Biuletyn Instytutu Spawalnictwa No. 3/2013

Transkrypt

BIULETYN
ISSN 2300-1674
INSTYTUTU SPAWALNICTWA
No. 3/2013
INSTITUTE OF WELDING BULLETIN
BIULETYN
INSTYTUTU SPAWALNICTWA
No. 3
BIMONTHLY
Volume 57
CONTENTS
•T. Pfeifer, L. Aucott, A. Żak – Testing of solidification cracking susceptibility
of MAG welded overlays on S355JR steel using a Transvarestraint test.......................
5
•W. Jamrozik, M. Fidali – Application of image fusion to identification
of welding imperfections............................................................................................12
•A. Sawicki - Damping factor function in AC electric arc models. Part 2:
Damping factor function in universal electric arc models with moderate cooling......... 22
•R. Krawczyk, P. Wojtas, K. Poch - Comparative assessment of selected
welding imperfections in VT, PT and MT methods........................................................ 31
•M. Stachurski – Non-destructive testing of helically welded pipes made of
thermomechanically rolled materials used for sending of combustibles........................ 37
This work is licenced under
Creative Commons Attribution-NonCommercial 3.0 License
INSTITUTE OF WELDING
The International Institute of Welding
and The European Federation for Welding,
Joining and Cutting member
Summaries of the articles
T. Pfeifer, L. Aucott, A. Żak Testing of solidification cracking
susceptibility of MAG welded
overlays on S355JR steel using a
Transvarestraint test
A. Sawicki - Damping factor
function in AC electric arc models.
Part 2: Damping factor function in
universal electric arc models with
moderate cooling
Application of the Transvarestraint test
to assessment of solidification cracking susceptibility of overlay welds made by MAG
method on S355JR steel has been presented.
A specific test rig enabling to apply different strain of the test piece with high speed
was designed. In the course of the research
the weld pool temperature history was registered, cracking qualification was done and
the examination of crack opening and HAZ
using scanning electron microscopy and light
microscopy was carried out. Solidification
cracking temperature range of this material
system was determined.
Nonlinear function of the heat processes
damping factor in thermal column plasma
have been introduced to modified mathematical models of the AC electric arc. This way,
universal arc models with moderate cooling
and constant geometric sizes, using static
characteristics, have been created. Similar
method has been used to hybrid arc models with variable column length connecting
Berger and Kulakov models. It has been proposed the introduction of nonlinear damping
factor function to modified Voronin model
using static characteristics. Thereby the extended range of current for application of
these models has been achieved.
W. Jamrozik, M. Fidali – Application
of image fusion to identification of
R. Krawczyk, P. Wojtas, K. Poch welding imperfections
Comparative assessment of selected
It has been presented the issues concerning welding imperfections in VT, PT
the application of images fusion registered in and MT methods
visible and infrared radiation for diagnostics
of MIG/MAG processes. The sequences of
visible and infrared images visualizing the
arc burning during performing the weld were
examined. Recorded during experiments image sequences were subjected to fusion and
next to analysis in order to obtain diagnostic
signals. On the basis of selected features of
diagnostic signals it was carried out the classification of the welding process state and its
results were compared with those obtained
for diagnostic signal features determined independently for visible and infrared images. The results show that the application of
images fusion enables to identify effectively various welding imperfections forming in
welding processes.
No. 3/2013
The issues connected with assessment of
sensitivity in selected NDT methods are considered. In particular, the attention has been
paid to the tests which make it possible to
detect any imperfection of surface, i.e. visual
testing, liquid-penetrant and magnetic-powder
method. The aim of the studies was to reveal
the analogy and differences between results
obtained from series of tests (VT, PT and MT).
M. Stachurski – Non-destructive
testing of helically welded pipes
made of thermomechanically
rolled materials used for sending
of combustibles
In the first part of the paper it has been
presented the short information on methods
BIULETYN INSTYTUTU SPAWALNICTWA
3
of fabrication of helically welded pipes used
for transporting of combustibles. In the second one it has been given the NDT methods
used during inspection of the pipes in production plants. In the next parts it has been
described the NDT methods for steel pipes
focusing first of all on visual, ultrasonic and
radiographic ones. The paper has been ended
with the information about prospects of gas
industry development in Poland and in the
world.
Biuletyn Instytutu Spawalnictwa
ISSN 2300-1674
Publisher:
Instytut Spawalnictwa (The Institute of Welding)
Editor-in-chief: Prof. Jan Pilarczyk
Managing editor: Alojzy Kajzerek
Language editor: R. Scott Henderson
Address:
ul. Bł. Czesława 16-18, 44-100 Gliwice, Poland
tel: +48 32 335 82 01(02); fax: +48 32 231 46 52
E-mail: [email protected];
[email protected];
[email protected]
www.bis.is.gliwice.pl
Prof. Jacek Senkara - Warsaw University of Technology,
Biuletyn Scientific Council:
Akademik Borys E. Paton - Institut Elektrosvarki im. E.O.
Patona, Kiev, Ukraine; Nacionalnaia Akademiia Nauk
Ukrainy (Chairman)
Prof. Luisa Countinho - European Federation for Welding,
Joining and Cutting, Lisbon, Portugal
dr Mike J. Russel - The Welding Institute (TWI),
Cambridge, England
Prof. Andrzej Klimpel - Silesian University of Technology,
Welding Department, Gliwice, Poland
Prof. Jan Pilarczyk - Instytut Spawalnictwa, Gliwice, Poland
Biuletyn Program Council:
External members:
Prof. Andrzej Ambroziak - Wrocław University
of Technology,
Prof. Andrzej Gruszczyk - Silesian University of Technology,
Prof. Andrzej Kolasa - Warsaw University of Technology,
Prof. Jerzy Łabanowski - Gdańsk University of Technology,
Prof. Zbigniew Mirski - Wrocław University of Technology,
Prof. Jerzy Nowacki - The West Pomeranian University
of Technology,
dr inż. Jan Plewniak - Częstochowa University of Technology,
4
Prof. Edmund Tasak - AGH University of Science
and Technology,
International members:
Prof. Peter Bernasovsky - Výskumný ústav zváračský Priemyselný institút SR, Bratislava, Slovakia
Prof. Alan Cocks - University of Oxford, England
dr Luca Costa - Istituto Italiano della Saldatura,
Genoa, Italy
Prof. Petar Darjanow - Technical University of Sofia,
Bulgaria
Prof. Dorin Dehelean - Romanian Welding Society,
Timisoara, Romania
Prof. Hongbiao Dong - University of Leicester, England
dr Lars Johansson - Swedish Welding Commission,
Stockholm, Sweden
Prof. Steffen Keitel - Gesellschaft für Schweißtechnik
International mbH, Duisburg, Halle, Germany
Eng. Peter Klamo - Výskumný ústav zváračský Priemyselný institút SR, Bratislava, Slovakia
Prof. Slobodan Kralj - Faculty of Mechanical Engineering
and Naval Architecture, University of Zagreb, Croatia
Akademik Leonid M. Łobanow - Institut Elektrosvarki
im. E.O. Patona, Kiev, Ukraine;
dr Cécile Mayer - International Institute of Welding,
Paris, France
Prof. Dr.-Ing. Hardy Mohrbacher - NiobelCon bvba,
Belgium
Prof. Ian Richardson - Delft University of Technology,
Netherlands
Mr Michel Rousseau - Institut de Soudure, Paris, France
Prof. Aleksander Zhelew - Schweisstechnische Lehr- und
Versuchsanstalt SLV-München Bulgarien GmbH, Sofia
Instytut Spawalnictwa members:
dr inż. Bogusław Czwórnóg;
dr hab. inż. Mirosław Łomozik prof. I.S.;
dr inż. Adam Pietras; dr inż. Piotr Sędek prof. I.S.;
dr hab. inż. Jacek Słania prof. I.S.;
dr hab. inż. Eugeniusz Turyk prof. I.S.
No. 3/2013
Research
Tomasz Pfeifer, Lee Aucott, Aleksander Żak
Testing of solidification cracking susceptibility
of MAG welded overlays on S355JR steel using
a Transvarestraint test
Abstract: Application of the Transvarestraint test to assessment of solidification
cracking susceptibility of overlay welds made by MAG method on S355JR steel has
been presented. A specific test rig enabling to apply different strain of the test piece
with high speed was designed. In the course of the research the weld pool temperature history was registered, cracking qualification was done and the examination of
crack opening and HAZ using scanning electron microscopy and light microscopy
was carried out. Solidification cracking temperature range of this material system
was determined
Keywords: welding, cracking, MAG, Transvarestraint test
Introduction
Solidification hot cracking of welded
joints is commonly regarded as one of the
most hazardous of various types of cracking.
For this reason the mechanisms of cracking and hot cracking resistance of different
structural materials are a subject of thorough
research. Such cracks, developing usually on
grain boundaries, are generated in a weld at
the end of a crystallisation process when a
metal being the mixture of a liquid and a solid body is torn as a result of deformations
related to weld material contraction accompanying crystallisation and self-cooling.
One of reliable tests enabling the assessment of steel susceptibility to solidification
cracking is the Transvarestraint test. The
methodology of this test has been developed
directly on the basis of the Varestraint test
whose precursors were Savage and Lundin
[1]. However, neither of these tests have
become standardised and, as a result, today
there are many different versions of them [2].
The purpose of research conducted using
the Transvarestraint test was to determine a
temperature range in which solidification
cracks are generated and develop. The range
is determined on the basis of the measurement of the maximum crack length (MCL)
generated in a sample subjected to a high-rate
strain after welding as well as on the basis
of a recorded temperature course in a weld
[3-8]. Most publications related to the Transvarestraint test describe tests conducted using
TIG welding without a filler metal. However,
the publication [9] refers to the possibility of
welding with a filler metal, which enabled determining the hot cracking resistance of a fill-
dr inż. Tomasz Pfeifer (Ph.D. Eng.) – Instytut Spawalnictwa, Zakład Technologii Spawalniczych
/Welding Technologies Department/; M.Sc. Eng. Lee Aucott – University of Leicester; mgr inż.
(M.Sc. Eng) Aleksander Żak – Frenzak ltd., Mikołów, Poland
No. 3/2013
5
er metal or of a specific material system (filler
metal + base metal). In such cases the use of
appropriate welding parameters, enabling the
proper overlay weld fusion into the base metal
is of great importance. This was the reason for
undertaking tests aimed at the determination
of hot cracking resistance and solidification
cracking temperature range using the Transvarestraint for the following material system:
electrode wire PN-EN ISO 14341-A-G3Si1
and steel S355JR (base metal).
The assessment of susceptibility to hot
cracking and the determination of solidification cracking temperature range of such a
material system is important from the Mint-Weld project implementation point of view
(contract no. NMP3-SL-2009-229108). The
tests involve the preparation of a reliable and
useful numerical model of the development
of interfaces during the welding and operation of welded joints. The model should take
into consideration molecular dynamics phenomena in the atomic scale (the properties
of interfaces, the chemical composition and
structure), phase models in the nano/micro
scale (the structure of grain boundaries and
the chemical composition of interface boundaries), mesoscopic models tracking the front
bracket
bending
block
rolls
of interfaces (crystal growth and grain size
distribution) as well as modelling using fluid
dynamics in the macroscopic scale (heat and
mass flow). The obtained results of the solidification cracking susceptibility assessment
tests constituted necessary and very important input data for modelling the development
of interphase boundaries in various scales.
The tests and their results presented below
are a part of the MintWeld project implemented at Instytut Spawalnictwa within the
confines of the 7th Framework Programme.
Course of tests and results
Transvarestraint test
The tests to determine the solidification
cracking temperature range involved the use
of a base metal in the form of 6 and 12mmthick S355JR steel plates (300 mm × 200 mm)
and an electrode wire PN-EN ISO 14341 – A
– G3Si1 with a diameter of 1 mm. A 180mmlong overlay weld was made on the aforesaid
plates, using a robotised standard arc MAG
method. In all the tests the following constant
welding parameter values were used: current
200 A, arc voltage 25 V and a welding rate of
15 cm/min. The Transvarestraint tests were
carried out with a device built especially for
this purpose in the Frenzak Sp. z o.o. company in Mikołów, one of the MintWeld project
hydraulic
actuator
Fig. 1. Device for Transvarestraint test and manner of mounting and deforming test piece
6
No. 3/2013
partners. The testing device is
presented in Figure 1. The test
plates were mounted in a repeatable manner under a fixing bracket on exchangeable
blocks with a provided arch
of appropriate radius (Fig. 1).
Next, using the MAG method and adopted parameters
a 180mm-long overlay weld
Fig. 2. Results of temperature measurements in liquid metal pool during
was made. During welding
surfacing in Transvarestraint test
the temperature of a weld pool
was measured by a D-type thermocouple transformation of the solidification process.
(tungsten-rhenium; dia. 0.25mm) directly in The temperature measurements revealed that
the liquid pool. After welding (welding arc for the 12mm-thick plate the cooling rate in
termination generated a signal for a bending the above range was approximately 40°C/s,
device) the plate was deformed by means of whereas for the 6mm-thick plate it amounted
a hydraulic actuator and bent on a steel block to approximately 19°C/s.
Solidification cracks appear in the last
perpendicularly in relation to the direction of
welding; the rate of bending being 24 mm/s. crystallisation phase, at a temperature below
Bend radiuses, made on a one edge of the which the whole material is fully solidified.
steel block were selected in a manner mak- The 6mm-thick plates revealed a significanting it possible to generate (on the surface) ly greater number of solidification cracks
a strain corresponding to a specific elonga- (Table 1) than in the 12-mm thick plates. The
tion percentage, i.e. 0.5%, 1%, 2%, 5%, 10% reason behind such a difference is the fact
and 20%. For this reason, for each S355JR for the same welding parameters the heat
steel plate 6 tests were carried out each time emission from the 6mm-thick plate is much
changing a steel block for another one, with slower. As a result, the weld remains significantly longer in the area above the solidifia different bend radius.
cation temperature; on the basis of Figure 2
Temperature measurements
it is possible to determine the cooling rate
Temperature measurements carried out at the liquidus–solidus temperature range. In
during welding revealed that the welds so- the aforesaid range the material is a mixture
lidified at approximately 1460°C (Fig. 2) in of liquid and solid phases, and therefore is
the case of both tested plate thicknesses. In significantly less susceptible to solidification
the case of the 12mm-thick plate the max- cracking. After solidification, a cooling rate
imum registered temperature of the liquid is approximately 74°C/s for the 12mm-thick
metal was approximately 1622°C, which plate and 51°C/s for the 6mm-thick plate.
significantly exceeds a liquidus temperature.
Figure 2, presenting the results of a temper- Quantitative analysis of cracks
ature measurement in the weld pool, shows
After the bend test conducted transversea relatively low temperature gradient from ly in relation to the direction of surfacing,
the maximum temperature to the solidifi- the analysis of cracks generated in the crater
cation temperature. Such a gradient results was carried out. Table 1 presents the results
from the heat release related to the phase obtained and Figure 3 shows the example of
No. 3/2013
7
Table 1. Results of analysis of solidification cracks in the craters of
overlay welds made during the Transvarestraint tests
Plate thickness, mm
6
Strain
%
Number MCL,
of cracks mm
12
TCL, Number MCL,
mm of cracks mm
0
0
0
0
0,5
0
0
0
1
1
10
10
2
2
11
14,5
5
5
21
36
10
16
20
36,5
20
34
21
40,5
Note:
MCL – Maximum Crack Length
TCL – Total Crack Length
0
0
0
1
1
3
4
0
0
0
8
3
8
9
0
0
0
8
3
19
26
Fig. 3. Solidification crack in the crater of
weld made on 12mm-thick plate with 20%
strain, used in Transvarestraint test
the crater in the overlay weld made on the
12mm-thick plate with a 20% strain in which
solidification cracks were observed.
Visual tests and crack length measurements revealed that an increase in the strain
degree is accompanied by an increase in the
number of cracks as well as an increase in
the maximum length of the solidification
crack. Figure 3 presents a longitudinal crack
in the weld axis formed at a significant strain.
The crack was generated at the interface of
grains, developing from the overlay weld
edge towards the main axis of the weld and
is a typical example of a solidification crack.
Fig. 4. Structure of S355JR steel base metal, visible
banded structure of ferrite and pearlite
8
TCL,
mm
The latter part of this publication presents
the results of the metallographic tests of this
crack and the areas adjacent to it.
Metallographic tests
Metallographic tests were carried out for
the test pieces cut out of the 12mm-thick
plates, which during surfacing were subjected to the greatest strain (20%). The tests
included observations with a light microscope and the structural analysis of the base
metal, heat affected zone and, in particular,
the overlay weld line of fusion into the base
metal. The tests also involved the observation of the crack surface using a scanning
microscope. Figure 4 presents the structure
of the base metal which was not exposed to
any welding-related heat impact, and thus
was free from any structural changes.
Figure 5a presents the structure of the
material approximately 1 mm away from
the fusion line, in the HAZ area adjacent
directly to the base metal. This area revealed only small changes in the microstructure (small amounts of bainite). Figure
5b presents the HAZ area near the fusion
line. During surfacing, the area reached a
temperature close to the melting point. The
No. 3/2013
HAZ has a coarse-grained structure com- significantly increase the likelihood of solidification cracking occurence.
posed mainly of ferrite and bainite.
At the next stage of the tests the surface
Figure 6 presents the microstructure in
the fusion line area both on the base met- of a crack in the overlay weld was examal side and on the overlay weld side. Dur- ined. The crack was sampled for a test piece
ing the Transvarestraint test the area was which underwent ultrasonic cleaning and
completely molten. Welding materials of a was observed by means of a scanning elecsimilar chemical composition is accompa- tron microscope FEI Sirion 200. Figure 7
nied by the epitaxial grain growth. In this presents the dendritic structure of the crack
process a liquid crystallises forming new surface. The structure indicates that the
grains directly from the solid phase from crack is a real solidification-type hot crack.
