Direct Numerical Simulation of a Statistically Stationary and Streamwise Periodic Turbulent Shear Layer
Victor Zendejas Lopez, California Institute of Technology
Turbulent shear layers are one of the most studied turbulent free shear flows and are commonly observed in a broad range of fields, including combustion, aerodynamics, oceanic and atmospheric flows. These shear layers form when two initially separated parallel streams with different velocities begin to mix. Unfortunately, most computer simulations of turbulent shear layers are computationally inefficient, requiring long streamwise domains due to the slow shear layer growth, resulting in high computational overhead.
In this work, we seek to develop a method for direct numerical simulation (DNS) of incompressible turbulent shear layers in a streamwise periodic domain using the Navier-Stokes equations, without the need of auxiliary simulations. We use anticipated self-similarity to solve the equations in a normalized coordinate system to allow for streamwise periodicity. This reduces the computational cost by decreasing the overall size of the streamwise domain and provides faster converging statistics for the turbulent shear layer.
This method extends the work of Ruan and Blanquart (2021), which originally applied this method to turbulent flat- plate boundary layers and demonstrated computational costs savings by an estimated one to two orders of magnitude. However, when developing this framework for the shear layer configuration, it is evident that a deeper understanding of the boundary layer formulation is required. This necessity and its implications are discussed.
Abstract Author(s): Victor Zendejas Lopez, Guillaume Blanquart