Abstract
The present study investigates the dynamics of merging flames, a prevalent phenomenon in reacting flow systems. We employ two different model order reduction techniques to uncover the underlying spatiotemporal features: Proper Orthogonal Decomposition (POD) and Dynamic Mode Decomposition (DMD). These approaches extract dominant flame modes, frequencies, and coherent structures within the reacting zone. We capture the intricate behavior of merging flames by conducting experiments with candle flames, which serve as a simplified yet representative model. By applying DMD to high-dimensional flame data, we successfully extracted coherent structures and dominant modes of variability. This allowed us to characterize the frequencies and spatial structures of the merging flame dynamics in more detail. Additionally, we utilized DMD to construct reduced-order models, which accurately captured the dominant flame modes. These models enable us to predict future flame behavior and design effective control strategies for flame stabilization. The combined use of DMD and POD enhances the analysis by comprehensively understanding the merging flame behavior. DMD uncovers the temporal evolution and characteristic frequencies, while POD identifies the spatial structures and energy distribution among different modes. This integrated approach allows us to construct reduced-order models that accurately capture the dominant flame modes and their corresponding frequencies, facilitating future predictions and control strategies for flame stabilization.