the base metal located underneath. The
a)
b)
grains are anisotropic
as they grow in accordance with a heat
flow direction. In this
manner a characteristic columnar structure is obtained.
Figure 6a presents
Fig. 5. a) Structure of material in S355JR steel HAZ,
a typical columnar
in area adjacent to base metal, coarse-grained structure,
b) Coarse-grained structure in HAZ adjacent to fusion line,
structure, in which it
visible bainite and small amounts of ferrite
is possible to observe
the
Widmanstätten
a)
b)
structure formed directly on the ferrite
of grain boundaries.
Figure 6b shows a
structure composed
mainly of acicular
ferrite and polygonal ferrite. Impurities
Fig. 6. Microstrucure in fusion line area,
present in the steel or a) view from base metal side: visible acicular ferrite and Widmanstätten structure,
weld metal such as
b) view from overlay weld side: acicular ferrite and polygonal structure
phosphorus and sulphur are usually sit- a)
uated between dendrites formed during
crystallisation.
For
this reason, impurities
Fig. 7. Results of tests of
usually concentrate in
solidification crack surface;
the weld axis at the
a) sample preparation manner;
end of the crystallib) image of tested surface obtained
sation process, which
from scanning electron microscope
No. 3/2013
9
Determination of solidification
cracking temperature range
In order to determine the temperature
range of solidification cracking in the overlay welds made on the 12mm-thick plates,
the longest crack recorded in all tests with
a various strain degree was divided by the
welding rate used in a given test. This was
done in order to determine a time interval
(crack length 9 mm, welding rate 2.5 mm/s,
crack formation time interval 3.6 s). Afterwards, the time interval was plotted on a diagram presenting weld pool temperature values recorded during surfacing. The diagram
enabled the determination of the temperature
range (above a solidification temperature) in
which a crack may develop when the weld
material remains a mixture of liquid and solid phases. The wider this range, the higher
the material susceptibility to solidification
cracking. Figure 8 presents the methodology
of determining a solidification cracking temperature range.
G3Si1 on S355JR steel, subjected to a strain
equal up to 20% elongation. The determined
temperature range amounts to 130°C. The
obtained value is significantly higher than
values presented in reference publications
(< 50°C). However, most of the described
tests aimed to determine the hot cracking resistance of a base metal. For this reason the
tests were conducted using the TIG method
without a filler metal. Instead, the base metal
underwent remelting. The higher SCTR values obtained in the tests refer to a specific material system (base metal + filler metal) and
may result from greater stress concentration.
Concluding remarks
The applied testing methodology and the
test rig used in the research made it possible
to carry out a modified Transvarestraint test
with a various strain degree and determine a
Solidification Cracking Temperature Range
during surfacing 6mm-thick and 12mm-thick
S355JR steel plates. The modification of the
Transvarestraint test consisted in MAG surfacing instead
of remelting a base metal by
means of the TIG method. The
modified test proved useful for
the assumed test objective, i.e.
to develop input data for welding process modelling in respect of developing interfaces.
The tests also enabled the development of a methodology
for liquid metal temperature
measurements during MAG
surfacing as well as made it
Fig. 8. Temperature of liquid metal pool and methodology of determining possible to record temperature
Solidification Cracking Temperature Range (SCTR) for overlay welds
changes during the liquid metmade of 12mm-thick S355JR steel plates
al pool solidification and weld
cooling.
The tests revealed that heat flow impeding
The tests revealed that the Transvarestraint test can be used to determine the So- conditions cause an overlay weld to remain a
lidification Cracking Temperature Range for mixture of liquid and solid phases far longer
MAG-welded overlay welds made with a wire than a material subjected to remelting only,
10
No. 3/2013
which favours hot cracking. The Transvarestraint test carried out for the 6mm-thick test
pieces revealed a significantly greater number of cracks than that involving the use of
the 12mm-thick test piece, applying the same
surfacing parameters and bending conditions.
The determined value of the Solidification
Cracking Temperature Range (SCTR) for the
MAG-surfaced 12mm-thick plate is significantly higher than the value determined in
the Transvarestraint test conducted with the
TIG method and without the addition of a
filler metal. The use of the MAG method and
making an overlay weld during a Transvarestraint tests increases the probability of hot
cracking due to the greater concentration of
stresses.
References
1. Savage, W., and Lundin, C. (1965). The
Varestraint Test. Welding Journal, (10).
2. Lippold, J. (2005). Recent Developments
in Weldability Testing for Advanced Materials. Joining of Advanced and Specialty Materials, (7).
3. Dupont, J., and Lippold, J., et al. (2009).
Welding metallurgy and weldability of
nickel-base alloys. Hoboken, N.J.: Wiley.
4. Shankar, V. (2003). Solidification cracking in austentic stainless steel welds.
No. 3/2013
Academy Proceedings in Engineering
Sciences. Indian Academy of Sciences.
5. Finton T., and Lippold J. (2004). Standardization of the Transvarestraint test.
Edison Welding Institute Summary Report Nr 04-05/2004
6. Kou, S. (2003). Welding metallurgy. Hoboken, N.J.: Wiley-Interscience.
7. Lippold, J., and Lin, W. (1996). “Weldability of commercial AlCu-Li alloys”,
paper presented at Aluminum Alloys –
Their Physical and Mechanical Properties ICAA5 Conference, Transtec Publications.
8. Nakata, K., and Matsuda, F., (1995).
Evaluation of ductility characteristics
and cracking susceptibility of Al alloys
during welding. Trans of JWRI, (1).
9. Scotti, A. (2004). Performance assessment of the (Trans) Varestraint test for
determining solidification cracking susceptibility when using welding processes
with filler metal. Measurement Science
and Technology, (11).
10.Keehan, E., et al, (2002). “Microstructural and mechanical effects of nickel and
manganese on high strength steel weld
metals”, paper presented at Trends in
Welding Research International Conference, Pine Mountain, Georgia, USA.
11
Wojciech Jamrozik, Marek Fidali
Application of image fusion to identification of welding
imperfections
Abstract: It has been presented the issues concerning the application of images fu-
sion registered in visible and infrared radiation for diagnostics of MIG/MAG processes. The sequences of visible and infrared images visualizing the arc burning during
performing the weld were examined. Recorded during experiments image sequences
were subjected to fusion and next to analysis in order to obtain diagnostic signals. On
the basis of selected features of diagnostic signals it was carried out the classification
of the welding process state and its results were compared with those obtained for
diagnostic signal features determined independently for visible and infrared images.
The results show that the application of images fusion enables to identify effectively
various welding imperfections forming in welding processes.
Keywords: MAG welding, welding imperfections, image fusion, identification of
welding process state classification
Introduction
Obtaining diagnostic signals is an important
stage of diagnosing, the primary purpose of
which is to recognise the state of an object or a
process. Diagnostic signals are courses of certain quantities in time, on the basis of which it
is possible to distinguish states. Diagnostic signals can simply be courses registered by means
of various sensors or they can be obtained by
processing and analysing such courses as the
courses of values of features in time. Real diagnostic tasks, carried out for objects or processes characterised by high complexity require
recording many process parameters. However,
it should be noted using all such parameters
to generate diagnostic signals, and further, to
recognise the state of an object or a process is
pointless as some signals carry no information
about state changes and may only bring information noise to a recognition process.
In welding process diagnostics, process
parameters are mostly registered in the form
of the time courses of the values of physical
quantities affecting the quality of a joint, e.g.
current, voltage, gas flow, sound flow etc. Recent years have seen a significant growth in the
importance of visual techniques in diagnosing, as well as those of welding processes [7],
which has created the possibility of recording
signals in the form of image sequences. Such
images are mainly recorded in the range of
visible light (TV) as well as in the middle infrared band (IR, with a wavelength from approximately 0.9 to 14 μm). The use of a thermographic camera, operating in the infrared
range enables a non-contact and multipoint
temperature measurement. Thanks to this solution it is possible to obtain information about
temperature distribution, mainly in the area of
a burning arc, weld pool and a cooling weld.
dr inż. Wojciech Jamrozik (Ph.D. Eng.), dr inż. Marek Fidali (Ph.D. Eng.) - Politechnika Śląska,
Wydział Mechaniczny Technologiczny, Instytut Podstaw Konstrukcji Maszyn/Silesian University
of Technology, Faculty of Mechanical Engineering, Institute of Fundamentals of Machinery Design
12
No. 3/2013
At the same time a given object can be observed by means of IR and TV cameras. In
such cases the redundant representation of
objects in an image is obtained. In addition
to information relevant from a welding process recognition point of view, contained in
IR and TV images, there is also some redundant information. The elimination of redundant information with simultaneous maintaining of all information important from a
welding imperfection recognition point of
view can be obtained by an image fusion,
i.e. a multi-stage technique of merging many
input images, presenting the same scene as
one output image [4]. Images obtained in this
way can constitute the basis for determining
diagnostic signals enabling efficient diagnosing of welding processes [3].
Image fusion
Image fusion is a technique which consists in forming a new synthetic image out of
at least two input images recorded by devices of different parameters and different locations. It is important that input signals should
represent a common scene (or its fragment).
The process of image fusion is a multi-stage
process and in most cases consists of three
basic stages: image matching, image aggregation and post-processing, consisting mainly in spatial filtration sharpening a resultant
image. Sometimes the third stage is omitted
and the whole fusion includes matching and
aggregation. There are three basic groups of
methods enabling image fusion, i.e. methods
operating on the level of pixels, methods operating on the level of features and methods
operating on a symbolic level also referred to
as a decision level.
In practice, the most commonly used fusion methods are those operating on the level of pixels. These methods assume the existence of a relationship between individual
pixels of all input images. Such an assumption requires performing a geometric matchNo. 3/2013
ing of input images. The results of an image
fusion largely depend on the quality of the
conducted matching.
Geometric matching, being the first stage
of image fusion, can be carried out in several
ways [5]. The most common is image matching, i.e. matching of areas characterised by
similar brightness or colour or matching
based on the characteristic features of an image, e.g. the edges or points of objects visible
in an image. Matching images registered in
visible light and in infrared is a complex task
as image contents differ, usually in pixel values and object shapes. An additional problem
can be the lack of an unambiguous relationship between object colours and greyness
shades in visible light and object colours
and greyness shades in infrared. Reference
publications contain descriptions of various
matching algorithms, yet comparative tests
[2], [5] of selected algorithms justify the application of hybrid algorithms, making use of
both the brightness distribution of areas in an
image and of edge-related information. One
of such methods described in the publication
[5] can be used with indexed images or images recorded in the scale of greyness shades.
This method uses edge orientation maps. On
the basis of these maps as well as on the basis of pixel brightness values it is possible to
create a 3D bar chart. An objective function
determines the uncertainty of matching objects located in input images corresponding
to edges as well as the shade of individual areas. Such a solution enables the obtainment
of a low level of matching errors.
The next stage of the fusion of images
is their aggregation aimed at joining previously matched input images into one output
image. Similarly, as in the case if matching
algorithms, there is also a significant number
of well described algorithms for the aggregation of images. The algorithms on the level
of pixels can be divided into four groups [7]:
algorithms operating on a model of a colour
13
space, statistical/numerical algorithms, algo- • wavelet transformation, which is similar
to pyramid decomposition methods, where
rithms making use of a large-scale decompothe main difference is the amount of inforsition as well as spectral and radiometric almation contained in the hierarchical strucgorithms. For the purpose of merging vision
ture of images. Pyramid transformations
and thermographic images of a welding arc,
lead to a redundant set of transformation
the authors considered mainly algorithms
factors, whereas a wavelet transform leads
using a large-scale transform, the functionto a non-redundant image representation.
ing of which is based on a general scheme of
successive operations presented in Figure 1. Within the confines of previously conducted
tests [3], the authors examined
Decomposed
and assessed various aggregaimage DA
Image A
tion algorithms for applicaAveraging
tion in the fusion of welding
Resultant
or selection
image DC arc vision and thermographic
images.
For the purpose of the tests
Decomposed
described in the article an agresultant
gregation algorithm based on
image DC
a shift invariant Haar wavelet
Image B
Decomposed
(SIH) [6] was used. Within
image DB
Fig. 1. Scheme of image aggregation
this method, unlike in a disusing large-scale transformation (on the basis of [1])
crete wavelet transform, sucAs is shown in Figure 1, aggregation us- cessive image copies are not subjected to
ing large-scale transformation consists of resolution reduction, which leads to obtaining a highly redundant image representation.
three stages [1]:
• decomposition, when each of the input This algorithm for joining information out of
images is transformed into a hierarchical decomposed input images uses Burt’s adapstructure, whose successive stages are ap- tive rule [1]. As a result it is possible to join
propriately transformed copies of an input images by averaging or selecting pixels depending on the changes of pixel brightness
image,
• joining of images, during which, by means variance in some vicinity of a given pixel.
of a proper fusion principle using the se- The use of this method, to a significant exlection or averaging of pixels, a resultant tent, enables maintaining information about
decomposed image representation is de- the shape, brightness distribution and temperature in the welding arc area and the weld
termined,
• synthesis, the result of which is a new syn- pool.
thetic image, created by using a transform
reverse to the one used at the decomposi- Welding process observation
tion stage.
The observation of a MAG welding proThe hierarchical structure is obtained by using: cess was carried out using an observation
• pyramid transformation, used for deter- head composed of an uncooled thermographmining the set of images, in which each ic camera (Infratec VarioCam Head) with a
of the images in a pyramid is two times bolometric detector of 640×480 px resolution
smaller than an image on a lower level of and a lens with a focal length f=50 mm as
the pyramid,
well as a CCD camera operating in the visible
14
No. 3/2013
light range (ImagingSource DMK21AF04),
with 640×480px resolution, equipped with a
lens with a focal length f=25 mm. The cameras were provided with filters protecting optical systems against spatters and limiting the
power of electromagnetic radiation emitted
during welding. Images were recorded synchronically at a frequency of 50 images/s.
The MAG welding tests were
carried out at the Department of
Welding of the Silesian Universi- Welding
ty of Technology, using a ro- current
[A]
botised station for rectilinear
MAG welding The station was
240
equipped with a welding tractor
AS14a-1200 made by the OZAS
company and provided with a set of supports
for fixing a welding torch as well as with a
semi-automatic welding machine TotalArc
5000 produced by Castolin. The testing station with the observation head is presented in
Figure 2.
Fig. 2. Mechanised station for MAG welding
During experiments butt joints of steel
S235JR (EN 10025-2) sheets with the dimensions of 300×150×5 mm were made.
Apart from situations when special cases
of imperfections were simulated, the edges of sheets were V-bevelled at an angle of
α=60°; the distance between the sheets was
b=1.0 mm.
No. 3/2013
The butt joints of sheets were made on
a copper plate equipped with mechanical
clamping of sheets being joined. The welding process involved the use of a solid electrode wire (Castolin CastoMag 45255) with
a diameter of 1.2 mm and a gas mixture M21
(82%Ar+18%CO2). Welding parameters are
presented in Table 1.
Table 1. MAG welding parameters
Wire Shielding
Arc
Welding
Exposed
feeding
gas
voltage
rate
length of
rate
flow rate
[V]
[cm/min]
wire [mm]
[m/min] [l/min]
25
32
7,4
15
15
While making welded joints various welding parameters were deliberately simulated
by destabilising arc using the following:
• changes in the voltage and/or current of an
electric arc,
• changes in a shielding gas flow rate,
• improper preparation of the surfaces of
sheets to be joined.
Welding imperfections were classified into
the following 8 condition classes:
• z0 – model process, no imperfections,
• z1 – momentary disappearance of a gas
shielding,
• z2 – atmospheric corrosion of sheets to be
joined,
• z3 – weld groove imperfections, simulated
by openings located at a specified distance
from one another, in the axis of the weld
groove,
• z4 – improper welding current different by
±20% against model parameters,
• z5 – improper distance between sheets being joined, amounting to 1.5 mm, 2.0 mm,
2.5 mm, 3.0 mm and 3.5 mm,
• z6 – improper arc voltage, different by
±15% against model parameters,
• z7 – changing geometry of the weld groove
due to various values of a bevelling angle:
0°, 15°, 30°, 45°, 60°.
15
Fig. 3. Exemplary model of welded joint:
a) view from the weld face b) view from the weld root
Fig. 4. Exemplary welded joint, with an imperfection caused by the momentary
disappearance of the gas shielding:
a) view from the weld face b) view from the weld root
Fig. 5: Examples of images presenting a welding arc recorded:
a) in visible radiation range b) in infrared
Fig. 6. Exemplary images presenting a welding arc obtained by
a) averaging b) aggregation of vision and thermographic images
