SPE 142075

Well Integrity Analysis in Gulf of Mexico Wells Using Passive Ultrasonic

Leak Detection Method
J.E. Johns, Seawell; F. Aloisio, Black Elk Energy; D.R. Mayfield, Murphy E&P Co.

Copyright 2011, Society of Petroleum Engineers

This paper was prepared for presentation at the SPE Production and Operations Symposium held in Oklahoma City, Oklahoma, USA, 2729 March 2011.
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
Well integrity management presents a wide variety of challenges for the industry today. With aging fields and
more complex completion techniques coming into play, more efficient methods of well diagnostics and
remediation are demanded. In the GOM, 45% of the wells have sustained casing pressure; therefore, the
importance of having a resource that can provide an effective, accurate method of leak detection is abundantly
clear.
Typical methods of leak detection today include the use of spinners, temperature tools and noise logs.
Mechanical means such as calipers and isolation packers are also employed. While effective for larger leaks,
these methods can produce nebulous results with smaller leaks and can be time consuming.
The frequency spectrum a leak produces is a function of differential pressure, leak magnitude, and leak
geometry. These properties determine whether the frequency is audible, ultrasonic, or both. Typically, smaller
volume leaks with a relatively high differential pressure will generate an ultrasonic signal. Based on this premise,
an ultrasonic logging tool was developed and proven that detects frequency spectrums typically produced by
leaks. The tool has a series of band pass filters which remove virtually all audible noise associated with tool
movement, allowing continuous logging. Because ultrasonic energy will pass through compressed gas and steel,
the tool can detect leaks in secondary barriers as well. Further, as ultrasonic energy attenuates quickly, the tool
locates leaks with a high degree of accuracy. Using this tool, leaks as small as 0.005 gpm have been quickly
located with an accuracy of a foot or less.
This paper will describe a down-hole ultrasonic leak detection tool and provide case histories of where the
ultrasonic leak detection tool was used to find leaks that other methods were unable to locate. A comparison of
the results from conventional leak detection methods will be discussed as well.
Introduction
Well integrity management is a continued and growing concern of all operators in the Gulf of Mexico. As
completions in the region continue to age and the complexities of the completions continue to increase, it stands
to reason that the occurrences of sustained casing pressure will be an ongoing and potentially increasing
problem. In order to efficiently remediate sources of sustained casing pressure one must first fully understand the
nature of the source or sources of communication. A full knowledge of the integrity issue can help the operator
avoid multiple failed repair attempts thereby increasing the overall efficiency of the remediation campaign.
Conventional logging techniques such as spinners, temperature logs and down-hole cameras are widely used
in the detection of leaks in offshore wells. If the leak rate is small, these devices are limited in their effectiveness.
In the case of multiple leaks, conventional logging techniques may detect the largest of the leaks but may miss
secondary leak paths. Mechanical methods are widely used as well but, while relatively inexpensive, can be time
consuming and can produce misleading results.

