Frequency Analysis and Predictive Identification of Flame Stability by Image Processing

Monitoring and characterization of combustion flames by digital image processing is an active research topic. This study experimentally investigates the feasibility of high speed visualization techniques for combustion instability monitoring in a swirl liquid-fueled lean combustor for different air/fuel ratios. Instability, in fact, is an unpleasant aspect in the combustive system that negatively impacts on combustion efficiency. This work investigates methods for extracting significant parameters using the geometrical and luminous data of the flame images; some flame features are related to the combustion regimes. The stability of the flame is identified using spectral and wavelet-based analysis of the pixel intensities of the flame images. In particular the most flame unstable regions were identified by analyzing the two dimensional maps of different physical quantities. The impact of the fuel/air ratio on the stability of the flame is investigated also by a Monochromator/Photomultiplier system (PMT). The results support the potential of the methods described for flame monitoring.