matched by means of the SIH method
16
Figures 3 and 4 present
an example of a model
joint and a joint formed
during the disappearance of a gas shielding.
The conducted observations enabled obtaining sequences of
thermographic images
and images recorded in
the visible light range.
The examples of images
showing a welding arc
in visible light and in
infrared are presented in
Figure 5.
The recorded sequences of vision and
thermographic images
underwent fusion. As
a result, sequences of
new, i.e. post-fusion images, were obtained.
Figure 6 presents examples of post-fusion
images, where the image in Figure 6a was obtained using the aforementioned aggregation
algorithm based on a
shift invariant discrete
wavelet transform modification. In turn, for
comparative purposes
Figure 6b presents the
effect of aggregation
consisting in averaging
the pixel values of input
images.
Both images contain
information from the vision and thermographic
images, yet as can be
easily seen the averaged
image is out of focus. In
No. 3/2013
addition, the contrast of the area representing
temperature distribution is reduced. The image obtained through SIH-based aggregation
is characterised by a much higher contrast
and sharpness of contours. The effect of the
thermographic image on the resultant image
is significant, which facilitates the description of the image both by means of measures
taking into account welding arc luminous intensity and those based on temperature distribution.
value image pixel. In the tests, the following parameters were taken into consideration
(Fig.7): left external width and right external
width (LEW, REW); left and right internal
width (LIW, RIW); central width (CW); left
and right half width (LHW, RHW) as well as
total width (TW). The values of determined
features were applied to determine diagnostic signals used for recognising the state of a
welding process.
Analysis of images after fusion
The new images created through fusion
were analysed in order to determine the set
of features used for recognising welding
process states. Three types of features were
determined for the specified area of interest
containing the welding arc area. The first
type includes first- and second-grade statistical features based on the vector of the bar
chart and Co-Occurrence Matrix (COM).
The features in question include the average, variance, coarseness, kurtosis, contrast,
correlation, entropy etc. The second type
includes topological features describing the
geometric properties of a previously binarised welding arc area. During the tests the
following topological features were taken
into consideration: shape factor, area, Feret
coefficient, circumference, Malinowska coefficient, longer diagonal and shorter diagonal. The third type was composed of features
resulting from the analysis of the horizontal
profile of the welding arc area. The horizontal profile passing through the weld pool area
has a pseudosymmetric character resulting
from welding process properties. Disorders
in the profile symmetry may indicate the instability of a welding process. This assumption made it possible to indicate the profile
features which can be the widths between the
abscissas of characteristic points, where characteristic points are the extrema of the horizontal profile passing through the maximum
No. 3/2013
Fig. 7. Manner of determining features basing on the
linear profile of post-fusion image pixel values
The diagnostic signals determined as a result of post-fusion image analysis were estimated using the following point statistical
features known in technical diagnostics: average value, root-mean-square value, peakto-peak value or kurtosis. The obtained numerical values of signal features were the
basis for the operation of recognising welding process states by way of classification.
Figure 8 presents the scheme of obtaining
models describing various welding process
states.
Recognition of welding process state
The recognition of a welding process state
involved the development of models describing each of the welding process states simulated during the experiment. Each model was
a one-dimensional vector consisting of the
values of the features determined for a given
welding process. The models were used to
form both a training set and a testing set, in
accordance with principles of building and
teaching classifiers.
17
image sequence
1
2
bf2
bf1
3
bfa
sf2
sf1
sfb
tf2
tf1
tfc
Diagnostic signals – courses of image feature values
band
cp1
1
cp2
bf1
...
statistical
cp16
cp1
cp2
sf1
...
topological
cp16
cp1
cp2
2
tf1
...
cp16
720
Classification
Fig 8. Scheme of obtaining models describing welding process
examples in the teaching and testing sets.
The process of state recognition was considered for two cases:
• classifier made an assignment to one of
the states from z0 to z7, which means that
both the examples, for which welding proceeded properly and those with simulated
disturbances leading to welding imperfections were considered;
• classifier made an assignment to one of
the states from z1 to z7, which means that
only the examples with simulated disturbances leading to welding imperfections
were considered; in other words a distinction between undesired states was made.
Prior to a state recognition, on the basis of
the set of models determined on the basis of
post-fusion images (FU), classification was
also performed for the models determined
separately for the thermographic images (IR)
(1)
and images recorded in the range of visible
=
radiation (TV). The models for the therwhere Ng is the number of properly recog- mographic images and vision images were
nised examples and N is the number of all built on the same principle as in the case of
For the purpose of state recognition it was
necessary to use a minimum distance classifier using the k-nearest neighbour algorithm, where k amounted to 7. The value of
parameter k was selected experimentally on
the basis of preliminary tests. The estimation of a classifier error was carried out using
n-fold cross-validation (leave-one-out). In
this method the process of training and testing is repeated N-times, where N is a number
of teaching examples. In each iteration the
teaching set has N-1 elements, whereas the
testing set has one element. In this method
each element of the set of teaching examples
is a testing example. The method leave-oneout was used because of a not too numerous
set of teaching examples. The relative number of correct assignments was used as a
measure of state recognition efficiency:
18
No. 3/2013
Figure 9a presents the comparison of the
post-fusion images. The classification on
the basis of separately determined features maximum, minimum and medium (marked
of the thermographic images and vision im- with a square) values of the classification efages made it possible to assess the usability ficiency for the diagnostic signals determined
of image fusion. Table 2 presents the values for individual sequences of images. The seof classification efficiency for considered quences of post-fusion images made it possiwelding process states obtained for select- ble to obtain the highest values of classificaed features and selected diagnostic signals tion efficiency eerr = 0.59, where the average
determined on the basis of the sequence of value of efficiency is also higher than in the
thermographic images (IR), vision images case of the diagnostic signals determined
(TV) and post-fusion images (FU)TV. The from the sequences of TV and IR images.
cases selected were those for which the aver- Taking into consideration only the states repage value of classification efficiency was the resenting the occurrence of welding imperhighest. The highest classification efficiency fections (Fig. 9b) one can notice that the best
was achieved for the diagnostic signal ob- results were obtained for the signals generattained by determining the average value of ed from the sequences of the post-fusion imthe signal containing the values of the total ages, eerr = 0.61. Very poor results were obwidth of a characteristic band (TW), calcu- tained for the use of the signals obtained from
lated for the sequence of images received as the sequences of the thermographic images
a result of the fusion (FU).
eerr = 0.44, with very similar average values.
While analysTable 2. Classification efficiency in recognising individually identified welding process
ing the achieved
states for the classifier of the maximum total efficiency
results one can
State
Mean Mean
also notice that
Signal
for the images
z0
z1
z2
z3
z4
z5
z6
z7 z0-z7 z1-z7
obtained as a re0,91 0,17 0,86 0,29 0,71 0,17 0,43 0,00 0,37 0,44
sult of fusion, the IR Variance
classification reTV RIW
0,91 0,67 0,86 0,29 0,57 0,33 0,71 0,14 0,51 0,56
sult for the state RMS
representing the
properly
con- FU TW
0,45 0,67 1,00 0,57 0,86 0,17 0,86 0,00 0,59 0,57
absolute mean
ducted welding
process is worse
than the classification results obtained on the
Using appropriate methods for the selecbasis of the sequence of IR and TV images tion of features it is possible to select feaonly. In the case of the states representing tures enabling (in many cases) error-free
imperfections, the assessment of post-fusion recognition of some states. Such an approach
images significantly improved classifica- makes it possible to increase the efficiency
tion efficiency. The states of special interest of classification and thereby identify states
are z5 and z7, for which very low results of which are most difficult to recognise. The
classifier efficiency obtained for the thermo- tests resulted in increasing the efficiency of
graphic images are directly reflected in the classification using groups of seven classiresults obtained for the classifiers based on fiers, each of which distinguished whether
the signals generated from the sequence of a given case belonged to one of the identipost-fusion images.
fied classes or not. Each classifier could opNo. 3/2013
19
tent of such signals. In
the case of a multi-mode
vision observation the reduction of the number of
sequences to be analysed,
and thus the limitation of
the number of determined
diagnostic signals can be
achieved by merging images using image fusion
methods.
Fig. 9. Classification efficiency values:
The conducted tests rea) during recognising all identified states,b) during recognising states
vealed that the sequences
with welding imperfections (set of signals; set of signals)
of images obtained as a
erate on a different previously determined result of the fusion of IR (thermographic)
feature. Table 3 presents the maximum clas- and TV (recorded in visible radiation) imsification values obtained for states under ages can be the basis for obtaining better
consideration. The error-free recognition of and more important diagnostic signals than
states z0, z1, z2 and z4 was obtained for the those which can be obtained by analysing
models created from the features of all con- the sequences of input images (IR and TV)
sidered diagnostic signals. States z6 and z7 for the fusion process. The manner of imcreated as a result of welding arc instability age analysis significantly affects the qualand improper weld groove geometry were ity of the obtained diagnostic signals. All
characterised by the lowest recognisability. imperfections such as, for instance, image
It was not possible to achieve error-free rec- mismatching related to possible momentary
and sudden displacements of cameras durognition for any of these states.
ing image acquisition signifTable 3. Classification efficiency in recognising individually identified
icantly deteriorate the final
welding process states (maximum values)
state recognition quality. It
State
should also be noted that the
Signal
z0
z1
z2
z3
z4
z5
z6
z7
selection of a feature used to
IR
1,00 1,00 1,00 0,86 1,00 1,00 0,86 0,86
describe the whole sequence
TV
1,00 1,00 1,00 0,86 1,00 0,67 0,71 0,71
significantly affects the final
FU
1,00 1,00 1,00 1,00 1,00 0,83 0,86 0,86
diagnostic outcome. The obtained results indicate that
Summary
the use of image fusion enables the efficient
Recognising state changes is the prima- identification of various welding imperfecry diagnostic objective often obtained by tion generated during welding.
way of model recognition. For this reason
obtaining proper diagnostic signals being
Research work financed
the basis for the generation of models is
from funds for science
an important, yet complicated task. Makin the years 2009-2012
ing use of too many diagnostic signals is
not advantageous either, and therefore it is
advisable to maximise the information cona)
20
b)
No. 3/2013
References
1. Burt, P., and Kolczynski, P (1993).“En- 5. Kim, Y. S., Lee, J. H., and Ra, J. B. (2008).
Multi-sensor image registration based on
hanced image capture through fusion”,
th
intensity and edge orientation information.
paper presented at Computer Vision 4
Pattern Recognition, 41, pp. 3356-3365.
International Conference, Proceedings,
6. Rockinger, O. (1997). “Image Sequence
pp. 173–182.
Fusion Using a Shift Invariant Wavelet
2. Jamrozik, W., and Fidali, M. (2011).
Transform”, paper presented at InternaMetody dopasowania termogramów i
tional Conference on Image Processing,
obrazów wizyjnych dla dynamicznie
Proceedings, vol.3, pp. 288-291.
zmieniającej się struktury obserwowanej
sceny. Measurement, Automation and 7. Nandhitha, N. M., Manoharan, N., Rani,
B.S., Venkataraman, B., Sundaram, P. K.,
Monitoring, 57 (10), pp.1134-1137.
and Ra, J.B. (2006). “Automatic Detec3. Jamrozik, W., and Fidali, M. (2012).
tion and Quantification of Incomplete
“Evaluation of the suitability of IR and
Penetration in TIG Welding Through
TV image aggregation algorithms for
Segmentation and Morphological Image
the purposes of welding process assessProcessing of Thermographs”, paper prement”, paper presented at 11th Quantitasented at National Seminar on Non-Detive InfraRed Thermography Conference,
structive Evaluation, Hyderabad, 7-9 DeNaples, Italy, 11-14 June.
cember, Proceedings, pp. 17-22.
4. Jamrozik, W., and Fidali M. (2011). Estimation of image fusion methods for purposes of vision monitoring of industrial
process. Measurement, Automation and
Monitoring, 57 (09), pp. 993-996.
No. 3/2013
21
Antoni Sawicki
Damping factor function in AC electric arc models.
Part 2: Damping factor function in universal electric arc
models with moderate cooling
Abstract: Nonlinear function of the heat processes damping factor in thermal col-
umn plasma have been introduced to modified mathematical models of the AC electric arc. This way, universal arc models with moderate cooling and constant geometric
sizes, using static characteristics, have been created. Similar method has been used
to hybrid arc models with variable column length connecting Berger and Kulakov
models. It has been proposed the introduction of nonlinear damping factor function
to modified Voronin model using static characteristics. Thereby the extended range
of current for application of these models has been achieved.
Keywords: electric arc, damping factor, time constant, Mayr model, Cassie model,
Berger model, Kulakov model, Voronin model, hybrid model
Introduction
Most of today’s arc and plasma-arc welding devices are designed as universal tools
used for joining and cutting elements of
various geometrical dimensions (thickness) and shapes made of various materials.
Such processes require the use of diverse
torches, electrodes, shielding gases, current pulses of various amplitudes, shape,
packing degree, polarity and frequency. In
variable operating conditions, the systems
of automatic control ensure stable electric
arc burning, high quality of technological
processes, high production efficiency, low
material consumption as well as minimum
noxiousness to personnel, the environment
etc. However, such versatility results in the
occurrence of a wide range of state variables, the appearance of various non-linear
static and dynamic characteristics of supply
systems and electric discharge.
Increasing advancements in computer-aided methods for designing electrotechnological and electrothermal welding devices require more and more accurate arc
discharge models, precisely reproducing
the non-linearity and dynamics of thermal
and electric processes. However, the earlier
imperfections of the experimental determination of the dynamic characteristics of arc
discharges are accompanied by imperfections in their mathematical modelling [1].
Despite emerging compromises between
the required accuracy of the reproduction
of physical processes and the ease of measurements, the simplicity of interpretation,
the low complexity and short computational
time, the most popular arc models still fail to
ensure appropriate precision. This is largely due to the wide range of exciting current
changes and external effects causing changes to column geometrical dimensions.
dr hab. inż. Antoni Sawicki (PH.D. Eng), professor at Częstochowa University of Technology Wydział Elektryczny /Faculty of Electrical Engineering/
22
No. 3/2013
This study presents universal electric arc
models with moderate cooling of the extended current range of approximations of static and dynamic characteristics. The objective of this work required the application of
modified mathematical models using static
characteristics and hybrid models, matching the Mayr and Cassie models. The use of
the non-linear damping factor function corresponding to experimentation results was
suggested [1] in the models.
In modelling electric arc dynamic characteristics a damping function (or its specific value – time constant) along with other
characteristics and dynamic parameters of a
specific model constitute the whole complex
of closely related quantities approximating
the course of real physical processes with a
pre-set accuracy. Due to random disorders,
obtaining repeatable arc experimentation results with a pre-defined accuracy is very difficult. Even more problematic is matching
several characteristics determined in various
separately conducted experiments (e.g. static
current-voltage characteristics, damping factor function) in a single model. This process
cannot be carried out by means of simple analytical or even numerical methods but requires the use of complex procedures for the
identification of electric arc mathematical
model characteristics.
Damping function in electric arc
models with moderate cooling using
static characteristics
arc voltage, which in the case of arc burning coaxially with electrodes corresponds
to electric field intensity Estat. The shapes of
typical static characteristics of electric field
voltage and intensity are presented in Figure
1. In the case of strong currents it is possible
to assume that static voltage Ustat does not
depend on current.
Ustat(l) = as + bsl(2)
According to formula (1), the generalised
power static characteristic is described by
the formula
Pstat=Ustat(l, I)·I=
= (as +bsl)·I+(cs+ dsl)·Ii-n
(3)
which in accordance with formula (2) in the
case of string currents adopts a simpler form:
Pstat=Ustat(l)·I=(as+bsl)·I(4)
If and arc length does not change (l = const),
the static voltage-current characteristics is
the following:
Ustat(I)=A+BI-n(5)
where A=as+bsl = const; B = cs+dsl = const.
Similarly, the power of an arc is a non-linear
current function:
Pstat(I)=Ustat·I=f(I)=AI+BI1-n(6)
and in the case of strong currents the dependence becomes linear
Pstat=Ustat·I=I·const(7)
A modified arc model by Mayr [2, 3] can
One of the most commonly known general be presented in a general conductance form
static characteristics of an electric arc with 1 dg
1  Pkol (t ) 
(8)
moderate cooling is the dependence provided g dt = θ (i(t ))  P (t ) − 1
 dys

