Researchers at Japan's Tokyo University of Science (TUS) and the Japan Aerospace Exploration Agency (JAXA) have designed a tool for detecting a precursor of thermoacoustic combustion oscillation, which damages combustion engines.
The researchers performed combustion experiments with varying fuel flow rates in a staged multisector combustor from JAXA, and fed the resulting data to a support vector machine (SVM) algorithm.
The SVM classified the combustion as stable, transitional, and combustion oscillations; the transitional state's pressure fluctuations are crucial to forecasting future combustion oscillations.
TUS' Hiroshi Gotoda said, "The methodology combining dynamical systems theory and machine learning can be useful for detecting predictive combustion oscillations in multisector combustors, such as those in aircraft engines."
From Tokyo University of Science (Japan)
View Full Article