*4.1.2. Results*

fer in transient mode exhibiting different oscillations. For large intermittency values, the oscillations are larger. Figure 18 shows that for the lift coefficients, the results from CFX dif‐

difference starts to appear near maximum lift. The k-ω SST intermittency model with γ=0.92, under predicts the lift coefficients as compared with the experimental results. The results with γ=0.94 predicts virtually identical results as compared to the OSU results. The model with γ= 0.96 predicts results that are sandwiched between the two experimental ones. Anal‐ ysis of the two figures brings us to the conclusion that the model with γ=0.94 provides re‐ sults very close to the DUT results. Therefore, we will compare the intermittency model

Figure 18 illustrates the drag and lift coefficients for different AoA using different transition‐

Figure 18 shows that the drag coefficients for the three models are very close until 180 after

models. For the lift coefficients, Figure 18 shows that the k-ω SST intermittency models pro‐

, the intermittency model does not provide good results. Hence, we conclude that the

, the γ-θ model over predicts

to 14<sup>0</sup>

for the two other

. The k-ω SST mod‐

. However, for

. For angles exceeding

. For the linear growth zone, the different results are close to each other. The

fer from 8.20

104 Advances in Wind Power

al models.

200

AoA greater than 20<sup>0</sup>

with γ=0.94 with the other transitional models.

**Figure 17.** Drag and lift coefficients for different AoA using different intermittency values

**Figure 18.** Drag and lift coefficients for different AoA using different transitional models

the experimental values whereas such phenomena appear only after 22.10

vide results closest to the experimental values for angles smaller than 14<sup>0</sup>

transitional model helps in obtaining better results for AoA smaller than 14<sup>0</sup>

, a purely turbulent model needs to be used.

which the results become clearly distinguishable. As from 20<sup>0</sup>

el under predicts the lift coefficients for angles ranging from 6<sup>0</sup>

In order to validate the quality of stall results, the latter are compared with OSU experimen‐ tal values and with Leishman-Beddoes model. Moreover, modelling of aeroelastic phenom‐ ena is computationally very demanding such that we have opted for an oscillation of 5.50 around 80 , 140 and 200 for a reduced frequency of k= <sup>ω</sup><sup>c</sup> 2U<sup>∞</sup> =0.026, where c is the length of the chord of the airfoil and U<sup>∞</sup> is the unperturbed flow velocity. From a structural point of view, the 0.457 m length profile will be submitted to an oscillation about an axis located at 25% of the chord. The results which follow illustrates the quality of our aeroelastic stall modelling at three different angles, all with a variation of 5.5 sin(w)\*t.
