Near real-time tracking of dynamic properties for standalone structural health monitoring systems

Automated modal parameter identification of civil engineering structures has been analyzed in a previous paper. An original algorithm, named LEONIDA, working in frequency domain, has been presented and a number of test cases have been discussed in order to point out advantages and drawbacks. It has been demonstrated that LEONIDA represents a promising and reliable tool, in particular for modal testing. Conversely, integration of such a procedure into a fully automated structural health monitoring (SHM) system has shown that it can be used as modal information engine, but length of record durations, amount of computational burden and response time lead to recognize that serious drawbacks and limitations exist for a class of applications, such as continuous monitoring of structures in seismically prone areas. In fact, a fast assessment of relevant structure health conditions in the early post-earthquake phase is becoming of interest in different European areas. In such a context, the statistical treatment of measured dynamic properties could be certainly useful, but it requires the collection of an extensive amount of local and global data in a short time. As a consequence, availability of reliable, robust and fairly fast data processing procedures for modal tracking is fundamental whenever really effective and useful SHM systems are adopted to support civil protection activities during seismic sequences. This applies mainly to strategic structures, whose health conditions must be rapidly assessed after any seismic event, in order to securely manage rescue operations. In the present paper, the main issues related to a fast, robust and reliable modal tracking for emergency management are outlined. Then, an automated modal tracking strategy for SHM applications in earthquake prone regions is described. It is based on the knowledge of the experimental mode shapes and a revised concept of spatial filtering. Results of sample applications of the proposed procedure refer to simulated data and to real measurements collected by a SHM system. The latter are representative of operational conditions and of the transient response due to the ground motion induced by the recent L'Aquila earthquake mainshock. Discussion of results will point out advantages and limitations of the data processing strategy. (C) 2011 Elsevier Ltd. All rights reserved.