betaFeedback

WorksPhysics and Astronomy

Numerical Investigation of Boundary-Layer Height and Actuation-Parameter Effects of a Circular Synthetic Jet Actuator in Crossflow

Abstract

Three-dimensional unsteady numerical simulations are performed to investigate the effects of blowing ratio CBC_B (0.85<U‾j/U∞<1.70.85 < \overline{U}_j/U_\infty < 1.7), stroke ratio L+L^+ (10.6<U‾j/(fd)<21.310.6 < \overline{U}_j /(fd) < 21.3), and boundary-layer height ratio D+D^+ (2.1<δ/d<8.02.1<δ/d<8.0) on circular synthetic jet actuator (SJA) performance in crossflow. Nine cases are examined at constant free-stream velocity U∞U_\infty, with systematic independent variation of averaged jet velocity U‾j\overline{U}_j, actuation frequency ff (200200-400 Hz400~\mathrm{Hz}), and boundary-layer momentum thickness Reynolds number (170<Reθ<740170<Re_θ<740) to isolate the individual effects of these parameters on a circular-nozzle SJA with fixed nozzle diameter dd in crossflow. Instantaneous vortical structures exhibited tilted vortex rings with a trailing vortex pair at low actuation frequency; closely packed expelled vortical structures for higher frequency SJAs, and the largest boundary-layer height ratio induced hairpin-like vortices. Near-wall tertiary vortices, which promote downwash and increase wall shear stress, remain coherent longer and have extended spanwise coverage for low D+D^+. Time-averaged boundary-layer profiles and skin-friction distributions reveal that SJAs with low to moderate D+D^+ have the greatest potential for separation control, maintaining increased near-wall momentum over extended streamwise distances.

Introduction

Synthetic jet actuators (SJAs) are zero-net-mass-flux (ZNMF) devices that have gained significant attention in active flow control due to their compactness, energy efficiency, and ability to operate without additional fluid supply systems[1, 2, 3, 4]. Unlike conventional steady jets that require continuous input from pressurized reservoirs or plumbing, SJAs operate with the surrounding bulk fluid, making them highly adaptable for embedded aerodynamic control applications. By transferring linear momentum into the flow via the periodic expulsion of vortex structures, SJAs have demonstrated effectiveness in diverse control scenarios such as separation control[5, 6, 7, 8, 9], fluid mixing enhancement[10, 11, 12, 13, 14], and turbulence manipulation[15]. Their ability to generate significant momentum flux while consuming minimal power makes them promising for low-Reynolds-number applications in UAVs[16, 17, 18], wind turbines[19, 20, 21], and compact heat exchangers[22, 23, 24, 25].

An SJA typically consists of a cavity with a vibrating diaphragm or piston and a neck leading to an orifice or slot exit on the control surface[26]. A schematic of an SJA in a crossflow boundary layer is shown in Figure 1; the crossflow has a free-stream velocity U∞U_{\infty} and boundary-layer thickness δ\delta. The actuator is characterized by an orifice diameter dd, neck height hh, and an overall actuator height denoted here as HsjaH_{\mathrm{sja}}. During each actuation cycle, the diaphragm alternates between ingestion and expulsion. When the expelled jet column escapes reingestion, a vortex ring forms and constitutes a synthetic jet (SJ). While there is no net mass addition to the flow, each expulsion transfers linear momentum to the external flow through the expelled vortical structures.

Schematic of a synthetic jet actuator in crossflow during the expulsion cycle

Figure 1. Schematic of a synthetic jet actuator in crossflow during the expulsion cycle.

The interaction between synthetic jets and crossflow involves three-dimensional mechanisms of entrainment, penetration, and shear-layer instabilities, generating coherent structures such as upstream horseshoe vortices, hairpins, and vortex rings[27, 28, 29, 30]. Depending on operating parameters, the jet column evolves into either a hairpin vortex or a tilted vortex ring with a trailing vortex pair and near-wall tertiary vortices, as shown in Figure 2. Key dimensionless parameters governing strength and penetration of the synthetic jet in a crossflow are the blowing ratio CB=U‾j/U∞C_B = \overline{U}_j/U_\infty, stroke ratio L+=U‾j/(fd)L^+ = \overline{U}_j/(fd), and momentum coefficient

Schematic of vortex structures formed by a synthetic jet actuator in boundary-layer crossflow

Figure 2. Schematic of vortex structures formed by a synthetic jet actuator in boundary-layer crossflow: (a) hairpin vortex (HV), or (b) tilted vortex ring (VR) with trailing vortex pair (TVP) and near-wall tertiary vortices (TV).

Cμ=ρjU‾j2ρ∞U∞2dθ0(1)C_\mu= \frac{\rho_j\overline{U}_j^{2}}{\rho_\infty U_\infty^{2}}\frac{d}{\theta_0} \tag*{(1)}

where U‾j\overline{U}_j is the average jet velocity during the expulsion cycle, defined as

U‾j=2TAn∫An∫0T/2Uj(t,An) dt dAn(2)\overline{U}_j = \frac{2}{T A_n}\int_{A_n}\int_{0}^{T/2} U_j(t,A_n)\,\mathrm{d}t\,\mathrm{d}A_n \tag*{(2)}

AnA_n is the jet exit area, TT the actuation period, ff the actuation frequency, and θ0\theta_0 the baseline momentum thickness. Prior experimental and numerical studies show strong sensitivity of structures to these parameters[30, 31, 32].

