Uncertainty of extreme wind and wave loads for marine renewable energy farms in hurricane-prone regions

Journal Article

Title: Uncertainty of extreme wind and wave loads for marine renewable energy farms in hurricane-prone regions

Author:

Kresning, B.; Hashemi, M.; Shirvani, A.; Hashemi, J.

Publication Date:

January 1, 2024

Journal:

Renewable Energy

Volume:

220

Publisher:

Elsevier

Affiliation:

University of Rhode Island, Shiraz University, Swinburne University of Technology

Collection Method:

Modeling

Operations:

Condition Monitoring, Survivability

Document Access

Website:

External Link

Citation

Kresning, B.; Hashemi, M.; Shirvani, A.; Hashemi, J. (2024). Uncertainty of extreme wind and wave loads for marine renewable energy farms in hurricane-prone regions. Renewable Energy, 220 https://doi.org/10.1016/j.renene.2023.119570

Abstract

The offshore wind industry is rapidly developing in a hurricane prone area off the US East Coast. For assessment of extreme environmental conditions, or hazard intensity measures, such as wind and wave loads (e.g., 50-year or 500-year wind speed) various methodologies including univariate (e.g., Generalized Extreme Value distribution; GEV) and multivariate (e.g., Inverse First Order Reliability Method, IFORM) analysis have been recommended. Further, due to lack of long-term observed data at a site, available observed/hindcast at nearby stations are commonly used which can lead to errors. The objective of this study is to better understand and quantify the level of uncertainty in extreme value analysis of wind/wave data for marine renewable energy sites in hurricane-prone regions. An area off the northeast of the US where several large projects have been planned was selected as a case study. Univariate and bivariate analyses of wind/wave data at several stations were carried out and results were compared.

For this study, Bivariate (IFORM) and univariate (GPD) methods resulted in an average difference of 4% and 6% for extreme wind and wave data, respectively. It has been demonstrated that these discrepancies still fall within the 95% confidence intervals of the standard GEV method.