Combustion modeling in the Lattice Boltzmann framework

The industry relies increasingly on numerical simulation for designing, improving, and even validating new combustion devices (engine, burner, furnace, etc.). Today, numerical combustion modelling relies almost exclusively on numerical codes solving the Navier-Stokes equations. 

The Lattice Boltzmann solvers are very different from these codes, intending to solve a discrete variant of the Boltzmann equation. This type of flow solver is progressing rapidly, however, in turbulent flows configurations. The results obtained with Lattice Boltzmann methods (LBM) have shown to be excellent for aerodynamic applications, motivating intensive development of new methods.

Lattice Boltzmann methods applied to industrial applications are recent, however, so few models are able to deal with multiphase flows, and almost none with reactive (combusting) flows. 

The development of combustion modelling within the LBM framework is the topic of this study, following our recent works. 

Extending the LBM capabilities to combustion requires a profound rethinking of existing methods developed within the Navier-Stokes framework. 

The team has recently made important steps in proving the LBM ability to tackle such flows, at a cost significantly reduced compared to classical results (divided by 5 or more), on relatively simple configurations.

The objective of this PhD proposal is to combine the LB method to turbulent combustion models. We will first focus on the various variants of the thickened flame model, investigating the possible extensions to make the most of the LBM solver.

Along with turbulent combustion modelling, the model will have to be compatible of multiple grid levels, complex boundary conditions and realistic geometry. 

Strong background in scientific computing (c++ preferred), reactive and/or multiphase flows.

English.

LBM, HPC, combustion

Intended Start date: October 2020.