Pre-averaging is often applied to wind turbine test data to improve correlation between wind speed and power output data. In the past, trial and error or intuition have been used in the selection of pre-averaging time and researchers and institutions have differed widely in their pre-averaging practice. In this paper a standardized method is proposed for selection of the optimum pre-averaging time. The method selects an averaging time such that the test data are low-pass-filtered at the same frequency as the response frequency of the test wind turbine/anemometer system. A theoretial method is provided for estimation of the wind system transfer function as a function of the anemometer location, rotor moment of inertia, the stiffness of the connection between the rotor and the electrical grid, hub height, rotor speed and wind speed. The method is based in proven theory, repeatable, easy to use and applicable to a wide range of wind turbines and test conditions. Results of the transfer function predictions are compared with the measured response of two wind systems. Agreement between the predicted and measured response is completely adequate for the purposes of the method. Example results of calculated averaging times are presented for several wind turbines. In addition, a case study is used to demonstrate the dramatic effects of test design and data analysis methods on the results of a power coefficient measurement.

This content is only available via PDF.
You do not currently have access to this content.