Welding steels
without hydrogen
cracking
SECOND EDITION (Revised)
N BAILEY, F R COE, T G GOOCH, P H M HART,
N JENKINS and R J PARGETER
WOODHEAD PUBLISHING LIMITED
Cambridge England
Published by Woodhead Publishing Limited, Abington Hall, Abington
Cambridge CB1 6AH, England
www.woodhead-publishing.com
First published 1973, The Welding Institute
Second edition, 1993, Abington Publishing (an imprint of Woodhead Publishing Limited)
and ASM International
Reprinted with revisions 2004, Woodhead Publishing Limited
© 1993 and 2004, Woodhead Publishing Limited
The authors have asserted their moral rights.
This book contains information obtained from authentic and highly regarded sources.
Reprinted material is quoted with permission, and sources are 'indicated. Reasonable
efforts have been made to publish reliable data and information, but the authors and the
publisher cannot assume responsibility for the validity of all materials. Neither the
authors nor the publisher, nor anyone else associated with this publication, shall be
liable for any loss, damage or liability directly or indirectly caused or alleged to be
caused by this book.
Neither this book nor any part may be reproduced or transmitted in any form or by
any means, electronic or mechanical, including photocopying, .microfllming and recording, or by any information storage or retrieval system, without permission in writing
from the publisher.
The consent of Woodhead Publishing Limited does not extend to copying for general distribution, for promotion, for creating new works, or for resale. ~ecific
permission must be
obtained in writing from Woodhead Publishing Limited for such copying.
Trademark notice: Product or corporate names may be trademarks or registered trademarks,
and are used only for identification and explanation, without intent to infringe.
British Library Cataloguing in Publication Data
A catalogue record for this book is available from the British Library.
ISBN 1 85573 014 6
Quotations at the head of each chapter are from:
'The Pilgrim's Progress - from this world to that which is to come' by John Bunyan. 1684
Typeset by SNP Best-set Typesetter Ltd., Hong Kong
Printed by TJI Digital. Padstow, Cornwall, England
Preface
Research is useless if the results obtained do not find practical application
and thus make their contribution to the improved efficiency of industrial
production and the consequent increase in the amenities of our daily life.
Sir William J Larke, KBE, 1947, Founder President of BWRA
Hydrogen cracking represents the most common problem encountered when welding steel structures. The major variables influencing
the incidence of cracking have been defined for many years and the
design of welding procedures is dominated by the need to incorporate appropriate safeguards. The previous edition of 'Welding
steels without hydrogen cracking' described these factors and further
presented nomograms on the basis of which cracking could be
reliably and economically avoided in steels of different types. The
success of this approach was such that, with only minor changes,
the appropriate part of the book formed the basis of the guidelines
in the British Standard BS 5135: 1974 'Arc welding of carbon and
carbon-manganese steels', and was retained in the latest, 1984
version.
Since publication of the first edition, significant changes have
taken place in steel compositions and production routes, many of
which have been intended primarily to obtain improved weldability,
especially in the sense of avoiding hydrogen cracking. In consequence, materials now commonly welded have compositions
outside those used from the derivation of the nomograms in the
first edition. Appropriate experimental work has been carried out to
define the effects of such material changes on cracking behaviour.
This edition has been produced to recognise both changes in steel
formulation and the body of data which now exists regarding cracking
sensitivity. In large part, the original format of 'Welding steels
without hydrogen cracking' remains untouched, indicative of the
soundness of the methodology presented. However, a number of
changes have been required and these are presented in this second
edition. In Chapter 4, dealing with welding procedures, these
include modifications to the nomograms for steels having carbon
viii
Preface
equivalents below 0.40 and the addition of a diagram to show conditions to avoid hydrogen cracking in C-Mn weld metals.
The original edition was produced with a major contribution from
F R Coe, with support from others cited in the preface. This revised
edition owes a considerable debt to Mr Coe. The revisions were
made most especially by N Bailey, T G Gooch, PH M Hart, N Jenkins
and R J Pargeter.
It is hoped that this second edition will be of assistance to all concerned in welding transformable steels, whether metallurgists,
welding or mechanical engineers, or designers. Because technology
is continually advancing, it is essential that new information should
be incorporated as soon as possible and, as was the case of the first
edition, TWI remains anxious to obtain practical feedback from
users of the book, both on its application to the practical situation,
and on new data that may become available.
Although hydrogen cracking is usually the major technological
problem to overcome when welding ferritic steels, the reader is also
recommended to study a companion volume, 'Weldability of ferritic
steels', which is being prepared as an introduction to the topic by
one of the authors of the present text, Norman Bailey. In addition to
a short chapter on hydrogen cracking, other topics related to fabrication cracking, the achievement of required properties and service
metallurgical problems, are covered.
T G Gooch
Head of Materials Department
It is now just over ten years since the publication of the second
edition of this book, and thirty since the original version. However,
hydrogen cracking remains a significant issue in the fabrication of
steel structures, and the clear and practical exposition of the subject
matter remains as relevant as when it was first published. Nevertheless, over the last ten years there have been developments in
various standards, and in particular, European standards have
moved to a universal description of welding conditions in terms of
heat input, rather than arc energy. The opportunity has therefore
been taken to revise and update the text and diagrams, and to bring
them into line with current practice. These revisions were principally made by Briony Lee, Richard Pargeter and Peter Hart.
PH M Hart
Manager, Metallurgy, Corrosion, Arcs & Surfacing Group
Chapter 1 Defining the problem
Would'st thou read Riddles, and their Explanation?
Just over 20 years ago, it was estimated that, in Britain alone, costs
amounting to £260 million were annually incurred as a result of
manufacturing problems directly attributable to welding. At least
£40 million of this total arose from the need to repair hydrogeninduced cracks at and adjacent to welds. Service failures due to
fatigue and brittle fracture of welded components cost industry a
further £140 million annually.
Current figures are, no doubt, a good deal higher for, although
much more is known about how to avoid hydrogen cracking when
welding conventional steels, new steels and unexpected problems
have arisen, so that further research and guidance has been needed.
Problems have also arisen recently because the difficult economic
climate has made industry ever keener to cut costs to maintain its
competitive edge. In doing so, corners have been cut which have
led to outbreaks of cracking which could have been avoided by
following the principles detailed in this book, albeit at somewhat
greater initial cost.
A significant number of these costly, and sometimes tragic,
failures originated at small pre-existing cracks in the heat-affected
zone (HAZ) of the parent steel adjacent to the weld. The most
common HAZ defects are those resulting from the presence of
hydrogen in the weld. The fracture of pressure vessels during
hydraulic testing and the collapse of structural steelwork in service
have both provided instances of failures in which hydrogen cracking
was involved.
This type of cracking often occurs some time after welding has
been completed and, although extensive, may be difficult to detect.
Thus a heavy responsibility is placed on the fabricator to match the
welding procedure with the material for each application so that
cracking does not occur. The incidence of failures suggests that this
has not always been successful. The data and techniques now
2
Welding steels without hydrogen cracking
provided in this book offer detailed guidance for deriving welding
procedures for avoiding hydrogen cracking.
As a topic for research, hydrogen-induced cracking has perhaps
received more attention than any other similar phenomenon but the
precise mechanism of embrittlement is still not fully understood in
terms of chemical and metallurgical reactions, which could be confirmed by direct experiment. Nonetheless, the conditions which collectively result in hydrogen cracking can be recognised and may be
stated simply as 'sufficient hydrogen and sufficient stress in a susceptible microstructure at a temperature usually below 150 aC.'
It is impossible to avoid producing the necessary stresses and temperatures in a cooling weld so measures to avoid cracking must rely
largely on control of hydrogen level, or control of microstructure, or
both. Once this is recognised it is possible to see how welding procedures can be derived from tables and diagrams indicating those combinations of conditions which achieve this control. Unfortunately,
there is no ideal welding diagram which is universally applicable and
a series of diagrams relating to different groups of steels is presented
in this book. The factors involved in hydrogen cracking and the way
these are manipulated to provide guidance for avoiding HAZ cracking
are dealt with in the first three chapters. Specific diagrams for different steel types are described in detail in Chapter 4, and Chapter 5 deals
with heat treatment for hydrogen removal. Appendix A summarises
current knowledge concerning hydrogen levels in welding, both with
different processes and different consumables, while Appendix B
recommends methods for laboratory measurement of hydrogen.
A number of British and other standards give guidance on preheat,
joint preparations, and other requirements. The more important of
these include:
BS 2633: 1987
BS 4570: 1985
BS EN 1011-2: 2001
ASME Code
ASME B31.3-2002 edition
ANSI!AWS D.1.1: 2002
Class 1 arc welding of ferritic steel
pipework for carrying fluids
Specification for fusion welding of
steel castings
Recommendations for welding of
metallic materials. Arc welding
of ferritic steels.
Section VIII: Division 1: 2003 ASME
Boiler and pressure vessel code
Petroleum refinery piping
Structural welding code - steel
In general, these specifications contain less detailed guidance than
the present book and care is sometimes needed in their use. The
3
Defining the problem
information presented in this book may thus be regarded as supplementing and amplifying the recommendations outlined in current
standards.
Hydrogen-induced cracking in welds
Hydrogen-induced cracking is also known as cold cracking, delayed
cracking or underbead cracking. Unfortunately, the term 'hydrogeninduced cracking', usually abbreviated as HIC, has been introduced
to designate cracking sometimes encountered in pipelines or vessels
as a result of hydrogen picked up in service; the term 'fabrication
hydrogen cracking' is therefore preferred for the subject of this book
whenever there may be any doubt. It also occurs in steels during
manufacture, during fabrication and in service. It is thus not confined to welding, but when it occurs as a result of welding the
cracks are sited either in the HAZ of the parent material or in the
weld metal itself. Brief descriptions are followed by a discussion of
the factors which are responsible for cracking and of the means by
which control may be achieved.
Cracking in the HAZ
Hydrogen-induced cracking occurs when the conditions outlined in
1-4 (below) occur simultaneously
Hydrogen is present to a sufficient degree
This is inevitably present, derived from moisture in the fluxes
used in welding and from other sources. It is absorbed by the
weld pool and some is transferred to the HAZ by diffusion.
2 Tensile stresses act on the weld
These arise inevitably from thermal contractions during cooling
and may be supplemented by other stresses developed as a
result of rigidity in the parts to be joined.
3 A susceptible HAZ microstructure is present
That part of the HAZ which experiences a high enough temperature for the parent steel to transform rapidly from ferrite to
austenite and back again produces microstructures which are
usually harder and more susceptible to hydrogen embrittlement
than other parts of the HAZ. Hydrogen cracks, when present, are
invariably found in these transformed regions.
4 A low temperature is reached
The greatest risk of cracking occurs when temperatures near
ambient are reached and cracking may thus take place several
1
4
Welding steels without hydrogen cracking
1.1 Hydrogen-induced cracks
in HAZs of (a) fillet and (b) butt
welds.
(b)
(a)
hours after welding has been completed. Cracking is unlikely to
occur in structural steels above about 150°C. and in any steel
above about 250°C.
Cracks in the HAZ are usually sited either at the weld toe. the
weld root or in an underbead position. These positions are shown
schematically for fillet welds and butt welds in Fig. 1.1. In fillet
welds. HAZ cracks are usually oriented along the weld length. but
in butt welds subsurface cracks can be transverse to the weld.
Hydrogen cracks examined in sections of a weld under the microscope may be intergranular, transgranular or a mixture with respect
to the transformed microstructures in which they lie.
Intergranular cracking is more common in harder. higher carbon
and more highly alloyed steels. Cracks may vary in length from a
few microns to several millimetres. Some typical HAZ cracks are
shown in the photomacrographs, Fig. 1.2.
1.2 Heat-affected zone in
C-Mn steel (a) hydrogen crack
at root of single-run fillet weld.
and (b) crack at toe of multipass
fillet weld.
Cracking in the weld metal
Hydrogen cracking can occur in the weld metal as well as in
the HAZ. Weld metal hydrogen cracks can be orientated longitudinally or transverse to the weld length, while in the transverse
orientation they can be either perpendicular or angled, typically at
approximately 45° (often referred to as chevron cracks), to the weld
surface. The cracks may be buried or may break the weld surface.
Under the microscope they are usually recognised as being predominantly transgranular, although in more alloyed deposits there
is an increasing proportion with intergranular morphologies. Typical
examples may be seen in Fig. 1.3.
The same factors which influence the risk of cracking in the HAZ
5
Defining the problem
also apply to weld metal, namely, stress, hydrogen, susceptible
microstructure and temperature. However, weld metal hydrogen
cracking can occur at much lower levels of weld metal hardness
than is generally the case for HAZ cracking.
Factors responsible for cracking and their control
The interaction between the factors responsible for cracking and
the ways in which control over them may be achieved can now be
discussed.
Hydrogen level
1.3 Weld metal hydrogen
cracks in (a) single-run manual
metal-arc fillet weld, (b) root
bead of Y groove welding test,
(c) submerged-arc weld
(longitudinal section).
During welding, hydrogen is absorbed by the weld pool from the
arc atmosphere (Fig. 1.4). During cooling, much of this hydrogen
escapes from the solidified bead by diffusion but some also diffuses
into the HAZ and the parent metal. The amount which does so
depends on several factors such as the original amount absorbed,
the size of the weld, the decreasing solubility and the timetemperature conditions of cooling (Fig. 1.5).
In general, the more hydrogen present in the metal the greater the
risk of cracking. Control over this hydrogen level may be achieved
either by minimising the amount initially absorbed or by ensuring
that sufficient is allowed to escape by diffusion before the weld
cools. Frequently a combination of both measures provides the best
practical solution. The principal sources of hydrogen in welding
consumables are:
Moisture in the coating of manual metal-arc electrodes, in the
flux used in submerged-arc welding or in flux-cored wires.
2 Any other hydrogenous compounds in the coating or flux.
3 Oil, dirt and grease either on the surface or trapped in the
surface layers of welding wires.
4 Hydrated oxide, e.g, rust, on the surface of welding wires.
1
The principal sources of hydrogen from the material to be welded
are:
1 Oil, grease, dirt, paint, rust, etc, on the surface and adjacent to
the weld preparation; these can break down to produce hydrogen
in the arc atmosphere.
2 Degreasing fluids used to clean surfaces before welding may
likewise break down to produce hydrogen.
3 Hydrogen from the parent steel, either remaining from the
original casting process (particularly in the interior of heavy
6
1.4 Amount of hydrogen
absorbed by molten weld pool
varies with concentration in
atmosphere surrounding arc.
Solubility at 1900°C.
Welding steels without hydrogen cracking
10
~
1;
l
l
5
.S
c
~
..,e
>,
::t
o
5
10
Hydrogen in arc atmosphere. %
1.5 Solubility of hydrogen in
weld metal decreases as
temperature falls.
25
~
fI""'"
Liquid Fe
20
1;
.~
:s
15
;11
:I
~
..,~
10
VYFe
>,
::t
5
:/~
.. 41'
o
500
./
a Fe
1000
1500
2000
Temperature. °c
sections), following service at high temperature and high hydrogen partial pressures, or as a result of corrosion processes, particularly sour [i.e, HzS) service.
In addition a small contribution to the total hydrogen pick-up
may arise from moisture in the ambient atmosphere; here absolute,
rather than relative. humidity appears to be important.
7
Defining the problem
The moisture in fluxes may be present as absorbed water, loosely
combined water of crystallisation or more firmly bound molecules
or trapped hydroxyl ions in the silicate structure.
All these forms can break down to produce hydrogen, and different materials and methods of manufacture produce different
moisture levels. The more firmly the moisture is bound the higher
the predrying treatment must be to remove it. The hydrogen generated by dirty wire and plate originates from drawing lubricants,
soaps, pickling and plating fluids, and hydrated oxides, e.g. rust,
persisting from earlier forming processes.
Direct laboratory measurement of the moisture or hydrogen level
of any consumable produces a result which is termed the potential
hydrogen level of the process. This term has been chosen because it
is known that not all the measured hydrogen is absorbed by the
weld pool. It is potentially available and, in general, the higher the
potential hydrogen level the higher will be the actual weld hydrogen. Other factors, however, may affect the extent to which potential
hydrogen appears as weld hydrogen. These can include resistance
heating of the wire stickout in continuous wire welding processes,
current type and polarity for submerged-arc welding, and the effect
of CO2 generation from carbonates in fluxes reducing the partial
pressure of hydrogen in the arc atmosphere.
The most commonly adopted way of characterising the hydrogen
level introduced by a given consumable is to make a measurement
of the weld hydrogen level, as described in Appendix B. These
measurements not only provide help in determining whether or not
the consumables used are introducing the minimum amount of
hydrogen to the weld pool but they also provide one of the input
parameters to the selection of welding procedure described in
Chapters 3 and 4. In addition, weld metal hydrogen measurements
provide a starting point for calculating times and temperatures for
removing hydrogen after welding (see Chapter 5). Typical hydrogen
levels for different consumables and welding processes are considered in Appendix A.
Control over the hydrogen potential of welding consumables
once they are received from the manufacturer depends on the conditions under which they are stored and used. Storerooms should
therefore be dry and warm to minimise moisture pick-up by electrodes and fluxes, and care should be taken to see that the welding
operation does not put oil, grease, moisture, etc, on to the welding
wire or moisture into electrode coatings and fluxes.
Hydrogen potential figures should be used with caution when
welding is carried out in hot, humid conditions. Such situations
8
Welding steels without hydrogen cracking
can lead to weld hydrogen levels somewhat higher than the standardised conditions under which most consumable manufacturers
assess their products. It is likely that the absolute, rather than the
relative, humidity is the controlling factor. Recent work with one
type of electrode designed to give diffusible hydrogen contents
below 3mLlI00g, typically gave 2mLlI00g in tests welded at
ambient conditions of 20°C and 40% RH (relative humidity).
However, tests welded under ambient conditions of 35°C and 95%
RH not only gave diffusible hydrogen values of about 3.8-4.8mLl
100g, but the electrode picked up sufficient moisture in a few hours
exposure to these conditions roughly to double these levels,
although exposure at 20°C/40% RH and 26.6°C/80% RH gave
negligible increase in weld hydrogen contents after 9 hours
exposure.'
The hydrogen potential of electrode coverings and fluxes can be
lowered in many cases by drying or baking, but the manufacturers'
recommendations should always be observed, and advice sought if
high baking temperatures are contemplated. In most electrode types,
exceeding the manufacturers' drying temperatures will increase the
fragility of the cover. In the case of basic coverings, over-heating can
also lead to loss of shielding and alloying elements, so that the resultant weld metal may not only be porous but may also not be of the
correct composition and properties. With other types of electrode,
the resultant loss of hydrogen (either as moisture or oxidised
cellulose) is likely to alter the welding characteristics, and particularly reduce penetration, especially with the cellulosic type,
which relies on the cellulose in the cover for its special welding
characteristics.
Most manufacturers now supply manual electrodes dried to very
low hydrogen levels and hermetically sealed in vacuum or dry argon
packaging. If the packaging remains intact (and it is obvious when
this is not the case), these electrodes can be used without further
drying, provided they are utilised within the period stated by the
manufacturer after opening, and provided the conditions of operation are within those specified by the manufacturer. Use of such electrodes is facilitated by recent improvements in the resistance of basic
electrodes to pick-up of moisture, and also by the use of small
packets containing only the number of electrodes which a welder is
likely to use in a shift or half a shift.
There have been moves in recent years to add a new, 'ultra-low',
hydrogen level. It has been suggested that this should cover consumables giving diffusible hydrogen levels below 3m1l100g
deposited metal. Indeed, the recently published EN 1011-2,
9
Defining the problem
'Welding - Recommendations for welding of metallic materials Part 2: Arc welding of ferritic steels' does include such a level.
The satisfactory use of such a level would require that any
possible increase in weld hydrogen levels as a result of welding in
hot, humid climatic conditions, which is likely to be more noticeable than with other types, is taken into account.
Stress level
1.6 An increase in size of root
gap raises the stress imposed
on the weld and causes
cracking even when hydrogen
levels and welding conditions
are held constant.
Stresses are developed by thermal contraction of the cooling weld
and these stresses must be accommodated by strain in the weld
metal. The presence of hydrogen appears to lower the stress level at
which cracking will occur. In rigid structures the natural contraction stresses are intensified because of the restraint imposed on the
weld by the different parts of the joint. These stresses are concentrated at the toe and root of the weld and also at notches constituted
by inclusions and other defects. The higher degrees of strain which
result produce higher risks of cracking for a given microstructure.
Alternatively, it can be considered that the same risk of cracking
exists but at a lower microstructural susceptibility.
The stress acting upon a weld is a function of weld size, joint
geometry, fit-up, external restraint and the yield strengths of the
parent steel and weld metal. At the moment, full quantification of
these factors is not possible, although in the case of fit-up it has been
found that the effect of root gaps of O.4mm and greater is to increase
markedly the risk of cracking (Fig. 1.6). Thus, in a joint where geometry and material have already been decided, stresses may be
reduced by using good fit-up and by selecting the lowest strength
weld metal allowable by the design. Hydrogen embrittlement is
strain-rate dependent and the risk of cracking is greatest at slow
strain rates. As the straining rate is, or course, low during the final
stages of cooling in the weld, the susceptibility to crack formation
is high at this time. (It should be noted that hydrogen embrittlement
is not normally revealed by high strain rate tests, i.e. impact tests
such as the Charpy test.)
Stress applied externally to a weld soon after completion, for
example during the lowering-off operation in pipeline welding, will
often be additional to that described above and may therefore temporarily increase the risk of cracking.
Type of microstructure
The parent metal adjacent to the weld is known as the HAZ and can
be seen clearly on a polished and etched section through the
10
Welding steels without hydrogen cracking
weld. This zone is raised to a high temperature during welding and
subsequent rapid cooling (quenching) by the surrounding parent
metal causes hardening. Close to the fusion boundary, the HAZ is
raised to a sufficiently high temperature to produce a coarse grain
size. This high temperature region, because of its coarse grain size,
is not only more hardenable but also less ductile than regions
further from the fusion boundary. It is thus the region in which the
greatest risk of cracking exists. As a general rule, for both carbonmanganese and low alloy steels, the harder the microstructure the
greater is the risk of cracking. Soft microstructures can tolerate
more hydrogen than hard before cracking occurs.
Soft microstructures can be obtained by using steel with low
contents of carbon and alloying elements (including manganese) to
reduce the hardenability of the HAZ. Additionally, the use of a
large weld bead, thin plate, and preheat will reduce the quenching
rate in the HAZ. After a bead has been deposited the HAZ can be
softened by tempering either as a result of subsequent weld runs or
by a postweld heat treatment (PWHT or stress relief).
It is possible to propose certain arbitrary critical hardness levels
below which a low risk of cracking exists. If welding procedures
can be selected in such a way that the resulting HAZ hardness
values (before reheating by any subsequent weld run) do not exceed
these critical levels, a basis for deriving welding procedures exists.
Any critical hardness level chosen is dependent on the hydrogen
An increase in HAZ
hardness level raises degree of
embrittlement but by different
amounts for different steel
types.
1.7
HAZ hardness
11
Defining the problem
level of the welding process used and the stresses which are operating and, although more data are needed, this concept can be used
in welding diagrams. Nonetheless, it must be recognised that the
tolerance to hydrogen at a given microstructural hardness varies
with steel type as shown in Fig. 1.7 while, even within the carbonmanganese family, critical hardness is influenced by composition,
decreasing at lower carbon contents.
The microstructure produced in any steel is essentially dependent
upon:
the
the
2 the
3 the
1
cooling rate through the transformation temperature range of
steel in question,
composition and the hardenability of the steel, and
(prior austenite) grain size before transformation.
The cooling rate is governed by the heat supplied during welding,
the initial temperature of the parts to be joined, their thickness and
their geometry. In arc welding, the heat supplied during welding is
characterised by the heat input which is defined throughout this
book as: Heat input = arc volts x welding current x arc efficiency/
welding speed (see Chapter 2). Control over cooling rate in a particular fabrication is therefore achieved by varying heat input and
preheat temperature.
The hardenability of the steel is governed by its composition, and
a useful way of describing hardenability is to assess the total contribution to it of all the elements that are present. This is done by
1.8 For each steel
composition, expressed in the
form of CE level, the hardness
of the microstructure depends
on weld cooling rate.
350 HV
'il
~
c
i
l
~
t---f----+---.:
Fast
Cooling
---------<.~
Cooling time
Slow
Cooling
12
Welding steels without hydrogen cracking
an empirical formula which defines a carbon equivalent (CE) value
and takes account of the important elements which are known to
affect hardenability, The formula used in this book is:
Mn Cr + Mo + V Ni + Cu
+--CE = C + - +
6
5
15
[1.1]
Its calculation and use are described in detail in Chapter 3. The
cooling rates which produce different microstructures of different
hardnesses are established by laboratory studies of each steel type,
using the cooling rates which the steel experiences during welding.
For carbon-manganese steels, a relationship between composition
(CE value), cooling rate and microstructural hardness level has
been established. This relationship, shown in Fig. 1.8, has been
used in constructing welding diagrams for these steels.
Several other carbon equivalent formulae have been proposed
from time to time.f Most of these have not been extensively used,
because of their unfamiliarity, their excessive complexity, or because
they were so close to the formula, [1.1] (frequently known as the
IIW formula), that there was no significant advantage in their use.
However, one formula developed in Japan for steels of low carbon
content, whose behaviour with regard to hydrogen cracking is not
well described by the IIW formula, is the Pcrn formula:
Si Mn Cu Ni Cr Mo V
Pcrn = C + 30+20+ 20 + 60+ 20+ 1"5+ 10+ 5B
[1.2]
Compared with the IIW formula, Pcrn gives an increased importance to carbon and adds the microalloying element boron. The
value of the boron factor is large, because boron is a light element
only added to steels in small quantities, but should be regarded
with some caution, because its use depends on the boron being in
an active state; some workers also think that the value of 5 is too
low and that a value of 10 may be nearer to the true value for active
boron.
The position with respect to weld metal microstructures is rather
more complex, because the toughness of weld metal (without
hydrogen present) does not depend on hardness in the same way as
does HAZ toughness. A tough weld metal microstructure-appears to
be better able to resist hydrogen cracking than one which is less
tough. However, toughness in weld metals can either be achieved
with a soft weld metal (softer than about 195 HV) or by a harder
weld metal with a fine acicular ferrite microstructure. Details of the
different types of microstructures in ferritic steel weld metals can
be found in Reference 3.
13
Notch tensile strength of
steel containing hydrogen
passes through minimum value
close to room temperature.
Defining the problem
1.9
H free, 0.5 mm/min
_
2
175
c~arHged,
~E
~
~
c
~
~
.@
•
."."
\
150
H 2 charged,S mmlmin / . . .
"+ ""........
0.5 mm/min
\.
"
\A
-,
~~
\\
125
..c
H2 charged,
0.05mmlmin
B
o
Z
..
\
100
-200
-100
..............
....-If
-.I
/
*
-II-~/.
•
I
AI
o
100
200
Temperature, °C
Temperature
Hydrogen embrittlement of ferritic steels occurs only at low temperatures, close to ambient, as shown in Fig. 1.9. It is therefore
possible to avoid cracking in a susceptible microstructure by maintaining it at a sufficiently high temperature by postheating until sufficient hydrogen has diffused away. The microstructure can also be
softened by tempering, to render it less susceptible, and this principle is employed in multipass welding and when using PWHT.
Provided the HAZ has completed its transformation from austenite, an increase of temperature increases the rate of diffusion of
hydrogen sufficiently to accelerate its removal from the weld. This
effect is particularly marked in the range 20-150°C, as shown in Fig.
1.10. Any measure which slows down the weld cooling rate is therefore helpful in reducing the hydrogen level. Preheat, for example,
by slowing the cooling rate, not only softens the microstructure but
also helps hydrogen to escape. As a result, higher HAZ hardness
levels can be tolerated without cracking than if preheat had not been
used.
For welds in those steels with hardenability so high that soft
microstructures cannot be produced at all, and where preheat cannot
remove sufficient hydrogen, a weld interpass temperature, or a
postheating temperature, high enough to avoid cracking must be
held for a sufficiently long time to allow hydrogen to diffuse
14
Welding steels without hydrogen cracking
1.10 Diffusion rate of
hydrogen through ferritic steel
decreases as temperature falls.
20
150
Temperature. "C
600
away before the weld cools. The times and temperatures for hydrogen removal from different thicknesses are dealt with in detail in
Chapter 5.
If PWHT cannot be employed, it is possible, in principle, to
soften the uppermost parent metal HAZ with a temper bead. This
method should not be used unless careful control can be exercised.
The temper bead is an extra weld run deposited so that its toe is a
small, fixed distance (typically 3 mm) from the fusion boundary
and it tempers the uppermost parent metal HAZ without creating a
new hardened region.
Although it is normally recommended that steels should be heated
above O°C for welding, mainly to avoid condensation of water
vapour on to the cold steel, it is possible to weld steel at very low
temperatures. A major disadvantage is that hydrogen takes very
much longer to diffuse out of the HAZ and weld metal. This means
that the normal time delay before inspection of about two days (to
ensure that any cracks develop before, instead of after, inspection)
must be extended considerably. It is also possible for a weld
deposited at a very low temperature to suffer hydrogen cracking
only after it has been brought to a temperature above O°C, particularly if it is stressed at the time.
15
Defining the problem
Detection and identification
Hydrogen cracks are normally very fine and difficult to detect, even
when they break the surface, which they often do not. For cracks
which are present at the surface, magnetic particle inspection (MPI)
is the preferred technique, although dye penetrant should be used
if a non-ferritic filler (e.g. an austenitic stainless steel or a nickel
alloy) has been used. The MPI technique has the advantage that
when using DC, cracks just below the surface can be detected.
For buried cracks, ultrasonic examination is greatly preferred to
radiography. Radiography is only capable of detecting cracks that
are relatively wide and are almost parallel to the incident beam.
However, even the use of ultrasonics is not always successful. If a
crack has formed in the root region of a two-sided multipass butt
weld, the residual stress pattern in the root on completion of
welding will usually be such that this region is in compression and
this will tend to force together the two faces of any cracks, making
them very difficult to detect until the weld has been give a PWHT,
thus relaxing the stresses. If it is suspected that transverse 45° weld
metal cracks (chevron cracks) may be present (Fig. 1.3(c», it is necessary to carry out an ultrasonic examination using a 45° probe in
both directions along the top of the weld seam. The usual 90° probe
is not capable of detecting such cracks.
A further common reason for not detecting cracks is when inadequate time is left for cracks to grow to a detectable size before examination. It is still uncertain how long an incubation period is
required for hydrogen cracks to start, and how long they can continue growing. Several periods have been proposed, varying from
overnight (16 hours) to 3 days (72 hours). There is some evidence
that when a plentiful supply of hydrogen is available, cracks start
forming fairly soon after the weldment has cooled to near ambient
temperature and grow to a detectable size within a few hours. That,
however, is not the problem area. The greatest delay probably occurs
when a weld is being made close to the crack/no-crack boundary so
that appreciable time is needed for hydrogen to diffuse to the sites
of cracking in quantities. In very cold conditions, i.e, below O°C,
longer times are necessary; above about 30°C it might be possible to
offer some relaxation in delay time. Further information on delay
times can be found in reference 4.
The difficulties of detecting buried cracks in as-welded joints
make it particularly important to use safe welding procedures, as
described subsequently in this book. The difficulties can also lead
to incorrect identification of the type of crack which has been
16
Welding steels without hydrogen cracking
detected: cracks found after PWHT tend to be classified as reheat
(stress relief) cracks, even if the steel is such as not to be susceptible to this type of cracking.
The identification of hydrogen cracks is not simple, as they can
occur in a variety of locations and orientations, and can be transgranular or intergranular, the latter being more likely if the steel or
weld metal is of the alloyed type, or exceptionally hard. In the HAZ,
hydrogen cracks are usually longitudinal to the weld (unless they
are extensions of transverse weld metal cracks) and usually have a
portion close to the fusion boundary. However, the cracks may divert
into the fine-grained HAZ, as in Fig. 1.2(b), or into the weld metal.
Crack surfaces in C and C-Mn steel HAZs are usually transgranular, being of a quasi-cleavage type, although some ductile tearing
(microvoid coalescence) may also be present. In alloyed steel HAZs,
intergranular cracking is the common type. Cracks will almost never
extend beyond the visible HAZ unless they have been heavily
stressed after welding: without post-weld stressing, cracks extending beyond the visible HAZ are likely to be lamellar tears. Unlike
reheat cracks, hydrogen cracks never show cavitation along the prior
austenite grain boundaries beyond the crack tips.
Similar comments are applicable to weld metal hydrogen cracks,
with the further observations that cracking in weld metals is frequently transverse to the weld and that in weld metals of the acicular ferrite type, cracking is most common in the primary ferrite (if
it is present) at the prior austenite grain boundaries, as in Fig. 1.3(a).
