WTIA – TECHNICAL NOTE 10 CHAPTER 1 PA G E 1

INTRODUCTION

1.1 General test data such as the Charpy V-Notch (CVN) energy. Some
industries prefer the use of COD, while in others it is a
The catastrophic and unexpected failure of a component rarely used parameter.
of production plant used to make a company’s product,
can have severe consequences from a health, safety and Early standards set limits on the proportion of the
financial point of view. The financial penalty may result value of K or CTOD relative to Kic (COD crit) and also
through a loss of production and immediate income as limits on yield or tensile strength were routinely employed
well as an ongoing loss as reduced market share, as a by structural standards. Work by the CEGB in the 1970s
result of the failure, is slowly regained from competitors. showed that the treatment of unstable fracture and plastic
This challenge is well understood by industry and to collapse criteria separately was not always conservative and
prevent this occurrence, NDE inspections are employed that failures may result within the supposedly safe limits
at regular intervals to monitor plant condition and reduce set for both unstable and plastic collapse failures. This lead
the incidence of unplanned outages. The ability of NDE to the publishing of the two criteria approach as seen in a
methods to find flaws before failure is dependent upon CEGB report “R/H/R6- Rev1” of 1977. Later revisions
inspection intervals and the mechanism of flaw creation. of this document (Rev 2 & Rev 3) saw the concept refined
While NDE is the first step in the assessment of the and expanded to include strain-hardening materials. The
condition of plant, it is generally seen as part of a more concept has gained general acceptance and is being included
comprehensive life assessment program of critical in more and more of the documents providing guidance
components. The full component assessment would in the fracture mechanics assessment of flaws.
utilise the information from NDE as well as metallurgical The approach uses a Failure Assessment Diagram or
and stress analyses to determine the remaining life of FAD (Chapter 5) which plots the ratio of unstable fracture
a cracked or uncracked component before it fails. The ratio of Ki/Kic called Kr or if CTOD is used δi/δic or δr
tool normally used to assess the importance of a crack on the “Y” axis and the plastic collapse ratio of applied
or flaw is Fracture Mechanics, which relates the size of stress to the plastic collapse stress or Sr on the “X” axis.
a flaw to the likelihood of its causing fracture in a given The diagram is separated into “safe” and “unsafe” regions
material under a given stress regime. The larger the flaw, by an acceptance curve, the shape of which varies with
generally, the lower the stress at which failure will occur. the assessment methodology.
Conversely, the lower the service stress the larger the
flaw which may exist without endangering the structure. Fracture mechanics can also be used to estimate the life
The fracture behaviour of steel as pertains to fracture of a component containing fatigue cracking (Chapter 6).
mechanics is discussed in Chapter 2.
1.2 Fracture Mechanics as a Tool in
Linear elastic fracture mechanics (Chapter 3 & 4) Engineering Design
defines a parameter Klc, the critical stress intensity factor,
which is a material property pertaining to the occurrence The design engineer can greatly improve the reliability
of brittle fracture. In many materials, however, yielding and serviceability of a structure or component by using
occurs before fracture, the fracture is not brittle and, fracture mechanics to ensure compatibility between critical
although failure may be catastrophic, it does not occur flaw sizes and the NDT process to be used in service.
as readily as use of Klc would imply. To make use of Once the initial stage of design has predicted the stress
this extra margin of safety available in tougher materials levels during use, the critical flaw size can be calculated.
Yielding Fracture Mechanics defines the CTOD and other This critical flaw size could be as small as one mm, for
parameters. These are not true material constants as they example, which might be below the threshold sensitivity
depend on material thickness. Their use, however, is an of the normal NDT process envisaged as economical for
order of magnitude improvement over traditional impact the structure. Fabrication inspection would not, therefore,
 PA G E 2 CHAPTER 1 WTIA – TECHNICAL NOTE 10

detect flaws which could cause catastrophic failure of assessment shows that it is benign. The estimated critical
the component or structure under service stresses. The flaw size may be used to assess whether the existing flaw
design engineer, having calculated a critical flaw size as will grow under the expected service stresses and cause
small as this, should therefore re-design to reduce stresses failure of the component or structure.
or increase toughness and thereby increase the critical
flaw size. At 10-20 mm, a flaw may be reliably detected This procedure applies to all structures, whether
eg. by ultrasonic testing in service, or during routine fracture mechanics was used in the design or not. It involves
maintenance shutdowns. Where constraints impose high estimation of the stresses acting on the flaw and of the
design stresses and consequently small critical flaw size, toughness of the material, before the actual calculation
the design engineer may elect to specify more sensitive can be made. The step-by-step assessment of the severity
or rigorous NDT methods or more frequent inspection at of the flaw is described in this Technical Note.
correspondingly higher expense.
1.4 The Accuracy of Fracture
1.3 Fracture Mechanics as a Tool in Mechanics Assessments
Flaw Assessment
The application of fracture mechanics is not an
Pressure vessel and structure codes such as AS 1210 exact science, the accuracy of the assessment being
and AS 1554 specify flaws in terms of their appearance on dependent upon the accuracy of the input data. Errors
radiographs. Such codes contain such all-encompassing in stress estimation, flaw measurement, and fracture
requirements as “no crack-like flaws shall be accepted”. toughness determination may compound, thus making
Radiography is relatively insensitive to crack-like flaws, prediction within 100% often difficult. However an
however, so other techniques are often adopted. The improvement in accuracy can be obtained by the use of
situation can arise, therefore, that flaws, un-detected correctly meshed finite element analysis (FEA) models,
by radiography, are found by ultrasonic testing and by using NDT methodology appropriate for the material and
their mere existence cause the inspector to reject the crack location and by obtaining several fracture toughness
vessel or structure under the appropriate code or require results tested to an approved method. Generally the most
it to be repaired. Many public authorities, aware of this difficult problem in a fracture mechanics assessment is the
problem, agree to accept such a flaw if a fracture mechanics determination of the stress, rather than the material fracture

Relationships Between Terms

Tough Material High Energy Absorption

Fracture Behaviour
Brittle Material Low Energy Absorption
Stable Fracture Slow Crack Growth
Material Behaviour
System Behaviour
Unstable Fracture Fast Crack Growth
Hard System Low Compliance

Soft System High Compliance

Fracture Mechanism and Appearance

Metallurgical Mechanism Apperance of Fracture Surface Legend

of Failure On Macro Scale On Micro Scale
Antonyms
Shear Fibrous Dimpled
Synonyms
Cleavage Crystalline Faceted

Grain Boundary Crystalline Intergranular

Separation

Figure 1.1 Schematic relationships between terms

 WTIA – TECHNICAL NOTE 10 CHAPTER 1 PA G E 3

toughness. On the other hand, toughness can usually be

estimated, or determined from tables, with an accuracy
of at worst ± 50%. At high stresses and large crack sizes
(the most dangerous conditions) the assessment is less
sensitive to material fracture toughness. In relatively
complex structures, such as the root penetration on a
pressure vessel weld, the stress field is extremely difficult
to determine it is frequently necessary to use the FEA
technique, with care to correctly allow for all stresses
and restraints. The fracture analysis may also have to
allow for secondary stresses, including any secondary
stresses which includes residual stresses from welding.
This procedure will be discussed in Section 4.3.

Quite often a high accuracy is not required in the

field. If the estimated critical crack size is 10 mm even a
5 mm crack cannot be accepted without other information
such as precise knowledge of its sub-critical growth rate
(see Chapter 6). Alternatively the critical crack size may
have a length of several metres, and even if this was
determined by inaccurate, but conservative methods a
useful judgement can be still be made if the present crack
size is for example, 10 mm.

The philosophy adopted in this note is to provide a

reference that gives an overview of the use of fracture
mechanics on flawed structures and those issues that
affect the integrity of fabricated structures.
PA G E 4 WTIA – TECHNICAL NOTE 10
 WTIA – TECHNICAL NOTE 10 CHAPTER 2 PA G E 5

FRACTURE BEHAVIOUR

2.1 Fracture of Materials the most brittle fractures in ferritic steel plate, shear lips
exist at the plate surface even if fracture is predominantly
2.1.1 Brittle and Ductile Fracture in Steel by cleavage.
Two distinct failure modes have long been identified Fracture mechanics does not apply to the mode of
in ferritic steel, shear and cleavage. Shear failure (Figs failure by overload with respect to yield or ultimate
2.2 and 2.3) occurs by void coalescence. It is fibrous in strength, which is accompanied by the visible warning
appearance and usually indicates high-energy absorption signs of buckling or necking (plastic collapse), although
and ductile behaviour. Cleavage failure, (Figs 2.4, 2.5 it will be seen in Chapter 5 that these two criteria are
and 2.6) either trans or intergranular, on the other hand, assessed together when a Failure Assessment Diagram
has a crystalline appearance and usually indicates low (FAD) is used. This is called the two criteria approach
energy absorption and brittle behaviour. Fractures can using approaches as seen in the BS 7910:1999.
form by a combination of these two modes and in all but
Methods that allow for greater design loads (for
example by increasing material strength and increasing
section thickness) tend to encourage catastrophic failure

Figure 2.1 (a) Fragments of a 100 mm thick steam

receiver tested to failure.

Figure 2.1 (b) Detail of one brittle fracture initiation point Figure 2.2 Shear fracture of COD test-specimen, showing
at a thrust block weld near the right end of the steam thinning due to plane stress deformation. Note chevron
receiver shown in (a). x 1/3 approx. notch and fatigue crack In upper portion of sample x1.17
 PA G E 6 CHAPTER 2 WTIA – TECHNICAL NOTE 10

Figure 2.3 SEM fractograph of the sample in Figure 2.2 x Figure 2.4 Cleavage fracture of a COD test specimen
100.17 showing shear lips at the sides. x1.17

by brittle fracture without warning, (Figure 2.1), should The fracture mechanics approach calculates the
a crack already be present. incremental growth of the crack for each load cycle. This
As ferritic steels are capable of both types of behaviour, information can subsequently be integrated to determine
the transition from one type of behaviour to the other is of crack extension over a number of cycles. Variable amplitude
considerable importance. It may be promoted by a relatively loading can be assessed using this method.
minor change in any one of the conditions discussed in
2.2.2 below. This brittle-ductile transition (Figure 2.7) or, 2.1.3 Stress Corrosion Cracking
more specifically, the temperature at which it occurs, is
sometimes used as a measure of the toughness of a steel. Stress corrosion cracking results from the combined
Like other estimates of toughness, however, there are effects of an aggressive environment and stress (either
precautions to be observed when using it. applied or residual) on a susceptible material. Provided the
environment and stress state remains unchanged, the rate of
A misleading fracture appearance can sometimes crack growth tends to be uniform. The significance of the
occur. A high inclusion content, for example, can crack can be determined from the applied stress intensity
greatly reduce the fracture energy absorbed even though factor. This value can be compared against a threshold value
the appearance remains fibrous. On the other hand a fine for the susceptibility to stress corrosion cracking (KISCC).
grain size can cause cleavage fracture, when it occurs, to The use of factor of safety may be appropriate. These
absorb considerable energy. Tests based on the appearance calculations can be made more complex by the applied
of a fracture surface, therefore, may not actually rank stress being derived from residual stresses, the magnitude
materials according to toughness, as intended. The more of which may be hard to accurately determine.
sophisticated technique of fracture mechanics predicts a
material’s resistance to failure without undue reliance on
2.1.4 Corrosion Fatigue
the appearance of the fracture surface.
The presence of an aggressive environment may
2.1.2 Fatigue Fracture of Materials increase the rate of fatigue crack propagation. The corrosion
Fatigue cracks can initiate from both planar and fatigue mechanism that leads to a change in the rate of
non-planar flaws, under the action of cyclic loads. A fatigue crack growth can be complex and care should be
fracture mechanics approach can be used to determine taken to obtain accurate data. This can be obtained by
the rate of growth of a fatigue crack provided the crack appropriate testing using fracture mechanics samples in the
dimensions, flaw location, magnitude of the cyclic stresses same environment and temperature and under the action of
and some material properties are known. For cracking the same stresses. Other factors including fluid flow rates,
from non-planar flaws the BS 7910:1999 uses a S/N dissimilar metals and electrochemical polarisation may
curve approach. be important and should be reproduced in the tests.
 WTIA – TECHNICAL NOTE 10 CHAPTER 2 PA G E 7

2.2 Factors Influencing Fracture

Behaviour
Attempts to relate the traditional Charpy-V Notch
impact energy to fracture behaviour of structures
almost always fail because of differences in scale, crack
sharpness, and strain rate. The CVN was only found to
be a useful indication of toughness if measured below the
minimum service temperature. This and the high strain
rate give a pessimistic or negative estimate of toughness,
to compensate for the spurious positive effects of small
specimen size and blunt notch discussed below. This
section, then, may be considered a list of precautions to
be taken when planning to carry out meaningful fracture
toughness tests.

