SPE 163723

Pressure Transient Analysis of Data from Permanent Downhole Gauges

M. M. Al-Buraikan, Saudi Aramco; H. S. Al-Hashim, KFUPM; M. B. Issaka, Saudi Aramco; H.Y. Al-Yousef,
KFUPM

Copyright 2013, Society of Petroleum Engineers

This paper was prepared for presentation at the 2013 SPE Digital Energy Conference and Exhibition held in The Woodlands, Texas, USA, 5–7 March 2013.

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

Abstract

In recent years there has been a significant growth in the number of Permanent Downhole Measurement System (PDHMS)
installations in oil and gas fields around the world, as PDHMS prices have been falling steadily and their reliability has also
been increasing.

However, the full benefits of this investment can only be realized when it is taken from simple surveillance and monitoring to
a source for reservoir characterization. This study assesses the systems and workflows that have been put in place to
transform the massive amounts of data, including pressure, flow rate and temperature, into actionable information to improve
field development and performance. This study presents a dynamic real-time well testing workflow for utilizing data from
from intelligent fields. It discusses the applicability of pressure transient analysis using I-Field data from Permanent
Downhole Gauges to characterize reservoir and well performance.

While seeking to develop an efficient workflow for the utilization of the real-time data for pressure transient analysis, the
study also seeks to assess the effects of producing time, flow rate simplification, and data interruptions on the reliability of
the analysis methods and results.

Actual real-time pressure data from PDHMS, and flow rate data Multi-Phase Flow Meters (MPFM) were used in the
analyses, to determine reservoir parameters and evaluate well performance. This study highlights some of the challenges in
using real-time data from PDHMS and MPFM. A commercially available software product was used to filter, and manage
the real-time data, as well as the modeling and analysis of the test data.

Introduction

During the past several years there has been a repid increase in the installation of permanent downhole measurement systems
(PDHMS) in oil and gas fields around the world. This rapid increase in installation of PDHMS is associated with the increase
in intelligent well completions and lunching more I-Fields (Issaka and Buraikan, 2010). This has been accelerated by the
steadly falling prices of these systems in addition to an increase in their ability to continuously monitor, record and instant
transfer of recorded data to the engineer’s disktops with good reliability. Even though many operating companies have been
investing heavily in the acquisition of real-time data from their fields, and more installations are scheduled in the coming
years, the full benefits of this investment can only be realized when it is taken from simple surveillance and monitoring to a
source for reservoir characterization and tools for better understanding reservoir performance. Systems and workflows need
to be put in place to transform the massive amounts of data—including pressure, flow rate, and temperature—into actionable
information to improve field development and performance.

PDHMSs are becoming an integral part of all new intelligent field (I-Field) development plans, but the actual real-time data
collected through these systems are not fully utilized. To evaluate and analyze the massive amount of data collected for
reservoir characterization and well performance evaluation, a dynamic real-time well testing workflow was developed using
data from actual I-Fields. Some data quality issues were encountered, and many benefits were realized. Also, a methodology
was developed for pressure transient analysis (PTA) utilizing I-Field data from permanent downhole gauges.
2 SPE 163723

With the wide adaptation of I-Fields, PDHMSs became a significant source of information to capture real-time reservoir
pressure response. PDHMSs, coupled with multiphase flow meters (MPFMs), can also provide far more value by translating
their data into reservoir characterization information. Occasional field and unscheduled well shut downs result in buildups
that may be called”free well tests,” which can be utilized to provide vital information about the field and well performance
evolution through time. An established workflow is required to manage the real-time data in terms of denoising, filtering,
storage, and retrieval, which are essential to attaining the maximum benefit from the instrumentation investments. Cases
were investigated to reveal the use of permanent downhole pressure gauges for reservoir characterization and well
performance evaluation with PTA and assess the rate effects on establishing reliable analysis. Actual real-time PDHMS and
MPFM data were used, reservoir, and well data were assessed and validated during which challenges were encountered.
These data were then analyzed to determine reservoir parameters and evaluate well performance. The Diamant® application
was used to filter and manage the real-time data, and the Saphir® application was used for modeling and analysis.

Historical Background

Probably the first paper to discuss the use of permanently installed bottomhole pressure gauges is by W.A. Nestlerode (1963).
Nestlerode aims to identify potential operation problems and to get some necessary reservoir data for effective control with
less than one pressure point per day. Athichanagorn et al. (1999) present a methodology and sequential steps for data
acquired with permanent downhole pressure gauges. A variety of wavelet algorithms are discussed to denoise the data and
provide reliable pressure transient identification. Ortiz et al. (2009) tested several wavelet denoising techniques on a large
volume of PDHMS data. The effect of several factors, such as wavelet type, threshold, and resolution level, were discussed.
The paper classified and rated each technique and variation for efficiency comparison. Chorneyko (2006) presented an
operational perspective of permanent downhole pressure gauges. The author presented practical cases of information obtained
and reservoir management decisions derived from the PDHMS. The author also highlighted the ever-increasing number of
gauge installations in his company, specifically, and across the industry as a whole. The author stressed the fact that such a
sizeable investment requires active stewardship to realize effective utilization. de Oliveira Silva and Kato (2004) presented a
successful case of utilizing a PDHMS to identify barrier and inter-reservoir connectivity, which eliminated the need for
workovers or survey services. This paper illustrates how a PDHMS can be a vital tool to achieve sound reservoir
management, completion, and production decisions. More can be attained with more improvement in PDHMS data
management and treatment tools.Horne (2007) discussed methods and algorithms to manage and interpret permanent
downhole pressure gauge data for the industry to make the best use of these abundant sources of data. He highlighted the
need to store the PDHMS data in a manner that allows efficient access and recovery. Also, he stressed that the problems are
not yet fully solved, and research in this area is appropriately active, where the need for a set of reliable automated algorithms
will be necessary to gain maximum advantage.

