High-fidelity numerical modelling of hydrogen-methane combustion

This thesis focuses on the development and validation of high-fidelity numerical models for turbulent H₂–CH₄ combustion in gas turbine applications. It investigates the predictive capabilities of the Dynamic Thickened Flame for LES (DTFLES) model, addressing challenges in multi-regime and multi-stream combustion. The work introduces virtual chemistry techniques to reduce computational cost while preserving key physical effects. A novel method for unburnt mixture reconstruction and an improved thickening strategy (T-DTFLES) are also proposed. The models are validated through comparisons with experiments and DNS, supporting the design of flexible, low-emission combustors.