FOR OPEN PUBLICATION

HOIS-G-028 Issue 2

Contents

1 INTRODUCTION ............................................................................................ 1

2 SCOPE ........................................................................................................... 2

3 ABBREVIATIONS AND SYMBOLS ............................................................... 3

4 STANDARDS .................................................................................................4

5 SCABS IN THE OIL AND GAS INDUSTRY ................................................... 4

6 NDT METHODS CURRENTLY AVAILABLE FOR SCAB INSPECTION ....... 5

7 TRIALS OF NDT METHODS FOR SCAB INSPECTION ............................... 5

8 GUIDANCE ON SELECTION OF NDT METHODS FOR SCAB INSPECTION

........................................................................................................................ 5
8.1 Overview of methods .................................................................................. 5
8.2 Methods for quantitative sizing of remaining ligament under scabs ............. 6
8.3 Localised or fine pitting ............................................................................... 9
8.3.1 Presence and likely effects on sizing methods .......................... 9
8.3.2 HOIS trials .............................................................................. 10
8.3.3 Double wall radiography for detection of presence of localised
pits .......................................................................................... 11

9 CONCLUSIONS ........................................................................................... 12

10 ACKNOWLEDGEMENTS ............................................................................ 13

11 REFERENCES ............................................................................................. 14

Appendices

APPENDIX 1 EXTERNAL CORROSION CHARACTERISTICS ............................ 16

APPENDIX 2 NDT METHODS ............................................................................... 38

IV
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

1 Introduction
External corrosion of carbon steel pipework is a major problem for oil and gas facility owners
and has led to loss of containment. As external degradation occurs, corrosion product or scale
usually builds up on the affected components. Removal of corrosion product by blasting
pressurised live process components has an associated hazard of loss of containment.

The requirement considered in this project is determination of the remaining ligament under
the corrosion product or scab, without the potentially hazardous removal of this material. As it
is generally not possible to propagate ultrasound through corrosion product, the presence of
the scab renders the area of corrosion inaccessible and standard 0° pulse-echo methods are
not applicable for measurement of the remaining thickness in badly corroded areas.

An initial 3-year HOIS project C12-01, involving extensive trials of NDT techniques for
inspection of externally corroded un-insulated pipes with the corrosion product in-situ, was
completed early in 2015 [1]. Although project C12-01 involved a significant number of trials on
a variety of specimens, it was found that there was no universally applicable technique capable
of accurately sizing the remaining wall thickness of externally corroded areas on all pipe
diameters and wall thicknesses. The morphology of the external corrosion was also found to
affect the results obtained with many of the techniques.

Given the widespread occurrence of externally corroded pressurised components within the
oil and gas industry, at the March 2015 meeting the HOIS members supported a further project
(C15-02) on this challenging area of NDT, with the objective of performing further trials of
techniques that showed promise from the earlier trials and any other developmental techniques
that had not been included in the earlier project. The main objective of C15-02 was to assess
further the performance and capabilities of two methods that had shown promising results in
the previous trials [2].

In March 2017, a further ambitious and strategic HOIS project (C17-01) on the inspection of
external corrosion with and without insulation was approved. This project was supported by
The Oil and Gas Technology Centre (now the Net Zero Technology Centre) and was designed
to develop and focus the HOIS resources and capabilities for conducting rigorously controlled
independent evaluation trials and to compare different inspection methods for these
challenging problems.

In addition to the CUI trial programme, project C17-01 included manufacture of additional test
samples to provide further examples of simulated external corrosion scabs, to be followed by
additional trials on these test samples and some of the earlier components, particularly the ex-
service ones. Project C17-01 also included the in-house design and build of an internal UT
scanner for corrosion mapping deployed from the inside of a pipe. The probe, scanning, data
collection and processing were in full accordance with the HOIS recommended practice for
precision UT thickness measurements [3]. This allowed accurate measurements of remaining
thickness to be made, and hence allowed benchmarking of the ex-service samples with
external corrosion scabs to allow comparisons to be made with the results of NDT trials
conducted within the present project, and those performed in the earlier project C15-02.

The C17-01 trial report [4] gave details of the samples used for the trials with a description of
the benchmarking techniques used to derive accurate measurements of remaining wall
thickness/wall loss for the corroded areas. The trials comprised further M-skip scans of the
additional manufactured components, dual frequency eddy currents, pulsed eddy currents
(PECT from Maxwell NDT) and the GUL QSR1 system. An analysis of the trial results was
reported, as well as any issues encountered during the trials. A guidance document for the

1
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2
inspection of un-insulated external corrosion scabs was developed in this project, informed by
all the HOIS trials undertaken since the first trials in project C12-01.

The current HOIS project (C19-01) has sought to evaluate the effectiveness of selected NDT
techniques in determining the minimum remaining ligament under scabs where the corrosion
includes localised pitting [5]. Two pipes were manufactured for this study, each containing
CNC-machined areas of wall loss containing localised pitting but with differing types of
morphology. This project was also supported by The Oil and Gas Technology Centre (now the
Net Zero Technology Centre).

