RP G103 2007

RECOMMENDED PRACTICE
DNV-RP-G103
NON-INTRUSIVE INSPECTION
OCTOBER 2007
DET NORSKE VERITAS

FOREWORD
DET NORSKE VERITAS (DNV) is an autonomous and independent foundation with the objectives of safeguarding life, prop-
erty and the environment, at sea and onshore. DNV undertakes classification, certification, and other verification and consultancy
services relating to quality of ships, offshore units and installations, and onshore industries worldwide, and carries out research
in relation to these functions.
DNV Offshore Codes consist of a three level hierarchy of documents:
— Offshore Service Specifications. Provide principles and procedures of DNV classification, certification, verification and con-
sultancy services.
— Offshore Standards. Provide technical provisions and acceptance criteria for general use by the offshore industry as well as
the technical basis for DNV offshore services.
— Recommended Practices. Provide proven technology and sound engineering practice as well as guidance for the higher level
Offshore Service Specifications and Offshore Standards.
DNV Offshore Codes are offered within the following areas:
A) Qualification, Quality and Safety Methodology
B) Materials Technology
C) Structures
D) Systems
E) Special Facilities
F) Pipelines and Risers
G) Asset Operation
H) Marine Operations
J) Wind Turbines
Amendments and Corrections

This document is valid until superseded by a new revision. Minor amendments and corrections will be published in a separate
document normally updated twice per year (April and October).
For a complete listing of the changes, see the “Amendments and Corrections” document located at:
http://webshop.dnv.com/global/, under category “Offshore Codes”.
The electronic web-versions of the DNV Offshore Codes will be regularly updated to include these amendments and corrections.
Comments may be sent by e-mail to [email protected]

For subscription orders or information about subscription terms, please use [email protected]
Comprehensive information about DNV services, research and publications can be found at http://www.dnv.com, or can be obtained from DNV, Veritas-
veien 1, NO-1322 Høvik, Norway; Tel +47 67 57 99 00, Fax +47 67 57 99 11.
© Det Norske Veritas. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, including pho-
tocopying and recording, without the prior written consent of Det Norske Veritas.
Computer Typesetting (FM+SGML) by Det Norske Veritas.

Printed in Norway
If any person suffers loss or damage which is proved to have been caused by any negligent act or omission of Det Norske Veritas, then Det Norske Veritas shall pay compensation to such person
for his proved direct loss or damage. However, the compensation shall not exceed an amount equal to ten times the fee charged for the service in question, provided that the maximum compen-
sation shall never exceed USD 2 million.
In this provision "Det Norske Veritas" shall mean the Foundation Det Norske Veritas as well as all its subsidiaries, directors, officers, employees, agents and any other acting on behalf of Det
Norske Veritas.
Recommended Practice DNV-RP-G103, October 2007
Introduction – Page 3
Executive summary
Pressure vessels and pressure systems are required to undergo fittings and connections associated with them, although the
periodic, statutory inspection to ensure continued safe and reli- document is equally applicable to forged/spun metal pressure
able operation. Traditionally this has been achieved by means vessels, which contain no welds, as well as storage tanks and
of an internal visual inspection (IVI), however, there can be large diameter pipelines. Although not intended to be covered
very high costs associated with shutting down a vessel (loss of by this document, some aspects of this recommended practice
production), isolating it and preparing it for entry. Indeed, may also be applicable to vessels manufactured from other
these costs can be much higher than the cost of the inspection materials and plant items other than pressurised equipment, in
itself. Furthermore, the mechanical disturbances involved in which case some of the general principles developed may well
preparing the vessel for internal inspection and reinstating it still apply.
may adversely affect its future performance. Finally, and by
no means least, man access may also be hazardous. There can, The recommended practice provides guidance for:
therefore, be significant advantages if inspections are per- i) determining when NII is appropriate in principle
formed from the outside of the vessel without breaking con-
tainment i.e. non-invasively. However, there needs to be a ii) information requirements to plan for NII
balance between achieving these benefits and obtaining the iii) defining the requirements for the NII method(s) to be
information required to ensure continued safe and reliable selected
operation.
While it may often be the preferred option, non-intrusive iv) selecting methods that meet the requirements
inspection (NII) represents a relatively new approach by com- v) evaluating the results of the inspection
parison to IVI and many engineers responsible for inspection
planning have yet to build up experience with and confidence vi) documentation requirements.
in its application. In addition, there are a wide variety of tech-
niques available, each with its own specific capabilities and
limitations. Acknowledgements
This has been recognised by industry and has led to the devel- This recommended practice is based on a number of preceding
opment of a number of guidance documents aimed at assisting documents and as such the permission of the previous authors
plant operators to plan and justify NII. This recommended to incorporate their work is hereby gratefully acknowledged.
practice is intended to bring these documents together under a In particular, we would like to thank members of the HOIS col-
single cover, and to provide a consistent and logical approach laborative project for their contributions and comments during
at all stages of the non-intrusive inspection process. the production of the recommended practice, Mitsui Babcock
This recommended practice is primarily intended for those for their permission to incorporate elements of the GSP235
with responsibilities in the planning, implementation and report covering aspects of the inspection planning procedure
acceptance of vessel inspections. It is aimed at the inspection and summary of inspection methods, and ESR Technology Ltd
of welded vessels constructed from metals, and related items, for compiling the original text for this recommended practice.
DET NORSKE VERITAS

Page 4 – Introduction
DET NORSKE VERITAS

Contents – Page 5
CONTENTS
1. INTRODUCTION .................................................. 7 5. ON-SITE INSPECTION ACTIVITIES............. 44

1.1 Background ..............................................................7 5.1 Preparation for Inspection ................................... 44
5.1.1 Individual Responsibilities ............................................... 44
1.2 Objectives of Non-Intrusive Inspection ................7
5.1.2 Preparation Tasks ............................................................. 44
1.3 Scope .........................................................................7 5.2 Performing the Inspection ................................... 45
1.4 Overview of the Recommended Practice...............8 5.3 Dealing with Non-conformances.......................... 45
1.5 Definitions ..............................................................10 5.4 Reporting of Results ............................................. 45
2. INTEGRITY REVIEW ....................................... 10 5.5 Demobilisation ....................................................... 46
2.1 General Approach .................................................10 6. EVALUATION OF INSPECTION .................... 46
2.2 Equipment Profile .................................................10 6.1 Introduction........................................................... 46
2.3 Risk Based Inspection Approaches......................11 6.2 Items to be checked for conformance.................. 46
2.4 Corrosion Risk Assessment ..................................11 6.3 Inspection Method................................................. 46
2.4.1 Corrosion Risk Assessment Types ................................... 11 6.3.1 Type A inspection............................................................. 47
2.5 Structural Integrity Assessment..........................12 6.3.2 Inspection Types B and C................................................. 47
2.6 Operational Experience ........................................12 6.4 Procedure............................................................... 49
6.4.1 Type A inspection............................................................. 49
3. THE DECISION GUIDANCE PROCESS......... 12 6.4.2 Type B inspection............................................................. 49
6.4.3 Type C inspection............................................................. 49
3.1 Introduction ..........................................................12
6.5 Data Quality........................................................... 49
3.2 Screening ................................................................12 6.5.1 Type A inspection............................................................. 49
3.3 High-level decision process ..................................13 6.5.2 Type B inspection............................................................. 49
3.3.1 Confidence in ability to predict types and locations of 6.5.3 Type C inspection............................................................. 49
degradation ....................................................................... 14 6.6 Location.................................................................. 49
3.3.2 Previous inspection effectiveness ..................................... 15
6.6.1 Type A inspection............................................................. 49
3.3.3 Severity and rate of degradation ....................................... 16
6.6.2 Type B inspection............................................................. 49
3.3.4 NII recommendation......................................................... 16
3.4 NII Decision Record ..............................................16 6.7 Coverage ................................................................ 49
3.5 Examples ................................................................16 6.7.1 Type A inspection............................................................. 49
3.5.1 Heat Exchanger Vessel .................................................... 16 6.7.2 Type B inspection............................................................. 50
3.5.2 Gas Receiver Vessel ........................................................ 16 6.7.3 Type C inspection............................................................. 50
3.5.3 Separator Vessel ............................................................... 17 6.8 Critical Non-conformance.................................... 50
3.5.4 Absorber Vessel ............................................................... 17
6.9 Reportable Indications and flaws ........................ 50
4. INSPECTION PLANNING................................. 18 6.9.1 Type A inspection............................................................. 50
4.1 Introduction ...........................................................18 6.9.2 Type B inspection............................................................. 50
4.2 The Planning Team ...............................................20 6.10 Examples ................................................................ 50
4.3 Inspection Strategy Type .....................................20 6.10.1 Type A inspection............................................................. 50
4.3.1 Type A Inspection............................................................. 21 6.10.2 Type B inspection............................................................. 51
4.3.2 Type B Inspection............................................................. 21 6.10.3 Type C inspection............................................................. 51
4.3.3 Type C Inspection............................................................. 21
4.3.4 Selection of Inspection Type ............................................ 21 7. INSPECTION INTERVAL................................. 51
4.4 Definition of Vessel Zones ....................................23 7.1 Discussion............................................................... 51
4.4.1 Identification of Zones ..................................................... 23 7.2 Comparison with IVI ............................................ 51
4.4.2 Example ............................................................................ 24
7.3 Detailed Assessment.............................................. 54
4.5 Definition of Degradation Type............................24
7.4 Examples ................................................................ 55
4.6 Inspection Effectiveness ........................................25 7.4.1 Glycol Contactor............................................................... 55
4.7 Required Inspection Effectiveness .......................26 7.4.2 Glycol Flash Drum ........................................................... 56
7.4.3 Vent Knock Out Drum ..................................................... 57
4.8 Coverage.................................................................27 7.4.4 HP Condensate Separator ................................................. 58
4.9 Selection of Inspection Method ............................27
4.9.1 Inspection Capability ........................................................ 27 8. REFERENCES..................................................... 60
4.9.2 Inspection Method Selection Flowcharts.......................... 27
4.9.3 Inspection Method Selection Criteria ............................... 28
APP. A REVIEW OF NON INTRUSIVE NDT
4.9.4 Statistical Methods............................................................ 29 METHODS......................................................................... 61
4.10 Preparation of Work-pack ..................................43 APP. B STATISTICAL APPROACHES TO NII......... 72
4.11 Inspection Plan Review .........................................43
DET NORSKE VERITAS

Page 6 – Contents
DET NORSKE VERITAS

Page 7
1. Introduction required and it is hoped to avoid shutting down the plant by

carrying out the inspection non-invasively. It is important to
1.1 Background clarify the objective of the non-intrusive inspection in advance,
as this may have an impact on the approach to the inspection.
Pressure vessels and pressure systems are required to undergo
periodic, statutory and other non-destructive testing to ensure Questions that should be addressed may include:
continued safe and reliable operation. This generally includes
a requirement to inspect vessels for possible internal degrada- — Is the inspection to complement an internal inspection pro-
tion. This is frequently achieved by means of an internal visual gramme?
inspection (IVI) coupled with the use of surface flaw detection — Is the inspection intended to replace an entire internal
methods such as dye penetrant inspection (DPI) and magnetic inspection or an internal inspection regime?
particle inspection (MPI), however, there can be very high Potential benefits of performing a non-intrusive inspection
costs associated with shutting down a vessel (loss of produc- include:
tion), isolating it and preparing for it for entry. Indeed, these
costs can be much higher than the cost of the inspection itself. — Avoids man access which can be hazardous (possibilities
Furthermore, the mechanical disturbances involved in prepar- of flammable or toxic residues which can be difficult to
ing the vessel for internal inspection and reinstating it may remove, adequate lighting may be difficult to achieve).
adversely affect its future performance. Finally, and by no — Planning for turnaround / shutdown. Identifying what
means least, man access may also be hazardous. remedial work is likely at the next turnaround? Carrying
There can, therefore, be significant advantages if inspections out non-intrusive inspection allows the preliminary
are performed from the outside of the vessel without breaking inspection to be made before the plant is shutdown, pro-
containment i.e. non-invasively. However, there needs to be a viding an opportunity for the turnaround to be shortened
balance between achieving these benefits and obtaining the by long-lead time planning and preparation (for repair and
information required to ensure continued safe and reliable maintenance based upon the NDT results) to be made in
operation. The acceptability and benefits of non-intrusive advance of the start of the turnaround.
inspection for a particular vessel will depend on a number of — Shortening the turnaround. Shutdown duration may be
factors including: reduced by carrying out most or all of the inspection work in
advance of the shutdown, allowing the turnaround to be
— vessel geometry restricted to mechanical work. This also simplifies planning.
— materials — Removal of requirement to break containment. No need to
— potential deterioration mechanisms and modes isolate, drain and purge the vessel. This may include partial
— locations and sizes of flaws of concern break of containment, for example access to water/coolant
— process side of a heat exchanger without breaking hydrocarbon con-
— historic inspection data tainment. This may lead to shortening of the turnaround.
— confidence in inspection capability — Minimises disturbances to the vessel which could create
— inspection costs. new problems.
— It may be possible to avoid the need to shut down the ves-
Techniques for non-intrusive inspection (NII) of equipment sel operation entirely. Inspections can be made on a differ-
are becoming increasingly sophisticated and more widely ent cycle from any other maintenance, or the inspections
available. While it may often be the preferred option, NII rep- may be made at reduced capacity or temperature, rather
resents a relatively new approach by comparison to IVI and than having to isolate, drain and purge the vessel.
many engineers responsible for inspection planning have yet to — Allows the inspection to be carried out when a potential
build up experience with and confidence in its application. For problem is identified, without interfering with other oper-
the purposes of this recommended practice, IVI is considered ations. This might occur when either routine surveillance
to include the use of surface flaw detection methods such as or unusual operating conditions suggest that damage
DPI and MPI where appropriate. In addition to a general lack might have occurred.
of experience, there are a wide variety of methods available to
non-intrusively inspect equipment, each with its own specific 1.3 Scope
capabilities and limitations. This recommended practice is primarily intended for those
This has been recognised by industry and a series of structured with responsibilities in the planning, implementation and
guidance documents have been developed Ref. [4], [5] and [6] acceptance of vessel inspections. It is aimed at the inspection
to assist plant operators to justify and plan NII. Each of these of welded vessels constructed from metals, and related items,
documents deals with a specific aspect of the NII process, and fittings and connections associated with them. The document
there are some minor inconsistencies in notation and approach. is equally applicable to forged/spun metal pressure vessels,
This recommended practice is intended to bring all three doc- which contain no welds, as well as storage tanks and large
uments together under a single cover, and to provide a consist- diameter pipelines.
ent and logical approach at all stages of the non-intrusive Although not intended to be covered by this document, some
inspection process. aspects of this recommended practice may also be applicable
Many of the recommendations in this document are not unique to vessels manufactured from other materials and plant items
to non-intrusive inspection since proper planning and adminis- other than pressurised equipment, in which case some of the
tration is also important for internal inspection. The document general principles developed may well still apply.
should be considered in conjunction with other relevant guide- The recommended practice provides guidance for:
lines such as those contained in the HSE Best Practice for NDT
document (Ref. [7]). 1) determining when NII is appropriate in principle
1.2 Objectives of Non-Intrusive Inspection 2) information requirements to plan for NII
It is essential to be clear about the reasons for performing a 3) defining the requirements for the NII method(s) to be
non-intrusive inspection. The decision to carry out non-intru- selected
sive inspection will normally depend on a number of different 4) selecting methods that meet the requirements
factors. A primary advantage is likely to be overall cost reduc-
tion, but this may arise not because an internal inspection is 5) evaluating the results of the inspection
replaced, but, for example, when an unscheduled inspection is 6) documentation requirements.
DET NORSKE VERITAS

Page 8
The scope is limited by the following constraints. 1.4 Overview of the Recommended Practice
— It is assumed that the date of the next inspection following Non-intrusive inspections generally require a more sophisti-
NII will be determined in the same way as with any other cated approach than internal inspections. The recommended
inspection, based on sound engineering judgement and an practice developed in this document recommends a systematic
understanding of the degradation rates and tolerance of the assessment of each item of equipment to be inspected using
equipment to degradation. This recommended practice NII. This is a staged process which sequentially considers:
therefore makes no attempt to influence the timing of the
next inspection, other than where the intended scope of the 1) when and where inspection is required
NII inspection has not been fully met. 2) whether NII is appropriate
— No consideration is given to the relative cost of different
inspection options and the guidance process considers 3) the inspection plan
technical issues only. 4) what inspection methods are appropriate
— The recommended practice is principally for application to
pressure vessels. 5) requirements during inspection
— The recommended practice does not address legislative 6) whether the inspection actually performed is adequate.
requirements, which may in certain countries preclude the
use of non-intrusive methods. The recommended practice A summary of the recommended NII process is shown dia-
considers only the technical aspects of the non-intrusive grammatically in Figure 1-1. The colours used on the flow
inspection planning process, and the user should confirm chart indicate the appropriate section of this recommended
that any inspection plan derived using this document satis- practice to which the action relates.
fies national legislative requirements.
— The recommended practice does not consider the impact It is important to recognise that a transfer to a non-intrusive
of external degradation mechanisms, for which inspec- inspection strategy is likely to require a step-change in the
tions are intrinsically non-intrusive and would be administration and execution of the inspection. In particular,
addressed by conventional assessment means. inspection methods are likely to be more elaborate when com-
pared with internal visual inspection. Therefore the inspection
It is also important to note that the recommended practice does must be controlled more rigorously, with the procedures (i.e.
not aim to comprehensively cover every aspect of planning an equipment, settings and reporting criteria) carefully scrutinised
inspection by NII but rather seeks to provide structured guid- and monitored at all stages in order to ensure that the inspec-
ance at key stages of the process. Although the document pro- tion objectives are met. As a consequence, it is likely that plan-
vides guidance on the general principles of non-intrusive ning of the inspection will need to be considerably more
inspection it is not intended to be prescriptive, and readers thorough (this is covered in Sec.4). In addition, the reporting
should assess each case point by point against their own crite- format must be precisely specified. If the results are not
ria, using the document as a guideline. The recommended requested in the correct form at the outset of the inspection or
practice cannot and is not intended to replace sound engineer- are inadequately reported, it can be difficult to transform the
ing and commercial judgement by competent personnel. data to the correct format, and useful information may be over-
looked or lost.
It should be recognised that with any scheme of examination
there is a finite probability of missing flaws or degradation
which could lead to failure. This applies to both invasive and
non-invasive inspection.
DET NORSKE VERITAS

Page 9
1. Integrity Review
(section 2)
Select pressure Integrity Prepare Determine date of

vessel for Review Equipment Profile next inspection
assessment e.g. RBI, CRA etc.
2. Decision Process
Alternative
(section 3)
inspection
approach required
No No
Is NII possible? Is NII possible?
Apply screening criteria Yes Apply high level decision Yes
(Figure 3-1) process
(Figure 3-2)
3. Planning Process
(section 4)
No
Prepare work- Does plan meet Review Inspection Prepare detailed
Yes Plans inspection plan
pack requirements ? e.g.Access, practicality (Figure 4-1)
4. Inspection
(section 5)
Yes
Inspection Alternative
Review inspection No
Perform inspection according to plan? inspection possible
results
Yes No
5. Evaluation
(section 6)
Evaluate Inspection meets Reduced interval Alternative
No No
inspection NII requirements? acceptable / inspection
(Figure 6-1) possible required
Yes Yes
Inspection Reduced
Return to complete inspection interval
start (Figures 7-1 / 7-2)
Figure 1-1
Overview of NII Procedure.
DET NORSKE VERITAS

Page 10
1.5 Definitions Internal Visual This is considered as an intrusive close visual

In the context of this Recommended Practice the following Inspection (IVI) examination of all internally accessible plate
material and, where applicable, conventional
definitions apply. In certain cases these definitions are the magnetic particle (MPI) or dye penetrant (DPI)
same as, or are based on, those which appear in other docu- inspection of welds. In the context of this Recom-
ments such as those of the British Personnel Certification in mended Practice, the term IVI is not intended to
Non-Destruction Testing (PCN) scheme. cover a less rigorous general visual examination.
Non-Intrusive This refers to any inspection performed from the
Area Specified zone where inspection will be carried Inspection (NII) outside of the vessel without having to break con-
out. In many instances this area will contain a tainment and/or not requiring vessel entry. It may
specific feature (e.g. weld) which is of particular be performed on-stream or off-stream.
interest. The terms “non-invasive” and “non-intrusive”
Capability Capability is used to qualitatively describe an are often used interchangeably.
NDT methods’ ability to detect (POD) and size Operator/ Qualified NDT personnel who execute the
flaws. technician inspection.
Certification Procedure used to demonstrate the qualification Probability of Probability of detecting a defined flaw type and
of NDT personnel in a method, level and indus- detection (POD) size in the area covered by the inspection method.
trial sector, and leading to the issue of a certifi-
cate. Procedure A written description of all essential parameters
and precautions to be observed when applying an
Competency Capability to perform a given task on the basis of NDT method to a specific test, following an
education, training, qualification and experience established standard, code or specification.
following objective assessment. To achieve the
appropriate level of competency might require a Qualification Evidence of training, professional knowledge,
team. skill and experience as well as physical fitness to
enable NDT personnel to properly perform NDT
Corrosion risk An assessment of the susceptibility of the struc- tasks, which satisfies the requirements of
assessment ture under investigation to all in-service degrada- EN 473 (Ref. [1]) and ISO 9712 (Ref. [2]), e.g.
(CRA) tion mechanisms that may affect it. The CRA is PCN (Ref. [3]), ASNT / TC1A.
not restricted to simply those degradation mecha-
nisms related to corrosion. Risk based Process of planning inspection requirements
inspection (RBI) through a detailed assessment of the relative
Coverage Defines the proportion of the structure or region probabilities of failure and their associated conse-
thereof under consideration that is actually sub- quences.
ject to inspection, i.e.
Work-pack A complete package of documents (procedures,
Area inspected drawings, standards etc.) relevant to the inspec-
Coverage = tion outlining scope and details of inspection to
Total area under consideration be performed.
Criticality A function of the risk associated with the
inspected equipment, incorporating likelihood of
degradation occurring and associated conse- 2. Integrity Review
quences.
Defect A defect is here taken to be a flaw which renders 2.1 General Approach
the equipment unfit for its specified service in its
current state. Internal visual inspection (IVI) remains widely used on the
Degradation Those mechanisms by which integrity of the pipe majority of vessels and several accepted procedures exist for
mechanism or vessel could potentially be impaired e.g. ero- the specification of inspection intervals to ensure safe opera-
sion, fatigue, creep, brittle fracture, wall loss etc. tion. Traditionally, such intervals have been specified on the
Effectiveness A qualitative measure of the probability of basis of legislative requirements, but increasingly plant owners
detecting flaws, taking coverage into account. take risk considerations into account in order to maximise the
Effectiveness = f (POD, Coverage). cost benefit of inspection. The decision to apply non-intrusive
Three effectiveness categories are used, (high,
inspection methods for a particular item of equipment can
medium and low), these being defined by com- depend critically on the type and extent of flaws or degradation
parison to the POD for visual inspection. High mechanisms expected. Non-intrusive inspection methods are
implies a higher POD, medium implies a broadly often slower and more expensive to apply than internal visual
similar POD and low implies a lower POD. inspections, so that 100% inspection of a vessel is often
Feature Specific part of area to be inspected i.e. weld, impractical. If non-intrusive inspection (NII) is to be used in
nozzle etc. conjunction with, or as an alternative to IVI, then the inspec-
Flaw The physical manifestation of a degradation tion requirements should be defined such that the risk levels
mechanism, in terms of cracking, pitting or wall are not increased. For any given vessel this effectively means
loss etc. that there should not be an increase in the probability of failure
Inspection body The organisation which manages the perform- (POF) when NII is used. Non-intrusive inspection therefore
ance of the NDT inspection (e.g. inspection ven- requires more careful consideration of the parts of the vessel to
dor) be inspected, the flaws/degradation to be detected and the
Inspection The plant owner’s representative with overall inspection methods to be applied than is normally the case for
manager responsibility for the inspection. internal (e.g. visual) inspections. Nonetheless, many of the
Inspection A specific way of applying a NDT method (e.g.
detailed planning considerations will remain the same for NII
Method Pulse echo, TOFD, Radiography etc.) as for IVI hence many of the traditional and more recent risk
based inspection (RBI) planning approaches will remain appli-
Inspection The leader of the site inspection team with overall cable.
supervisor responsibility for coordinating and supervising
the inspection. This section of the recommended practice is intended to provide
an overview of the various data requirements that are necessary in
order to be able to successfully plan for non-intrusive inspection.
2.2 Equipment Profile
The first stage of the non-intrusive inspection planning process
DET NORSKE VERITAS

Page 11
is to carry out a detailed review of the equipment. The intention Possible flaws/degradation
is to ensure that the inspection planners have a comprehensive Potential failure modes and effects. Possible flaw/degradation
understanding of the design, operation, current condition and locations. Possible flaw/degradation morphology/ sizes.
anticipated degradation mechanisms that may affect the equip- Reporting criteria. (See also Sec.2.3 and 2.4 below)
ment in service. This background information should be com-
piled into an equipment profile, which should form a part of the The following sections describe in brief some of the estab-
auditable document trail for the inspection. Background infor- lished techniques used to obtain some of this information.
mation that should be included in the equipment profile is as
follows: 2.3 Risk Based Inspection Approaches
Identity and design Current practice is increasingly to consider the risk presented
by any particular item of equipment to the plant [8] when
Vessel unique reference number, general arrangement draw- establishing an inspection schedule. Risk based inspection
ings, materials, current design basis (pressure, temperature, (RBI) processes consider the probability of a failure of the
corrosion allowance, cycling regime etc.) equipment and the associated consequences in order to deter-
Type of vessel and function mine an overall risk ranking. The frequency and extent of
e.g. separator, heat exchanger, boiler, storage tank, blowdown inspection required are then set on the basis of this risk rank-
vessel, reactor, etc. ing. Effective RBI depends heavily on a good knowledge of
the degradation mechanisms and their growth rates, the inspec-
Operation and service details tion history, operating conditions etc. – i.e. much of the knowl-
Process fluids and possible contaminants, operating tempera- edge that goes into RBI is the same as that required to enable
tures, operating pressures, loading and temperature cycles, justification and planning for NII.
transients, excursions outside normal operating envelope,
insulation systems. 2.4 Corrosion Risk Assessment
Detailed drawings Different non-intrusive inspection methods have different
capabilities for detecting and sizing flaws, and therefore the
Number and type of welds, longitudinal /circumferential shell inspection manager requires a more detailed knowledge of the
welds, welds on the domed end/dished end, nozzles, man- types and locations of flaws which may be present in a partic-
ways, construction details including saddles, supports and sup- ular item of equipment than is the case for an internal visual
port skirts, flanges, compensating plates, insulation etc. inspection. Hence, one of the key sources of information for
Modifications and repairs the NII assessment is the Corrosion Risk Assessment (CRA).
Has the vessel been modified since its original commission- In practice, most rigorous integrity management systems will
ing? Have any previous flaws or damage been removed or already incorporate a consideration of the type and likelihood
repaired (dates)? N.B. – it is important to maintain records of of degradation expected for each vessel, as this is also a
any such modifications or repairs. requirement for most RBI assessments.
Previous inspection results The corrosion risk assessment is a formal review of the degra-
dation mechanisms to which a particular plant item may be
Details of known/previously reported flaws/areas of degrada- susceptible, along with a determination of the anticipated deg-
tion. Inspection methods and coverage. radation rates. Typically the corrosion risk assessment will
General experience consider flaws or degradation such as:
Flaws/degradation/failures in other similar vessels (if availa- — general corrosion over the whole area
ble) — local corrosion
Complementary information — pitting
Information from corrosion monitors etc. providing evidence — erosion
on whether degradation is occurring. — general or preferential corrosion of welds (including
“grooving”)
Accessibility — loss of or damage to cladding or lining
General accessibility, access limitations. There may be scaf- — cracking under cladding or lining
folding requirements stated for each vessel. This sometimes — cracking in or near welds
accompanies the data held on vessels or is noted in drawings or — cracking at or near nozzles or other perforations
isometrics. Though this is primarily a cost issue, there may be — cracking at or from internal fitting welds, nubs etc.
occasions when there are overriding factors that preclude sat- — stress corrosion cracking in parent material
isfactory external access or the construction of suitable scaf- — hydrogen damage (e.g. blistering, stepwise cracking)
folding. — damage to seals, flange gasket, flange faces
Safety Requirements — damage to internals (not normally detectable by NII).
Details should be recorded of any safety requirements pertain-
In addition, it is important to consider the different and possi-
ing to the equipment. Examples of items to consider are listed
below, however any other safety related information consid- bly unusual flaw morphologies which can occur (e.g. micro-
ered pertinent to the inspection should also be recorded. biological influenced corrosion) since these aspects can
influence the selection and capability of non-intrusive inspec-
— general safety procedures tion methods.
— site-specific procedures including requirements for local
induction courses 2.4.1 Corrosion Risk Assessment Types
— job-specific risk assessments Corrosion risk assessments can be carried out at very different
— local safety review requirements for equipment (at stores/ levels within different organisations. Given the importance of
safety officer) the assessment to the NII planning process, it is therefore
— safety testing and certification requirements for equipment important to be clear about what level of assessment has been
– Portable Appliance Testing (PAT) certification etc. carried out. The following CRA Types have been defined in
— requirement for “inherently safe” equipment order to describe the nature and extent of the CRA carried out.
— restrictions on ionising radiations These CRA Types are used in the NII decision process later in
— personal protection equipment. this recommended practice (Sec.3).
DET NORSKE VERITAS

Page 12
CRA Type 1 which in turn will influence the effectiveness of inspection

required in order to provide assurance that failure will not
A basic assessment considering primarily existing inspection
results and expected degradation based on experience with occur in-service between scheduled inspections.
other similar vessels. This level of assessment is what would
exist as a minimum to meet inspection planning requirements
2.6 Operational Experience
by RBI. It would be carried out and reviewed by competent Where the user has extensive experience with the same or sim-
individuals from appropriate discipline groups including, for ilar vessels in the same or similar service, operational experi-
example, inspection, metallurgy/materials, process and engi- ence provides a very useful corroboration of the theoretical
neering. assessments described previously. Inspection histories outlin-
CRA Type 2 ing the types and sizes of any flaws found in service (or indeed
the absence of flaws), give a clear indication of what inspec-
A more detailed assessment providing documented consideration requirements are likely to be necessary. Of course, this is
tion of at least the following: only true if the inspection carried out is appropriate for the deg-
— the vessel's condition, based on previous inspections radation anticipated.
whenever these have been carried out
— the vessel's metallurgy, modification and repair history
— the process fluid composition and operating conditions 3. The decision guidance process
— corrosion management
— changes to any of the above factors that may affect inspec- 3.1 Introduction
tion requirements Having carried out the integrity review and obtained the nec-
— types of degradation essary information regarding the equipment condition at the
— growth rates for each type of degradation identified as of time of the last inspection, it is necessary to determine whether
concern (or at least some estimate related to severity, e.g. the equipment is intrinsically suitable for non-intrusive inspec-
a likelihood type value from a criticality assessment) tion, or whether an alternative, intrusive technique such as IVI
— the locations where each type of degradation is likely to be is necessary. This decision process is the second stage of the
active. NII planning process, shown in Figure 1-1. Through applica-
tion of a flow chart, the decision guidance process determines
The assessment should consider existing inspection results and whether NII should be considered for the inspection of a given
theoretical predictions. Theoretical predictions should be piece of equipment.
based on process and materials information.
The process can be broken down into two main stages, namely
CRA Type 3 screening and the high level decision process. The process is
A comprehensive assessment including an in-depth theoretical shown diagrammatically in Figure 3-1, and is described in
study based on process and materials information. The follow- more detail in the following sections.
ing issues need to be addressed:
3.2 Screening
— types of degradation
The purpose of this first stage is to rapidly identify those ves-
— growth rates for each type of degradation identified as of sels for which NII should not be considered or where the
concern
required information can not be obtained from such an inspec-
— the locations where each type of degradation is likely to be tion. The screening process is based on the user’s response to
active each of the following questions.
— upset conditions that can lead to accelerated degradation
— potential for incorrect identification or omission of degra- Is the vessel intrinsically suitable for NII?
dation mechanisms. Before proceeding any further with the NII decision process, it
is necessary to confirm that the equipment is intrinsically
A Type 3 assessment is carried out in detail on a vessel by ves- suited to inspection by non-intrusive means; that is that there
sel basis.
are no immediately obvious impediments to NII being under-
CRA Type 4 taken. These include factors such as where there is no access to
This meets the requirements of a Type 3 assessment with the vessel exterior, extreme surface temperatures, geometry
respect to the theoretical study but also includes consideration constraints and restrictions to access, as well as any require-
of inspection results from the vessel following at least one in- ment for inspection of internal fittings.
service inspection. Some interpretation/correlation of the pre- Has the vessel previously been inspected and is the history still
dicted degradation and condition found by inspection shall relevant?
also have been carried out.
Vessels with no previous in-service inspection history or for
2.5 Structural Integrity Assessment which there is reason that the inspection history may no longer
be relevant (due for example to a change in process conditions)
It is important to have an understanding of the equipment’s
ability to resist structural damage following degradation. In its should not normally be considered for NII. There are two pos-
simplest form, this can be knowledge of the margins against sible exceptions to this recommendation, as covered by the
operating conditions and corrosion allowance incorporated at next two screening questions.
the design stage. In certain circumstances however, a more Is the vessel designed specifically for NII?
comprehensive fitness for service assessment will have been
carried out (for example using API 579 [9) which will provide Where a vessel is designed specifically for inspection by NII,
information on, for example, critical crack dimensions. All such inspection should be considered from the outset or even
such information provides the inspection manager with evi- when conditions may have changed (while remaining within
dence as to how tolerant the equipment is to degradation, the design intent).
DET NORSKE VERITAS

