Study On Improved Line Sampling Methods For Structural Reliability And Sensitivity Analysis

In consideration of the complexity of the structural response caused by strong nonlinearity of the limit state function,it is necessary to use effective methods for reliability analysis.Monte Carlo simulation(MCS),which is widely used by engineers and researchers all around the world,is one of the most powerful methods to solve reliability problems.Among the advanced Monte Carlo simulation methods,line sampling(LS)is popular and has attracted much attention in recent years.Line sampling has a simple principle and is easy to oprate.The basic idea behind line sampling is to refine estimates obtained from the first order reliability method(FORM).In order to balance the efficiency and accuracy of the calculation,many improved line sampling methods have been proposed and applied to reliability analysis of different kind of problems.In this thesis,improved line sampling methods are proposed to help identifying the importance direction.Then,line sampling and proposed methods are used to analyzing seismic reliability and sensitivity of building structures.The main research contents are as follows:(1)Based on the basic principle,the program of line sampling is compiled with MATLAB.The efficiency and applicability of line sampling is demonstrated via numerical examples and simplified structural examples,including reliability and sensitivity analysis.Meanwhile,some problems in the process of line sampling calculation is discussed.(2)Combining the weight average simulation method(WASM)and markov chain Monte Carlo(MCMC)simulation with line sampling,an improved method is proposed to identify the importance direction and execute sampling.The efficiency and applicability of the proposed method is demonstrated via numerical examples and simplified structural examples.(3)Combining the genetic algorithm(GA)with line sampling,another improved method is proposed to identify the importance direction.The efficiency and applicability of the proposed method is demonstrated via numerical examples and simplified structural examples.(4)In order to improve the efficiency of structural reliability analysis,Kriging model is combined with line sampling and its' improved methods.Then,the hybrid reliability method is used to analyze the static and dynamic seismic reliability and sensitivity of a three-storey reinforced concrete frame structure.(5)Considering the uncertainty of load and structure,ultimate limit state function and deformability limit stare function for static seismic reliability problem are established respectively.The failure probability and sensitivity of the three-storey RC frame are calculated by proposed methods.The results are compared with MCS's.(6)The stochastic process model of earthquake ground motion is introduced as excitation.Considering the structural uncertainty,the deformability limit stare function under dynamic load is established.Both failure probability and sensitivity of the frame are calculated by proposed methods.The results are compared with MCS's.The study found that improved line sampling methods proposed in this thesis,which could achieve high accuracy of importance direction,have high accuracy and efficiency in solving structural reliability and sensitivity problems,and are well applicated in structural seismic reliability analysis.