Scanning for Downhole Corrosion - Schlumberger · Scanning for Downhole Corrosion Oilfield Review Spring 2010: ... design may call for cathodic protection, ... Koch et al, reference

42 Oilfield Review

Scanning for Downhole Corrosion

Oilfield Review Spring 2010: 22, no. 1. Copyright © 2010 Schlumberger.For help in preparation of this article, thanks to Nash Asrar, Richard Byrd and Martin Isaacs, Sugar Land, Texas, USA.EM Pipe Scanner and PS Platform are marks of Schlumberger.

Corrosion is one of the many ways that nature humbles human activity. It is a relentless process that, unchecked, renders our most marvelous constructions into little more than junk. How-ever, for our global economic well-being we rely on an infrastructure of metal in buildings, bridges, factories, vehicles and pipelines. The network of pipes leading from hydrocarbon-bear-ing strata deep underground to refineries—even to the burner tips in our homes—is critical for supplying the energy to fuel our economy. And so the battle against corrosion continues.

It is an expensive battle. In a massive study published in 2001, the direct total cost of corro-sion in the USA was calculated to be US $276 bil-lion per year, about 3.1% of the US gross domestic product (GDP).1 Costs worldwide are estimated to be a similar fraction of the global GDP, result-ing in a worldwide cost of about US $1.8 trillion.2 Within the USA corrosion costs in the E&P indus-try were estimated to be almost US $1.4 billion annually, comprising US $589 million for surface piping and facilities, US $463 million in downhole

tubing expenses and US $320 million in capital expenditures.3 Expenses and lost revenue result-ing from lost production and leakage were not included in these figures.

Corrosion is caused by several mechanisms, including electrochemical, chemical and mechan-ical effects.4 One way of mitigating this action is to substitute corrosion-resistant alloys, such as chro-mium steel instead of carbon steel. Another is to use a coating, the simplest of which is paint. A design may call for cathodic protection, which transfers the corrosive effect from essential struc-tural components to a nonessential, sacrificial piece of metal. This approach can also be achieved for large structures by supplying a DC current.

A primary element in the battle against corro-sion is monitoring. In addition to mitigating direct costs, corrosion monitoring also reduces risks to safety and the environment by detecting weak spots before they fail or leak. At the surface, monitoring can sometimes be done visually, but there are also tools designed to detect hidden metal loss due to corrosion. For downhole casing

1. Koch GH, Brongers MPH, Thompson NG, Virmani YP and Payer JH: “Corrosion Cost and Preventive Strategies in the United States,” Report FHWA-RD-01-156 prepared by CC Technologies Laboratories, Inc., for the US Federal Highway Administration (FHWA), Office of Infrastructure Research and Development (September 2001), http://www.corrosioncost.com/home.html (accessed February 3, 2010).

Irlec Alexandra AcuñaAlan MonsegueThe Hague, The Netherlands

Thilo M. BrillPrinceton, New Jersey, USA

Hilbrand GravenFrans MuldersGDF SUEZ E&P Nederland B.V.Zoetermeer, The Netherlands

Jean-Luc Le CalvezEdward A. NicholsFernando Zapata BermudezClamart, France

Dian M. NotoadinegoroBalikpapan, Indonesia

Ivan SofronovMoscow, Russia

2. Hays GF: “Now Is the Time,” Advanced Materials Research 95 (2010), http://www.scientific.net/AMR.95.-2.pdf (accessed February 3, 2010).

3. Koch et al, reference 1.4. For more on the basics of corrosion: Brondel D,

Edwards R, Hayman A, Hill D, Mehta S and Semerad T: “Corrosion in the Oil Industry,” Oilfield Review 6, no. 2 (April 1994): 4–18.

5. For more on corrosion measurement: Cased Hole Log Interpretation Principles/Applications. Houston: Schlumberger Educational Services, 1989.

Electromagnetic induction tools can be used to investigate corrosion and pitting in

downhole pipe. Using a combination of sensors, a new corrosion-monitoring tool

provides measurements of average pipe thickness and two-dimensional imaging of

the pipe wall to distinguish between internal and external damage. The tool also

delivers a qualitative measurement of metal loss in outer casing strings.

30039schD7R1 42-50.indd 1 5/1/10 5:33 PM

Spring 2010 43

and tubing strings, logging tools are the only means of monitoring.

