Machine learning-assisted early ignition prediction in a complex flow

Abstract Machine learning methods are used to improve the efficiency by which turbulence-resolved simulations predict whether a hydrogen jet in air crossflow will successfully ignite. The flush-mounted jet issues perpendicularly from the wall of a low-speed wind tunnel into a turbulent boundary layer wherein a laser-induced optical breakdown (LIB) hotspot is deposited. A detailed hydrogen chemical mechanism is used to model the radicals and any subsequent chemical reactions. A dielectric-barrier discharge actuator generates body forces and hydrogen radicals near the jet orifice. We focus on the success or not of the ignition based on LIB location. A challenge is that definitive determination of this requires long simulations, up to 440 µs after the LIB deposition. This is particularly expensive since multiple simulations are required to find the threshold. To reduce the computational effort, three short-time (91 µs) criteria are proposed, evaluated, and compared: a constructed criterion based on detailed observations of radicals near the stoichiometric surface and two machine learning approaches, each trained on 38 realizations. The constructed criterion provides a low-cost estimate of the ignition boundary that is unambiguous in 45 of the 50 training and test trials, so only 10% of them would need to be simulated longer. The trained neural networks correctly predict outcomes in all cases evaluated, with the more automated procedure—a convolutional neural network (CNN) trained on two-dimensional images—providing the most definitive outcome prediction. From the CNN, a sensitivity analysis is used to determine which kernel features from the two-dimensional input data, as defined by intermediate-layer network weights, are important for identifying ignition.

[1]  Hao Chen,et al.  3D Deeply Supervised Network for Automatic Liver Segmentation from CT Volumes , 2016, MICCAI.

[2]  Luca Massa,et al.  An integrated predictive simulation model for the plasma-assisted ignition of a fuel jet in a turbulent crossflow , 2016 .

[3]  Martin T. Hagan,et al.  Gauss-Newton approximation to Bayesian learning , 1997, Proceedings of International Conference on Neural Networks (ICNN'97).

[4]  Luca Massa,et al.  Plasma-combustion coupling in a dielectric-barrier discharge actuated fuel jet , 2017 .

[5]  Thomas Wiatowski,et al.  A Mathematical Theory of Deep Convolutional Neural Networks for Feature Extraction , 2015, IEEE Transactions on Information Theory.

[6]  S. Lele.  Compact finite difference schemes with spectral-like resolution , 1992 .

[7]  F. Williams,et al.  Recent advances in understanding of flammability characteristics of hydrogen , 2014 .

[8]  Brian D. Ziebart,et al.  A Deep Learning Approach to Identifying Shock Locations in Turbulent Combustion Tensor Fields , 2017 .

[9]  Marco Panesi,et al.  FEM simulation of laser-induced plasma breakdown experiments for combustion applications , 2017 .

[10]  S. H. Lam,et al.  Using CSP to Understand Complex Chemical Kinetics , 1993 .

[11]  Nasser M. Nasrabadi,et al.  Pattern Recognition and Machine Learning , 2006, Technometrics.

[12]  N. Aleksandrov,et al.  Plasma-assisted ignition and combustion , 2013 .

[13]  Martin A. Reno,et al.  Coefficients for calculating thermodynamic and transport properties of individual species , 1993 .

[14]  Hermann Ney,et al.  LSTM Neural Networks for Language Modeling , 2012, INTERSPEECH.

[15]  Daniel J. Bodony,et al.  Accuracy of the Simultaneous-Approximation-Term Boundary Condition for Time-Dependent Problems , 2010, J. Sci. Comput..

[16]  B. Strand.  Summation by parts for finite difference approximations for d/dx , 1994 .

[17]  B. Deshaies,et al.  Relative influences of convective and diffusive transports during spherical flame initiation , 1988 .

[18]  C. Demichelis.  Laser induced gas breakdown: A bibliographical review , 1969 .

[19]  C. F. Curtiss,et al.  Molecular Theory Of Gases And Liquids , 1954 .

[20]  Nitish Srivastava,et al.  Dropout: a simple way to prevent neural networks from overfitting , 2014, J. Mach. Learn. Res..

[21]  C. Law,et al.  Chemical explosive mode analysis for a turbulent lifted ethylene jet flame in highly-heated coflow , 2012 .

[22]  P. S. Tromans,et al.  An analysis of lewis number and flow effects on the ignition of premixed gases , 1988 .

[23]  Christopher M. Bishop,et al.  Neural networks for pattern recognition , 1995 .

[24]  Abdelkader Frendi,et al.  Dependence of Minimum Ignition Energy on Ignition Parameters , 1990 .

[25]  Martin T. Hagan,et al.  Neural network design , 1995 .

[26]  Hyungrok Do,et al.  Plasma assisted cavity flame ignition in supersonic flows , 2010 .

[27]  Samy Bengio,et al.  Show and tell: A neural image caption generator , 2014, 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[28]  Song Han,et al.  Deep Compression: Compressing Deep Neural Network with Pruning, Trained Quantization and Huffman Coding , 2015, ICLR.

[29]  Forman A. Williams,et al.  HYDROGEN–OXYGEN INDUCTION TIMES ABOVE CROSSOVER TEMPERATURES , 2004 .

[30]  David J. C. MacKay,et al.  Bayesian Interpolation , 1992, Neural Computation.

[31]  Fei-Fei Li,et al.  Large-Scale Video Classification with Convolutional Neural Networks , 2014, 2014 IEEE Conference on Computer Vision and Pattern Recognition.

[32]  M. Stuart,et al.  Dielectric Constant of Quartz as a Function of Frequency and Temperature , 1955 .

[33]  Grgoire Montavon,et al.  Neural Networks: Tricks of the Trade , 2012, Lecture Notes in Computer Science.

[34]  Miguel R. Visbal,et al.  High-Order Schemes for Navier-Stokes Equations: Algorithm and Implementation Into FDL3DI , 1998 .

[35]  Zhi-Hua Zhou,et al.  Ensemble Methods: Foundations and Algorithms , 2012 .

[36]  Yaozong Gao,et al.  Estimating CT Image from MRI Data Using 3D Fully Convolutional Networks , 2016, LABELS/DLMIA@MICCAI.

[37]  Akihiko Ito,et al.  Schlieren visualization of blast extinguishment with laser-induced breakdown , 2017 .

[38]  P. Glarborg,et al.  Chemically Reacting Flow : Theory and Practice , 2003 .

[39]  J J Hopfield,et al.  Neural networks and physical systems with emergent collective computational abilities. , 1982, Proceedings of the National Academy of Sciences of the United States of America.

[40]  Miguel R. Visbal,et al.  High-Order-Accurate Methods for Complex Unsteady Subsonic Flows , 1999 .

[41]  C. J. Butler,et al.  Measurements of the Concentrations of Free Hydrogen Atoms in Flames from Observations of Ions: Correlation of Burning Velocities with Concentrations of Free Hydrogen Atoms , 1998 .

[42]  T. Phuoc.  Laser-induced spark ignition fundamental and applications , 2006 .

[43]  Gregory S Elliott,et al.  Coaxial DBD Actuator Design for Control of a Hydrogen Diffusion Flame , 2016 .

[44]  Jürgen Schmidhuber,et al.  Long Short-Term Memory , 1997, Neural Computation.