Experimental Study of the Transient Behavior of Thermoacoustic Combustion Oscillations Using the Kurtosis Analysis of Dynamic Pressure

Combustion oscillations can cause severe damage to combustion systems and compromise the reliability of energy conversion devices. To achieve early identification of the transition toward combustion instability, this study proposes an adaptive kurtosis‐based detection method using dynamic pressure signals. The feasibility of temporal kurtosis ( T K ^∗ ) as a precursor indicator of combustion oscillations is investigated using experimental pressure data obtained from both a Rijke tube combustor and a swirl combustor. The statistical characteristics of kurtosis under stable combustion, intermittent oscillation, and limit‐cycle oscillation conditions are first analyzed, revealing the relationship between pressure signal distribution evolution and combustion stability variation. Furthermore, the temporal evolution of T K ^∗ during the formation of combustion oscillations is investigated, demonstrating that the increase of T K ^∗ is associated with the gradual excitation of thermoacoustic modes and the transition from stable combustion to instability. An adaptive sliding‐window algorithm based on the fluctuation characteristics of dynamic pressure signals is developed to improve the robustness and efficiency of online monitoring. The results demonstrate that temporal kurtosis can serve as a quantitative statistical indicator for tracking stability degradation and provide an effective criterion for real‐time assessment and early warning of combustion oscillations.