Fatigue reliability analysis of long span bridges with structural health monitoring systems

This study is to present a framework for assessing structural reliability in association with fatigue damage at critical locations of long span bridges by making use of in-situ wind and structural health monitoring systems. The Tsing Ma Bridge in Hong Kong, a long suspension bridge that carries both highway and railway, is taken as an example. The bridge has been continuously monitored for the past ten years by a system known as Wind and Structural Health Monitoring System (WASHMS), which measures the loadings and structural responses of the bridge. Railway, highway and wind loadings are considered to be major contributors to the fatigue damage of the bridge. Random parameters in probabilistic loading models are represented by mixture distribution functions and established by the expectation-maximization (EM) algorithm on the basis of the monitoring data acquired by WASHMS. Daily stochastic stress processes in response to railway, highway and wind loadings at critical locations are developed by virtue of the probabilistic loading models and the finite element stress analysis. With a Monte-Carlo simulation, a location-dependent probabilistic distribution function of daily stress ranges is then derived on the basis of a large number of virtual daily stochastic stress processes. Meanwhile, structural fatigue resistance is also considered as a random variable in the Miner's rule taking account uncertainties in fatigue life assessment. Finally, a limit state function of fatigue reliability is elaborated and a fatigue reliability index is estimated, depending on combined wind, and railway and highway loadings. It is found that the fatigue reliability index decreases nonlinearly with time, and it will be less than 3.7 after 95 years for some critical elements.