Petroleum & Coal

ISSN 1337-7027
Available online at www.vurup.sk/petroleum-coal
Petroleum & Coal 56(3) 332-338, 2014

SENSITIVITY & UNCERTAINTY ANALYSIS OF ORIGINAL OIL-IN-PLACE IN

CARBONATE RESERVOIR MODELING, A CASE STUDY
Arvin Khadem Samimi, Ghafoor Karimi
NIOC-Research Institute of Petroleum Industry, E-mail: [email protected]
Received May 5, 2014, Accepted June 10, 2014

Abstract
A systematic quantification of technical risks is crucial for decision makers to be able to compare or
rank possible projects of a global portfolio, or understand the risk level they are taking on any new
field development. A quantification of the impact of various subsurface uncertainties on economic may
also help justify the acquisition of further data, in order to reduce the uncertainty before major decisions are
made. All geostatistical methods used in estimation of reservoir parameters are inaccurate, modeling
of "estimation error" in form of uncertainty analysis is very important. Consequently, the application of
sensitivity analysis on stochastic models of reservoir horizons, petrophysical properties, and oil-water
contacts, also their effect on reserve, clearly shows any alteration in the reservoir geometry has significant
effect on the oil in place.
The studied reservoir is located at carbonate sequences of Ilam formation, Iran; it comprises Four Zones.
Zone 1-2 and zone 1-4 are main reserve and other zone have a little oil. Average porosity and water
saturation of the reservoir is about 25% and 43%. the uncertainty analysis result P90, P50 & P10 are
about 415*106 bbl, 600*106 bbl & 800*106 bbl.
Key Words: Original oil-in-place; Sensitivity; Uncertainty; Carbonate; Reservoir modeling.

1. Introduction
The decision to go for the project or not, or the choice between different field development
concepts, are made on mean values and not on deterministic reference values. Hence it is
important to estimate mean production, mean net present value, mean cash flow etc. for the
different alternatives and decisions. Understanding the volume uncertainty and estimating
the volume uncertainty range are important input for risk mitigation actions such as defining
flexibility in the field development plan and finding robust solutions that are optimal over a
range of outcomes rather than optimal on the reference case understanding of the field.
Uncertainty exists everywhere, such as (1) raw data measurements, (2) data processing
and interpretation,(3) structural modeling, (4) facies modeling, (5) petrophysical modeling
and (6)PVT analysis, which affects the ability to understand the reservoir behavior, making
reliable production forecasts and risk-free decisions.
Throughout the life of a hydrocarbon reservoir, from discovery to abandonment, a great
number of decisions depend on incomplete and uncertain information [2].Uncertainty is not
an inherent feature of our reservoirs; it is due to our lack of knowledge and understanding
about the reservoir [3]. This data shortage is mainly due to reservoir complexity and to the
high cost of data acquisition. Therefore, the prediction of oil-in-place mostly is uncertain [4].
2. Methodology
For volumetric computations, different types of volumes that can be computed are: the
gross rock volume (GRV), the net rock volume (NRV), the net porous volume (NPV), the
different fluid volumes (OIP and GIP), and the connected volumes using various connectivity
criteria. These volumes should be reported not only for the whole model but also for any regions
defined on the model.
Equation for Oil in Place at reservoir conditions (no gas):

A. K. Samimi, G. Karimi/Petroleum & Coal 56(3) 332-338, 2014

333

STOIIP GRV * NTG * * (1 Sw) .

The major elements here are the Gross Rock Volume (GRV), consisting of a structural piece:
horizons and faults - properly connected together - and fluid contacts; the Net-to-Gross (NTG);
the net Porosity (), and the Water Saturation(Sw).
Identifying which of these elements is more important than the others will help focus resources
on relevant issues, stay at the right level of technical detail, and in turn save time and money.
The main objective of this study is to uncertainty quantifying of the objective function. To
meet this goal, the following procedure should be done.

