Sensitivity of predicted bridge traffic load effects to the tails of truck weight distributions

In the last two decades, simulations have been used to predict the characteristic traffic load effects on bridges using Weigh-In-Motion (WIM) data. The recorded Gross Vehicle Weight (GVW) is usually modelled by multimodal distributions. The parameters for these distributions are generally obtained by using different goodness-of-fit tests where the entire recorded data is considered. These parameters are then used as the basis for the simulations. In this work, the sensitivity of the predicted traffic load effects to these fittings is investigated. Generally, moment at mid-span for different return periods can be determined from simulations based on three different assumptions. The first approach is to use empirical distribution functions, i.e., direct simulations using the recorded GVW data. The second approach is to use parametric distribution functions to represent GVW from the recorded data. In the third approach, developed here, semi-parametric distribution functions are used to model the distributions of GVW. From these, load effects corresponding to different return periods are calculated and compared. The results are shown to be highly sensitive to the assumption adopted.