The thesis aims to identify the most effective reliability analysis methods for offshore structural engineering applications, considering factors such as the dimensionality of the random variable space, non-linear structural responses, stochastic excitations, and the complexity of the resultant damage scenarios. By benchmarking and enhancing these techniques, the research provides engineers with practical tools for assessing reliability, identifying critical damage scenarios, and improving risk management.

Key contributions include the development of computational frameworks for largescale offshore structures, insights into the limitations and strengths of emerging methods, and strategies for enhancing computational efficiency using cloud computing environments. The findings highlight Subset Simulation as a robust method for handling structural non-linear responses and high-dimensional problems. Additionally, the study proposes improvements to MCMC algorithms involved in Subset Simulation, such as the adaptive conditional sampling algorithm, to improve robustness and efficiency. This work advances the state-of-the-art in reliability analysis. Also, it lays the foundation for future applications to other engineering domains and integration with industry practices. The results emphasize the importance of efficient and scalable methods for assuring the safety and reliability of aging offshore infrastructure.

For a copy of the thesis, please email inst.build.phd@build.aau.dk