In addition to these recent trials, HOIS performed a series of preliminary tests to investigate
the feasibility of using double-wall radiography for use as a complementary screening tool for
determining the possible presence of isolated pits under external corrosion scabs [6]. The
applicability of 3-D laser scanning and photogrammetry to characterise the height/topography
of the scabs was also investigated as a means for assisting with the interpretation of the
radiographs.

The present document provides updated guidance on the inspection of uninsulated external
corrosion scabs, based on that developed in project C17-01 and the trial results from the four
HOIS projects on external corrosion outlined above (C12-01, C15-01, C17-01 and C19-01),
and other information obtained during these HOIS projects.

This document gives in Section 5 and Appendix 1, a summary of relevant information on scabs
within the oil and gas industry. The main NDT methods currently available are summarised in
Section 6 and Appendix 2. Where available, performance and reliability information obtained
from the HOIS scab trials (as summarised in Section 7) is included in Appendix 2. Section 8
provides guidance on application of the methods depending on the inspection requirements,
including summary tables of method applicability.

2 Scope
This guidance is intended for areas of external corrosion that have been located by other
means (e.g. visual inspection) and for which removal of the corrosion product or scale is
considered to be hazardous, due for example to the component being in operation (‘live-line’).

Hence the requirement for these areas of external corrosion is taken to be measurement of
the remaining ligament by methods that do not involve removal or disturbance of the corrosion
product or scale. For the purposes of these guidelines, it is however assumed that direct
access can be obtained to the outside of the corroded area, which is not made inaccessible
by the presence of obstructions such as clamps, pipe supports, saddles etc.

The component material is limited to carbon steel and the degradation type is limited to loss of
wall due to corrosion. Other forms of degradation such as cracking are beyond the scope of
the present document.

2
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

3 Abbreviations and symbols

Term/symbol Meaning
AWT Averaged Wall Thickness is used in some PEC units, where the
average of the remaining wall thickness is determined over the
sensor footprint.
CR Computed Radiography, which involves usage of a flexible imaging
plate (IP) which is first exposed to radiation which generates a
latent image and then scanned using a laser scanner to provide a
digital radiograph that is displayed using a computer.
CWT Compensated Wall Thickness has been developed by Eddyfi and
uses a proprietary algorithm for correcting the remaining wall
thickness values for the effects of sensor footprint averaging.
DDA Digital detector array. Used for digital radiography, a DDA is usually
a 2-D array detector that provides a real-time or near real-time
digital radiograph which is displayed on a computer monitor.
EMAT Electromagnetic acoustic transducer
GWT Guided Wave Testing
ID Pipe inner diameter
95% LUS 95% LUS is a statistical parameter. It is a 95% safety limit against
undersizing. When the value of this parameter is added to a
reported size from an NDT inspection, then the resulting size has
a 95% confidence of not being undersized (see Section 5.6.5.2 of
Burch, 2015b and references therein).
Multi-skip/M-skip A pitch-catch ultrasonic inspection method using vertically
polarised shear waves (SV) and involving multiple reflections or
skips off the component ID and OD.
OD Pipe outside diameter
PA Phased Array
QSR-1 Quantitative Short-Range guided wave inspection system
developed by Guided Ultrasonics Ltd (GUL).
R Correlation coefficient. Used here to assess correlation between
two sets of values. A value of 1 indicates perfect
correlation/agreement. A value of zero indicates no correlation.
RMS Root-mean-square
σ Standard deviation – in this document refers to the standard
deviation of the differences between the sizing values from an NDT
method and the corresponding benchmark values.
SV Vertically polarised shear waves
wmax Maximum steel equivalent penetrated thickness (for radiography)
WT Uncorroded wall thickness

3
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

4 Standards
ASME BPVC-V 2021 Edition, July 2021 Article 21. Pulsed Eddy Current (PEC) Technique For
Corrosion Screening

ISO 20769:2018 Non-destructive testing. Radiographic inspection of corrosion and deposits in

pipes by X- and gamma rays. Part 1: Tangential radiographic inspection Part 2: Double wall
inspection.

ISO 20669:2017 Non-destructive testing – Pulsed eddy current testing of ferromagnetic

metallic material components

5 Scabs in the oil and gas industry

Available information on the external appearance of scabs, and their characteristics (including
morphology) is summarised in Appendix 1. Included in this appendix is a compilation of
available values for the ratio between scab height and wall loss.

On the basis of available information, it is unclear how the ratio between scab height and wall
loss is influenced by the corrosion conditions, and hence it is not possible to predict what an
appropriate value would be for a particular corroded area.

The available scab height and wall loss data does however suggest that a conservative “rule
of thumb” value to use would be to assume that the wall loss, w, is equal to the scab height, h
(see Figure A1-5 for definition of these values), but only provided there is good evidence that
the corrosion product has not been lost (e.g. the coating is still mainly intact).

