Needed: A Next Generation Combustion LES Code

Abstract

Advances in computers and Computational Fluid Dynamics (CFD) technology now make it possible to understand the detailed flow physics of unsteady flow and thermal behavior through the use of combustion Large Eddy Simulation (LES).  LES methods solve the time-accurate conservation equations for mass, momentum, energy, and chemical species over the energy containing large length scales.  For smaller dissipative scales, LES employs subgrid models to compute turbulence and turbulence chemistry interaction.  In this way, combustion LES calculations numerically resolve unsteady vortices and eddies that occur in shear layers, boundary layers, and acoustic pressures waves.  Using this advanced LES methodology, researchers aim to improve the predictive capability of combustion CFD tools; address issues associated with combustion dynamics and other unsteady events; and study design concepts proposed for the Integrated High-Performance Turbine Engine Technology, Joint Strike Fighter (JSF), Versatile Affordable Advanced Turbine Engine programs, and other advanced combustion systems for gas turbines.

Although CFD scientists have continuously improved combustion LES modeling in recent years, additional advances are necessary before engineers can reliably use LES in everyday combustor design.  Scientists must further improve subgrid models for spray, turbulence, and turbulence chemistry interaction, along with developing more efficient solution algorithms to reduce the turnaround time for associated calculations.  We propose to develop a next generation combustion LES code (the DREAM Code) starting from a clean sheet of paper.  The DREAM code will encompass the latest strategies of improving accuracy and reducing run times, including programming language (C++), massive parallelization, temporal adaptive meshing, fourth-order temporal and spatial differencing on arbitrary unstructured grids, implicit LES subgrid turbulence, etc.  Only by starting from scratch can the code’s architecture be assembled in a way to realize efficient use of advanced computing technologies.