2.2.1 Temperature
Ferritic steels exhibit the ductile-brittle transition
(from shear to cleavage behaviour) on cooling over a Figure 2.5 SEM fractograph of the central, crystal-line
temperature range which may be quite narrow. Every area of the sample in Figure 2.4 showing cleavage facets.
test of toughness shows this variation (Figure 2.7), but x3000.17
the transition temperature and the width of the transition
range determined vary from test to test. It depends on the
other variables mentioned below and cannot be predicted
without a database. KIC or COD values to be used must
therefore apply at or below the expected minimum service
temperature, or at several temperatures if a range of load/
temperature combinations is expected.

2.2.2 Thickness
Ductile materials such as the structural steels and
nonferrous alloys seem tough when a thin section is tested
and relatively brittle when a thick section is examined.
(Figures 2.8 and 2.11). The reason for this is embodied in
the important concepts of the Plane Strain and Plane Stress
modes of deformation. The latter is assisted by thinning
or necking and Plane Stress Fracture is distinguished
by Shear Lips at 45°. The thinner the metal, the greater
the fraction of the fracture surface is composed of shear
lips and the higher the apparent toughness. Plane strain Figure 2.6 SEM fractograph illustrating re-heat cracking,
predominates in constrained situations such as in thicker which occurs by grain boundary separation. x200.
sections and at the tip of sharp cracks. Because it is more
difficult, the plane strain yield point is higher and plane
strain flow is more likely to be replaced by brittle fracture.
The thicker the section, then, the lower the proportion of
plane stress deformation and the nearer to its Klc value
is the apparent fracture toughness of the material. This
Toughness

is quite separate from the expected metallurgical causes

for variations in properties of plates of the same material
rolled to different thicknesses. Figure 2.8 should strictly
be thought of as applying to a certain material thickness
from which specimens of different thicknesses were cut
and tested. These metallurgical differences however are
less marked than those depicted in Figure 2.8.
To gain meaningful fracture toughness data, then, it is Temperature
normally necessary to test a specimen of the same thickness
as the structural component in question. The constraint Figure 2.7 Schematic variation of toughness with
temperature in steel (compare with Figure 2.8).
conditions in the test will then simulate those existing in
 PA G E 8 CHAPTER 2 WTIA – TECHNICAL NOTE 10

the structure. Above a certain thickness, however, test porosity or slag inclusions (those easily detected by
piece and component need not be the same thickness. radiography) may be comparatively harmless.
2.2.3 Material Strength Crack sharpness has a corresponding effect on
experimental toughness measurement (Figure 2.9).
The ratio of the yield strength of a material to its A fatigue-cracked test piece may give rise to a lower
fracture toughness also affects the transition from plane fracture toughness than one with a machined notch. A
stress to plane strain (refer to section 2.2.2). A low ratio machined notch is not therefore used in fracture toughness
will promote plane stress deformation and increase the testing. Unfortunately, Charpy impact tests which have
size of shear lips. Subject to other factors such as inclusion been performed for many decades with machined notches
content, plane stress fractures generally imply a high level have not given estimates of the fracture toughness of
of fracture toughness. materials. Their use has instead been based on empirical
Any change in metallurgical condition such as heat relationships applying to well-known materials and
treatment or grain size which alters strength will therefore structures. The more laborious technique of extending a
alters fracture toughness. In fact two samples of the same machined notch with a fatigue crack is only now being
steel heat treated to the same strength by different routes used on full thickness test pieces to give reliable estimates
would differ in fracture toughness. of fracture toughness.
Most structural materials yield at the crack tip. This
2.2.4 Crack Sharpness
crack “blunting” increases their actual and measured
The sharper (and longer) a crack, the higher the local toughness by both reducing the stress concentration at the
increase in stress or “stress concentration” at its tip. The crack tip and by increasing the amount of energy absorbed
local stress may be many times the average stress, as the during fracture. If extensive plastic deformation occurs
stress concentration factor for a flaw of length 2a and tip then linear elastic fracture mechanics (LEFM) (Chapter
radius r is given by 3) cannot be used; the COD analysis has been developed
2 r
a 0.5
( /) for this important mode of cracking. Whichever fracture
toughness parameter is applied, its Resistance or R-curve is
The most dangerous cracks in structures are therefore used when it is desired to take crack blunting into account
sharp ones such as fatigue cracks, heat treatment cracks, (Appendix B).
hydrogen cracks and stress corrosion cracks, since these
cracks may be so sharp that their width at the tip is only 2.2.5 Crack Orientation
a few atoms. Flaws which are relatively blunt such as In anisotropic material such as forgings, extrusions or
corrosion pits or lack of fusion in a weld are less dangerous plates, crack orientation can have an important bearing on
provided that they are not long. Rounded flaws such as fracture toughness measured. This is generally because
of aligned particles of non-metallic inclusions (in steel)
or intermetallic constituents (in aluminium), which assist
the crack by nucleating voids. It follows that it is the void

200
Fracture touchness, Kc (MNm-3/2)

150
Toughness

100

50

0 0.1 0.2 0.3

1 1
(Root radius) /2 (mm /2)

Thickness
Crack sharpness
Material strength } increasing
0.001 0.005 0.01
0.006 mm
0.05
Root radius (mm)
0.10

Figure 2.8 Schematic variation of toughness with Figure 2.9 Schematic variation of toughness with notch
thickness, crack sharpness and material strength. root radius. Data taken as for that of a high strength steel.3
 WTIA – TECHNICAL NOTE 10 CHAPTER 2 PA G E 9

coalescence (shear) fracture mechanism which is affected. quantity of strain energy stored in it at a given stress. In
The six possible crack growth orientations in plate are the time scale of the rapid propagation of a crack, this
shown in Figure 2.10. Of these, ZX and ZY are those strain energy in the cracking member itself is almost the
most affected by inclusions. This does not prevent the only energy available for supplying the work of fracture
consideration of fracture toughness as a material property, required to sustain the cracking. If the energy is sufficient,
as many parameters such as UTS and elongation vary with cracking will continue. If not, the crack will arrest. This
direction of measurement in anisotropic materials. applies whether the cracking is by shear or by cleavage,
the difference lying only in the greater level of energy
2.2.6 Strain Rate to be provided in the case of a ductile tear. This concept
Different rates of application of a load can greatly of energy balance is the origin of all fracture toughness
affect the toughness measured. In steel, particularly, parameters. It will be mentioned specifically, however,
the flow stress varies with strain rate. Since flow stress in respect of the J Integral in Appendix C.
determines the relative amounts of plane strain and plane • A high pressure gas pipeline is an example of a soft
stress deformation, which influence toughness, toughness system with high compliance. Unstable crack growth
and transition temperature also depend on strain rate. For in a gas pipeline can go on for many kilometres
this reason the hammer velocity in the Charpy Test has catastrophically heaving the pipe out of the ground,
always been specified to eliminate strain rate as a variable. damaging property and endangering life. In this case
In a fracture toughness test, the rate of application of the the potential energy of the compressed gas is available
load is slow because of the difficulty of making accurate in addition to that in the strained material to provide
measurements at rapid rates. energy to do the work of fracturing the tough line pipe
steel.
2.2.7 System Compliance • A liquid-filled pressure vessel is an example of a
Fracture mechanics calculations basically predict hard system with low compliance and high stiffness.
whether or not a known flaw will grow. It is often of A crack here may be stable and cause a leak rather
equal importance, however, to know whether the system than be unstable and fracture the vessel in two.
compliance causes such crack growth to be stable NOTE: This is one reason why liquid-filled pipelines
or unstable. The compliance of a system defines the are safer than gas-filled pipelines.
Other reasons are:
(i) the latter are run at much higher pressures in order
ion
irect to compress the gas and achieve appreciable
llin gD mass flow.
X-Y Ro Z
Y-X (ii) The velocity of sound in a liquid is far greater
than in a gas and in fact is greater than the crack
Y X propagation rate in steel pipe. As a result, the
stress in a liquid pipeline is relieved ahead of a
Z-Y X-Z
rapid, unstable crack and the crack arrests.
• A weld in an un-stressed structure is an example of
a very hard system. The residual stresses caused by
Z-X cooling and contraction of the solidified weld metal
and heated parent metal are local and may be relieved
Y-Z by initiation and growth of a small crack. Once these
stresses are relieved there is no further driving force
for propagation of the crack.
Figure 2.10 Specimen orientation nomeclature.12 The effect of system compliance on type of crack
growth is defined in terms of the Resistance or R Curve
in Appendix C.
2.2.8 Welding
One of the greatest problems in using fracture
Toughness

mechanics to assess welded structures, results from the

Increasing thickness
inhomogeneity of the welded joint. It is normal for the weld
and /or strain rate metal, the heat affected zone (HAZ) and the parent material
to have different fracture properties, but also, these various
components of the joint do not behave independently. For
example, although the HAZ frequently contains the least
Temperature tough micro-structure, it is rather narrow and irregular and
any fracture may not be restricted to it. Thus the fracture
Figure 2.11 Schematic variation of toughness with behaviour is affected by the proportion of the crack front
temperature, thickness and/or strain rate
which passes through tougher regions.
 PA G E 1 0 CHAPTER 2 WTIA – TECHNICAL NOTE 10

Cracks have been observed to grow to a critical size

in one portion of a weld and propagate rapidly but then
arrest when they enter the parent plate. This implies a
marked increase in toughness in the plate because:
• the longer crack gives a greater stress intensity, and
• material resistance to a propagating crack is generally
less than to a stationary crack.
Likewise, a slow-growing crack could suddenly
accelerate when entering a less tough region of the
microstructure.
Figure 2.12 illustrates the irregular behaviour of
cracks in weld zones.
The residual stress field is another frequent cause of
crack path or crack front in welds. This is a particular
problem in Fracture Toughness Testing when establishing
a regular fatigue crack. In this case the weld is compressed
as a form of mechanical stress relief (Appendix B) prior
to fatigue cracking. This is accepted as an undesirable
necessity in order to achieve a valid test result.
There is unfortunately little data on the Fracture
Toughness values of weld metal, and even less for heat
affected zones. When the variety of filler metals, welding
techniques and cooling rates are considered it is unlikely
that the literature will provide the information needed for
the accurate Fracture Mechanics assessment of a welded
joint. It will usually be necessary to do the assessment for
the parent material and assume various factors of safety
to account for the estimated worst properties in the weld
region. If the application is critical it will be necessary to
take samples and measure the toughness of the various
weld zones: measurement techniques are described in
Appendix B.
Figure 2.12 A fatigue crack tip deliberately induced In the
Figure 2.13 (a)
HAZ (right) of a COD test piece, which deviated into the
The range of
weld metal in the COD test.17
microstructures
occurring in a
weld zone. x6.517

