One main objective of structural health monitoring (SHM) is to assess the performance of structures.The detection of structural damage at an early stage is essential for civil engineering infrastructures to prevent the occurrence of catastrophic failures and to reduce costs concerning maintenance.A long-term structural health monitoring (SHM) is usually considered for monitoring the structural performance level and for damage detection.This method has been widely considered for aircrafts, ships, offshore platforms, buildings, as well as for bridges.In this paper, the authors focus on the damage assessment problem based on a vibration-based detection approach specifically designed for a bridge in real environment and traffic conditions.For this purpose, a cluster-based approach is proposed to discriminate abnormal changes from normal changes in the structural behavior.Besides, symbolic data analysis (SDA) is introduced to process and analyze large amounts of data.At the same time, some novelty detection strategies using original symbolic objects and principal component analysis (PCA) are proposed to extract useful information related to structures from large amounts of data.The efficiency and reliability of this approach is checked using both simulation and real data for a road-rail bridge, the Adour Bridge, which is a steel road-rail bridge located in Bayonne, France, and managed by SNCF (National Corporation of French Railways).A long-term structural health monitoring (SHM) program of the Adour Bridge was decided until its demolition, to analyze the structural dynamic behavior and detect abnormal structural response under the current traffic loads.