Research Report 433: Safety Implications of TOFD For In-Manufacture Inspections

HSE
Health & Safety

Executive
Safety implications of TOFD for

in-manufacture inspections
Prepared by MitsuiBabcock Energy Limited

for the Health and Safety Executive 2006
RESEARCH REPORT 433
HSE
Health & Safety

Executive
Safety implications of TOFD for

in-manufacture inspections
N S Goujon
MitsuiBabcock Energy Limited
Porterfield Road
Renfrew
Renfrewshire
PA4 8DJ
This report describes the results from a project to evaluate the safety implications related to the European
project Effective application of TOFD method for weld inspection at the manufacturing stage of pressure
vessels (acronym: TOFDPROOF).
The TOFDPROOF project was performed by a consortium of European organisations including Mitsui
Babcock.
The project aimed at producing a coherent package of EU documents including procedures for applying
TOFD, acceptance criteria and recommendations for training and certification.
The work involved ultrasonic inspection of a number of test-specimens containing both synthetic and real
service-induced defects (mainly cracks in welds). The project concentrated on detection and sizing trials.
Recommendations based both on the results of this project and the views of the partners of the
TOFDPROOF project are provided and the safety issues related to the results of the TOFDPROOF project
are discussed.
This report and the work it describes were funded by the Health and Safety Executive (HSE). Its contents,
including any opinions and/or conclusions expressed, are those of the author alone and do not necessarily
reflect HSE policy.
HSE BOOKS
Crown copyright 2006

First published 2006
All rights reserved. No part of this publication may be
reproduced, stored in a retrieval system, or transmitted in
any form or by any means (electronic, mechanical,
photocopying, recording or otherwise) without the prior
written permission of the copyright owner.
Applications for reproduction should be made in writing to:
Licensing Division, Her Majesty's Stationery Office,
St Clements House, 2-16 Colegate, Norwich NR3 1BQ
or by e-mail to hmsolicensing@cabinet-office.x.gsi.gov.uk
ii
CONTENTS
DISTRIBUTION LIST.................................................................................................................................. i
AMENDMENT CONTROL.......................................................................................................................... i
CONTENTS .............................................................................................................................................. iii
SUMMARY................................................................................................................................................. v
1.
INTRODUCTION .................................................................................................1
2.
TOFDPROOF PROJECT ....................................................................................2
2.1
TOFDPROOF CONSORTIUM....................................................................................................... 2
2.2
TOFDPROOF WORK PACKAGES............................................................................................... 2
2.3
WORK RESULTS .......................................................................................................................... 4
2.3.1
2.3.2
2.3.3
2.3.4
2.3.5
2.3.6
2.3.7
Work Package WP1:

Work Package WP2:
Work Package WP3:
Work Package WP4:
Work Package WP5:
Work Package WP6:
Work Package WP7:
Trials organisation and justification of representative specimens .......... 4
Blind Trials and Performance Assessment ............................................ 8
Guidelines for training, qualification and certification ........................... 20
Acceptance criteria definition................................................................ 24
Economic analysis ................................................................................ 29
Exploitation and dissemination ............................................................. 33
Data storage, analysis and exchange .................................................. 35
3.
SAFETY RELATED ISSUES ............................................................................36
3.1
INTRODUCTION.......................................................................................................................... 36
3.2
ACCEPTANCE CRITERIA .......................................................................................................... 37
4.
CONCLUSIONS ................................................................................................39
5.
REFERENCES ..................................................................................................41
APPENDICES.......................................................................................................................................... 43
APPENDIX 1: RECOMMENDATIONS FOR APPLYING TOFD (FIELD OF APPLICATION,
STRENGTHS, WEAKNESSES) .............................................................................................................. 44
APPENDIX 2 EXAMPLE OF TOFD ERRORS FOR A 15MM THICK SAMPLE ................................... 59
iii
iv
SUMMARY
This final report describes the work carried out on the HSE funded project Safety
Implications of TOFD for In-Manufacture Inspections.
This report describes the results from a project to evaluate the safety implications
related to the European project Effective application of TOFD method for weld
inspection at the manufacturing stage of pressure vessels (acronym: TOFDPROOF).
The TOFDPROOF project was performed by a consortium of European organisations
including Mitsui Babcock.
The project aimed at producing a coherent package of EU documents including
procedures for applying TOFD, acceptance criteria and recommendations for training
and certification.
The work involved ultrasonic inspection of a number of test-specimens containing both
synthetic and real service-induced defects (mainly cracks in welds). The project
concentrated on detection and sizing trials.
Recommendations based both on the results of this project and the views of the
partners of the TOFDPROOF project are provided and the safety issues related to the
results of the TOFDPROOF project are discussed.
vi
1.
INTRODUCTION
This final report describes the work carried out on the HSE funded project Safety
Implications of TOFD for In-Manufacture Inspections.
Time of Flight Diffraction (TOFD) is an ultrasonic non-destructive testing method
which was originally developed as an accurate method for measuring the throughwall extent of defects which had been detected by more conventional methods such
as Pulse-Echo (PE) ultrasonics.
TOFD differs from conventional PE examinations in that it relies on tip diffraction from
the extremities of the defect for both detection and sizing. With the knowledge of the
wave velocity and the spatial relationships of the two probes, the position of the tips
of the flaw can be calculated very accurately.
The success in this field, combined with cheaper and more portable TOFD hardware,
has resulted in the acceptance within a wide range of industrial sectors (offshore,
petrochemical, chemical, defence, conventional power generation). However, it is
increasingly being proposed as a rapid search tool, not just for defect sizing (i.e. as
an alternative to pulse-echo inspection instead of as a complement to it).
TOFD has been reported as being a technique with high probability of detection for
both planar and volumetric defects and of excellent reproducibility and accuracy.
Although the economic advantages of using TOFD for manufacturing inspections are
clear (inspection speed is typically 3 times faster than conventional PE techniques)
there remain obstacles to its widespread application for general inspections. These
include concerns over detection capability, lack of objective guidance on applicability
and (particularly for manufacturing inspections) lack of agreed acceptance criteria.
To overcome these obstacles a European consortium submitted a proposal to the
European commission for partial funding of a project called TOFDPROOF. The
TOFDPROOF project aimed to develop procedures for applying TOFD, guidance on
applicability, acceptance criteria, and recommendations for training and certification.
The intention is that this will allow vessel manufacturers to use TOFD as a stand
alone method for weld inspection.
This report summarised the work performed on the TOFDPROOF project and covers
the evaluation of the implications of the TOFDPROOF project for UK manufacturing
inspections of safety critical components.
2.
TOFDPROOF PROJECT
2.1
TOFDPROOF CONSORTIUM
Eight European countries were represented in the TOFDPROOF consortium .
The consortium was a mix of industrial companies and research organisations. Major
Non Destructive Testing (NDT) service providers and two inspection and certification
notified bodies were represented to enable rapid acceptance and adoption of the
TOFD method throughout industry.
An overview of the consortium is given in the Table 1.
2.2
TOFDPROOF WORK PACKAGES

The TOFDPROOF project was divided into the following seven work packages:
WP1: Trials organisation and justification of representative specimens
WP2: Blind trials and performance assessing
WP3: Guidelines for training, qualification and certification
WP4: Acceptance criteria definition
WP5: Economic analysis
WP6: Exploitation and dissemination of results, link with CEN
WP7: Data storage, analysis and exchange
The objectives and the work performed for each of these work packages are
described in the next section.
Table 1 Overview of the consortium (extract from TOFDPROOF proposal)

Name
Country
Main mission / Business activity /Area of activity
Non profit company.

Institut de Soudure
Research works and expertise in applications of welding and allied
(IS)
techniques and provider of services and training.
France
(Project
Member of an accredited notified body for pressure vessels.
Co-ordinator)
Certification of NDT operators.
Sonovation are a high technology SME company offering
sophisticated structural integrity inspection.
Provider of NDT services, specifically (mechanised) UT and TOFD.
Sonovation
Netherlands Involved in TOFD inspection since 1988, instrumental in introduction
and development of TOFD, involved in trials and validations of TOFD
since 1990, initiated and participated in many projects on TOFD,
involved in national and international codes and standards
development.
A non profit company.
Contract research and consultancy in all methods of joining
technology and associated technologies.
Major worldwide provider of both NDT and welder training and
TWI Limited
UK
certification.
Provider of NDT field services including automated ultrasonics (P
scan, TOFD, phased array), ACFM, Lizard and conventional methods.
Many research projects and on-site inspections undertaken with
TOFD since 1979.
Major provider of power generation equipment.
Design, supply and construction of power plant world wide.
Mitsui Babcock
Major provider of NDT services.
UK
(MB)
On-site NDT inspection, expertise.
Expertise in development and qualification of automated ultrasonic
inspection techniques.
Staatliche
Major organisation relevant to risk-analysis, reliability and safety of
Materialprefungsa
materials, components, processes, structures and systems also
Germany
nstalt
component testing and validation of advanced engineering methods
as well as intelligent software systems.
(MPA Stuttgart)
NDE services, NDE in-service inspections, Engineering services,
Tecnatom S.A.
Spain
manufacturer of inspection systems
Research and development of NDE, NDE services, materials
VTT
Finland
research, assessment of structural integrity
Non profit company.
Instituto de
Portugal
Institute for welding, materials testing and allied techniques.
Soldadura e
Qualidade (ISQ)
Major supplier of NDE services.
Major technical control organisation, including power plants, nuclear
TUV
power plants and plants for process-technology.
Suddeutschland
Bau und Betrieb
Member of an accredited notified body for pressure vessels.
Germany
GmbH
Focus is put on inspection, testing and assessment of construction of
(TUV)
tanks and vessels, pipes, heat exchangers and structural steelwork.
2.3
W
ORK RESULTS
Each work package is discussed in turn in the following paragraphs.
2.3.1 Work Package WP1: Trials organisation and justification of representative

specimens
2.3.1.1 Objectives
This task was divided into three sub-tasks, that is, the identification of an appropriate
collection of welded specimens for the project, the preparation and agreement on
conventional NDT procedures and the organisation of a matrix of tests using various
NDT inspections and the manufacture of additional specimens if needed. Each of
these sub-tasks is considered in more detail below.
2.3.1.2 Work Performed
Review of Existing Specimens (Task 1.1)
The aim of this task was to establish a population of defects for practical trials. The
set of samples to be used in the project was selected from a database compiled by
the partners, supplemented by a series of manufactured samples. A review of the
original database and the proposals for additional samples was provided. The main
outputs are given below.
Ideally the specimens had to offer a wide variety of parameters against which to
compare time of flight diffraction and conventional NDT methods. Only simple butt
welded geometries were considered, such as flat plat and piping. The material was
limited to ferritic steel. The sample thicknesses had to be in the range between 6mm
and 100mm and both longitudinal and transverse defects were to be included.
The final selection aimed to provide a diverse group of defect types, positions and
dimensions spread over the range of material thickness. The group had also to offer
a fair balance between plate and pipe specimens.
The database contained more than 190 defects. Weaknesses in the database were
identified and additional defect samples were suggested.
Manufacturing of Specimens, Justification of the Collection and Design of the
Matrix of Tests (Task 1.2)
The set of samples used in the project were selected from the database compiled by
the partners and supplemented by a series of manufactured samples as discussed
above. The finalised database is made of 150 samples including 118 samples
selected from the original database and 32 additional manufactured samples. The
finalised database is given Table 2.
Having established the group of test pieces it was necessary to define the test matrix
i.e. the set of tests to be applied to each sample.
14
Crack
Porosity
Slag
Transverse
crack
Lack of
penetration
Transverse
LoF
Lack of IR
fusion
Volumetric
Over
penetration
TOTAL
10
Lack of
fusion
~10mm
~6mm
Defect
Type
19
11
~15mm
10
~20mm
28
~25mm
17
~30mm
~40mm
Sample thickness
~50mm
Table 2 Summary of the sample population
21
~60mm
~75mm
~100mm
150
12
12
17
15
36
44
TOTAL
Test Matrix
The test philosophy was to apply conventional pulse echo Ultrasonic Testing (UT),
conventional Radiographic Testing (RT) and TOFD to each specimen in accordance
with current European standards. To allow comparison, TOFD was applied by at
least two different teams and for the common thickness range 15 to 75mm three
teams applied TOFD. At least two different teams applied conventional tests to each
specimen. For the common thickness range 15 to 50 mm, two teams applied manual
UT and at least one of the radiographic methods allowed by European standards was
used. Samples less than 10mm thickness were not examined by manual ultrasonic
inspection but were examined by two independent X-Ray teams.
The following trials testing organisation was defined:
5 teams to perform TOFD testing
4 teams to perform UT testing
3 teams to perform X-ray testing
2 teams to perform ray testing
1 team to perform accelerator testing
Test Groups and Tests per Sample