Page 13
START
Select vessel for IVI / NII
Is the vessel intrinsically No

NII is not appropriate
suitable for NII?
Yes
Has vessel previously No Was vessel designed No Is vessel the same as No

others for which service NII is not appropriate
been inspected? specifically for NII ?
history exists?
Yes Yes Yes
Is operating history still

relevant? No
Yes
Is entry scheduled for other No Apply flow chart for high Yes Prepare NII inspection
level decision Can NII be considered
reasons? plan.
(Figure 3-2)
Yes No
Perform IVI NII is not appropriate
Figure 3-1
NII Screening Procedure
Is the vessel similar to others for which service history exists? Is entry scheduled for other reasons?
The intent of this question is to identify if there are other ves- When the vessel is to be opened for other reasons, advantage
sels whose inspection history may be directly relevant to the should be taken of the opportunity to perform an internal visual
vessel under consideration. Hence the answer can only be inspection. This does not mean that NII should not be done.
taken as yes for vessels: However, if it is intended to do NII in parallel with IVI then
this can be done without additional justification.
— substantially the same in terms of design, geometry, con-
struction and conditions of service (i.e. normally empty / 3.3 High-level decision process
full, etc.).
and The next step is to decide whether sufficient information exists
— for which no factors with potential to cause a difference in to plan the non-intrusive inspection and what inspection effec-
the nature, distribution or rate of degradation can be iden- tiveness is required. This requires consideration of how confi-
tified. dently potential flaw types and locations can be predicted, the
effectiveness of previous inspections, and the severity and rate
Similar vessels shall be taken to mean vessels substantially the of any known or predicted degradation. This is achieved using
same in function, geometry, design, material and construction. the high-level decision process, which determines whether NII
Similar service shall be taken to mean substantially the same is appropriate in principle, based largely on the use of the deci-
in each of chemistry, fractions and phase(s) of the vessel con- sion tree shown in Figure 3-2. This leads the user to a decision
tents, process type(s), flow rates and temperatures. When on whether NII is appropriate in principle based on the catego-
classing vessels as similar, justification must be provided. ries selected for each of three factors. The factors and criteria
DET NORSKE VERITAS

Page 14
for categorisation are covered below. sideration and their locations will depend on a number of
factors. Evidence can be considered from two main sources, as
3.3.1 Confidence in ability to predict types and locations described below:
of degradation
The decision on whether NII is appropriate in principle is theoretical: The nature of the integrity management systems
based to a large extent on confidence in being able to predict employed
all active degradation mechanisms and hence specify methods evidential : Evidence available from the same or similar ves-
capable of identifying the associated flaws. The ability to pre- sels in the same or similar service (as defined
dict degradation mechanisms relevant to the vessel under con- previously).
Confidence in ability to
Previous inspection Severity and rate of
predict type and NII Possible
effectiveness degradation
location of degradation.
High High High Yes
Medium Medium Medium Yes
Low Low Low Yes
High Yes
Medium Yes
Low Yes
High No
Medium Yes
Low Yes
High High Yes
Medium Medium Yes
Low Low Yes
High No
Medium Yes
Low Yes
High No
Medium No
Low Yes
High High No
Medium Medium Yes
Low Low Yes
High No
Medium No
Low Yes
High No
Medium No
Low No
Figure 3-2
High Level Decision Guidance Chart
DET NORSKE VERITAS

Page 15
When using evidential information as the basis for predicting whenever these have been carried out
types and locations of degradation, it is important that the pre- — the vessel's metallurgy, modification and repair his-
vious inspection results have been considered in detail at the tory
planning stage for the current inspection and that a Type 1 — the process fluid composition and operating condi-
CRA is in place as a minimum. The credibility of the evidence tions
is directly linked to the amount of evidence available - i.e. — corrosion management
number of vessels and previous inspections considered. — the degradation types, locations and rates
The confidence categories have therefore been defined as fol- — changes to any of the above factors that may affect
lows to consider the above factors. inspection requirements.
The integrity management plan ensures consideration of
High the above factors by a team of competent individuals from
Either of the following factors apply: appropriate discipline groups including, for example, in-
spection, metallurgy/materials, process and engineering.
1) a thorough assessment of potential degradation mecha- The CRA carried out as part of the integrity management
nisms gives confidence that all relevant mechanisms and activities described above should conform to the require-
their likely locations have been predicted. This assessment ments of Type 2.
has been carried out as part of the integrity management
activity by which inspection planning is conducted. This 2) experience, including inspection histories, is available for
includes as a minimum for each vessel, documented con- the same or similar vessels in the same or similar service
sideration of: covering not less than four inspections in total and not less
than a single inspection for the longest serving vessel con-
— the vessel's condition, based on previous inspections sidered in making the judgement. A detailed review of all
— the vessel's metallurgy, modification and repair his- previous inspection results has been performed. These
tory findings have been considered in predicting the type and
— the process fluid composition and operating condi- locations of flaws that may be expected in the current
tions inspection. In addition, a CRA shall have been carried out,
— corrosion management this conforming at least to the requirements of Type 1.
— the degradation types, locations and rates
— changes to any of the above factors that may affect Low
inspection requirements.
Justification for inclusion in the High or Medium categories, as
The integrity management plan ensures consideration of defined above, is not possible.
the above factors by a team of competent individuals from
appropriate discipline groups including, for example, in- 3.3.2 Previous inspection effectiveness
spection, metallurgy/materials, process and engineering. This is included as a further measure of the confidence in abil-
The CRA carried out as part of the integrity management ity to predict all relevant flaw types and is also used in defining
activities described above should conform to the require- the effectiveness required for methods used in the current
ments of Type 4. inspection as a means of ensuring the probability of failure is
2) the user has extensive experience with the same or similar managed. The intent is to compare the effectiveness of the last
vessels in the same or similar service and inspection histo- inspection carried out relative to that of a conventional internal
ries outlining the types of flaws existing are available. A visual inspection, i.e. consisting of close visual examination of
detailed review of all previous inspection results has been plate material and, where applicable, MPI or DPI.
performed. These findings have been considered in pre- The following category definitions apply:
dicting the type and locations of flaws that may be
expected in the current inspection. Note that extensive High
experience is taken to mean that data is available covering
at least eight inspections in total and not less than two The inspection performed has a better probability of detecting
inspections for the longest serving single vessel used in flaws of concern than does internal visual inspection.
making the judgement, at least one of which should have
been a close visual inspection (internal or external depend- Medium
ing on the nature of the degradation). In addition, a CRA The inspection performed offers a probability of detecting
shall have been carried out, this conforming at least to the flaws of concern broadly similar to that of internal visual
requirements of Type 1. inspection.
Note that a vessel for which the operating conditions have Low
changed since the previous inspection can not be classed as
High, except where the conditions can be shown to be more The inspection performed has lower probability of detecting
benign than previously. flaws of concern than does internal visual inspection.
In deciding which category to select, the user must consider the
Medium following:
Either of the following apply: 1) the probability of detection of the method(s) used with
1) a thorough assessment of potential degradation mecha- respect to flaws of concern. This is given by the effective-
nisms is considered likely to predict the majority of rele- ness assigned to each method for a given flaw type.
vant mechanisms and their locations but cannot fully rule 2) the inspection coverage and how this relates to the loca-
out the possibility of additional degradation mechanisms tions of flaws of concern.
being active. This assessment has been carried out as part
of the integrity management activity by which inspection Clearly, the user’s knowledge of what flaws are likely to be of
planning is conducted. It includes as a minimum for each concern and where these flaws may be located is important in
vessel, documented consideration of: assigning effectiveness. For example, if it is known that flaws
only occur in a particular region of the vessel and this region
— the vessel's condition, based on previous inspections was fully inspected with a highly effective method then a High
DET NORSKE VERITAS

Page 16
inspection effectiveness may be justified. Conversely, if the — process fluid (tube side): Water.
user has little understanding of what flaws might exist then a
high coverage inspection but using only a single method The vessel is designed to ASME VIII Div 1 with a 3.2 mm cor-
(which may be capable of detecting only one flaw type) may rosion allowance. The vessel forms part of a new process
result in a Low ranking. stream (Stream 2). This is similar in operation to Stream 1 (in
service for a period of 10 years) but has been designed for
The user can consider previous inspections on similar vessels greater efficiency and higher throughput. The vessel under
in similar service (as defined above) in determining the cate- consideration performs the function of two vessels (1-HX-05
gory to select. and 1-HX-06) on Stream 1. Both vessels have been subject to
3.3.3 Severity and rate of degradation inspection every three years. The first two inspection were by
IVI but the most recent was by NII, this being regarded as at
In making this judgement, the user will consider the worst least as effective as the IVI in identifying the main degradation
affected zone of the vessels and the following category defini- mechanism. The shells on both vessels are observed to suffer
tions apply. from mild internal corrosion. The average rate has been deter-
mined to be approximately 0.15 mm per year. The initial
High inspections showed that the corrosion rate does vary with posi-
The degradation and rate thereof is such that failure of the ves- tion in the vessel. This is believed to be related to the gas flow
sel or rejection based on inspection results can reasonably be rate and condensate impingement. On 1-HX-06, for example,
expected within the remaining plant lifetime. the first inspection revealed one area with localised loss of wall
thickness of 1.5 mm. An investigation carried out at the time
Medium suggested that the problem was occurring in a region of partic-
ularly high flow where water droplets were likely to be
The degradation and rate thereof is such as to be observable entrained. This was considered to be sufficient to remove the
during the plant lifetime but would not be expected to threaten protective hydrocarbon rich film that would normally be
the integrity of the vessel during this period. expected on the steel and lead to accelerated CO2 corrosion.
An additional diverter plate was installed during the shut-
Low down. Subsequent inspections revealed that this had resolved
There is no degradation expected or degradation is superficial. the problem and the region was no longer subject to higher cor-
rosion than the remainder of the vessel.
3.3.4 NII recommendation The design of 2-HX-05 is different from that of the two vessels
Having answered the above questions, in conjunction with the it replaces in requiring a higher flow rate and more complex
flowchart, the suitability of the equipment for inspection by gas flow path. In addition, the gas exit temperature is lower
NII can be read from the chart. than that for 1-HX-06, hence the rate of water condensation is
expected to be slightly higher. The feedstock for Stream 2 is
3.4 NII Decision Record nominally the same as for that for Stream 1 although it does
As with all matters related to plant safety, it is essential that an come from a different source.
auditable record is kept of all factors considered in making Application of the screening process (Figure 3-1):
decisions throughout the above decision guidance process.
Documentation shall cover as a minimum the following: — vessel has had previous inspection and history is still rele-
vant? No
1) statement of any changes occurring in process that may — vessel designed specifically for NII? No
affect the nature or rate of degradation — vessel is the same as others for which service history
2) inspection reports (if not already included in the Equip- exists?
ment Profile)
The response here should be No. The two vessels for which
3) justification for acceptance under screening criteria service history exists are similar in overall function but differ-
4) list of vessels considered to be the same as the one under ent in respect of design and process conditions and hence it is
consideration and justification that degradation can be possible that the type and rate of degradation might be differ-
expected to be the same ent. The No response at this point indicates that NII is not rec-
ommended for this vessel for the forthcoming inspection.
5) justification of selection of category for:
3.5.2 Gas Receiver Vessel
— confidence in ability to predict types and locations of
degradation A gas receiver vessel is due for its first inspection three years
— previous inspection effectiveness after entry to service. It one of ten vessels having the same
— severity and rate of degradation. design and function. Gas enters the vessels from a common line
and gas take-off is to a manifold. Five of the vessels have been
3.5 Examples in service for ten years and each has had three previous inspec-
A number of examples of application of the decision process tions, the last two of which have been by NII. Inspection reports
are presented in this section. are available for each inspection performed and these have been
reviewed in planning for the current inspection. The reports
3.5.1 Heat Exchanger Vessel indicate that mild internal corrosion has been found on the inte-
rior at the bottom of each vessel. The corrosion is fairly uniform
A heat exchanger vessel (2-HX-05) has been in service for two across this region in each of the vessels, with a maximum loss
years and is scheduled for its first inspection since commis- of wall of 1 mm on one of the vessels (the design incorporates
sioning. A decision is to be made whether this can be by NII or a corrosion allowance of 5 mm). This corrosion is expected as
whether an IVI should be performed. the gas is not completely dry and a certain amount of moisture
The following information is relevant: deposition occurs at the bottom of the vessel. The recent inspec-
tions by NII have included extensive wall thickness readings in
— material (shell and ends): Carbon steel the corroded regions and TOFD was undertaken on the vessel
— material (tubes): Stainless steel welds since fatigue cracking under the pressure cycling was
— process fluid (shell side): Wet hydrocarbon gas with some identified as a possible failure mechanism. None of the TOFD
condensation expected tests undertaken have identified any cracking.
DET NORSKE VERITAS

Page 17
A decision on whether the inspection can be by NII is required One can therefore proceed to apply the high level decision
as this is the preferred option. guidance flow chart (Figure 3-2):
Application of the screening process (Figure 3-1):
— confidence in ability to predict types and locations of deg-
— vessel has had previous inspection and history is still rele- radation. The reviews performed as part of the inspection
vant? No planning process are considered to fall within the option 1)
— vessel designed specifically for NII? No requirements for selection of this factor (see item 3.3.1,
category High). The High category is selected here given
— vessel is the same as others for which service history that the initial conclusions of the reviews have been vali-
exists? Yes, since the vessel is of the same design, is in the dated by the inspections carried out - i.e. there is now no
same service and there are no factors that would be expectation that further, as yet unidentified, mechanisms
expected to lead to differences in degradation. may be active
— entry is scheduled for other reasons? No.
— previous inspection effectiveness. The previous inspec-
Proceeding to apply the high level decision guidance flow tions by IVI can be taken as Medium
chart (Figure 3-2): — severity and rate of degradation. The internal inspections
performed did not reveal significant degradation but cor-
— confidence in ability to predict types and locations of deg- rosion (not expected to impact on vessel integrity during
radation. Given that there are a large number of vessels of the remaining lifetime) was observed on the exterior. A
the same design and function and that many of these ves- Medium ranking therefore applies here.
sels have long inspection histories, the user can directly
consider evidence available from these vessels in selecting Following Figure 3-2 with Medium, Medium, Medium, one
the category here. Considering the option 2) (category arrives at a recommendation that non-intrusive inspection is
High), in item 3.3.1, High is justified here possible in principal.
— previous inspection effectiveness. The previous inspec- This vessel has jacketed external insulation hence NII meeting
tions by NII on the vessels already in service can be taken the efficiency requirement may not be possible unless the insu-
as Medium, i.e. broadly equivalent to IVI lation is removed.
— severity and rate of degradation. Corrosion occurring at
the same rate as on the vessels in service would not be 3.5.4 Absorber Vessel
expected to threaten the integrity of the vessels during the
design lifetime, hence Medium is selected here. A gas sweetener drum operates by absorption of H2S in a zinc
oxide bed. The zinc oxide becomes depleted through operation
Following Figure 3-2 with High, Medium, Medium, one and requires replacement after some time in service. The rate
arrives at a recommendation that NII is possible in principle. of depletion depends on the nature of the gas and replacement
timing is determined by monitoring the effectiveness of the
3.5.3 Separator Vessel sweetening based on the downstream gas composition.
Replacement of the zinc oxide requires shut down of the vessel
A High Pressure Separator Vessel is due for an inspection. The and is a complex and expensive operation. Entry for internal
vessel has been in service for eleven years. Its first inspection visual examination requires removal of the zinc oxide beds and
was after three years in service. The interval to the next inspec- this involves a cost almost as high as that for replacement.
tion was set at four years and this interval has been retained to There are two vessels, referred to herein as A and B, running
the present inspection. A decision on whether this inspection in parallel and a single vessel is capable of sweetening the
can be by NII is required. entire gas throughput hence there are no production losses
The vessel is constructed of carbon steel that is clad internally associated with a shut down. Nevertheless the costs of
with a stainless steel liner. The vessel exterior is fully insulated removal/replacement mean that considerable savings can be
made if entry can be timed to coincide with replacement of the
and the insulation is protected by galvanised steel sheet. zinc oxide bed when it becomes necessary.
The integrity management plan for the plant includes technical The vessels fall under an integrity management plan that uses
reviews, considering the main threats to integrity, on a vessel a risk based approach to inspection planning. As part of this, an
by vessel basis. This has identified that corrosion and erosion operational risk assessment was performed to identify and
are the only degradation mechanism of concern here. External assess possible damage mechanisms. This indicated that inter-
corrosion (under the insulation) has been identified as having nal pitting type corrosion in the wet sour gas environment is
a higher probability than internal corrosion/erosion. However, the main threat to integrity. The rate of corrosion is expected to
if the carbon steel becomes exposed to the process fluids (after be moderate however, and wall loss would not be expected to
breakdown of the stainless steel cladding), corrosion can be exceed the corrosion allowance through the vessel lifetime.
very rapid hence integrity of the cladding is important. Corro- The Risk Based Inspection plan devised at the time of vessel
sion is more likely to be in the form of pitting than of a gener- commissioning recommended inspection by IVI at intervals of
alised nature. four years. After the first four years of operation, the zinc oxide
showed little degradation in performance and a decision was
Both of the previous inspections have included visual inspec- taken to delay the IVI but perform an interim NII. Both vessels
tion of the vessel interior. This included close visual examina- were subject to NII consisting of (a) point thickness readings
tion of all accessible internal surfaces. All areas inspected were at 42 shell/head and 24 nozzle locations for which pre-service
reported to be in good condition with no visible degradation. measurements had been made and (b) Time of Flight Diffrac-
This information has been considered in the most recent tech- tion testing over each of the main longitudinal and girth welds.
nical review undertaken in planning for the forthcoming The thickness gauge measurements did not show any notable
inspection. In order to check for signs of external corrosion, changes in wall thickness. No reportable flaws were identified
sections of insulation were removed during each of these by the TOFD inspection. Based on the results obtained, the
inspections. The insulation was seen to be dry in each case and inspection plan was revised to perform the first IVI after a fur-
mild localised corrosion (<0.5 mm) was observed. ther two years in operation (or sooner if a change of the zinc
oxide was required).
Application of the screening process (Figure 3-1):
After this two year period it is evident that the zinc oxide in
— vessel has had previous inspection and history is still rele- both vessels has considerable remaining life hence it would be
vant? Yes desirable to avoid opening either vessel. The decision guidance
— entry is scheduled for other reasons? No. process is to be followed to establish is this is possible.
DET NORSKE VERITAS

Page 18
Application of the screening process (Figure 3-1): 4. Inspection planning

— vessel has had previous inspection and history is still rele- 4.1 Introduction
vant? Yes (NII)
— entry is scheduled for other reasons? No. The principal objective of inspection planning is to establish a
cost effective strategy which provides a satisfactory level of
Proceed to apply the high level decision guidance flow chart confidence in the vessel’s safe and reliable operation until the
(Figure 3-2): next inspection. Inspection planning is a complex task that
demands consideration of a broad spectrum of issues, ranging
— confidence in ability to predict types and locations of deg- from detailed technical assessments of the impact of vessel
radation. The integrity management approach and opera- operating conditions on degradation through to budget plan-
tional risk assessments performed allow option 1) to be ning and allocation. As well as the technical considerations, it
considered in determining the category applicable here. is generally also necessary to satisfy certain legislative require-
While the operational risk assessment has considered in ments. While this aspect is not covered specifically in this rec-
detail the types of degradation likely, the findings cannot ommended practice, since the requirements are usually country
be taken to be fully validated by the inspections carried out specific, it remains an important part of the planning process.
to date. It is conceivable that further, unexpected, degrada- The non-intrusive inspection plan for a vessel defines which
tion mechanisms might not have been identified by the parts of the vessel should be inspected, what inspection meth-
inspection, given the limited extent of the NII carried out. ods should be used, and what coverage is required. This section
Hence a Medium ranking is assigned here of the recommended practice provides guidance on determin-
— previous inspection effectiveness. The previous inspec- ing the most effective non-intrusive inspection plan for a given
tions (on both vessels) were by NII. The thickness gauge vessel, in terms of establishing an appropriate balance between
inspection at fixed points is judged as having a markedly vessel integrity and inspection cost / duration. In particular it
lower probability that IVI of detecting the flaw type of addresses the question of how to establish an appropriate bal-
most concern (pitting). Hence the previous inspection ance between sensitive but relatively slow (and therefore
expensive) inspection methods, and rapid (and therefore less
effectiveness is taken as Low here expensive) but less sensitive screening methods.
— severity and rate of degradation. The thickness gauge
measurements did not reveal any clear loss of wall thick- Approaches to planning will vary from company to company
ness. However, the operational risk study did suggest cor- but, typically, the plan will define at least the following:
rosion can be expected, although not sufficient to exceed
the corrosion allowance through the vessel life. Hence a — timing for the inspection
Medium ranking applies here — type of inspection method(s) to be used
— following Figure 3-2 with Medium, Low, Medium, one — regions of the vessel to be inspected
establishes that NII is not recommended — shortlist of inspection service provider(s)
— based on the above recommendation, a decision was made — qualifications of inspectors, Reporting requirements
to perform IVI on one of the vessels and vessel B was — requirements from plant operations (e.g. shut down)
opened for inspection. This consisted of close visual — requirements from plant maintenance (e.g. set up scaffolding)
examination of all internal surfaces and MPI on all nozzle — safety requirements (equipment and personnel).
welds and approximately 20% of shell weld length. Most
of the internal surfaces showed some minor pitting, with a Clearly there are many additional considerations that go into
maximum depth of 0.75 mm. No crack like flaws were the overall inspection planning activity. In particular, consid-
identified in any of the welds eration should be given to the following:
— the zinc oxide in vessel A remains in good condition and — whether inspection to be performed on stream or off stream
it would be preferable to continue to operate on the basis — temperature during inspection (high temperature is likely
of having performed NII rather than opening the vessel for to be the main concern, though difficulties may arise from
IVI. The decision guidance process is to be followed in the build-up of ice on low temperature items).
determining whether this is a justifiable approach.
— whether the inspection is on the critical path of an outage?
Application of the decision guidance process: — whether the inspection is opportunity driven?
— requirement for comparison with previous/past inspections
— vessel has had previous inspection and history is still rele- — cost and time constraints.
vant? Yes (NII)
— entry is scheduled for other reasons? No The steps in planning and implementing a non-intrusive
inspection are the same, regardless of whether the inspection
Proceed to apply the high level decision guidance flow chart: requirement has been determined using a risk based inspection
(RBI) or more prescriptive approach. However, there can be a
— confidence in ability to predict types and locations of deg- high degree of synergy between non-intrusive inspection and
radation. Again the option 1) definitions are considered in RBI since both approaches require similar types of information
determining the category applicable. The same considera- to be available.
tions apply here as for vessel B but now the potential exist- The approach developed here is intended to be systematic,
ence of unexpected degradation mechanisms can be ruled thereby promoting a consistent approach between operators.
out based on the findings of the inspection on vessel B. Nonetheless, it does not remove the need for input and review by
Hence a High is assigned competent personnel and the results must be reviewed by com-
— previous inspection effectiveness. The inspection on ves- petent inspection planning personnel to check that the inspection
sel B was by IVI and the results are relevant to vessel A plan is broadly consistent with their engineering judgment.
hence Medium applies here
— severity and rate of degradation. The thickness gauge The main elements in devising an inspection plan are as follows:
measurements did not reveal any clear loss of wall thick- Identification and selection of the planning team.
ness but the internal visual inspection on vessel B showed
pitting up to 0.75 mm depth. Hence a Medium ranking The team to plan, prepare and execute the inspection must be
applies here. competent across a range of disciplines. These must be identi-
fied and appropriate personnel selected.
Following Figure 3-2 with High, Medium, Medium, one
arrives at a recommendation for that NII can be performed on Definition of the inspection strategy.
this vessel in principle. The planning approach will depend in part on the intended
DET NORSKE VERITAS

Page 19
inspection strategy, i.e. whether the inspection is intended to ing a particular combination of geometry, material, likelihood
confirm the absence of degradation, or to establish the depth or of degradation etc.
extent of known active degradation mechanisms. Selection of inspection methods and coverage.
Definition of vessel zones. Selection of appropriate inspection method(s) and coverage for
The vessel can be treated as one or more zones, each represent- each of the defined zones (some zones may require no inspection).
START
Select vessel for IVI/ NII
Determine Inspection
Identify Strategy Type NII Evaluation
Inspection Zones
(Figure 4-2)
Determine anticipated
degradation type .
Determine efficiency of
Determine required candidate inspection
Inspection Effectiveness
methods
(Figure 4-4) (Flowcharts)
Review the risk

Is a technique available No associated with
No Alternative Inspection
with efficiency >= the Is the risk acceptable?
performing inspection Required
requirement ?
with lower efficiency
Yes
Yes
No Review the risk No

Is the technique able to Alternative Inspection
size defects? associated with not Is the risk acceptable? Required
accurately sizing defects
Yes Yes
No
Have all relevant defect
types been considered ?
Yes
No Have all zones been

Decision Guidance
Equipment Profile
Records
considered ? (Section 2 .2)
(From Section 3.4)
Yes
Prepare Work -pack
Figure 4-1
Inspection Planning Flowchart
DET NORSKE VERITAS

Page 20
4.2 The Planning Team — location of degradation, preferential attack or more

random.
Inspection planning is generally conducted by a team consist-
ing of people with responsibilities in different areas (e.g. pro- b) Potential to threaten integrity (what resistance to degrada-
duction, process, metallurgy, inspection, maintenance), the tion is embodied in the design)
objective being to ensure that the inspection is effective within
the constraints imposed by sometimes conflicting internal — corrosion allowance
needs.
— critical crack depth.
The development of a non-intrusive inspection plan can be sig-
nificantly more complicated than is the case for an internal (e.g. c) Degree of assurance – feeds into the inspection perform-
visual) inspection. A multi-disciplinary approach is required ance requirements.
which is likely to involve appropriate experience in engineer-
ing/materials/ process operations/non-destructive testing. Any inspection program specified must also be able to give
The basic team required to assess the requirements for a non- some degree of assurance that unexpected damage mecha-
intrusive inspection should consist of personnel with compe- nisms are not occurring which might lead to failure of the com-
tencies in the following areas: ponent.
The above clearly rests on knowing what degradation mecha-
— general knowledge of the construction of containment ves- nism to inspect for – once this is determined (through corro-
sels and systems, materials and materials processing, fab- sion risk assessment [CRA] or historical evidence) the
rication processes etc. questions regarding assurance and potential threats to integrity
— corrosion or materials technology can be addressed through defining where to inspect and how to
— specific knowledge of the systems to be inspected, opera- inspect. These assessments must take into consideration the
tional history and ‘general knowledge’, (knowledge of the future operating conditions for the component, and not merely
working practices and history of the system, safety reflect past conditions. For example, where produced fluids are
requirements, and the likely conditions at the time of increasingly sour, it would be misleading to base the inspec-
inspection) tion requirements on a CRA which has assumed low sulphur
— non-destructive testing content, resulting in a downplaying of the importance of
— nominated person to coordinate the overall process. inspection for mechanisms such as sulphide stress cracking
(SSC).
Team members may have more than one of the specified skills; The starting point in assessing any deviation from the specified
it is not necessary for the team to contain individual specialists
in all of the above. inspection plan is an understanding of how the objectives of
the inspection might be affected, e.g. does the non-conform-
The most effective team is likely to be the smallest team that ance significantly compromise the ability of the inspection to
has the requisite skills, but the team should not be reduced meet the overall objectives. It is therefore important that the
excessively, as items are likely to be overlooked. Personnel to objectives be understood, bearing in mind that these may not
carry out any of these functions should be competent to assess be defined in detail in the inspection plan itself (as this defines
their own level of expertise in the selected area. One member the work scope for the inspection team).
of the team should take responsibility for the overall planning
process. As stated above, different vessels will clearly have different
inspection requirements, not just in the detail of the “what,
4.3 Inspection Strategy Type where, how”, but also in the basis of the approach. For exam-
ple, a vessel for which it has been determined that generalised
The objective of any inspection, at the highest level, is to give corrosion is the main degradation mechanism of concern but
a high degree of assurance that any degradation with potential with a low probability would be treated differently to a vessel
to threaten integrity is detected before the next inspection. in which weld cracking by stress corrosion, such as hydrogen
However, how it achieves that ambition will vary according to induced stress corrosion cracking (HISCC) or sulphide stress
the specific details of the item under consideration. For exam- corrosion cracking (SSCC), is seen a high likelihood. The dif-
ple, the inspection regime for a vessel where the corrosion risk ferences on this level can form a useful basis for categorising
assessment has shown no likely degradation mechanisms will they type of inspection in a way that facilitates subsequent
be very different from the inspection regime for a vessel where assessment of non-conformances. To this end, three inspection
stress corrosion cracking is predicted to be likely. types have been defined, as described in Table 4-1 below.
In the course of defining the inspection schedule the following There is a clear difference in emphasis in each of the inspection
three, closely linked, aspects must therefore be taken into con- types and this provides a useful framework for the establish-
sideration: ment of the inspection plan, as well, as later in the process,
treatment of any non-conformances. The three categories are
a) Degradation method: discussed in more detail in the following sections, and guid-
ance on the assignment of an inspection type to a particular
— nature of degradation, global wall thinning or cracking item is given in Sec. 4.3.4.
Table 4-1 Inspection Type Definitions

Type Definition
Degradation mechanism NOT expected to occur.
A Inspection is required to confirm there is no onset of the degradation mechanism.
Degradation mechanism expected, with low / medium progression.
Location of degradation can be predicted.
B Not anticipated to impact on vessel integrity in the medium term (typically at least 2 outage periods).
Inspection required to confirm CRA predictions.
Degradation expected with medium / high progression.
Location of degradation can not be predicted.
C MAY impact on vessel integrity in the medium term (two-outage timeframe).
Inspection required to confirm absence of flaws of critical size.
DET NORSKE VERITAS

Page 21
4.3.1 Type A Inspection zones where corrosion conditions may be different. For this
vessel a moderate coverage using corrosion mapping is appro-
Type A inspection applies in situations where there is a low priate, with a certain amount of coverage in each of the three
probability of degradation based on previous inspection his- zones. The coverage should be sufficient to allow a quantified
tory and / or CRA and if degradation is present it will tend to
be general or there is a high confidence that the most likely statistical assessment that can be used to demonstrate a high
areas for degradation can be identified. It is also intended to confidence in estimating the worst flaw. The accuracy and res-
provide a general screening for damage due to degradation olution of the inspection system need to be considered in the
mechanisms that might unknowingly be active. The purpose of same context. Note the shift in emphasis compared to Type A
this type of inspection is therefore primarily to confirm that – here one is using the information provided by the inspection
there is no degradation active. If degradation is found then fur- to say something about the worst degradation that might exist
ther steps are required to be taken. in the vessel (including the regions not inspected) where in the
Type A inspection the emphasis is on identifying the presence
Example of Type A (or confirming the absence) of degradation.
A vessel constructed in Duplex stainless steel and exposed to
produced water, oil and gas at moderate temperature. The most 4.3.3 Type C Inspection
likely mechanisms here would be chloride pitting of the shell Type C inspection applies when there is a reasonably high
and chloride SCC of the welds. The probability would typi- probability of degradation being present and/or degradation
cally be very low however, provided the chloride levels are not may be severe and/or degradation has no preferred locations.
excessive and the temperature is moderate. In this situation the This inspection will often apply at moderate/high coverage. Its
corrosion assessment would typically indicate that degrada- purpose is to give a high probability that any flaw with poten-
tion, if it does occur, will tend to be found at the bottom of the tial to threaten integrity is found directly. In the event that such
vessel (where there is contact with water). Provided exposure flaws are found, then further action is required in order to accu-
to water is similar, there are no other factors leading to prefer- rately size the flaws to allow integrity assessments to be carried
ential degradation. This means there is no need to do a high out. These more detailed, flaw specific inspections are outside
coverage inspection – a fairly small coverage can give confir- of the scope for the NII inspection. The purpose of NII is pri-
mation that degradation is not active. Its important however to marily to identify and locate these flaws.
ensure that the areas selected for coverage are likely to be rep-
resentative of the worst areas. If this is possible (based on the It is useful to illustrate, by way of examples, how the approach
findings of the corrosion assessment) then very low coverage to inspection on different vessels aligns to the above types.
may be acceptable 1). A key inspection performance require-
ment here is the ability to detect the presence of degradation, Example of Type C
even when it is in its early stages. A vessel constructed in carbon steel and in process conditions
for which the corrosion assessment identifies a potentially high
1) Note that there is a sound statistical foundation for allowing this type of corrosion rate that varies across the vessel with limited confi-
approach. A semi-quantitative Bayesian statistics approach can be used
to show that for a situation in which there is a high level of confidence in dence in ability to identify the worst areas. It is possible that
limited degradation, a lesser amount of inspection can still have signifi- corrosion might be severe over a small region with the remain-
cant influence on the "degree of assurance". der of the vessel being relatively unaffected. These conditions
mean that conventional statistical analysis may not apply. Here
4.3.2 Type B Inspection the inspection coverage must be high to ensure that the worst
Type B inspection applies when there is some degradation flaws are not missed. In this type of inspection coverage will
expected but it is not expected to be such as to threaten integ- often be more important than system accuracy.
rity in the medium term. Medium term is in this case taken as
being a period equivalent to at least two inspection intervals. 4.3.4 Selection of Inspection Type
This has been adopted as it allows for any missed flaws to be Guidance of the selection of a particular inspection type is
identified at the subsequent inspection, without threatening the given in Figure 4-2 below. This considers the likelihood,
integrity of the equipment. This inspection applies at a low/ extent and rate of degradation expected for a particular mech-
moderate coverage and its purpose is to provide sufficient anism (based on the corrosion risk assessment) to determine
information to allow quantified demonstration of the required the appropriate inspection type.
degree of assurance. If the results of the inspection do not
allow this then further action is taken. The inspection type categorisation is unique to a particular
degradation mechanism and may also vary from one location
Example of Type B
to another. It is important, therefore that it is re-evaluated for
A vessel constructed in carbon steel that is exposed to wet gas. each. For example, a vessel where the CRA has predicted both
Corrosion (pitting or more general) is expected but with a low general wall loss and localised SCC cracking mechanisms
corrosion rate. The corrosion assessment has defined three might be assessed as shown in Table 4-2.
Table 4-2 Example of Inspection Strategy Grading Process.

