Professional Documents
Culture Documents
1990
by
Richard. M. Bateman
ABSTRACT
1
SPWLA 31st Annual Logging Symposium June 24-27. 1990
INTRODUCTION
2
SPWLA31StAmudL0gging Symposium.June 2627.1990
When a logging tool passes through a thin bed it convolves the properties
of the thin bed with the properties of the surrounding beds. Effectively
the “picture” taken of the formation becomes “blurred”. Figure 1
illustrates this concept. A generalized logging tool smooths out sharp bed
boundaries and distorts center-bed readings. Two examples should
suffice to place in perspective the problem facing the log analyst:-
Under-estimation of reserves.
3
SPWL.A 31~ Annual Logging Symposium June 24-27.1990
It is not the purpose of this paper to suggest that the method described
here is the only valid method of analyzing logs run in thin beds. It is one
among many but it does fill a need to describe the envelope of possible
solutions when binary lithology is assumed and logging data is limited.
4
SPWLA 31st Annual Logging Symposium. June 24-27. 1990
sensitive to noise and user deflned threshold picks. The solution adopted
calls for a 4 step process including-
Filters with substantially sinusoidal shapes have been found useful. A long
filter, three times longer than the short filter, has been found most
appropriate. Depending on the rate at which the raw data is sampled in
depth these filters must be designed to maintain approximately the same
vertical sample of formation. For the long filter about 3 feet (0.91 meters)
of formation is required and for the short filter about 1 foot (0.31 meters)
is sufficient. If the actual sampling rate is 4 samples per foot then a short
filter of 5 weights and a long filter of 15 weights have been found to
optimize the process. The scheme can be used with 2, 4 or 10 samples
II
per foot data. Evidently better resolution can be expected with higher
sample rates.
When applying this algorithm the analyst can control the degree of
bedding introduced into the column logged. At the step where a scaled
version of the AV-Shale curve is compared to the original V-Shale curve
the choice of a scaling factor controls the degree of bedding. What
criteria can the analyst use to control the choice of this scaling factor in a
non-subjective fashion 7 An initial guide can be obtained by running some
simple statistical measures of the distribution of values in the AV-Shale
5
SPWLA 31s Annual Logging symposium, June?4-11.19w
curve. The standard deviation of this curve over a given depth interval has
been found useful as a guide to the analyst. However the ultimate arbiter
is the use of integrated V-Shale curves. If the scale factor has been
chosen correctly then the integration of the conventional V-Shale curve
with respect to depth should produce the same number of ‘V-Shale Feet”
as the integration of the binary shale beds. Figure 6 illustrates this
process. In track 1 an overlay is shown of the original V-Shale curve with
a scaled A V-Shale curve with an initial scale factor of 10. In track 2 the
resulting lithology track is shown together with a monitor curve that
shows the difference between the integrated V-Shale feet from the two
methods.
Notice that the monitor curve progressively gets more negative as the
computation and integration progresses from the top down. This implies
that too many sand beds have been introduced to the binary mixture and
not enough shale beds. The scaling factor is thus changed and track 3
shows a revised overlay of the scaled AV-Shale with the original V-Shale.
Track 4 shows the revised binary lithology track and the monitor curve
that now hovers respectfully around zero.
In the case of interest to the log analyst the first term in the equation,
representing the conductivity of the shale portion of the sand/shale
“sandwich”, is overwhelmingly larger than the second term, which
represents the conductivity of a hydrocarbon bearing sand bed. For
example a typical shale conductivity might be 1 mho/m but a typical “pay”
sand might have a conductivity of only 0.04 mho/m, a ratio of 25:l. This
opens the door for a pseudo conductivity curve to be generated from a
knowledge of the shale conductivity and the shale content of the
6
SPWLA 31st Annual Logging Symposium, June 24.27. 1990
The smoothing filter used to produce the pseudo conductivity curve from
the ‘CSHVSH’ curve is not critical to the success of this process as long as
it faithfully covers the closest vertical geometric factor set for the
conditions where the induction tool is used. One that covers I2 feet
(3.66 meters) of borehole has been found very satisfactory.
