E

1
EP Challenges for imaging
2
Outline

• EP targets
• Imaging objectives
• Emerging technologies
• State of the art imaging workflows
• Conclusion

3
EP Challenges

• Exploration findings in new - difficult/expensive
  - areas
• Deep / Ultra-deep offshore
• Deep reservoirs (11000m)
• Sub-salt / sub-basalt
• Complex geology (foothills, etc.)
• Onshore reservoirs
• Heavy oil
• Tight gas
• Gas and Oily shales
• Production focus on Improved Oil Recovery (IOR)
  through better prediction
• Onshore (current recovery factor 30)
• Low reservoir quality (weak porosity and/or
  permeability)
• HP-HT reservoirs
• Time-lapse seismic
• Semi-permanent / Permanent / Passive monitoring

4
EP Challenges

• Exploration findings in new - difficult/expensive
  - areas
• Deep / Ultra-deep offshore
• Deep reservoirs (11000m)
• Sub-salt / sub-basalt
• Complex geology (foothills, etc.)
• Onshore reservoirs
• Heavy oil
• Tight gas
• Gas and Oily shales
• Production focus on Improved Oil Recovery (IOR)
  through better prediction
• Onshore (current recovery factor 30)
• Low reservoir quality (weak porosity and/or
  permeability)
• HP-HT reservoirs
• Time-lapse seismic
• Semi-permanent / Permanent / Passive monitoring

5
Complex subsurface features

• Deep
• Sharp velocity contrasts
• Imaging below high velocity bodies (sub-salt,
  sub-basalt)
• Shallow anomalies (dunes, topography, permafrost,
  channels)
• Fault networks

6
Where is the salt base?
?
Arbitrary 2D line through large 3D PSDM volume
Close-Up
7
Imaging issues

• Illumination
• Resolution (sampling, frequency content)
• De-noising (migration, multiples,)
• Focusing (velocity model building)
• Positioning (anisotropy)
• Amplitude preservation (repeatability, seismic
  attributes)

8
Imaging issues

• Land and
• Marine
• acquisition
• geometries

• Illumination
• Resolution (sampling, frequency content)
• De-noising (migration, multiples,)
• Focusing (velocity model building)
• Positioning (anisotropy)
• Amplitude preservation (repeatability, seismic
  attributes)

9
Seismic acquisition needs
Increase the value of seismic for EP
from Veritas

• Improve quality
• Better seismic imaging
• - enhanced resolution (vertical)
• - quantitative reservoir attributes

from BGP

• Minimize cost
• Cost per trace must drop to improve
• survey characteristics for the same price

• Adapt to any environment
• Equipment to support efficient acquisition in
  all conditions

• Reduce turnaround time
• Seismic interpretation in time for drilling

10
Seismic acquisition answers
Increase quality productivity
High channel count crew

• More channels
• Large spread
• with dense RP sampling

• More records
• Increase source productivity,
• SP density

From Sercel
Vibrator array , Sinopec, Saudi Arabia

• More flexibility
• Deploy a seamless spread
• whatever the terrain conditions

From Sercel
11
Wide-azimuth 3D or 3D2
Current 3D geometry
3D2 geometry

• Basic 3D2 parameters for dense acquisition
• 12.512.5 m² bin size, four times more bins than
  current geometries.
• Very high trace density, ten times more than
  current with folds over 1000.
• Isotropic illumination at all offsets

12
Receiver arrays, today

• A 25-m group interval provides adequate signal
  sampling at frequencies up to 100 Hz for 95 of
  reservoirs (gt1000 m).
• Why receiver arrays of limited size?
• Although 25-m arrays do not significantly affect
  signal amplitudes at frequencies up to 100 Hz,
  they are needed to correctly sample surface waves
  and bring ambient noise to acceptable levels.

5 to 25 m
25 m
13
Examples of land 3D2 geometry
20 000-channel crew with 14 400 live 12.5 x 12.5
m² bin size
Deep reservoir
Medium depth reservoir
612 km² 30 lines of 480 channels Dx 25m Dy
200m 2 external shot point lines
66 km² 60 lines of 240 channels Dx 25m Dy
100m Centered shot point lines
14
More records slip sweep principle
A vibrator group sweeps without waiting for the
previous groups sweep to terminate
Continuous recording of composite signal
Correlation by reference signal
Slip time Time
shift between sweeps
Slip time
Time
PDO Vibroseis Technique
15
Effect of correlation on harmonic noise
0
16
Post slip-sweep processing
Reference Slip-sweep HPVA
HPVA technology for estimating the harmonic
content of each vibro-seis VP recorded with the
slip-sweep method, and then subtracting the
predicted noise from the field data.
From CGG
17
Point acquisition principle

• -- No source noise attenuation
• -- No ambient noise attenuation
• - Good coupling mandatory
• Less sensitivity per RP

-- Noise often aliased -- HF filtering (intra
array statics) - Azimuth dependant filtering
Noise well sampled HF content preserved
Isotropic recording
18
Marine systems
From D.Monk Apache
19
Why do we need wide-azimuth?

