In canonical jet-in-hot-coflow (JHC) burners that emulate Moderate or Intense Low-oxygen Dilution (MILD) combustion, all standard turbulence-chemistry interaction closures - steady flamelet, the Eddy Dissipation Concept (EDC), Lagrangian transported composition PDF (LPDF), and multi-environment Eulerian PDF (MEPDF) - reproduce mean velocity and major species but fail on the minor radicals CO and OH and their RMS fields, and their parameters require case-by-case tuning. The source assessment (De and Dongre, Flow, Turbulence and Combustion 94:439-478, 2015) and the subsequent literature trace this failure to six linked deficiencies: an unexplained minor-species prediction error; MEPDF ill-conditioning, realizability violations and boundary-condition ambiguity; Monte-Carlo statistical and bias errors in transported PDF methods; the absence of a general micro-mixing model prescription; no standardized inlet scalar-fluctuation boundary conditions; and no first-principles calibration of the EDC fine-structure constants at low Damköhler number. Later work (2015-2026) has delivered a generalised EDC validated on twelve JHC flames, direct-numerical-simulation-based a priori closure assessments, locally estimated mixing constants, and large-eddy-simulation (LES)/transported-PDF studies emphasizing resolved molecular diffusion - yet each of these issues remains contested or unresolved, and no existing framework predicts mean and RMS CO and OH across the Adelaide JHC database - the only JHC database with measured species fields - without case-specific tuning. The problem asks for such a predictive, non-tuned framework together with the attribution of the persistent minor-species failure.

Background

MILD (Moderate or Intense Low-oxygen Dilution) combustion is a combustion regime achieved by strongly diluting and preheating the reactants, typically by internal or external recirculation of exhaust gas, so that the reaction zone is distributed rather than confined to a thin flame front, peak temperatures are lowered, and thermal NOx formation is drastically reduced. Because of its low pollutant emissions and fuel flexibility it is attractive for industrial furnaces and gas turbines. The regime is studied in the laboratory with jet-in-hot-coflow (JHC) burners, in which a fuel jet issues into a hot, oxygen-lean coflow produced by a secondary burner: the Delft-JHC burner and the Adelaide JHC burner (flames HM1, HM2, HM3 with roughly 3, 6 and 9 percent oxygen in the coflow). The Adelaide flames' temperature and species fields, including CO and OH, were measured with laser-based techniques by Dally, Karpetis and Barlow (2002) and follow-up campaigns, while the Delft-JHC database provides velocity, temperature and turbulence measurements but no species data; jointly the two databases are the standard benchmarks for MILD combustion models. The central numerical difficulty is the turbulence-chemistry interaction (TCI) closure problem: in a turbulent reacting flow the mean chemical source term, a strongly nonlinear function of temperature and composition, cannot be evaluated from mean values alone, so a closure must represent the unresolved fluctuations. Classical closures were derived for conventional flames. The Eddy Dissipation Concept (EDC) of Magnussen partitions each computational cell into a 'fine-structure' region, treated as a perfectly stirred reactor where chemistry occurs, and surrounding fluid; the volume fraction and residence time of the fine structures are set by two constants derived from assumptions of isotropic, high-Reynolds-number turbulence and a fast-chemistry limit. Presumed-shape PDF and flamelet approaches (such as the steady flamelet model used in the source study) assume the local flame structure is that of a one-dimensional laminar flame, which fails when reaction zones are distributed. The Lagrangian transported composition PDF method (LPDF) represents the joint probability density function of composition with notional Monte-Carlo particles, so the chemical source term appears in closed form, but molecular mixing must be closed by a micro-mixing model: Interaction-by-Exchange-with-the-Mean (IEM, all scalars relax to the local mean at one rate), Coalescence-Dispersion (CD, randomly paired particles mix), or the Euclidean Minimum Spanning Tree (EMST) model, which mixes only neighboring particles in composition space and hence respects localness. The multi-environment Eulerian PDF method (MEPDF, a direct-quadrature-moment-type closure) discretizes the composition PDF into a few environments