Sunteți pe pagina 1din 48

Seismic Inversion and AVO applied

to Lithologic Prediction

Part 4 Shear Wave Analysis and
Inversion
4-2
Introduction
In our previous section on rock physics we
discussed fluid effects on P and S-wave velocity, and
density.
We then looked at post-stack inversion applied to P-
wave data.
In this section, we will look at various options for
acquiring, analyzing, and inverting S-wave data.
We will start by analyzing the models that were
created in the first section.
We will then look at the analysis of full S-wave data.
Finally, we will discuss converted wave, or PS wave
data.
4-3
(a) Wet model (b) Gas model
Recall that, in the rock physics section, we analyzed the two models
shown above. Model A consists of a wet sand, and Model B consists of
a gas-saturated sand. Specifically, we wanted to look at the effects of
the gas on the density, P-wave velocity, and S-wave velocity of the
saturated sand.
Our Two Models
4-4
(a) P-wave motion (b) S-wave motion
Since the direction of particle motion for a P-wave is in the same direction as
its wave movement, it will be more affected by a gas sand than the S-wave,
since the direction of particle motion for the S-wave is at right angles to the
direction of its wave movement.
P- and S-waves
4-5
Model Values
This was indeed found to be the case when we computed
the wet and gas cases using the Biot-Gassmann equations
in Part 1 of the course. The values were as follows, where
typical values for a shale have also been added.

Wet: V
P
= 2500 m/s, V
S
= 1250 m/s, = 2.11 g/cc, o = 0.33
Vp/Vs = 2.0
Gas: V
P
= 2000 m/s, V
S
= 1310 m/s, = 1.95 g/cc, o = 0.13
Vp/Vs = 1.53
Shale: V
P
= 2250 m/s, V
S
= 1125 m/s, = 2.0 g/cc, o = 0.33
Vp/Vs = 2.0
Notice that the P-wave velocity drops dramatically in the
gas sand, when compared to the wet sand, but the S-wave
velocity actually goes up.
4-6
V
P
V
S

depth
Surface
Shale
Shale
Gas Sand
Seismic
Raypath
As shown above, the seismic raypath is dependent on three
physical parameters: density (), P-wave velocity (V
P
), and S-wave
velocity (V
S
), which were discussed in the rock physics section.
The Vertical Incidence Seismic Raypath
4-7
Exercise 4-1 Traveltimes
On the previous slide, the vertical units were in depth. If they had been in
time, the arrival times for P and S waves would have been different. In
fact, as we will shortly see, there are three different traveltimes that we
can record: t
PP
, or P-wave down and P-wave up; t
SS
, or S-wave down
and S-wave up; and t
PS
, or P-wave down and S-wave up (this is called
the converted wave). Assuming that the gas sand in the previous slide is
at a depth of 2000 m and has a thickness of 20 m, and using the
velocities on the slide before the previous one, work out the following
traveltimes:

To base of shale: t
PP1
= To base of sand: t
PP2
=
t
PS1
= t
PS2
=
t
SS1
= t
SS2
=

Isochron: At
PP
= t
PP2
- t
PP1
=
At
PS
= t
PS2
- t
PS1
=
At
SS
= t
SS2
- t
SS1
=
1-8
The reflection coefficient
If the ray paths in the previous slide were at normal
incidence (i.e. vertical) the reflection coefficients for the P
and S-waves are as follows:
.
2
,
2
V V
V ,
2
V V
V
, , V V V , V V V
: where
,
V
V
2
1
V V
V V
R
,
V
V
2
1
V V
V V
R
1 2 1 S 2 S
S
1 P 2 P
P
1 2 1 S 2 S S 1 P 2 P P
S
S
1 S 1 2 S 2
1 S 1 2 S 2
0 S
P
P
1 P 1 2 P 2
1 P 1 2 P 2
0 P

