From Organic Geochemistry to Statistical Thermodynamics: the Development of Simulation Methods for the Petroleum Industry

This work is a synthesis of researches to which the author contributed successively in three areas of applied research. The first of these areas is petroleum organic geochemistry, which is aimed to understand the formation of petroleum reservoirs from detailed analyses of the composition of fossil fuels, experimental simulations in the laboratory by pyrolysis, and reconstructions of the geological history of the sedimentary basins concerned. These reconstructions, which imply the mathematical simulation of many coupled mechanisms such as thermal cracking, fluid flow in porous media, sedimentation, sedimentary rock ;compaction and inter-phase equilibria, were the subject of the author's main contributions in this field. The results obtained at IFP in the 1980s, as part of a multidisciplinary team project, provided the basis for numerous industrial applications in connection with the new discipline of basin models. The second of these areas is petroleum thermodynamics, which involves measuring and predicting the equilibrium and transport properties in crude oil and natural gas production and processing. The contribution of the work presented in this vast area was to characterize petroleum fluids at high pressure, ;both by experimental identification of the phase diagrams specific to certain HP-HT fluids, and the prediction of their properties by means of equations of state. It also allowed a more accurate understanding of how the phase equilibria between oil and gas were likely to alter the composition of petroleum fluids. The third area of study is statistical thermodynamics, whose applications have recently soared with the generalization of molecular simulation. Monte Carlo and molecular dynamics methods not only offer an understanding at microscopic scale, but also a more reliable prediction of the equilibrium and transport properties in many cases where classical macroscopic models are limited by their empirical character. By relying on this occasion on close collaboration with a university team, it was possible to implement improved algorithms and to develop a new intermolecular potential to meet the needs of the petroleum industry. Thus effective applications of molecular simulation materialized in the processing of H2S gases, separations by adsorption in zeolites, and the calculation of properties of commercially unavailable hydrocarbons. These three areas, which appear very different at first glance, display significant similarities, like their dominant articulation around physical chemistry, their need of new software, and the empirical optimization of parameters whereof a physical meaning is nonetheless expected