A quantitative study of the seismic time-amplitude ... · A quantitative study of the seismic...
Transcript of A quantitative study of the seismic time-amplitude ... · A quantitative study of the seismic...
GeoL. Soc. MaLaYJia, BuLLetin 56, December 1994; pp. 95-104
A quantitative study of the seismic time-amplitude reflection characteristics in an on field
NG TONG SANI, IDRUS MOHD SHUHUD2 AND LEONG LAP SAU3
1PETRONAS Carigali Sdn. Bhd. Wisma Peladang, Jalan Bukit Bintang
P.O. Box 12407 50776 Kuala Lumpur
2PETRONAS Research & Scientific Service Sdn. Bhd. Lot 1026, PKNS Industrial Estate
54200 Hulu Kelang 3School of Physics
Universiti Sains Malaysia 11800 Penang
Abstract: The seismic time-amplitude reflection characteristics of selected sandstone horizons in a recently developed oil field are examined for effects of thicknesses, continuity and bed quality. This study uses an integrated approach of well data calibration, forward seismic modelling and 3D seismic data set for interpretation. In this area, wireline logs indicate velocity to be a poor lithologic descriptor. The acoustic impedance at sand-shale interfaces could be accounted for by changes in the density instead. Gassmann's equation confirms the minor effect of velocity perturbation with gas. Forward amplitude modelling lD for sandstone encased in shale in the selected stratigraphic horizons permits values of tuning thicknesses to be ascertained for each lithologic units. This learning phase quantities subsequent reflection parameters and aids 3D seismic interpretation. Preliminary results suggest an east-west trending sandstone reservoir with thicker and better developed sandstone horizons towards the flanks of the anticlinal structure.
INTRODUCTION
The study area is located in Block PM-6, offshore Terengganu which is about 130 kms from the Terengganu coast (Fig. 1), Thirteen (13) exploration wells have been drilled in this oil field.
The objectives of this study are (i) to measure the changes in the waveshape of seismic reflection data associated with the reservoir and (ii) to interpret this trace-to-trace change with changes in their thicknesses and porosity extent.
An integrated approach of using well data, forward seismic modeling, and, a 3D seismic data set, are used in this study. A sandstone horizon have been selected to study its corresponding timeamplitude characteristics for effects of thicknesses and reservoir quality.
This project is ajoint research study by Petronas Carigali Sdn. Bhd., Petronas Petroleum Research Institute and the University Sains Malaysia.
WELL DATA
There are three (3) main lithologies identified
Pre.tenteu at GSM Petroleum Geology Seminar '93
namely coal, shale and sandstone. The sandstone is further classified into two different facies, namely flaser bedding sandstone (sf) and wavy bedding sandstone (sw) together with three (3) fluid types; gas, oil and water. The sonic and density data for the three lithologies, from all exploration wells, were extensively plotted to demonstrate differentiation of the various lithologies and to determine their trend and effects on the acoustic impedance. Two common geologic parameters, porosity and presence of gas in the sand have also been examined to determine their influence on the seismic amplitudes measured.
Velocity trend Velocity data are found scattered over a wide
range of values as shown in Figure 2. Coal has the lowest velocity range, between 2,000 m/s to 2,350 mis, but sand and shale have higher velocities which overlap with each other. This indicates velocity as a weak parameter for differentiating various lithologices, except for coal in this instance. There is only a minor lowering of velocities due to presence of gas in the reservoir.
94 NG TONG SAN, IDRUS MOHD SHUHUD AND LEONG LAp SAU
Density trend
Density data are well separated for different lithologies and shows good trend patterns as seen in Figure 3. Coal has the lowest density range of 1.2 g/cm3 to 1.8 g/cm3 followed by sandstone with density range between 2.1 g/cm3 to 2.4 g/cm3. Shale has the highest density range between 2.4 g/cm3 to 2.6 g/cm3. This indicated density can be used to differentiate lithology types.
Acoustic impedance trend Acoustic impedance data for all lithologic type
are quite well separated which is similar to the density trend (Fig. 4). This implies acoustic impedance is mainly influenced by the density of the rock type and emphasizes the importance of density as a major contributing factor that affects seismic reflection amplitudes. In the study area
the sandstone reservoirs have a lower acoustic impedance compared to shale.
