Mathematical Formulation Of Tools For Time-dependent Reliability Assessment Of Existing Structures

Existing aging structures and infrastrures may suffer from aggressive operating and environmental coditions during their service life. Such conditions may cause the deterioration of structural resistance fall below an acceptable level as specified for new ones. Reliability analysis is a widely-used tool for structural safety assessment dealing with the uncertainties associated with the structural resistance and loading process based on probabilistic and statistical theorems. It provides quantitive criterion for sturcutral design and optimization of maintanance meansures conducting reliability analysis of existing aging structures of interest.Realsitic resistance deteriroration model and load process model are the fundamentals for structural reliability analysis. In current researches, the probabilistic models of variables are often simplified with some assumptions. Such assumptions may result in significant difference between the theoretical reliability assessments and the realistic ones. This paper mainly focuses on the improvement of probabilistic models of random variables and the analysis methodology associated with the reliability analysis. In detail, it includes the followings.(1) The correlation in loading process is taken into account, the effect of which on structural time-dependent reliability is investigated using Monte Carlo simulation method. It was found that the reliability is underestimated if the load correlation is neglected and that larger load correlations result in greater uncerestimates of the reliability.(2) The time-variant properites of statisticas of loads(e.g., mean value and standard deviation) are considered when analyzing structural time-dependent reliability. An explicit formula is present, which is simple for calculation. It is shown that the traffic load change has great impact on structural reliability; the increment of mean load by 1.1% per year, or the increment of standard deviation of load by 1.3% per year, results in a twice failure probability in a service period of 20 years.(3) The effect of the time-variant occurrence frequency of wind load in structural time-dependent anti-wind reliability is investigated. An explicit formula is given for the analysis of reliability.(4) It utilizes the traffic load history information to update the time-dependent reliability for subsequent service years, and it shows that if over-weighted vehicles are observed in traffic history or the resistance has remarkable degradation, the updated reliability using the proposed method is a more reasonable indicator of bridge performance, while ignoring the service history information will result in an underestimated reliability.(5) The updating effect of traffic load history on prior structural resistance is investigated, and an explicit formula is presented for the updating of resistance. It shows that traffic load history and resistance degradation have significant effects on the reliability of existing bridges. If over-weighted vehicles are observed in traffic history or the resistance has remarkable degradation, the updated reliability using the proposed method is a more reasonable indicator of bridge performance, while ignoring the service history information will result in an underestimated reliability.