M
by Nottingham:
where g – column conductance; i – alternatcs + d s l
U stat (l , I ) = as + bs l +
(1) ing current intensity; Pkol – electric power
In
supplied to plasma column; Pdys – power of
where l – column length; I – direct current energy dissipated from column; θM(i) –
intensity; as – sum of near-cathode and near- damping function corresponding to the time
-anode voltage drop; bs – gradient of deflected of relaxation of heat processes. Electric
No. 3/2013
23
where A’ = bsl = const. The power of losses
in disequilibrium plasma PE = as |i| of very
thin near-electrode areas is taken into consideration separately. More often applied are
static voltage-current characteristics (1), (2)
or (5), and then
i2
Pdys (t ) = U stat (i ) ⋅ i =
(12)
G stat (i )
0
After substituting formula (12) to formula
(10) it is possible to obtain the modified Mayr
equations in the conductance form
1 dg
1  Gstat (i ) 
=
− 1
g dt θ M (i )  g

(13)
Fig. 1. Static characteristics of arc:
Ustat(I) - voltage-current, Estat(I) – electric field intensity
The shapes of the typical static characteristics of power and column conductance are
power supplied to thermal plasma is ex- presented in Figure 2.
pressed by the formula below:
i2
Pkol (t ) = u kol i =
g
(9)
where ukol – voltage drop on arc column. In
the classical Mayr model it is assumed that
Pdys(t) = const. In the range of stronger currents this condition cannot be met any longer
and usually the Cassie model is used instead
[3]. As heat dissipation processes react slowly to external disturbances, one can roughly assume that a power loss is principally
determined by static characteristics [4],
Pdys(t) ≈ P’stat(i(t)), i.e.

1 dg
1  ukol i
=
−
1
 '
(i ) 
g dt θ M (i )  Pstat
(10)
Fig. 2. Arc static characteristics:
Pstat(I) - power-current, Gstat(I) - conductance-current
At this moment, when conductance does
not change in time, the static characteristics of
an arc in this model has the following form:
where P’stat – takes into consideration power losses only in column plasma, without U (i ) = Pstat (i )
(14)
stat
i
near-electrode areas. The function of loss
power can be approximated by means of dei2
i2
pendences (6) or (7) allowing a generalised Gstat (i ) =
=
(15)
(
)
(
)
P
i
U
i
⋅
i
stat
stat
dynamic arc model to be obtained:

u i
1 dg
1 
(11)
=
 ' kol 1−n − 1
Thus, based on this it is possible to write
g dt θ M (i )  A i + B i

model (13) in a conductance form
24
No. 3/2013

i
1 dg
1 
=
−
1

'
(i ) 
g dt θ M (i )  g ⋅ U stat
(16)
• Cassie model
g (t ) ≈ g C (t ) =
u kol iC
dg C
, if i > I0 (21)
−
θ
C
U C2
dt
which on the basis of formulas (5) and (10)
where PM – constant power of the Mayr
leads to the generalised dependence
model; θM – time constant of the Mayr model

(0<θC<<θM); UC – constant voltage of the
i
1 dg
1 
1
=
−


(17)
Cassie model; θC – time constant of the Casg dt θ M (i )  g ⋅ A' + B i −n

sie model (0<θC<<10-3 s).
In comparison with the classical Mayr modFigure 3 presents the shapes of the typical
el [2], models (11) and (17) can significant- static characteristics of voltage and power.
ly more accurately reproduce the static and Adopted approximations in the form of the
dynamic characteristics of an arc in a wid- constant values of Cassie model voltage and
er range of current changes. This results not Mayr model power are marked against their
only from the variation of a time constant background of the aforesaid static characterobtaining the form of the damping function istics. Both models also assume the constant
θM(i), but also from applying a more accu- values of a damping factor, which is presentrate function approximating static voltage- ed in Figure 4 against the background of typical characteristics θ(i).
-current or power-current characteristics.
(
)
Damping function in arc TWV
hybrid model matching Mayr
and Cassie models
In the TWV hybrid model of an arc [5] the
fractions of currents flowing through two
parallel non-linear conductances, corresponding to the Mayr and Cassie models,
depend on their resultant value. Thus, it is
possible to write
 i2 
i (t ) = u kol g = u kol ⋅ g M exp − 2  +
 I0 
(18)

 i2 
+ u kol ⋅ g C ⋅ 1 − exp − 2  
 I0 

where I0 – limiting current between the Mayr
and Cassie models. Hence we obtain

 i2 
 i2 



g (t ) = g M (t ) ⋅ exp − 2  + g C (t ) ⋅ 1 − exp − 2  
 I0 
 I0 

(19)
Fig. 3. Static characteristics and dynamic parameters
of arc: Ustat(I) - voltage-current, Pstat(I) – power-current,
PM(I)=const – Mayr model power,
UC(I) = const – Cassie model voltage,
I0 – limiting current between the Mayr and Cassie models
On the basis of formulas (18)-(21) a hybrid model is created [5]
g = G min + [1 − ε (i )]
u kol i
i2
dg
(
)
i
+
− θ MC (i )
ε
(22)
2
PM
dt
UC
The model selections conditions are the
following:
where θMC(i) – damping factor function of
• Mayr model
2
i
dg M
, if i < I0 (20) TWV model.
g (t ) ≈ g M (t ) = M − θ M
In practical considerations it is usually asPM
dt
No. 3/2013
25
corresponds to an increase in thermal power
dQa dQa dl
dl
=
= ql
dt
dl dt
dt
Fig. 4. Characteristics and dynamic parameters of arc:
θ(i) – characteristics of damping factor,
θM –Mayr model time constant,
θC –Cassie model time constant
(25)
where ql – arc energy linear density. Therefore, in simplifying conditions [8] thermal
power required for the generation of additional plasma volume is roughly proportional
to the rate of a length increase. This is accompanied by some relaxation times resulting from gas thermal inertia and additional
cooling of a column. As the arc voltage grows
along with an increase in a column length, in
publication [8] the following approach to determining Cassie model voltage component
was suggested (21):
sumed that Gmin = 0 [5, 6]. Then, formula u2 (l) = al(26)
c
(22) can be written in a simplified form
where the parameter a is almost constant in
u kol i
u kol i 
1 dg
1 
(
)
(
)
[
]
=
1
−
+
−
1
i
ε
ε
i

