Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Accurate ignition detection of solid fuel particles using machine learning

Tao Li, Zhangke Liang, Andreas Dreizler, Benjamin Böhm

TL;DR

The paper tackles accurate ignition timing for homogeneous ignition of single coal particles in a laminar flow reactor by fusing high-speed OH-LIF/DBI diagnostics with convolutional neural networks. It compares threshold-based SAS detection to ResNet-18 and FPN architectures, showing that ResNet-18 trained on about 462 labeled events achieves ignition-delay predictions with $ITD$ near zero and minimal dispersion, outperforming threshold methods across particle sizes and atmospheres. The study demonstrates that hierarchical feature extraction from multi-parameter optical data enhances ignition detection and that increasing training data improves robustness, suggesting broad applicability to similar solid-fuel experiments. The approach provides a practical, data-driven path to reliable ignition timing in high-speed combustion diagnostics, enabling improved model validation and reactor design insights.

Abstract

In the present work, accurate determination of single-particle ignition is focused on using high-speed optical diagnostics combined with machine learning approaches. Ignition of individual particles in a laminar flow reactor are visualized by simultaneous 10 kHz OH-LIF and DBI measurements. Two coal particle sizes of 90-125μm and 160-200μm are investigated in conventional air and oxy-fuel conditions with increasing oxygen concentrations. Ignition delay times are first evaluated with threshold methods, revealing obvious deviations compared to the ground truth detected by the human eye. Then, residual networks (ResNet) and feature pyramidal networks (FPN) are trained on the ground truth and applied to predict the ignition time.~Both networks are capable of detecting ignition with significantly higher accuracy and precision. Besides, influences of input data and depth of networks on the prediction performance of a trained model are examined.~The current study shows that the hierarchical feature extraction of the convolutions networks clearly facilitates data evaluation for high-speed optical measurements and could be transferred to other solid fuel experiments with similar boundary conditions.

Accurate ignition detection of solid fuel particles using machine learning

TL;DR

The paper tackles accurate ignition timing for homogeneous ignition of single coal particles in a laminar flow reactor by fusing high-speed OH-LIF/DBI diagnostics with convolutional neural networks. It compares threshold-based SAS detection to ResNet-18 and FPN architectures, showing that ResNet-18 trained on about 462 labeled events achieves ignition-delay predictions with near zero and minimal dispersion, outperforming threshold methods across particle sizes and atmospheres. The study demonstrates that hierarchical feature extraction from multi-parameter optical data enhances ignition detection and that increasing training data improves robustness, suggesting broad applicability to similar solid-fuel experiments. The approach provides a practical, data-driven path to reliable ignition timing in high-speed combustion diagnostics, enabling improved model validation and reactor design insights.

Abstract

In the present work, accurate determination of single-particle ignition is focused on using high-speed optical diagnostics combined with machine learning approaches. Ignition of individual particles in a laminar flow reactor are visualized by simultaneous 10 kHz OH-LIF and DBI measurements. Two coal particle sizes of 90-125μm and 160-200μm are investigated in conventional air and oxy-fuel conditions with increasing oxygen concentrations. Ignition delay times are first evaluated with threshold methods, revealing obvious deviations compared to the ground truth detected by the human eye. Then, residual networks (ResNet) and feature pyramidal networks (FPN) are trained on the ground truth and applied to predict the ignition time.~Both networks are capable of detecting ignition with significantly higher accuracy and precision. Besides, influences of input data and depth of networks on the prediction performance of a trained model are examined.~The current study shows that the hierarchical feature extraction of the convolutions networks clearly facilitates data evaluation for high-speed optical measurements and could be transferred to other solid fuel experiments with similar boundary conditions.
Paper Structure (10 sections, 1 equation, 6 figures)

This paper contains 10 sections, 1 equation, 6 figures.

Table of Contents

  1. Introduction
  2. Experiments
  3. Data evaluation
  4. Results and Discussions
  5. Ignition detected by SAS
  6. Ignition detected by ResNets
  7. Ignition detected by FPN
  8. Conclusions
  9. Acknowledgements
  10. References

Figures (6)

  • Figure 1: A schematic experimental layout including optical diagnostics and the laminar flow reactor.
  • Figure 2: A time-resolved sequence of particle ignition with $t_\textrm{ign}$ given by the ground truth (manual labeling). (a) OH-LIF raw images. (b) binary OH-LIF images.
  • Figure 3: Comparison of ignition delay times by the SAS method $t_\textrm{i,SAS}$ and the manual label $t_\textrm{i,gt}$ for two particle sizes A and B in seven atmospheres.
  • Figure 4: Ignition time difference $t_\textrm{i,RN}$ - $t_\textrm{i,gt}$ by using ResNet-18 with the amount of particle events (a) $N_\textrm{ev}$ = 14, (b) $N_\textrm{ev}$ = 56, (c) $N_\textrm{ev}$ = 140, and (d) $N_\textrm{ev}$ = 462.
  • Figure 5: Ignition time difference $t_\textrm{i,FPN}$ - $t_\textrm{i,gt}$ using different ResNet models in the bottom-up pathway of FPN networks.
  • ...and 1 more figures