The four main types of corrosion-monitoring logging tools today are mechanical calipers, ultrasonic acoustic tools, cameras and electro-magnetic (EM) tools.5 Multifingered calipers are well-established tools for evaluating internal problems, but they provide no data about exter-nal corrosion and are affected by scale buildup on the inner wall. Ultrasonic measurements yield excellent pipe thickness information in a single casing string and have superior azimuthal resolution. However, ultrasonic tools are unable to operate in gas wells, through tight restrictions or on monocables, and their measurements can be disrupted by pipe roughness and excessive corrosion. Downhole cameras can also be used for corrosion detection, if the wellbore is filled with gas or another clear fluid.

EM corrosion-monitoring tools in use today rely on one of two physical principles: flux leak-age and electromagnetic induction. A flux leak-age tool uses a permanent or electromagnet to magnetize the pipe to near saturation. Near a pit, hole or corrosion patch, some of the magnetic flux leaks out of the metal; this flux leakage is detected by coils on the tool’s pad-mounted sen-sors. A flux leakage tool can sense defects on the inside or outside of the casing, but since the mag-net must be as close as possible to the pipe, a cas-ing examination requires operators to pull the tubing out of the hole. In addition, flux leakage tools are good at measuring sudden thickness changes, but they are not effective if the corro-sion is constant or varies slowly over a whole sec-tion of pipe.

The most recent Schlumberger EM induction sonde for corrosion monitoring is the EM Pipe Scanner tool. It has excellent vertical resolution and good thickness resolution, although the azimuthal resolution is not as high as that of ultrasonic measurements. The tool detects metal loss both inside and outside of casing as well as loss from an outer casing string when multiple strings are present. It can operate in any fluid, can be run on monocables and can pass through small restrictions.

This article describes the physics of EM induction as applied to this tool. Case studies from Indonesia and the Netherlands illustrate tool use in the field.

Tool PhysicsThe EM Pipe Scanner tool provides nondestruc-tive casing inspection using electromagnetic induction. Its principle of operation is similar to that of a transformer with losses. A transformer’s

primary coil generates a time-varying magnetic field that flows through a magnetic core to induce a voltage in its secondary coil. In comparison, the tool’s transmitter coil—acting as a primary coil—generates a magnetic field whose flux is guided by the casing; this magnetic flux induces a volt-age in a secondary or receiver coil.

The flux guide provided by the casing is lossy—energy is lost or dissipated in the medium—because of the currents induced in the casing metal. The tool measures these losses to determine geometrical, electrical and magnetic properties of the casing, including the presence of corrosion or pitting in the pipe.

The EM Pipe Scanner tool contains several EM transmitters and associated receivers. The

basic EM physics is the same for all transmitter-receiver pairs, but the responses differ because of the frequency of the signal and the transmit-ter-receiver spacing. The general aspects of the physics of EM induction are described next, followed by specific tool applications.

When a time-varying EM wave penetrates a conductive body, such as the steel pipe of tubing or casing, its magnitude decays exponentially. The rate of decay depends on the body’s conductivity and magnetic permeability and the frequency of the wave; the decay rate is characterized by a length called the skin depth, δ (above). The phase of the wave also changes as it passes through the conductor, a property that is useful in measuring the thickness of the material.

> Skin depth. When an EM field impinging from below encounters a conducting material such as the metal of a pipe (blue), the amplitude (red, top left) decreases exponentially with a characteristic rate given by the skin depth δ. An unattenuated signal (dashed gray) is shown for comparison. At the same time, the phase shifts almost linearly with distance of travel through the metal (top right). The phase can change more than 360°, as it does here. Skin depth varies greatly, depending on the medium (table, bottom). Air has properties close to those of a vacuum, which has an infinite skin depth at all frequencies. A conductive and ferromagnetic material, such as casing, has a short skin depth. All media other than a vacuum have shorter skin depths at higher frequencies. Resistivity is the inverse of conductivity (σ). Angular frequency ω is 2πf. The values used for µr and σ are typical for the various media.

Oilfield ReviewSpring 10PipeScanner Fig. 1ORSPRG10-PPSCN Fig. 1

Electric field amplitude, relative–100 0 100 0 90 180 270 360

Electric field phase shift, deg

5

4

3

2

1

0

–1

–2

5

4

3

2

1

0

–1

–2

Pipe wall

Dist

ance

in m

ultip

les

of s

kin

dept

h,δ

Dist

ance

in m

ultip

les

of s

kin

dept

h,δ

Oilfield ReviewSpring 10PipeScanner Fig. 1AORSPRG10-PPSCN Fig. 1A

Arbitrary Infinite01Vacuum

Brine 1 5010 22,500

2,2501,000

10 21

1,000 2

10

1,000

7

0.7

Copper 1

100Casing

60 x 106

5 x 106

Medium Frequency,f, Hz

Conductivity, , S/mσ

Relative Magnetic Permeability, μr

Skin Depth, , mmδ

30039schD7R1.indd 2 6/10/10 2:43 PM

44 Oilfield Review

The EM signal decay results from response currents—called eddy currents—created in the conductor. In the geometry of a circular pipe with a transmitter at its axis, the eddy current forms a closed current sheet flowing azimuthally within the pipe wall. The eddy current sets up a response EM field that acts to oppose the primary field from the transmitter. This attenuates the field much more rapidly than when no casing is present.