 (23) the wide range of current i changes. In turn,
g dt θ MC (i ) 
PM
gU C2

additional power pv(dl/dt) dissipated from
where the designation of a tapering function the column is determined by u2c g; due to diswas introduced
sipativity, the lack of even partial conservativity and the return of energy to the circuit
 i2 
ε (i ) = exp − 2 
(24) leads to the following dependence [8]:
 I0 
dl
 dl
The form of this function may vary and
>0
if
b
(27)
 dl   dt
depends on the type of an electrotechnologidt
pv   = 
cal device and its operating conditions [7].
dl
 dt 
Damping function in variable length
arc hybrid model matching Berger
and Kulakov models
0

if
dt
≤0
A modified Cassie equation with the
dl
variable value of voltage U C (t ) = U C  l , 
While considering the models of an elec dt 
tric arc with the variable length of a plasma enables obtaining the conductance form of
column it is possible to modify separate Cas- the Berger model [8]
sie and Mayr models retaining the constant
value of damping factors (time constants)


2
u kol
1 dg
1 
and current ranges preferred by them.
=
− 1
(28)
An increase in the geometrical dimensions g dt θ CB  u 2 (l ) + 1 p  dl  


v
 C
of the plasma column of a high-current arc is
g  dt  

accompanied by an increase in energy necessary for the generation of additional plasma where pv(dl/dt) - power needed for the genvolume. The assumed axial-cylindrical shape eration of additional plasma volume; θCB –
of a column and its tension by a length dl time constant of Cassie-Berger model.
26
No. 3/2013
Kulakov suggested the modification of a 1 dg = 1 [1 − ε (i )] u2kol + ε (i ) u kol − 1 (33)
l ⋅ E stat (i ) 
uC (l )
low-current arc model (16) taking into con- g dt θ BK (i ) 
sideration column length changes. The After taking into consideration approximafirst-degree Mayr-Kulakov model in the con- tions (26) and (30) the equation is as follows:
ductance form is the following [9]:
2
u kol
1 dg
1 
=
+
[1 − ε (i )]
 1 dl
i
1 dg
1 
(
)
g
dt
i
al
θ

BK
=
− 1 −

(29)

(34)
g dt θ MK  g ⋅ l ⋅ Estat (i )  l dt

u kol

+ ε (i )
− 1
where Estat(i) – non-linear static characteris−n
l ⋅ bs + d s i

tics of electric field intensity; θMK – time constant of the Mayr-Kulakov model. One of the The matching of the Berger and Kulakov
imperfections of approximation by means of models as well as using the non-linear functhis model is overlooking the impact of plas- tion of a damping factor makes it possible to
ma physical properties on conductance dy- use the hybrid model for simulating processes in electrotechnological devices operating
namics during column length changes.
Using the dependence approximating the with a wide range of current and an arc disstatic voltage-current characteristics (1) it is charge column length.
possible to determine the static characterisDamping function in Voronin model
tics of electric field intensity
of arc with variable geometrical
∂U stat (l , I )
ds
Estat (I ) =
= bs + n
(30) dimensions
∂l
I
The Voronin model makes it possible to
Then, the Kulakov model has the following form: take into consideration an external influence
exerted on the length and diameter of a cy

i
1 dg
1
1 dl
=
− 1 −

(31) lindrical column. In order to create such a
−n
g dt θ MK  g ⋅ l ⋅ bs + d s i
l
dt
model it is necessary to make a number of

simplifying assumptions [9, 10]. The model
The arc column hybrid model, taking into basis is a simplified equation of the thermal
consideration the changes of an arc length, balance of a column. It is assumed that the
matches models (28) and (29) in the manner dissipated power is proportional to the side
(23) by means of an appropriate tapering area of an arc:
function ε(i). Therefore the model has the P (l , S ) = p l 4πS
(35)
dysS
S
following form:
As a result, an arc model with variable geomet
2
u kol
1 dg
1 
rical dimensions S(t) and l(t) of the following
=
+
[1 − ε (i )]
1
dl
g dt θ BK (i ) 
 
general conductance form is obtained [10]:
uC2 (l ) + pv  

g  dt  (32)
2
(
(
)

u kol
1 dl
+ ε (i )
− 1 − ε (i )
g ⋅ l ⋅ E stat (i ) 
l dt
where θBK(i) – damping factor function of
the hybrid Berger-Kulakov model.
If a relatively low rate of arc length changes is assumed (dl/dt≈0), equation (32) can be
written as
No. 3/2013
)

1 dg
1  ukol i
=
− 1 +
g dt θ S (S )  PdysS (l , S ) 
−
g l 
1 dl 
1 + ln
+
l dt 
K g S 
+
g l 
1 dS 
1 − ln
S dt 
K g S 
(36)
27
θ (S (i )) ∝ S (i ) ∝ d (i ). In turn, in theoretical
deliberations [11] the relation θ (S (i )) ∝ S (i )
where a damping factor function is
θ S (S ) =
Q0 Sl
Q
= 0
PdysS (l , S ) pS
S
4π
(37) is assumed. However, experimental tests [5, 11]
−1
reveal almost a reverse tendency θ (i ) ∝ i ,
and Q0 – reference factor, J/m3; Kg – coeffi- which should be recognised as a real one,
cient of unitary conductance approximation, especially due to the fact that simplified cyS/m; l – length of arc column, m; pS – density lindrical arc models take into consideration
of power dissipated by the side surface of a only selected dominant heat processes, and
column, W/m2; S – area of arc cross section, entirely pass over gasodynamic and electrom2. All three parameters Q0, Kg, pS are deter- magnetic processes (e.g. effects of contracmined on the basis of experiments and are tion and gas pumping by a column).
Figure 5 presents a typical shape of charassumed to be constant quantities.
If the relative rate of arc length changes is acteristics d(i) and S(i). One can see that
low (dl/dt≈0), equation (36) can be written as there is no correlation of these quantities
with a damping factor function. To some ex
1 dg
1  ukol i
tent an increase in θ in areas where current
=
− 1 +
g dt θ S (S )  PdysS (l , S ) 
passes through zero can be explained by the
(38)
significant weakening of a contraction effect
g l 
1 dS 
1 − ln
+
and momentary plasma expansion.
S dt 
K S

g

Models (36) and (38), similarly as the Mayr
model, reproduce arc characteristics in the
low-current range relatively well. Therefore
they can be used to calculate processes in
devices with a relatively low temperature of
the area surrounding a discharge. Such conditions occur in open arc welding or during
melting of a charge at the initial stages of arc
furnace operation. While considering a free
or quasi-free arc, in which the area of column cross section is primarily determined by
the module of current value, using equation
(36) one obtains

ukol i
1 dg
1 
=
− 1 +
g dt θ S (i )  PdysS (l , S (i )) 
−
g l 
1 dl 
1 + ln
+
l dt 
K g S (i ) 
+
g l 
1 dS di 
1 − ln
S (i ) di dt 
K g S (i ) 
While considering the case of a free or
quasi-free welding arc, on the basis of experimentation and theoretical analysis [12]
on can adopt a dependence related to an arc
column diameter
d=k|i|m(40)
(39)
Adopted assumptions simplifying model (36)
lead to almost a linear dependence of a damping function (37) on an arc column diameter
28
Damping function in Voronin
modified arc model using static
characteristics
Fig. 5. Arc static and dynamic characteristics of:
θ(I) – damping factor, d(I) – column diameter,
S(I) – area of column cross section
No. 3/2013
As heat dissipation processes react slowly to rameter, i.e. the surface density of thermally
external disturbances, just as previously it is dissipated power:
−n
possible
to
roughly
assume
that
bs + d s i
Pdys(t) ≈ P’stat(i(t)), which on the basis of for- pS =
(45)
m −1
πk i
mula (35) leads to the following dependence:
PdysS (l , S (i )) = p S l 4πS (i ) = E stat (i ) ⋅ li
(41) Concluding remarks
Thus, the intensity of a field is expressed by
1. Introducing a damping function non-linthe following formula:
early dependent on current extends the possiEstat(i) = πpsk|i|m-1 sign(i) = (bs+ds|i|-n)·sign(i) bilities of applying dynamic arc mathematical
models using static characteristics.
(42)
2. Introducing a damping function nonHence, on the basis of formula (36) the fol- -linearly dependent on current strengthens
the versatility of arc hybrid models using
lowing dependence is obtained:
sub-models of a low and high-current arc.

1 dg
1 
i
3. The developed mathematical models

=
− 1 +
g dt θ S (i )  gEstat (i ) ⋅ l 
include a wide range of changes of current
and those of plasma column geometrical di1 dl 
g l 
(43) mensions. As a result, they can be used to
1 + ln
−
+
l dt 
K g S (i ) 
simulate processes in various arc electrotechnological and plasma devices (both weld1 dS (i ) di 
g l 
1 − ln
+
ing and electrothermal ones) with moderate
S (i ) di dt 
K g S (i ) 
cooling of a discharge area.
If compared with the Kulakov model (29)
this model significantly extends the possi- Research work financed from funds
bility of approximating characteristics as it for science in the years 2010-2013 as
takes into consideration not only the dynam- research project no. N N511 305038.
ics of length changes but also the dynamics
References
of column cross section changes.
Slow arc length changes (dl/dt ≈ 0) corre- 1. Sawicki, A. (2013). Damping Factor
spond to the following equation:
Function in AC Electrical Arc Models.
Part 1: Heat Process Relaxation Phenom
1 dg
1 
i

