|Title||Statistical estimation of extreme loads for the design of offshore wind turbines|
|Publication Type||Journal Article|
|Year of Publication||2015|
|Authors||Stewart GM, Lackner MA, Arwade SR, Hallowell S, Myers AT|
The International Electrotechnical Commission (IEC) design standard (IEC 61400-3) for offshore wind turbines includes a design load case which considers loads on the turbine during extreme conditions, when the wind turbine is not operating and the blades are feathered. The recommendation of this design standard is to simulate 6 one-hour periods, each with wind and wave fields modeled as random processes, and to calculate design loads as the mean of the maximum from the six simulations. Previous studies have investigated this recommendation for fixed bottom offshore wind turbines and raised concerns about the stability of the estimate of the design load calculated using these guidelines. The research presented in this paper calculates statistics of extreme design loads as a function of the number of one-hour simulations considered for three offshore wind turbine support structures: a fixed-bottom turbine supported by a monopile and two floating turbines, a spar buoy and a semi-submersible. The study considers one extreme design load, the moment at the tower base, one set of 50-year metocean conditions representive of the Northeast U.S. Atlantic coast, and five combinations of wind and wave models, including linear and nonlinear representations of irregular waves. The data generated from this study are used to assess the stability of the estimate of the extreme load for various numbers of one-hour simulations for each turbine type. The study shows that, for the considered metocean conditions and models, the monopile exhibits the least stability in the estimate of the extreme load and therefore requires more one-hour simulations to have comparable stability as the two floating support structures. Explanations for this result are provided along with a probabilistic formulation for estimating variability of design loads more efficiently, using fewer one-hour simulations.