Predicting daily frequency variations on a concrete arch dam

Continuous dynamic monitoring is becoming an increasingly valuable tool for assessing the structural condition of concrete dams, providing critical insight into their dynamic behav-iour and supporting the development of early-warning strategies, reinforcing the importance of monitoring natural frequencies as indicators of potential stiffness changes. However, natural frequencies are not only dependent on structural integrity, being also in-fluenced by operational and environmental conditions, which can mask subtle variations due to the occurrence of damage. For instance, in the context of dam monitoring, temperature and hydrostatic pressure due to the reservoir both play a significant role in the dynamic be-haviour of these structures, not only on seasonal cycles but also daily. Therefore, properly accounting for these influences is essential to obtain damage-sensitive features and thus, modelling tools capable of minimizing or isolating these effects play a crucial role in relia-ble condition assessment. The present work focuses on the development of predictive models aimed at capturing the evolution of natural frequencies on both seasonal and short-term timescales. Beyond charac-terizing long-period trends, particular attention is given to daily frequency variations, whose accurate prediction is fundamental for interpreting high-resolution monitoring data and en-hancing the detectability of abnormal structural behaviour in concrete arch dams.