Ouyang and Kikani (2002) discuss some improvement to PDHMS data processing. A new formula to automatically identify
pressure transient periods was proposed, and polytope regression for noise level identification and outlier removal techniques
were also discussed. Suzuki and Chorneyko (2009) presented a new method for automatic pressure buildup detection from
PDHMSs. The new method analyzes the pressure response for specific patterns and the change of specific pressure over the
change of a specific time window to identify the start of the buildup or drawdown. The authors tested this new method on
field cases with positive results. Liu, and Horne (2011) presented an interpretation approach for pressure and flow rate data
from permanent downhole gauges using data mining. The aim was to obtain a reservoir model using pressure and flow rate
data from the PDHMS and nonparametric data mining algorithm. Noisy synthetic data and the real field data were used to
test this approach. The method was able to recover the reservoir model successfully. Even at the extreme cases when the flow
rate data are very noisy and changing frequently, and in the absence of any shut-ins, the method was still able to extract the
reservoir models

Houzé et al., (2011) presented the use of a new generation of wavelets, allowing a more accurate processing. The objectives
were to reduce the volume of data without losing valuable information, remove outliers, and identify build-ups and re-
allocate production. This paper also presents a new method using tangents crossing, which successfully replaces the failing
wavelets to identify build-ups and re-allocate production.

In the literature, there is evident focus on PDHMS data denoising, filtration, and events detection. However, there are very
limited publications on the utilization of these data in reservoir characterization and well performance evaluation, which this
study focuses on.

Pressure Transient Analysis

In its most basic form, a well test records changes in downhole pressure that follow a change in flow rate. Often, downhole
pressures and temperatures, surface flow rates, and samples of produced fluids are obtained (Horne, 1990). The behavior of
reservoir fluids and their interactions with reservoir rock and completion and production systems must be thoroughly
SPE 163723 3

characterized to produce a reservoir efficiently. This characterization is accomplished through reservoir modeling, and well
test data provide a driving force for running model simulations. Reservoir models are developed on a framework of
geophysical, geological, and petrophysical data. Dynamic well test data are integrated into this static framework to simulate
and predict reservoir behavior. Data from PTA are particularly useful in detecting heterogeneities, permeability barriers,
structural boundaries, fractures, fluid contacts, and gradients that can be incorporated into the model. Once a reservoir model
is built, it is calibrated by comparing results of a test simulation against measured data to check its parameters. To achieve a
good match between real and modeled data, the user may need to fine-tune certain assumptions in the model concerning the
well and its reservoir, such as permeability, distance to a fault, or other such parameters. Production histories from wells in
this field are then entered into the model. Another simulation is carried out to model pressures at the wellbore and across the
reservoir. Simulation-derived fluid ratio and wellbore pressures are run through a history-matching process for comparison
with measured production ratios and pressures. It is not unusual for initial results to disagree, in which case the model
parameters are again changed. This iterative procedure continues until a good match is obtained between actual and simulated
results. The reservoir model can then be used in predicting future production, well location, and completion scenarios.

Perhaps one of the most useful applications of well test data is achieved through PTA. By generating a log-log plot of
measured pressure over time, when plotted along with the derivative of changing pressure, analysts are able to study pressure
changes in great detail. The derivative of the pressure change provides a characteristic signature of reservoir pressure
response to well testing that can be interpreted in terms of flow regimes, boundaries, permeability, formation damage,
heterogeneities, and reservoir volumes. PTA data, when integrated into these and other advanced interpretation techniques,
help production teams understand their reservoirs and achieve their engineering and business objectives (Horne, 1990).

Digital Oil Field of the Future

The exploration and production (E&P) industry has a long history of exploiting the growing power of digital technology. The
accelerating performance of digital devices (e.g., processors, storage, and bandwidth) is leading to waves of technological
innovation that promise significant new capabilities for E&P firms. As a result, the industry is standing on the crest of the
digital oil field of the future (DOFF), which will enable petro-professionals and field workers to benefit from total asset
awareness on the ability to monitor and manage all operational activities in real time or near real time, regardless of location
(Al-Dhubaib et al., 2008b).
It is believed that this state-of-the-art technology for finding, developing, and producing oil and gas will likely play an
important role in allowing E&P companies to realize the full economic potential of their assets. The primary benefits offered
by the DOFF includes: potential extra recovery (de Oliveira Silva and Kato, 2004), lower operating costs (Chorneyko, 2006),
increased production rates (de Oliveira Silva and Kato, 2004), reduction in capital costs by 5 to 10% over the next 3 to 5
years and additionally, real-time drilling technology is leading to reductions in drilling costs by 5 to 15%, as drilling
engineers are able to react to problems in a more informed and timely manner (Chorneyko, 2006).
These benefits do not exist independent of one another, but rather rely on a strong interdependence to achieve their maximum
gains. The technologies and processes that enable greater production volumes also control water and gas handling, strongly
influencing ultimate recovery capabilities. Discovering and exploiting additional reserves increases facility throughput and
the need for optimization.
DOFF benefits will not materialize simply by acquiring more and better data from all aspects of oil and gas operations.
Transformative work processes and organizational changes will likely be needed to take full advantage of the new
technologies, one of which is the utilization of real-time data from PDHMSs into the well testing and PTA process (Saleri et
al., 2006).