Figure 2.13 (b) The range of microstructures occurring in a weld zone.x200.17

 WTIA – TECHNICAL NOTE 10 CHAPTER 3 PA G E 11

LINEAR ELASTIC FRACTURE MECHANICS

Linear Elastic Fracture Mechanics (LEFM) is strictly Units of K thickness and/or strain rate.8
applicable only to materials in which cracks can grow with
very little plastic deformation. Apart from the obvious The Stress Intensity, Factor K may have the units
relevance to brittle materials implied in this statement, MPam1⁄2, MNm-3/2, kgmm-3/2 or ksi √in.
LEFM has also been remarkably successful in the analysis Note: 1 Pa = 1 Nm-2 and 1 MPa = 106Nm-2 = 1 Nmm-2
of fatigue crack growth and stress corrosion cracking
(SCC), and environmentally assisted cracking, (EAC). The Sl unit, MPam1⁄2 is preferred. The conversion
These modes of failure have consequently been placed factors are as follows:
on a more analytical footing. MPam1⁄2 = MNm-3/2 = 31.6 Nmm-3/2 = 3.22 kgmm-3/2
= 0.91 ksi√in
3.1 Stress Intensity Factors
When using equation (1) the crack length must be in
3.1.1 The Concept of Stress Intensity Factor the correct units, ie., metres for Sl units, mm for metric
and inches for British units.
A sharp crack-like flaw of length = 2a is assumed to be
present in an infinite plate and the stress intensity factor,
K, at the tip of the crack is calculated from 3.1.2 Critical Stress Intensity and Critical
Crack Size
K= (1)
The crack tip stress intensity at which the crack
Where σ is the elastic stress acting on the structure. commences to grow in a given plate is the critical stress
In the configuration shown in figure 3.1, where the intensity factor, denoted Kc, and this is sometimes called
crack is being opened by a tensile stress at right angles the fracture toughness of the plate. Plate is used rather
to it, the stress intensity factor is known as Kl. This is the than material as a deliberate implication of constant
common crack opening mode. thickness. In the thickness range where Kc varies, shear
lips are pronounced and, being at an angle to the applied
Other modes II and III refer to cracks subject to shear tensile stress, are not subject to true mode I crack opening.
stresses of particular types and the corresponding Kl and In the higher thickness-range where Kc is constant the
Klll are sometimes seen in the literature. They are not fracture is flat and the crack opening is truly mode 1. This
considered in this note. limiting value of Kc is therefore called Klc, and this is the
material property which may be used in design similarly to
the tensile strength or yield strength. However, even these
“material properties” can vary with test specimen size,
strain rate and method of determination. These variations
in yield strength, tensile strength or critical stress intensity
factor do not normally concern the design engineer.
W It should be remembered that:
• the crack must be sharp.
t • ‘K’ may still refer to the stress intensity factor at the
tip of a stationary crack, just as ‘stress’ may refer to the
2a stress that exists in a structure well before yielding;
• the values of Klc for a particular material in a particular
condition may be determined by standard tests (see
Appendix A) or may be found tabulated in various
handbooks;
Figure 3.1 Rectangular through-thickness crack in an • there are Australian laboratories equipped to determine
infinite plate.
fracture toughness21, although published tables may
 PA G E 1 2 CHAPTER 3 WTIA – TECHNICAL NOTE 10

contain the values sought by an engineer for a particular materials and sections that do not develop plane strain
material. Thus, with a knowledge of the service stress conditions at the tip of a crack. Plastic deformation will
of the structure, and the fracture toughness, KIC, of occur at the crack tip, and linear elastic fracture mechanics
the material, it is possible to calculate from equation cannot be applied. This problem can be overcome to some
(1) the maximum size of a safe crack in the structure. extent by semi-empirical corrections, as is explained below.
This is termed the ‘critical crack size’, 2ac. If linear elastic fracture mechanics is used inapprop-
The critical crack size is of extreme importance. As riately for relatively ductile materials, an excessively safe
discussed in Chapter 1: and perhaps uneconomic assessment will be made. On
– it gives the size of flaw to look for in a structure occasions this assessment could be used as a first estimate
and hence is the basis for specifying inspection of the worst possible case.
and non-destructive testing programs.
In fatigue crack growth and stress corrosion cracking
– it determines whether a known flaw is ‘safe’ or
in ductile or other material, cracks are very sharp, and
‘dangerous’. stresses are much lower than the yield point. The analysis
3.1.3 Thickness Limit for Valid use of KIC of these failure modes by LEFM as discussed in Section
3.3 and Section 3.4 is therefore rigorous and useful.
The thickness at which the value of Kc becomes
constant is given approximately by: 3.2.2 Reason for Use of Correction Factors
As in most fields of engineering, the basic equation
(2) (1) must be modified by assorted factors. Some of these
have a valid theoretical basis: others are simply empirical
For AS 1548-7-430 plate, a steel commonly used in correction factors. They arise because:
chemical engineering plant, this value becomes: a) The basic equation, K = , applies only to brittle
SI British materials that fracture like glass in plane strain. When it
t = 2.5 x 150 2
t = 2.5 x 137 2 is desired to use the equation for metals that show some
230 2 33.6 2 ductility a plasticity correction factor must be used.
= 1.1m = 42" b) The basic equation applies to a small crack in a very
wide member where the reduction in load bearing area
(*from tabulated KIC values) due to the crack is insignificant. When a component
of finite size is cracked, the effect of the crack on
Given a yield strength of 175 MPa it is possible to the stress distribution is taken into account with a
verify that this thickness minimum for the aluminium calibration factor.
alloy 5083 plate in example 1 is 51 mm. Below this c) The basic equation treats a simple through-thickness
thickness limit, use of KIC and equation (1) is inaccurate flaw of rectangular cross section in an infinite plate. If
but conservative. Measurement and use of Kc at the service the shape differs from this, a shape factor must be used.
thickness is more accurate. The thickness below which
use of Kc or KIC is so inaccurate as to be misleading is
y
given in Section 3.6 condition (i).
3.1.4 Dynamic Effect
Normally KIC is measured under pseudostatic conditions
using low strain rates. However KIC is normally strain rate
sensitive, and under impact loading a different value of KIC
(often termed KID, D – dynamic) is obtained. This value
is usually lower than KIC, but under some circumstances
it can be higher. r

Similarly, the value of K at which a rapidly moving θ

crack will stop, (KIA, A – arrest) is often lower than the
x
value at which the crack will commence movement.
For this reason, some codes (including the ASME Code)
require both initiation and arrest fracture toughness values.
3.2 Extension of Linear Elastic Fracture
Mechanics (LEFM) to Real Structures
3.2.1 The Applications and Limitations of LEFM
The example of the AS 1548 plate in Section 3.1.4
highlights one of the greatest deficiencies of plane strain
Figure. 3.2 Shape of the plastic zone ahead of a crack3
fracture toughness: most common structures are made of
 WTIA – TECHNICAL NOTE 10 CHAPTER 3 PA G E 1 3

d) If the stressed member is not flat eg. the sides or ends 1.0 a 0.5 0.4 0.3 0.2 0.1
of a pressure vessel, or the walls of a pipe, a ‘bulging 2c 0
factor’ must be used to account for the additional stress
concentration at the crack tip. 2c
e) Other correction factors can be found in the literature a
(especially the ASME Code)27 to allow for the depth of a B
B 0.1
a buried flaw, the presence of bending moments, etc.

3.2.3 Correction for Finite Structure B

Dimensions Calibration Factor
2a
The calibration factor, Y, is used in the basic equation, 0.01
usually together with the plasticity correction factor. 0.01 0.1 a 1.0
B
(3) Figure 3.3 Method for treatment of partial thickness flaws
The calibration factor, and the formula for calculating as through-thickness flaws.
the stress, σ, from the load and specimen geometry in
Figure 3.2 are given in Table 3.1. which they are found. The ASME Code treats this partial
thickness flaw, but other treatments discussed below refer
The notations describing the specimen geometry are:
only to through-wall flaws and so techniques have been
SE single edge, N notch, B bend (or dimension breadth),
developed to allow them to be applied to partial thickness
T tension, C compact or centre or cracked, D double, P
flaws.
partial, NR notched round.
The very simplest and least accurate treatment is to
Although these data refer to test specimens used
assume that a partial thickness flaw behaves the same as
to measure K, they can be used to determine critical
a shorter through-thickness flaw of the same flaw area,
fracture mechanics parameters in structures of similar
Figure 3.3.
geometry.
This also permits a solution for through-thickness
The value of ‘λ’ is inserted into the appropriate flaws that are of irregular shape. They can be equated
equation to rectangular through-thickness flaws of the same area,
3.2.4 Crack Shape Factor Figure 3.4.

The flaw shape factor takes into account the length to Another approach to the prediction of failure in
V width ratio of an elliptical flaw, and is most conveniently pressurised components is that used originally by Folias.28
determined graphically from Figures 3.5, 3.6 and 3.7. It is a technologically ‘old’ approach (circa 1963) but it is
often useful as it considers through-wall flaws in a range
Note that it is necessary to know the stress, σ before of geometrics, eg. pressurised spheres, tapered cylinders
the shape factor can be determined from Figure 3.6. etc., rather than just simple cylinders.
Most common flaws can be treated in this way because It is used to account for stress amplification at the
they will not penetrate completely through the section in ends of an axial flaw resulting from the outward radial

Table 3.1
Specimen Y Range
SENB4 1.93 – 3.07 λ + 14.53 λ2 – 25.11 λ3 + 25.80 λ4 0 ≤ λ ≤ 0.6
SENB8 1.96 – 2.75 λ + 13.66 λ2 – 23.98 λ3 + 25.22 λ4 0 ≤ λ ≤ 0.6
1.99 – 2.47 λ + 12.97 λ2 – 23.17 λ3 + 24.80 λ4 0 ≤ λ ≤ 0.6
SENB 2/ λ –1/2 (1 – λ) –3/2
3 0.5 ≤ λ ≤ 1.0
1.99 – 0.41 λ + 18.70 λ2 – 38.48 λ3 + 53.85 λ4 0 ≤ λ ≤ 0.6
SENT 1/ λ –1/2 (1 – λ) –3/2 (1 + 3λ)
2 0.3 ≤ λ ≤ 1.0
29.6 – 185.5 λ + 655.7 λ2 – 1017.0 λ3 + 638.9 λ4 0.3 ≤ λ ≤ 0.7
CT 1/ λ –1/2 (1 – λ) –3/2 (5 + 3λ)
2 0.8 ≤ λ ≤ 1.0
CC (1 – 0.025 λ2 + 0.06 λ4) π secant 1/2 0 ≤ λ ≤ 1.0
DEN √π (1.122 – 0.561 λ – 0.015 λ2 + 0.091 λ3) (1 – λ)–1/2 0 ≤ λ ≤ 1.0
NR √π (1.122 – 1.542 λ + 1.836 λ2 – 1.280 + 0.366 λ4) (1 – λ)–3/2 0 ≤ λ ≤ 1.0
8 5
PTCn √πΦ–1 [1 + Σ 0 ∫ nλn Σ 0 gn n] 0 ≤ λ ≤ 0.96
 PA G E 1 4 CHAPTER 3 WTIA – TECHNICAL NOTE 10

d A1 A2 The code explains the points simply and does not

t require extensive experience in Fracture Mechanics.
Briefly the steps involved are:
A1 = A 2 i) The flaw orientation and size are determined. Flaws
Figure 3.4 Approximation of an irregular flaw with a at an angle to the applied stress are projected on
rectangular flaw of equal area the normal plane. Multiple flaws are included in
an envelope and treated as a single large flaw.
deflection along the flaw. The effect increases as the crack ii) The stresses at the flaw location are determined.
length increases and as the pipe wall and radius decreases iii) The stress intensity factors are determined. The
(l = crack length) method used in the code is based on equation (1).
The failure criterion is of the form iv) The material properties are determined.
v) The critical flaw sizes are determined. These take
into account the maximum conceivable growth
or, when the crack length is corrected for plasticity, of the flaw by stress corrosion fatigue during the
remaining service lifetime.

l l

(4)
a 2a

3.3 The ASME Approach to Flaw Figure 3.5 Dimensions of surface and sub-surface flaws.
Assessment
The approaches adopted by the ASME Code has 0.5
(σm + σb)/σy
become the basis for the assessment of flaws in pressurised α
Ellipse aspect ration a

components for unstable fracture. 0.4 1.0

0.8
The approach developed the basic fracture toughness
0.3
equation (K = ) and added various semi-empirical 0.5
0.3
factors so that the results agree with experimental 0.2 0
verification.
As in the work of Folias the empirical factors have 0.1
been experimentally verified43,44 using machined artificial
flaws which are relatively blunt. In these cases the use of 0
these models to predict the failure stresses of fatigued or 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2
stress-corroded components might be dangerous. Flaw shape parameter, Q
The methods are to be commended, (however, as valid Figure 3.6 ASME Shape factors for equation (5).27
elasto-plastic fracture mechanics approaches are not yet
available) to the field engineer with limited computer
access and testing facilities. While the accuracy of these 2.0
approaches is questionable, they can give at least an order
15
05
Membrane correction factor, mm

1.9
0

of magnitude estimate. a
= 0.
= 0.

