Boundary layer flashback of non-swirling premixed flames: Mechanisms, fundamental research, and recent advances

Boundary layer flashback in premixed jet flames has been the subject of detailed experimental and numerical investigation since the 1940′s. The traditional approach for characterizing flashback has involved the critical velocity gradient concept, with higher values indicating a higher flashback propensity for a given situation. Recent studies in confined configurations have illustrated that a key assumption underlying the critical velocity gradient concept, namely a lack of interaction between the flame and the approaching flow, is fundamentally incorrect. However, for unconfined configurations, where this interaction is much less important, the critical velocity gradient concept is able to partially capture flashback characteristics. Historically, the critical velocity gradient concept predicts trends of flashback behavior in laminar configurations for a wide range of temperatures, pressures, and fuel compositions more consistently than in turbulent configurations. This is due in part to the fact that many laminar studies establish well behaved velocity conditions in the tube conveying the premixed reactants to the reaction zone. Yet many important practical systems are in the turbulent regime and cannot be approximated by a simplified analysis. Studies to date in either regime, while numerous, generally do not provide a comprehensive methodology for accounting for all parameters. Recent work has attempted to capture the effect of a large number of these parameters in the turbulent regime, with some emphasis on providing design tools that can be used to estimate flashback propensity in more general terms. These approaches have demonstrated reasonable performance for the limited data available at elevated temperature and pressure which are representative of important practical system such as lean premixed combustors for gas turbines. While progress has been made in the last few years relative to predicting flashback for practical systems with high Reynolds numbers, only limited data are available for developing and validating correlations. Open questions remain in terms of using detailed numerical simulations and complex reaction chemistry to predict flashback for unconfined flames. In addition, flame-wall interaction in terms of heat transfer, sensitivity to turbulence levels, the role of general velocity gradients (vs idealized fully developed flow), and the role of high pressure must be further evaluated.