Collaborative theoretical and computational study of jet noise control
Abstract
Despite intensive research, the reduction of jet engine noise by both active and passive means remains a theoretically and technologically challenging problem. The most effective means of jet noise reduction known to engineers is to increase the bypass-ratio. This technology, however, is not applicable to naval tactical aircraft, whose design is dictated by maneuverability, supersonicflight, and the requirement to perform short take-off and landing maneuvers from aircraft carrier decks. At the same time, basic research on jet noise reduction has focused on regimes and simplified geometries that are not directly relevant to military aviation. Active flow control, furthermore, often relies actuators that require prohibitively large amounts of energy, and control strategies that either rely on significantly simplified physical models, or black-box algorithm thatdo not leverage physical insights.Performers proposed research aims at breaking this paradigm by progressing and synthesizing two key technologies developed in previous ONR-supported research, and by leveraging a third developed by a separate ONR-funded activity at OSU. The first technology is large eddy simulation, which though decades of development has now reached a level of fidelity at which it accurately and reliably predicts the acoustic radiation from complex military-style nozzles at realistic conditions, and at a computational cost that brings design optimization within reach. The second technology is data- and operator-driven flow decomposition techniquesnamely spectral proper orthogonal decomposition (SPOD) and resolvent analysisthat afford insights into the physics of chaotic, fully turbulent flows from a dynamical systems perspective, even for complex geometries. The third technology is the latest generation of arc filament plasma actuators thatprovide control authority over turbulent, high-speed flows at moderate energy cost.Performer uses these technologies to push the envelope of both fundamental understanding of the physics of jet noise actuation, and of high-fidelity modeling of complex military-style nozzles including plasma actuation at the same time. With these ambitious goals in mind, we organize our effort into two converging thrusts. In the first thrust, we concentrate on supersonic jets from canonical,smooth converging-diverging nozzles, and use flow decomposition techniques to understand how passive nozzle modifications such as serrations (e.g. chevrons) and steady mass injection (e.g. microjets) alter the full spectrum of coherent turbulent motions and the mechanisms by which they radiate sound. Armed with LES databases for these jets, we propose a novel resolvent-analysis based optimization technique that will identify the structure of the near-nozzle t velocity and temperature fields that minimize far-field sound. In the second thrust, we directly model a complex, plasma-actuated twin rectangular jet associated with the OSU experiments. For this complex flow, and the extremely large dataset generated by LES, we propose a class of machine learningtechniques referred to as updating, streaming, or on-the-fly algorithms for computing the SPOD spectrum. This technique will enable us to obtain a detailed understanding of the flow physics of complex non-MOC nozzles, and in particular of the effects of plasma actuation on the acoustic far-field. The theoretical understanding gained in these studies will be synthesized inorder to propose specific actuation strategies in the companion experimental effort.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 29, 2020
- Source ID
- N000142012311
Entities
People
- Tim Colonius
Organizations
- California Institute of Technology
- Office of Naval Research
- United States Navy