=
− 1 +
ena, their Approximations and Measureg dt θ S (i )  gEstat (i ) ⋅ l 
ment. Biuletyn Instytutu Spawalnictwa,
(44)
57 (2), pp. 37-41.
1 dS (i ) di 
g l 
1 − ln
+
2. Ciok, Z. (1987). Modele matematyczne
S (i ) di dt 
K g S (i ) 
łuku łączeniowego. Warsaw: Warsaw
University of Technology.
If in formula (40) a typical value m= ⅔
[12] is assumed, using formula (42) the 3. Лесков, Г.И. (1970). Электрическая
−1 / 3
сварочная дуга. Изд-во Машиностроdependence Estat (i ) ∝ i is obtained. This
ение, Москва.
dependence corresponds to applied approximations of electric field intensity in a 4. Sawicki, A. (2012). Modele dynamiczne
łuku elektrycznego wykorzystujące chalow-current arc with moderate cooling of a
rakterystyki statyczne. Śląskie Wiadomocolumn [4]. In addition, formula (42) can
ści Elektryczne, (6), pp. 13-17.
be used to determine the basic model paNo. 3/2013
29
5. Tseng, K.J., Wang, Y., and Vilathgamuwa, D.M. (1996). “Development of a Dynamic Model of Electric Arc for Power
Electrics Simulations”, paper presented
at IEEE Industrial Applications Conference, IAS’96, Proceedings, vol. 4,
pp. 2173-2180.
6. Sawicki, A. (2012). Modified Habedank
and TWV hybrid models of the arc with
variable length for simulating processes
in electrical devices. Biuletyn Instytutu
Spawalnictwa, 56 (1), pp. 45-49.
7. Sawicki, A. (2012). Funkcje wagowe w
modelach hybrydowych łuku elektrycznego. Śląskie Wiadomości Elektryczne,
(5), pp. 15-19.
8. Berger, S. (2006). “Mathematical approach to model rapidly elongated
30
free-burning arcs in air in electric power
circuits”, paper presented at ICEC 2006,
6-9 June, Sendai, Japan.
9. Sawicki, A. (2012). Modelowanie łuku
elektrycznego o zaburzanej długości kolumny plazmowej. Śląskie Wiadomości
Elektryczne, (1), pp. 9-17.
10.Воронин, А.А. (2009). Повышение эффективности контактно-дугогасительных систем сильноточных коммутационных аппаратов с удлиняющейся
дугой. Автореферат дис. к.т.н. Самара.
11.Kalasek, V. (1971). Measurements of time
constants on cascade D.C. arc in nitrogen. Technische Hogeschool Eindhoven,
TH-Report 71-E-18.
12.Коперсак, В.М. (2011). Теорiя процесiв
зварювания-1. КПI, Киiiв
No. 3/2013
Ryszard Krawczyk, Piotr Wojtas, Karolina Poch
Comparative assessment of selected welding imperfections
in VT, PT and MT methods
Abstract: The issues connected with assessment of sensitivity in selected NDT
methods are considered. In particular, the attention has been paid to the tests which
make it possible to detect any imperfection of surface, i.e. visual testing, liquid-penetrant and magnetic-powder method. The aim of the studies was to reveal the analogy
and differences between results obtained from series of tests (VT, PT and MT).
Keywords: non-destructive testing, magnetic particle inspection (MT), penetration
testing (PT)
Introduction
Importance of surface NDT
The non-destructive testing of welding materials and welded joints is classified into two categories. The first category,
comprising tests enabling the detection of
surface imperfections is comprised of three
methods, i.e. visual testing (VT), penetration testing (PT) and magnetic particle testing (MT). The second category, comprising
non-destructive tests enabling the detection
of internal imperfections, often referred to
as volumetric imperfections, producing no
superficial symptoms or indications, includes two methods, i.e. ultrasonic testing
(UT) and radiographic testing (RT). The
testing and assessment of surface imperfections of welding materials and welded joints
are a very important element of technological and operating processes. Conducting
tests makes it possible to verify that materials or structural elements being tested do
not contain surface imperfections directly
compromising the safety of operation and
the possibility of further use of an element
being tested.
The visual test of welded joints is a basic
test which must be carried out irrespective
of what other tests will follow. The visual
test consists of a thorough inspection of te
element being tested in order to assess the
state of its surface. The conditions which
must be satisfied during VT are specified in
detail in standard PN-EN ISO 17637:2011
”Non-Destructive Testing of Welds – Visual
Testing of Fusion-Welded Joints”.
Penetration testing is based on the phenomenon of capillarity, i.e. water penetrating narrow spaces and rising within them despite inversely directed gravitational force.
This process makes it possible to detect even
the smallest discontinuities present inside
a material and coming up to its surface. In
turn, the essence of magnetic particle testing
is focused on two aspects, the first of which
involves introducing a magnetic field to the
object being tested, whereas the other refers
to the manner of interpreting revealed indications. Both elements are interdependent
as in order to assess the effectiveness of the
dr inż. Ryszard Krawczyk (Ph.D. Eng), mgr inż. Piotr Wojtas (M.Sc. Eng), mgr inż. Karolina
Poch (M.Sc. Eng) - Częstochowa University of Technology, Zakład Spawalnictwa /Welding
Technology Department /
No. 3/2013
31
magnetic field introduction it is necessary to
apply one of the methods used to identify
any detected imperfections. Each of these
testing methods (VT, PT and MT) is based
on different phenomena and is characterised
by different testing sensitivity. It is assumed
that VT is characterised by the lowest, PT
by higher and MT by the highest testing sensitivity. Having taken this into account, an
attempt aimed at comparing detection sensitivity of the aforesaid testing methods, particularly PT and MT, was undertaken.
Fig. 1. Testing station prepared for tests
Tests – preparation of sample
and testing station
wards, the sample underwent etching. The
processing of the sample aimed to decrease
the effect of weld reinforcements and significant roughness of the element, which could
significantly deteriorate testing accuracy.
The sizes of the sample used in tests were
220×130×8 mm.
In order to carry out non-destructive testing (VT, PT and MT) it was necessary to prepare a testing station provided with measuring equipment following related standards
(PN-EN 13927: 2009, PN-EN ISO 3452-4:
2006 and PN-EN ISO 9934-3: 2003). The
testing station was equipped with the following devices and materials:
• set of reagents for penetration testing:
Diffu-Therm penetrant, remover and
developer,
• device for magnetising an object being
tested – contour probe DA400S made by
the Parker company,
• set of reagents for magnetic particle testing: Diffu-Therm ferromagnetic coloured and fluorescent powder, priming paint
and remover,
• set of measuring instruments and additional devices:
◦◦ instrument for measuring tangent and
residual areas,
◦◦ instrument for measuring the intensity
of ultraviolet and visible light,
◦◦ Berthold magnetic field indicator,
◦◦ reference samples,
◦◦ ultraviolet lamp,
◦◦ straightedge, slide caliper,
◦◦ magnifying glass.
The objective of the conducted experiments was to reveal differences and relations
between the results of visual, penetration and
magnetic particle testing methods carried out
on the same object, i.e. the same imperfections. To this end, it was necessary to make
test V-preparation butt joints of 10mm-thick
S355 steel plates. During MAG welding the
arc energy was significantly reduced, resulting in the
VT TESTING
lack of side fusion on both
x
y
d
No.
fusion lines. The external
[mm] [mm] [mm]
surfaces of the sample were
1
124
4
1
subjected to grinding, by
2
173
4
0,3
means of which weld face
3
188
4
0,4
reinforcement and weld
Fig. 2. Characteristics of imperfections revealed during VT
root were removed. After32
No. 3/2013
The equipment and measuring devices
used in the tests were provided with check
certificates and the reagents had confirmed
use-by dates. In order to verify the suitability
of reagents, their sensitivity was additionally
tested on standards appropriate for a given
method. The prepared testing station is presented in Figure 1.
The experimental tests (VT, PT and MT)
were carried out according to a specified sequence. The first was a visual test consisting
of a thorough inspection of the test surface
both by the naked eye and using a magnifying glass 3x. The visual testing did not reveal any linear imperfections on the surface
of the sample, yet it revealed three blowholes
which due to weld face grinding are visible
as pores with the following diameters d1=1
mm, d2=0.3 mm and d3=0.4 mm. The characteristics of point imperfections are presented
in Figure 2.
The second stage consisted of penetration testing by means of the dye penetrant
method. The results of the test are presented
in Figure 3 showing the sample following
the test, the sample sketch with plotted indications and a table with full dimensional
characteristics.
The penetration testing revealed the presence of twenty one linear indications, the total length of which was 192 mm as well as
identified three non-linear indications with
the diameters of d10 =2.2 mm, d15 =0.8 mm
and d17 =1.0 mm respectively.
At the next stage fluorescent magnetic
particle tests (MT-F) were carried out. The
results of the tests are presented in Figure 4
showing the sample after the test, the sample
sketch with plotted indications and a table
with full dimensional characteristics.
The magnetic particle testing revealed the
presence of twenty linear indications, the total length of which was 270 mm.
At the final stage, magnetic particle testing
using the dye penetrant technique was carried
out. The test revealed that the priming paint
PT: dye penetrant method
No.
Fig. 3. View of the sample after PT,
sketch of the sample with plotted
indications and table with full
dimensional characteristics
No. 3/2013
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
x [mm]
24
44
48
78
87
98
106
120
124
123
130
141
144
152
172
180
187
44
66
92
112
130
170
213
y [mm]
5
5
5
5
5
5
5
5
5
3
5
5
5
5
4
5
4
-5
-5
-5
-5
-5
-5
-5
l [mm]
d [mm]
12
2
10
2
2
6
5
2
2
--9
2
2
13
--2
--4
4
18
16
36
40
3
------------------2,2
--------0,8
--1,0
--------------33
No.
Fig. 4. View of the sample after MT-F, sketch of the
sample with plotted indications and table with full dimensional characteristics
can close the outlet of irregularities preventing subsequent, e.g. penetration, tests. Figure
5 presents the results of the test.
The penetration testing revealed the presence of thirty eight linear indications, the total
length of which was 247 mm, and also identified three non-linear indications having the
diameters of d10 = 1.2 mm, d15 = 0.4 mm and
d17 = 0.5 mm respectively.
Analysis of results and summary
The analysis involved the comparative assessment of welding imperfections detected
during the following tests:
• penetration testing with the dye penetrant
technique (PT-D),
• magnetic particle testing with the fluorescent technique MT-F,
• magnetic particle testing with the dye penetrant technique (MT-D).
The comparison of the presented test results reveals that detected imperfections
and their detections differ in basic dimensions. Following the general assessment of
the test sample area it is possible to draw a
conclusion that MT-F was characterised by
34
x
[mm]
MT -F
y
l
[mm]
[mm]
d
[mm]
-- 21
5
16
-- 48
5
4
-- 56
5
6
-- 68
5
4
-- 86
5
3
-- 92
5
2
-- 98
5
3
-- 110
5
10
-- 134
5
5
-- 142
5
32
-- 176
5
7
-- 185
5
25
--
8
-5
4,5
-- 14
-5
4,5
-- 40
-5
16
-- 66
-5
11
-- 90
-5
5
-- 96
-5
2
-- 102
-5
34
-- 142
-5
76
the best detectability of linear imperfections.
In this testing the total length of indications
amounted to 270 mm, in MT-D 247 mm, and
in the penetration method (PT-D) 192 mm.
The best result obtained in MT-F is approximately 9% more advantageous than that
obtained in MT-B, and 40% better than the
result obtained in PT-D. It is largely due to
the fact that penetration testing enables the
detection of open imperfections which at the
same time are present on the surface whilst in
magnetic particle testing such requirements
do not have to be satisfied.
Only MT-F failed to uncover non-linear imperfections. The failure to detect these
discontinuities is strictly related to the following factors:
• shape of the imperfection – a small oval
shape of the discontinuity does not cause
sufficiently great interference in the magnetic field,
• intensity – a small number of point discontinuities do not cause sufficiently great
interference in the magnetic field either.
The other two testing methods, in addition
to detecting linear indications, also made it
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
No. 3/2013
MT-D
No.
Fig. 5. View of the sample after MT-D, sketch of the
sample with plotted indications and table with full
dimensional characteristics
possible to uncover non-linear indications.
In PT-D the diameter of obtained indications
is, on the average, 130% greater in relation to
actual imperfections detected in visual tests,
whereas the indications in MT-D are on the
average 25% greater than those observed in
VT. This leads to the conclusion that indications detected by means of penetration testing
do not present the real dimensions of imperfections but rather the indications obtained
by the penetrant spread around imperfections
on the developer surface. Comparing PT-D
and MT-D, in which both types of indications were detected, it is possible to state that
magnetic particle testing is characterised by
better detectability and that its sensitivity is
better by approximately 30%.
In conclusion it is possible to state that
testing techniques analysed in this study
(VT, PT and MT) significantly differ as to the
detectability of linear and non-linear imperfections. Therefore it is advisable to always
use at least two surface testing methods. The
best configuration includes visual testing followed by magnetic particle testing with the
use of the dye penetrant method.
No. 3/2013
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
x
[mm]
28
44
51
64
82
88
96
102
120
124
126
130
140
146
154
164
172
177
173
186
188
193
198
10
16
38
48
58
62
67
73
79
81
92
102
112
125
132
164
172
212
y
[mm]
5
5
5
5
5
5
5
5
5
4
5
5
5
5
5
5
5
5
3,5
5
4
5
5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
-5
l
[mm]
10
5
8
2
3
6
4
4
2
--2
8
4
3
7
5
3
4
--4
--3
6
4
4
8
7
2
3
4
5
1,5
1,5
8
8
12
6
30
6
38
6
d
[mm]
------------------1,2
----------------0,4
--0,5
----------------------------------------35
References
przemysłowych. Wydawnictwo Instytutu
1. Czuchryj, J., and Sikora, S. (2007).
Spawalnictwa w Gliwicach, Gliwice,
Podstawy badań penetracyjnych wyPoland.
robów przemysłowych. Wydawnictwo
Instytutu Spawalnictwa w Gliwicach, 3. Krawczyk, R., Poch, K. (2011). Assessment of sensitivity of selected techniques
Gliwice, Poland.
for magnetic particle inspection. Biuletyn
2. Czuchryj, J., Drop, M., Sikora, S., and
Instytutu Spawalnictwa w Gliwicach, 55
Szymański, A. (2011). Podstawy badań
(4), pp. 33-39.
magnetyczno–proszkowych wyrobów
36
No. 3/2013
Mariusz Stachurski
Non-destructive testing of helically welded pipes made of
thermomechanically rolled materials used for sending of
combustibles
Abstract: In the first part of the paper it has been presented the short information
on methods of fabrication of helically welded pipes used for transporting of combustibles. In the second one it has been given the NDT methods used during inspection
of the pipes in production plants. In the next parts it has been described the NDT
methods for steel pipes focusing first of all on visual, ultrasonic and radiographic
ones. The paper has been ended with the information about prospects of gas industry
development in Poland and in the world.
Keywords: welding, non-destructive testing, helially welded pipes
Introduction
The golden age of pipelines has come
with a permanently increasing demand for
raw materials; the excavation of which is an
enormously complicated undertaking. As of
today, the best solution related to this type
of transportation is a subterranean grid of
main pipelines made of steel pipes. Poland’s
gas distribution system has approximately 17 thousand kilometres of high-pressure
gas pipelines and approximately 79 thousand
kilometres of distribution network, 23 gas
pumping stations, and approximately 4 thousand 1st and 2nd degree pressure reducing and
measuring stations. A growing consumption
of energy combined with a growing demand
for energy carriers, in particular for natural
gas, shapes the development of the network of
industrial pipelines. Great distances between
the sources at which natural gas and oil are
excavated and the places of their consumption require the construction of transit pipelines characterised by very high operating
parameters, e.g. gas pipelines with a diameter of over 1000 mm and pressure of 10 MPa.
Such pipelines must be made of pipes characterised by a strength of over 690 MPa
(485MB, 555MB) and very good weldability. As a result, it is possible to observe the
development of low-alloy steels (C 0.16%,
Si 0.55%, Mn 1.9%), often with microadditions of Nb, Mo, or Ti. Such pipes are obtained by means of controller rolling referred
to as thermoplastic or thermomechanical
treatment. Pipes for the pipelines mentioned
above are produced following the regulations
contained in DIN standards or in their American (API5L) and European (EN 10208)
counterparts. Large-diameter steel line pipes
for transporting liquid and gaseous fuels are
used as the following:
• seamless pipes (S), obtained by means of
hot or cold working methods,
• welded pipes (W), produced by means of
electric induction welding (EW) or butt
welding (BW) as well as submerged arc
welding (SAW) with a longitudinal weld
(SAWL) or a helical weld (SAWH). Figure 1 presents the process of welding pipes
with a helical weld (two-stage). Pipes can
