Mijail Littin, doctoral student at the CORIA laboratory, will defend his thesis entitled “Characterizatiοn οf sοοt prοperties in laminar diffusiοn flames by cοupling οptical techniques: applicatiοn tο sustainable aerοnautic fuels” on Thursday, December 11 at 2 p.m. in the CORIA conference room (St Etienne du Rouvray).
The defense will be public and accessible remotely via the following link:
https://zoom.us/j/99809118740
Abstract :
This thesis addresses the need for precise soot modeling in next-generation propulsion systems by developing advanced non-intrusive diagnostics—combining laser-based and data-driven methods—to quantify key soot properties (volume fraction, radius of gyration, primary particle size, maturity, and composition) in laminar diffusion flames and realistic combustion environments.
Key Innovations:
Spline-based Abel Transform (SAT): Reconstructs axisymmetric soot fields from line-of-sight measurements, incorporating curvature regularization and corrections for signal-trapping effects (attenuation, self-absorption) to accurately retrieve soot temperature and scattering coefficients.
Multi-wavelength, multi-angle light scattering (MALS) paired with Rayleigh-Debye-Gans theory for fractal aggregates (RDG-FA) quantifies aggregate size distributions.
SAXS Interpretation: A novel light–particle interaction model enables direct assessment of primary particle size distributions using Small-Angle X-ray Scattering (SAXS).
Unified Approach: A physics-informed neural network (PINN) embeds radiative transfer and RDG-FA equations, allowing simultaneous inversion of extinction, scattering, and emission data without calibration constants. This reduces biases inherent in traditional methods by grounding the inversion in physical laws. Among others, this allows to report 2D maps of the optical index of soot.
Applied to Jet A-1 and HEFA–SPK flames, the framework reveals soot’s compositional evolution—from organic to amorphous and graphitic forms—along particle trajectories, driven by temperature and residence time. This establishes a foundation for modeling soot maturity.
The work provides a comprehensive framework for quantitative soot diagnostics, integrating optical methods (X-ray to visible) with physics-informed machine learning. It paves the way for unified models to describe soot formation, growth, and oxidation in practical combustion systems.