Machine learning applied to the characterization of fluids used in electric propulsion chains.

This thesis is part of the field of the evolution of thermal management of electric propulsion chains, a key element for improving the performance of electric vehicles in terms of power density. Automobile manufacturers are moving towards the use of a “single fluid” for managing the cooling and lubrication of the propulsion chain. However, the current challenge lies in formulating these fluids to withstand multiple physical stresses. In this context, we propose a thesis aimed at developing machine learning models to optimize the formulation of these heat transfer fluids based on their elemental and molecular characterization. This research work is distinguished by its approach aimed at understanding the impact of the formulation of fluids on their physicochemical properties. It aims to predict suitable formulations for target properties or to predict properties for a given formulation using supervised machine learning. IFPEN has already explored the application of machine learning on physicochemical property databases, using chemometric methods that link the chemical characteristics of fluids to the properties of interest. 
Thesis stages:
• Identification of the relevant physicochemical properties of fluids used in electric propulsion chains.
• Identification of the different types of fluids currently used in the field but also in the future.
• Definition then optimization of the most appropriate analysis techniques (LC-HRMS, GCxGC/MS, NMR, etc.) to characterize these fluids at the molecular and elementary level.
• Experimental characterization of fluids to evaluate the targeted physical properties (viscosity, conductivity, thermal capacity, electrical conductivity, etc.) and creation of a database.
• Development and validation of chemometric models based on experimental data to predict targeted properties based on formulations.

Keywords: Fluid formulation, Lubricants, Electric mobility, Analysis, Chemometrics

• Academic supervisor    Prof. Ludovic DUPONCHEL, LASIRE, ORCID : 0000-0002-7206-4498
• Doctoral School   ED104  SMRE, Université de Lille
• IFPEN supervisor    Dr Lucia GIARRACCA-MEHL, ORCID : 0000-0001-6238-268X
• PhD location    IFPEN, Lyon, France
• Duration and start date    3 years, starting in the fourth quarter 2024 (Novembre 4)
• Employer    IFPEN
• Academic requirements    Master en chimie analytique
• Language requirements    English level B2 (CEFR) Willingness to learn French 
• Other requirements    Knowledge of chemometrics or multivariate data processing

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