SPE-171488-MS

Identification and Modeling of Inter-Reservoir Cross-flow Using a Novel

Material Balance Method
E.T. Montague, LENS Petroleum Pty. Ltd.; Hafneh Airul, Brunei Shell Petroleum (BSP) Sdn. Bhd

Copyright 2014, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Asia Pacific Oil & Gas Conference and Exhibition held in Adelaide, Australia, 14 –16 October 2014.

This paper was selected for presentation by an SPE program committee following review of information contained in an abstract submitted by the author(s). Contents
of the paper have not been reviewed by the Society of Petroleum Engineers and are subject to correction by the author(s). The material does not necessarily reflect
any position of the Society of Petroleum Engineers, its officers, or members. Electronic reproduction, distribution, or storage of any part of this paper without the written
consent of the Society of Petroleum Engineers is prohibited. Permission to reproduce in print is restricted to an abstract of not more than 300 words; illustrations may
not be copied. The abstract must contain conspicuous acknowledgment of SPE copyright.

Abstract
A detailed material balance study of the Northern Area Deep reservoirs was conducted to address
anomalous trends identified from p/Z analyses. The material balance study was conducted using a novel
application of Petroleum Experts’ MBal software, and determined that previously unrecognised volumes
of gas were cross-flowing between reservoirs predominantly behind casing in the wells. The magnitude
of the cross-flow volumes significantly impacted well and reservoir performance, but were unable to be
modelled or quantified.
The approach that was adopted to model the inter-reservoir cross-flow using MBal as a “sub-surface
abacus”, whereby injection of gas into the reservoirs was introduced manually to represent influx due to
cross-flow. Additional production (on top of allocated produced volumes) was introduced to model the
corresponding efflux of gas. Manually maintaining a volumetric balance between the efflux and influx
profiles was difficult, but achievable, and provided an internally consistent method to model the complex
dynamics of this reservoir system.
Detailed simulation studies have been recently conducted on the Northern Area Deep reservoirs and
have complemented the findings from this material balance study. The MBal model has been updated
regularly with over 6 years of additional production and pressure data, and the model is still holding a
good history match. This has provided high confidence in the model’s robustness, and validates the
adopted methodology, which has broader applications to enable material balance modelling of “non-
geological” inter-reservoir cross-flow.
The purpose of this article is to present the methods and practices employed in this process and to show
how they can be applied to other fields/reservoirs.
Introduction
The Northern Area Deep reservoirs are part of a large, offshore oil and gas field located in the western
area of Brunei. The Northern Area Deep reservoirs are at the northern part of an anticline that plunges in
both directions and formed over an elongate shale ridge. The structure is compartmentalised by north-east
to south-west trending growth faults, forming sealing boundaries on the northern and southern areas of the
field. Fig. 1 shows that the Northern Area is divided into two main fault blocks; Block A to the south and
Block B to the north. A single counter-regional growth fault, the Northern Area Fault, separates Block A
2 SPE-171488-MS