Permanent Downhole Measurement Systems

Over the past 100 years, petroleum exploration and development has evolved from simple techniques, such as digging a hole
into a suspected reservoir, to complex production monitoring and control methods. The petroleum industry has since then
grown into a multibillion dollar industry. In addition, exploration and development techniques have become increasingly
complex as reserves have become more difficult to find. The need for accurate downhole data is now a necessity for
successful reservoir monitoring and production because the petroleum reservoirs available today are located in environments
that pose a great deal of technical challenges when it comes to development. Today, petroleum is found in offshore
environments, and those located on land still require improved oil recovery techniques in order to maximize profitability. To
successfully apply improved recovery techniques—such as water flooding, vertical lift performance, hydraulic fracturing,
etc.—accurate downhole data are required. Permanent sensors and monitoring systems help to ensure optimization of
reservoir monitoring and production techniques by providing the petroleum engineer with real-time data to make timely and
accurate decisions. Examples of such decisions are where to place perforations, how best to conduct a waterflood, and
whether to fracture a well or acidize it. Data acquired by permanent monitoring systems also enable the petroleum engineer to
diagnose problems such as plugged chokes, leaking valves, etc. This is why the study of sensors and permanent monitoring
systems is so important. A great deal of research is being done in the area of sensors to improve them and make them more
4 SPE 163723

effective. Permanent monitoring systems have become an important aspect of petroleum technology (Daungkaew et al.,
2000).
The need for accurate downhole data first led to the use of surface gauges. It was soon discovered that the data acquired with
surface gauges was not sufficiently accurate for oil recovery techniques. In order to acquire accurate data, wireline gauges
were invented. This led to further inventions, such as the downhole gauge, from which the permanent monitoring system
evolved.
Permanent sensors can be defined as measuring devices that make measurements by exhibiting changes in properties in
response to a measured variable such as pressure, temperature, flow rate, etc. They are devices that could be electrical,
mechanical, or in the form of an optical fiber. In the petroleum industry, sensors are used to measure physical variables
downhole. Measured variables include temperature, pressure, flow rate, density, viscosity, and electrical resistivity
(Omotosho, 2004). Permanent monitoring systems can greatly improve the decisions made during oil production and
reservoir development. However, traditional reservoir monitoring methods cannot be completely overruled. Rather, they can
be used in conjunction with permanent well monitoring systems as a reference for evaluating the accuracy of the system. That
way, a problem with the monitoring system can be easily detected. Permanent well monitoring systems acquire information
quicker than conventional methods of data acquisition. As measurements are made, information is relayed through various
means of telecommunication such as satellite communication. Information can reach the petroleum engineer responsible for
interpreting the data in real time (Reynolds, 1986).
Technology has made permanent well monitoring systems more than just data collectors. With the recent advent of various
kinds of software for communication and interpretation, a well monitoring system can now collect data and interpret it. In
some cases, it can take the necessary action required to control the well, such as closing a valve with minimal human
intervention. When a permanent well monitoring system is capable of taking certain actions—such as the shutting-in of a gas
lift valve based on the data collected by the system—with little or no human input, it is known as an intelligent completion.
The intelligent system is a well monitoring system that, to some extent, is able to manage field production with little need for
human intervention. It maximizes field production by analyzing the data collected with permanent sensors on a continuous
basis. The intelligent completion is designed to last throughout the life of the well, but in reality, 55% of most downhole
electronic sensors record failures within less than 4 years (Daungkaew, 2000).
To accomplish this form of automated field management, a feedback loop connects the well monitoring system to subsurface
controls. As the data are collected, they are interpreted, and the necessary action is taken without the need for an expensive
workover. Because of the high cost of intelligent completions and permanent monitoring systems in general, only offshore
wells usually justify such an expense because well interventions for such wells far exceed the cost of an intelligent
completion. In addition, high production rates from offshore wells meet similar conditions. Other wells where intelligent
completion technology may be applied are high-rate production land wells in very remote areas. Again, this is because a
workover operation in such wells will be more expensive than an intelligent completion, therefore making intelligent
completions an economic option. The ability to monitor downhole variables in real time provides a better picture of what is
actually happening downhole. In addition, the wealth of data collected is much greater than the data collected using methods
that are more traditional. All these factors help the production or reservoir engineer to make effective decisions during
production or reservoir development (de Oliveira Silva and Kato, 2004; Van Gisbergen and Vandeweijer, 2001).

Wavelet Filtration

The main challenge in the processing of permanent gauge data is to implement a smart filter that would drastically reduce the
number of data points without losing either high-frequency or low-frequency data. Since the raw data coming from the
permanent sensors is very large, and often accompanied by ‘noise’ and ‘outliers’, the data needs to be ‘de-noised’ and then
filtered down to a reasonable size for analysis and interpretation. While there are a number of different methods for
accomplishing this data cleansing and reduction, the one method that has proven to be widely successful is a wavelet-based
algorithm. The application of the wavelet algorithm can reduce pressure data from permanent downhole gauges down a
fraction of the original size, without losing any of the meaningful transient periods contained. The wavelet acts as a filter
with a threshold. Any noise below a certain level is considered noise and is filtered out, as in the case of noisy signals during
production. On the other hand, any noise above a certain level of threshold will be considered a representative break in the
data and will be preserved, as in the case of well shut in. Here, the break in the pressure data will act as a local, high-level
noise that should be preserved.

