This study examines the acoustic behavior in forward speeds of 0, 15, 30, and 60 knots of manned-size, multi-rotor, eVTOL aircraft in quadcopter, hexacopter, and octocopter configurations. The rotors are assumed to have constant RPM and are controlled through collective pitch, with orthogonal phasing between rotors. All configurations share the same disk loading and hover tip Mach, with the rotor radius decreasing and the RPM increasing as the number of rotors increase. The simulations use the Rensselaer Multicopter Analysis Code (RMAC) for the aerodynamic loads coupled with the PSUWOPWOP code for noise predictions at an observer hemisphere. From the results, it is shown that at higher forward flight speeds, where loading noise becomes more dominant, high elevation angles (below the vehicle) show peak noise for all configurations. Asymmetry in the number of outside advancing blades, such as on the plus quadcopter, causes higher noise levels for the left (advancing) side. In-plane directivity patterns that show large decreases due to signal cancellations in inter-boom noise are shown to progressively diminish at larger forward speeds, with lower reductions for the hexacopter and octocopter configurations. The total acoustic radiated power is also compared with the single rotor for all multicopter configurations. The comparison shows that the acoustic power versus the single rotor power is decreasing as we increase the number of rotors or the forward speed.

This study examines a hovering three-bladed two-rotor system in close operation to the ground. The rotor pair is oppositely phased and is examined for two heights, H/D = 1 and H/D = 0.5. Loads for these rotors are generated using the CFD solver AcuSolve, as well as the Rensselaer Multicopter Analysis Code (RMAC). The loads generated using CFD include aerodynamic interactions from inter-rotor effects and ground-rotor effects. These loads are coupled to the acoustic propagation PSU-WOPWOP code for acoustic predictions at an observer grid located at “ear height” from the ground. Rotors are also added below the ground to simulate perfect acoustic ground reflections. From the simulation results, the noise signals from the rotors were found to have a distinct directivity pattern of six locations of high noise and six locations of low noise caused by the rotor phasing. The introduction of aerodynamic interactions increases loading noise and has a significant effect on the high frequency noise content. These effects are amplified for the H/D = 0.5 case as the interactions are stronger. Ground reflections increased noise across the observer grid, and further enhanced the high frequency noise generated by the interactional aerodynamic loads.

Introduction V ISBAL and Rizzetta1 and Visbal et al.2 have recently performed large-eddy simulation (LES) calculations of turbulent channel flow and compressible isotropic turbulence decay without using any explicit subgrid-scale (SGS) model. In those simulations, spatial filtering was treated as an implicit SGS model. They also showed that use of an SGS model in those simulations did not produce results superior to those obtained without employing an SGS model. Bogey and Bailly3 also performed LES calculations for jet flows by using spatial filtering only. Moreover, they brought up the issue of the effects of the eddy–viscosity-based Smagorinsky SGS model on jet noise in yet another recent study.4,5 They showed that the high-frequency portion of the noise spectra was significantly suppressed by the eddy–viscosity. It is well understood that in turbulent flows the energy cascade is associated with a mean flux of energy that is directed from large scales toward small scales. The large scales contain the major part of the turbulent kinetic energy, and they continuously feed the turbulent kinetic energy via the cascade to the smallest eddies where it is dissipated. Because the grid resolution is too coarse to resolve all of the relevant length scales in an LES, the pile-up of energy at high wave numbers can be eliminated through the use of a spatial filter. Hence, the spatial filter can be thought of as an effective SGS model in an LES. In this study, we perform two jet simulations to study the impact of the SGS model on jet turbulence and far-field noise. The first simulation does not employ any explicit SGS model, but treats the spatial filter as an implicit SGS model. In the second simulation, we employ a localized version of the dynamic Smagorinsky model (DSM) (see Ref. 6) and keep all test case parameters the same as those in the first simulation. We examine the differences between the two simulations to quantify the effect of the SGS model on the near-field jet turbulence and the far-field noise. Although this study is similar to those done by Bogey and Bailly,4,5 our work additionally does a comparison of the two LES results with experimental jet noise spectra with the hope

