Abstract
Full-waveform inversion (FWI) is an appealing data-fitting approach to produce high-resolution seismic imaging of subsurface geological structures. Implementation of FWI in frequency domain usually requires spectral data of seismograms only at some frequencies, making it more convenient. Compared to the common approximations such as acoustic or viscoacoustic, 3D frequency-domain viscoelastic FWI is quite challenging due to the fact that more additional parameters are needed to be inverted resulting in highly increased computational cost and unwanted crosstalk between quality factors and wave velocities. We derived explicit Fréchet derivatives for sensitivity kernels for 3D frequency-domain viscoelastic FWI and validated it numerically. Then, we performed 3D multiparameter viscoelastic FWI in frequency domain in parallel environments. Our results demonstrate successful inversion of five key parameters-density, p-wave velocity, s-wave velocity, and associated quality factors-thus providing a preliminary validation of the feasibility and efficacy of our approach for multiparameter FWI.
| Original language | British English |
|---|---|
| Pages (from-to) | 1082-1086 |
| Number of pages | 5 |
| Journal | SEG Technical Program Expanded Abstracts |
| Volume | 2024-August |
| DOIs | |
| State | Published - 2024 |
| Event | 4th International Meeting for Applied Geoscience and Energy, IMAGE 2024 - Houston, United States Duration: 26 Aug 2024 → 29 Aug 2024 |
Keywords
- 3D
- frequency-domain
- full-waveform inversion
- parallel
- viscoelastic
Fingerprint
Dive into the research topics of 'Numerical experiments on 3D frequency domain multiparameter full-waveform inversion in viscoelastic media in parallel environments'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver