Deep artificial neural networks (ANNs) are used for modeling sub-grid scale mixing quantities such as the filtered density function (FDF) of the mixture fraction and the conditional scalar dissipation rate. A deep ANN with four hidden layers is trained with carrier-phase direct numerical simulations (CP-DNS) of turbulent spray combustion. A priori validation corroborates that ANN predictions of the mixture fraction FDF and the conditional scalar dissipation rate are in very good agreement with CP-DNS data. ANN modeled solutions show much better performance with a mean error of around 1%, which is one order of magnitude smaller than that of standard modeling approaches such as the β-FDF and its modified version. The predicted conditional scalar dissipation rate agrees very well with CP-DNS data over the entire mixture fraction space, whereas conventional models derived for pure gas phase combustion fail to describe ⟨N|ξ = η⟩ in regions with higher mixture fraction and low probability. In the second part of this paper, uncertainties associated with ANN predictions are analyzed. It is shown that a suitable selection of training sets can reduce the size of the necessary test database by ∼50% without compromising the accuracy. Feature importance analysis is used to analyze the importance of different combustion model parameters. While the droplet evaporating rate, the droplet number density, and the mixture fraction remain the dominant features, the influence of turbulence related parameters only becomes important if turbulence levels are sufficiently high.
REFERENCES
1.
P.
Jenny
, D.
Roekaerts
, and N.
Beishuizen
, “Modeling of turbulent dilute spray combustion
,” Prog. Combust. Sci. Technol.
38
, 846
–887
(2012
).2.
S.
Sreedhara
and K. Y.
Huh
, “Conditional statistics of nonreacting and reacting sprays in turbulent flows by direct numerical simulation
,” Proc. Combust. Inst.
31
, 2335
–2342
(2007
).3.
H.
Wang
, K.
Luo
, and J.
Fan
, “Direct numerical simulation and CMC (conditional moment closure) sub-model validation of spray combustion
,” Energy
46
, 606
–617
(2012
).4.
S.
Ukai
, A.
Kronenburg
, and O. T.
Stein
, “Large eddy simulation of dilute acetone spray flames using CMC coupled with tabulated chemistry
,” Proc. Combust. Inst.
35
, 1667
–1674
(2015
).5.
M. P.
Sitte
and E.
Mastorakos
, “Modelling of spray flames with doubly conditional moment closure
,” Flow, Turbul. Combust.
99
, 933
–954
(2017
).6.
7.
M.
Chrigui
, J.
Gounder
, A.
Sadiki
, A. R.
Masri
, and J.
Janicka
, “Partially premixed reacting acetone spray using LES and FGM tabulated chemistry
,” Combust. Flame
159
, 2718
–2741
(2012
).8.
S.
De
and S. H.
Kim
, “Large eddy simulation of dilute reacting sprays: Droplet evaporation and scalar mixing
,” Combust. Flame
160
, 2048
–2066
(2013
).9.
A.
Rittler
, L.
Deng
, I.
Wlokas
, and A. M.
Kempf
, “Large eddy simulations of nanoparticle synthesis from flame spray pyrolysis
,” Proc. Combust. Inst.
36
, 1077
–1087
(2017
).10.
Y.
Hu
and R.
Kurose
, “Nonpremixed and premixed flamelets LES of partially premixed spray flames using a two-phase transport equation of progress variable
,” Combust. Flame
188
, 227
–242
(2018
).11.
P.
Domingo-Alvarez
, P.
Bénard
, V.
Moureau
, G.
Lartigue
, and F.
Grisch
, “Impact of spray droplet distribution on the performances of a kerosene lean/premixed injector
,” Flow, Turbul. Combust.
104
, 421
–450
(2020
).12.
S.
Navarro-Martinez
, A.
Kronenburg
, and F.
Di Mare
, “Conditional moment closure for large eddy simulations
,” Flow, Turbul. Combust.
75
, 245
–274
(2005
).13.
V.
Raman
, H.
Pitsch
, and R.
Fox
, “Hybrid large-eddy simulation/Lagrangian filtered-density-function approach for simulating turbulent combustion
,” Combust. Flame
143
, 56
–78
(2005
).14.
A.
Triantafyllidis
and E.
Mastorakos
, “Implementation issues of the conditional moment closure model in large eddy simulations
,” Flow, Turbul. Combust.
84
, 481
–512
(2010
).15.
M.
Ihme
and H.
Pitsch
, “Prediction of extinction and reignition in nonpremixed turbulent flames using a flamelet/progress variable model: 2. Application in LES of sandia flames D and E
,” Combust. Flame
155
, 90
–107
(2008
).16.
H.-W.
Ge
and E.
Gutheil
, “Probability density function (PDF) simulation of turbulent spray flows
,” Atomisation Sprays
16
, 531
–542
(2006
).17.
H.-W.
