A SELF

TEACHING EXPERT SYSTEM FOR THE
ANALYSIS, DESIGN A
ND PREDICTION OF GAS PRODUCTION
F
ROM
UNCONVENTIONAL
GAS RESOURCES
Lawrence Berkeley National Laboratory
(LBNL)
Texas A&M University
(TAMU)
University of Houston
(UH)
Task 2.0

Technology Status
Assessment
The ensuing analysis encompasses assessments of the status of the various technologies
that are involved in the project.
A.
Development of a Data Abstraction, Reduction and/or
Compression (DARC) Technology
The problem with interpretation of 3

D
seismic and well

log data is that, although the
reflections in seismic data correspond well to geologic structure, and there is a general
belief that information is available in the data regarding rock type, porosity and so on,
particularly when well

log
data can be combined with seismic, the systems the data
represent are so complex and varied that it is impossible to come up with a complete
physical model that is precise enough to make accurate inferences. On the other hand, it
is known from experience
that trained experts
can
sift through the evidence well enough
to place drill holes with reasonable confidence.
Since humans are able to find correlations in data with reasonable success, most standard
packages such as Kingdom Suite, Geographix, Seiswor
ks or Charisma are largely data
visualization tools allowing the user to examine colored surfaces and cross

sections
through a 3

d volume. They advertise features such as
semi

automated horizon tracking
in which a particular feature can be followed throug
hout the volume,
integration of well

interpretations
, in which lithologies or other rock properties such as porosity that are
known from well

logs can be combined or overlaid with the data set. The real challenge
addressed by these packages is data manag
ement, since a medium sized 3

d seismic data
set can run into the 100s of gigabytes. Their job is to present data in such a way that
humans can make correlations and inferences and finally management decisions. Very
little is done automatically by the so
ftware.
When automatic correlation is attempted, it is usually through
attribute analysis
in which
particular functions of pre or post

stack data are computed such as Fourier envelopes,
average energy, reflection strength, and amplitude variation with off
set (AVO). The
desire is to use these as predictive variables for reservoir characterization. The reality is,
however, that attributes or combinations of attributes are usually used qualitatively by
matching them with known reservoir properties by eye un
til the picture looks right.
February 9, 2009
Page
2
There are, however, some attempts at quantitative correlation, by regression of attributes
against known reservoir properties, by selection of attributes according to physical or
empirical criteria, or by training neural netw
orks to output parameters of interest based on
an input suite of attributes. In the end, however, it is a human who makes the decision
as to the most effective combination of attributes and architecture of the learning system.
It is very difficult for p
eople to take more than a few variables into account and
intuitive
ly design an effective computer

learning algorithm.
The following illustrate several state

of

the

art methods for automatic, comprehensive
data interpretation.
A.1.
State of the art: T
he Smart Field project (Stanford University)
The Smart Field consortium at Stanford University includes a group of oil and service
companies: BP, Chevron, Conoco Phillips, Halliburton, Petrobras, Saudi Aramco, Shell,
Total, StatoilHydro, Norwegian Univers
ity of Science and Technology, BakerHughes
and Computer Modeling Group.
The term “Smart Field” refers to the real

time monitoring, model updating and optimal
control of oil and gas fields. The main aim is using optimization techniques to determine
the l
ocation, number and type of wells, in a preexisting or new field. Obviously these
techniques are based on several types of information acquired using different monitoring
techniques: seismic data, well log data, geology cores, production data (pressure,
te
mperature, stress and flow rates, etc).
Thus, this procedure has to be dynamical and iterative depending on the degree of
information. The implementation of such advanced technologies in large and complex oil
and gas fields requires several key developme
nts:
1.
Optimization techniques for determining initial field development and continuous
development over the life of the field. This includes determining where and when
to drill, the design of the wells to be drilled and the type of monitoring needed
Also, w
hat kind of monitoring will be needed and when? Risk assessment and
decision

making techniques will have to be developed for aiding management.
2.
Efficient and reliable techniques for data filtering and assimilation.
3.
Procedures to update and calibrate vario
us models (e.g., geological, relative
permeability, thermodynamic description of fluids, and multiphase flow in wells
and facilities) in order to improve future predictions.
4.
Very fast reservoir simulators and
suitable
proxies
for spanning
the space of
possibilities for finding the optimum.
5.
Effective integration of different types of geophysical data (e.g. seismic, gravity,
electro

magnetic, etc) into the reservoir updating and
optimization process.
6.
Integration of all of the above tools in the closed

loop system
(F
igure 1).
February 9, 2009
Page
3
A.2.
Focusing on methodology
: The role of the reduced Order Models for
Production Optimization
(M.A. Cardoso and L. Durlofsky, Stanford University).
Progr
ess in seismic and well log tools resulted in exceptionally detailed geological
models and, as a consequence, more complex reservoir flow models.
Maximizing the
hydrocarbon production or the Net Present Value (NPV) of a reservoir in a Smart Field
can be ac
complished by applying Optimal Control Theory. This usually requires many
simulations of the reservoir model. The time needed to calculate optimized controls
increases with the number of controls and with the complexity of the reservoir model.
Model order
reduction (MOR) techniques can be used to overcome this issue. MOR
generates simple models based on a detailed model of the system under study. A reduced
order model
uses a considerably smaller number of variables or states than the original model
to descr
ibe a particular system,
is relatively inexpensive to simulate, when compared to the original model.
Principal Component Analysis (PCA) one of the most popular model reduction
techniques for large scale models. The aim is to control the principal variab
les involved
on the optimization process, inferring the others by the correlation between them. The
main idea is to account for simplicity on the mathematical model, and thus on the expert
system.
Figure 1: A possible flowchart of a “smart field” syst
em.
February 9, 2009
Page
4
A.3.
The role of geostatistics
Geostatistics has been proved to be the right tool to account for heterogeneities on the
reservoir, through the conditional simulations techniques
that
are
able to generate
isomorphic reservoir model properties having a
prescribed spatial
correlation model and
honoring the hard information (wells). A
dditionally,
these techniques can be used to
account for uncertainty estimation (indicator kriging) and to generate Regionalized
variables taking into account some shape info
rmation coming from external sources, such
as the geophysical methods. Important variables such as porosities, permeabilities and
saturations can be geostatistically modeled using these tools. These conditional fields can
be input in a reservoir model tran
sfer function to account for production uncertainties.
Other applications can be imagined since in Petroleum engineering we deal we spatially
distributed processes.
A4.
The role of inverse problems
Inverse problems
(IP)
are always present in engineering a
nd are a key point to make the
methodology successful.
Global optimization algorithms are very important since they
are able to model processes described by a known physical model or based on empirical
relationships or even experience.
In particular,
Part
icle Swarm
Organization
(PSO)
methods, o
n which
the UCB team
i
s actively
involved, give
a good performance rate and
allows posterior analysis of uncertainty from the model

