Assessment of Steel Truss Bridges Using Distributed Strain Measurements

Abstract

Steel truss bridges are an important component of the railway network across North America. A majority of this infrastructure has already exceeded its 100 year service life, making the monitoring of these structures critical to ensuring they have adequate capacity to handle the increasing freight loads. Visual inspection remains the primary monitoring method because of its cost and time effectiveness. However it can be subjective, produce varying results between inspectors, and is unable to accurately quantify damage or behaviour. The use of distributed fibre optic sensors has the potential to overcome the shortcomings of visual inspection, along with the added benefit of capturing thousands of measurements with a similar effort required to obtain a single measurement with conventional strain gauges. The current research investigated the use of distributed fibre optic strain measurement to monitor the behaviour of steel truss bridges. A scaled down model truss based on an in service railway bridge was constructed in a laboratory setting. The truss was monitored with both fibre optic sensors and linear potentiometers. The connection conditions were varied by either removing or loosening bolts between vertical and diagonal members. A response change was detected by both sensor technologies when all bolts were loosened versus tightened. Varying degrees of corrosion and cracking damage were simulated on members, and the strain changes in the members were detected using the distributed strain sensors. An in service rail bridge was monitored and static distributed strain measurements were taken as a train of known weight was placed at various positions along the bridge span. The same train then travelled across the bridge at varying speeds as dynamic measurements were taken. Finally, in service measurements for passenger and freight trains were also taken. The static measurements were used to detect localized variations in strain due to changing member geometry as well as global behaviour. The dynamic strain measurements for a train travelling at 4 km/h showed good correlation to the static measurements, however all dynamic measurements gave inconclusive results beyond the first 10 m of sensing fibre. In service vehicles could not be monitored at their normal travelling speed.