Abstract

Detailed chemical kinetics calculations can be very computationally expensive, and so various approaches have been used to speed up combustion calculations. Deep neural networks (DNNs) are one promising approach that has seen significant development recently. Standard DNNs, however, do not necessarily follow physical constraints such as conservation of mass. Physics Informed Neural Networks (PINNs) are a class of neural networks that have physical laws embedded within the training process to create networks that follow those physical laws. A new PINN-based DNN approach to chemical kinetics modeling has been developed to make sure mass fraction predictions adhere to the conservation of atomic species. The approach also utilizes a mixture-of-experts (MOE) architecture where the data is distributed on multiple subnetworks followed by a softmax selective layer. The MOE architecture allows the different subnetworks to specialize in different thermochemical regimes, such as early stage ignition reactions or post-flame equilibrium chemistry, then the softmax layer smoothly transitions between the subnetwork predictions. This modeling approach was applied to the prediction of methane-air combustion using the GRI-Mech 3.0 as the reference mechanism. The training database was composed of data from 0D ignition delay simulations under initial conditions of 0.2–50 bar pressure, 500–2000 K temperature, an equivalence ratio between 0 and 2, and an N2-dilution percentage of up to 50%. A wide variety of network sizes and architectures of between 3 and 20 subnetworks and 6,600 to 77,000 neurons were tested. The resulting networks were able to predict 0D combustion simulations with similar accuracy and atomic mass conservation as standard kinetics solvers while having a 10-50× speedup in online evaluation time using CPUs, and on average over 200× when using a GPU.

References

1.
Echekki
,
T.
, and
Mastorakos
,
E.
, eds.,
2013
,
Turbulent Combustion Modeling: Advances, New Trends and Perspectives
,
Springer
,
Dordrecht, The Netherlands
.
2.
Lu
,
T.
, and
Law
,
C. K.
,
2009
, “
Toward Accommodating Realistic Fuel Chemistry in Large-Scale Computations
,”
Prog. Energy Combust. Sci.
,
35
(
2
), pp.
192
215
.10.1016/j.pecs.2008.10.002
3.
Schwer
,
D. A.
,
Tolsma
,
J. E.
,
Green
,
W. H.
, and
Barton
,
P. I.
,
2002
, “
On Upgrading the Numerics in Combustion Chemistry Codes
,”
Combust. Flame
,
128
(
3
), pp.
270
291
.10.1016/S0010-2180(01)00352-2
4.
McNenly
,
M. J.
,
Whitesides
,
R. A.
, and
Flowers
,
D. L.
,
2015
, “
Faster Solvers for Large Kinetic Mechanisms Using Adaptive Preconditioners
,”
Proc. Combust. Inst.
,
35
(
1
), pp.
581
587
.10.1016/j.proci.2014.05.113
5.
Pitsch
,
H.
,
Barths
,
H.
, and
Peters
,
N.
,
1996
, “
Three-Dimensional Modeling of Nox and Soot Formation in di-Diesel Engines Using Detailed Chemistry Based on the Interactive Flamelet Approach
,”
SAE
Paper No. 962057. 10.4271/962057
6.
D'Errico
,
G.
,
Lucchini
,
T.
,
Contino
,
F.
,
Jangi
,
M.
, and
Bai
,
X.-S.
,
2014
, “
Comparison of Well-Mixed and Multiple Representative Interactive Flamelet Approaches for Diesel Spray Combustion Modelling
,”
Combust. Theory Modell.
,
18
(
1
), pp.
65
88
.10.1080/13647830.2013.860238
7.
Ameen
,
M. M.
,
Kundu
,
P.
, and
Som
,
S.
,
2016
, “
Novel Tabulated Combustion Model Approach for Lifted Spray Flames With Large Eddy Simulations
,”
SAE Int. J. Engines
,
9
(
4
), pp.
2056
2065
.10.4271/2016-01-2194
8.
Kundu
,
P.
,
Ameen
,
M. M.
, and
Som
,
S.
,
2018
,
Recent Progress in Turbulent Combustion Modeling of Spray Flames Using Flamelet Models
,
Springer Singapore
,
Singapore
, pp.
477
512
.
9.