Location Damage Mechanism Inspection Type Comment
General wall loss Type A Low likelihood, general in extent and medium predicted
rate.
Above fluid level
Localised cracking Type A Low likelihood, localised but would be likely to occur at
welds, rate medium.
General wall loss Type B High likelihood, general in extent and medium pre-
dicted rate.
Below fluid level vessel walls
Localised cracking Type B Medium likelihood, localised but would be likely to
occur at welds, rate medium.
General wall loss Type B High likelihood, general in extent and medium pre-
dicted rate.
Nozzle connection
Localised cracking Type C High likelihood, localised but would be likely to occur
at welds, rate high.
DET NORSKE VERITAS

Page 22
Degradation Likelihood Degradation Extent Degradation Rate Inspection Type
Inspection High General High B
Medium B
Low A
Localised High C
(Clearly identifiable)
Medium B
Low B
Localised High C
(Random)
Medium C
Low C
Medium General High B
Medium A
Low A
Localised High C
Medium B
Low A
Localised High C
(Random)
Medium C
Low B
Low General High A
Medium A
Low A
Localised High B
Medium A
Low A
Localised High C
(Random)
Medium C
Low B
Figure 4-2
Selection of Inspection Type
The first stage of the assessment is to determine whether deg- Mechanism Likelihood
radation of the item is likely to occur. In making this judgment High
the assessor should consider the worst location of the item, and
take account of any previous inspection history or CRA assess- Previous inspection has shown the degradation mechanism to
ment. The following categories then apply: have taken place, or the CRA indicates that the mechanism
WILL take place during the remaining plant lifetime.
DET NORSKE VERITAS

Page 23
Medium and which should be subject to sample inspections. In the case

The CRA assessment has shown that the damage mechanism of sample inspection of a zone, it may be possible to use the
MAY take place during the equipment lifetime, but inspection results to make predictions about the condition of the un-
history to date has shown no evidence of degradation. inspected part of that zone, but the results may provide little or
no relevant information about the condition of other zones.
Low Examples of features which could be considered when divid-
No degradation expected or degradation is superficial. ing the vessel into zones include:
The next step is to consider the morphology of the damage — longitudinal welds
mechanism. For the purposes of the inspection categorisation, — circumferential welds
these can be defined as either — attachment welds
— nozzle welds
Degradation Extent — parent plate with little probability of corrosion/erosion
General — parent plate with medium probability of corrosion/erosion
— parent plate with the highest probability of corrosion/ero-
Covers corrosion or erosion, where the loss of wall thickness sion (within that particular vessel)
is uniform or varies slowly within the area under considera- — known corroded area in parent plate
tion. It also covers crack- like flaws that are numerous and — internal fixings with integrity impact.
closely spaced within the area considered.
Localised (Clearly identifiable) Note that these are examples only, and in practice the zones
should be determined by considering the detailed design, func-
This covers corrosion or erosion where the loss of wall thick- tion, operating conditions and history of the vessel. The under-
ness is localised or irregular within the area under considera- lying principle is that each individual zone should be
tion, pitting or localised cracking which is isolated and does “homogeneous” so that any given part of a zone is representa-
not merge with surrounding flaws. The principal feature is that tive of the rest of that zone in terms of likelihood of degrada-
the location of the cracking is well understood and predictable tion, type of possible degradation, tolerance to degradation and
– i.e. at the weld root, and inspection can be targeted in that type of inspection method(s) which can be applied. Inspection
location. method and the basis for sampling (if appropriate) can then be
Localised (Random) determined individually for each zone.
As above, but may occur anywhere with no particular prefer- Zones which are physically separate but otherwise similar can
ence. be considered as one, for the purpose of sampling e.g. in many
Finally, consideration is given to the rate of the degradation. cases no distinction need be made between spherical and cylin-
Again for simplicity, the assessor is required to categorise the drical shell regions.
rate under one of three headings: In order to simplify the process of defining the separate inspec-
tion zones, it is recommended that the three main aspects
Degradation Rate “design”, “operational” and “inspection history” should be
considered separately.
High
Design and manufacturing factors
The anticipated rate of degradation is such that failure of the
equipment or rejection based on inspection results can reason- The vessel is divided into different zones (categories) based on
ably be expected within the remaining equipment lifetime. loading conditions and tolerance to flaws. Examples include,
but are not limited to: shell plates, heads/dished ends, nozzles
Medium (set on, set through and forged nozzle designs might need to be
The anticipated degradation rate is such that it would be treated separately), inlets, man-ways, longitudinal welds, cir-
expected to be observable during the equipment lifetime, but cumferential welds, internal attachments, internal components.
would not be expected to threaten its integrity during that time. Separate parts might be included in the same zone, e.g. all noz-
Low zles might belong to the same zone, welds might be included
in the same zone regardless of welding process or geometry.
No degradation expected or degradation is superficial.
Operational factors
The resultant inspection type can then be read from the flow
chart. As with all aspects of the evaluation process, the result The vessel is divided into different zones reflecting the extent
thus obtained should be considered against engineering judge- to which different locations are known or expected to be
ment, and if necessary over-ruled, in which case a record of the affected by the operating and process conditions. Considera-
decision process should be maintained. tions include, but are not limited to: service fluids, inlet/outlet
locations, locations of mixed phases, high fluid flow rates and
4.4 Definition of Vessel Zones turbulence and impingement, vapour/condensation, bubbling/
cavitation, pressure/pressure cycling, loading, temperature,
4.4.1 Identification of Zones oxidizing atmosphere, aggressive abrasive content. This
Non-intrusive inspection methods have different capabilities requires detailed knowledge of the operational characteristics,
and limitations for different geometries and materials and can chemical nature of the fluids, metallurgy etc. When consider-
be heavily influenced by aspects such as degradation and flaw ing the zones corresponding to operational conditions, consid-
types. In addition, it is generally impractical to perform non- eration should be given to previous experience from that
intrusive inspection over the entire vessel surface. The vessel, experience from similar types of vessel operating under
approach suggested herein for the selection of methods is broadly similar conditions, and generic knowledge and experi-
therefore based on the concept of “zones” representing difference of how the vessel material behaves under the particular
ent combinations of likelihood of degradation, tolerance to operating conditions (process, temperature, etc.).
degradation and practicality of inspection. Factors such as Previous inspection factors
geometry, material, likelihood of degradation, type of degrada- The vessel is divided into different zones corresponding to the
tion and previous inspection results should all be considered effectiveness and results of previous inspections. Examples
when identifying these different zones. include, but are not limited to, regions where no previous in-
This then provides the basis for deciding which parts (zones) service inspections have been performed, regions subjected to
of the vessel should be inspected, by which inspection method, internal visual inspection, regions subject to ultrasonic thick-
DET NORSKE VERITAS

Page 24
ness measurements, regions reported to contain flaws / degra- for the following zones to be considered: A: Liquid zone, B:
dation etc. Interface zone, C: Vapour zone. Often there are specific issues
Note that further subdivisions of zones may be necessary, e.g. relating to any nozzles or connections to vessels, so these
regions covered by fire retardant lining, regions with limited should also be identified as individual zones, D: Inlet nozzle,
access to inspection surface due to adjacent pipework, etc. F: Outlet nozzle. This would normally be sufficient to assess
the vessel; however, if previous inspection records had identi-
4.4.2 Example fied a region of more advanced corrosion in a particular area of
The following example is used to illustrate the identification of the shell (for example within the vapour region caused by con-
zones corresponding to the different combinations. densation of vapour accelerating corrosion) than an additional
zone (designated “E” in Figure 4-3 below), which is a combi-
In a partially filled vessel the shell and weld submerged in and
out of the liquid and at the interface between liquid and vapour nation of shell, vapour region, previously detected flaws would
may corrode at different rates. This gives rise to a requirement be required.
C E
D (Inlet)
F (Outlet)
Zone Zone Identifier

Liquid phase A
Interface B
Vapour phase C
Inlet (liquid) D
Previously reported corrosion E
Outlet (mixed phases) F
Figure 4-3
Example of Vessel Zones
Each of the zones A to F would be considered individually for purposes of this recommended practice, localised loss of wall
suitability for inspection by non-intrusive means. thickness is also intended to include pitting.
4.5 Definition of Degradation Type Localised cracking.
The next stage is to identify the type of degradation expected This covers crack like flaws that are typically isolated and do
at each location. Degradation type should be defined according not merge with surrounding flaws. A fatigue crack initiated in
to its associated flaw morphology as this is what mostly influ- a region of stress concentration would typically be considered
ences the type of inspection method that will be applicable. In as localised cracking.
the majority of cases flaws can be categorised as one of the fol- Generalised cracking.
lowing, although judgement should always be applied as dif-
ferent mechanisms result in flaws with varying detectability This covers crack like flaws which are numerous and closely
using non-destructive methods. spaced in the region under consideration.
Generalised loss of wall thickness. In addition to the above major categories there are some less
commonly encountered flaw types including de-lamination
This covers corrosion or erosion where loss of wall thickness and blistering.
(LOWT) is uniform or varies slowly within the area under con-
sideration. The user may find it helpful to develop a matrix of features and
flaws to ensure that no combinations are overlooked. Table 4-3
Localised loss of wall thickness. gives an example matrix. Note that the list of features selected
This covers corrosion or erosion where the loss of wall is local- for inspection may not always include every feature on the ves-
ised or irregular within the area under consideration. For the sel. Depending on the coverage requirements, there may be
DET NORSKE VERITAS

Page 25
features that do not need to be included in the inspection. lihood of degradation is high, tolerance to further degradation
is low, and consequence of vessel failure is high. However less
Table 4-3 Example of matrix used to define feature and flaw expensive methods based on rapid (but reduced sensitivity)
combinations screening methods or sample inspections might be acceptable
Vessel Flaw type for other zones in this vessel, or for vessels where the conse-
Feature quence of failure is low. Some zones may require no inspec-
Localised General- Localised Generalised
LOWT ised LOWT cracking cracking tion.
Set on By way of example, if for a particular zone, the design is such
Nozzle N1 that tolerance to flaws is high, the likelihood of degradation is
Shell welds very low, and previous inspections have not detected any deg-
radation, then there may be a strong justification for deciding
Shell plate
that the inspection has a low criticality, and therefore perform-
Saddle ing only minimal inspection (or no inspection) of that zone. If
plate on the other hand flaw tolerance is low and the likelihood of
degradation is high, the inspection is critical, and obviously
4.6 Inspection Effectiveness there is an incentive to do a much more comprehensive inspec-
Having identified the different zones within the vessel, and tion in that particular zone.
defined the inspection strategy (from Sec.4.3) for each zone, A decision process based on a flowchart approach is presented
the inspection criticality should be determined on a zone by (Figure 4-4), which provides guidance on determining the
zone basis. The level of effectiveness which is appropriate for required inspection effectiveness, taking into account for each
a particular zone will depend on the likelihood of degradation, zone whether degradation has previously been detected, the
previous inspection results, tolerance to degradation, and the likelihood of degradation (using the principles of inspection
consequence of vessel failure. Clearly there is likely to be a grading), current tolerance to degradation and consequence of
strong incentive to perform a rigorous inspection (e.g. full cov- vessel failure. The required inspection effectiveness for a par-
erage using a sensitive inspection method which has a high ticular location is read off from the appropriate strand in the
probability of detecting flaws) for a part of a vessel where like- flowchart
Minimum inspection effectiveness

Inspection Grade Current tolerance to Consequence of Failure
( IP12 / IP13) degradation Low Medium High
Grade 0 Low High High High
Grade 1 Medium High High High
Grade 2 High Medium High High
Grade 3
Low High High High
Medium Medium High High
High Medium Medium High
Low Medium High High
Medium Low Medium High
High Low Medium Medium
Low Medium Medium High
Medium Low Medium Medium
High Low Low Medium
Figure 4-4
Inspection Effectiveness Flowchart
Inspection Grade Equipment may be allocated a Grade between 0 and 3 based on

The inspection grading system used here is taken from the the number of previous inspections and the rate and predicta-
principles of grading contained within the Energy Institute bility of deterioration based on the Energy Institute guidance
(formerly the Institute of Petroleum (IP)) model codes 12 and on the examination of pressure vessels and piping.
13. [10], [11].
DET NORSKE VERITAS

Page 26
Grade 0 Medium and Low categories used here. Alternatively, in the

Items for which: absence of other criteria, the following definitions should apply:
1) there is no historical evidence to support the judgement, or — Health, Safety and Environmental Consequences
2) the rate of deterioration is high, or High
3) the rate of deterioration is unpredictable. One or more fatalities or serious injuries requiring hospital
treatment, or major release of hazardous material or pollu-
Items newly installed in a system commence at Grade 0. tion
Grade 1
Medium
Items which have: Injury requiring hospital treatment, or release of hazardous
1) at least one previous examination at Grade 0; and material or pollution but with no significant effect off site.
2) which show a moderate rate of deterioration which is pre- Low
dictable. At most, minor injury with full recovery, or minimal re-
Grade 2 lease of hazardous material or pollution.
Items which have: — Business Interruption and Cost Consequences
In the absence of other criteria, the following definitions
1) at least one previous examination at either Grade 0 or could apply:
Grade 1; and
2) which show a low rate of deterioration which is predicta- High
ble. Major shutdown / turnaround or high costs of repair / re-
placement
Grade 3
Items which have either: Medium
Several days shutdown / turnaround or significant costs of
1) at least one examination at Grade 0 and one examination repair / replacement
at either Grade 1 or 2, and which show a low rate of dete-
rioration which is predictable; or Low
2) a negligible rate of deterioration in a stable service envi- Less than one day shutdown / turnaround or low costs of
ronment. repair / replacement
Current Tolerance to Degradation 4.7 Required Inspection Effectiveness
The current tolerance to degradation should be considered in The flowchart provides a “score” which represents the
terms of whether it is low, medium or high. required inspection effectiveness for the zone under consider-
In the absence of other criteria, the following definitions could ation. Effectiveness is here defined as a qualitative measure of
apply: the probability of detecting flaws, taking coverage into
account. Assuming a uniform flaw distribution,
Low Effectiveness = f(POD, ×Coverage).
The known or predicted degradation and rate thereof in the Three effectiveness categories are used, (high, medium and
zone under consideration are such that failure of the vessel (if low), these being defined by comparison to the effectiveness
no remedial action is taken) can reasonably be expected within for visual inspection. High implies a higher effectiveness than
the remaining plant lifetime. visual inspection, medium implies a broadly similar effective-
Medium ness and low implies a lower effectiveness.
The known or predicted degradation and rate thereof in the These requirements should be interpreted as follows.
zone under consideration are such as to be observable during Low: Spot checks. The inspection is performed at a number of
the plant lifetime but would not be expected to threaten the discrete locations within the zone, for example manual ultra-
vessel during this period or require remedial action. sonic thickness measurements at 500 mm intervals might be
High appropriate for monitoring general loss of wall thickness, or at
There is no degradation expected or degradation is superficial. a reduced interval for more localised corrosion/erosion.
Consequence of Failure Medium: 100% inspection of the planned inspection area using
a method which has medium efficiency (>70% probability of
The consequences of vessel failure must be considered when detection), or sample inspection using a method which has
determining what level of inspection is appropriate for each high efficiency (>90% probability of detection).
zone, ref. [12]. For example two vessels of similar design and
operating regime will experience similar degradation, however High: 100% inspection of the planned inspection area using a
if failure of one vessel resulted in the closure of the plant for an method which has high efficiency, i.e. expected to have a prob-
extended period whilst repairs are undertaken, whereas for the ability of detection exceeding 90% for the degradation or flaws
other vessel production could continue comparatively unaf- of concern.
fected, it is clear that the inspection of the former vessel would In this context, the actual area of the zone inspected should be
be specified more rigorously than that of the second. sufficient to allow meaningful extrapolation of the results to
In order to simplify the determination of consequence, it is use- the un-inspected parts of that zone.
ful to consider the health, safety and environmental conse- Note: The inspection requirements for a zone may need to be
quences separately from the purely business and cost increased if flaws or degradation are detected during the
consequences. Having done so, the overall consequence ranking inspection. For example if spot checks reveal greater loss of
should be taken as whichever of these is the higher. Conse- wall thickness than expected, or localised thinning, increased
quence rankings are generally determined as a part of the risk coverage is recommended to properly assess the extent of the
based methodology, and these can be mapped to the High, thinning. If cracks are detected during a sample inspection,
DET NORSKE VERITAS

Page 27
100% inspection of that zone (and other susceptible zones) fore important to consider the applicability of a particular
may be required. method on a zone by zone basis. The overall objective should
be to ensure that the integrity of each zone meets the minimum
4.8 Coverage level needed to ensure the continued integrity required for the
Before proceeding with selecting inspection methods, it is vessel as a whole.
important to establish the nature of inspection coverage
required. This, together with several other factors, has an influ- 4.9.2 Inspection Method Selection Flowcharts
ence on what features should be inspected, which in turn, influ- Different NDT methods have differing capabilities, strengths
ences what inspection methods can be selected. The decision and weaknesses. In order to facilitate the selection of the right
process does not seek to guide the user to determining the exact method for the application, a number of selection flow charts
coverage for each region of the vessel, this forms part of the have been developed for the most commonly used methods in
detailed inspection planning activity and is discussed in more the context of non-intrusive inspection.
detail in Sec.4.4. The charts are intended to assist the inspection manager to
A high level recommendation for coverage is provided below; determine the efficiency of a particular method in a given
the intent being that this is used to establish the framework into application, as defined by comparison to IVI, for a number of
which any detailed decisions on coverage shall conform. A commonly used inspection methods. The method capability is
coverage selectiveness requirement is defined, this being such classified in comparison with that of internal visual inspection,
as to ensure coverage consistent with the user’s ability to pre- taking into consideration the degradation type. Each method is
dict the sites of potential flaws of concern and hence direct given two capability scores, one for capability in detecting
inspection accordingly. This is determined according to the flaws, and the other for the method’s ability to provide quanti-
response selected for the Confidence in ability to predict types tative information regarding flaw size or wall thickness. For
and locations of degradation question of decision guidance both factors, three categories are defined, namely. When
process, Figure 3-2. selecting a method appropriate for the particular application,
A High confidence will justify a minimum of Targeted cover- the inspection manager should use the appropriate capability
age. Medium confidence will require a minimum of Targeted score (POD / sizing) to compare against the required minimum
plus Exploratory coverage and Low confidence will require inspection effectiveness, from Figure 4-4. For example, where
Global coverage. The coverage categories are as defined a Type A inspection is planned, the dominant requirement is
below: the ability to detect degradation, thus the POD portion of the
capability score would be most appropriate; whereas with a
Targeted type C inspection, the requirement is to determine whether
Inspection can be restricted to the sites where potential degra- degradation is within expected limits, thus a method with a
dation has been predicted. good ability to provide quantitative information is required and
the sizing portion of the capability rating would be more appro-
Targeted plus exploratory priate.
Sites where potential degradation has been predicted must be
inspected. In addition several further areas where it is not pos- High
sible to rule out the presence of flaws must also be inspected.
POD: The method has, in the conditions under consideration,
Global a higher probability of detecting the flaw type than
The entire area/feature of the vessel under consideration does IVI.
should be inspected. Sizing: The method is able to provide accurate, quantitative
information regarding flaw size or wall thickness.
As stated previously, the actual area inspected should be suffi-
cient to allow meaningful extrapolation to remaining areas of
the zone, which will be a function of the anticipated type of Medium
degradation. i.e. where general loss of wall thickness is
expected, it is possible to carry out spot checks, which will rap- POD: The method has, in the conditions under consideration,
idly confirm the current status, however where localised wall a probability of detecting the flaw type broadly similar
loss is expected, a more rigorous scan is required to obtain the to that of IVI.
same degree of confidence. Sizing: The method is able to provide some semi-quantitative
or comparative information regarding flaw size or wall
4.9 Selection of Inspection Method loss.
4.9.1 Inspection Capability
Although routine methods and generic procedures may be Low
appropriate for some zones (e.g. ultrasonic 0 degree wall thick-
ness measurement, ultrasonic examination of butt welds) in POD: The method has, in the conditions under consideration,
other instances (e.g. where access to inspection surfaces is a lower probability of detecting the flaw type than does
restricted; where complicated flaw orientations/morphologies IVI.
are sought; difficult material or geometry) it may be necessary Sizing: The method is able to provide only limited, generally
to apply specialised methods. The purpose of the inspection qualitative information regarding flaw size or wall
plan is to ensure that the correct methods are used in each con- thickness.
text (location, degradation type and anticipated extent) in order
to ensure the equipment integrity between inspections. A gen- In order to select a method meeting the minimum efficiency
eral description of the main inspection methods available for requirement, the user consults each of the flow charts, consid-
non-intrusive inspection of vessels, and their associated capa- ering each factor sequentially in the following order;
bilities is provided in Appendix A. A variety of additional Vessel feature → Flaw type → Surface → Temperature →
sources of information on methods exist, see for example Ref- Thickness → Access
erences [13], [14], [15], [16], [17], and internet based software The flow charts allow continuation in cases where the method
tools such as the HOIS Interactive Knowledge Base [18]. is applicable for the factor considered and terminate where it is
The nature of the degradation expected in each zone is likely not. In each case of termination, recommendations are made
to vary, as is the required inspection effectiveness. It is there- for alternatives that may have a better chance of success.
DET NORSKE VERITAS

Page 28
Branches that do not terminate prematurely end with the effi- — seam and girth butt welds
ciency rating for the method under the combination of condi- — cylindrical and spherical shells
tions specified. — set on and set through nozzles
— nozzle compensating plate (not shown on vessel but
For each method a flow chart is presented which takes due con- included)
sideration of vessel features, flaw type, external surface coat- — saddle plate (external support)
ing, temperature, wall thickness and access requirements in — weir plate (internal)
order to determine the likely method efficiency. A brief — lifting lug (external)
description of each of these factors is given below. — external stiffener.
Note that several of the features are external to the vessel
Features considered (compensating plate, saddle plate, lifting lug and external stiff-
The features for which guidance on the selection of an inspec- ener). The inspection method for these is recommended not in
tion method can be provided are limited to those on the base- terms of assessing their external condition but in terms of the
line vessel shown in Figure 4-5 and include: condition of the vessel interior adjacent to the feature.
Figure 4-5
Baseline Vessel Design (showing features considered).
Flaw types considered smooth and its body continuous so as to allow adequate ultra-
At this stage only four flaw types are considered in making rec- sonic transmission through conventional probes.
ommendations for inspection method. These are: Temperature
— generalised loss of wall thickness (e.g. corrosion/erosion) Temperature refers to the external surface temperature of the
— localised corrosion/erosion (e.g. pitting) feature under inspection, and therefore the temperature to
— multiple cracking (e.g. stress corrosion cracking) which the inspection equipment may be exposed.
— isolated cracking (e.g. fatigue cracking in welds). Wall thickness
External surface coating Wall thickness refers to the wall thickness of the feature to be
inspected (either vessel wall thickness or feature thickness as
The user selects from one of: applicable).
— bare metal Access requirements
— paint The access requirements are specified as the minimum length
— flame or thermal sprayed aluminium required (i.e. dimension parallel to the surface being inspected)
— zinc × the minimum height required (i.e. dimension perpendicular
— thermal insulation to the surface being inspected).
— passive fire protection.
4.9.3 Inspection Method Selection Criteria
For bare metal it is assumed, in all cases except for when con- The user works through the flow charts until a method is found
sidering magnetic flux exclusion, that the exposed surface is having at least the effectiveness rating determined in Sec.4.7.
sufficiently smooth and continuous for adequate ultrasonic The order in which the flow charts are consulted or, indeed,
transmission through conventional probes. Bare metal in a cor- whether all of the charts should be consulted in every instance
rosive environment will often need some cleaning up before cannot be prescribed. There will, for example, be cases in
being suitable for ultrasonic inspection. which the user has sufficient knowledge/experience with the
For paint it is assumed, in all cases except for when consider- inspection requirement under consideration to confidently
ing magnetic flux exclusion, that the coating is less than 2 mm specify an inspection method that will be suitable. In such
thick and in good condition, i.e. its surface is sufficiently cases only the flow chart for the selected method need be con-
DET NORSKE VERITAS

Page 29
sulted, with this serving as confirmation of the selection. There sideration.

will also be cases where the user can not easily identify suita-
ble methods and in these instances it may be necessary to con- Having identified a method which satisfies the efficiency cri-
sult each of the flow charts. It is possible that, for the terion, it is worth considering whether the chosen method has
combination of factors considered, several methods will turn the ability to size any flaws found during inspection. The out-
out to have more than the minimum efficiency required. Whilst come here might not affect the decision on whether or not to
under ideal circumstances the best method would be selected use NII for the inspection but allows early identification of the
for each zone, in such cases the decision on which of the approach that may have to be adopted if flaws are found.
acceptable methods to use will be tempered by non-technical
considerations such as cost and practicality. The process of method selection is repeated for each degrada-
tion type anticipated for the zone under consideration, and then
In addition, it is important to consider compatibility with pre- for each zone until a compete inspection plan is established for
vious inspections. In order to be able to monitor progression of the equipment.
any degradation, there is clearly a need for continuity between
inspections. The introduction of different inspection methods It is worth noting that, in following the high level decision
may complicate the comparison of results (particularly when guidance flowchart, Figure 3-2, the effectiveness of the previ-
changing from an invasive to a non-invasive inspection ous inspection has a direct effect on the requirements for the
regime). The impact of any such changes can be minimised inspection being planned and may determine whether NII is
with due care in specifying the inspection reporting criteria and recommended at all. Consequently, in planning for the current
format. inspection the user should consider the potential implications
The intent of the flow charts is to allow a rapid assessment of of the effectiveness likely to be achieved on the requirements
method capability in a particular set of circumstances. In meet- for future inspections.
ing this objective they are necessarily simple and can not con-
sider in depth all situations that might arise. There may be 4.9.4 Statistical Methods
circumstances in which capability is significantly degraded by In some cases, there is a requirement for the inspection plan-
factors not considered in the flow charts, e.g. unusual flaw ori- ning process to involve a more quantified assessment of the
entation, obscuration by other flaws etc. Consequently the user risks and likelihood of failure occurring, and the impact of
should regard the results of the flow charts as a guide only and inspection on mitigating those risks. In particular, in some
should review the results obtained for each case in the context industries, there is a requirement to demonstrate that the risk
of any factors that might play a role in degrading capability in exposure remains below some threshold value, typically quan-
practice. Furthermore, the flowcharts reflect the typical capa- tified in units of probability e.g. number of failures per 106
bility of the method described and do not address improve- years. In such cases, it is necessary to consider the statistical
ments in performance possible through the use of specialised probability that the inspections performed will have found any
or custom equipment and advanced techniques. degradation present, and that the most significant degradation
If no method meeting the minimum requirement is available has been accurately assessed. Reference 7 provides an insight
then it is recommended that the risk associated with not into how inspection reduces the risk of failure, where risk of
inspecting, or a reduced efficiency inspection of the feature for failure is a combination of likelihood of failure and conse-
the flaw type under consideration be reviewed (see Figure quence of failure (although Reference 7 specifically addresses
4-1). In order to show that the risk is acceptable, the review ultrasonic inspection, the principles of reducing risk by inspec-
must be comprehensive and fully documented. The review tion apply to NDT methods in general). Inspection can only
should consider as a minimum: reduce likelihood of failure, not consequence. For a vessel
where consequence of failure is high, inspection should have
— the probability of defects with potential to cause failure the potential to maintain likelihood of failure at a low level. If
existing in the feature under consideration the predicted likelihood of failure (without inspection) is high,
— the findings of previous inspections covering the feature then there needs to be high confidence in the capability and
under consideration reliability of the inspection method to detect (and correctly
— the type and extent of inspection in other regions where sentence) flaws or degradation of concern.
degradation can be expected to be similar to that for the
feature under consideration Further guidance on the use of statistical methods to plan
— the potential consequences of failure modes associated inspection activities and evaluate inspection results is given in
with the types of flaw expected in the feature under con- Appendix B of this document.
DET NORSKE VERITAS

Page 30
UT Thickness Gauge
Feature Defect Type Surface Temperature (°C) Thickness (mm ) Capability

N 1, 2
Is the weld
Butt Weld ground flat ?
POD H
Generalised LOWT Bare metal < 60 °C All sizes
Y
Y
Sizing H
Localised LOWT Paint HT equipment
Cylindrical shell 60 – 500 °C available ?
Multiple cracking Flame spray Al N
2
Spherical shell Localised cracking Zinc > 500 °C 4
Thermal insulation
Piping 3
Passive fire protection
< 4mm 2
Bare metal < 60 °C POD M
Y > 4mm
Paint HT equipment Sizing M
60 – 500 °C
Flame spray Al available?
N
Zinc > 500 °C 4