7
SPWLA 31sr Annual Logging Symposium. June 24.27. 1990
With the binary bedding resolved and the formation resistivity enhanced
the next step in the analysis is to estimate the porosity in the sand
laminae and to compute water saturation in those laminae. For these
purposes the outputs of the conventional shaly sand analysis conducted at
the 1 to 2 foot vertical resolution scale are used in a redistributed form.
As was shown in figure 3, it is possible transform an effective porosity,
0e, derived from the dispersed approach, to the true porosity of the
individual sand laminae, 0l_am.
Such are the vicissitudes of practical log analysis that equations of this
sort are prone to lead to undesirable results unless tempered by common
sense limits which readers are encouraged to deduce on their own. In
most cases where danger lurks the binary lithology processing will have
already eliminated the majority of the most dangerous levels in the well,
where, for example, Vshalecl.
DISPLAY FORMAT
To make the output of this chain of analysis steps of practical use to both
geological engineers and completion/production engineers a display
.format of the sort shown in Figure 9 may be employed.
8
SPWLA 31st Annual Logging Symposium, June X-27, 1990
FIELD EXAMPLES
Bed Resolution
Figure 11 shows a comparison between a binary lithology column
and an independently derived borehole image made from a multi-
electrode dipmeter tool. Considering that one is made from the
electrical properties of the formation, and the other from a V-Shale
curve derived from poro-lithology tools measuring entirely different
formation characteristics, the agreement is striking. The II
conventional logs were sampled every 3 inches and it would appear
from this example that beds as thin as 3 inches can be resolved by
the binary lithology approach.
Porosity Resolution
Figure 12 shows a section of a sand shale sequence for which
sidewall cores were available. Note the agreement between the
measured and calculated values for both porosity and permeability
at depth X312’.
9
Comparison between conventional shaly sand analysis & binny
analysis
Figure 13 shows a comparison between two shaly sand analyses run
in a laminated sand/shale sequence from the Vicksburg formation
in south Texas. For comparison purposes the conventional analysis
is shown on the left two tracks and the binary lithology analysis is
shown on the right two tracks. For each analysis a pay count was
made based on the same criteria for porosity and saturation. It is
notable that the binary analysis reveals more pay than the
conventional analysis.
CONCLUSIONS
5. The method does not require the use of any high resolution logging
device. A simple triple combo logging suite is sufficient. This opens up
the possibility of re-evaluating old wells for which conventional logging
data is available in the files.
10
SPWLA 31s Annual Logging Symposium, June 24-27, 1990
REFERENCES
II
11
SPWLA 31~ Annual Logging Sympxium. June 24.27. 1990
6 ”
12
SPWLA 31st Annual Logging Symposium. June 24-27.19W
t
Rt ct
I I
2.5 0.4
5 0.2
10 -VShale-, R Sand = 20 Q-m
co 0 I I I I
0 0.2 0.4 0.6 0.8 1.0
Seen as
13
SPWLA 31s Annual Logging Symposium he 24-11.1990
14
SPWLA 31% Annual Logging Sympsium. June 2627.1990