Information is also given on the morphologies of the wall loss for scabs on several examples
of ex-service pipes provided by HOIS members. Internal UT scans of scabs on two different
pipe sets, have shown significant differences between the morphologies of set 1 compared
with set 2.

It is particularly significant that examples of corroded areas in one of these pipe sets were
found to contain highly localised, and often isolated pits, superimposed on generalised wall
loss. These localised pits had significant additional wall losses of ~3-5mm compared with the
surrounding corrosion, over circumferential/axial extents of ~5-15mm.

In comparison, the scabs in the other pipe set did not contain these highly localised pits and
had irregularities in wall loss over significantly larger length scales (typically a few tens of
millimetres, although this varied from one scab to another).

The presence of highly localised pitting presents significant challenges for any NDT methods
that may be used to estimate the remaining wall thickness under these scabs, due to the
normally limited spatial resolution of the NDT methods. This can lead to significant sensor
footprint averaging effects, and/or failure to detect the fine pitting, both of which effects may
lead to significant underestimates of the wall loss/severity of the corrosion. As a result, the
latest HOIS trial programme focussed on the examination of scabs with localised pitting [5].

4
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

6 NDT methods currently available for scab inspection

Appendix 2 summarises current NDT methods that can, under appropriate conditions, provide
quantitative measurements of the remaining wall thickness under areas of external corrosion
covered with corrosion product.

7 Trials of NDT methods for scab inspection

Within the HOIS JIP there have been four sets of trials of NDT methods for scab inspection
[1][2][4][5]. The trials were generally conducted blind and were based on both ex-service and
manufactured components. Most trials were in the HOIS NICE facility, although in a few cases
the trials were conducted elsewhere. The following NDT methods were included:

1. Pulsed Eddy Current (Applus RTD Incotest, Shell PEC and Eddyfi Lyft, Maxwell
PECT)
2. Digital tangential radiography (HOIS multi-vendor trials)
3. Medium range guided waves (GUL, QSR1)
4. Dual frequency Eddy Currents (ETherNDE)
5. Multi-skip/M-skip® ultrasonic method (ESR)
6. Guided and SH-wave EMATs for screening (Baugh & Weedon & Sonomatic)
7. Back-scatter radiography (developmental) (Inversa Systems)
8. Pulsed X-ray radiography (Bilfinger)

Where available, for each NDT method a summary of the results from the trials are given in
Appendix 2.

8 Guidance on selection of NDT methods for scab

inspection

8.1 Overview of methods

Detailed information is given on NDT methods for scab inspection in Appendix 2.

The applicability of an NDT method for scab inspection depends on many factors including:

• Radiation hazards - can they be adequately managed?

• Does the method need to be applicable to pressure vessels?
• Does the inspection need to cover geometry changes (tees, supports, connections
etc.)?
• Are the pipe diameter and wall thickness within the applicability limits of the methods?
• The extent and morphology of the corroded area.
• The possible presence of highly localised fine pitting superimposed on more
generalised wall loss.

5
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

8.2 Methods for quantitative sizing of remaining ligament under scabs

Table 8-1 gives a summary of the information contained in Appendix 2 in a format that allows
the main factors listed above to be considered for all methods deemed to have potential for
quantitative sizing of the remaining ligaments under scabs. Note also the table footnotes.

® M-skip is a registered Trademark of ESR Technology Ltd

6
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2
Table 8-1: Summary of applicability of current methods for scab NDT

Complex Highly
Vessel Sizing accuracy
Method Pipe OD Pipe WT geometry localised Other constraints 2
applicability from HOIS trials 1
applicability pitting
Visual/Scab
Any Any High Y N Low Scab height to wall loss ratio varies widely.
height
Need to maintain a constant offset between
Dual frequency 3 the probe and the component surface.
≥ 1" ≥ 3mm Medium Y N Medium
Eddy currents Variations in magnetic properties of scab
material.
Scab extent < 18WT; adequate surface
Multi-skip/M-
≥4" ~ 15 – 30mm Low Y N High condition around scab needed for UT
skip®
probes.
Magnetic properties of scab material,
4 5 6 morphology of corrosion. Averages over
PEC ≥ 2" 3 – 102mm Medium Y N Medium
sensor footprint area, so can undersize
localised areas of corrosion.
At least 120° of good material needed at
same axial location as scab, to allow scanner
QSR1 6” - 24” 6-13mm (nominal) Low N N High
wheels and probes to be in contact with the
pipe surface.
For Ir192, max penetrated steel
Radiography Ionising radiation hazard, morphology of the
thickness <80mm – see also High N N6 Medium 7
(tangential) corrosion.
Figure 8-1.
Ionising radiation hazard. Liquid in pipe
45mm for Ir 192,
reduces contrast of radiograph, and
Radiography without liquid in
Any High N Y7 N/A 8 therefore the sensitivity to localised pitting.
(double wall) pipe (improved
Scab spallation would complicate
technique).
interpretation of radiograph.

