DI-UMONS : Dépôt institutionnel de l’université de Mons

Recherche transversale
Rechercher
(titres de publication, de périodique et noms de colloque inclus)
2021-05-25 - Travail avec promoteur/Doctorat - Anglais - 246 page(s)

Cordier Marie , "Large eddy simulation of MILD combustion, application to free-jets and a lab-scale furnace", soutenue le 2021-05-25

  • Codes CREF : Recherche énergétique (DI2290), Transfert de chaleur (DI2211), Combustion (DI2212)
  • Unités de recherche UMONS : Fluides-Machines (F702), Thermique et Combustion (F704)
  • Instituts UMONS : Institut de Recherche en Energétique (Energie)

Abstract(s) :

(Anglais) In the current environmental context, it is relevant to develop new industrial technologies to reduce pollutant emissions. MILD (Moderate or Intense Low oxygen Dilution) combustion, also called flameless or diluted combustion, is a combustion mode that allows to significantly reduce nitrogen oxides emissions. Hence, it is interesting to consider alternative burner designs to switch from classic combustion to MILD combustion. Such modifications must preliminarily be supported by numerical simulation. The purpose of this thesis is to develop and assess a methodology to simulate the reactive flow in industrial furnaces operating in MILD conditions, using Large Eddy Simulation (LES). Numerical simulation of turbulent combustion is challenging as it involves many physical coupled phenomena: turbulence, chemistry and radiation. Currently, the computational resources do not allow the use of meshes that are fine enough to solve these phenomena directly. Therefore, models are generally applied to compensate for the unresolved part of the flow, the combustion and the radiation. Regarding the flow dynamics, LES allows to solve the largest scales of the turbulence and a subgrid-scale model compensates for the effect of the smallest scales. The literature shows that RANS (Reynolds Averaged Navier-Stokes) has mainly been used for numerical simulation of industrial furnaces. More recently, and thanks to the computational power growth, LES investigations have become more widespread. We propose here to move from a RANS modelling to an LES approach and to assess the improvement of the numerical results. The methodology resulting from the investigations carried out during this work, proposed to simulate MILD furnaces, mainly consists in applying the LES and the Finite Rate Chemistry (FRC) equations combined with a skeletal kinetic scheme and a discrete ordinate method for radiation. The YALES2 code developed at CORIA, is used here to carry out high-performance computing (HPC). Two academic flames, the Delft and Adelaide Jets-in Hot-Coflow (JHC), are used to validate the methodology sequentially. After that, the entire method is applied to a 30kW furnace, fed with natural gas and operating in MILD combustion regime.