石油工程专业英语.docx

Reservoir Model Construction
High-frequency cycles were picked in 48 well with in the study area using
Porosity as a surrogate for rock fabric. A segmented of the resulting cycle
Stratigraphy in the middle upper Clear Fork is shown in Fig 48. In order

Fig.7.48.Cross section from a portion of the middle Clear Fork reservoir showing
Correlation of high-frequency cycles and flow layers based on porosity logs
to maintain the high- and low-permeability intervals, each high-frequency
cycle is divided into two rock-fabric flow layers, sandstone and mud-
dominated dolostones at the base and grain-dominated dolostones at the
top. The middle upper Clear Fork reservoir is divided into 21 high-
frequency cycles and 42 flow layers, and the lower Clear Fork is also divided
into 42 flow layers. The high-frequency cycles are the basic geological
elements, and the flow layers are the basic petrophysical elements for
constructing the reservoir model.
The high-frequency cycles and flows layers were correlated over the
study area, forming the basis for the reservoir model. A cross section of the
middle upper Clear Fork reservoir showing cycles and flow layers is illustrated
in Fig. 49. No cycles are described in the lower measures because it
is not part of the reservoir due to high water saturations typical of the transition
zone.For illustrative purposes, petrophysical properties distributed within the flow
layers using simple linear interpolation methods(Fig.50).

Fig. 7.49. Cross section illustrating the layer model for the middle Clear Fork reservoir
Showing 7 silt-based cycles labeled A-G, 14 carbonate cycles, and 21 rock-fabric flow layers.
Fig. 7.50. North-to-south Stratamodel cross section of the middle Clear Fork reservoir
Showing permeability distribution.

Flow Simulation Model
A 3-D reservoir flow simulation model of the study area was constructed
(Fig.51) using a method that links high-resolution sequence-stratigraphic
frameworks, porosity-permeability relations from core data ,outcrop-derived
models of small-scale spatial statistics, and a practical approach to
porosity-pemeability scaleup (Jennings et al.in press).Identification and

Fig.7.51.Tracer simulation results. a Tracer sweep pattern in the improved model
At one pore volume injection. b Tracer sweep pattern in the traditional model at
one pore volume injection

modeling of petrophysical layering are critical for waterflood performance
prediction. In this study the layering is based on high-frequency cycles and
rock-fabric flow layer.The large-scal component of petrophysical variability
is spatially organized into rock-fabric flow units with abrupt vertical
contrasts at flow-unit boundaries and gradual lateral transitions. The
flow-unit-scale petrophysical layering is laterally persistent at interwell
scales, leading to highly stratified reservoir behavior with rapid waterflood
sweep in the higher permeability layers, bypassing of the lower permeability
layers, minimal cross-flow between layers, and early water break-through.
Fluid-flow simulation was conducted to assess the benefits to reservoir
Performance prediction provided by the improved model described here
Compared with an existing model. The existing model was constructed
Without a high-resolution sequence-stratigraphic framework,and layering
Was assigned by proportioning layers between traditional stratigraphic
markers. This model will be referred to as the “traditional model.” The
model developed in this study using layers defined by rock fabrics and
high-frequency cycles will be referred to as the “improved model.” The
areal grid of the improved model was chosen to coincide exactly with that
of the traditional model, and the same set of well-log data was used for the
construction of both models. The same simplified set of well controls was
used in both models.
Injectivity and sweep were the aspects of reservoir performance addressed
in this study. Meaningful comparison of the injectivity and sweep
predictions of the two models was achieved by conducting single-phase
tracer injection simulations, avoiding the additional complications of waterflood
modeling. Thus, no initial saturation, residual saturation, or relative
permeability modeling was required. The single-phase fluid was modeled
as an incompressible liquid having a constant viscosity.
Detailed waterflood matching of the traditional model to historical
SWCF performance was conducted in a previous study. A good history
match was obtained by applying a kv/kh multiplier of 0.0002 to reduce
cross-flow between the layers and by increasing the horizontal permeabilities
by a factor of 2 to match reservoir pressure behavior. In this study the
improved model was matched to the traditional model, with both models
running the same simplified incompressible tracer displacement case, by
adjusting the same two parameters.
The kv/kh multiplier in the improved model was adjusted to obtain the
same sweep at one pore volume injection. However, the kv/kh multiplier required
to achieve this match was 0.02, two orders of magnitude larger than
the 0.0002 required in the traditional model and much closer to the moderate
flow-unit scale kv/kh ratio expected from typical whole-core data in carbonates.
This improvement in performance modeling was produced by the
improved representation of petrophysical layering in the model.
The tracer sweep patterns at one pore volume injection in the improved
model are stratigraphically organized into alternating high- and lowpermeability
flow units in the middle Clear Fork, and thin higher permeability
flow units near the top of the lower Clear Fork (Fig. 51a). These
sweep patterns were produced by the stratigraphically organized petrophysical
layering. The corresponding sweep patterns in the traditional
model are more random (Fig. 51b). The improved model also predicts
more injection in the southern portion of the model, relative to the injection
predicted by the traditional model, because of the subtle north-tosouth
porosity increase detected by the trend modeling portion of this
study. Careful comparison with reservoir performance data, outside of the
scope of this study, would be required to demonstrate that these sweep patterns
in the improved model constitute a superior representation of reservoir
behavior. Nevertheless, the sweep patterns are consistent with the
SWCF geological interpretation and are thus more satisfying