= 0.
= 0.

1.8 α α α α
a
a

a
a

Although the ASME Code refers to nuclear power a = 0.2

α

plant, it can be applied to other plant components where 1.7

(Surface flaws)

t
a high degree of reliability and safety are required. It does 1.6
not treat through-thickness flaws, perhaps because it refers 1.5 a = 0.25
to thick-walled pressure vessels, where a through-thickness
α

1.4
flaw would cause a leak and require repair whether it a = 0.3
1.3
made the vessel unsafe or not. Through-thickness flaws
α

have already been treated in the Folias approach (See 1.2

Section 3.2.6). 1.1 0.35 a 0.5
α

The section of the code dealing with fracture mechanics 1.0

is not very extensive and is located after the section dealing 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
with non-destructive testing. It is intended as a method Ratio of flaw depth to thickness, a
t
for assessing flaws that have been located during an NDT
examination, rather than as a method for design using Figure 3.7 ASME Membrane correction factor for surface
flaws.27
fracture mechanics.
 WTIA – TECHNICAL NOTE 10 CHAPTER 3 PA G E 1 5

The stress intensity factor Kl in the ASME Code is 3.4 Fatigue Crack Growth Data
defined:
Conventional fatigue crack data is derived from smooth
specimens and is given in the form of an S-N curve and
Where: an endurance limit. There is little attempt to separate
crack nucleation from crack growth. This practice grew
σm, σb = Membrane and bending stresses up because the known detrimental influence of surface
a = Minor half-diameter of embedded flaw; flaw depth flaws on fatigue life was not quantified.
for surface flaw (Figure 3.3.1) Fracture Mechanics, however, can treat the case of
Q = Flaw shape parameter to be determined using a known or postulated flaw in terms of stress intensity
σmm and the flaw geometry (Figure 3.3.2). Fluctuating or fatigue stressing gives a stress intensity
range at the tip of the crack.
Mm = Correction factor for membrane stresses
In terms of stress intensity, then, crack growth rate is
Mb = Correction factor for bending stresses predictable and is given by
It is to be noted that it is the depth ‘a’ or ‘2a’ which is
of overriding importance rather than the length of the flaw. where is the amount of crack growth per cycle
The correction factor for membrane stress has not been of fatigue loading, and ∆K is the variation in the crack
introduced before, but it is analogous to ‘Y’ presented tip stress intensity through each cycle. As the crack is
earlier and is a correction for geometry. It is derived growing, ∆K is increasing, hence ∆K i is used to apply
graphically, for surface flaws from Figure 3.7. to the ith cycle.
The ASME Code also includes a separate solution for If the loading in the ith cycle produces a maximum
Mm for subsurface flaws. stress intensity of 100 MPam1⁄2 and a minimum of 40
MPam1⁄2, then ∆Kj = 60 MPam1⁄2. The constants CO and n
A further correction factor for bending, Mb, is also must be determined.
determined graphically from Figure 3.8. Again, the code
For example, a 0.15 C, 0.55 Mn-Ni-Cr-Mo pressure
contains a solution for subsurface flaws. vessel steel will crack at the rate given by:
The ASME treatment also requires an examination of
the possible growth of the flaw by fatigue in the service
environment. and for a reactor pressure vessel steel ASTM (A533B) in
a water environment, the growth rate is the much more rapid:
Plastic collapse is not treated using a the failure
assessment diagram approach, however steps are being
taken to incorporate this into the code. We may now deal with a definite crack size 2ao as such
a as a known crack or a postulated crack of the minimum
detectable size. After a given number of cycles N, the
1.4 M∆ new crack length 2a is given by
β
α

a1-n/2 - ao1-n/2 = (1—n/2) Co (Y∆σ)n

1.2
t if ∆a is small, variation of ∆K i with a may be
neglected.
Bending correction factor, Mb

1.0 ie. ∆a = N Co(∆Ki)n

a/ 0.3
α

may be related to a given period of service and the

(Surface flaws)

0.1 revised value of the crack length is then used to determine

0.8
a/ 0.2
if the crack will be dangerous at this later stage in the
α

structure’s service life.

0.6 This type of approach is used to determine the
0.2 frequency of non-destructive testing examination. The
period is selected so that known or unknown flaws
0.4 0.3 cannot approach critical size in the interval between
Exact solution (β = 0°)
0.4 inspections.
Estimate (β = 0°) 0.5
0.2 Estimate (β = 90°) 3.5 Stress Corrosion Cracking
Most fracture toughness data is determined in harmless
0 environments at rates rapid enough to allow little likelihood
0 0.1 0.2 0.3 0.4 0.5 0.6 of environmental effects. If the test is conducted in a
Flaw depth to thickness ratio, a/t hostile environment, then lower values of Klc are obtained
in stress corrosion susceptible materials. Because these
Figure 3.8 ASME Bending correction factor for surface lower values are due to stress corrosion, they are termed
flaws.27
KISCC or KIEAC.
 PA G E 1 6 CHAPTER 3 WTIA – TECHNICAL NOTE 10

Table 3.2 Typical comparative values of KIC and KISCC

Metal KIC(MPa m1⁄2) KISCC (MPa m1⁄2) Environment
Aluminium 2024 T852 19 15 Sea water
Aluminium 7075 T6 25 8 Sea water
Maraging Steel Grade 300 80 8 NaCl solution
Steel 4340 69 6 Sea water

Generally the value of KISCC becomes lower the longer that may give higher stress intensities at the crack tip
the experimenter is prepared to wait. However, the values than those calculated using only the service stress state.
of KISCC published in the literature refer to a minimum Growth rates may also be enhanced by corrosion product
threshold, below which cracking will not occur, as shown wedging inside the crack.
in Figure 3.9.
The other types of EAC are also amenable to useful
The use of a stress intensity at the tip of an SCC crack analysis on the basis of da/dN vs ΔK curves as shown in
to specify the stress corrosion behaviour of a material is Figures 3.5.2-4.
much more exact than specifying a stress only. Stress
corrosion cracking is known to be unpredictable, but this 3.6 Choice of Linear Elastic vs
is usually due to variations in the crack initiation stage, Yielding Fracture Mechanics
where incubation times of up to 1000 hr are known.
In this Note, LEFM has been applied to thick
The values of KISCC must be measured on the material fabrications such as nuclear pressure vessels and is also
of interest, in the service environment, or obtained from relevant in fatigue and stress corrosion cracking. It has
the literature. A wide range of data is available. also been shown that linear elastic fracture mechanics and
It can be seen that the stress intensity at the tip of a its material parameter, Klc, can be used slightly outside its
recently discovered crack in a component can give an range of strictly correct application, resulting in excessively
indication of whether that crack could have been caused conservative answers to problems. It is advisable, however,
by SCC. However SCC may be caused by residual stresses to use Yielding Fracture Mechanics on steel if any of the
following conditions are fulfilled:
(i) If the plastic zone size (3.2.3) approximates or
100
100
exceeds any one of:

  
 
• the material thickness
75
K1(MPa m)

75 • the crack length

• the remaining ligament
Ksi in

50 (ii) If at the service temperature:

50
σy ≤ 450 MPa, t ≤ 75 mm and CVN ≥40J
25 25 (iii) If σy = 450 MPa, t ≤ 50 mm and the steel is used at
a temperature at least 20°C above the 20J Charpy
transition temperature.
2 3 4 5 10 100 1000 10,000
Time (min) (iv) If COD values are available (if these are small,
σc ≤ 0.1 mm, they are readily converted to KIC
Figure 3.9 Change in K level associated with subcritical values to check condition (i) above).
flaw growth. Regardless of initial K, failure in any sample
(v) If the service temperature/thickness combination
occurs at K = KIC
complies with Appendix D of AS 1554.

Aggressive
Aggressive
Aggressive
da da da
dN dN dN

Inert Inert Inert

KIC KISCC KISCC

∆K ∆K ∆K
Figure 3.10 True Figure 3.11 Stress
corrosion fatigue (TCF). corrosion fatigue (SCF). Figure 3.12 SCF on TCF
 WTIA – TECHNICAL NOTE 10 CHAPTER 4 PA G E 1 7

DATA REQUIREMENTS FOR A FRACTURE

MECHANICS ASSESSMENT

A fracture mechanics assessment is dependent upon hands of skilled operators can detect and size some flaws
the input from the NDT inspections, the metallurgical smaller than permissible flaw levels in many codes. In fact,
assessment. material properties and the stress analysis. NDT specified in codes is pitched at a level designed to
This methodology brings together the results from detect poor workmanship in welding, rather than eliminate
these technologies to determine whether a flaw affects flaws which may be critical on a fracture Mechanics
the fitness for purpose of the structure. If sufficient basis. It is necessary for the design engineer to design
information is available a prediction of the remaining for “Economically Detectable” critical flaw sizes rather
life of the component/structure containing the flaw can than to require the user of the structure to repeat regularly
be made. Each of the required inputs will be reviewed the exhaustive non-destructive examination carried out
briefly in turn: after fabrication. It is necessary, therefore, for the design
engineer to have an appreciation of the capabilities of NDT
4.1 Previous Plant History methods. Consultation with experienced NDT operators
The normal and any other pertinent operating conditions is advised before stresses are set in a structure.
of the component should be reviewed. This should include
the operating environment. Any previous incidents of a 4.3 Operating Stresses
similar nature may have been reported and investigated. The stress acting on the flaw is often the most difficult
Plant operators may be able to provide additional detail information to obtain in the flaw assessment due to the
of unusual operating conditions, which has assisted in presence of cracking at nozzles, reinforcement, ends
the formation of the flaw. Sometimes similarly failed of fillets welds, structural joints and misalignment as
components may be available or metallurgical replication well component geometries that may not be simple. An
of the surface typically using cellulose acetate films may be indication of these stresses may be obtained from design
required. These replicas can be taken during the component calculations, by experimental measurement including the
inspection and assist in later stages of the investigation. use of strain gauges and by analytical stress calculation,
Surface replicas need care during preparation to obtain a which may include the use of finite element analysis.
presentative microstructure for examination. For instance Some of the flaw assessments require an estimate of the
surface oxide or decarburised layers should be removed secondary (eg residual welding stresses) as well as the
before taking the replica. Previous flaw histories may primary stress state. Often additional stresses within a
give an insight into whether the represent difficulties are component occur due to temperature gradients resulting
heralding a number of similar generic failures. in thermal stresses.
4.2 Inspection to Determine the Flaw Stresses may be determined by:
Size, Location & Orientation • Analytical formulae, which are available for simple
The initial step in the assessment procedure is to configurations such as circular or elliptical changes
determine the size of the flaw, note the plate thickness of section.
and any geometric stress raisers or other abnormalities • Strain gauging is an electronic method of measuring
such as discolouration of the surface which may denote sub-yield strains where stress does not vary rapidly
overheating or environmental attack. This assessment in time or space or where its average over the size of
will normally require a visual examination and probably the gauge (5 x 10 mm, 2 x 5 mm) is of interest.
either/or magnetic particle inspection (MPI) or ultrasonic • Photoelastic models of the structure made of
(US) inspection. In pipes other inspection methods such polymers with appropriate optical properties give
as eddy current testing may be used. semi-quantitative estimation of stress and stress
Non-destructive testing techniques are becoming more concentration factor before and (with high speed
sophisticated than required in most codes. NDT in the photography) during fracture.
 PA G E 1 8 CHAPTER 4 WTIA – TECHNICAL NOTE 10

