SELECTING MAINTENANCE MONTHS FOR AN OFFSHORE WIND FARM USING A WAVE HINDCAST AND HIGH SURF WARNING DATABASE

Abstract

 The selection of months for offshore wind farm maintenance was discussed by using high surf advisory or warning information. High surf warnings and advisories from 2013 to 2022 were obtained from Disaster Prevention DB, and information whether warnings or advisories were announced every six hours (advisory information) was calculated at six locations. Significant wave heights from 2010 to 2020 by WAVEWATCH III were used to categorize an information comparable to the advisory information. Classification approaches by threshold (Ex1) and by using neural network (Ex2) were investigated to show that the accuracies by both approaches were about 0.9, except for one site. Monthly averages were calculated for the number of the daily advisories or warnings from the Disaster Prevention DB, Ex1 and Ex2. Some areas showed distinct seasonal variations in the monthly averages, suggesting that the cost advantage is greater when the monthly averages are smaller.