7.5.3 Fullerton Clear Fork Reservoir

The Fullerton Clear Fork study is an example of using stratigraphy to obtain
petrophysical class numbers where only gamma-ray/neutron logs are
available and multiple values are required because of the diversity of rock
fabrics. The Fullerton Clear Fork field in Andrews County, Texas (Fig. 52)
was discovered in 1942, and the Fullerton Clear Fork Unit formed in 1953.
The unit has produced 289 million barrels of oil from 1,250 wells and covers
an area of about 30,000 acres, or 47 square miles. Original oil in place
(OOIP) is estimated at between 1.6 and 1.9 billion stock-tank barrels
(BSTB), for a recovery efficiency of about 17%. The field produces from
permeability zones scattered over 500 ft of Permian-age Wichita and
Lower Clear Fork limestones and dolostones. The full report of this study
can be found at the Department of Energy Web site (Ruppel 2004)..

Vertical Succession of Depositional Textures

Twelve facies can be identified in the Fullerton reservoir succession on the
basis of grain type, grain size and sorting, fabric, depositional texture, and
lithology (Ruppel et al. 2004):

1. Peritidal Mudstone–Wackestone: generally dolomitized and most abundant
in the Wichita and locally in the Lower Clear Fork associated with
tidal-flat facies.
2. Clay-rich Carbonate Mudstone: generally thin and locally found associated
with peritidal mudstone-wackestone facies.
3. Exposed Tidal Flat: defined by fenestra, pisolites, mudcracks, microbial
laminations, and marked sea-level changes in the Wichita and Lower
Clear Fork.
4. Peloid Wackestone: a burrowed mud-dominated fabric deposited in a
low-energy subtidal setting

.
Fig.7.52.Location of Fullerton Clear Fork field

5. Peloid Packstone: a burrowed mud-dominated fabric with abundant
peloids (probably fecal pellets) deposited in a low-energy subtidal setting.
6. Peloid Grain-dominated Packstone: moderately well sorted peloids in
intergrain pore space deposited in a subtidal setting having moderate energy
levels.
7. Ooid-Peloid Grain-dominated Packstone-Grainstone: contains ooids and
skeletal grains, in addition to peloids, and is moderate (grain-dominated
packstone) to well (grainstone) sorted, suggesting moderate to high energy
levels.
8. Fusulinid Wackestone-Packstone: most abundant in the Lower Clear
Fork, found in sites suggesting water depths of 30 m of more, the deepest
water facies at Fullerton.
9. Skeletal Wackestone-Packstone: found mainly in the Lower Clear Fork,
containing mollusks and crinoids, suggesting a low-energy inner platform
setting.
10.Oncoid Wackestone-Packstone: abundant at the base of the Lower
Clear Fork through the entire field associated with fusulinids and other
faunas, suggesting an open-marine environment during flooding of the
platform.
11.Siltstone-Sandstone: restricted to the Tubb Formation that overlies the
Lower Clear Fork, but traces can often be found in peritidal and tidalflat
facies.
12.Lithoclastic Wackestone: thin beds of tidal-flat fabrics overlying tidalflat
facies.

