Traffic Load Simulation for Long-Span Suspension Bridges

Abstract

This study presents a procedure for simulating traffic loads for long-span suspension bridges based on the data recorded by the weigh-in-motion (WIM) system and the agent-based traffic flow microsimulation. The past traffic volume recorded by the WIM system is first used to forecast the future traffic volume in terms of the annual average hourly traffic (AAHT) volume and using the seasonal autoregressive integrated moving average (SARIMA) method so that both seasonality and growth trend effects can be considered. The WIM data are then analyzed and used to determine the probability models of major vehicle parameters, such as vehicle type, speed, axle weight, and axle spacing. The motions of individual vehicles on the bridge are finally simulated by using the agent-based traffic flow microsimulation, including both free flow and congested flow conditions, considering the drivers’ behavior of vehicle following and lane changing, and using the Monte Carlo simulation. The Tsing Ma suspension bridge in Hong Kong is taken as a case study to demonstrate the traffic load simulation. The results show that the proposed procedure is suitable for simulating traffic loads of a long-span suspension bridge. The results from the case study also show that simultaneous “long-term” and “short-term” traffic load simulation is important so that both seasonality and growth trend can be included.