Model-Based Optimal Active Control of Liquid-Fueled Combustion Systems

Advances in modeling combustor are obtained by investigating the shear-flow driven combustion instability. Shear-layer instability produces nonacoustic modes in shear flow. These modes can be predicted and matched to experimental data. These data are combined with a new tool based on the proper orthogonal decomposition (POD) method to yield the reduced-order models of combustion systems. These models shed light on the interactions among hydrodynamics, acoustics, and heat release. A recursive POD (RePOD) algorithm is developed to facilitate a stable online control design by providing suitable updates of the POD modes. Model-based posi-cast controllers produce favorable results as compared to phase-shift controllers using simulation studies on a large-scale combustor model. A theoretical framework for the investigation of shear-flow dynamics is found in modern linear stability analysis in which unstable modes have been characterized as connectivity unstable and absolute unstable.