Jabbal and Zhong[28] used dye visualization to reveal flow regimes under different CBC_B. At low CBC_B (<0.35< 0.35), expelled vortices form stretched hairpins that convect near the wall[29]. With increasing CBC_B, tilted vortex rings form and penetrate further[28], often accompanied by trailing vortex pairs and near-wall tertiary vortices that promote downwash[33, 34]. Ho et al.[35] used 3-D URANS to examine the influence of CBC_B (0.32–1.10) in turbulent boundary-layer crossflow and observed improved penetration at constant frequency, but a trade-off between mid-span wall-shear increase and spanwise control authority. Actuation frequency is also central to control effectiveness[36, 37]. Excessive frequency can reduce separation control due to insufficient momentum per stroke[38, 39]. In addition to CBC_B, L+L^{+} helps determine whether structures appear as stretched hairpins or well-defined vortex rings[28]. The effect of L+L^{+} (12–42) at constant CB=5C_B = 5 was studied numerically in[40], showing reduced penetration and changing CfC_f with increasing frequency. Boundary-layer characteristics further influence performance. Chaudhry and Zhong[41] examined laminar and turbulent layers with the same D+=δ/d=6D^{+} = \delta/d = 6 and observed hairpins, stretched rings, and tilted vortices in both. Higher CBC_B and L+L^{+} produced structures that persisted longer in the turbulent layer.

Table I summarizes previous experimental and numerical studies investigating circular synthetic jet actuators in boundary layer crossflow. The table highlights that while there is substantial prior work examining SJAs at low to moderate blowing ratios, limited studies have explored higher momentum configurations (CB>1C_B > 1) with extended stroke ratios (L+>10L^{+} > 10). Furthermore, most investigations maintain a fixed boundary layer height ratio, leaving the independent influence of D+D^{+} on SJA performance inadequately characterized. This gap poses a direct challenge for applying SJAs in flow control applications, where a small D+D^{+} requires a larger array of SJAs to achieve the same spanwise control authority, while a larger D+D^{+} can lead to significant alterations to the surface geometry.

Therefore, the objective of this study is to investigate the effects of blowing ratio (0.85<CB<1.70.85 < C_B < 1.7) and stroke ratio (10.6<L+<21.310.6 < L^{+} < 21.3) on SJA performance across different boundary-layer height ratios (2.1<D+<82.1 < D^{+} < 8) using three-dimensional URANS. Because SJAs often operate near the cavity’s Helmholtz resonance, CBC_B and L+L^{+} vary together experimentally, complicating isolation of their effects. A numerical approach allows independent boundary conditions and systematic variation. We examine the evolution of three-dimensional vortical structures and near-wall behavior across CBC_B, L+L^{+}, and D+D^{+} to inform SJA selection for flow control.

TABLE I: Summary of previous studies on circular-nozzle SJAs in boundary-layer crossflow.

Technique abbreviations: Dye Vis. = Dye Visualization, LDV = Laser Doppler Velocimetry, PIV = Particle Image Velocimetry, HWA = Hot-Wire Anemometry, DNS = Direct Numerical Simulation, LES = Large-Eddy Simulation, URANS = Unsteady Reynolds-Averaged Navier–Stokes.

ReferenceReθRe_\thetaD+D^{+}CBC_{B}L+L^{+}Technique
Zhong et al.[29]395–5472.7–3.70.06–0.70.56–1.4Dye Vis.
Shuster et al.[42]––1.121–2PIV
Schaeffler et al.[43]–3.30.56–1.3–LDV, 2D PIV
Dandois et al.[44]430031.4517URANS, LES
Wu & Leschziner[45]9204291LES
Jabbal & Zhong[28]–2.3–3.60.08–0.70.8–5.1Dye Vis.
Wu & Leschziner[46]240010291LES
Jabbal & Zhong[34]–40.27–0.541.6–2.72D PIV
Chaudhry & Zhong[47]32060.17–0.541.7–2.7PIV
Chaudhry & Zhong[41]32060.11–0.362.2–3.6Dye Vis.
Xia & Mohseni[3]85–14432.8 – 8.32.8–5.7HWA, PIV
Palumbo et al.[48]500–5500.23–0.250.10.22–0.85DNS
Palumbo et al.[15]5500.23–0.250.05–0.10.1–0.85DNS
Ho et al.[35]9007.250.32–1.103.9 – 13URANS
Ho et al.[40]–1.54.911.5 – 41URANS
Chhetri et al.[49]8957.750.654.1URANS
Current Study170–7402.1–8.00.85–1.710.6–21.3URANS

Table 1. Summary of previous studies on circular-nozzle SJAs in boundary-layer crossflow.

