A generic and self-adapting method for flame detection and thickening in the thickened flame model

A generic and self-adapting method for flame front detection and thickening is presented. This approach relies solely on geometric considerations and unlike previous thickening methods does not need any parameterization nor preliminary calibration. The detection process is based on the analysis of the curvature of a test function, associating a bell-curve shape to a flame front. Once the front is located, the front thickness is also evaluated from the test function, allowing (1) a thickening restricted to under-resolved flame regions, (2) a self-adapting thickening of the front. The thickening process is finally applied to the detected front, over a normal-to-the-flame distance, using a Lagrangian point-localization algorithm. The method was developed and implemented in an unstructured and massively parallel environment and is therefore directly usable for the computation of complex configurations. Three test cases are presented to validate the methodology, ranging from a one-dimensional laminar premixed flame to the VOLVO turbulent premixed flame. © 2019 The Combustion Institute