WF1 Course

Waterflood Manual Chapter 1.
Reservoir Aspect of Waterflooding
1. Reservoir Aspect of Waterflooding

1.1 Introduction
This chapter addresses the issues related to the reservoir aspects of waterflooding.
These issues can be grouped into three major topics: (1) when to water flood, (2)
waterflood design considerations, and (3) prediction/evaluation of oil recovery.
Waterflooding is a relatively mature technology and has been the subject of several
classical books. The focus of the chapter is to introduce new technologies applicable
to the waterflood process and incorporate them into the updated, background material
drawn mostly from the old waterflood manuals [Ref. 1-1]. For a more complete
description of the background material, the reader may consult classical references of
two waterflood monographs published by SPE [Ref. 1-2 and Ref. 1-3]. Inside Shell, a
training manual entitled "Supplemental Oil Recovery" by R.T. Miller of Shell Oil
Company [Ref. 1-4], together with an internal correlation for water drive recovery
efficiencies in Oman, [Ref. 1-5], also deserve highlighting.
The emphasis in this chapter is on waterflooding non-fractured reservoirs.
Waterflooding in naturally fractured reservoirs is a complex process with the majority
of experience in PDO (Oman) carbonate reservoirs. Efficient waterflood operation in
a fractured reservoir requires a good knowledge of fracture networks so as to use
them to enhance rather than short-circuit the flood. In addition, rock wettability has a
major impact on the performance. For general reference on the subject, an SIPM
report on the reservoir engineering aspects of naturally fractured reservoirs (which
includes waterflooding) may be consulted [Ref. 1-6].
Ref. 1-1 Niko, H. "Water Injection Manual", EP 93-2900, 1994.
(Electronic version not available)
Ref. 1-2 Craig, F.F.Jr. "The Reservoir Engineering Aspects of
Waterflooding" Third Print, Society of Petroleum Engineers, 1980,
ISBN 0-89520-202-6. (Electronic version not available)
Ref. 1-3 Willhite, G.P "Waterflooding" Society of Petroleum Engineers
Textbook Series Vol. 3, 1986, ISBN 1-55563-005-7.
Ref. 1-4 Miller, R.T. "Supplemental Oil Recovery" EP 89-0642 Shell Oil
Production training Series. (Electronic version not available)
Ref. 1-5 Vijver van de, W. "Guidelines for water drive recovery factors" EP
88-2307 1988. (Electronic version not available)
Ref. 1-6 Niko, H. "Modern Reservoir Engineering Techniques for Naturally
Fractured Reservoirs" EP 92-1870 State-of-the-art report, 1992
1.2 When to Waterflood?

1.2.1 Favorable reservoir conditions
Waterflooding has always been considered as a recovery process to enhance the oil
recovery beyond primary production. Although addition of reserves should be the
1/49
Waterflood Manual Chapter 1. Reservoir Aspect of Waterflooding
primary objective of a waterflood, it is also used as a method of disposing wastewater

and, in some situations, a method to combat well failures due to compaction and
subsidence of the formation, e.g., Belridge diatomite reservoir and Upper Green
reservoir of the Mars Basin [Ref. 1-7].
Prior to starting a waterflood, the best possible assessment of the primary recovery
mechanisms and reservoir architecture should be made. This requirement highlights
the dilemma many petroleum engineers are faced with, namely that both primary
recovery mechanisms and reservoir models become better defined as primary
production proceeds, but the attractiveness of waterflooding may suffer.
With regard to primary recovery mechanisms, the reservoir engineer should first
determine whether there is a strong natural water drive. In assessing the strength of
the natural water drive, permeability, aquifer volume connected to the reservoir,
effective compressibility, presence or absence of communication barriers (faults, tar
accumulations at the oil water contact), and non-sealing faults acting as water
conduits are important factors to consider. Note that bottom water drives can be as
strong as edge water drives (refer to the Auk Field in the UK North Sea; in this field
no significant edge water drive was expected in the main Zechstein reservoir,
however, it turned out that the field had a very efficient and strong bottom water drive
through the underlying Rotliegendes).
Ideal candidates for waterflooding are areally extensive highly under-saturated oil
reservoirs where it is not likely that a natural water drive will develop. Examples of
this type are some of the fields operated by Shell Expro in the North Sea, such as
Cormorant and Dunlin, that were waterflooded from the beginning.
Ref. 1-7 Coremans, J.W.A., ”Compaction calculation for MARS: A GeoMec
evaluation in assistance of well positioning and pressure
maintenance decision making”, EP 2001-5130, 2002.
1.2.2 Unfavorable reservoir conditions

An important primary recovery mechanism is gravity drainage where either a primary
gas cap is present or large amounts of gas are released from solution in the oil as the
pressure declines. Some thick reservoirs that have a large oil column or reservoirs
with a high dip, together with high horizontal and vertical permeability, can be
produced very efficiently under gravity drainage, thus eliminating the need for
waterflooding. Internal gas drive can further add to the high recovery efficiency under
gravity drainage. Some reservoirs may also have high initial reservoir pressures. At
these conditions, pressure maintenance by gas injection under miscible conditions
may be the best route to increase recovery efficiency rather than water injection,
particularly in the case of high dips.
Apart from reservoirs with strong natural water drive or an efficient gravity drainage
process, oil rims with bottom water and large gas caps are also poorly suited to water
injection. In the latter, water injection requires delicate GOR control in the producing
wells to prevent resaturation of the gas cap, first by oil and then by water.
Reservoir permeability assumes an added significance since it determines the
throughput of a waterflood process. It thus has an impact on the number of wells to
be drilled and the timing of the waterflood response, hence the economics of
waterflooding could be marginal in certain tight reservoirs. Water injectivity can be
2/49
improved by injecting at pressures higher than the fracturing pressure but with the
risks of uncontrolled fracturing, reduced sweep efficiency and/or loss of injection
outside the target zones.
High oil viscosity lowers the waterflood performance in two ways. First, the water is
more likely to channel through the oil because of the higher mobility of water than oil.
Second, the total throughput of the process is lowered just like in the case of low
permeability reservoirs. For a relatively high permeability reservoir, however, we can
increase the number of the producers (e.g., using 9-spot flood patterns) to compensate
for the low oil mobility. When the reservoir is relatively shallow (less than 2000 ft),
steam injection may become a preferred option because it provides very high recovery
efficiencies (proper application in the right environment can give recovery factors in
excess of 70%). Waterflooding in this kind of reservoir should be considered with
care because, once implemented, it could diminish the attractiveness of applying
steam injection in the future, as a lot of energy is needed to vaporize the cold water
already injected into the reservoir.
Reservoir heterogeneity significantly impacts the waterflood performance. Large
permeability contrasts and poor sand continuity can contribute to poor waterflood
sweep. Typical examples are the Ventura waterflood project in California USA,
which is predicted to ultimately recover about 30 % of the oil in place, and the Sirikit
waterflood in Thailand [Ref. 1-8], which showed a water breakthrough shortly after
commencement of water injection.
In case of naturally fractured reservoirs, a waterflood can be executed when the rock
matrix is relatively water wet, in which case capillary forces can help enhance the
entry of water from fracture to matrix. In some cases even with mixed-wettability,
floods are implemented taking advantage of the fracture orientation (e.g. Lekhwair).
In cases of highly oil-wet matrix, however, the capillary forces prevent the water
from entering the matrix, resulting in severe bypassing of oil and channeling of water
through the fracture system (see the early performance of the Natih waterflood in the
Fahud Field), [Ref. 1-9].
Ref. 1-8 Ketyungyoenwong, K.B., ”Sirikit waterflood phase I, VAR 5: Post
investment review, EP 2001-5136, 2001. (Electronic version not
available)
Ref. 1-9 O’Neill, Niel, ”Fahud field review: A switch from water to gas
injection”, SPE 15691, 1987.
1.2.3 Optimal reservoir pressure level

If reservoir conditions have been determined suitable for water injection, a decision
on when to start water injection has to be taken. Leaving out considerations on well
productivity, letting the pressure drop in an under-saturated oil reservoir has a
beneficial effect on the recovery by subsequent water injection if one assumes that the
residual oil saturation is not affected (fewer stock tank barrels left behind in the
waterflooded zone).
In saturated oil reservoirs, letting the reservoir pressure drop prior to waterflooding
produces two effects. A given reservoir oil volume subsequently trapped by water
represents more stock tank barrels than the case in which the pressure was
maintained. On the other hand, a drop in reservoir pressure below the bubble point
produces free gas saturations. A small free (critical) gas can lead to a significant
3/49
reduction of residual oil saturation to water. This is illustrated in Figure 1-1, which
shows laboratory results obtained for the Brent and Statfjord reservoirs indicating that
the presence of 10% gas saturation can reduce the residual oil saturation to water by
roughly the same amount.
In reality, the actual impact of gas saturation on the residual oil saturation will be
much less significant than that shown in Figure 1-1. The gas saturation in the oil
column will be limited to the critical gas saturation (typically about 5%) because the
remaining free gas will migrate up-dip through gravity forces. In addition, a portion
of the critical gas will be re-dissolved in the oil by the pressure build-up associated
with the water injection.
Allowing reservoir pressures to drop further and gas saturations to build up above the
critical saturation level prior to the water injection start should be avoided, as it can
lead to long fill-up periods, delays in response, unwanted oil productivity declines,
and increases in the oil viscosity. The latter has unfavourable effects on the mobility
ratio as will be shown later.
An additional factor affecting the pressure level is well bore integrity. When a
waterflood is implemented to arrest formation compaction and/or reservoir
subsidence, optimization of oil recovery becomes secondary.
Figure 1-1 Waterflood residual oil saturation as a function of free gas

saturation
4/49
1.3 Waterflood Design Considerations

1.3.1 Geological considerations
Successful identification, implementation and management of a waterflood
opportunity relies on a sound understanding of the geology and realistic modeling of
the reservoir. Quantitative measures frequently used to describe the impact of geology
on the wateflood process are:
• Heterogeneity, frequently quantified as Dykstra-Parsons coeffiecients, which
is the permeability contrast due to variations of rock facies, reservoir layering,
karsting, fractures, etc.
• Anisotropy, frequently measured as the ratio of kmax/kmin, which is the
preferred direction of fluid flow resulting from depositional architecture,
fractures, faults, etc.
• Drainability, which is the portion of the reservoir connected to the producers
and can be produced via primary depletion,
• Floodability, which is the portion of the reservoir connected to both the
producer and injector and can be produced via waterflooding.
Whilst, by definition, floodability is always smaller than drainability at the same well
spacing, both parameters increase with the decrease of well spacing [Ref. 1-10].
Reservoir heterogeneity can be addressed at different scales: from field scale (e.g.
channel orientation and connectivity) to plug scale (grain-size distribution, sediment
texture, facies changes, etc). Historically, most poor waterflood predictions are the
results of inaccurate heterogeneity descriptions in physical scales somewhat smaller
than well spacing and reservoir thickness (e.g. poor descriptions of reservoir layering,
continuity correlation, fracturing, thief zones, etc). Heterogeneities on a large field
scale are relatively easy to predict. Small, plug-scale heterogeneities do not correlate
across a long distance so their impacts on waterflood performance are less significant.
Key parameters, such as heterogeneity, anisotropy, drainability and floodability, are
intimately related to the perceived geological characteristics of a candidate reservoir
to be waterflooded: reservoir geometry, sand body orientation and distribution,
porosity and permeability distributions, faults, fractures, facies, primary or secondary
clay distributions, etc. In this section, key geological characteristics are presented
following the three major stages in the construction of geological model: structure
framework, depositional architecture, and reservoir properties including diagenetic
overprints (see Figure 1-2). Note that, depending on the maturity of the field, there
will be some uncertainties associated with the geological characteristics so a number
of conceptual geological (or static) models may be needed in a “green field”
development where the uncertainties of the key geological characteristics are still
large.
5/49
a. b. Proximal system?
Distal system?
Stacked channels?
Isolated channels?
c.
Figure 1-2 Reservoir structure framework (a), depositional architecture (b),

and diagenetic overprint (c) – three major elements in constructing
geological models for predicting waterflood performance
Ref. 1-10 Christman, P.G., ”Modeling the effect of infill drilling and pattern
modification in discontinuous reservoirs, SPE 27747
1.3.1.1 Structure framework
Key reservoir characteristics, such as reservoir dip, orientation and density of faults
and fractures, are associated with the overall structural framework – the first step in
geological modeling over which depositional architecture and reservoir properties
including diagenetic overprints will be built.
The overall morphology (shape) of the reservoir affects the fluid movement so it
needs to be modeled prior to embarking on a waterflood project. Gravity would be
expected to be the dominant force driving the fluid movement in a reservoir with high
dip angles, so a down-dip water injection would be the preferred waterflood scheme
rather than a pattern flood (see Section 1.3.2).
The reservoir depth, pore pressure and in-situ stresses also need to be understood in
choosing the most appropriate secondary recovery process (see Section 1.3.9) and to
optimise the timing (see Section 1.2.3) and flood pattern (see Section 1.3.2), should
waterflood be selected as the most appropriate secondary recovery method.
Seismically resolvable faults, and associated sub-seismic scale faults or fractures, are
crucial data to be gathered and analyzed before planning a waterflood project. The
structural complexity of the field will have a first-order impact on fluid flow, ultimate
recovery, development planning, and hence on the overall project profitability. Open
faults and fractures usually behave as preferential conduits for injected water due to
their very high transmissibilities compared to normal reservoir rocks. The
effectiveness of waterflooding in faulted/fractured reservoirs, therefore, depends upon
6/49
the geometry, distribution and orientation of high permeability conduits with respect
to the drainage and injection points. Water injection into an unfavorable network of
fractures and faults can result in very ineffective waterflooding since the water can
easily bypass the oil (see Figure 1-3). Closed fault and fracture systems can also form
flow barriers to injected water, resulting in poor waterflood sweep.
In closed fracture and fault systems, it is important to know the type of fault rock and
its static and dynamic behaviours. Cataclastic or cement filled faults will have
different impact on fluid dynamic than faults filled with water sensitive clays. The
latter, if originally discontinuous and not completely sealing, can react to water and
turn a flow barrier into a high transmissibility conduit.
Reservoirs with Extensive

