Intelligent Algorithm For Evaluating Rock/Soil Slope Stability And Reliability

Slope stability is one of most important theoretical and practical subjects in geomechanics and geotechnical engineering, to which classical soil mechanics has not yet provided satisfactory solutions in spite of its first and long endeavor. First of all, a new intelligent optimization algorithm, i.e., the adaptive genetic-tabu search algorithm (AGA-TS algorithm), is proposed in this dissertation by combining the adaptive genetic algorithm(AGA) with tabu search algorithm(TS algorithm). Then, the applications of AGA-TS algorithm in soil slope and jointed rock slope stability are elaborated through definite value and reliability analysis for safety factor. At last, some reliability problems of three dimensional soil slope are discussed. The main achievements of this dissertation are as follows:(1) The disadvantages of simple genetic algorithm (SGA) are firstly analyzed. The basic theory and computation steps of AGA and TS algorithm are following elaborated. Then, Based on two different kinds of optimal strategy, the adaptive genetic–tabu mutation search (AGA-TSM) algorithms and adaptive genetic-tabu random search (AGA-TRS) algorithms are respectively developed through the combination of AGA and TS (both of them belong to the adaptive genetic–tabu search (AGA-TS) algorithms). Simulating results show that the searching efficiency of AGA-TSM and AGA-TRS are both better than that of AGA and TS algorithm.(2) An improved method which is used to acquire the Hoek-Brown empirical parameters, i.e., the IBDRMP (Improved Blasting Damaged Rock Mechanical Parameters) method is taken forward. At the same time, using the IBDRMP-based Hoek-Brown criterion, improved stability analysis methods of jointed rock slope are stated in this dissertation. Then, AGA-TS(including AGA-TSM and AGA-TRS) algorithms are used to seek the critical slip surface and the corresponding minimal safety factor of jointed rock and soil slopes. Numerous example studies show that the critical slip surface yielded by the AGA-TS algorithm is more dangerous than those given by other conventional approaches.(3) By adopting limit equilibrium analysis methods(such as simple Bishop method, Morgenstern-Price method and their improved formulas for soil and jointed rock slopes ) in the proposed slope reliability algorithm instead of the finite element method numerical simulation in conventional response surface techniques, an approximate calculation method of slope reliability is established. Taking the minimum reliability index as the target function and on basis of AGA-TS algorithms, a global optimization method, which can locate the critical probabilistic slip surface associated with the minimum reliability index, is presented in this dissertation. At the same time, the compared computation results indicate that the critical slip surfaces sought respectively according to the minimum reliability indexβand the minimum safety factor FS are considerably different.(4) Through the development of two dimension Morgenstern-Price method, an improved three dimensional safety factor method, i.e., three dimensional Morgenstern- Price method, is proposed in this dissertation. By adopting the improved method in the proposed slope reliability algorithm instead of the finite element method numerical simulation in conventional response surface techniques, a new kind of response surface calculation method of three dimension slope reliability is established. At last, it is also discussed that the width of slip body, the probability distribution and correlativity of soil parameters are how to influence on the reliability index.