Gassmann model
A study carried out by Ng and Leong (1990) on reservoir characterization using the Gassmann model showed that the presence of gas in sandstone has only a minor effect on seismic amplitudes.
In Figure 5, computed velocities from Gassmann's equation assuming a clean sandstone (quartz) with bulk modulus of 35 x 1010 dyne/cm2,
indicate a large velocity drop as due to the initial presence of 2.5 percent gas. The large lowering of velocity due to gas is generally associated with anomalous high amplitude reflector on the seismic section and are commonly use as a direct hydrocarbon indicator by seismic interpretator. Subsequent calibration of the Gassmann model from
STUDY AREA
SINGAPORE
Figure 1. Location map of the study area.
GeoL. Soc. MaLaYdia, BuLLetin 36
~ " '" " ~ <:)...
" ~ ...... ~ ~
DEPTH (mss)
1,100
-
1,200
1,300
1,400
1,500
1,600
1,700 2,000
=
--:
=
2,200
* * • 6 * •
6 '"
- COAL
* SHALE 6 SST SF GAS
• SST SF OIL
o SST SF WATER
~ SSTSWGAS
• SST SW OIL
o SSTSWWATER
*
* * *
• *iIi o '"
* * *
o
* * *00.6*
* *
* * 0 * ** 0 0 00
'" * 0
'" '" * '" * '"
o
o
*
o
*
2,400 2,600 2,BOO 3,000
VELOCITY (M/SEC)
*
*
'" * *
'" 0* * *
* *
o * ;: * '"
* *
* o
* * *
o
3,200
Figure 2. Velocity trend plot of the coal, shale and sandstone.
3,400
DEPTH (mss)
1,100
1,200
1,300
1,400
1,500
1,600
1,700 1.2 1.4 1.6
- COAL
* SHALE
6 SST SF GAS
• SST SF OIL
0 SST SF WATER
- SSTSWGAS
• SSTSWOIL
0 SST SW WATER
1.B 2
DENSITY (glee)
* * **
* t* ,,~
* *",* ... - * * ,. 660
~* * ~61!1 6 *r* • .co .6 * * : ·1-: ;. "'* *; • • * *f * 6 '" ,f!<* • 0 *\ * .- (*** 1~00 0 ,f!< i;f*
.0 .f* "'# • • • I- l\l!<**", * .~o o ~O 0 ****1t t • 0 .00
ColO • ** >Il< *t, "** o 8 t. * ~ o 0 * *~*** 6 0 0
6 8 0 * * *",* * 0«'>0 *'* l iii
• t t*>Il< '\ * 00
0 * ~,( * .,00' * *
*** * * • 0 **** * 0
0 0 ** * * * 0 **
0 * lII'*# 0 * iii 0 6 0 * *>Ii< 0 *** 00 0 ,f .f
* []J]
0 0 '" ~ * "* ** * *>Il<
*** cP
* *
* o o
2.2 2.4 2.6
Figure 3. Density trend plot of coal, shale and sandstone. There is a wide separation between the three main lithologies.
» o c » ~
~ <: m
~ C Cl -< o "Tl
:t! m en m Ci5 s:: ('5 -i ~ m :I> s:: "0 .-=i c Cl m ::0 m "Tl
~ 5 z () :c » ::0
~ m ::0
~ ('5 en Z » z o ;= ::!! m .Cl
<0 CJ1
~ l'
~ !'
~ ~ ~ ~.
b:I !::
~ .... s· ~
DEPTH (mss) 1,100 .----
1,200
1,300
1,400
1,500
1,600
1,700 2,000
.:
--
-::
---
3,000 4,000
....
D.
D." ....
* * * *
* * * * * .D. >i<I* * • * * ·00- **~ o·
D. D. D. **" ~* D. '* **
* *
D.'('\... ~ * * ... .... * * ···tII ***** D., • ~ * •• ~ * * ,**
D. •• • 0 **:iff ..• 0 * * ....... O*OO**"*'>tIi*
•
D.