For noise and outlier removal, the wavelet algorithm determines the trend in the data and provides the user with an adjustable
band or ribbon around the data set. All data points outside this band are discarded as noise or outliers. Figure 1 shows an
example of wavelet denoising with the aid of the adjustable ribbon. The wavelet algorithm is also employed for smart data
filtration that ensures all significant events such as shut-ins are preserved. The wavelet data filtration algorithm requires that
the original data set be evenly-spaced. Otherwise, the system fills in any uneven spaces with interpolated data. This action
may lead to a situation where the filtered data set could potentially be much larger than the original raw data set. One way to
enure that this does not occur is to specify a reasonable initial sampling time spacing. This is a user-controlled parameter that
can be adjusted to fit the intended use of the data; whether for rate or pressure transient analysis. For pressure transient
SPE 163723 5

analysis purposes, the initial time spacing may be set to between 10 and 20 seconds. This will ensure that the high frequency
data are retained, especially at the instant of shut-in, at a reasonable computing cost. During the initial filtration setup, the
user has the option to use the system default filter settings or override them. These settings will automatically be applied to
any subsequent raw data received, and appended to the existing filtered data set. However, the user may return to any part of
the data history and locally repopulate any sequences of interest, such as shut-in periods required for pressure transient
analysis. This is possible because the original raw data is stored or ‘mirrored’ in a compressed format that is available on
demand. Figure 2 is an example of the application of the wavelet algorithm that reduced the pressure data size down to 5 %
of the original size, while maintaining the underlying pressure transient signals (Issaka and Buraikan, 2010).

Field Case Studies

Three case studies are presented and analysed, using real-time data from permanent downhole pressure gauges, for reservoir
characterization and evaluating well performance.

Field Case 1: Well X-1

This case presents downhole pressure and rate data recorded using PDHMS from horizontal Well X-1 completed in an I-Field
in Saudi Arabia. The field is a north-south–trending anticline 22 miles [37 km] long and 4 miles [7 km] wide. It was
discovered in April 1990 and consists of two productive sandstone reservoirs (Zone-X1 and Zone-X2).
Well X-1 was completed in June 2008 as a cased-hole horizontal producer in Zone-X1. Formation analysis and directional
logs for Well X-1 are shown in Figure 3. The well was equipped with six premium sand screens with an MP packer for zonal
isolation and PDHMS for reservoir monitoring at 7,662 ft. Table 1 presents reservoir rock and fluid properties, including
other well data.

PDHMS Data Analysis

Figure 4 shows a section of the PDHMS data, indicating a series of drawdown and buildup downhole pressure data, along
with flow rate readings after filtration using the wavelet algorithm. These data cover a period of 3,200 hours [133+ days],
with 5,540 pressure points and 1,480 rate points. It can be seen from this data set, that there are three distinct buildups.
Figure 5 shows the log-log diagnostic plot of the last pressure buildup (Buildup-3). This log-log diagnostic plot does not
show clearly the expected flow regimes from which the reservoir parameters can be obtained. For this reason the pressure
data during this buildup period was matched against a model of a horizontal well in an infinite-acting reservoir. The results
of the analysis of this buildup period revealed an average permeability of 8.72 md, a sking factor of -2.89, an effective well
length of 1025 ft and kz/kr of 0.0146. When we tried to match the entire flow and buildup periods with these parameters by
selectibg BU3 as a reference it produced a good match to BU3 but the match of the early flow and buildp periods was poor as
can be seen from Figure 6. For this reason, we reanalyzed the test by considering all flow and buildp periods using the same
model and obtained an excellent match for all flow and buildup periods as can be seen from Figure 7. The match was
obtained by honoring all flow and buildup periods with the last buildup (BU3) as the reference period. Similar match with the
same reservoir parameters was also obtained with the last production period before BU3 as a reference period. The result of
the anaylysis revealed even more realistic values of the reservoir parameters compared to the values obtained from the the
analysis of the single pressure buildup period (BU3) as can be seen from Table 2. The results indicate a flow capacity of 307
md-ft (an average permeability of 3.41 md), with an effective horizontal well length of 1518 ft out of a drilled horizontal
length of about 2000 ft, a mechanical skin value of -2.36 and kz/kr of 0.071. Given that no prior stimulation was conducted
on this well, the relatively high negative skin could possibly be attributed to geoskin or fracture..

Assessment of Data Interruption Effects

During the last buildup period, there was an error in pressure recording that resulted in a single pressure value being recorded
for almost 72 hours [3 days] (Figure 8). This period of ‘frozen’ data affects the shape of the pressure derivative curve, as
indicated by the scattered nature of the pressure derivative data during the late time (Figure 9). The ‘frozen’ data
notwithstanding, the analysis results can be corroborated by comparing the pressure derivative curves of the three previous
buildup periods. Figure 10 shows an overlay of the three buildups on the log-log diagnostic plot for Well X-1, indicating a
good agreement. An attempt was made to evaluate the impact on the pressure derivative curve, if the data error had not
occurred. Figure 11 shows the log-log diagnostic plot when the ‘frozen’ data are removed. The figure shows that the
pressure derivative data tends to rise at late time, indicating possible development of the linear flow period.

Field Case 2: Well Y-2

This field is a north-south trending asymmetrical anticline. The eastern flank is slightly steeper than the regional dip,
averaging approximately 20, while the western flank steepens to a maximum dip of 8.70. The length of the field is
approximately 90 km, and the width ranges from 5 to 17 km. There are three oil-bearing reservoirs: Reservoir-Y1, Reservoir-
Y2 and Reservoir-Y3, all of which are non-communicating carbonate formations. Well Y-2 was completed in May 2008 as
6 SPE 163723

an open-hole horizontal producer in Reservoir-Y1. Figure 12 shows the formation analysis logs Well Y-2. The well is
equipped with an electrical submersible pump (ESP) set at 4784 ft, and a Y-tool on 4½-in. tubing. It also has a PDHMS for
pressure monitoring. This well was later subjected to acid stimulation, at the interval of Reservoir-Y1.

PDHMS Data Analysis

Figure 13 shows the downhole pressure and flow rate readings after data filtration using the wavelet algorithm. These data
cover a period of 1,800 hours [75 days], with 12,541 pressure points and 574 rate points. Reservoir rock and fluid properties,
in addition to other well data are presented in Table 1.

