LAUSR

Abstract High-quality wind data are essential for the whole wind energy assessment process. At present, wind data obtained from numerical weather prediction models are regarded as the most promising alternative to overcome the multiple constraints of observation data. However, the associated numerical uncertainties may lead to incorrect results. New capabilities and strategies are highly required to improve the utilization of wind data obtained from numerical models and further to reduce their uncertainty. Our paper contributes to the development of an improved wind energy assessment method, which utilizes a set of wind speeds obtained from different single-valued Weather Research and Forecasting simulations to estimate the yearly wind speed distribution and power generation. First, the wind data were generated by running a set of configured numerical models. Second, the wind speed series were divided into segments (i.e., “waves”), which were clustered into several groups by the fuzzy C -means clustering. Third, we applied a Cuckoo search optimized induced ordered weighted average method, induced by the gray relationship to each wave group. The proposed method reduced the numerical uncertainties and had superior performance compared to other tested models. In this study, we demonstrated an improvement in the utilization of data obtained from numerical simulations and constructed a practical tool for real wind applications.