In relatively hard, alloyed weld metals (e.g. in the Cr-Mo steels) any
transverse weld metal cracking is usually perpendicular to the weld
surface, but in C-Mn weld metals the chevron type of cracking (Fig.
1.3(c)) at 45°C to the surface is common.
REFERENCES
1 Dickehut G et al: Oerlikon Schweissmitteilungen 199048 (125) 10-14.
2 Bailey N: Proc Symp 'Hardenability of steels'. Derby, TWI, May 1990.
3 Pargeter R J and Harrison P L: 'Compendium of weld metal microstructures and properties'. IIW Doc, 1985.
4 Pargeter R J: Welding Journal rAWS) 200382 (11) 321s-3298 and erratum
83 (1) ISs.
Chapter 2 Guidance on safe welding
procedures by graphical methods
Shall they who wrong begin yet rightly end?
Shall they at all have Safety for their friend?
This chapter describes the principles involved in using welding
diagrams to derive welding procedures for the avoidance of HAZ
cracking. The diagrams are introduced schematically and are given
in detail in Chapter 4. They have been constructed on the assumption
that they must be capable of indicating welding procedures which
will not lead to cracking, even in situations where the worst possible combinations of circumstances are encountered. For this
reason there will be many occasions when circumstances will not
be at their worst and a procedure less stringent than that indicated
by the diagram will be successful. The particular diagram to be
used depends largely on the type of steel to be welded. These
different steel types are listed in detail in Chapter 4 but basically a
division between those of low hardenability and those of high
hardenability is used. A CE level of 0.60 (see Chapter 3) is used as
the division between these two groups.
Low hardenability steels
The prediction method for low hardenability steels aims to avoid
hard HAZ microstructures. This is achieved by selecting welding
conditions which provide a sufficiently slow weld cooling rate. The
cooling rate that is necessary depends on the composition of the
steel and the amount of hydrogen introduced into the HAZ from
the weld metal. Although the thickness of the sections to be joined
does not influence the cooling rate needed to achieve a particular
microstructure, it does play an important role (as the 'combined
thickness') in defining the welding and preheat requirements to
achieve that cooling rate.
All these parameters are linked in a diagram of the type shown
schematically in Fig. 2.1. On the full diagram there are many lines
referring to different preheat temperatures and combined thick-
18
2.1 Prediction diagrams bring
together information on
material composition (CE
level), plate thickness, and joint
geometry (combined thickness)
so that preheat levels and heat
inputs can be selected for
successful welding.
Welding steels without hydrogen cracking
,-~~
~
/
V
""~ /
»rIPr>
J-
~tj
~
~
~
,;'
'"
~
/
,
V
~
)
\
V·#
vo~
:!:J"
'
i--:
"",."".
.... V
r
•
Heat input
Carbon equivalent
Basic diagram shown in
Fig. 2.1 amplified by adding
further CE axes to deal with
other steel compositions and
hydrogen levels and then
referring to tables linking
electrode size and weld run
length to heat input scale.
2.2
Combined thickness
4
Refer to Tables
2.1 - 2.4
Carbon equivalent
nesses, but for clarity only two of each are shown in Fig. 2.1. The
broad arrow shows how, starting with the CE value, preheat temperatures and heat input levels can be selected for a particular
combined thickness. The diagram can of course be traversed in the
19
Guidance on safe welding procedures by graphical methods
opposite direction, or from both sides to meet in the centre when
restrictions have to be placed on bead size and preheat level.
The same diagram is shown in slightly greater detail in Fig. 2.2 so
that the various steps required in its use can be illustrated, as
follows:
Step 1 Decide on CE axis appropriate to hydrogen level of process,
joint type, etc (Chapter 3 and Table 3.2).
Step 2 Decide steel composition, calculate CE value (Chapter 3)
and erect vertical into preheat portion of diagram.
Step 3 Decide combined thickness of joint in question (Chapter 3).
Step 4 Decide limitations on heat input, bead size, or electrode size
which can be used (Chapter 3). These limitations may arise
because of positional welding or because of a need to
achieve minimum toughness levels in weld metal or HAZ.
Step 5 Trace horizontal line to obtain required preheat level.
It may also be necessary to decide limitations on the level of
preheat and interpass temperature which can be used; for example,
welding manually within an enclosed space may preclude high
temperatures.
In Fig. 2.2 a heat input limitation has been shown and a vertical
line from this scale meets the particular combined thickness chosen.
From this point, horizontal movement meets the vertical from the
CE scale to identify the line TT, which gives the minimum preheat
temperature required. It should be noted that in making this construction the electrode size and run length are related to heat input
by means of Tables 2.1-2.4. Bead sizes less than this maximum may
be employed but only at the cost of increasing the preheat temperature and this can be checked in the diagram.
The term electrode efficiency (not to be confused with arc efficiency) in Tables 2.1-2.4 relates to the weight of metal deposited
from unit weight of electrode core wire; efficiencies over 100% refer
to electrodes whose coverings contain appreciable proportions of
iron powder which are incorporated into the weld pool.
If these are limits to the preheat temperature that can be used and
to maximum bead size a double construction is necessary, as shown
in Fig. 2.3. Only those heat inputs, electrode diameters, and run
lengths indicated by the arrows are permissible and then only at the
corresponding preheat temperatures which are shaded.
In some cases it is possible for a limitation on preheat to require a
minimum bead size which is larger than that which can be produced, particularly when welding in the overhead position. In such
circumstances the diagram indicates that it is necessary to use a
2.5
3.0
3.5
4.0
4.5
5.0
5.5
2.2
2.0
1.6
1.8
1.4
1.0
1.2
0.8
Heat input
kl/rnm
130
105
85
2.5mm
215
170
145
120
105
95
85
3.2mrn
335
270
225
190
165
150
135
120
105
90
4mrn
525
420
350
300
260
230
210
190
165
140
120
105
95
85
5mm
Run out length from 410mm of a 450mrn electrode of diameter:
Table 2.1 95% < electrode efficlency s 110%
600
500
430
375
335
300
275
240
200
170
150
135
120
110
6mm
555
475
415
370
330
300
265
220
190
165
150
135
120
-
6.3mm
Tables 2.1-2.3 show run out length for manual metal arc welding. Tables from BS EN 1011-2: 2002 are
reproduced with the permission of BSI under licence number 2003SKlo159. British Standards can be
obtained from BSI Customer Services, 389 Chiswick High Road, London W4 4AL. (Tel + 44 (0) 208996
9001).
-a
Ii':'"
~.
('l
Cl
('l
=
..,
~
~Q.
-g
~
;:
'IJ
('l)
('l)
'IJ
~.
Q.
-~
N
Q
0.8
1.0
1.2
1.4
1.6
1.8
2.0
2.2
2.5
3.0
3.5
4.0
4.5
5.0
5.5
Heat input
kl/rnm
-
-
-
-
-
-
250
200
165
140
125
110
100
90
150
120
100
85
-
3.2mm
2.5mm
-
-
-
385
310
260
220
195
170
155
140
125
105
90
4mm
605
485
405
345
300
270
240
220
195
160
140
120
110
95
90
5mm
Run out length from 410mm of a 450mm electrode of diameter:
Table 2.2 110% < electrode efficiency s 130%
580
500
435
385
350
315
280
230
200
175
155
140
125
6mm
550
480
425
385
350
305
255
220
190
170
155
140
6.3mm
=
l:lo
Ctl
~
-....
( Il
=
~
=
Q
Ctl
("l
(Il
l:lo
Q
~
!3
("l
"t:l
-
"'l
=
....='"
=
0Cl
'<
C'"
(Il
Ctl
"'l
=
l:lo
Ctl
("l
"'l
Q
"t:l
0Cl
I~
l:lo
....
=
o
N
...
4mm
500
400
330
285
250
220
200
180
160
135
115
100
90
3.2mm
320
255
215
180
160
140
130
115
100
85
625
520
445
390
345
310
285
250
210
180
155
140
125
115
5mm
Run out length from 410mm of a 450mm electrode of diameter:
560
500
450
410
360
300
255
225
200
180
165
6mm
620
550
495
450
395
330
285
245
220
200
180
6.3mm
where
L
is the consumed length of electrode (in mm) (normally the original length less 40mm for the stub end)
and
F
is a factor in kl/rnm" having a value depending on the electrode efficiency, as follows:
F = 0.0408
95% < efficiencv s 110%
F = 0.0472
110% < efficiencv s 130%
F = 0.0608
efficiency> 130 oio
Run out length (rnm) = (Electrode diameted x Lx F
Heat input
Note: The values given in the above tables relate to electrodes having an original length of 450mm. For other electrode lengths
the following expression may be used:
1.8
2.0
2.2
2.5
3.0
3.5
4.0
4.5
5.0
5.5
1.6
0.8
1.0
1.2
1.4
Heat input
kl/rnm
Table 2.3 Electrode efficiency> 130%
-
~.
~
o
=
o"S
=
~cc
Q..
~
-
~
=
~
( Il
if
(Il
IJQ
s:S·
~
N
N
23
Guidance on safe welding procedures by graphical methods
Table 2.4 Values of heat input for single run fillet welds (These values are to be
used only when fillet welds to a specified minimum leg length are required. In
other cases heat input is controlled by specifying electrode run lengths - Tables
2.1-2.3)
Minimum leg
length for
single run
fillet weld
mm
Manual metal-arc welding
Heat input for electrode efficiencies
"'110%
kl/rnrn
4
5
6
8
10
0.8
1.1
1.6
2.2
3.0
2.3 For a given steel and joint
geometry there is still some
flexibility when matching
preheat levels to welding
conditions.
>110%
"'130%
>130%
kl/rnm
kl/rnm
1.0
1.4
1.8
2.7
4.0
0.6
0.9
1.3
1.8
Tubular
cored arc
welding
kl/rnm
1.4
1.6
1.8
2.2
3.0
Combined thickness
Min
Max
~-*--+--t---t-----I""""'--
Heat input
Refer to Tables
2.1-2.4
----~
Carbon equivalent
process or a consumable of lower hydrogen level and thus move to
another CE axis.
Attention is drawn to the value of joint simulation testing as a
means of confirming selected procedures; this is discussed at the end
of the chapter.
24
Welding steels without hydrogen cracking
High hardenability steels
These steels form hard HAZ microstructures even at the slowest
cooling rates achieved in welding. There are three methods of
deriving welding procedures:
1 Temperature control
2 Isothermal transformation
3 The use of austenitic or nickel alloy weld metals
In general the first is more suitable for steels with lower carbon
contents (below 0.3%) and the second method is appropriate to
higher carbon steels. The third method is useful where circumstances
preclude the use of high preheating temperatures.
(1)
The temperature control method
This method depends on holding the weld at an elevated temperature,
in particular above that at which hydrogen cracking occurs, so that
removal of hydrogen by diffusion is accelerated. The temperature is
raised as a preheat, maintained during welding by specifying a
minimum interpass temperature, and, where further hydrogen
reduction is necessary, held after welding as a postheat. The temperature is selected by using a diagram of the type shown schematically in Fig. 2.4. From the carbon content an estimate of expected
HAZ hardness can be made for different steel types using the lower
half of the diagram. For each estimated hardness the preheat, inter-
2.4 For hardenable steels the
carbon content is used to select
preheat, interpass and postheat
temperatures.
.A'
I
I
I
High restraint
v:
~
/
/
.'
~
~
/
/
Low restraint
»< ~
I
I
Expected HAZ hardness
I
Steel type
.
/"
V
./
VI/;
V
/" /
....
~
./
3
25
Guidance on safe welding procedures by graphical methods
pass and postheat temperature is given in the upper part of the
diagram. The temperatures given schematically are those at which
fully hardened microstructures do nut crack at high hydrogen levels.
The temperature also depends on the restraint of the particular
joint and the hydrogen level; at present, this cannot be described
numerically. Judgement and previous experience must therefore be
applied when reading from these diagrams, but if no information
exists, joint simulation tests should be made.
The decision to maintain weld temperature as a postheat, together
with an estimate of the time for which it must be held, must be
based on a consideration of hydrogen concentration at the end of
welding and some critical concentration below which cracking will
not occur when the weld cools to ambient temperature. At present
there is little information on these critical concentrations although
it is possible to make cautious estimates. These estimates can then
be used to calculate postheating times for different joint thicknesses
and temperatures. Diagrams of the type shown schematically in Fig.
2.5 are available for this purpose and their use is described in detail
by means of worked examples in Chapter 5. The amount of hydrogen
removed (expressed as a percentage of the initial concentration) by
different holding times at a given temperature can be estimated for
different plate thicknesses as shown by the arrows in Fig. 2.5.
Maximum and minimum estimates are given in the diagram because
Extent of hydrogen
removal occurring during
different postweld heat
treatments can be estimated
from diagrams of this type;
temperature, time, and material
thickness are related in such
diagrams.
2.5
90 ....-
__
Max.
o
Time at temperature
-+ :
I
Nomogram other thickness
I
I
t7
26
Welding steels without hydrogen cracking
the rate of hydrogen diffusion depends on steel composition. Advice
on narrowing these limits by selecting the appropriate diffusion
rate for hydrogen is also given in Chapter 5.
As well as indicating times for postheating, these diagrams can
be used to assess any advantages (in terms of hydrogen removal) to
be gained by extending the interpass time at a controlled temperature
and/or reducing the weld bead size to reduce the diffusion path for
hydrogen. When welding very thick joints in particular, an extension
of interpass time can produce lower final hydrogen levels more
quickly than a long postheating time for the completed weld.
Using these procedures it should be possible to avoid hydrogen
cracking, but it must be remembered that the HAZ will still be hard
and its fracture toughness may be poor. The risk of stress-corrosion
cracking in service may also exist depending on the actual hardness
level and the service environment. At higher carbon levels it will
be advisable to temper (PWHTlthese hard microstructures by heating
typically to the range 550-650°C, provided the original tempering
temperature used is not exceeded. However, for these steels it is
quite likely that, at the selected preheat/interpass/postheat temperatures, transformation from austenite to martensite may be only
partially complete. This can be determined by reference to transformation diagrams, or the equations for predicting the Ms temperature in Chapter 4. Hence, when tempering is conducted directly
from the postheat temperature; any austenite retained and which
does not transform during tempering, is likely to produce hard
martensite on cooling out to room temperature, and a subsequent
further tempering operation is needed. For some steels this could
be avoided by reducing the temperature at the end of welding to a
lower value which allows nearly complete transformation to martensite to occur, at a temperature where cracking is unlikely, before
proceeding to an immediate tempering treatment. This temperature
to obtain nearly complete transformation should not generally be
lower than -180°C and can be determined, again by reference to
transformation diagrams and/or the equations in Chapter 4.
Joint simulation tests are of particular value in confirming welding
procedures for these steels where there may be insufficient data on
behaviour and critical hydrogen levels.
(2)
The isothermal transformation method
This method is particularly suitable for steels of higher carbon
content showing high hardenability, where it is required to produce
a softer HAZ microstructure than martensite directly after welding
and without recourse to postweld tempering.
27
Guidance on safe welding procedures by graphical methods
2.6 The transformation
characteristics of a steel can be
used to select heating times and
temperatures after welding so
that a relatively soft
microstructure can be obtained.
Start of transformation
Time
•
The method relies on a knowledge of the isothermal transformation
characteristics of the steel and these are usually found in a diagram
of the type shown schematically in Fig. 2.6. The object is to control
the cooling of the HAZ so that it transforms in an approximately
isothermal manner and produces a softer microstructure than martensite. In Fig. 2.6 the arrow superimposed on the transformation
diagram indicates the particular course of transformation to be
followed to produce this softer HAZ and shows that a minimum
preheat, interpass, and postheat temperature of 360°C ± 20°C should
be specified. Unless the transformation diagram has been assembled
from data involving austenitising temperatures higher than 1250°C
to produce a coarse grain size, the minimum holding time after
welding should be at least double that indicated on the diagram for
100% transformation.
It must be noted that, although the microstructures produced in
this way will be softer and tougher than martensite, they will
almost certainly be harder and less tough than the tempered martensitic structures produced by method (1) already described. Hence,
this method is the best to adopt when tempering cannot be used
(because it would perhaps interfere with the condition of other
parts of the fabrication), but, if the best level of toughness is desired,
a tempering treatment should be given. In these circumstances, the
principal advantage of method (2) over method (1) lies in the greater
ease in deciding how to control transformation by comparison with
the difficulty of deciding how much hydrogen needs to be removed.
28
Welding steels without hydrogen cracking
(3) The use of austenitic and nickel alloy weld metal
Where, for various reasons, it is not possible to use preheat temperatures greater than 150 oe, method (1) is severely restricted and
method (2) cannot be used. The only alternative is then to use a
combination of welding process and consumable which virtually
prevents the introduction of hydrogen into the HAZ and which
produces a weld metal insensitive to hydrogen. This is achieved by
the use of austenitic (stainless steel) or nickel alloy electrodes, solid
or flux-cored wires. Although both material types are sometimes
covered by the blanket term 'austenitic', and have face-centered
cubic microstructures (known as austenite when present in steels),
such terminology is incorrect, as there are important differences
between them. For example, nickel alloy electrodes are less prone
to give cracked martensitic regions along the fusion boundary, which
may occur with austenitic consumables. Nickel alloy electrodes,
unless selected to minimise the problem, may suffer solidification
(hot) cracking where austenitic electrodes would be immune; also
suitable nickel alloy consumables are not available to give the high
strength levels (e.g. up to 900N/mm 2 or more) possible with some
of the austenitic stainless steel types.
At ambient temperatures, both austenitic and nickel alloy weld
metals have much higher solubilities for hydrogen, much slower
hydrogen diffusion rates and very low sensitivities to hydrogen
embrittlement and cracking in comparison with ferritic weld metals.
The higher solubility means that once hydrogen has diffused (relatively rapidly) from the HAZ to the fusion boundary, it can easily
enter weld metal which has a high solubility for hydrogen. The
slow diffusion rate means that hydrogen which reaches the weld
metal close to the fusion boundary stays in its vicinity. It is, therefore,
advantageous to use consumables giving low weld metal hydrogen
contents to reduce the chance of local saturation of the weld metal
with hydrogen near the fusion boundary and consequent cracking.
In assessing the hydrogen levels of such consumables it is the
total (not the diffusible) weld metal hydrogen content which
is important and should be measured. This is because hydrogen
analysis of austenitic and nickel weld metals at ambient temperatures
would only measure very small amounts of diffusible hydrogen, as
diffusion rates in these materials are so slow that virtually all the
hydrogen present would be measured as 'residual', although it is
still active and diffusing, albeit at a very slow rate (see, for example,
Fig. 5.17).
With both austenitic stainless steel (and to a lesser extent with
29
Guidance on safe welding procedures by graphical methods
nickel alloy) weld metals, sufficient hydrogen may be left in the
HAZ to give cracking, particularly at the fusion boundary. Here
incomplete mixing often leads to small regions alloyed sufficiently
to transform them to hard martensite on cooling, but not sufficiently
alloyed to remain austenitic.
Normal HAZ cracking is also possible, particularly where the use
of large electrodes (more usual with the stainless steel types) results
in wide HAZs. These give long diffusion distances for hydrogen
which has diffused into the HAZ, when it was austenitic at high
temperatures (or may even have been present in the parent steel), to
diffuse back into the weld metal. Whether in the fusion boundary
or HAZ proper, cracking can usually be prevented by applying
some preheat (normally about 150°C) to increase the HAZ cooling
time and thus allow longer times for hydrogen to diffuse out of the
HAZ while it is still too warm to be embrittled by hydrogen. In all
cases, freedom from cracking will be more easily achieved if consumables giving low total weld metal hydrogen levels are used.
Only general guidance can be given on the choice of preheat
temperatures when using austenitic and nickel alloy fillers, as the
role of total weld metal hydrogen content has not been systematically
studied. The use of joint simulation tests is therefore strongly recommended, with a caution that sectioning and careful metallographic
examination is essential for their evaluation, because no nondestructive examination (NDE) method has been developed which
is capable of reliably detecting hydrogen cracks when non-matching
fillers are used.
The choice of method
One of the problems facing the fabricator at this stage is that of
making a logical choice of method: inspection of the 'flow diagram'
in Fig. 2.7 shows how to make this choice.
Joint simulation testing
Increased assurance that the selected procedure is safe, but not so
safe that it is unnecessarily expensive, can be gained only by carrying
out tests to estimate the margin of safety. Ideally, these would be
made on the fabrication itself, modifying the procedure in a series
of tests to establish the point of maximum safety with minimum
cost, but in most cases this would be an unacceptable exercise on
economic grounds. It is recommended however that, whenever a
high level of assurance is required, tests should be made to establish
30
2.7 Sequence of decisions
leading to final choice of
welding diagram.
Welding steels without hydrogen cracking
alternative action - __
Warning 'A':
In th is case the HAZ will be hard
and may have inadequate toughness
and low resistance to stress-corrosion
crack ing. Postweld tempering could
be em ployed bUI lime at temperature
shou Id be restricted to minimise carbon
migrarion at fusion boundaries.
Low
CE < 0.6
Steel
High
hardenability
CE:;;' 0.6
~
Can temperature,
control
method be used"
I
~
Welding
Possible ~
conditions f(,,~
,
soft HAZ'!
Possible
No
r-;--:~J
,
Can isothermal ~
transformation
method be used? No
Is maximum HAZ ~
Can austenitic or ~
Ni alloy weld
metal be used?
No
Is resistance to Yes fhydrogen-induced I - - No
stress-corrosion
cracking required'!
...
toughness
required'
test
I
~
t
Consult TWI
I
r
&l
I
Docs experience
suggest a
relaxed procedure? No
~
Joinl simulation
lest
~
Fail
I
I
Reconsider use Ves
of austenitic or
Ni alloy filler with No
very low
hydrogen potential
Joint simulation
No,
't'
I
~
'I
Fail
Proceed with isothermal
transformation method;
select lime and
temperature;
see Ch4, pp. 66-67
Proceed with use of
austenitic or Ni alloy
weld metal; see Ch4,
pp. 67-69.
Note Warning
"A' above.
Proceed with
temperature control
method and
tempering; see Ch4,
pp. 63-66
Proceed with
temperature control
method without
tempering,
see Ch4. 63 - 66
I
II
I
Proceed using ,
CM, pp. 48-62
Proceed
I
the margin of safety, and that these tests should reproduce as closely
as possible the effects of restraint, heat sink, and other factors which
occur in the actual fabrication.
An example illustrates a very simple case where a joint simulation test can create a high degree of assurance of success for a procedure which, incidentally, is less stringent than that indicated by
relevant sections of this book. It must be emphasised that this is an
example only and does not constitute a suggestion for general relaxation of recommendations.
Assume a fabricator is required to make a stiffened beam from a
C-Mn steel to BS EN 10 025: 1993 Grade Fe510D1. He has obtained
31
Guidance on safe welding procedures by graphical methods
2.8 General arrangement of
proposed fabrication;
dimensions in mm.
10 mm leg length
6 mm leg length
Table 2.5 Details of welds to be made in joint simulation test illustrated in Fig. 2.8
Joint
Combined thickness, *
mm
Minimum fillet leg length
(single bead), mm
(A) Web-to-flange
(B) Stiffener-to-flange
(C) Stiffener-to-web
75
70
10
6
35
6
* See Chapter 3, and Fig. 4.2.
plate to an agreed restricted CE of 0.43, some of which is at the
maximum of 0.43. He wishes to use the submerged-arc process with
a basic flux for the web-to-flange fillet welds and a basic covered
electrode for the stiffener-to-web fillet and the stiffener-to-flange
fillet. The general arrangement and sizes of the plates and welds are
shown in Fig. 2.8. The welds to be made are therefore as in Table
2.5. Reference to Fig. 4.2 (Chapter 4) and Table 2.4 provides the
recommendations that the welds should be made with the following
preheat temperatures: A: 20°C, B: 60°C, C: 20°C.
Only welds of type B therefore pose any problem for the fabricator.
The fabricator is convinced that he need not use any preheat for
the stiffener-to-flange welds. as he has made similar welds successfully on at least two occasions in the past on steel of this CE level.
Unfortunately, he has no documented evidence of this fact with
32
Welding steels without hydrogen cracking
2.9 Joint simulation test
geometry; the three stiffeners
are numbered in order of
welding.
3
2
which to reassure his customer. He decides that it will be to his
long-term advantage to carry out a joint simulation test and to keep
adequate records of this and successful fabrication as evidence
supporting proposed procedures for future constructions.
However, the fabricator is a cautious man and, although he is
confident that he will need no preheat, he designs his joint simulation
test so that he can include test welds made with some preheat. He
therefore makes the fabricated beam 2 m long, so that he can include
three stiffeners staggered on either side of the beam as shown in
Fig. 2.9. He also selects the plates which have the maximum eE of
0.43.
The first stiffener is welded with no preheat using the preferred
sequence of welding to web and to flange and taking care to allow
each single weld run to cool out just as if it were the first or only
weld run made on that stiffener before the end of the shift.
After a minimum of 48 hr has elapsed to allow any cracking to
develop, the second stiffener is welded using a local preheat of
50 0 e and allowing the area to cool back to the preheat temperature
between each single weld run.
After a further 48 hr the third and last stiffener is welded in the
same manner as the second but using a preheat of 100 o e. After a
further 48 hr the welds are examined, firstly by non-destructive
testing methods and then by selective sectioning to establish whether
cracking has occurred. This examination provides the fabricator
with evidence to justify his use of either no preheat at all or a level
below the recommended temperature of 60 "C. The evidence might.
of course, justify the original prediction of 60 e minimum.
0
Chapter 3 Selecting values for graphical
presentation
Here I have seen things rare and profitable:
Things pleasant, dreadful, things to make me stable
In what I have begun to take in hand.
In Chapter 2 the construction of graphs and nomograms for the
selection of suitable welding procedures for different steel types
was described, but before using the nomograms, which are given in
detail later, it is necessary to know how to select values for:
1
2
3
4
Carbon equivalent level
Combined thickness
Weld metal hydrogen level
Heat input
These values, which vary with each welding problem, provide
points of entry to the nomograms. This chapter provides the basis
for the choice of such values.
It is emphasised that the procedures in this book have been
devised for steels which themselves contain little or no hydrogen,
i.e. appreciably less than 1 ppm. Steels of thick section (particularly
heavy forgings) which have been recently manufactured and not
tempered (or otherwise heated after transformation from austenite)
for long times, steels which have been in service in high pressure,
high temperature hydrogen and steels which have been in sour
service are particularly at risk. Such steels may themselves contain
sufficient hydrogen, particularly away from their original free surfaces (e.g. when forgings have been bored out prior to welding in
the bore) to give rise to cracking when using otherwise safe procedures. Before welding, consideration should be given to heat
treating the steel after rough machining to reduce the hydrogen
level (using the guidance given in Chapter 5), lowering the hydrogen
introduced during welding by the use of consumables of very low
hydrogen potential and/or applying postheating after welding.
34
Welding steels without hydrogen cracking
Chemical composition
A knowledge of steel composition is essential for selecting the
appropriate welding diagram. Where test certificates are available
for the material to be welded, these should be consulted to determine
the CE appropriate to the steel. Where this is not available, the
specification for the steel, whether it be a national or international
standard or a maker's specification, is helpful and increasingly
these may provide data on the maximum CE values for particular
grades.
Some sources of information must be examined carefully to determine whether the composition quoted is 'typical' or 'maximum
permitted', whether it represents ladle analysis or product analysis,
and even whether the steel is a free-cutting grade. It is not unknown
for 'typical' values to be used for some elements and 'maximum'
values for others, in the same table!
In the absence of such detailed information it may be necessary
to use specification maxima for the particular steel type, but the
limits on composition may be so wide that a maximum CE value,
calculated on this basis, could lead to welding procedures which
would prove hopelessly uneconomic.
In cases where the test certificate, or steel specification, only
gives data on carbon and manganese contents, information for C: Mn
steels 1 suggests that an appropriate allowance for residual alloying
elements to be added to the CE would not exceed 0.03%. For steels
produced by an all-scrap route, this may be too low: values as high
as 0.08% have been found necessary, while 0.05% is a more typical
value.
In cases of doubt it is advisable to undertake sampling and analysis
of the steel to be welded. It is emphasised that any analytical
technique involving analysis of samples produced by machining is
liable to give high results for carbon (and consequently a risk of
specifying a higher preheat than necessary), unless the samples are
taken in a carefully controlled manner.
Segregation of elements may cause problems since it is the
chemical composition of the plate at the point where welding is to
be carried out which is important. Occasionally a plate may have &
significantly different composition from others from the same cast.
Segregation during solidification of both ingot and continuously
cast products can lead to compositional variations within one plate,
the centre thicknesses usually having higher alloy and impurity
contents than elsewhere. This segregation is normally reduced by
hot working so that castings, despite the homogenising heat treat-
35
Selecting values for graphical presentation
ments usually given, tend to exhibit more segregation than wrought
products.
The weldability diagrams detailed later have been devised to
determine safe procedures for steels having a maximum CE value
below that stated in the diagram. The values in the diagrams are
based on actual analyses so that the diagrams should indicate a
procedure with a very low risk of cracking. The risk of the predicted
procedure proving unsuccessful is increased when the CE value of
the steel to be welded is calculated only from a ladle analysis.
When such information on composition is not available, it is possible
to calculate a CE value from the specification maxima for the particular steel type, taking care that all significant elements in the
steel are specified. When only a typical composition is available,
the CE value must be calculated with due allowance for likely
variations above that typical composition.
Carbon equivalent level
The calculation of a CE level represents an attempt to describe
chemical composition by means of a single number in order to
show how changes in composition affect material behaviour. In
Chapter 2 it was explained that the likelihood of HAZ hydrogen
cracking tends to increase as the microstructure of the steel becomes
harder; thus, the CE formula describes how hardenability changes
with composition. There are many CE formulae, all of which are
empirical; the one used in this book (as discussed in Chapter 1)
is widely used, namely:
CE = C
Mn
6
+-+
Cr
+ Mo + V Ni + Cu
+--5
15
[3.1]
where the symbol for each element represents the content of that
element in weight per cent as determined by analysis or estimation,
as has been discussed. It is used as a point of entry into the
weldability diagrams to relate steel composition to the weld cooling
rates which are expected to produce specific levels of microstructural
hardness, in particular 350, 375, 400 and 450HV. However, the
HAZ hardness is not the only factor which determines the likelihood
of cracking; steel composition, the type of microstructure developed
and the level of preheat used are also important, as considered
further in Chapter 4.
Equation [3.1] was originally developed for semi-killed steels and
it has been shown that, for C-Mn-Si steels, hardenability is better
described by including in the equation an additional term, namely
36
Welding steels without hydrogen cracking
Si/6. However, it is also found that C-Mn-Si steels show the same
risk of cracking as otherwise similar semi-killed steels when the risk
is expressed as a function of cooling rate. The original form of [3.1]
is therefore retained for steels containing silicon.
It should be noted that when [3.1] is used for semi-killed steels it
is relating composition to hardenability, and when it is used for
silicon-containing steels it is relating composition to hardenability
and likelihood of cracking.
Precision of the CE formulae
Carbon equivalent levels are normally quoted to two decimal places
but the significance of the second place depends on the precision of
the original chemical analysis. The precision of the CE value can be
estimated from:
where the element symbol now refers to the precision of each individual analysis. For practical purposes the significance of the calculated CE level is important in deciding the point of entry to a
welding diagram.
The levels of analytical precision used in [3.2] should allow for the
differences obtained by operators in different laboratories if the CE
level is to be estimated from typical, specification, or ladle analyses.
A measure of this is the reproducibility limit, R, which is used in
British Standard 6200 for the Sampling and Analysis of Iron, Steel,
and other Ferrous Metals, and is defined as the 95% confidence limit
for results from two operators in different laboratories. It should be
remembered that compositions are not always based on the use of
such methods except in cases of dispute. If the CE level is calculated
from product analyses, the within-laboratory precisions of the particular analysis methods employed can be used in [3.2].
These precisions will generally be better than the corresponding
BS ones, because no inter-laboratory errors are included. Such estimates of precision for methods currently used at TWI are shown in
Table 3.1. When these figures are used in [3.2], the following results
are obtained:
CE level reproducibility (BS method)
to.01B
CE level reproducibility (TWI method) to.009.
There is thus little point in considering whether a CE level of 0.40
would be preferable to one of 0.42 when using a welding diagram,
37
Selecting values for graphical presentation
Table 3.1 Precision (as levels) of analytical methods: wt%
BS Methods
TWI methods
C
Mn
Cr
Mo
v
Ni
Cu
0.011
0.012
0.034
0.008
0.016
0.004
0.016
0.015
0.010
0.001
0.012
0.012
0.010
0.003
since the steel composition, even with product analysis, is not characterised with sufficient accuracy. This also explains why the diagrams given later in Fig. 4.3 are given for CE values 0.02 CE apart.