Footnotes
1
Excluding scabs with highly localised pitting
2
In addition to the factors listed, all methods are affected by corrosion morphology, by varying amounts, especially the presence of highly localised pitting, where present
3
Accuracy reduced for complex scab morphology, due to large sensor size
4
Minimum inspectable pipe OD may differ between manufacturers. Refer to manufacturer’s literature for more information.
5
Maximum wall thickness is dependent upon probe size and lift off. Refer to manufacturer’s literature for more information.
6
On one set of scabs, sensor footprint averaging effects led to ~20% undersizing for AWT values; improved accuracy for CWT (Lyft) and DWT (Incotest) measurements
7
Tangential radiography becomes unreliable for localised pitting (see ISO 20769:1 and Ref. [1]).
8
Double radiography is a method with potential for screening for the possible presence of localised corrosion pits, which if present, are likely to reduce the accuracy of the sizing methods given in this table. Double Wall
radiography does not in itself give any through-wall sizing information.

7
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2
Based on the information in the table above, Figure 8-1 shows the applicable pipe OD and WT
values for all the NDT methods except PEC and dual frequency EC. Given that PEC and dual
frequency EC are applicable for inspection of pipes with similar outside diameters and wall
thicknesses, a graphical representation was not produced.

The white areas in this figure represent pipes for which none of the NDT methods are currently
applicable, whereas the coloured or shaded regions represent where the different methods
can be applied, which may overlap in some cases.

Examination of Figure 8-1 shows that one or more NDT methods are applicable to the great
majority of combinations of pipe OD and WT. Note also that those methods shown as being
applicable to pipe diameters up to the maximum of 800mm given in the figure or in Table 8-1
are also applicable to larger diameter components, including vessels with higher ODs.

The three methods shown in Figure 8-1 (QSR1, Tangential RT and Multi-skip) are mainly
complementary since their ranges of applicability only overlap slightly. Tangential RT in
accordance with ISO 20769-1 should be used for small-bore pipes and some with larger
diameters but thin walls (maximum penetrated thickness at the tangent position of ~80mm if
using an Ir 192 source). QSR1 covers larger diameter thinner walled pipes up to 24", while
Multi-skip is applicable to the thicker-walled pipes (those with wall thicknesses greater than
about 15mm), without any limit on OD (applicable to vessels as well as large diameter pipes).
The PEC method covers most pipes with wall thickness up to 16mm (larger wall thicknesses
may also be examined but only by means of larger sensors that lead to increased sensor
footprint averaging effects). The Dual frequency EC method can, in principle, be applied to all
components with diameters of at least 1”, although the smallest diameter it has been trialled
on is 6”.

In considering the most effective method for quantitative sizing of scabs to use for a particular
scab, the information given in Table 8-1 and Appendix 2 should be considered carefully in
addition to the applicability data for OD and wall thickness given in Figure 8-1.

8
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

Figure 8-1: Approximate representation of the applicability of different NDT methods for scab sizing, excluding dual frequency EC &
PEC, for different pipe diameters and wall thicknesses. The black areas represent pipes that are thinner walled or thicker walled than
those normally manufactured (sch 10 used as lower limit and sch XXS or sch 160 as upper limit).

9
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

8.3 Localised or fine pitting

8.3.1 Presence and likely effects on sizing methods

One set of ex-service pipes included in the HOIS trial programmes showed the unexpected presence of
highly localised groups or clusters of multiple pits, superimposed on more general areas of wall loss (see
Appendix A1.3). The individual pits within these groups were detected by internal UT scanning and were
found to have average axial/circumferential extents of only ~5mm, with depths of up to 2-3mm [4].

The presence of these localised pits was not visually apparent from an external examination of the scab,
and their frequency of occurrence under external corrosion scabs within the energy industry is unknown,
as are the corrosion conditions leading to their formation.

The potential that these localised pits are present may increase the uncertainty in the values for remaining
ligament derived by current NDT methods since the spatial resolution of these methods is generally
substantially lower than the dimensions of the individual pits in the clusters. This limited resolution,
combined with sensor footprint averaging effects, is likely to make the localised pits difficult to detect and
may cause increased undersizing of the wall loss.

8.3.2 HOIS trials

The current HOIS project (C19-01) sought to evaluate the effectiveness of selected NDT techniques in
determining the minimum remaining ligament of scabs containing localised pitting [5]. Two pipes were
manufactured for this study, each containing CNC-machined areas of wall loss containing localised pitting
but with differing types of morphology, see Figure 8-2.

Both pipes had wall loss with morphologies based upon corrosion found in ex-service pipes. For one pipe,
the areas of wall loss were characterised by multiple localised pits, superimposed on more general areas
of wall loss. These were direct replicas of examples of corrosion scabs found in ex-service pipes, although
due to limitations of the modelling and CNC processes, it was not possible to replicate some of the finest
detail found in these pipes.

