Influence of pilot H[formula omitted] injection on methane-air swirled flame stabilization and acoustic response

Abstract

Large Eddy Simulation is used to investigate the effect of localized pilot H2 injection on the Flame Transfer Function (FTF) of a premixed CH4-air swirled flame. The response of a perfectly premixed methane-air flame is compared to the one with an additional pilot hydrogen injection that supplies 10% of the original power. Numerical simulations are validated against experiments in terms of global FTF values at selected forcing frequencies, acoustic pressure and velocity signals, CH* flame images and flame root position dynamics. The unforced cases are first considered showing that, when H2 is injected in the center, the flame becomes slightly lifted with a localized diffusion front above the pilot injection. As in the experiments, LES retrieves that hydrogen pilot injection leads to a global redistribution of the heat release rate towards the flame root due to higher burning rates which translates to an overall FTF gain reduction over the entire frequency range explored. Then, the flame acoustic response for the two injection strategies is scrutinized at two distinct forcing frequencies: 240 Hz where the FTF gain difference is maximum, and 590 Hz where the FTF phase shift is maximum. LES reveals that, despite H2 pilot injection does not modify the dynamics of the large vortical structures shed in the external shear layer of the swiling jet which are synchronized by the acoustic forcing, the redistribution of the heat release rate towards the flame base weakens their interaction with the flame tip, explaining partly the FTF gain reduction at the two selected frequencies. In addition to that, a marked axial movement of the internal recirculation zone is observed at 240 Hz during the forcing cycle. For the pilot injection, it leads to an oscillation of the lifted flame root while, for the CH4-air case, the flame anchoring point is not affected. This additional oscillation leads for the pilot case to heat release rate fluctuations acting in phase opposition with respect to those observed at the flame tip, generating a further drop of the FTF gain at this specific frequency. The increased burning rate at the flame root and the flame length reduction of the pilot hydrogen flame also affect the characteristic time lag of the flame response. For both frequencies f = 240 Hz and 590 Hz, the phase shift between the two injection strategies is proportional to the flame length reduction caused by hydrogen injection. These simulations confirm that pilot hydrogen injection is an efficient way to reduce the acoustic response of swirled flames over a large frequency bandwidth.