Efficient Numerical Methods For Structural Time-dependent Reliability Analysis And Cost Optimization

Engineering structural parameters often have uncertainties,including geometric parameters,material properties,and load excitation.The particular uncertainty of these parameters usually exists in the form of coupling,and changes with the evolution of the structural performance model,which directly affects the entire life cycle.Due to the existence of these uncertainties,there are some unnecessary errors in the time-varying reliability results of the degenerate structure and the cost optimization results based on the age replacement strategy.This has a significant impact on the development of optimal strategies for the lifecycle management of infrastructure systems such as roads,bridges,nuclear power plants and transmission lines,and can even lead to more serious accidents.This paper presents an effective Bayesian update method for time-dependent reliability analysis and Cost optimization of degraded structures.The method includes additional information such as measurement Data via Bayesian modeling to reduce estimation uncertainties.(1)In this paper,the accelerated corrosion process of pipeline flow in a power station is taken as the research object.The degradation model under structural strength is established according to the degradation data of pipeline wall thickness,and the time-varying reliability analysis of the pipeline is carried out.(2)A general posterior distribution model for Bayesian information updating of uncertain parameters is established,and the Bayesian posterior distribution sample is estimated by DRAM algorithm.The influence of observation data on the posterior distribution of parameters and reliability index is discussed with numerical examples.The Bayesian update method is applied to the parameter update of the pipeline degradation model.And the time-dependent reliability analysis results before and after the parameter update are compared.The shortcomings of DRAM algorithm considering the marginal distribution of parameters when establishing the proposed function are reviewed.(3)The Copula function is proposed to establish the proposed function of the DRAM algorithm.Compared to standard DRAM algorithm and MH algorithm,results have proven High accuracy of the proposed method for Bayesian posterior distribution problems.Monte Carlo Method is developed to evaluate structure responses given a reliability index.(4)Combined with the proposed improved algorithm,the cost optimization of preventive replacement maintenance for pipelines is based on three age-based replacement cost optimization models,and the optimization results before and after the update are compared.The results show that the introduction of observation data can effectively reduce the expected cost rate and optimal replacement time of pipeline preventive replacement maintenance.