solved by Eulerian transport equations; it avoids Monte-Carlo noise but introduces correction terms whose source terms are ill-conditioned and can violate realizability (the requirement that the approximated PDF remain non-negative) for multi-component systems, as established by Jaishree and Haworth (2012) and reported by De and Dongre. In MILD conditions the Damköhler number (the ratio of turbulent to chemical time scales) is of order unity or smaller, chemical time scales are long, and molecular (differential) diffusion of hydrogen-bearing and radical species competes with turbulent mixing, so several closure assumptions break at once. De and Dongre (2015) simulated the Delft-JHC flame at Re=4100 and the Adelaide flames HM1-HM3 at Re=10000 with two-dimensional Reynolds-averaged Navier-Stokes (RANS) equations using the steady flamelet, EDC, LPDF (with IEM, CD and EMST micro-mixing, a 19-species reduced mechanism, and about 30 particles per cell) and MEPDF models. Mean velocity and turbulent kinetic energy were captured well and LPDF reproduced most mean and RMS fields, but every model family failed on the minor species CO and OH; transported-PDF predictions required inlet temperature fluctuations (not just means) to reproduce RMS fields; and EDC improved only when its chemical time-scale constant was enlarged - an empirical adjustment without first-principles derivation. These observations define the unresolved modeling problem stated below.

Problem Statement

Make simulation of MILD combustion in the canonical jet-in-hot-coflow burners genuinely predictive, and identify the mechanism behind its best-known failure. Concretely: obtain, without case-specific parameter tuning, quantitatively accurate radial profiles of the mean and RMS of temperature at the measured axial stations of the standard experimental flame set - the Delft-JHC flame at Re=4100 and the Adelaide JHC flames HM1, HM2 and HM3 at Re=10000 with coflow oxygen levels of 3-9 percent - together with, for the Adelaide flames, where laser species measurements exist, quantitatively accurate radial profiles of the mean and RMS of the major species (CH4, H2, O2, H2O, CO2) and of the minor radicals CO and OH, and establish which physical or numerical mechanism causes the persistent CO and OH discrepancy. Here "quantitatively accurate" means agreement with the measurements within their experimental uncertainty at the measured axial stations, and "without case-specific parameter tuning" means that one and the same set of closure parameters - whether derived from physical reasoning or from a single calibration common to the whole flame set - is applied to all flames of the set, with no per-flame readjustment. The source assessment (De and Dongre, 2015) and the subsequent literature document six linked modeling challenges that frame this objective: (1) the unexplained failure of all tested closures to capture mean and RMS CO and OH, whose possible origins - the micro-mixing closure, the chemical mechanism, the turbulence parametrization, or inadequate inlet data - have not been disentangled; (2) the ill-conditioned source terms, realizability violations for multi-component systems, and ambiguous boundary-condition specification of the multi-environment Eulerian PDF formulation; (3) the Monte-Carlo statistical and bias errors of transported-PDF particle methods and their sensitivity to particle-tracking schemes and particle density; (4) the absence of a general, physically grounded prescription for selecting the micro-mixing model in the MILD regime and of understanding of its interaction with detailed chemistry including differential diffusion; (5) the lack of a standardized method to prescribe physically consistent inlet scalar-fluctuation spectra for JHC burners, which materially control RMS predictions; and (6) the lack of a first-principles calibration of the EDC fine-structure constants in low-Damköhler MILD flows, where enlarging the chemical time-scale constant improves predictions but has no first-principles justification. These six linked modeling challenges motivate the objective stated here. Obtain a closure, from any of the established model families or a new one, whose parameters follow from physical reasoning or from a single calibration common to the flame set and which meets the quantitative-accuracy requirement stated above across the flame set, and identify the physical or numerical error source responsible for the persistent CO and OH discrepancy and remedy it so that the quantitative-accuracy requirement is met.