Ramaekers
,
W. J. S.
,
van Oijen
,
J. A.
, and
de Goey
,
L. P. H.
,
2010
, “
A Priori Testing of Flamelet Generated Manifolds for Turbulent Partially Premixed Methane/Air Flames
,”
Flow, Turbul. Combust.
,
84
(
3
), pp.
439
458
.10.1007/s10494-009-9223-1
10.
Donini
,
A.
,
M. Bastiaans
,
R. J.
,
van Oijen
,
J. A.
, and
H. de Goey
,
L. P.
,
2017
, “
A 5-D Implementation of FGM for the Large Eddy Simulation of a Stratified Swirled Flame With Heat Loss in a Gas Turbine Combustor
,”
Flow, Turbul. Combust.
,
98
(
3
), pp.
887
922
.10.1007/s10494-016-9777-7
11.
Fiorina
,
B.
,
Gicquel
,
O.
,
Vervisch
,
L.
,
Carpentier
,
S.
, and
Darabiha
,
N.
,
2005
, “
Approximating the Chemical Structure of Partially Premixed and Diffusion Counterflow Flames Using Fpi Flamelet Tabulation
,”
Combust. Flame
,
140
(
3
), pp.
147
160
.10.1016/j.combustflame.2004.11.002
12.
Franzelli
,
B.
,
Fiorina
,
B.
, and
Darabiha
,
N.
,
2013
, “
A Tabulated Chemistry Method for Spray Combustion
,”
Proc. Combust. Inst.
,
34
(
1
), pp.
1659
1666
.10.1016/j.proci.2012.06.013
13.
Donoho
,
D. L.
,
2000
, “
High-Dimensional Data Analysis: The Curses and Blessings of Dimensionality
,”
AMS Conference on Mathematical Challenges of the 21st Century
, Los Angeles, CA, Aug. 6–12, pp.
1
33
.https://www.researchgate.net/publication/220049061_High-Dimensional_Data_Analysis_The_Curses_and_Blessings_of_Dimensionality
14.
Bai
,
X.-S.
,
2018
,
Numerical Simulation of Turbulent Combustion in Internal Combustion Engines
,
Springer Singapore
,
Singapore
, pp.
513
541
. Ch. 17.
15.
Christo
,
F.
,
Masri
,
A.
, and
Nebot
,
E.
,
1996
, “
Artificial Neural Network Implementation of Chemistry With PDF Simulation of H2/CO2 Flames
,”
Combust. Flame
,
106
(
4
), pp.
406
427
.10.1016/0010-2180(95)00250-2
16.
Christo
,
F. C.
,
Masri
,
A. R.
,
Nebot
,
E. M.
, and
Pope
,
S. B.
,
1996
, “
An Integrated Pdf/Neural Network Approach for Simulating Turbulent Reacting Systems
,”
Symp. (Int.) Combust.
,
26
(
1
), pp.
43
48
.10.1016/S0082-0784(96)80198-6
17.
Blasco
,
J.
,
Fueyo
,
N.
,
Dopazo
,
C.
, and
Ballester
,
J.
,
1998
, “
Modelling the Temporal Evolution of a Reduced Combustion Chemical System With an Artificial Neural Network
,”
Combust. Flame
,
113
(
1–2
), pp.
38
52
.10.1016/S0010-2180(97)00211-3
18.
Blasco
,
J.
,
Fueyo
,
N.
,
Dopazo
,
C.
, and
Chen
,
J.-Y.
,
2000
, “
A Self-Organizing-Map Approach to Chemistry Representation in Combustion Applications
,”
Combust. Theory Modell.
,
4
(
1
), pp.
61
76
.10.1088/1364-7830/4/1/304
19.
Sen
,
B. A.
, and
Menon
,
S.
,
2009
, “
Turbulent Premixed Flame Modeling Using Artificial Neural Networks Based Chemical Kinetics
,”
Proc. Combust. Inst.
,
32
(
1
), pp.
1605
1611
.10.1016/j.proci.2008.05.077
20.
Sen
,
B. A.
, and
Menon
,
S.
,
2010
, “
Linear Eddy Mixing Based Tabulation and Artificial Neural Networks for Large Eddy Simulations of Turbulent Flames
,”
Combust. Flame
,
157
(
1
), pp.
