Summary
It is often difficult to quantify the redevelopment potential of marginal
oil and gas fields because of a wide range of depositional environments,
variability in reservoir properties, a large number of wells, and limited
reservoir information. Evaluation of infill potential in these fields with
traditional simulation methods is time-consuming, labor-intensive, and
frequently cost-prohibitive. Without adequate assessment technology, some
unprofitable infill campaigns may be initiated, while other promising infill
campaigns may be terminated prematurely because of disappointing early
results.
In this paper, we present a simulation-based regression technique to assess
infill-drilling potential in marginal gas fields. With limited, basic reservoir
information, this technique first estimates the spatial distribution of
subsurface reservoir properties by rapid history matching of well production
data. We implemented a sequential regression algorithm to estimate, from
available flow-rate measurements, not only the permeability distribution, but
also the pore-volume distribution. Future production is forecasted and infill-
drilling potential is determined using the estimated permeability and
pore-volume distributions. Because the method uses an approximate reservoir
description, it identifies regions of the field with promising infill potential
rather than individual-infill-well locations.
The proposed technique provides rapid, reliable, and cost-effective
assessment of redevelopment potential in marginal gas fields. In this paper, we
first validate our approach using synthetic reservoir data. We then apply the
approach to the Second White Specks formation, Garden Plains field, western
Canada sedimentary basin. The prediction of infill potential in this gas field,
which has more than 700 wells, demonstrates the power and utility of the
proposed technique.
Introduction
Infill drilling can be an effective process that can accelerate field
development, increase field oil-gas-production rates, and add reserves. Infill
drilling, particularly in tight, marginal gas fields, represents a significant
opportunity to increase production rates and reserves. The more than 200,000
low-rate, or "stripper," gas wells in the US lower 48 states produce
approximately 10% of onshore gas production (Exploration and Production
Technologies). Before implementing any program to increase well density, the
potential of infill drilling must be assessed reliably to justify such infill
activities both technically and economically, especially for marginal fields.
Without a reliable evaluation of infill potential, some unprofitable infill
campaigns may be initiated, while other promising infill campaigns may be
terminated prematurely because of disappointing early results. The
quantification of infill drilling for marginal fields represents a significant
technical challenge, not only because of the complexity (the result of a wide
range of depositional environments and large variability in reservoir
properties) involved in the evaluation, but also because the evaluation often
has to deal with a large number of wells (hundreds to thousands of wells is not
uncommon), limited reservoir information, and time and budget constraints.
It is well recognized that the performance of infill wells depends on many
factors and their interaction. These factors include reservoir heterogeneity or
spatial variations of reservoir properties (such as permeability, net pay,
porosity, and gas saturation); well interference; and available reservoir
energy (pressure). The combination, or interaction, of these factors determines
how much oil-gas can be produced in the future. These factors and their
interaction must be well understood and adequately evaluated to provide a
reliable assessment of infill-drilling potential. To achieve this, assessment
technology must be able to address the previously mentioned engineering
concerns and, at the same time, must be cost-effective and able to handle the
practical issues of having a large number of wells and limited reservoir
information.
In general, an integrated reservoir study, which typically requires detailed
geological, geophysical, petrophysical, reservoir-engineering, and
reservoir-simulation studies, is the best way to evaluate infill-drilling
potential. Integrated reservoir studies can adequately address the engineering
concerns discussed previously. However, for marginal fields, high costs and the
lack of relevant reservoir data make such detailed studies difficult to perform
and to justify economically. One of the key tasks in a reservoir-simulation
study is to calibrate the reservoir model by matching the historical production
data. In a traditional simulation study, however, it is very challenging and
time-consuming to achieve a history match in a field with hundreds or thousands
of wells.
In the past decade, some authors have proposed statistical moving-window
techniques to provide a rapid and cost-effective assessment of infill-drilling
potential in large, tight gas fields (McCain et al. 1993; Voneiff and Cipolla
1996; Hudson et al. 2000). These techniques perform a statistical analysis of
production data using defined performance indicators, such as the best 12
consecutive months of production (best year) and the decline ratio for the best
year. The performance indicators serve as proxies for reservoir properties,
production response, and reservoir pressure. On the basis of the comparison of
indicators between old and new wells in areal “windows” throughout the field,
judgments are made regarding the interference between the existing wells and
whether pressure depletion is occurring. These judgments are then used to
estimate the potential for infill drilling in each window. While this
technology can be a useful screening tool to determine which regions of the
field have potential for infill drilling, the prediction errors for infill-well
performance can be quite significant; additionally, interference effects become
complicated and reservoir heterogeneity increases (Guan et al. 2002).
To improve upon moving-window methods, Gao and McVay (2004) proposed a
simulation-based regression approach for the rapid assessment of infill
potential in gas-well fields. Using readily available data (such as well
locations, production data, and fieldwide average values of porosity,
permeability, saturation, and net pay), their approach provides an estimate of
the heterogeneous permeability field by inverse modeling. Using the reservoir
model and the inverted permeability field, the performance of infill wells is
predicted. The authors show that this simulation-based regression approach
results in more accurate predictions of infill performance than do the
moving-window statistical methods (Guan et al. 2002). However, while their
reservoir model accounts for permeability heterogeneity, all other reservoir
properties remain fixed at their initial values. Thus, their model does not
condition the hydrocarbon-pore-volume distribution (resulting from the
porosity, net thickness, and hydrocarbon-saturation distributions) to the
production data. In marginal reservoirs, the initial geological model is often
based on incomplete data and analyses and, in some cases, may consist simply of
the constant average values of each reservoir property. Thus, the initial
geological model may possess significant errors. Consequently, while infill
performance may be predicted more accurately with the simulation-based
regression approach rather than with the moving-window methods, because the
reservoir model is not able to adjust the pore-volume distribution, predictions
for infill wells and existing wells may still contain large errors (Gao and
McVay 2004).
In this paper, we advance the simulation-based regression approach to the
rapid assessment of infill potential by presenting a new inverse-modeling
algorithm. This algorithm implements a sequential inversion of both
permeability and pore-volume distributions through an integration of production
data. Thus, starting with the estimated average reservoir properties (net
thickness, porosity, gas saturation, and permeability) or a prior geological
model, we can now obtain not only an estimate of the large-scale permeability
distribution, but also an estimate of the pore-volume distribution. In the
following sections, we first present the development of the sequential
inversion technique to estimate the distributions of permeability and pore
volume by conditioning them to historical production data. We then validate
this new technique with synthetic and field examples and, finally, demonstrate
its application to assess infill potential in the field example.
© 2008. Society of Petroleum Engineers
View full textPDF
(
2,398 KB
)
History
- Original manuscript received:
6 March 2006
- Meeting paper published:
15 May 2006
- Revised manuscript received:
2 July 2007
- Manuscript approved:
27 January 2008
- Version of record:
20 June 2008
View Cart
