Camera-based Bridge Health Monitoring

This project is about the development of novel damage indicators that will utilize new forms of data, specifically camera data of bridge deflections. The cameras will primarily monitor deflections and will scan at a frequency that can monitor the bridges’ vibrational behaviour.
The bridge monitoring problem will be addressed by a consortium of three: University of Central Florida (UCF) in the United States, University College Dublin (UCD) in Ireland and the Queens University of Belfast (QUB) in the United Kingdom. This vacancy is in University College Dublin and relates only to UCD’s part of the overall project. The student will be based in Dublin with occasional visits to the other countries. Two other UCD researchers will be working on the project (a PhD student and post-doctoral researcher) and the tasks will be shared within the group, according to each researcher’s skills.
To achieve the intended goal, three primary objectives need to be realized (i) Develop a new framework for structural identification using computer vision approaches (not the responsibility of UCD), (ii) Develop enhanced vibration-based damage indicators by improving existing indicators and developing new indicators, (iii) Update bridge safety information and evaluate remaining life using damage indicator data. To achieve these objectives, the project has been broken up into the four distinct work packages (WPs) shown below (WP leaders in brackets). University College Dublin (UCD) is only involved directly in WP2 and WP3 but may use some of the field measurements from WP4.
• WP1 New framework for structural identification using computer vision methods (UCF)
• WP2 Damage detection (UCD and UCF)
• WP3 Safety & remaining life evaluation (UCD)
• WP4 Laboratory testing and field trials (QUB)
Work Package WP2 – Damage Detection

Work Package WP2 – Exploiting Vehicle Class/Type Information in Bridge Damage Assessment
When the traffic data is broken up into the different vehicle types, the results become much more repeatable. In this task, a common vehicle type such as a semi-trailer, will be identified and bridge response data will be filtered to give responses to this type of vehicle only. At the same time, deflections will be measured at a point, using a camera mounted on a telescope, at high resolution (20 to 50 microns).

(a) Long Term Monitoring: For long term monitoring, data will be available over enough time to collect a ‘population’ of bridge responses to trucks of a given class. The dynamic response of a bridge to a truck is a complex vehicle/bridge dynamic interaction problem. In this project, we will consider the dynamic responses from a population of similar trucks. The research question will be to explore whether the dynamic response of a bridge to a population of trucks is repeatable and how sensitive is that dynamic response pattern to bridge damage.

(b) Short Term Monitoring: An approach will be explored where a short term monitoring with a camera can be efficiently conducted in the few hours that it takes to inspect the bridge. An Inverse Bridge Weigh-in-Motion (WIM) algorithm will be applied. Bridge WIM is the problem of finding truck weights from the bridge response. Inverse Bridge WIM is the problem of finding the bridge behavior given the response and the trucks’ axle weights. The bridge behaviour is described by the inferred influence line, i.e., the response to a unit point load. Any change in the inferred influence line is a good indication that the bridge has deteriorated, e.g., a bearing has seized.

Work Package WP3 – Safety & Remaining Life Evaluation
While damage indicators are important for detecting any change that could be due to damage, they may not provide information about the actual safety of the bridge that decision makers can immediately utilize. In this project, the link between damage indicator and bridge safety will be established. The end result will be a novel vision-based probabilistic load rating based on the new data gleaned from the camera-based SHM system.

This part of the project will involve the development of Reliability Models of the bridge (e.g. FORM or SORM) and using the camera-based data to update these models to reflect the bridge’s current condition.