Development of an AI-based strategy to accelerate the computation of chemical kinetics in Sustainable Aviation Fuel spray combustion

Air traffic is responsible for an increasingly significant share of global CO₂ emissions. The use of Sustainable Aviation Fuels (SAFs) offers a promising path toward reducing the carbon footprint of the aviation sector. Although SAFs exhibit physico-chemical properties similar to conventional kerosene, their combustion behavior can differ significantly, requiring adjustments and optimization of current gas turbines (GT). In this context, numerical simulation plays an essential role for the aerospace industry, enabling cost-effective design and optimization of GT systems. However, several challenges remain, particularly in achieving both accurate and computationally efficient simulations. One of the main bottlenecks is the numerical integration of chemical kinetics, which is computationally expensive due to the large number of species and reactions involved in SAF combustion. Recent advances have demonstrated that machine learning techniques, particularly neural networks, can significantly accelerate chemical kinetics computations. Nevertheless, most of these developments have focused on conventional hydrocarbons under purely gaseous conditions. In contrast, SAF combustion in GTs occurs in a multiphase regime, where complex interactions between liquid fuel droplets and the flame must be taken into account. The objective of this thesis is to extend machine learning-based chemical kinetics acceleration methods to the multiphase combustion of SAFs injected as sprays. A key challenge lies in generating a suitable training dataset that accurately reflects the operating conditions of industrial systems. The work will build upon a methodology previously developed at CORIA and IFPEN, based on the coupled simulation of interacting 0D reactors, which will be adapted to account for spray combustion dynamics. The research will initially focus on a canonical laboratory flame, before being applied to a configuration representative of an actual SAF-fueled aero-engine burner.

Keywords: Sustainable Aviation Fuel, Artificial Intelligence, Chemical kinetics, Computing acceleration

• Academic supervisor    Pr Luc VERVISCH, CORIA, ORCID : 0000-0003-0313-2060
• Doctoral School    ED591 PSIME, INSA Rouen Normandie
• IFPEN supervisor    Dr Cédric MEHL, ORCID : 0000-0003-2293-9281
• PhD location    IFPEN Rueil-Malmaison, France 
• Duration and start date    3 years, starting in the fourth quarter 2025 (Novembre 3)
• Employer    IFPEN
• Funding    Currently under instruction
• Academic requirements    University Master degree involving CFD, physics and/or numerical modelling
• Language requirements    English level B2 (CEFR)    
• Other requirements    Programming skills (Python, C++)

To apply, please send your cover letter and CV to the IFPEN supervisor indicated here below.