An efficient reliability approach using SVM and the first-order reliability method(FORM)was proposed to compute the reliability index and the corresponding probability of failure of the piles-reinforced slope.The uncertainties involved in both the soil parameters and the stabilizing pile parameters were taken into consideration in the proposed method.Two typical failure modes,namely the insufficient Fs of the reinforced slope and the insufficient flexural bearing capacity of the stabilizing piles,were investigated from perspective of probabilistic viewpoint.The influence of stabilizing pile parameters on the computed reliability indices and the probabilities of failure were studied.(5)The proposed method was applied for a typical landslide induced by inappropriate cut at the slope foot.The reassessment of this actual case using both the deterministic Fs method and the probabilistic method illustrated the feasibility of applying the proposed methods to compute the practical engineering.The conclusions of the thesis are summarized as below:(1)A design optimization method using SVM and uniform design for the slope reinforced with stabilizing piles was proposed.The uniform design solved the problem of the sampling points selection for SVM training.The SVM model that quantitatively depicted the relationship between the design parameters of pile-slope system and its mechanical response solved the problem of optimization design of complicated practical engineering which always requires numerical procedures.(2)A reliability method using SVM and FORM was proposed for evaluating the stability of the reinforced slope and the safety of the stabilizing piles.The problem that the explicit form of the limit state function is required in FORM was overcome by using the response surface model of SVM.The practicability and the effectiveness of the proposed method were illustrated via reassessment of a true landslide case.