Abstract
Abstract Gas turbine combustors are commonly operated with lean premix flames, allowing for high efficiencies and low emissions. These operating conditions are susceptible to thermoacoustic pulsations, originating from acoustic-flame coupling. To reveal this coupling, experiments or simulations of acoustically forced combustion systems are necessary, which are very challenging for real-scale applications. In this work we investigate the possibility to determine the flame response to acoustic forcing from snapshots of the unforced flow. This approach is based on three central hypothesis: first, the flame response is driven by flow fluctuations, second, these flow fluctuations are dominated by coherent structures driven by hydrodynamic instabilities, and third, these instabilities are driven by stochastic forcing of the background turbulence. As a consequence the dynamics in the natural flow should be low-rank and very similar to those of the acoustically forced system. In this work, the methodology is applied to experimental data of an industry-scale swirl combustor. A resolvent analysis is conducted based on the linearized Navier-Stokes equations to assure analytically the low-rank behavior of the flow dynamics. Then, these dynamics are extracted from flow snapshots using spectral proper orthogonal decomposition (SPOD). The extended SPOD is applied to determine the heat release rate fluctuations that are correlated with the flow dynamics. The low-rank flow and flame dynamics determined from the analytic and data-driven approach are then compared to the flow response determined from a classic phase average of the acoustically forced flow, which allow the research hypothesis to be evaluated. It is concluded that for the present combustor, the flow and flame dynamics are low-rank for a wider frequency range and the response to harmonic forcing can be determined quite accurately from unforced snapshots. The methodology further allows to isolate the frequency range where the flame response is predominantly driven by hydrodynamic instabilites.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have