Define
Uncertainty
Distribution and
Define

Ranges

Sensitivity

Uncertain

(2)

Analysis by

Parameters
(1)

Variable

STOIIP

(3)

Sensitivity
&
Generate

Uncertainty

Identify Most
Influence

Result,

Analysis
P10,P50

Parameter
by Tornado Chart

&

Uncertainty

P90

Analysis by

(6)

Mont-Carlo

(4)

sampler
(5)

Figure 1 Sensitivity & Uncertainty Analysis workflow of STOIIP, in six steps

3. Uncertain Parameters
When building a geological model one has to go through a thorough subsurface evaluation
including interpreting seismic data, performing time-to-depth conversion, interpreting log
data, establishing geological conceptual models, building a 3D reservoir model, and estimating
fluid contacts and fluid types in the different areas. Then it is important to acknowledge that
at each step there are uncertainties, both in the data that are used, the selection of which
data to include, and in the interpretation of the data. In an integrated uncertainty analysis it
is important to include all these uncertainties in the value chain from seismic interpretation
to volume and value calculation.
The uncertainties included in the uncertainty analysis of in-place volumes are:
1. Structural Uncertainty
2. Fluid contacts
3. Petrophysical Properties
4. PVT as Bo, Bg
3.1 Structure Uncertainty
The main sources of structural depth uncertainty include (1) velocity model; (2) time
interpretation pick; (3) isochore thickness; and (4) fault interpretation. Uncertainty in velocity
model can be defined by assigning probability distributions to the parameters of the model
(V0, k, etc). The input parameters can be constants, functions or maps. Velocity uncertainty
maps are particularly useful for incorporating realistic lateral variations such as increased
uncertainty away from well data and down dip. These maps can be based on well miss ties,

A. K. Samimi, G. Karimi/Petroleum & Coal 56(3) 332-338, 2014

334

analysis of stacking velocities and general time interpretation uncertainty is usually defined
by small time shifts of interpreted horizons. The time shift can either be a constant or can vary
laterally to reflect varying quality of the seismic response or increasing uncertainty over major
faults.
Isochore uncertainty is handled in a similar manner to the time interpretation uncertainty.
Constant thickness multipliers or thickness uncertainty maps are applied to base case isochore.
Horizon depth uncertainty also needs to be linked to the fault model and uncertainty in
velocities and depth conversion should be handled in a consistent manner for both depth horizons
and fault surfaces.

Figure 2 An uncertainty map has been added and removed (dotted lines) to a reference map
(thick continuous line)
3.2. Fluid Contact
It is important not only to include the uncertainty in the contact given the underlying assumptions, but also the uncertainty in the underlying assumptions.
1. If an oil-down-to is observed the maximum contact depth is the spill point. However, the
depth (and position) of the spill point is uncertain due to uncertainty in the depth
conversion model. Presence or absence of sealing faults may change the definition of spill
point drastically.
2. In some cases the gas-oil contact is observed and pressure measurements and fluid
samples are taken from the oil zone but no samples or measurements are made in the
water-zone. If the neighboring segment has pressure measurements in the water zone it is
assumed, as a reference case, that the water zones in these two segments are in pressure
communication. Then the oil-water contact can be estimated with the uncertainty defined by
the uncertainties in fluid density and pressure measurements. However, the assumption
that the water is in pressure communication should also be challenged and included in the
uncertainty evaluation giving a complex uncertainty picture and large spans.
3.3 Petrophysical Properties Uncertainty
Reservoir model porosities were derived from log porosities calibrated to core measurements.
Uncertainty resulting was obtained by considering an areal distribution of porosities. The
porosities considered are vertical averages over reservoir thickness. Porosity is a key component
of reservoir Pore Volume and hence of the resource base for the field. Since the porosity defines
the volumes of moveable fluids in the reservoir, measurements of the effective porosity of
the reservoir rock is also required. There are two important sources of porosity uncertainty.
The first is related to logging tool measurement, processing and interpretation. The second
is related to the sampling of wells. Additionally, the impact of internal porosity trends is
independent of the relationship of the trends to the structure and contacts, so they need to
be quantified separately.
As with porosity related uncertainty, there are two main sources of water saturation uncertainty.
3.4. PVT as Bo, Bg
Fluid samples are retrieved from most of the segments with wells. Here the uncertainties
in the expansion factors Bo and Bg are relatively small and mainly due to uncertainty in the
lab-results. The fluids change significantly across the field, and there is a large uncertainty