•• **'"' ~ * • • ~* '"'iii '!i<*4"'*
., 0 ~ '50" * 4*'!i<* o tP 0 ~"-** * * ~ 0 • ,p. *"- ** * *
<00
• 0 ~ ** * 00 0 (lI,\* W-~ * *
D. >t!' .iI<#. (j)0 0 * **,. *>j/I'
.0 ** * ~ ** •• 0 ** t \ ** *
00 0 ro * * '" o 00 0 °* t
• g. * ***>Ii*
o 0 * *\ _* ¥'O* *
,f
.. •
o
o LD * 0 0 *>Ii * * o *0
lID * *
.coAl.
* .sJ:!AL..E D. SST SF GAS
• SST SF OIL
o SST SF WATER
.... SST sw GAS
• SST SW OIL
o SST SW WATER
5,000 6,000
o 0
ACOUSTIC IMPEDANCE
o
,f * * * * * *
* * ** *
"'" o 0
** *
* o
7,000 8,000
Figure 4. Acoustic impedance is the product of velocity and density.
9,000
Vp(m/oec)
3,000
2,900
2,800 ~ $
2,700
2,800
2,500
2,400
2,300
2,200
************
r :. WEUOATA
COMPUTED 1 o
o 0
• ~ ••• * • 1:>' • ~ ••• * ••• * • ~B,!" •• • • • o
o o
• * •• * •••••• * ••• * •••••••••••••• ':.a'!" •• • • •
Unit K810tody'ne/~
I I I
10 20 30 40 50 60 70 80 90 100
Sg (%)
Figure 5. Calibration of the well data using Gassman model suggested the bulk modulus of sandstone is about 20 x 1010 dyne/cm2.
Vp(mJoec)
3~ rl------------------------------------------------------~
3,200
3,000
2,800
2,800
2~
2,200
2,000
l,aoo
1,800
•
•
~~ AAA ••••• AAA •••••••• •••••••••••••••••••••
~ 0.26
***************************************
~ 0.30
• •••••••••••••••••••••••••••••••••••••••
• • ~ 0.34
••••••••••••••••••••••••••••••••••••••
10 20 30 40 50 80 70 80 80 100
Sg(%)
Figure 6. An increase of porosity from 30% to 34% will lower the velocity in the same magnitude as in the case of presence of gas,
(0 0)
Z G>
d z G>
~ i5' ::IJ C en s: o ::c c en ::c c ::c c c ~ c r ::s z G>
~ -0
g? C
A QUANTITATIVE STUDY OF THE SEISMIC TIME-AMPLITUDE REFLECTION CHARACTERISTICS IN AN OIL FIELD 97
well data in the area shows that the bulk modulus of sandstone is about 25 x 1010 dyne/cm2. This suggests the presence of impurities in the sand. The corresponding drop in velocity is much less compared to the initial computation for clean sandstone.
An increase in porosity of 4 percent in the sandstone will lower the velocity in the same magnitude as in the case of presence of gas (Fig. 6). Thus a drop in velocity can be attributed either to the presence of gas or an increase in the porosity of the sandstone (Gardner et al., 1974). Therefore it is concluded that, in the study area, the presence of high reflection amplitude in the seismic data is more likely due to an improvement of porosity.
FORWARD SEISMIC MODELLING
Seismic modelling was carried out to determine the change of bed thickness on the seismic response (Mecker and Nath, 1977). Information from well data and our results of studies from it are incorporated in constructing the geological model of the Lower D Sand reservoir.
Stratigraphically, Lower D Sand is bounded at the top and bottom by shale. The seismic responses have been generated by convolving the reflectivity series with a 10-60 Hz bandpass zero phase wavelet which is similar in frequency content with the acquired seismic data (Domenico, 1974). Density and velocity parameters for each lithology are the average values obtained from well wireline results.
Synthetic seismogram A 10-60 Hz bandpass zero phase wavelet is
used for generating the synthetic seismograms. Figure 7 is an example from a typical well. There is a good correlation between the synthetic seismograms and the seismic section as shown in Figure 8.
Synthetic seismograms from all wells have been generated and tied to each individual well location on the seismic section to ensure that the correct seismic .reflectors corresponding to the top and bottom of the Lower D Sand reservoir have been correctly chosen.
