Session H21: Acoustics IV: Aeroacoustics

Acoustic structures in the near-field from clustered rocket nozzles

The plume and acoustic field produced by a cluster of two and four rocket nozzles is visualized by way of retroreflective shadowgraphy. Steady state and transient operations (startup/shutdown) were conducted in the fully-anechoic chamber and open jet facility of The University of Texas at Austin. The laboratory scale rocket nozzles comprise thrust-optimized parabolic contours, which during start-up, experience free shock separated flow, restricted shock separated flow and an end-effects regime prior to flowing full. Shadowgraphy images with synchronized surveys of the acoustic loads produced in close vicinity to the rocket clusters and wall static pressure profiles are first compared with several RANS simulations during steady operations. A Proper Orthogonal Decomposition of various regions in the shadowgraphy images is then performed to elucidate the prominent features residing in the supersonic annular flow region, the acoustic near field and the interaction zone that resides between the nozzle plumes. POD modes are used to detect propagation paths of the acoustic waves and shock cell structures in the supersonic shear layer. Spectral peak frequencies on the propagation paths are associated with the shock cell length, which are responsible for generating broadband shock noise.   

Properties and Localizations of Acoustic Sources in High Speed Jet

Jet noise has become one major concern for aircraft engine design in recent decades. The problem is to identify the near-field (NF) structures that produce far-field (FF) noise and develop noise control and reduction strategies. We developed an algorithm to identify the events that are responsible for NF and FF cross-correlations. Two sets of experimental data from Mach 0.6 jets are analyzed. They consist of 10kHz TRPIV measurement and pressure sampling in both near- and far-field. Several NF diagnostics (velocity, vorticity, Q criterion, etc.) are calculated to represent the 2D velocity fields. The main contributors between these NF diagnostics and FF pressure are extracted as Diagnostic-Microphone (DM) events. The NF localization of DM event clusters will be compared to the NF triangulation of MM events, which were acquired using FF signals alone. In the time-frequency domain, the events are short wave packets, distorted by ambient perturbations. As a result, the matching of DM to MM events at physical lags is particularly difficult. We will report on different algorithms using time, frequency and space information to improve the reliability of the matches. We will also relate the event localization to the NF flow fields that correspond to FF ``loud'' POD modes (Low et al. 2013 and Berger et al. 2014).   

End-effects-regime in full scale and lab scale rocket nozzles

Modern rockets utilize a thrust-optimized parabolic-contour design for their nozzles for its high performance and reliability. However, the evolving internal flow structures within these high area ratio rocket nozzles during start up generate a powerful amount of vibro-acoustic loads that act on the launch vehicle. Modern rockets must be designed to accommodate for these heavy loads or else risk a catastrophic failure. This study quantifies a particular moment referred to as the ``end-effects regime,'' or the largest source of vibro-acoustic loading during start-up [Nave {\&} Coffey, AIAA Paper 1973-1284]. Measurements from full scale ignitions are compared with aerodynamically scaled representations in a fully anechoic chamber. Laboratory scale data is then matched with both static and dynamic wall pressure measurements to capture the associating shock structures within the nozzle. The event generated during the ``end-effects regime'' was successfully reproduced in the both the lab-scale models, and was characterized in terms of its mean, variance and skewness, as well as the spectral properties of the signal obtained by way of time-frequency analyses.   

The Structure and Noise Reduction Capacity of Owl Down

Many species of owl rely on specialized plumage to reduce their self-noise levels and enable hunting in acoustic stealth. In contrast to the leading-edge comb and compliant trailing-edge fringe attributes of owls, the aeroacoustic impact of the fluffy down material on the upper wing surface remains largely speculative as a means to eliminate aerodynamic noise across a broad range of frequencies. Photographic analysis of the owl down reveals a unique forest-like structure, whereby the down fibers rise straight up from the wing surface and then bend into the flow direction to form a porous canopy, with an open area fraction of approximately 70\%. Experimental measurements demonstrate that the canopy feature reduces dramatically the turbulent pressure levels on the wing surface by up to 30dB, which affects the roughness noise characteristic of the down in a manner consistent with the theory of flows over and through vegetation. Mathematical models developed for the turbulence noise generation by the down fibers and for the mixing-layer instability above the porous canopy furnish a theoretical basis to understand the influence of the down geometric structure on its self-noise signature and noise suppression characteristics.   

Wall Modeled Large Eddy Simulation of Airfoil Trailing Edge Noise

Large eddy simulation (LES) of airfoil trailing edge noise has largely been restricted to low Reynolds numbers due to prohibitive computational cost. Wall modeled LES (WMLES) is a computationally cheaper alternative that makes full-scale Reynolds numbers relevant to large wind turbines accessible. A systematic investigation of trailing edge noise prediction using WMLES is conducted. Detailed comparisons are made with experimental data. The stress boundary condition from a wall model does not constrain the fluctuating velocity to vanish at the wall. This limitation has profound implications for trailing edge noise prediction. The simulation over-predicts the intensity of fluctuating wall pressure and far-field noise. An improved wall model formulation that minimizes the over-prediction of fluctuating wall pressure is proposed and carefully validated. The flow configurations chosen for the study are from the workshop on benchmark problems for airframe noise computations. The large eddy simulation database is used to examine the adequacy of scaling laws that quantify the dependence of trailing edge noise on Mach number, Reynolds number and angle of attack. Simplifying assumptions invoked in engineering approaches towards predicting trailing edge noise are critically evaluated.   