Thermal insulation
3
DET NORSKE VERITAS
POD M
Generalised LOWT Bare metal < 60 °C All sizes
Sizing M
Set-on nozzle Localised LOWT Paint HT equipment
Y
60 – 500 °C
Multiple cracking Flame spray Al available ? N
Set-through nozzle
2
Localised cracking Zinc > 500 °C 4
Thermal insulation
3
Compensating pad
Saddle plate < 4mm 2

Bare metal < 60 °C POD L
Lifting lug Not Suitable Paint Y > 4mm
HT equipment Sizing L
60 – 500 °C
External stiffener Flame spray Al available? N
Zinc > 500 °C 4

Weir plate Thermal insulation
3
Notes
1 Grind weld flat
2 Consider alternative technique
3 Consider removing coating
4 Equipment must be taken off- line for inspection
Typical probe size down to 20 mm x 20 mm x 20 mm

Manual 0° UT Mapping
Feature Defect Type Surface Temperature (°C ) Thickness (mm ) Capability
Is the weld N
1, 2
Y
< 2 mm 2
Generalised LOWT Bare metal < 60 °C POD M
Y > 2 mm
Localised LOWT Paint HT equipment Sizing M
Cylindrical shell 60 – 200 °C
Multiple cracking Flame spray Al available ?
N
2
Spherical shell Localised cracking Zinc > 200 °C 2, 4
Thermal insulation
Piping 3
< 2 mm 2
Generalised LOWT Bare metal < 60 °C POD L
Set-on nozzle Y > 2 mm
Localised LOWT Paint HT equipment Sizing L
60 – 200 °C available ?
Set-through nozzle Multiple cracking Flame spray Al N
DET NORSKE VERITAS
2
Localised cracking Zinc > 200 °C 2, 4
Thermal insulation
3
Saddle plate Passive fire protection
Lifting lug
External stiffener Not Suitable
Compensating pad

Weir plate
Notes
1 Grind weld flat
4 Consider taking equipment off line for inspection
Typical equipment size :

Mechanical scanner : from 300 mm x 300 mm x 50 mm high
Hand-held and camera : from 20 mm x 20 mm x 20 mm plus camera line of sight .
Page 31
Page 32
UT Corrosion Mapping
Is the weld N
1, 2
Y
< 2 mm 2
Generalised LOWT Bare metal < 60 °C POD H
Y > 2 mm
Localised LOWT Paint HT equipment Sizing H
Cylindrical shell 60 – 200 °C
Multiple cracking Flame spray Al available ?
N
2
Spherical shell Localised cracking Zinc > 200 °C 2, 4
Thermal insulation
Piping 3
< 2 mm 2
Localised LOWT Paint HT equipment Sizing M
Set-through nozzle Multiple cracking Flame spray Al N
DET NORSKE VERITAS
2
Thermal insulation
3
Saddle plate Passive fire protection
Lifting lug
External stiffener Not Suitable
Compensating pad
Weir plate
Notes
1 Grind weld flat
Typical equipment size :

Mechanical scanner : from 300 mm x 300 mm x 50 mm high
Hand-held and camera : from 20 mm x 20 mm x 20 mm plus camera line of sight .
UT Angled Pulse Echo
Feature Defect Type Surface Temperature (°C) Thickness (mm) Capability
Butt Weld < 8 mm 1

Generalised LOWT 1 Bare metal < 60 °C POD L
Cylindrical shell Y > 8 mm
Localised LOWT Paint HT equipment Sizing None
60 – 200 °C available?
Spherical shell Multiple cracking Flame spray Al N Some capability but other techniques
preferred.
Piping Thermal insulation
2
< 8 mm 4
Bare metal < 60 °C POD M
Y > 8 mm
Paint HT equipment Sizing M
60 – 200 °C
Flame spray Al available ?
N
Zinc > 200 °C 3, 1
Thermal insulation
2
DET NORSKE VERITAS
< 8 mm 1
Generalised LOWT 1 Bare metal < 60 °C POD L
Localised LOWT Paint HT equipment Sizing None
60 – 200 °C
Multiple cracking Flame spray Al available? N
Set-through nozzle Some capability but other techniques
preferred.
Thermal insulation
Compensating pad 2

Saddle plate
< 8 mm 4
Lifting lug Not Suitable Bare metal < 60 °C POD L
Paint Y > 8 mm
External stiffener HT equipment Sizing L
60 – 200 °C
Flame spray Al available ? N
Weir plate
Zinc > 200 °C 3, 1
Thermal insulation
2
Notes
4 Consider internal inspection
Typical probe size from 20 mm x 20 mm x 20 mm
Page 33
Page 34
Time of Flight Diffraction
< 8 mm 1
Butt Weld Generalised LOWT Bare metal < 60 °C POD H
Y > 8 mm
Cylindrical shell Localised LOWT Paint HT equipment Sizing H
Spherical shell Multiple cracking Flame spray Al N
Piping
Thermal insulation
1, 2
< 8 mm 1
Bare metal < 60 °C POD H
Y > 8 mm
Paint HT equipment Sizing H
60 – 200 °C
Flame spray Al available ? N TOFD is not effective for detection
and sizing of shallow cracks (<2 mm)
Zinc > 200 °C 1, 3 positioned near the front wall
Thermal insulation
1, 2
DET NORSKE VERITAS
< 8 mm 1
1 > 8 mm
Set-on nozzle Localised LOWT Paint Y
60 – 200 °C
HT equipment Sizing M
Multiple cracking Flame spray Al available ? TOFD is not effective for detection
Set-through nozzle N
and sizing of shallow cracks (<2 mm)
Localised cracking Zinc > 200 °C 1, 3 positioned near the front wall
Thermal insulation
Compensating pad 1, 2
Saddle plate
Lifting lug Not Suitable

External stiffener
Weir plate
Notes
3 Consider taking off line for inspection
TOFD requires two probes, one either side of the inspection volume.
Separation is dependant on the wall thickness
Typical probe size is 20 mm x 20 mm x 50 mm high
Medium Range UT (LORUS )
Butt Weld 1
Generalised LOWT 1 Bare metal < 60 °C < 4 mm
Cylindrical shell Localised LOWT Paint HT equipment Y POD L
60 – 250 °C 4 - 50 mm
Spherical shell Multiple cracking Flame spray Al available ? N Sizing None
Localised cracking Zinc > 250 °C > 50 mm Other techniques preferred .
Piping (general)
Y
Thermal insulation Access at
2 3 1
Passive fire protection edges ?
N
Piping (restricted )
1
Compensating pad Generalised LOWT 1 Bare metal < 60 °C < 4 mm
Localised LOWT Paint HT equipment Y POD L
External stiffener 60 – 250 °C 4 - 50 mm
Multiple cracking Flame spray Al available ? N Sizing None
Saddle plate Localised cracking > 250 °C > 50 mm
Zinc
Lifting lug Y
Thermal insulation
DET NORSKE VERITAS
Access at
edges?
2 3 1
N
1
Generalised LOWT 1 Bare metal < 60 °C < 4 mm
Set-on nozzle Localised LOWT Paint Y POD L
HT equipment
60 – 250 °C 4 - 50 mm
Set-through nozzle Multiple cracking Flame spray Al available ? N Sizing None
Localised cracking Zinc > 250 °C > 50 mm

Y
Weir plate Not Suitable Thermal insulation Access at
edges ?
2 3 1
N
Notes
1 Consider alternative technique .
3 Consider taking equipment off -line for inspection
Typical probe sizes from 50 mm x 50 mm x 50 mm
LORUS is particularly suited to locations with limited access , and pipe supports .
A range of up to 1m either side of the probe can be inspected in one pass .
Page 35
Page 36
Medium Range UT (CHIME )
Butt Weld
< 60 °C < 4 mm 1
Bare metal Y
Cylindrical shell HT equipment POD H
Paint 60 – 250 °C 4 – 50 mm
available ? Sizing L
Spherical shell Generalised LOWT Flame spray Al N
Other techniques preferred for
Localised LOWT > 250 °C > 50 mm 1 locations without access restrictions .
Piping Zinc
Multiple cracking Thermal insulation 1, 3
External stiffener Localised cracking 1, 2
Saddle plate
Lifting lug
< 60 °C < 4 mm 1
Bare metal Y
HT equipment POD M
Paint 60 – 250 °C 4 – 50 mm
available ?
Sizing L
Flame spray Al N
> 250 °C > 50 mm 1
DET NORSKE VERITAS
Zinc
Set-on nozzle Thermal insulation 1, 3
1, 2
Set-through nozzle
Not Suitable
Compensating pad
< 60 °C < 4 mm 1
Weir plate Bare metal Y
HT equipment POD M
Paint 60 – 250 °C 4 – 50 mm
available ?
Sizing L
Flame spray Al N
> 250 °C > 50 mm Defect perpendicular to UT beam .
Zinc
Thermal insulation 1, 3 POD NONE
1, 2
Passive fire protection Sizing NONE
Defect parallel to UT beam .
Notes
Typical probe sizes from 25 mm x 100 mm x 100 mm high
CHIME requires two probes positioned up to 1m apart .

Long Range UT (LRUT – Guided Wave )
Generalised LOWT 1 Bare metal POD L

< 125 °C < 50 mm
Localised LOWT Paint Sizing None
Piping > 50 mm
Multiple cracking Flame spray Al
1
Y
Butt Weld 1, 2
Passive fire protection edges ? N
Cylindrical shell
Spherical shell
Bare metal POD L
Compensating pad Not Suitable < 125 °C < 50 mm
Paint Sizing None
External stiffener > 50 mm Other techniques preferred .
Flame spray Al
Saddle plate Zinc > 125 °C 1, 3
1
Lifting lug Y
DET NORSKE VERITAS

edges ?
1, 2
Weir plate Passive fire protection N

Notes
Typically requires a ring of probes length 500 mm height 50 mm
Page 37
Page 38
Magnetic Flux Exclusion
Bare metal POD M

< 80 °C < 15 mm
Paint Sizing L
Generalised LOWT 1
Cylindrical shell Flame spray Al > 15 mm 1
Localised LOWT > 80 °C 1, 3
Zinc
Spherical shell Multiple cracking
1 Thermal insulation
Localised cracking 1, 2
Butt Weld
Set - on nozzle
Set - through nozzle
Compensating pad
Not Suitable
Saddle plate
Lifting lug
DET NORSKE VERITAS
External stiffener
Weir plate
Notes
Typical scanner size 300 mm x 200 mm x 200 mm high

Pulsed Eddy Current
< 70 °C < 6 mm 1
Bare metal Y POD H
HT equipment
Paint 70 - 500 °C 6 - 65 mm
Cylindrical shell Generalised LOWT
available ? Sizing M
N
Flame spray Al > 500 °C > 65 mm 1
Spherical shell Localised LOWT
Zinc
Multiple cracking Y
Piping Passive fire protection 1, 3
Localised cracking 1 Thickness
Thermal insulation 1, 2
< 100mm ? N
Butt Weld
Set-on nozzle
< 70 °C < 6 mm 1
Set-through nozzle Bare metal Y POD M
HT equipment
Paint 70 - 500 °C available ?
6 - 65 mm
Compensating pad Sizing L
N
Not Suitable Flame spray Al > 500 °C > 65 mm 1 Limit of detectability
˜ 20 mm x 20 mm
Saddle plate Zinc
Y
Passive fire protection 1, 3
DET NORSKE VERITAS
Lifting lug Thickness

Thermal insulation 1, 2
< 100 mm ? N
External stiffener
Weir plate

Notes
3 Consider taking equipment off line for inspection .
Particularly suitable for vessels and pipes with lagging or insulation ,

or thin metallic cladding made from aluminium , stainless steel or low alloy steel
Only suitable for use on low alloy steels .
Typical probe size 200 mm x 200 mm x 100 mm high , although specialist probes
Available down to 20 mm x 20 mm x 5 mm high
Page 39
Page 40
Saturation Low Frequency Eddy Current (SLOFEC )
Bare metal POD H

< 120 °C < 35 mm
Paint Sizing M
Cylindrical shell Generalised LOWT 1
Flame spray Al > 35 mm 1
Localised LOWT > 120 °C 1, 3
Spherical shell Zinc
Multiple cracking
1 Thermal insulation
Piping Localised cracking 1, 2
Butt Weld
Set-on nozzle
Set-through nozzle
Compensating pad
Not Suitable
Saddle plate
DET NORSKE VERITAS
Lifting lug
External stiffener
Weir plate
Notes
3 Consider taking equipment off line
Multi sensor scanner typically 150 mm x 150 mm x 200 mm

Passive Thermography
Bare metal
POD L
Paint < 15 mm
Non-contact Sizing None
Cylindrical shell Flame spray Al > 15 mm
Generalised LOWT
Zinc
Spherical shell Localised LOWT 1 1
Multiple cracking
Set-on nozzle Thermal insulation 1, 2
Localised cracking 1
Set-through nozzle
Butt Weld
Saddle plate
Lifting lug
DET NORSKE VERITAS
Not Suitable
External stiffener
Compensating pad
Weir plate

Notes
For use as an NII technique requires a process transient.
Page 41
Generally used to monitor insulation effectiveness .
Page 42
Radiography

Bare metal
POD H
Paint < 60 °C < 85 mm
Generalised LOWT Sizing H
Tangential thickness
Flame spray Al
Localised LOWT Special measures Use C60 or Betatron
Piping Zinc > 60 °C > 85 mm
Multiple cracking required source
Thermal insulation
Localised cracking
Bare metal
POD H
Paint < 60 °C < 50 mm
Sizing M
Wall thickness
Flame spray Al
Special measures Use C60 or Betatron
Zinc > 60 °C > 50 mm
required source
Thermal insulation
Bare metal
DET NORSKE VERITAS
POD M
Paint < 60 °C < 50 mm
Sizing L
Flame spray Al
Zinc > 60 °C > 50 mm
required source
Thermal insulation
Spherical shell
Cylindrical shell
Butt Weld Bare metal

POD M
Set-on nozzle Paint < 60 °C < 50 mm
Generalised LOWT 1 Sizing M
Flame spray Al Wall thickness
Set-through nozzle Localised LOWT Special measures Use C60 or Betatron
Zinc > 60 °C required
> 50 mm source
Compensating pad Multiple cracking
Localised cracking Thermal insulation
Lifting lug Passive fire protection
External stiffener
Bare metal
Saddle plate POD L
Paint < 60 °C < 50 mm
Sizing L
Flame spray Al Wall thickness
Zinc > 60 °C > 50 mm
required source
Generalised LOWT Thermal insulation

Localised LOWT Passive fire protection
Weir plate Notes
Multiple cracking
2 1 Consider thickness gauge or pulse echo UT
Localised cracking
2 Consider internal visual inspection
Requires access to both sides of equipment.

Typically requires 100 mm x 100 mm and 250 mm height for source.
Radiography not generally used for non-intrusive inspection of vessels due to long exposure
times required.
Page 43
4.10 Preparation of Work-pack — inspection procedures and method sheets for each method
The work-pack is a comprehensive package (electronic or — inspection procedures must normally be written or
paper based) of all relevant documentation necessary to per- approved by a person certified to Level 3/III in NDT
form the inspection. It is typically prepared by the Inspection — details of the standard and extent of surface prepara-
Body (organisation which will manage the performance of the tion
inspection), in conjunction with the owner in advance of the — specification of the temperature at which the inspec-
inspection. The work-pack will be based on the requirements tions are to be performed together with the operating
of the document vessel inspection plan which defines the over- conditions (if plant is live)
all inspection strategy – see Sec.4.3. (The vessel inspection
plan may be in the form of a specification which identifies the — required inspection coverage
zones to be inspected, the inspection methods and the coverage — reporting criteria and format (possibly in the form of a
required but will not include detailed procedures). The Inspec- reporting template)
tion Body should review this plan and compile (as a work- — acceptance reporting criteria and format
pack) all of the documents required for the inspection. Collab-
oration with other members of the NII planning team (see — well-defined thresholds or dimensions above which
Sec.4.2) may be useful. indications (interpreted as flaws) are recorded
Contents of the work-pack should include the following: — wherever possible acceptance criteria should be
defined
Equipment Profile — the reporting criteria and format should be specified in
Details of the equipment design and operation and previous detail. The format should facilitate repeatability of the
inspection history, as described in Sec.2.2, and including: inspection and facilitate comparison between past,
current and future inspection results
— identity and design
— type of vessel and function. — recording criteria
— operation and service details
— detailed drawings — when it is inappropriate to define a recording level
— modifications and repairs (e.g. corrosion mapping, thickness measurements)
— previous inspection results then any condition of particular interest (e.g. evidence
— general experience of wall thinning) should be defined.
— complementary information Method Statement
— accessibility
— safety limitations The Inspection Body may decide to produce a Method State-
— viable degradation mechanisms ment which summarises the inspection strategy and incorpo-
— anticipated degradation rates and extent rates or refers to the information described above. This will
help to make the work-pack a coherent, stand-alone document.
NII Decision Record In addition, this document could account for any areas where a
Details of the decision confirming that NII is appropriate for departure from the recommended Inspection Plan is unavoida-
the equipment (Sec.3.4) including: ble and describe and justify the remedial action. Other records
which should be included are:
— statement of any changes occurring in process that may
affect the nature or rate of degradation — training, qualification and certification requirements for
— inspection reports (if not already included in the Equip- inspection personnel
ment Profile). — names of the inspection team, copies of certificates
— justification for acceptance under screening criteria. — records of any pre-inspection development work and per-
— list of vessels considered to be the same as the one under sonnel training and qualification requirements where spe-
consideration and justification that degradation can be cialised methods not covered by the general certification
expected to be the same schemes (e.g. Ref. 3) shall be applied.
— justification of selection of category for: Programme
— confidence in ability to predict types and locations of An outline of the inspection programme should be included to
degradation enable the detailed planning of resources. This should also
— previous inspection effectiveness include advice on any parallel activities which may impact on
— severity and rate of degradation. the timing or performance of the inspection.
Inspection Plan Once prepared, the work-pack should be formally issued and
treated as a controlled document. Relevant documents which
This is the document produced by the NII planning team which are produced during the course of the inspection, e.g. inspec-
is the basis of the inspection. It should contain the following tion reports, should be incorporated in the work-pack, with
information: copies sent to all those issued with controlled copies of the
work-pack. When completed, the work-pack and results,
— details of the inspection strategy and objectives together with the analysis should become part of the vessel
— including the associated decision process inspection history records.
The master work-pack should be issued to the Inspection Man-
— datum referencing system ager who will assume responsibility for maintaining it. The
— complete list of equipment zones considered and their work-pack should be issued in advance of the inspection to
associated flaw types allow adequate time for inspection preparation.
— justification shall be provided for each feature not 4.11 Inspection Plan Review
being considered for inspection
The inspection plan forms the basis for most of the subsequent
— inspection criticality and effectiveness activity hence it is important that it is reviewed prior to imple-
mentation. It is good practice to include the inspection body
— including the associated decision process (including their inspection technicians) in this process to
DET NORSKE VERITAS

Page 44
ensure that all are fully aware of the issues. As a minimum, this other team members). The Inspection Supervisor should liaise
review should consider the following: between all parties and ensure good communication.
— are all degradation mechanisms identified in the CRA The Inspection Supervisor should ideally be certified at least to
being addressed EN473 Level 2 in the methods of NDT to be applied during the
— has the location of potential degradation been considered inspection. However, Level 3 certification is preferable, partic-
in the inspection plan ularly when the Inspection Supervisor is not participating
— is the inspection plan consistent with the inspection directly in the execution of the inspections. Where the method
Type(s) identified as applicable to be used is outwith a certification scheme (e.g. thermogra-
phy), the supervisor should be suitably experienced in the use
— is the inspection plan consistent with the results of previ- of the chosen method.
ous inspection(s)
— does the inspection plan include details on: Other members of Inspection Team
EN 473 Level 2 certification should normally be the minimum
— method requirement for site inspection. However, for NII the require-
— coverage ments are often more specialised than covered by the general
— procedures Level 2 certificate. The Inspection Body should ensure that the
— equipment personnel are suitably qualified, experienced and certified
— resolution where necessary for special applications.
— are there any shortcomings in terms of access, insula-
tion removal, presence of restrictions (e.g. cable trays 5.1.2 Preparation Tasks
etc.)
— reporting requirements. Preparation for the inspection should include the following:
Programming/planning
The review process should particularly address any zones
where the inspection effectiveness has been down-graded in Provision of adequate resources to meet the scope of work.
order to enable inspection by non-intrusive methods. In partic- Coordination with parallel activities which may impact on the
ular, the impact on plant risk and subsequent inspection inter- inspection.
val should be considered. Are there any restrictions to working in the area (over-the-side,
The inspection plans should be made available to the inspec- time constraints, fire watch requirements)?
tion body well in advance of the work being carried out. The Team selection
purpose of this is to allow identification of potential problem
areas in terms of methods and procedures specified. Likewise, Selection of personnel with the necessary experience, qualifi-
the inspection plan should be made available to platform per- cations and certification.
sonnel as soon as possible to allow checks on possible access Any job-specific training requirements should be considered
problems (rope access / scaffolding) to be made. It is highly and dealt with (e.g. specialised ultrasonic inspection methods
recommended that a brief visual review should be carried out such as TOFD). This issue has particular relevance to NII as
in order to confirm that access to the specified locations will be specialised methods not covered by standard certification
possible, and with sufficient space to enable efficient inspec- schemes may be required. Evidence of general NDT compe-
tions to be performed. tency (e.g. PCN level 2) is not necessarily evidence of compe-
The inspection body should, at this stage, be asked to provide tency in a specialised method and appropriate supplementary
evidence of other similar inspections successfully carried out training (and occasionally examination) may be appropriate.
if new methods or challenging requirements are being consid- Access
ered. In certain instances there may also be a need for the ven-
dor to carry out trials to demonstrate application of the Is the area to be inspected accessible for the method to be
proposed method and procedure e.g. using test blocks. In this employed? This includes adequate scaffolding, inspection area
eventuality, suitable evidence should be recorded and handed within reach and sufficient clearance for personnel to access,
to the inspection manager in order to maintain an auditable power supply available.
document trail. Requirements for removal of lagging and insulation and,
Detailed records of the review process should be retained and where appropriate, heat tracing to the required extent.
incorporated into the work-pack. Facilities
— adequate messing facilities?
5. On-site Inspection Activities — office space for producing reports, PC working?
— storage for equipment, including power for battery charging?
5.1 Preparation for Inspection — adequate protection from inclement weather conditions?
— adequate protection for equipment and personnel?
5.1.1 Individual Responsibilities — is the area sufficiently clean and dry?
Preparation for the inspection will require contributions from — any temperature considerations?
each of the following members of the inspection team: Surface Conditioning
Inspection Manager (the plant owner’s representative) Surface requirements stated to be checked by the local inspec-
To ensure that all of the parties are aware of what is expected tion supervisor and if not satisfactory then to be reported back
of them, and have access to all of the relevant information. To to the inspection manager and the plant operator. Remedial
process and act upon any feedback on the inspection work- action, if needed, will have to be sanctioned by the plant
pack. owner/operator, together with any making good of coatings
Inspection Supervisor (the leader of the site NDT Team) etc. This is a critical issue for NDT, as it has a major influence
on the reliability and quality of the inspection.
This is the key coordinating role. The Inspection Supervisor
has many critical responsibilities and there may be a need for Safety Issues
more than one supervisor for large scale inspections (or at least Team members should be properly qualified – site safety train-
for the Inspection Supervisor to delegate some of the tasks to ing is normally mandatory.
DET NORSKE VERITAS

Page 45
Permit requirements: Is the area a designated safe area and are alous measurements at the earliest opportunity.
permits required? Are there equipment safety check require- The Inspection team should practice good housekeeping both
ments? (most NDT equipment is not intrinsically safe) during the course of the inspections and on completion of the
Local safety induction courses or medical clearance may be inspection.
required prior to going on site. When the inspection is underway, responsibility for ensuring
Check any medical restrictions on personnel. the plan is implemented lies, in the first instance, with the
Pre-mobilisation Briefing inspection vendor. They should highlight as soon as possible
any potential problem areas. These might include:
In certain circumstances, particularly difficult or off-shore
inspections, it is beneficial to hold a pre-mobilisation briefing — access problems
in order to familiarise the personnel involved in the inspection — insulation not removed
with the particular aspects of the inspection. — problems with surface condition
Mobilisation — poor sensitivity or excess noise due to material condition
— faulty equipment
Equipment should be checked in advance of shipping to the — changes to procedure or methods used.
inspection site. Mobilisation to site/platform may require
advance shipping of the equipment. It may be advisable to pre- Similarly, platform inspection personnel should monitor
pare a checklist of ancillary items that may be needed (tools, progress of the inspection and check off work carried out
reporting materials, markers, spare consumables/IT consuma- against the plan. The work items can be checked off directly
bles), check that the requisite software is installed on comput- against the plan in terms of method, location and coverage. In
ers. addition, discussions while the inspection is being carried out
The team should ensure that the requirements for the equip- should include consideration of data quality and whether this
ment have been met (including calibration and certification) is regarded as acceptable by the technicians. The importance of
and that everything is in good condition, batteries charged, this regular reporting and feedback can not be overemphasised
PAT certification satisfactory etc. before packing and shipping as it allows corrective actions to be taken whilst the inspection
or mobilisation. teams are still on site.
Start-up Meeting The inspection team should be encouraged to produce interim
reports on an ongoing basis. These should describe as a mini-
All parties should meet prior to the inspection to ensure lines mum the work items completed on the plan, any anomalies
of communication are clear and all understand the inspection detected and any deviations from the plan.
requirements and objectives.
The Inspection Team should be familiarised with the inspec- 5.3 Dealing with Non-conformances
tion, procedures, components, safety issues in a briefing ses- Problems should be reported back to the integrity management
sion prior to the inspection. team as soon as they arise. The aim should be to ensure that,
Known areas of degradation should be identified and quanti- wherever possible, deviations from the plan are identified prior
fied to the NDT operator in advance of the inspection. (Change to or during the inspection rather than when the inspection is
in the extent of degradation may be as important as new areas complete. This allows direct assessment of the effects of the
of degradation) deviation and suitable alternatives that can be carried out as
part of the current inspection to be specified.
The role of each team member should be clearly defined.
The principles outlined in Sec.6 should be applied to dealing
5.2 Performing the Inspection with non-conformances at this stage. Every opportunity should
be taken to replace any inspection that can not be carried out to
The Inspection Manager (or an appropriate senior delegate) the plan with a substitute inspection, using a different method
should coordinate the Permit To Work system and liaise with or in a different location. An understanding of the Strategy
site personnel and the inspection team. Type for the particular inspection (see Sec.4.3) is important at
The Inspection Manager should monitor progress against the this stage since it affects how best to deal with the substitution.
programme and take appropriate action where necessary. Table 5-1 below summarises the approach recommended.
The NDT operators should comply with the agreed scope of Documents which evolve as the inspection progresses (e.g.
work, and inform the supervisor of any obstructions or anom- inspection reports) should be added to the work-pack.
Table 5-1 Summary of approach for defining actions, according to Inspection Type, for dealing with deviations
Inspection Type
Deviation
Type A Type B Type C
Area not inspected, e.g. Substitute with area subject to simi- Substitute with area subject to simi- Attempt to overcome access restric-
access problem lar conditions lar conditions tion. If not, highlight as a non-con-
formance to be dealt with after the
inspection is complete.
Poor technique perform- Attempt to correct and redo inspec- Attempt to correct and redo inspec- Attempt to correct and redo inspec-
ance, e.g. procedures not tion that is affected. If not then con- tion that is affected. If not then con- tion that is affected. If not then con-
followed, incorrect calibra- sider substitution by alternative sider sider
tion, poor surface condition method giving similar performance. (i) substitution by alternative (i) substitution by alternative
method giving similar performance method giving similar performance
or or
(ii) data analysis to check if results (ii) FFS type study to define mini-
are acceptable. mum performance requirement.
Indications potentially associated with in-service degradation 5.4 Reporting of Results

should also be reported to the integrity management team as The format for reporting NDT results will have been specified
soon as possible. Such indications may warrant additional by the Inspection Management Team and defined in the work-
inspection or a different approach (see Sec.7). pack.
DET NORSKE VERITAS

Page 46
Guidance is provided below. during the next inspection.

