Strategies for elastic full waveform inversion

• Author / Creator

  Matharu, Gian

• The advent of modern supercomputers, in conjunction with larger, more comprehensive datasets, has led to a paradigm shift in seismic imaging. Full waveform inversion is routinely employed
  as a tool to estimate subsurface properties of the Earth with high resolution. The method fits simulated waveforms to observed data by iteratively updating estimates of subsurface properties. While recent advances have fostered seismic imaging success in areas with complex subsurface geology, a variety of challenges persist. Underdeveloped topics include the estimation of multiple physical parameters, uncertainty quantification, robust convergence, and the incorporation of more complex physics.

  This thesis focuses on multi-parameter inversion in isotropic, elastic full waveform inversion. The transition from acoustic to elastic waveform inversion increases the computational cost, data complexity, and the ill-posed nature of the inverse problem. Estimating multiple independent subsurface parameters is challenging due to the limited, or overlapping, sensitivity of data to different parameters. In this thesis, I explore approaches to accelerate elastic full waveform inversion through simultaneous sources (Chapter 3) and second-order stochastic optimization (Chapter 4). Performance is assessed through controlled numerical experiments. Using the acoustic formulation, I present two forms of resolution/uncertainty analysis predicated on an approximation of the Hessian as a superposition of Kronecker products (Chapter 5). The final chapter compares applications of 2D acoustic and elastic full waveform inversion to a land dataset from the western Canadian basin (Chapter 6). I devise a workflow that includes data-preprocessing, initial model building and inversion.

• Subjects / Keywords

  • Non-linear optimization
  • Inverse problems
  • Seismic imaging
  • PDE-constrained optimization
  • Full waveform inversion

• Graduation date

  Fall 2020

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-yfey-v975

• License

  Permission is hereby granted to the University of Alberta Libraries to reproduce single copies of this thesis and to lend or sell such copies for private, scholarly or scientific research purposes only. Where the thesis is converted to, or otherwise made available in digital form, the University of Alberta will advise potential users of the thesis of these terms. The author reserves all other publication and other rights in association with the copyright in the thesis and, except as herein before provided, neither the thesis nor any substantial portion thereof may be printed or otherwise reproduced in any material form whatsoever without the author's prior written permission.