Research on the general method for extrapolating traffic load effects for highway bridges

Accurate knowledge of real traffic load levels is one of the prerequisites for structural safety assessments, and the common method used to extrapolate load effects depends on the analysis of effect crossing rate distribution according to Rice's Theory. But as proved by the analysis made in this paper from three perspectives (theoretical basis of the method, determination of fitting starting class interval and discreteness of extrapolated effects), the method is flawed in at least three respects: obvious deviation between the presumption of Rice's Theory and the real distribution of crossing rate, narrow application range (extremely high traffic density) and unreliability in extrapolation of effects. Besides, the method only estimates the effects of the return period, without a reliability analysis of randomness in sampling and thus an effective estimation of effect range containing true values. Therefore, this paper, after analysing the relation between return period and extreme value distribution of load effects, as well as the occurrence rule for return period effects, proposes a general method for extrapolating traffic load effects on highway bridges. The general method, with the characteristic of high utilization of sample data in effect extrapolation from crossing rate distribution, is wide in its scope of application; it can be used for traffic flows with different densities and influence lines with different forms and lengths, and provides a confidence interval for estimated effect values. The method thus represents a significant improvement to the reliability of extrapolating effects for different return periods.