Descriptors research for understanding and predicting the fluids oxidation stability by machine learning

Many fluids are used in the energy, transport and environmental sectors, for various applications ranging from the production of renewable energy to the mobility of people. These fluids are often complex mixtures whose components are mainly made up of hydrocarbon species. The complexity and chemistry of the compounds (hydrocarbons, alcohols, esters, etc.) of these fluids vary depending on the targeted application such as combustion, cooling, lubrication or even electrical insulation. Whatever the application considered, it is essential that the fluid considered preserves all its properties over time. Indeed, degradation by oxidation leads to the deterioration of the quality of the product, which can thus limit the efficiency of the system or even lead to failures. For this reason, it is necessary to have a better understanding of the key molecular parameters related to stability.
The proposed PhD thesis therefore aims to predict representative physicochemical properties of oxidation from descriptors of the fluid compounds structure or / and vibrational spectra. The objective is also to interpret the developed models (relevant descriptors, influencing factors) to better understand the stability of fluids. The first part of the thesis will be devoted to the acquisition of experimental data (spectroscopic and physico-chemical analyses). Secondly, this data will be exploited via computational methods for predictive modeling (machine learning, chemoinformatics and chemometrics). The chemoinformatics and chemometrics approaches are fundamentally similar, the only difference is the representation of fluids: in the first case, we explicitly consider the structure of the fluid components, in the second, we rely on the results of spectroscopy analyses. The originality of the subject lies in a coupling of these two approaches for the understanding and prediction of oxidation stability, and will allow the selected candidate to interact with experts from different fields within IFPEN and LASIRE in order to acquire solid skills.

Keywords: Liquid Phase Oxidation, Chemoinformatic, Chemiometry,

• Academic supervisor    Pr. Ludovic DUPONCHEL (ORCID : 0000-0002-7206-4498)  Pr. Cyril RUCKEBUSCH ( ORCID : 0000-0001-8120-4133) LAboratoire de Spectroscopie pour les Interactions, la Réactivité et l'Environnement (LASIRE) - University of Lille, France
• Doctoral School    ED104 SMRE, www.edsmre.univ-lille1.fr/
• IFPEN supervisor    Dr. GIARRACCA Lucia, Research engineer, Combustion Systems and Fuels Department, lucia.giarracca@ifpen.fr,
• PhD location    IFP Energies nouvelles, Rueil-Malmaison, France
• Duration and start date    3 years, starting in fourth quarter 2021
• Employer    IFP Energies nouvelles, Rueil-Malmaison, France
• Academic requirements    University Master degree in Chemical sciences or Chemical Engineering s
• Language requirements    Fluency in English, and in French or willingness to learn French
• Other requirements    Chemoinformatics, Chemometrics, Spectroscopy.