EXPERIMENTAL INVESTIGATION OF TURBULENT BOUNDARY LAYER FLASHBACK LIMITS FOR PREMIXED HYDROGEN-AIR FLAMES CONFINED IN DUCTS

The design of flashback-resistant premixed burners for hydrogen-rich fuels is strongly dependent on reliable turbulent boundary layer flashback limits, since this process can be the dominant failure type for mixtures with high burning velocities. So far the flashback data published in literature is based on tube burner experiments with unconfined flames. However this flame configuration may not be representative for the most critical design case, which is aflame being already present inside the duct geometry. In order to shed light on this potential misconception, boundary layer flashback limits have been measured for unconfined and confined flames in fully premixed hydrogen-air mixtures at atmospheric conditions. Two duct geometries were considered, a tube burner and a quasi-2D turbulent channel flow. Furthermore, two confined flame holding configurations were realized, a small backward-facing step inside the duct and a ceramic tile at high temperature, which was mounted flush with the duct wall. While the measured flashback limits for unconfined tube burner flames compare well with literature results, a confinement of the stable flame leads to a shift of the flashback limits towards higher critical velocity gradients, which are in good agreement between the tube burner and the quasi-2D channel setup. The underestimation of flashback propensity resulting from unconfined tube burner experiments emerges from the physical situation at the burner rim. Heat loss from the flame to the wall results in a quenching gap, which causes a radial leakage flow of fresh gases. This flow in turn tends to increase the quenching distance, since it constitutes an additional convective heat loss. On the one hand, the quenching gap reduces the local adverse pressure gradient on the boundary layer On the other hand, the flame base is pushed outward, which deters the flame from entering the boundary layer region inside the duct. The flashback limits of confined flames stabilized at backward-facing steps followed this interpretation, and experiments with a flush ceramic flame holder constituted the upper limit of flashback propensity. It is concluded that the distribution of the flame backpressure and the flame position itself are key parameters for the determination of meaningful turbulent boundary layer flashback limits. For a conservative design path, the present results obtained from confined flames should be considered instead of unconfined tube burner values.