The field of jet noise has emerged as one of the most active areas of research due to the recent introduction of stringent aircraft noise regulations. Computational aeroacoustics (CAA) in conjunction with the large eddy simulation (LES) technique shows a tremendous potential in improving our understanding of the noise generation mechanisms and thereby facilitates the development of effective noise reduction devices for jet engines. This potential can be realized on current petascale computing platforms provided that the LES code is capable of handling complex grid topologies with an outstanding scalability to many thousands of cores. The current work aims at developing a scalable, multiblock, structured-grid LES code based on a compact finite difference scheme formulated by a tridiagonal system of equations. The solution of such a system is computed by employing a scalable parallel solver that exhibits good scalability up to 91,125 processor cores. The implementation details of the multiblock topology that simplifies handling complex topologies are discussed. The present multiblock approach does not require overlapping grid points between the neighboring blocks and thus, eliminates routinely-used special treatments at the block boundaries that may compromise the simulation accuracy. The simulation turnaround time using the current LES tool for up to two-billion-grid-point case is compared with our previously developed LES tools to demonstrate the scalability of the code using as many as 91,125 processors. Flow and far-field noise data are reported for an unheated jet at a Reynolds number of 200,000 by utilizing three meshes of different grid resolutions with the finest mesh containing �100 million grid points. The sensitivity of the results to the shear layer thickness is also investigated using the high resolution grids.

Two modified compressibility corrections are introduced based on the notion that the reduction in the growth rate in compressible free shear flows levels off at a certain turbulence Mach number, as seen in experiments and direct numerical simulation (DNS). The OVERFLOW code is used to assess Sarkar’s compressibility correction and two modifications when used with the SST turbulence model. Test cases include an axisymmetric jet, a 3-D jet with crossflow, a generic boundary layer, and a generic mixing layer. The modified corrections do not disturb the beneficial behavior of the standard correction observed for the axisymmetric jets. For the 3-D cases, which have a higher Mach number, the modified corrections show significant improvement over the Sarkar and uncorrected models. The compressible boundary layer test case shows that there is still an underprediction of the wall skin friction coefficient with the compressibility corrections, but this can be solved by using a simple modification in the turbulence model, as is already done in OVERFLOW. Finally, analysis of the compressible mixing layer shows that the modified corrections exhibit the correct trend in limiting the reduction of the growth rate, but do not match the level of reduction seen with experimentally-obtained results.

The jet propulsion exhaust nozzle system is an integral part of an airbreathing gas turbine engine and critical to its overall performance. Challenges associated with the design and manufacturing of an exhaust nozzle increases with the cruise speed of the aircraft. The exhaust nozzle system for a supersonic cruise aircraft mandates additional features such as variable throat and exit area, jet noise suppression, and reverse thrust. In the past, in order to address this challenge, an ejector nozzle with clamshells was designed and fabricated. The experimental investigation and computation of the nozzle at low subsonic conditions showed the presence of a recirculation zone at the inner surface of the clamshells. The present work summarizes the computational simulation of the ejector nozzle with clamshells at flight conditions. Similar recirculation zones were predicted at the inner surface of the clamshells. Initiatives were taken to improve the ejector performance by the elimination of the recirculation zone. The current nozzle design was modified by the application of chevrons on the nozzle throat. A preliminary design and computational analysis of the ejector nozzle with clamshells and chevrons was carried out. Two design cases with different number of chevrons were implemented and their computational analysis was successfully carried out. It was observed that the nozzle flow features were improved because of enhanced mixing and the recirculation zone was decreased in its extent.