References

  1. Ashoke De and Akshay Dongre, 'Assessment of Turbulence-Chemistry Interaction Models in MILD Combustion Regime', Flow, Turbulence and Combustion 94:439-478 (2015), DOI 10.1007/s10494-014-9587-8, https://doi.org/10.1007/s10494-014-9587-8 (arXiv:2101.09198, https://arxiv.org/abs/2101.09198)
  2. B. B. Dally, A. N. Karpetis and R. S. Barlow, 'Structure of turbulent non-premixed jet flames in a diluted hot coflow', Proceedings of the Combustion Institute 29:1147-1154 (2002), DOI 10.1016/s1540-7489(02)80145-6, https://doi.org/10.1016/s1540-7489(02)80145-6
  3. J. Jaishree and D. C. Haworth, 'Comparisons of Lagrangian and Eulerian PDF methods in simulations of non-premixed turbulent jet flames with moderate-to-strong turbulence-chemistry interactions', Combustion Theory and Modelling 16:435-463 (2012), DOI 10.1080/13647830.2011.633349, https://doi.org/10.1080/13647830.2011.633349
  4. Alessandro Parente, Mohammad Rafi Malik, Francesco Contino, Alberto Cuoci and Bassam B. Dally, 'Extension of the Eddy Dissipation Concept for turbulence/chemistry interactions to MILD combustion', Fuel 163:98-111 (2016), DOI 10.1016/j.fuel.2015.09.020, https://doi.org/10.1016/j.fuel.2015.09.020
  5. M. J. Evans, P. R. Medwell and Z. F. Tian, 'Modeling Lifted Jet Flames in a Heated Coflow Using an Optimized Eddy Dissipation Concept Model', Combustion Science and Technology 187:1093-1109 (2015), DOI 10.1080/00102202.2014.1002836, https://doi.org/10.1080/00102202.2014.1002836
  6. M. J. Evans, C. Petre, P. R. Medwell and A. Parente, 'Generalisation of the eddy-dissipation concept for jet flames with low turbulence and low Damköhler number', Proceedings of the Combustion Institute 37:4497-4505 (2019), DOI 10.1016/j.proci.2018.06.017, https://doi.org/10.1016/j.proci.2018.06.017
  7. Amir Mardani, 'Optimization of the Eddy Dissipation Concept (EDC) model for turbulence-chemistry interactions under hot diluted combustion of CH4/H2', Fuel 191:114-129 (2017), DOI 10.1016/j.fuel.2016.11.056, https://doi.org/10.1016/j.fuel.2016.11.056
  8. Michał T. Lewandowski and Ivar S. Ertesvåg, 'Analysis of the Eddy Dissipation Concept formulation for MILD combustion modelling', Fuel 224:687-700 (2018), DOI 10.1016/j.fuel.2018.03.110, https://doi.org/10.1016/j.fuel.2018.03.110
  9. Michał T. Lewandowski, Alessandro Parente and Jacek Pozorski, 'Generalised Eddy Dissipation Concept for MILD combustion regime at low local Reynolds and Damköhler numbers. Part 1: Model framework development', Fuel 278:117743 (2020), DOI 10.1016/j.fuel.2020.117743, https://doi.org/10.1016/j.fuel.2020.117743
  10. Michał T. Lewandowski, Zhiyi Li, Alessandro Parente and Jacek Pozorski, 'Generalised Eddy Dissipation Concept for MILD combustion regime at low local Reynolds and Damköhler numbers. Part 2: Validation of the model', Fuel 278:117773 (2020), DOI 10.1016/j.fuel.2020.117773, https://doi.org/10.1016/j.fuel.2020.117773
  11. Ivar S. Ertesvåg, 'Scrutinizing proposed extensions to the Eddy Dissipation Concept (EDC) at low turbulence Reynolds numbers and low Damköhler numbers', Fuel 309:122032 (2022), DOI 10.1016/j.fuel.2021.122032, https://doi.org/10.1016/j.fuel.2021.122032
