Vibration signature effects on damping identification of a RC bridge under ambient vibrations

Experimental tests for dynamic identification of reinforced concrete (RC) bridges by means of Operational Modal Analysis (OMA) are increasingly used in common engineering practice. Nevertheless, especially when measurements are carried out under in-service conditions (i.e. under traffic induced vibrations), some drawbacks should be carefully considered, especially in the damping-ratio quantification. As a matter of fact, the estimation is affected by several factors: i) the length of the signals, ii) the non-stationarity of the input process, and iii) the dependence on the vibration amplitude. Even if the damping ratio is a key parameter in the bridge dynamics, a major part of these aspects has not been yet fully investigated to estimate reliable values. Starting from a dynamic test program on a short-span RC bridge with half-joints in Italy, this paper investigates the issues mentioned above for the damping ratio estimation of the first two modes focusing on: i) the influence of the signal length, ii) the effects of the signals properties, and iii) their correlation with the vibration amplitude. Both the Stochastic Subspace Identification technique fed with the signal covariance (SSI-cov) and the Random Decrement technique (RD) have been used to compute the damping ratios from the collected signals. This paper shows how a convergence of the results cannot be attained by simply increasing the sample size, suggesting that the nature of the vibration itself influences the damping values. A negative, although weak, correlation between the damping ratio and the power of the signals indicates that several factors play a crucial role in the damping estimation in the short span RC bridges with half-joints.