Developing a benchmark study for bridge monitoring

Abstract

Structural health monitoring is the process of implementing a continuous damage detection strategy to optimize the inspection and maintenance schedules of bridges, and extend their lifespans. One of the main challenges of automated damage detection is the lack of data on damaged states, which makes it difficult to validate new approaches in the research and development stage. To alleviate this problem, a monitoring campaign on a two‐span test bridge with defined defects is conducted and documented in this article. The bridge is a steel‐concrete composite structure with a length of 30 m, with two primary steel girders and a segmented concrete deck. The recorded data capture the long‐term ambient data from 18 test days and changing environmental conditions, as well as the short‐term ambient data and dynamic load tests from four damage scenarios with well‐defined damage extents. A mobile measurement system with numerous sensors is used for data acquisition. A shaker is placed on the bridge to excite white noise. The main goal of this article is to document the experimental procedure and perform preliminary plausibility checks on the measured data. First results demonstrate that system response data and environmental conditions are recorded reliably and that environmental effects significantly affect the long‐term measurements. Therefore, a suitable data set is provided as open‐source data for future studies on data normalization and automated damage detection.