The operation envelope of modern gas turbines is affected by thermoacoustically induced combustion oscillations. The understanding and development of active and passive means for their suppression is crucial for the design process and field introduction of new gas turbine combustion systems. Whereas the propagation of acoustic sound waves in gas turbine combustion systems has been well understood, the flame induced acoustic source terms are still a major topic of investigation. The dynamics of combustion processes can be analyzed by means of flame transfer functions which relate heat release fluctuations to velocity fluctuations caused by a flame. The purpose of this paper is to introduce and to validate a novel computational approach to reconstruct flame transfer functions based on unsteady excited RANS simulations and system identification. Resulting time series of velocity and heat release are then used to reconstruct the flame transfer function by application of a system identification method based on Wiener-Hopf formulation. CFD/SI approach has been applied to a typical gas turbine burner. 3D unsteady simulations have been performed and the flame transfer results have been validated by comparison to experimental data. In addition the method has been benchmarked to results obtained from sinusoidal excitations.

This content is only available via PDF.
You do not currently have access to this content.