Numerical Methods

Governing equations and turbulence model

The three-dimensional URANS simulations were conducted using OpenFOAM v2412[50]. The unsteady Reynolds-averaged mass and momentum equations are

∂Ui∂xi=0,(3)\frac{\partial U_i}{\partial x_i}=0, \tag*{(3)}
∂Ui∂t+Uj∂Ui∂xj=−1ρ∂P∂xi+ν∂2Ui∂xj∂xj−∂∂xj(ui′uj′‾),(4)\frac{\partial U_i}{\partial t}+U_j\frac{\partial U_i}{\partial x_j}=-\frac{1}{\rho}\frac{\partial P}{\partial x_i}+\nu\frac{\partial^2 U_i}{\partial x_j\partial x_j}-\frac{\partial}{\partial x_j}\left(\overline{u'_i u'_j}\right), \tag*{(4)}

where UiU_i are mean velocity components, ρ\rho is density, PP is mean pressure, and ui′uj′‾\overline{u'_i u'_j} are the Reynolds stresses. With the Boussinesq approximation,

−ui′uj′‾=νt(∂Ui∂xj+∂Uj∂xi)−23kδij,(5)-\overline{u'_i u'_j}=\nu_t\left(\frac{\partial U_i}{\partial x_j}+\frac{\partial U_j}{\partial x_i}\right)-\frac{2}{3}k\delta_{ij}, \tag*{(5)}

where νt\nu_t is eddy viscosity, δij\delta_{ij} the Kronecker delta, and kk turbulent kinetic energy. Following preliminary testing in prior work[35, 51], the low-Reynolds-number kk–ε\varepsilon model of Launder–Sharma[52] was selected; it models the viscous sublayer via damping functions, avoiding wall functions.

The discretized equations were solved with the pisoFoam solver (finite-volume method). Spatial discretization used second-order schemes for convective terms; first-order upwind was applied to turbulence quantities for stability. Time integration used first-order explicit Euler with adaptive time-stepping to maintain a Courant–Friedrichs–Lewy (CFL) number <0.9< 0.9. A convergence threshold of 10−610^{-6} was applied to all variables.

Model Setup, Boundary and Test Conditions

The computational domain for the SJA in crossflow is shown in Figure 3. The rectangular duct is 200d200d long, 20d20d wide, and 38d38d high; the SJA exit center is at x/d=0x/d=0, y/d=0y/d=0, mid-span z/d=0z/d=0, located 25d25d downstream of the inlet. A velocity inlet and pressure outlet were used; walls are no-slip; side boundaries are symmetry. Separate duct simulations generated the different inlet boundary-layer profiles.

Computational domain and boundary conditions for the SJA in crossflow

Figure 3. Computational domain and boundary conditions for the SJA in crossflow.

The SJA exit diameter is d=2d=2 mm. The cavity was not modeled; instead, the analytical Womersley solution for pulsating laminar pipe flow was applied at the neck inlet[15]:

v(r,t)=VjRe{[1−J0(i3/2Wo r)J0(i3/2Wo)]eiωt},(6)v(r,t)=V_j\mathop{\mathrm{Re}}\left\{\left[1-\frac{J_0\left(i^{3/2}Wo\,r\right)}{J_0\left(i^{3/2}Wo\right)}\right]e^{i\omega t}\right\}, \tag*{(6)}
CaseD+D^{+}ReθRe_{\theta}HHReτRe_{\tau}RejRe_{j}CBC_{B}ffL+L^{+}
A-15700.8520010.6
A-22.11703.14911501.720021.3
A-311501.740010.6
B-15700.8520010.6
B-24.14601.813511501.720021.3
B-311501.740010.6
C-15700.8520010.6
C-28.07401.526611501.720021.3
C-311501.740010.6

Table 2. Summary of parameters for the synthetic jet in crossflow.

where VjV_j is the maximum centerline velocity, ω\omega is the angular frequency, Wo=dω/4νWo = d\sqrt{\omega/4\nu} is the Womersley number, and J0J_0 is the zeroth-order Bessel function. Prior work[51] validated that, with sufficient modeled neck volume, this method reproduces whole-SJA (dynamic mesh) results at lower cost. Here, a neck height ratio h/d=15h/d = 15 is used to satisfy the neck-volume requirement and avoid nonphysical jet exit behavior.

Nine cases (Table II) examine the effects of boundary-layer height and SJA settings. The series are grouped by inlet D+D^{+}: Series A: D+=2.1D^{+}=2.1; Series B: D+=4.1D^{+}=4.1; Series C: D+=8.0D^{+}=8.0. Within each, case “1” is low momentum/low frequency; “2” is high momentum/low frequency; “3” is high momentum/high frequency.

Model Validation

The computational grid is shown in Figure 4. Inflation layers were applied to the top and bottom walls. An O-grid was used in the SJA neck and extended into the duct. Refinement is highest near the exit and immediately downstream, coarsening into the free stream. A grid-resolution analysis[53] on case C-3 used three meshes. Mesh refinement occurred along all axes, focusing on shear-layer and boundary-layer regions. Table III summarizes centerline velocity UclU_{cl} at peak expulsion and time-averaged displacement thickness δ∗\delta^{*} at x/d=10x/d=10 for the three meshes. Sampling began after 2 flow-through cycles (tU∞/l=2tU_{\infty}/l=2, where ll is the length of the duct), which corresponds to a dimensionless time based on jet diameter of t∗=tU∞/d=400t^{*}=tU_{\infty}/d=400. Time-averaged δ∗\delta^{*} was computed over five actuation cycles. Grid II was selected for accuracy versus cost. The medium mesh had 40 wall layers, first cell height 75 μ\mum, growth rate 1.07, and a dimensionless first cell wall distance of ymax⁡+<5y_{\max}^{+}<5.

