Thesis offer: reverse homogenization for the characterization of fault systems

map faults in the basement and understand their properties (CThat is, their geometry, their dimensions, their behavior as a barrier, their growth) is essential to guide decision-making in applications related to underground resources. todaytoday, the interpretations of the faults in depth come mainly from high-resolution models obtained by D-shape inversionComplete wave (FWI, Full WaveForm Inversion), in addition to migrated images. although the ease and precision of theimaging andseismic interpretation continues to progress, problems such as limited seismic data bandwidth, noise,approximation of the physical model and incomplete data coverage remain major challenges for theDetailed imaging of faults (for example, those whose rejection is lower than the seismic vertical resolution) (Dimmen, Rotevatn & Lecomte, 2023). Underestimating the associated uncertainties can lead to overly optimistic model-based forecasts and increase the financial risk of projects dBasement exploitation.

In recent work, Ruggiero, Cupillard & Caumon (2024) propose a two-scale approach that combines the FWI with a refinement inversion to probabilistically estimate the da fault (for example: length, dip, rejection). This approach, based on previous work (Hedjazian, Capdeville & Bodin, 2021; Santos et al., 2024), is called reverse homogenization. Assuming that the FWI provides a smooth representation of real structures, it aims to find all the finer scale models compatible with the FWI solution. As part of this project, we proposeApply this approach to a fault system, cis to say not to a single flaw, but to several, potentially connected. After having tested the method on a synthetic case in 2D, a real case and/or a 3D extension will be considered.