samples found on the low

misfit
area.
Other methods such as Genetic algorithms or
simulated annealing can be applied but they
are much slower. Some preliminary studies for the application of PSO to production data,
using well log information, cross hole tomography (transmission and diffraction) and
reservoir modeling techniques seems to
give very impressive results using a reduced set
of model parameters found after PCA. A major issue related to the use of global methods
is the curse of dimensionality.
Local optimization methods can provide for an initial guess, but they are not able t
o
account for uncertainty, which is a major issue in presence of noise when no prior
information is available.
Some of the inverse problems involved production data (history
matching) and other seismic and petrographic attributes (tomography, AVO, etc
.
).
These
last methods are important to account for constraints on the reservoir model.
Additionally,
it is important
to highlight the role of multi

scale inversion using wavelet
basis, which can help to solve complicated inverse problems in a
n
efficient way,
reducing
the number of involved parameters. In fact
,
the IP is solved for different scales and the
regularization techniques are based on the degree of available information.
A5.
The role of kernel methods
One of the main reasons why expert systems faile
d is that they are not able to learn from
experience, i.e
.
, they are based on a restrained and static set of rules that cannot be
changed. In this sense Kernel Methods
(KM)
and Support Vector Machines
(SVM)
can
February 9, 2009
Page
5
be used as the right tool to interpolate from
experience, when the kernel is rightly chosen.
There is work in progress
at UCB
to develop
a methodology to design the kernel. Kernel
methods can be used to solve, both, the forward and inverse problems.
KM and SVM can also be used to solve classificati
on problems and to learn relationships
between parameters. This is very important, because weak relations between parameters
are always needed. The success of our methodology can rely partly in learning these
relationships and to define the main clusters o
f variables and how to process them in
order to condition the final result.
A.6.
Accounting for uncertainty
Most of (all) the information involved in the modeling process has an associated
uncertainty, and this is important to input it on the modeling pro
cess in order to asses
s
risk.
Core data can be considered hard information, but
their
support is linear, i.e.,
they
are
only valid on the neighborhood of the borehole.
Log information can be considered
as certain information with a pdf
,
which can be larg
er or narrower
,
depending on the
method and on the acquisition process.
Geophysical information is soft, in the sense that noise is always present and is suited for
interpretation. Sometimes the uncertainty of one parameter (time) is transmitted to other
related parameter through an algebraic technique or mathematical method (including
inverse problems).
A major contribution of
the
methodology
proposed in this project
will
be to find a way to quantify and input (even from expertise) this uncertainty.
A.7
.
Graphical Models including Bayesian Networks
Graphical models can be used to represent complex physical and statistical systems.
These are graphs in which the nodes of the graph represent variables and the links
between them represent dependencies, prob
abilistic or physical, known or assumed. The
nodes can be observed (
data
) or
hidden. The links can be directed (Bayesian networks or
belief networks) or undirected (Markov models). The structure of the graph can be
known entirely or only partially. Ass
embling such a graph allows an expert to inject
knowledge in a way that is simple and intuitive and does not require complete
understanding of the dependencies system.
Most learning algorithms can be written out as graphical models where the learning
prob
lem is to use training data to solve for parameters in the links of the graphs (such as
parameters of probability distributions) and to figure out what the values of
the hidden
nodes must be.
Usually, the links have a notion of probability built into them
, and the
Graph is used not only to produce answers, but
also
to provide posterior distributions.
F
or instance, a “forwa
rd model” could be drawn out as:
February 9, 2009
Page
6
A Kalman filter is drawn
as:
Graphical models can be much more complicated than these examples a
nd still solved.
Tools such as the BLog (Bayesian Logic) language under developm
ent by MIT and UCB
give general

purpose frameworks developed at UCB for setting up and
solving such
problems.
Depending on the application and architecture of the model, th
ey can be
solved in many different ways. Inference on very large graphical models with many
nodes and links can be slow, however many techniques for approximating them exist,
notably sampling using Monte Carlo simulations and Variational methods in which
large
sums of random numbers are replaced by their means to reduce the number of nodes in
the graph. In some solution, it becomes apparent that the data cannot be explained by the
structure of the graph and new hidden nodes may be added.
A simple graph
ical model designed to solve the seismic data interpretation problem
would have data in its observed nodes: subsets of seismic traces, physically derived
attributes, and well log observables. It would have a set of hidden nodes representing
geologic param
eters such as lithology, porosity, and hydrocarbon content. The links
between the nodes could by physical, such as Darcy’s law linking porosity and
hydrocarbon flow. They could be statistical, such as the probability of a change in
bedding linkin
g one li
thology to the next. Alternatively,
they could be simply intuitive,
representing the expert’s opinion that one phenomena is linked to the next.
B.
Analysis of Production From Unconventional Tight Gas Systems
A thorough listing of the pertinent litera
ture can be found in the Section “References”
(Sections B.1, B.2 and B.3). Here we provide a summary of the most important findings.
B.1
.
A
nalysis of production
from
coalbed methane
systems
The early attempts on characterizing the production from coalbed
methane reservoirs
included analytical/semi

analytical solutions of the radial diffusivity equation with the
incorporation of a source term (
i.e
, the Langmuir desorption term) for the matrix/fracture
February 9, 2009
Page
7
transport (see Bayles and Reznik (1986), Ertekin and Su
ng (1989), and Kolesar
et al
(1990)). Production decline curves and decline type curves have also been employed in
the analysis of production data from coalbed methane reservoirs (Mohaghegh
et al
(1995), Seidle (2002), Mavor
et al
(2003), Aminian
et al
(2
004), and Rushing
et al
(2008)). Analysis of the pressure transient data of hydraulically fractured wells and
horizontal wells in coalbed methane reservoirs can also be performed using the solutions
proposed by Anbarci and Ertekin (1992) and Sarkar and Ra
jtar (1994). Recently,
Clarkson
et al
(2007) describe an integrated approach making use of production data
analysis (
PDA
), pressure transient analysis (
PTA
), analytical/numerical simulation, and
flowing material balance (
FMB
) in conjunction for analyzing
the production from single

phase (dry) coalbed methane reservoirs. This work has been extended to several possible
flow characteristics such as single

phase flow of water, and two

phase flow (Clarkson
et
al
(2008), and Clarkson (2008)) in coalbed methane
reservoirs.
B.2
.
A
nalysis of
gas
production
from
tight sand
reservoirs
Type curves have been a continually developing method to characterize the behavior of
production from tight gas wells (Thompson (1981), Maley (1985), Neal and Mian (1989),
Cox
et al
(1
996), Amini
et al
(2007), and Ilk
et al
(2008b)). The use of stimulation
treatments, primarily hydraulically fracturing, has led to the modification of methods
used to determine the production performance of tight gas wells (Sinha and Furlong
(1979), Hale
(1986), Medeiros (2007), and Medeiros
et al
(2007)). Fracture treatments
performed on tight gas wells have been analyzed to determine the treatment that results in
the optimal production performance from the reservoir (Roberts (1981), Warpinski
et al
(1990
), Ilk
et al
(2007), Cipolla
et al
(2008), and Warpinski
et al
(2008)). Current work
related to tight gas sands includes the characterization of low

permeability systems and
determining methods to estimate reserves (Ilk
et al
(2008a), Blasingame (2008), Ru
shing
et al
(2008), and Rushing
et al
(2007)).
B.3
.
A
nalysis of
gas
production
from
shale reservoirs
Early work focused on studying naturally fractured reservoir has provided to be the
foundation on which shale gas reservoirs are described (Warren
et al
(
1963), De Swaan
(1976), Najurieta (1980), Kucuk and Sawyer (1980), Cinco