Effects of sulphur
Free-cutting or free-machining grades of steel normally contain up
to 0.2% of sulphur, although up to 0.5% is possible. Other elements
such as lead and selenium may also be present. In these cases special
precautions are necessary and are outlined in Chapter 4.
Low sulphur contents, e.g. <0.015%, have been shown':" to
increase the risk of hydrogen cracking in one specific instance when
a C-Mn steel was being welded in a situation where a hardened HAZ
was likely.
Since that time, several investigations have studied the problem
and the majority have found that reducing the sulphur level in the
steel increased both the HAZ hardenability and the risk of cracking."
As the phenomenon is linked to the number of inclusions, and not
directly to the weight % of sulphur, quantitative guidance on the
possible increased risk of cracking when welding steels of low
sulphur content is not possible. The need for caution is greatest
when, as experienced originally," existing safe procedures for high
sulphur steel, e.g. >0.015%, are applied unaltered and unchecked to
steels of low sulphur content.
The change in procedure required for a low sulphur steel to one
of higher sulphur content of the same CE, may be as significant as
an increase in CE of 0.03, or an increase in preheat temperature of
50-100°C, depending on the steel and the welding circumstances.
In many cases, however, the change in risk of cracking is very small
and accommodated without significant change to the procedure.
Although the emphasis in recent years has been on the role of
low sulphur contents in promoting hydrogen cracking, it is also
possible for a sulphur content which is relatively high to give liquation cracks in the HAZ and for these minute cracks or hot tears to
act as potent nuclei for the nucleation of hydrogen cracks. Steels
38
Welding steels without hydrogen cracking
with Mn: S ratios below about 25: 1 are at increasing risk as their
sulphur levels exceed 0.010% and their carbon contents about 0.15%.
Welding dissimilar steels
When joining dissimilar steels, welding procedures should be
devised to avoid cracking either of them; this usually involves
selecting a procedure to suit the steel of higher CEo However, both
steels should be examined if one is of higher carbon but lower alloy
content than the other. This is particularly important for alloy
steels, or for thick plate, because fully hardened HAZs are likely to
be formed in both. The steel with the higher carbon content will
probably produce the harder HAZ and thus be subject to the greater
risk of cracking.
In the selection of consumables for a dissimilar steel joint, it is
only necessary to match the strength of the weaker steel. In fact, it
may be an advantage to use the lowest strength weld metal possible
to minimise the stresses on both HAZ and weld metal.
Hydrogen potential of the consumable
With mild steel and C-Mn steels the hydrogen potential of the
welding process must be known so that the welding diagram can be
entered via the correct CE axis. Detailed information on typical
hydrogen levels for different processes is given in Appendix A.
In the case of hardenable steels, low or very low hydrogen processes should always be used, even though the weldability diagrams
for such steels are based on medium and high hydrogen potential
consumables. At the low composition end of these diagrams the use
of low and very low hydrogen consumables should enable lower
preheat temperatures to be used, but this must be confirmed by
making a joint simulation test.
Selection of carbon equivalent axis
As explained earlier, the risk of cracking increases with increasing
hardness of the HAZ for a particular level of hydrogen and joint
restraint. For a given restraint, lowering the weld hydrogen level
allows harder HAZs to be tolerated without cracking, and this
forms the basis of the different CE axes for mild and C-Mn steels.
The following guidance on which CE axis to use is based on a
39
Selecting values for graphical presentation
medium restraint level, i.e. that generally appropriate to normal
structural steel welding:
Hydrogen level
(mL/100 g deposited metal)
CE
axis
>15
""15, >10
""10, >5
A
B
C
D
"..5
Higher restraint conditions (e.g. ring stiffeners, partial penetration
restrained welds, box-sectioned constructions, etc) require additional
precautions against hydrogen cracking. These could take the form
of increased preheat levels, perhaps by 25 -75°C, or use of the CE
axis corresponding to the next higher hydrogen level. Joint restraint
cannot yet be quantified in a practical manner and appropriate
precautions can only be established by experience or by joint
simulation testing. Table 3.2 provides some guidance on selection
of the CE axis for different joint types and restraint levels.
In assessing, which CE axis to use for different hydrogen levels
and types of joint, use was originally made of critical hardness
levels to avoid cracking, and these values were included in the
version of Table 3.2 in the first edition. However, because critical
hardness values depend on so many factors other than hydrogen
level, it has been decided to remove indications of hardness from
Table 3.2 and the relevant text.
Combined thickness of the joint
The combined thickness of a joint is the total thickness (mm) of the
plates meeting at the joint line. It is a parameter which describes
the size of the paths available for conduction of heat away from the
weld and thus enables the effect of material thickness and joint
geometry on cooling rate to be quantified. An increase of plate
thickness above a certain limit has no effect on the cooling rate and,
as that thickness also depends on preheat temperature and heat
input, broken lines have been drawn on the appropriate diagrams
for fillet welds at three different preheat levels to indicate the
maximum effective combined thickness to be used (see Fig. 4.2).
These lines correspond to a cooling rate at 300°C, the one used to
relate CE to the extent of hardening, and their positions would
change for cooling rates at different temperatures. The broken lines
Butts
Fillets
©
@
[I
®
©
e
®
@
D
©
©
D
EEl
®
®
A
EEl
EEl
@
@
C
®
®
EEl
r-,
e
..
©
©
®
©
@
@
®
@
EEl
®
®
EEl
Partial
penetration, Full
Normal
penetration, structural Machined High
Continuous Abutting
high
restraint Patch
fit
patch
fit
surface
surfaces Normal Misaligned restraint
Bead-on-plate
CE axis
Notes:
(a) Letters ringed refer to the 4 CE scales in Fig. 4.1-4.3 and enable the correct axis to be chosen for a particular combination of hydrogen
level, joint geometry and restraint.
(b) Dotted letters are tentative.
(c)
less sensitive than Scale A, EEl more sensitive than Scale D.
" See Appendix A.
Very low
(':;;5)
Low
(':;;10, >5)
e
Manual metal-arc electrodes
with other than basic
coverings
Manual metal-arc electrodes
of hydrogen controlled type
unless established as low or
very low
Submerged-arc process,
unless established as low or
very low"
Cored wire processes, unless
established as low or very
low"
Manual metal-arc electrodes
of hydrogen controlled type,
may require drying ~250°C
Submerged-arc and cored
wire processes ~
Manual metal-arc electrodes
of hydrogen controlled type,
may require drying ~350°C
Submerged-arc and cored
wire processes ~
Gas-shielded and TIG
processes with solid clean
wires
High
(>15)
Medium
(':;;15,
>10)
Examples of
corresponding processes and
consumables
Weld
hydrogen
level, mil
100g of
deposited
metal"
41
Selecting values for graphical presentation
in Fig. 4.2 correspond to the positions in Fig. 4.3 where the slightly
curved sloping lines become vertical.
Combined thicknesses greater than the values indicated by the
broken lines on Fig. 4.2 decrease the time to cool to temperatures
lower than 300°C, and hence increase the amount of hydrogen
retained at room temperature. As a result, some increase in preheat,
or the use of postheating, may be required for very large thicknesses
beyond the broken lines.
When calculating combined thickness, the thickness of each part
should be averaged over a distance of 75 mm from the weld line.
However, when using preheats above 100°C and/or low hydrogen
processes, account must be taken of any thickening beyond the
75 mm mark, since this may increase cooling rates at low temperatures when hydrogen is diffusing out of the joint. On the other hand,
if the part thins or terminates just beyond 75 mm, there is a possibility that less stringent welding conditions can be employed. In
both instances a joint simulation test should be conducted to establish safe conditions.
With directly opposed twin fillet welding the combined thickness is calculated as one-half of that of the joint welded one
side at a time. However, it is necessary to ensure that the two
arcs remain opposite each other if advantage is to be gained
from the less stringent procedure that is possible with this
method.
Heat input
This term is used to indicate the heat input to the plate during
welding. It is neither technically rigorous nor universally applicable in welding, but it has the merit of being easily calculated and
hence useful for those processes and procedures dealt with in this
book. Heat input, together with combined thickness and preheat
temperature, controls the cooling rate of the weld, the microstructure, and hence the susceptibility of the HAZ to hydrogen cracking.
The term 'heat input' should be reserved for the heat input to the
steel being welded and 'arc energy' for the energy supplied by the
arc (the difference being accounted for by the so-called 'arc efficiency'). The terms have been used indiscriminately by some, and
different standards & techniques have used each parameter. Only the
term 'heat input' is now used in European Standards and in the
nomograms in this text.
Arc energy (E) is calculated as follows:
42
Welding steels without hydrogen cracking
E=
v x A x 60(kJ/ mm)
100 xS
[3.3]
where V = arc voltage
A = welding current
S = welding speed or travel speed (mm/min)
Care must be taken to measure the arc voltage, which is likely to be
lower than the voltage indicated on the welding set because of
potential drop along leads, etc. The following approximate arc efficiency factors allow the calculation of heat input from arc energy:
Manual metal-arc
80%
Tungsten inert gas
60%
Submerged-arc
100%
Cas shielded metal-arc 80%
The reader should be warned that two types of presentation of heat
input information are possible. In the earlier edition of this book, in
BS 5135 and in many texts originating from English-speaking countries, arc energy is used (although it may be termed heat input),
unfactored by arc efficiency. It is important in such cases to know
which process was used to develop the data. For the present text
and BS 5135, MMA was used. Thus, for MMA or gas shielded metalarc welding, a value of Lkj/rnm relates to an arc energy of Lkj/rnm
which is equivalent to a heat input to the steel being welded of
~0.8kJ/mm.
For submerged-arc welding the arc energy of the process
should be factored by 80/100 to obtain a value that is to be used, so
that if a minimum heat input of 1 kl/rnm is required for MMA
welding, 0.8kJ/mm will be needed for submerged-arc. Conversely
for TIC welding, the required arc energy will be 1 x 80/60. i.e.
1.3 kj/mm. With the system frequently used in continental Europe,
and now embodied in BS EN 1011, and this edition of this book the
heat input value quoted is already factored.
For manually welded fillet and butt joints, heat input can be most
accurately controlled by means of the run length from one electrode.
The length of electrode used to make a unit length of weld is termed
the runout ratio, k, and is proportional to the heat input, E in kj/rnm.
Tables providing this conversion are included for use with the
welding diagrams (Tables 2.1-2.4). Results from many sources
confirm that a relationship exists between runout ratio, arc energy,
and electrode core diameter, d, in mm, for electrode coatings having
43
Selecting values for graphical presentation
different contents of iron powder. For iron-free and low iron powder
electrodes, giving electrode efficiencies not greater than 110%
according to BS EN 499: 1995 (e.g. basic electrodes of AWS class
E__16):
k = 1000E/51d 2
[3.4a]
For medium iron powder electrodes giving electrode efficiencies
above 110% and not greater than 130% (e.g. electrodes of AWS
classes E__14 and E__18):
k = 1000E/59d 2
[3.4b]
For high iron powder electrodes giving electrode efficiencies above
130% (e.g. electrodes of AWS classes E__24, E__27 and E__28):
k = 1000E/76d 2
[3.4c]
Tables 2.1-2.3 provide conversions from heat input values given in
the weldability diagrams to weld run lengths (the reciprocal of k,
the runout ratio) for different sizes of manual electrodes.
Control of fillet welds by leg length is possible but should be used
with care since different leg lengths can be obtained at the same heat
input in positional as opposed to flat welding, and also from basic
electrodes as distinct from rutile coated electrodes. Furthermore, the
minimum leg length is normally used for strength calculations and
inspection, whereas the mean leg length is most closely related to
the HAZ cooling rate. If standard sized fillets have to be deposited
and minimum leg lengths are specified, the heat input values collected in Table 2.4 can be used with the appropriate welding diagrams. When using Table 2.4 periodic checks on run lengths should
be made to ensure that the required heat input is being used.
When using twin or multi-arc welding techniques the individual
heat input values should be added together when a single weld pool
is maintained.
Preheat and interpass temperature
It has been assumed that local preheat is used, that it is applied along
the weld line, and that the temperature on both sides of the joint is
measured at least 75 mm from the weld line on the opposite side
of the plate to that being heated. If the temperature can be
44
Welding steels without hydrogen cracking
checked only on the heated side, the heat source should be removed
and sufficient time (1 min/25 mm thickness) allowed for temperature
equalisation before measurement. If general preheat is used, a lower
preheat temperature may be adequate, but this should be confirmed
by joint simulation tests. Conversely, a lower standard of preheating
would demand a higher temperature.
Postheat
The width of the band heated is not as important as ensuring that
the weld and HAZ do not fall below the specified temperature.
Fit-up
A poor joint fit-up, equivalent to a fillet weld root gap greater than
0.4 mm, has been assumed in developing the diagrams. If a better
joint fit-up can be guaranteed, as, for example, when welding relatively small components with machined preparations, there is a
strong possibility that less stringent procedures can still avoid hydrogen cracking. This possibility must also be confirmed by joint
simulation tests.
Misalignment
From the knowledge of fatigue cracking, it is known that misalignment between two plates butt welded together gives an increased
stress concentration. There is some evidence that there is a similar,
adverse effect on hydrogen cracking. This was noticed when the
two plates were misaligned by a few millimetres, but there is also
likely to be an effect if the two plates are out of line by a few
degrees.
Multirun welds
After the root pass has been made, the absence of a root gap reduces
the amount of strain in the upper weld runs so that some relaxation
of procedures may be possible. However, in thicker material, i.e.
50 mm and above, the strain on the upper runs increases because of
the build-up of residual stresses, and relaxation of procedures is
less possible.
When short welds, or short blocks of welding, are contemplated,
relaxation of procedures may be possible if overlying runs are
deposited before the lower runs have cooled to about 100°C. This
45
Selecting values for graphical presentation
technique, however, is not recommended for thick material since
excess hydrogen can be retained in the weld due to the short time
cycle at high temperatures.
Tack welds
It is important to ensure that tack welds are either made using the
same procedure as is used for the main welds, or that the procedures
for them are safe. They must be of adequate length, i.e. at least
50 mm, or four times the length of the thicker part, whichever is
less.
REFERENCES
1 Gledhill P K and GoultEL: 'The production of universal beams and
columns in high yield structural steels'. lSI Publication 104, 1967, 1827,243.
2 Smith N and Bagnall B I: 'The influence of sulphur on HAZ cracking of
C-Mn steel welds'. Brit Weld J 15 (2) 1968 63-9.
3 Hewitt J and Murray J D: 'The effect of sulphur on the production and
fabrication of C-Mn steel forgings'. Ibid, (4), 151-8.
4 Hart P H M: Welding in the World 23 (9/10) 1985 230-238.
Chapter 4 Welding procedures for different
steel types
This Book will make a Traveller of thee.
If by its Counsel thou wilt ruled be;
Yea, it will make the slothful active be.
The blind also delightful things to see.
So far, the factors responsible for hydrogen-induced cracking have
been described, the principles upon which safe welding procedures
can be based have been outlined, and the information required
before using a welding diagram has been listed. It is now possible
to explain in this chapter the selection of safe welding procedures
for different steel types.
The welding procedures derived in this book were intended to
give a high degree of protection against the risk of hydrogen cracking in weld HAZs. In 1974, when incorporation of the procedures
into the British Standard for welding structural steels (BS 5135*)
was discussed, comparison of the TWI data with the experience of
the British structural steel fabrication industry showed that the TWI
procedures were more conservative than had been used successfully
up to that time. It appeared that the discrepancy could be removed
if the TWI procedures were made appropriate to steels of 0.02 CE
higher, and this difference was put into the standard, so that the procedures for a steel of, say, 0.45 CE in the present text were equivalent to those of a steel of 0.47 CE in BS 5135: 1974. Although
experience since BS 5135 was introduced in 1974 has generally been
satisfactory (and the shift from TWI data has been maintained in BS
* BS 5135 was superseded by BS EN 1011-1 and 2 when these were published in
1998 and 2001 respectively. Recommendations on the derivation of procedures to
avoid hydrogen cracking are presented by two methods, method A and method B.
Method A is essentially that in BS 5135. with the following principal modifications and additions.
(1) Modifications to welding procedures for steels of low CE, as in the present text.
(2) A new CE scale (E), based on a maximum diffusible hydrogen level of
3ml/l00g deposited metal.
(3) Procedures for steels having CEs up to 0.70 for use with scale C, 0 and E.
(4) Heat input (as defined in chapter 3 and in BS EN 1011-1) is now used rather
than arc energy as an entry point into the diagrams.
47
Welding procedures for different steel types
EN 1011-2: 2001), the original TWI approach has been retained in
the present text. This is because the more conservative TWI procedures give less risk of small hydrogen cracks occurring, which may
be detected only when a high level of quality is needed and inspection standards are high.
For mild steels, C-Mn steels and low alloy steels with yield
strengths up to about 600N/mm2 , HAZ cracking is generally the
major form induced by hydrogen. Possible exceptions involve
welded joints where the design imposes a particularly high stress
concentration; for example in a partial penetration weld or a weld
made on to a permanent backing bar. Under such conditions the
symptoms may be those of weld metal hydrogen cracking, but close
examination often reveals that cracking has begun in the HAZ, even
though it has propagated through the weld metal.
Hence, the information given in Tables 4.1 and 4.2 and Fig. 4.1-4.3
is based on avoidance of hydrogen cracking in HAZs, and all experience shows that, for most fabrications, procedures based on these
will also be adequate for avoiding hydrogen cracking in the weld
metal. However, there are circumstances where cracking is more
likely in weld metal and specific precautions may then be necessary
to avoid that problem. This is particularly the case when little or no
preheat is indicated by the diagrams and tables, either because a
steel of fairly low CE «~0.45)
is being welded or because low heat
inputs are not being used. The risk of weld metal hydrogen cracking further increases when welding thick section steel (>~50 mm
thick) and when using alloyed weld metal (especially if the CE of
the weld metal exceeds that of the parent steel) or if a C-Mn weld
metal is used containing >~1.5%Mn.
As yet there are few comprehensive data to determine the preheat
levels necessary to avoid weld metal hydrogen cracking.
Recent work indicates that for non-alloyed C-Mn weld metal, in
contrast to the situation in the HAZ, the risk of weld metal hydrogen cracking can increase, particularly in multipass welds, with
increasing heat input. The normally envisaged relaxation of preheat
at higher heat inputs should therefore be reviewed carefully in such
situations.
For steels containing <-0.2%C and of yield strengths greater than
600N/mm 2 , welded with weld metal of matching or over-matching
yield strength, weld metal hydrogen cracking appears to be the predominant form. The precautions needed to prevent weld metal
cracking are generally more than sufficient to prevent HAZ hydrogen cracking. It has already been stated (Chapter 1) that the factors
responsible for weld metal cracking are essentially the same as those
48
Welding steels without hydrogen cracking
for HAZ cracking. The types of microstructure which appear susceptible in the weld metal are, of course, different from those in the
HAZ and for a given susceptibility have lower hardness values.
There are, as yet, few systematic data which can be used to select
safe welding procedures for steels of these higher strengths and in
general an empirical approach, supported by joint simulation
testing, is employed.
A welding procedure can be defined by specifying a heat input
calculated directly from welding parameters (Chapter 3). In manual
welding it is convenient to maintain a specified heat input by controlling the length of weld produced by a single electrode. Tables
2.1-2.3 give the relationships between heat input and run length for
electrodes containing different amounts of iron powder in their coatings. If, for strength purposes, a minimum fillet leg length is specified, the heat input values given in Table 2.4 should be used. It is
advisable to check the run lengths of such welds, particularly when
conditions begin to involve a small risk of cracking, because the relationship between heat input and leg length is affected by several
factors.
Mild steel
Approximate limits of composition:
Carbon not more than
Manganese not more than
Silicon not more than
0.25%
1.0%
0.5%
Mild steel is the first, and generally the most weldable class of
steel. It has very low hardenability and any suitable electrode may
be used. For thin section welds preheat is not necessary, and with
CE values less than 0.30 no control of bead size is needed for any
combined thickness. Overhead and vertical-down welding can be
performed successfully at 1.0 kj/rnm heat input on any thickness up
to a CE of 0.32. Horizontal-vertical welding at 1.4kJ/mm can be
undertaken up to a CE of 0.34. For mild steels generally, composition may not be known and a maximum CE of 0.38 must be assumed.
Table 4.1 summarises the heat inputs to be used to control bead size
when welding without preheat mild steels of assumed maximum CE
0.38.
For sections thicker than about 70 mm preheat may be necessary
and Table 4.2 allows the appropriate level to be selected. Low hydrogen electrodes are required only when plate thickness exceeds about
25 mm, or if a specific toughness level is required. However, the use
49
Welding procedures for different steel types
Table 4.1. Maximum thicknesses for welding mild steel without preheat (suitable
for a maximum CE of 0.38)
Heat input,
kJ/mm
Maximum combined thickness, mm
Hydrogen potential:
0.6
0.8
1.0
1.4
1.8
High (A)
Medium (B)
Low (C)
33
42
51
69
87
37
48
60
83
44
58
73
Very Low (D)
~2.2
Notes:
A dash indicates no limit.
2 Lower values of combined thickness may be required for highly
restrained joints.
1
Table 4.2 Maximum thicknesses for welding mild steel with preheat (suitable for a maximum CE of 0.38)
Preheat
temperature,
°C
75
100
Notes:
Heat input,
kl/mm
0.6
0.8
1.0
1.4
1.8
0.6
0.8
1.0
1.4
1.8
Maximum combined thickness, mm
Hydrogen potential:
High (A)
Medium (B)
Low (C)
41
51
62
82
45
60
74
58
79
48
63
78
59
76
Very Low (D)
1 A dash indicates no limit.
2 Lower values of combined thickness may be required for highly restrained joints.
of such electrodes or other low hydrogen processes may enable
preheat temperatures and weld bead sizes to be reduced.
Although a CE of 0.38 is generally the maximum for mild steel in
relatively thin section, for thicknesses above about 100mm a higher
level of CE may be encountered because of higher carbon and manganese contents. When the manganese content exceeds 1.2% the
50
Welding steels without hydrogen cracking
steel should be treated as a C-Mn steel and welding procedures
deduced as described below. When only the carbon content of thick
section is known the procedures referring to fully hardened mild
steel may be used as on pp. 61 to 62.
Carbon-manganese steels
Approximate limits of composition:
Carbon not greater than 0.25%
Manganese
1.0-1.7%
4.1 Welding procedures for CMn steels without preheat:
select electrode size and run
length from Tables 2.1-2.4.
0.4.1
0.35
0.3
0.4lJ
0.41
0.38
0.33
Carbon equivalent levels
0.4
.)l
0.43
0.45
0.47
0.41
0.43
0.45
0.43
0.38
0.41
CE axis from Table 3.2
0.49 - - C I
0.47 _ Ii ......
0.45 A ...
.
180 +---+t----ll-l-I---+-+---f---++----f
160 +---+t---1H---,I---+-+-+--+-+-....--l--L....-L........L-I
/
140 t--1t--r-t-11111-r----tif-7~~~~
I
11
U
/
OA7OAY 0.510.57
OAY 0.51 0.53 O.5Y
/
/
/[1'"'
'
'
'
'
'
.
/;//J~
//v~~~/
I /~V/)~
1~1Ia~
80+---+-hf----HL-h,t-t,h~A-:.~-+----1
O+---I---+---+---t-----+---I
o
2
3
4
5
6
Heat input. kJ/mm
O.m.550.57 -
Welding procedures for different steel types
51
Silicon not greater than
0.7%
Niobium or vanadium not greater than 0.1%
The second group of weldable steels comprises the weldable CMn steels, with or without microalloying additions of Nb, V and/or
Ti.
These steels are of relatively low hardenability and preheat is not
usually necessary for thin sections. Those thicknesses which can be
welded without preheat however depend on composition, hydrogen
level, and bead size; Fig. 4.1 relates these factors to define those
thicknesses which can be handled without preheat. In this diagram
the appropriate CE axis, A, B, C, or D, selected from Table 3.2,
e-
4.2 Welding procedures for
Mn steels with preheat.
Combined thickness = t. +t2 +1 3
II = average thickness
t,
t
2
~
over a length of 75mm
... 75mm'For directly opposed twin fillel
welds combined thickness = '/2 (I. +1 2 + (3)
I,
Combined thickness, mm
"-1j+-t
R ~o
/
3
./" V
///// V
1--7~/,y h-¥
/'/
/."...q
/i--7"q-----!o,L. /"=-----::>.L+--""~
1$// /
CE Scale A
o
./.........
0.3 0.35 0.40
1",,111" 1
B
C
~ //.
h/.
0.300.35 0.40
'""I"",
0.30 0.35 0.40
I
I
0.40
I
I
I
I
0.45
I
I
I
I
I
I
I
!
0.45
I
I
I
0.45
I
I
I
0.50
I
0.50
,
I
I
0.55
Carbon Equivalent
I
I
'"
0.50
I
I
I
I
I
I
I
0.55
15 20
/
/
/
25
/
30
/
40
,/
50
/
60
/
75
1/
100
/
/ / / / 1/ / / / / / 1/ 125
II/II 4~ 11/ / /
'!OW.-/.:'··V···
'/>--:'......
50
~gv,,'
0.50
"!
I
0.45
~
:..----
10
.
2
I
!
0.55
I
I
Ma,ximUm{nOprehe
100°C
combined thickness
valu1elo1beused'jith
Heal input (kJ/mm)
2~oC
4
6
52
Welding steels without hydrogen cracking
Minimum preheat temperature, "C
75 20 0
-
140
";
i
!
!
120
!
,
100
!
§
Ii
<l.l
c::
:s
~
o
80
"'c::"
Scale (see Table 3.2)
A
B
0.30
0.33 0.38
0.38 0.41
0.41 0.43
0.43 0.45
0.45 0.47
0.47 0.49
0.49 0.51
0.51 0.53
0.53 0.55
C
D
0.35
0.41
0.43
0.45
0.47
0.49
0.51
0.53
0.55
0.57
0.43
0.46
0.48
0.50
0.53
0.55
0.57
0.59
-
-
~
see (b)
~
see (c)
see (d)
see (e)
see (f)
see (g)
see (h)
see (i)
see (j)
~
~
~
~
~
~
~
<l.l
:E 60
E
0
U
40
II
II
,
20
(a)
4.3a-j Welding procedures for
C-Mn steels with selected CE
values for MMA, select
electrode sizes and run lengths
from Tables 2.1-2.4, for other
processes use factors from p.
42.
o
o
0.5
Heat input, kJ/mm
1.5
together with the maximum CE value for the steel in question, are
used to identify the line relating to the heat inputs for different combined thicknesses. Figure 4.2 is a comprehensive diagram showing
the weld sizes and preheat levels necessary when heat input must
be restricted, as in positional welding, and for thicker plates. It is
used in the manner described in Chapter 2 and links all the variables into one diagram. The same information is given again in Fig.
4.3a-j, but subdivided, so that the preheat temperature, combined
thickness, and heat input values at specific CE levels can be emphasised. This method of presentation was also used in Fig. 4.1.
Provided strength and toughness requirements can be met, low
hydrogen processes are not essential, although their use allows bead
sizes and preheat temperatures to be reduced. It is always advantageous to use the lowest strength consumable consistent with achieving the required strength and toughness. The likelihood of hydrogen
cracking occurring in this type of steel increases as carbon and man-
53
4.3
Welding procedures for different steel types
Continued
Minimum preheat temperature, °C
Minimum preheat temperature, °C
125 100 75
100 7'5 200
,
14v
;
120
12v
s
S 100
]'"
,
80
u
:£
20 0
s
S 100
]''""
u
-e
~ 60
I
;
~o
u 40
I
,
20
:£
]
-
~
80
60
40
20
;
0
(b)
o
200
1
Heat input, kJ/mm
150
12'5
100
2
0.5
75
20 0
160
140
,
;i 120
~
100
~
80
,
60
,,
40
20
(d)
1
1.5
2
2.5
3
Minimum preheat temperature. °C
-i
180
£
]
1
Heat input, kJ/mm
2
3
Heat input. kJ/mm
4
Welding steels without hydrogen cracking
54
4.3
Continued
ISO
200
125
100
75
o
20
Minimum preheat
temperature, °C
180
160
§
vi
'"
~
.~
-5
"0
ll)
140
120
100
c
80
u
60
~o
40
20
,
o
o
2
(e)
175
200
150
3
4
Heat input, kl/mm
125
100
5
75
6
20 Minimum preheat
temperature. °C
180
160
S
S
:i
c
ll)
o
140
I
120
-""
.~
-5 100
"0
c
ll)
:E
S
0
u
80
60
40
20
o
o
(f)
~
234
Heat input, kJ/mm
5
6
55
4.3 Continued
Welding procedures for different steel types
175
200
125
150
~
'"
~oS
I----
l-r-
~-
~
---r--
180
§
Minimum preheat
temperature, °C
100
160
75
140
o
20
120
100
]
~
80
60
40
-l
-'
20
o
o
(g)
200
2
3
5
4
6
Heat input, kJ/mm
175
150
125
Minimum preheat
temperature, °C
180
160
§
100
75
140
20
o
'" 120
'"
]
~ 100
l
~o
80
u 60
40
20
:
(h)
o
o
234
Heat input, kJ/mm
5
6
56
4.3
Welding steels without hydrogen cracking
Continued
200
200
175
125 ,
ISO
Minimum preheat
temperature, °C
180
160
E
E
V>
V>
OJ
J2o
100
140
75
120
o
:s
:E
"E
100
U
60
20
"0
OJ
80
0
40
20
234
(i)
6
5
Heat input, kJ/mm
,.....
200
200
150
,
Minimum preheat
temperature, °C
180
160
§
~
]
, 125
100
75
140
120
20
o
~ 100
"0
~
:E 80
,
E
(3
60
40
20
o
(j)
o
2
3
Heat input, kJ/mm
4
5
6
57
Tentative guidelines for
avoiding weld metal hydrogen
cracking when submerged-arc
welding C: Mn steels with 83
(l.5%Mn) wire and basic flux.
Multipass weld metal contains
approximately 0.8%C. 1.5%Mn
and has a CE of -0.36.
Interpass time 20 minutes,
hydrogen levels, thickness and
preheat/interpass temperatures
as shown. C - cracking found:
NC - cracking not found.
Interpass and preheat
temperatures were the same for
each test weld.
Welding procedures for different steel types
25 - w - - - - - r - - - - - - , - - - - - - - - - r - - - - - r - - - - - . ,
4.4
20 t - - - - - + - - - - + - - - I - - - I I - - - - + - - - - - i
8
15
t-----I-------+-+----+-+---I---+-------1
::J
E
NC
5 kJ/mm
IO.-----II----¥--+----+---.--I-----I
5t-----t----+-----I1----+-----;
NC
o
50
100
150
Preheat/interpass temperature,
200
250
°c
ganese contents are raised, and a watch should be kept for occasional
'rogue' plates of high composition. It is important to ensure that tack
welds are either made using the same procedure as for the main
welds, or that the procedures for them are chosen from Fig. 4.2 or
4.3.
When C-Mn steels of very thick section are to be welded and fully
hardened HAZs are likely to be formed, the procedures des-
58
Welding steels without hydrogen cracking
cribed under 'Medium carbon and carbon-manganese steels' in this
chapter can be used as an alternative to those summarised in Fig.
4.2.
Also, with thick sections, and in the other cases mentioned in p.
47, the risk of weld metal hydrogen cracking can be minimised by
taking account of the data given in Fig. 4.4. This figure is relevant
to multipass submerged-arc weld metals (made from S3 (1.5%Mn)
wire and a basic flux) containing about 1.5%Mn and having CE
values of about 0.36. The welds were made with controlled interpass
times of 20 minutes and preheat and interpass temperatures as
shown. Up to weld metal hydrogen contents of 5 mIllOO g, hydrogen
cracking in the weld metal was not a problem, but at higher levels,
heat input was the major factor. Unlike HAZ cracking, increasing
heat input increased the risk of cracking.
Lower carbon, lean alloy steels
Approximate limits of composition:
Carbon not greater than
0.20%
Manganese not greater than 1.5%
Other elements are present, but in amounts such that the CE does
not greatly exceed 0.60, i.e. insufficient to confer high hardenability:
in this group, hardenability is high if welding a combined thickness
of 25 mm at a heat input of 1.4 k]/mm fully hardens the HAZ. The
group is the third major group of weldable steels.
These steels are generally of relatively low hardenability and can
often be successfully welded with procedures similar to those used
for C-Mn steels. Some of them, however, show different susceptibilities to hydrogen cracking at given HAZ hardness levels as compared with C-Mn steels. For example, a molybdenum-boron steel
has a critical HAZ hardness of 375 HV instead of 350 HV when high
and medium hydrogen processes are used. On the other hand,
although differing in transformation behaviour, the Mn-Cr-Mo-V
steel resembles C-Mn steel in critical hardness values. The use of
Fig. 4.2 derived for C-Mn steels can therefore lead to either unsafe
or uneconomic procedures in some instances.