The reporting criteria and format should be specified in detail. 5.5 Demobilisation
The format should facilitate repeatability of the inspection and
comparison between past, current and future inspection results. The Inspection Supervisor should check that all parties are
aware of, and satisfied with, the completion of the inspection.
Performa reporting formats are recommended to optimise All equipment should be checked before packing and shipping.
repeatability. These should prompt the operator to enter the
same type of information recommended for any routine NDT A debriefing meeting with all parties is recommended. This
inspection. offers opportunity for feedback, lessons learned, and recom-
mendations for future inspections. Any such recommendations
Generally the following information should be included in the should be documented and stored appropriately, such that it
report: can be retrieved for consideration in future inspection planning
1) unique report no. stages.
2) client
3) contract name/number
6. Evaluation of Inspection
4) test date(s)
5) component - title/brief description of component under test 6.1 Introduction
6) dimensions - relevant dimensions of component under test When carrying out inspection by NII, the inspection plan will
have been devised with specific objectives aimed at ensuring
7) drawing no. that the integrity requirements for the equipment, typically as
8) surface condition defined in the RBI, CRA or hazard assessment, are satisfied.
The plan forms the basis for a number of subsequent activities
9) material and it is important to have a means of evaluating work carried
10) technical details about the equipment and inspection out by comparison to the plan. Where there are deviations from
11) item ref. - reference/identification no. of component under the plan these should be dealt with in a consistent way.
test This section of the procedure provides guidance on evaluation
12) inspection result - including sketches where appropriate of NII. It covers actions to be implemented prior to the inspec-
tion being carried out, during the inspection and on completion
13) procedure no. - inspection procedure number including of the inspection. Guidance on options for justifying non-con-
issue number formances and for actions following identification of critical
14) acceptance standard - including issue no. non-conformances is also provided. Finally it provides recom-
mendations for setting the next inspection interval, based on
15) test limitations and any remarks the existing inspection schedule or RBI assessment for the
16) clear identification of the ‘sample’ where a sample inspec- item, and the quality of the non-intrusive inspection carried
tion only has been requested. (for example, where an out.
entire component from a batch of similar components has
been tested, the unique identity no. of the component Evaluation forms an important activity at various stages in the
should be reported; where only a percentage of a compo- approach to NII. The stages at which it is most important are
nent has been tested the operator should report the size and highlighted in the flowchart below. This document provides
location of the ‘sample’ with reference to a datum system) further details on each of these areas.
17) operator’s name, signature and date of report 6.2 Items to be checked for conformance
18) operator’s certification details (e.g. certificate no., type, On completion of the inspection and delivery of the inspection
expiry date) reports, a thorough review of the reports should be carried out.
19) inspection supervisor’s/manager’s approval This should include consideration of the following for each
work item in the inspection plan:
20) client approval - if required
21) any other information required by the specified procedure. — method
— procedure
Alternatively the report format may be a customised software — data quality
application, particularly where repetitive statistical informa- — location
tion is to be recorded and analysed. Such software programmes — coverage.
should be designed to quickly highlight important measure-
ments. The approach recommended is summarised in Figure 6-1
below, and further explanation under each of the evaluation
The Inspection Supervisor should check the reports and the headings is given in the following sections. Where a non-con-
Inspection Manager should evaluate the results as early as pos- formance is identified, its impact on the effectiveness of the
sible. This will ensure early warning of any of the following inspection can be evaluated using the partial factors (RFQuality
situations and allow more time to take the necessary action: and RFCoverage) approach described in Sec.7. Where no non-
— any new significant indications conformances are identified, and the inspection has been spec-
— any significant changes in the inspection result ified in accordance with references [4] and [5], then the inspec-
— any factor which has restricted the inspection tion can be considered to be at least equivalent to IVI and the
— restrictions to the performance of the inspection in accord- re-inspection date set accordingly.
ance with the procedure 6.3 Inspection Method
— incorrect application of the procedure
— the need for supplementary inspections. Different methods have differing abilities to identify certain
types of flaw. When considering whether an alternative inspec-
The inspection reports should be reviewed and approved by the tion method is acceptable for a given inspection, it is important
Inspection Supervisor and Manager. The approved reports to consider the nature of the degradation expected, if any, and
should be incorporated in the work-pack. Finally, the work- the ability of the new method to detect and characterise it in
pack shall be archived and available for review prior to and comparison with the method originally specified.
DET NORSKE VERITAS

Page 47
6.3.1 Type A inspection data obtained from the new method is sufficient to demonstrate
A successful Type A inspection relies on a method that is suf- that the vessel is fit for service, then in most cases the deviation
ficiently sensitive to detect early signs of degradation. If the can be considered acceptable.
method applied has similar POD (to that of the method speci- Guidance note:
fied) for the type of flaws considered as possible in the CRA, There is one exception to this, i.e. when the technique and the
then it would normally be considered as acceptable. If, how- way it is applied is such that there is a strong systematic overes-
ever, the POD is worse then the deviation may be considered timation of wall thickness. This possibility should be carefully
as unacceptable or only justified after further investigation (see considered in carrying out any statistical analysis of the data.
Sec.7). ---e-n-d---of---G-u-i-d-a-n-c-e---n-o-t-e---
6.3.2 Inspection Types B and C The replacement method should be considered in relation to
For inspection types B and C, the method is required to suffi- the allowable defect sizes as determined by fitness for service
ciently characterize any degradation present to allow a fitness considerations. This will usually mean that some compromise
for service judgement to be made. Therefore, as long as the in performance for small flaws is acceptable.
DET NORSKE VERITAS

Page 48
Technique used in No POD of technique used No

NII Inspection complete accordance with equivalent to that
specification ? specified ?
Yes
Procedure in accordance No POD of procedure used No

with specification ? equivalent to that
specified ?
Yes
Inspection data quality as No POD be reduced as a Yes Determine revised quality

factor RFQuality
expected ? result ?
(Figure 7-1)
No
No Alternative locations of No
All locations inspected equivalent significance
inspected ?
Yes
Inspection coverage as No Alternative locations of No Determine revised

equivalent significance coverage factor RF Coverage
specified inspected ? (Figure 7-2)
Yes
Yes Evaluate using company

Annomalies identified
procedures
No
Determine next
inspection interval
( Section 6.2)
NII scope complete
Figure 6-1
Inspection Evaluation Flowchart
DET NORSKE VERITAS

Page 49
6.4 Procedure and therefore precise sizing of defects is not absolutely neces-
sary. Hence a reduction in data quality is often tolerable in a
The procedure used to carry out the inspection is likely to Type B inspection provided it does not affect the validity of the
affect a number of inspection performance parameters includ- statistical analysis. Changes in data quality that are likely to
ing POD. The resulting evaluation is therefore very similar to result in a situation where there is a strong systematic overes-
that for alternative inspection methods. timation of wall thickness or underestimation of crack length
6.4.1 Type A inspection will not be acceptable, however, as this affects the validity of
the statistical approach.
In a Type A inspection, deviations in the procedure that are
unlikely to impact negatively on the POD for small flaws (of 6.5.3 Type C inspection
the types identified as being possible in the CRA) would nor-
The data quality directly affects a number of inspection per-
mally be acceptable. However, when POD for such flaws is
likely to be negatively affected, the deviation will usually be formance parameters including POD. In a Type C inspection it
is usually the POD for relatively deep flaws that is the primary
unacceptable. In certain cases it may be possible to justify
concern and some compromise in the POD for smaller flaws is
deviations in procedure following a more detailed analysis (see
Sec.7). often acceptable. Hence if the data quality is such that it is
unlikely to severely affect the POD for larger flaw sizes
6.4.2 Type B inspection (approaching the FFS allowable) it can be regarded as accept-
able. When there is a negative impact on POD for flaws of this
In the case of inspections of type B, the defects to be identified size the deviation would normally be considered as unaccept-
are expected to be significantly smaller than the critical size, able. However, there may be situations in which the deviation
and therefore precise sizing of defects is not absolutely neces- can be shown to be justifiable on the basis of more detailed
sary. Hence a reduction in inspection performance is often tol- analysis.
erable in a Type B inspection provided it does not affect the
validity of the statistical analysis. Changes to the procedure 6.6 Location
that are likely to result in a situation where there is a strong sys-
tematic overestimation of wall thickness or underestimation of 6.6.1 Type A inspection
crack length will not be acceptable, however, as this affects the The locations selected in a Type A inspection will have been
validity of the statistical approach. determined as being representative of the worst regions in dif-
6.4.3 Type C inspection ferent zones of the vessel. If a location for inspection has been
moved (e.g. due to access restrictions) to somewhere subject to
The procedure used affects a number of inspection perform- similar process conditions and potential for degradation, then
ance parameters including POD. In a Type C inspection it is the deviation would normally be considered acceptable. If,
usually the POD for relatively large flaws that is the primary however, the conditions in the new location are likely to be dif-
concern and some compromise in the POD for smaller flaws is ferent from those in the location specified then the deviation
often acceptable. Hence, if the deviation in the procedure is would be considered unacceptable or only justifiable following
unlikely to severely affect the POD for larger flaw sizes further investigation (see Sec.7).
(approaching the “fitness for service (FFS)” allowable) it can
be regarded as acceptable. When there is a negative impact on 6.6.2 Type B inspection
POD for flaws of this size the deviation would normally be The locations selected for a Type B inspection will have been
considered as unacceptable. However, there may be situations determined as being representative of the worst regions in dif-
in which the deviation can shown to be justifiable on the basis ferent zones of the vessel with the objective being to gather
of more detailed analysis. (statistically) sufficient data from each zone to enable an esti-
6.5 Data Quality mation of the worst potential flaws in areas not inspected. If a
location for inspection has been moved (e.g. due to access
Identification of flaws relies on interpretation of the inspection restrictions) to somewhere subject to similar process conditions
data. When carrying out manual inspection this usually has to and potential for degradation, then the deviation would nor-
be done on-line as the data is collected. In this case the proce- mally be considered acceptable. If, however, the conditions in
dure provides the main means of assurance in data quality and the new location are likely to be different from those in the loca-
there is an expectation that qualified technicians will be able to tion specified then the deviation would be considered unaccept-
recognise when problems arise. When data collection is auto- able or only justifiable following a statistical analysis that
mated, a higher level of assurance is often possible since the shows sufficient data has been collected for the affected zone.
data can be examined in detail off-line. It is difficult to define
the limits of what constitutes acceptable data quality and this, 6.6.3 Type C inspection
in itself, is usually determined by the NDT method. However, In a Type C inspection high coverage over regions or features
there are a number of factors not directly defined in the proce- of concern is usually a primary requirement. This will nor-
dure (e.g. poor surface quality, system noise pick up) that can mally mean that there is little scope for accepting substitution
affect the data quality. by location. However, acceptability can sometimes be demon-
6.5.1 Type A inspection strated by more detailed analysis (see Sec.7).
There is an emphasis in a Type A inspection on the ability to 6.7 Coverage
detect signs of relatively minor degradation. Hence, reductions
in data quality that are unlikely to impact negatively on the 6.7.1 Type A inspection
POD for small flaws (of the types identified as being possible The coverage in a Type A inspection will have been deter-
in the CRA) would normally be acceptable. However, when mined to allow sufficient sampling of the worst regions in dif-
POD for such flaws is likely to be negatively affected, the devi- ferent zones of the vessel to give a high level of confidence that
ation will usually be unacceptable. In certain cases it may be if any degradation is active, its presence is detected. Deviations
possible to justify reductions in data quality following a more that locally reduce the coverage by a small amount will nor-
detailed analysis (see Sec.7). mally be acceptable. However, where coverage has been
reduced to the point where there is no inspection data from a
6.5.2 Type B inspection particular zone or feature of concern, then the deviation would
In the case of inspections of type B, the defects to be identified be seen as unacceptable or only justified following further
are expected to be significantly smaller than the critical size, investigation (see Sec.7).
DET NORSKE VERITAS

Page 50
6.7.2 Type B inspection a shortcoming in the corrosion assessment. In either event it is

The coverage in a Type B inspection will have been deter- important to gain an understanding of the underlying cause.
mined so as to ensure (statistically) sufficient data from each This will usually depend on a greater knowledge of the nature,
zone to enable an estimation of the worst potential flaws in size and locations of the degradation than offered by the initial
areas not inspected. The effects of reduced coverage normally inspection by NII. Furthermore, given that this inspection will
feed directly into the statistical analysis, with reduced cover- typically have relatively limited coverage it is important, from
age making it more difficult to demonstrate an acceptable con- a direct integrity perspective, to ensure that significantly worse
dition. Most instances of small amounts of reduced coverage degradation is not present in the areas not inspected.
can be dealt with on this basis. However, where coverage has Given the above concerns, in most instances of flaws being
been reduced to the point where there is no inspection data detected in a Type A inspection, additional inspection to be
from a particular zone or feature of concern, then the deviation implemented within a relatively short timeframe is warranted.
would be seen as unacceptable or only justified following fur- This inspection should seek to increase coverage and provide
ther investigation (see Sec.7). detailed information on the degradation. In many instances this
requirement is best met by internal visual inspection but there
6.7.3 Type C inspection may be situations in which further inspection by NII is appro-
In a Type C inspection high coverage over regions or features priate.
of concern is usually a primary requirement. This will nor-
mally mean that there is little scope for accepting reductions in 6.9.2 Type B inspection
coverage. However, acceptability for small reductions in cov- An unacceptable situation in a Type B inspection is one where
erage can usually be demonstrated by more detailed analysis the probability of failure (typically based on the potential for
(see Sec.7). flaws in the areas not inspected) is regarded as excessive. This
can be addressed by a number of means, as described below:
6.8 Critical Non-conformance
When the inspection includes a critical non-conformance, i.e. — additional inspection coverage using the same method to
it clearly does not meet the objectives of the inspection, some increase the size of data set (and reduced area for which
action will be required to redress the situation. Each case will data is not available)
be dealt with on its merits but, broadly speaking, the following — inspection with the same coverage but using a method will
options can be considered: less inherent variability in wall thickness readings. this
will reduce the estimates of the worst flaw sizes
— repeat as soon as possible the inspection work items to — inspection by NII or IVI with 100% coverage to identify
which the non-conformance relates directly the worst flaw size
— carry out internal visual inspection as soon as possible — additional, more sophisticated, data analysis including
— repeat part or all of the NII work-scope on a shorter inter- correlation of results from different locations.
val than would normally be applied
— carry out internal visual inspection on a shorter interval In many instances the first two options will be more cost effec-
than would normally be applied tive.
— apply an alternative inspection in the short term
— carry out regular monitoring of wall thickness over local- 6.9.3 Type C inspection
ised areas When flaws are found in a Type C inspection a detailed fitness
— place emphasis on demonstrating that the process is under for service assessment would normally be required unless the
control and conditions leading to excessive corrosion are flaws can be directly classified as superficial. It may often be
not present (e.g. monitor levels of CO2, H2S or regular the case that the information provided in the initial NII inspec-
checks on corrosion coupons). tion is not in itself sufficient to underpin the FFS study. Hence
some additional inspection aimed at more fully characterising
The particular circumstances will determine the best option the flaws may be needed. This inspection may be by NII or IVI
and the associated timeframes for implementation. depending on the circumstances. Further action going forward
would normally be defined using the findings of the FFS study
6.9 Reportable Indications and flaws as a basis.
In general the procedures relating to flaws are well covered by
international standards (such as API 579 [9] and BS7910 [19]) 6.10 Examples
as well as company standards, and will not be dealt with in
detail here. However, there are some aspects where the appli- 6.10.1 Type A inspection
cation of NII may lead to differences in approach under some Within a Type A inspection the key issue is the ability to detect
circumstances. the presence of degradation, even when it might be in its early
For example, where a Type A inspection has been specified stages. Hence the focus in the Type A inspection will tend to
(i.e. no degradation is predicted) however the inspection iden- be on inspection performance and this will have been used to
tifies degradation, albeit below the level that would require specify the inspection method.
action based on a fitness for service evaluation, then further Non-conformances that significantly affect the performance
investigation is required. (POD) of the inspection compared to that of the system speci-
As with integrity management approaches that rely on internal fied will therefore potentially seriously compromise the objec-
visual inspection, specific actions are required when flaws are tives. If a system with reduced performance is used this can be
found. The Inspection Type forms a useful basis on which to compensated for to some extent by increasing the coverage but
define the principles governing the nature of action to be taken a minimum performance requirement must still be in place.
when flaws are found. A brief summary is provided below. Non-conformances that relate to coverage are potentially less
serious provided it is clear that the reduction/change in cover-
6.9.1 Type A inspection age does not entail a significant reduction in the area inspected
A Type A inspection usually applies where degradation is con- within zones having similar operating characteristics. For
sidered unlikely and/or degradation is expected to be superfi- example, moving a region for inspection on the bottom of a
cial. Hence if degradation - beyond some small limit - is found horizontal vessel longitudinally by a few hundred millimetres
then it is likely to be sign of a process problem leading to cor- to allow easier access would be acceptable provided the condi-
rosion that is more rapid than expected or it is an indication of tions are no different to those in the area originally specified
DET NORSKE VERITAS

Page 51
for inspection. reduce some of the initial conservatism and allow the non-
In many cases it will be location more than total area covered conformance to be acceptable.
that is important, i.e. it may be acceptable to reduce coverage Although inspection performance may typically be less of a
in a region of concern rather than moving the inspection to concern than coverage in this type of inspection, it remains an
another location where conditions will be less onerous. important consideration. There will be certain minimum
6.10.2 Type B inspection requirements, usually related to tolerable flaw sizes, that must
be met. If the inspection system or implementation is such that
Within a Type B inspection the emphasis will be on gathering these requirements are not met then the only scope for demon-
sufficient information to allow a quantified statistical assess- strating acceptability will be to refine the definition of tolera-
ment that can be used to demonstrate a high confidence in esti- ble flaw dimensions.
mating the worst flaw that might exist (including the areas not
inspected). Inspection performance and coverage both affect
whether the information is sufficient.
If we accept that this type of inspection forms the basis for a
7. Inspection Interval
statistical analysis whose aim is to allow quantified assessment 7.1 Discussion
of vessel condition then this same statistical analysis can be
used as a check on the inspection itself. Uncertainties associ- Sec.6 provides guidance on defining acceptability of devia-
ated with inspection system performance (e.g. poor resolution) tions from the work-scope specified and how they can be over-
and coverage will tend to (but not always) drive and increase come or compensated for. Once a non-conformance has been
in the dimensions of the worst expected flaw. Hence, if the determined to be unacceptable there may be further action that
inspection system is inherently noisy or has poor resolution, it can be taken to show that it remains justifiable. This section
may be difficult to demonstrate an acceptable situation. Like- outlines briefly the types of action that may be appropriate.
wise if the coverage is very low it will be more difficult to This Decision Guidance Process does not usually require a
demonstrate an acceptable situation. Hence judging non-con- very comprehensive CRA and integrity assessment to be in
formances will generally be straightforward for this type of place before NII is deemed appropriate. It adopts a pragmatic
inspection. There are some potential complications associated approach consistent with the level of knowledge/information
with systematic errors in the inspection system (e.g. a consist- that will typically be associated with offshore pressure equip-
ent overestimation of thickness) but these can be dealt with by ment and, consequently, the inspection requirements are usu-
carefully considered analysis of the results. ally specified conservatively.
6.10.3 Type C inspection Guidance note:
The Decision Guidance Process aims for practical implementa-
Within a Type C inspection the emphasis is on ensuring a high tion in an efficient way. In most cases the benefits (in terms of
probability of detecting the worst flaw, recognising that degra- reduced workscope) of carrying out very comprehensive CRA
dation may be isolated to particular regions. In many situations and FFS studies on a vessel by vessel basis as a precursor to
the dimensions of a flaw that it is considered essential to detect inspection planning do not justify the cost of these activities.
(i.e. something with potential to grow to a size of concern
within the interval to the next inspection) will be large enough ---e-n-d---of---G-u-i-d-a-n-c-e---n-o-t-e---
to allow a reasonably high probability of detection (assuming There are two approaches that can be used to demonstrate that
the affected area is inspected). For example, in this type of the inspection carried out can be considered acceptable; firstly,
inspection, it may sometimes be sensible to use a method that comparison with internal visual inspection (IVI) and secondly,
trades off a slightly lower POD for small flaws against ability by a quantified (or semi-quantified) statistical assessment of
to cover large areas quickly. the probability of failure. Whilst the first of these is conceptu-
Since coverage will tend to be the key issue in this type of ally simple, it is of necessity conservative, and it may be pos-
inspection, non-conformances that affect coverage will often sible to gain further concession based on a more quantitative
be of most concern. There are a number of issues that affect the assessment. However, the effort and data requirements of a
acceptability of coverage related non-conformances. Several quantitative or semi-quantitative assessment are considerably
of these are highlighted below: greater. The two approaches are each discussed in more detail
in the following sections.
1) localised reduction in coverage due to local access restric-
tion, e.g. nozzles or pipework blocking access. In many In common will all other aspects of the NII process, it is imper-
circumstances this may be acceptable and is to be ative that a record is kept of all factors considered during the
expected. For corrosion type flaws the effects can be evaluation of the inspection. In particular, any decisions relat-
assessed relatively simply however by establishing the ing to the acceptance or rejection of any non-conformances
acceptable loss in wall thickness for the dimensions of the identified, and the subsequent adjustment of the inspection
area not inspected. If the metal loss elsewhere is well interval must be adequately recorded in order to allow for
within the acceptable wall loss, then it may be reasonable review and checking as necessary.
to surmise that it is unlikely that unacceptable wall loss
will have occurred in the un-inspected region; on the other
7.2 Comparison with IVI
hand, if measured wall loss is only just within limits else- Although the first approach is widely used - and forms the
where, it will be more difficult to justify the acceptability basis for this recommended practice - it can be difficult to inte-
of the un-inspected area. This type of approach should take grate in a detailed safety case since the performance of IVI is
into consideration the findings of the corrosion assess- not well defined. For this reason, the following flow charts
ment, i.e. what type of flaws are likely to be present. It have been developed. They guide the user in the event of a
will, for example, be less useful when microbial corrosion non-conformant inspection being carried out, providing a
of a type that may be very localised is active pragmatic approach to assessing the next inspection interval on
2) significant reductions in coverage over particular areas. the basis of the interval based on the relative effectiveness of
This will normally be unacceptable – justification here the inspection compared to IVI.
would be primarily on the basis of the corrosion review. The impact of any non-conformance in the NII process can
This might reconsider the possible nature of degradation in then be evaluated in terms of a reduction in inspection interval,
light of the results that are available. A more comprehen- for example that determined by RBI and based on internal vis-
sive corrosion assessment for the vessel concerned may ual inspection as follows:
DET NORSKE VERITAS

Page 52
The second function, the coverage factor RFCoverage is a func-

Interval NII = RFQuality × RFCoverage × Interval IVI tion of the actual coverage achieved in comparison with that
specified at the outset. Determination of the coverage factor is
Where described in the flowchart below, Figure 7-2. Here the issue is
to ensure that sufficient material has been inspected to enable
IntervalNII is the new inspection interval following NII conclusions to be drawn. This is particularly important where
the CRA predicts that corrosion will take place in a particular
RFQuality is the factor determined from Figure 7-1 area, for example at an inlet nozzle.
RFCoverage is the factor determined from Figure 7-2 By following the logic of the appropriate chart, the two factors
IntervalIVI is the inspection interval based on IVI inspection can be evaluated and the revised inspection interval deter-
mined as a proportion of the normal inspection interval based
Where there is no non-conformity, the appropriate factor can on RBI considerations combined with internal visual inspec-
be taken as 1.0, therefore a successfully completed NII inspection.
tion would have a new inspection interval equal to that for an
IVI inspection. It is important that any inspection interval determined on the
basis of the charts in Figures 7-1 and 7-2 should be reviewed
The quality factor RFQuality is a function of the ability of the by competent staff to ensure that all relevant facts have been
inspection carried out to identify and adequately quantify any adequately taken into account in setting the revised interval. In
flaws present in the inspected item. It combines an assessment particular, it should be noted that the factors embodied in the
of the inspection method, and the parameters and procedures charts are based on a judgemental assessment of the impact of
used as well as an assessment of the quality of the data a particular deficiency for “typical” situations. Where such sit-
obtained from the inspection. The assessment is by its nature uations do not exist, it is likely that the reduction factors would
judgemental and should be made by personnel experienced in need to be adjusted accordingly. The results of the review
the use and interpretation of the inspection methods specified process and any amendments to the inspection interval should
and used. be recorded and retained for QA audit purposes.
DET NORSKE VERITAS

Page 53
Inspection
No Treat as loss of coverage;

Is data obtained useable ?
Calculate RF Coverage from Figure 7-2
Yes
Yes No No No
Inspect further
Type A Inspection POD reduced < 10 % POD reduced < 25 % POD reduced < 50%
RF Quality = 0
No Yes Yes Yes
Inspection acceptable Interval reduced Interval reduced

RF Quality = 1.0 RFQuality = 0.75 RF Quality = 0 .5
Inspection to measure Yes Systematic over estimation

No Defects > 50% acceptable
Yes Investigate further
defect size of defect size limit found RF Quality = 0
No Yes No
Inspection acceptable
RFQuality = 1 .0
B Interval reduced
Inspection type
RF Quality = 0 .5
C Interval reduced
RFQuality = 0.25
Yes No Minimum measured wall No No Investigate further

Inspection to measure wall Systematic over estimation
thickness > 10% over Total variation < 2 σ RF Quality = 0
thickness of wall thickness
allowable minimum
Yes Yes Yes
Investigate further
Inspection acceptable
RF Quality = 0
RFQuality =1 .0
B Interval reduced
Inspection type RF Quality = 0.5
C Interval reduced
RF Quality = 0 .25
Figure 7-1
Evaluation of Method / Performance Non-conformances
DET NORSKE VERITAS

Page 54
Inspection
Area missed near feature

No No No No Types A /B / C
<10% missed <25% missed <50% missed Inspect alternative area
of concern?
RFCoverage = 0
Yes Yes Yes Yes
Types A /B Types A / B Types A /B

Inspection acceptable Interval reduced Interval reduced
RFCoverage =1.0 RFCoverage =0.75 RFCoverage = 0.5
Type C Type C Type C

Interval reduced Interval reduced Inspect alternative area
RF Coverage =0 .66 RFCoverage =0.33 RFCoverage =0
Area or feature similar to

No No No Types A / B /C
<10% missed <25% missed Inspect alternative area
that missed inspected? RFCoverage =0
Yes Yes Yes
Types A /B Types A / B
Interval reduced Interval reduced
RFCoverage =0.5 RFCoverage = 0.5
Type C Type C
Interval reduced Inspect alternative area
RF Coverage =0 .25 RF Coverage= 0
No No No Types A / B /C
<10 % missed <25% missed <50% missed Inspect alternative area
RFCoverage =0
Yes Yes Yes
Types A /B Types A /B Types A / B

Inspection acceptable Inspection acceptable Interval reduced
RFCoverage =1.0 RFCoverage =1.0 RFCoverage = 0.5
Type C Type C Type C

Interval reduced Interval reduced Inspect alternative area
RFCoverage =0.5 RF Coverage =0 .25 RF Coverage= 0
Figure 7-2
Evaluation of Non-conformance in Coverage / Location
7.3 Detailed Assessment that would consider the worst sizes of flaws expected, the dis-
Where it is not possible to justify the non-conformance on the tribution of flaws, growth rates and allowable flaw sizes.
basis of the above comparison with IVI, a quantitative or semi- Detailed guidance on how to assess fitness for service (or prob-
quantitative approach may enable a given non-conformance to ability of failure) of a vessel, given information on corrosion
be justified. These assessments cover three main areas, i.e. conditions and inspection performance, is beyond the scope of
— corrosion engineering this document. Guidance on fitness for service assessment is
— inspection performance covered in detail in a number of codes and standards (e.g.
— structural integrity. BS7910, API 579) and this can form the basis for a probabilis-
tic approach including consideration of inspection perform-
A reduced inspection work-scope would usually be justifiable ance measures. A number of approaches, different in terms of
following a very comprehensive CRA and integrity assessment detail, are applicable and will be appropriate in different cir-
DET NORSKE VERITAS

Page 55
cumstances. This document includes a brief summary of the Clearly, as with the entire NII process, inspection validation
key elements of the type of approach that can be adopted. requires engineering judgement to be used at all stages. All
The starting point for the assessment will typically be a available information should be used in order to assess the
detailed corrosion review. This would normally consider some value of the inspection carried out and sole reliance should not
or all of the following: be placed on the mechanistic approach presented in the guide-
lines. The examples quoted here, however, demonstrate that
— process conditions (chemistry, temperatures, pressures, under normal circumstances the approach will give appropriate
dew-points) assessments.
— materials The remaining sub-sections of this section describe the back-
— types of degradation (e.g. pitting, general corrosion ground and resulting evaluation for the various validation cases.
— locations of degradation
— growth rates 7.4.1 Glycol Contactor
— upset conditions Background information
— effects of controls including inhibitors
Inspection History
— monitoring activities
— inspection and maintenance history. Installation date: 1995
The aim of the corrosion study should be to provide an Date: Feb. 1995
informed view on the likely types of degradation, their loca- Type of inspection: Internal visual – baseline inspection.
tions, their distribution and the worst expected sizes at present Summary of findings: Nothing significant in unlined section.
and going forward over the interval to the next inspection. A holiday noted in coating near one of the nozzles in the lined
section. This was repaired.
An assessment of the performance of the inspection actually
carried out with respect to the types of flaw identified in the Date: May 1998
corrosion review is also usually necessary. The key aspect to Type of inspection: Internal visual.
be considered is the probability of detection for a range of flaw Summary of findings: Lining in good condition. Shell plates in
sizes and distributions. This should include, where possible, good condition, some corrosion deposits noted on surface near
consideration of the coverage and likely spatial distribution of middle section.
the flaws (e.g. is degradation likely to be localised or more ran-
domly distributed). Note that the emphasis will be different Expected degradation
according to inspection type however and the following Internal corrosion is primary mechanism of concern (CRA pre-
broadly applies: dicts 0.99 mm/yr loss for the contactor section and the corro-
sion rate in middle section just above the lined section may be
— in a type a inspection the key concern is on the ability to expected to more than double.). Inspection (IVI May 1998)
detect early signs of degradation. hence the probability of confirmed presence of corrosion deposits in middle section but
detection should relate to finding sufficient signs of degra- there was no measurable depth of corrosion loss.
dation to conclude that degradation is indeed present
(since this is what would trigger further action) External corrosion, Stress corrosion cracking and Wet H2S
— in a type b inspection there is a lesser emphasis on pod cracking identified as low probability. No cracking identified
directly. however, it remains essential that the inspection by inspections to date.
system is unlikely to systematically overestimate wall Failure of lining could potentially lead to rapid corrosion hence
thickness. for this reason, it is important to have an under- integrity of lining is very important.
standing of the likely back wall surface morphology and
any limitations this can impose on the recording of thick- Similar vessels
ness data (depending on scan increment for example) This is the only vessel of this type on the platform.
— in a type c inspection it is typically the ability to detect Specified IVI Work-scope
larger flaws that is the primary concern. for such inspec-
tions the flaws are potentially isolated so the pod should The work-scope for traditional internal visual inspections
related to isolated flaws rather than finding signs of degra- includes inspection of the entire surface of the pressure retain-
dation within a larger flaw population. ing envelope, including man-way doors and necks, nozzle
bores and nozzle welds. N.B. Some areas of the internal sur-
The integrity assessment brings together the corrosion and face are inaccessible due to the presence of internals which are
inspection information to develop an estimate of the probabil- not removed during shutdowns. Following the 1998 inspec-
ity of flaws, with potential to threaten integrity of the vessel, tion, the period was set at 5 years.
having been missed in the inspection. This relies on estimates NII Inspection Plan
of the allowable defect sizes (as defined by fitness for service
considerations) and comparison with the flaw sizes that may Following the HOIS decision guidance procedure, it was con-
have escaped detection during the inspection. cluded the vessel was suitable for inspection by NII given the
margin against perforation assuming conservative corrosion
The final outcome of the assessment should normally be an rates (approximately 5 years). The following inspection plan
estimate of the effects of the inspection actually carried out on was specified for unlined sections of the vessel:
the level of assurance. If this is not considered acceptable then
further action (usually in the form of inspection) will be Corrosion mapping at standard resolution of the following areas:
required in addressing the non-conformance.
1) vessel top head: 1 off 300 mm × 300 mm area adjacent to
7.4 Examples top nozzle.
(Region 1 of Figure 7-3)
The following example cases were used in order to validate the
evaluation approach described in this section. A number of 2) strake 6: 4 off 300 mm × 300 mm areas at cardinal posi-
actual NII assessments which had used the recommended prac- tions.
tice in establishing the validity and scope of inspection were (Region 2 of Figure 7-3)
reviewed. For each example, the steps in the evaluation are 3) strake 5: 4 off 300 mm × 300 mm areas at cardinal posi-
presented and the resulting reduction in inspection interval tions.
considered in comparison with engineering judgement. (Region 3 of Figure 7-3)
DET NORSKE VERITAS

Page 56
4) strake 4: 4 off 300 mm wide vertical bands from 200 mm 7.4.2 Glycol Flash Drum
below chimney tray to 200 mm above packing tray. Background information
(Region 4 of Figure 7-3)
Inspection History
5) strake 4: Circumferential band 300mm high centred on
packing tray. Start up date: 1996
(Region 5 of Figure 7-3).
Date: Jan. 1995
Type of inspection: Internal visual – baseline inspection.
Item 1 Summary of findings: Shell and head plates in good condition.
Summary of findings: No signs of corrosion reported.
Date: Oct. 2001
Type of inspection: External thermal survey.
Summary of findings: No significant problems.
Item 2
El+11750 Expected degradation
Internal corrosion, External corrosion, Stress corrosion crack-
Item 3 ing and Wet H2S cracking identified as low probability in
CRA. No corrosion or cracking observed during previous
El+8500
El+11630 inspections.
Similar vessels
El+6380
This is the only vessel of this type on the platform.
El+5434 Specified IVI Work-scope
The specified IVI work-scope was for internal visual inspec-
Item 4 tions of the vessel heads shell plates, man-way door, nozzle
bores and attachment welds.
Item 5 NII Inspection Plan
Following the HOIS decision guidance procedure, it was con-
cluded that the vessel was suitable for inspection by NII using
a method with medium efficiency as no mechanisms leading to
high probability of failure were identified either by previous
inspection or corrosion assessment. However, credible mecha-
nisms include:
— corrosion – some internal corrosion may be present. If so,
it is anticipated at the liquid level and would be expected
Figure 7-3 to be more advanced at the inlet end of the vessel. No par-
Glycol Contactor Inspection Locations
ticular susceptibility for corrosion of nozzles (except inlet
nozzle) or welds in preference to parent material
— stress corrosion cracking and Wet H2S cracking identified
In addition to the corrosion mapping, time of flight UT inspec- as low probability in CRA.
tion (TOFD) was carried out on the lined section and nozzles.
Inspection Evaluation The proposed NII work-scope was therefore as follows:
Determination of Inspection Type Vessel shell
The inspection type is determined following the guidance of — Corrosion mapping of one 300 mm × 300 mm area per
Figure 3-2 of the procedure. In this case, the corrosion map- shell plate at standard resolution at TDC.
ping inspection is intended to confirm that the extent of inter- — Corrosion mapping of 100% of top half of inlet side vessel
nal corrosion has remained within expected and acceptable head to 300 mm below centreline.
limits. — Corrosion mapping of a band 600 mm wide along cen-
Given that the CRA has predicted that internal corrosion will treline of vessel for full length of strake 1 (closest to inlet
take place and that corrosion debris was identified during the nozzle).
previous inspection, the flowchart is entered on the “high like- — Corrosion maps 600 mm high by 300 mm wide astride
lihood” of degradation. strake 2/3 circ. weld. Both sides of vessel at vessel cen-
treline.
For internal corrosion, the extent of degradation is typically — Corrosion maps 600 mm high by 300 mm wide, both sides
uniform, and “General” can be selected. The CRA predicted of strake 3, adjacent to strake 3/vessel head circ. weld at
corrosion rates in excess of 0.99mm/year in the event of coat- vessel centreline.
ing failure at the, at which rate integrity of the vessel is pre- — Corrosion map 600 mm high by 300 mm wide at centreline
dicted to be compromised in around 5 years. The degradation of vessel head, outlet end.
rate is therefore taken as “High”, and the inspection type is — TOFD inspection of all main shell circ. welds from
therefore given as “Type B”. 300 mm above vessel centreline to 300 mm below, both
Determination of Inspection Interval sides of vessel.
— TOFD inspection of all main vessel shell weld long/circ.
The inspection was carried out in full, using the intended tee intersections for 300 mm along each leg.
inspection methods. Following the evaluation flowchart (Fig-
ure 6-1), no reduction in the proposed future inspection inter- Nozzles
val is therefore required. This is clearly what would be
expected, and the evaluation procedure correctly leads to a sat- — Grey scale 0° axial B-scans at 90° intervals around inlet
isfactory conclusion. nozzle circumference or corrosion map of nozzle material.
DET NORSKE VERITAS