Well Y-2 was acid-stimulated in May 2010. First, two pressure buildup periods were selected for analysis (one pre-acid and
one post-acid), as highlighted on Figure 13. Figures 14 shows the log-log diagnostic curves of the pre-acid and post-acid
buildup data, plotted together. The pressure data for the two buildup periods were both matched to a model of a horizontal
well in an infinite homogeneous reservoir, with changing wellbore storage. Figure 15 shows the model match of the post-
acid buildup test. The analysis results of this post-stimulation buildup data indicate an effective horizontal length, Le of 1,816
ft, which is a significant improvement over the value of 1286 ft determined from the match of first pre-stimulation buildup
period. The results also show a skin factor of –0.228 compared with the pre-acid value of –0.10, and kz/kr of 0.42 compared
to 0.0819 obtained from the pre-acid.

The analysis was again repeated by reanalyzing the the flow and buildup periods before the acid stimulation and the flow and
buildup periods after the acid simulation. All flow and buildup periods before the acid stimulation wer included in the
analysis using an infinite homogeneous reservoir, with constant wellbore storage model and the first buildup period was used
as a reference. Figure 16 shows perfect match of all flow and buildp periods. The results of the analysis are presented in
Table 2.
For the case of post-acid, the last buildup period and the the two flow periods (before and after the last buildup) were
considered in the analysis of the post-acid data with the last buildup period used as a reference. Figure 17 shows an excellent
match of the last buildp period and the two flow periods before and after the last buildup period. The results of the analysis
are presented in Table 2. It can be seen from the results of the anaysis presented in Table 2 that the results are more
consistent than the results obtained by the single buildup period analysis. The results shows a skin value of -3.1 and effective
well length of 1,345 ft from the analysis of the post-acid compared to a positive skin of 4.9 and an effective well length of
1,182 ft from the pre-acid data.
The productivity index determined from the pre- and post-acid stimulation were found to be 13 STB/D/psi and 21 STB/D/psi,
respectively confirming the effetveness of the acid job. These results indicate a successful acid treatment, and highlight the
value of pressure transient analysis of PDHMS data in providing a reliable means of diagnosing the well and characterizing
the reservoir.

Data from Table 2 also show the flow capacity obtained from the pre-and post-acid analysis to be 13,700 md-ft. However,
careful inspection of the log-log diagnostic plot in Figure 14 indicates that late horizontal well flow regime was not well
established in either the pre- or the post-acid buildup tests. This creates some degree of uncertainty in the determination of
the flow capacity and a longer shut-in period is needed to confirm the late horizontal radial flow period which will result in
more confidence in the value of the flow capacity determined. The log-log diagnostic plot also shows an anomally during the
early parts of both buildup tests. This is more pronounced in the pre-acid buildup test, where the pressure decreases for some
time during the early part of the buildup test, before continuing to build up again. This anomalous phenomenon is very often
observed in wells equipped with Electical Submersible Pumps (ESP). The behavior is indicative of falling liquid level in the
well, when the pump is stopped for the buildup test. This was modeled as a changing wellbore storage in the single buildup
period anaysis. Figure 14 also indicates that the early-time data distortion is less severe in the post-acid test than in the pre-
acid. This is another indication of a successful acid treatment, due to quicker settling of the falling liquid level after the well
was stimulated.The above results indicate a successful acid treatment, and highlight the value of pressure transient analysis of
PDHMS data in providing reliable means of diagnosing the well and characterizing the reservoir.

Assessment of Rate Simplification Effects

Figure 18 shows the flow rate variation before the first pre-acid buildup period over a period of 30 days. It can be seen from
this figure that during this 30 days period there is an extreme production rate variation. However, the classical methods of
pressure buildup analysis (using semi-log graphs) were derived using the Superposition Theorem, based on the assumption of
a constant flow rate, or a series of constant flow rates prior to the shut-in period. In this study, we investigated the effect of
simplifying the variable rates, by using a single average value. Figure 19 shows the log-log plot of the first pre-acid buildup
period using the whole variable production rate history, compared to using a single average production rate value of 3,500
STB/Day. Figure 19 shows that the rate simplification has no significant effect on the behavior of the pressure derivative.
However, the pressure response is affected, indicating a lower pressure drop with the simplified rate. This indicate that the
lower skin factor will be obtained with the simplified rate.
SPE 163723 7

Field Case 3: Well Z-3

This field measures 23 km in length and 10 km in width. It was discovered in March 1940, with first production in 1961. The
field has eight oil-bearing reservoirs, producing varied crude grades. The depths of the reservoirs range from 8,600 ft to
10,000 ft. In general, the reservoir quality is moderate to good, especially in the crestal area. The reservoir quality
deteriorates toward the flank.

Well Z-3 was completed in June 1973 as vertical producer. It was re-completed in March 2007 as an open-hole vertical
producer in Reservoir-Z. Figure 20 shows the formation analysis logs for Well Z-3. The well was equipped with an ESP
and a PDHMS, for artificial lift and reservoir monitoring, respectively. The well was also subjected to an acid stimulation
prior to the buildup test.

PDHMS Data Analysis

Figure 21 shows the downhole pressure and rate readings after filtration using the wavelet algorithm. These data cover a
period of 5,500 hours [230 days] with 46,367 pressure points and 104 rate points. Reservoir rock and fluid properties, in
addition to other well data are presented in Table 1. It can be seen from this figure gauge malfunction at different times
during the above period. It also shows that the best buildup period is the last one.