Mesh configuration showing the symmetry-plane view and zoomed view of the jet exit

Figure 4. Mesh configuration: (a) symmetry-plane view (partial length shown) and (b) zoomed view of the jet exit.

GridTotal cellsUclU_{cl}UclU_{cl} Uncertainty (%)δx/d=10∗\delta^*_{x/d=10}δx/d=10∗\delta^*_{x/d=10} Uncertainty (%)
I2×1062 \times10^{6}15.8–4.97–
II5×1065 \times10^{6}15.61.54.391.9
III13×10613 \times10^{6}15.5<1<14.45<1<1

Table III. Grid properties for the mesh-sensitivity study (case C-3).

Baseline boundary layers for the three series of test cases were sampled at x/d=0x/d=0 (Figure 5). The Blasius profile is used for the lowest-ReRe inlet A; the T3A experimental data[54] is used to validate the highest-ReRe inlet C. The agreement between the present simulations and the experimental and analytical results validates the approach boundary-layer.

Boundary-layer profiles for the three inlets

Figure 5. Boundary-layer profiles for the three inlets, compared with analytical/experimental data[54] at similar Reynolds numbers.

Sampling of crossflow simulations began after 4.375 flow-through cycles (t∗=875t^* = 875). Probes near the jet exit and within the boundary layer confirmed highly periodic behavior with minimal phase fluctuations; therefore, phase averaging was not performed. Time-averaged fields were obtained by averaging every time step over five actuation cycles.

Results and Discussion

Instantaneous Flow Structure

QQ-criterion contours are used to visualize vortical structures, where QQ is the second invariant of the velocity-gradient tensor[55]. Chakraborty et al.[56] showed that different vortex-identification criteria are qualitatively consistent, validating the use of QQ as

Q=12(∥Ω∥2−∥S∥2)>0,(7)Q = \frac{1}{2}\left(\lVert\boldsymbol{\Omega}\rVert^{2} - \lVert\mathbf{S}\rVert^{2}\right) > 0, \tag*{(7)}

identifying rotation-dominated regions. Here, QQ is applied to instantaneous 3-D fields to visualize structures induced by the synthetic-jet boundary-layer interaction.

Figure 6 and Figure 7 show iso-surfaces at t∗=875t^{*} = 875, corresponding to the start of the ingestion cycle for all cases. Low-momentum cases (A-1, B-1, C-1) display a tilted vortex ring (VR) with trailing vortex pairs (TVP) extending toward the wall. With higher momentum (A-2, B-2, C-2), the VR penetrates further; upstream shear-layer interactions produce vortex loops[46, 57]. At high momentum and frequency (A-3, B-3, C-3), expelled structures cluster more closely streamwise. In C-3 (largest D+D^{+}), VRs break into hairpin-like structures away from the wall due to interference between consecutive structures; this is not observed in A-3 and B-3, where the jet experiences a more uniform crossflow. Top-down views (Figure 7) show slower convection of the primary VR with increasing D+D^{+} at low momentum; this trend diminishes at higher momentum/frequency.

Instantaneous Q-criterion iso-surfaces for Cases A-1 through C-3

Figure 6. Instantaneous QQ-criterion iso-surfaces (Q=0.1U∞2/d2Q = 0.1U_{\infty}^{2}/d^{2}) colored by spanwise vorticity at isometric view.

Instantaneous Q-criterion iso-surfaces for Cases A-1 through C-3

Figure 7. Instantaneous QQ-criterion iso-surfaces (Q=0.1U∞2/d2Q = 0.1U_{\infty}^{2}/d^{2}) colored by spanwise vorticity at xx–zz plane view.

Near-wall tertiary vortices (TV) increase near-wall momentum by inducing downwash and are relevant to separation control[34, 35]. In series A, prominent TVs convect downstream along the wall; at higher frequency (A-3), their spanwise footprint narrows, indicating reduced near-wall control authority. With increasing D+D^{+} (series B, C), TVs dissipate sooner, consistent with weaker near-wall momentum and reduced shear-layer interaction.

Figure 8 further probes spanwise control via streamwise vorticity and spanwise velocity. Plane locations (dashed lines) vary with jet momentum/frequency but are fixed with respect to D+D^{+} to

Instantaneous Q-criterion iso-surfaces and normalized vorticity and velocity fields for Cases A-1 through C-3

Figure 8. Instantaneous QQ-criterion iso-surfaces (Q=0.1U∞2/d2Q = 0.1U_{\infty}^{2}/d^{2}) colored by spanwise vorticity (xx–yy plane); normalized streamwise vorticity and spanwise velocity (yy–zz planes) at indicated x/dx/d.

capture the primary structures. In series A, A-2 shows the deepest penetration, while A-3 exhibits closer-spaced structures and a narrower near-wall spanwise impact than A-1/A-2. In series B and C, near-wall spanwise flow is weaker due to greater D+D^{+}.

Top-down planes (Figure 9) show classic vortex-pair signatures for low-momentum VRs at y/d=5y/d = 5. With higher momentum, the vorticity concentrates near mid-span and structures approach continuity at a higher frequency, resembling a steady jet in crossflow.

Instantaneous streamwise vorticity contours

Figure 9. Instantaneous streamwise vorticity contours at y/d=5y/d = 5.