If the pipe has a defect, such as one caused by corrosion or pitting, the eddy current can no longer form a closed sheet since it is forced to bypass the defect. This behavior is like that of water in a stream flowing around a rock in its path. The response EM field is altered by this anomalous flow path. Receivers located in multiple pads pressed against the inside of the casing respond to these perturbations in the current flow path. The output of the sensor pads supplies a 2D image from which engineers can assess the altered EM field that provides evidence of the damage.

The EM field generated by a transmitter coil extends throughout space to infinity. At physical boundaries within that space, such as the inner and outer walls of the pipe, the field from both sides must match. Because of this required match of the boundary conditions, the behavior

of the field in each region influences its behavior in all the others (below). The total EM field can be represented as a superposition of three con-stituent fields.

The first field is that of a transmitter in free space, that is, in the absence of any pipe. At a distance sufficiently far from the coil, this is the weakly attenuated field of a simple mag-netic dipole.

The second field is added by the presence of a pipe that is thick enough that any EM field pene-trating it is completely absorbed. This introduces the influence of the boundary condition at the inner surface of the pipe; the outer boundary plays no role in this field. The eddy currents induced inside the conductive pipe give rise to a secondary response field. It is opposed to the source field—out of phase by 180°—and has similar amplitude. As a result, the sum of the first and second fields—termed the direct field within the pipe—is weak and decays exponentially. This situation is similar to the case of propagating microwaves in a wave-guide: The frequencies used by the tool are beyond the cutoff frequency, so the signal attenuates rap-idly within the pipe.6 Axial attenuation is faster than radial attenuation.

The thick-pipe approximation is appropriate for high-frequency signals because the field decays rapidly within the metal and eddy cur-rents are localized near the inner surface of the pipe. The response field from such a signal, which is affected by the conductivity and permeability of the steel, can be detected by a receiver coil that is close to the transmitter.

Since the direct field is the sum of the free-space field and the induced secondary field of a thick pipe, it does not contain any informa-tion on the thickness of the pipe. It is the contri-bution of a third field—the indirect field—that provides this.

The indirect field is caused by the boundary condition at the outer pipe surface, which was neglected for the case of a thick pipe. At great distance this field also must satisfy the free-space boundary condition of a simple magnetic dipole. This third field can be seen, somewhat simplisti-cally, as relating to the reflection of the penetrat-ing EM field at the outer pipe surface. The indirect field is strongly attenuated by passage of the signal through the pipe metal, but it contains the sought-after information about the pipe thickness. This information comes from the phase change that is approximately linear with distance of travel within the pipe, as discussed previously in the description of skin depth.

> Lines of potential for transmitter coils inside a pipe from finite-difference models. Each color contour represents a one-decibel decrease in the potential voltage of the electric field created by a transmitter coil. A low-frequency signal (left) penetrates the pipe wall and decays slowly outside the pipe. Because of this, in the RFEC region at large offset from the transmitter, the main flow of energy (yellow arrows) goes through the pipe wall, along the outside of the pipe, then back through the wall to the inside of the pipe. In contrast, the direct signal within the pipe (black arrow) decays rapidly. A high-frequency signal (right) reaches the pipe wall but decays rapidly within the pipe. The signal and response in this NFEC region (black arrows) provide information about the properties of the metal on the inner surface of the pipe wall. The radial scale is greatly expanded in comparison to the axial scale, and the low- and high-frequency transmitter coils are of typical sizes for an EM corrosion-monitoring tool.

Remote-Field Eddy Current (RFEC)

Distance along pipe, in.0 25 50 75 100

Radi

us, i

n.

1

3

4

5

2

0

Transmittercoil

0 25 50 75 100

Near-Field Eddy Current (NFEC)

Distance along pipe, in.

Radi

us, i

n.

1

3

4

5

2

0

Transmittercoil

Pipe wall

Oilfield ReviewSpring 10PipeScanner Fig. 2ORSPRG10-PPSCN Fig. 2

30039schD7R1 42-50.indd 3 5/1/10 5:33 PM

Spring 2010 45

The skin depth is frequency dependent, so only low-frequency signals contribute to this indirect field. The low-frequency signal propagates beyond the outer wall of the pipe into material of lower conductivity, such as cement, rock, oil, brine or, in the case of coaxial pipes, a gas such as air. If there are multiple strings and the signal is strong enough, the signal will continue propagating through the other pipes. It will decay in the same manner as for the innermost string, acquiring a similar thickness-dependent phase shift.