Figure 2—Simple Cross-Sectional Area

Figure 1—Northern Area Field Overview

from Block B. A clay diapir is present at the western opening of Block A, and is interpreted to isolate the
Block A reservoirs from aquifer support from the west. Block B is interpreted to have aquifer support,
especially in the shallower horizons as the permeability increases with decreasing compaction.
Hydrocarbon reservoirs in the Northern Area are found in depths greater than 1000m. The Northern
Area Deep reservoirs that were the focus of the study are the DE, DF and DG reservoirs (Figs. 2 & 3).
Reservoir quality deteriorates with depth and towards the east and north-east reflecting the direction of
sediment deposition of these shallow marine, shore-face reservoirs. Block B reservoirs are predominately
oil reservoirs with large gas caps, with the sand open to the west, some aquifer support is expected from
this direction, and consist of the DE, DF & DG reservoirs. For the purposes of this study the Block A Deep
reservoirs consist of the DE and DF reservoirs, which are predominately gas reservoirs with some small
oil rims in the upper DE. No aquifer support is interpreted in Block A because of the diapir to the west,
the deterioration of reservoir quality to the east and the presence of the major growth faults sealing to the
north and south respectively.
Production began from Block B Deep reservoirs in 1969 and from the Block A Deep reservoirs in 1972.
Understanding the gas production history of the Northern Area Deep reservoirs is important because of
its significant influence on the production performance, in particular the interaction between Blocks A and
B. Fig. 4 summarises the gas production strategies for the Northern Area Deep gas reservoirs into 4
discreet time periods. The first period from 1969 to 1980 (time period A) had production from both Blocks
A & B, and then a gas conservation policy was employed on the Block B DE and DF reservoirs from 1981
to 1996 (time period B). This policy required a reduction in the Block B gas production rates to conserve
reservoir energy and maximise recovery from the oil rims, during this period Block A production
increased dramatically. In 1997, permission was granted to release the Block B DE and DF gas caps (time
period C). Shortly after in 2000 the Block A DC internal gas injection scheme was implemented, this
project involved cross-flowing gas from the Block A DE/DF reservoirs via two dedicated gas injection
wells. In order to maximize recovery from the DC oil reservoirs, production from the Block A Deep
reservoirs has been controlled and monitored as part of the conservation gas strategy (time period D).
Previous studies and their model results indicated the potential for fluid movement via lateral/vertical
cross-flow, but unfortunately due to the complex nature of the data set these were unable to be definitively
quantified or modelled. Consequently, these models were unable to provide a totally integrated picture of
the Northern Area Deep gas reservoirs, and this was recognised to result in significant uncertainties in the
model forecasts. Therefore, a requirement was identified to conduct a study that was able to definitively
SPE-171488-MS 3

Figure 3—Stratigraphic Section of Northern Area Deep reservoirs Figure 4 —Northern Area Production History & Production Strategies

identify the magnitude and direction of the inter reservoir cross-flow and then devise a method to model
it. The end result was the novel application of material balance model presented in this paper. Efflux is
modeled as additional production and influx is modeled as additional injection, with the material balance
tool (MBal) effectively being used as a “sub-surface abacus”.
The study utilised the following methodology:
● Review and QC pressure and production data, which identified anomalous well production
performance associated primarily with poor well integrity and cross-flow behind pipe as well as
misallocation of production.
● Conduct a first stage material balance analysis via p/Z interpretation to identify the magnitude and
timing of cross-flow at a high level which could then be used for the construction of the material
balance models.
● Conduct a second, more detailed stage of material balance analysis using MBal to calculate
influx/efflux profiles, and the cumulative efflux/influx gas volumes.
The end result was the first time an auditable and consistent methodology to calculate the current gas
in place volumes and reservoir pressures for these reservoirs which were then fed into the corporate gas
forecasting tool for calculating reserves and production forecasts. This methodology was endorsed by
internal and external auditors.

Pressure and Production Data Review

A vital step in the construction of the material balance model is a critical review of the data that is used
for the analyses; namely the PVT, pressure and production data. In the case of the Northern Deep gas
reservoirs there are uncertainties in the pressure and production data because the reservoirs are relatively
thin, stacked sands, and are in many wells under commingled production. Most of the pressure data
available for analyses are limited to static pressure surveys, which due to commingling, are not necessarily
representative of the individual reservoir pressure but are an average of multiple reservoirs. Production
allocation also carries uncertainty because of the limited amount of PLT data available.
Previous studies had used a mixture of static and RFT pressures resulting in a lot of data scatter, and
ambiguity in the interpretations. This was especially evident in the Block B DG1-4 sands (Block B-DG’s).
4 SPE-171488-MS

Figure 5—Block B-DG2’s Reservoir Pressure History

For this study a conscious decision was made to rely

primarily on RFT data, and use it to verify selected
non-comingled static pressure data as required to fill
in data gaps. The advantages of this were a signif-
icant reduction in data volumes resulting in clearer,
higher confidence reservoir pressure trends. Figs. 5
and 6 depict the results of the analyses from two DG
reservoirs, and they provide evidence for previously
unrecognised reservoir compartments. Another im-
portant observation is the occurrence of a large
influx of energy around 1980, which occurred si-
multaneously in all of the Block B-DG reservoirs.
Although the overpressures/differential depletion Figure 6 —Block B-DG4 Reservoir Pressure History

