|Title||Probabilistic Modeling of Joint Hurricane-induced Wind and Wave Hazards to Offshore Wind Farms on the Atlantic Coast|
|Publication Type||Conference Proceedings|
|Year of Publication||2013|
|Authors||Valamanesh V, Myers AT, Hajjar JF, Arwade SR|
|Conference Name||ICOSSAR 2013: 11th International Conference on Structural Safety & Reliability|
|Publisher||Taylor & Francis Group, London|
|Conference Location||New York, NY|
Offshore wind turbines are designed per the IEC 61400-3 standard to withstand dozens of load cases involving combinations of wind and wave-induced load effects. Three of these load cases consider extreme loading wherein a turbine with a parked rotor and feathered blades is expected to sustain combined wind and wave loads corresponding to a 50-year mean return period. For most locations along the Atlantic coast, this 50-year combination of wind and wave will be influenced by hurricanes. During offshore hurricanes and other storms, a multi-hour lag between the maximum winds and the maximum waves is typical and this has important implications for the development of a rational method to estimate the joint wind and wave hazard at a particular return period. The lag causes significant differences in the magnitudes of the wind and wave depending on whether the hazards are assessed independently, as is currently recommended by existing guidelines, or jointly. In this paper, we introduce a procedure based on multivariate annual extreme value distributions, the Nataf model and a joint exceedance condition to estimate contours of wind and wave with a constant mean return period. Using 24 years of wind and wave measurements from a NOAA buoy off the Atlantic coast of Florida, we present numerical results to assess the impact of two methods for selecting the joint annual maxima of wind and wave. In the first method, the annual maximum for the hourly wind speed is paired with the simultaneous value of the significant wave height and, in the second method, the annual maximum for the significant wave height is paired with the simultaneous value of the hourly wind speed. The contours of constant mean return period resulting from each method are significantly different. The 24 year period at this particular station includes measurements from five Atlantic hurricanes.