A A A

A A

A A


+
=
+
=
+
=
= = =
(

+ ~
+

=
(

+ ~
+

=
1-9
Exercise 4-2
Top Shale:
V
P1
= 2250 m/s
V
S1
= 1125 m/s

1
= 2.0 g/cc
Wet Sand:
V
P2
= 2500 m/s
V
S2
= 1250 m/s

2
= 2.11 g/cc
Base Shale:
V
P3
= 2250 m/s
V
S3
= 1125 m/s

3
= 2.0 g/cc
V
P
V
S
AV
P
AV
S
A
R
P0
R
S0

A
P
P
V
V A
S
S
V
V A
Compute the parameters for the wet sand interfaces using the approximate
formulae for the reflection coefficients:
1-10
Exercise 4-3
Top Shale:
V
P1
= 2250 m/s
V
S1
= 1125 m/s

1
= 2.0 g/cc
Gas sand:
V
P2
= 2000 m/s
V
S2
= 1300 m/s

2
= 1.94 g/cc
Base Shale:
V
P3
= 2250 m/s
V
S3
= 1125 m/s

3
= 2.0 g/cc
V
P
V
S
AV
P
AV
S
A
R
P0
R
S0

A
P
P
V
V A
S
S
V
V A
Compute the parameters for the gas sand interfaces using the approximate
formulae for the reflection coefficients:
4-11
Model Values
We also found in an exercise that the P and S-impedances for the
three cases were:

Z
Pgas
= 3900 m/s*g/cc Z
Sgas
= 2555 m/s*g/cc

Z
Pwet
= 5275 m/s*g/cc Z
Swet
= 2638 m/s*g/cc

Z
Pshale
= 4500 m/s*g/cc Z
Sshale
= 2250 m/s*g/cc

Using the above values, the P and S reflection coefficients for the gas
and wet cases, where the shale overlies the sand, are:

R
Pgas
= -0.071 R
Sgas
= 0.063

R
Pwet
= 0.079 R
Swet
= 0.079

An interesting thing to note about the reflection coefficients is that the
gas and wet cases for the P-waves show opposite polarity, whereas
the gas and wet cases for the S-waves show the same polarity.
4-12
The four figures on the next two slides show synthetic
zero-offset models of the four cases we have
considered: the P and S-wave responses of both the wet
case (Model A) and the gas case (Model B). (Note that
the parameter values have been changed slightly)

We have used a 25 Hz Ricker wavelet as the seismic
wavelet, and that this wavelet has a wavelength that is
less than the time thickness of the sand. Thus, we are
seeing tuning of the top and base responses.

The key thing to note is that the P-wave response
changes polarity in going from a wet to a gas sand, but
the S-wave response remains the same polarity.
Synthetic Models
4-13
(a) P-wave log, density and synthetic from model A
(b) S-wave log, density and synthetic from model A (note the different traveltimes).
4-14
(a) P-wave log, density and synthetic from model B
(b) S-wave log, density and synthetic from model B
4-15
The above diagram shows a schematic diagram of (a) P, or compressional,
waves, (b) SH, or horizontal shear-waves, and (c) SV, or vertical shear-waves,
where the S-waves have been generated using a shear wave source. This
recording approach, using multi-component geophones, was used over a gas
sand in Alberta to look for the presence of a gas sand. (Ensley, 1984)
P- and S-wave recording
(a)
(b) (c)
4-16
The above diagram shows recorded and processed seismic sections of (a) P, or
compressional, waves, and (b) SH, or horizontal shear-waves, over the Myrnham
gas field in Alberta. As predicted by the theory, the P-waves respond to the gas
sand whereas the S-waves do not, allowing us to predict the presence of the
gas. Note the different time arrivals in the two sections. The arrows indicate the
same events and the ellipses outline the anomaly. (Ensley, 1984)
P and SH-waves Gas Sand Example
(a) (b)
4-17
P and SH-waves Coal Example
(a) (b)
The above diagram shows recorded and processed seismic sections of (a) P, or
compressional, waves, and (b) SH, or horizontal shear-waves, over a false
bright spot due to a coal near the gas field in the previous slide. Note that the
P-waves and the S-waves both respond to the coal, allowing us to predict that
the bright-spot is not due to the presence of gas. Again, the arrows show
equivalent events, and the ellipses show the zone of interest. (Ensley, 1984)
4-18
Converted S-waves
The previous example used full S-wave recording, in
which S-waves were generated at the surface of the
earth using an S-wave vibrator, and the reflections
were recorded using multi-component geophones.
However, there is a simpler, and cheaper, way to
record S-wave information, as shown in the next slide.
If we use a P-wave source, and record the data at
different offsets using multi-component geophones,
we can record converted S-waves, and reflected P-
waves which contain some influence from the S-
waves.
4-19
Reflected
P-wave = R
P