Of the very many individual steels in this group, only a few have
been examined in detail and procedures for these are given, but
others can be treated as C-Mn steels or as fully hardened alloy
steels (see later in this chapter). In either case the presence of boron
should be noted because it makes the steel more hardenable, even
59
Welding procedures for different steel types
Minimumlocalpreheattemperature, "C
100
150
200 175
150
125
§
]
u
:s
-0
II
/V / / /
100
75
I /
~
~0
u
50
25
20
V
'/ /II
J
l VI
V/
V
(b) .....
2
/
I
/
~V /
(a)/
0
/
100
3
4
175
200
,)
J
20
V
/
I
/
I
J
II
/IJV/
II
150
10~'"
/V
If
rV
/
V
,/
/
~
125
100
20
~
(c) ....
2
3
4
2
3
4
Minimum heat input,kJ/mm
4.5 Welding procedures for
steels to BS 1501: Part 2: 1988271, -281 (nearest current
equivalent is BS EN 10028-6:
2003: P460Q) (Mn-Cr-Mo-V
steels with maximum CE of
0.62); (a) very low hydrogen
processes, Scale D, Table 3.2,
(b) low hydrogen processes,
Scale C, Table 3.2, and (c)
medium and high hydrogen
processes, Scales A and B,
Table 3.2; select electrode size
and run length from Tables
2.1-2.4.
in very small amounts. Although boron is not in the CE formula
[1.1], it is in the P crn formula [1.2] with a multiplying factor of 5.
Low hydrogen electrodes are normally required to produce the
necessary strength and toughness, although sound welds can be
made with rutile electrodes, especially in thin section.
For steels not listed individually it is advisable to consult TWI to
determine whether new information on weldability has become
available. If much welding of an unlisted steel is envisaged, some
experimental work to determine its transformation behaviour and
susceptibility to cracking is recommended. The cost of this in most
cases will be more than offset by savings either in unnecessary
preheat or in repair.
BS 1501: Part 2: 1988-271, -281 (Mn-Cr-Mo-V steels) (nearest
current equivalent is BS EN 10028-6: 2003: P460Q)
These steels are less hardenable than would be expected from their
CE levels. To reduce the HAZ hardness, therefore, lower preheats
are needed than for C-Mn steels. Procedures for welding this steel,
based on a maximum CE level of 0.62, may be obtained from Fig.
4.5. If fully hardened HAZs are expected an indication of suitable
preheat levels may be gained from curve L of Fig. 4.6.
60
Welding steels without hydrogen cracking
!;J 300 r - - - - - - , - - - , - - - - - - - , - - - , - - - - , - - - - . , - - - , - - - - - ,
i
!
8'" 250 I - - - - - + - - - - t - - - - + - - - ! - - - - . . = . , - - - + - - - - - - j - - - - + - - - - - - j
~
I
8. 200 I - - - - - + - - - - t - - - g
.i~
150
1
~ 100 I - - - - - l -....
E
.~
~
01----+---f----I----+----+---t---+-----1
0.4
.----+---+----+----1r----+----t---+--~4
Ii<
~
0.31-----+----t----+---4r----+--""'7""'-t---,,----t---,,------I
.~
Ii 0.21-----+---+----+-"7"'~
~
U
0.1 1-----+--:7""'--+-~"""'--+"7"''''---+_---+-----l----t------1
450
500
550
Expected HAZ hardness, HV
Diagram for selecting
minimum preheat, interpass
and postheat temperatures for
alloy and lean alloy steels
giving fully hardened HAZs.
Note: relaxation of preheat
temperatures may be possible
with low hydrogen processes
and thin sections, particularly
at low C contents. Confirmation
should be obtained by joint
simulation tests. For class of
steel refer to Table 4.3
4.6
600
650
700
ASTM A514 grade B (Tl type A steel]
Experimental work with the controlled thermal severity (CTS) test
indicates that a critical hardness of 400 HV can be accepted without
cracking using medium hydrogen (normally dried basic) electrodes
on this steel.' With matching strength electrodes of low hydrogen
level, a similar critical hardness level was obtained. Studies of the
hardenability of this steel suggest that the normal CE formula ([3.1],
Chapter 3) overestimates the effect of composition when selecting
the cooling rate to produce a 400 HV hardness level. It thus becomes
possible to use Fig. 4.2 directly for predicting welding procedures
for Tl type A steel as long as the CE scale, marked 'C', is taken
as referring to medium hydrogen mild steel electrodes and low
61
Welding procedures for different steel types
hydrogen matching strength electrodes (ef Table 3.2). The limited
nature of available data must be borne in mind when using Fig. 4.2
in this way. It is also necessary to observe the steelmaker's limits on
heat input and preheat to avoid reducing HAZ fracture toughness.
If fully hardened HAZs are expected, the curve labelled 'C-Mn steels'
in Fig. 4.7 may be used.
Medium carbon and carbon-manganese steels
Approximate limits of composition:
C steels
- Carbon 0.25-0.45%
Manganese not more than 1.0%
Silicon not more than 0.5%
C-Mn steels - Carbon 0.25-0.45%
Manganese 1.0-1.7%
Silicon not more than 0.5%
The steels in this group are mainly the C and C-Mn steels used for
general engineering purposes.
In thin section, medium carbon and C-Mn steels can be welded
using the procedures recommended for mild and C-Mn steels given
already in this chapter. Such procedures ensure that sufficiently
large weld beads are made so that the cooling rate is slow enough
to avoid certain HAZ hardness levels.
In thicker sections, however, fully hardened HAZs, which are very
susceptible to hydrogen cracking, are formed readily, and welding
procedures similar to those for alloy steels become necessary. Figure
4.7 enables a preheat, interpass, and postheat temperature to be
selected for both carbon and C-Mn steels. Entering the lower part of
the diagram at the appropriate carbon content and moving horizontally, the expected HAZ hardness level can be found from the bottom
scale. Vertical movement into the top half of the diagram reveals a
range of temperatures for different restraint conditions. The particular temperature selected for preheat, interpass, and postheat will
depend on the particular welding problem and it may be possible to
reduce this temperature by using low hydrogen processes, particularly if the carbon content is low and the section is thin. The suitability of the procedure chosen should be confirmed by joint
simulation tests.
With procedures of this type no control of weld bead size is normally necessary. However, when conditions are severe (for example,
high carbon content, thick plate, or high restraint levels) there is
some advantage in producing relatively small beads. These require
little time at the interpass temperature for much of their
62
Welding steels without hydrogen cracking
4.7 Diagram for selecting
minimum preheat, interpass,
and postheat temperatures
(carbon and C-Mn steels with
fully hardened HAZs).
Note: Relaxation of preheat
temperatures may be possible
with low hydrogen processes
and thin sections, particularly
at low C contents. Confirmation
should be sought by joint
simulation tests.
JJ
300
~
'High restraint"
==+----+----+----+----t-z""'"""-I----+---+----i
250
1
roo
3=+---+---+----l.<F""
1
150
.1
j 100
5E
.~" 0
~
0.4
~~
-;O~.3;t---+---t---+----+~,...,:::...----P~---+----+----1
I
c -:0-:.2+---t----+---:::;,...,:::...----j,,~---+---+----+----+-----I
t
u
250
300
350
400
450
500
550
Expected HAZ hardness, HV
600
650
700
hydrogen to be lost by diffusion. If adequate time for such diffusion
is not given, any remaining hydrogen will be trapped by succeeding
runs and can give rise to cracking. Suitable inter-run times can be
calculated by using the data and techniques given in Chapter 5, in
which is given an example of such a calculation where each run is
treated as an infinite cylinder. Steels of this type may also be
welded using the isothermal transformation method described later
in this chapter.
Alloy steels
Approximate limits of composition:
Carbon not greater than
Total alloying elements greater than
(excluding Mn)
0.45%
1.0%
63
Welding procedures for different steel types
The final group comprises the less weldable alloy steels, often used
at high temperatures or for general engineering purposes.
Most welding processes, with the possible exception of electroslag, produce hardened HAZs when applied to these steels.
Preheat will therefore be required in nearly all circumstances.
Basic coated, alloyed consumables are usually needed to achieve the
required strength and toughness and sometimes to provide resistance to creep or hydrogen attack at elevated temperature during
service. The weld metal is therefore more likely to be the site of
hydrogen cracking than is the HAZ. As it is often necessary to
restrict the weld bead size and even the preheat temperature in order
to achieve desired properties, such as adequate HAZ toughness, the
use of a low hydrogen process becomes doubly essential.
Of the many varieties in this group the Cr-Mo steels show a high
susceptibility to hydrogen cracking and are more likely than the
others to require a postheat treatment for hydrogen removal before
being allowed to cool out after welding. The same is also likely to
be true for steels of high manganese content (>1.7%).
Three methods of establishing welding procedures for highly
hardenable steels were described in Chapter 2. These involved the
use of:
1.
2.
3.
Temperature control
The isothermal transformation characteristics of the material
Austenitic or nickel alloy filler metal
It was explained that the first method is more applicable to steels
with lower carbon contents «0.3%) and the second is suitable for
higher carbon steels. The third method is used where conditions
prevent the use of high preheating temperatures. In this chapter each
method is detailed separately.
The temperature control method
As a first step the steel is placed in one of five grades by calculation
(as shown below) or by means of Table 4.3. These grades are based
on welding tests, on studies of the continuous cooling transformation behaviour of steels, and on an empirical formula" relating composition to maximum HAZ hardness:
HV = 90+1050C + [47Si + 75Mn +30Ni +31Cr]
[4.1]
The chemical symbols refer to the percentage of the element in the
steel, but this formua is not valid for low carbon low alloy steels
containing boron, although it appears to work for higher carbon
ri
Type A
Creuselso 47
D6AC
Durabele 900, 950
Durabele 1050
FV 520(B)
HY80
HY130
Ducol QT455
Maraging (18Ni-8Co)
Maraging (12Ni-5Cr)
Superelso 70
Dueol W30A, B
Commercial example
of steel type
* Normally welded with Ni alloy electrodes.
lMn-j-Ni-j-Cr-Mo-B
l~r~lNi-V
1!,Mn-l,Ni-V
lCr-j-Ni
lj-Cr-l+Mo(-}V)
1Cr-1Mo-j-V
14Cr-5Ni-Mo-Cu-Nb
2+Ni-l.!;Cr-J,Mo
5Ni-Cr-Mo-V
l lNi-1Cr...!.Mo-V
18Ni -8Co=5Mo-Ti-Al
12Ni-5Cr-3Mo-Ti-Al
l1Mn-1Ni-1Cr
l1Mn-1Cr-j-NHMo
l.Cr-Mo-B
3Mn
ll,Mn-2Ni(Co)
9Ni-4Co
12Cr-4Ni-Mo
5Cr-l,Mo
2.!;Cr-Mo
1 and liCr-+Mo
3Ni
9Ni
Mn(Ni)-Cr-Mo-V
l~Mn-Ni-Cr-Mo
l.l,cr-Al-Mo
3~r-Mo-V
3Ni-Cr
lCr-Mo
LCr-Mo
3Cr-Mo
1!,Mn-Ni-Mo
l+Ni-Cr-Mo
ljNi-Cr-Mo
2Ni-Cr-Mo
21,Ni-Cr-Mo
3Ni-Cr-Mo
4Ni-Cr-Mo
l~Ni-Cr
1.JMn-Mo
11,Mn-Mo
t Cr
lNi
Steel type
503-37,40,42
530-30, 32, 36, 40
605-30, 36, 37
608-37, 38
640-35,40
653 M31
708-37,40,42
709 M40
722 M24
785 M19
816 M40
817 M40
823 M30
826 M31, 40
830 M31
835 M30
897 M39
905 M31, 39
945-38, 40
BS 1501: Part 2: 1988-271, 281 BS EN 10083-6:2003
P460Q
BS 1501-503 BS EN 10083-4:2003 12Ni14
BS 1501: Part 2: 1988-510 BS EN 10083-4:2003
X7Ni9, X8Ni9
BS 1501: Part 2: 1988-620,621 BS EN 10083-2:2003
13CrMo4-5, 13CrMoSi5-5
BS 1501: Part 2: 1988-622 BS EN 10083-2:2003
10CrMo9-10, 12CrMo9-10
Old BS 970. Pt 2,
designation
Table 4.3 Grading alloy steels for use with Fig. 4.6 and 4.7
-
34CrNiMo6
30CrNiMo8
36NiCrMo16
K
L
L
M
L
K
L
K
M
L
L
L
C-Mn
L
L
M
M
L
L
C-Mn
i.
L
-
-
Observed
grading
42CrMo4
34Cr4, 37Cr4, 41Cr4
BS EN 100 83-1
designation
-
-
-
-
-
L
L
K
-
C-Mn
K
K
K
K
K
L
L
L
L
K
K
K
K
K
L
L
L
L
-
Estimated
grading
--
Q.
-=....~
..,
OC
il';'"
('l
('l
Ctl
=
..,
=
....
=
OC
Q.
0
'<
I;
0
I~
I:Il
Ctl
Ctl
I:Il
OC
I~
65
Welding procedures for different steel types
steels. The portion of [4.1] used for grading steels is that given
within the brackets and is described by a parameter F:
F
= 47Si + 75Mn + 30Ni + 31Cr
[4.2]
For values ofthis parameter, F, up to 115, a steel is graded 'carbon
steels'. For values from 116-145 the steel is graded 'carbonmanganese steels' and examples can be found in Table 4.3. For
these two grades the relationship between carbon content and
expected HAZ hardness is given in Fig. 4.7, and its subsequent use
to establish minimum preheat, interpass, and postheat temperatures
is exactly as described previously.
For higher values of the parameter, F, from 146-180, steels are
graded as 'K', from 181-225 as 'L', and for higher values as 'M'.
Examples of all these grades are to be found among the steels listed
in Table 4.3. The K, L, and M grades enable the appropriate lines
on Fig. 4.6 to be used to relate, for a particular steel type, carbon
content to the expected HAZ hardness. This figure is used exactly
as Fig. 4.7 described previously. Using the carbon content of the
steel (maximum specified, analysed level or the figure given on mill
certificates) the appropriate line is intersected horizontally in the
lower half of the figure. Vertical movement to the band in the upper
half of the figure then defines a minimum temperature for preheat,
interpass, and postheat. This temperature should not exceed the M,
temperature of the steel in question, and this point should be
checked (see 'High carbon, plain and alloy steels' later in this
chapter).
At low hardness values (below about 450 HV), low preheat and
interpass temperatures are predicted and postweld heating may not
be necessary. A further reduction of temperature may be possible if
very low hydrogen processes can be used, but this should be confirmed by joint simulation tests. For example, in maraging steels
graded 'M' and having a maximum carbon content of 0.02%, Fig.
4.6 would indicate a preheat temperature of 130°C for a highly
restrained weld. However, this temperature would not be compatible
with achieving adequate mechanical properties, so that a very low
hydrogen process is normally used to avoid the use of any preheat.
As the expected HAZ hardness increases, higher temperatures are
predicted and it becomes necessary to select appropriate times for
which postheat should be held to assist hydrogen removal. These
times may be obtained from the hydrogen removal curves which
are described and listed in Chapter 5. If the times involved appear
unacceptably long, the possibility of tempering before the weld
cools out can be considered. If this is not possible careful use of a
66
Welding steels without hydrogen cracking
temper bead technique can sometimes give acceptable results.
The Cr-Mo and Cr-Mo-V steels appear to be particularly susceptible to hydrogen cracking and, although temperatures higher than
those predicted from Fig. 4.6 are not normally required when using
a low hydrogen process, postheating and slow cooling after welding
are commonly used for such steels, for example those to BS EN 100
83-2: 2003 13CrM04-5, 13CrMoSi5-5, 10CrM09-10, 12CrM09-10.
With steels which give a hard HAZ two points should always be
noted:
(a)
An as-welded HAZ which contains appreciable amounts of
retained austenite produces hard, brittle martensite after a
single PWHT or tempering heat treatment. Hence, a second tempering treatment will be necessary.
(b) A check should be made to ensure that the selected preheat
temperature does not exceed the M, temperature of the steel; in
making this check, the M, temperature and the various degrees
of transformation may be estimated from formulae given later.
In addition, it must be realised that significant hydrogen loss from
the weld metal and HAZ will not take place until the temperature
is sufficiently below the M, for substantial transformation of the
austenite to have occurred. Highly alloyed materials such as the
martensitic 12%Cr creep resisting steels display M, points well
below 250°C and, to obtain hydrogen removal, must be cooled to
fairly low temperatures. This can greatly increase the risk of cracking. Hydrogen levels must be minimised, and careful consideration
should be given to the temperature to which the joint is cooled to
achieve transformation and hydrogen removal during the preheat
stage. To reduce postheat time, it may be advantageous to permit
cooling to the selected transformation temperature followed by
heating to a higher temperature giving faster hydrogen diffusion.
The use of isothermal transformation data
In the description of the principle of this method in Chapter 2 it was
explained that a knowledge of the transformation behaviour of a
steel made it possible to control the cooling of the weld HAZ and
so to produce certain preferred, i.e. less crack sensitive, microstructures. To use this technique it is necessary to know the isothermal
transformation characteristics of the steel. These may be obtained
from the steel-makers' data sheets or from one of the collections of
such data."? As explained in Chapter 2, a temperature is selected
67
Welding procedures for different steel types
which promotes transformation, usually to bainite, in a reasonable
time and over the temperature range which can be controlled in
practice. This selected temperature, which is, in fact, a preheat and
interpass temperature, must be held long enough after welding to
ensure complete transformation. It must also be recognised that adequate times will be twice as long as those indicated in diagrams
which have been prepared from specimens austenitised at temperatures less than 1250°C.
The use of austenitic and nickel alloy consumables
When the temperature control methods described above cannot be
used because of limitations on the preheat temperature, or if they
do not succeed in avoiding hydrogen cracking, the only alternative
remaining is to use a consumable which is itself insensitive to
hydrogen and which results in less hydrogen being left in the HAZ
after it has cooled to temperatures at which hydrogen cracking can
occur. Such consumables are provided by appropriate austenitic
stainless steels and nickel alloys.
However, both types require some preheat for the more difficult
steels they cannot be effectively stress relieved thermally, and welds
are difficult to inspect non-destructively for cracking. Austenitic
weld metals can be selected which are of higher strength than the
common suitable nickel alloys.
When selecting austenitic stainless steel or nickel alloy fillers, it
is necessary to ensure that dilution from the base steel can be satisfactorily accommodated. The normal choice of austenitic consumabIes for MMA welding is from the types 23Cr: 12Ni, 29Cr: 9Ni or
20Cr:9Ni:3Mo, e.g. from BS EN 1600: 1997. Grades 23 12L, 29 9 or
18 9 Mn Mo. The first named is most commonly used and is suitable for giving deposits containing sufficient ferrite to suppress
solidification (hot) cracking, with little or no martensite in the bulk
deposit. The 29Cr: 9Ni type may be preferred for high dilution runs
to avoid a fully austenitic deposit. However, in low dilution situations, the weld metal will contain a high ferrite level, perhaps as
high as 35%. Although this is of possible benefit in tolerating the
pick-up of sulphur from the parent steel, and also in giving a high
yield strength, this type of high ferrite weld deposit should not be
subjected to PWHT, since it will show marked embrittlement as a
result of the formation of the sigma phase during heat treatment.
Nickel alloy fillers have the advantage of lower coefficients of
thermal expansion than stainless steels, and this can reduce shrinkage strains and thus the risk of cracking in highly restrained
68
4.8 Guide to preheat
temperatures when using
austenitic manual metal-arc
electrodes at about 0.81.6k]/mm a) low restrain, e.g.
material thickness <30mm; b)
high restraint, e.g. material
thickness >30 mm.
Welding steels without hydrogen cracking
200
~--.,.---~-----.-----,----.--,
Increasing parent metal
hardenability or weld
metal hydrogen level
~
::l
~
~
100 1 - -__f---+---..........--f---+--+---1--t
~
Reducing parent metal
.
hardenability or weld
metal hydrogen leve I
~
~
c..
50 I--_ _+-__-+-~_~~~
o
0.1
0.2
0.3
0.4
0.5
Carbon content, %
joints. A number of Ni: Cr :Fe alloys are suitable, among them MMA
electrodes of the types covered by the American AWS specification
ANSI/AWS A5.11/A5.11M-97-NiCrFe-2 and E-NiCrFe-3. However,
they are more sensitive to solidification cracking and microfissuring
than the stainless steels, and may not be usable for high sulphur
steels. If PWHT is required, a nickel alloy filler may well become
the preferred choice to avoid intermetallic formation during the heat
treatment and consequent embrittlement.
When austenitic electrodes are used, preheat is not normally
required for steels containing up to O.2%C, although they are liable
to give hard regions of crack-sensitive alloyed martensite at the
fusion boundary. Such regions are prone to hydrogen cracking and
are very difficult to detect. At O.4%C and above, a minimum temperature of 150°C is required to prevent such cracking, as well as
normal HAZ cracking. Figure 4.8 gives guidance, showing schematically how hydrogen and restraint levels affect the degree of preheat
needed. Buttering the surfaces to be welded may reduce the level of
preheat necessary.
Although the technique is generally successful, hydrogen cracking can still occur in severe situations, and it is always advantageous
to reduce the hydrogen input to the joint by using covered electrodes
dried at a high temperature (following manufacturers' recommendations), or to use a gas-shielded process with solid wires of high
quality or cored wires known to give very low total hydrogen levels.
Diffusible hydrogen measurements are not useful on austenitic or
nickel alloy consumables. High heat input is often helpful, and techniques which give low dilution should be employed to minimise
69
Welding procedures for different steel types
the formation of martensite in the weld metal. Hard HAZs will
normally be produced and it is usually advantageous to temper
them, even if only by using a temper bead technique. Although it
tempers the HAZ, PWHT is usually ineffective in giving a high
degree of stress relief, because of the difference in coefficients
of thermal expansion between the austenitic weld metal and the
ferritic parent steel. In this respect, nickel alloys are likely to be
advantageous.
When welding 9%Ni steels for cryogenic applications, nickel
alloy fillers are used almost exclusively and preheat is not required.
Nickel alloys are also used for welding other steels without preheat
in specialised applications. These include repairs to heavy power
station plant in alloy steels of the Cr-Mo and Cr-Mo-V types and to
a limited extent in underwater wet welding. Although solidification
cracking can be a problem if the steel contains more than a trace of
sulphur and dilution levels are high, nickel alloy fillers give rise
to much less severe problems with hard martensite at the fusion
boundary and therefore are unlikely to require preheat.
Because it gives large contraction strains compared with ferritic
steels, it is, however, often recommended that large repairs are
filled with weld metal of the 18Cr: 10Ni type after buttering on to
the ferritic steels with 29Cr: 9Ni. It must be emphasised that ferritic
weld metal should never be deposited on to weld metals of the
stainless steel or nickel alloy types.
High carbon, plain and alloy steels
Approximate compositions:
Carbon greater than 0.45%
For steels of this type, welding procedures are similar to those
already described for alloy steels. When using the temperature
control method (see earlier in this chapter), the preheat temperature
should be below that corresponding to the M, temperature of the
material rather than that predicted from the expected HAZ hardness.
It should not exceed the M, temperature since there would then be
a risk of the HAZ remaining austenitic and retaining hydrogen
during the PWHT. Little hydrogen would be lost during the PWHT,
because of its low diffusion rate in austenite (see Fig. 5.17), and
most would remain during the subsequent transformation to martensite on cooling. This would complete the conditions necessary
for cracking. The same considerations apply when austenitic consumables are used.
70
Welding steels without hydrogen cracking
Double tempering is usually necessary after welding: the first
temper removes retained austenite but converts it to martensite
which can be removed only by a second temper. It is an advantage
to cool slowly after welding to a temperature as low as possible
below the M, (usually 50-70°C minimum) before reheating for
tempering.
M, temperatures can be obtained from the steelmakers' literature,
from the collections of isothermal transformation data,4- 9 or from
formulae based on composition, as set out below (see Ref. 3).
M,
= 539 - 432C - 30.4Mn - 17.7Ni - 12.1Cr - 7.5Mo
[4.3]
For steels containing between 2 and 5%Cr, the following empirical
formula is more useful:
M,
= 512 - 453C - 16.9Ni + 15Cr - 9.5Mo - 71.4(C x Mn)
[4.4]
- 67.6(C x Cr) + 217(C)z
For steels containing nominally 12% to 18%Cr (10), with carbon
below about 0.3%, the following relationship may be used:
M,
= 540 - 497C - 6.3Mn - 36.3Ni - 10.8Cr - 46.6Mo
[4.5]
It may be an advantage to allow the HAZ to partially transform
isothermally during the first tempering to minimise the subsequent
formation of martensite from retained austenite, provided that
the isothermal transformation product has adequate mechanical
properties. In this respect, [4.6]-[4.8]3 are useful for estimating the
temperatures which lead to different degrees of transformation in
the HAZ during cooling prior to the initial tempering.
M;
= k1
Mx
M;
= 538
-
361C - 39Mn - 19.5Ni - 39Cr - 28Mo
[4.6]
k z - 474C - 33Mn - 17Ni - 17Cr - 21Mo
[4.7]
- k 3(361C + 39Mn + 19.5Ni + 39Cr + 28Mo)
[4.8]
=
where M; refers to the temperature leading to various degrees of
transformation; the appropriate values for k., k z, and k 3 can be
found in Table 4.4.
In tool steels which are extremely notch sensitive, two possibilities exist. The first involves welding at a temperature as close as
possible to the tempering temperature, slowly cooling below the
Ms , and then double tempering. The second involves re-austenitising
the component, quenching into" a lead bath at a temperature at
which austenite is stable for a sufficient period of time, and then
welding at that temperature. After welding, the component is
slowly cooled to below the M, and then double tempered. Welding
71
Welding procedures for different steel types
Table 4.4 Values of the constants k, k 2 , k, in [4.61-[4.81 which allow the
calculation of temperatures for various degrees of transformation
x
Ms
M,o
MOll
Mm)
M""
k,
k2
k"
538
561
1.0
513
551
1.084
488
514
1.18
452
458
1.29
416
1.45
Mr
346
steels with alloy contents high enough for them to be essentially
austenitic, and welding cast irons" are outside the scope of this
book.
Machinable grades of steel
Such steels contain additions of sulphur within the range
0.10-0.50%, whilst lead and selenium may be present also. In PD
970: 2001 the free-cutting grades are numbered 212M36 216M44
214M15 and 606M36. Up to 0.12% of lead may be added to any of
the PD 970: 2001 steels by agreement. It should also be noted that
the high sulphur free-cutting carbon steels' contain manganese up
to a maximum of 1.2-1. 7%, and should therefore be treated as C-Mn
steels.
The presence of sulphur renders these steels liable to liquation
cracking in the HAZ, and such cracks, although usually short and
relatively harmless in themselves, can act as nuclei for hydrogen
cracks. For this reason, and because sulphur, lead, etc, is picked up
by the weld pool, machinable grades of steel should not be used in
joints where full strength is required. Basic consumables should be
used and low dilution welding techniques employed to minimise
sulphur pick-up. Bead sizes should be small to reduce the size of
liquation cracks. Apart from these precautions, procedures for
avoiding hydrogen cracking should be those appropriate to the
normal grades of these steels.
REFERENCES
1 Hart P H M: Weld Inst Members Report M/60172.
2 Brisson J et al: 'Study of underbead hardness in carbon and low alloy
steels'. Soudages et Tech Conn exes 22 (11/12) 1968 437-55.
3 Woolman J and Mottram R A: 'The mechanical and physical properties
of the British Standard En steels', Vol 1, 2, and 3. BISRA, London
1964/66/69.
4 Delbart G, Constant A, and Clerc A: 'Courbes de transformation des
aciers de fabrication Francaise', Vol 1-4. Inst. de Recherches de la
Siderurgie, St-Cermain-en-Laye, France.
72
Welding steels without hydrogen cracking
5 Wever F et al: 'Atlas zur Warmebehandlung der Stahle'. Max-PlanckInstitut fur Eisenforschung. Verlag Stahleisen, Dusseldorf 1954/56/58.
6 'Atlas of isothermal transformation diagrams'. US Steel Corp., Pittsburgh, USA 1951 and supplement 1953.
7 Atkins M: 'Atlas of continuous cooling transformation diagrams for
engineering steels', British Steel Corporation, Sheffield, 1977.
8 Roberts G and Cory R: 'Tool steels', 4th ed, ASM, 1980.
9 Vander Voort G F (ed): 'Atlas of time-temperature diagrams for irons
and steels', ASM, 1991.
10 Gooch T G: 'Welding murtensitic stainless steels'. TWI Res Bull 18
(12) 1977 343-9.
11 Cottrell C L M: 'Welding cast irons'. TWI, 1985,
Chapter 5 Removing hydrogen during
welding and heat treatment
Thus, I set Pen to Paper with delight,
And quickly had my thoughts in black and white.
For having now my Method by the end,
Still as I pull'd, it came: and so I penn'd
It down, until it came at last to be
For length and breadth the bigness that you see.
The techniques for predicting safe welding procedures already
described require, in a number of cases, an estimate to be made of
the extent to which hydrogen is removed from a weld by the
thermal treatments during and after welding. This chapter sets out
the method by which such estimates can be made.
The efficiency with which a given heat treatment is able to remove
hydrogen from a welded joint can be calculated when it is possible
to define the thickness of the material and a value for the diffusion
rate of hydrogen for the material and temperature concerned.
Hydrogen removal efficiency is calculated in terms of the time at
temperature necessary to reduce the hydrogen content at the centre
of the section thickness to any fraction of its original level. This
chapter thus contains a series of working graphs which provide
this information for a range of material thicknesses and a range
of temperatures in Fig. 5.2-5.15, the key to which is given in
Table 5.1. Any fabrication sequence can be broken down into units
of time-at-temperature so that the hydrogen removal efficiencies
throughout the history of a welded joint can be assessed rapidly
without the need for tedious calculation.
The basis upon which the working graphs have been constructed
and the factors to be considered when using them in practice are
discussed, and then illustrated by worked examples. In all the
examples it is assumed that hydrogen diffuses at approximately the
same rate in weld metal, HAZ and parent metal. The diffusion of
hydrogen from welds deposited using austenitic or nickel alloy
fillers is not considered.
74
Welding steels without hydrogen cracking
Table 5.1 Key to hydrogen removal curves given in Fig. 5.2-5.15
Joint
geometry
Simulated
by
Plate halfthickness,
mm
Butt weld
Infinite
plate
2.5
5
10
15
20
30
40
75
2.5
5
10
15
20
30
40
75
Fillet
weld
Infinite
cylinder
For heat treatment temperature. °C, of
100
use Fig. No.
150
200
250
300
400
650
5.2a
5.3a
5.4a
5.4b
5.4c
5.5a
5.5b
5.5c
5.6a
5.6b
5.6c
5.7a
5.7b
5.7c
5.8a
5.8b
5.8c
5.4d
5.5d
5.6d
5.7d
5.8d
5.4e
5.4f
5.11a
5.11b
5.11c
5.5e
5.5f
5.12a
5.12b
5.12c
5.6e
5.13a
5.13b
5.13c
5.7e
5.7f
5.8e
5.8f
5.14a
5.14b
5.14c
5.15a
5.15b
5.15c
5.11d
5.12d
5.13d
5.14d
5.15d
5.11e
5.11£
5.12e
5.12f
5.13e
5.13f
5.14e
5.14f
5.15e
5.15f
5.2b
5.3b
5.3c
5.2c
5.9a
5.9b
5.3d
5.10a
5.10b
5.lOc
5.9c
5.10d
Construction of hydrogen removal curves
Experiment shows that diffusion of hydrogen out of metals, under
most circumstances, obeys Fick's second law and can thus be
described mathematically by a general differential equation:
OC= D\7 2 C
ot
[5.1]
This law describes the relationship among hydrogen concentration
in the steel, C, time t, and spatial distribution characterised by \7 2 .
It is of similar form to the differential equation which describes
heat flow in solids.
Precise solutions of the general equation [5.1] for many different
boundary conditions have been given by Barrer.! However, for
order of magnitude calculations in practical problems, approximate
solutions are adequate since these are more easily handled and
relate to simpler geometric forms. Comprehensive sets of solutions
have been given by Russell- and by Darken and Curry." In both
cases tables and graphs were developed showing the form of
evolution curves for different geometric shapes.
