Towards extending the aircraft flight envelope by mitigating transonic airfoil buffet

Aircraft flight envelopes are defined to ensure safe operations within limits of speed, load factor, and atmospheric conditions, as operations outside may endanger structural integrity. In modern lightweight aircraft operating transonically, the high-speed boundary is frequently determined by aeroelastic instabilities. While classical flutter depends primarily on dynamic pressure and mode coupling, transonic aeroelastic problems involve Mach number and angle of attack and are closely linked to unsteady aerodynamics known as transonic buffet, wherein a shock on the suction side oscillates and induces unsteady loads. Competing hypotheses attribute buffet either to a shock–trailing-edge acoustic feedback loop or to time-varying shock-induced separation altering the pressure ratio across the shock. Regardless, buffet poses a severe operational hazard, restricting the high-speed flight envelope. Prior control strategies include passive and active approaches: structural damping, shock control bumps, vortex generators, and trailing-edge devices (deflectors/flaps), with various trade-offs in complexity, robustness, off-design sensitivity, drag penalties, and noise. Trailing edges are also key acoustic sources; porous trailing edges (PTEs) have demonstrated trailing-edge noise reduction. This work proposes and experimentally validates porous trailing edges as a simple, robust, low-cost, and low-sensitivity passive technology to attenuate transonic buffet while retaining or improving aerodynamic performance through specific design choices: an internal impermeable plate to prevent pressure-side/suction-side mass flux (lift loss) and a perforated surface skin to reduce roughness-induced drag.

The transonic buffet mechanism has been extensively investigated experimentally and numerically. Dominant hypotheses include: (1) a feedback loop in which vortical disturbances convect from the shock to the trailing edge, generate upstream-traveling acoustic waves, and interact with the shock; and (2) coupling between shock oscillations and time-varying shock-induced separation (boundary-layer "breathing"). Both mechanisms emphasize the shock–boundary-layer interaction and the role of trailing-edge regions. Control strategies explored include shock control bumps (effective but sensitive to shock position/strength and off-design), structural damping (adds weight), passive/active vortex generators (energize boundary layer but increase drag and are location-sensitive), and trailing-edge flaps/deflectors (effective in closed-loop but add complexity; fixed flaps can reduce oscillation amplitude). Trailing edges are also primary airframe noise sources; numerous studies show porous trailing edges reduce trailing-edge noise but may degrade lift or increase drag if mass flux and surface roughness effects are not mitigated. This study leverages that literature to design PTEs that target buffet attenuation with minimal aerodynamic penalty.

Experiments were conducted in the Trisonic Wind Tunnel at the Institute of Aerodynamics, RWTH Aachen University, in the transonic test section. The rigid model is based on the supercritical DRA 2303 airfoil (c = 150 mm, 14% thickness), mounted at fixed angle of attack α = 3.5°, with a zigzag trip at 5% chord to ensure turbulent boundary layer. Test conditions included fully developed buffet at Mach number Ma = 0.76; additional reference data at lower Ma were acquired to characterize pre-buffet conditions. The airfoil includes an exchangeable trailing-edge section enabling tests of: (i) a solid trailing edge (REF); and (ii) two porous trailing edges (PTE1, PTE2). PTE1 is a 3D lattice structure; PTE2 comprises stacked gyroid cubes. Both are 3D-printed in titanium via selective laser melting as single parts, with design features to minimize aerodynamic penalties: a solid impermeable plate along the centerline to block pressure–suction side mass flux and perforated surface skins to reduce surface roughness and turbulence production near the wall. The wind tunnel has adjustable side-walls, calibrated to follow streamlines to minimize wall interference. Flow generation is via a vacuum storage facility providing stable conditions for about 2 seconds per run, enabling approximately 85 buffet cycles per measurement. Synchronized diagnostics captured the shock motion and trailing-edge flow: high-speed Background-Oriented Schlieren (BOS) for density gradients (two Photron SA5 cameras, 8000 Hz, high-power LEDs, dot pattern) and planar high-speed Particle-Image Velocimetry (PIV) for velocity fields (two pco.dimax HS4 cameras, Darwin Duo 527-100-M laser at 4000 Hz; DEHS seeding; Scheimpflug setup). BOS data were processed using a deep optical flow network (RAFT-PIV) to obtain pixel-resolved density gradient fields and a 2D Noise-Assisted Multivariate Empirical Mode Decomposition (2D NA-MEMD) to enhance shock edge extraction and ensure temporal coherence. PIV data were evaluated with an in-house cross-correlation pipeline (multigrid 8–4–2, final 16×16 px windows, 75% overlap, iterative predictor–corrector, Gaussian peak fit, normalized median outlier detection). Shock positions were extracted from BOS; boundary-layer properties (thickness δ at 0.99U∞, velocity PDFs, Reynolds shear stress) and vortex metrics (swirling strength λ) were derived from PIV in the trailing-edge region x/c ∈ [0.8, 1]. Aerodynamic performance trends (lift and drag contributions) were inferred by reconstructing the pressure field from measurements: integrate BOS density gradients (weighted least squares) to obtain density; use PIV velocities and the energy equation to compute temperature; then apply the ideal gas law for pressure. The near-wall pressure distribution was validated against surface pressure taps from prior studies on the same airfoil in the same facility at Ma = 0.73, showing good agreement. Due to optical limitations, only portions of the airfoil surface (notably around and downstream of the shock on the suction side and along the porous trailing edge on the pressure side) were captured; lift and drag contributions from these regions were integrated and compared across configurations to assess performance trends.

