Abstract
We investigated the deterministic behavior of combustion instability in a lean premixed gas-turbine model combustor from the viewpoint of nonlinear forecasting. A genetic algorithm, which quantifies the predictability of time variation in pressure fluctuations, is applied as a nonlinear forecasting method in this work. The dynamic behavior in the combustion instability near lean blowout exhibits the fractional Brownian motion with anti-persistent, and intermittent oscillation represents chaos. This is clearly demonstrated by nonlinear forecasting based on genetic algorithm, which has not been addressed in the fields of combustion science and physics.