The test method applied to a given test-sample depended on the thickness of the
test-sample. The number of times a given method was applied was also dependant
on the sample thickness. The number of times a given sample was inspected by a
given method is illustrated Table 3.
Table 3 Number of methods/applications applied against sample thickness
Methods / applications
Sample
Thickness
TOFD
Manual UT
X Ray
Gamma
Accelerator
<10
10 to <15
15 to <40
40 to <50
50 to <75
75 to 100
Matrix of Tests
The matrix of tests defined which team was to examine which sample. The number of
tests to be applied by a given team was established by determining the total number
of tests for a given method and assuring the following:
Conventional methods (including UT and RT) and TOFD were to be applied

to each specimen in accordance with current European standards.
To allow comparisons, TOFD was applied by at least two different teams and
for the common thickness (range 15 to 75mm) three teams applied TOFD.
At least two different teams applied conventional tests to each specimen
(three were applied to specimens above 15mm thick).
The test method applied and the number of times a given method was applied
to a given test piece depended on its thickness.
Additionally, it was agreed that the following points had to be addressed in order to
preserve the integrity of the work:
Previous inspection results of the specimens were deliberately withheld.
A given partner was not allowed to carry out both TOFD and conventional
NDT.
A given partner was only allowed to apply one type of conventional testing
method (as far as possible). For practical reasons, the same team would
perform gamma-ray inspection and accelerator inspection on the specimens
60mm thick and over.
The results from the trials were only to be circulated to all the partners after
completion of all the trials.
An acceptable balance of workload between partners had to be achieved.
Design and agreement on ionising radiation and UT procedures (Task 1.3)
A UT manual procedure was written in accordance with EN 1714 [1] and EN 1713
[2]; as well as a series of ionising radiation procedures written in accordance with EN
1435 class B [3]. These procedures were reviewed by a working group including
most of the partners and were all approved.
2.3.1.3 Comments
It has to be stressed that the samples used for the study are only of simple geometry
(plate and pipe) and that the defects they contained are essentially manufactured
defects.
Some of the samples contained additional defects which were not included in the
study by choice or because they were not intended defects. In some instances, these
extra defects made the study more difficult when these defects interfered with the
selected defects. Moreover, one of the main difficulties encountered later on, related
to the matrix of the test samples, was the quality of information provided on the
defects including defects type, position and size. Wrong information but mainly
incomplete information made the data study difficult.
2.3.2
Work Package WP2: Blind Trials and Performance Assessment
2.3.2.1 Objectives
The Objectives were to carry out blind trials in order to assess the performance, the
influence of scan set-up, the reliability and reproducibility of the TOFD technique:
To compare TOFD performance (both in term of acceptance criteria and best
practice) with UT and RT, as applied according to the European standards
defined by CEN/TC 121 "Welding";
To define the field of application of TOFD, highlighting weaknesses and
strengths and the possible need to use TOFD in combination with other NDT
techniques;
To optimise the methodology of application in order to ensure reproducible
inspections with different pieces of equipment and inspectors;
To verify how TOFD allows for detection of transverse defects;
To perform sectioning or advanced NDT techniques when discrepancies were
observed between TOFD and conventional inspection.
The work package was divided into sub-packages, that is:
a) Design of TOFD procedure and reference blocks;
b) Experimental work: TOFD and conventional NDT procedures application;
c) Technical analysis;
d) Optimisation of TOFD procedure;
e) Recommendations.
The work carried out under each sub-work packages is discuss below.

a) Design and agreement on set of TOFD procedures and manufacturing of
test-blocks
TOFD procedure
A TOFD procedure was produced by IS and reviewed by the partners.
The choice of set-ups to be applied (including: probe centre separation, probe
frequency, etc,) in relation to the sample thickness was determined using the
results of beam coverage modelling.
200 cases in terms of thickness, probe diameter, probe frequency, refraction angle
were studied. A table of recommended values was produced, see Table 4.
Table 4 Recommended TOFD Set-up

Element
Beamdiameter
angle ()
(mm)
Beam
intersection
depth
Thickness
t (mm)
Number
of TOFD
set-ups
Depthrange
(mm)
Centre
frequency
(MHz)
6-10
0-t
15
70
2-3
2/3 of t
>10-15
0-t
15-10
70
2-3
2/3 of t
>15-35
0-t
10-5
70-60
2-6
2/3 of t
>35-50
0-t
5-3.5
70-60
3-6
2/3 of t
0-t/2
5-3.5
70-60
3-6
1/3 of t
t/2-t
5-3.5
60-45
6-12
5/6 of t for 60
or t for 45
>50-100
The classification of relevant indications was specified by analysing the following

features:
Disturbance of the lateral wave
Disturbance of the back wall echo
Pattern between lateral wave and back wall echo
Signal phase pattern between lateral wave and back wall echo
Mode converted signal after the first back wall reflection.
The indications were classified as follows:

Surface breaking discontinuities
Embedded discontinuities (point like, elongated without measurable height,
elongated with measurable height)
Unclassified.
Reference block
For the application of this procedure, a reference block had to be used. A report
describing the reference blocks was written by IS and reviewed by the other partners
involved in the task. Two new designs were proposed for the reference blocks, one
with side drilled holes and one with notches. Note that the reflectors machined in the
reference blocks were side drilled holes which were not connected to the scanning
surface (as suggested by ENV 583-6:2000 [4]).
b) Experimental work: Round Robin Trials (RRT)

The RRT were carried out using the NDT procedures developed during the project.
Conventional inspections
The data were compiled by the partners and stored into a data base. The data base
was available through the TOFDPROOF web site. The data was only available to the
TOFDPROOF partners.
TOFD inspections
The TOFD inspections were carried out by five TOFD inspection teams (IS,
SONOVATION, TWI, VTT, ISQ). The team who supplied the specimens were not
involved in the examination of their own specimens.
Following the round robin trials and the reporting of the results, the data were
collated and a review focusing on the cause of discrepancies of the results was
carried out by MBEL and TWI. TOFD images and the inspection settings used for the
round robin were requested to enable the identification of discrepancy of the results.
Selected results from the round robin trial were examined and the source of the
discrepancies between the intended defects and the inspection results was identified.
This was carried out in order to highlight the strengths and weaknesses of the TOFD
technique. The lack of information on some of the defects limited the analysis.
The results were provided in a discrepancies analysis report. The report includes a
table summarising the discrepancies between the TOFD results for the defects for
which sufficient information was provided.
The report discussion is divided into six groups including: general comments, effect
of the set-up, surface defects, transverse defects, sizing errors and acceptance
criteria. For each group, comments and recommendations are provided.
Recommendations on when complementary NDT techniques should be used are
also provided.
The results confirm some of the TOFD limitations and highlight the need for an
appropriate procedure, a skilled data analyst and for realistic acceptance criteria.
c) Technical analysis of the results
A technical analysis of results (performance, reproducibility) was carried out.
A template was defined for reporting the results and each partners completed their
template either offline (e.g. Excel) or on-line on the web site. In the course of data
gathering, more than 600 documents, including single test reports have been
gathered and processed.
Definition of the parameters to store for each NDT Technique was done for each
specific procedure. MPAs ALIAS software or optionally Access or Excel software
using links to images and files were used for data storage, processing and analysis.
The existing (MPA ALIAS system) tool was used as an initial platform for results
analysis. The database contains a list of specimens with basic information as well as
the information about intended and detected defects.
10
Based on the literature overview carried out and on the results of similar projects, it
was decided to perform the hit-miss data analysis using either a log-logistic function
or using the Weibull three parameter cumulative distribution.
Due to the fact that the Weibull distribution has already been used for structure
assessment and development of the acceptance criteria in other similar projects, it
was favoured by the project partners against the log-logistic or log-normal
approaches.
The assumed probability of detection (POD) curves for TOFD are based on a 3
parameter Weibull equation to describe the mean POD as a function of defect height,
POD (a) = 1 exp { - [( a - ) / ] }
Where:
: Scale parameter, calculated using selected mean POD values

(for given shape and threshold parameters)
: Shape parameter
: Threshold parameter, below which size no defects are detected
(POD=0)
It should be noted that a further comparison with accepted approaches in other NDT
areas (RT with lognormal cumulative distribution approach, for example) will be
needed in order to obtain better benchmarking results. Furthermore, one of the
results of the interaction between the project partners and other interested groups in
this area showed that they might be objections to the shape of the curves obtained in
the project. These objections were based on the comparison of the curves proposed
with those obtained in other projects using similar approaches for other NDT
techniques. However, for the purpose of data analysis and development of
acceptance criteria, the applied data analysis methods was considered to give more
than reasonable and feasible results, comparable with similar results obtained in
other related projects.
For the hit-miss data analysis (approach by which the inspection results can be
recorded only in terms of whether or not a flaw was found), a comparison of expected
results against obtained results has been performed.
The data fitting process has been performed using, in addition to pure
analytical/statistical data analysis, some expert judgement and taking into account
the results and experience of other projects, namely:
1. It has been assumed that the gamma parameter has the physical meaning of
the non-detection limit, that is, under the size of gamma (in millimetres) the
system is either not capable or it is not possible to detect the defect.
2. It is assumed that the curves have similar shapes (similarity principle),
therefore, an average value of beta parameter has been fixed as 0.65.
3. A lower bound curve can be constructed based on confidence level of 95%
and shape similarity.
The achieved mean POD curves are shown in Figure 1 to Figure 6, whereas Figure 7
shows the whole family of curves.
11
Figure 1 POD curve for plate thickness up to 6 mm
Figure 2 POD curve for plate thickness 6 to 15 mm
Figure 3 POD curve for plate thickness 15 to 25.4 mm
12
Figure 4 POD curve for plate thickness 25.4 to 40 mm
Figure 5 POD curve for plate thickness 40 to 60 mm
Figure 6 POD curve for plate thickness over 60 mm

13
Figure 7 Family of POD curves for defect height detection
Figure 8 Comparison of TOFD, RT and MUT, defect height, for a 25.4mm thick plate
The Figure 8 shows an example of overall NDT methods comparison for the methods
considered in the project (TOFD, RT, MUT) for a 25.4mm thick plate.
For sizing performance, one of the greatest drawbacks in the analysis process was
the lack of information on the true size of the defects which could have been provided
through sectioning (destructive confirmation of defects). The information may have
explained some gaps and discrepancies in the results obtained.
However, using the available data, and combining it with other available data
provided by Sonovation, the data analysis for sizing was performed; the summary of
the results is given in Table 5.
14
Table 5 Performances comparison

Height Mean
Error
h (mm)
Height Standard
Deviation
Length Mean
Error
l (mm)
Length
Standard
Deviation
TOFD
1.0*
1.9
4.5
6.5
RT
NA
NA
1.5
18
MUT
0.4
4
1.5
25
* Mean value on the full range of samples 6mm < t <100mm.
Note that the mean value obtained on another population of defects provided
a h of 0.3mm for TOFD after sectioning.
Other comparison performances for POD and false call rate (FCR) were carried out
and are summarised in Table 6. Table 6 also includes performances obtained during
the Dutch acceptance criteria project by the Dutch Quality Surveillance and NonDestructive Testing Society (KINT).
Table 6 Performances comparison POD & FCR
TOFDPROOF project
KINT project
POD
FCR
POD
FCR
TOFD
70-90%
<10%
82.4%
11.1%
RT
60-70%
NA
60.1%*
10.8%*
MUT
55-65%
NA
52.3%
22.7%
* Using -ray only.