Natural Fractures and Faults
(Challenging WF Candidate)
Key Questions to Ask
-Fractures open or closed
C-1
N
-Direction of open fractures FAULT
AT
-Fracture face mineralized or not IC O
-Frequency of fracture swarms C-4 PILAR SU
R
-Presence of fracture corridors C-5 C-1
-Rock wettability I
-Faults sealing, leaky, or major C-2
conduits for fluid flow C-3
C-4
MT GU
J/C A
C-5
Related Study Topics OL SA
ON R
E
G
MT UAS
-Tectonic history J/C A
OL RE G
ON M UA C-4
CA TJ SAR
•Structural complexity LIZ
AS /C O
LO
E C-5
N
- Burial history MT
C-6
J /C G
•Diagenesis OL UAS
ON AR
C-7
E E
CA GUASAR
•cementation LIZ
AS ON
MTJ/COL
•mineralisation
-Biodegradation CALIZAS
EN T
B A SA M
Figure 1-3 Challenges of waterflooding a natural-fractured and faulted

reservoir
The direction of maximum horizontal stress needs to be taken into consideration when
planning a waterflood project. Injecting water into faulted reservoirs can reactivate
some faults and induce thief zones. Understanding the present conditions of regional
and local stress fields in the subsurface can help us orient injector patterns and reduce
the risk of water breakthrough via induced fractures. (see Section 1.3.6) In some
instances, they can be the basic input to geo-mechanical modeling for predicting the
subsidence-induced well failures from depleting a reservoir consisting of very
compressible rocks and highly under-saturated oil.
1.3.1.2 Depositional architecture
Depositional architecture, encompassing the size, shape and distribution of reservoir

versus non-reservoir facies, can have a big impact on the fluid flow in a field. As
such, any conceptual geological models must capture the essence of this architecture,
encompassing factors likely to have a major impact on reservoir performance, and
address any associated uncertainty.
7/49
Predictive stratigraphic tools such as sequence stratigraphy is recommended, where

possible, to try to identify a range of scenarios of reservoir facies distribution. This
technique is applicable both to carbonate and clastic reservoirs and will help to
identify vertical and lateral facies changes. The use of field and reservoir analogues is
a good starting point of constructing a ‘base case’ subsurface realization. However,
the presence of multiple realizations, all honouring the same set of observed data, is
certainly possible as shown in Figure 1-4.
In clastic reservoirs represented either by fluvial deltaic or deep marine turbidite
systems, the shape, size, orientation and vertical distribution of channel-fill
sandstones determines the degree of connectivity and provides an indication of
complexity associated with anisotropy of the reservoir. Normally, the net-to-gross
ratio is a reasonable indication of the degree of vertical connectivity whereas an
accurate stratigraphic subdivision determines lateral correlation and thus lateral
connectivity. Vertical and lateral connectivity control overall floodability and
drainability of channel-dominated reservoirs and ultimately the effectiveness of
waterflooding.
Channel vertical and horizontal distribution, dip and connectivity determine the
number and spacing of injectors and producers. The injector design needs to take into
account the presence of vertical baffles/seals (e.g. cemented calcrete horizons, lower
shoreface shale or transgressive marine shale intercalations) and the internal
complexity of the reservoir such as the direction of sand degradation (petering out)
into shale of sand progradation. The latter is particularly important in reservoirs
developed in deltaic environments where facies heterogeneities among distributary
channels, long coastal drift upper shoreface sand and lower shoreface shale are areally
complex.
Fluvial architecture and waterflooding
Two alternative models can be produced from the
same starting same low net to gross fluvial
reservoir data (logs) depending on geologist’s
interpretation. Model 2 is based on new
supporting information (new logs, cores,
producing data) – (Moscariello, 2002)
5 KM
Model 1
A sound reservoir model and understanding of

internal architecture is key to plan an effective
waterflooding
Any waterflooding activity if planned and designed

to intervene in the lower unit based on Model 2
5 KM
would have been very ineffective for the same
lower straigraphical unit in Model 2
Model 2
Figure 1-4 Multiple subsurface realisation for fluvial reservoirs

In carbonate rocks, predictions of facies and subsequent diagenetic overprints are
critical to determine the degree of connectivity (and therefore flood efficiency)
8/49
because flow paths in carbonate reservoirs can be dominated by secondary porosity

(e.g. dissolution due to early and/or post depositional diagenetic process and
fractures). In addition, facies prediction is crucial to the rock permeability, which can
vary significantly among facies – sometimes even one or two orders of magnitude
even at the same porosity class. Accurate stratigraphical subdivision and property
distribution in each reservoir unit (porosity, horizontal and vertical permeabilities and,
in some instances, mechanical properties) are also needed to provide reasonable
waterflood predictions.
1.3.1.3 Reservoir properties including diagenetic overprints
The integration of existing data from cores, wireline logs, borehole images and,
sometimes, production data can provide good evaluations of property distributions
(e.g., porosity, permeability and facies) of reservoir and non-reservoir rocks.
Diagenetic events need to be identified in terms of 3D distribution of reservoir
properties. This is particularly relevant for carbonate rocks where diagenetic
overprints (dolomitization and karstification) can completely obliterate any primary
reservoir characteristics. The effect of dissolution processes at large scale in
carbonate rocks can create a secondary porosity network characterized by either
continuous or discontinuous large pore space that in turn control the advance of flood
fronts in the reservoir.
Early cementation or neoformation of clay minerals can affect the efficiency of
waterflooding. In particular, the presence of clay minerals reactive to water (swelling
clays such as montmorillonite) will reduce the water injectivity due to considerable
reduction of pore space.
Rock wettability (e.g. water wet or oil wet) can significantly affect waterflood
performance in natural fractured reservoirs. A water-wet, natural-fractured reservoir
can exhibit good waterflood performance like a normal matrix reservoir whilst an oil-
wet reservoir can exhibit immediate water breakthrough.
Occurrence of tight zones (e.g. cementation patches as in the Triassic of Cormorant
and Tern fields; smectite and kaolinite pore filling) due to secondary diagenetic
modifications also impact fluid flow in the reservoir. Depending upon the extent of
tight zones, this situation can lead to the formation of large areas of bypassed oil. On
the other hand, high permeability streaks can completely dominate the waterflood
process - leaving unswept oil in areas of poorer reservoir properties.
1.3.1.4 Aquifer
A correct estimate of aquifer geometry and volume connected to the reservoir is

essential in waterflood planning. The size of the aquifer and the path of aquifer
movement can be inferred by regional structural maps of the subsurface. This coupled
with the tectonic and burial history of the area of interest, will determine whether the
aquifer is at hydrostatic or overpressured conditions and therefore whether the natural
aquifer will provide pressure support. An accurate prediction of the primary
characteristics of aquifer rocks is important in understanding the dynamic behaviour
of the aquifer.
The presence of an open network of faults and fractures extending from the oil
reservoir to the aquifer (or even to separate formations) can provide preferential
9/49
conduits to injected water and can be a reason for low waterflood performance. On
the other hand, in the presence of a strong aquifer, it can also provide stronger support
than a conventional flank aquifer, as observed in the main area of the Al Huwaisah
field (see Section 1.4.5.2.4).
1.3.2 Peripheral versus pattern water injection

When making a choice between peripheral water injection (e.g. injection wells drilled
near the original oil/water contact) and pattern water injection (e.g. five spots, nine
spots), two main factors play a role. The first is sweep efficiency and the other is
injectivity.
In case of relatively thick reservoirs with good hydraulic communication vertically, or
reservoirs with high dip angles, it is always advisable to inject water near the original
oil/water contact, as this ensures the best vertical sweep efficiency due to gravity
segregation. A pattern water injection scheme in this case would result in water
slumping towards the base of the sand and flowing downdip without displacing much
oil. It may nevertheless be necessary to resort to adding water injection into the oil
leg, when adequate injectivity or drainage cannot be ensured. An example is the
Cormorant Field in the North Sea, where injectors had to be drilled into the oil
column in Block IV because of a high degree of faulting (lack of continuity in the
reservoir) and in Block II because of injectivity problems in water injection wells
drilled at the oil/water contact. The injectivity problems in Block II were associated
with water sensitive clays.
Opposite to flank water injection, crestal water injection has been tried in a number of
cases, such as MAW/X and Seria West in Brunei and Gamba/Ivinga in Gabon. These
projects were, however, mostly abandoned, in view of the excessive time it took to
fill up the secondary gas caps and the tendency of water to slump downdip.
Pattern water injection is best suited for reservoirs of low vertical hydraulic
communication, where little advantage can be expected from oil / water gravity
segregation. There are a number of possible flood patterns: 5-spot, 7-spot, 9-spot
patterns, and line drive, but, in most cases, the optimal flood patterns are dictated by
reservoir architecture such as fault pattern, areal heterogeneity and reservoir
anisotropy. From a reservoir flooding perspective the aim is to ensure reasonable
reservoir connectivity between the injector and the producer, but short-circuits
through fractures, thief zones and conducting faults should be avoided. Having taken
reservoir architecture into consideration, it may be preferable to orient the injector
rows along the direction of maximum horizontal stress to reduce the possibility of
early water breakthrough via induced fractures. Finally, there should be a balance
between productivity and injectivity through a proper choice of the producer-to-
injector ratio: 1:1, 1:1, 2:1, and 3:1 for line drive, 5-spot, 7-spot, and 9-spot
respectively (see Figure 1-5). A high producer-to-injector ratio is preferred for
reservoirs containing high viscosity oil (>10 cP) while, for reservoirs containing
typical light oil, 5-spot and line drive can provide the shortest project life for a given
well spacing. Analytical models have been developed to help optimize pattern
selections including the impact of various mobility ratios and well skins. [Ref. 1-11].
10/49
5-Spot Pattern 7-Spot Pattern

Producer-to-Injector Ratio = 1:1 Producer-to-Injector Ratio = 2:1
9-Spot Pattern Linedrive Pattern

Producer-to-Injector Ratio = 3:1 Producer-to-Injector Ratio = 1:1
Figure 1-5 Various waterflood patterns

