uAbstract
uFigures
uIntroduction
uBackground
uStructure and Deposition
uHistorical Production
uInterpretation of 3D
uConclusions
uReferences
uAbstract
uFigures
uIntroduction
uBackground
uStructure
and Deposition
uHistorical
Production
uInterpretation
of 3D
uConclusions
uReferences
uAbstract
uFigures
uIntroduction
uBackground
uStructure
and Deposition
uHistorical Production
uInterpretation
of 3D
uConclusions
uReferences
uAbstract
uFigures
uIntroduction
uBackground
uStructure
and Deposition
uHistorical
Production
uInterpretation
of 3D
uConclusions
uReferences
uAbstract
uFigures
uIntroduction
uBackground
uStructure and Deposition
uHistorical
Production
uInterpretation
of 3D
uConclusions
uReferences
uAbstract
uFigures
uIntroduction
uBackground
uStructure
and Deposition
uHistorical
Production
uInterpretation
of 3D
uConclusions
uReferences
|
Figure
Captions
Introduction
Back
in the late 30s, 3D wave propagation was investigated based on
multiple 2D lines in different directions and wave propagation
geometries (Rock, 1938). After the 50s, 3D  seismic reflection and
imaging were discussed based on physical 3D models. In the mid-70s,
the first 3D seismic survey was acquired in the Gulf of Mexico
(Dahm and Graebner, 1982). It was observed in the lab that a large
velocity change can occur in rocks with heavy oil if the oil is
replaced by steam (Nur et al., 1984). From the late 80s, the application
of time-lapse seismic was investigated (Robert and Terrance, 1987).
With hundreds of field applications in the last decade, time-lapse
seismic has become an industry-wide tool for reservoir monitoring
(Lumley, 2001).
However, offshore and onshore applications have not been evenly balanced.
Time-lapse seismic
has been successfully implemented in offshore areas, especially
in the North Sea and the GOM. It has also been successful in some
onshore areas where there are shallow, heavy oil reservoirs, especially
in Canada. Non-repeatable noise in time-lapse seismics at land fields
is a key issue. 3D legacy data always reflect differences in geometries,
acquisition directions, source/receiver types and their locations,
surface equipment, and surface conditions. As for processing, the
differences can include the processing flow, parameters, algorithms,
and the software system itself (Ross et al.,
1996). The non-repeatability from acquisition and processing
makes it difficult for 3D time-lapse seismics to be applied to onshore
fields, especially to thin interbedded reservoirs. Based on research
conducted at an onshore field, we propose a 3.5D seismic method to
obviate these problems.
Background of
Geology and Seismic
Geology
background: The
field is on a monoclinal structure that forms a lithological
trap for the reservoir, with depths between 3230-3480 m and sand
thicknesses ranging from 3 to 5 m. The field has been in production
since 1991, with water cut around 60% at present.
Seismic
acquisition and processing:
The 3D survey conducted in the
field is located in the northwest margin of the Jungar basin. The
surface is mainly covered with sand dunes of 3-20 m in height and
some farmland on the eastern side. The main parameters of the survey
geometry are as follows: Spread, 12 lines x 10 shots; Fold, 60; Bin
size, 12.5 m x 12.5 m. The processing workflow includes amplitude
preservation, source/receiver statistical deconvolution, velocity
picking, statics, and NMO+DMO+poststack migration. Compared with
the processing workflow of the conventional time-lapse seismic (Ross
et al., 1996), this workflow does not need to focus on the elimination
of non-repeatable noise from both acquisition and processing. This
indicates that a 3.5D seismic application could be more practical
compared to time-lapse seismics, possibly reducing the financial
risk of the entire process.
Reservoir Structure
and Depositional Study
Figure 1a
is a seismic
section across the reservoir, which is flattened at the bottom
of the Jurassic. From this section we can see that the deposition
before the Jurassic has an erosional period of exposure, with volcanic
activity in the deeper formation. The Jurassic formation contacts
its underlying formation as an angular unconformity. Above the
unconformity, the Jurassic formation starts to subside in the southern
part.
This
leads to a paleotopography with a high in northwest, and a low
in the southeast (see Figure 1b).
Sedimentary deposition starts in the early Jurassic (J1b). This
period has three sub-cycles of deposition as marked in Figure
1b
by the deep to light yellowish colors for the three superimposed
sedimentary units. The sediment source is from the northwest. After
this period, the southeast continues to subside and leads to the
formation of the mid-Jurassic deposition (J2x). J2x has two sub-cycles
of deposition that form two superimposed sedimentary units, with
the source in the northwest. In the late Jurassic (J3q) there
are two sub-cycles of deposition (see Figure 1b
marked with deep or light color) that are
thicker in the southeast than in the northwest. In the seismic
sections, a foreset reflection can be observed clearly in the northwest.
This indicates that the sediments are from the northwest.
Figure 1c
shows the present
structure from the seismic data. From Figure
1c
and the above structural evolution discussion, we conclude,
that with a good cap rock, the reservoir mainly has onlap and unconformity
lithological traps.
Based
on the structural and depositional evolution study, the seismic
waveform clustering attribute is generated as shown in Figure
2
. The mid-to-light blue colored region indicates a region
of alluvial deposition. An uplift (the white line in Figure
2) separates it from the adjacent depositional region. These
two zones form two major depositional regions. From the depositional
analysis, the J3q formation mainly forms onlap traps. On the other
hand, the J1b and J2x formations mainly form unconformity traps.