SPE 142075

For a number of years, a tool operating off the attributes of ultrasonic energy has been utilized for the
detection of leaks in virtually all pressure barriers present in a well. Prior to the introduction of this tool, very small
leaks were practically undetectable. Leaks behind pipe were for the most part impossible to locate unless they
were of a very large magnitude. Utilization of this tool alone or in conjunction with conventional technology has
allowed operators to ascertain an accurate and complete picture of integrity issues within their wells.
Tool Principles and Operation. Ultrasonic frequency generation by a leak is usually associated with
turbulent flow at the leak site. The frequency spectrum produced is a function of the rate, differential pressure
and leak path geometry. The leak media also plays a significant part in the frequency generated. The combination
of these variables will determine whether the frequency is audible, ultrasonic, or both. The ultrasonic logging tool
(Fig. 1) utilizes a sensor that detects a frequency spectrum, including those typically produced by leaks. The
signal is then processed on board the tool where frequency spectrums associated with mechanical noise are
removed before being transmitted to surface or stored in down-hole memory. Nearly all noise associated with tool
movement is filtered out. This allows the tool to be run at 30 fpm in comparison to older sound-based technology
which had to take readings while stationary. The tool is operated on a widely used telemetry platform allowing it
to be used in conjunction with an array of production logging tools. The tool may be run in memory mode for slick
line operations or real time (surface read out) for electric line operations.
The tools accuracy is supported by the characteristics of ultrasound. Ultrasound attenuates at a much higher
rate than audible sound; therefore, moving the tool a short distance from the source will yield a comparatively
large difference in signal strength. This allows the tool to pinpoint the sources of leaks very accurately, often
within inches. Ultrasonic energy will travel through steel which allows leaks to be detected in secondary barriers.
Ultrasonic energy travels very poorly in low pressure gas, therefore gas should be compressed for best logging
results. To detect this ultrasonic frequency spectrum, the leak must be kept active during the logging run.
Figure 2 is a typical log presentation showing three frequency windows of investigation. The three traces
presented are the total energy level (ALD A), a medium-high frequency range (ALD B), and a very high frequency
range (ALD C). These are unit-less measurements of ultrasound signal strength. A casing collar locator (CCL) is
presented for correlation purposes. The leak signature shown in Figure 2 is the first leak detected using the tool.
This is a 0.08 gpm leak with a differential pressure of 900-1200 psi.
Gulf of Mexico Case Histories
Federal regulations mandate that sustained casing pressures in Gulf of Mexico wells must be reported and
remediated. As completions continue to age in the region, occurrences of SCP will continue to be a regularly
combated issue. This section will cite specific case histories involving leak scenarios that have been addressed
in the region.
Injection Well - Packer Leak (Well A)
Well A (Fig. 3) is an active water injection well completed with 5.5 tubing inside 9-5/8 casing. The A annulus
began to build pressure while the well was under injection operations a few years after it was completed. The
pressure increase in the annulus indicated a leak ranging from 0.1 gpm to 0.9 gpm leak rate. The leak rate was
calculated using a fixed volume pressure calculation.
In order to activate the leak for the logging run, the well was brought on injection and the pressure in the
annulus was bled to zero psi. The annular pressure during the run pressured at an approximate rate of 200 psi in
twenty minutes. In order to assure that the leak was active over the course of the run, the annulus was shut in
and allowed to build pressure. The pressure build up in the annulus yielded an average calculated leak rate of
about 0.3 gpm. The annulus was periodically bled down in order to assure maximum differential pressure across
the leak. The injection pressure was maintained at 3,000 psi with a rate ranging from 15,000 to 20,000 BWPD.
This log was run with the tool in memory-mode on slick line. The tool string was run-in-hole to approximately
50 feet below the packer and with the conditions in the well as described above, the run was performed by
logging out of the hole at 30 feet per minute. The resulting log (Fig. 4) indicated a leak at the top of the packer.
Note how the signal is very sharp on the down-hole side of the leak. This is due to the acoustic dampening
caused when the tool string is in proximity to the leak. As the tool is removed from the leak area, the signal
travels further.
A patch was subsequently installed across the leak path. The data from the log enabled accurate placement of
the patch. Further, a full analysis of the integrity of the tubing string from the log assured no other leaks were
present.
For conventional logging tools, a 0.3 gallon per minute leak would be challenging to locate. This challenge
would be compounded if attempts were made to locate the leak while under injection operations. Because the
ultrasonic logging tool is insensitive to laminar flow, it may be used while under injection or production.

SPE 142075

Production Well - Tubing Leak (Well B)