Ref. 1-11 Hansen, C.E., ”Producer/injector ratio: The key to understand
pattern flow performance and optimizing waterflood design”, SPE
75140, 2002
Table 1-1 Waterflood projects in the Shell Group and major historical
waterflood projects operated by Shell
Project Type of waterflood Status as of mid
2003
Asab, Bab, BuHasa, Abu Dhabi Flank/Pattern Ongoing
Plutonia Cluster, Angola Flank VAR 4
Bijupira-Salema, Brazil Flank Start-up
Seria West, Brunei Crestal Abandoned
Seria RE*A, Brunei Pattern Ongoing
Champion Main, Brunei Flank Ongoing
Champion South East, Brunei Flank/Pattern VAR 3
Dan, Denmark Pattern (pilot) Ongoing
Skjold, Denmark Flank Ongoing
Gamba/Ivinga, Gabon Mostly crestal Abandoned
11/49
Barton, Sabah, Malaysia Flank VAR 3
Kinabalu, Sabah, Malaysia ESP Dumpflood Ongoing
Schoonebeek, Netherlands Flank Abandoned
Egwa, Nigeria Flank Ongoing
Forcados Yokri, Nigeria Flank VAR 3
Bonga, Nigeria Flank Start-up
Draugen, Norway Flank Ongoing
Statfjord, Norway Flank Ongoing
Fahud Matrix, Oman Pattern Ongoing
Lekhwair, Oman Pattern Ongoing
Marmul, Oman Pattern Ongoing
Saih Rawl, Oman Pattern Ongoing
Yibal, Oman Pattern Ongoing
Zauliyah, Oman Pattern Ongoing
Astokh, Sakhalin, Russia Flank Start-up
Piltun, Sakhalin, Russia Flank VAR 4
Salym, Russia Pattern VAR 4
Omar Field, Syria Flank Ongoing
Tanak Field, Syria Flank Ongoing
Sirikit West, Thailand Flank Start-up
Brent, UK Flank Finished
Cormorant, UK Partly flank partly pattern Ongoing
Dunlin, UK Flank Ongoing
Auk, UK Flank Ongoing
Fulmar, UK Flank Ongoing
Mars Unit, USA Flank Starting
Cognac Field, USA Flank and Crestal Ongoing
Ventura Field, Aera, USA Pattern Ongoing
12/49
Belridge Diatomite, Aera, USA Pattern Ongoing
Wasson San Andres, Hobbs, Pattern Converted to CO2

Wasson Clearfolk, Little Creek
– USA - Sold
Cedar Creek Anticline, USA Pattern Ongoing
1.3.3 Well spacing

The profitability of a waterflood project is the ultimate deciding factor for the well
spacing. A short project life is economically attractive but has to be justified against
the increased capital costs of additional wells. Reservoir factors such as permeability,
heterogeneity, connectivity, and compartmentalization play an important role. In all
cases, conventional well stimulation techniques (acid stimulations, acid fracturing and
sand fracturing, etc.) should always be considered. Injection above the fracturing
gradient is a possible option to increase water injectivity and reduce the number of
injectors, but the risks of uncontrolled fracturing, reduced sweep efficiency and
injection loss into non-target zones should be weighed against the benefits of this
option.
Many of the Group's operated (or advised) waterfloods are in relatively permeable
geologically complex fluvial sandstone reservoirs where sand continuity is a problem.
In many of the waterfloods in the Eocene sandstones in Venezuela (ex Shell, now
Maraven), well spacings of 650 m were common. Well patterns were usually 7 spots,
which was also the well pattern for primary development. It was recognized that infill
drilling would be required in these areally heterogeneous reservoirs to recover
bypassed oil. Typical North Sea reservoirs such as Brent were developed by 2 rows of
flank water injectors each of which was targeted to a certain zone of the reservoir.
The spacing of the Brent water injectors was roughly 1000 m. Whilst this system
provided sufficient injectivity, continuously recompleting watered out producers to
improve the waterflood efficiency were also employed in this geologically complex
reservoir.
The waterflood in the Lekhwair field in Oman is in tight carbonate zones of the
Lower Shuaiba and Kharaib. The original well arrangement was a multiple 5 spot
with well spacings of 425 m. A classical calculation on how this well spacing could
be increased (and thus the number of wells reduced) by allowing a limited amount of
fracturing in the water injectors and opting for a 9 spot was reported [Ref. 1-12]. The
expansion of the Lekhwair waterflood from a pilot to full field implementation
(largely based on ideas described in Ref. 1-13), however, brought reservoir features to
light which had hitherto not been recognized, namely that faulting provided
preferential flow paths for the injection water. This was confirmed by 3D seismic and
horizontal appraisal wells in the flood area. The development pattern was modified to
a line drive by converting some producers to injectors and infilling producers along
the production line.
In summary, it can be said that either simple reservoir-engineering calculations or,
preferably, an integrated reservoir study can help to determine an initial estimate of
well spacing, but flexibility must be maintained to cope with unexpected reservoir
heterogeneities.
13/49
Ref. 1-12 Koning, E.J.L. "Waterflooding under fracturing conditions"

Doctoral Thesis, TU Delft, C161 88-001, 1988. (Electronic version
not available)
Ref. 1-13 Dikken, B. "Waterflood-Induced Hydraulic Fractures in Dan Field
Phase 1: Results of a Scouting Study." Report RKER.88.196
November 1988. (Electronic version not available)
1.3.4 Well type: the use of horizontal or multiple laterals

Advances in drilling technologies enable long (multi) laterals to be drilled relatively
fast and economically. The technologies have been applied extensively in several
Oman fields (e.g., Yibal, Al Huwaisah, Lekhwair, SaihRawl, etc.) but the results so
far have been mixed and, on the average, poorer than the initial expectations. The
benefit of using long laterals is to provide high initial production but the disadvantage
is the high sensitivity to reservoir factors such as high permeability streaks, fractures
and faults. The longer reservoir sections are drilled, the more likely it is that such
unfavorable zones will be encountered. The problem becomes even worse if long
horizontal wells are used for both injection and production. Furthermore, PLT
logging and well workovers are very difficult in long laterals so reservoir surveillance
and flood remediation are almost impossible in this situation. The low likelihood of
proper waterflood management will certainly result in low waterflood recovery.
Based on these results, lateral technologies should only be applied with care. The
following generic recommendations should be considered:
1. Understand the reservoir architecture well enough to ensure low likelihood of
short-circuits between wells.
2. Build enough heterogeneity in the simulation models and don’t rely on simple
analytical models.
3. Consider drilling short laterals.
4. Consider a well spacing much larger than the lateral length.
5. Include future monitor logging and work-over in the well design.
6. Consider using smart well technology to improve the ability to manage the
flood.
Ref. 1-14 discusses situations where multilaterals can improve the economics over
horizontal wells in waterfloods.
Ref. 1-14 Economic Comparisons of Multilateral and Horizontal Wells in
Water-Drive Reservoirs, Nestor Rivera, SPE 80927.
1.3.5 Reservoir souring

Reservoir souring, which is the process of generation, migration and production of
H2S due to seawater injection, has manifested itself in oilfields for decades. On a
global basis, a significant proportion of waterfloods have turned initially sweet
reservoirs sour. Different theories on souring mechanisms exist, but industry
consensus has it that microbiological processes, attributed to SRBs (Sulfate Reducing
Bacteria), taking place inside the reservoir are responsible for souring.
14/49
H2S generated in previously sweet fields poses a severe threat, particularly when its
generation has not been foreseen. H2S levels in fields that have gone sour range from
5 to thousands of ppm - causing major concerns on facility integrity, quality of sales
products and HSE aspects. Within the Shell Group, reservoir souring has become a
fact of daily life in several oil fields in the North Sea and the Gulf of Mexico. The
emergence of significant H2S levels, however, invariably drives up production costs
and reduces revenue, and hence is certain to have a negative business impact.
Whilst there is consensus on souring being a biogenic phenomenon, the actual
microbiological processes that take place are still subject to lively, ongoing debate. It
is broadly accepted, that stimulating SRB colonization requires:
1. Sulfate (electron donor, present at very high levels in seawater).
2. Carbon energy source (e.g. Volatile Fatty Acids, VFA).
3. Nutrients for reproduction (e.g. phosphorus, nitrogen, trace elements).
4. Suitable temperature regime (reservoir temperatures from 35 to 95 oC)
In the absence of universally accepted and/or proven prediction models, risks must be
identified and steps taken to manage potential problems. Natural factors that will limit
bacterial growth are high formation water salinity and high reservoir temperature. In
order to provide controls for managing the souring that may eventually take place,
mitigation options must be considered. The following technologies exist:
1. Kill the SRB’s by applying effective biocides.
2. Reduce sulfate levels in feed water.
3. Introduce feedstock (nitrates) to allow competing micro-organisms to flourish.
If the reservoir sours, then measures must be taken such as injecting H2S scanvenger
downhole in the producers and/or change out to sour service completion material.
Reservoir souring is an important issue in waterflood. Detailed descriptions of this
topic can be found in Section 2.4.3.
1.3.6 Injection under fracturing conditions

Fractured water injectors are quite a common feature in many waterfloods. They may
be created intentionally through well stimulations or unintentionally due to pore
plugging by impurities in the injection water or due to thermal effects.
There is ample evidence available for the presence of waterflood-induced fractures in
waterfloods worldwide. In the North Sea, formation breakdown gradients as low as
0.65 psi/ft have been recorded after prolonged periods of water injection (due to
thermal effects). NAM's de Lier flank water injection project showed evidence of
long fractures communicating directly between injectors and certain producers. The
causes of fracturing in the Lake Maracaibo water injection project has been ascribed
to plugging by suspended solids in the injection water.
Many floods now intentionally employ fracturing conditions (i.e., operating the
injectors above the fracture propagation pressure) as part of the field development
plan, since the increased injection potential may enable a significant reduction in the
number of injectors drilled. Furthermore, water purity requirements may be relaxed
providing a significant reduction in facilities costs. These benefits are particularly
15/49
attractive in offshore operations where well slots and platform space are limited and
the costs of well workovers are very high.
However, when injecting under the fracturing conditions, once fracture lengths reach
a significant fraction of the well spacing, there may be a significant influence on
sweep efficiency. Furthermore, if the fracture does not extend over the full height of
the formation, non-uniform sweep in the vertical sense may be encountered, which
may alter field development plans. Therefore, an accurate prediction of fracture
length and height is required. In Section 2.5.3, basic physical processes behind
fracturing are outlined, currently available technologies to predict 3D fracture
extension are noted and data collection requirements are discussed.
In designing a waterfood project it is important to be aware of the uncertainties of
fracture predictions. Despite recent advances in the simulation technology, fracture
simulations are very sensitive to the uncertainties of reservoir geology, rock
mechanical properties and in-situ stress conditions, as the fracturing process itself is
inherently unstable. It should be remembered that the goal is to flood oil residing in
the reservoir (there is very little oil in the fractures), so the presence of fractures could
impose a negative impact on waterflood recovery. Waterflood throughput can be
increased by an injector fracture (negative skin) but the same effect can be achieved
by a conventional fracture stimulation process rather than relying on injecting above
the fracture gradient. The cleanup costs saved by injecting above the fracture
gradient may not be significant in the on-land operations – a simple bull-head of acid
and solvent down the wellbore or coil tubing to clean out the fills/fragments in the
injector bottom hole.
Detailed descriptions on basic physics of matrix and fractured injection can be found
in Section 2.5.2 and Section 2.5.3, respectively. A generic procedure for selecting an
appropriate water injection scheme and its associated surveillance programmes are
described in Section 6.2.1. Relative merits of three possible injection schemes,
namely, matrix injection, injection below fracture propagation pressure, and injection
above fracture propagation pressure, are described in Sections 6.2.2, 6.2.3, and 6.2.4,
respectively.
1.3.7 Waterflood pilot

Waterflood pilots are conducted to allow improved quantification, and thus reduction,
of risks for a larger-scale project where expenditure will be much greater. Major
uncertainties in a waterflood development are related to reservoir geology:
drainability, floodability, anisotropy, heterogeneity, fractures, faults, etc. Historically,
many waterflood pilots (e.g., Lekhwair, Fahud and Al Huwaisah) were conducted
mainly for this purpose. The Al Huwaisah water injection pilot in the southwestern
area of the field employs two water injectors, converted from two existing vertical
producers and re-completed in the water leg. After more than one year of water
injection, significant increases in oil production indicated the stability of the flood
fronts and supported the credibility of simulator predictions based on latest geological
realizations of the reservoir. Results of the pilot helped to justify additional water
injection in the water leg and retain horizontal wells already drilled near the top of the
reservoir as producers.
Frequently, water injectivity, long-term deliverability of the water sources, and water
treatment are also included in the pilot scope to minimize surprises in the full-scale
16/49
implementation. The high, sustained water injectivity (nearly 9000 bwpd per well) of
the Al Huwaisah pilot indicated the presence of high permeability rock and / or the
abundance network of natural fractures to effectively distribute water areally and thus
eliminated the need to adopt a more aggressive fractured injection policy.
For general discussions of the economic values of a waterflood pilot, please refer to
Section 7.2.1 of the manual.
1.3.8 Water-alternating-gas injection

