Waveform inversion in acoustic orthorhombic media with a practical set of parameters

Nabil Masmoudi, Tariq Ali Alkhalifah

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Scopus citations

Abstract

Full-waveform inversion (FWI) in anisotropic media is overall challenging, mainly because of the large computational cost, especially in 3D, and the potential trade-offs between the model parameters needed to describe such a media. We propose an efficient 3D FWI implementation for orthorhombic anisotropy under the acoustic assumption. Our modeling is based on solving the pseudo-differential orthorhombic wave equation split into a differential operator and a scalar one. The modeling is computationally efficient and free of shear wave artifacts. Using the adjoint state method, we derive the gradients with respect to a practical set of parameters describing the acoustic orthorhombic model, made of one velocity and five dimensionless parameters. This parameterization allows us to use a multi-stage model inversion strategy based on the continuity of the scattering potential of the parameters as we go from higher symmetry anisotropy to lower ones. We apply the proposed approach on a modified SEG-EAGE overthrust synthetic model. The quality of the inverted model suggest that we may recover only 4 parameters, with different resolution scales depending on the scattering potential of these parameters.
Original languageEnglish (US)
Title of host publicationSEG Technical Program Expanded Abstracts 2017
PublisherSociety of Exploration Geophysicists
DOIs
StatePublished - Aug 17 2017

Bibliographical note

KAUST Repository Item: Exported on 2020-10-01
Acknowledgements: We would like to thank KAUST for financial support and SWAG members for many useful discussions. For computer time, this research used the resources of the Supercomputing Laboratory in KAUST.

Fingerprint

Dive into the research topics of 'Waveform inversion in acoustic orthorhombic media with a practical set of parameters'. Together they form a unique fingerprint.

Cite this