Abstract
In the present study, soft computing methods are employed for thermoacoustic simulation. A ducted Burke-Schumann diffusion flame is used as the heat source for a horizontal duct. First, a dynamic model is constructed from the input-output data sets (velocity forcing - heat release) generated from the Burke-Schumann flame using Comsol. An efficient and cheap model of heat source is obtained using dynamic fuzzy identification. The full thermoacoustic system is simulated in a time domain with the Galerkin method using the identified heat source model. Finally, dynamic neural networks are utilized for obtaining a dynamic fit for a set of operating conditions for the acoustic velocity at the heater location. The overall agreement between the outputs of the soft computing tools (fuzzy and neural network tools) with the Comsol and Galerkin solver is found to be satisfactory.