Tech Lunch: Full Bandwidth FWI - Mar 9th
Online Only, you must pre-register to receive access instructions.
NOTE: You Must Be Logged In to Register.
Speaker: Tatiana Kalinicheva, Fullwave at Imperial College London
Co-Authors: M. Warner and F. Mancini, Imperial College London
FWI is widely used to generate high-resolution models of sub-surface physical properties, especially p-wave velocity. Although FWI-generated velocity models have a variety of possible applications, they are most often used as part of a wider velocity-building workflow for subsequent pre-stack depth migration using Kirchhoff or wave-equation based migration. However, in the early development of FWI, a key strategic objective was also to replace conventional seismic processing, velocity-model building and migration entirely by running FWI on raw data to full seismic bandwidth. In the event, the full capabilities of FWI have not yet been exploited in the way that its early champions envisioned. In this paper, we demonstrate the practicality of their vision applied to deep-water marine field data. In principle, with enough compute power and with a sufficiently accurate inversion algorithm, there should be no need to process seismic data. Instead it should be possible to create a model of the subsurface that predicts the unprocessed raw seismic data to within the noise level, and that consequently acts as a direct replacement for a conventional PSDM volume. In this paper, we demonstrate exactly that outcome, applying FWI at 100 Hz directly to unprocessed field data from a towed-streamer marine dataset. We differentiate the resultant FWI velocity model vertically in space and compare the result of this directly with the results of conventional PSDM. Our full-bandwidth FWI result is broader band than the PSDM section at both high and low frequencies and appears to be better resolved vertically. When run on the cloud, this approach is capable of generating a full high-quality PSDM volume for narrow-azimuth towed-streamer data within a few days of data acquisition.
Speaker Biography: Tatiana Kalinicheva, Fullwave at Imperial College London
Tatiana Kalinicheva is a PhD graduate from Fullwave research group at Imperial College London. Her PhD was focused on application of FWI to “difficult” datasets, where inversion seemed to produce the wrong answer or failed completely. Before that she has completed a master's degree in Geophysics at Imperial and a bachelor degree in Geoscience at Moscow State University. She has a keen interest in depth model building and has previously given a talk about 2D vs 3D FWI at SEG 2017.
**Access information will be distributed after registration closes.
Price List:
|
Pre-Registered |
Member
|
Free |
| Non-Member |
$10 |
** Please be aware that ALL registrations must go through 'Checkout', even if your total is zero. Registration is not confirmed until you complete the checkout process.