False Call Probability analysis was not performed for RT and MUT due to the lack of
relevant data and no possibility to objectively verify the results by sectioning the
samples.
d) Optimisation of TOFD procedures and application
The TOFD procedure developed for the trials was updated to reflect the remarks and
conclusions from the discrepancy analysis report and the comments from the
partners involved in the TOFD examinations. A final report of the procedure was
produced. The main items modified are listed below:
When using the recommended TOFD set-ups presented in Table 4, the
maximum recommended probe frequency should be considered first. Lower
frequencies may be used if the required sensitivity setting cannot be achieved
with higher frequencies.
The techniques to measure length and height were more clearly defined. This
part of the procedure has also been added in the proposed acceptance
criteria and in the recommendation for applying TOFD (see Appendix 1).
When the specific pattern (replication of mode converted signals) shown in
Figure 10 is observed one can conclude that there is evidence of transverse
defects.
15
The end of time window shall be at least 1 s after the 1st mode converted
signal.
Transverse defects
Specific pattern
Figure 10 TOFD image of a non-parallel scan of a sample containing 3 transverse defects.

e) Recommendations for applying TOFD
Taking account of the discrepancies identified during the analysis of the round robin
trials TOFD data, recommendations for applying TOFD and on when a
complementary NDT technique could be required, were provided.
A report on recommendations, report No 2-31-D-2004-02-1 [5] given in Appendix 1
(the report can also be downloaded from the TOFDPROOF web site:
http://www.mpa-lifetech.de/TOFD) was written by MB and TWI and includes
comments from the other TOFDPROOF partners.
The recommendations were transmitted to CEN/TC 121, CEN/TC 54, CEN/TC 138
and EPERC.
The recommendations (complete with observations) were provided for seven
categories, including: procedure, identification of reference marks, set-up,
classification of indications, evaluation of indications, personnel qualification and
acceptance criteria. For each category, comments and recommendations were
provided. A summary is provided hereafter.
16
A specific procedure shall be written in accordance with the guidelines given

in ENV 583-6 [4] and DD CEN/TS 14751 [6] for each individual type of
inspection.
The identification and inscription of reference marks (including datum on the
component and reference point on the inspection probes array) is critical to
allow repeatability of the inspection and results comparison.
Where a defect extends before and after a datum, an extra scan covering the
entire defect area (without a break) should be carried out in order to provide a
more accurate value of the defect length.
Where a large number of point-like indications have been detected which
creates a cluster of indications, the inspection should be supported by
another NDT technique.
It became clearer after the review of the analysis that high frequency small
crystal diameter probes (15MHz, 3mm) should be recommended for the
inspection of thin samples (up to 15mm) especially when the weld surfaces
are as-welded, an alternative choice (e.g., 10MHz) may not be appropriate.
The presence of the lateral wave and the back wall echo restrict the
inspection zone and small defects in these zones can be missed. High
frequency (15MHz) probes provided a better means of detection and
evaluation of defects which are surface breaking or just below the surface.
A separate root scan should be considered.
Inspection from both surfaces is recommended when access permits.
Inspection for transverse indications can be limited especially when the weld
cap is still present. The normal TOFD configuration is not optimised for
transverse defect inspection. However, the presence of some mode
converted echoes may suggest that transverse defects could be present.
Transverse indications were in general not reported or wrongly reported as
point- like or as longitudinal defects. When transverse defects are expected
or/and when indications on the TOFD image suggest the presence of such
defects (especially if the weld is as-welded), additional technique(s) should be
used.
The indications shall be classified into categories clearly defined in the
inspection procedure.
For interpretation of the images, initial analysis has to be carried out on
unprocessed data. Straightening and removal (for lateral wave and back wall
echo) tools can be use for subsequent analysis e.g. confirmation of
presence/absence of surface defects.
The technique used to determine the indications dimensions should be clearly
defined.
In addition, guidelines on how to determine defect length, depth and height were
provided.
17
The results confirm some of the TOFD limitations and highlight the need for an
appropriate procedure, a skilled data analyst and for realistic acceptance criteria.
2.3.2.3 Comments
General
Most of the specimens used during the round robin trials contained additional defects
which were not included in the matrix of defects. The lack of information / wrong
information provided by the owner of the sample used for the round robin trials as
well as the presence of additional unwanted defects restricted the study.
Technical analysis of the results
The technical analysis of the results provided was not as detailed as expected. This
is probably partly due to the comments above; however, the results from the
conventional inspections were not presented in any detail. In order to analyse the
large quantity of data, MPA used their ALIAS software. The results provided are
limited to Table 5 and Table 6 and the example presented Figure 8.
It is not clear how the information from the round robin was treated. Having carried
out a preliminary analysis of the TOFD data it is not clear how much of the results
from the round robin were used for the construction of the POD curves. The
preliminary analysis carried out highlights the difficulty of identifying the defects
selected for the study, mainly due to: the inconsistency of the referencing of the
TOFD scans, the presence of unwanted defects and the lack of information on the
defects. The conclusions are therefore unlikely to be based on the full population of
defects. Moreover, the PODs (Figures 1 to 8) relate to defects as small as 0.5mm, it
is not clear how those dimensions are known.
Note that the dimensions used for the analysis are based on manufacturing values.
No sectioning was performed on any of the specimens even when discrepancies
between the TOFD results (and/or NDT techniques) were observed.
Recommendations
The recommendations provided in recommendations for applying TOFD report [5]
were based on the preliminary analysis of the TOFD data. The report provides
important information. However, an important recommendation was omitted in the
final TOFDPROOF report, that is, a separate root scan should be considered. It is
well recognised that one of the main limitations of TOFD is the capability of detecting
small defects at the root area especially when the root bead is present. It is therefore
important to concentrate on this area, by using specific TOFD scans but more
appropriately to use additional NDT techniques.
TOFD procedure
The Optimisation of the TOFD procedures was carried out using the comments from
the TOFDPROOF partners. Although the procedure is specific to the TOFDPROOF
round robin, it can be used as an example for other inspections.
The procedure was supplied to the Technical Committee CEN/TC 121 Welding and
the information was used to update the technique specification CEN/TC 121/SC
5/WG 2 N 146 document Welding Use of time-of-flight diffraction technique
(TOFD) for testing of welds and the draft European standard DD CEN/TS
18
14751:2004 document Welding Use of time-of-flight diffraction technique (TOFD)

for examination of welds [6]. These documents were approved after the completion
of the TOFDPROOF project.
The TOFD procedure used for the TOFDPROOF round robin trial has a lot of
similarity with the DD CEN/TS 14751:2004 document (being developed at similar
time). The scope of this project did not include the review of the CEN standard
however, comparisons and comments on the preparation for examination section(s)
of the documents are presented hereafter.
The main differences are that the CEN standard extends its scope to in-service
inspections and to thicker specimens (note that the standard does not provide an
upper thickness limit but provides recommended TOFD set-ups for thicknesses up to
300mm).
The type of calibration blocks recommended under both the TOFDPROOF procedure
and the DD CEN/TS 14751:2004, does not include one of the two type of blocks
proposed under ENV 583-6:2000. The European pre-standard DD ENV 583-6
proposed a choice of blocks containing two types of diffractors:
Machined notches, open to the scanning surface of the reference block; or
Side drilled holes with a diameter of at least twice the wavelength of the
nominal frequency of the probes utilised in the inspection. The holes were to
be cut to the scanning surface in order to block the direct reflection from the
top of the hole.
It is not clear why the block containing the second type of diffractor is not included
into the recommendations. Especially because the manufacture of this type of
diffractor is easier than the 60 notches and allows greater reliability in calibration
than the reference block containing side-drilled holes without being cut to the surface.
Both documents state that the sensitivity setting shall be performed in accordance
with ENV 583-6:2000 which specifies that the settings in the region of the timebase
after the arrival of the lateral wave should be set to approximately 5% of the
amplitude scale. However, both documents favour the setting with reference to the
lateral wave. The sensitivity should be set such that the amplitude of the lateral wave
is between 40% and 80% full screen height (FSH). Then, in the case where the use
of the lateral wave is not appropriate the sensitivity shall be set such that the
amplitude of the backwall signal is between 18dB and 30dB above FSH. If neither
lateral wave nor backwall signal is appropriate, the sensitivity should be set such that
the material grain noise is between 5% and 10% FSH.
The variation between the two documents are:
Under the TOFDPROOF procedure the sensitivity check shall be performed
on a reference block, under the CEN standard the sensitivity shall be set on
the test object.
Under the TOFDPROOF procedure an additional note is provided when the
sensitivity is set with reference to the lateral wave, that is, if the noise level
exceeds 10% FSH the lateral wave amplitude shall be decreased accordingly.
Both documents cover simple geometry in plates, pipes and vessels where both the
weld and the parent material are low alloy carbon steel. However, it is likely that the
19
standard will also be used as a reference for other material. Moreover, it seems more
appropriate to use the noise level as a preferred setting and if the material grain
noise is not appropriate (lateral wave or backwall signal not appropriate e.g.,
saturation of the lateral wave) use the lateral wave or the backwall signal (as per
ENV 583-6:2000). Too high amplitude of the lateral wave or the backwall echo may
increase difficulties in identifying small defects at these areas.
As the statement a sensitivity check shall be performed at least every four hours and
after completion of the examination covers both pre-service and in-service
inspections for the CEN standard, It should be noted that if the inspection is applied
to radiation environments the four hours may not be always practical from an ALARP
point of view (e.g. during the inspection of a large vessel).
Furthermore, the sensitivity and range corrections checking proposed by both
documents, suggest that all examinations shall be repeated since the last valid check
if any deviation on the sensitivity greater than 6dB is observed. It has to be noted that
local deviations are frequently observed and local deviations of more than 6dB
should not be considered as a source of rejection of acquired data. The sensitivity
variation should mainly be used to verify the proper working condition of the
equipment.
Due to the similarity between the TOFD procedure used under the TOFDPROOF
project and the DD CEN/TS 14751:2004 document, The TOFDPROOF round robin
trials could have been a good opportunity to test the capability of the new EN
technical specification. Unfortunately, there were evidences that the partners
involved in the round robin trials did not always follow the TOFD procedure.
2.3.3
Work Package WP3: Guidelines for training, qualification and certification
2.3.3.1 Objectives
This work package was to define a framework for operators qualification and
certification and assessing the influence of the objectiveness of the inspector
interpreting the results.
A.
Interactive guideline
Interactive training guidelines for interpreting TOFD images were produced. They
were designed as a computer software programme. It contains a database of more
than 25 typical TOFD images. TOFD images used for the software were provided by
the partners.
The guidelines are divided into three parts, providing:
Basic information on TOFD
TOFD image assessment methodology
TOFD indication classification and sizing methodology
20
The last part doesnt take into account the optimising sizing techniques
recommended by the optimised TOFD procedure presented in Work package 2;
and the individual A-scans of a given TOFD image were not made available.
As a part of the training, the inspector learns to assess the quality of the TOFD
image. Unsatisfactory TOFD images are provided with explanation on the cause of
the problems. This section also provides a way of identifying defects on a TOFD
image and to determine their type and dimensions (length and height).
The interactive training software can be used on line free of charge on the web site:
http://www.mpa-lifetech.de/TOFD.
B.
Objectivity assessment
As well as the training part, the software also contains a testing part that has been
used to assess the objectivity of the inspectors interpretation of TOFD images.
Inspectors from the TOFD partners have taken part in the objectivity assessment,
and the results of this exercise have been analysed and reported.
The discrepancies seen in the results, during the objectivity assessment, have been
fed back to the guidelines designer to enable the interactive training guidelines to be
improved and updated.
The conclusions from the survey were:
The greater degree of confusion was observed in the defects type E
(elongated without measurable height) and D (point like).
For most of the defects it was observed that there was a tendency to
undersize defect length and over size defect height.
In order to improve the operator capability for defect characterisation and
sizing, it is recommended to increase the number of defects available for the
training.
C.
Recommendations for TOFD training and certification
Two documents containing recommendations for TOFD training and certification

have been produced (reports [7] & [8] respectively). These have taken account of
existing certification schemes in the UK (administered by PCN) and The Netherlands
(administered by SKO).
The documents contain:
details of the recommended training syllabus for TOFD levels 1, 2 and 3,
the necessary experience and pre-qualifications required, and
a recommended format and marking scheme for the theoretical and practical
parts of the examinations.
The recommendations have been transmitted to one of the technical committee for
standardisation, CEN/TC 138 (NDT). They can be downloaded from the
TOFDPROOF website (http://www.mpa-lifetech.de/TOFD).
21
One of the pre-requirement for the certification described in the document is that it
can only be available to holders of current, valid ultrasonic weld testing certification.
The certification is initially valid for a period of 5 years. The candidates are required
to demonstrate that they meet the minimum supplementary training and certification
requirements presented in Table 7 before they will be allowed to take TOFD
examinations.
Table 7 Minimum training and experience requirements
Minimum Training Requirements
Level 1
3 Days (24 Hours)
Level 2
5 Days (40 Hours)*
Level 2 (direct)
8 Days (64 Hours)
Minimum Experience Requirements

Level 1
3 Month
Level 2
6 Months
* Contains additional modules covering interpretation of

TOFD images and instruction (procedure) writing.
2.3.3.3 Comments
The number of defects provided in the testing study is limited. They were no
opposite surface discontinuities in the catalogue of defects included. From
experience these defects may be difficult to identify/characterised and are often
wrongly sized in both depth and length.
The tool is very limited as the A-scans are not available and the curser available may
not be adequate for an accurate sizing performance. Moreover, information on the
defects used for the survey is based on intended defect type and dimensions, with no
confirmation from sectioning.
Although the training tool in the TOFDPROOF website is useful, it is very basic and
only offers a restrictive number of defects which are relatively easy to interpret.
However, it still provides a good introduction to TOFD images and data analysis;
besides, the web training tool was never designed to replace more sophisticated
software available with the TOFD equipment.
Training in TOFD is critical as training and experience in TOFD are essential for an
appropriate application of the TOFD technique.
The recommendations for TOFD training and certification proposed under the
TOFDPROOF project are similar to the BINDT specific requirements for the
certification of personnel engaged in ultrasonic time of flight diffraction testing of
linear butt welds in ferritic steel which was issued in January 2002 [9]. The variations
between the PCN/GEN document and the TOFDPROOF document could be due to
the later issue of the TOFDPROOF document.
22
The main differences are:

Durations for training and experience as summarised in Table 8 below. The
TOFDPROOF figures presented were modified numerous times throughout
the project. The new requirement figures seem more practical.
Table 8 Comparison of durations for training and experience
TOFDPROOF requirements
PCN/GEN requirements
Minimum Training Requirements

Level 1
24 Hours
40 Hours
Level 2
40 Hours
40 Hours
Level 2 (direct)
64 Hours
80 Hours
Minimum Experience Requirements

Level 1
3 Months
3 Months
Level 2
6 Months
9 Months
Level 2 (direct)
9 Months
12 Months
The syllabus presented by the TOFDPROOF project is generally less

demanding for a level 1.
Two items are allocated to the level 3 syllabus with TOFDPROOF and to the
level 1 syllabus for the PCN/GEN, that is:
o TOFD on complex geometry
o The use of Synthetic Aperture Focussing Technique (SAFT)
These two items should certainly not be included into the level 1 syllabus. It is
believed that for a level 1 only simple analysis and geometry should be
covered. It can however be argued that these two items could have been
included into the level 2 syllabus.
Additional requirements to those presented in the TOFDPROOF recommendations
are being requested by training centre. Some examples are given below:
An eye test must be passed (near vision and color) to be eligible for the
PCN/SNT Levels 1 to 3 examination.
The supervised work experience is to be obtained either before the
examination or within 12 months after passing the exam.
The TOFD Level 1 & 2 examination includes a theory paper as well as
specific paper and a practical exam.
Overall, the recommendations for TOFD training and certification presented by the
TOFDPROOF project provide an acceptable compromise and are considered more
practical.
23
2.3.4
Work Package WP4: Acceptance criteria definition
2.3.4.1 Objectives
The TOFDPROOF objectives of this package were: to develop justified acceptance
criteria; design severity levels to achieve an acceptable repair rate and integrity level
and to validate the acceptance criteria on real structures.
The integrity level of a structure when using the developed TOFD acceptance criteria
were planned to be equal to or better than the achieved integrity level when utilising
the currently applied NDT methods.
The work package was divided into three tasks: a literature survey, the design of
acceptance criteria and the validation of the acceptance criteria.
This package has the most safety issues.
a
Literature survey
The first task was to carry out a literature survey to review all existing projects and
studies relevant to the development of acceptance criteria for TOFD. The literature
survey was carried out by TWI.
The main reviewed projects were:
The development of acceptance criteria for time-of-flight diffraction
examination method KINT project in Netherlands.
The concluded limits of detection provided in the TOFDPROOF final report
(although it is not clear where in the source literature these figures come
from) were the following:
o h=0.5 mm for thickness t=7mm,
o h=1mm for a thickness t=100mm,
o h=1.5mm for a thickness t >200mm.
Evaluation of acceptance criteria for the ultrasonic time-of-flight diffraction
(TOFD) technique HSE project in UK; this project included an independent
assessment of KINT proposal based on TOFD responses from existing TWI
data base thickness range 10mm to 93mm. TOFD rejects a similar number
of flaws to radiography according to British standards and UT according to EN
standards.
ASME B & PV Code case 2235 [10] in USA, this code case allows the
replacement of radiography by ultrasonics for thickness above 12.7mm. The
code case requires manufacture of a calibration block containing at least two
planar defects. Acceptance criteria are tabulated and specify maximum
acceptable height as a fraction of the thickness for a given range of flaw
aspect (height/length ratio),
The acceptance criteria developed under TOFDPROOF are intended to be
integrated into the existing European standardisation scheme (see Table 9).
24
Table 9 European standard

NDT Method
NDT procedure
for welds
Acceptance levels
for welds
VT
EN 970
EN 25817
RT
EN 1435
EN 12517
UT
EN 1714
EN 1712
PT
EN 571-1
EN 1289
MT
EN 1290
EN 1291
TOFD
XPCEN/TS 14751
The European standardisation is organised in different Technical Committees (TC)

entrusted to write standards. To perform NDT of welds for pressure vessels, it is in
fact necessary to use the standards issued by CEN/TC 138 (NDT), CEN/TC 121
(Welds), CEN/TC 54 (Pressure vessels). CEN/TC 138 provides general
requirements for the application of a given NDT method. CEN/TC 121 adds
complementary instructions for testing welds giving the detailed methodology to
follow for detecting the flaws and if required to perform their characterization.
CEN/TC 121 proposes various testing and acceptance levels. CEN/TC 54 adds
specific requirements for testing welds on pressure vessels and specifies the
acceptance level to apply for pressure vessels.
The general philosophy of EN ISO 5817 [11] has been adopted.
EN 25817 [12] contains three quality levels B, C and D where B is the most stringent.
Quality level C can be used for pressure vessel not submitted to fatigue, creep, and
design with no reduced safety factor, otherwise quality level B shall be used.
However, the physical flaw descriptions of EN ISO 5817 do not equate to the
capabilities of different NDT techniques to detect and discriminate certain types of
imperfections.
So CEN/TC 121 and CEN/TC 54 have agreed that there is a fundamental difference
between Quality level and Acceptance level. The former limits the size and the
number of imperfections in a given weld, including their physical size, while the latter
limits the size and/or number of indications in a given weld using a given NDT
technique. Thus acceptance levels are established in accordance with the capability
of individual techniques to detect and discriminate certain types of imperfections and
to determine their size in relation to the quality requirements.
EN 12062 [13] is used as a starting point or as a method transfer function. This
standard defines the interface between quality levels in EN ISO 5817 and
acceptance levels of indications as per the EN standards.
25
The main conclusions from the literature survey were:

The acceptance criteria developed in the KINT project Development of
Acceptance Criteria for TOFD examination method are a good starting point
to compare the reject rate between TOFD and conventional NDT method
The types of flaws rejected by applying TOFD acceptance criteria should be
compared to those rejected by applying conventional NDT techniques.
ASME Code Case 2235 is an important alternative and should be compared
to the KINT acceptance criteria proposal.
An acceptance criteria should be included in the European standardisation
scheme.
b
Design of acceptance criteria

Acceptance criteria in accordance with EN 25817 were designed and proposed
tables designed by the TOFD experts of the consortium, meeting the various levels
as defined in EN 25817 and EN 12062 have been produced. The proposal consists
of TOFD acceptance criteria for three acceptance levels related to the quality levels
according to EN 25817 (see Table 10).
Table 10 Acceptance levels
Quality level according
to EN 25817
Acceptance level
B (Stringent)
C (Intermediate)
D (Moderate)
The proposed acceptance criteria are linked to the capability of TOFD to classify
defects as specified in DD CEN/TS 14751 and illustrated by Figure 9.
l
p
h
t
Embedded discontinuity
Surface breaking discontinuity
Figure 9 TOFD capabilities concerning defect classification.
26
For single discontinuities, the proposed acceptance criteria approved by the

TOFDPROOF consortium are given in Table 11 to Table 13.
Where:
l: length of an discontinuity
h: height of an discontinuity
dd: nominal wall thickness
Table 11 Acceptance criteria for acceptance level 1

Maximum allowable length (Lmax)
if h < h2 or h3.
Surface breaking
Embedded
Maximum allowable
height (h1) when
L > Lmax
Thickness range
Lmax [mm]
h3 [mm]
h2 [mm]
h1 [mm]
6mm< dd 15mm
0.75 x dd
1.5
15mm< dd 50mm
0.75 x dd
50mm< dd 100mm
40
2.5
dd > 100mm
50

if h < h2 or h3.
Surface breaking
Embedded
Maximum allowable
height (h1) when
L > Lmax
Thickness range
Lmax [mm]
h3 [mm]
h2 [mm]
h1 [mm]
6mm< dd 15mm
dd
15mm< dd 50mm
dd
50mm< dd 100mm
50
dd > 100mm
60

if h < h2 or h3.
Surface breaking Embedded
Maximum allowable
height (h1) when
L > Lmax
Thickness range
Lmax [mm]
h3 [mm]
h2 [mm]
h1 [mm]
6mm< dd 15mm
ddx1.5 (max. 20)
2.5
4.5
15mm< dd 50mm ddx1.5 (max. 60)

50mm< dd 100mm
60
dd > 100mm
75
27
Rules for how to treat groups of discontinuities are also provided in the proposed
acceptance criteria. These rules depend on the acceptance level.
Additionnal information on groups of discontinuities was presented during the
TOFDPROOF seminar in Villepinte. The information is provided below:
Indications shall be considered as a group if:
o The distance between two indications along the weld is less than the
length of the longest indication.
o The distance between two indications in the thickness direction of the
weld is less than the height of the highest indication.
The sum of the lengths of the individual indications measured along the weld
over a length of 12 dd shall be less or equal to:
o Level 1: 3.5 dd with a maximum of 150mm
o Level 2: 4 dd with a maximum of 200mm
o Level 3: 4.5 dd with a maximum of 250mm
Indications that do not fulfill the requirements mentioned above should be
considered as single indications.
A maximum accumulated length is given for discontinuities: 10% of the total weld
length with a maximum of 500mm.
c
Rejection rate comparison

In addition, work has been carried out to compare the TOFD rejection rate when
applying the present acceptance criteria proposal to RT and UT rejection rates when
applied according to CEN standards . To this end the different NDT techniques were
simulated in a probabilistic model and fracture mechanics assessments were
performed for different sample geometries, loading conditions, material data and
defect configuration.
In these simulations a structure with a population of defects is inspected with TOFD
and other NDT techniques. Depending on the performance of the applied NDT
technique and on corresponding acceptance criteria, a certain percentage of the total
number of defects will be rejected. The number and types of rejectable defects
largely depend upon the applied NDT technique.
By comparing the calculated reject rates and calculated failure probabilities, the
acceptance criteria for TOFD were justified in such a manner that a structure
inspected with TOFD would have the same safety level (or probability of failure) as a
conventionally inspected structure.
The methodology used in the probabilistic comparison and the fracture mechanics
assessment was based upon the results from the KINT project
With use of the ALIAS software the probability of detection (POD) and the population
distribution of flaws (PDF) were computed from the available TOFDPROOF
database collected during the Round Robin Trials. The parameters determined from
the TOFDPROOF database were used to validate the proposed TOFD acceptance
criteria.
28
It was demonstrated that the final proposal for TOFD acceptance criteria meets the
stated conditions:
A better or equal probability of failure compared to conventional NDT.
An equal or lesser percentage of rejections for TOFD compared to
conventional NDT.
In order to validate the acceptance criteria, an on site validation was scheduled. To
complete this task, a TOFD inspection was carried out on pressure vessels at the
manufacturing stage and the results were compared with conventional NDT
methods. Unfortunately (for the TOFDPROOF project) during this exercise, the
quality of the welds was found to be high (by each of the NDT methods). Therefore,
due to the lack of a representative population it has been concluded that it was not
possible to perform a realistic on site acceptance criteria validation.
2.3.4.3 Comments
The acceptance levels tables (Table 11 to Table 13) for acceptance level 1 to 3,
have not been modified since the last TOFDPROOF acceptance criteria draft report.
The comments raised by MB on the draft, at the time, are therefore still applicable:
The criteria assume that surface breaking cracks height measurement can be
achieved with a resolution of at least 0.5mm which is doubtful.
It would be difficult in practice to resolve edges or measure defects accurately
enough to apply the criteria proposed.
It is doubtful whether the criteria make much sense from a structural integrity
viewpoint, e.g. if one can accept 2mm x 38mm in 50mm thick weld then why
not 3 x 50mm in weld thickness 100mm.
Discontinuities as a group are briefly mentioned in the TOFDPROOF acceptance
criteria: a maximum accumulated length is given for discontinuities: 10% of the total
weld length with a maximum of 500mm. However, it is believed by Mitsui Babcock
that the maximum accumulated length of 500mm is not justified. The maximum value
should only be related to the weld length as very long welds could be under
investigation.
2.3.5
Work Package WP5: Economic analysis
2.3.5.1 Objectives
The objective was to carry out an economic analysis of the TOFD technique
compared with conventional NDT methods. The economic analysis was carried out
using the TOFD acceptance criteria proposed for each level.
The comparison was to take into account:
The direct cost of the inspection (day or night working) based on an average
rate for the inspector, the equipment, consumable, etc;
The indirect cost of the inspection caused by for example: false calls and time
to analyse the results;
29
The cost of purchase and maintenance of the equipment and the cost of
training of the personnel;
The manufacturing disturbance caused by the NDT applied;
The productivity gains.
2.3.5.2 Cost Comparison
The economic analysis of TOFD was carried out by TV and reported to the
TOFDPROOF partners. The following paragraphs were extracted from the report.
Case studies
Six case studies were distributed to the partners. For each case study, direct cost
and indirect cost were required for manual UT, TOFD and -ray inspections.
Each partner involved in the consortium was asked to supply the estimated average
costs in their own country for inspection of several industrial components by TOFD,
RT and UT. For this purpose representative case studies were defined:
1. Vessel of 2m diameter; thickness 36mm, 50mm, 100mm, length 10m; number
of welds to be inspected: 3 circumferential welds and 2 longitudinal welds.
The cost per meter of weld inspected was to be supplied for each of
configuration.
2. Piping of 500mm diameter, thickness 6mm, 15mm and 36mm; number of
welds to be inspected: 10 welds. The cost per circumferential weld was to be
supplied for each of configuration.
The costs of a TOFD inspection was compared with the direct and indirect costs
generated when using:
Manual UT applied according EN 1714 and EN 1713 (acceptance criteria
according EN 1712 [14] level 2 + no planar imperfection accepted)
X-rays applied according EN 1435 [3] (acceptance criteria according EN
12517 level 1 and 2 [15] & table 6.5.3.2.1 from Pr EN 13445 [16]).
The effect of the POD and FCR on the resulting cost of an inspection was quantified
using decision trees, taking into account the rejection rate and the probability of
failure. An example of probability of failure using this method was studied. The
decision tree input are given in Table 14. The calculation was carried out using a real
model with cost for the inspection, repair and consequential loss of a failure. The
results are given in Table 15 and illustrated by Figure .
30
Table 14 Decision tree input

