Professional Documents
Culture Documents
1) The Objectives place. By using a fixed group of people, local expertise can be
The responsible reservoir engineer defines the objectives of acquired relatively quickly and information is dissipated
the test. Being a producing gas field, the requirement is mostly efficiently.
determination of reservoir pressure and inflow performance
per layer. Additional issues include mainly the identification b) Standardisation
of water producing sands in commingled wells and the At the core of the interpretation processes are the reporting
completion integrity. Templates. There are two types: one for the PBU’s and one for
the MRPL’s. These templates combine the results of the
2) Type of Survey interpretation with all other relevant data: geological
This depends on the completion and test objectives. Wells in parameters, wireline logging analysis, production data, well
which a single reservoir is produced are surveyed with a and perforation history and results from previous pressure
classical Pressure Build Up survey (PBU). In case of very low measurements like wireline logging surveys and PBU/MRPL
influx or a dead well, a Static Gradient Survey is run mainly to surveys if any.
assess the wellbore pressure and wellbore content. Wells, Results and reservoir parameters are put into perspective by
which produce commingled from multiple reservoirs the use of specific Quality Indicators eg: A (excellent), B
simultaneously, are tested with a multi-rate production logging (good) and C (doubtful). These indicators facilitate comparing
survey (MRPL). Discussions in this paper are limited to the results of different surveys in the same or different wells
PBU’s and MRPL’s. and reservoirs.
The pros and cons of these surveys are: Once a report is fully validated by all the concerned parties,
PBU. Advantage: high (potential) accuracy in the the data are incorporated into a global database. The
determination of pressure, permeability and skin. consistency of the calculations and the interpretations as well
Disadvantage: not relevant for multi-layer reservoirs due to as the quality indicators allows fast and efficient correlation or
crossflow. Even when the pressures of a build-up seem to be data research for any particular requirement.
interpretable, there remains substantial uncertainty about the The process of interpretations and analysis and the use of the
relevance of the derived parameters due to questions like the templates are described in two userguides: one for the PBU
impact of crossflow and which layer is producing what. and one for the MRPL, ensuring that everyone uses the same
MRPL. Advantage: pressures and inflow performance of method for analysis. The content of the two guides is changing
multi-layer completions. Disadvantage: no distinction between over time as the process is refined and new knowledge about
permeability and skin, generally greater uncertainty in average specific issues comes available.
reservoir pressures compared to those derived from a PBU on
a single layer. c) Assessment of MRPL data
As the majority of the wells are producing commingled, Multi
3) Data Acquisition Rate Production Logging is the most appropriate tool to obtain
The survey is conducted according to a standard work reservoir pressure and rates.
program. Standardisation facilitates planning and Before embarking on the extensive MRPL campaign for
interpretation of the results, as it limits the number of variables monobore and standard completions, TFE conducted a study
in the process. Thus comparison of results from different to assess the validity of the results obtained from the MRPL:
surveys is simplified, as is the ongoing optimisation of 1. On the same well and within a very short time span, two
procedures. Close co-operation between the TFE personnel MRPL’s were done and interpreted by two different
and the Service Company responsible for the data acquisition Service Companies. A full set of MRPL consists of three
is considered to be of high importance to achieve consistency producing rates and one shut-in pass.
in the data quality and to optimise the different stages in the The tools had very similar response, but there were
process. differences in the interpretation, mainly due to a lack of
Tunu field knowledge and sharp understanding of well
4) Interpretations behavior. To fully compare these MRPL, both surveys
a) People were re-analysed by in-house interpretation engineers and
Interpretations are done by a dedicated group of experienced now the results of both surveys were very similar (see
engineers from TotalFinaElf and Kappa. The data is analysed figure 2). The differences in the initial results are
using industry standard software packages. These engineers therefore due to the different interpreters and the
are independent of the Service Company responsible for the difference in their methodology, not the tools and their
data acquisition and are located in the TFE office in sensors.
Balikpapan - Indonesia. This set-up allows them to have an 2. The reservoir pressures of each zone was compared with
independent view on uncertainty and limits associated with the the results of a single layer PBU for 2 wells. The reservoir
acquired data and objectives. Specific emphasis is placed on pressures from the Selective Inflow Performance curve
the collection of the knowledge of all involved people (eg. was within 200 psi (5%) of the PBU derived pressure.
geology, reservoir, well servicing, and production) in one
SPE72114 RESERVOIR MONITORING METHODOLOGY FOR A GIANT GAS FIELD 3
3. A methodology was developed to assess the source of Performance factor (c factor) and type of reservoir model
water production from individual layers. The very low (homogeneous, boundaries, layering etc.). The integration of
water hold-up values require a in-depth analysis, these results in the template allows an automatic estimation of
including a detailed study of the historical production the connected volume.