The signal outside the pipe (or pipes) is dom-inated by the field set up by the eddy currents in the pipe metal. Because of the approximate dipole behavior of the field, the signal decays as the inverse cube of the distance traveled. This is a significantly smaller decay than that experi-enced by the direct signal inside the pipe. Thus, with use of a low-frequency signal and long trans-mitter-receiver spacing, the direct field may be much smaller than the indirect field at the receiver position. For the geometry of the tool and the low-frequency signals used, that spacing is about twice the diameter of the pipe.

Since the direct field is so small at this dis-tance from the transmitter, the path of energy flow follows the indirect field. The field decayed while traveling from the coil to the inner pipe

wall. It then decayed exponentially passing through the metal (and had its phase shifted in that traversal). In the medium outside the pipe it decayed by the inverse cube of the distance trav-elled. The field experienced a second attenuation and phase shift as it passed through the pipe metal to the receiver coil, which measures an induced voltage.7

In practice, the tool signal is normalized by a measurement in air to cancel out geometry and tool effects. This leaves a normalized signal that has been attenuated by the product of the expo-nential decay in the metal (including the phase shift) and constant geometrical factors. Metal loss from pitting or corrosion affects both the phase shift and the attenuation detected at the receiver coil (above).

The physical behavior of the field, given the geometry of coils inside a conducting pipe, pro-vides a neat division into two regions and two frequency ranges, each of which has a relatively easy-to-interpret measurement. With a short transmitter-receiver offset a high-frequency signal can be used to investigate the properties of the inner wall of the pipe. This configuration measures the direct field from the eddy currents in the pipe near the receiver coil. This is termed the near-field eddy current (NFEC) region.

A long transmitter-receiver offset with a low-frequency signal investigates what is called the remote-field eddy current (RFEC) region. This region is dominated by the indirect field, which involves the signal path described previously: The path goes through the pipe metal twice in its traversal from transmitter to receiver. That pas-sage through the metal generates both signal attenuation and a phase shift.

Between the RFEC and the NFEC lies the transition region. Both the direct and indirect signals influence the field here, and the interpre-tation may be quite complex. For that reason, commercial induction-tool designs for corrosion detection avoid placing receivers in this region.

6. At certain frequencies a waveguide such as a metal pipe transmits EM signals with little loss. This range is bounded by the upper and lower cutoff frequencies; signals beyond those cutoffs decay exponentially with distance.

7. Although it seems counterintuitive to be able to measure pipe thickness using a source and receiver that are both inside the pipe, the physics is well-defined. The energy flux, as indicated by the Poynting vector, flows approxi mately radially outward through the pipe wall at the transmitter, along the outside wall of the pipe, then approximately radially inward again, providing the receiver is more than about two pipe diameters from the transmitter. For an example of a finite-element analysis: Lord W, Sun YS, Udpa SS and Nath S: “A Finite Element Study of the Remote Field Eddy Current Phenomenon,” IEEE Transactions on Magnetics 24, no. 1 (January 1988): 435–438.

> Response to a low-frequency source in a grooved pipe. Lines of electric potential (left) and phase (right) are perturbed by grooves on the inner (white box at 10 in.) and outer (white box at 90 in.) surfaces of the pipe wall. Both the potential and phase perturbations within the pipe where measurements are made are identical for the inside and the outside grooves.

Oilfield ReviewSpring 10PipeScanner Fig.2AORSPRG10-PPSCN Fig. 2A

Pipe wall

Potential

Distance along pipe, in.0 25 50 75 100

Radi

us, i

n.

1

3

4

5

2

0

Transmittercoil

0 25 50 75 100

Phase

Distance along pipe, in.

Radi

us, i

n.

1

3

4

5

2

0

Transmittercoil

30039schD7R1 42-50.indd 4 5/1/10 5:33 PM

46 Oilfield Review

Applying Principles to MeasurementThe EM Pipe Scanner tool takes advantage of both the skin-depth effect and the difference in signal between the near and remote regions to make four distinct measurements (left). The first determines the casing electrical and magnetic properties, referred to as impedance or Z proper-ties. The transmitter sends high-frequency sig-nals to the pipe and back to receivers mounted on the tool mandrel at a short offset, making this an NFEC measurement. The second measures the average thickness of the metal normalized by the skin depth; it uses a low-frequency signal in the RFEC region. The final two measurements are 2D images of the pipe using 18 pad sensors pressed against the inner wall of the pipe. One image uses low-frequency signals in the RFEC region to obtain 2D thickness information. The other uses high-frequency, NFEC signals to discriminate inner-wall features from those elsewhere.