Well B (Fig. 5) is an active gas and oil production well completed with a 4-1/2 inch tubing string inside 9-7/8
casing. The well had begun to exhibit signs of tubing to A annulus communication during production operations.
Pressure testing indicated that the source of communication existed above the subsurface safety valve set at
9,723 MD, potentially in one of the gas lift mandrels. With the safety valve closed, the A annulus and tubing
could be bled to zero pressure. With the safety valve open, the A annulus pressure tracked the tubing pressure.
As the fluid level was unknown, a leak rate could not be calculated accurately. Assuming some range of fluid
levels and that the leak was gas based on bleeds from the A annulus, the leak range could be between 0.03 and
2.8 cu. ft/min.
In order to evaluate the tubing integrity above the safety valve in this well, the ultrasonic leak detection tool
was run in memory mode on slick line. In addition, a conventional production logging string from the slick line
provider was run in conjunction with the tool. The string included a caged full bore spinner, continuous flow meter
spinner, pressure sensor and temperature tool along with a casing collar locator (Fig. 6).
In order to keep the leak active during the course of the log, the safety valve was closed thereby trapping the
shut-in pressure above this device. The A annulus was then bled as low as possible to the flare line thereby
generating the differential to drive the leak. As a contingency, it was planned to pump dead crude into the tubing
to maintain differential if required. In general, the tubing pressure averaged around 1,200 psi and the A annulus
averaged about 150 psi.
A down pass at thirty feet per minute was performed followed by an up pass at the same line speed. The
pressure conditions in the well are shown in Figure 7 along with pertinent events during the logging run. The leak
was detected on the down pass (Fig. 8) and the up pass (Fig. 9). In both cases the ultrasonic leak detection tool
showed the exact location of the leak at a tubing joint collar at 3,074 MD. While there is a mild temperature
deflection in this area, this indication would not be enough information upon which to base decisions toward
remediation efforts. The spinner data has been omitted from these log examples because neither tool functioned
over the course of the run. Reviewing the signal scale on the leak detection tool shows that the signal of the leak
during the up pass was much higher than that of the down pass. The pressure data indicates that about 150 psi
more differential was present during the down pass than during the up pass. The pressure tool, however,
indicates the reverse. It is believed that the introduction of fluid into the tubing string increased the differential
pressure due to hydrostatic head at this location. Additionally, acoustic energy travels better in fluids than in
gasses, and fluid or mixed media has more of a propensity to generate turbulent flow at the leak site. These
factors likely influenced the signal strength at this region.
Had the leak been located at the gas lift mandrel, the operator would have had an option to replace the
devices in order to remedy the tubing to annulus communication. As the leak was at a tubing joint collar, a patch
would have been necessary to fix the leak. Because this would have limited access to critical wellbore
components deeper in the well due to the restricted ID of the patch, a decision was made to pull and replace the
tubing. When the tubing was removed, a failed connection was located at the joint located by the leak detection
tool. See (Fig. 10). Upon close examination of the tool joint, (Fig. 11) signs of erosion could be seen at the failure
point.
Production Well - Multiple Tubing Leaks (Well C)
Well C (Fig. 12) is an active gas production well completed with a 2-7/8 inch tubing string inside 9-5/8 casing.
The well had been recompleted and upon being brought on production experienced pressure build up in the
annulus beyond thermal expansion considerations. The build up rate in the annulus measured approximately 50
to 100 psi in thirty minutes. This translated to a leak rate of approximately 0.18 gallons per minute.
In order to evaluate the tubing integrity the ultrasonic leak detection tool was run in memory mode on slick line.
As was the case with Well B, a conventional production logging string was used in the same string. (Fig. 6). The
leak in this case was activated by using the shut in tubing pressure to drive the leak. The shut in tubing pressure
remained at 6,850 psi during the course of the log and the A annulus was vented to a bleed tank. The pressure
during the bleed never went below 47 psi. The annulus was shut in momentarily in order to assure the leak was
active. The resulting calculated leak rate based on the pressure build up was 0.193 gpm. The logging pass was
performed by running the tool string in the hole to approximately 50 below the packer set at 13,342 MD and
logging up at 30 feet per minute. This would allow the packer to be surveyed during the pass.
After completing the run, downloading and evaluating the data, it was found that a minimum of four leaks were
present. These leaks were present at 9,832 MD, 8,274 MD, 7,199 MD and 5,015 MD. Figure 13 shows an
example of one of the major leak points. The other tools in the string produced either no evidence or no
conclusive evidence of leak activity. By decreasing the scale of the leak detection tool thereby elevating the
resolution, it was found that possible secondary leaks existed around the main leak signatures (Fig. 14). This
condition existed at two large leak signals at 7,199 MD and 8,274 MD indicating that a total of four definite leaks
and four possible leaks were in the tubing string. The leaks in the string were quickly and accurately located
where conventional tools would have been ineffective and time consuming.