dr inż. Mariusz Stachurski (Ph.D. Eng) - Bureau Veritas Poland ltd.
No. 3/2013
37
also be welded using the combination of
gas-shielded welding of root runs and
submerged arc welding of filling layers
(COW) [1].
Second stage submerged arc
welding (with three automatic welding machines
simultaneously)
First stage tack weld –
MAG welding
Fig. 1. Two-stage helical welding of pipes used by the
German company of Salzgitter Mannesmann Grossrohr
GmbH [5]
This article presents non-destructive testing of thermomechanically rolled helically submerged arc welded (SAWH) pipes
L485MB and L555MB. These pipes were selected as they are typical and manufactured
in many steelworks both in Poland and overseas. Following their manufacture the pipes
undergo insulation application, usually with
polyethylene or polypropylene and are painted inside with epoxy paints.
Non-destructive testing used during
pipe manufacture
General information
The basic purpose of non-destructive testing (NDT) is to assess the condition of a material or welded joint and, on this basis, to
issue an opinion related to the quality, durability and safe operation of a structure. The
effectiveness and quality of non-destructive
testing depends on numerous factors:
• personnel competence based on adequate
training,
• experience,
• inspection procedures,
• equipment,
• environment in which testing is carried out,
• psychological pressure, to which NDT
personnel are exposed,
• working time during the day,
38
• equipment features,
• supervision and how it is exercised,
• applied standards, requirements and
guidelines,
• systems applied for the certification of
personnel, procedures and equipment [1].
The basic NDT methods used in the production of pipes are the following:
• visual testing – VT,
• radiographic testing of pipe ends – RT,
• automatic fluoroscopic testing of a whole
welded joint,
• ultrasonic testing of a burden material in
the form of a strip unwound from a coil
(laminar imperfection test) – UT,
• automatic ultrasonic testing of a whole
welded joint – UT,
• manual ultrasonic testing of a welded joint
on a pipe end – UT,
• manual ultrasonic testing of a welded joint
being repaired – UT.
In addition, in order to determine precisely the size of welding imperfections, also
other NDT methods, i.e. penetration testing
(PT) and magnetic particle testing (MT),
can be used.
The basic standard describing the selection of testing manners and assessment criteria is standard PN-EN 10208-2 “Steel pipes
for pipelines for combustible fluids – Technical delivery conditions – Part 2: Pipes of requirements class B”. The standards describing testing manners and assessment criteria
are the following:
• PN-EN 10246-9 Non-destructive testing
of Steel Tubes – Part 9: Automatic Ultrasonic Testing of the Weld Seam of Submerged Arc Welded Steel Tubes for the
Detection of Longitudinal and/or Transverse Imperfections;
of Steel Tubes – Part 10: Radiographic
Testing of of the Weld Seam of Automatic Fusion Arc Welded Steel Tubes for the
Detection of Imperfections;
No. 3/2013
of Steel Tubes – Part 15: Automatic Ultrasonic Testing of Strip/Plate Used in Manufacture of Welded Steel Tubes for the Detection of Laminar Imperfections;
of Steel Tubes – Part 16: Automatic Ultrasonic Testing of the Area Adjacent to the
Weld Seam of Welded Steel Tubes for the
Detection of Laminar Imperfections;
of Steel Tubes – Part 17: Ultrasonic Testing of Tube Ends of Seamless and Welded
Steel Tubes for the Detection of Laminar
Imperfections;
• PN-EN ISO 10893-6 Non-destructive
testing of Steel Tubes – Part 6: Radiographic Testing of the Weld Seam of
Welded Steel Tubes for the Detection of
Imperfections;
• PN-EN ISO 10893-8 Non-destructive testing of Steel Tubes – Part 8: Automatic Ultrasonic Testing of Seamless and Welded
Steel Tubes for the Detection of Laminar
Imperfections;
• PN-EN ISO 10893-9 Non-destructive
testing of Steel Tubes – Part 9: Automated Ultrasonic Testing for the Detection of
Laminar Imperfections in Strip/Plate Used
in the Manufacture of Welded Steel Tubes;
Table 1. Non-destructive tests used in manufacture of pipes [2]
Types of imperfections
detected during pipe
manufacture
Residual magnetism
on pipe ends
Laminar imperfection
on pipe ends
Longitudinal/transverse
imperfections in welded
joint
Laminar imperfections
on pipe body
Laminar imperfections
on strip edge where
transverse weld was
made
Non-destructive testing
of weld in pipe ends
(areas not tested previously) / testing of areas
being repaired
No. 3/2013
Testing status
Obligatory
Optional
Obligatory
Optional
Optional
Obligatory
Types of tests, requirements and acceptance levels
Gaussmeter using Hall effect or equivalent; max. 30 gausses,
random tests
Ultrasonic testing according to PN-EN 10246-17 or PN-EN
ISO 10893-8, inspection area: max. 6 mm from pipe ends on
circumference
Ultrasonic testing according to EN 10246-9 or PN-EN ISO
10893-11, acceptance level U2/U2H or calibration method
”two lamb” (also for welds joining strip ends in helically welded pipes)
Radiographic testing according to PN-EN 10246-10 or PN-EN
ISO 10893-6, image quality class R1, used for T-type joints of
helically welded pipes
ISO 10893-9, acceptance level U2
ISO 10893-9 or alternatively ultrasonic testing according to
PN-EN 10246-16 or PN-EN ISO 10893-8, acceptance level
U2
Ultrasonic testing according to PN-EN 10246-9 or PN-EN ISO
10893-11 for longitudinal imperfections, acceptance level U2/U2H
or (if not specified otherwise)
Radiographic testing according to PN-EN 10246-10 or
PN-EN ISO 10893-6, image quality class R1for longitudinal
imperfections
and
Ultrasonic testing according to PN-EN 10246-9, (PN-EN
ISO 10893-11) or radiographic testing according to PN-EN
10246-10 (PN-EN ISO 10893-6) for transverse imperfections
39
• PN-EN ISO 10893-11 Non-destructive testing of Steel Tubes – Part 11: Automated Ultrasonic Testing of the Weld Seam of Welded
Steel Tubes for the Detection of Longitudinal and/or Transverse Imperfections.
The standards of the PN-EN 10246 series
have been now replaced by the standards
of the PN-EN ISO 10893 series. The basic
product standard PN-EN 10208-2 (version in
effect) still indicates older standards as those
in effect. As a result, all the greater production plants continue to use these standards in
the manufacturing of pipes. For this reason,
in all cases this study refers to both standards
as recommended for use. As far as ultrasonic manual testing is concerned, the standards
applied are the basic standards concerned
with the manner of the ultrasonic manual
testing of welded joints and their assessment
and are not discussed in this study. A detailed
list of non-destructive tests applied in pipe
manufacture is presented in Table 1.
The residual magnetism at the ends of
each pipe, parallel in relation to the pipe axis,
should not exceed 30 G (3mT). The measurement of magnetism should be carried out on
a random basis on the butting face of the pipe
end, by means of a calibrated meter (gaussmeter) using the Hall effect or by means of
equivalent devices.
All non-destructive tests of welded joints
should follow a hydraulic test of the pipe.
Visual testing (VT)
In accordance with the requirements of
standard PN-EN 10208-2 each pipe should
undergo a visual inspection of the entire
outer surface. The inner surface of the pipe
should be inspected on each side of the pipe
if its outer diameter is 610 mm or less and in
100% of the entire inner surface if the outer
diameter of the pipe is 610 mm or greater.
Visual testing should be carried out by adequately trained personnel and with proper
40
lighting of the surface being viewed (the
light intensity should amount to a minimum
of 300 lx).
Pipes should be free from imperfections
as a finished product. The appearance of the
outer and inner surfaces of the pipe should be
typical for the production process and heat
treatment applied. No surface imperfections
requiring removal should be visible.
Below are presented manners of dealing
with imperfections detected on the outer
surface:
• imperfections with a depth equal to or below 12.5% of a required wall thickness,
which do not reduce the nominal thickness
of the pipe in the place of their occurrence
should be treated as allowed imperfections
(they can remain on the pipe surface or, in
accordance with the manufacturer’s decision, can be removed by cosmetic grinding);
• imperfections with a depth exceeding
12.5% of a required wall thickness, which
do not reduce the nominal thickness of the
pipe in the place of their occurrence are
classified as imperfections and should be
removed by grinding (after grinding it is
necessary to apply a proper NDT method,
e.g. MT, in order to check the pipe surface
after grinding);
• imperfections which reduce the nominal
thickness of the pipe in the place of their
occurrence should be treated as unallowed
imperfections and repaired by welding or
the part of the pipe containing such imperfections should be cut out (the required
length of the time should be maintained).
If the aforesaid solution is impossible,
a pipe with an unallowed imperfection
should be rejected.
Geometrical deformations of the cylindrical
contour of the pipe should not exceed the
following:
• 3 mm (flattening, metal excess, coldformed indentations with gentle edges);
• 6 mm (other indentations).
No. 3/2013
Outer imperfections can be removed only by
machining or grinding. The thickness of a
pipe wall in the areas being repaired cannot
be less than the required nominal wall thickness. All machined/ground areas should gently pass into the contour of the pipe [2].
Ultrasonic testing (UT)
Ultrasonic testing of strip (pipe body) for
laminar imperfections
The first type of ultrasonic testing used
during pipe checks is testing for laminar imperfections of the strip surface. In this case
it is possible to accept ultrasonic testing carried out at the sheet manufacturer’s or test
a (flat-formed) material used in the production of pipes, acting in accordance with the
requirements of standards PN-EN 10246-15
or PN-EN ISO 10893-9 and acceptance level
U2. Single laminar imperfections or laminar
imperfections in groups exceeding acceptance level U2 are unallowed [2, 6, 11]. An
example of a device for strip ultrasonic testing is presented in Figure 2.
Fig. 2. Device applied in UT of strips for the presence
of laminar imperfections, used in the Salzgitter Mannesmann Grossrohr GmbH compay, Germany [5]
No. 3/2013
Ultrasonic testing for the detection
of longitudinal and transverse imperfections
in welded joints
In order to ensure full NDT of welded joints it is required that helically welded pipes should undergo ultrasonic testing
along their whole length. The test objective
includes both longitudinal and transverse imperfections. Ultrasonic testing should be carried out in accordance with the requirements
of standard PN-EN 10246-9 or PN-EN ISO
10893-11 [2]. As a rule, UT of pipes follows
the completion of all the main operations of
a production process. In order to ensure the
reliability of ultrasonic testing, pipes should
be sufficiently straight and free from foreign
matter.
In order to detect longitudinal and transverse imperfections, helical welds should
undergo ultrasonic testing. In both cases the
testing should be conducted in two opposite
beam propagation directions. During the
test the converter unit should remain in the
proper position in relation to a weld so that
the entire weld could be searched through. A
searching rate should not oscillate by more
than 10% than the adopted basic rate. Depending on the thickness and roughness of
the pipe surface, the frequency used in UT
should be contained within a 1 MHz÷15 MHz
range. The maximum width of each single
converter, measured parallel to the main
axis of the pipe, should amount to 25 mm.
By means of an ultrasonic monitor connected with a marking and/or sorting system, the
equipment should enable the classification of
pipes as accepted or rejected. If it is not possible to test welds on pipe ends by means of
automatic ultrasonic testing equipment, the
testing of pipe ends should be conducted by
the manufacturer using either manual ultrasonic testing or radiographic testing.
Ultrasonic equipment for detecting longitudinal imperfections should be calibrated by means of four longitudinal grooves,
41
A reference groove should be of an
two on the outer surface and two in the inner
surface of pipe masters and/or by means of ”N-type” (Fig. 4). The sides of the groove
a datum hole (Fig. 3). Converters used for should be parallel to each other and the botdetecting transverse imperfections should tom should be perpendicular to the sides.
be calibrated by means of a hole and/or two
grooves transverse in relation to the weld,
one on the outside and one inside a sample
being tested. The decision whether to choose
grooves or a hole is at the manufacturer’s
discretion.
Fig. 4. “N-groove” [3, 9].
w – width, d - depth
Fig. 3. Arrangement of reference grooves and datum hole [3, 9]
1. longitudinal outer grooves
2. submerged arc made weld
3. pipe master or pipe section
4. right-through hole
5. longitudinal inner grooves
A test sample should have the same nominal diameter, wall thickness, surface roughness, heat treatment state and similar acoustic properties (e.g. wave propagation velocity
and wave damping coefficient) as the pipe
being tested. The manufacturer should have
the possibility of removing the inner and outer weld run in accordance with the pipe curvature. The outer and inner grooves as well
as the datum hole should be sufficiently distant from the pipe ends and from each other
so that clearly separate indications of signals
coming from the grooves and the hole can be
obtained.
42
Longitudinal reference grooves should be
located in the parent metal near the edge of the
weld and in parallel to the weld run (Fig. 3).
A reference groove should be electrodischarge machined or made with another method. The dimensions of reference grooves
should be the following:
a)width, w, (Fig. 4) of a reference groove
should not exceed 1.5 mm;
b)depth, d, (Fig. 4) should be as specified in
Table 2, subject to the following restrictions:
◦◦ minimum groove depth: 0.3 mm for
pipes of categories U2 and U3 and 0.5
mm for pipes of category U4;
◦◦ maximum groove depth: 2.0 mm for
pipes of categories U2 and U3 and 3
mm for pipes of category U4;
c)depth tolerance for a reference groove
should amount to ± 15% depth or ± 0.05
mm, whichever value is greater;
d)length of reference grooves should be
equal to at least double width of each single converter, but not more than 50 mm.
A datum hole should be drilled through the
wall of a test sample, in the weld centre, perpendicularly to the sample surface (Fig. 3).
No. 3/2013
The diameter of the drill used to make the datum hole should be selected in accordance with
Table 3. The diameter of a datum hole should
be checked and cannot exceed the nominal diameter of a drill by more than 0.2 mm.
testing sensitivity by 3 dB (to compensate
equipment indications), it is necessary to
carry out equipment calibration again as
well as to check all the pipes tested since the
previous calibration verification.
Table 2. Designation of acceptance level and related
depth of reference groove [3, 9]
Table 3. Designation of acceptance level and related
drill diameter [3, 9]
Acceptance
level
U2
U3
U4
U5
Groove depth in % of nominal
wall thickness
5.0
10.0
12.5
15.0
The equipment should be calibrated so
that it would ensure (e.g. in three subsequent
passes of the test sample through the equipment) obtaining clearly distinguishable signals of the reference standard. The peak value of the amplitude of these signals should
be used for adjusting the level of the gate/
alarm monitor in the equipment. The rate
of the test sample travel in relation to the
set of ultrasonic converters during the verification of calibration should be the same
as during the tests of a product. Calibration
can be verified by means of a semi-dynamic
method. During the testing of pipes with the
same nominal diameter, wall thickness, and
made of the same steel grade, it is necessary
to verify the calibration of the equipment
at regular intervals. The equipment calibration is verified by putting the pipe master
through the testing equipment. The verification of calibration should be carried out at
least once every four hours as well as each
time the equipment operator changes. Calibration should also be verified at the beginning and end of production. In addition, the
equipment should be calibrated if any of the
parameters set during the initial calibration
have been altered. If while verifying calibration during production, calibration requirements are not satisfied even after increasing
No. 3/2013
Acceptance
level
Drill diameter
U2H
1.6
U3H
3.2
U4H
4.0
Following testing, a pipe revealing signals
below the monitor/gate sensitivity threshold
should be classified as meeting the requirements. And a pipe having signals equal to
or greater than the monitor/gate sensitivity
threshold should be classified as questioned
or, at the manufacturer’s request, should
undergo another test. If during the subsequent test of the same pipe no signal equal
to or greater than the monitor/gate sensitivity threshold was obtained, the pipe should
be regarded as meeting the requirements.
Depending on the requirements of a product standard, pipes questioned during testing
should be dealt with in accordance with one
or several following manners:
a)Subject to an agreement between the purchaser and the manufacturer, the area
which has been questioned should be tested by means of another NDT technique or
method (usually radiographic), on the basis of agreed acceptance criteria;
