A pseudospectral method for direct numerical simulation of low-Mach, variable-density, turbulent flows

TL;DR

This paper introduces a new pseudospectral algorithm for direct numerical simulation of low-Mach, variable-density turbulent flows, extending existing methods to handle large density ratios with high accuracy and stability.

Contribution

It develops a novel pseudospectral method that removes pressure solve, enabling efficient, stable simulations of variable-density flows with large density ratios.

Findings

Achieves second-order temporal accuracy.

Demonstrates stability for density ratios up to ~25.7.

Provides an efficient matrix-free iterative solution.

Abstract

A novel algorithm for the direct numerical simulation of the variable-density, low-Mach Navier-Stokes equations extending the method of Kim, Moin, and Moser (1987) for incompressible flow is presented here. A Fourier representation is employed in the two homogeneous spatial directions and a number of discretizations can be used in the inhomogeneous direction. The momentum is decomposed into divergence- and curl-free portions which allows the momentum equations to be rewritten, removing the need to solve for the pressure. The temporal discretization is based on an explicit, segregated Runge-Kutta method and the scalar equations are reformulated to directly address the redundancy of the equation of state and the mass conservation equation. An efficient, matrix-free, iterative solution of the resulting equations allows for second-order accuracy in time and numerical stability for large density ratios, which is demonstrated for ratios up to $\sim 25.7$.