Observed State-Dependent Adaptive Control for Thermoacoustic Instability in Aeroengine Oscillating Combustion with Unknown Thermal Inertia

Thermoacoustic instability is the main form of combustion instability in aeroengine. For the thermoacoustic system with coupling between sound field and unknown nonlinear flame heat release, an active control method based on the fully actuated system method and a time-delay neural network (NN) state observer is proposed for one acoustic mode. Firstly, a time-delay NN observer is proposed, which addresses the shortcomings of the linear observer in observing nonlinear systems and extends the NN observer to time-delay nonlinear systems. Subsequently, the finite covering lemma is combined with the radial basis neural network (RBF NN) to approximate the unknown nonlinear heat release with the unknown thermal inertia. Then, based on the fully actuated system approach, an adaptive active control strategy is proposed, which ensures that the pressure and velocity of the thermoacoustic system, as well as the observation errors of the pressure and velocity, converge to a small neighborhood about the origin. Finally, the feasibility and effectiveness of the adaptive active control method based on the time-delay NN observer are verified through the coupling of the Culick acoustic motion model and the nonlinear flame response to form the thermoacoustic system.