Nonlinear response of lean-premixed multi-jet turbulent hydrogen/natural gas/air flames

Fuel-flexible gas turbines supplied with hydrogen-blended natural gas (NG) are expected to play a pivotal role in balancing intermittent production of electricity from renewable sources and therefore to contribute to decarbonizing the power generation sector. Despite substantial progress in developing H 2 -fueled combustors, the acoustic response of H 2 flames, which governs the onset of thermoacoustic instabilities remains difficult to predict. In particular, the nonlinear heat release response of multi-jet turbulent H 2 /NG/air flames to high-amplitude acoustic forcing is still largely unknown. To address this problem, we perform flame describing function (FDF) measurements over a broad range of forcing frequencies and amplitudes, and levels of H 2 enrichment, based on the purely acoustic method using multiple microphones. Our results show that when subjected to sufficiently high-amplitude forcing, the multi-jet flames exhibit a nonlinear acoustic response, and the saturation amplitude heavily depends on the forcing frequency and the level of H 2 enrichment. Based on an experimental FDF dataset of pure H 2 flames and a distributed time delay (DTD) modeling, we demonstrate that the convective time delay of the dominant upstream disturbance and its spread are amplitude-dependent. A stability analysis using a thermoacoustic network modeling coupled with the DTD models enables us to successfully predict the onset of the instability. The present work thus provides an experimental FDF dataset and important insights into the nonlinear response of multi-jet turbulent H 2 and H 2 /NG flames. Novelty and significance statement The novelty of this study is twofold. First, we characterize the nonlinear heat release response of lean-premixed multi-jet turbulent flames over a broad range of levels of hydrogen enrichment and forcing conditions. We then model the frequency- and amplitude-dependent heat release response of pure H 2 jet flames, based on the present flame describing function (FDF) dataset and distributed time delay modeling (DTD) approach. Second, we demonstrate that incorporating the identified DTD model into a thermoacoustic network modeling enables us to reliably assess the thermoacoustic stability of the multi-jet H 2 combustor. These findings fill a key research gap in the scarcity of H 2 FDF data, and constitute a solid basis for predicting the thermoacoustic stability of H 2 combustion systems comprising arrays of multi-jet nozzles, which are currently attracting intense interest in the gas turbine industry.