Figure 5.1, taken from Russell's paper, shows diagrammatically
the percentage of the original hydrogen concentration left at the
centre of various geometric shapes as time, t, progresses. Time
appears in the dimensionless expression, Dt/LZ, which forms the
abscissa: D is the appropriate overall diffusivity coefficient for the
75
Removing hydrogen during welding and heat treatment
Table 5.2 Loss of hydrogen from different specimen geometries as a function of time
(after Russel-)
Dt/V
Original hydrogen remaining at centre, 0/0
Infinite
plate
thickness
= 2L
0.01
0.02
0.04
0.05
0.06
0.08
0.10
0.16
0.20
0.24
0.30
0.32
0.40
0.50
0.60
0.70
0.80
0.90
1.00
1.20
1.40
1.50
1.60
2.00
2.40
2.80
3.20
3.60
92.0
84.1
69.4
63.0
57.1
46.8
38.5
23.5
14.3
5.3
2.0
0.7
0.3
0.10
0.04
0.01
Infinite
cylinder,
radius
=L
Finite
cylinder,
radius
=L
Square rod,
dia. = length
= 2L
Cube,
side
= 2L
96.0
92.0
84.0
80.0
76.1
68.6
61.5
52.5
66.8
57.9
34.7
23.6
41.4
28.5
19.6
10.4
25.4
13.6
10.9
6.2
3.4
1.9
1.1
0.6
0.3
0.1
0.03
0.02
0.01
4.5
2.0
0.9
15.5
9.4
5.8
6.5
3.1
1.5
0.2
2.2
0.3
0.03
0.8
0.3
0.1
0.08
0.02
temperature and material under consideration, and L, the critical
dimension of the solid, represents the half-thickness of plate, rod,
or cube, or alternatively, the radius of cylindrical forms. Since Dis
quoted in cm'' sec"? care must be taken to use the appropriate units
for t and L. For practical purposes it is possible to treat as infinite
plates both rectangular slabs whose width and breadth each exceed
three times the thickness and rectangular sections where the ratio
of sides exceeds 2: 1, as long as it is the hydrogen concentration
at the geometric centre that is required. Actual values for the
percentage of hydrogen remaining at various values of Dt/L 2 are
listed in Table 5.2, which again has been derived from Russell's
paper. The amounts of hydrogen remaining after a particular heat
treatment, or alternatively, the time and temperature required to
76
Loss of hydrogen from
different specimen geometries
as a function of time (after
Russell''). - - infinite plate;
- - infinite square rod;
- - - infinite cylindrical rod;
- - - finite tube;
finite
cylinder.
Welding steels without hydrogen cracking
5.1
100
90
Ifl. 80
~8
.
70
0Ll
c:: 60
:~
~
~
e
"0
50
40
E
e;; 30
.S
~,
\\
:~\\\\
' \.
\\
\
".\
\~
\.
\\
'C
o 20
...
, \~
....
'
"'"
....
.,,\~
,
\,
"
.........
................
-.~, <,<,
........
...................... <,
10
o
-,
".\
"\\'
'.\
-,
0Ll
-,
-;.....-::-:.
0.2
~;.;~ .......
~.-::
~.-.,.
0.6
......
--0.8
1.0
reduce the hydrogen concentration to a chosen low level, can now
be easily calculated.
The initial weld metal hydrogen concentration is not known at
this stage but is defined as the 100% level. For a particular temperature, D is known, and for a specific welded detail, L (mm) is
known, since it is the half-thickness of a plate or the radius of a
cylinder. The value of Dt/U can now be calculated for a number
of different values of time, t. Estimates of percentage hydrogen
remaining in the weld are obtained from the values of Dt/U and
Fig. 5.1. Thus, it is now possible to plot a new curve relating
percentage hydrogen remaining to the time, t, which is specific for
the particular value of D and L employed. Since, at any given
temperature, D can have numerous values (depending on material
composition and condition), the calculation is repeated for both the
maximum and minimum possible values of D for ferritic material to
produce a band for the hydrogen removal curve at each temperature
and thickness. The behaviour of any material in practice is unlikely
to fall outside this band so that its limits can be taken to represent
the most optimistic and the most pessimistic estimates of hydrogen
removal efficiency.
The calculation is repeated for the complete range of temperatures
and thicknesses; the results being the collection of curves shown as
Fig. 5.2-5.15.
77
Removing hydrogen during welding and heat treatment
80 I--+-;-++--'
70 I--+-;-++--t-'
60 1--+--14+--+--'
50 1--+--14+--+--\-'
401--+-;4+--+---'t--~
30 I--+-+-+-+----lr---+--+-Iti
20 I--+-+-++--+----ll--t+-
]
:~
101---1--1f--++--;--t-+t-t--
o
.................1.I.Iou....;
0'--~-J...L.I..u.LU._-'-
10-
2
2 3 45
2
Max total evolution lime
Min total evolution,time
5.2
Infinite plate 100°C.
5.3
Infinite plate 150 °C.
10
5
L = 2,5
15
20
30
40
50 75mm
90 r---r--r=
ll\!
i
~
OJ
00
0::
'a
'0/
e
e
0::
'"
a
~
..<:
]
'60
'C
0
80
70
60
50
40
30
20
10
2
3 45
Max total evolution time
Min total evolution time
2
34 5
1
10
2
I2 3 4 5
:~~
L = 2.5
5
10
15
102 I 2 3 4 5
20
30
40 50
103
75mm
2
78
90
Welding steels without hydrogen cracking
r-""lE£'!ml!!!1El
80 J--f---'.
~
8
70
1;1
60 1---i--+-+1f-
:(
50
e
40 J--+-+-++---+'"
I
I----i--i~
30
t--t-+-t+--'-~
t--t-+--H--t--W.EIa
1
20 ~+-~+----J.-._
;5
1Ot--+-t--H--t--t-'!!ftI;
'60
o............1.............&...1.1.1.0l.Io...................1...
3 4 5
10- 2
Max total evolution time
Min total evolution time
5.4
10
5
L = 2.5
1520
30 40
50
75 mm
Infinite plate 200°C.
5.5 Infinite plate 250°C,
90
80
Ill!
g
8
60
:
50
70
Oi
'a
.~
40
c::
u
30
e
!-=
1'60
'C
0
20
10
0
10- 2
10- 1
Max total evolution time
103
2
, I 10
10
Min total evolution time
L = 2,5
5
10
15
20
3040
50
75 mm
2
345
79
Removing hydrogen during welding and heat treatment
90
80
70 r---+--+3Wii
60 r---+--+-+
50 I----1HH_+_
.~
.~
e
f
J
o
40
1---1--1-+-+-~
30 1----1f-H_+_-
20 "--+-+-t-+--+.0
1--+--+--+-+--+-.......
O _ _..........................._..I.o.lpililll
10- 2
2
3.45
2
345
Max total evolution time
Min total evolution time
L
= 2.5
5
10
15
20
30
40
50 75 mm
5.6 Infinite plate 300 "C,
5.7
Infinite plate 400 "C,
90
80
70
60
501---+-
40
I---+-+-~
30
20 I---+-+-++10
1---+-+-1-+--
2 345 10-\ I2
10- 2
Max total evolution time .
2 345
Min total evolution time I
L = 2.5
5
10
15 20
30
40
50
75 mm
80
Welding steels without hydrogen cracking
90
80
70
60
50
301--+-HiH
20 t--+-+~
10
o
10- 2
2 3 45
10-
1
Max total evolution time.
Min total evolution time
L = 2.5
5.8
Infinite plate 650°C.
5.9
Infinite cylinder 100 -c.
30
1520
10
5
40 50
75 mm
70t--t~-++
60 t--t~-++--...:
50 t--t-t-++--
401---t-H+--f30 II--+-+-~f--+--
20 t-~-1I-t-+--+-10 t---It-HH--+--+--
10- 2
10- 1
Max total evolution time
Min total evolution time
L
=
2.5
5
10
15
20
30
40
50 75 mm
81
Removing hydrogen during welding and heat treatment
90
~
e
80
;:
70
'OJ
60
.~
50
8
gp
'c
.,...
5tlIl
40
-e
30
]
20
e
>.
.c
tlIl
'1:
0
10
0
10-2
10- 1
Max total evolution time
Min total evolution time
L = 2.5
5.10
Infinite cylinder 150 "C.
5.11
Infinite cylinder 200 "C.
5
10
15
20
30
40
50 75 mm
90
80
~
i
u
'OJ
gp
'c
.~
2:!
.,"
e
]:
]
tlIl
'1:
0
70
60
50
40
30
20
10
0
10-2
10- 1
Max total evolution time
Min total evolution time
L = 2.5
5
10
15 20
30 40 50
75 mm
Welding steels without hydrogen cracking
82
90
80
70
50
I---+-+--\M
40 .---1--11--1-\
30
1----+--+-++
20
1---1-1-4-+--
10 .---lHI--I-+--
o ~"""'-.L.'-"'' ' ' ' ~.r..-_
10-·,
2 3 4 5
10- 1
2
Max total evolution time
I
34 5
I
1 12345
2345
II
I 1
I I
5
10
10 1
I
'I
Time, hr I
2 345
12345
2
10 t •
I ,I 1
:
Min total evolution time
L = 2.5
5.12
Infinite cylinder 250 o e.
5.13
Infinite cylinder 300 oe.
15
20
30
40
10
2 3,45
50
75 mm
90
80
70
60
50
~---11--
40
~---11--1-'
30
1---+-+-+
20 I----+--+_+_+_
10
.---l--l-++--
o '----'--'-.L...u..........L.\.-....;
2
10- 2
1
Max total evolution time I
Min total evolution time
34 5
I
I
I I
I
121 345
I
I 1 Time, hr l
LZI:Z~~~I
L = 2.5
5
10
15
1
I I
1
20
30
40
102
I
I I
50
2
75 mm
3 4 5
I
Removing hydrogen during welding and heat treatment
83
90
~
i
W
til
~
80
70
60
'S
.~
so
c
40
'tl
30
~
'60
20
0
10
t
>.
.c
'1:
0
10- 2
103
Max total evolution time
Min total evolution time
L = 2.5
10
5.14
Infinite cylinder 400°C.
5.15
Infinite cylinder 650°C.
IS
20
30 40
SO
75 mm
90
~
i
til
80
70
60
~
'S
SO
c
e
40
-e
~
30
.~
t
!
'1:
0
20
10
0
10- 2
345
I
I
I
I
I
tz' ;7
L=2.5
3 45
10
i
Max total evolution time
IS
20
30 40 50
75 mm
84
Welding steels without hydrogen cracking
Simplification of weld joint geometry
It should be noted that Fig. 5.2-5.8 refer to hydrogen removal from
infinite plates and Fig. 5.9-5.15 refer to infinite cylinders. These
two simple geometric forms have been chosen for several reasons.
Firstly, solutions of the general equation for diffusion are readily
available for these forms which, because of the simple boundary
conditions involved, yield expressions which are easily handled
when repetitive calculations are required. Secondly, as can be seen
from Fig. 5.1 in which Dt/L 2 v. t has been plotted for several
geometric forms, infinite plate geometry produces a distinct curve,
but other geometries yield relationships very close to that shown by
an infinite cylinder, which therefore can be taken to represent a
mean form. This step is further justified because a range of D values
is used to produce a hydrogen removal 'band', and this band also
encompasses deviations from infinite cylinder geometry which may
exist in a practical case. Finally, it is possible to show that any type
of welded joint can be simulated by either infinite plate or infinite
cylinder geometry for the purposes of discussing heat flow and
hydrogen diffusion. Although such an approximation oversimplifies
real behaviour, it has been used successfully in other areas of study
and is adequate for order of magnitude calculations such as are
dealt with here.
When using Fig. 5.2-5.15 it is therefore necessary to decide
whether the welded joint under consideration is best simulated by
plate or cylinder geometry.
The majority of possible joint forms for welded details are defined
in BS 499 : Pt 1 : 1991. With only a few exceptions, all joints are
completed by making either a butt or a fillet weld. In general, butt
welds are used to join components lying in one plane and finally
present two opposite paths in the same plane for hydrogen diffusion. Butt welds may thus be simulated by plane plate geometry.
Fillet welds on the other hand are used to link components lying in
planes at right angles, thereby presenting diffusion paths which are
radial to the source of hydrogen.
Fillet welds are thus best simulated by cylindrical geometry.
Certain configurations do not fit this simple subdivision: T butt,
cruciform butt and corner butt welds, for example, are best considered as fillets simulated by cylindrical geometry. No ri,8id subdivisions can be made in such cases and each problem must be
considered individually taking into account material thickness.
Roberts and Wells" have shown that in a single-pass weld with a
'characteristic' of between 0.25 and 1.0, the thermal cycle approxi-
85
Removing hydrogen during welding and heat treatment
mates to that in infinite plate if the actual plate has a half width of
ten times the bead width. The weld characteristic is defined as Vdl
4a where V is welding speed, d is bead width, and a is the thermal
diffusivity of the material. Values of this characteristic lying between
0.25 and 1.0 cover the majority of welding conditions, so that this
criterion may be used to determine the plate widths for which the
infinite plate assumption is applicable to partial penetration butt
welds.
Rectangular slabs whose width and length each exceed three
times the thickness can also be considered as infinite plates. The
plane plate simplification can thus be extended to fully penetrating
butt welds using this criterion.
In a single bead fillet weld (i.e. simulated by cylindrical geometry)
the fusion boundary occupies between 60 and 70% of the total
circumference of the bead and diffusive release of hydrogen will be
predominantly radial. The justification in these terms becomes less
clear in the case of fillet joints consisting of multirun welds. However, from Fig. 5.1 it can be deduced that the ratio between the
removal efficiencies for plate and cylinder is approximately 0.6.
This is paralleled in the concept of combined thickness", used in
Chapter 3 and Fig. 4.2 to distinguish between the relative cooling
potentials in welds with butt and fillet geometries. Here again the
ratio between the two forms is approximately 0.6.
In a study designed to compare experimentally measured temperatures in multirun welds with those predicted by classical heat
flow theory, Rykalin" found that agreement was good in butt welds
where plane plate symmetry had been used in the heat flow argument. The same treatment for T, lap, and cruciform joints produced
discrepancies which could, however, be removed by applying a
correction factor of 0.6 to the calculated heat input. Such a correction would have the same effect as applying cylindrical geometry in
the calculations of predicted temperatures, and it is significant that
these were necessary only for fillet welds.
As a general working guide the following recommendations can
be made:
For simple butt welds: assume infinite plate geometry, and use
Fig. 5.2-5.8.
For fillet welds and T-butt welds: assume infinite cylinder geometry,
and use Fig. 5.9-5.15.
For complex joint configurations try both simplifications or consult
TWI.
86
Welding steels without hydrogen cracking
Material thickness
Hydrogen removal curves have been constructed using plate halfthickness, L(mm). When using the curves in practice for a butt
welded joint, simulated by infinite plate geometry, the choice of
value for L is easy, as long as the two plates being joined are of
similar thickness. Similarly for a fillet welded joint, simulated by
cylindrical geometry, L (although this time, representing the radius
of a cylinder for the purposes of calculation) can be taken as the
half-thickness of the plates constituting the legs of the fillet.
Difficulties arise when the two plates in a butt weld differ greatly
in thickness or when a fillet weld cross-section is large in relation
to plate thicknesses. No general rules can be formulated to deal
with every possibility and where such difficulties occur, judgement
must be exercised. A feature of the graphs, however, is that for a
given heat-treatment temperature the effect of changing thickness
or geometry on hydrogen removal time can be quickly established
by comparing curves. For example, there will be a point when the
difference in thickness between the two plates in a butt weld
becomes so great that the joint is better simulated by cylindrical
geometry, because hydrogen diffusion is able to proceed in a predominantly radial fashion. The existence of removal curves for both
situations enables this possibility to be quickly investigated.
Each curve on a graph will be seen to be labelled with the halfthickness, L, to which it refers, but to avoid overlapping not all
the curves are drawn in. Those that are not drawn can easily be
identified by inspection, since within each group the curves are
parallel. In addition, below each graph a scale is drawn locating the
position on the time axis for 100% removal, * i.e. none of original
hydrogen remaining. For example, in Fig. 5.2 the scale positions for
L = 2.5 mm coincide with the feet of the curves for L = 2.5 mrn,
Similarly the scale positions for L = 5 mm identify the feet of the
curves for this value which, however, are not drawn but which
would produce a parallel band. Hydrogen removal efficiencies can
therefore be obtained for all the material thicknesses listed, either
directly or by interpolation.
Heat treatment temperature and the choice of value for D
It will be noted that each hydrogen removal curve consists of a
band and that the lower the temperature ;the wider the band. This
• In fact, -99% removal, as the curves in Fig. 5.1 approach the abscissa
asymptotically.
87
Schematic illustration of
the use of hydrogen removal
curves. At time txhr the
percentage of original hydrogen
remaining at the centre of the
material will be between Hz
max and Hz min.
5.16
Removing hydrogen during welding and heat treatment
Heat treatment temperature T, °c
Material half-thickness. L,mm
Time, hr - - - - .
arises because the diffusion rate of hydrogen, D, varies not only
with temperature but also with material composition and condition.
It is, however, possible to state with some confidence both the
maximum and minimum values of D at a particular temperature.
By using these extremes the maximum and minimum hydrogen
removal times can be calculated for each set of conditions and, as a
result, bands appear in Fig. 5.2-5.15. Each band may be interpreted
as showing both the most optimistic and the most pessimistic times
required to remove any fraction of the original hydrogen level for
a specified temperature and thickness. Conversely, the probable
maximum and minimum levels of hydrogen remaining after a fixed
time can be assessed; see, for example, Fig. 5.16. Seven heat treatment temperatures have been dealt with in this way and are listed
in Table 5.1. It is extremely unlikely that the D value of a steel will
be such that the time predicted will fall outside a band on the
graph and, in general, at each temperature shown, the centre of
each band represents the most likely behaviour. However, particuis such that
larly at low temperatures, the width of each ba~d
the possible hydrogen removal time indicated extends over one
order of magnitude and it may no longer be possible to say that
the most likely behaviour is predicted by a mean value. It then
becomes important to define more precisely the D value of the steel
concerned.
88
Variation of overall
diffusivity coefficient, D, with
temperature.
5.17
Welding steels without hydrogen cracking
lOOO/T where T is temperature,
4
3
"K
2
o 10 2030405060 70 80
100
150
Temperature, °c
200 250 300,350400 500 600
The way in which the diffusivity coefficient varies with temperature is shown in Fig. 5.17. This figure can be used to select values
of D for all future calculations. Table 5.3 lists approximate values of
the overall diffusivity coefficient for a number of materials typical
350
400
450
470
550
600
650
20
30
35
40
45
50
60
65
70
80
°C
Temperature,
2.5
X
10- 4
1.6 x 10- 4
1.4 x 10- 4
X
X
2.3
7.0
10- 6
10- 6
10- 6
1.8 x 104
1.4 x 10- 4
X
0.003C
0.0045
0.004C
0.0058
1.8
Fe-6Cr
alloy,
BI8RA
pure Fe,
1.0
6.4
7.6
8.5
3.3
2.0
X
X
X
X
X
X
104
10- 5
10- 5
10- 6
10- 6
10- 6
0.06C1.5Mn
steel,
0.012S
1.1
8.6
6.9
7.8
X
X
X
X
10- 4
10- 5
10- 5
10- 7
3.4 x 10- 7
8.9 x 10- 8
5Cr-MoVoW
steel,
0.34C
0.012S
X
10- 9
9.9
7.5
2.0
X
X
X
10- 5
10- 5
10- 7
5.2 x 10- 8
5.6
0.0068
Fe-0.75C
alloy,
1.0
X
X
X
1.2
X
X
X
1.3
5.8
X
X
7.5
2.9
10- 4
10- 5
X
X
X
10- 6 1.7
7.5
10- 7 2.8
X
X
X
X
10- 6
10- 7
10- 7
10- 8
0.0088
0.2C-2Ni
steel,
10- 8 5.9
8.2 x 10- 5
10- 6 1.4
10- 7
10- 7 2.9
10- 7 7.8
0.0055
0.0068
X
0.4C-1Ni
steel,
0.2C-1Ni
steel,
10- 9 1.0
Freecutting
steel,
0.07C
0.255
10- 9 3.0
Freecutting
steel,
0.09C
0.238
10- 7 6.8
O.lOC
0.018S
Rimming
steel,
Table 5.3 Variation of overall diffusivity coefficient, D [em- sec-'), with composition (approximate values only)
90
Welding steels without hydrogen cracking
5.18 Effect of carbon, X, and
sulphur, ., on overall
diffusivity coefficient for
hydrogen at 20°C.
10- 5
I
~
"'S
o
5
4
3
2
o
0.02 0.04 0.06 0.08 0 1 0.12 0.14 0.16 0.18 0.2
Sulphur, %
1\
0.22 0.24 0.26 0.28 0 3
-
\.
.,
10- 6
\
'\.
"
...
::e...
"-,
........
.-
..........
.......
~ 10- 8
g
r-..... ...........
5
4
3
2
10- 9
...........
I'-...
- - -.......... ........
<,
...
;-..
e_
....
Carbon, %
o
0.1
0.2 0.3 0.4
0.5
0.6 0.7
0.8 0.9
1.0
1.1 1.2
1.3 1.4 t.5
of those investigated. This table gives rounded values only, with
no estimate of precision, but results in research show that, over the
whole range of D values, the magnitude of two standard deviations
is about 10% of the mean value quoted.
Although the range of different materials is relatively small, certain trends can be detected in Table 5.3, and these are used as a
basis for selecting values for the band in Fig. 5.17. A pure vacuummelted iron tends to show relatively high values of D within the
band and, with very low carbon and sulphur levels, D may be as
high as 10- 6 cm 2 sec-t. Increasing amounts of carbon and sulphur
exert a strong depressive effect on the overall diffusivity coefficient
at a given temperature. Carbon is the most potent depressant and
the effect is most marked at low temperatures. The situation at
20°C is shown in Fig. 5.18. Here values of D for different steels
have been plotted against percentage carbon and percentage sulphur
separately, but using scales so that the graphs coincide. The effects
appear to be additive and suggest that with more results a simple
formula relating carbon and sulphur content could be found which
would place all the points on a smooth curve. Figure 5.18 indicates
I
91
Removing hydrogen during welding and heat treatment
already that materials with carbon greater than 0.1% and sulphur
greater than 0.02% are unlikely to have D (20°) values greater than
10- 7 em" sec" ", The effect of additions of alloying elements such as
manganese nickel, chromium, molybdenum, etc, is also to lower
the D value, but to a much smaller extent than carbon and sulphur.
Other work has examined the effects of chemical composition,
grain size, microstructure, dislocations, inclusions and microvoids
on the overall diffusivity value at selected temperatures. References
to the more important of these studies, together with full details of
the experimental procedures for determining D values, have been
reported by Coe and Moreton." Although such features of a material
undoubtedly influence the specific value of D at any temperature,
the variations which occur still fall within the envelope of Fig. 5.17
so that, in cases of uncertainty, a maximum and minimum D value
can be chosen and calculations made to define both optimistic and
pessimistic conditions.
No curves have been provided for ambient temperature evolution.
The width of a curve for a particular plate thickness at a given
temperature is directly proportional to the spread of diffusivity
coefficients at that temperature, as shown in Fig. 5.17. Therefore
any curves which could be drawn at room temperature would
consist of a series of very wide bands which would give very
disparate pairs of maximum and minimum hydrogen removal times.
However, it is possible to construct hydrogen removal curves for
ambient temperatures, using the procedure described below. In
such cases it is necessary to assess each problem individually,
considering material composition and the probable overall diffusivity coefficient so that more information about plate and consumables is required.
1 From a knowledge of plate and weld composition define a new
range of possible diffusivity coefficients, D, within the limits
shown in Fig. 5.17 at 20°C.
For example, a low carbon mild steel with a very 'clean'
matrix might have limits:
D~~
= 2 x 10- 6cm 2sec- I
Dii3~
= 10- 7 cm 2 sec- I
A low alloy material with 0.35%C might have:
D~~
=2 X
Dii3~
=
10- 7 em" sec- I
10- 8 cm 2 sec- 1
92
Welding steels without hydrogen cracking
2 Also from Fig. 5.17 read off Dmax and Dmin at some other, higher
temperature where the band is narrower and where curves are
available, Fig. 5.2-5.15.
For example, at 200°C, for any material:
3
D~ao;c
=5
D~~oc
= 1.5
X
10- 5 ern" sec- 1
X
10- 5 crrr' sec- 1
For a given plate thickness and weld configuration the evolution
time is directly proportional to the diffusivity coefficient. Therefore the following proportionality constants, Cmax and Cmin' may
be defined to be used in conjunction with the 200°C curves:
ZOO°C
Dmax
Cmax = DZo0C
max
Cmin
ZO,O°C
Dmill
= DZo'0C
mm
For the low alloy material described:
5 X 10- 5
Cmax =
7
2 X 10Cmin =
4
1.5
X
=
10- 5
10- 8
250
= 1500
From Fig. 5.4 (infinite plate) or Fig. 5.11 (infinite cylinder) an
estimate of the ambient temperature removal time can be
obtained from the 200°C curves.
For example, with an infinite plate configuration with L =
20 mm, time for total hydrogen removal at 20°C will be between
t max and tmin (hr):
t~~
= Cmax X 35
= 250 X 35hr
= 8750hr
t~:~
= Cmin X 100
= 1500 X 100 hr
= 150 OOOhr
Therefore the time for total evolution from the centre of a 40 mm
plate for the low alloy material will be between about one and
seventeen years.
Calculations of hydrogen removal at ambient temperature are of
little practical value for thick sections, and this statement is borne
out by the results of Worked Example No.1, later in this chapter.
Choice of value for total original hydrogen level
Each of the main graphs, Fig. 5.2-5.15, records the percentage of
the original hydrogen remaining at the centre of the weld after
various times at the given temperature.
Since hydrogen is removed by diffusion, its distribution will
93
Removing hydrogen during welding and heat treatment
be non-uniform and its level will be lower at points nearer the
surface. It would have been possible to use percentages on the
ordinates which related to mean hydrogen levels through the thickness, but presentation would have become more complicated and
would have resulted in over-optimistic estimates of removal times,
especially from thick sections. The approach that has been adopted
possibly errs in the opposite direction but, in doing so, compensates
a tendency to select an over-optimistic D value when using the
curves in practice.
The values on each ordinate, being in percentage units, apply
irrespective of whether the original hydrogen level is high or low.
The curves themselves take the same form since diffusive movement of hydrogen is dependent on geometry and temperature but
not on total concentration. Although the curves are valid whatever
original level exists initially, it is obviously of great importance to
know what the 100% level stands for in each practical case. For
example, if it is found necessary to reduce hydrogen levels to less
than 5 ml/100 g, this could require 80% removal in a welding
process producing 20 ml/1oo g originally, but only 50% removal in
a process producing 10 ml/100 g originally; the time at temperature
would clearly be different in each case.
It would be desirable to know, firstly, the initial hydrogen levels
in practice, particularly those occurring in multirun deposits, and
secondly, the critical hydrogen levels below which the risk of
cracking is minimised. It must be admitted that so far it has proved
impossible to measure these levels. The weld metal hydrogen
levels listed in Appendix A refer to single bead deposits produced
by a standard test method; their prime function is to compare
qualities of various consumables. In production welding, the
weld bead would not be quenched and would lose an appreciable
amount of hydrogen in the first few minutes after the arc had
passed and before the next bead had been deposited (see Worked
Example No.2). Additionally, the weld metal hydrogen content
would not be that given by the standard test (Appendix B) for
deposited metal, but should be derived from the fused metal
weight, allowing for dilution by parent steel or previously deposited weld metal. In multirun deposits the situation is even more
complicated, bearing in mind the presence of hydrogen in the
preceding beads.
The levels listed in Appendix A cannot therefore serve as a valid
starting point for hydrogen removal calculations.
There are however several pieces of evidence which enable ruleof-thumb procedures to be prepared. Attempts to measure hydrogen
/
94
Welding steels without hydrogen cracking
levels in multirun welds have suggested that the level after welding
is less than 30% of the figure given by the standard single bead test
which refers only to deposited weld metal. This level is compatible
with a dilution of say 40%, with a loss of 50% of the hydrogen
from the original fused metal concentration. At the same time, in
laboratory work employing hydrogen-charged specimens to study
embrittlement mechanisms, levels of 5 mll100 g and less are sufficient to provoke embrittlement. Work on hairline cracking in
forgings has shown that critical levels of hydrogen lie in the range
2-5m1l100g. It can therefore be argued that hydrogen levels in
real welding situations are likely to be only 50% of those listed in
Appendix A and that such a value should be used as the starting
point for hydrogen removal calculations. In addition, reduction
of this initial hydrogen content to a level of a fraction of a millilitre of hydrogen per 100 g should ensure that sufficient hydrogen is removed to minimise the risk of cracking in the majority
of cases.
It cannot be too strongly emphasised that these are intended to
be conservative and rule-of-thumb criteria, and they should be
adopted only after a careful appraisal of all other relevant factors as
explained elsewhere in this book. Despite these reservations, experience over a period of several years at TWI in dealing with this type
of problem suggests that they represent the best recommendations
which can be made at present.
Use of hydrogen removal curves in practice
It is often useful to make a preliminary and very rough estimate of
heat treatment times to indicate whether hours, tens of hours, or
hundreds of hours are likely to be involved, and this can be done
by using Fig. 5.19a which refers to infinite plate geometry, i.e, butt
welds, and Fig. 5.19b which refers to cylindrical geometry, i.e, fillet
welds. For different plate half-thicknesses and different heat treatment temperatures the approximate times to remove 75% of the
original hydrogen can be read from these graphs. In constructing
each graph a mean value of the diffusivity coefficient of hydrogen
at each temperature has been assumed. If the steel composition is
such that the D value departs from a mean level, as discussed
earlier, the times indicated in these graphs will not be very accurate,
particularly at temperatures below 170 "C.
Figure 5.19 refers to the removal of 75% of the hydrogen. In some
cases a 50%, or even a 25%, removal may prove to be all that is
required to reduce the likelihood of cracking. The times required
95
Heat treatment temperature.
100
e-,
f-
ISO 200
~ ~
1"0..
°c
s:: "'" "
2SO 300 400
.....
~'"I".
r-o
Removing hydrogen during welding and heat treatment
650
"~
i"""'oo.........
"-.
100
I
SO
l'.
I" ~ ..... ""'
""""-. ........ ~
~" ........
...... ...... """"-.......... .... r-; ~ i'-.................
.......
<,
!'ooo .....
~
"'"
r-; ...."..... t--.... t::::: ~ !'o o~ .... ;::: --,
r-,
r':
""""-. ~ ~
~ t:::::
........
........ ........
I-
f-- H
~
I
1000
(a)
SOO
100
Heat treatment temperature.
100
~ ............. ....
ISO
.....
~
II
SO
1000
....
...... ::""00..........
r""'-- .................. ~~ ...... .... i'-......
...... r-....: ....... ........ ........
.......
........... ~
.......
<, ........
1"'- ...... ~
1"'- ......
~
, I'....::~
<,
I
I
SOO
100
I.
ll:
]
~
2
I •
0.1
0.5
~
~
SO
........
10
20
....r-oi=:::
-..,: ~
""
"" ......
....
~
-c- .............. <,
'""'" ~
..........
-.
~-.
.......
<, ...............
I
5
e
e
...i
r--... """"-.
II
SO
-- 100
IT
II I I
I I
6SO
Infinite cylinder geometry (fillet welds)
III
5
5
<o!.
Time. hr
.
°c
.......
.... "-..
......
.... .......: ~ ~,
•• I
10
200 2SO300 400
.....
~iV
1111 •
I
J
~
1"'- .... ....
i". ......... ............... ~
Infinite plate geometry (butt welds)
10
20
~~
f-
~
...i
~
10
......
..........
~ ...... ....
........
....
0.5
....
...... """-:' .................. ..........
....... """"'" ~
-=5
~
,-.
...... ....
vi
B
:§
<o!.
]
~
ll:
2
0.1
Time. hr
(b)
5.19 Approximate times for removing 75% of original hydrogen: (a) butt welds. (b) fillet welds.
Note: These diagrams may be used for making estimates of heat treatment times and temperatures for
removing 75% of the original hydrogen from different steel thicknesses. The lines have been calculated on the
basis of average values of the diffusivity coefficient for hydrogen at each temperature. Since the coefficient
varies with steel composition. particularly at the lower temperatures. the lines drawn may not represent real
behaviour for all cases. In addition, it may not be necessary to remove as much as 75% of the hydrogen
present in every instance. The diagrams should be used only for preliminary estimates, and the full
techniques described in the worked examples shoud be employed before any final decision is made.
to achieve these lower degrees of hydrogen removal can also be
estimated from the figure as follows:
1
Time for 50% removal is approximately one-half of indicated
time for 75%.
96
Welding steels without hydrogen cracking
2 Time for 25% removal is approximately one-quarter of indicated
time for 75%.