Z-properties measurement—The electro-magnetic properties of the pipe must be known to interpret other tool measurements. Two trans-mitter-receiver offsets of 1.5 and 2.5 in. [3.81 and 6.35 cm] are available; the operator selects which to use based on the pipe diameter. The Z-properties system transmits three signals rang-ing from medium to high frequency, each having a skin depth small enough that the signal does not penetrate far into the pipe wall. The resulting measurement is a function of two quantities: the

Oilfield ReviewSpring 10PipeScanner Fig. 3ORSPRG10-PPSCN Fig. 3

Sensitiveregion

Discriminatortransmitter, TH

Tool outerdiameter

PipePad

receiver, RP

Skin-depth decay

Average thickness

2D thickness

2D discrimination

RZ

RZ

TZ

RLS

RLL

RLS

RLL

RLL

RLS

TL

TL

RP RP

TH

RP

TL

RLS

RLL

TH

RP

Z properties

RZ

RZ

TZ

τ =ωσ

μ1 1μ ID0

d, Coil configurations for the EM Pipe Scanner tool. The tool makes four measurements. The Z-properties measurement (bottom) uses a transmitter (TZ) operating at three frequencies and one of two nearby receivers (RZ).The response signal can be used to determine a quantity, τ, that is a function of the pipe ID, the angular frequency ω, and the EM properties of the pipe metal: the permeability µ and the conductivity σ. The term µ0 is the constant permeability of free space. The average thickness d is determined from the low-frequency transmitter (TL) in the middle of the tool, along with two receivers above and two below the transmitter (lower middle). Two low-frequency receivers (RLL) are termed long-spacing receivers and two are termed short-spacing receivers (RLS), but all are in the RFEC region. The phase shift of the signal—which is a function of skin depth δ—as it goes through the pipe at the transmitter and again at each receiver is used to determine the EM thickness of the pipe d/δ. Near the top of the tool 18 caliper arms press pad receivers (RP) against the inside of the pipe. Combining these sensors with the low-frequency signal from the transmitter (TL) at the middle of the tool provides a 2D thickness measurement (upper middle). The 18 sensors are also used with a high-frequency discriminator transmitter (TH) located on the tool mandrel in line with the sensor pads (top). The high-frequency signal does not penetrate the pipe wall, so this part of the tool provides a 2D map that discriminates damage on the inside wall from other signals.

30039schD7R1 42-50.indd 5 5/1/10 5:33 PM

Spring 2010 47

inside pipe diameter and the square root of the ratio of the magnetic permeability and conductiv-ity of the metal. A physical model helps define the geometry and EM properties as a solution of an inverse problem.

Average EM thickness—In the RFEC region the phase change of a low-frequency signal is almost a linear function of the thickness of the pipe wall, expressed as a ratio of the actual thick-ness d to the skin depth, or d/δ. As the signal passes through the pipe at the transmitter, the phase change is proportional to d/δ, then as it passes back through the pipe at the receiver, the phase changes again proportionally to d/δ. Because of the cylindrical symmetry with trans-mitter and receiver at the center of the pipe, the thickness measurement is an average over the cir-cumference at the two locations. For multiple cas-ing strings the result is qualitative, but the thickness measurement can be compared with those of past and future runs to indicate changes.

With inclusion of the Z-properties measure-ment the thickness of a single string can be cal-culated from either the conductivity of the pipe or its magnetic permeability. The conductivity depends on the pipe chemistry and is typically constant for a given pipe joint and even for a majority of joints in a well, since they often come from one manufacturing run. A computation based on conductivity provides the basic mea-surement of thickness. In contrast, the magnetic permeability is highly variable, so derivation of the thickness based on permeability is used as a quality-control measure.

Thickness is measured at a user-selected fre-quency. The operator typically chooses a signal at 8.75 Hz for multiple strings, at 17.5 or 35 Hz for single strings, and at 70 Hz for chromium-steel strings. Processing combines data from multiple receivers, all at offsets sufficient to be in the RFEC region, to remove ghosts.8 Although the thickness is almost a linear function of the phase shift, more-accurate values are obtained by inver-sion modeling to account for nonlinearity.

