Studies in multicomponent seismic data processing and Kronecker least-squares reverse time migration

• Author / Creator

  Gao, Wenlei

• Traditionally, seismic data processing considers data composed mainly of compressional waves. In recent years, we have witnessed the rapid development of multicomponent seismic exploration in the oil and gas industry. Multicomponent seismic data contain both compressional (P) and shear (S) wave modes permitting improving processes associated with lithology identification, fluid discrimination and fracture-stress characterization. Unfortunately, deficiencies of land multicomponent seismic data, such as the low-quality of horizontal components, inadequate spatial sampling of the elastic wavefield and misalignment of time-domain PP-wave and PS-wave images, pose unique challenges for conventional seismic processing and interpretation workflows. This thesis addresses several essential processing steps for multicomponent seismic data. Multilinear (tensor) algebra is proposed as a means to denoise and regularize onshore multicomponent data. For this purpose, I adopted the Candecomp/Parafac (CP) tensor decomposition to represent seismic prestack volumes. The decomposition is used for denoising and 5D tensor reconstruction. Randomization techniques are utilized to reduce the computational cost of the CP decomposition and making it an efficient noise attenuation tool for large multi-dimensional seismic datasets. The process of aligning PS-wave events to their corresponding PP-wave responses is called registration, which is also studied in this thesis. Registration is posed as a non-linear constrained optimization problem. I also proposed a processing flow where 5D reconstruction via low-rank tensor completion is applied to prestack PP-wave and PS-wave data independently to enhance the data quality before registration. In the second part of my thesis, I proposed a new method for least-squares reverse time migration. Inspired by techniques associated with the matrix and tensor completion, I approximate the Hessian matrix of the least-squares reverse time migration problem as the superposition of Kronecker products. The latter leads to an efficient least-squares reverse time migration algorithm that operates in the image domain.

• Subjects / Keywords

  • Least-squares reverse time migration
  • Multicomponent seismic data processing

• Graduation date

  Fall 2020

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-smhh-2g43

• 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.