Detecting Weakened Highway and Railroad Bridge Substructures at Deck Level

The ability to detect damage in substructures of highway and railway bridges using modal vibration techniques performed at the superstructure deck level is investigated as a means to improve nondestructive testing evaluation in cases where visual inspection is difficult or impossible. Methodology initiated in a 2012 NCITEC project is extended here first through a limited experimental study of lab scale models of substructure subsystems in a variety of configurations. Tests are conducted on a shake table to assess variations in frequency and temporal dynamic response characteristics for the different configurations, and frequencies are compared with those obtained by simplified finite element analysis of the subsystems. Detailed finite element analysis is then performed to characterize the dynamic characteristics of two full scale three-span highway bridges accessible to the project team. The two bridges have similar superstructures consisting of composite steel girder decks but lie in different geologic formations such that one was designed with a deep foundation system and the other with a shallow one. Fixed base models highlight the significant difference in fundamental frequencies for the two structural systems even when the different foundations are not considered. Soil-structure interaction models are developed to incorporate the soil and foundation elements and account for scour conditions that are the subject of a recently completed 2013 project. The deep foundation or flexible system is used to characterize the effect of soil-structure interaction and the influence of damage scenarios on the dynamic characteristics. Damage scenarios are considered that consist of material deterioration in the form of softening modeled as reductions of the elastic modulus in various substructure elements of one of two central piers. By virtue of the companion study, a damage scenario is also considered in which symmetric scour of a stream bed occurs between the two central piers. The material deterioration scenarios do not produce noticeable changes in the modal frequencies of the flexible system whereas the scour scenario produces changes in modes involving horizontal movement of the deck mass that are potentially significant enough to be detectable by measurements made at deck level.