II
15
SPWLA 31st Amual Logging Symposium. June 24-27.1990
16
SPwtA 3lstANludLogging Sympsium. June 24-27.1990
II
17
X600
18
SPWLA 31st Annual Logging Symposium. June 24-27. 1990
WELL : LWINATEDtl 1
DATE : a-HA&SO @ 11:14:2a
DKFTH PHIIAM aNL4l HCFVIM CXllHCPVL PAM CNtPAYUl PEFM&l CuMKHLAn
____-___
0547.00 11.710 100.000 ,000 .004 .OoO .oQo .045 .m
8547.25 12.380 94.645 .ooo .OOO . 000 .OoO .075 .OOO
6547.50 15.573 53.237 .016 .01a .250 .250 .944 .236
8547.75 14.822 55.751 .ola .ola .250 .250 .640 .236
a54a.00 13.750 72.317 .ola .01a .250 .250 .242 .236
8548.25 13.907 66.835 .016 .016 .250 .250 .304 .236
8548.50 13.971 63.991 .ola .016 .250 .250 .341 .236
8548.75 14.542 60.675 .ola .ola .250 .250 .4a2 .236
8549.00 12.681 82.245 .ola .ola ,250 .250 .127 .236
6549.25 .ooo 100.000 .OOO .ola .OOO .250 .OOl .236
8549.50 .OoO 100.000 .ooo .ola .OOO .250 .OOl .236
8549.15 .ooo 100.000 .ooo .ola .OQO .250 .OOl .236
8550.00 .ooo 100.000 .ooo .ola .OQO .250 .OOl .236
8550.25 16.363 53.410 .OlS .037 .250 .500 1.262 .551
8550.50 19.613 32.063 .033 .071 .250 .750 10.383 3.147
8550.75 19.979 31.99s ,034 .105 .250 l.OQO 11.645 6.058
8551.00 19.600 34.793 .032 .137 .250 1.250 9.332 a.391
8551.25 .OOO 100.000 .OoO ,137 000 1.250 ,001 6.391
6551.50 .ooo 1oo.ocml .OoO .137 1.250 .OQl a.391
6551.75 .oOO 100.000 . 000 .137 . 000 1.250 .OQl a.391
6552.00 .OoO 100.000 000 .137 .OOO 1.250 .OOl 6.391
8552.25 la.142 39.291 .02a .164 .250 1.500 4.331 9.474
6552.50 17.265 41.438 .025 .19O -250 1.750 2.692 10.197
8552.75 15.115 57.012 .016 .206 .250 2.000 .aaa 10.369
8553.00 16.437 42.762 .024 .229 .250 2.250 2.020 10.674
6553.25 la.535 26.676 .OM .263 .250 2.500 10.525 13.505
8553.50 17.655 38.648 .021 .290 .250 2.750 3.801 14.455
6553.75 19.928 26.615 .037 .327 .250 3.000 16.579 la.600
6554.00 19.262 31.007 .033 .360 .250 3.250 9.962 21.091
6554.25 20.215 25.224 .038 .3sa .250 3.500 20.111 26.118
8554.50 22.430 16.382 .046 .444 .250 3.750 70.669 43.766
6554.75 16.283 46.106 .024 .467 .250 4.000 3.026 44.542
6555.00 .m 100.000 .000 .467 .OOO 4.000 .OOl 44.542
8555.25 22.011 29.873 .039 .506 .250 4.250 23.694 50.516
8555.50 21.637 28.289 .039 .545 .250 4.500 24.042 56.526
a555.75 23.219 22.146 .045 .590 .250 4.750 59.909 71.504
8556.00 23.022 35.134 .037 .627 .250 5.000 22.613 77.157
6556.25 .OOO 100.000 .OOO .627 000 5.000 .OOl 77.157
8556.50 .ooo 100.000 .OoO .627 . 000 5.000 .OOl 77.157
6556.75 23.334 23.107 .045 .672 .250 5.250 56.885 91.328
6557.00 21.514 27.953 .039 .711 .250 5.500 23.793 97.277
8557.25 21.693 29.168 .036 .749 .250 5.750 22.967 103.019
6557.50 21.264 31.758 .036 .786 .250 6.000 17.264 107.340
6557.75 20.896 33.262 .035 .a21 .250 6.250 14.106 110.867
8558.00 21.856 26.431 .040 .a61 .250 6.500 29.251 116.179
8558.25 19.237 42.013 .02a .689 .250 6.750 5.384 119.525
6558.50 20.559 37.230 .032 .921 .250 7.000 10.216 122.079
6558.75 22.663 23.658 -043 .964 .250 7.250 44.629 133.237
6559.00 16.638 47.40s .025 .969 .250 7.500 3.497 134.111
8559.25 000 100.000 .000 .969 . 000 7.500 .OOl 134.111
20
SPWLA 31s~ Annual Logging Symposium. June 24-27. 1990
>
>
>
t
>
>
-L 1
-- 3
-*6
>
>
II
21
SPWLA 31s Annual Logging Sppsiurn, Iune 24-17. 1990
I-
L.- L1
22
SPWLA 31st Annual Logging Symposium, June 24.27. 1990
II
23