• Limitations of conventional 3D marine
  acquisition
• Single or, at most, narrow azimuth
• Not symmetrically sampled in the in-line and
  cross-line directions
• Unbalanced receiver and shot densities
• Hence, the wave-field is not fully sampled

As a result, complex structures are not optimally
imaged with conventional 3D
20
What do we want ?

• Isotropic distribution of source-receiver pairs
• Full target illumination in complex areas, among
  a high multiplicity of rays some will carry the
  useful signal
• Proper input to tomography optimal angular
  illumination redundancy
• Improved imaging of multi-azimuth dipping
  structures
• Extraction of rock elastic properties including
  fracture
• Isotropic wave-field sampling in the receiver and
  shot-point domains
• Isotropic (3D) FK/Radon filtering of unwanted
  waves
• 3D Surface Multiple Attenuation
• Wave-equation imaging - 3D pre-stack datuming
• 3D Surface scattered noise attenuation (during
  the stacking or focussing process)

21
Limitations of conventional acquisition

In line / Cross line ratio is 6 to 20
In line 9000m
Cross line 450m
Current algorithms and workflows adapted /
designed for almost single-azimuth acquisitions
22
Conventional perpendicular acquisitions
Courtesy of BP
23
Wide-azimuth streamer acquisitions
New algorithms and workflows are needed to take
full advantage of new acquisition patterns
24
Wide-azimuth means
OPERATIONAL SET-UP
9000m
One 10-streamer vessel, 90 km
900m

One or more dual source vessels
25
Wide-azimuth methodology
2
1
Y 5000 m
Y 4000 m
Y 3000 m
Y 2000 m
Y 1000 m
3
4
18,000 metres
The super shot pattern is repeated in the
cross-line direction to build the total CMP fold
of multiplicity. The fold will depend on the
nominal distance between source lines.
Receiver template
5,000 metres
Super Shot Point Gather
50 Receiver Lines
26
Comparing narrow and wide-azimuth

• Conventional geometry
• Full fold area 1000 sq.km.
• One recording source vessel
• Receiver spread 10 x 9000 m 9 sq.km.
• Fold 80
• Azimuths /- 3 degrees
• Infill level 33 (optimistic)
• Base duration 34 days
• Base cost 1

• Wide-azimuth geometry
• Full fold area 1000 sq.km.
• One recording two source vs.
• Receiver spread 50 x 2 x 9000 m 90 sq.km.
• Fold 40 x 2 x 5 400
• Azimuths /- 90 degrees
• Infill level 0 (assumption)
• Duration 129 days x 4
• Estimated cost x 6 / 6.5

27
More channels for higher resolution
from M. Lansley, PGS Onshore, The Leading Edge,
October 2004
28
VARG base Cretaceous amplitude map
Wide azimuth for higher resolution
from S. Hegna and D. Gaus, PGS Geophysical, EAGE,
2003
29
Wide azimuth for detecting anisotropy

• Anisotropy is detected by wide azimuthal
  distribution and can be compensated for
• From this compensation we get better imaging and
  info about fracture orientation density

North South North
Near Far
Offset sorted CMP
Azimuth sorted CMP
from M. Williams E. Jenner The Leading Edge,
Aug.02
30
Sources with broad signal bandwidth

• Lower frequencies for deeper targets
• Limitations are more on the source side
• Piezoelectric hydrophones in marine down to 0.5Hz
  (wave height measure helps de-ghosting)
• MEMS based digital accelerometers in land (linear
  response on 0-800Hz)
• Land sources
• Longer explosions improves conversion into
  elastic wave and broadens bandwidth with pick
  amplitude shifted towards lower frequencies
• Vibrators with bigger mass broadens bandwidth
  broadened at both extremities (5-250Hz) and
  allows point source
• Marine sources
• Air guns bandwidth driven by bubble period and
  depth (ghost). low frequency (5Hz)

31
New acquisition geometries

• More channels and shots (higher fold, dense and
  regular sampling wave-field sampling)
• Wide azimuth (isotropic wave-field sampling)
• Broader signal bandwidth and lower frequencies
  (resolution)

near future is 2D
32
Imaging issues

• Illumination
• Resolution (sampling, frequency content)
• De-noising (migration, multiples,)
• Focusing (velocity model building)
• Positioning (anisotropy)
• Amplitude preservation (repeatability, seismic
  attributes)

33
Wide-azimuth seismic processing

• Move away from offset cube pre-processing and
  azimuth sector processing
• One-pass solutions preferably on shot gathers
• De-noising and preserving low frequencies
• De-aliasing
• Surface-related multiple attenuation
• Internal multiple attenuation