Jet Trajectory

To examine the effect of D+D^{+} on trajectory and penetration, the vortex-pair centers for A-1, B-1, and C-1 are plotted in Figure 10. High-momentum/frequency cases are omitted due to complex symmetry-plane vorticity. The trajectory for C-1 rises more rapidly but plateaus sooner, consistent with Figure 7. Time-averaged streamlines superimposed on transverse-velocity contours are shown in Figure 11. The streamlines originating from the jet exit are evaluated up to x/d=20x/d = 20. Overall, penetration does not vary strongly with D+D^{+}, with A-2 an exception showing lower penetration than B-2/C-2.

Normalized vortex-center trajectories for cases A-1, B-1, and C-1

Figure 10. Normalized vortex-center trajectories for cases A-1, B-1, and C-1.

Time-averaged normalized transverse velocity along the symmetry plane with time-averaged streamlines superimposed

Figure 11. Time-averaged normalized transverse velocity along the symmetry plane with time-averaged streamlines superimposed.

Time-Averaged Boundary-Layer Profile

Figure 12 shows time-averaged streamwise velocity U‾x/U∞\overline{U}_x/U_\infty at several x/dx/d. For series A, all SJAs increase near-wall momentum (y/d∈[0,0.5]y/d \in[0,0.5]), beneficial for separation delay, with a compensating outer deficit (y/d∈[0.5,10]y/d \in[0.5,10]). The outer deficit recovers but remains visible at x/d=30x/d = 30; near-wall profiles converge by x/d=30x/d = 30. For series B and C, behavior is similar, but the near-wall increase is weaker; by x/d=30x/d = 30, the near-wall profiles are close to baseline while an outer deficit remains.

Time-averaged normalized streamwise velocity profiles along the symmetry plane

Figure 12. Time-averaged normalized streamwise velocity along the symmetry plane.

Time-averaged transverse velocity U‾y/U∞\overline{U}_y/U_\infty is presented in Figure 13, which shows downwash near the wall (y/d<1y/d < 1) at x/d=1.5x/d = 1.5 for all actuated cases, consistent with TVs. In series A, the downwash persists to x/d=10x/d = 10; in series B and C, it decays by x/d=5x/d = 5. Away from the wall (y/d>1y/d > 1), a primary peak develops; a minor secondary peak near the exit for B, C decays by x/d=5x/d = 5.

Time-averaged normalized transverse velocity along the symmetry plane

Figure 13. Time-averaged normalized transverse velocity along the symmetry plane.

The magnitude and location of the peak transverse velocity are extracted and plotted in Figure 14. Peak transverse velocity decays rapidly for all cases from x/d=1.5x/d = 1.5 to 5, most sharply at high frequency (A-3, B-3, C-3). In series A, the magnitudes of A-2 and A-3 converge after x/d=20x/d = 20, but their peak locations differ: for A-3, the peak occurs at y/d=6.45y/d = 6.45 at x/d=20x/d = 20 and moves to y/d=4.8y/d = 4.8 by x/d=30x/d = 30, whereas A-2 moves farther from the wall with xx. In series B and C, peak magnitudes and locations vary modestly, except B-3 peaks farther from the wall than B-2 at x/d=30x/d = 30. Thus D+D^{+} is most influential at low D+D^{+} (≲2\lesssim2); moderate–high D+D^{+} (≥4\ge4) has less effect on transverse profiles.

Downstream decay and wall-normal location of maximum streamwise velocity for series A, B, and C

Figure 14. (a) Downstream decay of U‾y,max⁡/U∞\overline{U}_{y,\max}/U_{\infty} against x/dx/d; (b) Wall-normal location y/dy/d of U‾y,max⁡\overline{U}_{y,\max} against x/dx/d.

Time-averaged kˉ\bar{k} (series B, C) increases in the near field (x/d∈[1.5,10]x/d \in[1.5,10]) for all actuated cases, with stronger sensitivity to jet momentum than to D+D^{+} or frequency. As structures convect downstream, the SJA effect persists through the boundary layer, with a near-wall kˉ\bar{k} deficit appearing for x/d<20x/d < 20–3030.

Time-averaged turbulent kinetic energy along the symmetry plane for series B and C

Figure 15. Time-averaged turbulent kinetic energy along the symmetry plane for series B and C.

Skin-Friction Coefficient Profile

Figure 16 shows that series A produces large increases in mid-span CfC_f downstream of the jet exit. Case A-2 shows two additional high-CfC_f regions at z/d=±5z/d = \pm5 not attributable to near-wall hairpins; this suggests excitation of Tollmien–Schlichting (TS) waves. The baseline for series A has Reδ∗≈500Re_{\delta_*} \approx500 (critical Blasius Reδ∗≈520Re_{\delta_*} \approx520). Following[15], the reduced frequency F+=f/frefF^+ = f/f_{\mathrm{ref}} with fref=0.752U∞/df_{\mathrm{ref}} = 0.752U_\infty/d gives F+≈1.06F^+ \approx1.06 for A-1/A-2, consistent with TS amplification; the effect is weaker for A-1 due to lower jet momentum. TS amplification is not observed in B-2 or C-2 due to larger Reδ∗Re_{\delta_*}.

For B-1 and C-1, CfC_f increases only within x/d≲10x/d \lesssim10; farther downstream, it drops below baseline. Higher momentum (B-2, C-2) extends the region of increased CfC_f but introduces spanwise deficit regions at moderate–high CB35C_B^{35}. Higher frequency (B-3, C-3) yields a modest additional mid-span increase; however, the effect decays rapidly and CfC_f falls below baseline by x/d≳25x/d \gtrsim25, underscoring potential adverse far-field impacts in both spanwise and streamwise directions.

