Dynamic Characterization of Pulse Combustion by Image Series Processing

Abstract

This paper proposed a flame image processing method to characterize the dynamic progress of pulse combustion at different operation conditions. Experiments were carried out using a Helmholtz pulse combustor with a range of gas supply pressures that determine the operation conditions of the combustor. The spatial distribution, time-domain, and frequency-domain characteristics were proposed to characterize the dynamic behaviors of pulse combustion. Image segmentation was first adopted to divide the original image into the flame region and background. The center of gravity, non-uniformity, dispersity, and normalized flame area were then extracted to quantify the spatial distribution characteristics of the flame. Variations of the center of gravity, non-uniformity, dispersity, and normalized flame area with time were also obtained to describe the temporal features of the flame. The fast Fourier transform-based analysis was implemented to attain the frequency-domain features of the flame from the time series of the normalized flame area. Results show that the proposed method is effective in characterizing the spatial, temporal, and spectral features of the dynamic process in a pulse combustor. The promising results indicate that the proposed method can be used to monitor the operation conditions of the combustor, including its safety and stability, and provide comprehensive reference data for in-depth understanding of pulse combustion.