On the performance of stability margin measures for thermoacoustic instabilities in turbulent combustion systems

Recently, the search for a stability margin quantity has been extended to the field of nonlinear time series analysis, yielding a set of measures showing a smooth transition between the stable and unstable regime of a turbulent combustion system. So far, the performance of these measures to presage instability has been tested on a few laboratory setups, without mutual comparison of the different approaches. In this work a variety of measures are tested on a laboratory scale burner for validation and comparison. An atmospheric, partially premixed methane-fueled bluff-body stabilized burner is operated at 20 kW, showing stable and limit cycle oscillation operation for varying equivalence ratio. Focus is on an early and robust detection of an impending instability, suitable for on-line monitoring. Two data sets of acoustic pressure are considered, with different acoustic damping. Monotonicity and unambiguity of the stability margins are investigated. Additionally a surrogate data analysis is conducted, showing evidence of a nonlinear underlying dynamical system. However, these nonlinearities are found to have no significant influence on the predictions by methods that are based on nonlinear time series analysis.