PSI - Issue 64

Marco Pirrò et al. / Procedia Structural Integrity 64 (2024) 669–676 Author name / Structural Integrity Procedia 00 (2019) 000–000

671

3

concrete blocks, separated by vertical contraction joints (Pereira et al., 2021). Baixo Sabor dam hosts two turbines at the power plant adjacent to the structure, that rotate at 214 rpm. The full storage level of the reservoir is at 234 m asl, and the first complete filling has been reached in January 2016. A dynamic monitoring system, installed and active since December 2015, is composed by 20 uniaxial force balance accelerometers, radially installed over the upper, second and third galleries (Fig. 2). The monitoring system continuously records acceleration time series in the 20 instrumented points with a sampling rate of 50 Hz, and saves a file every 30 min. The paper presents the results based on the analysis of accelerations from the 12 sensors disposed along the crest of the dam, during the first two years of monitoring, i.e. from 01/12/2015 to 01/12/2017.

Fig. 1. Baixo Sabor arch dam (left) and power plant (right).

Fig. 2. Sensor layout of Baixo Sabor dynamic monitoring system.

3. Automated modal identification and tracking

In the two years of monitoring, i.e. from 01/12/2015 to 01/12/2017, 31326 datasets have been collected and processed: the modal parameters are automatically identified using MATLAB software package called DYMOND (Pirrò, 2021), which implements the well-known Covariance-driven Stochastic Subspace Identification (SSI-COV) algorithm (Peeters and De Roeck, 1999). In particular, an automated algorithm based on stabilization diagram (Fig. 3) is implemented for the SSI-COV modal identification (Pirrò, 2021); the adopted SSI parameters are the following: (a) the time lag needed to fill the block Toeplitz matrix was set equal to 40; (b) the minimum and maximum model order have been set equal to 20, and 120, respectively; (c) the maximum allowable damping ratio and Modal Phase Collinearity (MPC) (Pappa et al., 1993) were set equal to 5% and 0.80, respectively. From the stabilization diagram of Fig. 3, computed using the data collected on 01/12/2015 00:00 – 00:30, 6 vibration modes are identified in the range of 2–7 Hz; the estimated natural frequencies and mode shapes (Fig. 4) are used as reference values for the subsequent tracking procedure. In fact, since the SSI-COV procedure has to be automatically applied on the 2-years monitoring, the modal tracking procedure proposed in Cabboi et al. (2017) has been implemented and applied in order to track the evolution of the modal characteristics of the dam over time. The modal tracking procedure aims to automatically collect all the identified modes from each dataset, tracking only the ones that share similar modal characteristics with the reference modes (Fig. 4).