For the other pipe, the underlying morphology was different with larger scale variations in wall loss, based
on a different set of ex-service pipes. Single localised pits were then added to these CNC areas using
mechanical methods.

For both pipes, to simulate the presence of natural corrosion product, the areas of wall loss were covered
with synthetic scab material that consisted of magnetite granules (which is ferrimagnetic) embedded in
modelling clay.

10
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

Figure 8-2: Examples of morphologies of the CNC areas of wall loss introduced into the test
pipes. Left – multiple localised pits superimposed on a larger scale area of wall loss. Right – area
with larger scale wall loss, with a single localised pit introduced.

Of the different methods trialled, QSR1 from GUL showed the smallest overall differences from the
benchmark values for both pipe samples. M-Skip exhibited similar results to QSR1 for the pipe with the
multiple localised pit clusters, but larger differences on the sample containing the additional single localised
pits. This may have been caused either by the lack of detection of the additional pits, or the lack of both
axial and circumferential measurements for three of the four scabs. Of the electromagnetic methods, the
PECA-HR (with CWT) exhibited smaller differences from the benchmark when compared with the Maxwell
PECT (with AWT) values. The dual frequency EC method was generally comparable to the PECA-HR
results.

Generally, it is likely that most, if not all, of the methods trialled failed to detect the additional single localised
pits, which resulted in the observed undersizing. However for the pipe containing areas with multiple
localised pits, the mean undersizing with QSR1 and M-skip was reduced. In addition, the mean undersizing
for the eddy current methods was reduced for the pipe with the multiple localised pits, an effect which may
be attributed to reduced sensor footprint averaging effects, compared with the single localised pits.
8.3.3 Double wall radiography for detection of presence of localised pits

In this trial programme, double-wall radiography was also investigated as a method with potential for use
as a complementary screening tool for determining the possible presence of localised pits under external
corrosion scabs.

The basis of this method is that the isolated fine pitting will cause less radiation attenuation than the
surrounding areas and that this will then be visible as areas (indications) on the corresponding radiographs.
Knowledge of the presence or absence of isolated pitting from double wall radiography could then
potentially be used to assess the likely uncertainty in the results from other NDT methods that are known
to undersize the wall loss when isolated pits are present. If double wall radiography shows no evidence of
isolated localised pitting, this would give increased confidence in the reliability of the NDT methods for
determination of the remaining ligament. Conversely, if evidence of localised pitting is obtained on the
radiographs, this would increase the likelihood that the results from the NDT method significantly undersize
the wall loss present.

In recent trials [6], based on a small sample size, it was demonstrated that the use of double wall
radiography has potential for this purpose, but more work is required to determine whether it would be
effective in the field. The applicability of 3-D laser scanning and photogrammetry to characterise scab

11
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2
height/topography was also investigated and both were found to be a useful tool when interpreting
radiographs from an external corrosion scab.

12
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

9 Conclusions
This guidance document aims to provide operators and NDT service providers with current relevant
information on the available NDT methods for in-service determination of the minimum remaining ligament
for external corrosion scabs on live lines where it is hazardous to remove the scale/corrosion product to
obtain direct access to the underlying steel surface. This should lead to a more consistent approach
across the industry.

The available scab height and wall loss data shows wide variations in the values for service-induced
corrosion, without any reliable information on the most reliable value to use. The available values suggest,
however, that a mostly conservative “rule of thumb” for estimates would be to assume that the wall loss is
equal to the scab height (see Figure A1-5 for definition of these values), i.e. considering a 1:1 ratio. It
should be recognised that the data set on which this is based is limited and is thought to be primarily based
on results from offshore, North Sea operations. Hence the 1:1 ratio can not be taken as universally
conservative and should only be considered, as appropriate, as a basis for estimates where the resulting
wall loss is not approaching an integrity limit and there is also good evidence that corrosion product has
not been lost from the scab.

Accurate measurement of the remaining wall thickness under scabs is a challenging requirement for NDT
methods, as the presence of the scab material makes the outer steel surface inaccessible. In addition, the
scab itself can take many different shapes, with varying circumferential and axial extents and profiles,
depending on the time over which it has developed and the corrosion mechanisms.

Factors that influence the applicability of NDT methods for scab minimum ligament determination include
the following:

• Pipe diameter and wall thickness

• Component geometry
• Extent of the corroded area in the axial and circumferential directions
• Accessibility to the corroded area and space around it
• Surface condition surrounding the corroded area
• Morphology of the corrosion, especially the presence of highly localised pitting

The methods included in the HOIS trials that gave the highest sizing accuracies (of around ±10% of the
uncorroded wall thickness), under appropriate conditions, in test components lacking highly localised
corrosion, were:

• Tangential Computed Radiography (for thin-walled, small bore pipes only, but note limited
circumferential coverage and issues stated in ISO 20769 part 1).
• QSR1 for pipes with wall thicknesses in the range 6-13mm, and diameters in the range 6” – 24”.
• Multi-skip/Multi-skip for thick-walled pipes only.
• Eddyfi Lyft PEC with CWT.