A. K. Samimi, G. Karimi/Petroleum & Coal 56(3) 332-338, 2014

335

in segments with no fluid samples. This uncertainty is estimated using information from the
drilled wells and the variation in fluid properties across the field.
4. Sensitivity analysis
Design sensitivity analysis plays a critical role in inverse and identification studies, as well
as numerical optimization, and reliability analyses. Before Uncertainty analysis and determine
most influenced uncertain parameters, it is important to make sensitivity analysis. Sensitivity
analysis is frequently performed to gain a better understanding of the influence of variables
or parameters on the distributions of uncertainty. The purpose of sensitivity analysis is to
identify those parameters and/or processes that strongly influence simulated results of the
given model, and further analyze the trends of these correlations. Tornado chart is sensitivity
analysis result and show that the major elements. Petrel offers two separate sensitivity tasks
in the Uncertainty and optimization process:
4.1 Sensitivity by variable (Uncertain, SEED)
The variable-based sensitivity of the given model is determined by successively selecting
one variable at a time from the set of all uncertain variables, and changing its value while
keeping the others fixed at their Base values. This is done for each uncertain variable in turn,
so the total No. of runs thus equals to the number of uncertain or SEED variables multiplied
by the No. of samples per variable that has been entered by the user. Note that Expressiontype variables are not directly sampled to test their sensitivity because they may depend on
other $-variables that are tested.
4.2 Sensitivity by process
The process-based sensitivity of the given model is determined by altering the variables
of one process at a time. This differs from Sensitivity by variable as several parameters may
be defined within a single process. When this task is chosen, the total No. of runs equals to
the number of parameterized processes multiplied by the value entered in the No. of samples. If
multiple variables are defined in one process, all these variables will be active together.
5. Case study sensitivity & uncertainty analysis
Spatial 3D modelling was performed with the use of the Petrel software. This case data
only one exploration well with full sets logs and three fluid samples. Structural model was
created by interpreted UGC with 100 x 100 m grid increment. Five proportional inner layers
were set, in order to divide the model into over 297460 cells. The studied reservoir is located
at carbonate sequences of Ilam formation, Iran; it comprises Four Zones. Zone 1-2 and zone
1-4 are main reserve and other zone have a little oil. In this case oil-water contact is oil-downto and the Spill Point value were selected to be -3475 m. Average porosity was counted from
Deterministic(Moving Average) porosity model. The same procedure was used to count water
saturation and permeability distribution. Average porosity, water saturation and permeability
of the reservoir are about 25%, 43% and 28 md.
Basic volume parameters were calculated by the Volume calculations module in Petrel.
These parameters include Bulk volume, Net volume, Pore volume and Hydrocarbons Per Volume
(oil in our case) (HCPVo).
Using the Uncertainty and optimization module in Petrel, first define uncertain parameter
such as Structure, oil-water contact, petrophysical model(porosity & water saturation) and Bo.
Second define uncertain parameter distribution and ranges, third analyzed sensitivity by
variable and obtain more influenced parameter in volumetric uncertainty. For oil-water contact
select triangular distribution and define three depth, spill point(-3475 TVDSS), base Ilam
reservoir(-3391 TVDSS) and reference depth (-3440 TVDSS) to reduce volumetric uncertainty.
Using Monte-Carlo sampler(Latin-hypercube sampling) with 2000 No. of samples from their
assigned distributions was used for this purpose.
Table 1 contains a summary of the input distributions type and their ranges for case Model.
Low and high values of factors were obtained.
Implementation of Sensitivity analysis on case model using uncertain parameter described
before would yield the tornado chart of figure 3 where it is confirmed that oil-water Contact
(WOC) and petrophysical parameters (porosity and water saturation model) have largest

A. K. Samimi, G. Karimi/Petroleum & Coal 56(3) 332-338, 2014

336

impact on our original oil-in-place calculation. Tornado charts are great to visually summarize
information and more precisely the impact of various plotted parameters. They are excellent
decision-support tools for that reason.
Table 1 Uncertainty and sensitivity analysis variable distribution and ranges

Monte-Carlo sampler(Latin-hypercube sampling) with 2000 No. of samples from their assigned
distributions was used for uncertainty analysis. It investigates the effect of the various variables
simultaneously in a series of experimental runs.

Figure 3 Sensitivity analysis result in tornado chart

A specific combination of properties (input variables) at different levels that make up multiple
realizations of the volumetric reports. The results were analyzed to obtain the relationship
between the input variables and the output responses.
The results can be shown in histograms (figure 4) and table (table 2Table 2) in which
probability level of 10%, probability level of 50% (reality-oriented model) and probability
level of 90% have been identified and CDF curve can be depicted. This way we can study
performance of that parameter on the amount of oil.
Table 2 Volumetric uncertainty analysis result
P90

