Output-only system identification of a large post-tensioned concrete bridge using hybrid excitation was conducted. To boost ambient excitations, dynamic forcing applied by two light-weight electrodynamic shakers was adopted for hybrid testing. The bridge modal characteristics were identified using three output-only identification methods: auto-regressive time series model, eigensystem realization algorithm with observer/Kalman identification and data-driven stochastic subspace identification method. The accuracy and efficacy of the system identification algorithms when applied to hybrid dynamic responses were investigated and benchmarked against the results obtained from pure ambient vibration testing. More modes were identified from the hybrid excitation data compared to using only ambient vibrations. The study demonstrated that using output-only identification algorithms on vibration responses produced by hybrid excitations can be used for extracting more accurate and reliable modal parameters with enhanced accuracy and reliability for large-scale structures due to enhanced excitation energy and better coverage of relevant frequency bands.
Read full abstract