• Finite element analysis is carried out by computer on Such stresses can be high enough to cause actual
a hypothetical net of elements into which the structure plastic deformation (in pressure vessel ends, pipemaking,
is divided. welding) in which cases residual stresses may be small
• X-ray diffraction stressometer measures stresses in or may approximate the yield point of the metal. The
surface layers of metal to which there is access. principal purpose of stress relief or PWHT is to reduce
these stresses, which may exceed the average service
Finite element analysis is increasingly being used stress multiplied by the SCF.
for stress assessments of both elastic and plastic stress
situations. The cheapest form of this assessment involves a PWHT may also slightly toughen the weld metal
two dimensional analysis (2D) with “lower order” elements. or HAZ although this would not be necessary if correct
A more accurate analysis for the same number of elements welding preheat and heat input are used. Such a heat
would generally result if “higher order” elements are used treatment also prevents (if before welding) or removes
in a changing stress field. To compensate for this a larger (if after welding) strain age embrittlement in steels which
number of lower order elements are often used. A feel have undergone plastic deformation and are susceptible
for the number of elements required can be obtained by to strain ageing. In cold formed sections in which brittle
looking at the stress contour plots to see if there are large fracture must be prevented, then, stress relief or PWHT
stress differences within a single element. Higher order is necessary.
elements can handle larger stress changes within an element As there is a general inverse relationship between
than the lower order. If a 2D model can be used because yield point and fracture toughness there is a fairly sharply
of the geometry of the component and its loading and defined level of material strength, depending on thickness,
boundary conditions this is an accurate form of assessment. above which stresses equal to the yield strength become
Some components of circular form can be model using dangerous. In this area, heat treatment is mandatory, as
‘axysymmetric” elements to give an accurate result. If the is seen in pressure vessel codes such as AS 1210. In such
component or its loading or boundary conditions are not codes detailed instructions are given for carrying out PW
adequately described by a 2D model a more costly three
dimensional analysis may need to be carried out. In determining the applied stresses, the termal stresses
should be included as these have proven extremely important
Finite element analysis can also model temperature in two significant failures in Australia in the past few years.
gradients and hence thermal stress distributions and
analysis vibrational problems. If possible experimental 4.4 Flaw Mechanism
confirmation of stresses (usually by strain gauging) is
an advantage. As the life assessment methodology is dependent upon
the likely failure mechanism, it is of the upmost importance
Where the stress information is to be used in an to correctly identify the damage mechanism. This will
assessment method the stress data is generally best put usually involve a metallurgical assessment involving a
in the format of a maximum principle profile through the metallurgical examination of the metal microstructure of
wall thickness or at least the magnitude of the stresses at failure surface (if available). Any replicas of the metal
the surfaces. Von Mises stress plots have some interest, but surface prepared during the inspection phase of the
are not generally used as input into later creep or fracture assessment would be examined. Replicas at the surface
mechanics assessments. These methods generally break condition of the component, metallurgical samples have
the stress data into primary and secondary stresses and the benefit of looking below the surface and may provide
these stresses into membrane and bending components additional relevant information. For instance the effects
with allowance for stress concentrators. of localised overheating may be confined to the surface
of the component.
Residual stresses and post weld heat treatment (PWHT)
should be considered as these stresses can influence the A visual examination of the component will frequently
fracture mechanics calculations. These stresses may arise provide some insight into the mechanism.
from:
• Fitting and jigging. 4.5 Environment
• Warm work (rolling of vessel strakes). The operating environment can have adverse affects
on the rate of crack propagation and failure mechanism
• Cold work (cold forming of vessel ends, manufacture
and must therefore be considered. The calculation of the
of roll formed structural sections).
rate of crack propagation is dependent upon the choice
• Heat treatment (phase change, differential heating of failure mechanism. The environment not only includes
or cooling. Note heat treatment can reduce stresses; the presence of corrosive chemicals, but also the effects
PWHT is carried out with this express intention). of temperature and the combined effects of these acting
• Welding (contraction of weld metal, expansion and in conjunction with the stresses.
contraction of parent metal during heating and cooling,
choice of too strong a weld metal). 4.6 Material Properties
• Straightening after welding (including the effect of The material properties required depends upon the
first pressure test on misaligned joints). assessment carried out. The material specification will
 WTIA – TECHNICAL NOTE 10 CHAPTER 4 PA G E 1 9

often provide minimum material properties including suitable material for testing. If this also is unavailable, small
yield, tensile strength, ductility and sometimes impact circular samples may be removed and a 50 mm diameter
strength. In order to determine the material specification nozzle welded in their place. AS 1210 permits such nozzles
a chemical analysis may be required. Sometimes other to be fitted without compensation pads or PWHT. If the
material properties are required including the critical value vessel cannot be shut down for this work, hot tapping
of the stress intensity factor, fatigue crack growth threshold techniques may be used to safely weld a small nozzle onto
and crack growth constants. The material properties are a pressurised shell; a device then removes a small sample
generally found from either the specification, test plate without releasing the pressure of the vessel.
records or by mechanical testing.
4.7 Risk Based Assessments
Initially the published literature should be reviewed The overall assessment of the component should
for fracture toughness data. Conservative estimates of consider the consequences of failure of the component.
fracture toughness can be obtained from Charpy data. If the consequences of failure are recognised as high,
However, if no published literature can be found, and generally far greater effort is put into ensuring the accuracy
there is no duplicate material to test, or the history of the of the calculations and defining the input data. Where the
structure is not known sufficiently accurately to prepare consequence of failure is less severe conservative estimates
a duplicate, material must be removed from the structure. for the input data are often used. Further reading on this
It is often the case that modifications to structures yield can be seen in reference.

1000MPa
200 KIC
MP
am 1
500 /2
100
MP
am 1
/2

200 40M
Pa
m1 /2
100MPa
20M
Stress, σ Pa
m1 /2
50 10M
Pa
m1 /2

20

10MPa
1mm 2 5 10mm 20 50 100mm 200 500 1m
Flaw size, a

Procedure Qualification

Ultrasonics

Radiography

}
Dye Penetrant
Surface
Magnetic Particle cracks
only
Eddy Current

Figure 4.1 Relationship between stress, flaw size, and KIC (equation 1), showing useful ranges of application of NDT techniques.
 PA G E 2 0 WTIA – TECHNICAL NOTE 10

4.8 Assessment 4.8.3 Crack extension by Fatigue – Methods

of Assessment
4.8.1 General
Once the material properties, flaw size, mechanism Generally there are two main methods of assessment
and stresses are known, the life assessment can be carried utilising either a S/N curve or fracture mechanics approach.
out. This will determine whether the flaw has reached In both instances the fatigue assessments are concerned
an “unsafe” or “critical” size and may also be used to
with the cyclic loads applied to the component. The steady
calculate the remaining life of a component before the
critical crack size is reached. This is normally performed state load can some influence in both analyses.
by determining the total life and deducting the life fraction
already used. If one of the most common forms of crack The S/N approach looks at the fatigue performance
extension such as fatigue cracking is taken as an example, of a similar component under a range of stress ranges
the cyclic stress range is required. This will probably and the number of cycles to failure of life is obtained.
require knowing the operational history as input into the This approach may use a number of factors for various
theoretical calculation of damage. In addition materials geometric. material and loading conditions. Guidance on
data is required in the form of crack growth constants for this method is given in AS 4100:1991 or BS 7608:1993.
input into the fatigue crack growth rate equation.
4.8.2 Unstable Fracture & Plastic Collapse The more recent fracture mechanics approach generally
In practice an assessment method similar to that found uses the Paris equation, which is based upon the incremental
in BS 7910:1999 or ASME Section XI is used. There extension of the crack for each application of the cyclic
are several references providing solutions for the stress load. Guidance on this method is found in the BS 7910:
intensity factor for different geometries. 1999 or ASME Section XI.
 WTIA – TECHNICAL NOTE 10 CHAPTER 5 PA G E 2 1

THE USE OF THE BS 7910 : 1999

FRACTURE MECHANICS ASSESSMENT
METHODOLOGY

5.1 Introduction as the plate thickness. The distance between indications

is also required to determine whether two neighbouring
The document looks at assessment primarily for indications will interact. For embedded indications the
fracture and fatigue (Chapter 6) with guidance on other depth beneath the surface is important, as an indication
failure mechanisms including stress corrosion cracking, may be reclassified as a surface flaw if it is close to the
corrosion fatigue, buckling and creep. The document uses surface. Likewise, deep surface flaws may be reclassified
a three level approach for the assessment of fracture using as through thickness flaws if very close to the back surface.
a two-criterion approach involving unstable fracture and An example of the test for interaction of flaws is seen in
plastic collapse. The results of the unstable fracture and Figure 5.2
plastic collapse calculations are plotted on a Failure
Assessment Diagram or FAD. Result points representing 5.2.2 Stresses
a flaw under certain operating conditions, lying beneath
the assessment line are declared as being safe, while those The stresses are defined within the document in terms
lying above the assessment line are regarded as being of membrane, bending, secondary and peak stresses. These
unsafe. Allowance is made in the document for crack are defined as follows:
growth under cyclic loading with a fatigue assessment • A membrane stress (Pm) is the component of a
methodology which offers two methods of assessment, uniformly distributed stress which is the average
a “simplified” procedure using quality categories and a value of the stress across the section that is necessary
“general” approach. to satisfy the laws of equilibrium of any internal or
If input data for the assessment of a crack is limited the external forces or moments.
first level of assessment for unstable fracture is suggested. • A bending stress (Pb) is the component arising from
It is the most simple and conservative assessment procedure an imposed force which varies across the section.
which enables an assessment with the minimum amount
of data, the second and third levels become progressively • A secondary stresses (Q) are self equilibrating stresses
more demanding for the information and accuracy required. within the structure that exist even without the influence
Each of the three levels has its own different shaped failure of external forces. Stresses arising from welding are a
assessment diagram or FAD. Most assessments would use good example. These stresses can be reduced by stress
either a level 1 or 2 assessment. The level 3 assessment relief operations.
is used primarily for strain hardening materials such as • A peak stress (F) arises from stress concentrations
austenitic stainless steel and where extensive material’s such as occur at local discontinuities and is added to
data is available. the other stresses. These stresses can arise due to gross
structural discontinuities such as a hot spot stress in
5.2 Data Input Into BS 7910:1999 tubular structures or stress concentration at nozzles,
Assessment misalignment or deviation from the intended shape
or local stress concentrations such as would occur at
The Assessment requires the information on the crack a hole, notch or sharp corner.
size, component geometry, stresses and material properties.
Each of these requirements will be looked at in turn. The addition of these stresses to form a total stress
profile is seen in figure (Figure 5.3) . This total stress
5.2.1 Crack Size and Component profile is used in the calculation. The primary stress
component lying normal to the indication is resolved and
Geometry the secondary stress assessed. The stress distribution is
This is the data provided by NDE. Each flaw is defined separated into membrane and bending components using
by its length and if known, its depth (Figure 5.1) as well a stress linearisation technique shown in Figure 5.4
 PA G E 2 2 CHAPTER 5 WTIA – TECHNICAL NOTE 10