Water-alternating-gas (WAG) injection is an attractive option to increase oil recovery
above what can be achieved by waterflooding alone. When injected at sufficiently
high pressure, the injected gas is miscible with the oil. Field results indicate that
incremental recovery by the miscible WAG process can be 10% of the original oil in
place or higher. The miscible WAG injection process has been successfully applied
in many fields, as a tertiary or even as a secondary recovery process applied right after
primary depletion.
At low pressure, gas is not miscible with oil. As such the injected gas forms a separate
phase from the oil phase, bypasses the oil phase (due to high gas mobility), and
quickly reaches the producer. The displacement efficiency of an immiscible WAG is
therefore much lower than the miscible WAG, although the benefits for an immiscible
WAG can come from opportunities to improve the macro sweep efficiency due to the
architecture of the flooded reservoir. A recent study predicted improved vertical
sweep (above waterflooding only) by the immiscible WAG injection in the Al
Huwaisah reservoir where the gas was predicted to sweep the region bypassed by
waterflooding due to water coning [Ref. 1-15].
Whether the WAG process is economically attractive depends largely on the value of
the injected gas. Under most situations, development costs (per barrel of oil
recovered) of a WAG process is higher than waterflooding, but there are special
situations when WAG injection could be attractive:
1. Hydrocarbon gas stranded due to transportation costs and lack of local gas
market, e.g., Prudhoe Bay field of Alaska USA and Salym field in West
Siberia Russia.
2. Produced gas containing high concentrations of CO2 making the gas less
valuable, e.g., the giant K5 gas reservoir (60% CO2) in Malaysia.
3. Operating policy emphasizing high ultimate oil recovery rather than short-
term production, e.g., ADCO at Abu Dhabi.
4. Presence of pure CO2 sources, e.g., McElmo Dome and Bravo Dome in USA.
With increasing emphasis on limiting the emissions of green house gases to the air,
water-alterating-CO2 injection will become more attractive in the future.
For general references to the miscible WAG process, a SPE monograph by Stalkup
may be consulted [Ref. 1-16]. Many published papers on CO2 projects in West Texas
USA may also be consulted. A general methodology for scouting WAG potential in
a typical reservoir is also reported [Ref. 1-17]
Ref. 1-15 Mollinger, A.M., ”Water alternating gas development options for
Al Huwaisah”, EP 2002-5052, 2002
17/49
Ref. 1-16 Stalkup, F.I., “Miscible displacement”, SPE monograph series,

volume 8, C173 84-001. (Electronic version not available)
Ref. 1-17 Hsu, C.F. and Keating, J.M., ”A scoping evaluation of the miscible
potential in the Al Huwaisah field, North Oman, EPNL Special
Issue - EP Solutions, 2003
1.3.9 Other EOR considerations

It is important to consider all possible EOR options as an alternative to waterflood, or
as a follow-up option, before embarking on a waterflood development. This
consideration will ensure maximization of oil recovery from a reservoir. In other
words, the life-cycle development plan for the reservoir needs to have been
considered before committing to a waterflood development. If any future
development option has to be compromised for the waterflood decision, present
values and possibilities of success of the lost opportunities have to be evaluated.
Figure 1-6 illustrates various recovery methods applicable to various stages of the
field development. Note that, in some situations, it may be attractive to skip the
secondary recovery and go straight to a tertiary recovery - yielding an accelerated
recovery from the reservoir. The incremental recovery from a miscible gas process
is usually independent of the start-up time of gas injection, as indicated in many
miscible projects of the West Texas USA. However, the project profitability may be
higher on the present-value basis if the project is implemented earlier on, provided
enough reservoir data has been collected from the waterflood performance. Delayed
applications of steam injection after waterflooding, nevertheless, could be very
unattractive because of the additional heat required to vaporize the large amount of
injected water. Normally, it would be prudent to skip waterflood when considering
steam injection. This practice has been applied to many heavy oil reservoirs in
California, USA.
18/49
Oil
Oil Re
Recove
covery
ry Me
Methods
thods
Primary
Primary Recovery
Recovery
Natural
Natural Flow
Flow Artificial
Artificial Lift
Lift
Conve ntiona l
Re cove ry
Secondary
Secondary Recovery
Recovery
Pressure
Pressure
Waterflood
Waterflood Maintenance
Maintenance
Tertiary
Tertiary Recovery
Recovery
Enha nce d Oil

Re cove ry
The
Therma
rmall Ga
Gass Che
Chemica
micall
Inje
Inje ction
ction
Injection
Figure 1-6 Oil recovery methods applicable to various stages of the reservoir
development
Candidate EOR Processes Additional Factors to Consider

Under Various Reservoir Conditions Selections of Candidate EOR process
• Bottom Water Injection

Oil Viscosity (cp) – For high oil-column height (>100 m) w/o strong aquifer
1 10 100 1000 10000 100000 – Less sensitive to natural fractures and reservoir heterogeneity
0 – Ultimate RE about 50%
Immiscible
• Pattern Water Injection
Gas Injection
– Sensitive to fractures and reservoir heterogeneity
2000 Steam Injection
– Ultimate RE up to 50%
• Immiscible Gas Injection
4000
– For high oil-column height (>100 m) w/o strong aquifer
Reservoir Depth (ft
– Could be sensitive to heterogeneity for layered reservoir

– Natural fractures could help the process
6000 – Ultimate recovery > 50%
Miscible CO2 Pattern Water Injection • Steam Injection
or HC Gas or Bottom Water Injection – For high permeability reservoirs (> 500 mD)
Injection
8000 – For relatively thick gross interval (> 20 m) with good N/G (>0.6)
– Natural fractures could help the process
Miscible – Ultimate RE > 70%
10000
Nitrogen • Miscible Injection Processes
Injection – Very sensitive to reservoir heterogeneity and fractures
12000
– Ultimate RE > 60%
Figure 1-7 Appropriate reservoir conditions for applying various recovery

methods
Reservoir conditions suitable for applying various recovery methods are summarized
in Figure 1-7. These EOR processes are the subjects of many published books and
literature [Ref. 1-18 and Ref. 1-16].
Ref. 1-18 Prats, M., “Thermal Recovery”, SPE monograph series, volume 7
(No electronic version available).
19/49
1.4 Waterflood Oil Recovery Predictions

1.4.1 General guideline
The economic success of a waterflood project depends on the additional oil recovery
it can achieve relative to increased costs over primary development. Historically,
classical techniques, such as those developed by Buckley-Leverett, Dietz, Stiles, and
Dykstra-Parsons, were frequently used to predict waterflood recovery. While these
analytical techniques/theories provide important insights to the process of
waterflooding they include virtually no geology. Currently, the preferred predictive
method is an integrated full-field study incorporating all petroleum-engineering
disciplines, facilitated with the state-of-the-art computers, geological modeling
software, and fast dynamic reservoir simulators as described in Section 1.4.5.
Section 1.4.2 describes the crucial fluid data and rock properties needed for
waterflood predictions. The basic fluid properties of oil, gas and water can be
obtained from a standard PVT analysis. Key rock properties for waterflood
predictions are the relative permeability and capillary pressure. Sections 1.4.2.2.6,
1.4.2.2.7, and 1.4.2.2.8 describe the appropriate process of preparing rock plugs
(aged) and the preferred measurement techniques (steady-state and/or centrifuge) to
obtain good relative-permeability measurements. These three sections are included
for those who have a specific interest in relative permeability data and the appropriate
measurement techniques to obtain these data.
Section 1.4.3 summarizes a number of classical techniques. The section is prepared
for reservoir engineers who are interested in the fundamentals of fluid displacement
(microscopic) and sweep (macroscopic) in simple geology. Although most of these
techniques are dated and of limited applications, an in-depth understanding of these
techniques can help reservoir engineers to understand the physics of waterflooding
which is essential for successful use of reservoir simulators.
Section 1.4.4 summarizes the techniques of scoping dynamic simulation study. While
3D full-field simulations with sufficient geological details are always preferred, there
are situations that quick dynamic simulations encompassing simplified geology might
also be helpful. A 2D cross-sectional simulation may be used to simulate the channel
architecture (Section 1.4.4.1) with enhanced geological complexities in the vertical
cross-section. Such a model is valuable for predicting the impact of channel
architecture on the sweep, as exhaustive 3D simulations with enough geological
details might be impractical. In Section 1.4.4.2, 2D models are constructed to model
waterflooding in typical turbidite sand packages, yielding “dynamic pseudo relative
permeability curves” for use in the 3D simulations. Section 1.4.4.3 discusses the use
of 2D areal simulations and their applicability using either streamline or conventional
simulators to quickly identify areas poorly swept by the waterflood. To study the
impact of induced fractures on waterflood performance, simulations resembling
idealized patterns are frequently employed (see Section 1.4.4.4).
In summary, an integrated study involving all petroleum-engineering disciplines is
the best approach to predict waterflood recovery. Within the scope of the integrated
study, 3D full-field dynamic simulations are always preferred although dynamic
simulations with limited geological details can also add value. In all cases, reservoir
engineers require a good understanding of the classical techniques/theories of
waterflooding to judge the credibility of the forecasts.
20/49
1.4.2 Basic properties

1.4.2.1 Fluid properties
Oil fluid properties needed for the prediction of waterflood recovery include oil
viscosity (µ), oil formation volume factor (Bo), solution gas-to-oil ratio (Rs), and the
gas formation volume factor (Bg) at average reservoir condition of waterflooding. For
reservoirs that have produced through primary depletion, variations of these
properties with pressure are needed to establish the reservoir conditions at the on-set
of the water flood. Typically, these data are needed in the PVT input to a black-oil
reservoir simulator. In addition, the water density and viscosity at reservoir
conditions are also needed.
1.4.2.2 Rock wettability, relative permeability and capillary pressure
This section describes the rock and fluid properties key to waterfood predictions.
They include wettability, relative permeability, and capillary pressure. There are a
significant number of publications in the open literature that address this topic area.
For example, Anderson [Ref. 1-19] published a series of SPE review papers between
1986-87, which discuss wettability and its impact on petrophysical and fluid flow
properties. However, much of the existing literature no longer reflect the more recent
knowledge and experience, particularly with regard to the development of mixed-
wettability in oil reservoirs and its impact on relative permeability and imbibition
capillary pressure curves. This section aims to provide a summary of the current
knowledge and Shell Group best practices.
The prediction of waterflood displacement efficiency requires an accurate knowledge
of basic flow parameters, such as oil/water relative permeabilities, capillary pressures,
and residual oil saturation. A multi-disciplinary integrated approach is required in the
core sample selection process to cover all predominate reservoir rock types. X-ray CT
scanning on whole core and core plugs is recommended for sample screening and
selection. Group recommended procedures should be applied in sample selection,
preparation, wettability restoration, flow experiments and data interpretation.
1.4.2.2.1 Wettability
Wettability describes the relative preference of a rock to be covered by a certain
phase. Rock is defined to be water-wet if the rock has (much) more affinity for water
than for oil. In that case, a major part of the rock surface in the pores will be covered
with a water layer. Clearly, wettability will be affected by the minerals present in the
pores. Clean sandstone or quartz is extremely water-wet, but sandstone reservoir rock
is usually found to be mixed-wet. Mixed wettability means that some pores are water
wet and other pores are oil wet. Carbonates are found to be more oil-wet than clastics:
a major part of the rock surface in the pores is then covered with oil. It should be
stressed that in practice extreme water-wetness or extreme oil-wetness is rare.
Basic reservoir properties like relative permeability, capillary pressure and resistivity
depend strongly on wettability. It is therefore important that laboratory experiments in
which these properties are measured are carried out on samples whose wettability is
representative of the reservoir from which they are taken. Since no well-established
techniques exist at present for downhole measurement of the wettability of reservoir
21/49
material, it is not possible to state categorically whether a certain wettability is

representative in a given case [Ref. 1-20].
1.4.2.2.2 Residual saturations

As mentioned above, in an extremely water-wet rock, the surface is covered with
water. In that case oil (or gas) will be located in the centre of the pores. In the
laboratory, when water-filled water-wet rock is brought to irreducible or connate
water saturation (Scw) by primary drainage, this water will remain a continuous phase
covering the pore walls. During a subsequent water-drive to produce the oil from
water-wet rock, a significant amount of oil eventually will remain capillary trapped,
floating as disconnected blobs in the centre of the pores (Figure 1-8). The residual oil
saturation Sor is determined by the topology of the pore space and is usually high:
around 30% and up. Conversely, if the system is strongly oil-wet, then a low
theoretical residual oil saturation of as low as 10% or less are expected due to film
flow (Figure 1-8). Real reservoir crude oil/ brine/rock systems generally fall in
between the above extreme water-wet and oil-wet cases and show a mixed wet
behaviour [Ref. 1-20, Ref. 1-21].
1.4.2.2.3 Capillary pressure