It must be strongly emphasised that the use of Fig. 5.19 gives only a
preliminary assessment of each problem and does not provide a
complete basis for fabrication decisions. This figure should be used
to supplement, but not replace, the full technique which takes into
account the particular details of each problem.
The more detailed use of the graphs and tables given in this
chapter to establish suitable heat treatment times and temperatures
is best illustrated by means of worked examples. Two such examples are given which are based on typical industrial problems that
have been referred to TWI in the past.
Worked Example No. 1
914 rad
(a)
(b)
(c)
5.20 Detail of Worked
Example no. 1; (a) actual detail,
(b) simplification to plane plate
geometry, and (c) simplification
to cylindrical geometry;
dimensions in mm,
Nozzles are to be fitted into a thick walled low alloy steel pressure
vessel by manual metal-arc welding using basic electrodes. The
detail is shown diagrammatically in Fig. 5.20a. The specification
requires a preheat of 150°C applied locally before and during
welding, and the total welding time for several such nozzles is
estimated at 50 hours. After completion of welding, the vessel is
given a PWHT at 650°C for approximately six hours. The waiting
time at ambient temperature before stress relieving is on average
two weeks. It is necessary to decide whether the weld metal contains
significant amounts of hydrogen after this treatment.
1 Simplification of joint geometry
For this particular detail, simplification of the geometry is difficult.
At first sight, because there are three paths of diffusion, it would
appear that the joint could be simulated by cylindrical symmetry.
However, since the radius of the vessel is large and since the nozzle
wall thickness is equal to the half-thickness of the vessel wall, a
reasonable approximation could be made by considering diffusion
out of a plate of half-thickness 75 mm. For the purposes of the
worked example both approximations will be investigated; they are
illustrated in Fig. 5.20b and c. Accordingly, in Table 5.4, where
estimates are collected, headings for both plate and cylinder simplifications are given.
2 Choice of the value for D
The material of the vessel is a low alloy steel for which the precise values of D at various temperatures are not known. Values
representing maximum, most probable, and minimum diffusion
97
Removing hydrogen during welding and heat treatment
Table 5.4 Estimates of the percentage of the original hydrogen remaining at the centre of the weld after various heat
treatments for Worked Example No. 1.
Heat treatment
time and
temperature
Estimate of original hydrogen remaining, %
Cylinder geometry
Plate geometry
Optimistic
Most
probable
Pessimistic
Optimistic
Most
probable
Pessimistic
50hr at 150°C
(Fig. 5.3d & 10d)
336 hr at 20°C
(see text)
6hr at 650°C
(Fig. 5.8f & 15t)
16 hr at 650°C
(Fig. 5.8f & 15t)
80
90
>90
66
82
>90
Cumulative effect
50hr at 150°C +
6hr at 650°C
Or 50hr at 150°C
+ 6hr at 650°C
Ineffective in removing hydrogen
81
88
>90
68
75
87
51
64
75
30
40
64
65
79
>80
45
62
>78
41
58
>68
20
33
>58
rates can be chosen on the basis of the information given earlier.
These lead to the three subheadings, 'optimistic', 'most probable',
and 'pessimistic' for the estimates of hydrogen remaining in the
weld listed in Table 5.4.
3 Use of the hydrogen removal curves
In Table 5.4 three heat treatments are considered. The first refers to
50 hr at 150°C and lists the percentages of hydrogen likely to remain
after welding is completed. Figures 5.3 and 5.10 are used to deal
with plate and cylinder geometry, respectively, at 150°C. On each
figure, band (d) is used to correspond to 75 mm, half-thickness of
the material. At the 50 hr mark on the abscissa a vertical line cuts
the bottom, midpoint, and top of this band to give three ordinate
values of percentage hydrogen remaining. These values provide the
'optimistic', 'most probable', and 'pessimistic' estimates listed in
Table 5.4.
The effect of two weeks (336 hr) at room temperature is calculated
using the proportionality constants relating room temperature evolution to that at a higher temperature, as set out on pp. 91-92. From
the 336 hours at 20°C, optimistic and pessimistic estimates can be
obtained for the equivalent times at, say, 200°C. For a low alloy
steel as considered on pp. 91-92, the optimistic equivalent time for
hydrogen removal at 200°C is:
98
Welding steels without hydrogen cracking
time at 20°C
Cm ax
336hr
250
=--
= 1.3 hr
From the curves in Fig. 5.4 (infinite plate) or Fig. 5.11 (infinite
cylinder), it may be seen that 1.3 hr at 200°C (equivalent to 336 hr
at 20°C) is ineffective in removing hydrogen.
Similarly, a pessimistic equivalent time for hydrogen removal at
200°C is:
=
time at 20°C
=
336hr
1500
Cm in
= 0.2hr
This will also have a negligible effect on hydrogen in both the plate
and cylinder cases.
The PWHT of 6 hr at 650°C is handled in the same way as the
150°C treatments (above) using Fig. 5.8 and 5.15.
Since rapid hydrogen removal is encouraged by higher temperatures, and since such a large vessel will also take a considerable
time to heat up and cool down during stress relieving, a further
estimate is made in Table 5.4 by postulating that as much as 18 hr
altogether may be spent at or near 650°C. For this final estimate Fig.
5.8 and 5.15 are again used but using the 18 hr instead of the 6 hr
vertical.
4 Interpretation of results
Actual behaviour in practice is likely to produce the hydrogen
levels listed in the subheading 'most probable' in Table 5.4. The
other two estimates represent the likely extremes of behaviour and
would occur only if the material possessed unusually high or low
D values at these temperatures. Thus, only the 'most probable'
estimates would be considered at this stage, recognising that similar
arguments could be developed for the extreme values if the problem
demanded them.
Each estimate refers to a heat treatment considered individually.
The cumulative effect of successive heat treatments can be assessed
as:
99
Removing hydrogen during welding and heat treatment
For plate geometry, original hydrogen level
= 100%
after 50 hr at 150°C = 90%
after 6 hr at 650°C = 0.88 x 90%
= 79%
Thus, for this sequence, 79% of the original hydrogen remains after
stress relieving. The last part of Table 5.4 lists results of this type
for both geometries.
The effect of employing different simplifications of the joint
geometry is clearly demonstrated. The second sequence of heat
treatments, where it is assumed that 18 hr are spent at or near
650°C, is probably more realistic in view of the size and thickness
of the vessel, and it is also likely that real behaviour will tend to
compromise between plate and cylinder geometries. Between 40
and 50% of the original hydrogen present will therefore remain at
the centre of the weld after fabrication is complete.
The only remaining problem is to decide whether 40-50% of the
original level represents a dangerous concentration. Information on
the hydrogen introduced during welding, together with a knowledge of the embrittlement susceptibility of the steel being welded,
is needed at this point. Some advice on these questions has been
given above.
This worked example illustrates the use in practice of the removal
curves given in the main text. Clearly the problem could be dealt
with in greater detail by considering the extremes of likely behaviour
indicated in the first part of Table 5.4. In addition, the sequence of
heat treatments could be further subdivided to correspond more
precisely to actual fabrication procedures and to include the effect
on the preheat level of heating cycles due to welding. The effect
of a local postweld heat treatment above 150°C could also be
examined.
Regardless of the degree of detail introduced into the problem,
the techniques of calculation remain the same and all the necessary
removal curves are provided in this book.
Worked Example No.2
A typical detail of thermal history which could be of importance in
a calculation of hydrogen removal is the interpass time in multirun
weldments. For example, a short multirun weld in a V preparation
could be made with a negligible time between consecutive runs,
thus resulting in minimal hydrogen loss from the weld area as a
whole, while the welding is proceeding.
100
Welding steels without hydrogen cracking
Table 5.5 Estimation of the percentage of the original hydrogen removed from
individual beads of a multirun weld during the interpass period; Worked Example
No.2
Conditions
Maximum
hydrogen removal
Minimum
hydrogen removal
60%
25%
18-30%
[i.e. 70-82% remains)
8-13%
[i.e. 87-92% remains)
10 mm diameter cylinder
L = 5mm,
T = 250°C,
t = 0.17hr;
First bead:
use Fig. 5.12b
Subsequent beads:
assume loss from a surface of
only 30-50% of that employed
in Fig. 5.12b
[i.e, 40% remains)
[i.e, 75% remains)
Suppose, however, multirun welds were made around the circumference of a large pipe so that there was a waiting time of 10 min
and an interpass temperature of 250°C; then, at a given point on the
pipe, before the next bead was laid over the last, a considerable
amount of hydrogen could be lost during this interval. This hydrogen lost to the atmosphere, or transferred to underlying layers, is
difficult to calculate because of the varying temperature and thickness of the overall weld as the joint is built up.
Nonetheless, an approximate and conservative estimate of the
hydrogen loss during the time taken to make the weld can be made
by calculating the amount lost from an individual weld bead during
the 10 min interpass time at the interpass temperature of 250°C.
For the purpose of estimation, it is assumed that each weld bead
is a 10mm diameter cylinder. It is also noted that, whilst the
hydrogen removal from the first bead indicated in Fig. 5.12b does
take place, some of the hydrogen removed, possibly 50%, is effectively transferred to the adjacent plate. The hydrogen removed from
the first bead is given in Table 5.5. Subsequent weld beads will
have the same initial hydrogen content as the first bead, but will
lose less hydrogen into the underlying material since this already
contains some hydrogen. Hence the loss of hydrogen from the
subsequent beads will largely be restricted to the free metal surface,
typically 30-50% of the cylinder surface. The net removal of
hydrogen will therefore be only about 30-50% of that lost from the
first bead. The hydrogen removal estimated on this basis for the
second and subsequent beads is 'given in Table 5.5.
101
Removing hydrogen during welding and heat treatment
This calculation becomes the first step in estimating hydrogen
removal during the welding operation. Any subsequent calculations
refer to the loss of hydrogen from the centre of the completed weld,
and the complete depth of weld metal, as shown in Example 1.
REFERENCES
1 Barrer R M: 'Diffusion in and through solids'. Cambridge Univ Press,
Cambridge 1951.
2 Russell T F: 'Some mathematical considerations on the heating and
cooling of steel'. Iron and Steel Inst, Special Report No. 14, 1936, 14987.
3 Darken L S and Gurry R W: 'Physical chemistry of metals'. McGraw-Hill
Book Co Inc, New York 1953.
4 Roberts D K and Wells A A: 'Fusion welding of aluminium alloys, Pt V
- A mathematical examination of the effect of bounding planes on the
temperature distribution due to welding.' Brit Weld J 1 (12) 1954 55360.
5 Burdekin F M: 'Heat treatment of welded structures'. Weld Inst 1969.
6 Rykalin N N: 'Calculating the parameters of the thermal cycle in the
parent metal during multipass arc welding'. Izvest Akad Nauk SSSR,
Technical Sciences Section, No.1 and 2, 1950, 233-48.
7 Coe F R and Moreton Mrs J: TWI Members Report M/49/70, June 1970.
Appendix A Typical hydrogen levels
Introduction
The distinction between potential hydrogen levels and weld hydrogen levels was explained in Chapter 1, and this appendix provides
the required details. The procedures for making such measurements
are described in Appendix B. It is emphasised that both this discussion of typical hydrogen levels and also prediction procedures
explained in the main text are based on:
the combustion methods for determining flux moisture contents,
the single bead, rapid quench method either with mercury as
the collecting fluid at room temperature or with a direct analysis
at 650°C for the measurement of weld hydrogen levels, and
3 the encapsulation method for determination of the hydrogen
content of welding consumables by analysis at 700°C.
1
2
Other testing procedures are used to produce information on
potential and weld hydrogen levels because of the demand for
simpler methods for production control purposes. Such methods
may not necessarily afford adequate precision, and it is recommended
that any decisions regarding welding procedure or comparison of
consumables should be based on information provided by the standard tests, or methods calibrated to them.
Potential hydrogen levels
This type of measurement is intended to reveal the amount of
hydrogen that is potentially available for absorption by the weld
pool during welding. It therefore provides a means of characterising
the quality of a consumable with respect to hydrogen before that
consumable is put to use. In electrode coatings and submerged-arc
fluxes the moisture content is measured. For flux-cored and solid
wires the hydrogen content of an encapsulated sample is measured.
The potential hydrogen levels revealed by such tests for different
consumables are compared below and an indication of the typical
range of values in each case is given.
103
Appendix A
"i~
Covered electrodes
.,. cellulosic type
Covered electrodes
rutile types
Covered electrodes,
basic type, as-received
and re-baked
Agglomerated
submerged-arc fluxes
Fused submerged-arc fluxes
0.02
At Typical potential
hydrogen levels (moisture
levels in coatings and fluxes).
(Heights of curves indicate
relative frequency of test results
for each type of consumable.)
0.05
0.1
0.2
0.5
1.0
2.0
'Moisture' content, % by weight
5.0
10.0
Figure A1 compares moistures levels of electrode coatings and
submerged-arc fluxes. (The histograms show the relative frequency
of measurements over the total range.) Typical levels of moisture in
both agglomerated and fused submerged-arc fluxes are shown, but
it is emphasised that in this process the filler wire is an additional
source of potential hydrogen.
There are relatively few data points for cellulosic and rutile
covered electrodes, since these are of inherently high moisture
content and have not been widely studied in terms of hydrogen
control. Moreover, it is not possible to make satisfactory measurements of moisture in coatings containing a high proportion of
cellulose (which itself contains a high proportion of hydrogen).
Potential measurements on welding wires are not widely made
nowadays on a routine basis, and so current data for these are not
presented. With open arc welding processes, e.g. manual metal-arc,
it has been shown that the atmospheric moisture content can
influence the level of weld hydrogen produced. In general, the
increase in weld hydrogen from atmospheric conditions is small,
about 1-2m1l100g deposited metal, and clearly this will have a
greater influence when weld hydrogen levels are low, i.e, <5 mll
100 g, than when they are above 10 m1l100 g. Particular care should
be given when assigning a weld hydrogen level to a given consumable when the consumable naturally gives very low hydrogen
levels and when it is known that welding is likely to take place in
atmospheres of high humidity, particularly when the ambient temperature is also high, because more water can be held in air at
higher temperatures.
104
Appendix A
Weld metal hydrogen levels
This type of measurement is intended to reveal the amount of
hydrogen absorbed by the weld metal during welding. It is much
more commonly used than potential hydrogen measurements for
characterising the welding consumable, under standard conditions,
in terms of its subsequent weld hydrogen level. It requires a single
bead to be deposited under carefully standardised conditions (see
Appendix B for details) and to be quenched immediately the 'weld'
is complete. It is thus clear that the hydrogen levels measured tend
to represent maxima and do not necessarily represent the levels
persisting in welded joints. However, as long as test conditions are
standardised, the levels measured provide a scale of values which
allows a decision to be made, using previous experience, on whether too much hydrogen is entering the weld. Such decisions are
strictly applicable only within each welding process, since the
specimen size, welding current, and thermal cycle vary slightly
from process to process and these factors influence the absorption,
retention, and release of hydrogen.
Nevertheless weld hydrogen measurements are being increasingly
introduced into standards as a means of classifying consumables
in terms of weld hydrogen levels. The value of weld hydrogen
so obtained allows the consumable to be fitted into one of the
hydrogen levels used in the present text, and thus serves as an
entry point into the method described earlier for deriving guidance
on welding conditions to avoid hydrogen cracking. It should be
noted that it has become normal practice to use diffusible hydrogen
values, which are expressed in terms of the weight of deposited
A2 Typical weld hydrogen
levels.
Notes: Cellulosic electrodes
cannot be shown on this
diagram, as they normally give
weld hydrogen levels between
70 and 100ml/l00g. Heights of
curves indicate relative
frequency of test results.
Very low Low Medium
High
t---+--t---+--.----,---r-----l
Covered electrodes,
rutile coating
Cored wires
Submerged-arc
Covered electrodes,
~~!!l!Il!Il_+---+---I---+-----J
o
5
10
IS
basic coating
20
25
30
Weld hydrogen, mLlIOOg of deposited metal
105
Appendix A
metal in the standard tests described later. This type of value is
used throughout, unless otherwise stated.
Typical values obtained for different welding processes and consumable types are shown diagrammatically in Fig. A2. The lower
values are, in general, obtained with cleaner wires and drier fluxes
and electrode coverings. Baking the electrode or flux, where appropriate, will also, in general, give lower hydrogen levels.
It must be emphasised that the aim of Fig. A2 is merely to
indicate the ranges of typical values, and is no guarantee that
particular levels will be achieved. In general, the cleaner the wire
the lower the weld hydrogen level for continuous welding processes
General relationships
between potential hydrogen
and weld metal hydrogen
levels.
A3
1.0
0.6
0.5
0.4
200 0.3
.~ 100
~
90
'- 80
o 70
bl)
60
§ 50
]
40
]
30
c
~
].
:§
C
~
20
101----"'"
~
7
6
5t--7i'#fiii!iif;ififi!
4
3
2.....----1---+------'1----+--------+---1
High
2
4 5 6
8
10 12 14 IS 16.18
20 22 24
Weld hydrogen level, mLlIOOg of deposit
26
28 30 '32 '34
106
Appendix A
without fluxes, but the type of flux, and its moisture content, have
a major impact on weld hydrogen levels. Thus fused submerged
fluxes usually give lower hydrogen contents than agglomerated
fluxes unless the latter have been manufactured by a route intended
to give particularly low levels of hydrogen.
It is explained in the main text that when predicting safe welding
procedures it is frequently necessary to decide whether the welding
process chosen or the particular consumable used produce low,
medium, or high hydrogen weld metal. It was also pointed out at
the beginning of this section that it is impossible to define specific
levels which ensure low risks. Nevertheless, it is essential to
indicate the ranges of values which are considered to be low,
medium or high, and these are drawn at the foot of Fig. A3. Thus,
for example, when a process is referred to as 'low hydrogen' elsewhere in this book the implication is that the standard test would
show a hydrogen level of 5-10mLll00g, but a 'very low hydrogen'
level would be indicated by results less than 5 mL/l00 g. Figure A3
then gives some indication of the quality of consumables necessary
to achieve this level.
Interpretation of results and their use in practice
A study of the figures in this appendix indicates which welding
processes may be described as low hydrogen and also which condition of a consumable will ensure this. Measures, such as the
selection of clean wires and the use of dry flux, for improving the
quality of welding consumables have already been mentioned in
the main text and this appendix provides information against
which to gauge the success of these measures in practice. The
testing procedures defined in Appendix B must, of course, be used.
In many cases, as is explained in the main text, a reduction in the
hydrogen potential of the welding process alone will not be sufficient to eliminate the risk of cracking. In such cases, measures for
removing hydrogen during and after welding must be considered,
these are dealt with in Chapter 5.
Appendix B Techniques of hydrogen
measurement
1
The determination of moisture in electrode coatings and
welding fluxes
Principle
The sample is ignited in a stream of pure dry oxygen, and the water
expelled or formed by combustion in this operation is collected in a
suitable absorbent, and weighed. Alternatively, the water determination may be made by physical or chemical methods, provided
that these are calibrated against the method of wei~hing.
These methods are not capable of detecting any hydroxyl ions
which may be trapped in the silicate lattice in fused fluxes.
Sampling
The electrode or flux to be tested shall be given such treatment
before sampling as has been agreed upon or specified by the manufacturer, e.g. predrying under specified conditions, or storing under
specified conditions.
A sample of sufficient size for two separate determinations (about
10 g total) shall be taken from the coating of two separate electrodes
by bending the electrode rods, or by means of pliers or other
suitable tools. In the case of fluxes, a representative sample of
smaller weight shall be taken, avoiding the top layer of flux and any
fines from the bottom of the flux container.
The sample shall be transferred to a glass-stoppered weighing
bottle immediately; the operation of weighing shall be performed
with the stopper in position.
Apparatus
The equipment (Fig. B1) shall consist of:
(a)
A supply of oxygen and a purification train to give a blank
determination less than 0.5 mgH 20/hr. For less pure oxygen, a
108
Appendix B
Oxygen purification
train if necessary
Combustion furnace
Oxygen supply
Translucent silica
furnace tube
Combustion boat
Absorption bulbs o- 30 g each
containing magnesium perchlorate
81 Apparatus for
determination of moisture in
electrode coatings and fluxes.
catalyst, e.g. CuO at about 350°C, and a water-absorption bulb
in series shall be employed. The drying agent shall be the
same as that used in the analysis.
(b) A furnace capable of maintaining a temperature of not less
than 900°C and not more than 1000°C throughout the combustion zone.
(c) A silica tube, preferably of the translucent type, and combustion boats of silica, porcelain, or nickel. Before use the
combustion boats are ignited in air at a temperature about
100°C above that employed in the analysis and then stored in
a desiccator.
(d) If fluorides are present in the sample they may be trapped by
covering the specimen in the boat with ignited magnesium
oxide.
(e) Two absorption bulbs containing magnesium perchlorate, each
of a size not exceeding 30 g weight in the filled condition. The
bulbs are connected to the exit end of the furnace by a twoway stopcock for alternate use.
Procedure
The blank value for the apparatus using oxygen at 100 mllmin is
determined under normal operating conditions with an empty boat
in the combustion tube. Normal blank values are in the range 0.1 to
0.2 mg HzO/hr; blanks higher than 0.5 mg HzO/hr indicate that
oxygen purification is required.
A weighed sample of about 5 g is transferred from the weighing
bottle to a previously ignited boat, care being taken to avoid
109
Appendix B
unnecessary exposure to the atmosphere during the process of
transfer. The boat is placed in the combustion zone without delay
and left there for 30 min in a stream of oxygen at 100 mLimin.
The increase in weight of the absorption bulb is determined on
an analytical balance, and the blank correction is subtracted. The
net increase is reported as grams of total water per 100 g coating, or,
alternatively as mL hydrogen (STP) per 100 g coating. The conversion
factor is 1245 mLig water.
Reproducibility of procedure
The precision to be expected in the application of the recommended
procedure may be represented by a standard deviation of the order
of ±10% when applied to a predried sample. In the range tested,
from about 0.2 to about Ig water/l00g coating, this corresponds to
standard deviations in the range from 0.02 to 0.10g water/l00g
coating. When undried electrodes or fluxes are used, the results are
likely to be less consistent.
2
The determination of shielding gas moisture content
Under normal circumstances, the shielding gas makes a very minor
and insignificant contribution to the hydrogen potential of a welding
process. However, the presence of several metres of on-line plastic
tubing may introduce moisture by diffusion from air through the
tubing wall, or a nearly empty gas cylinder may contain atypically
high moisture contents. If information is required, it is important to
take readings on a dynamic system at the gas flow rate to be used in
welding and with the normal gas delivery lines in place. An
electrolytic hygrometer is the most suitable instrument for this type
of measurement. Results should be expressed as ppm of moisture
in the gas. The relation between moisture content, dewpoint and
relative humidity is plotted in Fig. H2.
The water vapour concentrations plotted in Fig. H2 are expressed
in terms of vpm and mg/L and related to both dewpoint and relative
humidity. They refer specifically to a gas at DoC and are independent
of the nature of the gas. For most practical purposes Fig. H2 can
be considered to represent the relationships at room temperature
(20°C), except for relative humidity, and for this expression an
additional ordinate in Fig. H2 gives values for any gas at about
room temperature (20°C).
If the expression of water vapour content in weight concentration
units as opposed to volume (vpm) is required, the nature of the
110
Appendix B
82 Relationship between
moisture content (vpm),
dewpoint and relative humidity
for any gas at O°C and 760torr.
s.o
I
22 ..... S3
/
/
/
4.3 1--16.6'
*-
2.2 I=-S.3
o~N
*-
'iii
·s~
:I
.c
·i'"
~
~
0.43
C-1. 66 -
=I
0.22 =-0.S3~
I
'iii
.~
V
~:s
.c
OJ>'"
/
~
I
0.022 -O.OS
-so
2.h
1-
I
II
/
r 0.6
0.4
I
-=
::
r--
0.1
2
O.OS- 10
~
t:
s
~r-
ti
>
-e
e
1?
o~
E
e,
9
0.2
e
os
I
3
O.S- 10
os
I
"c::l
II
0.043 ----0.166
I
9
I
"il
:=
/
/
0.4-
I
/
~ I
~.b
I
- 0.06 :
o~
ti
~
1>0
>,
e
os
I
-0.04-
'-
I
E
0
B
e
- 0.02
0
u
I
0.01
O.OOS- 10
/
0.006 :
a
~
<Il
'15
~
0.004
I
I
- 0.002
~
-60
-40
Dewpoinl of gas,
-20
I
0.001
o
1
+20
°c
particular gas in question must be considered. The following simple
conversion table provides factors for rapid calculation:
Conversion table:
0.000045
0.000040
0.000089
0.000064
0.000056
111
Appendix B
3
The determination of the hydrogen potential of welding
consumables by encapsulation
Theory
The sources of hydrogen in steel wires are hydrated oxides and
entrapped drawing compounds. Considering the oxides, equations
may be written of the type:
2Fe + 3H20
~
Fe203 + 3H2
From the free energy values for the oxidation of the metal and the
formation of water, the extent of the reduction can be calculated at
any temperature using the Van't Hoff equation. At equilibrium, at
any given temperature, the partial pressure of hydrogen in a closed
system (such as a mild steel capsule) will be governed by the
equilibrium constant of the system. In the nickel and iron alloys
the production of hydrogen is thermodynamically unfavourable
and increasing the temperature of the reaction does not improve
the situation. Kinetic factors also influence the efficiency of reduction. For iron alloys, the continuous removal of the reaction product, that is hydrogen, by diffusion through the walls of the mild
steel vessel may outweigh the unfavourable equilibrium position,
although there is a danger that the testing time may be controlled
by the rate of reduction and prove longer than the time for diffusion.
Encapsulation techniques
Capsules are fabricated from mild steel which will allow fast
hydrogen effusion at elevated temperatures. The geometry and
dimensions can be varied to suit the facilities available for both
sealing and analysis. Alternative capsule designs are shown in Fig.
B3. Types (a) and (d) require sealing by resistance welding, type (b)
has a press fit lid which is sealed by autogenous T1G or plasma
welding while keeping the capsule cool, and type (c) has a screw
lid with a degassed steel sealing washer. Design (d) is the most
widely used, having the advantage that no machining of the steel
tubing is required. In all cases, capsule thickness is limited to about
1.5 mm maximum to minimise hydrogen effusion time through the
capsule walls. Whichever design is adopted, the capsule and seal
must be capable of withstanding an internal pressure of approximately 500kN/m 2 (5 bar) developed during testing. In all cases,
most particularly at low hydrogen levels, it is important to run
blank tests on empty capsules.
112
83 Alternative forms of
capsule.
Appendix B
V 7 7 77 A
r»
(7).
1/
1/
1/
1/
1/
1.1
1.1
w
~:::::~
v
00
~
(d)
Using type (d) capsules the procedure is:
Capsules must be thoroughly degassed in vacuum for 1 hour at
800°C before use and allowed to cool in vacuum. Subsequent
handling must be with tweezers or gloved hands to avoid
contamination.
2 When sampling a reel of wire, the outside layer should be
avoided, since this may be contaminated by handling, or contact
with packaging. It is good practice to cut 25 mm lengths of wire
and collect them directly in a glass container. The wire must not
be handled. If correlation is sought with weld metal hydrogen
content, the wire sample should be taken at the nozzle tip after
it has been through the wire feed mechanism and guide tubes.
3 Wire samples are weighed and stored in a desiccator.
4 For a typical size of capsule, a partition is projection welded
halfway along a 150 mm length of 13 mm tubing, and five 25 mm
lengths of wire of known weight are loaded into the tube on
either side of the weld.
5 Air is expelled by a blast of dry argon while the final sealing
welds are made approximately 38 mm on either side of the
first. The two capsules so formed are cut from the tube with a
fine clean hacksaw and stored in a desiccator while awaiting
analysis.
1
Analysis of encapsulated specimens
Because the steel walls of the capsule act as a filter, allowing only
hydrogen to be evolved for measurements, both vacuum heating
and inert gas carrier methods are applicable. An analysis tempera-
113
Appendix B
ture of 1000°C is satisfactory. Fusion techniques are unsuitable
because of the large volume of gas sealed into the capsule with the
sample.
4
The determination of diffusible hydrogen in ferritic steel
weld metals
Scope
The method described is intended for assessment of diffusible
hydrogen in ferritic weld metal deposited by common arc welding
processes (see Note 1, p. 122). The techniques used are embodied in
international standard BS EN ISO 3690: 2001, and are incorporated
in many national standards, e.g. BS EN 499: 1995. Before use, the
text of the appropriate standard should always be consulted.
Principle of method
The consumables to be tested are used to deposit a single weld bead
which is rapidly quenched. Both the welding and quenching
processes are carefully controlled. The specimen so produced is
maintained at room temperature for a sufficient time to release its
content of diffusible hydrogen, which is measured by volumetric
methods and reported on unit mass of either deposited or fused
metal (Note 2).
The total hydrogen content is defined for the present purposes as
the sum of the diffusible and residual contents, the latter being determined by hot extraction at 650°C under vacuum or in a carrier gas
as described below. The results may be expressed in terms of the
hydrogen content per unit weight of either deposited metal [i.e. the
weld metal deposited on to the test specimen during the test) or
fused metal (i.e. the metal deposited plus the parent metal in the test
sample which has been melted by the welding operation).
Materials required for test
(a) Parent plate material
The testpiece assembly shall be prepared from a mild steel containing not more than 0.20%C, 0.35%Si, 0.05%S. Prior to use the testpiece shall be degassed at a temperature and time sufficient to
remove all hydrogen, e.g. 650°C for 1 hour. If these conditions are
not suitable, reference should be made to Fig. 5.2-5.8 to select an
alternative heat treatment.
114
Welding fixtures and
testpiece assemblies for
hydrogen sampling of weld
deposits; dimensions in mm:
a) For heat input levels up to
2 kJ/mm; b) For heat input
levels from 2 to 4 kl/mrn: (i)
Fixture and test assembly:
length 'I' depends on process
and heat input; (ii) Use of
copper foil to maintain
constant flux depth with
submerged-arc welding.
B4
Appendix B
140
I~
A
.Copper block
I
Water cooling
channels may be used
I
I
45
30
45
[C(j
-«(c((CCCCCC~·
100
approximately
(a)
A
A-A
Testpiece assembly
Testpiece ~
assembly Iiillill
Water cooling--channels
Materials A - Copper
B - Mild Steel
(b (i)
Flux
Imm copper foil, 40mm x 300mm
II-
Welding fixture
Testpiece assembly
(b (iij)
(b) Consumables
Consumables to be tested shall be prepared according to the information wanted from the results of the test. For example, manual
metal arc electrodes or submerged-arc flux would normally be dried
115
Appendix B
Testpiece assembly
showing deposited bead.
B5
Run-off piece
according to the manufacturer's instructions: however, different
conditions might be used in an experimental study of the effect of
consumable treatment on weld metal hydrogen content. When a
hydrogen classification test is being carried out, the requirements of
the relevant standard should be followed.
(c) Welding fixtures
Copper jigs, as shown in Fig. B4, are used for the alignment and
clamping of the testpiece assembly illustrated in Fig. B5, The choice
of fixture and testpiece assembly depends on the arc energy involved
and the resultant weld pool size.
Apparatus for the determination of diffusible hydrogen
An example of a gas burette for the measurement of cold extracted
gas is shown in Fig. B6. Burettes of other designs may be employed,
provided that the following requirements are fulfilled:
(a) Mercury is to be used as a confining liquid.
(b) Using the two-way stopcock, it should be possible to maintain
the sample under vacuum for a brief period to prevent transfer
to the capillary of foreign gases trapped on the fractured surface
of the sample. In burettes consisting of a single limb, this may
be achieved through manipulation of the mercury level and
the stopcock, any gas released during the brief period of
surface degassing being swept out of the burette before the
measurements.
(The technique used to transfer the sample to the capillary
limb of the V-tube burette in Fig. B6 without ingress of air is
illustrated in a TWI video.)
116
Burette of Y-tube design
with stopcock for collecting
diffusible hydrogen;
dimensions in mm.
Appendix B
D6
To air
-fi'---""""""J
=5[]U
To vacuum
29/32 socket
Two way glass vacuum stopcock
~
29i32oo~
w15".2"~-
Elevation of side arm in direction A
(c) It should be possible to read the volume of collected gas with
a precision of at least 0.01 mL (STP) using a cathetometer
as illustrated in Fig. B7 to measure the mercury level to an
accuracy of ±0.02 mm.
Preparation of test specimen
Collection and
measurement of room
temperature diffusible
hydrogen.
D7
1 Testpiece assembly
Triplicate sets of testpieces shall be welded for each brand of
electrode to be tested. Beads of 100 mm overall length shall be
deposited along the centreline ofthe testpiece assembly. No burningoff prior to the testing is allowed.
117
Appendix B
The testpiece assembly shown in Fig. B4 consists of run-on and
run-off pieces and a central sample section (Fig. B5). The determinations shall be made using the entire length of this section.
