Introduction to Multiple Attenuation Methods

Introduction to Multiple
Attenuation Methods
GP 210
Basic Earth Imaging
Dec 5th, 2012
Prepared by Mandy Wong
Overview
  Introduction
  Types of multiples
  Multiple removal methods
1.  Predictive Deconvolution
2.  Fk filtering
3.  Hyperbolic Radon filtering
4.  Surface-related multiple elimination (SRME)
  Summary
Introduction
•  Multiples are seismic arrival that have more than one reflections
or scattering
•  There are many types of multiples with special names
(1) Source ghost
(2) Receiver ghost
(3) Mirror signal
• A type of receiver ghost that involves ocean bottom receiver
• It can easily be used as signal
(4) Surface-related multiples
•  event with at least one reflection off the sea-surface
(4) Surface-related multiples
•  event with at least one reflection off the sea-surface
(5) Water-column reverberations
• a class of surface-related multiples
• Reflections only between the sea-surface and the seabed.
(6) Internal multiples
•  event with no reflection off the sea-surface
(7) Peg-leg multiples
•  A multiple reflection involving successive reflection at different
interfaces so that its travel path is not symmetric
Multiple removal methods
1.  Predictive Deconvolution
2.  F-k filtering
3.  Parabolic Radon Transform
4.  Surface-related multiple elimination
(SRME)
Predictive Deconvolution
• Remove short-period multiples (most notably from relatively flat, shallow
water-bottom)
• The periodicity of the multiples is exploited to design a filter that
removes the predictable part of the wavelet (multiples), leaving only its
non-predictable part (signal)
Zero-offset gather
After prediction decon
Predictive Deconvolution
To suppress multiples choose a lag corresponding to the two-way traveltime of the multiples
Zero-offset gather
After prediction decon
Predictive Deconvolution
Pros
Cons
•  Computationally affordable
•  Good for shallow water
reverberation
•  1D model
•  For dipping reflector, multiples are
not periodic
Fk-filtering
Primaries and multiples exhibit different hyperbolic moveout in CMPs
CMP gather from Lab 5
Fk-filtering
Primaries and multiples exhibit different hyperbolic moveout in CMPs
Which one is primary?
And multiple?
CMP gather from Lab 5
Fk-filtering
NMO with the Vrms between the
primaries and the multiples
NMO with the Vrms of the primaries
Fk-filtering
Up- and down-going event can be separated in the f-k domain
Fk-filtering
Pros
Cons
•  Computationally affordable
•  Good for simple subsurface
•  Fail at near offset
•  Insufficient for complex subsurface
•  Require velocity model
Hyperbolic Radon Transform
Radon Transform
Pros
Cons
•  Computationally affordable
•  Good for simple subsurface
• Insufficient for complex subsurface
•  Require velocity model
Surface Related Multiple Elimination (SRME)
X1
X2
Surface Related Multiple Elimination (SRME)
X1
X2
X1*X2 (convolution): predict multiples of path S1-R1-R2
What is Convolution?
Source:http://en.wikipedia.org/wiki/Cross-correlation
Surface Related Multiple Elimination (SRME)
X1
X2
X1*X2
(convolution)
-
Surface Related Multiple Elimination (SRME)
X1
X2
t1
X1*X2
(convolution)
t1
Surface Related Multiple Elimination (SRME)
X1
X2
t2
X1*X2
(convolution)
t2
Surface Related Multiple Elimination (SRME)
X1
X2
+
X1*X2
(convolution)
+
Surface Related Multiple Elimination (SRME)
Surface Related Multiple Elimination (SRME)
To estimate the multiples coming from point A to C, convolve the following
•  common shot gather (shot at A)
•  common receiver gather (receiver at B)
And then sum all the contributions
C
A
Verschuur (2009)
Surface Related Multiple Elimination (SRME)
To estimate the multiples coming from point A to C, convolve the following
•  common shot gather (shot at A)
•  common receiver gather (receiver at B)
And then sum all the contributions
Verschuur (2009)
Surface Related Multiple Elimination (SRME)
To estimate the multiples coming from point A to C, convolve the following
•  common shot gather (shot at A)
•  common receiver gather (receiver at B)
And then sum all the contributions
Verschuur (2009)
Surface Related Multiple Elimination (SRME)
Figure 5:Surface bounce lie outside of acquisition
geometry.
Surface Related Multiple Elimination (SRME)
Pros
Cons
•  Require no subsurface info
•  Can eliminate complex surfacerelated multiples
•  Require dense and regular
acquisition geometry
Summary
Methods Pros Predictive Deconvolution  Computationally  1D model affordable  For dipping reflector, not  Good for shallow water periodic reverberation F‐k Filtering  ‐Computationally affordable  ‐ Good for simple subsurface  Fail at near offset  Insufficient for complex subsurface  Require velocity model Radon Transform  ‐Computationally affordable  ‐ Good for simple subsurface  Insufficient for complex subsurface  Require velocity model Surface Related  Require no subsurface Multiple Elimination info (SRME)  Effectively converging Cons   Require dense and regular acquisition geometry. References
Alvarez G F, Attenuation of Multiples in Image Space, PhD Thesis, Stanford
University, 2007 (Figure 5)
Cao Z.H., Analysis and Application of Radon Transform, MSc Thesis, Univ of
Calgary, 2006 (Figure 1, 4)
Peacock K L and Treitel S, Predictive Deconvolution: Theory and Practice,
Geophysics, 34 (1969) (Figure 3)
Verschuur, D. J., Berkhout, A. J. and Wapenaar, C. P. A., Adaptive surfacerelated multiple elimination, Geophysics, 57, 1166-1177, 1992
Weglein A B, Multiple attenuation: an overview of recent advances and the
road ahead, The Leading Edge, 18 (1999), 40-44