Analysis of Intrinsic Instabilities
in Premixed Hydrogen Flames
For my master’s thesis, I investigated how lean premixed hydrogen flames respond to strain and intrinsic instabilities, with the broader goal of understanding how these factors affect hydrogen’s potential as a clean aviation fuel. While hydrogen avoids CO₂ emissions, it still forms nitrogen oxides (NOx), which are subject to increasingly strict aviation regulations. Reducing NOx while keeping the combustion stable is therefore critical to the future of hydrogen-powered flight.
Why does this problem matter?
Previous research showed that in a counterflow arrangement (shown on the right)—where cold premixed reactants enter from one side and a stream of hot combustion products enters from the other—the flame is strained into a thin sheet, which reduces NOx formation (by suppressing certain NOx formation pathways). This suggests a promising approach toward regulatory compliance for low-emission hydrogen combustion.
However, flames also exhibit intrinsic instabilities: small perturbations of the flame front that can arise naturally from the reaction–diffusion balance. These can grow into corrugations or wrinkles, altering flame shape and ultimately influencing both stability and emissions. The video to the right shows an example of how this occurs in a free flame, where small perturbations were introduced in the CFD simulation to initiate the instability, since they would not otherwise emerge in a perfect DNS setup.
Technical Hurdles Faced
During early simulation runs, the pressure field exhibited fluctuations at consecutive time steps, indicating that pressure wave reflections were occurring due to fixed inlet and outlet boundaries. These reflections led to insufficient dissipation of acoustic waves, causing their accumulation within the domain and eventually resulting in disruptive oscillations. This phenomenon is illustrated by the two GIFs to the left: one displays the fluctuating pressure field at consecutive time steps, while the other illustrates a simplified one-dimensional pressure (acoustic) wave reflecting off a fixed boundary point to highlight how such wave reflections contribute to oscillations in the simulation domain. These observations underlined the need for advanced boundary treatments to maintain physically consistent and stable pressure behaviour—an issue I addressed using Navier–Stokes Characteristic Boundary Conditions (NSCBCs).
Initialising the Simulation Data
To quantify how instabilities evolve, I programmed routines that introduced controlled sinusoidal perturbations onto steady flame surface solutions, as shown in a GIF to the right. By varying both the amplitude and wavelength of these disturbances, this automated system enabled a comprehensive parametric evaluation of instability growth across multiple strain rates. The results of these simulations were carefully validated through mesh sensitivity analyses, confirming the convergence of instability growth measurements and reinforcing the robustness of the findings, also displayed in two of the figures to the right.
Results Visualised
Dynamic behaviour of the flamefront under different perturbation conditions and strain rates was visualised in animated sequences (videos) to the left, capturing how increasing tangential strain consistently dampened flame corrugation and instability growth. Comparative analysis against free-flame cases from the literature revealed that disturbance wavelength had a stronger influence on flame stability than amplitude—an insight critical for designing practical hydrogen combustors.
Together, these stages—from literature review, simulation methodology, implementation of appropriate boundary conditions, analysis of results, to rigorous validation and dynamic visualisation—highlight the scientific rigor and technical problem-solving depth of the thesis. The figures included throughout the project illustrate this comprehensive approach, offering clear evidence of my ability to carry out advanced CFD research relevant to stable, low-emission hydrogen combustion systems.