2D thickness imaging—A high-resolution thickness image is obtained by 18 sensor pads pressed against the inner pipe surface, incorpo-rating the same low-frequency transmitter as that used for the average EM thickness measure-ment (above). Each pad is sensitive to the nearby pipe thickness, sampling an azimuthal area extending about 0.5 in. [1.27 cm] on either side of the pad. Coverage of the inner pipe surface by the tool depends on the pipe diameter and weight. The minimum pipe size that can be accessed is 27/8-in. OD, and 100% coverage of a single string is possible for up to 7-in. OD for heavyweight pipe. The tool can make accurate measurements in a maximum pipe ID of 95/8 in.

8. A ghost is a duplicate of the signal generated by a defect. It results from the traversal of the signal through the pipe at both the transmitter and receiver locations. Thus, a defect is indicated once when the transmitter passes it and again when the receiver passes it. Use of several offset receivers allows addition and subtraction of logs to remove the ghost.

> Configuration of 18 arms with sensors. Wellsite sensor experts examine and service the sensor pads after a logging run.

Oilfield ReviewSpring 10PipeScanner Fig. 4ORSPRG10-PPSCN Fig. 4

30039schD7R1 42-50.indd 6 5/1/10 5:33 PM

48 Oilfield Review

, Uncorroded pipe in a double string. This log illustrates the response of an uncorroded interval of pipe, measured using both the EM Pipe Scanner tool and the PS Platform multifinger-caliper imaging tool (PMIT). The 2D thickness display (Track 4) has been normalized by subtracting the average measurement of all 18 sensors from each sensor measurement. Other than background noise, only the casing collars are present as horizontal bands of darker colors. The 2D discriminator image (Track 5) is relatively featureless, as is the PMIT radii image (Track 3), except for some indications of casing collars. Track 1 contains the ID measurements from both tools, which agree—within 5%—with the nominal value. The casing properties measurement (Track 1, gold) is almost constant through this section. The average EM thickness measurement (Track 1, green) and the 2D thickness image (Track 4) detected a collar on the outer casing string at X,583 m, which was not detected by the other measurements, including the casing collar locator (CCL, Track 2).

Oilfield ReviewSpring 10PipeScanner Fig. 5ORSPRG10-PPSCN Fig. 5

Outercasingcollar

Radii MinusAverage (Caliper)

EM Thickness

Casing Inner Diameter (EM)

Maximum Internal Radius (Caliper)

Minimum Internal Radius (Caliper)

Average Internal Radius (Caliper)

Casing Properties (EM)

RelativeBearing

Tension

CableSpeed

Gamma Ray

Discriminated CCL2D Discriminator

MinusAverage (EM)

2D ThicknessMinus

Average (EM)20 200

1.85 2.35in.

1.85 2.35in.

1.85 2.35in.

3.7 4.7in.

0 1.5in.

0 1,200m/h

10,000 lbf

0 360deg

X,580

X,600

9 –1V

0 150gAPI

Depth,m

0

X,590

, Corrosion at perforations in Kampung Baru field in a well producing natural gas with H2S. The 2D thickness image (Track 6) clearly shows metal loss (reds) below X,Y15 ft, while the 2D discriminator log (Track 7) shows only the perforations (browns). This observation indicates that the loss is on the outside wall of the casing. Higher in the interval shown, the log responses are evidence of casing collars and pipe manufacturing patterns: Pipe is manufactured from flat steel and then rolled and welded, creating seams that are seen by pipe-analysis tools.

Oilfield ReviewSpring 10PipeScanner Fig. 6ORSPRG10-PPSCN Fig. 6

Casing InnerDiameter

EMThickness

CasingProperties

Double-CoilAmplitude

Double-CoilPhase

2D ThicknessMinus Average

2D DiscriminatorMinus AverageDepth,

ft

in. dB deg in. in.in.4 6 0 0.4 0 10 –30 0 0 270 –0.2 0.2 –0.15 0.15

X,X40

X,X60

X,X80

X,Y00

X,Y20

X,Y40

X,Y60

30039schD7R1 42-50.indd 7 5/1/10 5:33 PM

Spring 2010 49

9. “HALFYR: EWC: Energy World Corporation Half Year Ended 31 December—Directors’ Report,” March 2, 2009, http://www.findata.co.nz/Markets/NZX/14125/HALFYR_EWC_Energy_World_Corporation_Half_Year_Ended_31_December.htm (accessed February 22, 2010).

In multiple-string casing designs the tool measurement includes all strings (out to its signal-to-noise limit) but is most influenced by the innermost string. Like the average thickness measurement, the 2D thickness image is based on the phase shift of the signal as it passes through the metal of the pipe wall or walls. It can be normalized by subtracting the average of all 18 measurements at that location. This removes thickness variation that is uniform around the pipe, such as that caused by a casing collar (pre-vious page, top).

