Sensor-based fatigue verification of bridge components using monitoring-informed stress ranges and partial safety factor calibration

In the semi-probabilistic approach of structural design, partial safety factors on the action side are calibrated at code level to account for uncertainties in actions and resistance. For fatigue limit states, these uncertainties are typically embedded in conservative traffic load models and fatigue strength curves. With the increasing availability of structural health monitoring data, it becomes possible to base fatigue verification on monitoring-informed stress ranges rather than purely code-based assumptions. This contribution presents a methodology for sensor-based fatigue verification of bridge components and for the calibration of monitoring-informed partial safety factors on the action side. As input, time-dependent probability distributions of stress ranges at fatigue-critical details are assumed to be available from a monitoring-based load and stress identification procedure. The updated structural uncertainty and the identified axle-load uncertainty are propagated through a stress-response model and a cumulative fatigue-damage model, such that the action side is represented by measurement-based stochastic fatigue damage and the resistance side is described by a deterministic fatigue resistance model. The time-dependent probabilities of fatigue failure are evaluated by simulation-based reliability methods and compared with target reliability levels. Using the First Order Reliability Method (FORM), partial safety factors for fatigue verification on the action side are derived such that the semi-probabilistic design format remains applicable while explicitly accounting for sensor-based information on actual load effects. The proposed approach provides a step towards monitoring-informed fatigue design and life prediction for bridge structures.