A damage detection study of a bridge using bypassing vehicles and computational intelligence

Abstract

Structural integrity is an ubiquitous topic in research and daily living. Conventional monitoring strategies follow a preventive maintenance plan or require permanent installation of sensors. These steps are time and cost-inefficient taking into account the high number of bridges to be monitored. In this study, a bridge structure is passively monitored using the acceleration signal of a bypassing vehicle. Thereby, the signal is measured on the vehicle. The health status of the bridge can be evaluated after each pass, which allows a continuous monitoring of the structure. A numerical model of a bridge and a bypassing car is carried out to prove the feasibility of this approach and to create a large database with different damage locations and vehicle velocities. Moreover, an artificial neural network is trained on that database to detect damaged bridge structures. The proposed method proves the feasibility of monitoring the structural integrity of bridges and supports conventional maintenance methods. Therefore, the combination of active and passive monitoring strategies paves the way to efficient and cheap infrastructure monitoring strategies.