observed in WELL-D & WELL-DST1 (Fig. 7) have

long been recognised, they were assumed to be a
result of a stratigraphic pinchout and various sedimentological models had been proposed. However, the
higher, differential pressures observed in WELL-C (~8000 kPa) (Figs.6 & 7) which is close to the
depleted, main producing core of the field, had not been integrated into previous models. The impact of
the WELL-C RFT points is highly significant because it required a re-interpretation of the presence of
pressure compartments in the Block B-DG reservoirs. A detailed stratigraphic and structural review could
not identify any faulting or pinch-outs between WELL-C and the more depleted wells in the core area to
the south. Further consideration resulted in a new interpretation that reservoir compartments were caused
by clay filled sub-seismic faults and fractures sourced from underlying overpressured formations, and this
also explains the higher pressures at WELL-D/-DST1. Fig. 7 depicts a likely location for a compartment
boundary as a clay filled extension of the smaller scale faults within Block B, and that also fits with the
differential pressures observed between SWA-C and the producing well further to the south.
RFT data also revealed pre-production depletion in the Block B-DG3 providing evidence that zonal
isolation in these wells had been compromised at the start of production due to poor well integrity.
Numerous anomalies in the static pressure data observed throughout the life of the wells are also evidence
of compromised well/completion integrity. Identifying these phenomena places even more importance on
using open hole RFT data to generate high confidence pressure trends.
The scatter in the pressure data for the shallower DE/DF reservoirs was less pronounced than the Block
B-DG’s and plotting RFT and static pressure together corroborated with the static pressure trends.
SPE-171488-MS 5

Figure 7—Northern Area DF3 Top Structure Map

Figure 8 —B-DF3 p/Z Interpretation (corrected production)

The production data QC was performed at a high level since it is difficult and time consuming to check
the validity of the production splits and allocations between the reservoirs because there are many
wells/zones in the study area. However, this high level pass revealed a significant misallocation of gas
volumes in two Block B Deep Gas wells; WELL-G and WELL-H. Additional zones were added to these
wells in September 2000 via perforating the DG reservoirs. However, after September 2000 the corporate
production database had the entire production from these wells allocated to the newly added DG
reservoirs. The impact of this is that significant gas volumes had been under allocated to the B-DE/DF
reservoirs and over allocated to the B-DG’s. PLT surveys are planned for these wells but preliminary
calculations indicate an over- allocation of a few Bcm of gas to the B-DG reservoirs. This has had a
significant impact on the material balance analyses of the Block B reservoirs (Fig. 8).
Wells with anomalous production performances were also reviewed in more detail, for example a
number of wells exhibited sudden jumps in GOR. The outcome of this review is that these data are
considered valid, and in fact ended up being one of the criteria used to identify wells with integrity
problems.
6 SPE-171488-MS

Figure 9 —A-DF3 p/Z Interpretation

Material Balance Analyses

The material balance analyses were conducted in

two stages. Firstly, a p/Z interpretation to observe
the magnitude and direction of the inter reservoir
cross-flow occurring between the Northern Area
Deep reservoirs. This information was then used to
guide the construction of the MBal models which
were then used to calculate the gas influx/efflux
profiles and provide a more robust estimate of the
current in place gas volumes.
Stage 1: p/Z Interpretation
The p/Z interpretation was conducted manually on
each reservoir unit, i.e. A-DF1, A-DF2 etc., exam- Figure 10 —Block A & B DF3 Pressure and Gas Production History
ples of the hand annotated interpretations of se-
lected reservoirs are depicted in Figs. 8, 9, 11 and
12. It is acknowledged that there are limitations to this method in reservoirs where significant oil
production has occurred (eg. B-DG’s). Not all the reservoirs analysed are shown but only a select few
regarded to provide key insights into the behaviour of the Northern Area Deep reservoir system.
The most important finding from the p/Z analysis is illustrated in Figs. 8 and 9, which identifies
connectivity between Block A and B via the DF3 reservoir. In the B-DF3 the initial p/Z trend yields a GIIP
close to the 3D static model volume. Fig. 8 shows a clear downward deviation in the p/Z trend due to an
efflux of gas equivalent to approximate 30% of B-DF3 GIIP from the reservoir which occurred during the
Block B gas conservation period (time period B, 1981-1996). Fig. 9 shows an influx of the same amount
of gas into the A-DF3 at the same time as the efflux from B-DF3 occurred. Geological and geophysical
data confirms the presence of a flow path around the tip of the Northern Area fault, because there is still
good DF3 sand development to the east of where the Northern Area fault throw dies out (Fig. 7). The DF3
is the only Northern Area Deep reservoir where the inter-reservoir cross-flow is interpreted to be
geological in nature. One noteworthy observation is that WELL-K which is the well closest to the tip of
the Northern Area Fault (Fig. 7), is consistently the first well to show deviation from the previous pressure
SPE-171488-MS 7