66.106 sm3

415.106 bbl

P50

96.106 sm3

600.106 bbl

P10

127.106 sm3

800.106 bbl

A. K. Samimi, G. Karimi/Petroleum & Coal 56(3) 332-338, 2014

337

Figure 4 STOIIP CDF curve of different runs for Ilam reservoir volumetric calculation
6. Conclusion & recommendation
Sensitivity analysis on case model using uncertain parameter described that oil-water contact
(WOC) and petrophysical parameters(porosity and water saturation model) have largest impact
on our original oil in place calculation.
The results can be shown in histograms in which probability level of 10%, probability level
of 50% (reality-oriented model) and probability level of 90% have been identified and CDF curve
can be depicted. This way we can study performance of that parameter on the amount of oil.
It is necessary to rapidly know what actions are required to reduce their uncertainty to an
acceptable level. For a reservoir modeling point of view, that would mean: (i) refining the
interpretation, (ii) refining the model, or (iii) gathering more data because interpretation and
modeling uncertainty cannot be refined any further with existing information.
References
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]

Charles, T., Gumn, J.M.: "Experience with the Quantification of Subsurface Uncertainties," in SPE Asia Pacific Oil and Gas Conference, Jakarta, Indonesia, 2001.
Corre, Thore, B. Feraudy, V. P. de and G.Vincent, "Integrated Uncertainty Assessment
For Project Evaluation and Risk Analysis," in SPE European Petroleum, Paris, France,
2000.
Gringarten, E.:"Uncertainty Assessment in 3D Reservoir Modeling: an Integrated
Approach," Earth Decision Sciences.
Boer, L. den, J. Poncet, P. Biver, P. Henriquel and V. Marlot, "A New Technique To
Achieve History Match Using a Probabilistic Approach," in SPE Annual Technical
Conference, San Antonio, Texas, U.S.A, 2006.
Colina, J. D. L.: "Integrated Static Model Uncertainty Analysis of the Lower Ben Nevis I
Reservoir at the Hibernia Field," in Chevron Canada Resources,, Calgary, AB, Canada,
2007.
Samson, P., Gumen, J.-M., Robbe, O., Feraudy, V. D., Rossi, T., Larsonneur, J.-L.,
Bez, M., Bourdat, M., Larue, D.:, "3D Modeling and Reservoir Uncertainties: A Case
Study," in Chevron Petroleum Technology.
Caumon, G., Mallet, J.:"3D Stratigraphic models: representation and stochastic
modelling," in International Congress, Belgium, 2006.
Soleng, H. H.,. Rivens, J. C., Gjerde, J.:"Structural uncertainty modelling and the
representation of faults as staircases," in 9thEuropean Conference of the Mathematics
of Oil Recovery, 2004.
Rivens, J. C., Otterlei, C., Zachariassen, E., Dart, C., Sjholm, J.: "A 3D stochastic
model integrating depth, fault and property uncertainty for planning robust wells,

A. K. Samimi, G. Karimi/Petroleum & Coal 56(3) 332-338, 2014

[11]
[11]
[12]
[13]
[14]
[15]
[16]

338

Njord Field, offshore Norway," Petroleum Geoscience, vol. 11, no. 1, pp. 57-65, 2005.
Mohaghegh, S. D., Hafez, H.:"Uncertainty Analysis of a Giant Oil Field in the Middle
East Using Surrogate Reservoir Model," in International Petroleum Exhibition and
Conference, Abu Dhabi, 2006.
Schlumberger, "Petrel Introduction Course," Seismic-to-Simulation Software Petrel
introduction course 2008, 2011.
Lelliott, M. R., Cave, M.R. Wealthall, G.P.: "A structured approach to the measurement of
uncertainty in 3D geological models, Quarterly journal of Engineering Geology and
Hydrogeology," vol. 42, pp. 95-105, 2009.
Zebalza-Mezghani, I., Manceau, E., Feraille, M., Jourdan, A.:"Uncertainty management:
from geological scenarios to production scheme optimization, " Journal of Petroleum
Science and Engineering, vol. 44, pp. 11-25, 2004.
Sahni, A.: "Case Studies of Uncertainty Analysis in the Seismic to Reservoir Simulation
Workflow," in SPE Annual Technical Conference , Denver, Colorado, U.S.A, 2003.
Bueno, J. F., Drummond, R. D., Vidal, A. C., Sancevero, S. S.: "Constraining uncertainty in
volumetric estimation: A case study from Namorado Field, Brazil," Journal of Petroleum
Science and Engineering, vol.77, pp. 200-208, May 2011.
Pocaterra, S. S. B.: Comparison of sampling methods for uncertainty evaluation in
reservoir flow predictions, University of Tulsa, 2000.