Page 57
East
15 16 17
11
1,7 3,9 5,18 13

12 2,8 4,10 6,19 14
Figure 7-4
Inspection Plan for Glycol Flash Drum
Inspection Evaluation therefore, the quality factor is evaluated using Figure 7-1. In
Determination of Inspection Type this case, the inspection is a type B inspection to determine that
any cracks present are below critical size (i.e. the inspection
The inspection type is determined separately for the two deg- objective is to determine the size of defects). UT does not
radation methods considered, internal corrosion and cracking, imply a systematic over estimation of defect size, and as no
using the flowchart in Figure 3-2 of the procedure. flaws were found during the inspection (let alone defects in
As the CRA identified no corrosion mechanisms leading to a excess of 50% of critical), this results in a quality factor of 0.5
high probability of degradation, the flow chart is entered at the for the cracking mechanism inspection.
“Low likelihood” spine. Corrosion is generally either “Gen- In both cases, corrosion and cracking, the evaluation has deter-
eral” or “Localised” in extent. In order to present a conserva- mined that further inspection should be carried out at a reduced
tive assessment, in this case “Localised” has been assumed. interval. In this case, the guidance suggests that the interval
The corrosion rate is taken as “low” given the results of the should be half that intended based on the original inspection
CRA, which gives an inspection type of “Type A” for the cor- scope. In practice, further inspection of the vessel was sched-
rosion mapping. uled for the following year. Once again, the evaluation proce-
For the cracking mechanisms, the CRA predicted a “Low like- dure has arrived at a suggested interval which is in agreement
lihood”. By their nature, stress corrosion and wet H2S cracking with that actually determined for the plant on the basis of engi-
mechanisms principally occur in regions of highest stress, and neering judgment alone.
thus the location is “Likelihood – clearly defined”. Cracking
can proceed rapidly, impacting on vessel integrity within its 7.4.3 Vent Knock Out Drum
planned life, thus “High rate” is chosen. This gives and inspec- Background information
tion “Type B” for the TOFD inspection.
Inspection History
Determination of Inspection Interval
Start up date: 1996
A number of the corrosion mapping inspections on the north-
ern side of the vessel were unable to be carried out for reasons Date: Jan. 1995
of accessibility. Following the flowchart in Figure 6-1 of the Type of inspection: Internal visual, video only – baseline in-
procedure, the inspection was carried out using the correct spection.
methods and procedures, thus the quality factor is taken as 1, Summary of findings: Some rust staining identified, but no ev-
however the coverage factor needs to be determined using Fig- idence of significant wall loss. No evidence of cracking.
ure 7-2.
The un-inspected regions (Regions 7, 8, 9, 10, 18 and 19) are Expected degradation
expected to experience similar conditions to the corresponding Internal corrosion, External corrosion, Stress corrosion crack-
regions on the southern side of the vessel (i.e. Regions 1, 2, 3, ing and Wet H2S cracking identified as low probability in
4, 5, 6). The scans do however cover directly the primary zone OCA. No cracking observed during inspections but evidence
of concern on the end, i.e. the liquid line, and a large proportion of corrosion (rust staining) observed during internal video sur-
of the total head area and the results presented do not show any vey. No significant loss of wall however in UT thickness meas-
signs of significant corrosion. Following the flowchart in Fig- urements.
ure 7-2 the coverage factor is determined as 0.5 as follows:
Similar vessels
The area missed is deemed to include areas of concern (i.e. liq-
uid line) but areas of similar susceptibility have been success- This is the only vessel of this type on the platform.
fully inspected with no indications. The total area of those not IVI Work-scope
inspected is estimated to be between 25 and 50% of the origi-
nally specified zones. For a type A inspection, this gives a cov- The specified IVI work-scope was for internal visual inspec-
tions of the vessel heads shell plates, man-way door, nozzle
erage factor of 0.5.
bores and attachment welds.
Similarly, a number of the planned TOFD inspections were not
able to be carried out; however in this case, manual UT inspec- NII Inspection Plan
tions were carried out instead. In evaluating this inspection, The HOIS decision guidance procedure concludes that the ves-
DET NORSKE VERITAS

Page 58
sel was suitable for inspection by NII using a method with 7.4.4 HP Condensate Separator
medium efficiency as the CRA indicated only minor corrosion Background information
would take place and that sufficient corrosion allowance
Inspection History
existed. Credible degradation mechanisms for the vessel
include: There is no evidence service induced degradation.
Expected degradation
The vessel is constructed in carbon steel overlaid with roll clad
it is anticipated at the liquid level and would be expected 316 stainless steel.
to be more advanced at the inlet end of the vessel. No par-
ticular susceptibility for corrosion of nozzles (except inlet The nominal operating temperature is moderate (19°C).
nozzle) or welds in preference to parent material The vessel is not insulated.
— stress corrosion cracking and Wet H2S cracking identified Degradation of the stainless steel cladding is unlikely in the
as low probability in CRA. process conditions (the threshold temperature for chloride pit-
ting/SCC (approx 200°C) is significantly higher than the oper-
The proposed NII work-scope was therefore as follows: ating temperature. Breakdown of the cladding would lead to
exposure of the underlying carbon steel to the process condi-
— UT corrosion mapping of the vessel shell and wall thick- tions and corrosion would then be expected. This is low likeli-
ness measurement to repeat base-line survey. hood but the inspection should seek to demonstrate that the
— Manual UT inspection of accessible areas of shell long/ cladding remains intact.
circ. welds for cracking at the internal surface. NII Inspection Plan
Following the HOIS decision guidance procedure, it was con-
Inspection Evaluation cluded that the vessel was suitable for inspection by NII using
Determination of Inspection Type a method with medium efficiency primarily to confirm integ-
rity of the lining, although degradation is considered to be neg-
The inspection type is determined separately for the two deg- ligible if the lining remains intact. Credible damage
radation methods considered, internal corrosion and cracking, mechanisms include:
using the flowchart in Figure 3-2.
As previous inspection indicated the presence of rust staining, it is anticipated at the liquid level and would be expected
the flow chart is entered at the “High likelihood” spine. Corro- to be more advanced at the inlet end of the vessel. No par-
sion is predicted by the CRA to be “General” in extent, but ticular susceptibility for corrosion of nozzles (except inlet
with a “low” corrosion rate. This gives an inspection type of nozzle) or welds in preference to parent material
“Type A” for the corrosion mapping. — stress corrosion cracking and Wet H2S cracking identified
as low probability in CRA.
For the cracking mechanisms, the CRA predicted a “Low like-
lihood”. By their nature, stress corrosion and wet H2S cracking The proposed NII work-scope was therefore as shown in Fig-
ure 7-5:
mechanisms principally occur in regions of highest stress, and
thus the location is “Likelihood – clearly defined”. Cracking 1) corrosion mapping (standard resolution) and manual UT
can proceed rapidly, impacting on vessel integrity within its for signs of pitting and SCC of 4 bands centred on the bot-
planned life, thus “High rate” is chosen. This gives and inspec- tom of the vessel and covering the lower 30 degrees of cir-
tion “Type B” for the weld inspections. cumference
Determination of Inspection Interval 2) corrosion mapping (standard resolution) and manual UT
of bands approximately 300 mm wide covering the lower
A total of thirty two scan areas were inspected on the shell of portion of both sides of the vessel to approximately 300
the Vent Knock Out Drum. However, the bottom head could mm above the vessel centreline. The bands should extend
not be mapped due to the curvature of the head and restricted from the weir position to approximately 300 mm upstream
access. A manual inspection of this region was carried out and of the weir (i.e. towards C2)
pitting to 2 mm deep was reported. 3) corrosion mapping (standard resolution) and manual UT
of regions extending approximately 300 mm longitudi-
In this case it is considered that the inspection actually carried nally and 600 mm circumferentially at the top of Strakes
out would have a broadly similar capability to identify degra- 1, 2 and 3. Each region should be located lengthwise
dation as the automated procedure originally defined, and approximately centrally on the strake
therefore the inspection interval does not need to be modified 4) circ. welds C1 and C2 to be inspected 50% by 45 degree
in the light of the change in method. The flaws identified, the manual UT and TOFD. Must include regions at bottom,
2mm deep pitting, should be evaluated using company proce- mid-side and top of vessel
dures and other appropriate codes such as API 579 [9]. In this 5) the bore of nozzles N1, N2, N3, N4, N5A, K1C, K2C,
case, the flaws were within acceptable limits and no further K4B, K3B, A1A and A1B to be inspected for signs of pit-
action was required. Therefore, the inspection interval could ting / SCC using ultrasonic scanning at each of the four
be safely maintained at 3 years. cardinal positions.
DET NORSKE VERITAS

Page 59
Figure 7-5 erage factor needs to be determined using Figure 4-3.

Inspection Plan for Condensate Separator
The un-inspected region is expected to experience similar con-
ditions to the other two top surface areas which were correctly
Inspection Evaluation inspected, although the top surface is not considered to be an
area of particular concern. None of the scans carried out on the
Determination of Inspection Type
vessel identified any areas of concern.
The inspection type is determined separately for the two deg-
radation methods considered, internal corrosion and cracking, The coverage factor is calculated using the flowchart
using the flowchart in Figure 3-2. (Figure 4-3). As the top surface is not considered to be of par-
ticular concern, the area of missed inspection is compared to
As the CRA identified no corrosion mechanisms leading to a the total area inspected for the vessel. In this case <10% of the
high probability of degradation, the flow chart is entered at the total planned inspection. The coverage factor is therefore
“Low likelihood” spine. Corrosion is generally either “Gen- determined as 1 and no reduction in inspection interval is
eral” or “Localised” in extent. In order to present a conserva-
tive assessment, in this case “Localised” has been assumed. required.
The corrosion rate is taken as “low” given the results of the Had the top surface been considered to be a high risk area,
CRA, which gives an inspection type of “Type A” for the cor- however, then the missed area would have been determined
rosion mapping. against the total equivalent area. In this case one out of three
For the cracking mechanisms, the CRA predicted a “Low like- equivalent areas designated for inspection at the top of the ves-
lihood”. By their nature, stress corrosion and wet H2S cracking sel was not inspected, therefore approximately 33% of the
mechanisms principally occur in regions of highest stress, and intended coverage was missed. The coverage factor for this
thus the location is “Likelihood – clearly defined”. Cracking case is determined to be 0.5 (an area of concern, but with sim-
can proceed rapidly, impacting on vessel integrity within its ilar features having been inspected and <50% missed). Further
planned life, thus “High rate” is chosen. This gives and inspec- inspection is therefore recommended at a reduced interval in
tion “Type B” for the TOFD inspection. order to maintain safety levels.
Determination of Inspection Interval This case therefore illustrates the importance of engineering
For reasons of accessibility, one area at the top of the separator judgement in the evaluation process. Clearly differing conclu-
vessel was not able to be inspected according to the original sions can be drawn through different interpretation of the
inspection plan. Following the flowchart in Figure 3-2, the results. It is therefore important that the supervising inspection
inspection was carried out using the correct methods and pro- engineer records all stages of the evaluation process for audit-
cedures, thus the quality factor is taken as 1, however the cov- ing purposes and future reference.
DET NORSKE VERITAS

Page 60
8. References 14) National NDT Centre POD database, AEA Technology

Engineering Solutions, 1998.
1) EN 473 General Principles for Qualification and Certifica- 15) Silk M.G., Stoneham A.M and Temple J.A.G.,
tion of NDT personnel. The reliability of non-destructive inspection, Adam
2) ISO 9712 Non-destructive testing - Qualification and cer- Hilger, Bristol, 1987.
tification of personnel. 16) Non-destructive Testing Handbook American Society for
3) PCN/GEN General Requirements for Qualification and Non-destructive Testing, Second Edition, 1996.
PCN Certification of NDT Personnel. British Institute of
Non-Destructive Testing. 17) Non-destructive evaluation (NDE) capabilities data-book
Non-destructive Testing Information Analysis Center
4) Decision Guidance Process for Non-Invasive Inspection: (NTIAC) Texas Research Institute, 1997.
User Guide, HOIS 2000 Report ref. HOIS2000 C5/3
September 2004. 18) Interactive Knowledge Base (IKB) on NDT, HOIS Mem-
ber Version:
5) Recommended Practice for Non-Invasive Inspections www.ndt.esrtechnology.com/HOIS/IKB/Application2000
GSP 235 Mitsui Babcock, 2002. or Demo Version of IKB on HOIS Public website:
6) Guidance on the Evaluation of Non-Invasive Inspection, http://www.ndt.esrtechnology.com/hois/ikb/demo.
HOIS 2000 Report ref. HOIS2000 C5/4 November 2005.
19) Guide on Methods for Assessing the Acceptability of
7) Best Practice for the Procurement and Conduct of Non- Flaws in Structures
Destructive Testing. BS7910: 1997 British Standards Institution, BSI, London.
Part I: Manual Ultrasonic Inspection
HSE Gas and Process Safety Technology Division, 20) "Best Practice for the Procurement and Conduct of Non-
November 2000. www.hse.gov.uk/dst/ndt.pdf. Destructive Testing Part 1: Manual Ultrasonic Inspec-
tion", HSE, November 2000.
8) Best practice for risk based inspection as part of plant
integrity management, 21) "Best Practice for the Procurement and Conduct of Non-
HSE Contract Research Report 363/2001 prepared by Destructive Testing Part 1: Magnetic Particle and Dye
TWI and Royal & Sun Alliance Engineering, Penetrant Inspection", HSE, April 2002.
Health and Safety Executive, 2001. 22) API 510: Pressure Vessel Inspection Code, 8th Ed., Amer-
9) API 579 Recommended Practice: Fitness for Service ican Petroleum Institute, 1997
American Petroleum Institute, 2000. 23) Discussion with P Horrocks regarding AEA Technology
10) Energy Institute Model code of safe practice. Pressure ves- RBI principles, January 2001
sels examination, Part 12. (1993). 24) Comments from D Ritchie regarding Phillips Petroleum
11) Energy Institute Model code of safe practice. Pressure pip- methods for determination of inspection interval, March
ing systems examination, Part 13. (1993). 1999
12) DNV-RP-G101 "Risk Based Inspection of Offshore 25) Example adapted from "Introduction to Life Prediction of
Topside Static Mechanical Equipment". Industrial Plant Materials" edited by M. Kowaka. Allerton
13) Wall M. and Highmore P.J., 'A comparison of visual and Press 1994
non-invasive methods for vessel and tank internal inspec- 26) Probabilistic Models for Optimising Defect Detection in
tion', HOIS(98)P8, A report produced for HOIS by AEA LPG Welds G. A. Georgiou, Proceedings of the British
Technology, 1998. Institute of NDT Conference 2000, pp 168 - 173.
DET NORSKE VERITAS

Page 61
APPENDIX A
REVIEW OF NON INTRUSIVE NDT METHODS
A.1 Introduction mined measurement position can be recorded. The transducer

is scanned manually over the surface and the thickness read-
This review gives a brief description of the principles, capabil- ings are stored on disk. After the scanning is finished, the data
ities and limitations of NDT methods which might be consid- are plotted in a wall thickness map. Each thickness level can be
ered as non-invasive inspection methods, including: colour coded and wall thinning by corrosion or erosion is more
1) Ultrasonic Testing readily recognised than by manual inspection. High reproduc-
ibility (typically within 0.3 mm wall loss) enables accurate
2) Eddy Current Testing monitoring and calculation of corrosion rates.
3) Liquid Penetrant Wall thickness mapping can be applied in-service at tempera-
4) Magnetic Particle Inspection tures up to about 250°C using special high temperature probes
and couplant. Wall thinning, pitting corrosion, flow accelera-
5) Magnetic Flux Leakage tion corrosion hydrogen induced corrosion and hot hydrogen
6) Thermography attack can be detected and imaged.
7) Radiography Recently there have been significant advances in the technol-
ogy used for corrosion mapping, which has greatly increased
8) Backscatter or Compton Imaging scan speeds and reduced noise levels, allowing more reliable
9) Acoustic Emission detection of small pitting type flaws. These systems are based
10) Remote Visual Inspection on high speed automated scanning systems (crawlers etc.),
which are well suited to scanning large, unobstructed areas
11) Shearography. such as sections of vessel shells between nozzles etc.
A.2 Ultrasonic Testing For NII of vessels, corrosion mapping is a widely used method
for detection and sizing of internal corrosion/erosion flaws
A.2.1 Conventional Ultrasonic Testing (both localised and generalised loss of wall).
Conventional ultrasonic testing is based on the generation of A.2.3 Time of Flight Diffraction (TOFD)
ultrasonic beams by means of probes containing a piezoelec-
tric element excited by an electrical pulse. The piezoelectric TOFD (Time of Flight Diffraction) is an advanced ultrasonic
element vibrates and generates mechanical waves of frequency inspection method which can simultaneously detect and size
typically in the range 1 MHz – 10 MHz. The sound wave prop- indications. TOFD sizing in the through wall direction is based
agates into the specimen and discontinuities make the waves on the measurement of signal arrival times which is inherently
reflect back to the same or to a different transducer. more accurate than methods based on amplitude. With TOFD,
through wall sizing accuracy is typically 1 mm or better.
Conventional ultrasonic transducers can generate a variety of
angle beams which can provide sufficient coverage for the It provides reproducible fingerprints which makes it suitable
inspection of the whole thickness of a specimen. for condition monitoring. Initial manufacturing flaws can be
monitored and service induced flaws detected and progres-
Ultrasonic inspection can be used for the monitoring of corro- sively monitored. TOFD examination only requires external
sion by measuring the wall- thickness of the specimen. It can access to the object to be inspected. TOFD can be applied on
also detect and size pits. hot structure up to 200°C using special transducers. Weld
A poor surface finish, thick paint or inspection at high or low inspection of heavy wall pressure vessels (up to 300 mm wall
temperature may cause problems for ultrasonics (although thickness) has been carried out. Nozzle and flange welds (com-
more appropriate transducers can be used for particular condi- plex geometry) can be inspected with prior computer simula-
tion, e.g. specially designed high temperature transducers can tion modelling to aid inspection planning and result evaluation.
be used to inspect hot surfaces). For NII applications, TOFD is often used for inspection of
welds for typical welding flaws, including weld root erosion/
A.2.2 Ultrasonic Imaging corrosion, cracks, lack of fusion etc. Recently, there have also
Ultrasonic imaging systems can be use to improve reliability in been NII applications where TOFD has been used as a rapid
detecting and/or sizing flaws. scanning method for detection and sizing of backwall pitting
type wall loss flaws, as an alternative to more conventional
A.2.2.1 Automated Ultrasonic Imaging corrosion mapping.
Ultrasonic imaging systems integrate microprocessor technol- A.2.4 Creeping Head Inspection (CHIME)
ogy with non-destructive testing methods and can perform
inspections on vessels and piping from the external surface. The CHIME method (Creeping Head Wave Inspection
Wall thickness up to 600 mm can be inspected for both fabri- Method) consists of the transmission of an ultrasonic signal
cation and service induced cracking, wall thinning or clad between two probes that are placed a distance apart (up to one
delamination in a single pass. The full RF waveforms are col- metre) scanning a parallel (or near parallel) surface.
lected and reliable repeatability is achieved. Multiple testing The transmitter fires a pulse of ultrasound that gives rise to
allows analysis to be performed while the inspection is carried creeping waves propagating along both the surfaces, and criti-
out. cal angle head waves (shear waves) are generated at all parts
Applications include erosion/corrosion detection and monitor- between the surfaces. As a result, the full wall volume between
ing, fatigue crack detection and sizing, hydrogen blistering and the probes is covered. The coverage is achieved with little
stepwise cracking and stress oriented hydrogen induced crack- attenuation, allowing the transmitting and receiving probes to
ing. be well separated compared to traditional inspection.
Wall thickness up to 40mm at a standard operational frequency
A.2.2.2 Ultrasonic Thickness (Corrosion) Mapping of c. 2 MHz and inspection widths (distance between the two
In ultrasonic wall thickness mapping systems, a transducer is probes) of up to one metre have been established. The CHIME
linked to a computer so that thickness data for each predeter- scans can provide information about thickness of the sample
DET NORSKE VERITAS

Page 62
(by measuring the time interval between successive peaks) or possible to measure accurately the through wall dimension of
about any flaw which locally changes the thickness. corrosion, or determine whether it is at the inner or outer sur-
Flaws due to corrosion or cracking, reduce the CHIME signal face. The surface should be free of impurities and loose depos-
magnitude and can change the arrival times of the signal peaks its to provide sufficient acoustic coupling at the transducer test
independent of their location between the transmitter and point. The method is used to detect corrosion at locations with
receiver. limited access, such as: insulated and (with certain restrictions)
clamped pipes, annular plates in storage tanks or under rein-
The method has tolerance to in-service surface conditions and forcement plates. The interpretation of the output signals is of
most coatings. It is sensitive to the presence of corrosion on prime importance as the signals which reach the transducer
internal and external surfaces and provides an estimation of the include those from welds, supports and other geometrical fea-
extent of the corrosion area and nominal wall thickness in areas tures and would superimposed over the signals from corrosion
between corrosion. Some approximate information on the area.
through-wall extent of a flaw is also obtained, by the amount
of the CHIME signal loss produced. “LORUS” (LOng Range UltraSonics) was developed by RTD.
Inspection results are reported in high-resolution colour maps
CHIME has been tested on samples containing general and iso- (presented in 3D views: top, side and end images) and cor-
lated corrosion, stress corrosion cracking, hydrogen and has roded areas are listed by angular position and extent with a
demonstrated tolerance to surface conditions such as general position accuracy of ±10mm. The LORUS system can be
roughness or thin layers of coating. applied to components in the thickness range 6-25 mm. Well-
For NII applications, CHIME is most often applied to sections bonded paint coating is acceptable. Heavy corrosion may
of components made inaccessible to conventional ultrasonic obscure detection of corrosion at larger distances from the
methods by geometry, such as pipe supports and pipe clamps. probe in the plate.
The method also has potential for fast screening for corrosion Trials have been performed to investigate the capability of
pits over large areas of vessel shells. LORUS to detect corrosion inside a vessel underneath a nozzle
A.2.5 M-Skip reinforcing plate (which prevents detection using conventional
0° ultrasonic inspection). The trials were performed using two
M-skip method is a new 'medium' range pitch-catch ultrasonic test-pieces manufactured by MBEL. One test-piece was a noz-
inspection technique developed within the HOIS programme zle welded into a square plate (representing the vessel) and
since 2005. The method has some similarities to CHIME, but containing a 90 mm wide (radial extent) reinforcing plate. The
provides more quantitative information on wall thickness and test-piece contained eleven simulated corrosion sites at various
wall loss via analysis of the arrival times of signals. locations under the reinforcing plate. The other test-piece was
M-skip is a variant of the well known TOFD method but is simplified geometry manufactured from flat plate and contain-
based on angled shear wave probes, separated by the distances ing four simulated corrosion sites. All fifteen simulated corro-
required to inspect wider pipe supports and clamps. The sig- sion sites were detected.
nals travelling between the probes then typically experience
many reflections, or skips, between the front and back compo- A.2.6.2 Lamb/Guided Wave Methods
nent surfaces. Lamb waves are "guided" or "plate" waves which can exist in
Measurements of the arrival times of the different skip signals plates or pipes where the thickness is the same order of magni-
can be used to determine the average wall thickness between tude as the ultrasound wavelength. The wave motion couples
the probes. For areas of wall loss on either the front-wall or both surfaces of the plate together, such that the plate or pipe
back-wall surfaces, the signal arrival times which involve acts as an acoustic waveguide, allowing the propagation of a
reflection/scattering off the defect will be reduced. Measure- series of different Lamb-wave modes, whose velocities depend
ments of the differences in the arrival times between the wall- on the plate thickness and wavelength.
loss signals and the skip signals allows the maximum depth of This method is generally applied using a ring of probes which
wall loss to be derived, assuming the number of reflections are clamped around the outside of a pipe. The method is pulse-
from the area of wall loss is known. If this number of reflec- echo and the transducer array is configured such that the test
tions is unknown, the assumption of a single defect reflection may be carried out in one direction along the pipe and then in
provides a conservative estimate of the actual depth of the wall the other. Long lengths (ca. 30 m - 50 m or more) of pipe may
loss. be examined in both directions from a single test point,
To date, M-skip has been applied to carbon steel plates and although certain pipe coatings (especially soft coatings such as
pipes with wall thicknesses in the range 7 to about 35 mm, with bitumen or tape wrap) can severely attenuate the Lamb waves,
probe separations up to at least one metre. Typical probe fre- and hence reduce the effective range of the method.
quencies are 3-5 MHz. Wall loss flaws are detected by the reflected waves they gen-
M-skip has been applied to carbon steel components contain- erate. For reliable detection, a wall loss flaw needs to reduce
ing generalised and isolated corrosion. For NII, M-skip is the cross sectional area of the pipe by around 2-5% (this varies
applicable to sections of components made inaccessible to depending on equipment used and general conditions of the
conventional ultrasonic methods by geometry, such as pipe pipe to be tested). The amplitude of the received flaw echo pro-
supports and pipe clamps. The method also has potential for vides some information on the overall flaw extent, but not on
fast screening for corrosion pits over large areas of vessel the maximum wall loss.
shells or long lengths of pipework.
The detection capabilities at locations close to (or within) pipe
A.2.6 Long Range Ultrasonics features (e.g. under simple supports) depend on the equipment
used and the operator skills. In some case detection capabilities
A.2.6.1 LORUS are hindered by the presence of pipe features and the user shall
refer to the manufacturers' recommendation.
The principle of the method consists of a high sensitivity angle
beam probe used in multi-skip mode to completely flood the For NII applications, Lamb/Guided inspection is generally used
wall of the component being inspected. Any changes of the as a fast screening method for the inspection of long lengths of
thickness of the pipe at the outer or inner surface can result in straight pipework, although recent advances allow some
reflection of some of the ultrasound back to the transducer. inspection around bends. Guided wave techniques have been
Ranges of over a metre can be achieved. The amplitude of the used to rapidly inspect insulated and clad pipeline corrosion,
response provides an indication of severity although it is not and the results correlated well with subsequent investigations.
DET NORSKE VERITAS

Page 63
Lamb waves can also be generated in flat or large curvature and several millimetres for austenitic material. Advanced eddy
plates, but these methods are much more developmental than current methods can significantly increase penetration depth as
the systems for pipe inspection. Other developments have been described below.
used for permanent monitoring of pipes in hazardous areas,
and for inspection of heat exchanger or boiler tubes. A.3.1 Pulsed Eddy Current Testing
A.2.7 Electromagnetic Acoustic Transmission (EMATs) In the pulsed eddy current method, a coil is supplied with a
short pulse. The resulting eddy current pulse propagates in the
Electromagnetic acoustic transducers (EMATs) are a rela- specimen as a heavily attenuated wave of electromagnetic
tively new method for generating and receiving ultrasound energy with a phase velocity depending on the material and the
which offer certain advantages compared to conventional (pie- frequency. For material thickness measurement, a detection
zoelectric) probes. However they have much lower sensitivity coil can be placed close to the transmitter with suitable shield-
than conventional probes and are much more expensive. ing. Because the wave propagation is highly dispersive,
An EMAT consists of a flat pancake coil of wire carrying a changes in pulse shape occur with distance into the specimen
radio frequency current and a magnet which produces a steady and by measuring the time-amplitude characteristics of the
magnetic field. The radio frequency current induces eddy cur- received pulse and comparing them with the time of occur-
rents in the surface of the specimen and their interaction with rence of certain signal features from similar calibration tests
the magnetic field results in Lorentz forces or magnetic forces via a computer, flaw depth can be determined. A change in
(predominantly magnetostrictive) which cause the specimen wall thickness indicates the presence of corrosion or erosion.
surface to vibrate in accordance with the applied radio fre- The probe does not need to be in contact with the component
quency current. The ultrasonic waves which have been gener- and can therefore be applied through insulation, and also thick
ated at the specimen surface propagate into the specimen in the coatings or layers of deposit as long as they are non-conductive
normal way. They can be detected after reflection from a flaw, and non-magnetic.
for example, by the same transducer or by a separate one using This method can be used for in-service inspection, has good
a reciprocal process. Depending on the direction of the mag- repeatability (±0.1 mm), is transportable and robust, operates
netic field the EMAT can generate shear (radially, horizontally over a wide range of climatic conditions, can be applied on
or vertically polarised), compression or Rayleigh waves. components with temperatures of -100°C to +500°C (above
EMAT ultrasonic can be a non-contact method since there is these temperatures, the use of the system is prohibited due to
no need of fluid couplant and it can be used on rough, scaled the reduction of magnetic permeability), measures through any
or painted surfaces, although any lift-off from the specimen non-conductive and non-magnetic material up to 100 mm
surface reduces the sensitivity still further. thickness (in favourable circumstances even up to 150 mm).
Ferritic cladding up to a certain thickness (at present 0.8 mm)
EMAT probes are now available from certain large probe man- does not influence accuracy and the transducer can be up to 30
ufacturers for use with conventional manual flaw detectors. metres from the base unit.
However these are intended for a special application, involving
thickness gauging of carbon steel components with a magnet- However, typical equipment is limited to inspection of wall
ite layer which has formed through operation at high tempera- thicknesses below 40 mm and pipe diameters over 100 mm, it
tures. The presence of this magnetite layer greatly enhances only detects general corrosion or erosion over relatively large
the efficiency of ultrasound generation and detection, via the areas, does not detect small isolated pits and can be influenced
magnetostrictive effect. For all other steel components, with- by the presence of large metal masses nearby.
out this coating, EMATs remain a research method only.
A.3.2 Saturated Low Frequency Eddy Current (SLOFEC)
EMATs can operate on hot metals, since EMATs are relatively
easy to keep below 100°C with simple water cooling. They can This method is based on the eddy current method with DC
be used up to 700°C if they are brought into momentary con- magnetisation. A magnetising yoke containing an electromag-
tact with the sample and up to 460°C in constant contact. net is used to generate a strong magnetic field in the material
Thicknesses from 2 mm up to 150 mm or higher can be under test. An eddy current probe senses component metal loss
inspected. EMATs can be used to measure wall thickness at due to the local concentration of the magnetic field.
high temperature with an accuracy of ±0.5 mm. EMATs can If the coil is used in differential mode then localised wall loss
also be used on low temperature surfaces (down to liquid nitro- such as pitting can be detected. General wall thinning can be
gen temperature). measured by using the system in absolute mode.
However, when the EMAT is used as a transmitter, problems A saturated low frequency eddy current (“SLOFEC”) system
emerge due to low sensitivity. One recent method used to solve has been developed by a German company (Kontroll Technik).
this problem is to use a laser to generate ultrasonic waves and SLOFEC can provide an improved performance for some
an EMAT as receiver; the EMAT-LASER or EMAT-EMAT applications where classic magnetic flux leakage (MFL) is
method is of high cost compared to more traditional methods, used, e.g. the inspection of tank floors and pipe inspection.
but has proved to be effective for certain specific applications.
Two major advantages of SLOFEC compared to classic mag-
EMATs have been used to measure the wall thickness of steel netic flux leakage are that the maximum wall thickness which
galvanising bottles at their normal operating temperatures that can be inspected is around 30-35 mm, compared to around
are in excess of 450°C. The kettles are typically fabricated 10-15 mm for MFL. Recent studies also suggest that SLOFEC
from 50 mm thick steel. The study indicates that it is possible has improved sensitivity to small pitting flaws than MFL. The
to measure the steel wall thickness to within ±0.5 mm. data from SLOFEC is stored in digital form, and analysis of the
EMATs are not typically currently widely applied for NII signal characteristics provides information on the location of
applications. wall loss flaws (backwall/frontwall). Some information on the
extent of an indications wall loss is available, but this is less
A.3 Eddy Current Testing quantitative than that provided by a method such ultrasonic
Eddy current testing is based on inducing electrical currents in corrosion mapping.
the material being inspected and observing the interaction For NII, SLOFEC has considerable potential for the rapid
between these currents and the material. detection/screening of corrosion in vessel shells and pipework.
Conventional eddy current testing methods are not suitable for A.4 Liquid Penetrant Inspection
volumetric inspection of vessels since typical penetration
depth of the eddy currents is less than 1mm for ferritic material Liquid penetrant techniques, including dye and magnetic par-
DET NORSKE VERITAS