Figure 11—Block A- DE3 p/Z Interpretation

Figure 12—Block B- DE2 p/Z Interpretation

Figure 13—Schematic of cross-flow behind pipe

8 SPE-171488-MS

Figure 14 —B-DE GOR and Pressure vs Time

trends. All the other Northern Area Deep reservoirs

pinch out or become very thin with poor quality
near the tip of the Northern Area fault, so they are
interpreted to provide poor or no connectivity be-
tween Blocks A & B at this location. Fault seal
calculations suggest that where the Northern Area
fault is present it is sealing.
The reason gas cross-flowed from B-DF3 to
A-DF3 during 1981-1996 is because of significant
pressure differentials generated by the Block B con-
servation gas strategy and subsequent Block A
blowdown. Fig. 10 depicts the relative reduction in
A-DF3 pressures to B-DF3, the pressure differen-
tials generated resulted in gas efflux from B-DF3
and subsequent influx into A-DF3.
Fig. 10 also shows the changes evident from the
late 1990’s on, with the Block B gas caps placed on Figure 15—Blocks A & B-DE/DF MBal model configuration
full production in 1997 (time period C) and then the
implementation of the Block A conservation gas
policy (time period D, 2000⫹). At that time the pressure gradient, and hence the direction of gas influx
switched from Block A to Block B. Fig. 8 clearly shows that gas is now effluxing from the A-DF3 into
B-DF3, and that most of the gas that flowed into Block A has cross-flowed back into Block B.
Selected p/Z plots are included to describe the characteristics of each group of remaining reservoirs,
i.e. those reservoirs apart from Block A&B-DF3’s. These are summarised below:
A-DE/DF’s- A typical p/Z plot for these reservoirs is depicted in Fig. 11 showing a clear trend of
increasing GIP (gas influx) during the Block B conservation gas strategy (time period B). During the
period when the Block A gas conservation policy was employed and Block B gas blowdown (time period
D) there was a sharp increase in gas efflux and corresponding drop in pressure.
SPE-171488-MS 9

Figure 16 —B-DG1 Initial History Match

Figure 17—B-DG1 Final History Match & Injection Profile

B-DE/DF’s- A typical p/Z plot for these reservoirs is depicted in Fig. 12 showing a clear trend of
increasing GIP (gas influx) commencing around 1980 and continuing throughout the Block B conserva-
tion gas strategy (time period B).
B-DG’s- Typical pressure vs time plots for these reservoirs are depicted in Figs. 5 and 6, showing a
clear trend of increasing GIP (gas influx) starting from 1980 as discussed previously. No p/Z interpre-
tations were conducted on the B-DG’s because of the large amount of oil production and scatter in the
pressure data.
10 SPE-171488-MS

Figure 18 —B-DF Initial History Match

Figure 19 —B-DF Final History Match & Injection Profile

These analyses clearly identify a significant volume of gas influx into the Block B DE & DG reservoirs
around 1980. The coincidental timing of these influxes is regarded as highly significant because these
reservoirs are 300-400 metres apart and there are very few open completions sharing these zones. Any gas
influx into the Block B-DF3 would be masked by the large amount of efflux from the Block B-DF3 to
the A-DF3 during time period B (Fig. 10). It is also interesting to note that the deviations in the p/Z trends
in the Block A DE-DF reservoirs occur later in the shallower reservoirs, indicating an order to the influx
relating to vertical distance from the A-DF3. The interpretation is that there is a large degree of vertical
cross-flow occurring.
SPE-171488-MS 11

Figure 20 —B-DG1 Production Adjusted for X-flow into B-DE/DF

Figure 21—A-DE Initial History Match

‘Previous studies had interpreted the gas influxes into the A-DE1-3/DF1-2 and B-DE1-3/DF1-2 to be
due predominantly to lateral cross-flow through “leaks” in the Northern Area fault since these reservoirs
are not geologically connected around the tip of the Northern Area fault. However, this theory did not
honour all the available data such as the sealing nature of the fault and in some instances where the
pressure differentials were inconsistent with the inferred flow directions.
However, this study has identified that the majority of gas influx is due to vertical cross-flow
predominantly from cross-flow behind pipe as depicted schematically in Fig. 13, in combination with
open perforations and poor completion integrity. Some of the evidence to support the vertical cross-flow
concept is as follows:
12 SPE-171488-MS