5.2.3 Material Properties Some industries refer the use of crack opening
The tensile and yield (proof) strength of the steel is displacement (δ) in the determination of the critical flaw
required for the assessment. The fracture toughness or an size. For the Level 1 assessment the BS 7910:1999 uses
estimate is also required. the following expressions:
The accuracy of the assessment is only as good as the
input data. The indication dimensions is dependent upon
the NDE, the primary stress distribution or range upon where σy is the materials yield strength
engineering calculations, either analytical or numerical
using procedures such as finite element analysis or The value of √δr is used in the assessment plot
measurement such as with strain gauges. The tensile where
properties may be available from test certificates or by
testing and the fracture toughness is usually found by
testing, although estimates can be made from Charpy V
notch data. The other axis of the FAD is the plastic collapse ratio
(Sr). For a level 1 assessment this established using the
5.3 The Assessment for Unstable following formula:
Fracture & Plastic Collapse.
5.3.1 Introduction
Depending upon the quality of the data available and the where
accuracy of the determination required, varying levels of
assessment are available. The first two levels of assessment σ ref is obtained from an appropriate reference stress
are described in this document following. Level 1 is the solution.
simplest and most conservative assessment level. σf is the flow strength of the material which has been
The assessment requires separate individual calcu- assumed to be the average of the yield and tensile strengths
lations for both unstable fracture and plastic collapse. upto a maximum of 1.2 times the yield strength.
The value of the stress intensity factor (Ki) which is the
result of the unstable fracture calculations is divided by 5.3.3 Level 2 Assessment of Flaws
the material’s fracture toughness to give a ratio called Kr. The Level 2 assessment requires a greater understanding
The net section stress which is calculated during the plastic of the stress profile and hence the stress profile should be
collapse analysis is divided by the material’s flow stress to estimated using a method which takes this into account. The
give a value of Sr. These values of Kr and Sr are plotted on method allows for the use of either worst case estimates
the Failure Assessment Diagram (FAD). Each assessment of the input data or the use of partial safety factors on
level uses its own differently shaped FAD. The FAD for the stress, flaw size and toughness values, which vary
a level 1 assessment is seen in Figure 5.5 . The position depending upon the consequences of failure.
of the assessment point on the FAD determines whether
an indication is unsafe (lies outside the safe region in the The stress intensity factor may be obtained from solu-
FAD) or safe (lies inside the safe region in the FAD). It tions found in handbooks, numerical analysis methods or
is again emphasised that the accuracy of the assessment weight function techniques, but where this is done the
is only as accurate as the input data. method and results require documentation. The use of
worst case estimates or partial safety factors is required.
5.3.2 Level 1 Assessment of Flaws Alternatively stress intensity factor may be obtained
from the linearised stress distribution using the general
The estimation of the stress intensity factor ratio for equation:
level 1 is determined by first calculating the stress intensity
factor from the following expression:
where
a is half the flaw length for full thickness and embedded
where flaws and the full flaw length for surface flaws.
σ1 is the maximum tensile stress (Pm+Pb+Q+F) which Yσ is (Yσ)p + (Yσ)s
is assumed to be uniform across the section.
where for primary stresses
Mm is the flaw shape factor
(Yσ)p = Mƒw [ktmM kmM mPm+
φ is the elliptical integral
+ ktbMkbMb{Pb + (km – 1)Pm}]
2a is the crack length
Kr is the ratio of the stress intensity factors KI to the and for secondary stresses
fracture toughness, Kmat (Yσ)s = M mQ m+M bQ m
In the presence of stress concentrators any applied
primary and thermal induced primary stresses should be
 WTIA – TECHNICAL NOTE 10 CHAPTER 5 PA G E 2 3

magnified by the appropriate stress concentration factors. of the primary and secondary stresses, such that
This should take into account material misalignment or
geometric stress concentrations. Secondary stresses need
not be increased by this factor. The values of M, Jw, Mm
and Mb are determined graphically within the standard.
The values of Mkm and Mkb for cracks at a weld toe are The solution for ρ is given graphically in the
dependent upon two non-dimensional ratios, being ratio document.
of the total surface attachment length between the weld
toes and the material thickness and the ratio of the crack
Where a CTOD approach is utilised the following
depth to the main material thickness. For embedded flaws
relationship is relevant:
these parameters are taken as unity.
Where secondary stresses are present the additional The level 2 assessment methodology to estimate plastic
adjustment “ρ” is required to allow for plasticity interactions collapse is the same as described for method 1.

Figure 5.1 Flaw dimensions (reproduced with permission from BS 7910: 1999).
 PA G E 2 4 CHAPTER 5 WTIA – TECHNICAL NOTE 10

(a) Coplanar

Figure 5.2. Planar flaw interactions (reproduced with permission from BS 7910: 1999).
 WTIA – TECHNICAL NOTE 10 CHAPTER 5 PA G E 2 5

(b) Non-coplanar

Figure 5.2. Planar flaw interactions (continued) (reproduced with permission from BS 7910: 1999).
PA G E 2 6 CHAPTER 5 WTIA – TECHNICAL NOTE 10

(c) Non-coplanar

Figure 5.2. Planar flaw interactions (continued)

(reproduced with permission from BS 7910: 1999).
 WTIA – TECHNICAL NOTE 10 CHAPTER 5 PA G E 2 7

Figure 5.3. Schematic representation of stress distribution across section (reproduced with permission from BS 7910: 1999).
 PA G E 2 8 CHAPTER 5 WTIA – TECHNICAL NOTE 10

Figure 5.4. Linearisation of stress distributions (reproduced with permission from BS 7910: 1999).
WTIA – TECHNICAL NOTE 10 CHAPTER 5 PA G E 2 9

Figure 5.5. Level 1A FAD (reproduced with permission from BS 7910: 1999).
PA G E 3 0 WTIA – TECHNICAL NOTE 10
 WTIA – TECHNICAL NOTE 10 CHAPTER 6 PA G E 3 1

FATIGUE ASSESSMENTS USING

BS 7910 : 1999

6.1 Introduction of the minimum top maximum stress) they are needed for
the fatigue calculation, however they do not form part of
When an indication is found to be safe using the FAD the cyclic stress included in the fatigue calculation.
consideration should be given to the possibility of crack The stress range is determined from the difference
extension to the critical flaw size. Thus the flaw is tolerable between the maximum and minimum through thickness
at its present size, but under operating conditions may stress profiles. Where fatigue crack growth is near the
grow to the maximum tolerable flaw size. One mechanism fatigue threshold (ΔKo) the applied stress ratio becomes
that causes crack extension is fatigue under the action of significant as it influences ΔKo.
cyclic loading. The BS 7910:1999 offers guidance on the
likelihood of crack growth under cyclic load conditions The fatigue life is obtained by integrating the fatigue
and if crack growth occurs, an estimate of the time for equation, which for partial thickness flaws is:
it to propagate to the maximum tolerable flaw size. The
establishment of the crack growth curve is important in
estimating the remaining life of the flaw and in establishing where
inspection intervals. a is the crack depth
Δσm is the range of the membrane stress
The fatigue calculations are based upon the Paris
component
equation (crack growth equation), whichm
is an exponential
equation of the form da/dN = CΔK where da/dN is the Δσb is the range of the bending stress component
rate of fatigue crack growth, C and m are fatigue constants φ, Mm and Mb are previously defined. An additional stress
and ΔK is the stress intensity factor range. The value of correction factor (Mk) is allowed for the in the PD for the
ΔK is calculated as in the unstable fracture assessments stress concentration at the weld toe.
discussed above, except the value of the cyclic load is
Allowance should be made in the calculation of
used. The residual stress can affect the value of the fatigue
∆K for the flaws in curved shells subjected to pressure
threshold below which no fatigue crack growth occurs.
loadings.
The overall life is calculated by the integration of the
crack growth equation. The growth of the crack for one cycle should be
estimated for one cycle due to the ΔK calculated should
The BS 7910:1999 provides two methods to carry be determined from the crack growth equation. The
out fatigue assessments. These methods are called the dimension of the crack depth (“a” for a surface crack &
SIMPLE and the GENERAL techniques. The simple “2a” for an embedded crack) and “2c” should be increased
method uses a graphical technique based upon quality by the amount of crack growth. The stress intensity factor
categories. The quality categories also correlate to S/N range for the new crack shape should now be estimated and
curves found in BS 7608:1993. Allowance is made for the amount of new crack growth for the next load cycle
stress relieved structures by modifying the stress range of determined. This process is repeated until the number of
the quality categories. The general method gives guidance load cycles is completed or the maximum tolerable flaw
to the user on how to integrate the Paris equation. The two size is reached.
assessment techniques will be described in turn.
6.3 Simplified Fatigue Assessment
6.2 The General Method of Estimating Methodology (Using Quality
Fatigue Crack Propagation Categories)
The fatigue assessment is carried out using the nominal In this procedure the flaws are assessed by comparison
applied stress range acting on the section at the flaw of S-N curves which represent the actual and required
location. As residual stresses affect the stress ratio (ratio fatigue strengths of the flawed component. Ten weld
 PA G E 3 2 CHAPTER 6 WTIA – TECHNICAL NOTE 10

categories (labelled Q1 to Q10) each represented by a S-N of the variable stress ranges (equivalent to the use of
curve is available in this procedure (Figure 6.1) miners rule) using the expression:
Each of the curves lie parallel to each other with a 1/3
slope of -1/3 in the log Δσ vs log N plot. Thus each curve
is of the form
Δσ3N = constant
The required quality category is that next corresponding
However for convenience each curve is described in to the stress range next above the calculated S value.
terms of a stress range (S), which corresponds to the fatigue The quality category of the flaw has to be determined for
life of the particular S-N curve at an arbitrarily chosen life comparison against the required quality category. In order
of 105 cycles. The required weld category is determined to do this the actual flaw dimensions (a & c) are converted
from the service conditions (ie stress range and number of to an effective flaw parameter ā i, which corresponds to an
cycles) expected by the flaws component or by reference extended straight fronted crack (a/2C = 0). The maximum
to design details since several of the quality category S-N tolerable value of the flaw parameter is next calculated,
curves are the S-N curves for weld design details D to W (ā max), this corresponds the maximum tolerable flaw length
in BS 3761 at the 97.7% probability of survival. The actual to which the crack is allowed to grow. This will most likely
design category of the flawed component is determined by be the limiting flaw size calculated from the FAD. Using
calculation and reference to Table 6.1. These curves are the Figure 6.2 the value of S (Smax) is found when using
based upon the as-welded steel structures that contain high the value of ā max and the plate thickness (B). Similarly
levels of residual stresses and thus is independent of the when the value of ā i is used as an input a corresponding
load ratio R. For non-planar flaws, where the component value of Si is obtained.
is stress relieved larger maximum allowable inclusion The value of S used in the assessment is obtained
sizes are allowed. from the expression:
For constant amplitude loading, the required quality S = (Si3 - Sm3)1/3
category is determined by selecting the quality category The actual quality category for the flaw is next below
S-N curve providing a life greater than that required for the value of S in Table 6.1
the stress range and cyclic life. Where the stresses are If this value is the same or higher than the required
variable, the value of S is determined by summing effects category, the flaw is acceptable.

Table 6.1 Details of quality category S-N curves (reproduced with permission from BS 7910: 1999).

Quality category Constant in equation Equivalent BS design Stress range, S, for 2 x 106 cycles
of curve Δσ3N = constant class BS 7608:1993 Steels Aluminium Alloys
(values for steel) N/mm2 N/mm2
Q1 1.52 x 1012 D 91 30
Q2 1.04 x 1012 E 80 27
Q3 6.33 x 1011 F 68 23
Q4 4.31 x 1011 F2 60 20
Q5 2.50 x 1011 G 50 17
Q6 1.58 x 1011 W 43 14
Q7 1.00 x 1011 – 37 12
Q8 6.14 x 1010 – 32 10
Q9 3.89 x 1010 – 27 9
Q10 2.38 x 1010 – 23 8
WTIA – TECHNICAL NOTE 10 CHAPTER 6 PA G E 3 3

Operational stress range, Δσ aluminium alloys, N/mm2

Operational stress range, Δσ C & CMn steels, N/mm2

Figure 6.1 Quality category S-N curves (reproduced with permission from BS 7910: 1999).
PA G E 3 4 CHAPTER 6 WTIA – TECHNICAL NOTE 10

(a) Relation between actual flaw dimensions and effective flaw parameter.

