Detection and prevention of blowout in a lean premixed gas-turbine model combustor using the concept of dynamical system theory

• DOI: 10.1016/j.proci.2014.07.014
• Publication Date: 2014-08-19T04:00:49Z
• Authors: Shohei Domen; Hiroshi Gotoda; Taku Kuriyama; Yuta Okuno; Shigeru Tachibana
• Keywords: Recurrence quantification analysis

Abstract We propose a novel online method of detecting lean blowout (LBO) in a laboratory-scale premixed-type gas-turbine model combustor from the viewpoint of dynamical system theory that enables us to quantify the complexity in combustion instability. The multiscale entropy method and nonlinear forecasting method, which have not been previously adopted for the analysis of combustion instability, demonstrate that combustion instability close to LBO exhibits much more complicated stochastic dynamics than fractional Brownian motion. The permutation entropy, which can be estimated by quantifying the rank order pattern of components in phase space vectors constructed from the short time-series data of pressure fluctuations, is used as a control variable to prevent LBO. The online method proposed in this work has the potential to control the complex combustion state close to LBO.