Direct numerical simulation and reduced-order modeling of the sound-induced flow through a cavity-backed circular under a turbulent boundary layer

Commercial jet aircraft generate undesirable noise from several sources, with the engines being the most dominant sources at take-off and major contributors at all other stages of flight. Acoustic liners, which are perforated sheets of metal or composite mounted within the engine, have been an effective means of reducing internal engine noise from the fan, compressor, combustor, and turbine but their performance suffers when subjected to a turbulent grazing flow or to high-amplitude incident sound due to poorly understood interactions between the liner orifices and the exterior flow. Through the use of direct numerical simulations, the flow-orifice interaction is examined numerically, quantified, and modeled over a range of conditions that includes current and envisioned uses of acoustic liners and with detail that exceeds experimental capabilities. A new time-domain model of acoustic liners is developed that extends currently-available reduced-order models to more complex flow conditions but is still efficient for use at the design stage.   

Computation of noise from separated flows using large eddy simulation

We investigate noise production from turbulent flow over bluff bodies using the Ffowcs-Williams and Hawkings (FW-H) acoustic analogy. We propose a dynamic end cap methodology to account for volumetric contributions to the far-field sound within the context of the FW-H acoustic analogy. The quadrupole source terms are correlated over multiple planes to obtain a convection velocity that is then used to determine a corrective convective flux at the FW-H porous surface. The proposed approach is first demonstrated for a convecting potential vortex. It is then applied to compute the noise from a cylinder at Re$_D$=89k, and a 45 degree beveled trailing edge at Re$_c$=1.9M. We compare our results for base flow and acoustic data to available computations and experiments. We demonstrate insensitivity of the end cap correction approach to end plane location and spacing, discuss the effect of dynamic convection velocity, and show better performance than commonly used end cap corrections. Finally, we discuss some physical mechanisms that generate the far-field noise.   

Cross-stream ejection in the inter-wheel region of aircraft landing gears

The reduction of aircraft noise is an important challenge currently faced by aircraft manufacturers. During approach and landing, the landing gears contribute a significant proportion of the aircraft generated noise. It is therefore critical that the key noise sources be identified and understood in order for effective mitigation methods to be developed. For a simplified two-wheel nose landing gear, a strong cross stream flow ejection phenomena has been observed to occur in the inter-wheel region in presence of wheel wells. The location and orientation of these flow ejections causes highly unsteady, three dimensional flow between the wheels that may impinge on other landing gear components, thereby potentially acting as a significant noise generator. The effects of changing the inter-wheel geometry (inter-wheel spacing, the wheel well depth and main strut geometry) upon the cross-stream ejection behaviour has been experimentally investigated using both qualitative flow visualisation and quantitative PIV techniques. A summary of the key results will be presented for the three main geometrical parameters under examination and the application of these findings to real life landing gears will be discussed.   

Sound-turbulence interaction in transonic boundary layers

Acoustic wave scattering in a transonic boundary layer is investigated through a novel approach. Instead of simulating directly the interaction of an incoming oblique acoustic wave with a turbulent boundary layer, suitable Dirichlet conditions are imposed at the wall to reproduce only the reflected wave resulting from the interaction of the incident wave with the boundary layer. The method is first validated using the laminar boundary layer profiles in a parallel flow approximation. For this scattering problem an exact inviscid solution can be found in the frequency domain which requires numerical solution of an ODE. The Dirichlet conditions are imposed in a high-fidelity unstructured compressible flow solver for Large Eddy Simulation (LES), CharLES$^{\mathrm{x}}$. The acoustic field of the reflected wave is then solved and the interaction between the boundary layer and sound scattering can be studied.   

An exact and dual-consistent formulation for high-order discretization of the compressible viscous flow equations

Finite-difference operators satisfying a summation-by-parts property enable discretization of PDEs such that the adjoint of the discretization is consistent with the continuous-adjoint equation. The advantages of this include smooth discrete-adjoint fields that converge with mesh refinement and superconvergence of linear functionals. We present a high-order dual-consistent discretization of the compressible flow equations with temperature-dependent viscosity and Fourier heat conduction in generalized curvilinear coordinates. We demonstrate dual-consistency for aeroacoustic control of a mixing layer by verifying superconvergence and show that the accuracy of the gradient is only limited by computing precision. We anticipate dual-consistency to play a key role in compressible turbulence control, for which the continuous-adjoint method, despite being robust, introduces adjoint-field errors that grow exponentially. Our dual-consistent formulation can leverage this robustness, while simultaneously providing an exact sensitivity gradient. We also present a strategy for extending dual-consistency to temporal discretization and show that it leads to implicit multi-stage schemes. Our formulation readily extends to multi-block grids through penalty-like enforcement of interface conditions.