b)The questioned area should be cut out. The
manufacturer should guarantee that the
whole questioned area has been removed;
c)The pipe should be classified as not satisfying the requirements [3, 9].
Testing of helically welded pipes is carried
out for acceptance level U2/U2H, taking into
account the following inspection criteria:
43
a)the maximum groove depth should amount
to 2.0 mm;
b)Using inner and outer longitudinal
grooves in the middle of the weld run for
the purpose of equipment calibration is
unallowed;
c)In the case of acceptance level U2, for the
purpose of detecting transverse imperfections, as an alternative to the datum hole
one may use the inner and outer grooves
positioned at a right angle and centred
in the weld. In such a situation both inner and outer weld reinforcements should
be ground evenly with the pipe contour,
in the direct vicinity and on both sides of
reference grooves. The grooves should be
sufficiently distant from each other in a
longitudinal direction. They should also
be sufficiently distant from any other reinforcements. Such an approach is necessary to obtain a clear and distinguishable
ultrasonic signal response. In order to determine the level of the gate/alarm of the
measuring equipment it is necessary to use
the complete amplitude of a signal from
each of these grooves.
d)For acceptance level U2 and by prior
agreement, as an alternative to using reference grooves for calibration purposes, it
is possible to use inner and outer grooves
of a constant depth and increase testing
sensitivity using electronic methods (by
increasing dB amplification). In this case,
also referred to as the „two lamb method”,
the depth of grooves should be twice as
big as the length of an ultrasonic wave.
A required increase in testing sensitivity
should depend on a pipe wall thickness. The
manufacturer should adequately assure the
purchaser that obtained testing sensitivity is
basically equivalent to that obtained while
using grooves for acceptance level U2. In
the case of helically welded pipes the entire
length of the weld joining strip edges should
also undergo ultrasonic testing. Testing sen44
sitivity and parameters should be the same as
during the initial testing of the pipe helical
weld. In addition, cruciform joints, in which
the weld ends of strip edges meet the helical
weld should undergo radiographic testing [2].
Following the tests, the manufacturer
should provide the ordering party with a report containing at least the following information:
a)reference to standards PN-EN 10246-9 or
PN-EN ISO 10893-11 and
PN-EN 10208-2;
b)date of test report;
c)acceptance level;
d)declaration of conformity;
e)product specification, i.e. steel grade and
dimensions;
f) type and details of testing technique;
g)description of a master/standard [3, 9].
An example of a device for UT of helically
welded pipes is presented in Figure 5.
Fig. 5. Automatic multichannel device for ultrasonic
testing of helically welded pipe weld and strip-joining
weld used in the Salzgitter Mannesmann Grossrohr
GmbH company, Germany[5]
Laminar imperfections in strip area adjacent
to welded joint
In such a case it is possible to accept ultrasonic testing carried out at steel sheet manufacturer’s or carry out post-weld testing at
the pipe manufacturer’s, following the requirements of standard PN-EN 10246-16 or
PN-EN ISO 10893-8 and acceptance level
No. 3/2013
U2. The tests are conducted within a 15mmwide zone along both longitudinal edges of a
helical weld. In the case of transverse welds
joining the edge of a strip testing is conducted along areas adjacent to this butt weld. Requirements concerned with ultrasonic testing
of laminar imperfections in the area adjacent
to the weld are dealt with in standard PNEN 10246-15 (PN-EN ISO 10893-9) or PNEN 10246-16 (PN-EN ISO 10893-8). Single
laminar imperfections or imperfections in
groups exceeding acceptance level U2 are
not allowed [2].
Non-destructive testing of helically welded
joints on pipe ends and in areas being
repaired
Sections of welded joints on pipe ends
which cannot be tested by means of automatic ultrasonic equipment and repaired areas of
welds should undergo the following tests:
a)in order to detect longitudinal imperfections – manual or semiautomatic ultrasonic
testing using the same testing parameters
and sensitivity as in the case of the automatic testing of the whole helical weld.
Radiographic testing is also possible.
b)in order to detect transverse imperfections
– manual or semiautomatic ultrasonic testing using the same testing parameters and
sensitivity as in the case of the automatic
testing of the whole helical weld. Radiographic testing is also possible.
c)It is also necessary to carry out testing
for the laminar imperfections of the pipe
body areas on pipe ends which cannot be
checked by means of automatic ultrasonic
equipment. In such a case it is also possible to carry out manual ultrasonic testing using the same testing parameters and
sensitivity as those used while testing areas adjacent to the weld. Laminar imperfections with a length ≥ 6 mm occurring
in girth direction are unallowed if they are
present 25 mm away from each pipe end;
No. 3/2013
d)While carrying out manual ultrasonic
testing a scanning rate should not exceed
150 mm/s [2].
Radiographic and radioscopic testing
(RT)
Another testing method used in the manufacture of helically welded pipes is radiographic testing, usually carried out on pipes
after the completion of all principal production operations. In order to ensure an appropriate testing class, pipes intended for tests
should be sufficiently straight and free from
foreign matter impurities. The surfaces of
welds and adjacent parent metal should be
free from foreign matter and surface imperfections as they could impede the interpretation of radiograms. Grinding surfaces is
allowed if it can ensure obtaining a surface
condition acceptable for tests. If a weld reinforcement must be removed, it is advisable to place markers (usually lead arrows) in
order to identify the weld location in a radiogram. In order to ensure the unambiguous
identification of a given weld section, each
section should be provided with identification symbols (usually lead letters) so that
their images can be visible in a radiogram.
It is necessary to apply permanent marking
on the pipe surface from the side of the radiation source. This is done in order to provide
reference points for the accurate assignment
of each radiogram location. If it is impossible to strike markers due to the type of a
product and/or its intended operating conditions, it is necessary to provide other appropriate means of radiogram assignment, e.g.
by marking with paint or preparing accurate
sketches. In order to ensure that each part of
the weld undergoes testing, during X-raying
a longer weld section with several films it
is necessary that neighbouring films should
overlap over a minimum length of 10 mm.
Longitudinal or helical welds of pipes
should undergo radiographic testing with
45
x-rays. The use of radioscopic methods is allowed (Fig. 6), yet only in cases when the
manufacturer can demonstrate their appropriate sensitivity.
Fig. 6. Example of device for radioscopic testing
of welded joints
There are two image quality classes, i.e.
R1 and R2:
• class R1 (class B): radiographic testing
technique by means of X-radiation of
heightened sensitivity;
• class R2 (class A): radiographic testing
technique by means of X-radiation of normal sensitivity.
For image quality class R1 it is necessary
to use at least fine-grained films (at least C4
class), whereas for image quality class R2
one should use at least medium-grained films
(at least C5 class). For both image quality
classes (R1 and R2), the thickness of front
intensifying screens should be between 0.02
mm and 0.25 mm. Rear intensifying screens
can have other thicknesses. Fluorescent intensifying screens should not be used.
The amount of backscattered and internal
X-radiation should be limited to a minimum.
When in doubt as to the efficiency of protection against backscattered radiation, it is advisable to fix a characteristic sign (usually a
1.5mm-thick letter B) behind the film holder
or film frame and make a radiogram in a nor46
mal manner. The density of the symbol image
in the radiogram lower than the background
density indicates that the protection against
backscattered radiation is insufficient and that
additional precautions should be used. A radiation beam should be directed onto the centre
of a weld section being tested and should be
perpendicular to the pipe surface at this point.
A length subjected to an assessment should be
so that the thickness X-rayed at the ends of the
usable length of a radiogram does not exceed
the thickness X-rayed in the radiogram centre by more than 10% for image quality class
R1 and by more than 20% for image quality
class R2. It is necessary to use a technique of
X-raying through one wall. If, due to the size
of the object, such a technique cannot be used,
it is allowed to use a technique of X-raying
through two walls (by prior agreement). The
distance between the film and the weld surface should be as small as possible.
The minimum distance between the source
and the sample ,f, should be selected using
appropriate formulas or diagrams. The conditions of exposition should be so that the
density of a radiogram in the tested area of
weld material free from imperfections is not
lower than 2.0 for image quality class R1 (2.3
for image quality class B) and not lower than
1.7 for image quality class R2 (2.0 for image
quality class B).
The quality of an image should be determined using an image quality indicator (IQI)
made of mild steel, grade specified in standards PN-EN 462-1 (PN-ISO 19232-1) and PNEN 462-2 (PN-ISO 19232-2). An IQI should
be placed on the surface from the radiation
source side or next to the weld or, in the case
of a wire-type IQI – across the weld. Image
quality indicators should be placed from the
film side only if the surface from the source
side is inaccessible. In such cases it is necessary to place a letter “F” next to an IQI and
make a note of this change of the procedure in
the test report [4, 10].
No. 3/2013
diameter T/3 (T – wall thickness), whichever is smaller;
c)The sum of the diameters of all acceptable single imperfections in each section
of a weld with the following dimensions:
a length of 150 mm or 12 T, whichever
is smaller, should not exceed the smaller
of the following values: 6.0 mm or 0.5 T,
where the distance between single inclusions should be smaller than 4 T;
d)Single longitudinal slag inclusions with
a length of 12.0 mm or 1 T, whichever
is smaller or with a width not exceeding
1.6 mm are acceptable. The total length
of all acceptable single imperfections in
each section of a weld with the following
dimensions: a length of 150 mm or 12 T,
whichever is smaller, should not exceed
12.0 mm, where the distance between single inclusions should be smaller than 4 T;
e)Single undercuts of any length and with a
maximum depth of 0.4 mm are acceptable;
f) In addition, single undercuts with a maximum length of T/2 and a maximum depth
of 0.8 mm which does not exceed 10%
of a specified wall thickness are allowed
provided that there are not more than two
such undercuts on a weld section with a
length of 300 mm and that all such undercuts will be levelled;
g)All undercuts exceeding the above values
should be repaired; in other cases the area
under suspicion should be cut out or the
pipe should be rejected;
h) All undercuts, of any
length and depth, on the inTable 4. Requirements for radiographic testing sensitivity and image
quality class R1 according to PN-EN 10246-10 and PN-EN ISO 10893-6 [2] ner and outer side of the
weld, overlapping in the lonWall thickness
Required visibility
gitudinal direction are unacOver [mm] Up to [mm] Hole diameter [mm] Wire diameter [mm]
ceptable [2].
10
0.40
0.16
4.5
Pipes not exceeding allowed
16
0.50
0.20
10.0
values/indications should be
25
0.63
0.25
16.0
classified as passing the test.
32
0.80
0.32
25.0
Pipes exceeding allowed val40
1.00
0.40
32.0
ues/indications should be clas-
The radiographic testing of helically
welded pipes is carried out following the requirements of standard PN-EN 10246-10 or
PN-EN ISO 10893-6, image quality class
R1, and in accordance with the following
conditions:
a)The conformity with sensitivity-related
requirements presented in Table 4, determined on the basis of a base metal should
be verified by means of a wire-type image
quality indicator, in accordance with standard PN-EN 462-1 (PN-ISO 19232-1) or by
means of an equivalent hole-type IQI according to PN-EN 462-2 (PN-ISO 19232-2);
b)Radiographic testing should be carried
out using only X-radiation, fine-grained
and high contrast films as well as lead
screens. It is possible to use fluoroscopic
methods, yet only when the manufacturer
can demonstrate their equivalence with a
method based on X-ray films;
c)The optical density of an X-ray photograph cannot be lower than 2.0 and should
be selected so that the density in the thickest area of the weld is not lower than 1.5
and that the maximum contrast for a given
film type is maintained.
Requirements related to the assessment of
welded joints are the following:
a)Cracks, no joint penetration and incomplete fusion are unacceptable;
b)Single slag inclusions and blowholes are
acceptable if characterised by the following dimensions: size up to 3.0 mm or a
No. 3/2013
47
sified as questioned. Questioned pipes should
be dealt with in one of the following manners:
a)A questioned area should be levelled by
means of an appropriate method. In order
to ensure that the imperfection has been
removed entirely, after checking that the
remaining thickness is still within a tolerance range, the area should undergo another check using magnetic particle testing or
penetration testing. Afterwards, the pipe
should be classified as passing the test. If
the recommended removal of the faulty
area has decreased the thickness below
the acceptable value, the questioned area
should be repaired by welding carried out
in accordance with an approved welding
procedure specification. After that, the repaired area should undergo radiographic
testing following the requirements of PNEN 10246-10 or PN-EN ISO 10893-6;
b)A questioned area should be cut out. The
manufacturer should ensure that the whole
questioned section has been cut out;
c)A pipe should be classified as failing to
meet the requirements.
In required cases the manufacturer should
write a test report containing at least the following information:
a)reference to PN-EN 10246-10 or PN-EN
ISO 10893-6 and PN-EN 10208-2;
b)test report date;
c)declaration of conformity;
d)designation of a product providing grade
and dimensions;
e)type of and detailed information about the
testing technique used;
f) each deviation from specified procedures,
agreed or not agreed;
g)image quality class;
h)operator’s name, signature and certificate
number [4, 10].
velopment owing to new gas pipelines under
construction both in Poland and abroad. The
production of these pipes requires the use of
materials characterised by the highest mechanical properties in a given material group.
NDT devices are usually provided with modern and recently developed technical solutions. Investment-related expectations for
the years to come make the manufacture of
welded pipes the area of interest for investors and manufacturers all over the world.
References
1. Michałowski, W., and Trzop, S., et al.
(2006). Rurociągi dalekiego zasięgu. Warsaw: Wydawnictwo Fundacja Odysseum.
2. PN-EN 10208-2:2011 Steel pipes for
pipelines for combustible fluids - Technical delivery conditions - Part 2: Pipes of
requirement class B
3. PN-EN 10246-9:2004 Non-destructive
testing of steel tubes - Part 9: Automatic ultrasonic testing of the weld seam of
submerged arc welded steel tubes for the
detection of longitudinal and/or transverse imperfections
testing of steel tubes - Part 10: Radiographic testing of the weld seam of automatic fusion arc welded steel tubes for
the detection of imperfections
5. Presentation of Salzgitter Mannesmann
Grossrohr GmBH - Salzgitter 2012.
testing of steel tubes - Part 15: Automatic ultrasonic testing of strip/plate used in
the manufacture of welded steel tubes for
the detection of laminar imperfections
testing of steel tubes - Part 16: Automatic
ultrasonic testing of the area adjacent to
the weld seam of welded steel tubes for
Conclusion
the detection of laminar imperfections
In recent years the manufacture of heli- 8. PN-EN 10246-17:2002 Non-destructive
testing of steel tubes - Part 17: Ultrasoncally welded pipes has seen an intensive de48
No. 3/2013
ic testing of tube ends of seamless and 11.PN-EN ISO 10893-9:2011 Non-destructive testing of steel tubes - Part 9: Autowelded steel tubes for the detection of
mated ultrasonic testing for the detection
of laminar imperfections in strip/plate
9. PN-EN ISO 10893-6:2011 Non-destrucused for the manufacture of welded steel
tive testing of steel tubes - Part 6: Ratubes
diographic testing of the weld seam of
welded steel tubes for the detection of 12.PN-EN ISO 10893-11:2011 Non-destructive testing of steel tubes - Part 11: Autoimperfections
mated ultrasonic testing of the weld seam
10.PN-EN ISO 10893-8:2011 Non-destrucof welded steel tubes for the detection of
tive testing of steel tubes - Part 8: Autolongitudinal and/or transverse imperfecmated ultrasonic testing of seamless and
tions
welded steel tubes for the detection of
No. 3/2013
49

Biuletyn Instytutu Spawalnictwa No. 3/2013

Transkrypt

Podobne dokumenty

Welding Center - mostostal zabrze sa

m/v Baltic