Hybrid algorithm for the detection of turbulent flame fronts

Area of Science:

• Combustion science
• Fluid dynamics
• Optical diagnostics

Background:

• Planar Laser-Induced Fluorescence (PLIF) is crucial for combustion research.
• Low signal-to-noise ratios and complex flame structures challenge accurate flame front detection.
• Existing methods often struggle with noisy data and computational efficiency.

Purpose of the Study:

• To develop a robust and efficient algorithm for flame front detection in low signal-to-noise PLIF images.
• To improve the accuracy and speed of flame front identification in turbulent flames.
• To enable better experimental studies of highly wrinkled turbulent flames.

Main Methods:

• A hybrid, unsupervised approach combining adaptive contrast enhancement, segmentation, and edge detection.
• Image preprocessing to improve quality before segmentation.
• Segmentation highlights the general flame shape; edge detection refines the flame front boundary.

Main Results:

• High accuracy achieved in detecting flame fronts in finely wrinkled turbulent hydrogen-enriched flames.
• Significant improvements in computational speed (order of magnitude) compared to existing methods.
• Demonstrated robustness in the presence of noise and flame structure variability.

Conclusions:

• The developed algorithm offers a low-cost and accurate solution for flame front detection.
• It is particularly suitable for analyzing turbulent flames with intense wrinkling and low signal-to-noise ratios.
• This method has significant potential for advancing experimental combustion research.