Figure 6.2 Assessment of surface flaws in flat material (no weld toe or other stress raiser) in bending for
simplified procedure (reproduced with permission from BS 7910: 1999).
 WTIA – TECHNICAL NOTE 10 CHAPTER 7 PA G E 3 5

ASSESSMENT OF FAILURE OF OTHER

MECHANISMS
USING FRACTURE MECHANICS

Other failure or crack growth mechanisms can occur in an environmental chamber using samples described in
than are amenable to assessment using fracture mechanics. ASTM E399 or BS 7448.
These include corrosion fatigue, stress corrosion cracking
and environmental assisted cracking. The BS 7910:1999 The rate of stress corrosion cracking tends to be more
provides additional assistance in the assessment of cracks time dependent than crack size dependent and hence the rate
growing by these mechanisms. of crack growth is often described in terms of da/dt. Da/dt
is typically plotted against the stress intensity factor, K.
The presence of stress corrosion cracking can be detected
The rate of fatigue crack growth in an aggressive
by a metallurgical inspection, possibly using replication.
environment is generally greater than that in a benign
environment. This possible acceleration in fatigue crack The significance of a flaw is determined by
growth rate has to be accounted for in any fracture comparison of the applied stress intensity factor with
assessment, or else the assessment may nit be conservative. the threshold stress intensity factor for susceptibility to
Reliable data may be available from the literature as to stress corrosion cracking or KISCC. If is normal to inlcude
the effects of the specific environment on the material in a safety factor where the value of Kiscc used is obtained
question, noting that the environment includes temperature, by experimentation. It should be noted that stress corrosion
electro-potential (or presence of dissimilar materials), cracking can be very sensitive to small changes in the
oxidation levels, crack shape, flow rates and specific environment that can be caused by local concentration
loading conditions, including the “R” ratio. If this data of the aggressive agent. Any experimental testing should
is not available, testing should be carried out to ascertain take this into consideration. The testing can be carried out
the relevant rate of crack growth. This is best carried out to the test procedures of ISO 7539-6.
PA G E 3 6 WTIA – TECHNICAL NOTE 10
 WTIA – TECHNICAL NOTE 10 CHAPTER 8 PA G E 3 7

OTHER APPLICATIONS FOR FRACTURE MECHANICS

8.1 The Designer objections to repair welding even allowing for removal of
whole welds and repeating them – apart from the obvious
The design engineer will know at some stage in the financial and contractual considerations.
iterative process of balancing stresses and flaw sizes, what
toughness would appear appropriate. Alternatively he may • The job will delay other arrangements, causing financial
be restricted to a certain material for reasons of corrosion loss far beyond its own value.
resistance, etc. In either case, the material supplier can • The repair may have to be welded out of position,
assist with fracture toughness data in the form of Charpy, causing, through welder discomfort and electrode
Klc or COD test results on prospective material. In some capabilities, further flaws.
cases eg. when quantity permits or consequence of failure • The weld will be more highly restrained, increasing the
dictates, the designer may order special material and specify risk of hot or cold cracking, lamellar tearing. etc.
a fracture toughness minimum for it. • A second PWHT may cause deterioration of
The designer may also calculate that a vessel will neighbouring parent metal or weld metal properties.
“leak before break”4. In this way, leakage of the contents There is strong incentive, therefore to:
of a vessel will warn the operator to shut down before • detect flaws as early as possible.
fatigue, etc., causes the through-thickness crack to become
• assess flaws in terms of risk to the structure.
critical.

8.2 The Fabricator 8.3 The User

At an early stage of fabrication, it is relatively easy to The maintenance engineer, responsible for trouble-free
repair flaws which may occur in structures being welded operation of equipment with a minimum of unscheduled
according to a code. NDT is often not applied, however, halts to production, relies heavily on NDT for assessing
until all welding is finished and the metal has cooled from the continuing reliability of his plant. Fracture mechanics
its PWHT temperature. At this stage there are certain provides a means of determining the significance of flaws
in the plant.
PA G E 3 8 WTIA – TECHNICAL NOTE 10
 WTIA – TECHNICAL NOTE 10 CHAPTER 9 PA G E 3 9

BIBLIOGRAPHY

9.1 General Reading 16.Nibbering J. I. W., “Some Observations on COD Testing

1.“Deformation and Fracture Mechanics of Engineering (With Particular Reference to IIW—Doc. X-90178)”
Materials, R. W. Hertzberg, Wiley N.Y. 1976. IIW-Doc. X-957-80.
2.Welding Handbook. Seventh Edition Vol. 1, 17.Henderson I. D., and Kunze H. D., “Final Report
Fundamentals of Welding P. 170 AWS. on Reactor Safety Project RS102-01 – Narrow Gap
Welding of Heavy Walled Reactor Pressure Vessel
3.Knott J. F., “Fundamentals of Fracture Mechanics”
Steels”, March 1977.
– Butterworths –London 1973.
4.Rolfe S. T. and Barsom J. M., “Fracture and Fatigue 18.Brown, W. F. & Strawley, J. E., ASTM STP
Control in Structures” – Prentice Hall – Englewood 410,1966.
Cliffs, N.J., 1977. 19.Irwin, G. R., J. Appl. Mech. 1962, 29, p. 651.
5.Biggs W. D., “The Brittle Fracture of Steel” 20.Orange, T. W., in “Fracture Analysis” 1973 ASTM
– MacDonald and Evans Ltd., London, 1960. STP 560, p. 122.
6.Gillemot L. F., Criterion of Crack Initiation and 21.Fracture Toughness Testing Facilities in Australia. A
Spreading” –Doc. IIW-X-928-79. survey carried out by The Department of Productivity,
7.Ritter J. C., “Fracture Mechanics of Yielding Materials” 1979.
– Journal of the Australasian Institute of Metals, Vol. 22.Lowes, J. M., and Fearnehough, G. D., “The detection
22, No. 2, June 1977. of slow crack growth in crack opening displacement
8.Chipperfield, C. G., “Fracture Mechanics” – Metals specimens using an electrical potential method” Eng.
Australasia Vol. 12, No. 2 March, 1980. Fract. Mech. 3, 1971, pp. 103-108.
9.Egan, G. R., “Compatibility of Linear Elastic (Klc) 23.Wessel, E. T., “Linear elastic fracture mechanics for
and General Yielding (COD) Fracture Mechanics” thick-walled steel pressure vessels: material property
Engineering Fracture Mechanics, 1973, Vol. 5, Pp. considerations” Practical Fracture Mechanics for
167-185. Structural Steel ed. R. W. Nichols, UK Atomic Energy
Authority, 1969.
9.2 Testing, Specifications and
Applications 9.3 Sources of Fracture Toughness
10.ASTM E 399-94 “Plane Strain Fracture Toughness Data
Testing of Metallic Materials.” 24.“Aerospace Structural Metals Handbook” 1972 (or
11.BS 5447:1977 “Methods of Test for Plane Strain more recent updates), Belfour Stulen Inc., N.Y. Note
Fracture Toughness of Metallic Materials”. that this handbook is also good for data on mechanical
12.BS 5762:1979 “Methods for Crack Opening properties and includes many properties not normally
Displacement (COD) Testing”. available.
13.Dawes M. G., “Contemporary Measurements of Weld Knotts’ & Hertzberg’s books (2, 1) tabulate a few “typical”
Metal Fracture Toughness” Wl Research Bulletin, Vol. toughness values.
14, June 1973. 25.Publications in International Journal of Fracture
14.“Progress in Flaw Growth and Fracture Toughness Mechanics.
Testing” ASTM STP 536. 26.Banks E. E., “A Fracture Mechanics Assessment of
15.Fletcher L., “Practical COD Fracture Toughness the HAZ Properties of Australian Structural Steels”
Measurement and Evaluation” Australian Welding – Australian Welding Journal Vol. 1 8, No. 5 September/
Journal, Vol. 23, No. 4, July/August/ September, 1979. October, 1974.
 PA G E 4 0 WTIA – TECHNICAL NOTE 10

9.4 Assessment and Applications 40.Nichols, R. W., “Developments in Pressure Vessel

27.ASME Boiler and Pressure Vessel Code Section Technology – 1”. Applied Science Publishers Ltd.
XI “In Service Inspection of Nuclear Power Plant – London, 1979.
Components”. 1974. Article A-3000, p. 117. 41.“Consideration of Fracture Mechanics Analysis and
28.Folias, E. S., Int. Jnl. Fracture. Mechanics. 1965, 1, Flaw Dimension Measurement for the Trans-Alaska
p. 104- 113. Oil Pipeline Girth Welds” NBS Vols. 1 & 2, Oct. 1976,
Reports DOT/MTB/OPSO—76/02 and 03.
29.Hahn, G. T., Sarrate, M. & Rosenfield, A. R., Int. Jnl.
Fracture Mechanics 1969, 5, 187. 42.Stanley, P. – Editor, “Fracture Mechanics in Engineering
Practice” – Applied Science Publishers Ltd. – London,
30.Duffy, A. R., Eiber, R. J. and Maxey, W. A., “Recent 1977.
Work on Flaw Behaviour in Pressure Vessels”,
Symposium on Fracture Toughness Concepts, UKAEA 43.A. R. Duffy, R. J. Eiber and W. A. Maxey. “Recent Work
April, 1969. on Flaw Behaviour in Pressure Vessels”, Symposium on
Fracture Toughness Concepts, UKAEA April, 1969.
31.Jaeger T. A. – Editor, Transactions of the 3rd
International Conference on “Structural Mechanics 9.5 Partial Thickness Flaws
in Reactor Technology – Vol. 3 Reactor Vessels,
Part G, Structural Analysis of Steel Reactor Pressure 44.J. W. Kiefner, W. A. Maxey, R. J. Eiber and A. R.
Vessels.” Duffy, “Failure Stress Levels of Flaws in Pressurised
Cylinders”, ASTM STP 536, 1973, p 461.
32.PD6493 Guidance on some methods for the derivation
of acceptance levels for flaws in fusion welded joints, 9.6 Risk Based Assessments
HMSO London, 1980.
45.“Reliability Centred Maintenance” John Maubray 2nd
33.Rooke, D. P. and Cartwright, D. J. “Compendium of Ed, Butterworth/Heinemann, Melbourne 1997
Stress Intensity Factors” HMSO London, 1976.
34.“Proposed Assessment Methods for Flaws with Respect 9.7 Relevant Standards
to Failure by Brittle Fracture” – Welding in the World,
46.AS 1210 – 1997 SAA Pressure Vessels.
Vol. 13, No. 1 /2, 1975.
35.Frederick G. and Salkin R. V., “Fracture Mechanics 47.AS 1210 Supp 1 – 1990 Unfired Pressure Vessels
Assessment of Steel Plates to Brittle Fracture” – –Advanced Design and Construction.
Fracture Mechanics Conference, 3rd-5th May, 1971, 48.AS 1210 Supp 2 – 1999 Pressure Vessels – Cold-
London. Stretched Austenitic Stainless Steel Vessels.
36.Dolby R. E., and Archer G. L., “The Assessment of 49.RUL PE.10-2001 Rulings to Pressure Equipment
Heat Affected Zone Fracture Toughness” – Fracture Standards –Ruling 10: advanced Design Vessels – Use
Mechanics Conference, 3rd-5th May, 1971, London. of Joggled Joints.
Paper 53/71. 50.RUL PE. 11-2001 Rulings to Pressure Equipment
37.Burdekin F. M., and Dawes, M. G., “Practical Use of Standards – Ruling 11: Advanced Design Vessels
Linear Elastic and Yielding Fracture Mechanics with – Design Strength Valves.
Particular Reference to Pressure Vessels” – Fracture 51.AS/NZS 1554 – 1995 Structural Steel Welding
Mechanics Conference, 3rd-5th May, 1971, London.
Paper C5/71. 52.AS 1665 – 1992 Welding of Aluminium Structures.
38.Gillemot, L. F., “Criterion of Crack Initiation and 53.BS 7910:1999 Guide on Methods for Assessing the
Spreading” – Doc. IIW-X-928-79. Acceptability of Flaws in Metallic Structures
39.Wilkowski, G. M. and Eiber, R. J., “Evaluation of 54.BS 7448: Pt 1: 1991 Fracture Mechanics Toughness
the Inherent Safety Factors in the Tentative API Girth Tests.
Weld Flaw Tolerance Approach” – Battelle Final Report 55.AS/NZS 3788:2001 Pressure Equipment –In-Service
Project PR-3-101, June 29, 1979. Inspection
 WTIA – TECHNICAL NOTE 10 APPENDIX A PA G E 4 1

