GNGTS 2015 - Atti del 34° Convegno Nazionale

GNGTS 2015 S essione 2.3 253 foundation and reservoir water level (Calcina et al. , 2014). Experimental methods can be used to have a control over general dynamic parameters calculated through dynamic analysis; ambient vibration tests can save a considerable amount of money in comparison to other techniques, such as those performed with mechanical vibrodyne. In the past many ambient vibration tests have been performed on arch concrete dams, but the pertinence of such tests on dams with more complex or mixed structural behavior (as in this case) have not been validated yet. The dam analyzed in this work, in fact, presents a mixed structure: arch and gravity. The ambient vibration test was performed on 7 points on the crest and inside the three internal galleries. All the measures were referred to the one made on a bedrock outcrop at the side of the dam. Unfortunately, even though the measures were executed on a short time (4 hours) in order to avoid changing in the conditions throughout the testing period, a strong wind raised in the middle of the survey, probably influencing it. Considering the Eq. 1 proposed by Priscu et al. (1985): (1) where f is the natural fundamental frequency and H the dam height (meters), the theoretical fundamental frequency of the dam, if it was considered as an arch dam, is equal to 5.2 Hz. According to ambient vibration tests, instead, the main frequency results equal to 4.6 Hz (Fig. 1, panel c)). The difference between the theoretical and experimental frequency values can be explained considering that the structure is a mixed arch-gravity structure and that the age of the structure may have affected it. Moreover, Eq. 1 neglects the crest length and thickness and the water level of the reservoir (Calcina et al. , 2014). In Fig. 1 panel a) the deformation shape of the dam for the fundamental mode is shown. That was obtained projecting the amplification value of the main peak for every test point on the crest. In the ambient vibration analysis of dams, an important role is played by the vertical component, since the rocking effect results can be relevant. Therefore the only peak in the H/H curve that can be considered as structural behavior are the ones to which also corresponds a peak on the V/V curve. This lead to the decision of performing also an analysis of vibrational directivity. Del Gaudio et al. (2008) demonstrated as techniques based on the analysis of horizontal-to-vertical spectral ratios (HVSR) can be usefully employed in investigations of directivity properties of site response. In this work, the technique is instead applied to the structure, using the XNSR Matlab code (Mucciarelli et al. , 2013) for the detection of the azimuthal and tilt angle of structure vibration, verifying therefore the contribution of rocking phenomena. In Figs. 1f and 1g, the size of the bullets represents the amplification of the HVSR peaks, calculated with a certain combination of azimuth and tilt angle; the color, instead, represents the frequency at which the peaks occur. In Fig 1, panel f, the results are shown for the measure acquired on the bedrock outcrop. As it was supposed to be, no directionality is evident, and the frequency of the main bullets is 1 Hz. In Fig. 1 panel g, instead, the results for the measure on the crest of the dam are shown. In this case, the cluster of bullets with a major size, around an angle of 5 degrees of tilt in respect to the vertical plan and approximately 45° of azimuth angle, corresponds to the fundamental vibration mode of 4.7 Hz. Therefore, the first mode shows, predictably, a rocking phenomenon. The other colors in the graph are ascribable to higher periods, more complex to figure out. The influence of the reservoir water level cannot be ignored (Calcina et al. , 2014). The water reservoir level was almost reaching the maximum at the time of the survey. Hence, it would be important to repeat the tests with a lower water level. Oil tank. The calculation of the seismic response in the case of a tank is fairly complex. In fact, the structural period needs to be combined with the one of the liquid content. It was shown that a part of the liquid moves in long-period sloshing motion, while the rest moves rigidly with