Diluted micromix combustion technology is an effective way to realize stable H₂ flames with low NO_x emission, and the turbulent flame structures of a micromix burner are experimentally revealed in this work with corrected OH-PLIF images.

In this study, a pattern recognition model is proposed to differentiate the flame state of a scramjet using an artificial neural network. The flame images are obtained from a scramjet ground test utilizing planar laser-induced fluorescence (PLIF). By extracting basic features, Hu moments and Zernike moments, the preprocessed flame images are compressed to mine valuable information. In order to reduce redundant input features and improve the efficiency of model operation, the partial least squares (PLS) method is introduced for feature screening and fusion. Then, a back propagation neural network (BPNN) model for multi-flame classification is established and analyzed. Finally, the flame states are determined by comparing the probabilities of different states. In order to optimize the recognition performance, the fusion features are studied and discussed. Experimental results show that when the filtered 11-dimensional features are used as input, the average recognition rate for the four states can reach 97.4%. These results demonstrate the significant potential of integrating PLIF and advanced data analysis methods, thereby broadening their application to intricate combustion fields.

A flame’s structural feature is a crucial parameter required to comprehensively understand the interaction between turbulence and flames. The generation and evolution processes of the structure feature have rarely been investigated in lean blowout (LBO) flame instability states. Hence, to understand the precursor features of the LBO flame, this work employed high-speed OH-PLIF measurements to acquire time-series LBO flame images and developed a novel feature extraction method based on a deep neural network to quantify the LBO features in real time. Meanwhile, we proposed a deep neural network segmentation method based on a tri-map called the Fire-MatteFormer, and conducted a statistical analysis on flame surface features, primarily holes. The statistical analysis results determined the relationship between the life cycle of holes (from generation to disappearance) and their area, perimeter, and total number. The trained Fire-MatteFormer model was found to represent a viable method for determining flame features in the detection of incipient LBO instability conditions. Overall, the model shows significant promise in ascertaining local flame structure features.