performance (rates, BSW, zone status, well head Two quality indicators are used to assess the results of the
pressures etc.) PBU analysis, one for the accuracy of the derived permeability
4. The viability of a multilayer transient test analysis using and a second indicator for the average reservoir pressure. The
the rate de-convolution method was assessed. With this Quality Indicator for the permeability is dependent on the
type of analysis, the permeability-thickness and skin can model’s match with the pressure data, i.e. reliability of the
be established for each layer. Although this method can model, quality of the pressure data etc. This value is based on
work in stable producing wells, the acquired sensor data engineering judgement. The Quality Indicator for the average
in our gas producing wells is too noisy to achieve reliable reservoir pressure is done automatically by comparing the
results. estimated average reservoir pressures from different reservoir
models and the latest measured pressure value in the Build-
Special Issues and Limitations of the MRPL are: Up.
a) For standard completions whereby several layers produce An example of the first page of the PBU template report is
via one entry port into the tubing, the flow can only be shown in figure 4.
assessed for the combined set of perforations. Except for
the bottom zone where the perforations flow directly into b) MRPL
the production casing (see figure 3). The MRPL data of all four Multi-Rate surveys are interpreted
b) The bottom zone in a standard completion is 7". In case of together using the same production analysis software
small gas inflow, it is likely to have a standing water (Emeraude). Results are then imported into the report template
column with gas bubbling through. The resulting data is and Selective Inflow Performance curves are plotted using
often so noisy that no reliable interpretation of small either surface rates or the rates as calculated by the production
producing sands can be done. The resulting density, logging program versus measured or pseudo pressures. This
capacitance and spinner fluctuations make interpretation gives an average reservoir pressure and a Productivity Index
especially challenging per layer. The different plots give the range of reservoir
c) The flow parameters from low productivity sands often pressure, therefore a range of uncertainty.
have large uncertainties. The actual contribution of the The consistency of the Selective Inflow Performance plot is a
layer is calculated by subtracting the flow values from quality check of the accuracy/limits of the interpretation (see
above and below the interval. As these values have their figure 5).
own uncertainty, the relative error in the difference is
proportionally very large. For this type of reservoir, the The following stage is an assessment of the connected volume
only way to obtain a rough estimate of reservoir pressure per layer.
is based on the shut-in pass: the reservoir pressure is
higher / lower than bottom hole flowing pressure if this c) Well performance history matching
reservoir is producing or taking fluids in shut-in ‘Predic’, a TotalFinaElf ‘multi-layer material balance’
conditions. software is used to perform the history match to assess
d) Even though the water detection methodology has been connected volume layer-wise on a well basis.
improved, it is often difficult to identify the source of
Inputs per layer are:
water when the water hold-up values in the measured part
of the conduit are very small. Typically the water hold-up • Initial reservoir pressure (wireline logging data)
is less than 3% in a gas filled tubing, and the sensor • porosity, net pay and depth
measurement can vary between different flowrates. • Present reservoir pressure (MRPL)
Alternatively, in the case of a standing water column, • Present flow contribution (MRPL)
where limited amounts of water flow into a fully loaded • Production data - well production history
casing section, the source of water might only be Matching parameters are:
determined using secondary qualitative information, e.g. • Split of cumulative production per layer
temperature, variations in liquid level during rates change
• Productivity Index per layer
and variations in density.
History match done on:
5) Integration • Well production profile
a) PBU: • Reservoir pressure per layer at MRPL time
All interpretations are done using the same transient analysis • Flow contribution per layer at MRPL time
software (Saphir). The parameters derived are average • Cumulative production per layer at MRPL time
reservoir pressure, permeability-thickness, skin, Inflow • Previous surveys (PBU, MRPL) on the well
4 I. POTAPIEFF, J. DE WIT SPE 72114
This manual history match is an iterative process to get the their own dynamic survey. The sum of these connected
best match on the performances with the observed production volumes is then compared with the volume from the
data (see figure 6 – 7). The present methodolgy has shortened geological model.
the history matching duration to less than one day per well.
This method assumes that the well behaviour is independent of This step is fundamental to improve the field knowledge and
other wells. This is a reasonable assumption as long as the especially to detect heterogeneities. Furthermore, it allows a
wells connected to the same layers have been producing under better estimate of the depleted sands for locating the future
the same conditions, i.e. the drainage area of the well under wells drilled; infill well locations can be proposed in non-
observation has remained constant over time (steady state drained areas.
behavior).
Quality indicators are assigned to the reservoir pressure
estimated layer-wise, the IPR of the reservoir and the Conclusions
connected volume derived after the history match. The MRPL logging is the most suitable tool for reservoir
management with commingled production like Tunu field. The
Once the history match is done, a production forecast can be methodology developed with the MRPL allows first deriving
performed to assess the impact of proposed future well reservoir pressure and production layerwise, then assessing the
servicing operations. connected volume per layer per well.
