An analytical solution to traffic loads on long span bridges

Abstract

The paper deals with the prediction of extreme values of traffic loads and load effects of long span bridges. The traffic states are idealized a series of spatially stationary stochastic fields, which correspond to free, congested and full stop traffic states. For each traffic state, the Poisson point field is invoked to model the spatial distribution of vehicles. The vehicle density is then determined initially by the empirical formula based on microsimulation and modified further by a Poisson point model with a limited minimum space lag. With the modified vehicle density and influence functions of the bridge response, the expressions of the mean values and variances of the traffic loads or load effects are derived based on Poisson model, which are very simple compared with Ditlevsen’ formulas. Because the bridge loads are the sum of a large number of independent vehicle weights, the normal distribution is a reasonable model for loads and load effects. Correspondingly, the long term characteristic values and distribution functions of traffic loads and load effects are estimated easily with the derived mean and variance . The numerical analyses show that all results calculated are very close to those by the popular microsimulation methods. Hence, an analytical method is developed for predicting the traffic loads and load effects, which can be easily used to develop a design code for a long span bridge. Moreover, the analytical method based on Poisson models incorporated with the normal distribution of loads or load effects would make it more easily combine the traffic loads with wave and wind loadings for a floating bridge.