The lattice-Boltzmann method (LBM) was used to study the far-field noise generated from a Mach, M(j)=0.4, unheated turbulent axisymmetric jet. A commercial code based on the LBM kernel was used to simulate the turbulent flow exhausting from a pipe which is 10 jet radii in length. Near-field flow results such as jet centerline velocity decay rates and turbulence intensities were in agreement with experimental results and results from comparable LES studies. The predicted far field sound pressure levels were within 2 dB from published experimental results. Weak unphysical tones were present at high frequency in the computed radiated sound pressure spectra. These tones are believed to be due to spurious sound wave reflections at boundaries between regions of varying voxel resolution. These "VR tones" did not appear to bias the underlying broadband noise spectrum, and they did not affect the overall levels significantly. The LBM appears to be a viable approach, comparable in accuracy to large eddy simulations, for the problem considered. The main advantages of this approach over Navier-Stokes based finite difference schemes may be a reduced computational cost, ease of including the nozzle in the computational domain, and ease of investigating nozzles with complex shapes.

An economically and technically viable supersonic airliner needs substantial new technology to ensure that such an aircraft meets airport noise, sonic boom and emission requirements. The noise requirement calls for a variable-cycle propulsion system with a variable-area nozzle. As an initiative to address this challenge, a new supersonic ejector nozzle with clamshell doors is proposed as a noise suppression jet engine exhaust system. A design table driven, parametric nozzle geometry was designed. The experimental and numerical studies of its flow field were carried out. Cases with and without clamshells were considered and their mean velocity flow fields were compared. The experimental investigation involved the testing of the nozzle in a wind tunnel, and the measurements were taken using a seven-hole probe mounted on an automated 2-axis traverse instrument. Various flow visualization techniques, such as the fluorescent oil flow, surface sediment traces, smoke wand, and surface tuft were also implemented to capture the flow physics on the inside of the clamshells. Numerical simulations were performed using the commercially available finite volume-based computational fluid dynamic code FLUENT. The computational results are in good agreement with the experimental measurements at different axial locations downstream of the nozzle throat. A zone of flow separation and recirculation is captured at the inner surface of the clamshells in both the experimental investigation and numerical computation.

The OVERFLOW code is used for computations of supersonic jets. Corrections for compressibility and temperature effects are evaluated based on comparisons with experimental data. Test cases include isothermal and heated, perfectly expanded and underexpanded axisymmetric jets with design Mach numbers around two. Also, a 3-D jet in an angled crossflow at a design Mach number of three is evaluated. The SST turbulence model is chosen for the flows examined in this study. The Sarkar, Zeman, and Wilcox compressibility corrections and Abdol-Hamid’s temperature correction are assessed when used with the SST model. Results for the axisymmetric jets show that using a compressibility correction is important for a good prediction of the jet development. The effect of the temperature correction is not as significant. For the 3-D jet in crossflow, conclusions are not as decisive. The velocity decay is underpredicted when using the compressibility correction. The temperature correction behaves as it did for the axisymmetric jet. The overall recommendation is that the corrections must be used with caution, especially when approaching high temperature and Mach number regimes.

Large-eddy simulations (LES) of a circular supersonic fully expanded jet and an underexpanded jet are performed. Both jets operate at a design Mach number 1.95, and the corresponding fully expanded Mach number for the underexpanded jet is 2.20. Since the simulations do not include the nozzle geometry explicitly, the efiects of the in∞ow conditions (such as the forcing modes and shear layer thickness) on the near fleld statistics and far fleld noise are investigated. The near fleld data are collected from the LES and then the far fleld noise is computed by the Ffowcs Williams-Hawkings (FWH) surface integral method. The results show that removing the lower forcing modes increases the centerline velocity decay rate, the peak turbulence intensities, and the overall sound pressure level. However, reducing the in∞ow shear layer thickness has the opposite efiect and achieves a better agreement with the experiments in both near fleld statistics and farfleld noise. Similar trends hold for the underexpanded jet in near fleld statistics; however, the number of shock cells is underpredicted and is insensitive to the in∞ow conditions. This may be due to contamination by spurious numerical ∞uctuations caused by the shocks. Further simulations including shock capturing schemes are needed.

