The Research On Methods Of Reliability Index And Partial Coefficient Calculation Of Retaining Structure

Reliability index and partial coefficient are two important indicators for limit state design. Because of their calculation methods varied, the accuracy of the results vary. At present, people normally use Monte Carlo (Monte Carlo, here after refers to M-C) method to calculate reliability index and use general separation method to calculate partial coefficient in reliability study. While both methods are relatively accurate, there are still some errors in practical use. In response to this situation, this paper makes a systematic exposition and a comparative analysis for these methods. It also makes an error analysis for M-C method and general separation method. The main contents include:(1)The paper systematically expounded the basic principles and computational processes of the methods of reliability index calculation, which includes first order second moment (FOSM) mean value method, design point method, JC method, distribution value method, M-C method, and the methods of partial coefficient calculation, which includes 0.75 linear separation method of Lind, power series expansion method, distribution value method, general separation method. All the computational methods are calculated and comparative analyzed in the case of gravity retaining wall in general area, and paper presents the advantages, disadvantages and selection principles of these methods.(2)For the current used M-C method, this paper respectively elaborates the impact on the result of reliability index calculation, from aspects of "random variable ", "statistical methods", "earth pressure calculation", " inactivation judgment ", "correlation between variables". Furthermore, it analyzes the error sources in current reliability index calculation.(3)For the general separation method of partial coefficient calculation currently in use, the paper respectively elaborates the impact on the result of partial coefficients calculation, from aspects of "separation times" and "correlation between variables". Moreover, it analyzes the error sources in current partial coefficient calculation, and gives partial coefficient formula in the case of related variables.(4)The self-consistent process of reliability index and partial coefficient calculations, and the key issues of self-consistent are proposed. In addition, this paper demonstrates the self-consistency of reliability index and partial coefficient calculations in the form of numerical examples.