Predicting Wind Speed for Wind Energy; Progress of the WINDENG Project

摘要

The suitability of the computer model MM5 for predicting wind speed, and hence wind energy, is investigated by performing simulations for different geographical regions. The focus is on wind speed in the lowest 200 m of the planetary boundary layer (PBL). The dependency of the simulated wind speed on PBL parameterization and atmospheric stability is studied. The smallest deviation between measured and simulated wind speed, averaged over a three-day period, is 1% and occurs for an off-shore simulation with unstable stratification. The largest deviations of 31% and 20% occur with orographically structured terrain, stable stratification and weak synoptic forcing. The results suggest that unstable conditions are simulated with better accuracy by MM5. Changes of the PBL scheme cause wind speed variations between 9% and 40% of the average wind speed. None of the PBL schemes is clearly the best and their performance can strongly vary for different conditions. Nevertheless, the Mellor-Yamada-Janjic (ETA) and the Blackadar PBL parameterization (BLK) schemes seem to be the most suitable schemes for wind energy applications.Additionally, MM5 was successfully adapted for idealised, stationary simulations in order to calculate a wind-climatology for Sardinia using a statistical-dynamical downscaling approach.