Figure 22—A-DF Initial History Match

Figure 23—A-DF Final History Match & Injection Profile

● CBL logs often show poor cement bond relating to large washouts resulting from drilling through
these poorly consolidated reservoirs.
● Temperature anomalies behind casing (Joule-Thomson cooling effects) from shut-in PLT surveys
indicating cross-flow behind pipe.
● Problems with well integrity resulting in cross-flow behind pipe is indicated by the pre-production
depletion observed in some RFT data
● PLT data confirms that vertical cross-flow has occurred from the higher pressured A-DF3 into the
shallower reservoirs via open perforations.
SPE-171488-MS 13

Figure 24 —A-DE Final History Match & Injection Profile

Figure 25—A-DC2 Initial History Match

● Anomalous well and reservoir production performance such as steep GOR trends coinciding with
increases in pressure as depicted for the B-DE reservoir in Fig. 14.
● The unconsolidated nature of these sands across the entire field has resulted in widespread
occurrence of poor hole conditions and the ability to place cement to achieve long-term zonal
isolation has proved problematic.
Fig. 14 illustrates another compounding the cross-flow issue and that is wells that suddenly produce
with high GOR’s were shut-in for gas cap management. The consequence of this is that without this higher
pressured gas flowing through the wellbores to surface increases the gas volumes being diverted into the
reservoirs via cross-flow.
14 SPE-171488-MS

Figure 26 —A-DC2 Final History Match & Injection Profile

Figure 27—Illustration of BHP history for the Block A-DC oil reservoirs.

With the well integrity model now firmly in mind, the inter-reservoir cross-flows identified from the
p/Z analyses can now be interpreted with more clarity; anomalous deviations in pressure are interpreted
to be caused by vertical cross-flow via behind pipe and/or from compromised well integrity/open
perforations. For example, the pressure deviation in the A-DE3 (Fig. 11) is now interpreted to be caused
by the vertical cross-flow of gas from the higher pressured A-DF3. This process continues to occur until
the year 2000 as gas continues to charge the A-DF3 reservoir via the B-DF3 geological connection and
then cross-flows vertically into these shallower reservoirs to maintain equilibrium. After 2000, there is a
reversal of direction in gas flow (time period D) with the implementation of the Block A conservation gas
strategy, and is illustrated clearly in the change in relative block pressures in Fig. 10. The implications of
this can be seen in the late time downward deviation in Fig. 11.
Increase/deviations in reservoir pressures are observed in most of the Block B upper reservoirs (B-DE1
to B-DF2). The most obvious is the B-DE2, as depicted in Fig. 12, showing an increase in reservoir
SPE-171488-MS 15

pressure in the late 1970’s. There is no evidence of compartmentalisation in the B-DE2 and only moderate
aquifer support, indicating that reservoirs are most likely being recharged from a higher pressure source,
which is most likely gas cross-flowing vertically from deeper, higher pressure reservoirs. Fig. 14
demonstrates clearly that these pressure deviations coincide with anomalously large increases in well
GOR’s that are not necessarily related to the structural elevation/location of the well, i.e. downdip wells
are breaking though gas whereas updip wells are still producing oil at low GOR’s.
Although the vertical cross-flow mechanism can be used to explain how gas volume/mass transfer
occurs vertically between segregated reservoirs, another mechanism is required to explain the source of
major energy influx seen in these reservoirs around 1980, in particular the Block B DG’s. The key
observations that provide the clues to explain this are the presence of compartments discussed previously
(Figs. 5 & 6), and that the initial pressure declines in the B-DG’s indicate that the initially accessed GIP
are only 60% of the total static model volumes. The initial pressure response is that of the core of
production wells within the main compartment only accessing a portion of the total reservoir GIIP. Then
around 1980, the higher pressured compartments within the B-DG gas reservoirs begin communicating
with the more depleted core area by a combination of the fault break-down and loss of well/completion
integrity. Regardless of the specific mechanisms, the net effect of mechanical/geological barriers being
breached is to connect reservoirs with large pressure differentials (⬎10,000kPa), and then higher pressure
gas cross-flows into the more depleted, overlying reservoirs. This is interpreted to be the most likely
explanation for the almost simultaneous increase in reservoir pressures from the B-DG4 to the B-DE.
Now that the inter-reservoir cross-flows identified from the p/Z responses could be explained and
understood, the next challenge was how to model this phenomenon in a material balance model (MBal)
and generate plausible forecasts.

