Density reconstruction in multiparameter elastic full-waveform inversion

Sun, Minao; Yang, Jizhong; Dong, Lu; Liu, Yuzhu; Huang, Chao

doi:10.1088/1742-2140/aa93b0

Cited by 15 publications

(9 citation statements)

References 41 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tarantola et al Tarantola (1986) proposed an inversion strategy that the P-velocity, S-velocity, and density will be estimated in sequence, according to the sensibility to the different wavelength information. Some two-stage inversion strategies were proposed later Jeong et al (2012); Sun et al (2017). In addition, some studies Jeong et al (2012); Guo and Alkhalifah (2017); Sun et al (2017) have proven that the P-velocity is the easiest parameter to estimate, and the density is opposed.…”

Section: Proposed Elasinvnetmentioning

confidence: 99%

“…Some two-stage inversion strategies were proposed later Jeong et al (2012); Sun et al (2017). In addition, some studies Jeong et al (2012); Guo and Alkhalifah (2017); Sun et al (2017) have proven that the P-velocity is the easiest parameter to estimate, and the density is opposed.…”

Section: Proposed Elasinvnetmentioning

confidence: 99%

“…Others Sun et al (2017); Jeong et al (2012); Guo and Alkhalifah (2017); Sun et al (2017) also noted the different complexity to inverse three elastic parameters and suggested utilizing the prior information for the more difficult task.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Multi-Task Deep Learning for Multi-Parameter Elastic Inversion

Li,

Jiang,

Yang

et al. 2024

Preprint

View full text Add to dashboard Cite

Elastic waveform inversion plays a vital role in estimating the Earth's subsurface property. The inversion of multiple elastic parameters from observation data has been regarded as a cutting-edge and challenging issue due to its severe non-linearity and ill-posed nature. Deep learning approaches have recently demonstrated incredible potential in simulating non-linear mapping and made remarkable achievements in geophysical inversion. In this work, we consider multi-parameter elastic inversion as a multi-task learning problem and propose to accomplish the three tasks sequentially by a deep neural network with sequential decoders, which we name ElasInvNet. Specifically, we reconstruct the three elastic parameters, P-velocity, S-velocity, and density, one after the other, and use the features from the former task as the prior information to assist the later tasks' reconstruction. We verified the effectiveness of the proposed ElasInvNet through comprehensive comparisons and ablation studies.

show abstract

Section: Proposed Elasinvnetmentioning

confidence: 99%

Section: Proposed Elasinvnetmentioning

confidence: 99%

See 1 more Smart Citation

Multi-Task Deep Learning for Multi-Parameter Elastic Inversion

Li,

Jiang,

Yang

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…It uses all the available dynamic information in the seismograms (amplitudes, travel times and frequencies) to obtain high‐resolution images of the physical properties of the subsurface. It has been applied to a variety of model types, including viscoacoustic media (Gauthier et al., 1986; Keating & Innanen, 2019; Operto et al., 2004), viscoelastic media (Brossier, 2011; Brossier et al., 2009; Mora, 1987; Sun et al., 2017) and even poroelastic media (Barros et al., 2010; Yang & Malcolm, 2021), by iteratively minimizing the misfit between the simulated and observed seismograms (Virieux & Operto, 2009; Virieux et al., 2017). Although FWI was originally proposed in the time domain (Tarantola, 1984), it has also been developed in the frequency domain (Pratt et al., 1998; Shin et al., 2007; Sirgue & Pratt, 2004; Zhou & Greenhalgh, 2003) as well as in the Laplace domain (Ha & Shin, 2012; Shin & Ho Cha, 2009).…”

Section: Introductionmentioning

confidence: 99%

“…However, 3‐D FWI is still quite expensive not only for the data acquisition but also in terms of computer cost (memory storage and runtime). Multi‐parameter FWI presents a challenge and requires a novel scheme to overcome the drawback of trade‐offs or crosstalk between the different parameters (Keating & Innanen, 2019; Sun et al., 2017; Yang & Malcolm, 2021).…”

Section: Introductionmentioning

confidence: 99%

Numerical Fréchet derivatives of the displacement tensor for 2.5‐D frequency‐domain seismic full‐waveform inversion in viscoelastic TTI media

Yang,

Zhou,

Engsig

et al. 2023

Near Surface Geophysics

View full text Add to dashboard Cite

Derivatives of the displacement tensor with respect to the independent model parameters of the subsurface, also called Fréchet derivatives (or sensitivity kernels), are a key ingredient for seismic full‐waveform inversion with a local‐search optimization algorithm. They provide a quantitative measure of the expected changes in the seismograms due to perturbations of the subsurface model parameters for a given survey geometry. Since 2.5‐D wavefield modeling involves a real point source in a 2‐D geological model with 3D (spherical) wave properties, it yields synthetic data much closer to the actual practical field data than the commonly used 2‐D wave simulation does, which uses an unrealistic line source in which the waves spread cylindrically. Based on our recently developed general 2.5‐D wavefield modeling scheme, we apply the perturbation method to obtain explicit analytic expressions for the derivatives of the displacement tensor for 2.5‐D/2‐D frequency‐domain seismic full‐waveform inversion in general viscoelastic anisotropic media. We then demonstrate the numerical calculations of all these derivatives in two common cases: (i) viscoelastic isotropic and (ii) viscoelastic tilted transversely isotropic (TTI) solids. Examples of the differing sensitivity patterns for the various derivatives are investigated and compared for four different homogeneous models involving 2‐D and 2.5‐D modeling. Also, the numerical results are verified against the analytic solutions for homogeneous models. We further validate the numerical derivatives in a 2‐D heterogeneous viscoelastic TTI case by conducting a synthetic data experiment of frequency‐domain full‐waveform inversion to individually recover the twelve independent model parameters (density, dip angle, five elastic moduli, and five corresponding Q‐factors) in a simple model comprising an anomalous square box target embedded in a uniform background. Another 2.5‐D multi‐target model experiment presenting impacts from four common seismic surveying geometries validates the Fréchet derivatives again.This article is protected by copyright. All rights reserved

show abstract