21.04.2014 Views

these simulation numerique et modelisation de l'ecoulement autour ...

these simulation numerique et modelisation de l'ecoulement autour ...

these simulation numerique et modelisation de l'ecoulement autour ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

18 S. Men<strong>de</strong>z and F. Nicoud<br />

Figure 11. Time-averaged quantities from Run C on the centreline plane (zoom over the hole<br />

region) (a): Contours and isolines of the time-averaged pressure. (b): Contours and isolines<br />

of the time-averaged velocity magnitu<strong>de</strong> |V |. White arrows show the flow direction in both<br />

channels.<br />

is observed just downstream of the j<strong>et</strong>, near the wall (3). This separation is known to<br />

appear for relatively high blowing ratios and is responsible for a key feature of this type<br />

of flow, the entrainment phenomenon, visualised in § 4.4 and 4.5.<br />

Further in the primary main stream, the j<strong>et</strong> loses its strength by mixing with the main<br />

flow (4). Note that due to the periodic configuration, the j<strong>et</strong> that goes out of the domain<br />

reenters by the other si<strong>de</strong> (5). Figure 11 strongly suggests that the shape of the micro-j<strong>et</strong><br />

(1–4) is influenced by the aspiration si<strong>de</strong> and that computing only the injection si<strong>de</strong><br />

would be questionable.<br />

4.2. Flow on the suction si<strong>de</strong><br />

Figure 12 presents the structure of the flow on the suction si<strong>de</strong> by displaying, in a horizontal<br />

plane located 0.5 d un<strong>de</strong>r the suction wall, contours and isolines of the three components<br />

of the time-averaged velocity in figures 12(a–c) and contours of the Q criterion<br />

(Hunt, Wray & Moin 1988) calculated from the time-averaged velocity in figure 12(d).<br />

The acceleration of fluid entering the hole can be seen in figure 12(b), on the timeaveraged<br />

vertical velocity field, which is very inhomogeneous. The spatial-averaged vertical<br />

velocity over the horizontal plane is 0.02 V j but locally in the cutting plane, it reaches<br />

0.3 V j . Note also that the maximum of the vertical velocity is not centred un<strong>de</strong>r the<br />

hole inl<strong>et</strong> but is located downstream of the centre. This can be related to the pressure<br />

gradients observed in figure 11(a): the maximum pressure variations are observed at the<br />

downstream edge of the hole inl<strong>et</strong>.<br />

The suction through the hole influences the three components of the velocity: figure<br />

12(a) shows its effect on the streamwise velocity: un<strong>de</strong>r the upstream edge of the<br />

hole, the aspiration induces a small acceleration and un<strong>de</strong>r the downstream edge of the<br />

hole, a <strong>de</strong>celeration. Near the plate, negative values of the streamwise velocity are even<br />

observed, showing that the fluid turns back to enter the hole. The aspiration makes the<br />

fluid come from all si<strong>de</strong>s of the hole, as observed on the time-averaged spanwise velocity<br />

field (figure 12c), which shows how the fluid comes from lateral si<strong>de</strong>s. The streamwise<br />

velocity field also shows the presence of two bands of low velocity on each si<strong>de</strong> of the<br />

hole. The lateral aspiration visualised in figure 12(c) creates a velocity <strong>de</strong>ficit on both<br />

si<strong>de</strong>s of the hole. Figure 12(d) presents isocontours of Q criterion (Hunt <strong>et</strong> al. 1988).

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!