34
Focusing state of the art near future
Velocity model building is a three stage
iterative process
2nd generation tomography Kirchhoff
migration Automatic RMO picking, Dense xn, yn,
zn, cnii Ray-based tomography
WE based method One-way wave-equation migration
of p-gathers Mean semblance maximization. No
picking, P-gather file Wave-equation
1st generation tomography Kirchhoff
migration Manual RMO picking, Sparse xn, yn,
zn, cn Ray-based tomography
35
Simple view of underdetermined problem
narrow azimuth angular redundancy (q 45)
36
Simple view of underdetermined problem
wide azimuth angular redundancy
37
Simple view of underdetermined problem
Medical tomography angular redundancy
38
Simple view of underdetermined problem
Building a linear equation
39
Simple view of underdetermined problem
Building a linear equation
40
Simple view of underdetermined problem
Medical tomography
41
Ray-based VMB GoM example
Example of workflow designed for Gulf of Mexico
42
Ray-based VMB observable quantities
CIP Gather RMO
For each locally coherent event
43
Ray-based VMB Data Conditioning
44
Ray-based VMB volumetric skeleton
45
Ray-based VMB automated RMO pick.
46
Ray-based VMB RMO pick filtering
47
Ray-based VMB volumetric dip picking
48
Ray-based VMB RMO High-Grading
Local semblance (RMO)
Structural semblance
49
Ray-based VMB VelTracer 3D tomo

• 3D Finite Offset Tomographic Inversion
• True 3D Depth Tomography using ray tracing
• All of the RMO information used (full offset)
• Non-linear Inversion
• TTI capable
• 64-bit multi-CPU Linux implementation

1 Pass !
50
Ray-based VMB VelTracer 3D tomo
VelTracer Internal Workflow
invariants TIMESR GRADTSR OFFSETSR AZIMUTHSR
initial PreSDM
RMO picking (OFFSET, RMO, DIP)
kinematic de-migration
invariants TIMESR GRADTSR OFFSETSR AZIMUTHSR
51
Ray-based VMB VelTracer kinematics

• RMO Pick Z(h)
• Dip

Demigration to surface consistent invariants
52
Ray-based VMB VelTracer criterion
RMO corresponds to spatial misalignment of image
facets for a local reflector
53
Ray-based VMB Inversion QC

• QC
• Predicted RMO attribute
• Gamma attribute
• Velocity
• CIP Gather flatness actual RMO
• Pre-SDM Image

54
Ray-based VMB RMO and gamma
GAMMA Initial Model
RMO Initial Model
55
Ray-based VMB CIP Gather Flatness
Pre-SDM Initial Model
56
Ray-based VMB sediment velocity
Velocity Initial Model
Pre-SDM Initial Model
salt mask
57
Ray-based VMB Velocity PreSDM
58
Marmousi using one-way WE migration

• Synthetic dataset recomputed using one-way
  wave-equation
• Realistic geometry and signal bandwidth
• 240 shots
• 96 receivers per shot, streamer geometry, max
  offset 2.4 km
• Frequency range fmin 5 Hz, fmax 40 Hz
• Initial model 1D, linear gradient 1500 m/s ?
  3000 m/s

59
Marmousi using one-way WE migration
initial velocity model
60
Marmousi using one-way WE migration
final velocity model
61
Marmousi using one-way WE migration
exact velocity model
62
Marmousi using one-way WE migration
final velocity model
63
Marmousi using one-way WE migration
initial migration
64
Marmousi using one-way WE migration
final migration
65
Marmousi using one-way WE migration
exact migration
66
Marmousi using one-way WE migration
initial CIP p-gathers
67
Marmousi using one-way WE migration
final CIP p-gathers
68
Marmousi using one-way WE migration
exact CIP p-gathers
69
Marmousi using waveform Inversion
70
Marmousi using waveform Inversion
71
Marmousi using waveform Inversion
72
Marmousi using waveform Inversion
73
BP model using waveform Inversion
from G.Pratt
exact model
74
Imaging issues

• Illumination
• Resolution (sampling, frequency content)
• De-noising (migration, multiples,)
• Focusing (velocity model building)
• Positioning (anisotropy)
• Amplitude preservation (repeatability, seismic
  attributes)

75
TTI velocity model building
76
Fast salt body delineation
77
Fast salt body delineation
78
Fast salt body delineation
79
Imaging issues

• Illumination
• Resolution (sampling, frequency content)
• De-noising (migration, multiples,)
• Focusing (velocity model building)
• Positioning (anisotropy)
• Amplitude preservation (repeatability, seismic
  attributes)

80
Time-lapse in depth
4D difference (PreSDM)
Structure (PreSDM)
from Norsk Hydro and CGG
81
Conclusion

• Incremental improvement of imaging through
• Denser acquisition geometry
• Wide azimuth geometry
• Broader signal bandwidth
• One-way WE-based velocity model building
• 3D Global offsets, Full waveform inversion

82
Acknowledgment

• Many thanks to Denis Mougenot, Volker Dirks,
  Gilles Lambaré, Robert Soubaras, Bruno Gratacos,
  Serge Zimine, Jean-Jacques Postel, Michel Manin
  and Robert Taylor who helped me prepare this
  presentation
• CGG for giving permission to show data
• WesternGeco, Veritas DGC, PGS, Sercel, Liaohe
  Geophysical and BashNeftGeofisika for used images
  from WEB-site / litterature
• Dave Monk (Apache Corp.), Williams E. Jenner
  for use of published images
• BP for data used by EAGE 2005 depth imaging
  benchmark