The performance of two popular turbulence models, the Spalart-Allmaras model and Menter s SST model, and one relatively new model, Olsen & Coakley s Lag model, are evaluated using the OVERFLOWcode. Turbulent shock-boundary layer interaction predictions are evaluated with three different experimental datasets: a series of 2D compression ramps at Mach 2.87, a series of 2D compression ramps at Mach 2.94, and an axisymmetric coneflare at Mach 11. The experimental datasets include flows with no separation, moderate separation, and significant separation, and use several different experimental measurement techniques (including laser doppler velocimetry (LDV), pitot-probe measurement, inclined hot-wire probe measurement, preston tube skin friction measurement, and surface pressure measurement). Additionally, the OVERFLOW solutions are compared to the solutions of a second CFD code, DPLR. The predictions for weak shock-boundary layer interactions are in reasonable agreement with the experimental data. For strong shock-boundary layer interactions, all of the turbulence models overpredict the separation size and fail to predict the correct skin friction recovery distribution. In most cases, surface pressure predictions show too much upstream influence, however including the tunnel side-wall boundary layers in the computation improves the separation predictions.

Through the use of Lighthill’s acoustic analogy, the aim of this paper is to investigate the noise sources of turbulent heated round jets using previously simulated Large Eddy Simulation (LES) data. Two heated and one unheated jet are considered to study the e ects of heating on the noise source contributions to the far-field. Firstly, the computed overall sound pressure level (OASPL) and spectra are in good agreement with the prediction obtained from the porous Ffowcs Williams-Hawkings (FWH) surface integral method. Like the FWH prediction, however, the computed OASPL over-predicts the experiments by approximately 3dB but the trends agree reasonably well with the experimental results. Through decomposition of the Lighthill source term we obtain such sources as shear, self and entropy noise. An important finding is that when a high speed subsonic compressible jet is heated while keeping the ambient jet Mach number constant, significant cancellations occur in the far-field between the shear and entropy noise. In addition, heating a jet reduces the intensity of the nonlinear self noise terms compared to an unheated jet. For a low speed heated jet, the main contributing source is the entropy noise source while the shear and self noise sources hardly contribute to the far-field noise.

Improvements in computing speed over the past decade have made Large Eddy Simulation (LES) an attractive tool to study jet noise. In this regard, high-order compact flnite difierence schemes along with high-order fllters are used extensively in LES, especially for aeroacoustics problems, since these schemes not only have very high accuracy and spectral-like resolution, but also low-dispersion and difiusion errors. Due to the implicit nature of compact schemes, one technique of parallelization is based on the data transposition strategy. However, the computational resources that are spent on the communication between processors can be as high as 40%. This paper proposes an alternative parallelization methodology based on the Schur complement technique to address the substantial communication overhead of the transposition strategy. Computed results for 1-D and 2-D Linearized Euler Equation (LEE) test cases compare well to the corresponding exact solutions. The 3-D LES test case via the Schur complement shows on average a 3% and 1.5% difierence in turbulent intensity and far-fleld acoustics, respectively, compared to results of a previously simulated single-block methodology using data transposition. However, the parallel 3-D Schur complement case is roughly 2 times slower compared to the single-block case. The relatively poor performance is mainly due to the one dimensional partitioning applied. A three dimensional decomposition is expected to remedy the poor performance and will be carried out in the future.

The capability of the OVERFLOW code to accurately compute high-speed turbulent boundary layers and turbulent shock-boundary layer interactions is being evaluated. Configurations being investigated include a Mach 2.87 flat plate to compare experimental velocity profiles and boundary layer growth, a Mach 6 flat plate to compare experimental surface heat transfer,a direct numerical simulation (DNS) at Mach 2.25 for turbulent quantities, and several Mach 3 compression ramps to compare computations of shock-boundary layer interactions to experimental laser doppler velocimetry (LDV) data and hot-wire data. The present paper describes outlines the study and presents preliminary results for two of the flat plate cases and two small-angle compression corner test cases.