For a giant field with such complexity, standardisation,
The interpretation engineer, the reservoir engineer, the building local expertise and flexibility to adjust methods to
geologists and the well servicing department discuss the optmise the process are the key tool to properly cope with the
performance of difficult wells and the relevant interpretation amount of data.
results. Final conclusions are then drawn and The use of templates optimizes the reservoir monitoring by
recommendations made. consistent interpretations, by reducing the interpretation time,
by estimating the uncertainties and it allows integration of all
Following these recommendations, well servicing operations the relevant informations into a global database.
done have confirmed the interpretations. Few well examples: This process drives the identification of production and
- Confirmation of water source: isolation with a bridge plug reserves optimisation activities such as reperforation,
(+ 10 mmscf/d). additionnal perforations, acid, fracturation, water shut-off, grid
- Confirmation of high skins obtained: successful and infill well locations…
reperforation (+ 9 mmscf/d) and fracturation operation (+
9 mmscf/d). References:
1. C. A. Ehlig-Economides: ”testing and interpretation in
layered reservoirs”. September 1987. SPE 17089
6) Integration with Geological model 2. C. A. Ehlig-Economides: ”testing and interpretation in
So far, 350 reservoir pressures have been obtained from layered reservoirs”. SPE formation evaluation. September
MRPL’s and PBU’s data. These dynamic reservoir pressures 1993. SPE 20923
layerwise derived from PBU’s and MRPL’s are then included 3. Kucuk F, Karakas M and Ayestaran L: “Well testing for
in the geological model in order to improve the geological layered reservoirs”, SPE formation evaluation (Aug. 1986)
accumulation geometry. 342
For a given accumulation, the integration of static and 4. H. Vogelij, M. Leach and P. Kapteyn: “Multilayer
dynamic reservoir pressure versus time allows a very good flowmeter testing combined with 3D field modeleing to
understanding of the unit once all the wells have been fully enhance the Ameland field development”, SPE formation
processed (see figures 8 and 9). evaluation (Mar. 1993). SPE 20897
For a given accumulation, there is an estimation of the
connected volume per well once all the producing wells have
SPE72114 RESERVOIR MONITORING METHODOLOGY FOR A GIANT GAS FIELD 5
C o m pan y 1
Z on e 2
3000
C o m pan y 2
2800
2600
Z on e 3
Pressure (psi)
2400
2200
Z on e 1
2000
c o m m ing le d p ro d u c tio n
1800
-5 0 5 Q gas (M M scf/d ) 10 15 20
Figure 2: Comparison of Selective Inflow Performance curves on the same well at the same time
6 I. POTAPIEFF, J. DE WIT SPE 72114
Zone 1
Zone 2
Zone 3
7” casing
I. G ENERAL DAT A W eighted av erage v alues from the sev eral sands
Reserv oir res. 1 res. 2 T otal
Field : TUNU av erage phi [% ]
W ell : well 1 av erage Sw [% ]
Zone : 2 Av erage Vsh (% )
Reserv oir(s) : reserv oir 1 net pay [m ] (% )
Test date : 03/06/00 initial pressure: [psia]
Ref. G auge : m bRT/SS gauge… nb reserv oirs
top perfo : m bRT/SS G eological layer
bottom perfo : m bRT/SS Initial CG R
II. O BJECT IVE O F T HE BPT
Determ ine kh, skin and reserv oir pressure
III. FLO W PERIO D DAT A
Period DD 1 DD2 DD3 BU
Choke ../64 "
Duration hrs
W ell Head Flowing Pressure psia
W ell Head Flowing Tem perature °F
Separator pressure psia
Separator Tem perature °F
Q gas M M scf/d
Q condensate stb/d
Q water stb/d
Condensate yield stb/M M scf
W GR stb/M M scf
SG gas (air = 1 )
Condensate density g/cm ³
Bottom hole flowing pressure psia Present salinity
last PBU @ gauge depth psia ppm
Pm ax during PBU psia
M axim um bottom hole flowing tem perature °F
Q uality indicators: Best M atch Alternatives
IV. RESULT S kh & skin A Param eter M odel
Reserv oir pressure Bbarriers (m )
A v ery good - B good - C doubtfull kg.h m D.m Kcorrel
Inflow perform ance A&B A= B= kg mD
Assum ing n=1 the v alue for c = scf/psi^2 S total
Reservoir Pressure at gauge depth Pini (Saphir) dS/dQ (M M scf/d)-1
P* (superpos) Skin at q=0
last PBU Pav. res top gauge
av erage pressure at gauge depth psia top res.
gradient to top res. g/cm 3 D eq.
Cum ulative production bcf up to Jun-00 BPT date G IIP bcf
bcf up to Jun-01 last value G p Jun-00
eq. R. drainage
V. CO NCLUSIO NS / CO MMENT S
3500
re se rvo ir 1
re se rvo ir 2
3250
re se rvo ir 3
B H F P at to p reservo ir(p si
re se rvo ir 4
3000
2750
2500
2250
2000
-4 0 4 8 )
Q g (m m scf/d 12 16 20
6000
In itia l p re ssu re (M D T )
re se rvo ir 1
5000
re se rvo ir 2
re se rvo ir 3
4000
re se rvo ir p re ssu re s fro m M R P L re se rvo ir 4
P /Z
3000
2000
1000
0
0 5 10 15 20 25 30 35 40
C u m u la tive p ro d u ctio n (b cf)
80
60
P ressu re at d atu m (p sia)
5000
50
40
4500
30
20
4000
M D t m e as u re m e n ts
M R P L - P B U m e as u re m e n ts 10
3500 0
Feb-97 A ug-97 M ar-98 S ep-98 A pr-99 O ct-99 M ay-00 D ec-00 Jun-01