62
74
.10.1016/j.combustflame.2009.06.005
21.
Chatzopoulos
,
A.
, and
Rigopoulos
,
S.
,
2013
, “
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
(
1
), pp.
1465
1473
.10.1016/j.proci.2012.06.057
22.
Franke
,
L. L.
,
Chatzopoulos
,
A. K.
, and
Rigopoulos
,
S.
,
2017
, “
Tabulation of Combustion Chemistry Via Artificial Neural Networks (Anns): Methodology and Application to LES-PDF Simulation of Sydney Flame l
,”
Combust. Flame
,
185
, pp.
245
260
.10.1016/j.combustflame.2017.07.014
23.
Owoyele
,
O.
,
Kundu
,
P.
, and
Pal
,
P.
,
2019
, “Efficient Bifurcation and Parameterization of Multi-Dimensional Combustion Manifolds Using Deep Mixture of Experts:
An a Priori Study
”.https://www.researchgate.net/publication/336796865_Efficient_bifurcation_and_parameterization_of_multi-dimensional_combustion_manifolds_using_deep_mixture_of_experts_an_a_priori_study
24.
Owoyele
,
O.
,
Kundu
,
P.
,
Ameen
,
M. M.
,
Echekki
,
T.
, and
Som
,
S.
,
2020
, “
Application of Deep Artificial Neural Networks to Multi-Dimensional Flamelet Libraries and Spray Flames
,”
Int. J. Engine Res.
,
21
(
1
), pp.
151
168
.10.1177/1468087419837770
25.
Niemeyer
,
K. E.
, and
Sung
,
C.-J.
,
2014
, “
Accelerating Moderately Stiff Chemical Kinetics in Reactive-Flow Simulations Using Gpus
,”
J. Comput. Phys.
,
256
, pp.
854
871
.10.1016/j.jcp.2013.09.025
26.
Ji
,
W.
, and
Deng
,
S.
,
2021
, “
Autonomous Discovery of Unknown Reaction Pathways From Data by Chemical Reaction Neural Network
,”
J. Phys. Chem. A
,
125
(
4
), pp.
1082
1092
.10.1021/acs.jpca.0c09316
27.
Owoyele
,
O.
, and
Pal
,
P.
,
2022
, “
Chemnode: A Neural Ordinary Differential Equations Framework for Efficient Chemical Kinetic Solvers
,”
Energy AI
,
7
, p.
100118
.10.1016/j.egyai.2021.100118
28.
Zhang
,
Y.
,
Xu
,
S.
,
Zhong
,
S.
,
Bai
,
X.-S.
,
Wang
,
H.
, and
Yao
,
M.
,
2020
, “
Large Eddy Simulation of Spray Combustion Using Flamelet Generated Manifolds Combined With Artificial Neural Networks
,”
Energy AI
,
2
, p.
100021
.10.1016/j.egyai.2020.100021
29.
Wan
,
K.
,
Barnaud
,
C.
,
Vervisch
,
L.
, and
Domingo
,
P.
,
2020
, “
Chemistry Reduction Using Machine Learning Trained From Non-Premixed Micro-Mixing Modeling: Application to Dns of a Syngas Turbulent Oxy-Flame With Side-Wall Effects
,”
Combust. Flame
,
220
, pp.
119
129
.10.1016/j.combustflame.2020.06.008
30.
Wan
,
K.
,
Hartl
,
S.
,
Vervisch
,
L.
,
Domingo
,
P.
,
Barlow
,
R. S.
, and
Hasse
,
C.
,
2020
, “
Combustion Regime Identification From Machine Learning Trained by Raman/Rayleigh Line Measurements
,”
Combust. Flame
,
219
, pp.
268
274
.10.1016/j.combustflame.2020.05.024
31.
Barwey
,
S.
,
Prakash
,
S.
,
Hassanaly
,
M.
, and
Raman
,
V.
,
2021
, “
Data-Driven Classification and Modeling of Combustion Regimes in Detonation Waves
,”
Flow, Turbul. Combust.
,
106
(
4
), pp.
1065
1089
.10.1007/s10494-020-00176-4
32.
Barwey
,
S.
, and
Raman
,
V.
