Show simple item record

dc.contributor.authorIturrarán-Viveros, Ursula
dc.contributor.authorMuñoz-García, Andrés M.
dc.contributor.authorCastillo, Octavio
dc.contributor.authorShukla, Khemraj
dc.contributor.otherBarcelona Supercomputing Center
dc.date.accessioned2021-02-04T08:54:13Z
dc.date.available2021-02-04T08:54:13Z
dc.date.issued2021-02
dc.identifier.citationIturrarán-Viveros, U. [et al.]. Machine learning as a seismic prior velocity model building method for full-waveform inversion: a case study from Colombia. "Pure and Applied Geophysics", Febrer 2021,
dc.identifier.issn0033-4553
dc.identifier.urihttp://hdl.handle.net/2117/336859
dc.description.abstractWe use machine learning algorithms (artificial neural networks, ANNs) to estimate petrophysical models at seismic scale combining well-log information, seismic data and seismic attributes. The resulting petrophysical images are the prior inputs in the process of full-waveform inversion (FWI). We calculate seismic attributes from a stacked reflected 2-D seismic section and then train ANNs to approximate the following petrophysical parameters: P-wave velocity (Vp), density (ρ) and volume of clay (Vclay). We extend the use of the Vclay by constraining it with the well lithology and we establish two classes: sands and shales. Consequently, machine learning allows us to build an initial estimate of the earth property model (Vp), which is iteratively refined to produce a synthetic seismogram that matches the observed seismic data. We apply the 1-D Kennett method as a forward modeling tool to create synthetic data with the images of Vp, ρ and the thickness of layers (sands or shales) obtained with the ANNs. A nonlinear least-squares inversion algorithm minimizes the residual (or misfit) between observed and synthetic full-waveform data, which improves the Vp resolution. In order to show the advantage of using the ANN velocity model as the initial velocity model for the inversion, we compare the results obtained with the ANNs and two other initial velocity models. One of these alternative initial velocity models is computed via P-wave impedance, and the other is achieved by velocity semblance analysis: root-mean-square velocity (RMS). The results are in good agreement when we use ρ and Vp obtained by ANNs. However, the results are poor and the synthetic data do not match the real acquired data when using the semblance velocity model and the ρ from the well log (constant for the entire 2-D section). Nevertheless, the results improve when including ρ, the layered structure driven by the Vclay (both obtained with ANNs) and the semblance velocity model. When doing inversion starting with the initial Vp model estimated using the P-wave impedance, there is some gain of the final Vp with respect to the RMS initial Vp. To assess the quality of the inversion of Vp, we use the information for two available wells and compare the final Vp obtained with ANNs and the final Vp computed with the P-wave impedance. This shows the benefit of employing ANNs estimations as prior models during the inversion process to obtain a final Vp that is in agreement with the geology and with the seismic and well-log data. To illustrate the computation of the final velocity model via FWI, we provide an algorithm with the detailed steps and its corresponding GitHub code.
dc.description.sponsorshipWe are grateful for the kind comments and suggestions from two anonymous reviewers. We thank Ecopetrol S.A. within the research program 0266-2013 Colciencias Ecopetrol, for providing the well and seismic data used in this research. We would like to thank J.O. Parra for fruitful discussion regarding the inversion using nonlinear least-squares. This work was partially supported by CONACYT México under Ciencia de Frontera project number 6655, PROINNOVA projects 251910, 241763; DGAPA-UNAM project number IN107720; funding from the European Union’s Horizon 2020 Programme under the ENERXICO Project (www.enerxico-project.eu), Grant agreement no. 828947; CONACYT-SENER-Hidrocarburos Grant agreement B-S-69926. This project has also received funding from the European Union’s Horizon 2020 programme under the Marie Sklodowska-Curie Grant agreement no. 777778. Part of this work was done while one of the authors (U.I-V.) was on a sabbatical leave from the National University of México at the Barcelona Super Computing Center in Barcelona Spain as a visiting researcher with support from DGAPA-UNAM. Furthermore, O.C.-R. has been 65% cofinanced by the European Regional Development Fund (ERDF) through the Interreg V-A Spain-France-Andorra program (POCTEFA2014-2020). POCTEFA aims to reinforce the economic and social integration of the French-Spanish-Andorran border. Its support is focused on developing economic, social and environmental cross-border activities through joint strategies favoring sustainable territorial development.
dc.format.extent26 p.
dc.language.isoeng
dc.publisherSpringer Nature
dc.rightsAttribution 3.0 Spain
dc.rightsAttribution 4.0 International (CC BY 4.0)
dc.rights.urihttp://creativecommons.org/licenses/by/3.0/es/
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/.
dc.subjectÀrees temàtiques de la UPC::Enginyeria civil::Geotècnia
dc.subject.lcshMachine learning
dc.subject.lcshNeural networks (Computer science)
dc.subject.lcshSeismic engineering data report
dc.subject.otherArtificial neural networks
dc.subject.otherFull-waveform inversion
dc.subject.otherVelocity-model building
dc.subject.otherNonlinear least squares
dc.subject.otherSeismic attributes
dc.titleMachine learning as a seismic prior velocity model building method for full-waveform inversion: a case study from Colombia
dc.typeArticle
dc.subject.lemacTerratrèmols -- Amèrica del Sud
dc.subject.lemacColòmbia
dc.identifier.doi10.1007/s00024-021-02655-9
dc.description.peerreviewedPeer Reviewed
dc.relation.publisherversionhttps://link.springer.com/article/10.1007/s00024-021-02655-9
dc.rights.accessOpen Access
dc.description.versionPostprint (published version)
dc.relation.projectidinfo:eu-repo/grantAgreement/EC/H2020/828947/EU/Supercomputing and Energy for Mexico/ENERXICO
dc.relation.projectidinfo:eu-repo/grantAgreement/EC/H2020/777778/EU/Multiscale Inversion of Porous Rock Physics using High-Performance Simulators: Bridging the Gap between Mathematics and Geophysics/MATHROCKS
local.citation.publicationNamePure and Applied Geophysics


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record

Attribution 3.0 Spain
Except where otherwise noted, content on this work is licensed under a Creative Commons license : Attribution 3.0 Spain