Ge
and E.
Gutheil
, “Simulation of a turbulent spray flame using coupled PDF gas phase and spray flamelet modeling
,” Combust. Flame
153
, 173
–185
(2008
).18.
E.
Madadi-Kandjani
, R. O.
Fox
, and A.
Passalacqua
, “Application of the Fokker-Planck molecular mixing model to turbulent scalar mixing using moment methods
,” Phys. Fluids
29
, 065109
(2017
).19.
N.
Kim
, K.
Jung
, and Y.
Kim
, “Multi-environment pdf modeling for n-dodecane spray combustion processes using tabulated chemistry
,” Combust. Flame
192
, 205
–220
(2018
).20.
E. E.
O’Brien
and T. L.
Jiang
, “The conditional dissipation rate of an initially binary scalar in homogeneous turbulence
,” Phys. Fluids A
3
, 3121
–3123
(1991
).21.
I. S.
Kim
and E.
Mastorakos
, “Simulations of turbulent non-premixed counterflow flames with first-order conditional moment closure
,” Flow, Turbul. Combust.
76
, 133
–162
(2006
).22.
A. J. M.
Buckrell
and C. B.
Devaud
, “Investigation of mixing models and conditional moment closure applied to autoignition of hydrogen jets
,” Flow, Turbul. Combust.
90
, 621
–644
(2013
).23.
A.
Garmory
and E.
Mastorakos
, “Capturing localised extinction in sandia flame F with LES-CMC
,” Proc. Combust. Inst.
33
, 1673
–1680
(2011
).24.
S. S.
Girimaji
, “On the modeling of scalar diffusion in isotropic turbulence
,” Phys. Fluids A
4
, 2529
–2537
(1992
).25.
V.
Eswaran
and S. B.
Pope
, “Direct numerical simulations of the turbulent mixing of a passive scalar
,” Phys. Fluids
31
, 506
–520
(1988
).26.
S. B.
Pope
, “Mapping closures for turbulent mixing and reaction
,” Theor. Comput. Fluid Dyn.
2
, 255
–270
(1991
).27.
A.
Kronenburg
, R. W.
Bilger
, and J. H.
Kent
, “Computation of conditional average scalar dissipation in turbulent jet diffusion flames
,” Flow, Turbul. Combust.
64
, 145
–159
(2000
).28.
C. B.
Devaud
, R. W.
Bilger
, and T.
Liu
, “A new method of modeling the conditional scalar dissipation rate
,” Phys. Fluids
16
, 2004
–2011
(2004
).29.
A. Y.
Klimenko
and S. B.
Pope
, “The modelling of turbulent reacting flows based on multiple mapping conditioning
,” Phys. Fluids
15
(7
), 1907
–1925
(2003
).30.
P.
Vaishnavi
and A.
Kronenburg
, “Multiple mapping conditioning of velocity in turbulent jet flames
,” Combust. Flame
157
, 1863
–1865
(2010
).31.
K.
Vogiatzaki
, A.
Kronenburg
, M. J.
Cleary
, and J. H.
Kent
, “Multiple mapping conditioning of turbulent jet diffusion flames
,” Proc. Combust. Inst.
32
, 1679
–1685
(2009
).32.
K.
Vogiatzaki
, M. J.
Cleary
, A.
Kronenburg
, and J. H.
Kent
, “Modeling of scalar mixing in turbulent jet flames by multiple mapping conditioning
,” Phys. Fluids
21
, 025105
(2009
).33.
J.
Réveillon
and L.
Vervisch
, “Spray vaporization in nonpremixed turbulent combustion modeling: A single droplet model
,” Combust. Flame
121
, 75
–90
(2000
).34.
A. Y.
Klimenko
and R. W.
Bilger
, “Conditional moment closure for turbulent combustion
,” Prog. Energy Combust. Sci.
25
, 595
–688
(1999
).35.
B.
Wang
, A.
Kronenburg
, and O. T.
Stein
, “A new perspective on modelling passive scalar conditional mixing statistics in turbulent spray flames
,” Combust. Flame
208
, 376
–387
(2019
).36.
J.
Seo
and K. Y.
Huh
, “Analysis of combustion regimes and conditional statistics of autoigniting turbulent n-heptane sprays
,” Proc. Combust. Inst.
33
, 2127
–2134
(2011
).37.
Y. M.
Wright
, G.
DePaola
, K.
Boulouchos
, and E.
Mastorakos
, “Simulations of spray autoignition and flame establishment with two-dimensional CMC
,” Combust. Flame
143
, 402
–419
(2005
).38.
G.
Borghesi
, E.
Mastorakos
, and R. S.
Cant
, “Complex chemistry DNS of n-heptane spray autoignition at high pressure and intermediate temperature conditions
,” Combust. Flame
160
, 1254
–1275
(2013
).39.
