CCSE Combustion

CCSE Combustion Team

Turbulent Vflame The combustion team in CCSE develops algorithms and software to simulate multiscale, multiphysics combustion applications in theoretical and practical industrial configurations on modern high-performance supercomputing hardware. The work has been continuously supported over 25 years by the US DOE applied math and high performance computing programs, including the Exacale Computing Project. Our algorithms, software and analysis tools are available at GitHub via the Pele suite, and feature:

Conservative, finite-volume discretizations
Block-structured adaptive mesh refinement, built on the AMReX library
Robust interface-capturing techniques for shocks and turbulence interactions
Approximate projection methods for low Mach number flows
Multiphysics coupling via generalized spectral deferred corrections
Arbitrary domain geometries via Embedded Boundary methods (in progress...)

CCSE Staff

Active Collaborators

Research

We are actively developing the Pele suite of software for low Mach and compressible reacting flows. The Pele codes are open-source, and available at GitHib for free download. We also conduct research in a number of areas of combustion science, most often with external collaborators, and are always looking to expand this list into new areas. Please contact Marc Day if you'd like to explore new opportunities. Below, is a sampling of the projects under active development using the Pele software.

	Pele Suite Development The Pele suite is supported by the Exascale Computing Project. The CCSE Combustion Team is focused primarily on the low Mach code, PeleLM, and associated analysis tools.
	Compressible and low Mach Premixed Flames Laminar and turbulent premixed flames in idealized and laboratory-relevant configurations, focusing on the interaction of turbulence and combustion chemistry and the impact of Lewis number.
	Turbulent Diffusion Flames Turbulent diffusion flames have many practical applications. Our work is focused on the jet, and multi-pulsed burning of diesel fuel surrogates.
	Multi-brachial Flames Triple-flames, and penta-brachial flames are idealizations of the flame stabilizations in more complex scenarios, and exhibit many features of the local balances of chemistry, flow straining and species diffusion in an environment much more ammenable to analysis.
	Electric-field Assisted Combustion Flames naturally produce a small amount of charged species, and these can be manipulated externally by applying electric fields across the reaction zone. There are a range of interesting, and potentially practical applications of this phenomena.
	Uncertainty Quantification, Model Fitting, Parameter Estimation Modeling plays a key role in combustion simulations: from inputs specification to the simulation procedure and analysis of the results. MCMC, particle filters and black-box optimization are used explore and build reduced, low-dimensional models of complex reacting flow systems.
	Discretizations and Multiphysics Coupling Detailed effects of various discretization approaches on computed turbulent spectra; approaches to couple physical processes that exhibit a wide range of temporal and spatial scales.
	Wildfires and Pool Fires Simulating fire at much larger scales introduces an entirely new set of complications. These flames exhibit large-scale pulsing and vortical patterns and complex propagation.
	Machine learning in Combustion Simulation and Analysis Machine learning approaches can be useful for constructing reduced order models of complex systems.