The ranges of applicability of these methods in terms of pipe wall thickness and diameter are largely
complementary, so that at least one of these methods is available for the majority of pipes used in the oil
and gas industry. The eddy current based methods trialled gave somewhat lower sizing accuracy, with a
tendency to undersize due to sensor footprint averaging effects. However, on one set of pipes with realistic
but simulated areas of wall loss, the CWT values provided by the Eddyfi Lyft pulsed eddy current
equipment provided a comparable level of accuracy.

Recent internal UT benchmarking information from one set of ex-service pipes has shown the unexpected
presence of some highly localised pitting, superimposed on more general areas of wall loss. These pits,
which had average axial/circumferential extents of only ~5mm presented substantial challenges to all the

13
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2
currently available NDT methods. This was caused by the limited spatial resolution of these methods,
which makes the localised pits difficult to detect and affects the ability to measure the actual minimum
ligament thickness. The result can be a tendency to undersize the remaining ligament thickness.

The presence of these localised pits was not visually apparent from an external examination of the scab,
and their frequency of occurrence in pipework within the energy industry is unknown, as are the corrosion
conditions leading to their formation. The possibility that these localised pits might be present increases
the uncertainty in the values for remaining ligament derived by current NDT methods.

In recent trials [6], based on a small sample size, it was demonstrated that double-wall radiography is a
method with potential for use as a complementary screening tool for determining the possible presence of
localised pits under external corrosion scabs. If double wall radiography shows no evidence for isolated
localised pitting, this would give increased confidence in the reliability of the NDT methods for
determination of the remaining ligament. Conversely, if evidence of localised pitting is obtained on the
radiographs, this would increase the likelihood that the results from the NDT method significantly undersize
the wall loss present.

The applicability of 3-D laser scanning and photogrammetry to characterise scab height/topography was
also investigated and both were found to be a useful tool when interpreting radiographs from an external
corrosion scab, where irregularities in the external scab profile, such as selective spallation of areas of the
scab, can complicate radiograph interpretation.

10 Acknowledgements
HOIS members are thanked for their contributions and the NZTC (formerly the OGTC) are thanked for co-
funding the most recent projects.

Those companies that participated in the various HOIS external corrosion (scab) trials that took place
within four separate projects since 2012 are thanked for their contributions. These included Baugh &
Weedon, Bilfinger, Eddyfi, ETherNDE, Inversa Systems, Guided Ultrasonics Ltd, Maxwell NDT,
Oceaneering and Sonomatic. The contributions made by the participants in HOIS digital radiography trials
are also acknowledged, especially the equipment vendors (Duerr NDT, FujiFilm and GE) and ApplusRTD
who hosted some of these trials and providing all the necessary facilities.

The project champions Billy Mackay (Total) and Dennis Zaal (Shell) provided useful guidance and support
throughout the duration of the most recent project. The contributions from the project champions from the
previous HOIS projects are also recognised.

14
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

11 References
1. Burch S F, Collett N J, Kitchener A and Peramatzis H R. Final report on the HOIS external corrosion
NDT trials project (C12-01). HOIS(14)R2 Issue 1, January 2015

2. Burch S F, Collett N J and Crook L. Results of further external corrosion NDT trials HOIS-R-001 Issue
1, August 2017.

3. Burch S F. HOIS recommended practice for precision thickness measurements for corrosion
monitoring HOIS-RP-003 Issue 1 November 2017.

4. Burch S F and N J Collett. Results of the HOIS/OGTC external corrosion NDT trials, HOIS-R-021
Issue 1, October 2019

5. Jones M, Burch S F. Determination of remaining ligament under external corrosion scabs and
influence of scab magnetic properties in EC techniques (C19-01), HOIS-R-044 Issue 1, October 2021.

6. Jones M, Burch S F. Double Wall Radiography of External Corrosion Scabs (C19-01), HOIS-N-047
Issue 1, October 2021.

7. Burch S F, Collett N J, Terpstra S, M V Hoekstra. M-skip: a quantitative technique for the measurement
of wall loss in inaccessible components, Insight Vol. 49 No. 4, April 2007.

8. Burch S F (2015). Recommended Practice for the in-service inspection of wall loss in pipes by digital
radiography HOIS RP 1 Issue 3.1, April 2015.

