An examination of key aerodynamic modelling issues raised by the NREL blind comparison

Abstract

AbstractIn May 2000 the US National Renewable Energy Laboratories (NREL) invited the international community to provide aerodynamic calculations for a range of test cases on its Unsteady Aerodynamics Experiment (UAE) wind turbine. Prior to this, NREL had collected test data on the turbine for these same cases in the NASA Ames wind tunnel. The calculations were conducted in the blind, i.e. without knowledge of the measured data, and subsequently compared with the data. This article describes some recent work undertaken in the aftermath of this ‘Blind Comparison’. The data set collected in the NASA Ames tunnel represents a unique opportunity for aerodynamic modellers to enhance the capability of prediction schemes by comparison with ‘clean’ aerodynamic data. In this article the sensitivity of the results predicted by the Glasgow University prescribed wake model (HAWTDAWG) to a range of parameters is examined with reference to the measured data. The sensitivity of the calculations to the input blade section aerodynamic data is demonstrated. It is also shown that, despite differences in the measured and predicted results, the inclusion of a dynamic stall model generally improves the quality of predictions at large yaw angles even when the model is not optimized for wind turbine flows. Finally, it is shown that careful selection of appropriate tower shadow velocity deficits can result in predictions that agree well with the measured data in terms of the depth and duration of the response. Copyright © 2002 John Wiley & Sons, Ltd.