Time-averaged skin-friction coefficient on the lower wall for Cases A-1, A-2, A-3, B-1, B-2, B-3, C-1, C-2, and C-3

Figure 16. Time-averaged skin-friction coefficient on the lower wall; black dashed contour marks 100Cf=−0.5100 C_f=-0.5.

Conclusion

We independently varied boundary-layer height ratio (D+≡δ/dD^{+} \equiv\delta/d), blowing ratio (CBC_{B}), and actuation frequency (ff) across nine three-dimensional URANS cases to isolate their individual influence on circular synthetic-jet/crossflow interactions. Three robust trends emerge.

Actuation frequency governs structure packing and jet penetration. Raising ff from 200 Hz to 400 Hz clusters expelled vortex rings in the streamwise direction and reduced jet penetration depth by 15–25 % in the thickest layer (D+=8D^{+} = 8). In this regime, consecutive rings interfere and promote a transition from tilted vortex rings to hairpin-like structures farther from the wall. The associated transverse-velocity peaks decay to less than 10 % of their near-field magnitude within x/d≈10x/d \approx10, indicating diminished sustained control authority at high frequency.

Boundary-layer height modulates near-wall coherence. At low D+≈2.1D^{+} \approx2.1, near-wall tertiary vortices persist beyond x/d=20x/d = 20 and span Δz/d>4\Delta z/d > 4, maintaining increased near-wall momentum over long distances. At moderate-to-high D+≥4.1D^{+} \ge4.1, these structures dissipate by x/d≈12x/d \approx12–15 and near-wall profiles recover to baseline conditions by x/d≈30x/d \approx30.

Mean-flow and wall-shear responses reflect a trade-off. Increasing CBC_{B} from 0.85 to 1.7 boosts mid-span skin-friction coefficient (CfC_{f}) by up to 120 % immediately downstream (x/d<5x/d < 5). However, the spanwise footprint narrows, with off-center deficits emerging at ∣z/d∣>3|z/d| > 3. An incidental Tollmien–Schlichting amplification occurs in one low-D+D^{+}, low-ff configuration at reduced frequency F+≈1.06F^{+} \approx1.06, consistent with transitional receptivity.

Practical guidance. For flat-plate separation control at Reθ∼102Re_{\theta} \sim10^{2}–10310^{3}, low-to-moderate D+≤4D^{+} \le4 combined with the lower end of the tested frequency range (f≈200f \approx200 Hz) proves most effective for sustained, spanwise-broad near-wall momentum increase. When D+D^{+} is large, raising CBC_{B} helps locally at mid-span but delivers reduced spanwise coverage and faster downstream decay.

Funding

This research was funded by the Natural Sciences and Engineering Research Council of Canada (NSERC) grant number RGPIN-2022-03071 and the Digital Research Alliance of Canada (4752).

Acknowledgments

Computations were performed at the SciNet HPC Consortium.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

