|
|
|
|||||||||||||||||
| JOURNAL HOME | HELP | CONTACT PUBLISHER | SUBSCRIBE | ARCHIVE | SEARCH | TABLE OF CONTENTS |
NORSAR, Kjeller, Norway
Sintef Petroleum Research, Trondheim, Norway
Department of Earth Science, Bergen, Norway
Corresponding author: isabelle.lecomte@norsar.com
| The first 20% of the full text of this article appears below. |
Model-based analysis of seismic data, now recognized as a key to a better understanding of images of subsurface structure, began in the 1960s. But these models were inflexible, primarily due to hardware and software limitations, and generally were simple homogeneous isotropic horizontal layersi.e., with only vertical velocity variations.
Various corrections to these models were quickly introduced to compensate for layer dips. However, many corrections were a posteriori, at various stages of the processing, without consulting the original model. Such approaches may be dangerous because we may lose control of the physics of the problem and end up with data about which our knowledge is limited: Do they display real structures or imaging artifacts? Are the amplitudes valid? These questions become harder to answer with successive generations of models. Earth models should be realistic from the beginningi.e., ideally as 3D heterogeneous, anisotropic, and anelastic structures! Modeling is then necessary to better understand wave propagation and get realistic data to help develop new processing techniques. But modeling in such realistic, and therefore complex models requires appropriate tools.
Two techniques dominate modeling in 2D and 3D: ray-tracing (RT) and finite-difference (FD). Both solve the same problem, i.e., (an)elastic wave propagation in the earth, but they solve it differently. The differences in approach are quite significant and, as a result, each has advantages and drawbacks. We therefore hope to demonstrate in this paper that it may be more effective to combine these approaches for seismic modeling of reservoirs and that we may significantly improve our results by using the two different methods in an alternating fashion.
| Review of ray tracing and finite-difference modeling |
|---|
| JOURNAL HOME | HELP | CONTACT PUBLISHER | SUBSCRIBE | ARCHIVE | SEARCH | TABLE OF CONTENTS |