No abstract

Understanding of the stall delay phenomenon on wind turbines remains, to this day, incomplete. A correct modelling of this phenomenon, which results from three-dimensional rotational effects, is essential in order to make reliable wind turbine simulations on the basis of two-dimensional airfoil data, such as with the widely used blade element momentum method. The present study addresses this issue by testing six existing models intended to correct for stall delay effects, namely those developed by Snel et al., Chaviaropoulos and Hansen, Raj, Bak et al., Corrigan and Schillings and Lindenburg. For this purpose, the models are implemented into a lifting-line-prescribed wake vortex scheme. Forces along the blades as well as power and root flap bending moment in a head-on flow configuration are predicted based on these models, and are compared to wind tunnel data from NREL's phase VI experiment. While load over-prediction in the presence of stall is in general observed from the use of the different models, significant differences between the models are still seen. Local over-prediction is generally seen in the tip region, while discrepancies are obtained, even at low wind speed, for the root flap bending moment. The results obtained are discussed in terms of deficiencies and strengths of the current correction schemes, and from there a basis is provided for the development of improved correction models.lift coeffi cients compared to the corresponding two-dimensional (2-D) case, and by a delay of the occurrence of fl ow separation to higher angles of attack.Ever since this phenomenon was fi rst observed by Himmelskamp on propeller blades, 3 it has captured the attention of many researchers in the helicopter and wind turbine fi elds, where computational, theoretical and experimental investigations have been performed. Among others, McCroskey, 4 and Savino and Nyland, 5 have performed fl ow vizualizations on rotating blades, investigating the separated fl ow along the blade. Milborrow 6 performed an analytical evaluation of existing experimental data to try and understand the mechanisms responsible for the stall delay phenomenon. Madsen and Christensen, 7 as well as Ronsten, 8 performed pressure measurements on rotating and non-rotating blades in order to quantify the importance of three-dimensional (3-D) fl ow effects. Sørensen et al.,9 in addition to Narramore and Vermeland, 10 used CFD computations to provide insight into 3-D rotational effects. Dumitrescu and Cardos 11 investigated the 3-D effects on the laminar boundary layer, trying to shed light on the stall delay phenomenon. Very recently, Schreck et al. 12 used CFD computations coupled with experimental data from the NREL phase VI experiment 13 to characterize the aerodynamic features in the rotating blade boundary layer, as well as to deduce mechanisms responsible for the rotational augmentation.The stall delay phenomenon is, however, still far from being completely understood. An illustration of this comes from the blind comparison exercise 14 that followed ...

No abstract

This paper reports an extensive program of forced and free vibration tests on a single circular cylinder moving mainly perpendicularly to a uniform current. For both free and forced vibration tests, two cases were investigated: one in which the cylinder was restrained in the in-line direction and the other in which it was supported on suitable springs. The cross-flow vibrational response and hydrodynamic forces on the cylinder were measured. Large variations of motion frequency in the “lock-in” range were found from the free vibration tests. This leads to two different definitions of reduced velocity, namely, a so-called nominal reduced velocity based on one reference frequency and the true reduced velocity based on the actual vibration frequency. When different results are compared, the true reduced velocity should be used. The forced vibration tests showed, as may be expected, that the transverse force in the “lock-in” range on the average will add energy to the cylinder at moderate motion amplitudes and subtract energy at large amplitudes. Some conditions resulting in a steady-state vibration of a flexibly mounted cylinder were analyzed. The actual force traces also show very large and apparently random deviations from the average force amplitude. The results from the forced and the free vibration tests are consistent with each other if the true reduced velocity and reduced amplitude are the same.

No abstract