Ley and Samaniego (1982),
Serra
et al
(1983), and Curtis (2002)). In addition, dual

porosity models have been
developed to describe shale gas plays (Carlson and Mercer (1991), Hazlett
et al
(1986),
and Javadpour
et al
(2007)). Gas desorption has proved to be essential to understanding
the productive capacity of shale gas reservoirs (Bumb and McKee (1988)), and the
development of shale gas plays including the Barnett Shale and the Antri
m Shale has
resulted in improved production analysis methods and a greater understanding of shale
gas reservoirs (Zuber
et al
(1994), Frantz
et al
(2005), Pollastro (2007), Bowker (2007),
Lewis and Hughes (2008), and Mattar
et al
(2008)).
C.
Quantificati
on of Uncertainty
February 9, 2009
Page
8
Confidence intervals on predicted system behavior are necessary for design and
optimization of any engineering system. They are also useful from a scientific point of
view, as model validation with respect to experimental measurements re
quires careful
measures of uncertainty in both experimental data and computational predictions. In
general, uncertainty in computational results can be due to both model and parametric
uncertainty. This discussion deals with only the latter. Empirical phys
ical parameters
used in the construction of computational models are inherently uncertain as a result of
unavoidable measurement errors, knowledge gaps in the understanding of subsurface
geology, and the inherent heterogeneity of geological systems. Uncert
ainty quantification
(UQ) goes hand

in

hand with the data reduction and compression needs of this project, as
methods that parameterize data or system variables in terms of (possibly uncertain)
parameters also provide a way to create response surfaces that
represent complex data
sets via analytical expressions.
In general, the propagation of parametric uncertainty can be studied using Monte Carlo
(MC) simulations by direct sampling assumed or known distributions of model
parameters over many iterations. F
or large data sets or complex simulations that
consume considerable computational resources, this method may be very costly and
inefficient. In addition, this approach does not readily provide information about the
sensitivity of model outputs to specific
parametric uncertainties. Conventional sensitivity
analysis may shed some light on first

order parametric dependencies, but will not
propagate nonlinear interactions through the model of interest or represent those
interactions in a response surface mode
l.
An alternate approach is based on a spectral stochastic description of uncertain
parameters and field quantities using polynomial chaos (PC) expansions for the stochastic
representation of uncertainty. Polynomial chaos is a member of the set of homogen
eous
chaos, first defined by Wiener (Wiener, 1938). Ghanem and Spanos (1991) implemented
a spectral PC expansion in terms of Hermite polynomials of Gaussian basis functions in a
finite element method. This was applied in the modeling of transport in poro
us media
(Ghanem, 1998), solid mechanics and structural applications. Le Maitre and collaborators
(2002) extended the application of these techniques to thermo

fluid applications in the
context of low Mach number flow. Debusschere (2004) used PC for UQ in
the context of
electrochemical flow in microfluidic systems, while Reagan (2003; 2005) studied UQ in
the context of chemically reacting flow. Later work expands the projection scheme into a
multi

resolution approach using multi

wavelets (LeMaitre, 2007) cr
eating a generalized
spectral decomposition and solution strategy for stochastic problems and a robust means
of expressing real

world data and system responses in a compact form.
In the most basic application, using MC sampling of the stochastic parameter
s, the
corresponding solutions of the deterministic system are evaluated and projected onto the
PC basis to compute the spectral mode coefficients. These coefficients are then used to
construct probability density functions (PDFs) of the solution, to infe
r sensitivity to
various parametric uncertainties, and to highlight the dominant sources of uncertainty.
This non

intrusive spectral projection (NISP) approach (LeMaitre, 2002; Reagan, 2003),
February 9, 2009
Page
9
has the advantage of being applicable to legacy codes or models
, which are run with
varying parameters to compute the statistics and spectral mode values via sampling. It
also provides an easy method to sample large sets of existing data to generate a spectral
representation.
For computationally intensive problems, l
arge

scale MC Sampling of many deterministic
runs may not be practical. For example, the problem of homogeneous ignition chemistry
using a reduced model of 8 uncertain reaction rate preexponentials and five uncertain
enthalpies of formation requires over
20,000 individual deterministic evaluations of the
model to adequately sample the full stochastic space and reach convergence (Reagan,
2003). Increasing the order of the PC expansion to more accurately represent the
resultant output PDFs and/or study larg
er sets of uncertain parameters can dramatically
increase the number of samples required to achieve convergence. More complex models
containing a greater number of reactions increase both the required sampling and the time
required to compute individual r
ealizations. As a result, more complex problems demand
a more efficient approach.
An “intrusive” spectral/pseudospectral methodology allows the direct incorporation of
spectral stochastic information into the basic formulation of a model (Le Maitre, 200
4;
Reagan, 2004; 2005). If it is possible to reformulate the governing equations for a
particular problem, numerical efficiency can be gained by creating a purpose

built
spectral code. In general, these implementations of spectral PC expansions involve (1
) the
introduction of a new stochastic dimension for each uncertain parameter in the problem,
(2) the expansion of parameters and field quantities using PC in terms of these stochastic
dimensions, (3) the substitution of these expansions in the governing e
quations and their
reformulation using a Galerkin projection procedure into equations for the stochastic
mode strengths, and (4) the solution of this larger system of equations and the
reconstruction of the field quantities of the solution based on their P
C expansions in
terms of the computed stochastic modes. The computational effort required to solve this
system can be many times smaller than that required to generate many MC realizations,
depending on system nonlinearities and the necessary spectral ord
er. Libraries of spectral
“overloaded operators” allow for quick conversion of deterministic codes to PC

expanded operations.
It is important to note that, while polynomial chaos

based uncertainty quantification
provides sensitivity information, it actual
ly goes well beyond sensitivity analysis to
propagate the full probabilistic representation of the model inputs to the model outputs.
Depending on the chosen order of the PC expansion, intrusive spectral methods also
provide higher

order sensitivity inform
ation. Higher

order effects are not lumped into a
single coefficient, but are considered independently and in terms of parameter

parameter
interactions. The result is a complete polynomial chaos representation of the complex
system response. This respons
e surface reduces the complex behavior of a given variable
to an analytic expression in terms of not only all dependent parameters, and also accounts
for the uncertainty associated with each of those parameters and the relative contribution
of each paramet
er to the total uncertainty.
February 9, 2009
Page
10
D.
Data Integration and Joint Hydrogeological