POD
82%
84%
60%
52%
0%
TOFD
Automated PE
-ray
Manual UT
No inspection
Probability of defect present
Price of repair
Consequential loss
Probability of failure if defect
detected
Remark
FCR
11%
14%
11%
23%
0%
Cost [C1]
1000
10,000
500
250
0
5%
1K
500K
not 50%
Maximum of two repairs allowed
* Price [C1] = Inspection price
Table 15 Comparison of decision tree results

TOTAL Cost
TOFD
Automated PE
-ray
Manual UT
No inspection
Actual
TOFD = 1
3,543
11,287
6,279.5
8,331
12500
1
3.2
1.8
2.4
3.5
Probability of Failure
Actual
TOFD = 1
(x10-3)
5.1
1
4.9
1
11
2.2
16
3
25
4.9
14000
3,0E-02
Price [C1]
12000
Total Costs
2,5E-02
Probability of failure
10000
2,0E-02
8000
1,5E-02
6000
1,0E-02
4000
5,0E-03
2000
0,0E+00
0
TOFD
PE Meander
G-Ray
Manual UT
No Inspection
Figure 11 Comparison of cost and probability of failure

As illustrated by Figure 11, TOFD inspection appeared to be the cheapest inspection
method and not the most expensive one as it is generally assumed.
31
Conclusions from the cost comparison report

TOFD is more competitive for high amount of inspections to be performed
(per meters of welds) especially for medium and high thicknesses.
The economic advantages of applying TOFD increase with the thickness of
the material inspected.
TOFD is even more competitive when taking into account the better
performance in PoD resulting in fewer problems during in-service inspection.
During recent years, due to the better availability of TOFD, the ever
increasing speed of computing power and experience being built up, the price
for inspection with TOFD has come down. Due to increased emphasis on
radiation safety, the price of radiography has still increasing direct and indirect
costs. These factors have pushed the break-even point in relation to the
material thickness even further down.
If all the factors are taken into account, it can be shown that TOFD can be
cheaper than conventional NDT methods.
2.3.5.3 Comments on the cost comparison report
It is not clear to what extend the report took into account the results of the case
studies.
Although most of the comments presented in the report (summarised in section
2.3.5.2 above) are to some extent true, the cost of TOFD is still calculated in respect
to the company facility and expertise. Most companies would not be specialised in
TOFD inspection and would therefore perform the inspection at higher cost to cover
for new equipment and specific transducers. Specialised TOFD companies, however,
would have an advantage as they would already own the TOFD equipment with a
variety of probes and scanners as well as have appropriate knowledge and
experience and would therefore be able to provide inspection at a lower cost. The
case studies certainly highlight this point. Most of the companies who participated in
the case studies provided higher total cost for TOFD than conventional UT, with Xray being the most costly inspection out of the three inspection techniques. The lower
cost for the TOFD inspection was provided by a specialised company.
However, even if the TOFD technique is flexible, it is more suitable for smooth
surfaces with straight forward geometry. Rough or uneven surfaces make it difficult to
maintain both probes on the scanning surface and can generate poor data quality.
Difficult geometry can be scanned but the interpretation of the data becomes more
complicated and may require complicated software.
The preparation time for the TOFD technique should not be much different to the
preparation time needed for the manual inspection.
32
2.3.6
Work Package WP6: Exploitation and dissemination
2.3.6.1 Objectives
The objectives were to:
Supply the results obtained, the proposed guidelines and the proposed
acceptance criteria to all parties involved in NDT inspection in the European
Union;
Get feedback through specific seminars, publication of papers and Web site,
in order to obtain the widest consensus on the guidelines and proposed
acceptance criteria;
Convey to the relevant CEN Technical Committees the revised guidelines and
proposal of acceptance criteria taking into account the observations received.
The main deliverables of the programme which required exploitation were: the
acceptance criteria, the analysis of TOFD performances, guidelines for TOFD
interpretation (CD-ROM), recommendations for training and certification and
recommendations for TOFD application.
The results were transmitted to CEN and related European networks. Some
members of the consortium were involved in National, European and International
standardisation activity. Three partners were member of the CEN/TC 121 ad hoc
group dealing with TOFD standardisation for welds.
Specific National (Lisbon, 03/2005) and International workshops (Paris,
04/2005) were organised both having attracted numerous participants from all
Europe (about 70 attended the Paris workshop).
The TOFDPROOF project specific Web site hosted by MPA (http://www.mpalifetech.de/tofd) was created.
Maintaining permanent link with CEN - TOFD ad hoc group CEN/TC 121
through several project partners and permanent observers.
o RIMAP and FITNET networks respective chairmen were kept informed
about TOFDPROOF progress (Annual progress reports).
o Dr Hecht, chairman of the CEN/TC 121 TOFD ad hoc group, Dr Ewert
(BAM) and Dr E. Zeelenberg (Lloyd Register DK) were permanent
Observers.
CEN/TC 121 adopted WI 00121377 Welding Use of time-of-flight diffraction
technique (TOFD) for testing of welds as a Technical Specification including
parts of the TOFD procedure written in task 2.1.
Various publication have been presented:
o
In-Service Inspection and Life Management of Pressure Equipment,

TOFDPROOF project presentation at the EPERC workshop, Didier Flott,
5th October 2001;
33
o Treatment of Uncertainties in Determination of Acceptance Criteria for

