Aboveground storage tanks (ASTs) have suffered severe damage during past storm events, resulting in the release of hazardous chemicals in the environment. For instance, more than 30 million liters of oil products were spilled during Hurricane Katrina and Rita in 2005. Despite the evident vulnerability of ASTs, the literature is currently lacking comprehensive studies evaluating the performance of ASTs during multi-hazard storm events. To address this gap, this thesis offers new methods, tools, models, and frameworks to assess the structural behavior and structural vulnerability of ASTs subjected to multi-hazard storm conditions, as well as to support risk assessment and mitigation of ASTs located in coastal regions.

First, a series of finite element models are developed to estimate storm loads on ASTs and investigate the structural behavior of ASTs under storm surge, wave, wind, debris impacts, and rainfall loads. Opportunities to reduce the computational complexity and cost of the derived numerical models are also explored using surrogate modeling techniques and assessing the validity of static analyses for dynamic phenomena. A general methodology is then posed to perform fragility assessments of ASTs under concurrent or multi-hazard storm loads using the derived numerical models, a statistical sampling method, and logistic regression. With this methodology, the first comprehensive fragility assessments are performed for (i) ASTs subjected to concurrent surge, wave, and wind loads; (ii) ASTs subjected to waterborne debris impacts; and (iii) floating roof ASTs subjected to rainwater accumulations.

This study also proposes frameworks to perform risk assessments of ASTs located in coastal regions. A first framework is developed for large-scale regional assessments of ASTs subjected to surge, wave, and wind loads. As a proof of concept, a scenario-based assessment and, for the first time, a probabilistic risk assessment are performed for a case study region, the Houston Ship Channel. Useful metrics, such as expected spill volumes and annual probabilities of failure, are obtained from the risk assessments. A second risk assessment framework is also developed to estimate the likelihood of debris impacts and damage due to such impacts for small-scale AST terminals located near known debris sources; this framework is illustrated again for a case study terminal along the Houston Ship Channel. Moving from a purely engineering perspective, the results of the risk assessments are also coupled with social vulnerability modeling to explore community impacts. Building from the risk assessment frameworks, an integrated model of built-human-natural systems is also posed to perform a comprehensive assessment of procedural, structural, and protective mitigation strategies. Since no single mitigation strategy appears optimal, a tool is also developed to optimally select and combine mitigation strategies to achieve a given performance target and propose cost-effective solutions, while considering social impacts. Finally, forensic investigations of AST failures during Hurricane Harvey are performed to highlight the viability of the derived fragility models to understand the causes and mechanisms behind AST failures and further evaluate the effectiveness of mitigation strategies.

Results obtained throughout this thesis demonstrate that the derived fragility models are efficient tools to perform rapid screening of vulnerable ASTs in industrial regions, and evaluate the viability of mitigation strategies to reduce this vulnerability. Insights obtained from the fragility and risk assessments reveal that neglecting the multi-hazard nature of storms, as existing studies have done, can lead to a significant underestimation of vulnerability and risks. Results of the assessments also indicate that small size ASTs are generally more vulnerable to loads such as wave, wind, debris impacts, and rainfall, and that floating roof ASTs do not appear to be vulnerable during rainfall events unless they are already damaged or their drain system is inefficient. Moreover, results show that simple mitigation strategies such as anchoring ASTs to the ground or filling them with liquid could greatly reduce the likelihood of AST failures and spills during storms. Lastly, this thesis illustrates how using surrogate model and statistical learning techniques can facilitate and reduce the computational complexity of fragility and risk assessments, particularly in multi-hazard settings.

Overall, this thesis provides methods, tools, and insights essential to understand, evaluate and mitigate the vulnerability of a key component of energy infrastructure and support stakeholders in doing so. Furthermore, this thesis offers as strong foundation for future vulnerability and risk assessment of other coastal structures and systems.