Capillary pressure pc is defined as the pressure difference between the non-wetting
phase and the wetting phase as a function of the (wetting phase) saturation. In
reservoir engineering of waterflooding, pc, particularly the imbibition capillary
pressure is of importance. Note that an Hg-air measurement will result in the primary
drainage curve and that this data is only used in the initialisation of a simulation
model (although there are special cases where imbibition capillary pressure may be
needed in initilisation). Figure 1-9 shows typical imbibition capillary pressure curves
for a water-oil system in a porous rock of different wettabilities from water-wet,
mixed-wet to oil-wet. The full capillary pressure curve consists of three branches: a
primary drainage, a primary imbibition and a second drainage branch. In
waterflooding, imbibition capillary pressure is in general needed as input to reservoir
simulations to model water displacing oil. In tight oil-wet carbonate reservoirs,
imbibition capillary pressure is generally expected to play a crucial role in waterflood
recovery performance.
22/49
AT END OF IMBIBITION, OIL IS TRAPPED AS ISOLATED BLOBS IN LARGER PORES. HIGH Sor, LOW kw
OIL REMAINS CONNECTED AND CAN DRAIN TO LOW SATURATION. WATER OCCUPIES LARGEST PORES.
LOW Sor, HIGH kw
Figure 1-8 Residual oil distribution for a water-wet (top) and a mixed-wet
(bottom) system
23/49
Wettability - Effect on imbibition capillary pressure
Pc
Water-wet
Mixed-wet
0 1
Sw
Oil-wet
Figure 1-9 Effect of wettability on water/oil imbibition capillary pressure
1.4.2.2.4 Relative permeabiltiy

Relative permeability curves are commonly fitted with the following Corey equations:
no
 1 − S w − Sor 
kro ( S w ) = kro ( Scw ). 
 1 − Scw − Sor 
nw
 S − Scw 
krw ( S w ) = krw ( Sor ). w 
 1 − Scw − Sor 
where Sor and Scw are the residual oil saturation and the connate water saturation
(fraction of pore volume), kro(Scw) and krw(Sor) are the end-point relative
permeabilities to oil and water, no and nw are the so-called Corey exponents for oil
and water, respectively.
The Corey model proves to be adequate to fit most experimental data. It has six
parameters, four defining the relative permeability curve limits, and two giving the
curvature between the end-points (i.e. the shape of the curves). Figure 1-10 shows the
impact of wettability on water/oil relative permeability and imbibition capillary
pressure characteristics.
24/49
Wettability & relative permeability
water - wet
kro 1
water-wet
10-1
oil-wet,
mixed-wet
10-2
oil wet 10-3

Sor
10-4
10-5
0 1
Sw
Figure 1-10 Effect of wettability on water/oil relative permeability
1.4.2.2.5 Impact on waterflood recovery predictions

In the past (pre-1995) relative permeabilities were measured in the laboratory on
cleaned (unaged) water-wet samples. However, some 10 years ago it was accepted
that reliable reservoir relative permeabilities can only be measured if the in-situ
wettability is properly mimicked in the laboratory, by aging of the samples with crude
oil, prior to the actual measurements [Ref. 1-20]. Hence, in the more recent
experiments (post-1995), cleaned reservoir cores were first saturated with brine and
then drained to the initial water saturation using crude oil. Then, the cores were aged
under reservoir temperature over a period of time (typically 4 weeks) to restore
reservoir wettability. As shown in Figure 1-10, old water/oil relative permeability and
imbibition capillary pressure measurements performed on cleaned, unaged samples do
not give representative results that reflect reservoir wettability. The following Table 2
lists some Group field examples of waterflood residual oil saturations comparing new
data (based on wettability restoration procedures) with old (water-wet on cleaned
cores) [Ref. 1-20, Ref. 1-21]. It should also be recognized that, although new data
suggest lower residual oil saturations, relatively higher pore-volumes of water
injection are needed to achieve high recoveries due to increased Corey exponents of
the oil relative permeability associated with mixed/oil wettability as shown in Figure
1-10.
As more and more special core analysis measurement are carried out at representative
reservoir conditions, it becomes clear that reservoir rocks which are non-water wet
and thus have low residual oil saturations to water, are more common than hitherto
thought. Some typical examples include the Brent and Draugen fields where very low
residual oil saturations were measured. Moreover, in many ongoing waterfloods
where the ideas regarding residual oil saturations have changed over the past years,
volumetric sweep efficiencies will have to be reassessed (lower residual oil
saturations means less volumetric sweep, i.e. more bypassing).
25/49
Field Examples of Lower Residual Oil Saturations
Current Old
Brent 15% 28%
Dunlin 15 25-30
Schiehallion 14 29
Draugen 10-15 -
Rabi 20 -
Nimr 15-25 -
Amin 10-20 -
San Francisco 10 >=40
Lekhwair 5 28
Maui 10 28
Table 2 Field Examples of Lower Residual Oil SaturationsPreferred SCAL

measurement technique
It is important to apply the best measurement technique available. For instance, many
measurements in the past employed the so-called Welge, or unsteady state, technique.
Recent work has shown that this technique is highly inaccurate and nearly always
produces too high residual oil saturations because of insufficient flooding. The
centrifuge technique can overcome this problem. Within Shell, a combination of
steady-state with in-situ saturation monitoring (for mid-saturation range) and
centrifuge (for end-point saturation range) techniques is recommended to determine
water-oil relative permeabilities, including the residual oil saturation and the shape
of the saturation functions. The measurements should be carried out on core plugs
under "imbibition" (displacement of oil by water down to residual oil saturation), but
measurements under first "drainage" (displacement of water by oil down to
irreducible water saturation) and second "drainage" (displacement of the injected
water by oil) may also be required. Analytical data interpretation procedures are often
found inadequate to derive relative permeability and capillary pressure curves from
SCAL experiments. Numerical simulation is recommended to unravel the combined
effects of relative permeability and capillary pressure and derive the multiphase flow
properties consistently while honouring the actual experimental conditions [Ref. 1-20,
Ref. 1-21]. Special experimental design is needed for more complex displacement
processes, such as WAG that involves 3-phase relative permeability and saturation
hysteresis effect.
1.4.2.2.7 Core sample preparation

It is important to carry out these measurements as close to reservoir conditions as
possible, representative of original wettability of the reservoir rock/fluid system. This
can be approached in two ways. First, the tests can be done on native state
(preserved) cores. Second, tests can be done on core samples, which have first been
cleaned and then restored to original conditions. The so-called restored state
technique consists of cleaning the core samples, saturating them using brine and crude
26/49
oil and subsequently ageing them. Ageing means keeping the cores in contact with
crude and brine for periods lasting for several weeks. Within Shell, the restored state
technique by ageing is preferred, as native state cores may have been inadequately
preserved (limited amounts of oxygen may have entered the core and oxidation of
certain minerals or constituents of the crude may have altered the wettability of the
core).
To judge the wettability of a core, two wettability indicators can be derived from the
capillary pressure tests. The first one is the so called modified United States Bureau of
Mines (USBM) index, defined as the logarithm of the ratio of the areas under the
secondary drainage and imbibition capillary pressure curves. Typically, a strongly
water wet core has a USBM index greater than one, whilst a strongly oil wet core has
a USBM index less than minus one. A second wettability indicator is the so-called
AMOTT number. The AMOTT number for water is defined as the ratio of water
entering the core spontaneously to total water invasion during the "imbibition" cycle.
Similarly, the AMOTT number for oil is the ratio of the amount of oil entering the
core spontaneously to total oil invasion during the second drainage cycle (shown in
Figure 1-11). More details on the actual equations and descriptions can be found in
[Ref. 1-25]. The NMR wettability index is still in the research phase (confidential),
and no detail can yet be given.
Wettability - measurements
Wettability indices Pc
z USBM ( 10Log(A/B) )
z Amott ( x/y )
A Sw
No down-hole measurement
technique available 0 B 1
z Core preservation
x
z Wettability restoration
Clean to water-wet state
Age in crude for several weeks
Figure 1-11 Wettability measurements using the AMOTT and USBM methods.
Although the centrifuge method is a proven technique for measuring capillary
pressures (except in the case of imbibition in strongly water wet rock), the so-called
CAPRICI technique introduced by Shell offers a powerful alternative. In CAPRICI
the core sample is placed between an oil and water wet membrane, thus enabling the
continuous recording of all branches of the capillary pressure curve by first flooding
with oil and then by water (shown in Figure 1-12). By monitoring continuously the
drainage and imbibition capillary pressure curves and the corresponding resistivity
indexes during the ageing process, it is possible to assess how close the wettability
restoration approaches equilibrium.
27/49
CAPRICI: technology to assess the degree of

wettability restoration
140
120
Continuous injection of 100
reservoir fluids
Po-Pw (kPa)
80
60
40
20
oil-wet
filter 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Brine saturation
∆P core 100 1
plug 10-1
80
10-2
60
10-3
water-wet
10-4
filter 40
10-5
20
10-6
0 10-7
0 0 .2 0 .4 0 .6 0 .8 1 0 0 .2 0 .4 0 .6 0 .8 1
brine saturation brine saturation
Figure 1-12 CAPRI for measuring complete capillary pressure loops and assess
the degree of wettability restoration during ageing
As a typical example of measuring wetting indicators, Figure 1-13 shows some results
of wettability indices. In this work, centrifuge measurements were carried out on
"native state" (as received)) cores, using brine and crude oil; cleaned cores, using
brine and a refined oil; and "restored state" cores after ageing, again using brine and
crude oil. It indicates that the "native state'' cores were oil wet and that, after a
thorough cleaning procedure the cores containing brine and refined oil has become
largely water wet. Ageing changes the conditions of the cores back to oil-wet. Figure
1-14 shows an example of Brent relative permeability measured following the
recommend procedures of wettability restoration by ageing versus the “old’ data
obtained on cleaned and water-wet cores.
1.4.2.2.8 Proper coring procedure

Since coring is a critical first step in any basic and special core analysis, whenever
possible, coring and core analysis specialists in Shell should be consulted on the
proper procedures and recommended contractors for taking cores from the reservoir
of interest, in particular with regard to unconsolidated core materials. Several
documentations on the best practices may also be consulted [Ref. 1-23].
28/49
Effect of cleaning
Amott - water flood Amott - oil flood
1
0.9
0.8
0.7
Amott indices
0.6
0.5
0.4
0.3
0.2
0.1
0
2855 2869 2900 2922 2855 2869 2900 2922 2855 2869 2900 2922
Copyright 2001 SIEP B.V.

As received Cleaned Aged
Shell Learning
Figure 1-13 Comparison of wettability indices
Relative permeability for a Brent rock type

(solid lines - after wettability restoration; dotted lines - cleaned and water-wet )
0 .1
0 .0 1
0 .0 0 1
0 .0 0 0 1
0.15
0 .0 0 0 0 1
0 .0 0 0 0 0 1
0.28
0 .0 0 0 0 0 0 1
0 .0 0 0 0 0 0 0 1
0 0 .2 0 .4 0 .6 0 .8 1
W a te r s a tu r a ti o n ( fr a c ti o n )
Figure 1-14 Typical relative permeability curves for aged Brent cores compared
to the corresponding cleaned and water-wet “old” data.
1.4.2.2.9 Database and correlations

