An assessment of turbulence models for linear hydrodynamic stability analysis of strongly swirling jets

TL;DR

This paper evaluates turbulence models for linear stability analysis of strongly swirling jets, finding that anisotropic eddy viscosity models improve accuracy over isotropic models in predicting flow stability characteristics.

Contribution

It demonstrates that linear stability analysis can be accurate for turbulent flows with anisotropic eddy viscosities, even when URANS models with isotropic viscosities fail.

Findings

Anisotropic eddy viscosity models match experimental data better.

Isotropic eddy viscosity models from $k- extepsilon$ do not predict mean flow or stability accurately.

Linear stability analysis can be effective without URANS-based mean flow predictions.

Abstract

Linear stability analysis has proven to be a useful tool in the analysis of dominant coherent structures, such as the von K\'{a}rm\'{a}n vortex street and the global spiral mode associated with the vortex breakdown of swirling jets. In recent years, linear stability analysis has been applied successfully to turbulent time-mean flows, instead of laminar base-flows, \textcolor{black}{which requires turbulent models that account for the interaction of the turbulent field with the coherent structures. To retain the stability equations of laminar flows, the Boussinesq approximation with a spatially nonuniform but isotropic eddy viscosity is typically employed. In this work we assess the applicability of this concept to turbulent strongly swirling jets, a class of flows that is particularly unsuited for isotropic eddy viscosity models. Indeed we find that unsteady RANS simulations only match with experiments with a Reynolds stress model that accounts for an anisotropic eddy viscosity. However, linear stability analysis of the mean flow is shown to accurately predict the global mode growth rate and frequency if the employed isotropic eddy viscosity represents a least-squares approximation of the anisotropic eddy viscosity. Viscosities derived from the $k-\epsilon$ model did not achieve a good prediction of the mean flow nor did they allow for accurate stability calculations. We conclude from this study that linear stability analysis can be accurate for flows with strongly anisotropic turbulent viscosity and the capability of the Boussinesq approximation in terms of URANS-based mean flow prediction is not a prerequisite.