Geophysical
Inversion Technology
Site

specific information about (1) formation properties and (2) the initial system state
needs to be incorporated into numerical flow and transp
ort models that are used for
predictive simulations of reservoir conditions in support of well siting decisions. This
information may be available either in a relatively direct form (e.g., from local

scale core
data, geophysical logs, hydraulic tests) or n
eeds to be inferred from secondary
information (e.g., tracer experiments, production data, geophysical data). Each data type
may be sensitive to different physical properties and may sample different spatial scales.
Much uncertainty remains in knowing how
the various data types are related to each
other and to the geological heterogeneity that controls production.
The standard practice is to develop a reservoir model iteratively. Local

scale borehole
measurements are used to develop an understanding of hy
drogeological heterogeneity.
An initial geological model is obtained by simply interpolating hydrogeological data
between wellbores. When the horizontal correlation length of the property of interest is
large compared to the wellbore spacing, such a proced
ure might provide adequate
information about geological heterogeneity. However, most often, wellbore spacing is
much greater than the length scale of heterogeneity, and such an approach is inadequate.
Therefore, tomographic images created from geophysical
data (mainly seismic and
electrical) are used to delineate the major geologic features of the reservoir. These images
show spatial distributions of geophysical (rather than hydrogeological) attributes that are
related to subsurface properties. Typically, a
simple petrophysical relationship between a
geophysical survey attribute and a property measured at a borehole is established, and
then used to map the property throughout the reservoir. Geostatistical methods (spatial
interpolation or statistical simulat
ion) may be used to add smaller

scale heterogeneity,
conditioned on local

scale borehole measurements.
The scarcity of direct hydrogeological borehole data, and the inherent ambiguity in
relating geophysical attributes to reservoir properties are weakness
es of the standard
approach. Data integration and joint inversion techniques are approaches developed to
overcome these weaknesses by (1) making use of all available data, (2) combining
complementary information to reduce the ill

posedness of the inverse p
roblem, (3)
including soft information into the analysis, and (4) jointly characterizing the reservoir
structure and system state. While comprehensive data integration has the theoretical
appeal of solving the complex estimation

identification problem need
ed for accurate
reservoir characterization, substantial challenges need to be overcome. These challenges
are related to (1) up

and downscaling issues related to data with vastly different support
scales, (2) systematic errors in data, (3) systematic error
s in simulation results due to
model simplifications, (4) accurate simulation of coupled thermal

hydrological

mechanical

biogeochemical processes, (5) description, characterization, and simulation
of multi

scale heterogeneity consisting of stochastic and d
eterministic features, and (6)
computation requirements.
February 9, 2009
Page
11
Advances have been made on several aspects, including the development of (1) scaling
laws, (2) advanced error models and inversion methods, (3) sophisticated process
simulators, (4) data integration
, analysis and visualization, and (5) faster hardware and
more efficient numerical algorithms.
Many of these advances are made in separate fields
of science and technology
, and are discussed in other sections of this document.
I
ntegration of these simulat
ion, characterization, visualization, and optimization
techniques is
the key in
address
ing
the challenge of optimal reservoir management
, and is
the focus of this project
.
E.
D
ecision

Making, Soft

Computing and Expert System Methodology
As mentioned in
the proposal, the proposed development relies on "a multi

scale
approach to the overall problem of field development that can tame complexity, mitigate
uncertainty, and coordinate information

processing from a multitude of sources and
varied formats." The
schematic shown in the proposal (repeated in
Figure
) captures the
essence of the proposed development.
What has allowed the recent emphasis on improved decision making at each level is the
increasingly abundant
availability of real

time data (at the corresponding time scale), data
transmission technology (borehole to surface and beyond through wired or wireless
technologies), actuators (e.g. remotely activated valves in smart wells) and software
(Bieker, et al., 2007)
. The terms "intelligent" or "smart", among others, have been used
extensively in recent yea
rs to characterize completions, wells, reservoirs, fields and
anything else associated with hydrocarbon production that makes use of related new
technologies
(Konopczynski and Ajayi, 2008)
. A proliferation of conferences and
workshops on the subject during
the last decade reflects that trend
(SPE, 2009)
.
The rate of adoption of new technologies in the field has been d
ebated extensively. It is
generally assessed that new technologies could be adopted faster, provided certain
requirements are met
(Daneshy, 2005, Daneshy and Donnelly, 2004, SPE Emerging
Technology Workshop, 2009, Rajan and Krome, 2008)
, although the adoption rate
seems
to be accelerating recently. In fact, it is one of the main objectives of this RPSEA project
to contribute towards accelerating the development and adoption of new technologies for
tight gas production.
February 9, 2009
Page
12
Figure
2
. Gas field development and operati
ons hierarchy
In the context of tight gas, a clustering of the levels in
Figure
into three basic levels has
been identified in literature:
E.1.
Top level (before or at early production)
Relevant questions for decision making as
ked at this level are associated with reservoir
geology, well drilling and completion, such as
Based on seismic and/or core data, what can be expected from a reservoir?
Where should wells be placed, following what geometry, and in what order?
How should st
imulation be performed and what is its value?
Relevant models that can help make decisions at this level are coarse, in order to handle
the considerable uncertainty inherent in all predictions. Typical models are based on
some form of decline

curve analy
sis or its refinements, multivariable statistics, or crude
mass

balance formulations
(Ehlig

Economides, 2002, Valkó, 2009)
.
Because uncertainty is high at early stages of tight

gas field development, an explicit and
reliable assessment and treatment of uncertainty is important. This is addressed by the
LBNL
group participating in
this project
(see Section C)
.
E.2.
Medium level (Production planning):
Relevant questions asked at this level are associated with optimizing production by using
existing or additional resources. Typical decision making questions are
What has been learn
ed from production up to the current point that can be used in
the future?
Should additional wells be drilled, and if so where and how?
How should production be optimization over the next couple of years?
February 9, 2009
Page
13
Typical models used to answer the above questions
are in the form of a reservoir
simulator, frequently augmented by hydraulics or other surface facilities simulators. The
activity of the simulator "learning" from past experience is known as "history matching",
on which there is abundant literature.
Beca
use computations using reservoir simulators are time

consuming, there have been a
number of approaches aiming to generate decisions (e.g. about optimal production)
efficiently. They rely either on simplifying the decision making process (e.g. using a
movi
ng horizon optimization approach
(Nikolaou, et al., 2006)
) or on using a simplified
reservoir model that is accurate enough to generate optimal solutions reliably
(Awasthi,
et al., 2008)
. Further, solutions relying on optimal control theory have been proposed,
but they rely on software development not much less complicated than a reservoir
s
imulator
(Sarma, et al., 2005)
.
E.3.
Low level (well test analysis and production)
Questions at this level are related to daily operations and are multiface
ted. They are not
the main focus of this RPSEA project, but they are nevertheless related to it and will be
influenced by the RPSEA project's findings. Typical issues may be related to well test
analysis
(Gringarten, 2006, Hi
te, et al., 2006, Mochizuki, et al., 2006)
. Because daily
operations involve a number of repetitive tasks executed by humans, computer aids that
might not necessarily outperform those humans but would nevertheless remove the
burden of routine would be we
lcome, because they would enable personnel to focus on
issues that would be best resolved by direct, expert intervention.
References
Section
B.1.
1.
Ayers Jr., W.B. 2003. Coalbed Methane in the Fruitland Formation, San Juan
Basin, Western United States: A Gi
ant Unconventional Gas Play. In
Giant Oil
and Gas Fields of the Decade 1990
–
1999,
ed. M.T. Halbouty, Chap. 10, 159
–
188.
Houston: AAPG Memoir, AAPG.
2.
Aminian, K., Ameri, S., Bhavsar, A., Sanchez, M., and Garcia, A. 2004. Type
Curves for Coalbed Methane Produ
ction Prediction. Paper SPE 91482 presented
at the SPE Eastern Regional Meeting, Charleston, West Virginia, 15
–
17
September.
3.
Aminian, K., Ameri, S., Bhavsar, A.B., and Lakshminarayanan, S. 2005. Type
Curves for Production Prediction and Evaluation of Coalb
ed Methane Reservoirs.
Paper SPE 97957 presented at the SPE Eastern Regional Meeting, Morgantown,
W.V., 14
–
16 September.
4.
Anbarci, K. and Ertekin, T. 1992.
Pressure Transient Behavior of Fractured Wells
in Coalbed Reservoirs. Paper SPE 24703 presented at th
e SPE Annual Technical
Conference and Exhibition, Washington, DC, 4
–
7 October.
5.
Bayles, G.A. and Reznik, A.A. 1986. Pressure Transient Analysis of Methane
Production From Coal Beds: An Analytical Approach. Paper SPE 15226
February 9, 2009
Page
14
presented at the SPE Unconventional
Gas Technology Symposium, Louisville,
Kentucky, 18
–
21 May.
6.
Clarkson, C.R., Bustin, R.M., and Seidle, J.P. 2007. Production

Data Analysis of
Single

Phase (Gas) Coalbed

Methane Wells.
SPE Reservoir Evaluation &
Engineering
(June 2007): 312
–
331. SPE

100313

P
A.
7.
Clarkson, C.R., Jordan, C.L., Gierhart, R.R., and Seidle, J.P. 2008. Production
Data Analysis of Coalbed

Methane Wells.
SPE Reservoir Evaluation &
Engineering
(April 2008): 311
–
325. SPE

107705

PA.
8.
Clarkson, C.R. 2008. Case Study: Production Data and Pre
ssure Transient
Analysis of Horeshoe Canyon CBM Wells. Paper SPE 114485 presented at the
CIPC/SPE Gas Technology Symposium Joint Conference, Calgary, Alberta, 16
–
19 June.
9.
Connell, L.D. and Lu, M. 2007. A Dual

Porosity Model for Gas Reservoir Flow
Incorpora
ting Adsorption Behavior
—
Part II. Numerical Algorithm and Example
Applications.
Transport in Porous Media
69
(2): 139
–
158.
10.
Ertekin, T. and Sung, W. 1989. Pressure Transient Analysis of Coal Seams in the
Presence of Multi
–
Mechanistic Flow and Sorption Phe
nomena. Paper SPE 19102
presented the SPE Gas Technology Symposium, Dallas, Texas, 7
–
9 June.
11.
Kolesar, J.E., Ertekin, T., and Obut, S.T. 1990. The Unsteady

State Nature of
Sorption and Diffusion Phenomena in the Micropore Structure of Coal: Part 1
—
Theory an
d Mathematical Formulation.
SPE Formation Evaluation
(March 1990):
81
–
88. SPE

15233

PA.
12.
Kolesar, J.E., Ertekin, T., and Obut, S.T. 1990. The Unsteady

State Nature of
Sorption and Diffusion Phenomena in the Micropore Structure of Coal: Part 2
—
Solution.
SPE
Formation Evaluation
(March 1990): 89
–
97. SPE

19398

PA.
13.
Lu, M. and Connell, L.D. 2007. A Dual

Porosity Model for Gas Reservoir Flow
Incorporating Adsorption Behavior
—
Part I. Theoretical Development and
Asymptotic Analyses.
Transport in Porous Media
68
(2
): 153
–
173.
14.
Mavor, M.J., Russell, B., and Pratt, T.J. 2003. Powder River Basin Ft. Union Coal
Reservoir Properties and Production Decline Analysis. Paper SPE 84427
presented at the SPE Annual Technical Conference and Exhibition, Denver,
Colorado, 5
–
8 Octob
er.
15.
Mohaghegh, S., Bilgesu, H.I., and Ertekin, T. 1995. Production Decline Curves
for Low

Pressure Gas Reservoirs Undergoing Simultaneous Water Production.
SPE Formation Evaluation
(March 1995): 57
–
62. SPE

21269

PA.
16.
Munoz, J.D., Chen, H.Y., and Teufel, L.W
. 2001. A Type
–
Curve Based
Spreadsheet Program for History Matching and Forecasting Tight
–
Gas
Production. Paper SPE 71067 presented at the SPE Rocky Mountain Petroleum
Technology Conference, Keystone, Colorado, 21
–
23 May.
17.
Pinzon, C.L. and Patterson, J. 200
4. Production Analysis of Coalbed Wells Using
Analytical Transient Solutions. Paper SPE 91447 presented at the SPE Eastern
Regional Meeting, Charleston, West Virginia, 15
–
17 September.
18.
Sarkar, P. and Rajtar, J. 1994. Horizontal Well Transient Pressure Test
ing in
Coalbed Reservoirs. Paper SPE 26995 presented at the Latin American/Caribbean
Petroleum Engineering Conference, Buenos Aires, Argentina, 27
–
29 April.
February 9, 2009
Page
15
19.
Seidle, J.P, Kutas, G.M., and Krase, L.D. 1991. Pressure Falloff Tests of New
Coal Wells. Paper SPE
21809 presented at the Rocky Mountain Regional Meeting
and Low
–
Permeability Reservoirs Symposium, Denver, Colorado, 15
–
17 April.
20.
Seidle, J.P. 2002. Coal Well Decline Behavior and Drainage Areas: Theory and
Practice. Paper SPE 75519 presented at the SPE Ga
s Technology Symposium,
Calgary, Alberta, 30 April
–
2 May.
21.
Soot, P.M. 1992. Coalbed Methane Well Production Forecasting. Paper SPE
24359 presented at the SPE Rocky Mountain Regional Meeting, Casper,
Wyoming, 18
–
21 May.
22.
Rushing, J.A., Blasingame, T.A., Johns
ton, J.L., and Lee, W.J. 1991. Slug Testing
in Multiple Coal Seams Intersected by a Single, Vertical Fracture. Paper SPE
22945 presented at the SPE Annual Technical Conference and Exhibition, Dallas,
Texas, 6
–
9 October.
23.
Rushing, J.A., Perego, A.D., and Bla
singame, T.A. 2008. Applicability of the
Arps Rate
–
Time Relationships for Evaluating Decline Behavior and Ultimate Gas
Recovery of Coalbed Methane Wells. Paper SPE 114514 presented at the
CIPC/SPE Gas Technology Symposium Joint Conference, Calgary, Alberta
, 16
–
19 June.
Section
B.2.
1.
Al

Kobaisi, M., Ozkan, E., and Kazemi, H. 2005. A Hybrid Numerical/Analytical
Model of a Finite

Condutivity Vertical Fracture Intercepted by a Horizontal Well.
SPE Reservoir Evaluation & Engineering
(Aug. 2006): 345
–
355. SPE

920
40

PA.
2.
Amini, S., Ilk, D., and Blasingame, T.A. 2007. Evaluation of the Elliptical Flow
Period for Hydraulically

Fractured Wells in Tight Gas Sands
—
Theoretical
Aspects and Practical Considerations. Paper SPE 106308 presented at the SPE
Hydraulic Fracturi
ng Technology Conference, College Station, Texas, 29

31
January.
3.
Anderson, D.M., Stotts, G.W.J., Mattar, L., Ilk, D., and Blasingame, T.A. 2006.
Production Data Analysis
—
Challenges, Pitfalls, Diagnostics. Paper SPE 102048
presented at the SPE Annual Tech
nical Conference and Exhibition, San Antonio,
Texas, 24

