A NUMERICAL STUDY ON THE APPLICABILITY OF VBI SYSTEM IDENTIFICATION USING VEHICLE VIBRATION DATA FROM MULTIPLE RUNS

Abstract

One of the primary screening methods to evaluate the damage probability and health of bridges is the VBISI(Vehicle-Bridge Interaction System Identification) method, which simultaneously estimates vehicle and bridge mechanical parameters(mass, damping, stiffness) and road surface roughness based only on vehicle vibration and position data. In the ideal case without noise, the method has been shown to have excellent efficiency and accuracy in mathematical models. However, in the presence of noise, the estimation accuracy of mechanical parameters decreases. In this study, we propose a new algorithm for estimating mechanical parameters of vehicles and bridges from big data, and attempt to improve the estimation accuracy against noise. The obtained results show that VBISI method with multiple runs improves the estimation accuracy compared to the models in the previous studies. However, it has not yet reached the level of application to field experiments. Further improvement of the estimation accuracy and its application to field experiments are considered in the future.