Prediction of bridge maximum load effects under growing traffic using non-stationary bayesian method

Abstract

The past decades have witnessed a considerable growth of road traffic as result of economic developments and technological advances. The prediction of the maximum bridge traffic load effects (LEs) under the growing traffic can provide valuable information for bridge design and condition assessment. However, most previous studies assumed that the traffic is a stationary process when extrapolating the maximum traffic LEs. In order to more accurately predict the maximum traffic LEs, a Bayesian framework for predicting non-stationary extreme traffic LEs of bridges subject to growing traffic is presented in this study. Long-term traffic LEs are simulated using Monte Carlo simulations and influence line analyses considering three types of traffic growth including the growth of the truck volume, the proportion of heavy vehicles, and the truck weight. The non-stationary Bayesian method is applied to predict the maximum traffic LEs during the bridge lifespan using the simulated traffic LEs. The influence of the traffic growth on the bridge safety is investigated. The results obtained can provide references for the decision making on regulation changes and bridge management.