Stochastic Vehicle Load Simulation Method for Large Span Bridges Based on Weigh-in-Motion Data

The purpose of this study is to present a method for the simulation of vehicle loads on long span bridges based on the data recorded by the Weigh In Motion (WIM) system. Due to the strong randomness of the traffic loads on large span bridges, the traditional research methods have significant limitations in accurately analyzing the characteristics of the random traffic loads, which in turn affects the assessment of the bridge structural conditions. On the basis of the traffic information collected from a bridge's WIM system, traffic weight and speed of different vehicle types have been counted. A multi-lane random vehicle load simulation program based on MATLAB is developed based on the basic ideas of the vehicle following model and the lane-changing model to complete the simulation of random traffic flow. It is mainly composed of the following aspects: (1) Seven vehicle types were classified and the proportion of vehicle types in each lane was investigated, based on the actual collection of vehicle data from the studied bridge. The probability density distribution function of the vehicle speed and vehicle weight was obtained by fitting the data with the mixed normal distribution and the mixed log normal distribution. (2) Based on the statistical distribution model of the measured vehicle parameters, the Monte Carlo method is used to generate the important parameters such as vehicle type, vehicle weight, vehicle speed and so on by sampling, and finally to form the random vehicle load sequence under the specified flow rate. (3) By introducing the following model, the rules of acceleration and deceleration of vehicles are specified. Thus, the simulation of traffic flow can be more realistic. Besides, a lane-changing model is established considering that lane-changing behaviour occurs in real vehicles.