A physically based nonlinear model of combustion instability and active control

This paper addresses the ubiquitous limit cycle dynamics that occur in combustion processes in the context of thermoacoustic excitation. We present a finite-dimensional nonlinear model that is derived from the physics of heat release and its interactions with acoustics. We show that this model is capable of exhibiting limit cycle under a range of operating conditions and flow rates. Properties of the nonlinear model are verified numerically using the partial differential equation models, as well as experimentally on a bench top combustor rig. The linear control based on /spl Hscr//sub 2/-optimization is shown to be effective for a large class of initial conditions. Neural control is shown to lead to limit cycle suppression in combustion systems in the presence of modelling errors.