Predictions of the cycle-to-cycle aerodynamic loads on a yawed wind turbine blade under stalled conditions using a 3D empirical stochastic model

Abstract

This paper investigates a new approach to model the stochastic variations in the

aerodynamic loads on yawed wind turbines experienced at high angles of attack. The method

applies the one-dimensional Langevin equation in conjunction with known mean and standard

deviation values for the lift and drag data. The method is validated using the experimental data

from the NREL Phase VI rotor in which the mean and standard deviation values for the lift and

drag are derived through the combined use of blade pressure measurements and a free-wake

vortex model. Given that direct blade pressure measurements are used, 3D flow effects arising

from the co-existence of dynamic stall and stall delay are taken into account. The model is an

important step towards verification of several assumptions characterized as the estimated

standard deviation, Gaussian white noise of the data and the estimated drift and diffusion

coefficients of the Langevin equation. The results using the proposed assumptions lead to a

good agreement with measurements over a wide range of operating conditions. This provides

motivation to implement a general fully independent theoretical stochastic model within a rotor

aerodynamics model, such as the free-wake vortex or blade-element momentum code, whereby

the mean lift and drag coefficients can be estimated using 2D aerofoil data with correction

models for 3D dynamic stall and stall delay phenomena, while the corresponding standard

derivations are estimated through CFD.