A large-scale experiment on the seven span Flossgraben Bridge in operation to investigate damage localization utilizing state projection estimation error (SP2E)

A three-day large-scale experiment was carried out on the 358 m long, seven-span Flossgraben Bridge in Germany to investigate Structural Health Monitoring (SHM) under operational conditions. Reversible structural changes were introduced by positioning cargo trucks at different locations, simulating damage-equivalent mass alterations while one traffic lane remained open. Output-only acceleration data was collected from 56 nearly equidistant sensors along the bridge, complemented by environmental measurements. The study applied the State Projection Estimation Error (SP2E) method, which relies on stochastic subspace identification and H ∞ -optimization of state space models, supported by the Autonomous Model Order Selection (AMOS) framework. The results demonstrate that SP2E successfully detects and localizes induced structural changes while the structure is in operation. The published dataset and experimental description provide a benchmark for validating SHM methods on large-scale structures in operation. • Large-scale experiment on the 358 m Flossgraben Bridge in operation. • Reversible mass-induced damages simulated using cargo trucks in two spans. • SP2E method validated for damage localization under ambient excitations. • Large amount of measurement data provided (56 acceleration sensors + environment).