PhD Thesis on 3D elastic seismic imaging
Wednesday 26 August 2009 by Jean Virieux

Université Joseph Fourier Observatoire des Sciences de l’Univers de Grenoble

Ecole doctorale « T.U.E »

Allocations de recherche : campagne 2009-2010

Proposition d’un sujet de thèse

Title: 3D elastic high resolution seismic imaging [download].


PhD promotors: Stéphane Operto (GEOAZUR), Anne Sieminski (LGIT) & Jean Virieux (LGIT)

Research laboratory : Laboratoire de Géophysique Interne et Tectonophysique


PhD subject (15 lines) : The interpretation of the waveform allows the reconstruction of crustal structures with a resolution of half the wavelength. 3D acoustic applications start to be achieved for well-identified targets as the oil field VALHALL in the North Sea.

We propose in this research investigation the elastic reconstruction of the P-wave and S-wave velocity in order to have better physical insights of sampled media by waves, especially concerning fluid contents of such target. The VALHALL target will the selected object after a training of forward methods for wave propagation and optimisation methods for the reconstruction of the medium.

The expected method for the forward modelling is a method of time integration called the Discontinuous Galerkin method which is quite suited for computer parallelisation. The compute code exists and the student should be trained in managing the meshing as well as the different criterion for running the modelling.

The optimisation method is a local method based on techniques derived from the Newton method or the Gauss-Newton method. A particular attention will be put for the non-diagonal estimation of the inverse of the Hessian matrix which is crucial for the estimation of the two parameters. The computer code should be improved from the one built up during another thesis.

The application to the real dataset of VALHALL is the objective of this thesis and requires, therefore, not only skilfulness in the high performance computing challenge but also in the interpretation of the reconstructed parameters and the fitness of the recorded data. Fluid characterisation will be of importance as one of expected conclusions.


Interne collaborators: E. Chaljub (LGIT), A. Ribodetti (GEOAZUR) and R. Brossier (LGIT)


Industrial support: consortium SEISCOPE


For more information, please contact: