Extreme event risk assessment for offshore systems design and operation in harsh environments

Arif, Mohammad Hizbul Bahar (2023) Extreme event risk assessment for offshore systems design and operation in harsh environments. Doctoral (PhD) thesis, Memorial University of Newfoundland.

[English] PDF - Accepted Version Available under License - The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission. Download (3MB)

Abstract

Operations of the offshore systems in harsh environments require better understanding, precise assessment, and effective management of risks. The harsh environmental conditions, such as strong ocean currents, extreme wave conditions, complex subsurface geology, frigid temperatures, and icebergs, exert extreme load on the offshore systems. Environmental factors are interconnected, and when they occur at a higher rate or in extreme conditions, they are likely to cause a catastrophic event. Such scenarios are prone to occur in the current changing conditions of climate. Assessment of extreme loads that may cause a rare event situation is critical to define risk scenarios. This study focuses on the assessment of these extreme event risk scenarios. By integrating extreme load and its likelihood of occurring, this research investigates the current state of knowledge in extreme event risk analysis. The extreme load consideration task considers three dominating aspects: stationary and non-stationary conditions; univariate and multivariate analysis; and dependence of the variables. This study also focuses on the flexible risk-based design methodology that integrates the traditional Extreme Value Theory (EVT) with climate change. The key environmental parameters considered in this study are iceberg speed, wind speed, and wave height. The developed methodologies use the above parameters from the Atlantic Continental Shelf, specifically the Flemish Pass basin, Grand Bank, and the Jeanne d’Arc basin. Due to limited data for certain environmental phenomena, such as large iceberg data in the Flemish Pass basin, the iceberg load assessment problem is treated as a rare event scenario. Traditional methods, including Peak Over Threshold (POT) based Generalized Pareto Distribution (GPD) and Block Maxima (BM) based Generalized Extreme Value (GEV), were found to be inadequate to capture the present-day extreme characteristics in the rare event cases. As an alternative, this study proposes and validates the use of POT-based Heavy Right Tail Distribution (HRTD) for iceberg load analysis at the Flemish Pass basin. The research also observes that Maximum Likelihood Estimator (MLE) provides a biased estimate for model parameter estimation in rare event scenarios, whereas the Hill, SmooHill, and Bayesian approaches offer better estimates. The methodology is extended to multivariate settings to capture extreme dependencies using extreme value copula function for investigating rare event risk profiles. The proposed low-resolution risk profile methodology offers a more efficient and cost-effective alternative to computationally expensive numerical models in the offshore domain. Climate change is observed to have an impact on the correlation between various environmental factors, including wind speed and wave height. Because of climate change, 100-year events are becoming more frequent. Consequently, the study adopts a 1000-year time frame to adjust for the increasing frequency of 100-year events under the influence of climate change, enabling predictions beyond standard lifetimes. The conditional return level function is utilized to construct rare events return level predictions under climate change threats. Finally, a non-stationary process is considered to generate a dynamic risk profile. Outcomes of this research provide a clear understanding of how climate change affects the Newfoundland offshore region. By incorporating predicted extreme loads and their likelihood of occurring, the traditional EVT-based methodologies are combined with adaptable risk-based design methodologies. The proposed dynamic, flexible, and small-scale (0.10×0.10 latitude/longitude grid) risk assessment methodology aids in offshore design decision-making for safer design and operation.

Actions (login required)

View Item