3D Multi-parameters Full Waveform Inversion for challenging 3D elastic land targets

par Phuong-thu Trinh

Projet de thèse en Sciences de la Terre et de l'Univers et de l'Environnement

Sous la direction de Jean Virieux et de Ludovic Metivier.

Thèses en préparation à Grenoble Alpes , dans le cadre de Terre, Univers, Environnement , en partenariat avec Institut des Sciences de la Terre (laboratoire) depuis le 01-10-2015 .


  • Résumé

    This PhD project aims to tackle 3D elastic full waveform inversion for challenging land targets. This project will be done in the frame of the SEISCOPE consortium, and directly supported by TOTAL. Full Waveform Inversion appears to be a promising technology for quantitative multi-parameters seismic tomography. While the original idea has been proposed more than 30 years ago (Tarantola, 1984), the method gained popularity during the last 15 years, thanks from one hand, to the dramatic improvement of the data quality and azimuth coverage, and from the other hand, to the increase of the available computing power. However, most of these applications are done under the acoustic approximation of the Earth, in a velocity model building fashion in order to reconstruct high resolution P-wave velocity models (and sometimes an additional garbage collector parameter) for improved further migration. Also, these applications are often restricted to marine environments, or land targets for which elastic effects can be mitigated, to conform the acoustic approximation of the wave propagation. The industry now faces the challenging issue of inversion in complex land areas, in which elastic effects are significant in the wave-field and cannot be approximated anymore. Among the challenging land issues, the inversion algorithms need now to be able to deal with complex topographies, weathering zones, foothills environments and high contrasts (karst, carbonates and shales) in the near surface. Even if some recent approaches have been proposed to mitigate the free-surface effects (Plessix and Perez Solano, 2015), surface wave also strongly dominate the wave-field. In order to tackle this challenging project, and depending on the data that will be made available by TOTAL, the following issues could be tackled during the PhD: Efficient modeling tool, Surface wave, Hierarchical inversion schemes, Multi-component analysis

  • Titre traduit

    3D Multi-parameters Full Waveform Inversion for challenging 3D elastic land targets


  • Résumé

    This PhD project aims to tackle 3D elastic full waveform inversion for challenging land targets. This project will be done in the frame of the SEISCOPE consortium, and directly supported by TOTAL. Full Waveform Inversion appears to be a promising technology for quantitative multi-parameters seismic tomography. While the original idea has been proposed more than 30 years ago (Tarantola, 1984), the method gained popularity during the last 15 years, thanks from one hand, to the dramatic improvement of the data quality and azimuth coverage, and from the other hand, to the increase of the available computing power. However, most of these applications are done under the acoustic approximation of the Earth, in a velocity model building fashion in order to reconstruct high resolution P-wave velocity models (and sometimes an additional garbage collector parameter) for improved further migration. Also, these applications are often restricted to marine environments, or land targets for which elastic effects can be mitigated, to conform the acoustic approximation of the wave propagation. The industry now faces the challenging issue of inversion in complex land areas, in which elastic effects are significant in the wave-field and cannot be approximated anymore. Among the challenging land issues, the inversion algorithms need now to be able to deal with complex topographies, weathering zones, foothills environments and high contrasts (karst, carbonates and shales) in the near surface. Even if some recent approaches have been proposed to mitigate the free-surface effects (Plessix and Perez Solano, 2015), surface wave also strongly dominate the wave-field. In order to tackle this challenging project, and depending on the data that will be made available by TOTAL, the following issues could be tackled during the PhD: Efficient modeling tool, Surface wave, Hierarchical inversion schemes, Multi-component analysis