Reflected
S-wave
Transmitted
P-wave
Incident
P-wave
Transmitted
S-wave

Mode Conversion of an Incident P-wave
V
P1
, V
S1
,
1

V
P2
, V
S2
,
2
Consider the interface between two geologic horizons of differing P
and S-wave velocity and density and an incident P-wave at angle u
i
.
This will produce both P and S reflected and transmitted waves, as
shown above. These are SV waves in the in-line direction.
u
i

|
r

u
r

u
t

|
t

4-20
Utilizing mode conversion
But how do we utilize mode conversion?
There are actually two ways:
Record the converted S-waves using multi-component
receivers (in the X and Z direction).
Interpret the amplitudes of the P-waves as a function of offset,
or angle, which contain implied information about the S-waves.
This is called the AVO (Amplitude versus Offset) method, and
will be discussed in subsequent parts of the course.
When we record the converted waves, we need to be
very careful in their processing and interpretation, as
will be shown next.
In the AVO method, we can make use of the Zoeppritz
equations, to extract pseudo S-wave information from
P-wave reflections at different offsets.
4-21
Converted wave analysis
Before looking at a converted wave interpretation, we
will discuss the steps involved in converted wave
analysis, using a dataset from Alberta.
The most difficult part of converted wave interpretation
is in interpreting events on the PP and PS sections that
come from the same geological horizon but have
different arrival times and amplitudes.
As we will see, there are two ways to correct for these
problems:
(1) Use the well log velocities and perform modeling at the
wells.
(2) Use seismic pick analysis.
4-22
Initial multi-component display
(a) (b)
Let us consider the data shown above, where (a) shows PP data and
(b) shows PS data. Although this data is over the same part of the
subsurface, it is hard to correlate between the two sections due to time
and amplitude differences.
4-23
Converted display assuming Vp/Vs =2
(a) (b)
This slide again shows (a) PP data and (b) PS data. However, now
the PS data has been converted to PP time assuming that the Vp/Vs
ratio is equal to two. The fit is better, but still not very good.
4-24
P wave log correlation
We have now correlated the P-wave log at the log intersection on the
PP data. Notice the good tie on the right, where the blue trace is the
synthetic, and the red trace is the seismic trace.
4-25
PS log correlation
We have now correlated the P and S-wave logs at the log intersection
on the PS data. Again, notice the good tie on the right, where the blue
trace is the synthetic, and the red trace is the seismic trace.
4-26
PP and PS extracted wavelets
(a) (b)
(c) (d)
The wavelets on the
previous synthetics were
extracted from the
seismic data and are
shown on the left, where
(a) shows the wavelet
extracted from the PP
section, (b) shows the
amplitude spectrum of
the PP wavelet, (c)
shows the wavelet
extracted from the PS
section, and (d) shows
the amplitude spectrum
of the PS wavelet.
Notice the difference in
frequency content.
4-27
Synthetic to seismic correlation
The display above shows the offset synthetics computed from the well
logs and using the wavelets shown in the previous slide. We will be
discussing offset synthetics in the next section, but for now simply