Accurate relative permeability and capillary pressure data are essential for almost all
reservoir engineering calculations. Unfortunately, there is often a lack of reliable
experimental data. This can either be due to the general unavailability of core
material, for example in exploration, or to unreliable core measurements and/or
interpretation. General correlations based on analogue databases can be useful for the
following purposes [Ref. 1-25, Ref. 1-26]: to provide data in case no core data is
29/49
available (yet) for the field under investigation, to provide a quality check of core
measurements with respect to expectations based on the correlations, and to provide a
range of possible values for sensitivity studies.
A spreadsheet correlation for predicting the capillary pressure and relative
permeability of unconsolidated sands has been developed by Doe [Ref. 1-22, Ref.
1-24]. The predictive correlations are derived from more than 160 centrifuge
measurements performed in Houston, USA during the period of 1993 through 1997.
Half of these data are US Gulf coast unconsolidated sands. A few steady-state
measurements are also included in the database. The method includes an option to
adjust relative permeability curves for consolidated sands. The spreadsheet model can
be quite useful for making intelligent estimations when actual measurements are
lacking, or for confirming the measured data when its validity is in question.
More recent work in Rijswijk [Ref. 1-25, Ref. 1-26] has led to the compilation and
review of sandstone and carbonate relative permeability databases of Shell Group for
a number of reservoirs. For carbonate reservoirs, there was no clear correlation of
relative permeability vs. porosity and permeability, and an average set of Corey
parameters with the range of uncertainties was given [Ref. 1-26]. This database is
being updated though, and correlations may still be found when more data has been
added. For the sandstone reservoir rock samples, obtained on “aged” samples (i.e.
with restored wettability) and obtained by centrifuge/steady-state techniques, the
special core analysis data were subjected to a careful review to identify consistent and
reliable data. Moderate correlations have been found between Corey parameters and
permeability over porosity, for the higher permeability (K > 200 mD) sandstones. The
correlations were made based on an average mixed-wet characteristic for the
sandstone dataset since direct wettability measurements were generally not available
for the available database [Ref. 1-25]. Further insight should be obtained in the pore-
scale physics behind the relationships between relative permeability and rock/fluid
parameters from the ongoing Pore Network Modelling research [Ref. 1-27].
In summary, relative permeabilities should be measured in the laboratory under
representative reservoir conditions, using recommended current best practice
procedures [Ref. 1-20, Ref. 1-21]. In particular, the samples should be “aged”, i.e.
restored to in-situ wettability, a combination of centrifuge and steady-state equipment
should be used (i.e. unsteady-state measurements are deemed unreliable), and
corrections should be made numerically for possible measurement artefacts, such as
possible capillary end-effects and/or too short measurement time. In the absence of
reliable SCAL data, database and spreadsheet correlations provide a useful alternative
to gain an initial estimate of relative permeability trends and uncertainties. It is
important to note that a significant range of uncertainty exists for this type of
empirical correlations. Therefore, these predictions should be used with some caution.
They are best used in initial sensitivity studies, and should be checked whenever
possible against reliable laboratory SCAL measurements for the specific field under
investigation.
Further information on SCAL data acquisition and applications can be found via the
Shell team website: sww.siep.shell.com/CoReS.
Ref. 1-19 Anderson, W.G., “Wettability literature survey“, SPE 16471, JPT
1987.
30/49
Ref. 1-20 Boom, W., Boutkan, V., Frens, A.M., Kokkedee, J.A. and Maas,
J.G.: ”Manual course PW15: Interpretation of Special Core
Analysis Data”, RKMR.95.081. (Electronic version not available)
Ref. 1-21 Jing, X.D., Doe, P.H., Smits, R.M.M “SCAL - Measurement,
Interpretation and Application of Special Core Analysis Data”,
Shell Open University Course Manual. (Electronic version not
available)
Ref. 1-22 Brown, D.C. and Doe, P.H., “Relative Permeability Refresher”,
PT3360, Phase III, Petroleum Engineering, Technical Training
Series. (Electronic version not available)
Ref. 1-23 Rosen, R.L., Schipper, B “Unconsolidated core analysis best
practices: you get one chance to get it right”, Link to
http://sww.sww.siep.com/CoReS.
Ref. 1-24 Doe, P.H. “Relative permeability spreadsheet”
Ref. 1-25 Smits, R.M.M., Jing, X.D. “A relative permeability database for
water/oil imbibition in sandstone reservoirs’, EP 2003-5041, in
preparation. (Electronic version not available)
Ref. 1-26 Smits, R.M.M., “A database for relative permeabilities of carbonate
reservoir rock samples”, EP 2001-5484.
Ref. 1-27 Valvatne, P.H., “Predictive pore-scale network modeling”, EP
2003-5148.
1.4.3 Analytical techniques for waterflood prediction

Historically, many analytical techniques were developed for waterflood predictions.
Classical techniques developed by Buckley-Leverett, Dietz, Stiles, and Dykstra-
Parsons and physical situations that these techniques are applicable to are described in
[Ref. 1-29], [Ref. 1-2], and [Ref. 1-3]. The readers are encouraged to consult these
references to learn the essential background of waterflood theory. Key results of
these classical theories are described in Section 1.4.3.1, Section 1.4.3.2, and Section
1.4.3.3. Within Shell, a number of analytical and semi-analytical techniques have
been developed to improve upon these classical theories: Buckley-Leverett model
with gravity effect Section 1.4.3.4, improved Dietz model with saturation transition
Section 1.4.3.5, and Stiles model applied to 5-spot patterns Section 1.4.3.6.
1.4.3.1 Buckley Leverett Model
The Buckley-Leverett model was developed to predict one-dimensional frontal

displacement of oil by water. The model is very useful for predicting the microscopic
displacement efficiency of waterflooding including the effect of relative permeability
and fluid viscosity. Water-cut at water breakthrough and subsequent increases in
water-cut can be predicted using the so-called “fractional flow curves” constructed
from the relative permeability and fluid viscosity. The oil recovery as a function of
the cumulative water injected can then be predicted. Although the Buckley-Leverett
model is best suited for the small-scale applications, it is also a good approximation of
waterflooding in a moderately permeable reservoir containing relatively light oil
31/49
where the reservoir layer is thinner than the capillary transition zone and the flow
instability due to viscous fingering is not present. Key findings from this model are:
(i) While the residual oil saturation to the waterflood (Sorw) determines the ultimate
recovery by waterflooding, the shapes of the relative permeability curves have a big
impact on the speed and the economics of the process. The higher the Corey
exponent of the oil relative permeability (e.g. an oil-wet reservoir), the more pore
volume of water injection is needed to achieve ultimate recovery. (ii) All other
conditions being equal, the higher the oil viscosity, the higher the water cut at
breakthrough and the slower the oil recovery, resulting in a large amount of water
recycling even in a relatively homogeneous reservoir (see Figure 1-15).
Short Transition Length
Light
Water
Oil
Residual So
Initial So
To Waterflood
Water Heavy
Oil
Long Transition Length
Figure 1-15 Buckley Lverett model showing the effect of oil viscosity on
waterflood displacement efficiency
1.4.3.2 Dietz Model
In the case of small capillary pressures such as in a high permeability, sandstone

reservoir, gravity force tends to segregate oil from injected water in most part of the
reservoir. Dietz has investigated this type of displacement and, in particular, the
conditions under which the displacement is stable. He has developed a formula for
calculating the tilt angle of the displacement front relative to the bedding of the
reservoir as a function of the dimensionless gravity number, dip-angle of the reservoir
and the mobility ratio. The formula can be used to calculate the critical injection rate
above which injected water will under-run the oil and form a water tongue. One of
the key findings is that the displacement is unconditionally stable in a tilted reservoir
when the mobility ratio is less than or equal to one (i.e., water is not more mobile than
oil).
32/49
Figure 1-16 Stable displacement in the Dietz model
1.4.3.3 Stiles and Dykstra-Parsons Models
Stiles and Dykstra-Parsons developed the first generation of analytical techniques

useful for forecasting waterflood recovery in a layered reservoir, assuming a piston-
like displacement occurring in each non-communicating layer. Although rarely used
in their original forms, the basic concept and approach have been adopted as the bases
of many analytical methods developed later on, e.g., a model described in [Ref. 1-28]
which replaced the piston-like displacement by a Buckley-Leverett frontal
displacement, and a recent model, which replaced the linear flow by the radial flow in
the 5-spot situation [Ref. 1-34]. The key finding of this class of analytical techniques
is the sensitivity of the waterflood performance to the permeability contrast between
layers - normally quantified as the Dykstra-Parsons coefficient or alternatively as the
Lorentz coefficient. A fast waterflood breakthrough is expected in a reservoir with
high Dykstra-Parsons coefficient and poor waterflood performance is therefore
expected (see Figure 1-17).
33/49
Homogeneous Reservoir Heterogeneous Reservoir
Water Oil Water Oil
Low Dykstra -Parsons High Dykstra -Parsons

Coefficient Coefficient
Figure 1-17 Stiles model showing the impact of reservoir heterogeneity on

waterflood sweep efficiency
Ref. 1-28 Snyder, R.W. and Ramey, H.J.,Jr., ”Application of Buckley
Leverett displacement theory to non-communicating layered
systems”, Journal of Petroleum Technology, page 1500-1506,
November 1967. (Electronic version not available)
Ref. 1-29 Dake, L Fundamentals of Reservoir Engineering Elsevier Scientific
Publishing Company, Amsterdam 1978. (Electronic version not
available)
1.4.3.4 Buckley-Leverett model with gravity effect
In this type of calculation, gravity along the bedding plane is taken into account
whilst capillary pressure gradients in this direction are ignored. In order to ensure the
validity of the calculation, one first has to ascertain that there is no saturation change
over the height. This is the case if the length of the capillary transition zone greatly
exceeds the height of the formation. Second, displacement should not be subject to
small-scale instabilities leading to distortions in the displacement front ("viscous
fingering").
A typical example of the recovery efficiency calculation for a linear system consisting
of one injector on one end and a producer on the other is shown in Figure 1-18. In
this case the oil /water relative permeability curves have been generated using the
Corey formula, which correlate the oil and water relative permeability curves using
power law equations with exponents no and nw for the oil and water – as described in
Section 1.4.2.2.4.
In Figure 1-18, G denotes the ratio of gravity to viscous forces in the direction along
the bedding plane, which in this case takes a value of 0.2. M represents the endpoint
mobility ratio, being equal to the mobility of water at residual oil saturation divided
by the mobility of oil at connate water saturation. As Figure 1-18 shows, the more
curved the oil relative permeability line becomes (higher no, with no defined in
Section 1.4.2.2.4), the lower the recovery of oil becomes for a given level of
injection. For either exponent value, however, the large number of injection pore
volumes required is mainly due to the unfavorable endpoint mobility ratio M of 10 -
34/49
representing roughly the situation of waterflooding a 50 cP viscous oil. For further

information, the readers are encouraged to consult [Ref. 1-30]
Ref. 1-30 Production Handbook, Volume 4, EP 2000-5800
G = 0.2, no = 3
G = 0.2, no = 7
Figure 1-18: Waterflood efficiency using Buckley Leverett model with gravity
effect (M = 10, n w = 2, with nw defined in Section 1.4.2.2.4) in an
unfavorable situation
1.4.3.5 Improved Dietz model with saturation transition
To look at the vertical sweep efficiency in homogeneous thick sand layers where the
capillary forces are small and oil saturations are immediately reduced to residual
behind the flood front the classical theory of Dietz is available to predict vertical
sweep efficiencies under conditions of both stable and unstable displacement. A
situation reflecting stable displacement where the oil/water interface is moving at an
angle b (b>a) is depicted in Figure 1-16. Note that the conditions for displacement
become unstable when the injection rate exceeds the so-called critical rate.
A more comprehensive 2D displacement theory is available [Ref. 1-31] which
combines the concepts of Dietz with those of Buckley-Leverett in a consistent way. In
other words, this theory enables the calculation of the detailed shape of the edge of the
displacement front together with the saturation transition behind it. Results from this
more comprehensive model are given in Figure 1-19 and Figure 1-20 for conditions of
stable displacement. Figure 1-19 shows the interface angle of the displacement front
plotted against injection velocity relative to the so called Dietz critical velocity, for
endpoint mobility ratio equal to 2, curved relative permeabilities and a formation tilt
angle of 30 degrees. If the conventional Dietz theory were applicable, see lower
straight line in Figure 1-19 and Figure 1-20, the (sharp) frontal interface would be
horizontal (frontal tilt angle equal to formation dip) at injection velocity zero and
parallel to the formation (reflecting unstable conditions) when the velocity becomes
equal to the critical velocity. The new theory (middle curve developed by Schulte)
indicates that, due to the saturation gradient and relative permeability effects behind
the front, unstable displacement would in fact not occur in this particular case
although the endpoint mobility ratio is unfavourable. The upper curve in Figure 1-19
shows calculated frontal interface angles using an earlier approximation (developed
35/49
by Ypma) to the solutions presented in [Ref. 1-31]. It turns out that this approximation
is reasonably good in the case at hand. In the other case (Figure 1-19), where the dip
angle equals 10 degrees and all other conditions are identical, the results of
approximate solution would be incorrect but those of the Dietz theory are virtually
identical to the ones obtained by the more comprehensive model. In conclusion, when
calculating conditions of stable displacement and displacement front angles, the
validity of simple models such as Dietz should first be verified.
Figure 1-19 Stability angle vs. water injection rate

(M = 2, a = 30, no = 5, nw = 4)
36/49
Figure 1-20 Stability angle vs. water injection rate