References

  1. [1]M. Amitay, B. Smith, and A. Glezer, “Aerodynamic flow control using synthetic jet technology,” in 36th AIAA Aerospace Sciences Meeting and Exhibit (1998) p. 208.DOI
  2. [2]B. L. Smith and A. Glezer, “The formation and evolution of synthetic jets,” Physics of Fluids 10, 2281–2297 (1998).DOI
  3. [3]X. Xia and K. Mohseni, “Flow characterization and modeling of strong round synthetic jets in crossflow,” AIAA Journal 55, 389–402 (2017).DOI
  4. [4]A. Glezer and M. Amitay, “Synthetic jets,” Annual Review of Fluid Mechanics 34, 503–529 (2002).
  5. [5]F. De Vanna, M. R. Mcconnell, J. Knight, and J. M. Buick, “Boundary layer separation from a curved backward-facing step using improved delayed detached-eddy simulation,” Fluids 10, 145 (2025).
  6. [6]J. Yen and N. A. Ahmed, “Parametric study of dynamic stall flow field with synthetic jet actuation,” Journal of Fluids Engineering, Transactions of the ASME 134, 071104 (2012).DOI
  7. [7]M. Tadjfar and D. Kamari, “Optimization of flow control parameters over sd7003 airfoil with synthetic jet actuator,” Journal of Fluids Engineering, Transactions of the ASME 142, 021301 (2020).DOI
  8. [8]A. Machado, K. Xu, and P. E. Sullivan, “Spanwise control authority of synthetic jets on a stalled airfoil,” Physics of Fluids 36, 064113 (2024).DOI
  9. [9]W. Jia and H. Xu, “Effect of synthetic jets actuator parameters on deep reinforcement learning-based flow control performance in a square cylinder,” Physics of Fluids 36, 084113 (2024).DOI
  10. [10]B. Zhang, H. Liu, Y. Li, H. Liu, and J. Dong, “Experimental study of coaxial jets mixing enhancement using synthetic jets,” Applied Sciences 11, 1–13 (2021).DOI
  11. [11]W. He, Z. b. Luo, X. Deng, C. Peng, Q. Liu, T. x. Gao, P. Cheng, Y. Zhou, and W. q. Peng, “Numerical study on the atomization mechanism and energy characteristics of synthetic jet/dual synthetic jets,” Applied Energy 346, 121345 (2023).
  12. [12]12W. Huang, Z. b. Du, L. Yan, and Z. x. Xia, “Supersonic mixing in airbreathing propulsion systems for hypersonic flights,” Progress in Aerospace Sciences 109, 100545 (2019).DOI
  13. [13]13W. Huang, “Transverse jet in supersonic crossflows,” Aerospace Science and Technology 50, 183–195 (2016).DOI
  14. [14]14W. Huang and L. Yan, “Progress in research on mixing techniques for transverse injection flow fields in supersonic crossflows,” Journal of Zhejiang University-SCIENCE A 14, 554–564 (2013).
  15. [15]15A. Palumbo, O. Semeraro, J. C. Robinet, and L. De Luca, “Boundary layer transition induced by low-speed synthetic jets,” Physics of Fluids 34, 124104 (2022).DOI
  16. [16]16M. Ja’fari, F. J. Shojae, and A. J. Jaworski, “Synthetic jet actuators: Overview and applications,” International Journal of Thermofluids 20, 100438 (2023).DOI
  17. [17]17K. Xu, P. Lavoie, and P. E. Sullivan, “On the effect of varying momentum coefficient on separation control for a thick symmetric airfoil,” Physics of Fluids 37, 104120 (2025).DOI
  18. [18]18S. Nakamura, K. Ogawara, and H. Shingin, “Numerical study on control of flow into aileron gaps using psjas for uavs with side-force fins,” in AIAA SCITECH 2025 Forum (2025) p. 1872.
  19. [19]19V. Maldonado, M. Boucher, R. Ostman, and M. Amitay, “Active vibration control of a wind turbine blade using synthetic jets,” International Journal of Flow Control 1, 227–237 (2009).DOI
  20. [20]20F. Aguirre-Villegas, B. Miranda-Godoy, and L. Silva-Llanca, “Harnessing the jet-flap effect: Enhancing lift with synthetic jets,” International Journal of Mechanical Sciences 304, 110677 (2025).
  21. [21]21M. Saemian and J. M. Bergadà, “Active flow control actuators on wind turbines; comprehensive review,” Ocean Engineering 339, 121991 (2025).DOI
  22. [22]22A. Arshad, M. Jabbal, and Y. Yan, “Synthetic jet actuators for heat transfer enhancement – a critical review,” International Journal of Heat and Mass Transfer 146, 118815 (2020).DOI
  23. [23]23Y.-w. Lyu, J.-w. Tan, J.-z. Zhang, J.-y. Zhang, and F.-m. Wang, “Active heat transfer enhancement roles by an acoustic actuator integration into square array of continuous jets in the presence of crossflow,” International Journal of Heat and Mass Transfer 240, 126568 (2025).DOI
  24. [24]24K. N. D. Hammond, M. O. Sandoval, N. Naeem, and E. E. Essel, “On the turbulent flow characteristics of wall-attaching synthetic jets at varying offset height ratios,” Physics of Fluids 37, 115105 (2025).DOI
  25. [25]25S. Dutta and Shantanu, “Effects of the synthetic jet on the flow field and heat transfer over a surface-mounted square rib,” Experimental Thermal and Fluid Science 139, 110708 (2022).DOI
  26. [26]L. N. Cattafesta and M. Sheplak, “Actuators for active flow control,” Annual Review of Fluid Mechanics 43, 247–272 (2011).DOI
  27. [27]M. Jabbal, J. Wu, and S. Zhong, “The performance of round synthetic jets in quiescent flow,” The Aeronautical Journal 110, 385–393 (2006).DOI
  28. [28]M. Jabbal and S. Zhong, “The near wall effect of synthetic jets in a boundary layer,” International Journal of Heat and Fluid Flow 29, 119–130 (2008).
  29. [29]S. Zhong, F. Millet, and N. J. Wood, “The behaviour of circular synthetic jets in a laminar boundary layer,” The Aeronautical Journal 109, 461–470 (2005).
  30. [30]A. Crook and N. J. Wood, “Measurements and visualisations of synthetic jets,” in 36th AIAA Aerospace Sciences Meeting and Exhibit (2001).
  31. [31]H. Tang and S. Zhong, “Modelling of the characteristics of synthetic jet actuators,” in 35th AIAA Fluid Dynamics Conference and Exhibit (2005).DOI