The testpiece dimensions indicated in Fig. B4 shall be maintained
within the limits ±0.25 mm; however, each set comprising run-on
and run-off pieces and the central section shall be finished in one
operation of grinding so as to ensure a uniform width. The central
specimens shall be marked and weighed to the nearest 10 mg.
2 Welding
The temperature of the jig shall be 25 ± 5°C prior to testing. The
use of water cooling channels in the copper blocks simplifies this
aim, and facilitates a rapid turnround time for each sample. The
welding conditions selected depend on the objective of carrying out
the test.
Normally, the welding current shall be the maximum stated by
the manufacturer less 15 A, the machine setting being controlled
within a tolerance of ±5A (Note 3). The speed of welding shall be
adjusted to an electrode consumption between 12 and 13 mm per
10 mm bead length. The time spent in deposition shall be noted.
Immediately after the extinction of the arc, the jig is released
and
I
the testpiece assembly is quenched in iced water and subsequently
in alcohol or acetone saturated with solid carbon dioxide. Using
brief intermittent periods of cooling to maintain a sub-zero temperature, the sample pieces are vigorously wire brushed to remove
all traces of slag and surface oxide, and are broken apart while cold.
The samples may now be stored at dry ice or liquid nitrogen
temperatures until required for analysis.
Test procedure
1 Preparation of specimen for analysis
The sample is transferred from the cold storage to the Y-tube as
quickly as possible using the following procedure, which typically
takes 90 sec:
(a) Remove the sample from cold storage using clean tongs.
(b) Raise the temperature of the sample to just above O°C by
immersion in water until the ice skin first melts.
(c) Wash the sample with acetone to remove water.
(d) Wash the sample with petroleum ether (80-100°C fraction) to
remove acetone.
(e) Dry the sample in warm air for a few seconds.
118
Appendix B
Place the sample in the V-tube and evacuate the air above the
sample via the two-way stopcock.
(g) Using a magnet, transfer the sample to the mercury in the
capillary limb of the Y-tube: this is done whilst maintaining
the vacuum.
(f)
Analytical procedure (diffusible hydrogen)
The sample is maintained under reduced pressure in the Y-tube at
25 ± 5°C until hydrogen evolution ceases, when the volume of
hydrogen collected is measured. If a classification test is being
undertaken, then the specified hydrogen collection time is used
(e.g. 72 hr in BS 6693: 1986 Part 1), measuring the hydrogen after
this time rather than when evolution ceases. After measurement
of the collected hydrogen, the sample is removed from the V-tube
and reweighed to the nearest 10 mg to determine the weight of
deposited metal.
If it is also desired to determine the amount of fused metal,
the cross-section of the bead shall be measured on tracings or
photographs of the fractured faces. These measurements are made
after hot extraction if the residual hydrogen content is also to be
assessed.
2
Calculations and expression of results
The volume measured is converted to O°C and 760 mm. This
volume, divided by one-hundredth of the gain in weight, is the
diffusible hydrogen content, in mL/100 g deposited metal.
The fused metal weight is obtained by multiplication of the gain
in weight by the average ratio (fused metal/deposit area) determined
on the two faces of the specimen, and the diffusible hydrogen
content, in mL/100 g fused metal, is found by dividing the gas
volume (STP) by one-hundredth of this weight.
Each separate content shall be reported, together with the average
for each of the three samples tested, the time of collection, the
welding conditions and time of welding, the name, type and
condition of the consumable. The atmospheric temperature and
humidity (absolute or relative) at the time the weld samples were
deposited should also be reported.
Reproducibility (Note 5)
Because of the high mobility of hydrogen in steel, particularly at
elevated temperatures, the procedures laid down should be strictly
119
Appendix B
followed to ensure constant sampling conditions. A TWI Study
Group reported in 1982 on the reproducibility of the measurement
of diffusible hydrogen in manual metal-arc weld deposits. It was
concluded that the precision of the Y-tube measurement was good;
however the variability of results was ±10% of the hydrogen content
at the level of 10mL/100g of deposited metal, and arose from variations in individual electrodes or in welding procedure.
5
Alternative rapid tests for determination of diffusible
hydrogen levels
Scope
BS EN ISO 3690: 2001 allows for tests other than the mercury test
to be used. A different type of test can be performed if the results
are proven to be reproducible and in line with the results found
during mercury tests at room temperature. Thus it may be permissible for samples to be heated to increase the effusion rate of hydrogen in order to gain results more quickly. A comprehensive study
performed at TWI* to measure diffusible hydrogen content from
shielded metal are, flux cored wire and submerged arc welding has
shown that temperatures up to 400 0 e may be used. These results,
generated using carrier gas (both static collection and constant flow),
and vacuum hot extraction, agreed satisfactorily with the standard
room temperature mercury test. However, the risks of errors arising
from evolution of trapped hydrogen, reaction of evolving hydrogen
with surface oxide, and reaction of any moisture in the carrier gas
with the sample increased at higher temperatures.
Principle of method
The welds, see Figures B4 and B5, and their production, storage and
handling procedures are the same as those for the mercury standard
test. The difference is that a typical procedure for rapid assessment
of diffusible hydrogen content is to apply 150 0 e over 6 hours. Extraction at 400 0 e for 30-40 minutes is also possible, although higher
temperatures than these risk the release of significant amounts of
residual hydrogen. The removal of surface oxide is imperative for
extraction above 300 0 e to avoid its reaction with evolved hydrogen,
and dry carrier gas is essential at higher temperatures. Any method
may be used provided a proper correlation in terms of accuracy and
* N Jenkins, PH M Hart, D H Parker, An Evaluation of Rapid Methods for Diffusible
Weld Hydrogen, Welding Research (supplement to the Welding Journal), January
1997, pp. 1s-10s.
120
Appendix B
reproducibility has been established with the primary mercury
method. Calculations and expressions of results are as for the standard mercury test. The notes on reproducibility also apply.
6
The determination of residual or total hydrogen in
ferritic steel weld metals (Note 6)
Scope
Two methods are described for measuring the total or residual
hydrogen contents of ferritic weld metal: the residual hydrogen
content is that proportion remaining after measuring of diffusible
hydrogen as described above.
Principle of method
Single weld bead samples are produced from the consumables being
tested using the same approach as employed for assessment of
diffusible hydrogen. The samples are heated to 650°C to extract the
hydrogen, which is measured volumetrically using vacuum apparatus or an inert carrier gas technique. The samples can be produced
specifically for measurement of total hydrogen: alternatively, they
can be employed to determine diffusible hydrogen, followed by hot
extraction of the residual hydrogen.
Materials and test specimen preparation
The requirements for the base plates for the testpiece assemblies, for
the consumables under test and for the welding fixtures are exactly
the same as for the determination of diffusible hydrogen. Welding
is also carried out using the same procedure, with storage of
samples at dry ice or liquid nitrogen temperature until required for
analysis.
Apparatus and analysis technique
1 Vacuum hot extraction
As shown in Fig. Bd, the apparatus incorporates a furnace in which
the sample is heated to 650°C in vacuum, typically of <0.05 x 10- 3
Torr. Extracted hydrogen is transferred by a mercury diffusion pump
to an evacuated, calibrated analysis volume. After an initial pressure
measurement, hydrogen is removed from the analysis volume
through a heated palladium/silver osmosis tube, and the pressure of
121
Gauge
Appendix B
Gauge
Palladium tube
Expansion volumes
Furnace
10000001
Cold trap
pump out
Mercury diffusion pump
Vacuum hot extraction
equipment for the
determination of hydrogen.
D8
the residual gases is measured. The volume of hydrogen extracted
from the sample is calculated from the differential pressure
measurements.
The apparatus illustrated schematically in Fig. B8 is normally constructed of Pyrex glassware using standard size cone and socket
joints. Experience of laboratory glass blowing and handling is
required to assemble and operate the equipment.
2 Carrier gas hot extraction
The sample is heated in a furnace at 650°C in a stream of argon
carrier gas. The carrier gas removes hydrogen extracted from the
sample, and takes it through a katharometer cell. This measures the
change in thermal conductivity of the carrier as the argon/hydrogen
mixture passes through. The output of the cell is integrated, and the
measurement obtained is converted to a volume of hydrogen by reference to calibration derived from the injection into the carrier of
known volumes of hydrogen. For successful operation of the technique, it is essential to remove all traces of moisture from the carrier
gas and to maintain a constant carrier gas flow rate.
The equipment required is shown schematically in Fig. B9.
122
Appendix B
Gas out
I
Gas now indicator
Molecular sieve drier
Ar in
Carrier gas instrument for
determination of hydrogen.
89
BIBLIOGRAPHY
BS EN ISO 3690: 2001 Welding and allied processes. Determination of
hydrogen content in ferritic steel arc weld metal.
Coe F R: BWRA Bulla (1) January 1967 8-11, and IIW Doc-II-A-194-66.
Coe F R: Iron and Steel Inst Special Report, 1962, No 73, 111-22.
Coe F R, Moreton Mrs J and Parker D H: Weld Inst Report, Misc!13/68 and
Coe F R Metal Constr 1 (2) 1969 108--09.
Gayley C T and Wooding W H: Weld J 29 (8) 1950 629-35.
IIS/IIW-314-68. Weld in the World 7 (1) 19696-14.
ISO standard 3690-1977. 'Welding - determination of hydrogen in deposited
weld metal arising from the use of covered electrodes for welding and low
alloy steel'.
Jenkins N et al: TWI Report 9327, September 1985 and IIW Doc IIA 576-82.
Jenkins N, Dye C R and Parker D H: IIW Doc IIA 1019-84.
Jenkins N and Greenfield A A: British Steel Corporation Report No
CDLlCCAC/50174.
Johannes J P: Weld Res Council Bull (68) April 1961 8-15.
Laycock C: British Steel Corporation Report No MG/CC/549/71.
Video, 'Measurement of diffusible hydrogen in test welds using ISO 3690'.
TWI video recording, 1985.
NOTES
1 This technique was devised specifically for testing covered electrodes but
has been adapted to examine gas shielded welding processes with solid
and flux-cored wires and also the submerged-arc process. It is important
to note that results are not necessarily directly comparable between
processes (see Appendix A). A sketch of the copper jig used for the
submerged-arc process is given in Fig. B4b.
2 It is possible to quote the weld metal hydrogen in terms of mL hydrogen
gas at STP per 100g of fused metal, but this can be done only ifthe volume
of fused metal in the analysis sample is known. An estimate of fused metal
volume can be derived from examination of the cross-section of the weld
deposit as revealed on the ends of the sample by either optical metallography or an image analysing instrument. Fused metal hydrogen measurements may be valuable in investigating the effects of varying welding
123
Appendix B
conditions and process which may give different degrees of dilution of
parent steel into the weld metal, for example. However, classification
systems are based on controlled specific welding conditions and hydrogen values given in terms of deposited metal. This is because the weight
of deposited metal can be measured more accurately than the weight of
fused metal can be estimated.
3 With continuous wire welding processes, it is particularly important to
maintain a constant stickout when results need to be critically compared.
There is evidence that increasing the stickout reduces the hydrogen level
as a result of the increased resistive heating driving off more hydrogen
from the wire before it is melted.
4 The correct treatment of samples before hydrogen analysis is vital in order
to avoid loss of hydrogen or high surface blanks.
With weld metal deposits it is important to remove all slag from specimens before analysis in order to avoid spuriously large hydrogen figures.
This should be carried out by a brisk wire brushing treatment while the
specimen is still cold. Samples should then be degreased with acetone
and petroleum ether, and dried under vacuum or in a stream of dry gas.
The whole treatment should not take more than 2 min before the sample
is introduced into the analysis system, to avoid loss of hydrogen into the
atmosphere.
Even with a very clean sample, hydrogen evolution at room temperature could take in excess of 10 days. The presence of a surface oxide film
will further delay evolution, and low results may be obtained. If an oxidised sample is heated to 650°C to measure residual or total hydrogen,
low results will ensue because of the reaction of hydrogen with the oxide
to form water.
Vigorous abrasion of specimens should be avoided, since hydrogen
absorption by the specimen may result. It has been found that abrasion
of a mild steel surface by 120 grade silicon carbide paper resulted in
0.0049 ml.zcm" of hydrogen being absorbed by the steel; during grinding,
surface temperatures as high as 2000°C may be obtained and small cracks
and retained austenite may be induced near the surface. Cracks could
retain moisture and wash liquid, which would break down to hydrogen
during analysis, whilst the very high local temperatures indicate that as
a liquid interface might exist at the cutting face, this, followed by the
extremely rapid cooling rate, could account for the pick-up of hydrogen
during grinding.
5 The reproducibility and accuracy of various techniques for total hydrogen analysis have been discussed at length by the BISRA Study Group on
'Methods of analysis for hydrogen'. It is generally agreed that, with reliable carrier gas and vacuum hot extraction analysis equipment, the variation within sets of hydrogen results on plate, weld, or consumables is
inherent in the material itself or is a feature of the sampling procedure,
and is not a function of the apparatus.
6 The technique can be adapted to the analysis of the hydrogen contents
of austenitic and nickel alloy consumables used to weld ferritic steels.
However, only total hydrogen values are meaningful for these materials,
as diffusion rates at ambient temperature are so slow (see Fig. 5.17) that
a 'diffusible hydrogen' value would be meaningless. There will be a need
to use a higher collection temperature than with ferritic steels.
Glossary
This Glossary contains a number of terms which have been used in
the book. Those extracted from BS 499, Part 1: 1991, 'Welding terms
and symbols' are marked by an asterisk". They are reproduced by
permission of the British Standards Institute, 2 Park Street, London
W1A 2BS, from whom copies of the standard may be obtained.
Where two paragraphs are given, only the first gives the definition
in BS 499.
Arc energy*
The amount of heat provided by the welding process per unit length
of weld, defined
VI
-
w
X10- 3
where w is the welding speed in mm/s.
Arc energy or heat input (q.v.) can be used to estimate the cooling
cycle of a weld and its HAZ in order to estimate the microstructure
of the HAZ and, hence its risk of suffering hydrogen cracking. When
multiarc techniques are used, the separate arc energies should be
added together, provided that a single weld pool is produced.
Austenite (austenitic)
The solid solution of carbon in the face-centered cubic crystal structure of iron. Austenite is only stable at high temperatures in ferritic
steels but may be retained by fast cooling (retained austenite). In
austenitic stainless steels it is stabilised at ambient and lower
temperatures by the nickel in the steels.
Bainite (bainitic)
A transformation product of austenite formed during moderately
slow cooling, and comprising a mixture of ferrite (q.v.) and iron
carbide.
125
Glossary
Baking
Often used when MMA electrodes and submerged-arc fluxes are
dried at high temperatures, e.g, above about 250°C, to remove
moisture. When fluxes are dried at temperatures above about 800°C
the term pre-sintering is sometimes used.
Basic electrode* (basicity)
A covered electrode in which the covering is based on calcium
carbonate and fluoride.
A basic (submerged-arc) flux or MMA coating contains a high
proportion of basic oxides, and gives a weld metal of relatively low
oxygen content, which is usually associated with good weld metal
toughness. With MMA electrodes, the term is almost synonymous
with hydrogen controlled (q.v.); conversely it is not true that basic
submerged-arc fluxes give lower weld hydrogen contents than other
types.
Carbon equivalent (CE) value
A number, calculated from the chemical composition of the steel by
means of an empirical formula, which summarises some aspect of
the steel's behaviour, particularly its hardenability or its resistance
to hydrogen cracking. The term may also be used for weld metals.
Cellulosic electrode * (cellulose)
A covered electrode in which the covering contains a high proportion
of cellulose.
Cellulose contains a high proportion of hydrogen, which helps to
confer high penetration during welding.
CO2 welding* (see MAG welding)
Metal arc welding in which a bare wire electrode is used, the arc
and molten pool being shielded with carbon dioxide.
Combined thickness
The sum of the thicknesses (in millimetres) of the metal through
which the heat produced during welding can flow away from the
joint. In texts before about 1970, the term thermal severity number
(TSN), measured in 1/4 inch (6.35 mm) units was used.
Continuous cooling transformation (CCT)
The transformation of austenite to lower temperature transformation
products (martensite, bainite, pearlite, ferrite etc) in a steel which is
being steadily cooled, as in a weld HAZ, as opposed to being
126
Glossary
transformed at a constant temperature (isothermal transformation).
Diagrams of the CCT and isothermal transformation behaviour for
many types of steel are available in the literature.
Controlled thermal severity (CTS) test
One of several laboratory tests to investigate hydrogen cracking.
The combined thickness (thermal severity), welding process, weld
hydrogen level. preheat and heat input can be varied independently.
The test is a fillet weld with a deliberate root gap and a moderate
restraint level; results of CTS tests provided the basis of most of the
welding procedure diagrams in the present text.
Cored wire*
A consumable electrode having a core of flux or other material.
It comprises an outer sheath of steel (usually mild steel) which
contains flux and/or metallic ingredients. The sheath may be a fine
hollow tube or a strip wrapped round the core. Cored wires may be
used for MAG, MIG or submerged-arc welding.
Critical hardness
The lowest maximum HAZ hardness at which hydrogen cracking is
found in the HAZ under the particular test conditions used. The
critical hardness varies with weld hydrogen content, welding process, preheat level, joint geometry and steel type.
Deposited metal"
Filler metal after it has become part of a welded joint.
The increase in weight of a parent material due to the deposition
of filler metal is particularly used in tests to measure weld hydrogen
contents.
Diffusible hydrogen
That portion of the hydrogen content of a ferritic steel which will
be released by diffusion when the steel is held for several days
at ambient temperature (more strictly, at 25 ± 5 Qq.
Dip transfer"
A method of metal-arc welding in which fused particles of the
electrode wire in contact with the molten weld pool are detached
from the electrode in rapid succession by the short-circuit current
which develops every time the wire touches the weld pool.
127
Glossary
Drying
Heating of MMA electrodes or flux to remove moisture. The term is
sometimes restricted to low temperature heating up to about 250°C,
which will remove only loosely bound water (see baking).
Ferrite (ferritic)
The solid solution of carbon in the body-centered cubic crystal
structure of iron. The phase is stable at low temperatures and also,
in some ferritic steels, at temperatures close to the melting temperature, when it is called a-ferrite.
Fit-up
The gap between the two faces of a joint immediately prior to
welding. The terms close fit, machined fit, normal fit and poor fit
are sometimes used.
Flux-cored wire
A type of cored wire (q.v.) in which most of the core is of nonmetallic, fluxing ingredients, although some iron powder, deoxidants
and metallic alloying additions may also be present. Such a wire
may give weld hydrogen levels (q.v.) ranging from very low to
medium.
Fracture toughness (see toughness)
The ability of a material containing a sharp defect, such as a crack
tip, to resist fracture (or crack growth) when stressed. The fracture
toughness of steels (and weld metals) is reduced when the temperature" is reduced and also near ambient temperature when they
contain hydrogen." Parameters measured include I<e, the critical
stress intensity at the crack tip, and ac , the critical crack tip opening
displacement (CTOD):
Fused metal
That part of a welded joint which has been melted during welding.
It comprises added weld metal and fused parent metal. The term is
particularly used in tests to measure weld hydrogen contents and it
is usually estimated from the area of fused metal in a weld crosssection which has been ground and etched.
Fusion face *
The portion of a surface, or of an edge, that is to be fused in making
a fusion weld.
128
Glossary
Globular transfer*
Metal transfer which takes place as globules of diameter substantially
larger than that of the consumable electrode from which they are
transferred.
Grain size
The mean diameter of the grains which comprise the microstructure
of a material. When used of ferritic steels, the term often refers to
the prior austenite grain size, i.e. the grain size of the austenite
immediately before it transformed on cooling. Various expressions
are used to quantify prior austenite grain size, one of which (the
ASTM system) uses a single number, 0 being coarse and 10 fine.
The grain size of transformation products is more difficult to characterise; the terms ferrite grain size, martensite lath size, etc, are used.
Hardenability
The tendency of a ferritic steel to produce a hard microstructure as
a result of heating and rapid cooling, typically in the HAZ during
welding. The hardenability of a steel increases as its content of
alloying elements (and, to a smaller degree, carbon) is increased
(see Carbon equivalent value).
Hardness (hardness value)
The resistance of a material to deformation, usually measured by
pressing an indenter into the surface. Of the several types of hardness
measurement, the Vickers hardness test (HV) is most commonly
used in the context of welding, as the small size of the impression
produced by the indenter is suitable for measuring local variations
in hardness found in a weld HAZ, whose hardness varies over short
distances. When measuring HAZ hardness, it is important to make
sufficient measurements, particularly in the region close to the
fusion boundary, to obtain a realistic idea of the maximum hardness
of the HAZ. In connection with hydrogen cracking during welding,
the hardness of the HAZ associated with a weld bead before tempering (either by a subsequent weld pass or by PWHT) is the
important parameter.
Heat-affected zone (HAZ)*
That part of the parent metal that is metallurgically affected by the
heat of welding or thermal cutting, but not melted.
In the context of ferritic steels, the term is often used for that
portion of the HAZ (strictly, the visible HAZ) which is visible to
the naked eye on a macro section of a weld, i.e. a section which has
129
Glossary
been ground and etched. Close to the fusion boundary, where the
temperature had approached the melting point of the steel, appreciable coarsening of the prior austenite grain size occurs.
Heat input
The amount of heat supplied to the parent metal by welding. The
heat results from the energy of the welding arc (arc energy, q.v.),
reduced by the efficiency of transfer of the heat (arc efficiency).
Humidity
The amount of moisture in the air at any time. It can be expressed
in two ways; the relative humidity, which is the humidity expressed
as a percentage of the maximum amount of moisture it is possible
for air to contain at the particular temperature and pressure of
measurement. The absolute humidity is expressed as the weight of
moisture in a unit volume of air. The latter is probably the more significant as regards the amount of hydrogen likely to be introduced
from the atmosphere into a weld pool during welding.
Hydrogen concentration units
The concentration of hydrogen in a metal is expressed as parts per
million by mass (ppm). In weld metal it is the convention to use
the expression millilitres of hydrogen at standard temperature and
pressure (STP) per one hundred grams of deposited (or fused) metal
(mL/100g).
lmL/l00g == O.89ppm
A convention has been proposed that: for deposited metal, H D = mL(STP)/100g;
for fused metal,
H F = ppm.
Unfortunately, another convention has used the terms HD , HT and HR
for diffusible, total and residual hydrogen, respectively.
Hydrogen controlled electrode*
A covered electrode that, when used correctly, produces less than a
specific amount of diffusible hydrogen in the deposit.
Hydrogen embrittlement
Ferritic steels (and other metals) are embrittled to an increasing
degree at temperatures around ambient when containing increasing
amounts of hydrogen. Embrittlement is most severe at slow strain
130
Glossary
rates and is most readily apparent in steels with relatively poor
toughness; in HAZ material, this is synonymous with high hardness.
Impurity (impurity elements)
Elements usually of a non-metallic nature which are not intentionally
added to a metal. In ferritic steels the usual ones analysed for are
sulphur and phosphorus, together with nitrogen, oxygen and, in
some cases, arsenic, antimony and tin. Impurities can affect the
properties of steel and make it more prone to various types of
embrittlement and cracking, particularly during welding.
Inclusions (non-metallic inclusions)
Small particles within the steel consisting of a mixture of deoxidation products (oxides) and sulphides. These can act as 'traps' for
hydrogen, and hence reduce its diffusion rate, particularly at and
below -150 "C, Inclusions can also act as nuclei for the formation
of higher temperature (and softer) transformation products when
steel is cooled from high temperatures, thus reducing the hardenability of a steel.
Interpass temperature *
In a multirun weld, the temperature of the weld and adjacent
parent metal immediately prior to the application of the next run.
In welding procedures it is possible to specify a minimum and/or
a maximum interpass temperature.
Iron powder electrode *
A covered electrode in which the covering contains a high proportion
of iron in the form of a powder which acts as additional filler metal.
Iron powder can also be added to the weld pool during submergedarc welding. Strictly, the term should be applied only to pure iron
powders; when containing alloying and/or deoxidant additions, the
term metal powder (q.v.) should be used.
Isothermal transformation (see continuous cooling transformation)
Joint *
A connection where the individual components, suitably prepared
and assembled, are joined by welding or brazing.
131
Glossary
foint simulation test
A welding test conducted before beginning a major fabrication to
confirm that the selected welding procedure is safe. The test weld
is designed so that material, welding consumables, joint geometry,
restraint, fit-up, etc, closely resemble the details of the actual fabrication. After welding, the joints are sectioned and examined for
cracks.
Leg length *
The distance from the actual or projected intersection of the fusion
faces and the toes of a fillet weld, measured across the fusion face.
Relationships between heat input and minimum leg length are
different from those between heat input and measured leg length.
Liquation (liquation cracking)
The partial fusion of minute regions rich in impurity elements
(particularly sulphur) in the HAZ close to the fusion boundary.
When liquation leads to cracking, the cracks are very small, but
they can initiate hydrogen cracking.
MAG welding" (see also MIG welding)
Gas-shielded metal arc welding using a consumable wire electrode
where the shielding gas is provided by a shroud of active or noninert gas or mixture of gases.
The shielding gas is 'active' (hence MAG, metal active gas) because
the shielding gas contains oxygen or CO2 , often with an inert gas
such as Ar and/or He. The term CO2 welding [q.v.] is used when
CO2 alone is used as the shielding gas. A solid or a cored wire may
be used, although with the latter, the term cored wire welding [q.v.]
is more common.
Maraging
A heat treatment given to some complex nickel steels (so-called
maraging steels) of low carbon content which are martensitic at
room temperature, even with slow cooling. The treatment strengthens
the steel by an ageing process which involves the precipitation of
intermetallic compounds.
Metal cored wire
A cored wire [q.v.) in which the core is mainly of iron powder and
other metallic ingredients; it will normally be of very low hydrogen
potential.
132
Glossary
Metal powder
An admixture of iron and other metal powders (alloying and/
or deoxidising) used either as an ingredient of MMA electrode
coverings, or as an addition made during submerged-arc welding
(see iron powder).
MIG welding" (see also MilG welding)
Gas-shielded metal arc welding using a consumable wire electrode
where the shielding is provided by a shroud of inert gas.
The term was originally used (and sometimes still is used) for
both MIG and MAG welding (q.v.).
M s temperature (also MF , Myo, etc)
The temperature at which the transformation from austenite to
martensite begins (Ms), finishes (MF ) , is 90% complete (Mgo), etc.
Pearlite [pearlitic]
A transformation product resulting from slow cooling from austenite,
comprising a lamellar microstructure of ferrite and iron carbide.
Postheat
The application of heat to a weld, immediately after welding and
before cooling out, so as to maintain the minimum interpass temperature, or to raise it to some higher value, in order to increase the
rate of diffusion of hydrogen out of the weldment.
Post-weld heat treatment (PWHT)
Heat treatment given after welding to reduce residual stresses (stress
relief heat treatment) and/or to temper hard regions, usually in the
HAZ. The temperatures are usually between about 550 and 750 o e,
i.e. below the temperature at which the steel and weld metal start
transforming to austenite.
Preheating temperature* (preheat)
The temperature immediately prior to the commencement of welding
resulting from the heating of the parent metal in the region of the
weld.
Normally, a minimum preheat temperature is required; when it is
necessary to achieve particular levels of toughness and/or strength
in the HAZ or weld metal a maximum value may also be required.
Preheat may be local or general. It is recommended that the former
should be measured at least 75 mm from the weld line.
133
Glossary
Residual (alloying) elements
Elements, which usually originate from alloyed steel scrap and
non-ferrous metals in the steel charge, of the type which can be
added to steels as alloying elements. They include nickel, chromium,
molybdenum, copper, niobium and vanadium. Residual elements
can affect the CE value of the steel and also other properties.
Residual hydrogen
That portion of the hydrogen content of a ferritic steel which will
not be released by diffusion at ambient temperature (more particularly, at 25 ± 5 "C), but may be extracted at higher temperatures, e.g.
650°C. It is believed that this hydrogen is trapped in molecular
form, either in the lattice, in voids (particularly in association with
non-metallic inclusions) or in chemical combination with other
elements, such as carbon (when methane is formed).
Residual welding stress* (residual stress)
Stress remaining in a metal part or structure as a result of welding.
The stresses in a weldment result because the contraction of the
hot weld is hindered by the restraint (q.v.) of the joint. Residual
stresses tend to be tensile at the weld (at the weld surface in a
multipass weld) and are balanced by compressive stresses in the
parent steel (or in the root regions of a multipass weld).
Restraint
The extent to which the parts to be welded are secured to prevent
all controllable movement during and after welding. The parent
steel, unless extremely thin and narrow, provides sufficient restraint
to result in a weld containing residual stresses (q.v.).
Run length
The length of a weld bead deposited from the usable portion of a
single MMA electrode, i.e. from the length of the electrode less
40mm.
Runout ratio
The reciprocal of the run length (q.v.), i.e. the length of MMA
electrode required to deposit unit length of weld bead and which is
proportional to the heat input, divided by the square of the electrode
core wire diameter.
134
Glossary
Rutile electrode *
A covered electrode in which the covering contains a high proportion
of titanium dioxide.
A rutile electrode normally gives high weld hydrogen contents.
Sour service
Service where a material is in contact with a liquid containing
appreciable amounts of hydrogen sulphide (H2S), particularly when
the solution is acid.
Spray transfer*
Metal transfer which takes place as a rapidly projected stream of
droplets of diameter no larger than the consumable electrode from
which they are transferred.
Stickout
The length of wire in a continuous wire welding process between
the contact tip and the arc. A longer stickout results in resistive
heating of the welding wire in this region, which may be sufficient
to drive off a part of any hydrogenous material embedded in the
wire surface.
Stress corrosion cracking (SCC)
A form of cracking which can occur in a wide range of materials
caused by the combined action of stress and a chemically aggressive
environment. In ferritic steels, certain forms of stress corrosion
cracking are the result of embrittlement by hydrogen introduced
chemically from the environment and residual stresses resulting
from welding. Restrictions on the hardness of steel and weldments
are made to avoid the problem.
Stress relief heat treatment (see Post-weld heat treatment)
Temper (tempering)
The process of heating a steel, usually in the quenched condition,
to a temperature below that at which transformation to austenite
starts. Tempering (which occurs during PWHT, q.v.], usually (but
not always) causes softening. During welding, successive weld beads
give some tempering of the underlying runs.
Temper bead
A run of weld metal deposited on the surface of a weld with the
aim of tempering the HAZ of the final run of weld metal on the
135
Glossary
parent steel. It is important to ensure that the edge of the temper
bead is at a small fixed distance from the toe of the weld, otherwise
the temper bead will produce its own (hard) HAZ on the parent
steel. The temper bead may, after deposition, be ground off.
Therlllalconrraction
When a weld solidifies, it undergoes a volume contraction. Further
contraction occurs while it cools out, except that the transformation
from austenite to lower temperature transformation products gives
an expansion, counterbalancing some of the previous contraction.
Thermal contraction gives rise to distortion and residual stresses,
the balance between the two depending on the level of restraint
(q.v.).
TIG-welding*
Inert-gas arc welding using a pure or activated tungsten electrode
where the shielding is provided by a shroud of inert gas.
Toe *
The boundary between a weld face and the parent metal or between
runs. Note the term 'toe' should always be qualified according to
whether it applies to the complete weld or to individual runs.
Total hydrogen
The hydrogen content of a metal, i.e. diffusible plus residual hydrogen (q.v.].
Toughness (notch toughness, see fracture toughness)
The ability of a material to resist deformation and fracture in the
presence of a notch. The most common toughness test is the Charpy
impact test which, because of its high speed of straining, is normally
unable to detect the presence of hydrogen embrittlement in ferritic
steels.
Transformation
The change of crystal structure which occurs in some materials in
which different crystal structures are stable over different temperature (and pressure) ranges. In ferritic steels, the most important
transformation is from the high temperature form, austenite, to
lower temperature transformation products, such as ferrite, pearlite,
bainite, martensite and, in weld metals, acicular ferrite and ferrite
with aligned second phase.
136
Glossary
Weldability
The ease with which a metal can be welded to give a joint free from
unacceptable imperfections, particularly cracks, and having the
required properties.
Weld characteristic
An expression of the efficiency of heating during welding based on
welding speed, weld bead width and thermal conductivity. A low
value, less than 0.1, denotes a low heat input rate, with low cooling
rates and marked distortion. Values above 1.0 will give high cooling
rates on small welds but little distortion.
Weld hydrogen concentration
A laboratory measurement made on a test weld to determine the
amount of hydrogen absorbed and retained by weld metal deposited
using controlled welding, cooling and sampling conditions.
Weld hydrogen level
Arbitrary levels of weld hydrogen concentration into which a consumable in a particular condition can be placed in order to devise
welding procedures intended to avoid hydrogen cracking.