2D discrimination imaging—The 2D thick-ness image does not distinguish between defects on the inside or the outside of the inner tubing. To obtain that measurement, the tool has a high-frequency (8-kHz) transmitter mounted on the tool mandrel at the center of the ring of 18 sensor pads. The high-frequency signal barely pene-trates the metal of the pipe, so the response detected by this NFEC signal is strictly from the inner surface of the pipe, immediately adjacent to the pads. Thus, if an anomaly appears on the 2D thickness image but not on the 2D discrimina-tor image, it cannot be on the inner wall of the pipe. The 2D discriminator image can also be nor-malized by subtracting the azimuthal average.

Finding Corroded PipeEnergy Equity Epic operates the onshore Kam-pung Baru gas field in Sulawesi, Indonesia. The produced gas contains both carbon dioxide [CO2] and hydrogen sulfide [H2S]; the stream is treated to remove water and these corrosive gases at a central processing facility before transport to a power plant.9 The field has three producing wells that have been operating for 12 years. Because of the potential for pipe corro-sion caused by H2S in the gas stream, the wells in the field were logged using the EM Pipe Scanner tool and PS Platform multifinger imaging tool (PMIT) to assess corrosion.

In one interval the logs indicated substantial corrosion in a perforated zone (previous page, bottom). The 2D thickness image from the EM Pipe Scanner tool clearly showed metal loss, while the 2D discriminator log showed only the perforations and no metal loss. This combina-tion indicates that corrosive fluids are removing metal from the outside.

In another interval in the same well the EM Pipe Scanner average-thickness measurement revealed metal loss from the outer string of 95/8-in. casing (right). Neither the high-frequency 2D

> Evidence of metal loss in outer casing. The logged section has 41/2-in. tubing and 95/8-in. casing (well diagram, left). The EM computed thickness of the double string of pipe is significantly less than nominal above X40 m (Track 1), but there is no evidence of loss on the 2D discriminator log (Track 5), indicating the loss is not on the inside wall of the tubing. The EM computed thickness curve also shows metal loss from X83 to Y50 m, which also is not evident on the 2D discriminator log. In addition, the PMIT caliper log (not shown) indicated no metal loss from the inner surface of the 41/2-in. tubing. The log response is interpreted as loss of thickness in the outer wall of the 95/8-in. casing in these sections. In Track 4 the thickness change is represented as the proportional phase-angle change.

GammaRay

0 150gAPI 0 in. 1

EM Computed Thickness

–5 0dB

Double-CoilLong-Spacing

Amplitude

Double-CoilLong-Spacing

Phase

0 deg 360

2DThickness

deg

2DDiscriminator

Tension1,000 0lbf

3 V –1

CCL

–5 0dB

Double-CoilShort-Spacing

Amplitude

Double-CoilShort-Spacing

Phase

0 deg 360

–55 0 55 0.943 1.069

X00

X50

Y50

Y00

Depth,m

133/8 in.72 lbm/ft

12.347-in. ID0.514-in. thickness

95/8 in.53.5 lbm/ft8.535-in. ID

0.545-in. thickness

41/2 in.12.75 lbm/ft3.958-in. ID

0.271-in. thickness

Oilfield ReviewSpring 10PipeScanner Fig. 7ORSPRG10-PPSCN Fig. 7

30039schD7R1 42-50.indd 8 5/1/10 5:33 PM

50 Oilfield Review

discriminator log nor the caliper log indicated metal loss in this interval.

Results from this tool run clearly demon-strated that the tubing in one well was too thin to pull it safely, so the company is drilling a replace-ment well nearby.

Scale in PipesGDF SUEZ E&P Nederland B.V. operates the K12-B gas field located about 150 km [93 mi] northwest of Amsterdam in the Dutch sector of the North Sea. The field was discovered in 1982.10 About 13% of the produced gas is CO2. The separated CO2 from the K12-B platform is currently being rein-jected into the K12-B6 well, the first site in the world to return CO2 to the same reservoir from which it originated.11 The injection project is being studied by the Nederlandse Organisatie voor Toegepast Natuurwetenschappelijk Onderzoek (TNO, the Netherlands Organization for Applied Scientific Research) as part of several Dutch and European Union projects on CO2 injection.

The subject well was used for gas production from 1991 to 1999 and then was shut in for sev-eral years. The operator began CO2 injection in January 2005. Since the injection is a pilot for larger-scale CO2 injection, the important issue of well integrity may extend beyond the life of the nearly depleted field.