Page 64
ticle variations are low cost methods and are easy to apply and Many of these systems rely upon the use of an adjustable
used to detect surface breaking flaws such as cracks, laps and threshold or amplitude gate to provide the detection of corro-
porosity. Large areas can be inspected but liquid penetrant is a sion in real-time, while some of the more advanced systems,
slow process in terms of application and flaw indication. through the use of both advanced electronics and signal
The principle of liquid penetrant method: first, the surfaces to processing, are able to provide corrosion-maps of inspected
be tested are cleaned – pre-cleaned to remove scale etc., areas similar to the C-scan representation of ultrasonic data.
degreased and dried. The chosen penetrant (solutions of col- Maximum wall thickness which can be inspected is around
oured or fluorescent dyes in oil-based liquids) is applied to the 10-15 mm, and some information is provided on the extent of
surface of the specimen and allowed time to soak into discon- the wall loss, but MFL is less quantitative than methods such
tinuities. The excess penetrant is removed (by water, solvent, as ultrasonic corrosion mapping.
water and detergent or by emulsifier). A developer is then MFL can be used for NII, and crawlers have been developed
applied. After a period of time and if a discontinuity is present,
for inspection of vertical walls, as well as the more conven-
the penetrant seeps out the developer at the discontinuity posi- tional floor scanners.
tion, causing a marked local reduction in developer contrast.
The surface is examined under appropriate viewing conditions. A.7 Thermography
Finally, the surface is clean to prevent corrosion etc.
Thermography is based on the measurement of the heat distri-
The sensitivity of penetrant inspection process can be very bution across a surface. The effect of flaws on thermal conduc-
high and it is possible to detect very small cracks having open- tivity and emissivity of test materials can be analysed by a
ing widths of about 1μm. thermographic method.
Liquid Penetrant inspection is only applicable to the accessible Passive thermography investigates the heat distribution of a
component surface, and is not therefore an NII method. structure with a special infrared camera and looks for hot spots,
A.5 Magnetic Particle Inspection which could for example be due to loss of wall flaws in a ves-
sel/pipe containing hot products, or a breakdown in thermal
Magnetic particle inspection (MPI) is used for the detection of insulation.
surface and near-surface flaws in ferromagnetic materials. It is
one of the most extensively used electromagnetic methods in An alternative method, known as transient or pulsed thermog-
industry as it is easy to apply and provides a direct visual indi- raphy involves inducing a heat pulse into a component by sud-
cation of surface breaking crack. denly raising the temperature of the component surface locally,
using an external heat source (e.g. flash tube). Within the
The method involves the magnetisation of the component HOIS programme, an alternative heat source, using induction
either locally or overall, by application of a permanent magnet, heating was developed to allow the more effective inspection
electromagnet or electric current. A magnetic field is produced of thicker steel components.
inside the material. The magnetic field becomes distorted by
the presence of a flaw causing a local magnetic flux leakage. The heat pulse will diffuse into the material. The rate at which
the heat front is subsequently dissipated depends on the struc-
Ferromagnetic particles in the form of dry powder or sus- ture and flaw content below the surface. The surface tempera-
pended in a liquid (daylight visible or UV fluorescent parti- ture of the structure is monitored using an infrared camera.
cles) are sprayed onto the surface of the specimen to reveal the Anomalies in the temperature distribution reveal the presence
leakage field. The particles are attracted by the magnetic field of flaws or local changes in thermal conductivity due to e.g.
and accumulate in the vicinity of the flaw, which is subse- breakdown in insulation. The method is made more convenient
quently made visible. by using a video recorder or computer to store the rapidly
The minimum angle between the magnetic field and the detec- changing temperature pattern after the structure surface is
tion of imperfection is 30°. The optimum sensitivity is reached heated. Then, flaws in conducting materials that have only a
when the magnetic field is perpendicular to the imperfection transient effect on the temperature distribution may be
orientation. Therefore, the flaws must lie between 30° and 90° detected. Quantification of anomaly size is more usually
to the magnetic field in order to be detected. The magnetic field accomplished by application of a second non-destructive eval-
is therefore applied in two directions at right angles to each uation method as thermography can estimate size only on thin
other. outer layers.
MPI inspection is only applicable to the accessible component A new method known as lock-in thermography is reported to
surface, and is not therefore an NII method. be much quicker, more accurate and be able to probe deeper
than conventional thermographic inspection methods. Lock-in
A.6 Magnetic Flux Leakage thermography is a method which uses a modulated source to
The component is locally magnetised and depending upon the energise the material under test. However, this method is still
level of induced flux density, magnetic flux leakage due to limited for detection of flaws of several millimetres below a
both near and far surface flaws is detected by the voltage material surface. To date, tests have been carried out at fre-
induced in a detector coil or a Hall-effect element which is tra- quencies which allowed depth ranges of 10 mm to be covered.
versed over the surface of the specimen. The method is not lim- Transient thermography can be used as a rapid, non contacting
ited to surface-breaking or near-surface flaws, although it method for detection of sub-surface flaws or voids. However,
becomes increasingly sensitive to far-surface flaws with it becomes less effective in detection of sub-surface flaws as
increasing levels of magnetisation. Also, the output from the the depth of the flaw increases. Typically for detection, the
detector can be amplified, filtered, digitised, etc., and stored to width of a flaw must be greater than its depth below the inspec-
produce automated inspection systems. Multi-element and dif- tion surface. Laboratory tests of the induction heater developed
ferential probes can be used, and inspection speeds can be very in the HOIS programme showed detection of flaws (diameters
high. 15-30 mm) at depths of 10 mm below the inspection surface.
The method is finding increasing use in the petrochemical Smaller diameter flaws could be detected at smaller depths.
industries for providing high-speed inspections of storage tank
floors, as well as carbon steel pipes. These systems utilise A.8 Radiography
either permanent or electromagnets to provide localised near- Radiographic inspection is a process of testing materials using
magnetic saturation coupled with induction-coil or Hall Effect penetrating radiation - either electromagnetic radiation of very
sensor arrays for detecting anomalous flux leakage caused by short wavelength (X-rays, gamma-rays), or particulate radia-
the presence of corrosion flaws (both near and far surface). tion (neutron radiation). Radiography is based on the differen-
DET NORSKE VERITAS

Page 65
tial absorption of the radiation by the material being inspected. A.9 Backscatter (Compton’s) Imaging
In passing through the material some of the radiation is atten-
uated depending on the thickness and the density of the mate- Compton scattering tomography is a relatively new method for
industrial non-destructive testing, making use of γ or X-ray
rial. The unabsorbed radiation that passes through the test-
photons scattered inside the component under examination.
piece can be recorded on an imaging medium, such as film or The source emits a finely collimated beam into the material
more recently the film-less plates used in computed radiogra- and a finely collimated detector allows measurement of the
phy. In general, radiography can detect only features which number of photons that are scattered. The detected signal is
have an appreciable thickness (loss of material) in a direction largely composed of singly scattered photons from the volume
parallel to the radiation beam. Tight cracks in thick sections defined by the intersection of the incident beam and the accept-
might not be detected if they are not favourably oriented to the ance solid angle of the detection collimator. The amount of
radiation beam. scattered radiation produced is directly related to the electron
In service site radiography is generally carried out using iso- density of the material, which corresponds well to the physical
tope gamma ray sources, mainly Iridium 192, but sometime the density of the material. The major advantage offered by back-
lower energy Selenium 75 source is used for higher contrast scatter imaging is that it can be implemented from one side of
images on thinner wall components. For penetrated thick- the object.
nesses greater than about 85 mm, Cobalt 60 can be used, but It can provide information even if the object is very massive,
with increased safety issues. The highest energy source cur- or does not permit X-ray transmission through it, or does not
rently available for in-service applications is a Betatron (elec- permit access to the opposite side. The measurement can be
tron accelerator). performed in the presence of protective coatings and because
One radiographic method for inspecting large diameter pipe- the source and the detector are placed on the same side of the
work, and even small vessels is a double-wall single test-piece, large components filled (with oil for example) can
image(DWSI) method, where only the section of the wall that be inspected at speeds much greater than can be done with con-
is furthest from the source contributes to the radiographic ventional through-transmission radiography. Compton back-
image, although the radiation penetrates both walls. This is scatter imaging is a non contact method and can therefore be
achieved by placing the source very close to one wall. Note applicable at elevated temperatures.
that DWSI radiography is not suitable for vessels of diameter In Compton scatter imaging a three-dimensional image is built
greater than about 1 - 1.5 metres, or where internals obstruct up point by point. The time necessary to obtain an image of a
the beam. Also, the presence of liquid products in the vessel particular volume of interest in the object under study is a func-
further restricts the application of radiography, due to the tion of the size of this volume. The scanning rate varies
increased attenuation in the liquid. For water filled compo- inversely with the square of the inspection volume. If larger
nents, the additional attenuation obtained is approximately voids are of interest, greater scanning rates are possible and
equivalent to an additional penetrated steel thickness of ID/7, therefore shorter scanning times can be used although there is
where ID is the component internal diameter/dimension. loss of spatial resolution. A detector array is usually used to
cover the entire thickness of specimen traversed by the source
Alternative methods for inspection of smaller diameter pipes beam, for each position of the source. A feasibility study on the
and the sections nozzles protruding from vessels, include the use of Compton backscatter gamma-ray tomography for
double wall double image method (DWDI) where both pipe underwater inspections offshore shows that the use of 10 Ci of
walls contribute to the radiographic image, and also impor- 60Co should provide a scanning rate of about 35 cm2min-1. The
tantly the tangential method which gives a direct image of the study showed that this method is capable of detecting 5mm
pipe wall, allowing the extent of any loss of wall flaws to be cubical flaws positioned at a depth of 32 mm of steel. There-
measured directly from the image (provided appropriate dis- fore, this method could be an attractive tool to detect corrosion/
tance calibration methods are used), erosion of the inner surface of vessels. However, this is an
Raised surface temperature increases difficulties as the radio- expensive method and it might need to be adapted for particu-
graphic film cannot be used in contact with the sample surface lar applications.
at temperatures above ≈40°C. The film therefore needs to be The major drawbacks of Compton scatter imaging systems has
insulated at higher temperature. However, introducing insula- been their high capital costs and very slow scan speeds, which
tion increases the sample to film distance and can therefore have precluded their application to all but the highest value
increase the image un-sharpness (blurring of the image), components.
although this effect is likely to be small unless the insulation There have been however some recent developments to this
thicknesses are very large. method, especially in the USA, and systems are now being
Recent developments in the field of in service radiography used for certain specialised large area scanning applications
include the use of collimated sources which allow greatly (e.g. foam thermal insulation on NASA’s Space Shuttle).
reduced controlled areas, and allow radiography to continue This is not currently a recognised NII method in the oil and gas
without plant turnaround, or the need for after hours work. industry, but there may be some potential for applications such
In addition, computed radiography is now being increasing as CUI inspection.
used. With this technology, traditional film is replaced by film- A.10 Acoustic Emission (AE)
less plates, which are exposed in a similar manner to film but
are then read out (scanned) using a laser scanner, and the Acoustic emissions are pulses of elastic strain energy released
resulting digital image is displayed on a computer monitor. spontaneously during deformation of materials. When a body
Various computerised enhancement and analysis routines can suddenly deforms locally and relieves local stresses, a burst of
then be applied. Because film-less plate are more sensitive to energy is emitted.
penetrating radiation than film, computed radiography leads to AE can be released by a number of causes such as plastic
reduced exposure times, and hence faster inspection. deformation, inclusion cracking, crack growth, but also by cor-
In addition, the computer based analysis routines greatly facil- rosion, phases changes and impact. External factors such as
itate quantitative wall loss measurements using the tangential mechanical impacts, friction, machinery vibration, welding
method. Analysis of image grey level information can also be operations, can also produce acoustic emission.
used to estimate loss of wall in the direction parallel with the These emissions propagate through the specimen and are
radiation beam, provided appropriate calibration methods are detected by sensors placed on the surface of the specimen,
used, although this is a less well established method at present. which in turn convert the energy into electrical signals. These
DET NORSKE VERITAS

Page 66
are amplified, stored, processed and displayed. Acoustic emis- optic boro-scope carries visual information through fibre-optic
sion examination is non-directional as AE propagates in spher- cables each of which makes up a picture element of the final
ical wave fronts. Therefore, a sensor located anywhere in the image. Articulation controls allow the user to manipulate the
vicinity of an acoustic emission source can detect the resulting end of the scope in the interior of the structure. Special protec-
acoustic emission. A large volume of a structure can be moni- tive coatings allow the boro-scope to operate while submerged
tored at once; it is not necessary to examine specific regions. in liquid. They are typically available in diameters from 1.4 to
Since only limited access is required, “active” discontinuities 13 mm and in lengths up to 12 m.
in areas inaccessible to the more traditional non-destructive Rigid and flexible boro-scopes are available in a wide variety
methods can be detected. Location of the source of the acoustic of standard and customised designs and several factors can
emission can be accomplished by several search units posi- influence the selection of a scope for a particular application.
tioned over the surface area to be tested and monitoring the These factors include focusing, illumination, magnification,
time of arrival of the signals to the various search-unit loca- working length, direction of view and environment.
tions. Because of the high velocity of sound and the relatively
close spacing of search units on a steel vessel, time resolutions Both types of boro-scopes can be manufactured to withstand a
must be made in microseconds to locate the source within a variety of environments. Rigid endo-scopes as standard will
centimetre. In most cases, inspection requirements are such accept temperatures from -30°C up to 150°C and will withstand
that data must be available in a short period of time. Therefore pressure up to 6 bars. Flexible endo-scopes will work in temper-
most systems of this type utilise a computer for handling and atures up to 80°C and will withstand pressures of between 2 and
displaying the data. 3 bars. Boro-scopes can be tailor made to meet different speci-
The in-service inspection of a pressure vessel may consist of fications such as length (up to 18 m), diameter (down to 10 mm),
monitoring during periodic proof testing, during normal pres- lighting system and the use of a jacket protection to allow the
sure cycles or continuously during normal operation. When the system to be used at higher temperature (up to 1600°C). How-
vessel is pressurised to a level less than that to which it has been ever, the price of the system is related to these options and
previously subjected, little or no acoustic emission occurs. would increase rapidly with the quality of performance.
Therefore, on subsequent pressurisations no AE will be gener- Boro-scopes provide a means of checking in-service flaws in a
ated unless a crack has extended in service because of corrosion variety of equipment and are of particular interest for the build-
or fatigue. On pressurising after crack growth, the stress system up of corrosion and scale on the internal surfaces of for exam-
at the enlarged crack will be changed from that previously, and ple heat exchangers.
further emission will be obtained. The detection of in-service
stress corrosion cracking is reported to be a relatively straight- Video-scopes are similar to fibre-optic boro-scopes. The
forward application as stress corrosion cracking produces copi- video-scope involves the electronic transmission of colour or
ous AE. However, the success depends on the particular black and white images to a video monitor. The advantage of
material tested. Some materials are quiet almost up to the point these devices is their length. Due to light attenuation fibre-
of failure. Slow, continuous crack growth mechanisms such as optic boro-scopes are limited in length, but some video-scopes
active path corrosion are not detectable in themselves, but if can travel as far as 30 m to the inspection site. Other advan-
general yielding has not occurred, they may be detectable tages of the video-scope compared to the fibre-optic boroscope
through associated plastic zone growth. are that the display can help reduce eye fatigue, there is no hon-
eycomb pattern on irregular picture distortion and the elec-
The main advantages of the method are that it allows large area tronic form of the image signal allows digital image
coverage, growing discontinuities can be detected, location of enhancement and the potential for integration with automatic
discontinuities is possible, inaccessible area inspection, on- inspection systems. It produces generally higher resolution.
line application (even at high temperature) and non-invasive. Video cameras are more sensitive to temperature and will only
However, not all discontinuities emit detectable acoustic emis- generally work at temperatures from 0°C to 45°C.
sion, loading methods must be analysed to ensure they pro-
mote detectable crack growth, many factors can obscure A.12 Shearography
acoustic signals (geometry, materials, construction, noise), Shearography is a promising relatively new NDT method. It is a
discontinuity size not determinable and special sensors have to non-contact laser based interferometry system that is used to detect
be used at high temperature (high temperature sensors are areas of stress concentration caused by anomalies in materials.
available to operate up to 260°C).
The method senses out-of-plane surface displacement of an
A.11 Remote Visual Inspection object in response to an applied load. An laser beam is used to
coherently illuminate the test sample. The light reflected from
Remote visual inspection using equipment such as boro- the test sample is collected by a photo lens and imaged through
scopes, fibre optic boro-scopes and video-scopes is a very use- an image shearing interferometer onto the sensor array of a
ful tool which can give information on the condition of pipes, CCD video camera. The image shearing interferometer pro-
turbines, heat exchangers, and other critical machinery without duces a double image on the CCD array. The second image is
costly and time consuming disassembly. offset from the first one. The interferogram recorded by the
A boroscope is a long, tubular optical device that illuminates CCD camera is thus indicative of motion toward the camera
and allows the inspection of surfaces inside narrow tubes or over an interval of the image shearing distance in the plane of
difficult to reach chambers. the test sample surface, or the relative slope of the test sample
surface. Data are represented in the form of a fringe pattern
Rigid boro-scopes are generally limited to applications with a produced by comparing two states of the test sample, one
straight-line path between the observer and the area to be before and the other after a load is applied.
observed – an orbital scan allows the user to view flaws in a
360 degrees arc. The lengths and diameters of the probes can Shearography can be applied to a variety of materials, includ-
be tailored for optimum results. Typical sizes range in length ing steel, aluminium, plastic and composite. Shearography is a
from 0.15 m up to 30 m and in diameters from 0.9 to 70 mm. remote, non-destructive, non-contact and very fast method.
Magnification is usually 3 to 4 times although magnifications It can perform measurements in hazardous conditions such as
up to 50 times are available. high temperature, plasma and nuclear radiation. Laser Shearog-
Flexible boro-scopes are used when there is no straight pas- raphy is very sensitive to slight changes in surface stain due to
sageway to the point of observation. There are two types of sub-surface flaws. It can map changes in strains to 0.1 micros-
flexible boroscope which are flexible fibre-scopes and video- train at video frame rates. However, Shearography is not cur-
scopes with a CCD image sensor at the end. The flexible fibre- rently used as an NII method within the oil and gas industry.
DET NORSKE VERITAS

Page 67
A.13 Inspection Method Capabilities

Table A-1 Capabilities for Non-intrusive Inspection Methods
Inspection efficiency
Flaw type Inspection method (Low, Medium, High)
Detection Through wall Sizing
General wall thickness loss
UT Pulse-echo 0 deg. compression wave H H
UT mapping (C-Scan) H H
UT EMAT M M
Phased Array H H
M-Skip H H
Pulsed Eddy Current H M
Film Radiography M L
Computed Radiography (tangential on pipes) M M
Magnetic Flux Leakage L -
Saturated Low Frequency Eddy Current L -
Local wall thickness loss, pitting
UT Pulse-echo 0 deg. compression wave H M
UT mapping (C-Scan) H H
UT TOFD H H
Phased Array H H
M-Skip H M
Film Radiography H L
Compton Backscatter H L
Computed Radiography (double wall with computerised analysis of H M
image grey levels)
Magnetic Flux Leakage M L
UT CHIME H L
Pulsed Thermography M L
UT Long Range (Lamb Wave) M -
Shearography L -
Saturated Low Frequency Eddy Current H L
Blisters and embedded horizontal cracks, delamination
UT Pulse-echo 0 deg. compression wave H H
UT mapping (C-scan) H H
Phased Array H H
Shearography L -
UT TOFD H H
Compton Backscatter M L
Acoustic Emission L -
Surface breaking cracks
Eddy Current ACFM (inspection surface only) M M
Phased Array H H
Liquid Penetrant (inspection surface only) H -
Magnetic Particle Inspection (inspection surface only) H -
UT Pulse-echo Shear wave (backwall & frontwall) H M
UT TOFD (backwall & frontwall) H H
Guided Wave M -
Acoustic Emission (backwall & frontwall) M -
Embedded cracks
UT Pulse-echo Shear wave H M
UT TOFD H H
Phased Array H M
Film Radiography M -
Computed Radiography M -
Acoustic Emission M -
DET NORSKE VERITAS

Page 68
Table A-1 Capabilities for Non-intrusive Inspection Methods (Continued)

Inspection efficiency
Flaw type Inspection method (Low, Medium, High)
Detection Through wall Sizing
Embedded volumetric voids
UT Pulse-echo 0 deg. compression wave M -
UT Pulse-echo Shear wave H L
UT TOFD H H
Phased Array H L
Film Radiography H L
Shearography L -
Computed Radiography H M
Anomalies
Thermography M L
Weld root erosion
UT TOFD H H
UT Pulse-echo Shear wave M M
UT Pulse-echo 0 deg. compression wave (if weld cap removed) H M
Phased Array M M
UT mapping (C-Scan) (if weld cap removed) H H
Film Radiography (double wall method) M L
Computed Radiography (tangential on pipes) H M
- = Not applicable
DET NORSKE VERITAS

Table A-2 Capabilities of Non-intrusive Inspection Methods
Inspection Method Wall Material Tempera- Surface Finish Sensitivity / Accuracy Productivity Method Flaw Flaw type Vessel fea- Access Limitations / Comments
thickness ture min. detecta- / Repeat- Maturity location ture appli- restriction Testing req.’s
[mm] Range ble flaw ability cation
1. UT Pulse- 6-300 Reduced Up to uniform coat- 1 mm-5 mm ±3mm Weld 20m/ high ID Surface Capability small probe Well-bonded coating
echo Shear capability 250°C ing up to 1.5 depending on mean day POD and embed- for weld can be allowed.
wave for auste- using spe- mm geometry error for values ded crack, inspection Vessel features such as
nitic and cial Roughness weld pub- volume depends on supports, saddles and
duplex probes 6.3μm max. inspec- lished geometry reinforcement plates
welds Free of scale, tion and UT cause non-inspectable
slag, rust, oil procedure areas for spot inspec-
and grease at tion.
probe location.
2. UT Pulse- 2-300 Reduced Up to uniform coat- 0.5mm WT digital spot 1 000/ high ID/OD general and small probe Well-bonded coating
echo 0°.com- capability 250°C ing up to 1.5 typical, thickness day POD local WTL, can be allowed.
pression for auste- using spe- mm depending on gauge slow values volume, Vessel features such as
wave nitic and cial Roughness thickness ±0.1mm pub- blistering, supports, saddles and
duplex probes 6.3μm max. ideal, lished reinforcement plates
welds Higher Free of scale, ±0.5mm cause non-inspectable
tempera- slag, rust, oil typical areas for spot inspec-
tures pos- and grease at tion.
DET NORSKE VERITAS
sible for probe location.

spot
checks
3. UT corrosion 2-300 As pulse- Up to As pulse-echo 0.5mm WT ± 0.1 mm 2-3 m2, High ID local WTL, Surface probe and Patches of 250 × 500
mapping echo UT 125°C UT typical, ideal 8-12 m2 paint- POD blistering encoder or mm per scan. Coloured
C-Scan 0° (longer depending on ± brush, values mecha- WT map.
compression contact) thickness 0.5mmty Up to 16 m2 pub- nized vehi- Post inspection data
wave pical mechanized lished cle on manipulation when
magnetic Pscan is used.

enforced
wheels
4. UT TOFD 8-300 Restricted As pulse- As pulse-echo 3mm depend- ±0.5 mm Weld 40m/ high ID, Embedded corner scooter or Fast weld scanning.
to fine echo UT UT ing on geome- ideal, day POD embed- and sur- welds, guiding belt Post inspection data
grain try ±2 mm values ded face crack, complex interpretation. Dead
material typical pub- volume geometries zone up to 2 – 4 mm
lished below the surface
5. UT CHIME Up to 40 As pulse- Current As pulse-echo Min. c. 10% of 10 mm on Scan speed c. Medium no ID/ general and
reinforce- manipula- No quantitative signal,
echo UT applica- UT WT. lateral position 1m/min. POD OD dis- local WTL,ment tor and degree of damage can
tions up to dimension > values crimina- cracking plates, sad- probe be assessed.
c. 60°C 15 mm pub- tion dles, sup-
lished ports. Also
shell
screening
6. UT Long All pipes Normally -25 to Similar to Min. 5 – 9% of Typically 1 Medium no ID/ local WTL, Pipes – typ- Probes form Works best on straight
Range from 2 to carbon 125°C. pulse-echo UT. pipe wall cir- km /day POD OD dis- cracking ically ring around pipes. Probe ring needs
(Lamb 48 inc steel pipes Special Bare metal cumferential values crimina- straight pipe. to be at leasty1 m from
Wave) diameter only equip- best. area pub- tion (few Access to nearest girth weld.
ment up lished bends) 0.5 m of Some pipe coatings
to160°C bare pipe limit range of method
Page 69
needed (e.g. Bitumasic)
Page 70
Table A-2 Capabilities of Non-intrusive Inspection Methods (Continued)
7. UT EMAT 10-150 As pulse- -200°C to Rough surface As pulse-echo As pulse- As pulse-echo low ID As pulse- As pulse- small probe Sensitivity much lower
echo UT 460°C acceptable in UT under echo UT UT echo UT echo UT than conventional UT
principle but ideal circum- probes. Main applica-
reduces sensi- stances, typi- tion is for wall thickness
tivity even fur- cally much measurements on mag-
ther poorer netite coated compo-
nents,
Non-contact UT possi-
ble
8. Eddy Cur- N/A. All up to coating Crack depth > ±3mm Medium High OD only crack welds, sur- small probe Multi element surface
rent ACFM 150°C allowed with 1 mm and undercoat- face lim- scan coils, possible
using spe- restrictions length >10mm ing ited crack depth measure-
cial ideal, depth ment.
probes >3mm and Rust and ferromagnetic
length >20mm debris may influence the
typical inspection. Flaw free
sample needed for cali-
bration.
DET NORSKE VERITAS
9. Pulsed Eddy 6-60 low alloy -100°C to non contact, 5% WT Accu- 1,000 points/ medium When general surface probe 100 - Detect general corro-
Current (through C steel, 500°C through insula- (relative racy 5% day POD insu- WTL, vol- spot 250 mm sion or erosion if area
max. 150 restricted contact tion measurement) WT values lated, no ume exceeds (500 mm2), no
mm insu- by ferro- tempera- Repeata- pub- ID/OD pitting, influence of
lation) magnetic ture max. bility 2% lished discrim- nearby metal mass
sheeting 70°C WT ination
10. Saturated Up to 30 - all, wall up to 60°C non contact, 15% WT 10 – 50 m2 / Medium ID/OD local WTL, Surface heavy Fast large area scan-
Low Fre- 35 thickness max. 8 mm day discrim- volume device due ning. Down to1.4 mm3
quency Eddy for ferro- coating ination to magnets volume wall loss can be
Current magnetic, possible detected under favoura-
reduced ble conditions
capability
for other
but same
as EC.
11. Liquid Pene- N/A. all, non- 10°C to Free of scale, Anything visi- - 20 m per day high OD only surface all geome- N/A. Very clean surface
trant porous 50°C slag, rust, oil, ble POD crack try’s needed. Less reliable
(incl. dye grease or paint. values than MPI for ferromag-
and fluores- pub- netic material
cent) lished
12. Magnetic N/A. ferromag- up to Free of scale, Anything visi- - 40 m per day high OD only surface depending yoke size Contrast paint can be
Particle netic 100°C slag, rust, oil, ble POD crack on yoke 300 mm used. Dry test ‘ink’ for
Inspection grease or paint. values high temperatures.
(incl. fluo- Smooth sur- pub-
rescent) face increases lished
reliability
13. Magnetic 4-10 ferromag- up to max. 3 mm min. 30% WT - 10 –50 m2 / High no ID/ local WTL Surface medium Fast large area scan.
Flux Leak- netic 60°C coating, clean or 20mm3 day POD OD dis- size device Wall and pipescan, not
age surface values crimina- 300 × 300 floorscan
pub- tion mm
lished
Table A-2 Capabilities of Non-intrusive Inspection Methods (Continued)
14. Pulsed Ther- Surface N/A. Non-con- Depends on the Temperature - Depends on high OD, anomalies Surface N/A. Screening for anoma-
mography tact surface emis- variations of application screen- lies. Depends on the sur-
method, sivity 0.2ºC at 1 m; ing for face emissivity of the
-20 to 10°C at 100 internal specimen and the dis-
1000°C m. flaws tance. Filters above
500°C. Line-of-sight
required.
15. Film Radiog- Ir 192 Up all max. 40°C N/A. 2%WT slow depends high ID/OD local and critical WT 2 sided Double wall technique
raphy to c.100 on access and embed- general locations, access limited to vessels diam-
pene- radiation ded WTL, welds, noz- eter below 1.5 m. Tan-
trated safety regula- embedded zles, gential technique, on
thickness. tions. Can be volumetric stream possible. Radia-
Higher speeded up by flaws tion safety restrictions.
with use of SCAR
Co60. or SafeRAD
containers
16. Computed –Similar all N/A. 2% WT Generally medium ID/OD local and As Film 2 sided On stream. Radiation
Radiography to film faster than general Radiogra- access, safety restrictions.
DET NORSKE VERITAS
radiogra- film radiogra- WTL phy manipula-

phy phy subject to tor move-
sensitivity. ment
Depends on
access and
radiation
safety regula-
tions
17. Acoustic All Normal As 1 and 2 Only detects Informa- whole vessel medium ID/OD propagat- Whole ves- small extremely sensitive to