notice that the PS-wave synthetic has zero amplitude at zero offset.
PS-wave offset synthetic PP-wave offset synthetic
4-28
Seismic tie assuming that Vp/Vs = 2.0
(a) (b)
This slide again shows (a) PP data and (b) PS data, converted to PP
time assuming that the Vp/Vs ratio is equal to 2. We have spliced in
the synthetics using the correct velocities. Notice the misfit.
4-29
P - PS seismic and synthetic ties with well
log derived velocities
(a) (b)
This slide again shows (a) PP data and (b) PS data. However, now the PS
data has been converted to PP time using the Vp/Vs ratio from the logs.
The fit is very good at the wells but the sections dont match laterally.
4-30
PP and PS horizon picks
This slide shows a more extensive section of the (a) PP data and (b) PS
data. To correct for laterally varying velocities, we have picked the major
events on both sections, using the picks from the logs.
(a)
(b)
4-31
Horizon matching
This slide again shows the (a) PP data and (b) PS data. Now, the horizons
have been matched by computing a laterally varying Vp/Vs ratio.
(a)
(b)
4-32
Vp/Vs ratio from horizon match
This slide shows the laterally varying Vp/Vs ratio that was computed using
the horizon picks in the previous slide.
4-33
Vp/Vs Ratio maps
By applying this technique to all of the lines in the 3D volume, a map of
Vp/Vs ratios can be computed. The maps above show the change in
Vp/Vs ratio between different pairs of events shown in the previous slides.
4-34
Converted-wave case study
Let us now see how the previous analysis can be
applied in a field example.
For our case study, we will go back to the
Blackfoot example considered in the last part of
the course.
Recall that this case study involved the
delineation of a Lower Cretaceous channel sand
system.
We will start by re-displaying several of the slides
from the previous section, including the PP
section.
We will then look at the PS converted wave data
to see what can be added to the interpretation.
4-35
Blackfoot case study
A repeat from the previous section of the schematic stratigraphy of
the Blackfoot area, showing three different incised valleys. The
relative age is also indicated, where 30 is oldest and 40 is youngest.
(Dufour et al.)
4-36
Blackfoot case study
Another look at the index map from the previous section showing
seismic cross-line 95, and two east-west cross-sections. The
wells are also indicated.
4-37
Blackfoot case study
A repeat from the last section of seismic cross-line 95 from the PP
data, showing a clear indication of the three valleys. (Dufour et al.)
4-38
Blackfoot case study
Seismic cross-line 95 from the PS data. Note that resolution is not as
good as the PP data and shows only a single valley. (Dufour et al.)
4-39
Blackfoot case study
A comparison of the (a) PP data, and (b) PS data from line 95. The lack
of resolution in the PS data is now clear. (Dufour et al.)
(a) (b)
4-40
Blackfoot case study
In this case study, seismic amplitude inversion was
not performed on the PS data.

Instead, the authors extracted information about the
V
P
/V
S
ratio using the seismic time picks, which can be
thought of as a type of inversion. The formula used
was:
V
P
/V
S
= 2(At
PS
/At
PP
) 1, where At
PS
is the
PS isochron and At
PP
is the PP isochron.

