Parameterization of the SAFT thermodynamic model using machine learning

Status

Open

Scientific disciplines

Chemical Sciences

Research direction

Applied Physico-chemistry and Mechanics

Affiliate site

Rueil-Malmaison

The use of biomass is a promising alternative for the synthesis of various families of high value-added chemicals, which can be used for the synthesis of new products. Products stemming from biomass are mainly composed of oxygenated molecules, but the models and methods traditionally used in the chemical industry have historically been developed to predict properties of hydrocarbons. The design of new manufacturing processes with bio-based molecules requires methods for rapid and accurate estimation of the physicochemical properties of such molecules.
Thermodynamic models - equation of state (EoS), in particular the EoS PC-SAFT (for “Perturbed Chain Statistical Associating Fluid Theory”) and its derivatives are widely used to compute the properties of fluids for targeted conditions of temperature and pressure. These models are based on long-established statistical thermodynamics which ensures good extrapolation capacities, but they require parameters specific of the fluid to be studied. Group contributions have been developed to predict some of these parameters, thus highlighting the link between the structure of the considered fluid components and the values of associated parameters.
In this PhD subject, we propose to implement methods based on data science to develop new approaches to parameterize the PC-SAFT EoS, and thus extend its field of application to the area of chemical processes. The envisaged combination of machine learning and a thermodynamic model aims at improving the predictive capacity of each of these models taken individually. Several approaches will be investigated: (i) approaches consisting in predicting the EoS parameters by means of QSPR (quantitative structure-property relationship) based models, and (ii) an approach mixing physics and machine learning.
The candidate will be an employee of Fives ProSim, and the work carried out at IFPEN in close collaboration with research groups at Université de Lorraine.

Keywords: Equation of states, Machine Learning

  • Academic supervisor    Pr. Jean-Charles De HEMPTINNE (ORCID : 0000-0003-1607-3960), Research engineer, Thermodynamics and Molecular Simulation Department.
  • Doctoral School    ED388 Chimie Physique et Chimie Analytique de Paris Centre, ed388.upmc.fr/
  • IFPEN supervisor    Dr. Benoît CRETON (ORCID: 0000-0002-3287-877X), Research engineer, Thermodynamics and Molecular Simulation Department.
  • PhD location    IFP Energies nouvelles, Rueil-Malmaison, France
  • Duration and start date    3 years, starting from early 2025
  • Employer    Fives ProSim, Toulouse, France.
  • Academic requirements    University master’s degree in Statistics/Data sciences, Chemical/Physical sciences or Chemical Engineering 
  • Language requirements    Fluency in English, and in French or willingness to learn French
  • Other requirements    Thermodynamics, Chemoinformatics, Machine Learning.

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

Contact
Encadrant IFPEN :
Dr. Benoît CRETON