Stage 2: MBal Modelling

The challenge in modelling the vertical cross-flow is that at the time of this study the transmissibility links
in MBal were fixed or constant The vertical cross-flow/mass transfer between the reservoirs is due
primarily to mechanical issues and was observed to be increasing over time, as an increasing number of
wells developed integrity problems over time, and wells being opened or shut-in. There is also the
possibility that the cross-flow in individual wells may get worse as cement bonds continue to break down.
However, since the transmissibility link in MBal is a constant this presented a challenge on how to model
these reservoirs where large volumes of vertical cross-flow had been identified.
The novel concept developed in this study was to force MBal to model influx and efflux from these
reservoirs as additional injection and production volumes respectively. The initial pressure declines were
well matched calibrated to the mapped OGIP volumes (except for the deeper Block B DG reservoirs) and
providing a solid starting point before the inter-reservoir cross-flow commenced from 1980 onwards. The
QC’d pressure data became essential at this point as depicted in Fig. 17 where an injection gas profile was
iteratively adjusted to honour the pressure data for the B-DG1 reservoir. This concept uses MBal
effectively as a ‘sub-surface abacus’, manually moving volumes from one reservoir to the other whilst
honouring the overall material balance of the system, and consistency with pressure gradients. This
method is an iterative approach but for the first time provided the ability to track volumes, timing and
directions of inter-reservoir cross-flow.
Fig. 15 shows the configuration of the main MBal model. The Block A and B-DE1 to DF3 reservoirs
were reduced to 4 MBal tanks in order to try and simplify this complex problem; the A-DE, A-DF, B-DE
and B-DF. Pressure trends for the amalgamated tanks were derived from a volume weighted average of
the individual reservoirs. The A-DF and B-DF tanks were plumbed together by means of a transmissibility
link to model the geological connection and subsequent flow around the tip of the Northern Area fault.
The B-DG’s were modelled separately as individual tanks, the B-DG1, -DG2, -DG3 and -DG4.
16 SPE-171488-MS

Figure 28 —Material balance model match post 2006. The red bar showing the period after the conservation scheme was halted.

Fig. 16 shows the initial GIIP is well matched for the B-DG1 but then deteriorates around 1980 as
additional energy influx commences into this reservoir. However, this can be modelled by injecting
additional gas volumes into the MBal model, and can then a good history match can be achieved as shown
in Fig.17. In this example without additional gas volumes, there are insufficient gas volumes in the MBal
tank, and also a bust with the 3D static volumes.
Fig. 18 depicts the B-DF initial MBal history match and it shows that the model pressures begin to drop
below the historical measurements during the 1980’s. This is the same pattern observed in the B-DG’s,
and is explained by the additional gas influx in the B-DG’s is also cross-flowing vertically up into the
B-DF. The additional complexities with modelling inter-reservoir cross-flow of this system is that
efflux/influx rates and directions between the vertically communicating reservoirs are changing over time;
driven by variations in the differential pressures caused by the different production strategies. Fig. 19
shows that large volumes of gas influx in the 1980’s were required to history match the B-DF MBal
model; note that this gas is also feeding into the A-DF via the Northern Area fault tip transmissibility link
(Fig. 15).
The source of this gas is from the underlying, compartmentalised B-DG reservoirs, which as discussed
previously also had large volumes of gas influx at this time. By the 1980’s all reservoirs in the Northern
Area Deep reservoirs are in effective pressure communication through a combination of geological and
well integrity/wellbore factors.
Since the B-DG reservoirs are identified as the main source of the gas influx into the B-DE & B-DF
tanks, then equivalent efflux volumes needs to be modelled from the B-DG ‘source tanks’. Fig. 20
illustrates an example from the B-DG1 where additional gas production models the efflux volume that is
contributing to the gas influx into the B-DE/DF. Following the MBal history matching in Block B, the
SPE-171488-MS 17

