Reliability-based design of offshore structures for oil and gas applications

Okoro, Aghatise (2023) Reliability-based design of offshore structures for oil and gas applications. Doctoral (PhD) thesis, Memorial University of Newfoundland.

[img] [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 (4MB)


Offshore structures are complex in their structural and functional form and operate in a harsh and uncertain environment with complex interactions between ocean variables. Consequently, the ocean environment presents a high risk to these structures hence the need to develop an efficient and reliable design. Therefore, the need for a design that effectively: captures complex ocean parameter interactions, reduces the computational burden in structural response determination, quantifies the structure's ability to bounce back when faced with disruptive events, and minimizes cost under uncertainty at the desired safety levels of the asset is critical. A robust offshore structural design under uncertainty is essential for the safety of life, asset, and the environment during oil and gas exploration and production activities. This thesis presents improved methods for the effective reliability-based design of offshore structures. First, a framework is developed to capture the dependency of multivariate environmental ocean variables using vine copula and its impact on the reliability assessment of offshore structural systems. The model was tested using a cantilever beam and applied to an offshore jacket structure. The comparative results from the jacket structure and cantilever problem reveals that failure probability considering dependence between ocean variables is closer to the reference value than when variables are independent or modeled with a Gaussian copula. The outcome shows the importance of capturing nonlinearity and tail dependence between ocean variables in reliability evaluation. Secondly, the effectiveness of a hybrid metamodel, which is a combination of two commonly and independently used methods, Kriging and Polynomial Chaos Expansions (PCE), is investigated for offshore structural response determination and reliability studies. The hybrid metamodel herein, called (APCKKm-MCS) is constructed from an adaptive process with multiple enrichment of Experimental Design (ED). The hybrid approach was tested on simple non-linear functions, a truss bar, and an offshore deepwater Steel Catenary Riser (SCR). The study's outcome revealed that APCKKm-MCS produced a high predictive response capacity, reduced model evaluation, and shorter computing time during reliability evaluation than the single enrichment case (APCK-MCS) and the adaptive ordinary Kriging case (AK-MCS) considered. In addition, a novel framework is developed for the resilience quantification of offshore structures in terms of their time-varying reliability, adaptability, and maintainability. The developed framework was demonstrated using an internally corroded pipeline segment subject to disruptive events of leak, burst, and rupture. The framework captured the resilience index of the natural gas pipeline for its design life, and its sensitivity analysis revealed the influence of the pipe wall thickness and corrosion depth growth rate on the resilience of the pipeline. The framework provides a quantitative approach to determine the resilience of offshore structures and ascertain their critical influencing parameters for effective decision-making. Finally, a methodology for optimal structural design under uncertainty considering the dependency of environmental variables with the implementation of a hybrid metamodel in the inner loop of a nested optimization problem is presented and demonstrated on a steel column function and a segmented SCR. The study showed different decision outcomes for various vine tree configurations in the dependence modeling for the steel column function noting the importance of choosing the appropriate variable order in the vine tree for optimal design under uncertainty. Also, the research reveals the suitability of adaptive PCK for the inner loop reliability phase for a double-loop structural optimization due to its high predictive capacity and observed relatively low cross-validation error. The method shows the importance of effective dependence modeling of environmental ocean variables in structural cost minimization and selecting optimal structural design variables under uncertainty. From the research outcomes, considering multivariate dependence between ocean variables using vine copula and utilizing multiple enrichment hybrid metamodels in response evaluation for reliability and optimal design assessment of offshore structures could better predict their failure probability and enhance a safer structural design. In addition, the resilience quantification framework developed provides a vital decision-making tool for offshore structural systems' design and integrity management. The research into high dimensional dependence modeling of offshore structures using vine copula, comparative study of sampling strategies required for the hybrid (Kriging and PCE) metamodel construction, dependence-based structural resilience quantification, and multiobjective dependence-based structural optimization under uncertainty are among areas proposed for future investigation.

Item Type: Thesis (Doctoral (PhD))
Item ID: 15821
Additional Information: Includes bibliographical references
Keywords: structural safety, reliability, resilience, structural optimization, offshore structures, dependence modeling, copula functions
Department(s): Engineering and Applied Science, Faculty of
Date: February 2023
Date Type: Submission
Digital Object Identifier (DOI):
Library of Congress Subject Heading: Offshore structures--Design and construction ; Copulas (Mathematical statistics); Dependence (Statistics)

Actions (login required)

View Item View Item


Downloads per month over the past year

View more statistics