A comparative study of damage identification of a steel bridge girder using numerical modal analysis

Abstract

A steel bridge girder is one of the crucial structural components in transportation infrastructure. It is typically exposed to corrosion over time, leading to a reduction in steel thickness and deterioration of structural
performance. Early detection and localization of such damage are essential for ensuring structural safety and
longevity. This study presents a comparative numerical investigation on vibration-based damage identification
for a steel bridge girder using modal analysis. The purpose is to identify damage in a steel girder, considering
the effects of damage size and location, meshing, and element type in the numerical simulation. A simply supported steel girder with a span length of 15.2 m was selected for the analysis using both beam and solid
element models in Abaqus/CAE. Damage was modeled as an equivalent reduction in Young’s modulus to represent the stiffness loss associated with corrosion or cracking, rather than explicitly simulating geometric section loss. The reduction in Young’s modulus was introduced at the support, 1/4-span, and 1/2-span, while the influence of mesh density and element type was also investigated. Three damage indices, namely Modal Curvature Damage Index (MCDI), Modal Strain Energy Damage Index (MSEDI), and Modal Flexibility Damage Index (MFDI), were then evaluated and compared. The results showed that damage at 1/2-span produced the most significant reduction in natural frequencies, while multiple damages showed cumulative effects. For MCDI and MSEDI, Modes 1 and 3 offered reliable damage localization, whereas other modes demonstrated complexity due to multiple curvature effects. MFDI provided more effective damage localization when all mode shapes and natural frequencies were considered. Mesh density affected damage quantification but not localization. Damage quantification varied between three damage indices for different mesh densities. It was also found that beam elements were computationally efficient, while solid elements provided greater flexibility for representing localized stiffness changes. This study provided a simple and practical guidance for the Structural Health Monitoring modeling of steel bridge girders. However, it should be noted that this study is limited to Young's modulus reduction as an equivalent local stiffness loss. To better capture the damage, actual geometry modeling of corrosion or cracking should be further studied.