  12. Zhiyi Li and Alessandro Parente, 'A Review of the Numerical Investigations of Jet-In-Hot-Coflow Burner With Reactor-Based Models', Frontiers in Mechanical Engineering 6:512501 (2020), DOI 10.3389/fmech.2020.512501, https://doi.org/10.3389/fmech.2020.512501
  13. Hua Zhou, Tianwei Yang, Bassam Dally and Zhuyin Ren, 'LES/TPDF investigation of the role of reaction and diffusion timescales in the stabilization of a jet-in-hot-coflow CH4/H2 flame', Combustion and Flame 211:477-492 (2020), DOI 10.1016/j.combustflame.2019.10.021, https://doi.org/10.1016/j.combustflame.2019.10.021
  14. Lorenzo Giuntini, Matteo Savarese, Chiara Galletti and Alessandro Parente, 'Local estimation of the mixing model constant for transported PDF combustion models', Fuel 357:130022 (2024), DOI 10.1016/j.fuel.2023.130022, https://doi.org/10.1016/j.fuel.2023.130022
  15. Sangtae Jeon and Yongmo Kim, 'Multi-environment Probability Density Function Modeling for Turbulent CH4 Flames under Moderate or Intense Low-Oxygen Dilution Combustion Conditions with Recirculated Flue Gases', Energy & Fuels 31:8685-8697 (2017), DOI 10.1021/acs.energyfuels.7b01060, https://doi.org/10.1021/acs.energyfuels.7b01060
  16. S. B. Nourani Najafi, J. A. van Oijen, H. B. Levinsky and A. V. Mokhov, '(Non)Equilibrium of OH and Differential Transport in MILD Combustion: Measured and Computed OH Fractions in a Laminar Methane/Nitrogen Jet in Hot Coflow', Energy & Fuels 35:6798-6806 (2021), DOI 10.1021/acs.energyfuels.1c00583, https://doi.org/10.1021/acs.energyfuels.1c00583
  17. Marie Cordier, Pierre Bénard, Paul Lybaert, Ward De Paepe and Laurent Bricteux, 'On the Need for Turbulence Chemistry Interaction Modelling in Highly Resolved Large-Eddy Simulations of Mild Combustion', Flow, Turbulence and Combustion 108:509-538 (2022), DOI 10.1007/s10494-021-00282-X, https://doi.org/10.1007/s10494-021-00282-X
  18. Arthur Péquin, Michael J. Evans, Alfonso Chinnici, Paul R. Medwell and Alessandro Parente, 'The reactor-based perspective on finite-rate chemistry in turbulent reacting flows: A review from traditional to low-emission combustion', Applications in Energy and Combustion Science 16:100201 (2023), DOI 10.1016/j.jaecs.2023.100201, https://doi.org/10.1016/j.jaecs.2023.100201
  19. Hazem S. A. M. Awad, Khalil Abo-Amsha, Umair Ahmed, Nedunchezhian Swaminathan and Nilanjan Chakraborty, 'A Priori Direct Numerical Simulation Assessment of MILD Combustion Modelling in the Context of Reynolds Averaged Navier-Stokes Simulations', Flow, Turbulence and Combustion 111:799-823 (2023), DOI 10.1007/s10494-023-00422-5, https://doi.org/10.1007/s10494-023-00422-5
  20. Yudong Wang, Fang Wang, Sinan Lei, Jie Jin, Dasheng Wei, Feng Feng and Peng Liao, 'A turbulence-chemistry recursive mixing model for TPDF simulations and its appraisal', Chinese Journal of Aeronautics 39:104123 (2026), DOI 10.1016/j.cja.2026.104123, https://doi.org/10.1016/j.cja.2026.104123