SEG 3-D Elastic Salt Model

1
SEG 3-D Elastic Salt Model

• Biondo Biondi, Bee Bednar, Arthur Chang leading
  SEG committee work
• John Anderson is XOM contact
• Meeting to discuss desirable features
• Thursday, June 23, 2005, GW3-933
• 900 am to 1030 am
• Computational effort will take a long time
• Current task is to define a suitable model

2
3-D SEG Acoustic Salt Model
snap shot at 781 ms
source
snap shot at 1450 ms
3
Acoustic (18 Hz peak)SEG Salt model (2 data sets)

• 45 shot data set
• 5 lines
• 9 shots/line
• 201 by 201 receiver grid per shot (40 m spacing,
  wide azimuth, source at center of grid, 40401
  receivers/shot)
• ideal for shot record migration

• 4800 shot data(C3-NA)
• 50 lines, 160 m cross-line spacing
• 96 shots / line, 80 m shot spacing
• 8 cables, 40 m group interval within a cable
• 68 receivers / cable
• 544 receivers / shot
• simulates marine acquisition

4
Single raw 3-D shot record (8 streamers, C3-NA)
Numerical dispersion
Original SEG SALT model data
Each streamer has 68 receivers 40 m apart. The
entire survey is 50 lines with 96 shots per line.
Time sample rate is 0.008 s, with 625 samples
per trace
5
Computational Size of Problem

• For 3-D acoustic SEG Salt model
• Using model parameters for data without
  dispersion
• 441 shot wide-azimuth data set
• 21 lines of 21 shots each, receivers at every
  grid point
• 650 GB if SEGY, 400 CPU days on old hardware
• Aimed at ideal conditions for shot migration
• For 3-D elastic model (similar to acoustic model)
• scale compute time by roughly 144
• Vs 0.5 Vp requires finer grid by factor of 2
• Computation time scale factor is 1624
• 3 components x 3 terms factor of 9
• Data set volume grows by factor of 3
• Doubling bandwidth requires factor of 1624

6
Marmousi II 2-D Elastic Model (University of
Houston)
Vp
Low p-wave velocity simulating hydrocarbons to
give AVO response
Flat spot on target
Vs
AVO modeling requires Vp, Vs, and Density
Density
Synthetic data have 80 Hz bandwidth
7
SEG 3-D Elastic Salt Model

• Key desirable features
• (1) smooth and rugose (both deep notches and
  chirp signal) Top of Salt (TOS) components
• (2) shallow salt with impedance match to give
  large P-S conversions
• (3) deeper salt matched for P-P
• (4) multiple salt bodies, one obscuring portions
  of the other
• (5) compaction model for subsalt region honoring
  differences between salt and sediment overburdens

8
SEG 3-D Elastic Salt Model

• Key desirable features
• (6) overpressure zone for part of subsalt zone
• (7) sediment profile with some AVO target
  anomalies
• (8) reservoir zones with compartmentalization
• (9) variations in the Base Of Salt (BOS) (steep
  ramp to flat, gentle ramp to flat)
• (10) subsalt sediments that truncate steeply
  against the BOS

9
SEG 3-D Elastic Salt Model

• Key desirable features
• (11) realistic salt tectonics including faulting
  and structures in the sediments corresponding to
  deep salt withdrawal and slip interfaces
• (12) deep carbonate with rift faults near bottom
  of section.
• (13) components for calibrating image quality
• deep horizontal reflector at bottom
• isolated point diffractors

10
Sediment structures are related to salt tectonics
Allochthonous Salt
Autochthonous Salt
11
(No Transcript)
12
Salt
Salt
Autochthonous salt weld
13
(No Transcript)
14
Realistic Plan

• Begin with a fully elastic model, progress as
  compute capacity grows
• Acoustic model data set
• Elastic multicomponent data set
• Anisotropic multicomponent data set
• Begin collecting data over subsets of the model
• Subsets of the model could target different
  geologic objectives
• Over time merge surveys to cover entire model
• Surface, OBC, and VSP data
• Both absorbing and reflecting surface boundary
  conditions
• Zones of very dense sampling
• Could we have an equivalent physical model done
  at Delft or University of Houston or elsewhere?
• Can we get elastic physical models?