From our earlier discussion of P and S-waves, we
know to expect that the V
P
/V
S
ratio should go down
when we encounter a gas sand, since V
P
goes down
but V
S
goes up slightly.
4-41
Exercise 4-4 Vp/Vs ratio
Using the isochrons computed in exercise 4-1, and the
formula on the previous slide, compute the Vp/Vs ratio
for the gas sand example of slide 4-5, and show that this
method gives an accurate answer.
4-42
Blackfoot case study
Extracted amplitude slices from the (a) PP data, extracted from the
upper valley (40), and (b) PS data, extracted from the Glauconitic
channel. The white outlines shown the outline of the valley and the
anomalous amplitudes are defined by the red outlines. (Dufour et al.)
(a) (b)
4-43
Blackfoot case study
Computed V
P
/V
S
ratio slices the (a) Mannville-Wabamun interval, and (b)
top of Glauconitic-incised valley-Wabamun interval. The white outlines
shown the outline of the valley. Notice the good match of the
anomalously low V
P
/V
S
ratios to the productive wells. (Dufour et al.)
(a) (b)
4-44
Conclusions
In this section, we have discussed the use of
recorded shear wave sections for the computation of
reservoir parameter change.
Our first example showed how we could differentiate
a gas sand bright-spot from a coal bright-spot
using SH wave generation and multi-component
recording.
We then discussed the use of converted wave data,
where the PS conversion (which is an SV wave) is
recorded using multi-component geophones.
We showed how to integrate the PP and PS
recorded section to produce a Vp/Vs estimate and
then showed a case study in which this technique
was used to explore for channel sands.
4-45
Exercise 4-1 Answers

To base of shale: t
PP1
= 1778 ms To base of sand: t
PP2
=1798 ms
t
PS1
= 2667 ms t
PS2
= 2692 ms
t
SS1
= 3556 ms t
PS2
= 3586 ms

Isochron: At
PP
= t
PP2
- t
PP1
= 20 ms
At
PS
= t
PS2
- t
PS1
= 25 ms
At
SS
= t
SS2
- t
SS1
= 31 ms
4-46
Exercise 4-2 Answers
Top Shale:
V
P1
= 2250 m/s
V
S1
= 1125 m/s

1
= 2.0 g/cc
Wet Sand:
V
P2
= 2500 m/s
V
S2
= 1250 m/s

2
= 2.11 g/cc
Base Shale:
V
P3
= 2250 m/s
V
S3
= 1125 m/s

3
= 2.0 g/cc
V
P
V
S
AV
P
AV
S
A
R
P0
R
S0

A
P
P
V
V A
S
S
V
V A
2375 1187.5
2.06 250 125 0.11 0.105 0.105 0.05 0.079 0.079
2375 1187.5
2.06 - 250 - 125 - 0.11 -.105 -.105 -.05 -.079 -.079
In the following table, we have computed the parameters for the wet sand
interfaces using the approximate formulae for the reflection coefficients:
Question:
Why do you think R
P0
and R
S0
are identical?
4-47
Exercise 4-3 Answers
Top Shale:
V
P1
= 2250 m/s
V
S1
= 1125 m/s

1
= 2.0 g/cc
Gas sand:
V
P2
= 2000 m/s
V
S2
= 1300 m/s

2
= 1.94 g/cc
Base Shale:
V
P3
= 2250 m/s
V
S3
= 1125 m/s

3
= 2.0 g/cc
V
P
V
S
AV
P
AV
S
A
R
P0
R
S0

A
P
P
V
V A
S
S
V
V A
2125 1212.5
1.97 - 250 175 -0.06 -.118 0.144 -.03 -.074 0.057
2125 1212.5
1.97 250 - 175 0.06 0.118 -.144 0.03 0.074 -.057
In the following table, we have computed the parameters for the gas sand
interfaces using the approximate formulae for the reflection coefficients:
Questions:
(1) Why are R
P0
and R
S0
different now?
(2) How can the polarity of the two reflection
coefficients help us identify the gas sand?
4-48
Exercise 4-4 Answer
Recall that V
P
/V
S
= 2000/1310 =1.53. Also, recall that At
PS
= 25 ms and At
PP
= 20 ms.

V
P
/V
S
= 2(At
PS
/At
PP
) 1 = 2(25/20) -1 = 1.5

Note that the fact that we computed a value of 1.5 rather
than 1.53 is due to the fact that we rounded-off the
traveltimes to the closest millisecond. If we had used
more accuracy, the velocity ratio would have been
computed as 1.53

S-ar putea să vă placă și