  32. [32]S. Zhong, L. Garcillan, and N. J. Wood, “Dye visualisation of inclined and skewed synthetic jets in a cross flow,” The Aeronautical Journal 109, 147–155 (2005).DOI
  33. [33]J. Zhou and S. Zhong, “Numerical simulation of the interaction of a circular synthetic jet with a laminar boundary layer,” in 4th AIAA Flow Control Conference (2008).
  34. [34]M. Jabbal and S. Zhong, “Particle image velocimetry measurements of the interaction of synthetic jets with a zero-pressure gradient laminar boundary layer,” Physics of Fluids 22, 063603 (2010).DOI
  35. [35]H. H. Ho, E. E. Essel, and P. E. Sullivan, “The interactions of a circular synthetic jet with a turbulent crossflow,” Physics of Fluids 34, 075108 (2022).DOI
  36. [36]D. Greenblatt and I. J. Wygnanski, “Control of flow separation by periodic excitation,” Progress in Aerospace Sciences 36, 487–545 (2000).DOI
  37. [37]M. Amitay and A. Glezer, “Role of actuation frequency in controlled flow reattachment over a stalled airfoil,” AIAA Journal 40, 209–216 (2012).
  38. [38]W. Zhang and R. Samtaney, “A direct numerical simulation investigation of the synthetic jet frequency effects on separation control of low-re flow past an airfoil,” Physics of Fluids 27, 055101 (2015).DOI
  39. [39]M. Kim, E. E. Essel, and P. E. Sullivan, “Effect of varying frequency of a synthetic jet on flow separation over an airfoil,” Physics of Fluids 34, 015122 (2022).DOI
  40. [40]H. H. Ho, A. Machado, and P. E. Sullivan, “Computational and experimental study of spanwise synthetic jet flow control,” in 13th International Symposium on Turbulence and Shear Flow Phenomena (TSFP13) (2024).
  41. [41]I. A. Chaudhry and S. Zhong, “A single circular synthetic jet issued into turbulent boundary layer,” Journal of Visualization 17, 101–111 (2014).DOI
  42. [42]J. Shuster, R. Pink, D. McEligot, and D. Smith, “The interaction of a circular synthetic jet with a cross-flow boundary layer,” in 35th AIAA Fluid Dynamics Conference and Exhibit (2005) p. 4749.DOI
  43. [43]N. W. Schaeffler and L. N. Jenkins, “Isolated synthetic jet in crossflow: experimental protocols for a validation dataset,” AIAA journal 44, 2846–2856 (2006).DOI
  44. [44]J. Dandois, E. Garnier, and P. Sagaut, “Unsteady simulation of synthetic jet in a crossflow,” AIAA journal 44, 225–238 (2006).DOI
  45. [45]D. K. Wu and M. A. Leschziner, “Large-eddy simulations of synthetic jets in stagnant surroundings and turbulent cross-flow,” in IUTAM Symposium on Flow Control and MEMS: Proceedings of the IUTAM Symposium held at the Royal Geographical Society, 19–22 September 2006, hosted by Imperial College, London, England (Springer, 2008) pp. 127–134.DOI
  46. [46]D. K. L. Wu and M. A. Leschziner, “Large-eddy simulations of circular synthetic jets in quiescent surroundings and in turbulent cross-flow,” International Journal of Heat and Fluid Flow 30, 421–434 (2009).DOI
  47. [47]I. A. Chaudhry and S. Zhong, “The evolution of synthesised vortices in turbulent boundary layer,” Journal of Turbulence 14, 1–18 (2013).DOI
  48. [48]A. Palumbo, O. Semeraro, J.-C. Robinet, and L. de Luca, “Receptivity to synthetic jet actuation in boundary layer flows,” in AIAA Scitech 2020 Forum (2020) p. 0099.DOI
  49. [49]A. Chhetri, N. F. Ouedraogo, H. H. Ho, L. Drummond, and E. E. Essel, “Influence of nozzle convergence angle on the flow characteristics of synthetic jets in a turbulent crossflow boundary layer,” Journal of Fluids Engineering 147, 081203 (2025).DOI
  50. [50]H. Jasak, “Openfoam: Open source cfd in research and industry,” International Journal of Naval Architecture and Ocean Engineering 1, 89–94 (2009).DOI
  51. [51]H. H. Ho, E. E. Essel, and P. Sullivan, “Improving 3d synthetic jet modeling in a crossflow,” Journal of Fluids Engineering 146, 031302 (2024).arxiv.org/abs/2311.10072
  52. [52]B. E. Launder and B. I. Sharma, “Application of the energy-dissipation model of turbulence to the calculation of flow near a spinning disc,” Letters in Heat and Mass Transfer 1, 131–137 (1974).
  53. [53]53 I. B. Celik, U. Ghia, P. J. Roache, C. J. Freitas, H. Coleman, and P. E. Raad, “Procedure for estimation and reporting of uncertainty due to discretization in cfd applications,” Journal of Fluids Engineering, Transactions of the ASME 130, 078001 (2008).DOI
  54. [54]54 P. E. Roach and D. H. Brierley, “The influence of a turbulent free-stream on zero pressure gradient transitional boundary layer development. Part I: Test cases T3A and T3B,” in Numerical Simulation of Unsteady Flows and Transition to Turbulence, edited by O. Pironneau, W. Rodi, I. L. Ryhming, A. M. Savill, and T. V. Truong (Cambridge University Press, 1992) pp. 319–347.
  55. [55]55 J. C. R. Hunt, A. A. Wray, and P. Moin, “Eddies, streams, and convergence zones in turbulent flows,” Center for Turbulence Research Proceedings of the Summer Program , 193–208 (1988).
  56. [56]56 P. Chakraborty, S. Balachandar, and R. J. Adrian, “On the relationships between local vortex identification schemes,” Journal of Fluid Mechanics 535, 189–214 (2005).
  57. [57]57 B. R. Ravi, R. Mittal, and F. M. Najjar, “Study of three-dimensional synthetic jet flowfields using direct numerical simulation,” in 42nd AIAA Aerospace Sciences Meeting and Exhibit (2004).DOI

Paper details

vlnc:W_vuvac9vh
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Not recorded for this paper

View statement
Looking for related papers…
Contents