| Railroad bridges form an integral part of railway infrastructure in the USA carrying approximately 40 % of the ton-miles of freight. The US Department of Transportation (DOT) forecasts current rail tonnage to increase up to 88 % by 2035. Within the railway network, a bridge occurs every 1.4 miles of track, on average, making them critical elements. In an effort to accommodate safely the need for increased load carrying capacity, the Federal Railroad Association (FRA) announced a regulation in 2010 that the bridge owners must report annual inspection of all the bridges. Until now, visual inspection has been the most prevalent practice in monitoring this infrastructure, while high-cost and unreliability can limit the efficiency and accuracy of such assessments. With recent advances in sensing technology, structural health monitoring can be a promising solution for providing a reliable and inexpensive ways for assessing the bridges. Nonetheless, because damage is a local phenomenon, to be able to detect/ monitor existing/potential damage, densely deployed sensors are required, which is inefficient and still expensive. Alternatively, model-based monitoring strategies can be adopted to identify a critical element from a numerical model that has been calibrated with measured field data. However, this approach has been widely adopted and applied for highway bridges, while railroad bridges have received comparably less attention. The main reason for the limited number of studies is due, in part, to fundamental differences between the loading being applied to highway bridges and railroad bridges. Usually, the mass of the vehicles crossing highway bridges is assumed to be relatively small compared to the mass of the bridge itself; as a result, the mass of the vehicles are often neglecting in the problem. In contrast, the mass of a train crossing a railroad bridge can be as large as the mass of the bridge itself. Moreover, trains are typically composed of an engine, followed by multiple cars resulting in a nearly deterministic moving mass/load being applied to the bridge that varies with speed. As a consequence, numerous models have been developed to understand the dynamic response of bridges under in-service train loads, but most fail to provide a simple, yet flexible, representation of the salient features of the responses of the bridge.;The objective of this research is to develop appropriate modeling and monitoring techniques for railroad bridges toward understanding the dynamic responses under a moving train. To achieve the research objective, the following issues are considered specifically. For modeling, a simple, yet effective, model is developed to capture salient features of the bridge responses under a moving train. A new hybrid model is then proposed, which is a flexible and efficient tool for estimating bridge responses for arbitrary train configurations and speeds. For monitoring, measured field data is used to validate the performance of the numerical model. Further, interpretation of the proposed models showed that those models are efficient tools for predicting response of the bridge under undesirable and local phenomena, such as fatigue and resonance. Finally, fundamental software, hardware, and algorithm components are developed for providing synchronized sensing for geographically distributed networks, as can be found in railroad bridges. The results of this research successfully demonstrate the potentials of using wirelessly measured data to perform model development and calibration that will lead to better understanding the dynamic responses of railroad bridges and to provide an effective tool for prediction of bridge response for arbitrary train configurations and speeds. |