• Reference buffet characteristics (REF): Shock oscillation amplitude about 6% chord; dominant buffet frequency f ≈ 180 Hz with Strouhal number St ≈ 0.109; standard deviation of shock position std(x/c) = 0.0159; time-averaged shock location ⟨x/c⟩ = 0.4620.
• Buffet attenuation with PTEs: Both PTE1 and PTE2 substantially damp shock oscillations and eliminate the buffet frequency peak in power spectra.
  • PTE1: ⟨x/c⟩ = 0.4717 (2.1% downstream shift); std(x/c) = 0.0075 (52.8% reduction vs REF).
  • PTE2: ⟨x/c⟩ = 0.4665 (1.0% downstream shift); std(x/c) = 0.0088 (44.7% reduction vs REF).
• Boundary-layer modifications in x/c ∈ [0.8, 1]:
  • Thickness δ/c (median): PTE1 = 0.1069 (+10.2%); PTE2 = 0.0988 (+1.9%); REF = 0.0970.
  • Std(δ/c): PTE1 = 0.0238 (−23.5%); PTE2 = 0.0208 (−33.1%); REF = 0.0311.
  • Reynolds shear stress ⟨−u′v′⟩ (arbitrary units from PIV): PTE1 = 198.8 (+2.2%); PTE2 = 193.4 (~−0.6%); REF = 194.5.
  • Vortical structure strength λ (swirling strength metric): PTE1 = 13331 (−3.3%); PTE2 = 13078 (−5.1%); REF = 13787.
  • Velocity PDFs show with PTEs: increased wall-normal extent of reversed flow (larger recirculation), reduced boundary-layer breathing, and vertical velocity distributions more symmetric about zero. Streamwise velocity PDFs in PTE cases resemble pre-buffet distributions of REF (Ma ≤ 0.72), indicating a shift toward a more stable state.
• Aerodynamic performance (pressure-derived contributions in measured regions):
  • PTE1: Lift coefficient contribution −7.58%; drag coefficient contribution +2.55% vs REF.
  • PTE2: Lift coefficient contribution +1.35%; drag coefficient contribution −0.07% (negligible) vs REF.
• Mechanism: Porous surfaces allow transpiration, reducing local kinetic energy, thickening the recirculation region, and increasing boundary-layer resistance to disturbances. This reduces boundary-layer breathing and stabilizes pressure fluctuations downstream of the shock, and the enlarged recirculation region acts as a buffer at the shock foot. Together, these effects damp shock motion, mitigating buffet.

The experiments show that porous trailing edges effectively mitigate transonic airfoil buffet by modifying the separated boundary layer downstream of the shock. By reducing boundary-layer breathing and providing a buffer region at the shock foot, PTEs disrupt the self-sustaining mechanisms driving shock oscillations. This addresses the core challenge of extending the high-speed limit of the flight envelope by alleviating an aerodynamic instability often linked to aeroelastic issues, thereby potentially improving flight safety. Importantly, the choice of porous architecture is critical for aerodynamic performance: the lattice-based PTE1 increases boundary-layer thickness and turbulence slightly, leading to lift loss and drag increase, while the gyroid-based PTE2 achieves strong buffet damping with minimal and favorable impacts on lift and drag. These findings highlight a promising passive technology that, unlike many other buffet control methods, can simultaneously offer aeroacoustic benefits and reduced sensitivity to off-design shock positions.

This study experimentally demonstrates that porous trailing edges substantially attenuate transonic airfoil buffet, offering a path to extend the aircraft flight envelope toward higher speeds while also reducing trailing-edge noise. Two PTE designs were evaluated: both damped shock oscillations markedly, and the gyroid-based PTE2 preserved aerodynamic performance with a slight lift gain and negligible drag change, establishing design guidelines for practical implementation. The physical mechanism is attributed to transpiration-induced boundary-layer thickening, reduced boundary-layer breathing, and a buffering recirculation region stabilizing the shock–boundary-layer interaction. Future work should: (i) perform parametric studies of porous architectures and permeabilities enabled by additive manufacturing; (ii) quantify effects on buffet onset/offset boundaries across Mach number and angle of attack; (iii) integrate force balances or surface pressure measurements on non-porous regions to obtain full-airfoil loads, and/or use high-fidelity simulations incorporating porous media models; (iv) explore machine-learning-based data assimilation and causal discovery to refine understanding; and (v) extend to aeroelastic, 3D wing configurations and assess interactions with aircraft components.

• Only two porous trailing edge designs were tested; maximum attainable mitigation and optimal porosity/architecture remain unknown.
• Experiments used a rigid, 2D airfoil model; aeroelastic coupling and 3D wing effects (sweep, tip, nacelle, fuselage interactions) were not included.
• Aerodynamic performance (lift/drag) was inferred from partial surface pressure reconstructions due to limited optical access; no direct force balance was available, and pressure taps could not be embedded in porous sections.
• Measurements covered specific transonic conditions with a fixed angle of attack; broader off-design behavior and buffet onset/offset shifts were not mapped.
• Short run durations (order of seconds) limited the number of cycles captured; while sufficient for dominant statistics, very low-frequency phenomena were not assessed.
• Reconstruction relies on assumptions (e.g., ideal gas, isentropic regions upstream/downstream of the shock) and spatial interpolation between BOS and PIV grids.