Here peak and trough reflectors in the synthetic seismogram correspond to the black (peak) and white (trough) reflectors in the seismic data, respectively.
10 seismic modelling This modelling experiment generates a sequence
of synthetic seismic trace from a model of the Lower D Sands. Figure 9 shows the reflection geometries corresponding to a thickening of the sandstone (p = 2.11 g/cm3, V = 2,418 mls) surrounded by shale (p
December 1994
= 2.42 g/cm3, V = 2,932.6 mls). This is explained by accompanying quantitative plot Figure 10. This plot shows the variation of measured peak-to-trough thickness or apparent thickness (in msec) and relative amplitude as a function of the actual thickness (in meter).
The maximum amplitude occurs at 11 meters, which is called the tuning thickness of the sand, and corresponds to the location where the peak-totrough separation time becomes invariant. Below this point, the actual thickness is linearly related to the relative amplitude whereas above it, the point is linearly related to the peak-to-trough separation time.
In summary, peak-to-trough separation can be used to estimate bed thickness for units thicker than the tuning thickness (thick bed), and, below turning thickness (thin bed), the amplitude response can be used to measure the thickness instead.
3D SEISMIC INTERPRETATION
The inferences and results derived from the learning phase of the first part of this study are used next to carry out the interpretation of the 3D seismic data. Seismic reflectors which correspond to the top and base of the Lower D Sand which have been confirmed by the simulated synthetic seismograms, are picked, and correlated, to generate time structure maps, isochron maps and amplitude maps at each level using seismic interpretation workstation.
Structure The time structure maps at the top Lower D
Sand is shown in Figure 11. Correlation and interpretation (tracking) of the seismic reflector was executed using the AUTO mode. This map indicates the Dulang structure as an elongated anticline plunging towards the west with complex faulting at the crestal area where most of the exploration wells have been drilled. The main fault trend is in the east-west direction while secondary fault trend is in the northwest-southeast direction. Wells results suggested that the main east-west fault trend to divided the hydrocarbon accumulations is a northern and southern system.
Isochron The isochron map as shown in Figure 12 is
generated by subtracting the time structure map values of the base Lower D reflector from the top Lower D reflector. It indicate that thicker isochron is found especially on both of the two structural flanks, and, that the general sand geometry is trending in the east-west direction. Thin bed thickness are shown to exist in the southeastern
98 NG TONG SAN, IDRUS MOHD SHUHUD AND LEONG LAP SAU
DEPT TIME GR DEN VELO C I T'r RCCJUSTIC RC RRW DEPT M SEC GRPI G/C3 MT / S IMPEDRNCE SERIES SYNTHETIC FT
00
:~ 0 00 0 0 0 =ooooooo~ ("\J 0 ("\J
00 0 o 00 0 0 0 =00000000 . 0 0
~ . 0 0 0 0 0 =0000000...-0 0 0 0
0 (\J - en -- C\J ('r") -.;;r Ln lQY)~l1"H.D["'--.COO1-- I
I. 05 1 000- -3281
1 0 5 0 - I . 1 0 -r-:-;....;+;--:..-;..-'-t-t+-'--'-+¥.-'+if+--H-H++++J.++-+++-+++++++++++--{.----1- - 3445
11 00 - -3609 I. 15
1150 -
1200 - -393 7
1250 - -4101
J 300- -4255
Figure 7. Synthetic seismogram of a typical well in the study area. The zone of interest is at 1.17 to 1.19 sec TWT.
Figure 8. The Lower D Sand at the well location correlated to seismic data at 1.14 to 1.16 sec TWT.
A QUANTITATIVE STUDY OF THE SEISMIC TIME-AMPLITUDE REFLECTION CHARACTERISTICS IN AN OIL FIELD 99
1m 2m 4m
30
--. 0 Q) 25 rJl
E "-'"
rt.l 20 rt.l r£I Z ~ U 15 1-4
~ E-t
~ 10 r£I r::t= < O-t 5 O-t <
O· 0
Decemher 1994
8m 8m
1-0 SEISMIC MODELLING
SANDSTONE
10m 11m 12m 15m
Figure 9. I-D seismic modelling of a sand encased in shale.