Figure 22 shows the log-log diagnostic plot of the last buildup period using pressure data from the PDHMS. The test data
was matched against a vertical well model with constant wellbore storage, in a radial composite reservoir. The pressure
derivative data exhibit a sharp upward turn at late times, indicating possible degradation in reservoir quality, some boundary
effect, or interference from a nearby well. Attempts to simulate possible interference from nearby wells did not yield a match
of the pressure history. Also, the geology of the area does not indicate the existence of faults or boundaries. Therefore, the
rise in the derivative was attributed to degradation in reservoir quality. This is supported by evidence from a nearby well, Z-
105, completed in the same reservoir. Well Z-105 also indicates a similar sharp upward turn in the pressure derivative at late
times as shown in Figure 23. The analysis was conducted using the complete rate history.

The results indicate a good match of the buildup data, with a permeability contrast at approximately 1.5 km away from Well
Z-3. The mobility ratio, (k/μ of inner zone over k/μ of outer zone), and the Diffusivity ratio, (k/∅.μ.ct of innerzone over
k/∅.μ.ct of outer zone) are found to be equal to 40.3 and 2.18, respectively. The flow capacity for this well is estimated at
123,000md.ft, while the productivity index is 13 STB/D/psi. The results also indicate a skin factor of 1.3. Table 2 presents
the complete analysis results for this case.

Assessment of Producing Time Effects

As with the quality of pressure measurements, a focus on flow rate history will yield better model identifications and correct
well and reservoir parameter estimations. This is particularly true with the pressure derivative method, which is a very
powerful tool, but at the same time is more sensitive to the accuracy and completeness of rate data preceeding the buildup or
fall-off periods. The need to take into account the entire rate history has been investigated in the past, and several methods
have been proposed to account for production history effects, such as Horner equivalent time or the introduction of a
modified rate and time when the production has not stabilized prior to shut-in (Daungkaew, 2000).
Using Horner equivalent time, the rate history is simplified and reduced to one single drawdown, with a rate, qlast equal to the
last rate value prior to shut-in, and with a duration, tpe, equal to the correct cumulative production, Q, divided by that last
production rate:
24Q
t pe 
qlast
In this case tpe is calculated to be equal to 4,812.29 hours. Figure 24 presnts the log-log diagnostic plot of Well Z-3, with
Horner equivalent time. Cinco-Ley and Samaniego (1989) suggest a new rate history approximation that combines a Horner
equivalent time with a detailed history for some part of the production history before the test. Using this approach,
Daungkaew et al. (2000) investigated different combinations and concluded that the rates should be described accurately for
the last 40% of the cumulative production, with the first 60% being approximated with a Horner equivalent time, 40% + tpe.
Applying this to the Well Z-3 case resulted in the log-log plot shown in Figure 25.

Due to the nature of the permanent data capture that exists in the digital oilfields, which avails the whole production history,
and, in this case, an extended shut-in period of more than 16 days, it is possible to investigate the effects of eliminating part
of the production history using either the Horner or Cinco-Ley and Samaniego approaches. Figure 26 shows an overlay of
the log-log diagnostic plots of the base case, the Horner, and the Cinco-Ley and Samaniego approaches. The figure hardly
shows any noticeable differences between the three approaches. Only when we consider a far shorter rate history—in this
case, equivalent to the buildup period—that we notice a difference. Figure 27 shows this effect compared with the previous
8 SPE 163723

approximations. It can therefore be concluded that the Horner equivalent time approximation is reasonably valid for buildup
analysis of data preceeded by a long and varied production history.

Conclusions

In this study, we have presented a dynamic pressure transient analysis workflow to utilize data from permanent downhole
gauges availed by the new “Intelligent Field” initiatives in the industry. These workflows leverage the massive investments in
instrumentation, data acquisition, and storage to enable dynamic characterization of hydrocarbon reservoirs. Several
challenges have been faced to establish reliable analysis using these PDHMSs data have been discussed in this study;
including producing time, flow rate simplification, and data interruptions. Adequate utilization of this data in pressure
transient analysis will result in significant cost savings due to a reduction in the number of planned shut-ins for pressure
buildup tests which would normally require well shut-in and, consequently, loss of production from a few days to a few
weeks. In addition, the use of long production history data captured by PDHMSs enables the determination of reliable
reservoir parameters.

Acknowledgemt

The authors would like to thank KFUPM and Saudi Aramco for their support and permission to present the paper at the
Digital Energy Conference.

References

Aghar, H., Carie, M., Elshahawi, H., Gomez, J.R., Saeedi, J., Young, C., Pinguet, B., Swainson,
K., Takla, E., and Theuveny, B. 2007. The Expanding Scope of Well Testing. Oilfield Review19
(1): 44–59.

Al-Dhubaib, T.A., Almadi, S.M., Shenqiti, M.S., and Mansour, A.M. 2008a. I-Field Data
Acquisition and Delivery Infrastructure: Case Study. Paper SPE 112201 presented at the SPE
Intelligent Energy Conference and Exhibition, Amsterdam, The Netherlands, 25–27 February.
doi: 10.2118/112201-MS.

Al-Dhubaib, T.A., Issaka, M.B., Barghouty, M.F., Mubarak, S., Dowais, A.H., Shenqiti, M.S.,
and Ansari, N.H. 2008b. Saudi Aramco Intelligent Field Development Approach: Building the
Surveillance Layer. Paper SPE 112106 presented at the SPE Intelligent Energy Conference and
Exhibition, Amsterdam, The Netherlands, 25–27 February. doi: 10.2118/112106-MS.

Athichanagorn, S., Horne, R.N., and Kikani, J. 1999. Processing and Interpretation of Long-
Term Data Acquired from Permanent Pressure Gauges.SPE Res Eval&Eng5 (5): 384–391.
SPE-80287-PA.doi:10.2118/80287-PA.