27 September.
4.
Arevalo
–
Villagran, J.A., Ganpule, S.V., Wattenbarger, R.A., Samaniego
–
Verduzco, F., Yanez
–
Mondragon, M., and Serrano
–
Lozano, J.R. 2002. Analysis
of Long
–
Term Performance in Tight Gas Wel
ls: Field Examples. Paper SPE
74360 presented at the SPE International Petroleum Conference and Exhibition,
Villahermosa, Mexico, 10
–
12 February.
5.
Blasingame, T.A. 2008. The Characteristic Flow Behavior of Low

Permeability
Reservoir Systems. Paper SPE 11416
8 presented at the SPE Unconventional
Reservoirs Conference, Keystone, Colorado, 10

12 February.
6.
Cipolla, C.L., Warpinski, N.R., Mayerhofer, M.J., Lolon, E.P., and Vincent, M.C.
2008. The Relationship Between Fracture Complexity, Reservoir Properties, and
Fracture Treatment Design. Paper SPE 115769 presented at the SPE Annual
Technical Conference and Exhibition, Denver, Colorado, 21

24 September.
February 9, 2009
Page
16
7.
Cox, D.O., Kuuskraa, V.A., and Hansen, J.T. 1996. Advanced Type Curve
Analysis for Low Permeability Gas Reservoi
rs. Paper SPE 35595 presented at the
Gas Technology Conference, Calgary, Alberta, 28 April
–
1 May.
8.
Cox, S.A., Sutton, R.P., Stoltz, R.P., and Knobloch, T. 2005. Determination of
Effective Drainage Area for Tight Gas Wells. Paper SPE 98035 presented at the
S
PE Eastern Regional Meeting, Morgantown, West Virginia, 14
–
16 September.
9.
El
–
Banbi, A.H. and Wattenbarger, R.A. 1998. Analysis of Linear Flow in Gas
Well Production. Paper SPE 39972 presented at the SPE Gas Technology
Symposium, Calgary, Alberta, 15
–
18 Marc
h.
10.
Hager, C.J. and Jones, J.R. 2001. Analyzing Flowing Production Data with
Standard Pressure Transient Methods. Paper SPE 71033 presented at the SPE
Rocky Mountain Petroleum Technology Conference, Keystone, Colorado, 21
–
23
May.
11.
Hale, B.W. 1986. Analysis o
f Tight Gas Well Production Histories in the Rocky
Mountains.
SPE Production Engineering
(July 1986): 310
–
322. SPE

11639

PA.
12.
Hale, B.W., Firth, C.H., Hansen, R.C., Murphy, M.J., and Woodcox, S.C. 1989.
Postbuildup Drawdown Analysis of Tight Gas Wells in th
e Rocky Mountains.
SPE Reservoir Engineering
(Aug. 1989): 288
–
296. SPE

16804

PA.
13.
Ilk, D., Rushing, J.A., Sullivan, R.B., and Blasingame, T.A. 2007. Evaluating the
Impact of Waterfrac Technologies on Gas Recovery Efficiency: Case Studies
Using Elliptical Fl
ow Production Data Analysis. Paper SPE 110187 presented at
the SPE Annual Technical Conference and Exhibition, Anaheim, California, 11

14 November.
14.
Ilk, D., Perego, A.D., Rushing, J.A., and Blasingame, T.A. 2008. Integrating
Multiple Production Analysis Te
chniques to Assess Tight Gas Sand Reserves:
Defining a New Paradigm for Industry Best Practices. Paper SPE 114947
presented at the CIPC/SPE Gas Technology Symposium, Calgary, Alberta, 16

19
June.
15.
Ilk, D., Rushing, J.A., Perego, A.D., and Blasingame, T.A. 2
008. Exponential vs.
Hyperbolic Decline in Tight Gas Sands
—
Understanding the Origin and
Implications for Reserve Estimates Using Arps’ Decline Curves. Paper SPE
116731 presented at the SPE Annual Technical Conference and Exhibition,
Denver, Colorado, 21

24 September.
16.
Maley, S. 1985. The Use of Conventional Decline Curve Analysis in Tight Gas
Well Applications. Paper SPE 13898 presented at the SPE/DOE Low
Permeability Gas Reservoirs, Denver, Colorado, 19
–
22 May.
17.
Medeiros Jr., F., Ozkan, E., KAzemi, H. 2006
. A Semianalytical, Pressure

Transient Model for Horizontal and Multilateral Wells in Composite, Layered,
and Compartmentaliezed Reservoirs. Paper SPE 102834 presented at the SPE
Annual Technical Conference and Exhibition, San Antonio, Texas, 24

27
Septemb
er.
18.
Medeiros, F., Ozkan, E., and Kazemi, H. 2007. Productivity and Drainage Area of
Fractured Horizontal Wells in Tight Gas Reservoirs. Paper SPE 108110 presented
at the Rocky Mountain Oil & Gas Technology Symposium, Denver, Colorado,
16

18 April.
February 9, 2009
Page
17
19.
Medeiros
, F., Kurtoglu, B., Ozkan, E., and Kazemi, H. 2007. Analysis of
Production Data from Hydraulically Fractured Horizontal Wells in Tight,
Heterogeneous Formations. Paper SPE 110848 presented at the SPE Annual
Technical Conference and Exhibition, Anaheim, Cal
ifornia, 11

14 November.
20.
Neal, D.B. and Mian, M.A. 1989. Early

Time Gas Production Forecasting
Technique Improves Reserves and Reservoir Description.
SPE Formation
Evaluation
(March 1989): 25
–
32. SPE

15432

PA.
21.
Sinha, M.K. and Furlong, K.D. 1979. Gas Well D
eliverability Prediction
—
For
Hydraulically Fractured (Vertical) Wells in Tight Reservoirs. Paper SPE 7948
presented at the SPE Symposium on Low
–
Permeability Gas Reservoirs, Denver,
Colorado, 20
–
22 May.
22.
Roberts, C.N. 1981. Fracture Optimization in a Tight
Gas Play: Muddy “J”
Formation, Wattenberg Field, Colorado. Paper SPE 9851 presented at the
SPE/DOE Low Permeability Symposium, Denver, Colorado, 27
–
29 May.
23.
Rushing, J.A., Newsham, K.E., and Blasingame, T.A. 2008. Rock Typing
—
Keys to Understanding Produc
tivity in Tight Gas Sands. Paper SPE 114164
presented at the SPE Unconventional Reservoirs Conference, Keystone,
Colorado, 10

12 February.
24.
Rushing, J.A., Perego, A.D., Sullivan, R.B., and Blasingame, T.A. 2007.
Estimating Reserves in Tight Gas Sands at HP/
HT Reservoir Conditions: Use and
Misuse of an Aprls Decline Curve Methodology. Paper SPE 109625 presented at
the SPE Annual Technical Conference and Exhibition, Anaheim, California, 11

