In this study, novel deep learning models based on high-speed flame images are proposed to diagnose the combustion instability of a gas turbine. Two different network layers that can be combined with any existing backbone network are established—(1) An early-fusion layer that can learn to extract the power spectral density of subsequent image frames, which is time-invariant under certain conditions. (2) A late-fusion layer which combines the outputs of a backbone network at different time steps to predict the current combustion state. The performance of the proposed models is validated by the dataset of high speed flame images, which have been obtained in a gas turbine combustor during the transient process from stable condition to unstable condition and vice versa. Excellent performance is achieved for all test cases with high accuracy of 95.1–98.6% and a short processing time of 5.2–12.2 ms. Interestingly, simply increasing the number of input images is as competitive as combining the proposed early-fusion layer to a backbone network. In addition, using handcrafted weights for the late-fusion layer is shown to be more effective than using learned weights. From the results, the best combination is selected as the ResNet-18 model combined with our proposed fusion layers over 16 time-steps. The proposed deep learning method is proven as a potential tool for combustion instability identification and expected to be a promising tool for combustion instability prediction as well.
This study proposes and analyzes a novel methodology that can effectively detect multi-mode combustion instability (CI) in a gas turbine combustor. The experiment is conducted in a model gas turbine combustor, and dynamic pressure (DP) and flame images are examined during the transition from stable to unstable flame, which is driven by changing fuel compositions. As a powerful technique for early detection of CI in multi-mode as well as in single mode, a new filter bank (FB) method based on spectral analysis of DP is proposed. Sequential processing using a triangular filter with Mel-scaling and a Hamming window is applied to increase the accuracy of the FB method, and the instability criterion is determined by calculating the magnitude of FB components. The performance of the FB method is compared with that of two conventional methods that are based on the root-mean-squared DP and temporal kurtosis. From the results, the FB method shows comparable performance in detection speed, sensitivity, and accuracy with other parameters. In addition, the FB components enable the analysis of various frequencies and multi-mode frequencies. Therefore, the FB method can be considered as an additional prognosis tool to determine the multi-mode CI in a monitoring system for gas turbine combustors.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.