The seismic assessment of existing concrete gravity dams: FE model uncertainty quantification and reduction.

The implementation of resilience-enhancing strategies on existing concrete gravity dams is a task of primary importance for the society. This aim can be achieved by estimating the risk of concrete dams against multi-hazards and by improving the structural control. Focusing the attention only on the seismic hazard, numerical models assume great importance due to the lack of case studies. However, for the same reason, numerical models are characterised by a high level of uncertainty which must be reduced by exploiting all available information. In this way reliable predictive models of the structural behaviour can be built, thus improving the seismic fragility estimation and the dam control. In this context, the observations recorded by the monitoring systems are a powerful source of information. In this thesis two Bayesian frameworks for Structural Health Monitoring (SHM) of existing concrete gravity dams are proposed. On the one hand, the first proposed framework is defined for static SHM, so the dam displacements are considered as Quantity of Interest (QI). On the other hand, a dynamic SHM framework is defined by assuming the modal characteristics of the system as QI. In this second case an innovative numerical algorithm is proposed to solve the well-known mode matching problem without using the concept of system mode shapes or objective functions. Finally, a procedure based on the Optimal Bayesian Experimental Design is proposed in order to design the devices layout by optimizing the probability of damage detection. In all the three procedures the general Polynomial Chaos Expansion (gPCE) is widely used in order to strongly reduce the computational burden, thus making possible the application of the proposed procedure even without High Performance Computing (HPC). Two real large concrete gravity dams are analysed in order to show the effectiveness of the proposed procedures in the real world. In the first part of the thesis an extended literature review on the fragility assessment of concrete gravity dams and the application of SHM is presented. Afterwards, the statistical tools used for the definition of the proposed procedures are introduced. Finally, before the presentation of SHM frameworks, the main sources of uncertainties in the numerical analysis of concrete gravity dams are discussed in order to quantify their effects on the model outputs.