MODEL SANDSTONE
20m
/ RELATIVE AYP~E
'" TWO-WAY TIYE .
5 10 15 20 ACTUAL THICKNESS (m)
Figure 10. The 'tuning' thickness occur at eleven meters.
25m
30
25
r£I t:l
20 ~ E-t ~
H O-t ~
15 < r£I ~
10 E-t j r£I r::t=
5
0 25
~ <::> r---
~ !'
~ f ~.
~ ;:::
~ ~ ;:,
~ 0-.
1050.00C
zj .1 070 .000
~ . 1090.000
1 ' 1110.000 ;;;:
· 1130.000
· 1250.000
·1270.000
· 1390.000
' 14 10.000
' 1430.000
· 1450.000
· 1470.000
· 1530.000
· 1550.000
TIME
~
Figure 11. Time structure map of the top Lower D reflector.
· 9 .00000
,~,
. 12 .00000
·15.00000
·18.00000
'21.00000
'l .<
·24.00000
·27.00000
· 30.00000
MAP VALUE
I{~,. 6~J " ~6" ' lB· · .. ~ . · i~~l;
To J, f ~i' ".J .... !iI" ''"l.
.... ' ~ .
;
i\
I,' =O~L'6
.. :.
Figure 12. Seismic isochron map ofthe Lower D Sand.
~
....... o o
Z Gl
6 2 Gl (J) :._2
0-:D C en s:: o :r: o (J) :r: c :r: c o :.-2 o , m o 2 Gl
~ -0 (J) :.c
~ " '" ~ <:::--'" .., .......
~ ~
AMPLITUDE
·0.00
· 1300.00
· 2600.00
· 3900.00
·5200.00
· 6500.00
·7800.00
'9100.00
· 10400.00
·11700.00
· 13000.00
Figure 13. Seismic amplitude map of the top Lower D reflector.
AMPLITUDE
--13000.00
- -11500.00
- -10000.00
- -8500.00
- -7000.00
- -5500.00
;i( -4000 .00
~ ~~ .,
:d_ - 2500 .00
. -1000.00
Figure 14. Seismic amplitude map ofthe base Lower D reflector.
:t> o c: :t> Z -l
~ <: m (J) -l c: o -< o "Tl -l I m (J) m en ;:: o -l ~ m ;;. ;:: -u r =i c: o m :0 m "Tl r m (") -l 6 z (") I :t> :0 :t> (") -l m :0 en -l o (J)
Z :t> Z
o ;= ., m r o
....... o .......
102 NG TONG SAN, IORUS MOHO SHUHUO AND LEONG LAp SAU
TWT(msec)
30
25
20
15
10
I I 10
" . ....
I 15
..................
... "" .' ..............
I 20
THICKNESS (m)
.!" '/
I I 25 30 35
Figure 15. This plot established a linear relationship between seismic isochron and reservoir thickness.
KMS 0 2 KMS
SEISMIC AMPLITUDE
10,000
9,000
8,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
24 25 26 27 28 29 30 3 1 32 33 34 3~
POROSITY ( '" )
Figure 17. Assuming a linear relationship between seismic amplitude and reservoir porosity, the error could be ±2% porosity unit.
IB.O 22.0
~ 0 . 0
@ 0 15.0 IB.O
C::::J:::P 0
€!D 0 12. 14. ~ ... , ..
6.
Figure 16. Thick reservoirs are found at the flank ofthe structure whereas thin reservoirs are mainly concentrated in the southern and northwestern corner.
GeoL. Soc. MaLaYdia, BuLLetin 36
A QUANTITATIVE STUDY OF THE SEISMIC TIME-AMPLITUDE REFLECTION CHARACTERISTICS IN AN OIL FIELD 103
and northwestern corner of the structure.
Amplitude map
Amplitude maps of the two reflectors corresponding to the top and base of Lower D Sand are shown in Figures 13 and 14. These maps are direct output from the seismic interpretation workstation. They indicate high amplitude towards the west and low amplitude toward the east of the anticlinal crest and is probably due to relative high and low porosity beds respectively. The amplitude maps highlighted a east-west trend. By comparing Figures 13 and 14, they indicated a similar trend even though they are extracted from two difference seismic reflectors.