SPE 142075

Because too many leaks existed in the string to be effectively and efficiently repaired, the newly installed
tubing string was pulled and replaced. An inspection of the suspect tool joints located by the leak detection tool
revealed defects in sixteen out of sixteen joints inspected. Flaws such as corrosion (Fig. 15), major pitting on
sealing surfaces and threads (Fig. 16) and corrosion on couplings (Fig. 17) were noted.
Conclusions
The ultrasonic leak detection method is an efficient and economic option in detecting leaks as small as
.024 gpm.
Utilizing the tool alone or in concert with conventional logging technology can provide a
comprehensive assessment of wellbore barrier integrity.
Using the data provided by the tool allows operators to better plan and execute remediation
operations.
The tool detects leaks which are undetectable by conventional means.
Acknowledgements
This paper reflects the views of the authors, and does not necessarily reflect the views of the authors respective
employers or other working interest owners.
Nomenclature
Ft
in.
MD
Psi
gpm
CCL
GLM
SCP
SRO
BWPD

= Feet
= Inch
= Measured Depth
= Pounds (f) per Square Inch
=gallons per minute
=Casing Collar Locator
=Gas Lift Mandrel
=Sustained Casing Pressure
=Surface Read Out
=Barrels of Water Per Day

References
1.

2.

3.

J.E. Johns, C.G. Blount, J.C. Dethlefs, J.Y. Julian, M.J. Loveland, M.L. McConnell, G.L. Schwartz; Applied
Ultrasonic Technology in Wellbore Leak Detection and Case Histories in Alaska North Slope Wells paper
SPE 102815 presented at the 2006 SPE Annual Technical Conference and Exhibition held in San Antonio,
Texas, U.S.A., 2427 September 2006.
J.Y. Julian, G. E. King, J. E. Johns, J. K. Sack, D. B. Robertson; Applied Ultrasonic Technology in Wellbore
Leak Detection and Case Histories in Alaska North Slope Wells paper SPE 108906 presented at the 2007
International Oil Conference and Exhibition in Mexico held in Veracruz, Mexico, 2730 June 2007.
M. A. Farooqui, A. S. Al-Reyahi, K.K. Nasr; Application of Ultrasonic Technology for Well Leak Detection
paper IPTC 11583-MS presented at the International Petroleum Technology Conference, 2007.

SI Metric Conversion Factors

Bbl x 1.590
E - 01 = m3
Ft x 3.048
E - 01 = m
in. x 2.54
E + 00 = cm
Lbf x 4.448 222 E + 00 = N
Psi x 6.894 757 E - 03 = Mpa

SPE 142075

Figures
Sensor

Figure 1. Ultrasonic logging tool.

Figure 2. Sample log. This is a leak in a chemical injection mandrel detected while the well was on injection. The leak rate was 0.08
gpm calculated from pressure build up in the annulus.

SPE 142075

Figure 3. Well A. High-volume water injection well completed with 5-1/2 tubing.

SPE 142075

Figure 4. Packer leak located while injecting up to 20,000 BWPD.

SPE 142075

Figure 5. Well B. Active producing well completed with 4-1/2 tubing and multiple GLMs.

SPE 142075

Figure 6. Memory logging string used in wells B and C. Includes a conventional cased hole logging string and the ultrasonic leak
detection tool.

Figure 7. Pressure conditions in Well B during logging and selected events.

10

SPE 142075

Figure 8. Well B downward logging pass.

Figure 9. Well B upward logging pass. Note the difference in signal strength from ultrasonic leak detection tool.

SPE 142075

11

Figure 10. Failed tubing connection. Well B.

Figure 11. Failed tubing connection, Well B. Signs of erosion indicating leak path in the connection.

12

SPE 142075

Figure 12. Well C. Production well recently recompleted with 2-7/8 tubing string.

SPE 142075

13

Figure 13. Well C major leak point. One of eight leaks located in the well located by the ultrasonic logging tool. Conventional casedhole logging data included.

Figure 14. Expanded view of leak detection tool signal showing secondary leak paths in other tubing joints.

14

SPE 142075

Figure 15. Corrosion located in the tubing string recovered from Well C.

Figure 16. Major pitting on the sealing surface and threads of the tubing string from Well C.

SPE 142075

15

Figure 17. Internal corrosion located on a coupling from Well C.