9. Burch S F (2014b). HOIS Safety Notice: Application of radiography to measure the remaining wall
thickness of external corrosion, available from https://www.esrtechnology.com/index.php/centres-of-
excellence/hois/safety-notice. A shorter version was also published in March 2015 on the Step
Change in Safety website at https://www.stepchangeinsafety.net/safer-conversations/safety-
alerts/unreliable-measurements-remaining-wall-thickness-pipes-some-forms

15
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

Appendices

APPENDIX 1 EXTERNAL CORROSION CHARACTERISTICS ........................... 16

APPENDIX 2 NDT METHODS .............................................................................. 38

16
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

Appendix 1 External corrosion characteristics

CONTENTS
A1.1 Visual inspection and external corrosion appearance 17

A1.2 Scab height to wall loss ratios 20

A1.3 Examples of external corrosion morphologies 26

A1.4 Photogrammetry and Laser Scanning 35

17
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

A1.1 Visual inspection and external corrosion appearance

A1.1.1 Visual Inspection

A detailed or close visual inspection (CVI) is generally the method used to locate areas of external
corrosion, or scabs.

It is recommended that detailed photographic records are kept of the visual examination results. The
availability of good photographic records is important to assist in the interpretation of any subsequent NDT
results obtained. Some assessments can be made from a visual examination of the severity of the
corrosion and its general characteristics.

A1.1.2 Typical corrosion stages

In the early stages of corrosion, a blister will form underneath the coating consisting of the corrosion
product which forms over the corroded substrate. Examples of blisters with intact coatings are given in
Figure A1-1.

Figure A1-1: Examples of early stage external corrosion which has formed blisters underneath
the coating. Note the cracks in the coating in the upper left and lower examples.

18
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2
As the corrosion progresses, and the corrosion product height increases, cracks can start to appear in the
coating, as shown in the upper left and lower blister on Figure A1-1.

More extensive cracking in the coating and partial loss of the coating may then occur, as illustrated in
Figure A1-2.

Figure A1-2: Example showing more extensive cracks in the coating and partial loss of the coating

At some point, the height of the external corrosion scab is sufficient to displace the coating which may then
be lost, revealing the corrosion product itself, as illustrated in Figure A1-3. Once the corrosion product has
been exposed it may then start to be lost (or spalled); it cannot be guaranteed that all the product remains
in-situ.

Figure A1-3: Example of external corrosion showing loss of coating

The morphology of external corrosion can be highly variable, with some components having a relatively
high density of separate and distinct corroded areas. Depending on the density and stage of the corrosion,
these isolated scabs can start to merge.

19
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

Hence, as illustrated in Figure A1-4, in advanced cases of corrosion, the area affected can be generalised,
and extended over a considerable area.

Figure A1-4: Example of widely distributed external corrosion with partial loss of coating

20
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

A1.2 Scab height to wall loss ratios

A1.2.1 Definitions and available datasets

One widely considered method for estimating the extent of the wall loss underneath a corrosion scab is to
first measure the height, h, of the corrosion scab above the surrounding uncorroded surface of the
component. To estimate the extent of the wall loss, w, the height, h, is divided by a factor, F, i.e.

W ~ h/F (A1.1)

Note that in this definition, both the wall loss, w, and the scab height, h, are measured relative to the
uncorroded surface (see Figure A1-5). Hence the maximum thickness of scab is given by the sum of h
and w.

Figure A1-5: Diagram showing definition of scab height (h) and wall loss (w) values

Some data sets detailing the scab height to wall loss ratio for service-induced areas of corrosion were
provided by current and previous HOIS members. Operator 1 originally provided measured scab heights
and wall loss for two different categories of corrosion, namely ‘pitting’ and ‘general wall loss’. There was,
however, limited information available with regards to the accuracy of these measurements, what
components they originated from, the service history of the components and whether or not the scabs
were original or had partially (or fully) spalled at any point during service life. Operator 2 supplied a much
smaller data set containing five measurements from an oil export pipeline.

The reliability of the values in these historical datasets is unknown. For example, a few of the Operator 1
“general wall loss” values for F exceeded 20, with one being implausibly high at 80.

For the remaining three datasets (Operator 2 10” sch 20, Operator 1 6” sch 40 and Operator 1 6” sch 160),
the benchmarking was performed by ESR Technology which resulted in greater confidence in the reliability
of the measured scab height and wall loss values. To do this, ESR Technology accurately measured the
heights of these natural corrosion scabs. These values were combined with careful measurements of wall
loss, obtained either by precision internal UT scanning, or direct external measurements of wall loss,
following removal of the corrosion product by blasting. These benchmarking methods then gave reliable
measurements for the factor, F, as defined in equation (A1.1) above.

One significant complication with this approach can arise if some of the corrosion product has been lost
from the corroded area, which can easily occur once the coating has been lost (see for example Figure
A1-3). This was not however considered to be a major issue for the scabs benchmarked by ESR
Technology.

21
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

A1.2.2 Values from available datasets

A compilation of all currently available data sets for scab height and wall loss is shown in Figure A1-6.