In the most area, the amplitude map of the top Lower D reflector have a higher value compare to the amplitude map of the base Lower D reflector. This seismic attribute characteristics is consistence with the results we highlighted in the synthetic seismogram and ID seismic modelling.
SAND THICKNESS PREDICTION
Seismic isochron values (workstation output) ofthe Lower D Sand, observed at the vicinity of the well location are plotted against the actual thickness which have been calibrated by the well results. The plot for the Lower D Sand are shown in Figure 15. It suggests a linear relationship between seismic
isochron and reservoir thickness. Some ofthe well locations have to be shifted towards the south due to different navigation system used in well location and 3D seismic acquisition
Estimation and delineation of the thick beds (defined here as a unit thicker than the tuning thickness) are carried out by using the isochron maps and well calibration plot. Thick beds (Fig. 16) which are interpreted to have thicknesses of between 10 to 36 meters, are widely distributed all over the Dulang structure. The thin beds of the Lower D Sand was interpreted to have a thickness of less than 10 meters (corresponded to 9 msec TWT). These beds are generally located in the northwestern and southeastern corner of the structure. Two areas were interpreted to have reservoir thicker than 22 m but it had to be further proven by well results. Over the crestal area there are numerous pocket of thick and thin beds reservoirs.
RESERVOIR POROSITY PREDICTION
Seismic amplitude values (workstation output) correlated to the top of Lower D Sand observed at the vicinity of the well location are plotted against the average porosity of the sand reservoir (Fig. 17). Assuming a linear relationship between seismic amplitude and reservoir porosity, the error in predicting reservoir porosity from seismic amplitude
Figure 18. Porosity of the Lower D sandstone is predicted to range from 24% to 36%.
December 1994
104 NG TONG SAN, IDRUS MOHD SHUHUD AND LEONG LAp SAU
map is ±2% porosity unit. The error in estimation ofthe reservoir porosity is most probably due to the effect of scattering in the velocity data as discussed earlier. The result of this porosity prediction is shown in Figure 18. Relative high porosity is found in the west of the anticlinal crest. In the south eastern and north western corner of the map where low porosity (low amplitude) is compounded by thin bed (less than 9 msec isochron) it could imply a shale-out region of the Lower D Sand.
RECOMMENDATIONS
The following further work are recommended: a. Recalibration of the seismic attributes to bed
thickness and porosity incoryorating the additional development drilling results and updated sedimentological model.
b. Further work on multi seismic attributes calibration through neural network analysis.
CONCLUSIONS
We conclude that: a. Seismic attributes (time, amplitude etc.) can be
used to predict bed thickness and porosity. An
integrated approach of using well data, seismic modelling techniques and 3D seismic data interpretation can aid in the prediction of reservoir geometry with a higher degree of accuracy.
b. Seismic attributes study should provide geophysical constraints to sedimentological model.
REFERENCES DOMENICO, S.N., 1974. Effect of Water Saturation on Seismic
Reflectivity of Sand Reservoirs Encased in Shale. Geophysics, 39, 759-769.
GARDNER, G.H.F., GARDNER, L.W. AND GREGORY, A.R., 1974. Formation Velocity and Density - the Diagnostic basics for Stratigraphic Traps. Geophysics, 39, 770-780.
MECKER, L.D., JR., AND NATH, A.K., 1977. Geological Considerations for Stratigraphic Modelling and Interpretation, AAPG Memoir 26,417-438.
NG, T.S., AND LEONG, L.S., 1990. Application of Gassmann's Model in Hydrocarbon Bearing Reservoir Characterisation. Petroleum Geology Seminar 1990 (Abstract, Geological Society of Malaysia, Kuala Lumpur, 3.
WIDESS M.B., 1973. How Thin is a Thin Bed? Geophysics, 38, 1176-1180.
---------.~.~~-.~.---------
Manuscript received 8 April 1994
GeoL. Soc. lI1aLaYJia, BuLletin 36