Inlet conditions for wave packet models in turbulent jets based on eigenmode decomposition of large eddy simulation data

Abstract

This paper makes contributions towards reduced-order models of wave packets in supersonic, turbulent jets. Wave packets are large-scale turbulent structures that are correlated and advected over distances that are large compared to the integral scales of turbulence, i.e., many jet diameters at the lowest frequencies. They are thought to be responsible for the peak noise radiated at shallow angles to the jet axis. Linear wave packet models based on the Parabolized Stability Equations (PSE) have been shown in the past to be in excellent agreement with statistical structures educed from experimental pressure and velocity data in subsonic jets. Here, we extend these models to supersonic jets and validate them using a Large Eddy Simulation (LES) database for an isothermal and a moderately heated Mach 1.5 turbulent jets. For supersonic jets, inlet conditions for PSE models are ambiguous, as a parallel flow stability analysis shows several unstable modes at the inlet cross section. We develop a bi-orthogonal decomposition and project the LES data onto the relevant families of instability waves. These serve as inlet conditions, including the amplitude and shape functions, for PSE solutions which are then favorably compared to the near-field pressure fields educed from LES.