Trade and other names
used in the text
Name
Company
United States Steel Corporation
Corten
Creuselso Schneider-Creusot
BSC General Steels Division
Ducol
Durahete BSC General Steels Division
FV 520B Firth-Vickers Ltd
US Navy abbreviation denoting high yield strength (in
HY
ksi)
Maraging See Glossary
Superelso Schneider-Creusot
United States Steel Corporation
Tl
Selected bibliography of
TWI papers on hydrogen
cracking in welding
1 Baker R G and Watkinson F: 'The effect of hydrogen on the welding of
low alloy steels', Special Report no, 73, Iron and Steel Institute, 1961,
123-132,
2 Baker R G, Watkinson F and Newman R P: 'The metallurgical implications of welding practice as related to low alloy steels', Proc 2nd
Commonwealth Welding Conf, Inst of Welding, 1966, 125-131.
3 Graville B A, Baker R G and Watkinson F: 'Effect of temperature and
strain rate on hydrogen embrittlement of steel'. Brit Weld J 14 (6) 1967
337-342.
4 Watkinson F and Baker R G: 'Welding of steel to BS 968: 1962, a
dilatometric investigation', Brit Weld J 14 (11) 1967 603-613.
5 Graville B A: 'Effect of fit-up on HAZ cold cracking'. Brit Weld J 15 (4)
1968 183-190,
6 Graville B A: 'Effect of hydrogen concentration on hydrogen embrittlement'. Brit Weld J 15 (4) 1968 191-195.
7 Watkinson F: 'Practical implications of good weldability', Metals and
Materials 2 (7) 1968 217-222.
8 Watkinson F: 'Hydrogen cracking in high strength steel weld metals'.
Welding J Res Supp 48 (9) 1969 417s-424s,
9 Bailey N: 'Welding procedures for low alloy steels'. TWI Report Series,
July 1970.
10 Bailey N: 'Welding C:Mn steels'. Metal Constr 2 (10) 1970442-446.
11 Bailey N: 'Determination of welding procedures for 9%Ni and 18%Ni
maraging steels', Welding Intl Res and Dev 1 (3) 1971.
12 Dolby R E and Cane M W F: 'A new mechanical test approach for
assessing susceptibility to hydrogen induced HAZ cracking'. Metal
Constr 3 (9) 1971 351-354,
13 Coe F R: 'The avoidance of hydrogen cracking in welding'. Proc 1st Int
Symp, [ap Welding Soc, Tokyo, Nov 1971, 2, Paper 11B5,
14 Dolby R E: 'Heat-treatment aspects of joining structural and pressure
vessel steels', Proc Iron and Steel Inst Conf 'Heat treatment aspects of
metal joining processes', London, Dec 1971, 13-17, discn 54-55.
15 Bailey N: 'The establishment of safe welding procedures for steels',
Welding J Res Suppl 51 (4) 1972 169s-177s,
16 Boniszewski T and Watkinson F: 'Effect of weld microstructures
on hydrogen-induced cracking in transformable steels', Metals and
Materials 7 (2 & 3) 1973 90-96 & 145-151.
17 Gooch T G and Cane M W F: 'Fractographic observations of hydrogen
induced cracking in steel'. Proc Spring Meeting, Instn of Metallurgists,
139
Selected bibliography
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
'The practical implications of fracture mechanics', Newcastle, March
1973, Paper 10, 95-102.
Dolby R E, Hart P H M, Bailey N and Farrar J C M: 'Material aspects
controlling weld defects in offshore structures'. Proc Offshore Technology Conf, Houston, Apr-May 1973, Paper OTC 1908, 823-834,
Pub Amer Inst Mining, Met and Pet Engineers, 1973.
Cane M W F, Dolby R E and Baker R G: 'A slow bend test for HAZ
hydrogen cracking susceptibility and its correlation with welding
experience'. Welding Res Int13 (2) 1973 1-16.
Watkinson F: 'The welding of D6AC steel'. Welding Res lnt 3 (3) 1973
26-58.
Dolby R E: 'Welding abroad: fracture of welds in high strength steels'.
Proc Wksp 'Welding research opportunities', Arlington, Va, Jan 1974,
43-52, Pub Office of Naval Res, Dept of the [US] Navy, 1974.
Watkinson F: 'The new welding standard for C and C-Mn (structural)
steels'. Metal Constr 7 (1) 1974 26-28.
Watkinson F, Farrar J C M, Hart PH M and Widgery D J: 'The effect of
parent steel inclusions on the incidence of cracking in arc welds'. Proc
Conf 'Inclusions and their effects on steel', Leeds, Sept 1974, Paper 10!
1. British Steel Corpn.
Watkinson F and Bodger P: 'The advantages of low hydrogen welding
(processes and consumables) for higher strength steels'. Welding Res
lnt 5 (3) 1975 34-62.
Hart P H M and Watkinson F: 'Weld metal implant test ranks Cr-Mo
hydrogen cracking resistance'. Welding J Res Supp 54 (9) 1975 288s295s.
Hart P H M, Dolby R E, Bailey N and Widgery D J: 'The weldability of
microalloyed steels', Preprints 'Microalloying '75', Washington, Oct
1975, Session 3, 41-51.
Boulton C F and Bailey N: 'The influence of geometric weld defects on
structural integrity'. Proc TWI Int Conf 'Fabrication and reliability of
welded process plant', London, Nov 1976, paper 18.
Hart PH M: 'Some limitations of the implant cracking test for predicting
hydrogen cracking behaviour in welds'. Proc Int Conf 'Welding of
HSLA (microalloyed) structural steels', Rome, Nov 1976, 102-105,
discn 126-139. Pub ASM 1978.
Hart P H M: 'Material composition versus weld properties'. TWI Intl
Seminar, 'Pipeline Welding', London, Feb 1977, Paper 11. Also 'Recent
developments in pipeline welding', Ch 12, 57-62. Pub TWI, 1979.
Harasawa H and Hart P H M: 'A preliminary evaluation of the effect of
restraint on HAZ cold cracking of butt welds in three C-Mn steels'.
Welding Res Intl8 (2) 1978 129-164.
Hart P H M: 'Low sulphur levels in CMn steels and their effect on HAZ
hardenability and hydrogen cracking'. Proc TWI Intl Conf 'Trends in
steels and consumables for welding', London, Nov 1978, Paper 20,2153, discn 603-608.
Bailey N: 'Hydrogen cracking and austenitic electrodes'. Metal Constr
10 (12) 1978 580-583.
Dolby R E: 'Cracking in weldments of structural steels'. Proc S African
Natl Conf 'Fracture '79', Johannesburg, Nov 1979, 295-305. Pub Instn
of Metallurgists (SA branch).
140
Selected bibliography
34 Coe F R: 'The diffusion of hydrogen during welding', Proc 2nd JIM Inti
Symp 'Hydrogen in metals', Minakami, Nov 1979, 655-658. Pub Japan
Inst of Metals, 1980.
35 Gooch T G: 'Repair welding [of "hard to weld" ferritic steels) with
austenitic stainless steel MMA electrodes', Metal Constr 12 (11) 1980
622-677,
36 Hart P H M: 'The influence of REM treatment of steel on some aspects
of weldability when using the stovepipe welding technique for line
pipe girth welds', Proc Welding Inst Canada inti Conf 'Pipeline and
energy plant piping: Design and technology', Calgary, Nov 1980, 235245, Pub Pergamon Press.
37 Boothby P J, Hart P H M: 'Welding 0,45% V line pipe steels', Metal
Constr 13 (9) 1981 560-565,
38 Hart P H M: 'Hydrogen in steel - its significance for welding', Proc
Instn of Metallurgists Spring Conf 'Hydrogen in steel', Bath, Apr 1982,
93-98, discn 99-100.
39 Hart PH M: 'Hydrogen cracking in ferritic steel weld metals', Proc 3rd
ISMCM Intl Congress, 'Hydrogen and materials', Paris, June 1982, 2,
Paper J14, 1013-1018.
40 Hart P H M and Jones A R: 'The prediction of safe welding conditions
for C-Mn steels', Proc 1st ASM Intl Conf 'Current solutions to hydrogen
problems in steels', Nov 1982, 137-141,
41 Dolby R E: 'Steels for offshore construction', Proc TWI Conf 'Towards
rational and economic fabrication of offshore structures - Overcoming
the obstacles', London, Nov 1984, 65-74, discn 109-120,
42 Hart P H M: 'Some aspects of steel inclusions and residual elements
on weldability', Proc Conf 'Inclusions and residuals in steels: Effects
on fabrication and service behaviour', Ottawa, Mar 1985, 435-454.
Pub Canadian Govt Pub I Centre for CANMET and CSIRA, Also Metal
Canstr 18 (10) 1986 610-616,
43 Boothby P J: 'Predicting hardness in steel HAZs'. Metal Constr 17 (6)
1985 363-366,
44 Hart P H M: 'The influence of steel cleanliness on HAZ hydrogen
cracking: the present position'. Welding in the World/Soudage dans le
Monde 23 (9-10) 1985 230-239. Also Doc IIW/IIS828-85 (ex Doc IX1308-84),
45 Hart P H M: 'Resistance to hydrogen cracking in steel weld metals'.
Welding J Res Supp 65 (1) 1986 14s-22s,
46 Gooch, T G: 'Stainless steel consumables for dissimilar metal welds'.
In 'Welding dissimilar metals', ed N Bailey, Ch 3, 13-15, Pub TWI,
1986.
47 Pargeter R J: 'Field weldability of high strength pipeline steels', AGA
Report L51 507, Oct 1986. Pub American Gas Assocn, Arlington Va.
48 Hart P H M: 'Effects of steel inclusions and residual elements on
weldability'. Metal Canstr 18 (10) 1986 610-616.
49 Bailey N: 'High quality submerged arc welding with metal powders',
Proc CEC Info Day, 'Improved welding productivity with modern steels',
Luxembourg, Oct 1986. Pub CEC L-2985, Luxembourg.
50 Bailey N, Middleton K W, Casse A and Gaillard R: 'Improved productivity through metal powder additions to submerged arc welds',
CEC Report EUR-ll 180,1987. Pub L-2985, Luxembourg, CEe.
141
Selected bibliography
51
52
53
54
55
56
57
58
Hart P H M and Harrison P L: 'Compositional parameters for HAZ
cracking and hardening in C-Mn steels'. Welding I Res Supp 66 (10)
1987 31Os-322s. Also Rivista Italian a della Saldatura 42 (2) 1990
139-159; Welding Intl 5 (7) 1991 521-536,
Hart P H M, Matharu I S and Jones A R: 'The influence of reduced
carbon equivalent on HAZ cracking in structural steels', Proc 7th Inti
OMAE Conf 'Offshore mechanics and arctic engineering', Houston,
Feb 1988. Pub ASME.
Harrison P L and Farrar R A: 'Application of CCT diagrams for welding
steels'. Intl Materials Rev 34 (1) 1989 35-51,
Pargeter R J: 'The weldability of steels used in jack-up drilling platforms', Marine Struct 2 (3-5) 1989 255-264.
Hart P H M: 'Trends in welding materials for offshore structures', Proc
Intl Conf 'EVALMAT 89 - Evaluation of materials performance in
severe environments', Kobe, Nov 1989, 2, 811-825. Pub Iron and Steel
Inst of Japan.
Pargeter R J: 'Understanding how to weld without hydrogen cracking',
SAF Seminar 'Reduction of the welding costs using coated electrodes:
Towards the suppression of the drying and the decrease of preheating',
L'Epine, Oct 1990, Publ TWI. Also Welding and Metal Fabn 59 (1)
199119-20.
Terasaki T, Hall G T and Pargeter R J: 'Cooling time and predictive
equation for estimating hydrogen diffusion in CTS test welds', Trans
lap Welding Soc 22 (1) 1991 52-56,
Pargeter R J and Hart P H M: 'Welding and properties of welds in 450550 N/mm'' yield steels'. Proc 10th Intl OMAE Conf 'Offshore Mechanics
and Arctic Engineering', Stavanger, June 1991, IlIA, 81-91. Pub ASME.
Index
abrasion, influence on measured
hydrogen, 122
ACt temperature, 27
acicular ferrite, 12, 16
alloying elements in steel,
effect on hydrogen diffusion, 91
analysis, steel, 34, 36
arc,
atmosphere, hydrogen in, 6
efficiency, 41, 42
different welding processes, 42
energy, see also Heat input, 41,
42
voltage, 42
atmosphere, 6-9, 103
austenite, 13, 26, 28, 70
retained, 26, 66, 69-71
austenitic consumables
hydrogen analysis, 68, 123
preheat with, 29, 68
selection of, 67
austenitic stainless steel, 71
weld metal, 24, 28, 29, 63, 67-69
austenitic weld metal,
ferrite level, 67
pick up of sulphur, 67
thermal expansion coefficients,
67
austenitising temperature, 27
backing bars, permanent, 47
bainite, isothermal transformation
to, 67
baking, 8
electrodes Ifl uxes,
effect on weld hydrogen level,
105
basic electrodes, 31,43, 52, 60, 63,
71, 104, 105
bead size, see weld bead size,
block welding, multipass, 44
boron in steel, 12, 58, 59, 63
boundary fusion, 10
brittle fracture, 1
butt weld, geometry for hydrogen
removal, 84, 85
buttering, with austenitic
consumables, to reduce
preheat level, 68
carbon content, 24, 38, 49
high,24
low, 10, 24
carbon equivalent, see also CE, 11,
12,18,23,34-37
effect of silicon, 35, 36
precision of, 36, 37
carbon in steel, effect on hydrogen
diffusion, 90, 91
carbonates, 7
carrier gas hot extraction,
to determine residual hydrogen,
113, 120, 121
to determine total hydrogen, 120,
121
cast irons, 71
cavitation, creep, in reheat cracking,
16
CE axis, selection of, 19, 23, 38-40
CE scale, see CE axis,
cellulosic electrodes, 8, 103-105
characteristics, electrode, 8
Charpy test, 9
chevron cracks, 4, 15, 16
climatic conditions, effect on weld
hydrogen, 7-9, 103
CO 2 , 7, 110
CO 2 welding, see MAG welding
cold cracking, see hydrogen
cracking
143
Index
combined thickness, 17-19,23,39,
41, 51
variable thickness at joint, 41
composition, of steel, 34, 35
conditions for hydrogen cracking,
3-13
consumable,
storage, 7
use, 7
contraction
stress, 9
thermal, stress from, 3, 9, 69
cooling,
rapid,10
rMe, 11-13, 17,39,43
time, 11
cored wires (tubular wires), 23, 28,
68,102-105
heat inputs for fillet welds,
23
potential hydrogen levels, 102,
103
corrosion, 6
covered electrodes, see also MMA
electrodes, 104, 105
crack,
growth, slow, I, 15
size, 4
cracking,
chevron, 4, 15, 16
hydrogen, factors and conditions
for, 3-13
liquation, effect on hydrogen
cracking, 37, 38, 71
reheat, 16
solidification, 28, 67, 68
stress relief, 16
weld metal, 4, 5, 12, 15, 16, 47,
48,57,58,63
cracks,
intergranular, 4, 16
transgranular, 4, 16
transverse, 4, 16
weld metal, 4, 5, 15, 16
creep cavitation, in reheat cracking,
16
creep resistance, 63, 66
critical hardness, 10, 35, 39
current polarity, 7
current type, 7
degassing of steel, for use in
hydrogen analysis, 112, 113
degreasing fluid, 5
delay time for inspection, 1, 3, 14,
15,32
deposited metal, hydrogen content
related to, 113, 118
diffusible hydrogen,
determination, 113-119
extraction over mercury, 115-119
reproducibility, 118, 119, 123
sampling for, 115-117
diffusion, of hydrogen, 13, 14, 28,
73-100
dilution, 67
dirt,S
dislocations in steel effect on
hydrogen diffusion, 91
dissimilar steels, welding of, 38
double tempering, 26, 70
drying,
electrode, 8, 115
electrode overheating during, 8
flux, 8, 115
dye penetrant inspection, 15
efficiency MMA electrodes, 19-22
electrode, coating/covering,S, 8,
102, 107
electrode, MMA, size, run length
and heat input, 19-22
electrode, MMA, low hydrogen, see
Electrodes, basic
electrode moisture content,
determination by combustion,
107-109
sampling for, 107
Electrodes,
basic, 43, 49, 60, 63, 71, 103-105
cellulosic, 8, 103-105
MMA, efficiency, 19-22
MMA, run length, 19-22
rutile, 103-105
electroslag welding, 63
embrittlement, hydrogen,
mechanism of, 2
temperature of, 13
external stress, 9
factors for hydrogen cracking, 3-13
144
Index
failures, service, cost of, 1
fatigue, 1
ferrite,
acicular, 12, 16
in austenitic steel weld metal, 67
primary, 16
ferritic weld metal, not onto
austenitic or Ni alloy, 69
Fick's law of diffusion, 74
fillet weld,
geometry for hydrogen removal,
84, 85
leg length,
for control of heat input, 19, 23,
43
minimum for design strength,
43
single run, 23, 85
twin, directly opposed, 41
fit-up, 9, 44
flux, 3, 5, 7, 8
moisture content,
determination by combustion,
102, 107-109
sampling for, 107
flux-cored wires, 28, 102, 104, 105,
121, 122
forgings, heavy section, 33
formula,
CE, 12, 35, 36
CE, IIW, 12
Pe m , 12,
fused metal, hydrogen content
related to, 113, 118, 122
fusion boundary,
cracking, non-ferritic weld
metals, 28, 29, 68
hydrogen cracking at, 28, 29, 68
incomplete mixing near, 29
gas-shielded metal arc welding, see
individual process
geometric shape, effect on hydrogen
removal, 74-76, 86
grain boundaries, prior austenite, 16
grain size, 10, 11
effect on hydrogen diffusion, 91
grease, 5
half thickness, for hydrogen
removal,86
hardenability, 10, 11, 13, 17, 24, 36,
48, 51, 58, 60
hardness,
critical, 10, 35, 39
HAZ, 10, 13, 26, 27, 35, 63-65
HAZ,
hardness, 10, 13, 26, 27, 35,
63-65
maximum, estimation from
composition, 24, 25,61-65
microstructure, soft, 10, 26, 27
toughness, 12, 26, 27
heat input, 11, 18, 19,41-43
effect on weld metal hydrogen
cracking, 47, 58
for multi-arc welding, 43
for twin arc welding, 43
limitation of, 19, 52
relation to electrode runout ratio,
42,43
relation to fillet weld leg length,
19,23,43
relation to weld run length,
19-22,43
heat treatment,
for hydrogen removal, 73-101
post-weld, 10,13,14,16,26,27,
67, 69, 70
stress relief, 10, 67, 69
heat-affected zone, see HAZ
HIC, from hydrogen picked up in
service, 3
holding,
temperature, for isothermal
transformation, 27, 67
time, for isothermal
transformation, 27, 67
humidity, influence on weld
hydrogen, 6, 7, 9, 103
hydrogen,
analysis, 107-123
austenitic steel weld metal, 28,
68, 123
direct analysis, 102
encapsulated samples, 102,
111-113
nickel alloy weld metal, 28, 68,
123
over mercury, 102, 115, 116,
118
sampling for, 112, 115, 117, 119
145
Index
single bead, rapid quench
method,102
attack, resistance to, 63
cracking,
at fusion boundary, 28, 29, 68
effect of liquation cracking on,
37, 38, 71
weld metal, 4, 5, 12, 15, 16,47,
48, 57, 58, 63
weld metal, extending from
HAZ cracking, 47
cracks, identification of, 15, 16
diffusible,
determination of, 113 -119
extraction over mercury,
115-119
sampling for, 115-117
diffusion, 13, 14, 28, 73-100
effect of steel composition, 26,
87-91
in austenitic steel weld metal,
28, 73
in nickel alloy weld metal, 28,
73
diffusivity, D,
for different steels, 89-91
selection of value for H
removal, 86-92
variation with temperature, 14,
88-91
distribution, non-uniformity of,
93
embrittlement,
effect of strain rate, 9
mechanism of, 2
from parent steel, 5, 33
in arc atmosphere, 6
induced cracking (HIG), from
hydrogen picked up in
service, 3
level,
diffusible, 28, 104-106,
113-119
for cracking, choice of value
for, 25, 29, 94
original, choice of value for
removal estimation, 92-94
potential, 7, 8, 102, 103
residual, 28, 119-121
total, 28, 119-121
'ultra-low', 8, 9, 46
weld, 5, 7, 18, 19, 23, 38,
102-105
levels, typical, 102-106
potential, 102, 103
weld metal, 104-106
potential, 7, 8, 38, 102, 103
removal, 13, 14, 73-100
at ambient temperature, 91, 92,
97,98
at interpass, 26, 61, 62, 99, 100
by diffusion, 13, 14, 24, 25, 65,
73-100
curves, choice of value of
diffusivity, 86-92
curves, practical use, 94-100
curves, simplified, 94-96
during welding, 26, 59, 62, 99,
100
influence of steel thickness, 25,
86
selection of heating
temperature, 86-92
worked examples, 96-100
residual, determination of,
119-121
solubility, 5, 6
austenitic stainless steel, 28
ferritic steels, 5, 6
nickel alloys, 28
sources of, 5-7
total, determination of, 119-121
hydrogenous compounds, 5
hydroxyl ions, in silicate lattice, 7,
107
identification of hydrogen cracks,
15, 16
IIW CE formula, 12
inclusions,
effect on hydrogen diffusion, 91
hardenability and cracking, 37
notch effect of, 9
incubation time for cracking, 1, 3,
15, 32
inspection,
dye penetrant, 15
for chevron cracks, 15
for cracks, non-ferritic fillers, 15,
29, 67
MPI,15
radiography, 15
146
Index
time delay, I, 3, 14, 15
ultrasonic, 15
interpass,
temperature, 13,43,44,61,62,
99, 100
time, 26, 58, 62, 99, 100
interrun, see interpass
iron powder MMA electrodes, 19,
43
isothermal transformation,
characteristics, 27, 67
to control cracking. 24, 26, 27, 63,
66,67
joint simulation test, 9, 23, 26,
29-32,65
lamellar tearing, 16
lead, in free-cutting steels, 71
leg length,
for control of heat input, 19, 23,
43
minimum for design, 43
single run fillet, 23
liquation cracking,
effect on hydrogen cracking, 37,
38,71
effect of Mn: S ratio, 37, 38
low hydrogen electrodes, see also
electrodes, basic, 8, 46
low temperatures, welding at. 14,
15
lowering-off, in pipeline welding, 9
lubricant. drawing, 7
machined preparations, fit-up of, 44
MAG/MIG welding,
arc efficiency. 42
typical weld hydrogen levels,
104, lOS, 121, 122
magnetic particle inspection, 15
martensite,
in austenitic stainless steel weld
metal, 67, 69
near weld metal fusion boundary,
28,68
transformation, start temperature
(M s ), 26, 27. 66, 69-71
mechanism of hydrogen
embrittlement, 2
mercury, collection of hydrogen
over, 102, 115, 116, 118
microstructure, 11, 12
hardness. 12
HAZ, soft, 10, 26, 27
soft, 10, 13, 26
susceptible, 13
HAZ, 3, 13
microvoids in steel, effect on
hydrogen diffusion, 91
minimum leg length,
for control of heat input, 19, 23,
43
for design strength, 43
misalignment effect on cracking, 44
MMA electrodes.
austenitic stainless steel, 67
nickel alloy, 68
potential hydrogen levels, 102,
103
run-out ratio and heat input, 43
MMA welding arc efficiency, 42
MMA welds, heat input for fillets,
23
Mn: S ratio, effect on liquation
cracking, 37, 38
moisture content determination.
fluxes and electrodes, 107-109
shielding gases, 109, 110
moisture pick-up, resistance to. 8
MPI,15
Ms, see transformation start
Multi-arc welding. heat input for,
43
Multirun welds,
effect on root gap, 44
hydrogen level in, 94
NDE for cracks, non-ferritic fillers.
29, 68
nickel alloy consumables,
for repair of power station plant.
69
hydrogen measurement, 68, 123
selection of, 68
nickel alloy weld metal, 24, 28, 29,
63,67-69
solidification cracking of, 28, 68
thermal expansion coefficients,
67
147
Index
notch,
effect of inclusions, 9
effect on cracking, 9
oil,5
original hydrogen level, choice of
value for removal
estimation, 92-94
overheating during electrode
drying, 8
oxide,
hydrated, 5, 7, 111
influence on measured hydrogen,
122
paint, 5
parent steel,
hydrogen from, 5, 33
strength level, 9, 47
partial penetration welds, 47
Pcm formula, 12
penetration, 8
pickling fluid, 7
pipeline welding, 9, 100
plasma welding, 111
plating fluid, 7
positional welding, 19, 48, 52
post-weld heat treatment (PWHT),
10,13,14,16,26,69,70
avoidance of, 27
with austenitic and Ni
consumables, 67, 69
postheating,
temperature, 13, 25, 26, 44, 65
time, 25, 26, 65, 66
potential hydrogen,
determination by encapsulation,
111-113
levels,
cored wires, 102, 103
MMA electrodes, 102, 103
solid wires, 102, 103
submerged are, 102, 103
precision,
analysis, 36
of CE formula, 36, 37
preheat,
diagram, 18,23, 51-60, 62,68
general,44
limitation, 19
local, 43, 44
measurement of, 44
temperature, 11, 13, 18, 19, 23,
24,43,44
high, avoidance or reduction
of, 24, 28,48, 51, 61, 63,
65,67
pressure vessels, 1
primary ferrite, 16
production control, hydrogen
measurements for, 102
quenching, 10
lead bath, 70
radiography, 15
reheat cracks, 16
removal efficiency,
of hydrogen, 73, 86
related to weld geometry, 86
removal of hydrogen,
at ambient temperature, 91, 92,
97,98
at interpass temperature, 26, 61,
62,99,100
by diffusion, 13, 14, 25, 65, 73100
repair of hydrogen cracks, cost, 1
residual alloying elements, effect on
CE,34
residual hydrogen, determination
of,119-121
resistance welding, 111
restraint, 9, 24, 25, 38, 39, 61, 67
influence on choice of CE axis,
38, 39
retained austenite, 26, 66, 69-71
root gap, 9, 44
run length, weld, MMA, relation to
heat input, 19-22,43
run-out ratio, MMA relation to heat
input, 42, 43
rust, 5, 7
rutile electrodes, 43, 103-105
safe welding procedures, selection
of,46-48
samples,
for hydrogen analysis, 112, 115,
117,119
148
Index
for moisture analysis, 107, 109
segregation, in steel, 34, 35
selection of safe welding
procedures, 46-48
selenium, in free-cutting steels, 71
shielding gas moisture,
determination, 109, 110
short weld lengths, influence on
cracking, 44
sigma phase, formation during
PWHT,67
silicon in steel, effect on
hardenability, 35, 36
slag, influence on measured
hydrogen, 122
soap, wire drawing, 7, 111
solid wires, potential hydrogen
levels, 102, 103
solidification cracking, in nickel
alloy weld metal, 28, 68
solubility, hydrogen in steel, 5, 6
sour service, H 2S, 6
standards,
British, 2, 30, 36,42, 46, 47, 59,
64,66,67,71,84,113,121,
124
other, 2, 41, 46, 60, 68, 113, 121
steel,
9%Ni,69
analysis, 33, 34
composition, 33-35
effect on hydrogen diffusion,
27,87-91
high Mn, 10, 63
mild, 38, 47-50,111,113
mill sheet, 34
segregation in, 34, 35
specifications, 34
strength level, 38,47
test certificates, 34
thickness,
for hydrogen removal, 86
see also combined thickness
typical composition, 34
steels,
alloy, 58-60, 62-71, 89
austenitic, 70, 71
carbon, 16, 48, 49, 61, 62, 65,
69-71
C-Mn, 10-12, 16, 30-32, 35, 38,
47,49-58,61-63,65,71
chromium, 66, 70
Cr-Mo, 10, 16, 63, 66, 69, 70
Cr-Mo-V, 66, 69
cryogenic, 69
dissimilar, welding of, 38
free-cutting, 34, 37, 71, 89
hardenable, hydrogen level for
welding, 38
heavy (thick) section, 33,47-49,
57,58,61
high carbon, 38, 61, 63-71
high hardenability, 24-30, 38,
61-71
hydrogen content of, 5, 33
low alloy, 10, 47, 58-71
low hardenability, 17-23, 30,
48-61
lower carbon, lean alloy, 58-61
machinable, 34, 37, 71, 89
maraging, 65
medium C, 61, 62
medium C, C-Mn, 61, 62
microalloyed, 50, 51
Mn-Cr-Mo-V, 58, 59
nickel, 10, 69
Ni-Cr-Mo, 10
rimming, 89
semi-killed, 35, 36
structural fabrication, 46
tool, 70
welding dissimilar, 38
steelwork structural, 1
stickout, influence on measured
hydrogen, 7, 122
strain rate, effect on hydrogen
embrittlement, 9
stress,
concentration, 9, 47
external,9
from rigidity (restraint), 3
level,9
relief,
cracking, 16
with austenitic and Ni
consumables, 67, 69
stressing, post-weld, 16
submerged arc fluxes,
agglomerated, 103, 106
fused, 103, 106
potential hydrogen levels, 102,
103
149
Index
submerged arc welding,
arc efficiency, 42
weld metal cracking, 57, 58
submerged arc weld hydrogen
levels, 58, 104-106, 121,
122
sulphur,
effect on inclusions, 37
in steel,
effect on hardenability, 37
effect on hydrogen cracking,
time delay for inspection, 1, 3, 14,
32
tool steels, 70
total hydrogen, determination of,
119-121
toughness,
HAZ, 12, 26, 27
weld metal, 12
transformation, isothermal, 24, 26,
27, 63, 66, 67, 70
37, 38, 71
temperature,
finish of, 27, 70, 71
start of, 26, 27, 66, 69, 70
transverse 45° cracking, see
chevron cracking
tubular wires, see also cored wires,
67, 68
twin arc welding, heat input for, 43
twin fillet welds, directly opposed,
37,38,71
effect on hydrogen diffusion,
91, 92
effect on liquation cracking,
pick up by austenitic and
nickel alloy weld metal,
tack welds, 45, 57
temper bead, 14, 65, 66
temperature, austenitising, 27
control of cracking, 3, 24-26,
59-66
for hydrogen removal, 26, 61, 62,
86-92, 97-100
holding for isothermal
transformation, 27, 67
interpass, 13, 24-27,43,45
of hydrogen embrittlement, 1, 13
tempering, 10, 13, 26, 27, 65, 66, 70
double, 26, 70
test,
joint simulation, 9, 23, 26,
29-32, 65
procedure/procedural, 9
thermal contraction, stress from, 3,
9
thermal expansion coefficients,
nickel alloy, and stainless
steel weld metal, 67
thick steel, build up of hydrogen
when welding, 45
thickness, steel,
effect on hydrogen removal, 25,
23
41
typical composition steel, 33-35
typical weld hydrogen levels, MAGI
MIG,104,105, 121,122
ultrasonic,
examination, for chevron cracks,
15
inspection, 15
underbead, location for cracks, 4
underwater wet welding, 69
vacuum hot extraction,
to determine residual hydrogen,
113, 119, 120
to determine total hydrogen, 119,
120
vacuum packaging, of electrodes, 8
voltage, are, 42
water of crystallisation, 7
weld bead size, 9, 19-23
weld bead, large, 10
weld,
characteristic, 84, 85
geometry, simplification of, 84,
85, 96
73-100
hydrogen level, factors affecting,
51
hydrogen levels, typical, MAG/
see also combined thickness, 18,
TIG welding, 111
arc efficiency, 42
7
MIG, 104, 105, 121, 122
metal, alloyed, 16, 47
150
Index
austenitic stainless steel, 24,
28, 29, 63, 67 -69
thermal expansion
coefficients, 67
C-Mn, 16,47, 58
hydrogen cracking, 4, 5, 12, 15,
16,47,48,57,58,63
effect of heat input, 47, 58
extending from HAZ
cracking, 47
nickel alloy, 24, 28, 29, 63,
67-69
soft, 12
strength, 9, 28
toughness, 12
root, notch at, 4, 9
run length, 19-23,43
relation to heat input, 19-22
size, 9, 19-23
toe, notch at, 4, 9
welding,
at low temperatures, 14, 15
characteristics, effect of
hydrogen, 8
electroslag, 63
multipass, 13,44,94
pipeline, 9, 100
positional, 19, 48, 52
MAG/MIG arc efficiency, 42
MAG/MIG typical hydrogen
levels, 104, 105, 121, 122
welds,
partial penetration, 47
tack, 45, 57
wet welding, 69
wire,
cleanliness, effect on weld
hydrogen level, 105
drawing compounds, source of
hydrogen, 7, 111
resistance heating of, 7, 122
solid welding, 68, 103, 112, 121,
122