Bridge Analysis, Design and Repair
342520 - Multi-Scale Finite Element Model Development for Long-Term Condition Assessment of Vertical Lift Bridge
Friday, April 20
3:30 PM - 5:00 PM
Early stage objective information relating to the structural integrity of civil infrastructure elements such as bridges is increasingly critical for bridge engineers and managers concerned with long-term structural condition assessment. Collecting this information through sensor networks to capture and structural models to predict the dynamic and static responses of the bridge is more practical of late given advancements in technology and communication. This information will proactively assess and maintain the bridge performance, reduce long-term maintenance costs, and advance sustainable development of smart bridges around these issues. However, in real-time monitoring of new bridges, the complexity of the structure, as well as the range of loading conditions, including traffic and environmental loading, may result in missing the vulnerable locations of damage due to a limited number of sensors. The structural performance relationship between the critical locations and the instrumented components could be established via a global Finite Element (FE) model containing full details of the bridge to be analyzed. In this paper, the newly reconstructed Memorial Bridge is considered as a case study to monitor dynamic changes of structural elements under different loading condition. The bridge is a vertical lift bridge, connecting Portsmouth, NH. to Kittery, ME, which is instrumented with a sensor network consisting of strain rosettes, accelerometers, tiltmeters and a weather station that is permanently deployed at the most critical locations along the bridge. This paper addresses challenges associated with the bridge structural response in non-instrumented components via a full detailed 3D FE model made of shell elements. The initial analysis of the model reveals less important components for reducing the scale of such elements into 2D beam elements, illustrating a promising approach for developing a novel multiscale model. It is found that the multiple scale model has the accuracy of full shell model while becoming more practical as beam element models in terms of computational efforts. The results indicate that the model is in good agreement with monitoring data collected from the bridge.