GNGTS 2018 - 37° Convegno Nazionale

GNGTS 2018 S essione 2.2 383 uniform and describe both cases of positive correlation and cases of negative correlation. In all cases, the maximum reported frequency variations induced by temperature are of 3.5 %. Here we present the case of a residential mixed RC and steel structure, the natural frequencies of which show a clearly negative correlation with temperature in two years observation, as expected from the basic studies on the material properties. Not only is this correlation negative but it is also very large compared to what has been described in the literature so far. We found that the first flexion mode of the building shifts its frequency of 8%with less than 30°C summer- winter temperature difference and higher modes can shift even more (up to 23%, in the case studied, though generally around 10-12%). This is very important because it shows that it is not possible to infer any possible damage or to derive conclusions about the structure health, as long as the physiological behaviour under any operative conditions, including the low-temperature and the high-temperature season, has been assessed. When considering the diurnal thermal variations of the building under examination, one could hastily claim that the structure frequencies show a positive correlation with temperature, with a delay of 2-3 hours. This was the conclusion also in Guiller et al. (2016), equation 9, where they estimated 100-200 min as heat diffusion time for a 25 cm thick wall, and was indirectly the conclusion also in Clinton et al. (2006), Herak and Herak (2010); Guiller et al. (2016); Mikael et al. (2014). However, these models, that might possibly apply to a single wall, largely underestimate the law requirements (e.g. DM 26/06/2015 in Italy) and the behaviour of real structures, where several layers compose a wall (usually including also an air layer that strongly increases the heat diffusion time). The typical heat diffusion time in standard buildings has to be considered in the range of 8-12 hours. A correlation analysis on our dataset shows that the building response to the thermal fluctuations occurs in 8-9 hours and, by considering this time-lag, we find that the diurnal frequency variations of the building are also negatively correlated with temperature, contrary to what hurried first-glance interpretations could suggest. In conclusion, it is generally acknowledged that the temperature-induced variations in the modal frequencies of residential buildings can be as large as 1-3%. Here we illustrate the case of a building which shows temperature-induced frequency variations much larger than those documented so far (larger than 10% in general and as large as 23% for some higher modes). According to the theory, frequency and temperature should be negatively correlated. We suggest a possible interpretation bias that might suggest to reconsider the positive correlations claimed so far and we stress out the importance of knowing the behaviour of the structures under their full (yearly) operative range, of using always the same processing methods to compare the results and of using appropriately the number of significant digits in presenting the experimental results. All these issues play a non-neglectable role in establishing if the building natural frequencies wander over time due to physiological or non-physiological events. References Clinton J.F. Bradford S., Heaton T.H., Favela J., 2106. The observed wander of the natural frequencies in a structure, Bull. Seism. Soc. Am. , 96, 237-257. Guillier B., Chatelain J.-L., Perfettini H., Oubaiche E.H., Voisin C., Bensalem R., Machane D., Hellel M., 2016., Building frequency fluctuations from continuous monitoring of ambient vibrations and their relationship to temperature variations, Bull. Earthquake Eng. , 14, 2213–2227. Herak, M., and Herak, D., 2010. Continuous monitoring of dynamic parameters of the DGFSM building (Zagreb, Croatia), Bull. Earthquake Eng. , 8, 657–669. Liu H., Wang X., Jiao Y., 2016. Effect of temperature variation on modal frequency of reinforced concrete slab and beam in col regions, Shock and Vibration , ID 4792786, 17 pp. Mikael A., Guéguen P., Bard P.-Y., Roux P., Langlais M., 2014. The Analysis of Long-Term Frequency and Damping Wandering in Buildings Using the Random Decrement Technique, Bull. Seism. Soc. Am. , 103, 236-246.