Next generation jet noise models for complex-geometry nozzles
Abstract
Next generation jet noise models for complex-geometry nozzles The goal of this research is the development of accurate, next-generation reduced-order models for the mixing noise of supersonic jets issuing from complex nozzle geometries representative of UCAV and other tactical aircraft. Such nozzles are well beyond the scope of existing empirical models built primarily to predict subsonic noise from round nozzles. Long term, the new models would permit optimization of nozzle characteristics for minimum noise based on (rapidly computable) Reynolds-average Navier-Stokes simulations together with reliable, physics-based noise prediction. Buoyed by recent discoveries regarding the physics of wavepacket structures, the dominant source of mixing noise in turbulent jets, and by a revolutionary approach to predicting them via a spatial marching (parabolization) technique called the one-way Euler (OWE) equations, we propose here a number of steps that are required to develop the new noise models. First, we will perform large eddy simulations (LES) of co-annular and rectangular nozzles to produce databases that complement existing ones for single-stream round jets and provide crucially needed data for model development. The LES employ state-of-the-art algorithms such as resolution of internal-nozzle boundary-layer modeling and advanced mesh generation/adaptation techniques. Based on past work, the LES can be expected to produce a realistic database from which the mean flow field, turbulent fluctuations, and far-field sound can be reliably evaluated at every point in space. Armed with the data, we will extend our OWE method, previously developed for axisymmetric, subsonic flows to supersonic jets issuing from non-axisymmetric nozzles by adopting a planemarching technique. A unique feature compared to previous work on linear models for wavepacket-generated noise is the use of stochastically forced solutions. This stochastic forcing arises from small-scale turbulence that continually “jitters” the wavepackets and alters the acoustic efficiency with which they radiate. In the first instance, the stochastic forcing will be educed from the LES data in order to validate the OWE models. To pose universal stochastic forcing models, we will also perform a so-called resolvent (or input/output) analysis of the jets. Resolvent computations seek a low-rank description of the transfer function between distributed inputs associated with turbulent fluctuations and the far-field sound. We will compare OWE models forced using the LES data with those forced by resolvent modes and propose stochastic forcing models that lead to accurate representations of wavepackets and their radiated noise.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 23, 2016
- Source ID
- N000141612445
Entities
People
- Tim Colonius
Organizations
- California Institute of Technology
- Office of Naval Research
- United States Navy