Robust output-only modal identification and monitoring of buildings in the presence of dynamic interactions for rapid post-earthquake emergency management

A common assumption in the theoretical background of Operational Modal Analysis (OMA) methods concerns the input, which is not measured and consists of a Gaussian white noise characterized by a flat spectrum in the bandwidth of interest. Even if a number of applications have shown that OMA works also when the input spectrum is smooth enough, if this assumption is not fulfilled, the modal identification process may become more complex or even fail due to dominant frequency components of the input appearing in response spectra. Thus, specific techniques and strategies have to be applied to sort out structural dynamic properties from frequencies of the input. It is often possible to identify such non-structural frequencies in the case of spurious harmonics due to rotating equipment, but they can be erroneously identified as structural or even bias the estimates, for instance when such harmonics are close to structural modes. About civil structures, spurious frequencies can be due to rotating machines but also to interaction with adjacent constructions. In the present paper, attention is mainly focused on the issue of modes from surrounding structures appearing in the response of the structure under investigation. Thus, natural frequencies of a sample structure are mixed together with those due to interaction with similar surrounding structures, and its response spectra are characterized by several close peaks in a narrow bandwidth. In order to discriminate modal from spurious frequencies, an integrated use of output-only techniques is illustrated and applied to a real test case. An approach aiming at robust evaluation and monitoring of dynamic properties for fast post-earthquake emergency support is proposed and the experimental field validation is discussed in detail.