Chorneyko, D.M. 2006. Real-Time Reservoir Surveillance Utilizing Permanent Downhole

Pressures - An Operator’s Experience. Paper SPE 103213 presented at the SPE Annual Technical
Conference and Exhibition, San Antonio, Texas, 24–27 September. doi: 10.2118/103213-MS.

Cinco-Ley, H. and Samaniego, V.F. 1989.Use and Misuse of the Superposition Time Function in
Well Test Analysis. Paper SPE 19817 presented at the 64th Annual Technical Conference and
Exhibition, San Antonio, Texas, USA, 8–11 October. doi: 10.2118/19817-MS.

Daungkaew, S., Hollaender, F., and Gringarten, A.C. 2000.Frequently Asked Questions in Well
Test Analysis. Paper SPE 63077 presented at the SPE Annual Technical Conference and
Exhibition, Dallas, Texas, USA, 1–4 October. doi: 10.2118/63077-MS.
SPE 163723 9

de Oliveira Silva, M.I. and Kato, E.T. 2004. Reservoir Management Optimization Using
Permanent Downhole Gauge Data. Paper SPE 90973 presented at the SPE Annual Technical
Conference and Exhibition, Houston, Texas, USA, 26–29 September. doi: 10.2118/90973-MS.

Horne, R.N. 1990. Modern Well Test Analysis: A Computer-Aided Approach. Palo Alto,
California: Petroway.

Horne, R.N. 2007. Listening to the Reservoir—Interpreting Data from Permanent Downhole
Gauges. J Pet Technol59 (12): 78–86. SPE-103513-MS.

Horner, D.R. 1951.Pressure Build-Up in Wells. Paper WPC 4135 presented at the 3rd World
Petroleum Congress, The Hague, The Netherlands, May 28– June 6.

Houzé, O., Viturat, D., and Fjaere, O.S. 2008. Dynamic Flow Analysis. KAPPA Engineering.

Issaka, M.B. and M.M. Al-Buraikan. 2010. Implementation of Dynamic Well Testing
Workflows in Saudi Aramco Intelligent Fields. Paper SPE 128374 presented at the 2010 SPE
Intelligent Energy Conference and Exhibition, Utrecht, The Netherlands. 23-25 March 2010.

Miller, C.C., Dyes, A.B., and Hutchinson Jr., C.A. 1950. The Estimation of Permeability and
Reservoir Pressure From Bottom Hole Pressure Build-Up Characteristics. SPE-950091-G. Trans.
AIME 189: 91–104.

Nestlerode, W.A. 1963. The Use of Pressure Data from Permanently Installed Bottom Hole
Pressure Gauges. Paper SPE 590 presented at the SPE Rocky Mountain Joint Regional Meeting,
Denver, Colorado, USA, May 27–28. doi: 10.2118/590-MS.

Omotosho, R.J. 2004. Permanent Downhole Sensors in Today’s Petroleum Industry: Current
Trends, Problems and Case Studies. MS Thesis, U. of Texas, Texas.

Ortiz, C.E.P., Auguiar, R.B., and Pires, A.P. 2009. Wavelet Filtering of Permanent Downhole
Gauge Data. Paper SPE 123028 presented at the SPE Latin America and Caribbean Petroleum
Conference, Catagena, Colombia, 31 May–3 June. doi: 10.2118/123028-MS.

Ouyang, L.B. and Kikani, J. 2002. Improving Permanent Downhole Gauge (PDG) Data
Processing via Wavelet Analysis. Paper SPE 78290 presented at the 13th Europe Petroleum
Conference, Aberdeen, Scotland, UK, 29–31 October. doi: 10.2118/78290-MS.

Reynolds Jr., A.C. 1986. Errors Involved in Pressure Buildup Analysis. SPE-15903-MS.
Saleri, N.G., Al-Kaabi, A.O., and Muallem, A.S. 2006.Haradh III: A Milestone for Smart Fields.
J PetTechnol58 (11).

Suzuki, S. and Chorneyko, D. 2009.Automatic Detection of Pressure-Buildup Intervals from

Permanent Downhole Pressure Data Using Filter Convolution. Paper SPE 125240 presented at
the SPE Annual Technical Conference and Exhibition, New Orleans, Louisiana, USA, 4–7
October. doi: 10.2118/125240-MS.

Van Gisbergen, S.J.C.H.M and Vandeweijer, A.A.H. 2001.Reliability Analysis of Permanent

Downhole Monitoring Systems.SPE Drill &Compl16 (1): 60–63. SPE-57057-PA.doi:
10.2118/57057-PA.
10 SPE 163723

 
Table 1 Reservoir rock and fluid properties and well data
 
Parameter  Case 1 (Well X‐1)  Case 2 (Well Y‐2)  Case 3 (Well Z‐3) 

Wellbore radius, rw, ft  0.3  0.3542  0.33 

Average reservoir thichness, h, ft  90  110  210 

Average reservoir porosity, ϕ  0.17  0.18  0.18 

Oil FVF, bbl/STB  1.66  1.168  1.19 

Oil viscosity, μo, cp  0.362  1.83  0.75 

Total compressibility, ct, psi‐1  22.2x10‐6 11.1x10‐6  17.6x10‐6

Table 2 Model, Reservoir and Boundary Parameters Estimated

Parameter  Case 1 (Well X‐1)  Case 2 (Well Y‐2)  Case 3 (Well Z‐3) 

Pre‐Acid   Post‐Acid 

Skin factor  –2.36 4.9 –3.1 1.33 

khtotal , md.ft  307  13,700 13,700  123,000 
kavg, md  3.41 125 125 587 
Pi , psi  3,429 1,682 1,636 4,510
Effective well length Lw, ft  1,518 1,182 1,345 ‐ 
Zw, ft  75.42 55 55  ‐ 
kz/kr  0.071 0.51 0.51 ‐ 

