Damage detection in a real truss bridge using Hilbert-Huang transform of transient vibrations

Structural Health Monitoring (SHM) in bridges is an important field, which implements increasingly innovative damage detection strategies for both theoretical developments and laboratory-scale implementations. However, is very important to consider the studies of bridges in real conditions. This study presents the damage identification of a real bridge using Hilbert-Huang transform and the most recent advance in the empirical mode decomposition method (EMD) called Improvements on Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN). This technique is used to decompose vibration data into intrinsic mode functions (IMF). Then, the Marginal Hilbert spectrum and phase difference were two damage features studied on a bridge where artificial damage was imposed. These artificial damage levels were applied sequentially and the vibration data was obtained by the passage of a recreation vehicle. As results, the ICEEMDAN method and proposed damage indicators demonstrate good performance to detect and locate damages under transient vibration loads on a real bridge.