Table A1.1 provides the maximum and minimum scab height to wall loss ratios for the data sets, in addition
to the best fit scab height to wall loss ratios, F. The level of scatter is indicated by the values given for the
correlation coefficients, R2, and the root-mean-square (RMS) differences from the best fit straight lines
(constrained to pass through the origin).

The correlation coefficient is one for perfect correlation and zero for no correlation. Hence an R2 coefficient
near to one would be indicative of a reliable method for estimating wall loss from a straightforward
measurement of scab height. The available data shows substantial scatter in most, but not all, cases.

22
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

Figure A1-6: Graph showing scab height and wall loss data from all available sources

23
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2

Table A1.1: Relationship between defect wall loss and scale height for available datasets

Max. scab Min. scab Best fit scab RMS

R2
height to height to height to wall difference
Data Set correlation
wall loss wall loss loss ratio from best fit
coefficient
ratio (h/w) ratio (h/w) (h/w) (mm)
Operator 1: “pitting” 1 8 0.8 2 2.0 0.58

Operator 1: “general
80 1 5.3 0.9 0.25
wall loss” 1
Operator 2: oil export
4 1.1 2.6 1.1 N/A
pipe samples
Operator 2: 10” sch 20
2.6 1.5 2.1 0.24 0.97
Cooling Medium 30°C.
Operator 1: 6” sch 40
pipe samples (only 2 1.8 1.2 1.6 N/A
points)
Operator 1: 6” sch 160
thick walled pipe 2.8 0.8 1.3 2.3 N/A
samples
Operator 3 (only 1
1.8 1.8 1.8 N/A N/A
point)

Footnote:
1
Historical data with limited information on origin and reliability

24
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2
The dataset showing the least scatter was that of 10” sch 20 pipes obtained and benchmarked
by HOIS from Operator 2. The pipework formed part of the Cooling Medium system which had
a low operating pressure of at most approx. 10bar, and operating at about 30°C. It was on the
extremities of an offshore platform and the top section where the scabs formed were not visible
from walkways. The scabs formed apparently from drips of water from a structural beam above
it. For these scabs, the best fit scab height to wall loss factor F was 2.1 with a small (0.24mm)
root-mean-square difference from the best fit. Hence for these scabs, measurement of the
scab height, followed by division by 2.1, gave an accurate estimate of the wall loss.

It was found in the recent HOIS C19-01 project that of the HOIS members that estimated the
wall loss by measuring the scab height, the ratios used ranged from 2.5:1 to 10:1. There are
other anecdotal accounts of factors that are found to work well for other components
experiencing common corrosion conditions, but additional well documented, published data on
these factors has not been found. Appeals to HOIS members and to non-HOIS members (via
LinkedIn) resulted in additional data from one HOIS member.

However, based on general feedback from industry, it appears that a predictable relationship
between scab height and wall loss is the exception rather than the rule, as evidenced by the
much larger scatter shown by the other datasets included in Table A1.1 and Figure A1-6. For
example, the dataset for the Operator 1 6” sch 160 pipes showed factors, F, that ranged from
0.8 to 2.8 (ESR benchmarking).

The minimum scab height to wall loss ratio within the available datasets was 0.8 (within the
Operator 1 pitting data and Operator 1 6” sch 160 thick walled pipe samples).

A1.2.3 Summary
There is widespread interest in the potential applicability of measurements of scab height as a
basis for estimation of wall loss, as this is a relatively straightforward, fast, and low-cost
method. If reliable estimates could be obtained in this way, there would be no need to consider
further application of the NDT methods described elsewhere in this document (Appendix 2).

However, it is also widely perceived that the method is inherently unreliable, and it is unclear
what factor to use when converting from scab height to wall loss.

An additional major complication is that some of the corrosion product may have been lost if
the original coating is no longer present.

Under certain conditions, it appears that corrosion with a consistent scab height to wall loss
ratio can be obtained, over a wide range of corrosion severities (e.g. the factor, F, of 2.1 found
for the 10” sch 20 pipes provided to HOIS by Operator 2) – see Figure A1-6 and Table A1.1.

On the basis of available information, it is however unclear how the factor F is influenced by
the corrosion conditions, and hence it is not possible to predict what an appropriate value
would be for a particular corroded area.

The available scab height and wall loss data does however suggest that a mostly conservative
“rule of thumb” value to use would be F =1, i.e. the wall loss, w, is equal to the scab height, h
(see Figure A1-5 for definition of these values), i.e. considering a 1:1 ratio. It should be
recognised that the data set on which this is based is limited and is thought to be primarily
based on results from offshore, North Sea operations. Hence the 1:1 ratio can not be taken as
universally conservative and should only be considered, as appropriate, as a basis for
estimates where the resulting wall loss is not approaching an integrity limit and there is also

25
FOR OPEN PUBLICATION
HOIS-G-028 Issue 2
good evidence that corrosion product has not been lost from the scab (i.e. the corroded area
is still effectively a ‘blister’ under the coating – see the examples shown in Figure A1-1).

26