,
2021
, “
A Neural Network-Inspired Matrix Formulation of Chemical Kinetics for Acceleration on GPUS
,”
Energies
,
14
(
9
), p.
2710
.10.3390/en14092710
33.
Malpica Galassi
,
R.
,
Ciottoli
,
P. P.
,
Valorani
,
M.
, and
Im
,
H. G.
,
2022
, “
An Adaptive Time-Integration Scheme for Stiff Chemistry Based on Computational Singular Perturbation and Artificial Neural Networks
,”
J. Comput. Phys.
,
451
, p.
110875
.10.1016/j.jcp.2021.110875
34.
Raissi
,
M.
,
Perdikaris
,
P.
, and
Karniadakis
,
G.
,
2019
, “
Physics-Informed Neural Networks: A Deep Learning Framework for Solving Forward and Inverse Problems Involving Nonlinear Partial Differential Equations
,”
J. Comput. Phys.
,
378
, pp.
686
707
.10.1016/j.jcp.2018.10.045
35.
Smith
,
G. P.
,
Golden
,
D. M.
,
Frenklach
,
M.
,
Moriarty
,
N. W.
,
Eiteneer
,
B.
,
Goldenberg
,
M.
,
Bowman
,
C. T.
,
Hanson
,
R. K.
,
Song
,
S.
,
William
,
C.
,
Gardiner
,
J.
,
Lissianski
,
V. V.
, and
Qin
,
Z.
, “GRI-Mech 3.0,” accessed Apr. 15, 2022, http://combustion.berkeley.edu/gri-mech/
36.
Goodwin
,
D. G.
,
Speth
,
R. L.
,
Moffat
,
H. K.
, and
Weber
,
B. W.
,
2018
, “
Cantera: An Object-Oriented Software Toolkit for Chemical Kinetics, Thermodynamics, and Transport Processes
,” Version 2.4, accessed Apr. 15, 2022, https://www.cantera.org/
37.
Sola
,
J.
, and
Sevilla
,
J.
,
1997
, “
Importance of Input Data Normalization for the Application of Neural Networks to Complex Industrial Problems
,”
IEEE Trans. Nucl. Sci.
,
44
(
3
), pp.
1464
1468
.10.1109/23.589532
38.
Reza
,
E.
,
Ehsanollah
,
K.
, and
Reza
,
Y. M.
,
2007
, “
Face Detection Using Mixture of MLP Experts
,”
Neural Process. Lett.
,
26
(
1
), pp.
69
82
.10.1007/s11063-007-9043-z
39.
Jacobs
,
R. A.
,
Jordan
,
M. I.
,
Nowlan
,
S. J.
, and
Hinton
,
G. E.
,
1991
, “
Adaptive Mixtures of Local Experts
,”
Neural Comput.
,
3
(
1
), pp.
79
87
.10.1162/neco.1991.3.1.79
40.
Jacobs
,
R. A.
,
Jordan
,
M. I.
, and
Barto
,
A. G.
,
1991
, “
Task Decomposition Through Competition in a Modular Connectionist Architecture: The What and Where Vision Tasks
,”
Cognitive Sci.
,
15
(
2
), pp.
219
250
.10.1207/s15516709cog1502_2
41.
Raissi
,
M.
,
Perdikaris
,
P.
, and
Karniadakis
,
G. E.
,
2017
, “
Physics Informed Deep Learning (Part I): Data-Driven Solutions of Nonlinear Partial Differential Equations
,”
arXiv:1711.10561
.10.48550/arXiv.1711.10561
42.
Raissi
,
M.
,
Perdikaris
,
P.
, and
Karniadakis
,
G. E.
,
2017
, “
Physics Informed Deep Learning (Part II): Data-Driven Discovery of Nonlinear Partial Differential Equations
,”
arXiv:1711.10566
.10.48550/arXiv.1711.10566
43.
Abadi
,
M.
,
Agarwal
,
A.
,
Barham
,
P.
,
Brevdo
,
E.
,
Chen
,
Z.
,
Citro
,
C.
,
Corrado
,
G. S.
, and
et al
.,
2015
, “
TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems
,” accessed Apr. 15, 2022, https://www.tensorflow.org/
You do not currently have access to this content.