This thesis introduces a two-dimensional spectral finite element formulation tailored for the dynamic analysis of concrete gravity dams. While the Finite Element Method (FEM) is widely utilized for dynamic structural analysis, its application to large structures often demands substantial computational resources and time. To address this challenge, alternative modelling techniques known as Spectral Finite Element Methods (SFEMs) have been developed over the past decades to enhance computational efficiency.
The objective of this thesis is to develop a computationally efficient analysis procedure for dynamic analysis of large structures like concrete gravity dams and then apply the developed procedure for assessing the behaviour of dams subjected to seismic ground motions as well as deterioration effects like alkali-aggregate reactions. 
The Time Domain-based SFEM (TDSFEM) was chosen for dynamic time history analysis of concrete gravity dams due to its advantages over the Frequency Domain-based SFEM (FDSFEM). TDSFEM is particularly effective in handling irregular geometries and finite domains, making it a better choice for complex structural analyses. Sensitivity analyses and convergence studies were conducted using both 4-noded and 9-noded elements, revealing TDSFEM’s superior computational efficiency, especially when higher-order elements are employed.
Comparative studies between TDSFEM and conventional FEM highlighted TDSFEM's advantage in reducing computational time while maintaining accuracy in large-scale dynamic analyses. Modal analysis and time history analysis indicated that TDSFEM, when used with higher-order elements, could be a practical alternative to conventional FEM, offering substantial time savings without sacrificing precision.
The study also investigated damage detection methods based on modal parameters, including displacement, curvature, and strain energy. Among these, modal strain energy emerged as the most effective and reliable indicator for identifying and localizing damage. Additionally, TDSFEM was applied to model the behaviour and failure modes of FRP-reinforced concrete deep beams, with results closely matching experimental data, further demonstrating its efficiency.
Moreover, the thesis introduced a simplified thermo-mechanical approach for modelling deterioration effects, such as alkali-aggregate reactions (AAR), providing a novel alternative to existing chemical reaction-based models. TDSFEM is thus presented as a viable, efficient method for analysing large structures, offering significant time savings and effective applications in damage detection and deterioration modelling.