Historical Production
Data Study
Obviously,
the preceding structural and depositional discussion, which is
based only on high quality 3D seismic data, cannot characterize
reservoir dynamics. The key to 3.5D seismic is to further characterize
the reservoir using dynamic data such as historical production
data. Figure 3
is the spatial evolution
of cumulative oil production. The well locations and cumulative
oil production map of the early stage (1990) show that the field
development starts in the south. It then expands to the north with
time, and continues with the drilling of most of the production
wells until 1994. At that point, most wells with high cumulative
oil production are in the south (the big red circles in Figure
3
, 1994). Starting in 1996, the cumulative oil production
in the north increases every year. At the same time, the production
in the south starts to decline. It also shows that water production
starts to increase in the south. In 2006, the cumulative production
in the south is still slightly higher than that in the north, and
the cumulative production in the west is higher than that in the
east. Also, Figure 3
shows that
the wells near the field boundaries have relatively lower oil production,
which implies that they have encountered the oil/water contact. Figure 4
further shows the evolution of the cumulative
water production from each well. From the maps before 1994, observe
that the high water production areas are mainly located in the
south and south-east of the field (the blue areas in Figure
4). This implies that the water invasion is from the south-eastern
direction. The variations from 1996 to present give further evidence.
In addition, a water invasion path in the north of the field can
also be identified. By 2006, the field had been divided into three
low water-cut oil production regions (the red areas in Figure
4) and three high water-cut regions (the blue areas
in Figure
4). Comparing the evolution of the cumulative oil production
(Figure 3) with the water production (Figure 4), we see
that there was a higher cumulative oil production in the southeast
in the early stages of field development. At present, most regions
have become high water-cut areas except for the red areas shown
in Figure 4
.
Interpretation
of 3D with Historical Production Data
Based
on the interpretation of 3D seismic data and of the dynamic data, Figure 5a
shows the combined result of overlaying
the dynamic data with 3D seismic amplitudes. As shown in Figure 5a
, the low water producers are located in
regions where the seismic amplitudes are high, and the high water
producers are located in low amplitude regions in 2006. This indicates
a high correlation between the water producers and seismic amplitudes. However,
as shown in Figure 5b
(3D seismic
amplitude with cumulative oil before 2006), wells with high cumulative
oil are not located at places of high 3D seismic amplitude. Wells
with low cumulative oil do not always overlie regions of low seismic
amplitude either. This suggests that the high-resolution 3D seismic
data acquired in 2006 has been changed by fluid substitution. Thus,
through this study, 3D seismic data have been characterized dynamically.
To differentiate from time-lapse seismic, this can be called a
3.5D seismic approach. With the structural study of the 3D
seismic data and further integration of historical production data,
the aquifer invasion is mapped and the remaining potential areas
are identified in Figure 6
. The
water invaded from the southeast. In the north, two aquifers invaded
along two paths in the J2x formation. The unexplainable Well A,
as mentioned before, is located in a different sedimentary region
than the field’s major producing region. The area of the remaining
potential in the Well A region is estimated to be 1.5 km2. The
questionable Well B is located in the same facies as that of the
major production region, but it produces from a different sand
body. The potential for this region has an area of 1.1 km2.
Conclusions
Based
on a case study, a technique of 3D exploration for remaining oil
using historical production data (3.5D) has been proposed and demonstrated.
Compared with the workflow of time-lapse seismics, whose success
depends on reservoir conditions, production mode and seismic repeatability,
the workflow of our method is more practical, especially for onshore
fields with thin interbedded reservoirs. The results show that
reservoir dynamics can be well characterized with the seismic data
acquired in producing fields, by means of effective amplitude preservation
processing, structural and sedimentary studies, and integration
of 3D and historical production data. This addresses the repeatability
requirement for 3D time-lapse seismic and the requirement for good
legacy 3D seismic data. In turn, it reduces the cost of exploring
for remaining oil. The multi-disciplinary integration of seismics,
geology, and reservoir engineering will greatly improve the 3.5D
seismic technique and make it more effective.
References
Dahm, C.G. and R.J. Graebner, 1982,
Field development with three-dimensional seismic methods in the
Gulf of Thailand-A case history: Geophysics, V. 47/2, p. 149-176.
Lumley,
D.E., 2001, Time-lapse seismic reservoir monitoring: Geophysics,
V. 66/1, p. 50-53.
Nur,
A, C. Tosaya, and D.V. Thanh, 1984, Seismic monitoring of thermal
enhanced oil recovery processes: 54th Annual International Meeting
Society Exploration Geophysists, Expanded Abstracts, Session RS.6.
Robert
J.G. and J.F. Terrance, 1987, Three-dimensional seismic monitoring
of an enhanced oil recovery process: Geophysics, V. 52/9, p. 1175-1187.
Rock,
S.M, 1938, Three dimensional reflection control: Geophysics, V.
3/4, p. 340-348.
Ross, C.P, G.B.
Cunningham, and D.P. Weber, 1996, Inside the cross-equalization
black box: The Leading Edge, v. 15, p. 1233-1240.
Return
to top.
|
Print this page
Click to view article in PDF format.
