Non-Dissipative LES Design Tool for Turbulent Combustion and Acoustics Analysis of Gas Turbine Combustors

Combustion LES is a high fidelity CFD method that can analyze transient/acoustic challenges that previously were impossible to analyze with conventional CFD. However, excessive calculation times hinder the use of LES as an effective design tool. To make LES more practical, we propose to develop and validate a next generation combustion LES code that can solve 10M cell grids in less than one day, i.e. a 100 times speed-up over current codes. We will exploit the full potential of the latest emerging technologies to develop a new code (named the DREAM code) starting from scratch, rather than being handicapped with older architecture and computer/software technology. We will focus on developing a framework with superior computational efficiency (run time reduction), robustness, and accuracy. In the Phase II effort, we will use/develop the following advanced methods/capabilities in the DREAM code: 1) advanced programming language (C++) and practices; 2) massively parallel architecture, 3) fourth order or higher temporal and spatial differencing on unstructured meshes; 4) adaptive mesh refinement (static and dynamic); 5) dynamic load balancing; 6) coupled solver; 7) practical and robust LES subgrid models for turbulence, turbulent-combustion interaction and chemistry; and 8) Lagrangian spray tracking with atomization and vaporization models. Time/costs constraints will limit validation in Phase II. Extensive validation and application are planned for future Phase IIE and/or Phase III efforts.