Page 17 - Volume 9, Issue 3

66 Fig.14.Top:EnginenozzlecorrugationsdevelopedbySeineretal. andPSUfluidic
insert design, from Ref. 65. Bottom: Change in OASPL (negative number indicates reduction) due to three fluidic inserts for two different azimuthal angles.
worthwhile to summarize related concurrent research that may result in further understanding and possible reduction of military jet noise. Described are the development of advanced numerical simulation capabilities and application to the study of skewed waveform generation, wave packet modeling and control, and the development of a fluidic noz- zle “corrugation.”
Modeling Jet Skewness using Large-Eddy Simulations
Numerical modeling of military-style jet flows can be used to probe the flow features for the physics of jet noise production and examine the impact of nozzle design changes on the acoustic field. However, simulation of jet turbulence and noise generation is complicated, in part because of the vastly varying scales needing to be resolved to obtain the dynamic properties of the turbulence responsible for broad- band noise generation and associated memory and computa- tional requirements. In one approach, a compressible-flow large eddy simulation (LES) directly resolves the large-scale turbulence and then uses a sub-grid model to account for the fine-scale features within the jet plume. With the incorpora- tion of unstructured mesh capabilities and advancements in massively parallel, high-performance computing (HPC), LES is emerging as an accurate yet cost-effective computational tool for first-principles prediction of turbulent jets from complex military-style nozzles and their acoustic fields.
Researchers at Cascade Technologies and Stanford University have sought to improve understanding and devel- op predictive capabilities for propulsive jet aeroacoustics, through high-fidelity physics-based simulations with an unstructured LES framework known as “Charles.” In past
studies, Charles has been used to investigate wide-ranging jet
configurations, including various nozzle geometries57,58 with
58,59 60
chevrons and faceted military-style nozzles. In these
studies, calculations are carried out routinely on tens of thou- sands of processors at various HPC facilities.
Charles was recently used to reach a new HPC milestone when it ran on over 1 million cores in January 2013 during “Early Science” testing of the new Sequoia supercomputer at Lawrence Livermore National Laboratory. The jet noise cal- culation was performed for a heated supersonic jet from a military-style nozzle and is currently being used59 to under- stand how such jets emit the skewed pressure waveforms described previously in this article. The left plot in Figure 13 is an LES snapshot of the temperature field inside the jet plume and the instantaneous pressure field. At the right is an analysis of the spatial variation of the skewness of the unsteady pressure field inside and outside the shear layer, corroborating the F-35 AA results in Figure 10 that positive pressure skewness is produced at the source. The statistical properties of the pressure time derivative from the numerical data are also being analyzed. These results help illustrate the recent advances in the numerical modeling of jet aeroa- coustics and should yield an improved understanding of the source mechanisms in supersonic jet noise.
Modeling and Control of Wave Packets
A primary concern in developing jet noise reduction technologies for tactical aircraft is the requirement that air- craft performance is not impacted. Research46-40,61 conducted at California Institute of Technology and United Technologies Research Center aims at achieving significant jet noise reduction on tactical aircraft without impacting existing engine cycles. This requires active noise control techniques that target the peak low-frequency, aft-angle sound emissions associated with the most energetic large- scale structures. As described previously, the low-frequency large-scale mixing noise generation comprises radiating wave packets that are relatively coherent over multiple characteris- tic wavelengths. The present effort is aimed at improving understanding of how these wave packets responds to forcing (i.e. excitation by an external disturbances such as a second- ary unsteady jet) over a range of frequencies, waveforms, and actuation amplitudes.
The approach builds on successfully characterizing near- field pressure wave packets that are quantitatively related to both large-scale turbulent structures and far-field sound. These instability-wave models directly predict the evolution and radiation of the large-scale flow structures based solely on inputs available from experimental data or computational fluid dynamics codes. The resulting reduced-order models have already been validated for the unforced supersonic unheated and heated turbulent jets that were tested. By injecting unsteady flow disturbances with a harmonic com- ponent through two actuator jets near the nozzle lip, and by adjusting their relative phases, the excitation of the wave packets can be manipulated, thereby impacting the sound radiation. To date, the peak-radiation-angle noise from a per- fectly-expanded Mach 1.5 heated jet has been reduced by as
16 Acoustics Today, July 2013