Building long-wavelength velocity for salt structure using stochastic full waveform inversion with deep autoencoder based model reduction
Zhaoqi Gao,Zhibin Pan,Jinghuai Gao,Zongben Xu
DOI: https://doi.org/10.1190/segam2019-3215572.1
2019-01-01
SEG Technical Program Expanded Abstracts
Abstract:PreviousNext No AccessSEG Technical Program Expanded Abstracts 2019Building long-wavelength velocity for salt structure using stochastic full waveform inversion with deep autoencoder based model reductionAuthors: Zhaoqi GaoZhibin PanJinghuai GaoZongben XuZhaoqi GaoNational Engineering Laboratory for Offshore Oil Exploration, Xi’an Jiaotong University, ChinaSearch for more papers by this author, Zhibin PanNational Engineering Laboratory for Offshore Oil Exploration, Xi’an Jiaotong University, ChinaSearch for more papers by this author, Jinghuai GaoNational Engineering Laboratory for Offshore Oil Exploration, Xi’an Jiaotong University, ChinaSchool of Electronic and Information Engineering, Xi’an Jiaotong University, ChinaSearch for more papers by this author, and Zongben XuSchool of Mathematics and Statistics, Xi’an Jiaotong University, ChinaSearch for more papers by this authorhttps://doi.org/10.1190/segam2019-3215572.1 SectionsAboutPDF/ePub ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InRedditEmail AbstractBuilding velocity model for salt structure remains challenging because the strong heterogeneity of medium. Full waveform inversion (FWI) is such a technique that enables us to build a high-resolution velocity model for salt structure. However, it needs an accurate enough initial model to prevent cycle-skipping during data fitting. In this work, we propose a stochastic FWI method based on global optimization to build such an initial model. To mitigate the “curse of dimensionality” problem of global optimization, we embed a deep learning technique called deep autoencoder into the proposed method. Benefiting from the dimensionality reduction characteristic of deep autoencoder, the proposed method can transform the original large model dimensional FWI problem into a low model dimensional one that can be effectively optimized by global optimization. Numerical results verify that the proposed method can build a reasonable long-wavelength velocity model for salt structure that can then be used as an initial model for FWI.Presentation Date: Wednesday, September 18, 2019Session Start Time: 1:50 PMPresentation Time: 2:40 PMLocation: Poster Station 9Presentation Type: PosterKeywords: full-waveform inversion, salt dome, machine learningPermalink: https://doi.org/10.1190/segam2019-3215572.1FiguresReferencesRelatedDetailsCited byML-descent: An optimization algorithm for full-waveform inversion using machine learningBingbing Sun and Tariq Alkhalifah21 October 2020 | GEOPHYSICS, Vol. 85, No. 6 SEG Technical Program Expanded Abstracts 2019ISSN (print):1052-3812 ISSN (online):1949-4645Copyright: 2019 Pages: 5407 publication data© 2019 Published in electronic format with permission by the Society of Exploration GeophysicistsPublisher:Society of Exploration Geophysicists HistoryPublished Online: 10 Aug 2019 CITATION INFORMATION Zhaoqi Gao, Zhibin Pan, Jinghuai Gao, and Zongben Xu, (2019), "Building long-wavelength velocity for salt structure using stochastic full waveform inversion with deep autoencoder based model reduction," SEG Technical Program Expanded Abstracts : 1680-1684. https://doi.org/10.1190/segam2019-3215572.1 Plain-Language Summary Keywordsfull-waveform inversionsalt domemachine learningPDF DownloadLoading ...