Ultrasonic Testing, by A. Jovanovic, D, Balos, MPA Stuttgart, R. Kauer,
TV Sddeutschland Mnchen Paper presented at the UNCERT-AM
conference on Management of Uncertainties in Mechanical Testing &
Inspection on October 8, 2003, Stuttgart, Germany;
o Communication on TOFDPROOF objectives and achievements at the
International Conference on Pressure Vessels ESOPE 2004 Paris,
October 2004, D. Flott, D. Chauveau and al.
o TOFDPROOF workshop, 14th March 2005, Lisbon, Portugal, organised by
ISQ.
o Communication at the national COFREND conference (May 2005)
"Description and results of the European project TOFDPROOF", D. Flott,
D. Chauveau - Institut de Soudure
o TOFDPROOF workshop, 27th May 2005, Villepinte, France, organised by
IS and all partners.
TOFD acceptance criteria was on the agenda of the CEN/TC 121 Berlin meeting in
May 2005. The previous year, the Dutch Normalisation Committee submitted the
Dutch National standard NEN 1822 (Acceptance Criteria for the TOFD technique).
However, the document prepared under the TOFDPROOF project was more in line
with the latest developments and with the other European Standards. It was
suggested by Jan Verkooijen (Sonovation), that the TOFDPROOF document should
be used as an European Standard as a starting point.
The Acceptance Standard is now accepted as a new Work Item. Five countries are
participating in the work, that is: France, Germany, Netherlands, United Kingdom and
Finland.
This proposal is the starting point for the CEN standardisation. The first meeting of
the ad hoc group was scheduled in autumn 2005.
2.3.6.3 Comments
Two out of the main deliverables of the programme were not disseminated as public
documents through the TOFDPROOF web site, that is:
Analysis of performances and reproducibility of TOFD inspection with
conventional NDT
Design and proposal of acceptance criteria with several severity levels.
The information from the TOFDPROOF project was fed to a number of TC groups
and it is clear that the results from the TOFDPROOF project were taken into
consideration. Draft reports were submitted as references for new EU standards.
Work still needs to be carried out to finalise these documents.
34
2.3.7
Work Package WP7: Data storage, analysis and exchange
2.3.7.1 Objectives
This work package concentrates on exchanging information between the partners
(reports, minutes from meetings, etc.) through the web site created under this work
package. The web site also enables access to the data collected during the project
for analysis.
A wide collection of data was to be analysed by specific tools developed within the
work package in order to highlight weaknesses and strengths of the TOFD
inspection.
This work package concentrates on exchanging information between the partners
including reports, minutes from meetings, etc. the exchange was carried out through
the web site created under this work package: http://www.mpa-lifetech.de/tofd/. The
web site also enables access to the data collected during the project for analysis
once the data collection was completed.
The existing tools (MPAs ALIAS system) were used as initial platform for results
analysis. Based on the literature overview and on the results of similar projects, it
was decided to perform the hit-miss data analysis using the Weibull three parameter
cumulative distributions (as suggested by KINT report).
It was concluded by the work package leader that for the purpose of data analysis
and development of acceptance criteria, the applied data analysis methods gave
more than reasonable and feasible results, comparable with similar results obtained
in other related projects.
Furthermore, the TOFD training tool was made available through the TOFDPROOF
web site to the general public free of charge. In order to have a better overview of the
general public interest, it was decided by the consortium to require registration for the
use of the training tool. After approximately 3 months after the training tool was made
available, the number of registered users has reached 50, ranging mostly from
European countries, nevertheless participants from country such as US, Canada,
China and India have also been registered, thus showing that the interest for the
project results is not only limited to EU countries.
2.3.7.3 Comments
Although the website was useful to access information from the project, the
information was fed to the website very slowly and some documents were not made
available until the very end of the project (well after the round robin trials were
completed). Some of the newer issued reports are still not available through the site,
at the time of writing this current report.
35
3.
SAFETY RELATED ISSUES
3.1
INTRODUCTION
Mitsui Babcock previously reported significant limitations of the TOFD technique in a
previous project Critical evaluation of TOFD for search scanning carried out for the
HSE. Mitsui Babcock highlighted situations where TOFD limited for detection and it
was concluded that, even if TOFD is a good complementary NDT technique for
defect detection, there are circumstances where it should not be used as a stand
alone search method.
There is a tendency to favour TOFD as a replacement to more traditional NDT
methods, especially radiography (e.g. Code case 2235 of the ASME Boiler and
Pressure Vessel Code). One of the main disadvantages of radiography is the
potential hazard to health associated with the ionising radiations which are the basis
of the method. Ultrasonic inspection does not have any significant inherent safety
issues and therefore can be more attractive to apply than radiography. However, it is
important that the associated safety and economic advantages are not gained at the
expense of reduced confidence in weld integrity.
Although it should be possible to minimise the deficiencies of TOFD using optimised
parameter settings it might not always be considered as the inspection could become
costly.
When considering TOFD as a stand-alone technique for the inspection of safety
critical components, the following limitations should be kept in mind:
The diffracted tip wave is relatively small in amplitude so the sensitivity of the
NDT needs to be high which can then lead to false calls;
Other techniques should be applied to cover the near and rear surface
regions.
Near surface defects could be missed due to the presence of the lateral
wave. If such defects are of concern, then additional techniques should be
performed e.g., ultrasonic pulse echo using shear wave at full skip, especially
to investigate the cap area of an undressed weld.
Rear surface defects could be missed due to the presence of the backwall
echo. This is even more likely when the root bead is still present. It was
demonstrated during previous research (including previous work for the HSE
[17]) and during the TOFDPROOF project that defects up to 4mm height
could be missed when using the TOFD technique. Due to the uncertainty of
defect detection in the weld root, it was concluded that complementary
methods should be use.
As the weld thickness increases so does the number of probe separations
which are required to cover the inspection volume.
The technique requires optimization for the defects of concern.
Skilled operators are required to operate the equipment and interpret the
images.
36
Resolution between the defect tips may be difficult to achieve which can lead
to misinterpretation of defect.
The results from the TOFDPROOF project round robin trials confirm some of the
TOFD limitations and highlight the need for an appropriate procedure, a skilled data
analyst and for realistic acceptance criteria.
3.2
ACCEPTANCE CRITERIA
An acceptance criteria document is clearly required if TOFD is to be used as a stand
alone inspection method. Carefully specified acceptance criteria are required to
ensure component integrity without unnecessary rejection e.g. due to innocuous
defects or false calls.
The application of the TOFD standards (BS 7706:1993 and DD ENV 583-6:2000) has
been slowed down by the lack of appropriate defect acceptance criteria. The need of
such criteria is apparent.
The criteria prepared under the TOFDPROOF project have now been presented to
CEN/TC121/SC5B and CEN/TC54/WGE for adoption as a work item in order to
integrate them in EN standards.
The acceptance criteria document presented is considered to provide some
unrealistic targets as it assumes a very high capability of detection for search
scanning which is not always achievable mainly at the dead zone area. Small
surface breaking defects may be just detectable or/and wrongly sized.
The results from the TOFDPROOF trials indicate that the height of upper surface
defects is difficult to determine and in many cases these are inaccurately sized.
Where the apparent extent of the lateral wave is greater than or equal to the
acceptable height of a surface defect, it may be appropriate to regard all detected
upper surface defects as possibly rejectable regardless of their measured height and
then use a complementary technique to determine acceptability.
The accuracy of TOFD depends on defect location, probe arrangement and
specimen geometry. The rule of thumb which Mitsui Babcock are aware of and which
was concluded from independent trials is that an accuracy of 2% of the specimen
thickness is obtainable for dedicated sizing scans and 3% for search scans under
ideal conditions.
Whilst this may be the case, errors previously calculated by Mitsui Babcock (shown
in Appendix 2) are less optimistic. Taking the example in Appendix 2, a 3mm high
backwall defect in a 15mm thick specimen. Adding up the errors, as shown in Table 1
of Appendix 2, provides a worst-case estimate. In practice worst-case errors for all
aspects at the same time are unlikely so that there is a reasonable prospect of
justifying an error tolerance of around 2mm.
Acceptance levels should be established in accordance with the capability of
individual techniques to detect and discriminate certain types of imperfections and to
determine their size in relation to the quality requirements. However, the proposed
acceptance levels, do not differentiation between defect types. TOFD can not easily
differentiate between planar and non planar embedded defects. In practice resolving
defect tips may not always be achievable and small cracks could be misinterpreted
37
as slag. Similarly, slag has been reported as embedded linear defects when
diffracted signals were observed from the upper and lower edges of the indications.
It is therefore important to consider the maximum allowance height and identify
whether it as an achievable value. It is why Mitsui Babcock have reservations over
the acceptance level criteria proposed under the TOFDPROOF project. The
reservations are mainly over the following points:
The criteria assume that a surface breaking flaw height measurement can be
achieved with a resolution of at least 0.5mm;
The criteria assume that surface breaking flaw can be resolved from the
lateral/backwall echo with a resolution as little as 1.5mm (for thin samples).
The criteria assume that embedded defect edges can be resolved within
2mm.
The criteria do not appear to make much sense from a structural integrity
viewpoint, e.g. if one can accept 2mm x 38mm in 50mm thick weld then why
not 3mm x 50mm in weld thickness 100mm.
According to the tables, the maximum allowable height when the length of the
defect exceeds the thickness of the specimen is 1mm for specimens up to
50mm thick for acceptance level 1 and 2. Since it is not considered possible
to measure height less than 1mm by TOFD, this implies that most of the
defects detected (if not all) would be rejected for welds of thickness below
50mm if defect length exceeds the wall thickness.
Moreover, the criteria presented under the TOFDPROOF project are for inmanufacture inspections, however there were discussions on extending their
applications to in-service inspection. The criteria would be even more inappropriate
for in-service inspection as factors such as surface finish/irregularity, accessibility,
mismatch of parent material and environment would have even more influence on the
image quality achievable. Further, the TOFD technique is often used as a monitoring
technique for defects previously detected and analysed by another NDT technique
using the relevant acceptance criteria. These defects can be in some circumstances
much larger than those specified in the acceptance criteria proposed.
38
4.
CONCLUSIONS
Mitsui Babcock was a partner in the European Collaborative Project TOFDPROOF,
which was intended to provide guidance on the scope and capability of TOFD
procedures and acceptance criteria for the manufacturing inspection of pressure
vessels. Since the conclusions and recommendations could influence the approach
adopted in the UK, Mitsui Babcock has been reviewing the safety implications of
TOFDPROOF for the HSE. The work has involved keeping the HSE aware of
TOFDPROOF progress, reviewing TOFDPROOF results and recommendations
which could impact on component safety, and providing feedback to the
TOFDPROOF consortium on HSE/Mitsui Babcock views.
The TOFDPROOF project was aimed at promoting TOFD and to create documents
which could be used as references for future standards. The project achieved its
main goal, the results of the project being communicated to parties involved in NDT
inspection and a series of documents produced during the project were used as a
base or contribution for recent and future European Standards.
However, it is not clear how/if some of the conclusions were directly produced from
the data collected during the project and how much the results from traditional NDT
were used to compare the capability of the TOFD. Nevertheless, the TOFDPROOF
project has increased awareness of TOFD capabilities and limitations and provides
useful guidelines on the technique.
It is likely that TOFD will continue to be used as a stand alone search technique.
However, the TOFDPROOF project highlighted the danger of missing defects, in
particular in the dead zones, or misinterpreting the results.
The technique suffers from limited coverage resulting from two inspection dead
zones: the first dead zone at the near surface resulting from the lateral wave and the
second at the backwall resulting from the width of the backwall reflection. Coverage
of the whole inspection zone must be achieved. Specific investigations in these areas
should be applied. The best approach could be to use the TOFD technique combined
with the pulse echo technique in order to eliminate the uncertainty of a lack of
coverage.
Moreover, if indications are detected, an additional technique should be applied to
allow for defect characterisation, as a misinterpretation of a defect could lead to
wrongly accepting the defect.
The project showed that missing a defect may not only be due to the capability of the
technique but can be linked to the use of the wrong/inappropriate TOFD setups
or/and poor analysis. The project highlighted the need for specific and full training
and experience.
TOFD can be regarded as a cheap technique mainly because it appears to many as
a one scan inspection, though this can be misleading as different probe types and
arrangements should be used to enable minimise lack of coverage (additional scans
may also be needed to investigate suspicious indications). The reduction of
recommended scans should not be justified on the basic of cost reduction if this
reduces component integrity for safety critical components. The result of the survey
carried out on the cost comparison of the NDT techniques still identified TOFD as the
cheapest technique (for those specialising in the TOFD technique).
39
The main area of concern, however, was considered to be the proposed acceptance
criteria. Until now, one of the most recognised needs for the TOFD technique was
acceptance criteria. The final issue of the TOFDPROOF acceptance criteria
document has recently been accepted as a starting point for CEN standardisation.
The document is considered to present some unrealistic targets as it assumes a very
high capability of detection for search scanning which is not always achievable
mainly at the dead zone area. In general, It would be difficult in practice to resolve
edges or measure defects accurately enough to apply the criteria proposed and it is
doubtful whether the criteria make much sense from a structural integrity viewpoint,
e.g. it accepts 2mm x 38mm in 50mm thick weld but not 3 x 50mm in weld thickness
100mm.
40
5.
REFERENCES
The documents listed below are directly referred to within this report.
[1].
BS EN 1714:1998 Non-destructive testing of welded joints Ultrasonic

testing of welded joints.
[2].
BS EN 1713:1998 Non-destructive testing of welds Ultrasonic testing

Characterization of indications in welds.
[3].
BS EN 1435:1997 Non-destructive examination of welds Radiographic

examination of welded joints.
[4].
DD ENV 583-6:2000 - Non-destructive testing ultrasonic examination

part 6: time-of-flight diffraction technique as a method for detection and
sizing of discontinuities.
[5].
NS Goujon & BW Kenzie, Recommendations for Applying TOFD (Field of

Application, Strengths & Weaknesses), Issue 1, Report No. 2-31-D-200402-1, report for the TOFDPROOF project, January 2005.
[6].
DD CEN/TS 14751: 2004 - Welding Use of time-of-flight diffraction

technique (TOFD) for examination of welds, technique specification.
[7].
BW Kenzie, Recommendations for TOFD Certification, issue 2, TWI

report for the TOFDPROOF project, March 2005.
[8].
BW Kenzie, Recommendations for TOFD training, issue 1, TWI report for

the TOFDPROOF project, February 2005.
[9].
Specific requirements for the certification of personnel engaged in

ultrasonic time of flight diffraction testing of linear butt welds in ferritic
steel, PCN/GEN Appendix C4.1 Issue 1 Rev B, 25 January 2002.
[10].
Cases of ASME Boiler and Pressure Vessel Code Code Case 2235-4:
Use of Ultrasonic Examination in Lieu of Radiography, Section I and
Section VIII, Divisions 1 and 2.
[11].
BS EN ISO 5817:2003 Welding Fusion welded joints in steel, nickel,

titanium and their alloys (beam welding excluded) - Quality levels for
imperfections.
[12].
BS EN 25817:1992 Arc-welded joints in steel Guidance on quality

levels for imperfections (superseded).
[13].
BS EN 12062:1998 Non-destructive examination of welds General

rules for metallic materials.
[14].
BS EN 1712:1997 Non-destructive testing of weld Ultrasonic testing of

welded joints Acceptance levels.
[15].
BS EN 12517:1998 Non-destructive testing of welds Radiographic

testing of welded joints Acceptance levels.
41
[16].
Pr EN 13445 Unfired pressure vessels Inspection and testing.
[17].
NS Goujon & BWO. Shepherd Critical evaluation of TOFD for search

scanning, Report No71/00/037, 13 December 2004 (HSE research report).
42
APPENDICES
43
APPENDIX 1: RECOMMENDATIONS FOR APPLYING TOFD

(FIELD OF APPLICATION, STRENGTHS, WEAKNESSES)
44
TOFDPROOF
Document Name:
RECOMMENDATIONS FOR APPLYING

TOFD (FIELD OF APPLICATION,
STRENGTHS, WEAKNESSES)
Document Date:
Document Owner
Document Author/s:
Document approved by:
Task / Deliverable Number:
Issue:
Status
Document Reference Number:
TOFDPROOF project n
2005-01-11
Mitsui Babcock / TWI
NS Goujon & BW Kenzie

WP2 / 31
1
2-31-D-2005-02-1
G6RD-CT-2001-00626
Institut de Soudure
IS Service
Sonovation
TWI Limited
Mitsui Babcock Technology Centre
Staatliche Materialprfungsanstalt Stuttgart
Tecnatom S. A
VTT
Instituto de Soldadura e Qualidade
TV Sddeutschland Bau und Betrieb GmbH
NL
UK
UK
SP
FIN
PT
COPYRIGHT 2005 The TOFDPROOF Consortium

This document may not be copied, reproduced, or modified in whole or in part for any purpose without written
permission from the TOFDPROOF Consortium. In addition, to such written permission to copy, acknowledgement
of the authors of the document and all applicable portions of the copyright notice must be clearly referenced.
All rights reserved
45
CONTENTS
CONTENTS ........................................................................................................................................................ 46
1.
INTRODUCTION ................................................................................................................................. 47
2.
2.1
2.2
2.3
2.4
2.5
2.6
2.7
RECOMMENDATIONS BASED ON THE TOFDPROOF RRT RESULTS ................................. 47
Procedure ........................................................................................................................................ 47
Identification of reference marks .................................................................................................... 48
Set-up.............................................................................................................................................. 48
Classification of indications............................................................................................................ 50
Evaluation of indications ................................................................................................................ 51
Personnel qualification ................................................................................................................... 56
Acceptance criteria ......................................................................................................................... 56
3.
ADDITIONAL RECOMMENDATIONS............................................................................................ 57
4.
CONCLUSIONS.................................................................................................................................... 57
5.
REFERENCES ...................................................................................................................................... 58
46
1.
INTRODUCTION
Following the round robin trial (RRT) exercise and the reporting of the results, the
data were collated and a review focusing on the causes of discrepancies in the
results was carried out by MBEL and TWI. This was in order to highlight the strengths
and weaknesses of the TOFD technique. Report No 2-18-D-2004-01-1 [1] presents
the discrepancy analysis of the round robin trial results.
In order to complete work package WP2, recommendations were written to explain
when a complementary NDT technique is recommended. This report proposes
recommendations for applying TOFD. The recommendations are based on the
TOFDPROOF project round robin results and the difficulties identified during the
study. Additional considerations must be taken into account when different material
and component geometry are under study. These recommendations will be
transmitted to CEN TC121, TC54, TC138 and EPERC.
2.
R
ECOMMENDATIONS BASED ON THE TOFDPROOF RRT
RESULTS
The following recommendations are based on the TOFDPROOF project round robin
trials results and the discrepancy analysis report No 2-18-D-2004-01-1. They are
provided for seven categories, including:

Procedure;
Identification of reference marks;
Set-up;
Classification of indications;
Evaluation of indications;
Personnel qualification;
Acceptance criteria.
For each group, comments and recommendations are provided.
2.1
P
rocedure
A specific procedure shall be written in accordance with the guidelines given in
ENV 583-6 [2] and PrCEN/TS 14751 [3] for each individual type of inspection.
The TOFD procedure written under the TOFDPROOF project Procedure for
TOFD Inspection of Welds used for the Round Robin Trials report No. 2-21-Q2002-01-4 [4] can be used as an example.
The procedure shall include the following information:
SCOPE

GENERAL

REQUIREMENTS
o Description of the component
o Restrictions
o Surface preparation
o Weld profile
o Coverage
o Possible defects
o Inspection conditions (temperature, lighting, etc.)
o Examination level
47
REFERENCE

DOCUMENTS
PERSONNEL

REQUIREMENTS & QUALIFICATION
EQUIPMENT

REQUIREMENTS
o General requirements
o Equipment
o Scanning mechanism
o Probes
o Reference block
o Couplant
IDENTIFICATION OF REFERENCE MARKS
CALIBRATION

& SETTINGS
o Choice of probes and Probe Centre Separation (PCS)
o Sensitivity setting
o Time calibration
o Time window
o Scan resolution setting
o Verification of the setting
WELD INSPECTION
o longitudinal defects
o transverse defects
DATA ANALYSIS
o Interpretation and analysis of TOFD images

o Assessing the quality of the TOFD image
o Classification and evaluation of indications
DATA STORAGE
REPORTING

2.2
Identification of reference marks

The identification of reference marks (including datum on the component and
reference point on the inspection probes array) is critical to allow repeatability of the
inspection and results comparison.
The reference point on the probe array (usually the middle of the back face of one of
the probes) shall be clearly defined in the procedure and the reference marks on the
component should be clearly visible.
2.3
Set-up
Care should be taken to choose appropriate combinations of parameters.
The capability to cover the thickness range of interest must be demonstrated on a
reference block.
Sensitivity setting
Setting of an adequate sensitivity is essential to enable the detection of weak
diffracted signals and at the same time avoiding overloading the system with nonrelevant signals. The inspection teams must make sure that the lateral wave and
the BWE is not saturated to investigate for possible surface breaking defects.
Selection of probes and probe configuration
48
Always use the most suitable probes for the component and for the type of
defects under investigation. The choice of the type of probes (including:
frequency, crystal diameter and angle) to be used for an inspection can be
critical (especially for defect characterisation). Selection of probes and probe
configuration for full coverage of the complete weld thickness should follow
the recommendations provided in Table 1. The values given in Table 1 relate
to the reviewed values used for the TOFDPROOF project round robin trial.
Thickness
t (mm)
Minimum
number of
DepthFrequency Beam angle Crystal size
Beam
TOFD
range (mm)
(MHz)
()
(mm)
intersection
set-up(s)*
6-10
0-t
15
70
2-3
2/3t
>10-15
0-t
15-10
70
2-3
2/3t
>15-35
0-t
10-5
70-60
2-6
2/3t
>35-50
0-t
5-3.5
70-60
3-6
2/3t
>50-100
0-t/2
5-3.5
70-60
3-6
1/3t
5/6t for 60
or t for 45
* Note that the number of TOFD set-up(s) given in Table 1 is the minimum number of TOFD
set-up(s) recommended.
Table 1: probes set-up versus thickness
t/2-t
5-3.5
60-45
6-12
Other probe types and configuration than those given in Table 1 can be used
after demonstration on an appropriate calibration/reference block, see
Appendix 1 of reference [4].
The probe frequency used has to be high enough to achieve the best possible
resolution taking care of achieving the required sensitivity setting.
High frequency, small crystal diameter probes (15MHz, 3mm) are preferable
for the inspection of thin samples (up to 15mm) especially when the weld
surfaces are as-welded. An alternative choice (e.g. 10MHz) may not be
appropriate.
The probe frequency used has to be high enough to achieve the best possible
resolution. However frequencies at the lower end of the bands defined in
Table 1 may be used if the required sensitivity setting cannot be achieved
with higher frequencies.
Consistency in the inspection method has to be adopted. It is essential that the

inspection team follows the procedure provided, or provides a justification for any
variation between the work carried out and the procedure.
TOFD images are commonly represented by a grey scale. This is to allow for
better contrast and to permit the identification of indications. It is recommended
that this approach be used to allow for repeatability and consistency between
inspection teams (although other approaches can be used as long as they are
well understood).
49
The time window for data collection shall be extended to at least 1s beyond the
first mode converted BWE, in order to study possible defect mode converted
echoes. Note CEN/TS 14751 states that the time window shall at least cover the
depth range covered in Table 1, however, defect information may be provided by
mode converted echoes (e.g. transverse cracks) and therefore it is important to
extend the window to allow collection of mode converted echoes, when
appropriate.
Additional scanning
2.4
For wide welds (especially for as-welded and double-V weld preparation), at
least two offset scans must be considered to achieve the whole weld body
inspection coverage, one at each side of the weld centre line.
A separate root scan should be considered. The resolution of the backwall

echo and the defect tip of a small root defect is difficult, especially when the
root bead is still present. The presence of a weld root can prevent a clear
break of the back-wall (with the exception of the larger defects). Indication of
the presence of defects may be given by the waves arriving after the backwall echo. However, small defects can still be missed, or misinterpreted as
point reflectors.
Inspection from both surfaces is recommended when access permits. This is

especially important for thick as-welded components where satisfactory
coverage of the near surface is not achievable. When inspection from the
internal surface is not possible, it is recommended to use a complementary
NDT technique.
Inspection for transverse indications can be limited especially when the weld
cap is present. The normal TOFD configuration is not optimised for transverse
defect inspection. When transverse defects are expected and the weld is aswelded, additional NDT technique(s) should be used.
Classification of indications
The indications shall be classified into categories clearly defined in the inspection
procedure. The following categories are recommended:
Surface breaking indication (at scanning surface, at opposite surface and 100%
through-wall)
Surface breaking at the scanning surface: this type of indication shows up as
either a weakening, deviation or loss of lateral wave (not always observed)
and an elongated pattern generated by the signal from the lower edge of the
indication. The lower edge can be hidden by the lateral wave, but generally a
pattern can be observed in the mode-converted part of the image. For small
indications, only a slight shift of the lateral wave towards longer time-of-flight
may be observed.
Surface breaking at the opposite surface: this type of indication shows up as
either a weakening, deviation or a loss of the backwall signal (not always
observed) and an elongated pattern generated by the signal emitted from the
upper edge of the indication.
Through-wall indication: this type of indication shows up as a loss or
weakening of both the lateral wave and the backwall signal.
50
Embedded indications (point-like, elongated with a measurable height or without

a measurable height)
Embedded point-like indication: the most common pattern characterised by a
single arc shaped curve fitting the theoretical hyperbolic curve corresponding
to the depth of the indication. This pattern is mostly produced by a pore, but it
can also be generated by the edges of a transverse crack.
Embedded elongated indication with non measurable height: the indication
appears as an elongated pattern corresponding to an apparent upper edge
signal (approximately in phase with the backwall).
Embedded elongated indication with a measurable height: the indication
appears as two separate elongated patterns located at different positions in
depth, corresponding to the upper and lower edges of the indication.
Transverse indications
Transverse indication: can be surface breaking or embedded. The signal from
the upper and lower edges of a transverse crack may appear as a point-like
defect.
Uncategorised
Uncategorised indications: all indications that cannot be properly classified
into one of the above categories.
2.5
Evaluation of indications
Any feature, which is not due to geometry and appears as an indication on the TOFD
image, shall be investigated to the extent that it can be evaluated in terms of
acceptance criteria.
2.5.1
Interpretation of TOFD images

Initial analysis has to be carried out on unprocessed data. Straightening and removal
(for lateral wave and BWE) tools can be use for subsequent analysis e.g.
confirmation of presence/absence of surface defects.
Surface defects
It is well known that one of the limitations of the TOFD technique is the surface
inspection (upper and rear surfaces). The presence of the lateral wave and the
backwall echo restrict the inspection zone. Small defects in these zones can be
missed. Surface defects are more difficult to detect especially when the weld is in the
as-welded state. Sizing errors are also more likely to occur.
Care must be taken when surface indications are observed. Additionnal scanning
may be needed with more appropriate probes type and arrangements.
The RRT results indicate that the height of near surface defects is difficult to
determine and in many cases these are inaccurately sized. Where the apparent
extent of the lateral wave is greater than or equal to the acceptable height of a
surface defect, it may be appropriate to consider all detected upper surface defects
as rejectable regardless of their measured height and confirm results with another
NDT technique.
Embedded defects
51
A number of defects can be wrongly reported as linear if the resolution of the defect
tips cannot be achieved. The probe frequency used has to be high enough to achieve
the required resolution.
Where a large number of point-like indications have been detected that creates a
cluster of indications, that could mask the presence of a more serious defect, the
inspection should be supported by another NDT technique.
Transverse defects
The presence of some mode converted echoes associated with a point like indication
may suggest that transverse defects could be present. The normal TOFD
configuration is not optimised for transverse defect inspection. When transverse
defects are expected or/and when indications on the TOFD image suggest the
presence of such defects (especially if the weld is as-welded), additional NDT
technique(s) should be used.
2.5.2
Determination of indications dimensions

Determination of length
Indications with a length equal to or less than the probes beam width will appear as a
single hyperbolic shaped arc (point-like discontinuity).
For elongated discontinuities with or without a measurable height, depending upon
the type of indication, a technique for length sizing shall be selected from the
following:
Length sizing of linear indications:

This type of indication does not have length measurement characteristics
which change significantly in the through-wall direction, i.e. embedded defects
like slag and lack of fusion.
A hyperbolic cursor, shaped to fit the arc produced by a point-like flaw, is
fitted to the indication. The cursor is fitted at both extremities of the indication
and the difference between the measured positions of the turning points on
the cursors provides the length of the indication, see Figure 1.
52
Figure 1: Length sizing by fitting arc-shaped cursors

If the hyperbolic cursors do not fit the extremities of the indication, the 6 dB
drop method shall be used. The maximum amplitude (where the reflector
extends across the full width of the ultrasonic beam) shall be determined
using the cursor. The extremities of the indication shall be identified where the
amplitude provided by the cursor has fallen by half, see Figure 2.
Figure 2: Length sizing by the 6 dB drop method.
53
Length sizing of extended parabola-like indications:

This type of indication has length measurement characteristics which change
significantly in the through-wall direction, e.g. surface breaking defects like
cracks.
A cursor, shaped to fit the arc produced by a point-like flaw, is positioned at
either end of the indication at a time delay of one third of the indication
penetration. The distance moved between the cursor positions at each end of
the indication is taken to represent the length of the indication, see Figure 3.
Figure 3: length sizing of extended parabola-like indications.
Determination of depth and height

The depth and the height of the indication shall be determined as follows:
Assuming that the ultrasonic energy enters and leaves the specimen at the
probes index points and that the discontinuity is mid-way between the two
probes, the depth of the defect can be given by:
d = [ c2(t - to) - S ] 1/2
Where: c is the ultrasonic velocity
t is the transit time
to is the total time delay in the probe shoes
d is the depth of the tip of the discontinuity
S is the mid-distance between the ultrasonic probes index points
To prevent errors that may arise from the estimation of probe delay and probe
centre separation distance, the depth d shall be calculated, where possible, from
the time of flight differences, T, between the lateral wave and the diffracted
pulse or between the backwall echo and the diffracted pulse. Moreover, in order
to reduce the error related to time measurement, the measurement shall be done
54
from the A-scan and by choosing a consistent position on the waveforms. It is

recommended to use one of the methods described below (see Figure 4).
Method 1: by measuring the transit time to the rising signal.

Method 2: by measuring the transit time to the first maximum.
Method 3: by measuring the transit time to the peak amplitude.
Lateral wave or
defect lower-tip
Positions for measuring the transit time
Backwall or
defect upper-tip
Method 1
Method 2
Method 3
Figure 4: Position of the cursor for time measurement Methods 1, 2 & 3.

