Long-term monitoring of a PSC box girder bridge: Operational modal analysis, data normalization and structural modification assessment

► A novelty detection based on unsupervised learning techniques is developed. ► Neural networks are used to filter temperature effects on modal parameters. ► Application is made for the monitoring of a box girder bridge before and after strengthening. ► The random decrement technique is used as an operational modal analysis method. ► Symbolic data analysis and clustering methods are combined to assess strengthening efficiency.