Thermoacoustic Instability prediction based on time-harmonic linearized acoustic propagation models

1.1. Context: Thermoacoustic instabilities arise from the interaction of acoustic waves and unsteady heat release. Such instabilities must be controlled because they can produce excessive noise and may even result in massive structural damage. This is particularly true for thermoacoustic combustion instabilities encountered in many industrial applications such as domestic burners, gas turbines or rocket engines. In these cases, the heat source is the flame enclosed in a chamber that acts as an acoustic resonator. Sound is generally reflected back to the flame such that a feedback loop is established. If a positive correlation between the sound field in the chamber and the fluctuations of the heat release is established, a self-excited feedback instability may occur. It can result in excessive heat and mechanical loads that threaten the structural integrity of the combustors. The design of appropriate combustion chamber combined with acoustic mitigation devices (i.e acoustic liners) is therefore required to prevent the emergence of thermoacoustic instabilities. 1.2. Objective: The PhD work will focus on the development of time-harmonic acoustic propagation models for the prediction of thermoacoustic instabilities. While high-fidelity Computational Fluid Dynamics (CFD) methods, based either on Direct Numerical Simulation or Large-Eddy Simulation, may allow to accurately predict the complex physical phenomena at play (including multi-phase flows, chemical reactions, heat transfers and their non-linear interactions), they are too computationally expensive to be used at early stage design in industry. This thesis will thus focus on the development of lower- CPU linearized acoustic propagation models. These models include a simplified description of the flame dynamics (obtained either from CFD or experiments) and may be employed to examine multiple geometries and flow configurations at a reasonable computational cost. In particular, Linearized Navier-Stokes equations formulated in the frequency domain will be investigated. A special emphasis will be put on the accurate modelling of acoustic liners for the design of efficient mitigation devices for large-scale industry-relevant combustors.