inter-reservoir cross-flow concept was modelled as gas flowing sequentially from the higher pressured
B-DG compartments, then into the lower pressure B-DG’s, and then cross-flowing upwards into the B-DF.
The key with this method is to keep a track of the influx and efflux volumes, to ensure that the balance
between the reservoirs is maintained, and that cross-flow directions are consistent with pressure gradients.
Another volume balance requirement is to cross-check that the B-DG influx volumes sourced ‘internally’
from the B-DG higher pressured compartments do not violate static in place volumes. The total B-DG gas
in place volumes (GIP ⫹ efflux) matches with the in place static GIIP volumes.
The main influx/efflux volume movements in the B-DG’s occur in the upper sands (B-DG1 & B-DG2)
because of their proximity to the main pressure sink in the B-DF3 and indirectly to the A-DF3 via the fault
tip communication.
Although the internal B-DG influx rates and cross-flow volumes are significant, it is worth considering
that the B-DG reservoir pressures around 1980 were lower than the Block B-DE/DF and that the reservoir
pressure of the undepleted B-DG compartments would be approx. at near initial pressure condition. The
pressure difference between the undepleted compartments and the shallow depleted reservoirs are
upwards of 10000kPa. Such differential pressures, combined with numerous wells that create avenues for
pressure communication via behind casing flow provides the ideal conditions for significant volumes of
gas to migrate to the shallower reservoirs.
With the Block B reservoirs history matched and underpinning the major cross-flow and influx
volumes of the Northern Area Deep gas reservoirs, it is now time to focus on the Block A reservoirs. Figs.
21 & 22 shows the initial history matches of the A-DE & A-DF MBal tanks and that the pressure matches
begin to deviate from history during the 1980’s. This is consistent with the deviations from history
observed in the Block B initial MBal matches as discussed previously.
With the A-DF and B-DF reservoirs in direct communication in MBal via the transmissibility link, the
large volumes of gas influx modelled into the B-DF (Fig. 19) had a corresponding positive impact on the
improving the quality of the A-DF history match. As Fig. 23 illustrates the gas influx volumes into the
B-DF were sufficient to history match the A-DF pressures during the 1980’s. However, this did not assist
in improving the A-DE match because the tanks are not linked in the MBal model.
To history match the A-DE requires modelling vertical cross-flow between the A-DF and A-DE tanks,
utilising the same methodology as employed in the Block B reservoirs, i.e. influx as additional injection
and efflux as additional production. Fig. 24 depicts the final history match of the A-DE tank through
injecting/influxing gas volumes which honors influx volumes and timings calculated from the p/Z
analysis.
One further complication of modelling the Block A reservoirs results from the implementation of an
internal gas injection (IGI) scheme in 2000 where gas is deliberately cross-flowed via open perforations
from the A-DE and -DF reservoirs to gas-flood the overlying A-DC oil reservoirs. This is the reason why
the A-DE/DF reservoirs commenced a conservation gas policy after the year 2000 (time period D).
Therefore, in order to produce a fully integrated history match of the Block A-DE/DF reservoirs, the
Block A-DC1 and -DC2 reservoirs must also be history matched to model the transfer of gas volumes
between these reservoirs. What was interesting to note that both the A-DC1 and –DC2 required significant
volumes of gas influx during the 1970’s to history match the pressures as shown in Figs. 25, where there
was insufficient gas volumes in the model.
The initial match results (Fig. 25) require additional gas volumes which must be provided by the
cross-flow behind pipe concept because there are no completions that are commingled between the
A-DC’s and the A-DE/DF. Indications of cross-flow behind pipe are evident in the anomalous perfor-
mances of a number of A-DC production wells, for example:
– 50% increase in PI for a high producing well at the same time when the GOR increased rapidly.
18 SPE-171488-MS

