Updating prediction of fatigue reliability index of railway bridges using structural monitoring data and updated load histories

The assessment of fatigue consumption and the remaining lifetime of structural components is affected by considerable uncertainties on the side of the traffic loads, fatigue resistance and structural response. The purpose of the presented work was to develop methods for dealing with these uncertainties, as well as methods for improving the accuracy of assessment with the use of additional data.
Within the research project Assets4Rail, a structural monitoring system was installed on a railway bridge located on a local track in Austria. The system consisted of strain sensors, acceleration sensors and inclinometers. It was used to measure the bridge response during train passages with known axle loads in course of a test with controlled conditions. This data was used to calibrate the structural model and develop probabilistic methods for fatigue assessment. Influence lines at fatigue-critical locations were evaluated from measured bridge strain response including their uncertainty. Further uncertainties considered in the assessment include the load histories and the fatigue resistance.
The results showed the largest contribution by evaluation of model uncertainties from monitoring data. The effect of model updating was also considerable, but less significant. Further increase of estimation accuracy is achieved using section-specific traffic data. Whereas wayside monitoring data represent the reference scenario, the use of traffic management data provides a usable alternative.