14 November.
25.
Stright Jr., D.H. and Gordon, J.I. 1983. Decline Curve Ana
lysis in Fractured Low
Permeability Gas Wells in the Piceance Basin. Paper SPE 11640 presented at the
SPE/DOE Symposium on Low Permeability, Denver, Colorado, 14
–
16 March.
26.
Thompson, J.K. 1981. Use of Constant Pressure, Finite Capacity Type Curves for
Perfo
rmance Prediction of Fractured Wells in Low
–
Permeability Reservoirs.
Paper SPE 9839 presented at the SPE/DOE Low Permeability Symposium,
Denver, Colorado, 27
–
29 May.
27.
Warpinski, N.R., Mayerhofer, M.J., Vincent, M.C., Cipolla, C.L., and Lolon, E.P.
2008. Sti
mulating Unconventional Reservoirs: Maximizing Network Growth
while Optimizing Fracture Conductivity. Paper SPE 114173 presented at the SPE
Unconventional Reservoirs Conference, Keystone, Colorado, 10

12 February.
28.
Warpinski, N.R., Kramm, R.C., Heinze, J.R.
, and Waltman, C.K. 2005.
Comparison of Single
–
and Dual
–
Array Microseismic Mapping Techniques in the
Barnett Shale. Paper SPE 95568 presented at the SPE Annual Technical
Conference and Exhibition, Dallas, Texas, 9
–
12 October.
29.
Warpinski, N.R., Branagan, P.
T., Sattler, A.R., Cipola, C.L., Lorenz, J.C., and
Thorne, B.J. 1990. Case Study of a Stimulation Experiement in a Fluvial, Tight

Sandstone Gas Reservoir.
SPE Production Engineering
(Nov. 1990): 403
–
410.
SPE

18258

PA.
Section
B.3
February 9, 2009
Page
18
1.
Amini, S. 2007. Developme
nt and Application of the Method of Distributed
Volumetric Sources to the Problem of Unsteady

State Fluid Flow in Reservoirs.
PhD dissertation, Texas A&M U., College Station, Texas.
2.
Barenblatt, G.I., Zheltov, I.P., and Kochina, I.N. 1960. Basic Concepts in
the
Theory of Seepage of Homogeneous Liquids in Fissured Rocks [Strata].
PMM
24
(5): 852
–
864.
3.
Bowker, K.A. 2007. Barnett Shale Gas Production, Fort Worth Basin: Issues and
Discussion.
AAPG Bulletin
91
(4): 523
–
533.
4.
Bumb, A.C. and McKee, C.R. 1988. Gas
–
Wel
l Testing in the Presence of
Desorption for Coalbed Methane and Devonian Shale.
SPE Formation Evaluation
(March 1988): 179
–
185. SPE

15227

PA.
5.
Carlson, E.S. and Mercer, J.C. 1991. Devonian Shale Gas Production:
Mechanisms and Simple Models.
JPT
(April 1991)
: 476
–
482. SPE

19311

PA.
6.
Chen, C.C. and Raghavan, R. 1996. A Multiply
–
Fractured Horizontal Well in a
Rectangular Drainage Region.
SPE Journal
2
(Dec. 1997): 455
–
465. SPE

37072

PA.
7.
Cinco

Ley, H. and Samaniego, F. 1982. Pressure Transient Analysis for Natura
lly
Fractured Reservoirs. Paper SPE 11026 presented at the SPE Annual Technical
Conference and Exhibition, New Orleans, Louisiana, 26
–
29 September.
8.
Curtis, J.B. 2002. Fractured Shale

Gas Systems.
AAPG Bulletin
86
(11): 1921
–
1938.
9.
De Swaan, A.O. 1976. Analy
tic Solutions for Determining Naturally Fractured
Reservoir Properties by Well Testing.
Society of Petroleum Engineers Journal
(June 1976): 117
–
122. SPE

5346

PA.
10.
Frantz, J.H., Gatens III, J.M, and Hopkins, C.W. 1997. Using a Multilayer
Reservoir Model to D
escribe a Hydraulically Fractured, Low
–
Permeability Shale
Reservoir. Paper SPE 24885 presented at the SPE Annual Technical Conference
and Exhibition, Washington, DC, 4
–
7 October.
11.
Frantz Jr., J.H., Fairchild Jr., N.R., Dube, H.G., Campbell, S.M., Christians
en,
G.E., and Olszewski, A.J. 1999. Evaluating Reservoir Production Mechanisms
and Hydraulic Fracture Geometry in the Lewis Shale, San Juan Basin. Paper SPE
56552 presented at the SPE Annual Technical Conference and Exhibition,
Houston, Texas, 3
–
6 October.
12.
Frantz Jr., J.H., Williamson, J.R., Sawyer, W.K., Johnston, D., Waters, G.,
Moore, L.P., MacDonald, R.J., Pearcy, M., Ganpule, S.V., and March, K.S. 2005.
Evaluating Barnett Shale Production Performance Using an Integrated Approach.
Paper SPE 96917 presen
ted at the SPE Annual Technical Conference and
Exhibition, Dallas, Texas, 9
–
12 October.
13.
Hazlett, W.G., Lee, W.J., Narahara, G.M., and Gatens III, J.M. 1986. Production
Data Analysis Type Curves for Devonian Shales. Paper SPE 15934 presented at
the SPE East
ern Regional Meeting, Columbus, Ohio, 12
–
14 November.
14.
Javadpour, F., Fisher, D., and Unsworth, M. 2007. Nanoscale Gas Flow in Shale
Gas Sediments.
JCPT
46
(10): 55
–
61.
February 9, 2009
Page
19
15.
Koziar, G. 1984. Devonian Shale Production Data Analysis. Paper SPE 12871
presented at t
he SPE/DOE/GRI Unconventional Gas Recovery Symposium,
Pittsburgh, Pennsylvania, 13
–
15 May.
16.
Kucuk, F. and Sawyer, K. 1980. Transient Flow in Naturally Fractured Reservoirs
and its Application to Devonian Gas Shales. Paper SPE 9397 presented at the SPE
Annua
l Technical Conference and Exhibition, Dallas, Texas, 21
–
24 September.
17.
Lewis, A.M. and Hughes, R.G. 2008. Production Data Analysis of Shale Gas
Reservoirs. Paper SPE 116688 presented at the SPE Annual Technical
Conference and Exhibition, Denver, Colorado,
21
–
24 September.
18.
Mattar, L., Gault, B., Morad, K., Clarkson, C.R., Freeman, C.M., Ilk, D., and
Blasingame, T.A. 2008. Production Analysis and Forecasting of Shale Gas
Reservoirs: Case History
–
Based Approach. Paper SPE 119897 presented at the
SPE Shale Gas
Production Conference, Fort Worth, Texas, 16
–
18 November.
19.
Najurieta, H.L. 1980. A Theory for Pressure Transient Analysis in Naturally
Fractured Reservoirs.
JPT
(June 1980): 1241
–
1250. SPE

6017

PA
20.
Ogunsanya, B.O. 2005. A Physically Consistent Solution for D
escribing the
Transient Response of Hydraulically Fractured and Horizontal Wells. PhD
dissertation, Texas Tech U., Lubbock, Texas.
21.
Pollastro, R.M. 2007. Total Petroleum System Assessment of Undiscovered
Resources in the Giant Barnett Shale Continuous (Unco
nventional) Gas
Accumulation, Fort Worth Basin, Texas.
AAPG Bulletin
91
(4): 551
–
578.
22.
Raghavan, R.S., Chen. C.C., and Agarwal, B. 1994. An Analysis of Horizontal
Wells Intercepted by Multiple Fractures.
SPE Journal
2
(Sept. 1997): 235
–
245.
SPE

