Knowledge of the dynamic behavior of complex buildings subjected to near‐fault earthquakes may be enriched by valuable information obtained through rapid onsite dynamic testing to aid in the design of appropriate retrofitting interventions. Through a case study, the article demonstrates the enhancement in comprehension of the structural behavior obtained by means of a 2‐day testing campaign conducted on a complex building of the Engineering Faculty, Edifice A, which was heavily damaged in the 2009 L'Aquila earthquake. The onsite testing was carried out with a network of 13 accelerometers opportunely located to identify the dynamic characteristics of the structure by means of ambient noise‐induced vibration. Enhanced Frequency Domain Decomposition (EFDD) and Stochastic Subspace Identification (SSI) output‐only procedures were both used to identify the main modal parameters of two substructures of the building. The modal characteristics were used to update a finite element model representing small amplitude vibrations of the damaged structures. The direct comparison of the identified modal features with finite element models, in which damaged member locations are determined by onsite visual observations, has permitted identification of a model representative of the structural behavior of the building in the immediate postearthquake conditions.

The entire structural identification process, starting from measurement of the experimental response to ambient vibration, and concluding with the assessment of a representative dynamic model, depends upon an informed selection of appropriate data treatment techniques. The paper compares different procedures for identifying both modal and physical models. Four approaches (EFFD, ERA, SSI, and TFIE) for output-only modal identification are discussed. Their performance is evaluated using experimental data extracted from the ambient vibration response of a three-dimensional frame. The discussion of the identification results throws light on the intrinsic characteristics of each procedure, while the particular features of the structure, presenting unexpected local modes, demonstrate the necessity of a hierarchical treatment of the available information, based on a solid engineering knowledge. The calibration of a physical model is approached using an exact inverse procedure, based on a parametric analytical model, or automatic techniques of error minimization, based on finite element multi-models. The achievement of a good relative agreement, measured through performance-based indexes, increases the general confidence on the results independently obtained from each modal identification and model updating method. Copyright © 2011 John Wiley & Sons, Ltd.