Dynamic identification of a reinforced concrete structure by means of modal assurance distribution

For time-varying systems in structural engineering, such as bridges with vehicular traffic and structures during construction or seismic events, the application of traditional vibration-based identification methods may not be admissible, since the hypotheses of stationarity of the signal may be far from the real situation. Moreover, in some cases, real-time approaches to visualize the identified parameters are desirable to allow for a rapid decision-making process, which assumes the utmost importance for the early warning in structures near to excavation or demolition sites and strategic infrastructures after seismic events. Time-frequency representations have been largely used for the instantaneous identification of linear time-varying systems. However, traditional algorithms generally suffer from issues related to the identification of closely-spaced modes, which make the modal decomposition a non-trivial task. Moreover, by tracking only natural frequencies, some variations of the dynamic behavior could not be perceived, or the variations induced by other environmental and operational conditions could prevail over those due to a modification in the structural properties (e.g., due to ongoing damage). In this paper, a Decomposition Algorithm based on Modal Assurance (DAMA) is applied to separate modal responses, in order to allow a near real-time identification of modal parameters of structures with time-varying features. In particular, a Modal Assurance Distribution (MAD) obtained from the analysis of multi-variate signals is used to track the variations of modal parameters, considering both natural frequencies and modal shapes. The DAMA is applied to a full-scale structure subjected to a series of dynamic tests performed in different structural conditions. The analysis of multi-variate signals consisting of acceleration recordings collected at different locations of the structure allows the online identification of varying dynamic parameters, which can be used to assess the structural state of health.