27652

PA.
23.
Se
rra, K., Reynolds, A.C., and Raghavan, R. 1983. New Pressure Transient
Analysis Methods for Naturally Fractured Reservoirs.
JPT
(Dec. 1983): 2271
–
2283. SPE

10780

PA.
24.
Valko, P.P. and Amini, S. 2007. The Method of Distributed Volumetric Sources
for Calculati
ng the Transient and Pseudosteady
–
State Productivity of Complex
Well
–
Fracture Configurations. Paper SPE 106279 presented at the SPE Hydraulic
Fracturing Technology Conference, College Station, Texas, 29
–
31 January.
25.
van Kruysdijk, C.P.J.W. and Dullaert, G.M
. 1989. A Boundary Element Solution
of the Transient Pressure Response of Multiply Fractured Horizontal Wells. Paper
presented at the 2nd European Conference on the Mathematics of Oil Recovery,
Cambridge, England.
26.
Warren, J.E. and Root, P.J.1963. The Behav
ior of Naturally Fractured Reservoirs.
Society of Petroleum Engineers Journal
(Sept. 1963): 245
–
255. SPE

426

PA.
27.
Zuber, M.D., Frantz Jr., J.H., and Gattens III, J.M. 1994. Reservoir
Characterization and Production Forecasting for Antrim Shale Wells: An
Int
egrated Reservoir Analysis Methodology. Paper SPE 28606 presented at the
SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana, 25
–
28 September.
Section C
February 9, 2009
Page
20
1.
Debusschere, B.J., Najm, H.N., Pebay, P.P., Knio, O.M., Ghanem, R.G., and
LeMaitre,
O.P., 2004,
SIAM J. Sci. Comp.
, 26, 2, 698

719.
2.
Ghanem, R.G., and Spanos, P.D., 1991,
Stochastic Finite Elements: A Spectral
Approach
, Springer Verlag, New York.
3.
Ghanem, R.G., 1998,
Comp. Meth. Appl. Mech. Eng.,
158, 199

220.
4.
LeMaitre, O.P., Reagan, M.T.,
Najm, H.N., Ghanem, R.G., and Knio, O.M., 2002,
J. Comp. Phys.
, 181, 9

44.
5.
LeMaitre, O.P., Reagan, M.T., Debusschere, B.J., Najm, H.N., Ghanem, R.G.,
and O.M. Knio, 2004,
SIAM J. Sci. Comp.
, 26, 2, 375

394.
6.
LeMaitre, O.P, Najm, H.N., Pebay, P.P., Ghanem,
R.G., Knio, O.M., 2007,
SIAM
J. Sci. Comput.
19, 864

889.
7.
Reagan, M.T., Najm, H.N., Ghanem, R.G., and O.M. Knio, 2003,
Combustion
and Flame
132, 3, 545

555.
8.
Reagan, M.T., Najm, H.N., Ghanem, R.G., Debusschere, B.D., Knio, O.M., and
LeMaître, O.P, 2004,
Co
mbust. Theory and Modeling
, 8, 3, 607

632.
9.
Reagan, M.T., Najm, H.N., Pebay, P.P., Knio, O.M., and R.G. Ghanem, 2005
, Int.
J. Chem. Kinetics
, 37, 6, 368

382.
10.
Wiener, N., 1938,
Amer. J. Math,
60, 897

936.
Section
E
1.
Bieker, H.P., Slupp
haug, O., and Johansen, T.A., 2007.
Title Real

Time
Production Optimization of Oil and Gas Production Systems: A Technology
Survey
. in: SPE 99446

PA
2.
Konopczynski, M.R. and Ajayi, A., 2008.
Reservoir Surveillance, Production
Optimisation and Smart Wor
kflows for Smart Fields
—
A Guide for Developing
and Implementing Reservoir Management Philosophies and Operating Guidelines
in Next Generation Fields
. in: SPE 112244

MS
3.
SPE. 2009 [cited 2009]; Available from:
www.spe
.org
.
4.
Daneshy, A., 2005.
Industry View.
Petroleum Technology Transfer Council
Network News,
11
(1): 12.
5.
Daneshy, A. and Donnelly, J., 2004.
A JPT Roundtable: The Funding and Uptake
of New Upstream Technology.
Journal of Petroleum Technology: 28

30.
6.
SPE Emerging Technology Workshop, 2009.
Delivering and Using Emerging
Technology to Make Money in E&P
. San Antonio, TX
7.
Rajan, S. and Krome, J.D., 2008.
Intelligent Oil Field of the Future: Will the
Future Be Too Late?
in
SPE Annual Technical Conferen
ce and Exhibition
.
Denver, Colorado, USA: SPE 115030

MS
8.
Ehlig

Economides, C.A., 2002.
Recipe for Success in Ultradeep Water
. in
SPE
Annual Technical Conference and Exhibition
. San Antonio, TX: SPE 77625
9.
Valkó, P.P., 2009.
Assigning value to stimulati
on in the Barnett Shale: a
simultaneous analysis of 7000 plus production histories and well completion
records
. in
2009 SPE Hydraulic Fracturing Technology Conference
. Woodlands,
TX: SPE 119369
10.
Nikolaou, M., Cullick, A.S., and Saputelli, L., 2006.
Prod
uction Optimization
–
A
Moving

Horizon Approach
. in
SPE 99358, Intelligent Energy 2006
. Amsterdam
February 9, 2009
Page
21
11.
Awasthi, A., Sankaran, S., Nikolaou, M., Saputelli, L., and Mijares, G., 2008.
Short

Term Production Optimization by Automated Adaptive Modeling and
Contr
ol
: SPE 112239

MS
12.
Sarma, P., Aziz, K., and Durlofsky, L., 2005.
Implementation of Adjoint Solution
for Optimal Control of Smart Wells
. in
2005 SPE Reservoir Simulation
Symposium, paper SPE 92864
. Houston, TX
13.
Gringarten, A.C., 2006.
From Straight L
ines to Deconvolution: The Evolution of
the State of the Art in Well Test Analysis
. in
SPE Annual Technical Conference
and Exhibition
. SPE Annual Technical Conference and Exhibition: SPE 102079

MS
14.
Hite, J.R., Farid, K., Blanton, M.L., Gard, F., Crawley
, C., Menda, D.B.H.,
Dotterer, R.A., Saputelli, L., Deaton, D.F., and Sankaran, S., 2006.
Preparing a
Business Case for Technology Investment in Production Operations
. in
SPE
Annual Technical Conference and Exhibition
. San Antonio, Texas, USA: SPE
103248

M
S
15.
Mochizuki, S., Saputelli, L.A., Kabir, C.S., Cramer, R., Lochmann, M.J., Reese,
R.D., Harms, L.K., Sisk, C.D., Hite, J.R., and Escorcia, A., 2006.
Real

Time
Optimization: Classification and Assessment.
SPE Production & Operations,
21
(4): 455

466.
Comments 0
Log in to post a comment