To study the response of an aquifer system on a regional scale to massive CO2 injection, it is necessary to model the pressure response in detail near the injector while also considering the propagation of pressure perturbations over several tens of kilometers and potential interferences with other subsurface uses. This requires the ability to represent the aquifer system at a regional scale, referred here as the basin.
Thus, during this apprenticeship, we aim to address the issue of geological CO2 storage from the perspective of multi-scale numerical modeling (from the basin to the reservoir) and the multi-use of the subsurface. Based on these numerical models, we aim to predict the propagation of CO2 in underground formations, identify potential risks associated with aquifer system pressurization, potential leaks, or interference with other subsurface uses (CO2, geothermal energy, strategic aquifers, etc.). While the modeling is initially based on the basin scale, the representation of heterogeneities at this scale—crucial for predicting system disturbances—is often too coarse, representative of a large scale, and therefore insufficient to represent the heterogeneities of the porous medium that will impact the response near the well.

The objective here is to improve the representation of heterogeneities at all considered scales (basin and reservoir) by relying on geostatistical methods and complementary data. Then, by simulating different injection scenarios on one or more defined heterogeneity models, the goal is to assess the contribution of the method for predicting storage performance on a regional scale.

The work will involve:

• Reviewing the distribution of petrophysical properties in the Paris Basin model (including well constraints).
• Developing a Python code aimed at generating reservoir-scale heterogeneity from basin-scale data using statistical methods (downscaling step).
• Applying the code to introduce heterogeneity into several hypothetical CO2 storage zones extracted from the Paris Basin.
• Designing and simulating two or three different CO2 injection scenarios, including multiple CO2 injections, and comparing the results with those obtained from simulations using the initial basin petrophysical properties.

Methods:

• Geostatistical methods to better constrain the basin model using well data.
• Python programming (application of statistical methods for downscaling).
• Flow simulation in porous media.
• LGR (local grid refinement) method for coupling basin and reservoir scale models.

• Strong background in Python programming.
• Basic understanding of fluid mechanics or petrophysics.
• Good knowledge in geology and good understanding of physical processes.
• Rigor, curiosity, and motivation for innovation in an applied research context.

Master's degree student (M2 level) or final-year engineering student.
Ideally, student wishing to join IFP School: Subsurface Technologies for Sustainable Energies program, or Petroleum Engineering & Project Development program. For more information and to apply: https://www.ifp-school.com/en.

Beginner accepted.