M.
Ihme
, C.
Schmitt
, and H.
Pitsch
, “Optimal artificial neural networks and tabulation methods for chemistry representation in LES of a bluff-body swirl-stabilized flame
,” Proc. Combust. Inst.
32
, 1527
–1535
(2009
).40.
A. K.
Chatzopoulos
and S.
Rigopoulos
, “A chemistry tabulation approach via rate-controlled constrained equilibrium (RCCE) and artificial neural networks (ANNs), with application to turbulent non-premixed CH4/H2/N2 flames
,” Proc. Combust. Inst.
34
, 1465
–1473
(2013
).41.
L. L. C.
Franke
, A. K.
Chatzopoulos
, and S.
Rigopoulos
, “Tabulation of combustion chemistry via artificial neural networks (ANNs): Methodology and application to LES-PDF simulation of Sydney Flame L
,” Combust. Flame
185
, 245
–260
(2017
).42.
R.
Ranade
, G.
Li
, S.
Li
, and T.
Echekki
, “An efficient machine-learning approach for PDF tabulation in turbulent combustion closure
,” Combust. Sci. Technol.
2019
, 1
–20
.43.
O.
Owoyele
, P.
Kundu
, M. M.
Ameen
, T.
Echekki
, and S.
Som
, “Application of deep artificial neural networks to multi-dimensional flamelet libraries and spray flames
,” Int. J. Engine Res.
21
, 151
–168
(2020
).44.
A.
Seltz
, P.
Domingo
, L.
Vervisch
, and Z. M.
Nikolaou
, “Direct mapping from LES resolved scales to filtered-flame generated manifolds using convolutional neural networks
,” Combust. Flame
210
, 71
–82
(2019
).45.
J.
Xing
, K.
Luo
, H.
Pitsch
, H.
Wang
, Y.
Bai
, C.
Zhao
, and J.
Fan
, “Predicting kinetic parameters for coal devolatilization by means of artificial neural networks
,” Proc. Combust. Inst.
37
, 2943
–2950
(2019
).46.
M.
Gamahara
and Y.
Hattori
, “Searching for turbulence models by artificial neural network
,” Phys. Rev. Fluids
2
, 054604
(2017
).47.
Z.
Wang
, K.
Luo
, D.
Li
, J.
Tan
, and J.
Fan
, “Investigations of data-driven closure for subgrid-scale stress in large-eddy simulation
,” Phys. Fluids
30
, 125101
(2018
).48.
J.
Ling
, A.
Kurzawski
, and J.
Templeton
, “Reynolds averaged turbulence modelling using deep neural networks with embedded invariance
,” J. Fluid Mech.
807
, 155
–166
(2016
).49.
J.
Ling
and J.
Templeton
, “Evaluation of machine learning algorithms for prediction of regions of high Reynolds averaged Navier–Stokes uncertainty
,” Phys. Fluids
27
, 085103
(2015
).50.
M.
Bode
, M.
Gauding
, K.
Kleinheinz
, and H.
Pitsch
, “Deep learning at scale for subgrid modeling in turbulent flows
,” arXiv:1910.00928 (2019
).51.
A.
Moreau
, O.
Teytaud
, and J. P.
Bertoglio
, “Optimal estimation for large-eddy simulation of turbulence and application to the analysis of subgrid models
,” Phys. Fluids
18
, 105101
(2006
).52.
L.
Berger
, K.
Kleinheinz
, A.
Attili
, F.
Bisetti
, H.
Pitsch
, and M. E.
Mueller
, “Numerically accurate computational techniques for optimal estimator analyses of multi-parameter models
,” Combust. Theor. Modell.
22
, 480
–504
(2018
).53.
S.
Yao
, B.
Wang
, A.
Kronenburg
, and O. T.
Stein
, “Conditional scalar dissipation rate modeling for turbulent spray flames using artificial neural networks
,” Proc. Combust. Inst.
(unpublished) (2020
).54.
R. S.
Miller
and J.
Bellan
, “Direct numerical simulation of a confined three-dimensional gas mixing layer with one evaporating hydrocarbon-droplet-laden stream
,” J. Fluid Mech.
384
, 293
–338
(1999
).55.
P.
Schroll
, A. P.
Wandel
, R. S.
Cant
, and E.
Mastorakos
, “Direct numerical simulations of autoignition in turbulent two-phase flows
,” Proc. Combust. Inst.
32
, 2275
–2282
(2009
).56.
P.
Domingo
, L.
Vervisch
, and J.
Réveillon
, “DNS analysis of partially premixed combustion in spray and gaseous turbulent flame-bases stabilized in hot air
,” Combust. Flame
140
, 172
–195
(2005
).57.
D. H.
Wacks
and N.
Chakraborty
, “Flame structure and propagation in turbulent flame-droplet interaction: A direct numerical simulation analysis
,” Flow, Turbul. Combust.