When in contact with water, CO2 can be cor-rosive to the 13% chromium steel used in this well’s tubing. Although the CO2 now being injected is dry, Well K12-B6 occasionally pro-duced water while it was a production well. GDF SUEZ performs annual pipe-integrity studies to monitor for potential problems. Multifinger caliper surveys by a third party began showing anomalous results, with the measured pipe ID increasing and then decreasing with repeat sur-veys.12 Coverage by the caliper fingers was only 25% to 30% of the 41/2-in. OD tubing. The operator opted to switch to the PMIT in combination with the EM Pipe Scanner service to obtain increased coverage of the inner surface.

The resulting survey indicated the tubing was still in good condition in terms of corrosion, but the log showed the presence of scale. Buildup of scale inside pipe affects corrosion-monitoring tools differently. Calipers will ride along scale, indicating an ID that is too small. The effect on an EM-based measurement depends on the com-position of the scale itself. In the case of noncon-ducting, nonmagnetic scale such as calcium carbonate, there is no effect unless the buildup is

thick enough that the distance from the sensors to the pipe wall affects the 2D resolution.13

In this case the PMIT and EM Pipe Scanner measurements diverged (left). Along this interval there was also a strong increase in the gamma ray signal, indicative of a buildup of scale that includes naturally occurring radioactive mate-rial. The operator plans to obtain scrapings from this interval to verify the indication of scale, and to rerun the combination of monitoring tools. Engineers want to ascertain the type of scale and determine if it developed in the past, when the well was on production, or if it is occurring during injection of the dry CO2, and if so, how. The result from the upcoming monitoring run will also determine whether the company should continue monitoring annually or switch to every other year.

Scanning for ProblemsCorrosion doesn’t stop eating away at metals until there is nothing left for it to consume. Regardless of how hard engineers attempt to hold it at bay, it is relentless and will exploit any opportunity. Corrosion monitoring provides assurance that mitigation efforts are succeeding or tracks the progress of corrosion when they are not.

The EM Pipe Scanner tool is the newest Schlumberger induction tool for monitoring cas-ing conditions. Its combination of measurements allows quantitative evaluation of pipe thickness in single strings of casing. The 2D imaging capa-bilities indicate the spread of corrosion or pit-ting, and whether this is occurring on the inside or the outside of the casing. In multiple strings the tool is qualitative, since the EM characteris-tics of the outer pipe cannot be evaluated in situ.

Adding the dimension of time through repeat surveys allows determination of the progression of corrosion. This gives an operator the informa-tion needed to decide between replacing or repairing tubulars, or continuing to operate a well when it is safe to do so.

Although advances in metallurgy, coatings and equipment designs are being made, the basic methods to control corrosion have not changed in many years. The battle to defeat corrosion continues to challenge engineers to their utmost, and monitoring using equipment such as the EM Pipe Scanner service is an impor-tant tool in their arsenal for assessing the integ-rity of infrastructure. —MAA

> Indication of scale in 41/2-in. pipe. Below about 2,033 m, the measurement of inner radius from the PMIT caliper tool (Track 2, black) agrees with the EM Pipe Scanner ID measurement (blue). Above that point the EM measurement continues to indicate the same ID, but the caliper tool indicates a smaller radius. The large increase in gamma ray response (Track 1) is interpreted as resulting from a buildup of scale containing naturally occurring radioactive material.

0 3,000gAPI

Gamma Ray

Casing NominalInner Radius

in.

Casing InnerDiameter (EM)

in.3.5

Average InternalRadius (Caliper)

in.

1.75 2.25

4.5

1.75 2.25

Depth,m

1,625

1,750

1,875

2,000

2,125

2,250

2,375

2,500

Oilfield ReviewSpring 10PipeScanner Fig. 9ORSPRG10-PPSCN Fig. 9

10. Vandeweijer VP, Van der Meer LGH, Hofstee C, D’Hoore D and Mulders F: “CO2 Storage and Enhanced Gas Recovery at K12-B,” paper R041, presented at the 71st EAGE Conference and Exhibition, Amsterdam, June 8–11, 2009.

11. van der Meer LGH, Kreft E, Geel CR, D’Hoore D and Hartman J: “CO2 Storage and Testing Enhanced Gas Recovery in the K12-B Reservoir,” presented at the 23rd World Gas Conference, Amsterdam, June 5–9, 2006.

Vandeweijer et al, reference 10.12. Vandeweijer et al, reference 10.13. For more on scale problems: Crabtree M, Eslinger D,

Fletcher P, Miller M, Johnson A and King G: “Fighting Scale—Removal and Prevention,” Oilfield Review 11, no. 3 (Autumn 1999): 30–45.

30039schD7R1 42-50.indd 9 5/1/10 5:33 PM