Assessing confidence intervals of extreme traffic loads

Xiao Yi Zhou (1) , Franziska Schmidt (1) , Bernard Jacob (2)

1 IFSTTAR/SOA - Département Structures et Ouvrages d'Art
2 IFSTTAR/DS - Direction scientifique

For short- to medium-span bridges, truck gross vehicle weights (GVW) and axle loads are of great importance to assess extreme load effects, as extreme load events are obtained by combining the extremes of GVWs or axle loads. In this paper, three known prediction methods are applied, a normal distribution fitting to the sample distribution upper tail, the block maxima method and the peaks-over-threshold method, to extrapolate extreme traffic loads. Moreover, two methods, the delta method and the profile likelihood method, are introduced to assess the confidence in these extrapolations. Results show that generalized Pareto distribution based peaks-over-threshold method is the best among these three methods to assess the extreme traffic loads. WIM data, gross vehicle weight, return period, return level, confidence interval, delta method, profile likelihood method, generalized extreme value distribution, generalized Pareto distribution

Mots clés: PESAGE EN MARCHE; INTERVALLE DE CONFIANCE; VALEUR EXTREME

Domaines: Mécanique des structures [physics.class-ph]; Mécanique des structures [physics.class-ph]

Xiao Yi Zhou, Franziska Schmidt, Bernard Jacob. Assessing confidence intervals of extreme traffic loads. International Conference on Weigh-In-Motion ICWIM6, Jun 2012, France. 8p. ⟨hal-00950523⟩