Sampling Wind Data for Mean Wind Speed Estimation

View More View Less
  • 1 Wind Energy Laboratory, College of Engineering, Wichita State University, Wichita, KS 67208
© Get Permissions Rent on DeepDyve
Restricted access

Abstract

Two sampling techniques are applied to wind data at 3 h intervals for six stations in the Great Plains region in the United States in order to investigate the reduction in the number of data needed to estimate the mean wind speed. One-in-k sampling of the daily averages of 0600 and 1500 LT data indicates that it is necessary to sample only 4% of the total 3 h data to attain a 95% confidence interval of 5% of the true mean value. Random sampling of all 3 h data shows that ∼3.4% of the total data is needed to obtain the same accuracy.

Abstract

Two sampling techniques are applied to wind data at 3 h intervals for six stations in the Great Plains region in the United States in order to investigate the reduction in the number of data needed to estimate the mean wind speed. One-in-k sampling of the daily averages of 0600 and 1500 LT data indicates that it is necessary to sample only 4% of the total 3 h data to attain a 95% confidence interval of 5% of the true mean value. Random sampling of all 3 h data shows that ∼3.4% of the total data is needed to obtain the same accuracy.

Save