96
, 1053
–1081
(2016
).58.
Q.
Khan
, S.
Baek
, and H.
Ghassemi
, “On the autoignition and combustion characteristics of kerosene droplets at elevated pressure and temperature
,” Combust. Sci. Technol.
179
(12
), 2437
–2451
(2007
).59.
F.
Wang
, R.
Liu
, M.
Li
, J.
Yao
, and J.
Jin
, “Kerosene evaporation rate in high temperature air stationary and convective environment
,” Fuel
211
, 582
–590
(2018
).60.
A.
Giusti
, M. P.
Sitte
, G.
Borghesi
, and E.
Mastorakos
, “Numerical investigation of kerosene single droplet ignition at high-altitude relight conditions
,” Fuel
225
, 663
–670
(2018
).61.
W. P.
Jones
and R. P.
Lindstedt
, “Global reaction schemes for hydrocarbon combustion
,” Combust. Flame
73
, 233
–249
(1988
).62.
B.
Wang
, A.
Kronenburg
, and O. T.
Stein
, “Modelling sub-grid passive scalar statistics in moderately dense evaporating sprays
,” Flow, Turbul. Combust.
103
, 519
–535
(2019
).63.
64.
65.
J. A.
Blasco
, N.
Fueyo
, C.
Dopazo
, and J.
Ballester
, “Modelling the temporal evolution of a reduced combustion chemical system with an artificial neural network
,” Combust. Flame
113
, 38
–52
(1998
).66.
R. W.
Stineman
, “A consistently well-behaved method of interpolation
,” Creative Comput.
6
(7
), 54
–57
(1980
).67.
S.
Ioffe
and C.
Szegedy
, “Batch normalization: Accelerating deep network training by reducing internal covariate shift
,” arXiv:1502.03167 (2015
).68.
F.
Pedregosa
, G.
Varoquaux
, A.
Gramfort
, V.
Michel
, B.
Thirion
, O.
Grisel
, M.
Blondel
, P.
Prettenhofer
, R.
Weiss
, V.
Dubourg
, J.
Vanderplas
, A.
Passos
, D.
Cournapeau
, M.
Brucher
, M.
Perrot
, and E.
Duchesnay
, “Scikit-learn: Machine learning in Python
,” J. Mach. Learn. Res.
12
, 2825
–2830
(2011
).69.
M.
Feurer
, K.
Eggensperger
, S.
Falkner
, M.
Lindauer
, and F.
Hutter
, “Practical automated machine learning for the automl challenge 2018
,” in International Workshop on Automatic Machine Learning at ICML
, 2018
.70.
I.
Guyon
(2019
)et al., “Analysis of the AutoML Challenge Series 2015–2018
,” in , The Springer Series on Challenges in Machine Learning, edited by F.
Hutter
, L.
Kotthoff
, and J.
Vanschoren
(Springer
, Cham
, 2019
).71.
C.
Jiménez
, F.
Ducros
, B.
Cuenot
, and B.
Bédat
, “Subgrid scale variance and dissipation of a scalar field in large eddy simulations
,” Phys. Fluids
13
, 1748
–1754
(2001
).72.
C.
Pera
, J.
Réveillon
, L.
Vervisch
, and P.
Domingo
, “Modeling subgrid scale mixture fraction variance in LES of evaporating spray
,” Combust. Flame
146
, 635
–648
(2006
).73.
J.
Floyd
and A. M.
Kempf
, “Computed tomography of chemiluminescence (CTC): High resolution and instantaneous 3D measurements of a matrix burner
,” Proc. Combust. Inst.
33
, 751
–758
(2011
).74.
A.
Géron
, Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow: Concepts, Tools, and Techniques to Build Intelligent Systems
(O’Reilly Media
, 2019
).75.
T.
Parr
, K.
Turgutlu
, C.
Csiszar
, and J.
Howard
, “Beware default random forest importances
” (2018
), https://explained.ai/rf-importance; accessed June 2020.76.
M.
Klein
, N.
Chakraborty
, and S.
Ketterl
, “A comparison of strategies for direct numerical simulation of turbulence chemistry interaction in generic planar turbulent premixed flames
,” Flow, Turbul. Combust.
99
, 955
–971
(2017
).77.
D. H.
Wacks
, N.
Chakraborty
, and E.
Mastorakos
, “Statistical analysis of turbulent flame-droplet interaction: A direct numerical simulation study
,” Flow, Turbul. Combust.
96
, 573
–607
(2016
).78.
79.
G. M.
Faeth
, “Current status of droplet and liquid combustion
,” Prog. Energy Combust. Sci.
3
(4
), 191
–224
(1977
), ISSN: 0360-1285.© 2020 Author(s).
2020
Author(s)
You do not currently have access to this content.