(M = 2, a = 10, no = 5, nw = 4)
Ref. 1-31 Schulte, A.M. "Analytical modelling of 2D flow in gravity-stable
secondary gas or water drives" EP 88-1609 1988. (Electronic
version not available)
1.4.3.6 Stiles model for 5-spot pattern
In cases of heterogeneous layer packages with no vertical communication, models

due to Stiles [Ref. 1-32] and Dykstra-Parsons [Ref. 1-33] can be used to estimate
vertical sweep efficiency. In both cases, it is assumed that saturations jump abruptly
from maximum water saturation to connate water saturation across the flood front
(i.e. piston-like displacement) and relative-permeability functions are identical in all
layers. The method of Dykstra-Parsons presents useful correlations of recovery
efficiency versus a so-called “Dykstra-Parsons coefficient”, which reflects the
variation in porosity and absolute permeability over the reservoir thickness.
Gardner developed a semi-analytical technique that replaces the piston-like
displacement in the Stiles technique by a radial frontal displacement in a pattern-flood
situation [Ref. 1-34]. The technique can be used to provide quick forecasts of oil
recovery of a waterflood development using an idealized 5-spot pattern with uniform
reservoir properties. Gardner’s technique has been implemented in a WATERFLD
spreadsheet, where the impact of Dykstra-Parsons coefficient, water-oil relative
permeabilities, viscosities, and initial gas saturation on the waterflood performance
can be evaluated [Ref. 1-35].
Ref. 1-32 Stiles, W.E. “Use of permeability distribution in water flood
calculations”, Trans., AIME, Vol. 186, Page 9-13, 1948. (Electronic
version not available)
37/49
Ref. 1-33 Dykstra, H. and Parson, R.L. “The prediction of oil recovery by
water flood”, Secondary Oil Recovery in the United States, 2nd
edition, API, page 160-174, 1950. (Electronic version not available)
Ref. 1-34 Gardner, J.W., “Quick forecasts of waterflood performance in a
layered 5-spot with non-zero initial gas saturation”, EP 93-2361,
1993.
Ref. 1-35 Link to WATERFLD.xls spreadsheet.
1.4.4 Scoping dynamic simulation for waterflood prediction

This section illustrates the applications of dynamic simulations incorporating limited
geology. Although these models cannot provide direct forecasts of waterflood
recovery, they can be used to (1) address some specific issues related to waterflooding
(such as fracturing), (2) generate appropriate relative permeability curves for use in
full-field simulations, and (3) explore the basic mechanisms of displacement
processes. When up-scaled properly, reasonable estimations of the full-scale water
performance may also be achieved.
1.4.4.1 2D cross-sectional models
In cases of large vertical heterogeneities, it is necessary to resort to reservoir

simulation to estimate vertical sweep efficiencies. As an example of a very complex
flood front in the case of flooding perpendicular to channels, Figure 1-21 shows the
simulated results for a cross section in the Brent reservoir.
Figure 1-21 Example of a complex fluid front in a Brent reservoir simulation
38/49
Figure 1-22 Summary diagram illustrating the impact of different variables on

waterflood performance on ‘Brent-type’ and ‘VLA-type’ sand-
body permeability distributions
A useful summary of the effect of heterogeneities on vertical recovery efficiency is
given in Figure 1-22 which shows that heterogeneities as encountered in many of the
Group's reservoirs can easily reduce vertical sweep efficiencies by a factor 2 vis-a-vis
the homogeneous case [Ref. 1-36]. An extreme case would be that of a thief zone
where most of the water is taken up by a highly permeable, areally extensive, sub-
layer.
Vertical sweep efficiencies should never be looked at in isolation in these complex
reservoirs, as it may turn out that communication could take place in the (ignored)
lateral third dimension. However, vertical sweep efficiency estimates can
nevertheless be useful as a learning step.
39/49
Ref. 1-36 Livera, S.E. ,Herweijer, J.C. , O'Meara, D.J.,Weeda, H.C. , Roest,
J.A. "Displacement modelling in clastic sedimentary units" Paper
given at the 1988 Group Geological/Petrophysical Conference.
1.4.4.2 2D models to generate pseudo relative permeability curves
Strictly speaking, relative permeability data from the core plug measurements
(described in Section 1.4.2.2) are applicable to predict waterflooding in the small,
inch-scale rocks whilst the grid sizes in the reservoir simulations are in the order of
tens or even hundreds of feet. In the past when the computing power was limited,
reservoir engineers would use the analytical or simulated results of the vertical sweep
efficiency to calibrate relative permeability curves for use in the areal simulation.
Today, even with much improved computing power, 2D (sometime even 3D) fine-
grid models, which encompass enough small-scale geological details, are still
frequently used to upscale relative permeability data measured in small, core-plug
scales to large scales (say 50 -100 meters) so that the entire reservoir can be
efficiently simulated. A typical example of the latter case is illustrated later in this
section.
The use of rock curves (which are the relative permeability data measured from the
special core analysis) is acceptable if the capillary pressures are high enough to
prevent significant fluid segregations (by gravity forces) within a simulation layer. A
typical example of this situation is the simulation of relatively tight carbonates (5 mD)
where capillary transition zones (say 20-40 ft) are larger than the typical grid layer
thickness of about 10 ft.
When simulating a high permeability reservoir, typically a sandstone reservoir, non-
uniform fluid distributions (due to gravity segregation) invalidate the use of rock
curves. In one extreme case, where the fluids redistribute very quickly to reach
vertical equilibrium within the typical time step (days) of reservoir simulations, one
can impose the vertical equilibrium assumption to predict fluid saturations and then
use the saturation profiles to construct the so-called “pseudo” relative permeability
curves. Analytical tools for this purpose have been developed [Ref. 1-37]. Typically,
the resulting pseudo curves have exactly the same end-point saturations and end-point
relative permeabilities as the rock curves but with smaller Corey exponents. For the
case of zero capillary pressure in a single sand, the relative permeability curves
become linear.
When vertical equilibrium is not valid, a 2D simulation model can be constructed with
sufficient geological detail similar to those described in Section 1.4.4.1, and the
simulator can be used to calibrate relative permeability curves. The pseudo curves
constructed from this approach are referred to as “dynamic pseudo” in the literature.
A number of 2D geological realizations have been developed to represent “typical”
turbidite sand packages (based on observations of sand / shale distributions from
outcrops and from many Mars Basin cores) [Ref. 1-38]. Simulated results (see Figure
1-23) suggest an efficient waterflood process, which benefits greatly from the up-dip
migration of oil films along the bottom of the disconnected inter-bedded shales. The
predicted water breakthrough profiles were used to calibrate pseudo relative
permeability curves (See Figure 1-24) to be used in field-scale simulations. The
relatively lower Corey exponents of the pseudo curves (than the rock curves) resulted
40/49
in more favorable waterflood predictions (higher water injectivity and faster recovery)
than using the rock curves.
Realization 2B: End of WF

Avg. Bed Thickness = 4.84’
Kv/Kh = 0.002
Figure 1-23 Simulated oil saturation in a turbite sand package during

waterflooding
Water-Oil Pseudo Development Water-Oil Relative Perm

Water Flood Prediction
350 1.4
1.2
300
Krw or Krow
1
BOPD
0.8
250
0.6
200 0.4
0.2
150
0
50 70 90 110 130 150
0 0.5 1
Months
Water Sat
Realization 1 Realization 2
Realization 2A Realization 3 Water (Rock) Oil (Rock)
1D Pseudo 1D Rock Water (Pseudo) Oil (Pseudo)
Figure 1-24 Development of dynamic pseudo relative permeabilities by matching

water breakthrough profiles predicted from fine grid 2D cross-
sectional simulations
Ref. 1-37 Link to Relperm_Doe.xls
Ref. 1-38 Hsu, C.F., “Fine-grid simulations of primary depletion and water
flood process in typical Mars-Basin turbidite sand packages”
41/49
1.4.4.3 2D areal simulations via streamline model and advanced

streamline simulation
Although full-scale reservoir simulation is the preferred method, the use of a

streamline model as a precursor is sometimes desirable for quick evaluations of the
pattern performance and in case of large uncertainties in the geological model. An
example of such a streamline simulator is described in [Ref. 1-39]. The results of
such a simulation (see Figure 1-25) can give valuable first indications of the likely
areas in the reservoir where oil is bypassed – an additional injector is needed in the
poorly swept area. Most of the features of the special purpose simulator reported
under [Ref. 1-39] are presently also available in the MoReS simulation system.
Advanced streamline simulators have been developed recently [Ref. 1-40] to account
for fluid compressibility and complicated reservoir geology. This type of simulator
can simulate very large models (close to one million cells) with a computing speed
many fold faster than a conventional finite-difference simulator. Although such
simulators have been applied to predict fluid flows in very complicated geological
settings (with multiple realizations), practical aspects of this modeling approach have
yet to be demonstrated due to limited modeling capabilities for fluid properties, well
control, relative permeability, capillary pressure, plus many other factors.
Ref. 1-39 Weeda, H.C. "Tracer flow simulator TRACKSIM Introductory
users guide" RKMR 89.076, 1989. (Electronic version not
available)
Ref. 1-40 Link to FRONTSIM website at
http://www.sis.slb.com/content/software/simulation/frontsim.asp
Poorly
Swept
Area
Figure 1-25 Streamlines from two injection wells in blue triangles
42/49
1.4.4.4 2D models of idealized flood pattern with induced fractures
An important factor in areal sweep efficiency considerations is the presence of

waterflood induced growing hydraulic fractures. Based on studies with special
simulators, it can generally be said that they have little influence on the waterflood
sweep efficiency provided: (1) fractures are relatively short compared to the interwell
spacing and (2) the fracture propagation rate in the reservoir does not exceed a typical
flood-front velocity prevailing under unfractured conditions.
As an illustration, a result of a simulation is shown indicating the influence of a long
fracture on waterflood sweep efficiency. The case refers to a homogeneous highly
permeable sandstone reservoir resembling North Sea conditions with a hypothetical
well spacing of 250 m. Producer in position 1 (Figure 1-26) refers to a linedrive
pattern with the fracture growing directly from the injector to producer, position 3
refers to another linedrive pattern with the fracture growing in a direction
perpendicular to the line connecting injector with producer, and position 2 refers to a
five-spot pattern with the fracture growing at an angle of 45 degrees vis-a-vis the line
connecting injector and producer. Conditions in this simulation were set such that a
long thermal fracture over nearly 90% of the inter-well distance developed, resulting
in a significant difference in recovery efficiency of 20% MHCPV (Moveable
HydroCarbon Pore-Volume) between the best and the worst case (Figure 1-27). In
most thermally fractured injectors, fractures are, however, much shorter, resulting in
insignificant effects on sweep efficiency but an improved injectivity.
Figure 1-26 Areal view of grid system with possible injector/producer

arrangements
43/49
Figure 1-27 Impact of producer position on cumulative oil production when

injecting under fracturing conditions
1.4.5 Multi-discipline full-field integrated study for waterflood

prediction
1.4.5.1 General Considerations
This section describes the preferred approach for predicting waterflood recovery. An
integrated full-field study involving multi-disciplines has been demonstrated, in
virtually all instances, to be one of the best practices for both reservoir
characterization and management. Usually, 3D reservoir simulations are key steps in
waterflood development studies. The study can formulate field development and re-
development plans and can help to identify improvement opportunities, including
waterflooding. The preferred process of an integrated reservoir study is described
generically elsewhere [Ref. 1-41]. Here are, however, some suggestions that could
make waterflood modeling more efficient:
• Define up-front the specific business objectives of the study and construct
models with only sufficient complexities to achieve these business objectives.
Judge the appropriate models to use: a sector model, a full field model, or
something in between - including multiple realizations to quantify the impact
of the key uncertainties in the reservoir. Accept that some reservoirs are
unpredictable due to geological complexities.
• Gauge when the study has reached the stage of diminishing return in achieving
the business objectives. For example, be prepared to accept 80% solutions, if
further fine-tuning will take excessive time and not impact business decisions.
• Have a clear understanding of the lower risks of developing a brown field as
opposed to a green field. For example: the number of geological realizations
44/49
can be reduced by the production history; a single realization may be

acceptable if calibrated to sufficient production history.
• Choose appropriate simulation software for the study. Focus on applying
sufficient physics, fast turn-around, easy of use, good display, and portability
of the simulation work in case of staff turn-around.
• Have a clear understanding of the basic recovery mechanisms and be able to
relate the modeling results to the basic physics of the process.
• Focus on the physics and engineering instead of IT aspects of the simulation
study.
• Preserve enough reservoir heterogeneity in the up-scaling process
• Use appropriate well constraints in forecasting.
Ref. 1-41 Link to IRM web site at http://sww.global-
networks.shell.com/forums/networks/dispatch.cgi/IRM/showFolde
r/100108/def/def/fb45
1.4.5.2 Al Huwaisah Field Case

