Chapter 14: Machine Learning of Combustion LES Models from Reacting Direct Numerical Simulation

Research output: Chapter in Book/Report/Conference proceedingChapterpeer-review

11 Scopus Citations


In this chapter we demonstrate how supervised deep learning techniques can be used to construct models for the filtered progress variable source term necessary for large eddy simulation (LES). The source data for the model is a direct numerical simulation (DNS) of a reacting flow in a low swirl burner configuration. Filtered quantities taken from the DNS data are used to train a deep neural network (DNN)-based model. An efficient data sampling strategy was devised to ensure that a uniform representation of all the states observed in the filtered DNS data are equally present in the training dataset.A-priori testing of the DNN-based model highlights the representative power of DNN to accurately reproduce the filtered reaction progress variable source term over a range of scales and various flame regimes as seen in an industrial burner.

Original languageAmerican English
Title of host publicationData Analysis for Direct Numerical Simulations of Turbulent Combustion
Subtitle of host publicationFrom Equation-Based Analysis to Machine Learning
EditorsH. Pitsch, A. Attili
PublisherSpringer International Publishing
Number of pages20
ISBN (Electronic)9783030447182
ISBN (Print)9783030447175
StatePublished - 2020

Bibliographical note

Publisher Copyright:
© Springer Nature Switzerland AG 2020.

NREL Publication Number

  • NREL/CH-2C00-74457


  • combustion LES models
  • machine learning
  • reacting direct numerical simulation


Dive into the research topics of 'Chapter 14: Machine Learning of Combustion LES Models from Reacting Direct Numerical Simulation'. Together they form a unique fingerprint.

Cite this