GNGTS 2015 - Atti del 34° Convegno Nazionale

252 GNGTS 2015 S essione 2.3 Ambient vibration tests on energy production and distribution facilities: three case studies G. Massolino, D. Sandron, A. Rebez, M. Mucciarelli OGS - Istituto Nazionale di Oceanografia e di Geofisica Sperimentale, Trieste, Italy Introduction. In this study the ambient noise tests performed on three different kinds of energy related industrial structures are described: a concrete dam, an oil tank and a hangar. Since the failure of these structures could induce serious damage to people, property and the environment, it is fundamental to perform their dynamic characterization and operational monitoring. In fact, boundary conditions, geometrical and material properties may be different from the design characteristics, due to structure exposure or even to construction mistakes. The modal tests that can be performed on a structure for the determination of dynamic parameters, such as natural frequencies and damping, are of two kinds: forced vibration or ambient vibration. The former is very expensive and complicated, especially on big constructions, but (being the input known) has the advantage of allowing the determination of force response relationships, leading to more accurate modal proprieties definition (Bukenya et al. , 2012). The latter, on the contrary, is cost-effective, non-destructive, quick and easy to perform, since it doesn’t require any artificial source. Between these non-invasive techniques for damage detection and monitoring of structural performance of existing structures, there are the “Operational Modal Analysis” (OMA) or “Noise Input Modal Analysis” (NIMA), consisting in the estimation of natural frequencies, mode shapes and damping ratios using only the ambient noise as source input (Sevim et al. , 2013). Since the vibration amplitudes are small (Ivanović et al. , 2000), ambient vibration tests can describe only the linear behavior of structures. However, it can be an effective way to have a high degree of confidence that the structure’s model used is a realistic representation of the physical structure (Ventura et al. , 2005). Therefore, in this work the test results were compared, when possible, to theoretical formulations and Finite Element Models (FEM). Usually, ambient vibration tests are applied to the determination of dynamic characteristics of existing civil buildings but it can be a powerful tool also for industrial facilities related to the production and distribution of energy. In fact, in this case, it is fundamental to avoid interference on the operability. Moreover, a crucial application of the technique is the possibility to make a comparative analysis between the modal frequencies under different load conditions and before and after an earthquake. As a consequence, if a large number of strategical structures were characterized in this way, immediately after the occurrence of an earthquake it could be possible to estimate the damage entity and plan a safety intervention. Ambient vibration can be used also to detect asymmetries or rotation plans in structures. In fact, in this work it is shown how it is possible to determine the azimuthal and tilt angle of the rocking rotational direction for the dam, or the stiffness centre and torsional component for the hangar. Unfortunately, it won’t be possible to give detailed characteristics of the analyzed structures because of industrial confidentiality. Methods. The ambient noise surveys have been performed with tri-axial digital high resolution tromograph Tromino, Micromed ® . The horizontal axes were always orientated in accordance to the structural axes: in the direction of the curvature radius for the dam, tangential to the border for the tank and parallel to the main planar dimension for the hangar. Due to limited availability of instruments (3 Tromino stations), measures were not synchronized, although all the measures have been referred to an ambient noise measure on bedrock or free-field. Ambient noise was recorded for 12 minutes on the structures and 30 minutes on free-field soil, at a sampling rate of 128 Hz. In the analysis, the smoothing was fixed to 3% for the measures on structures, 10% for those on free-field soil. Results. Dam. The estimation of the actual dynamic response to loads in the case of dams is quite complicated, as it depends on several factors, such as design input, interaction with the