Evaluating the Performance of Protection Beams Subject to Overheight Vehicular Impacts Using Analytical and Machine Learning–Based Methods

Abstract

Overheight truck collision with a bridge superstructure can cause extensive structural damage and result in severe traffic congestion. Placing protection beams in front of the fascia girder is one of the common countermeasures to guard against such events. However, little research has been done on this topic in the past and there are no standard guidelines for designing these protection beams. High-fidelity numerical simulations are conducted to formulate a demand model for protection beams subjected to overheight truck impact scenarios involving three different trailer models (representing soft, semirigid, and rigid impactors), a range of truck weights and velocities, and a number of beam sections and lengths. A validated demand model is proposed to represent the loading pulses caused by the different types of impactors. Two performance levels for the protection beam are suggested. Simulation data are used to propose a reliable machine-learning (ML) model for the classification of beam performance as a function of the parameters of the impacting truck and protection beam. The ML model and demand function are complementary to each other and enable rapid and accurate assessments, respectively, of the performance of protection beams during impact scenarios.