Emission maximum local to probes growing flaws tion on embed- ing cracks, sel probes environmental influ-
60°C position ded crack/cor- ences
rosion initi-
ation
Remarks:
Material: Carbon steel, stainless steel (including Inconel), duplex
Productivity: speed under normal conditions
Method maturity: experience with method, (low/ medium/ high) and POD values known
Flaw location: OD is detect indications on far side, ID is near (probe) side, discrimination of ID/OD flaws possible, embedded flaws can be found
Flaw type: damage description, main character of method. WTL is wall thickness loss. Local WTL is pitting. Cracks can be surface breaking or embedded.
Vessel feature application: type of features suitable for inspection. Welds: butt, corner or complex geometry. Surface with full area access (inspected area is equal to probe access area). Volume means inspected material volume
and probe contact location are not necessarily the same
Access (restrictions): seen from the NDT method, the required access area. In most cases determined by the probe size (detached units).
POD References:
1. “Reliability assessment for containers of hazardous material RACH”, HSE Offshore Technology Report, 2000/095.
2. “Reliability of liquid penetrant and magnetic particle inspection", by P. Kauppinen and J Sillanpaa, Proceedings of the 10th International Conference on NDE in the Nuclear and Pressure Vessel Industries, 1990, Ed M J Whittle,
J E Doherty and K Iida, ASM International.
3. "Selection of NDT for FFP Final Report for Study 1", T Bouma et al, 1996 NIL summary report (no. FF 96-43)
4. "Reliable ultrasonic inspection in theory and in practice: sizing capability of time-of-flight diffraction", J A G Temple, 3rd European Conference on Non-destructive Testing, Florence, 15 - 18 October, 1984
5. W D Dover & J R Rudlin, 1997, "Inspection reliability trials for offshore applications", pp 225-236, Proceedings of an European-American Workshop on Determination of Reliability and Validation Methods on NDE, Berlin,
Germany, 1997, published by DGZfp
Page 71
6. G A Raine and M C Lugg, 1996, "ROV inspection of welds - a reality", Insight, Vol 38, no 5, pp 346-350.
7. PISC II Report No. 5 "Evaluation of the PISC-II trials results", Final Issue dated September 1986
Page 72
APPENDIX B
STATISTICAL APPROACHES TO NII
B.1 Use of Statistical Methods in NII performed. An inspection with a 100% POD of defects larger
than the acceptable limit would reduce the probability failure
B.1.1 Background to zero. Hence the probability distribution, as defined by p(x),
A range of statistical methods can be used to help quantify var- before and after the inspection might appear as shown in
ious aspects of the NII inspection procedure. This appendix Figure B-1.
presents some of the principal methods that can be used to Note that in practice there would be some modification of the
assist in inspection scheduling, as well as the evaluation of
inspection data. curve for flaws smaller than the acceptable size due to impre-
cision in the inspection method giving rise to false calls. This,
The first section discusses a statistically based approach to set- however, will not be considered here and it will be assumed for
ting the inspection interval in order to ensure safe operation. the present purposes that the inspection method has a 0% false
Two examples of statistical methods which can be useful tools call rate.
for evaluation of inspection data from non-intrusive inspection
are also provided.
The first method is extreme value statistical analysis, which
p(x)
can be useful for predicting what the maximum corrosion pit After inspection
depth in a vessel zone is, based on a sample inspection for pit-
ting within that zone. It can also be used to help determine
Before
what the sample size should be in order to achieve a particular
level of confidence in the conclusions. Area under curve defines
The second example covers the use of statistical methods to

determine what the probability of detecting at least one flaw Defect size
(and therefore revealing that a particular degradation mecha- Acceptance limit
Critical
nism is active) as a function of sample size and how wide-
spread the degradation is. Figure B-1
Effect of Inspection with POD = 1 on Probability Distribution
B.2 Statistical Basis for Inspection Scheduling
Traditionally, inspection intervals are often specified on the Few inspections will provide a 100% POD hence the case
basis of legislative requirements. More recently however, risk
considerations have been taken into account. If non-intrusive shown in Figure B-1 is not typical. More commonly the POD
inspection (NII) is to be used in conjunction with or as an alter- will be less than one and may be a function of flaw size. Hence
native to IVI then a sensible basis for the determination of the inspection will have the effect of reducing in some way the
NII requirements would be to ensure that the risk levels are not probability of defects larger than the acceptable size rather
increased. For any given vessel this effectively means that than eliminating it. A more typical case is thus as illustrated in
there should not be an increase in the probability of failure Figure B-2.
(POF) when NII is used.
The approach suggested here aims to compare the nature of
p(x)
evolution of the probability of failure under different assumed After inspection

inspection regimes. This comparison is then used in develop-
ing a basis for the determination of an acceptable NII strategy.
Before inspection
B.2.1 Probability of Failure Area under curve defines POF

The probability of failure is determined as the probability that
a defect will exist with a size larger than the critical size. Given Defect size
Acceptance limit
a known probability distribution for the sizes of the flaw pop- Critical
ulation and a known critical defect size, the probability of fail- Figure B-2
ure can be calculated as the area under the distribution curve Effect of Inspection with POD < 1 on Probability Distribution
for defects larger than the critical size.
In many cases the critical defect size will not vary with time,
however, the probability distribution will typically be time For a probability distribution before the inspection given by
dependent as a result of flaw growth by a variety of potential pb ( x) and a probability of detection defined by POD( x) ,
mechanisms. The nature of growth with time therefore has a the probability distribution after inspection is given by
direct influence on the probability of failure. With growth,
there is a greater probability of defects exceeding the critical
size hence the probability of failure will tend to increase with pa ( x ) = pb (x ) for x ≤ xaccept
time.
(1)
Inspection, aimed at identifying and sizing flaws, along with pa (x ) = (1 − POD(x )) ⋅ pb (x )
for x > xaccept
appropriate action (e.g. repair, rejection) taken when defects
exceed a certain size will have the effect of reducing the prob-
ability of failure. This is a consequence of the probability dis- The probability of failure before the inspection is given by
tribution being modified for defect sizes larger than a defined
acceptable limit. ∞
The nature of the reduction in probability of failure is directly

related to the probability of detection (POD) of the inspection
POFb = ∫
xcrit
pb ( x)dx (2)
DET NORSKE VERITAS

Page 73
and after inspection by approximated by

∞
f i = p ( xi )δ x (6)
POFa = ∫ (1 − POD( x)) p ( x)dx
xcrit
b (3)
In order to determine the flaw size distribution S(x) at the end
Clearly, the greater the POD of the inspection, the lower the of the period, the growth of the entire population of flaws must
probability of failure after the inspection. This lower probabil- be considered, this being achieved by summation of the indi-
ity is retained in part up to the time of the next inspection vidual distributions as follows
through the modifying effect on the probability distribution.
The way the latter evolves between inspections however
depends very much on the nature of flaw growth in addition to S ( x) = ∑ F ( x, xi ) = ∑ p( xi )G ( x, xi + μ , σ )δ x (7)
i i
its form directly after the inspection. Growth means that the
flaw distribution is time dependent and hence so is the proba- In the limit as δ x → 0 , the above becomes
bility of failure, which will tend to increase during time inter-
vals between inspections. The shorter these intervals (i.e. the ∞
greater the frequency of inspection) the lower the probability S ( x) = ∫ p ( y )G ( x, y + μ , σ ) dy (8)

of failure will be. In addition to being a function of flaw growth 0
in between inspections, the form of the flaw distribution after With no inspection at the end of the period, the starting distri-
an inspection is also dependent on the probability of detection bution for the next period is given directly by S(x) and this can
and the margin between acceptable and critical defect sizes. be used in the integration of equation (2) to obtain the distribu-
Hence it can be concluded that the probability of failure is pri- tion at the end of the period.
marily a function of time (which defines the growth proc-
esses), the inspection interval, the probability of detection of When an inspection is carried out, the distribution before the
the inspection and the margins between acceptable and critical inspection is S(x) and after the inspection it is modified by the
defect sizes. reduction in the probability of defects having size larger than
acceptable, as follows
Since the growth processes are very often dominated by the
equipment design and operating conditions, control over the S a ( x) = S ( x) for x ≤ xaccept
probability of failure is typically only possible through varia- (9)
tion of the inspection interval, the probability of detection and
S a ( x) = (1 − POD ( x )) S ( x) for x > xaccept
the margins between acceptable and critical defect sizes. The
definition of acceptable defect size is often out of the opera- The above values are then used to define the flaw distribution
tor’s hands hence safe management relies largely on a suitable at the start of the next time period. By recursively updating the
specification of inspection interval and method (as the latter flaw distribution in this way, the probability of failure can be
defines the POD). tracked throughout the life of a vessel and the effects of differ-
In order to gain some insight into how the interval and POD ent inspection strategies (e.g. variations in the interval between
may affect probability of failure it is useful to consider some inspections and the probability of detection) can be studied.
representative examples. Figure B-3 shows some typical results which illustrate the
The case to be considered assumes a growth process in which effect of variations in the inspection interval and the probabil-
the growth rate is, on average, constant but allows for some ity of detection.
variation in rate as defined by a normal distribution. This The uppermost solid curve shows the situation when there is
would be typical of corrosion type processes. no inspection performed and, as expected, the probability of
In this scenario, the growth of any flaw over a specified time failure increases continuously with time. Moving downwards,
interval is given by a normal distribution, this being defined by the next solid curve shows the evolution of POF for the case of
inspections with POD=0.5 at every tenth time period. The
⎛ ( x − μ )2 ⎞ reduction in POF following each inspection is clearly visible.
⎜ ⎟
G ( x, μ , σ ) =
1 ⎜ 2σ 2 ⎟
⋅ e⎝ ⎠ (4) This reduction is observed to nearly match the rise in POF over
2πσ the time since the preceding inspection hence, on average, the
POF is retained nearly constant under this inspection strategy.
where The figure shows that a similar average POF can be maintained
by adopting an alternative strategy in which the inspection
μ is the average growth
interval is reduced to 5 time periods and the POD also reduced
σ is the standard deviation of the growth (to 0.3).
For flaws whose initial sizes are in the range xi to xi + δx, the The curves for POD=0.5 with an interval of 5 time periods and
size distribution at the end of the specified interval is given by POD=0.8 with an interval of 10 time periods show, as
expected, that the effect increasing POD is to lower the POF.
F ( x, xi ) = fi G ( x, xi + μ , σ ) (5) Again it is clear that a given nominal POF at any time can be
achieved by variation of either POD or inspection interval.
where fi is the proportion of the total flaw population made up There will however be differences within the time between
of flaws covering the initial size range. For an initial flaw pop- inspections and a very long inspection interval may lead to an
ulation distribution given by the continuous function p(x), the unacceptably high POF even if preceded by a high POD
proportion of flaws covering the range xi to xi + δx can be inspection.
DET NORSKE VERITAS

Page 74
1.0E+00
1.0E-01
1.0E-02 No Inspection
1.0E-03
POD=0.5 Int=10
1.0E-04
POD=0.5 Int=5
POD=0.3 Int=5
1.0E-05
POD=0.8 Int=10
1.0E-06
1.0E-07
0 10 20 30 40 50 60 70
T i me
Figure B-3
Effect of Inspection Interval and POD on Probability of Failure
B.2.2 Practical considerations does not take risk into account in a systematic way. Risk based
In practice, due to a large number of uncertainties, it will often inspection (RBI) is becoming increasingly accepted as offering
be very difficult to quantify the probability of failure, as out- a technically sound yet cost effective means of managing plant
lined above, with sufficient confidence. Hence such an safety. Risk is determined as the product of the probability of
approach may not always be adequate in providing reliable failure and the consequences of failure and RBI aims to ensure
guidance for the definition of inspection interval and type. that risk is consistently maintained at an acceptably low level.
There is however a need that these be defined in a systematic This can be achieved by a variety of means but the primary var-
manner and this is addressed in a variety of codes of practice iables are inspection interval and inspection type. When inter-
(see for examples Refs. [10, 11, 22]) and many organisations nal visual inspection is relied upon, control of risk comes down
have developed their own guidance documents. to specification of an appropriate inspection interval, i.e. high
risk items demand a short interval while longer inspection
Given that internal visual inspection has historically been the intervals are acceptable for low risk items.
primary method, most industry accepted approaches a directed
at this type of inspection. However, the principles remain Different guidelines have been developed for the specification
applicable to the specification of NII. A common approach is of inspection interval according to risk. In many cases (e.g.
to base the inspection interval for IVI on the Inspection Grade Refs. [23, 24]) the interval is not specified according to risk
assigned to the vessel. Guidance on the determination of alone but also takes into account the inspection grading. The
Inspection Grade may vary according to the code used how- interval is then defined according to the combination of risk
ever it effectively depends on: ranking and inspection grade. Typically, rules might be as
shown in Table B-1. (Note that these rules are not generic and
— the history of previous inspection intervals may vary for different applications, risk ranking definitions
— the confidence in knowledge of the degradation rates etc.)
— the maximum degradation rate considered possible.
Table B-1 Inspection intervals (months) according to
Grade 0 is the most severe and demands the shortest inspection
interval and the interval increases with each Grade until it is a Criticality and Grade
maximum for Grade 3. In order for a vessel to advance a grade Inspection Grade
it should typically have at least one inspection at the preceding Criticality
grade. This is where the history of previous inspection inter- 0 1 2 3
vals plays a role in determining grading. 1 12 36 N/A N/A
Reduced confidence in knowledge of the degradation rates has
the effect of reducing inspection grade (and hence also short- 2 24 36 72 N/A
ening the inspection interval). This effectively takes a conserv-
ative approach to reducing probability of failure when faced 3 24 48 72 96
with uncertainty. 4 36 48 84 120
Likewise the inspection grade reduces with increasing degra-
dation rate. This is aimed at ensuring the probability of failure 5 36 48 96 144
remains more or less constant, even for different deterioration
mechanisms. In the table Criticality defines the risk ranking with
Specification of inspection interval according to grade only Criticality = 1 being the highest and Criticality = 5 being the
DET NORSKE VERITAS

Page 75
lowest risk rankings respectively. In the above rules, it would degradation present. If the item is now inspected, using a
be specified that Criticality 1 items cannot be assigned a grade method with a 90% probability of detecting degradation, with
higher than 1 and Criticality 2 items cannot be assigned a grade no false calls, and no degradation is reported, Bayes’ theorem
higher than 2. can be used to determine the revised probability that there is in
Items having high risk and low inspection grade (top left) have fact no degradation.
short inspection intervals while low risk, high grade (bottom In this case the event “B” is that no degradation is reported, and
right) items have longer intervals. Taking into consideration the “cause” Aj is no degradation present. However another pos-
the combination of risk and grade in this way may be regarded sible cause is that degradation is present (but missed).
as doubly accounting for risk. This is true only to the extent The notation p(A | B) means “the probability of A, given B”
that the grade is a measure of the probability of failure. How-
ever, grade does not consider the consequence of failure in any According to Bayes’s theorem :
way. In addition, risk may not, in general, consider the history (probability no degradation present | none reported) =
of previous inspection intervals. Hence taking a combination
of risk and grade into account does allow for a more complete (prob. none reported | none present) × (prob. none present)
basis for the specification of inspection interval and results in
greater differentiation in the requirements for different vessels. ————————————————————————
In a sense using the combination refines the scale of interval [(prob. none reported | none present × prob. none present)
that can be specified compared to the case when only one or the + (prob. none reported | present × prob. present)]
other is considered.
It is important to recognise that the type of inspection has to be =
considered in determining risk. An inspection having high 1 × 0.75
effectiveness will lead to a reduction in risk and vice versa. ___________________
In most cases inspection intervals will already have been spec-
ified assuming that IVI will be carried out. Once it has been (1 × 0.75) + (0.1 × 0.25)
determined that NII is possible and applicable at a given
inspection, the interval following the NII can then be adjusted = 0.968
taking into consideration the inspection effectiveness. i.e. the confidence that no degradation is present has therefore
The flow charts set up for use as guidance in selecting an increased from 75% to around 97%.
appropriate NII method assign efficiency rankings to each Bayes’ theorem can also be used to support decisions based on
method according to the application. The ranking is defined expert judgement even where precise quantitative data on
such that, for areas covered by the inspection: inspection effectiveness or likelihood of degradation does not
High the NII method has a markedly higher POD than IVI exist. For example expert judgement could be used to assign a
Medium the NII method has broadly similar POD to IVI category to inspection effectiveness such as “very high”,
“medium high”, “medium” etc. where each category is taken
Low the NII method has a markedly lower POD than IVI. to represent a band such as “better than 90%”, “between 75%
In keeping with the goal of maintaining a constant long term and 90%” etc. A similar approach could be adopted to describe
probability of failure, the decision process uses the effective- initial likelihood of degradation (or degradation rate). The
ness of the previous inspection as an input to determining the average within each band (e.g. 95% for very high) could then
minimum efficiency required of the current inspection. Credit be used to apply Bayes’s theorem. Depending on the band
can be taken for a high effectiveness inspection by allowing a within which the “answer” lay, the revised likelihood of deg-
low effectiveness for the following inspection (assuming of radation (or degradation rate) could then be interpreted in the
course that other factors considered allow it). Alternatively, an original linguistic terms (very high, high etc.)
inspection having a low effectiveness may need to be followed Note that care should be taken when using probability of detec-
by one having a high effectiveness in order to maintain the tion (POD) data in Bayes’ theorem in cases where a number of
desired probability of failure. discrete flaws (e.g. cracks) may be present. A method with a
Given that at this stage of the project the timing of the inspec- POD of 90% will have a 90% probability of detecting each
tion is taken as given, all that can be specified is the method individual flaw, but a much lower probability of detecting all
efficiency requirement. It is clear, however, that control of the of the flaws.
probability of failure could be more effective by allowing for
changes in interval according to the method used. This would B.4 Extreme Value Statistical Analysis
make for greater flexibility in developing and adapting an NII There is a common requirement to monitor the condition of
strategy to meet plant specific requirements. For example, it plant for reasons of safety or in order to plan the repair or
could allow the user to specify more frequent lower efficiency replacement of plant components during their working life. For
(and perhaps lower cost) NII while still maintaining a compa- example a chemical plant may have several vessels that require
rable probability of failure to what would be achieved with IVI examination to assess the severity of internal corrosion
at longer intervals. present. This could be done by carrying out a detailed inspec-
tion of the inside of a vessel, and mapping the areas of corro-
B.3 Bayes’ Theorem sion. This would inevitably mean that the plant would need to
Bayes’ theorem provides a logical way of updating incomplete be shut down in order to carry out such an inspection. An alter-
knowledge (e.g. 75% confident no degradation present) based native would be to carry out an ultrasonic thickness survey of
on a test or observation which itself may not be 100% reliable the entire vessel from the outside surface. The advantage of
(e.g. inspection method detects degradation 90% of the time). using such a method would be that the plant may not need to
be opened or even shut down in order to carry out the inspec-
Bayes’ theorem states that given that event B has occurred, the tion. However, the vessel may be large, requiring a long time
probability that it was due to cause Aj is equal to the probabil- to carry out the ultrasonic survey. Another way to assess the
ity that Aj should produce that event times the probability that corrosion present would be to carry out a survey on a repre-
Aj should occur in the first place, all divided by a scaling factor sentative area of the vessel and use statistical techniques to pre-
which is the sum of such terms over all causes. dict the condition of the entire vessel from the results of the
For example, if it is assumed that before inspection of a partic- representative area. The advantages of using this technique
ular plant item, there is a 75% confidence that there was no would be a saving in time and therefore cost of inspection.
DET NORSKE VERITAS

Page 76
The corrosion process can be complicated, with several corro- be classified into two broad categories: uniform and non-uni-
sion mechanisms taking place simultaneously. To fully under- form (or localised) corrosion. Illustrations of these two catego-
stand the mechanisms taking place can require extensive ries are shown in Figure B-4.
testing and analysis in the laboratory. However, corrosion can
Figure B-4
Corrosion Categories
When a statistical sample is taken a distribution of results is With a non-uniform corrosion process we would be much
produced from which certain things can be measured, such as more interested in the most extreme values from our ultrasonic
the mean and the variance (or Standard Deviation) of the dis- survey, as relying on average or mean values may lead us to be
tribution. If we applied this method of sampling to our vessel too optimistic in our assessment of plant condition. The branch
corrosion problem we could gain valuable information about of statistics which deals with the analysis of these extreme val-
the severity of corrosion within the vessel and we would also ues is “Extreme Value Statistical Analysis” (Ref. 25).
have quantitative information with which to compare previous
and future inspections. This type of analysis is useful where the This statistical analysis will be described by way of an exam-
corrosion within the vessel is known to be uniform. In which ple, where we are required to assess the deepest extent of cor-
case we would be more interested in the average or mean depth rosion on the inside surface of a large pressure vessel using
of corrosion. ultrasonic thickness measurements. We know that the corro-
sion occurring in the inside of the vessel is consistent over the
If we carried out an ultrasonic thickness survey on a vessel entire inner surface but the corrosion is non-uniform in that we
with uniform corrosion we would expect to find a Gaussian or know that severe corrosion pitting is occurring. Therefore we
Normal type of distribution from our measurement data. The can use our extreme value techniques to predict the deepest
Gaussian or Normal type of distribution is illustrated in corrosion likely to be present on the vessel.
Figure B-5. Our measurements would show a spread of results
symmetrically about the mean value. B.4.1 Obtaining the Data
We could use statistical techniques to measure the mean value When carrying out a survey for Extreme Value Statistical
and calculate the standard deviation or variance of the distribu- Analysis we place more statistical relevance to the extreme
tion. This would provide us with a measure of the spread of maximum (or extreme minimum) values measured. In terms of
values in our distribution. a normal statistical distribution we are in effect analysing the
tail of the normal distribution and using statistical techniques
to derive probability functions which we can use to gain quan-
titative information about the most extreme corrosion occur-
ring in the component.
Let us examine the case where we need to carry out a survey of
the non-uniform corrosion pitting (or other severe corrosion
processes) within a large pressure vessel. We know that the
corrosion is non-uniform as there is severe corrosion present
generally over the entire inner surface of the vessel. One way
of obtaining a statistical sample for extreme value statistical
analysis would be as follows:
Figure B-5 — choose a representative area of the vessel, say a convenient
Gaussian or Normal Distribution area of 1 m × 2.1 m. (Note: the area should be “represent-
ative” of the condition of the whole vessel or of the part of
the vessel being considered)
In the case of non-uniform corrosion we would be more likely — split the examination area into sub-areas, say 100 mm ×
to be interested in the deepest extent of the corrosion present, 100 mm squares
as a single through wall pit will cause the component to leak. — carry out a detailed ultrasonic examination of each sub-
If we used the statistical tools we used on the uniform corro- area and record only the most extreme value for pitting
sion we may underestimate the seriousness of the deepest cor- depth measured in each sub-area.
rosion present on the component. A different technique needs
to be used in this case; one where we concentrate our measure- We now have a sample which we can analyse using extreme
ment and statistical techniques on the most extreme depths of value techniques. There are 210 values which we can display
corrosion. in the form of a histogram, see Figure B-6. The x axis measures
Some corrosion processes occur which are non-uniform where increasing depth of pitting and the y axis records quantity i.e.
local areas can be subject to extreme degradation (e.g. pitting). number of readings of this depth of corrosion.
DET NORSKE VERITAS

Page 77
Figure B-6
Histogram of Results of Ultrasonic Thickness Survey (maximum pit depths)
B.4.2 Processing the Data x : Random Variable (corrosion depth)

The histogram showing our results is called a “frequency distri- λ : Location Parameter (= statistical mode)
bution”. Because it has been produced from a process which α : Scale Parameter (related to variance)
concentrates on extreme values however, it does not display the
normal or Gaussian symmetrical pattern. The extreme values Our first step in processing our data must be to check if the data
recorded, tend to trail off at the right hand side of the distribu- we have recorded will be able to be modelled by a Gumbel dis-
tion. In order to process our data we need to find a function tribution function. Processing of the data using PC spreadsheet
which will model our distribution. A distribution function called techniques can make this straightforward, even for large
a “double exponential” or “Gumbel” distribution is commonly amounts of data.
used to model such a “maximum” distribution. We should create a table which contains our results as illus-
The Gumbel distribution function takes the form: trated in Table B-2.
⎡ ⎛ x − λ ⎞⎤ ; - ∝ < x <∝ (1)

F1 ( x ) = exp⎢− exp⎜ − ⎟ (10)
⎣ ⎝ α ⎠⎥⎦
Table B-2 Example data for Extreme Value Statistical Analysis

Local Maxima of Pitting/Corro- Frequency A Probability Density Function Cumulative Distribution Function
sion (number) f = A/(N + 1)* F
(mm) F=Σf
0.2 19 0.090 0.090
0.4 40 0.190 0.280
0.6 43 0.203 0.483
0.8 43 0.203 0.686
1.0 31 0.138 0.834
1.2 15 0.071 0.905
1.4 12 0.057 0.962
1.6 4 0.019 0.981
1.8 2 0.010 0.991
2.0 1 0.004 0.995
Total (N) 210
* Using the “Average Rank Method”. See referenced text.
Table B-2 includes two columns containing useful information “Cumulative Distribution Function” (see Figure B-8). These
on our distribution which we will present graphically: The functions are derived from our recorded data.
“Probability Density Function” (see Figure B-7) and the
DET NORSKE VERITAS

Page 78
Figure B-7
Probability Density Function
Figure B-8
Cumulative Probability Function
In order to check that our Gumbel model is a valid description This has been done for our results presented in Table B-2 and
of the distribution it is necessary for us to plot the function: is shown in Figure B-9.
Y = − ln[− ln (F1 ( x ))] (11)
DET NORSKE VERITAS

Page 79
Figure B-9
Verification of Gumbel Function
Figure B-9 clearly shows a good linear relationship for the determined from the graph;. λ is the slope of the graph and α
plotted function Y. This linearity indicates that the assumption corresponds to the intercept at y = 0. Knowing parameters λ
of a Gumbel function being a good statistical model is valid for and α allows us to construct our model of the distribution using
our data. The parameters λ and α from equation (10) can be the Gumbel function, see equation (10).
Figure B-10
Comparison of CDF with Gumbel Function
Figure B-10 shows a comparison of our Gumbel function with many examples of its application.
F(x). If our model matches closely our data plot then we can be
reasonably confident in our statistical model and the probabil- B.4.3 Effect of Paint
ities we draw from it. In this case, the graph shows a reasona-
bly good match, and the results of our statistical analysis can Care should be taken when applying extreme value statistical
be used to predict the probability of there being a particular analysis to ultrasonic thickness data measured through painted
maximum depth of corrosion present in the vessel (or area of surfaces. Even if the paint is well adhered, there are likely to
the vessel which the sample represents). be slight changes in paint thickness. Since the velocity of ultra-
sound in paint is typically 40% of the velocity through steel, a
If we use the cumulative probability graph (Table B-8) this variation in paint thickness of 0.4 mm within the sample may
graph illustrates the probabilities of finding corrosion of a par- be misinterpreted as a variation of 1mm in steel thickness if a
ticular maximum depth. For example there is a 99.5% proba- correction is not applied. If extreme value statistical analysis is
bility that the deepest corrosion present is 2 mm deep. Or
conversely, there is a 0.5% probability of there being corrosion then applied to this uncorrected data, the maximum variation
greater than 2 mm. in component wall thickness in the un-inspected part of the
zone could then be predicted as being significantly greater than
The use of this technique can be a very useful tool in the 1mm, even if there was no corrosion or erosion and the thick-
assessment of non-uniform corrosion, especially when quanti- ness of the steel component was perfectly uniform.
tative information is needed quickly and efficiently. Extreme
value statistical techniques can be used for a wide range of Various commercially available instruments are available for
applications and there is good reference material available measuring the thickness of paint coatings including the
which describes the theory behind the technique and provides “banana gauge” which works on a simple magnetic principle.
DET NORSKE VERITAS

Page 80
B.5 Effects of Sample Size B.5.1 Analysis Methods

It is important to understand whether inspection will be able to Assume a weld has a uniform random distribution of flaws
find at least one instance of any flaw present. Clearly this will within its length. This random distribution can be represented
depend on the population of such flaws, the effectiveness of as a percentage of defective weld structure. If a flaw will be
the inspection method and the sample size. For localised crack- detected when the region it is in is inspected (assumes 100%
ing, spot checks are inappropriate and while sample inspec- detection capability), the problem is to determine the probabil-
tions may help to determine whether a certain type of cracking ity of detecting at least 1 flaw when only inspecting a percent-
mechanism is active (Ref. 26), they are unlikely to provide use- age of the weld volume (and therefore e.g. being alerted to the
ful information on how many cracks are present in the zone or fact that a particular degradation mechanism is active)
what their sizes are.
If a vessel zone is genuinely homogeneous, then statistical POI: Probability of Inclusion of a Flaw
methods may be useful in determining sample size, and in Cov: Percentage Coverage of the Weld Volume.
extrapolating the inspection results from the sample to the Dd: Flaw Distribution as a Percentage.
whole of that zone. The confidence in the validity of sample
inspections depends on the size of the sample compared to the To assist in calculating the POI, it is assumed that the weld vol-
size of the zone. The larger the sample size, the higher the con- ume is divided into 100 discrete units. The probability that a
fidence in the extrapolated results. given unit contains a flaw is: Pd = Dd / 100%.
The validity of spot or sample inspections also depends on the Probability that a unit contains no flaw: P0 = 1 - Pd
degradation mechanism and the population of flaws present. POI = 1-PON where PON is the Probability of Detecting No
For a zone susceptible only to uniform corrosion, a few spot Flaws
checks of wall thickness may provide highly reliable informa-
tion on the extent of degradation in the whole of that zone. Using the above definitions and assumptions a number of mod-
Repeat measurements at the same locations during future els could be used to determine the POI:
inspections can then be used to assess corrosion rates.
B.5.1.1 Simple Evaluation:
If spot or sample checks reveal unexpected degradation, or
more severe degradation than anticipated, then increased cov- Using the following data evaluation, a formula can be deter-
erage (possibly up to 100%) is likely to be required. mined and extrapolated to 100% coverage.
Pd % 1 2 3 4
P0 % 99 98 97 96
Cov % PON PON PON PON
0 1 1 1 1
99 98 97 96
1
100 100 100 100
99 98 98 97 97 96 96 95
2 ⋅ ⋅ ⋅ ⋅
100 99 100 99 100 99 100 99
99 98 97 98 97 96 97 96 95 96
95 94
3 ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅ ⋅
100 99 98 100 99 98 100 99 98 100 99 98
99 98 97 96 98 97 96 95 97 96 95 94 96
95 94 93
4 ⋅⋅ ⋅ ⋅
⋅ ⋅ ⋅⋅ ⋅ ⋅ ⋅⋅
100 99 98 97 100 99 98 97 100 99 98 97 100 99 98 97
General Formula B.5.1.3 Poisson Distribution:
(100 − Pd )!
(100 − Pd − Cov)! ⎛ e−μ ⋅ μ r ⎞
PON = POI = 1 − PON = 1 − ⎜⎜ ⎟⎟
100!
(100 − Cov)! ⎝ r! ⎠
where
B.5.1.2 Binomial Distribution:
μ = nPd
n = Cov
⎛ n! ⎞ n −r r
POI = 1 − PON = 1 − ⎜⎜ ⎟ ⋅ P0 ⋅ Pd
r )!r! ⎟⎠
r = 0 flaws to detect
⎝ (n −
Graphs for all three analysis methods are presented.
where:
Generally all three graphs exhibit similar trends.
n = Cov
The binomial and Poisson relationship do not provide reliable
r = 0 flaws to detect results for high coverage of welds with a low flaw distribution,
By altering n from 0 to 100% coverage the POI rises from 0 to since the graphs do not predict a probability of 1 at 100% cov-
1 in a curve. erage of the weld with 4% uniform flaw distribution.
A more rigorous treatment, which also addresses the probabil- the reference below.
ity of detection (POD) of the inspection method, is provided in
DET NORSKE VERITAS

Page 81
Figure B-11
Probability of Detecting 1 Defect Given a Percentage coverage of a Defective Weld.
Simple Distribution
Figure B-12
Using Binomial Distribution Model
DET NORSKE VERITAS

Page 82
Figure B-13
Using Poisson Distribution Model
DET NORSKE VERITAS

RP G103 2007

Uploaded by

Copyright:

Available Formats

RP G103 2007

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

RP G103 2007

Uploaded by

Copyright:

Available Formats

RECOMMENDED PRACTICE

DET NORSKE VERITAS

Amendments and Corrections

Comments may be sent by e-mail to [email protected]

Computer Typesetting (FM+SGML) by Det Norske Veritas.

DET NORSKE VERITAS

DET NORSKE VERITAS

1. INTRODUCTION .................................................. 7 5. ON-SITE INSPECTION ACTIVITIES............. 44

DET NORSKE VERITAS

DET NORSKE VERITAS

1. Introduction required and it is hoped to avoid shutting down the plant by

DET NORSKE VERITAS

DET NORSKE VERITAS

Select pressure Integrity Prepare Determine date of

DET NORSKE VERITAS

1.5 Definitions Internal Visual This is considered as an intrusive close visual

DET NORSKE VERITAS

DET NORSKE VERITAS

CRA Type 1 which in turn will influence the effectiveness of inspection

DET NORSKE VERITAS

Is the vessel intrinsically No

Has vessel previously No Was vessel designed No Is vessel the same as No

Yes Yes Yes

Is operating history still

Perform IVI NII is not appropriate

DET NORSKE VERITAS

High High High Yes

Medium Medium Medium Yes

Low Low Low Yes

High High Yes

Medium Medium Yes

Low Low Yes

Medium Medium Yes

Low Low Yes

DET NORSKE VERITAS

DET NORSKE VERITAS

DET NORSKE VERITAS

DET NORSKE VERITAS

Application of the screening process (Figure 3-1): 4. Inspection planning

DET NORSKE VERITAS

Review the risk

No Review the risk No

No Have all zones been

Prepare Work -pack

DET NORSKE VERITAS

4.2 The Planning Team — location of degradation, preferential attack or more

Table 4-1 Inspection Type Definitions

DET NORSKE VERITAS

Table 4-2 Example of Inspection Strategy Grading Process.

DET NORSKE VERITAS

Degradation Likelihood Degradation Extent Degradation Rate Inspection Type

Inspection High General High B

Medium General High B

Low General High A

DET NORSKE VERITAS

Medium and which should be subject to sample inspections. In the case

DET NORSKE VERITAS

Zone Zone Identifier

DET NORSKE VERITAS