Scientists of Lithuanian Energy Institute have been developing an open-source turbulent combustion solver flameFoam, based on OpenFOAM toolkit. OpenFOAM is the most popular open-source CFD solution. It is a broadly used toolkit that allows custom implementations of finite volume method-based solvers of partial differential equation systems. OpenFOAM solvers have been developed and applied for the simulations of a wide range of problems.

The stable and development versions of flameFoam are built on the OpenFOAM toolkit version 9. flameFoam is a comparatively new solver and still in the development phase and requires model refinement and validation.  Code to code analysis and comparisons with experiments showed that flameFoam mostly corresponds to the state-of-the-art currently achieved in the field.

Initially flameFoam has been developed for practically relevant simulations, therefore only simplified combustion modeling approach was implemented – in the stable version combustion is modeled using Turbulent Flame speed Closure (TFC) approach, applicable with RANS turbulence modeling, and turbulent flame speed correlations (Bradley, Bray and Zimont).

However, recently flameFoam capabilities have been expanded further.

First, TFC model lacks support for the simulation of (quasi-)laminar combustion regime. Therefore, additional model, Extended TFC, addressing slow combustion regime and turbulent flame development, has been implemented and is currently undergoing testing.

Second, to support more detailed studies, combustion model usable with LES turbulence had to be added. LES combustion model based on flame surface density approach with Charlette correlation for SGS flame wrinkling factor was selected and implemented into flameFoam.

Third, in the stable version, laminar flame velocity can be set by the user (constant) or estimated using Malet correlation for lean mixtures. To expand flameFoam application domain, additional method for laminar burning velocity (LBV) estimation was needed. To address this need LEI scientists developed an artificial neural network (ANN) of LBV. It has been trained on the experimental data available from open literature and implemented in flameFoam. Initial calculations showed no significant deviations from other methods or experiments and no significant reduction in computation speed.

Further development and validation of flameFoam is currently on-going at LEI.