Comprehensive fragility quantification framework for marine structures under hazards

Ships are subjected to various hazards throughout their service life. These include natural and man-made hazards such as storms, ice impact, collision, and grounding. Exposure to these hazards can cause significant damage to the structural and non-structural systems of the ship, potentially leading to adverse economic and/or environmental consequences. Quantitative assessment of the susceptibility to damage under extreme events can improve the safety and integrity of ships. Additionally, quantifying ship vulnerability against hazards informs the design and maintenance processes, leading to a more robust and resilient system. In this regard, traditional reliability analysis evaluates the structural performance under a single design event and does not quantify the performance at different intensities of the demand. Such quantification provides a broader overview of structural performance and allows for more comprehensive risk analysis. In this context, fragility can be utilized to establish comprehensive risk quantification and management approaches under various intensities of an extreme event. The research presented herein proposes a novel, comprehensive framework for quantifying the fragility of ship hulls under extreme operational conditions. High-fidelity nonlinear finite element (FE) analyses are utilized to quantify the ultimate strength of the ship hull and the degradation of the hull performance under cyclic loading arising from previous hazard exposure. Artificial neural networks (ANNs) are used to improve the computational efficiency of simulation-based probabilistic analysis. Next, the research examines the possibility of utilizing analytical functions, such as the lognormal cumulative distribution function (CDF), to model the ship fragility. The availability of such analytical models can significantly simplify the fragility quantification problem. To develop this model, probabilistic demand analyses are conducted on multiple ships within a specific category, such as tankers. Prediction models are developed to estimate key statistical descriptors of the engineering demand parameters (EDPs) (i.e., median and dispersion) directly from the operational conditions. These models aim to facilitate the computational approaches associated with demand and enable more efficient and direct fragility quantification for specific vessel types. Lastly, the dissertation proposes a novel framework for quantifying the functionality and long-term resilience of ship hulls subjected to damage. The developed approach provides a continuous measure of vessel performance in terms of functionality and resilience rather than a binary pass/fail evaluation. Functionality is measured in terms of the maximum allowable operating speed for given sea states that ensure compliance with acceptable structural performance criteria (i.e., target reliability threshold). The framework probabilistically quantifies the ultimate and residual flexural capacities of both intact and damaged hulls using Monte Carlo simulation. Reliability-based operational profiles are established to identify safe and unsafe combinations of ship speed and wave height, offering guidance for ship operators in post-damage conditions.