1.4.5.2.1 Introduction
In the following sections a joint PDO-SEPTAR study of the Al Huwaisah field in
North Oman is summarized to illustrate the proper study process and the potential
impact of integrated study work.
The billion-barrel Al Huwaisah field (STOIIP: 244 MMm3) is a large, low relief
faulted dip closure with a 40-meter thin oil column. The oil is low viscosity (1cP),
high API (38°) and highly undersaturated. The field was discovered in 1969 and has
produced 17% of its STOIIP to date from the Aptian Shuaiba Formation (carbonates).
Faulting and fracturing has a major impact on fluid movement through a significant
portion of the reservoir. Production behaviour is highly variable with initial net oil
rates from 0-1800m3/d and 0-100% initial water cut. As of 2002, the field average
water cut was greater than 90%.
Historically the Al Huwaisah development has been subdivided into 4 sectors: Main
Area, Southwest Area, Eastern Flank and Eastern Satellite (see Figure 1-28). As of
late 2002, only producers were being drilled in the Main Area and Eastern Satellite.
Pressure support through an aquifer from the west and via fracture corridors is strong
in the western parts of the Main Area but is insufficient in the Southwest Area and the
eastern parts of the Main Area (Figure 1-28). A water injection pilot has been
initiated in the Southwest Area since the first quarter of 2002. Production response to
the water injection has been observed in the area. In 2001, a joint field study by PDO
and SEPTAR was initiated to formulate an optimized field development plan
(including field-wide water injection) to achieve a higher ultimate recovery.
Ref. 1-42 Grondin, et. al, ”Al Huwaisah redevelopment: Static and dynamic
modeling to December 2002”, EP 2003-332, 2003, still in
preparation. (Electronic version not available)
45/49
OWC at 1476 mTVSS
FWL at 1490 m TVSS

Northeast Area
Contour interval 10 m
East Flank
Eas
tern
Sat
ellit
Main Area e
Southwest Area
Figure 1-28 Al Huwaisah reservoir top structures and current development

areas
1.4.5.2.2 Full-field static reservoir model – Al Huwaisah field example

The main reservoir facies in the Shuaiba Formation of the Al Huwaisah Field is
accumulated on a rudist bearing shelf margin some 4 to 8 km wide. They comprise
aggrading rudist mounds and thickets 1 to 4 km in diameter. Locally high-energy
channels (up to 800m wide and 70m deep) and sand waves are developed. Poor to
non-reservoir mud dominated, lagoonal deposits, with isolated rudist mounds, are
present to the east and south of the main field area. Variable bedding dip magnitude
and azimuth are seen in FMI and vertical dipmeter data. High dips (30 degrees +)
towards the lagoonal area are observed. A remaining uncertainty is whether the
dipping beds represent large-scale rudist margin flanks or are smaller scale flank and
cross-bedded deposits within meter scale depositional cycles. Porosities are between
8-26% and permeabilities are typically 10 to 100's mD in high-energy facies and
vuggy intervals.
The field has 40km+ FMI data from horizontal wells and is covered by 3D seismic of
various vintages (1990-2001). The field generally has a low background fracture
density (av. 0.14 per meter), however there are discrete fracture corridors, which are
usually fault related. There are two well defined fault / fracture sets with NE-SW and
NW-SE orientations respectively. The NE-SW fault / fracture trend, parallel to the
present day regional maximum horizontal stress direction, is dominant. Fractures
provide a conduit for bottom water and are associated with losses during drilling.
There is a close relationship between drilling losses and calculated fracture intensities.
The most intense faulting and fracturing is seen in the western part of the Main Field
Area. The lowest fracture densities are recorded in the South West area and no major
fracture corridors have been identified there.
46/49
The carbonate rocks in Al Huwaisah have been classified into 6 types of lithofacies,
ranging from high quality grainstone to almost non-reservoir (algal) mudstone.
Permeability to porosity correlations have been developed from limited core data but,
unfortunately, the capillary pressure data are too scattered to be useful. Alternatively,
logged saturations are used to derive “pseudo” capillary pressures for each rock facies
and they are used to initialize the oil-saturation distributions in the model.
Petrel was used to construct the static model. Three field-scale models have been
developed, encompassing, respectively, the Eastern Satellite, the Main Area, and the
South Western Area. The static model honors all logged / core data at the wellbores.
Seismic data have been incorporated to constrain reservoir structure and rock
porosities whilst seismic discontinuities have been used to indicate the likely presence
of fractures and faults. Fracture swarms have been identified from the FMI logs and
they are correlated deterministically to yield the fracture corridors in the reservoir.
1.4.5.2.3 Up-scaling – Al Huwaisah field example

All geological details are included in the static modeling on a 40m x 40m x 1 m grid
system. The detailed geological model was then up-scaled (using Reduce++) to 80m
x 80m areally and variable grid thickness vertically, ranging from 4 m to 12 m thick.
The up-scaled static model for the Eastern Satellite area (Figure 1-29) shows the
complicated rock facies and the presence of a deep channel, which has imposed a
significant impact on the production behavior in the area.
Channel
Lithofacies
Figure 1-29 Static reservoir for the Eastern Satellite of Al Huwaisah Field.
(Good rocks have lower lithoface index)
It is very important that geological complexities be retained in the up-scaling process.
In the past, improper up-scaling practices smeared out permeability contrasts and
yielded inappropriate dynamic models, which in tern resulted in optimistic waterflood
predictions. Here are some recommended practices. First, all flow barriers, fracture
corridors, and major geological features (channels, fore-reef, lagoon, mudstone, etc.)
should be retained. Second, the porosity-to-permeability transformation should be
performed in the smallest scale – here in every logged interval of 0.5 ft. Then, at the
Petrel grid containing a well, the horizontal permeability is calculated from the
47/49
arithmetic mean of the logged permeabilities (for every 0.5 ft) within the 1-meter
interval (Petrel grid thickness) whilst the vertical permeability is calculated using the
harmonic mean. The use of harmonic mean can ensure the preservation of vertical
flow barriers in the up-scaled Petrel model. When populating grid properties (phi, k,
facies index, etc.) in Petrel (either via Krigging or Gaussian simulation), proper
horizontal ranges need to be employed. (The use of short ranges will make the model
too “random”. On the other hand, long ranges will make the model too “layer-cake”)
Finally, to upscale from the Petrel model to the dynamic simulation model, a flow-
based upscaling (available in both Reduce++ and Petrel) is always preferred.
Section 1.4.4.2 describes the proper up-scaling process for relative permeability data.
In the past, dynamic simulations employing “pseudo functions”, generated to mimic
Stiles-type flows in layered reservoirs with contrasting permeabilities, were reported
in the literature. With improved computing power, however, this practice should be
avoided - the high permeability layers should be explicitly incorporated in the
dynamic simulation model.
1.4.5.2.4 Full-field dynamic reservoir simulation – Al Huwaisah field

example
The first step in the dynamic simulations was to calibrate the model against all
available reservoir and production data through a history-matching process. It is
important to adjust the reservoir parameters within the uncertainty ranges and for the
modeler to provide feedback to the other PE disciplines within the study team about
the changes. All members in the study team should be in good agreement with the
“knobs” tuned during the history matching. The key knobs in obtaining the Al
Huwaisah history match were the areal distributions of vertical permeability and the
fracture corridors. The Kv/Kh values were adjusted within the range of 0.01 up to 0.5
– reflecting the variability of the geological settings from 30-degree dipping rudist
cycles in a portion of the field to the lagoon deposits in the back reef. The areal
distributions of Kv/Kh have been supported by the geological interpretations of the
field.
The endpoints and the Corey exponents of the water-oil relative permeability were
also adjusted somewhat. The adjustments were applied globally to the entire field and
do not even attempt to provide local fine-tunes even though they could provide better
well-by-well history matches. It is important not to alter the relative permeability
near the well simply to provide a better history match of any particular well, since
such a model provides virtually no predictive capabilities.
One of the key results from the history match is that the major fracture corridors
(created by the faults), running from NE toward SW in the middle of the Main area,
could have been the major paths of the water influx from the bottom water zones - a
finding supported by the seismic data showing the extension of the main faults
continuing into the deeper Hawar and Kharaib formation and the potential open
fractures based on regional stress field. Without the presence of the fracture
corridors, the flank aquifers are not able to keep up with the liquid withdrawals from
the reservoir, resulting in the low-pressure “sinks” in the back-reef of the Main Area
and the entire South Western Area, and possibly the Eastern Satellite Area.
The models were used to simulate various development scenarios - including water
injection. Because of the gravity segregation of water and oil, a pattern waterflood
was found to be ineffective except at close well spacing (100 meters), which
48/49
unfortunately cannot be justified because of the thin oil column. Instead, simulation
results indicate that an edge water injection scheme (Figure 1-30) appears to be able
to provide enough pressure support to many horizontal producers already drilled in
the field although interior injectors have been planned in strategic locations to
accelerate the waterflood responses. The plan was also supported by the results of the
waterflood pilot as described in Section 1.3.7.
Likely WF Development Patterns

Al Huwaisah SW Area
Water Injector
Figure 1-30 Likely well locations of the waterflood development in the Al

Huwaisah SW area using 300-meter well spacing (Color scale used
to indicate top structures of the reservoir)
49/49

WF1 Course

Uploaded by

Copyright:

Available Formats

WF1 Course

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

WF1 Course

Uploaded by

Copyright:

Available Formats

What are the three major topics addressed in the chapter?

What are the three major topics addressed in the chapter?

What are some of the factors that should be considered before starting a waterflood project?

What are some of the factors that should be considered before starting a waterflood project?

Waterflood Manual Chapter 1.

Reservoir Aspect of Waterflooding

1. Reservoir Aspect of Waterflooding

1.2 When to Waterflood?

primary objective of a waterflood, it is also used as a method of disposing wastewater

1.2.2 Unfavorable reservoir conditions

1.2.3 Optimal reservoir pressure level

Figure 1-1 Waterflood residual oil saturation as a function of free gas

1.3 Waterflood Design Considerations

Figure 1-2 Reservoir structure framework (a), depositional architecture (b),

1.3.1.1 Structure framework

Reservoirs with Extensive

-Presence of fracture corridors C-5 C-1

Figure 1-3 Challenges of waterflooding a natural-fractured and faulted

1.3.1.2 Depositional architecture

Depositional architecture, encompassing the size, shape and distribution of reservoir

Predictive stratigraphic tools such as sequence stratigraphy is recommended, where

A sound reservoir model and understanding of

Any waterflooding activity if planned and designed

Figure 1-4 Multiple subsurface realisation for fluvial reservoirs

because flow paths in carbonate reservoirs can be dominated by secondary porosity

1.3.1.3 Reservoir properties including diagenetic overprints

A correct estimate of aquifer geometry and volume connected to the reservoir is

1.3.2 Peripheral versus pattern water injection

5-Spot Pattern 7-Spot Pattern

9-Spot Pattern Linedrive Pattern

Figure 1-5 Various waterflood patterns

Asab, Bab, BuHasa, Abu Dhabi Flank/Pattern Ongoing

Plutonia Cluster, Angola Flank VAR 4

Bijupira-Salema, Brazil Flank Start-up

Seria West, Brunei Crestal Abandoned

Seria RE*A, Brunei Pattern Ongoing

Champion Main, Brunei Flank Ongoing

Champion South East, Brunei Flank/Pattern VAR 3

Dan, Denmark Pattern (pilot) Ongoing

Skjold, Denmark Flank Ongoing

Gamba/Ivinga, Gabon Mostly crestal Abandoned

Barton, Sabah, Malaysia Flank VAR 3

Kinabalu, Sabah, Malaysia ESP Dumpflood Ongoing

Schoonebeek, Netherlands Flank Abandoned

Egwa, Nigeria Flank Ongoing

Forcados Yokri, Nigeria Flank VAR 3

Bonga, Nigeria Flank Start-up

Draugen, Norway Flank Ongoing

Statfjord, Norway Flank Ongoing

Fahud Matrix, Oman Pattern Ongoing

Lekhwair, Oman Pattern Ongoing

Marmul, Oman Pattern Ongoing

Saih Rawl, Oman Pattern Ongoing

Yibal, Oman Pattern Ongoing

Zauliyah, Oman Pattern Ongoing

Astokh, Sakhalin, Russia Flank Start-up

Piltun, Sakhalin, Russia Flank VAR 4

Salym, Russia Pattern VAR 4

Omar Field, Syria Flank Ongoing