All depth measurements should be performed after straightening of the position
of the lateral wave or the backwall echo.
The height of a near surface breaking discontinuity is determined by measuring
the distance between the near surface and the lower-tip diffraction signal from the
indication.
The height of a rear surface breaking discontinuity is determined by measuring
the distance between the rear surface and the upper-tip diffraction signal from the
indication.
The height of an embedded discontinuity is determined by the difference in depth
between the upper-tip and lower-tip diffraction. For indications displaying varying
depth along their length, the height is determined at the position along the length
of the discontinuity where the difference is the greatest.
Other measurement methods such as those proposed by ENV 583 part 6 can be
used as long as a justification is provided.
55
2.5.3
Sizing errors
Errors in reported height and length measurements for the RRT were reported.
In general, the discrepancies in length measurement related to the intended
values were significant. The standard deviation of the errors in the reported
length measurements for the RRT was 11.7mm. The errors may have been partly
related to differences between the intended values and the real values. However,
sizing errors also varied between inspection teams. On this basis, length
measurement from TOFD techniques should therefore be treated with caution.
The variations in reported height measurement may have been related to the
teams choice of the variables used to linearise the TOFD results (such as: the
reference time to the lateral wave, the reference time to the backwall echo, the
velocity and component thickness). The standard deviation of the errors in the
reported height measurements for the RRT was 2.0mm. It is important to use a
defined measurement technique.
The sizing method used to determine the defect dimensions (height and length)
should be clearly defined in the inspection procedure. This is in order to provide a
repeatable measurement technique and to allow comparison between inspection
teams and repeat inspections. The measurement techniques used for the calibration
and on the actual component should be consistent.
2.6
Personnel qualification
As the detection and sizing performance were highly dependant on the inspection
team, it was concluded that the training and experience of the inspection
personnel is critical.
In addition to a general knowledge of ultrasonic weld inspection, all key personnel
should be experienced in TOFD inspections.
At least one of the inspection personnel should be familiar with preparation of
written test instructions, final off-line analysis of data and be qualified to approve
the final inspection report. This inspection personnel should be certified as a
minimum to level 2 in accordance with EN 473 [5] or equivalent in ultrasonic
testing for the relevant industrial sector.
In cases where the above minimum qualifications are not considered adequate,
job-specific training should be carried out.
2.7
Acceptance criteria
Carefully specified acceptance criteria are required to ensure component integrity
without unnecessary rejection e.g. due to innocuous defects or false calls.
56
3.
ADDITIONAL RECOMMENDATIONS
Some TOFD technique limitations and recommendations were identified under
another project [6]:
A better response is obtained when the included angle between the probes is
120. Experimental results confirms the theory.
Offset-scans, that is, scans parallel to the weld-axis, where the beam intersection
point is not on the centre-line of the weld, should be carried out (especially for
thick X-shaped welds to ensure detection of toe cracks at the surface opposite
the scanning surface). Omitting offset-scans could lead to depth position errors,
e.g. indications will tend to be plotted deeper than their true through-wall location.
The operators must ensure the proper coverage of the area of interest. TOFD
can be limited by the geometry of the sample or by an obstruction limiting the
scanning area. For example: at the ends of long seams adjacent to
circumferential seams (require grinding); inspection of mismatch pipe to pipe weld
and material of small wall thickness t such as t 10mm thick. When the required
coverage is not achieved by TOFD, additional NDT techniques are required to
complete the inspection.
Existence of a dead zone of the order of 2-3mm below the scanning surface.
This problem also occurs at the back-wall but the extent of dead zone may vary.
TOFD is not reliable for detecting surface defects of height less than 4mm.
Experimental results showed that root defects with a depth of less than 4mm are
easily missed or misinterpreted. Moreover, the difficulty of detection increases
with the defect offset position relative to the weld centre line.
4.
CONCLUSIONS
This document provides recommendations for applying TOFD.
The recommendations are provided for seven categories, including: procedure,
identification of reference marks, set-up, classification of indications, evaluation of
indications, personnel qualification and acceptance criteria. For each group,
comments and recommendations are provided.
The results from the TOFDPROOF project round robin trials confirm some of the
TOFD limitations and highlight the need for an appropriate procedure, a skilled data
analyst and for realistic acceptance criteria.
57
5.
R
EFERENCES
[1]
NS Goujon & BW Kenzie, Discrepancy Analysis of the Round Robin Trial

Results, TOFDPROOF report No 2-18-D-2004-01-1.
[2]
ENV 583-6: Non destructive testing Ultrasonic examination. Part 6: Time of

flight diffraction technique as a method for defect detection and sizing.
[3]
PrCEN/TS 14751: Welding Use of time of flight diffraction technique for

examination of welds.
[4]
D Flott, Procedure for TOFD Inspection of welds used for the Round Robin
Trials TOFDPROOF report No 2-21-Q-2002-01-4.
[5]
EN 473: Qualification and certification of NDT personnel General principles.
[6]
JM Farley, NS Goujon & BWO. Shepherd Critical evaluation of TOFD for

search scanning, 16th WCNDT 2004, Montreal, Canada.
58
APPENDIX 2 EXAMPLE OF TOFD ERRORS FOR A 15MM THICK SAMPLE
59
1.
COMPARISON OF SIZING METHODS AVAILABLE
As an example, consider a 15mm butt weld containing a defect tip 3mm above
backwall (hence depth of defect is 12mm from scan surface) scanned with TOFD
probes such that their index points are separated by 60 mm (= pair of 63 probes).
Assume velocity of 5.9 mm.s-1.
Two TOFD sizing methods are available:
a) Measure the difference in arrival time of the defect tip echo and the backwall
echo.
b) Measure the difference in arrival time of the defect tip echo and the lateral wave.
Both methods have drawbacks.
Measurement relative to the backwall echo is affected by uncertainty as to the profile
of the backwall over the weld. The large difference in amplitude between the BWE
and the tip echo makes it difficult to measure the time difference between the two
signals accurately.
Measurement relative to the lateral wave depends on the lateral wave being
successfully propagated through the weld. This should not be a problem for ferritic
welds. For austenitic welds, propagation through the weld parallel to the pipe axis is
likely to be affected by high attenuation. There is also a possibility of the beam axis
following a curved path which could introduce significant error into the size estimate.
Mismatch will complicate depth estimates relative to a lateral wave.
The effect of individual errors is presented in a series of graphs below. In each case,
the vertical axis represents the depth estimate which would result as a function of the
value of the error in the particular variable from its correct value.
The errors detailed in the following pages are summarised below:
Measurement wrt
BWE
(mm)
Measurement wrt
Lateral Wave
(mm)
Probe separation +/- 2 mm
0.2
0.4
Timing +/- 1 cycle (0.2 s)
1.7
1.7
Velocity +/- 0.2 mm.s-1
0.1
0.2
Thickness +/- 2 mm
0.5
N/A
Total
2.5
2.3
Error component
Table 1: Summary of the errors

Hence errors are of the same order for both methods for a small defect at the
backwall. The most important error component is the timing error. This can
potentially be reduced particularly in the case of referencing relative to the lateral
wave.
Adding up the errors as shown provides a worst case estimate. In practice worst
case errors for all components at the same time are unlikely so that there is a
reasonable prospect of justifying an error tolerance of 2 mm or less.
60
2.
DEPTH MEASUREMENT RELATIVE TO BACKWALL ECHO
2.1
Variation in reported depth with probe separation

An error of +/- 2 mm in measuring
probe separation gives a +/- 0.2 mm
error in depth.
Note that the
separation is measured between the

index points so the separation is
unlikely to be measured more

accurately than +/- 2 mm.
12.5
12.5
12.25
d( s
)
12
11.75
11.5 11.5
2.2
3
3
Variation in reported depth with time of flight

Time-of-flight errors will be of the
order of the cycle length. It is often
claimed that timing errors can be

reduced to a small fraction of a
cycle but in practice there can be
uncertainty as to which cycle to

measure from (especially when
referencing relative to a back wall).

If we use a 5 MHz probe then one
cycle is 0.2 s giving an error of +/1.7 mm.
15
15
14
13
d( t
)
12
11
10
0.4
0.3
0.2
0.1
0.1
0.2
0.3
0.4
2.3
Variation in reported depth with velocity

The velocity will be uncertain but
variation from standard values is
unlikely to be more than +/- 0.2
mm.s-1 assuming we are dealing
with wrought piping. Errors due to
this uncertainty are only +/-0.1 mm.
12.2
12.2
12.1
d( v )
12
11.9
11.8 11.8
0.2
0.2
61
0.1
0
v
0.4
0.4
0.1
0.2
0.2
2.4
Variation in reported depth with wall thickness

The wall thickness will be uncertain
since the form of the backwall close
to the weld will be uncertain.

Tolerances may be up to about +/- 2
mm. This would affect the depth

measurement from the surface
significantly but in practice the

objective will be to estimate depth
from the inside surface. When this
is taken into account the error in

throughwall dimension would be
reduced to about +/- 0.5 mm.
15
15
14
13
d( T )
12
11
10
DEPTH MEASUREMENT RELATIVE TO LATERAL WAVE
3.1
Variation in reported depth with probe separation

An error of +/- 2 mm in measuring
probe separation gives a +/- 0.4 mm
error in depth.
Note that the
separation is measured between the

index points so the separation is
unlikely to be measured more

accurately than +/- 2 mm.
12.5
12.5
12.25
d( s
)
12
11.75
11.5 11.5
3.2
Variation in reported depth with time of flight

Time of flight errors may be of the
order of the cycle length. It is often
claimed that timing errors can be
reduced to a small fraction of a cycle
but in practice there can be
uncertainty as to which cycle to
measure from (though higher
precision say by a factor of 5 is
often achievable when referencing to
a lateral wave). If we use a 5 MHz
probe then one cycle is 0.2 s giving
an error of +/- 1.7 mm.
15
15
14
13
d( t )
12
11
10
9
0.4
0.4
62
0.3
0.2
0.1
0
t
0.1
0.2
0.3
0.4
0.4
3.3
Variation in reported depth with velocity

The velocity will be uncertain but
variation from standard values is
unlikely to be more than +/-0.2mm.s-1
assuming we are dealing with wrought
piping. Errors due to this uncertainty
are only +/-0.2 mm.
12.2
12.1
d( v )
12
11.9
11.8 11.8
0.1
0.2
63
0
v
0.1
0.2
RR 433

Research Report 433: Safety Implications of TOFD For In-Manufacture Inspections

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Research Report 433: Safety Implications of TOFD For In-Manufacture Inspections

Transféré par

Droits d'auteur :

Formats disponibles

HSE

Health & Safety

Safety implications of TOFD for

Prepared by MitsuiBabcock Energy Limited

RESEARCH REPORT 433

Health & Safety

Safety implications of TOFD for

Crown copyright 2006

reproduced, stored in a retrieval system, or transmitted in

any form or by any means (electronic, mechanical,

photocopying, recording or otherwise) without the prior

written permission of the copyright owner.

Applications for reproduction should be made in writing to:

Licensing Division, Her Majesty's Stationery Office,

St Clements House, 2-16 Colegate, Norwich NR3 1BQ

CONTENTS .............................................................................................................................................. iii

TOFDPROOF PROJECT ....................................................................................2

TOFDPROOF WORK PACKAGES............................................................................................... 2

WORK RESULTS .......................................................................................................................... 4

Work Package WP1:

Trials organisation and justification of representative specimens .......... 4

Blind Trials and Performance Assessment ............................................ 8

Guidelines for training, qualification and certification ........................... 20

Acceptance criteria definition................................................................ 24

Economic analysis ................................................................................ 29

Exploitation and dissemination ............................................................. 33

Data storage, analysis and exchange .................................................. 35

SAFETY RELATED ISSUES ............................................................................36

ACCEPTANCE CRITERIA .......................................................................................................... 37

APPENDIX 1: RECOMMENDATIONS FOR APPLYING TOFD (FIELD OF APPLICATION,

STRENGTHS, WEAKNESSES) .............................................................................................................. 44

APPENDIX 2 EXAMPLE OF TOFD ERRORS FOR A 15MM THICK SAMPLE ................................... 59

TOFDPROOF WORK PACKAGES

WP1: Trials organisation and justification of representative specimens

WP2: Blind trials and performance assessing

WP3: Guidelines for training, qualification and certification

WP4: Acceptance criteria definition

WP5: Economic analysis

WP6: Exploitation and dissemination of results, link with CEN

WP7: Data storage, analysis and exchange

Table 1 Overview of the consortium (extract from TOFDPROOF proposal)

Main mission / Business activity /Area of activity

Non profit company.

2.3.1 Work Package WP1: Trials organisation and justification of representative

Table 2 Summary of the sample population

5 teams to perform TOFD testing

4 teams to perform UT testing

3 teams to perform X-ray testing

2 teams to perform ray testing

1 team to perform accelerator testing

Test Groups and Tests per Sample

Conventional methods (including UT and RT) and TOFD were to be applied

Work Package WP2: Blind Trials and Performance Assessment

2.3.2.2 Work Performed

Table 4 Recommended TOFD Set-up

The classification of relevant indications was specified by analysing the following

Mode converted signal after the first back wall reflection.

The indications were classified as follows:

b) Experimental work: Round Robin Trials (RRT)

: Scale parameter, calculated using selected mean POD values

Figure 1 POD curve for plate thickness up to 6 mm

Figure 2 POD curve for plate thickness 6 to 15 mm

Figure 3 POD curve for plate thickness 15 to 25.4 mm