The producing interval of the Fullerton field is divided into sequences
and cycles based on vertical successions of depositional facies from core
description. Two sequences are defined: Leonardian 1 (L 1) and Leonardian
2 (L 2). Most of the reservoir is found in the L 2 sequence, which
is divided into four high-frequency sequences (HFS): Leonardian 2.0 (HFS
L 2.0), Leonardian 2.1 (HFS L 2.1), Leonardian 2.2 (HFS L 2.2), and Leonardian
2.3 (HFS L 2.3) (Fig. 53).
The Wichita consists of a diverse assemblage of peritidal and tidal-flat
facies that group into the highstand leg of sequence L 1 and the transgressive
leg of sequence L 2 (Fig. 54). The highstand leg of L 1 represents the
landward tidal-flat equivalent of the basinward outer platform facies of the
Abo Formation, and the transgressive leg of L 2 represents the landward
tidal-flat equivalent of the basal Lower Clear Fork subtidal facies. Evidence
of karst is found in a few cores below the L 1 – L 2 boundary in the
middle Wichita (Fig. 54). Intervals of polymict breccia of at least 25 ft to
as much as 60 ft in thickness are present in one core. Their discontinuous
nature and association with other features indicative of karst processes
suggest that they originated as cave-fill deposits.
Sequence L 2 is subdivided into four high-frequency sequences (L 2.0,
L 2.1, L 2.2, L 2.3) (Fig. 54). HFS L 2.0 documents the initial flooding of
the platform following exposure at the end of L 1 time. In the field area it
consists of peritidal and tidal-flat facies of the upper Wichita. HFS L 2.1
forms the base of the Lower Clear Fork and consists of a basal section of
transgressive subtidal platform facies and an upper section of highstand
tidal-flat facies. It represents the sharp change from peritidal deposition of
the Wichita to subtidal deposition of the basal Lower Clear Fork. HFS L
2.2 is similar to HFS L 2.1 in consisting of a basal transgressive leg composed
of backstepping tidal-flat facies, a middle leg composed dominantly
of subtidal facies, and an uppermost highstand leg composed of tidal-flat

Fig.7.53.Fullerton field type log showing formation, sequences,high-frequency
cycles, and flow layers(Ruppel 2004)

facies. HFS L 2.3 is composed of tidal-flat-capped restricted subtidal cycles
in the field area and is capped by the siliciclastic Tubb Formation.
The fundamental goal of cycle stratigraphy is to develop a correlation
framework based on time-equivalent surfaces. These surfaces form the basic
correlations for constructing the reservoir model and define highfrequency
cycles (HFC’s). Because the Wichita is composed of peritidal
and tidal-flat facies, cycles are difficult to define and correlate. Only one
bed of good subtidal facies was found. Most of the correlations were based
on porosity and limestone-dolostone layering, the porous intervals being
dolostone and the dense intervals being mostly limestone and occasionally
dense dolostone. It is assumed that each dolostone bed was formed by hypersaline
reflux flowing down from a tidal-flat into peritidal facies.

Fig.7.54.Schematic cross section of Fullerton field showing formations,sequences,
And general facies distribution(Ruppel,2004)

Therefore, the dolostone beds mark the tops of the HFC’s (Fig. 55). Using
this approach, we divided the Wichita into 10 HFS’s labeled W1 – W5 and
W8 – W12. The interval between W5 and W8 has little porosity and was
not subdivided.
HFS L 2.1 is divided into seven high-frequency cycles (Fig. 53). These
cycles are labeled LC4 – LC10. The lower cycles are transgressive and
typically grade upward from fusulinid and oncoid mud-dominated facies to
better sorted peloid-rich and tidal-flat caps. The upper cycles are highstand
and are typically composed of peloidal facies at their bases and grain-rich
peloid- or ooid-bearing facies at their tops. The top two cycles are composed
of peritidal and tidal-flat facies.
Cycle definition was difficult in the dolostones of HFS L.2.2 because
of low porosity. However, three HFC’s were proposed on the basis of a
rare limestone core that consisted of three upward-shoaling successions.
These successions are labeled LC 1 – LC 3. The three cycles were subdivided
into eight autocycles for the purposes of full field mapping, but these
subdivisions were not used in the simulation model.
High-frequency cyclicity is most readily definable in HFS L.2.3. These
rocks, which are characterized by a tidal-flat-capped subtidal cycle, are

Fig.7.55.Well log of Wichita formation showing high-frequency cycles and flow
Layers based primarily on porosity and lithology

more easily correlated than HFC’s in L 2.1 or L 2.2. However, this HFS is
not considered to be part of the reservoir and is not included in the reservoir
model.