Thermoacoustic Instabilities in Annular Combustors
Lead Research Organisation:
Imperial College London
Department Name: Mechanical Engineering
Abstract
Next generation combustors are able to produce signifcantly lower pollutants, however as a consequence of the operating conditions are particularly susceptible to a class of instabilities known as thermoacoustic instabilities. The latter is the resonant coupling in a positive feedback loop between sound and unsteady heat release which can lead to continuously increasing oscillations and potential structural failure of the engine. An annular combustor has unique unstable thermoacoustic modes, particularly in the azimuthal direction which are significantly less understood. This PhD looks at the underlying mechanisms for instability triggering, growth and even mode bifurcation in time. Tools employed are numerical in nature, such as LES/DNS, dynamic mode decomposition, phase space reconstruction, reduced order models etc.
Organisations
People |
ORCID iD |
| Salvador Navarro-Martinez (Primary Supervisor) | |
| Omer Rathore (Student) |
Publications
Studentship Projects
| Project Reference | Relationship | Related To | Start | End | Student Name |
|---|---|---|---|---|---|
| EP/N509486/1 | 01/10/2016 | 31/03/2022 | |||
| 2168694 | Studentship | EP/N509486/1 | 01/10/2017 | 30/06/2021 | Omer Rathore |
| Description | Thermoacoustic instabilities (where flame heat release and "noise" couple together) are one of the limiting factors on the adoption of low emissions, cleaner engine gas turbine combustors. A primary cause of the complexity is the presence of multiple flames in real combustion chambers and their interaction. One of the key research areas with initial work was to simulate two neighbouring flames undergoing acoustic forcing across a wide range of frequencies. Pinch off at the flame tip was observed across a range of scales and shown to be a noise source which could further destabilise the base flame. The presence of secondary reaction zones was also observed and suggested to be a potentially significant source of noise, that is naturally not captured with the use of simplified chemistry. The results were presented at the European Combustion Meeting 2019 (ECM19). Current work includes extending the previous simulations to a twin burner case study inspired from recent experiments. Most high order, compressible flow simulation codes suffer from the presence of spurious numerical noise at the computational boundaries. This is particularly crucial to minimise for thermoacoustic simulations as tainting of the acoustic field can lead to non-physical results. Various suggested methods found from literature were implemented with focus being on non-reflective characteristic boundary conditions. Contrary to literature, it was found that retention of chemical source terms in the ghost cell treatment brought no significant improvement to the system. The open question of how far away boundary treatments should be from regions of physical interest as well as the choice of user-defined parameters was also discussed and presented at the Numerical Combustion Conference 2019 (NC19). Currently work is also being done to understand the indirect noise generation from acceleration of compositional inhomogenities through a nozzle, as is often the case downstream of the combustor in a gas turbine engine. This work is being undertaken in collaboration with colleagues at Cambrige University who have run the experiments for us to verify simulations against. Initial finding have stressed the importance of choosing the correct relaxation parameters in the characteristic boundary treatment. |
| Exploitation Route | The work on boundary conditions is of relevance to any compressible flow solvers since it is a problem that has been observed for quite some time yet still poses several open questions. As a general overview, ideally the computational boundary should be situated as close to the region of interest as possible, however the closer it is placed the greater domain/boundary interactions are likely to occur. Meanwhile the work on flame-flame interaction is of particular use to designers of gas turbines, particularly the low pollutant generating next-gen turbines that are prone to thermoacoustic modes. Use of the modified LES-ATF model could also be of interest to the general combustion community as it offers the potential for significant computational savings. |
| Sectors | Aerospace, Defence and Marine,Energy |