Figure 1 Noise and outlier removal using wavelet algorithm with adjustable ribbon.
SPE 163723 11

Figure 2 Wavelet filtration reduces raw pressure data size down to 5% of original size (After Issakaand Buraikan, 2010)

Figure 3 Formation analysis and directional logs for Well X-1

12 SPE 163723

3300
Frozen data

3100
Pressure [psia]

2900

2700

BUP 1 BUP 2 BUP 3

2500
Liquid rate [STB/D]

3000

2000

1000

0
-3200 -3000 -2800 -2600 -2400 -2200 -2000 -1800 -1600 -1400 -1200 -1000 -800 -600
Time [hr]
Figure 4 Rate and pressure data obtained from well X-1

100
Pressure [psi]

10

1E-3 0.01 0.1 1 10 100

Time [hr]

Figure 5 Log-log analysis of the last buildup (BUP 3) for Well X-1
SPE 163723 13

3300

3100
Pressure [psia]

2900

2700

2500

2300
Liquid rate [STB/D]

3000

2000

1000

0
-3200 -3000 -2800 -2600 -2400 -2200 -2000 -1800 -1600 -1400 -1200 -1000 -800 -600
Time [hr]

Figure 6 History matching of pressure data for Well X-1 based on analysis of BU3.

Figure 7 History matching of pressure data for Well X-1by considering all flow and buildup periods.
14 SPE 163723

3300

3100
Pressure [psia]

2900

2700

2500

2300
Liquid rate [STB/D]

3000

2000

1000

0
-1250 -1200 -1150 -1100 -1050 -1000 -950 -900 -850 -800 -750 -700 -650
Time [hr]
Figure 8 Well X-1 frozen data during build-up 3.
Pressure[psi]

100

10 100
Time [hr]
Figure 9 Effect of data interruption (frozen pressure data) on Well X-1 log-log analysis
SPE 163723 15

BUP1

BUP2

BUP3
100
Pressure [psi]

10

1E-3 0.01 0.1 1 10 100

Time [hr]
Figure 10 Overlay of the 3 buildups on the log-log diagnostic plot for Well X-1
Pressure [psi]

100

10 100
Time [hr]

Figure 11 Effect of data interruption (frozen pressure data) on Well X-1 log-log analysis after removal of the “frozen” data
16 SPE 163723

Arab-D

Figure 12 Y-2 well trajectory, formation tops, and formation analysis logs

1650
Pre-acid
Post-acid
1550
Pressure [psia]

1450

1350

1250
Liquid rate [STB/D]

8000

4000

0
18700 18800 18900 19000 19100 19200 19300 19400 19500 19600 19700 19800
Time [hr]
History plot (Pressure [psia], Liquid rate [STB/D] vs Time [hr])
Figure 13 Rate and pressure data obtained from well Y-2
SPE 163723 17

Post-Acid

Pre-Acid

100
Pressure [psi]

10

1E-3 0.01 0.1 1 10 100

Time [hr]

Figure 14 Comparison of pre- and post-acid log-log analysis for WellY-2 as indicated on figure 13.

100
Pressure [psi]

10

1E-3 0.01 0.1 1 10

Time [hr]
Figure 15 Post-acid log-log analysis for Well Y-2, with model match
18 SPE 163723

Figure16 Pre-acid history matching of pressure data for Well Y-2 using all pre-acid flow and buildup periods.

Figure17 Post-acid history matching of pressure data for Well Y-2 using all Post-acid flow and buildup periods.
SPE 163723 19

Avg. = 3,500 bbl

Figure 18 Rate changes on Well Y-2 before the first buildup period of Pre-acid

K-414_3500 - Analysis 1 (ref)

K-414_base - Pre-Acid

100
Pressure [psi]

10

1
1E-3 0.01 0.1 1 10 100
Time [hr]

Figure 19 Well Y-2 log-log analysis overlay of original case vs. simplified rate
20 SPE 163723

Perforation Interval

Figure 20 Well Z-3 formation analysis logs

SPE 163723 21

4280

4230
Frozen data
Pressure [psia]

4180 Frozen data

4130

4080
Liquid rate [STB/D]

10000

5000

0
-7600 -7200 -6800 -6400 -6000 -5600 -5200 -4800 -4400 -4000 -3600 -3200
Time [hr]
Figure 21 Rate and pressure data obtained from Well Z-3

100
Pressure [psi]

10

0.01 0.1 1 10 100

Time [hr]
Figure 22 Matching of the last buildup data from Well Z-3 with a radial composite model
22 SPE 163723

Z-3

Z-105

Figure 23 Comparison between Well Z-3 log-log analyses with nearby Well Z-105
SPE 163723 23

100
Pressure [psi]

10

0.01 0.1 1 10 100

Time [hr]

Figure 24 Log-log analysis of Well Z-3 with Horner approximation

100
Pressure [psi]

10

0.01 0.1 1 10 100

Time [hr]

Figure 25 Log-log analysis of Well Z-3, with the Cinco-Ley–and–Samaniego approximation 40% + tpe
24 SPE 163723

A-25 - Cinco - Composite (ref)

A-25 - Composite
A-25_Horner - Composite

100
Pressure [psi]

10

1
0.01 0.1 1 10 100
Tim e [hr]

Figure 26 Well Z-3 log-log analysis of the base case, Horner, and Cinco-Ley–and–Samaniego approaches

A-25 - Cinco - Composite (ref)

A-25 - Composite
A-25_Horner - Composite
A-25_short history - Composite

100
P re s s u re [p s i]

10

1
1E-3 0.01 0.1 1 10 100
Time [hr]

Figure 27 Well Z-3 log-log analysis of the base case, Horner, and Cinco-Ley–and–Samaniego
approaches, compared w