Figure 29 —Northern Area Material Balance Montage

SPE-171488-MS 19

– Sharp increases in the GOR for oil wells, which resulted in the wells being shut-in due to them
gassing out.
The Block A-DC MBal models were history matched by iteratively generating a gas injection profile to
match the A-DC reservoir pressures. The matched gas injection profile was then added to the Block
A-DE/-DF cumulative gas production volumes to model the effect of the additional production or efflux
from the A-DE/-DF reservoirs.
With the Block A gas conservation scheme that started during the IGI period from the year 2000,
coinciding with the Block B gas cap blowdown, there is late time reversal of flow direction which
manifests itself as the late time increase in the A-DF pressures (Fig. 22), and the late time sharp drop in
the A-DE pressures (Fig. 11). These pressures are history matched by late-time gas injection into the A-DF
(Fig. 23), and a corresponding efflux from the A-DE tank. With the pressures in the B-DF now lower than
the A-DF (Fig. 10) gas is now cross-flowing back from A-DF3 to the B-DF3 around the tip of the
Northern Area fault. Fig. 8 shows that at the time of the study that the direction of the cross-flow has
reversed and that significant volumes of gas have cross-flowed back into the B-DF3. The gas influx rates
from the B-DG’s into the B-DF are declining as the pressures between these reservoirs continue to
equalise, and conversely the influx rates from the A-DF into the B-DF via flow around the tip of the
Northern Area fault has increased.

2010 Onwards
The material balance proposed has been continuously updated since its creation in 2006. In 2010, the
Block A gas conservation scheme has been stopped. This is due to the reducing benefits/cost of
maintaining IGI compared to the deeper non-associated gas that is in greater demand. Since then, gas wells
have been flowing freely from Block A and Block B while maintaining the IGI scheme. With even with
conservation scheme halted, material balance shows the IGI scheme has still managed to sustain pressure
for the DC oil reservoirs (Fig. 27). This is also due to the fact that wells have been subsequently closed
in due to gassing out as the DC oil reservoirs are approaching the end of field life, hence voidage
replacement is approximately 1 even with reduced internal gas injection rate. Another effect of this is the
strong reduction of pressure from Block A due to high gas offtake. As a result, Block A and Block B have
established equilibrium with negligible inter-block crossflows (Fig. 28).
The MBal models are still providing a solid history match over 7 years after their creation (Fig. 28),
and are still being used as the basis for the operator’s forecasts. A detailed field and simulation study was
recently conducted for the Northern Area Deep reservoirs and concluded that the cross-flow concepts used
in the 2006 material balance study were valid. These factors combine to give confidence that the method
employed to generate these MBal models are robust and valid.

Conclusions and Recommendations

The Northern Area Deep reservoirs are a complex system with production performance dominated by
cross-flow between reservoirs, combined with compartmentalisation in the Block B-DG’s. A lateral
sand-sand connection in the DF3 resulted in cross-flow between Blocks A and B, and the rest of the
reservoirs are connected vertically via compromised well/completion integrity, commingled production
zones and the two internal gas injection wells in the A-DC’s. The production performance was history
matched by applying a novel method in MBal by adding and subtracting gas volumes to model
influx/efflux, between reservoirs/compartments; in effect a “sub-surface abacus”. The final MBal
matches, relative influx/efflux volumes, timing and directions are depicted in a summary montage in Fig.
29.
Conducting the preliminary p/Z analysis proved to be invaluable step in identifying cross-flow volumes
and directions and helped provide context to assist during the MBal modeling stage.
20 SPE-171488-MS

Although the history matching results are acknowledged to be non-unique, by honouring the pressure
data and maintaining volume balances between reservoirs the method is considered valid for determin-
istically calculating the GIIP’s and reserves. Despite these uncertainties it was the first time a fully
integrated model was generated for these reservoirs through the development of an auditable technique to
model the vertical cross-flow between the reservoirs.
This method (and corresponding results) was audited internally and externally and validated in 2006.
Subsequent history matching updates and reservoir studies have further validated these models and their
conclusions. It was recommended by Shell International Exploration and Production (SIEP) as a Practice
Worth Replicating (PWR). At the time of the original study, using MBal as a “sub-surface abacus” was
acknowledged by both SIEP and the software vendor (Petroleum Experts) to be the first time this method
had been used in modeling inter-reservoir cross-flow. This method is recommended to be used in
situations where non geological inter-reservoir cross-flow is occurring, provided the appropriate checks
and volume balances are applied. Two other studies have been successfully conducted in other fields by
the author utilising this method of modeling inter-reservoir cross-flow.
For the purpose of this paper, all confidential information has been removed by the authors. Data, maps,
well logs and well names have been modified so that no sensitive information is made public.