Understanding the effects of free-stream turbulence (FST) and surface roughness on the flow around wind turbine blades is imperative in the quest for higher wind turbine efficiency, specially under stall conditions. While many investigations have focused on the aerodynamic loads on wind turbine airfoils, there are no studies that examine the effects of free-stream turbulence and surface roughness on the velocity field around a wind turbine airfoil. Hence, the aim of this investigation is to study the influence of high levels of FST on the flow around smooth and rough surfaces with pressure gradients. Moreover, of great importance in this study is the examination of how the length scales of turbulence and surface roughness interact in the flow over wind turbine airfoils to affect flow separation.
Particle Image Velocimetry measurements were performed to analyze the overall flow around a S809 wind turbine blade. Results indicate that when the flow is fully attached, free-stream turbulence significantly decreases aerodynamic efficiency by 82%, yielding to higher loads and fatigue on the blades. On the contrary, when the flow is separated, the effect is reversed and aerodynamic performance is slightly improved (i.e., by 5%) by the presence of the free-stream turbulence. Analysis of the mean flow over the suction surface shows that, under stall conditions, free-stream turbulence delays separation, and surface roughness advances separation. Interestingly, the highly non-linear interaction between free-stream turbulence and surface roughness results in the further advancement of separation.
Of particular interest is the study of the region closer to the wall (i.e., the boundary layer), where the flow interacts with both the surface of the blade and the free-stream. Turbulent boundary layer experiments subject to an external favorable pressure gradient (FPG) were performed to study the influence of FST, surface roughness and external pressure gradient (present around the wind turbine blade as a consequence of its geometry) on the behavior of turbulent boundary layers and to identify and quantify the length scales that are affected by these external conditions. Laser Doppler and hot-wire anemometry measurements, for smooth and rough surfaces, confirmed that FST and FPG cause a reduction in the wake of the boundary layer. Moreover, results show a discrepancy in the behavior of the stream-wise and wall-normal variances due to free-stream turbulence. As a result, the addition of FST increases the anisotropy in the body of the boundary layer. For FPG flows, a budget analysis of the Reynolds stresses shows that turbulent transport and pressure strain terms are responsible for the increase in the stream-wise Reynolds stress component when FST is present. Second-order structure functions and energy spectra are examined to identify and quantify which turbulence length-scales contribute mostly to the increased anisotropy, and to compare these effects to the case of a zero pressure gradient (ZPG) boundary layer. For ZPG flows, it is shown that the anisotropy created by adding nearly isotropic turbulence in the free-stream resides mostly in the larger scales of the flow, in a range between r/δ95 = 3 and 10. With an imposed FPG, the effect of FST resides in the very-largest length scales of the flow, r ≥ 4.3δ95, corresponding to scales of the same size, and even larger, than the integral scale of the outer free-stream turbulence. However, the free-stream turbulence is not increasing the anisotropy to the extent that it did for the ZPG case.
The effects of surface roughness on the different length scales of the flow, when a FPG and additional levels of FST are present, are also examined. Second-order structure functions and energy spectra analysis suggests that for highly turbulent favorable pressure gradient flows, the effect of roughness at the surface is felt, not only by the small length scales of the flow, but also by large (e.g. r ≈ 0.55δ95) and, even, very-large scales (r ≤ 11δ95). This observation is even more evident when additional levels of turbulence are present in the free-stream. Again, the complex dynamics between the length scales of the flow, observed in the study of the flow over the S809 wind turbine blade, are confirmed here.
In general, all of the external conditions analyzed have an impact on the largest and more energized scales of the flow. Therefore, free-stream turbulence and surface roughness should be taken into account when modeling the flow over wind turbine blades. This will certainly reduce the uncertainties and inaccuracies that currently result from the modeling, specially under separated flow conditions, due to the exclusion of free-stream turbulence from the analysis.