Combustion instability poses a negative impact on the combustor’s performance and structural durability of both land-based and aircraft gas turbines. Hence, there is a need for early detection of combustion instabilities not only for performance monitoring and fault diagnosis, but also for initiating efficient decision and control in combustion dynamics.

• The technology provides a robust mechanism for detecting the onset of combustion instability in a combustion system. It uses a set of fiber optic sensors, namely fiber optic bundles and Fiber Bragg Gratings (FBG, Fig.1) to simultaneously sense chemiluminescence, pressure and temperature respectively, which are a measure of heat release and acoustics of the combustors.
• Further, the dynamic data is analyzed over a fast symbolic time-series built upon a generalized D-Markov machine which derives the cross entropy, which is then used to indicate the onset of combustion instability.
• A decline in the entropy rate is experienced when there is a transition to instability. Therefore, as soon as a decay in entropy rate is noted, the impending instability can be predicted and appropriate control measures can be initiated in the combustion system.