DEFINITIONS AND NOTATIONS

GLOSSARY OF TERMS & SYMBOLS USED IN CRITICAL: referred to stress, crack size at which a
THIS TECHNICAL NOTE major change is about to take place with increase in value,
TERMS RELATING TO FRACTURE eg. at a given stress, the critical crack size is the maximum
which can exist without causing fracture. SUBCRITICAL,
The interrelationships between some of the following then is less than this size, stress, etc.
terms are illustrated schematically in Figure 1.1
CRYSTALLINE: the appearance of a fracture surface
BRITTLE: a non-specific term describing a fracture when light is reflected brightly from FACETS so that
in which there has been little or no deformation prior to they can clearly be seen. Cleavage fractures in structural
separation. The antonym of BRITTLE is DUCTILE in steel often have a highly crystalline appearance. In high
which there is considerable deformation before fracture strength steel a cleavage fracture may not clearly show
but the division between the two is quite arbitrary facets to the naked eye and is not readily recognisable
depending on the amount of deformation considered to as crystalline.
be significant and the accuracy and type of apparatus
used to measure it. FLAW: A harmful FLAW ie. one that must be removed
CLEAVAGE: the metallurgical mechanism of because it endangers the structure and/or is not permitted
fracture of a crystalline material whereby the fracture in the relevant CODE.
follows planes, or sets of planes, in the crystalline lattice DIMPLED: see COALESCENCE OF VOIDS
Figures 2.4. CLEAVAGE fractures usually absorb V
very little energy as the crack propagates. A cleavage DUCTILE: a non-specific term describing a fracture
fracture surface may contain areas of SHEAR fracture. prior to which there has been severe deformation. See
The relative amounts of cleavage and shear fracture can BRITTLE.
strictly speaking only be established by metallographic
DUCTILE TEAR: a ductile fracture of macroscopic
examination.
size.
COALESCENCE OF VOIDS: one of the metallurgical
mechanism of fracture whereby shear fractures are FACETED: see CRYSTALLINE
generated by the flow of metal in the vicinity of voids, FAD Failure assessment diagram
usually generated from inclusions, such that the voids
grow and join together. The fracture surface then appears FAST FRACTURE: refers to the situation where the
DIMPLED (Figure 2.2). elastic or potential energy stored in a structure exceeds
CODE: is an Australian or other Standard detailing the energy absorbed in crack propagation. Consequently
methods for carrying out design, fabrication, operation a crack, once initiated, continues without further energy
or other engineering functions. input in the form of increased load, stress, strain, pressure,
etc. This type of fracture is therefore more common in,
COMPLIANCE: is the deformation of a structure per but not exclusive to BRITTLE materials.
unit force. Its units are m/N. it is the inverse of STIFFNESS
or deflection of a structure per unit force on it. Its units FIBROUS: the silky appearance of the surface of a
are therefore m/N. The COMPLIANCE affects the crack that has failed by shear and does not show bright
fracture behaviour because the higher the compliance crystalline facets.
(ie. the SOFTER the system) the greater the elastic;
FLAW: a discontinuity in a weld zone or other part of
strain or potential energy stored as a given load which
a structure. No detrimental effect is necessarily implied
encourages UNSTABLE FRACTURE. Conversely, a
in this term. See FLAW.
HARD system stores relatively little energy, which makes
STABLE FRACTURE more likely (see 2.2.7). Figure 1.1 Schematic relationships between terms.
 PA G E 4 2 APPENDIX A WTIA – TECHNICAL NOTE 10

FRACTOGRAPH: photographic record of the a propagating crack. It is sometimes extended to the case
appearance of a fracture surface. when the supply of energy just falls below that required
for propagation and the crack is arrested. A fast, long,
GRAIN BOUNDARY EMBRITTLEMENT a pheno-
ductile tear in a pipeline is strictly a stable crack when
menon causing BRITTLE fracture with a characteristic
the conditions at every point are uniform.
INTERGRANULAR appearance, Figure 2.6.
TOUGH: A non-specific term indicating fracture with
HARD: see COMPLIANCE. high-energy absorption or a material which fractures in
INTERGRANULAR: describes fracture caused this manner. In metals this usually occurs by SHEAR,
by stress corrosion or GRAIN BOUNDARY ie. plastic deformation. The fracture appearance is silky
EMBRITTLEMENT. or FIBROUS.
Kr: fracture ratio of the applied elastic K value to Abbreviations used in this Note
Kmat AS* Australian Standard
NON DESTRUCTIVE TESTING: testing by methods ASME* American Society of Mechanical
that do not destroy the object. Engineers
ASTM* American Society for Testing Materials
PLANE STRAIN: the condition of deformation in
which. all strain occurs parallel to one plane. As no BS* British Standard
strain occurs perpendicular to this plane, no thinning or COD Crack (tip) opening displacement
thickening occurs and subsidiary stresses arise or must CVN Charpy V Notch — a “notch toughness” test
be applied to conform to this condition. Plain strain is expressed in energy per cm2.
therefore associated with high RESTRAINT. It can not EAC Environmentally Assisted Cracking
occur in thin shells. HAZ Heat Affected Zone
PLANE STRESS: the condition of deformation in LEFM Linear Elastic Fracture Mechanics
which all stresses are parallel to one plane. The material NDT Non-destructive Testing
is therefore free to thin (or neck) down or thicken. SCC Stress Corrosion Cracking
PLASTIC CONSTRAINT: severe RESTRAINT SCF Stress Concentration Factor
such as triaxiality of stresses in thick material or in SEM Scanning Electron Microscope
T or Cruciform joints, which encourages plane strain SI* The International Standards Organisation
deformation. See AWRA TN 6. System of Units
POP-IN: an initial increment of crack growth SMYS Specified Minimum Yield Strength
sometimes observed in fracture toughness testing. See UTS Ultimate Tensile Strength
Appendix B. YFM Yielding Fracture Mechanics
RADIOGRAPHY: a non-destructive method in * A standardisation organisation or its designation
internal examination in which metal or other objects
Symbols
are irradiated with X- or gamma radiation. Differences
in thickness, density, or absorption, caused by internal 2a crack length
discontinuities, are apparent in the shadow image either 2ac critical crack length
on a fluorescent screen or on photographic film placed c other crack dimension m
behind the object.
Co fatigue crack growth rate coefficient
RESTRAINT: refers to STIFFNESS of a system d crack depth
usually with respect to the stresses arising from welding.
e base of natural logarithms
SHEAR FRACTURE: the metallographic mechanism E Young’s modulus
whereby a crack is generated within a crystalline body by
g acceleration due to gravity (9.8 m sec-2)
the flow caused by shearing forces set up within the body.
G fracture toughness parameter
SLOW FRACTURE: occurs when crack propagation
K stress intensity factor
requires more energy than is available from the potential
energy of the system. Crack propagation therefore only Kl plane strain stress intensity factor
occurs with, and at a comparable rate to, the application Kc critical stress intensity factor
of load, stress, strain, etc. This type of fracture is more Klc critical plane strain stress intensity factor
common in tough materials.
KISCC stress corrosion cracking limit Pam 1⁄2 see 3.5
SOFT: see COMPLIANCE. KR cracking resistance stress intensity factor
Sr: load ratio of applied load to flow strength load. M,Mb,Mm,Mp bending correction factors, see 3.2, 3.3
STABLE FRACTURE: occurs when the energy n fatigue crack growth rate exponent
supply is continually just equal to the energy absorbed by p pressure
 WTIA – TECHNICAL NOTE 10 APPENDIX A PA G E 4 3

P load ∆ε difference in the strain range

Pm,PQ see 3.3 εy yield strain
Q flaw shape factor (see 3.3) λ (lambda) ratio of crack dimension to material
t material thickness dimension
r radius of plastic zone at crack tip υ (nu) Poisson’s Ratio (usually = œ)
R radius of pressure vessel π (pi) ratio of circumference of circle to its diameter
Vg clip gauge opening σ (sigma) stress
Y calibration factor σb bending stress
z height of clip gauge attachment knife edges σp failure stress due to partial thickness flaw
δ (delta) crack opening displacement(m) σu UTS
δc,δi,δm critical COD, COD at pop in, COD at max σy yield strength
load σ flow stress = (σy + σu)/2
∆ (delta) difference operator Ø diameter
∆K difference in the stress intensity factor range φ, φc (phi) nondimensional COD, its critical value
∆σ difference in the stress range (epsilon) strain
PA G E 4 4 WTIA – TECHNICAL NOTE 10
 WTIA – TECHNICAL NOTE 10 APPENDIX B PA G E 4 5

MEASUREMENT OF FRACTURE TOUGHNESS

PHYSICAL MEASUREMENT OF compliance also varies with cross-head position and

FRACTURE TOUGHNESS thickness and width of loading fixtures. The testing
system compliance also depends on specimen size.
B.1 Pre-cracking Facility Thus some enormous machines have been built to test
wide plates at eg. SMEC and BHP-MRL to duplicate
The major impediment to the establishment of a fracture behaviour to real structures. A very hard machine would
toughness testing capability is the necessity to pre-crack represent a liquid-filled pressure vessel. An ideally soft
the test specimens by fatigue. The fatigue load may be machine would simulate the conditions at the crack front
applied by any convenient means. The most common in a gas pipeline.
methods use either servo-hydraulics, or natural frequency
vibrations from electro magnetic apparatus. Higher cycling The use of soft testing systems will in general give
rates are possible with the latter form. A dedicated servo conservative results; that is, will encourage unstable
hydraulic machine may be constructed for the purpose fracture rather than stable fracture in a test piece. In
with a nominal capacity of 100 kN at up to 30 Hz. testing the more ductile materials, where toughness
increases with crack growth, it is more desirable to use
Cycling rate at a given load would depend upon the a hard machine to avoid the unstable fracture which
deflection. Typical fatigue load values would be 60 kN might be caused by using a soft machine. For detection
to 2 kN at 20 Hz. of initiation, the characteristics of the testing machine
This equipment would allow pre-cracking of about 6 are unimportant and an ordinary hydraulic machine is
specimens per day. With automatic compliance monitoring, satisfactory. “Pop-in” is a form of crack initiation which
changes in deflection which occur with crack growth can appears slightly differently on the two types of machine.
be used to shut the system down at pre-set intervals thus On a hard machine it is associated with a definite drop in
obviating the need for constant supervision. load (Figure. B.4. 2 ii) while on a soft machine it is much
less marked (Figure. B.4.2 iv).
The principle of using a dedicated pre-cracking
machine is strongly recommended. If a servo hydraulic B.2.2 Clip Gauges
testing machines with dynamic capabilities is used for pre-
cracking it means that a very expensive machine is tied Clip gauges for the measurement of crack mouth
up for long periods in pre-cracking and is unavailable for opening displacement, and hence the derivation of crack
general use. A good quality optical binocular microscope tip opening displacement in Klc and COD tests can be
with a calibrated eye piece graticule is necessary for purchased or homemade. They consist of strain gauged
observation and measurement of crack length. 500 ohm wire
resistance
B.2 Testing Equipment strain guage

B.2.1 Testing Machine

Fatigue crack Tension
A universal testing machine of 500 kN capacity will
enable the performance of three point bending B x 2B Compression
Spacer
COD tests in material up to about 75 mm thick. block

Ideally, the compliance of the testing machine should 0.45-0.55 W Tension

match that of the structure represented in the test. A
hydraulic machine is soft, and a mechanical machine is Gauge mounted on testpiece
hard. A servo-hydraulic machine in load-control mode is
softer than when in displacement-control mode. Machine Figure B.2.1 Design and assembly of clip gauge.3
 PA G E 4 6 APPENDIX B WTIA – TECHNICAL NOTE 10

double cantilever beams (Figure. B.3). The beams should location. In that case the specimen is only supported on
be constructed from a material which is capable or large a single roller, and special jigging is necessary in order
elastic deformations and should preferably have a low to be able to lower it onto that roller in the exact position
elastic modulus eg. titanium. Proprietary gauges with required for the test.
outputs of only 12 mV/mm give satisfactory performance The decision as to whether a conditioning bath should be
even on small COD specimens with short cracks. used during the test is best left to the laboratory concerned.
B.2.3 Non-ambient Temperature Testing It should be borne in mind however that the accurate
measurement of surface temperatures which are greatly
The orientation in which testing is performed is an different from the surrounding air poses many difficulties,
important consideration, ie. crack uppermost, or the other and may need special thermocouple probe design.
way round, if testing is to be performed at other than
ambient temperature bath. If it is performed in a bath, then B.3 The Determination of Klc
it is necessary that the crack be at the top so that the clip
gauge can be free of the effects of the bath fluid. If this is Standards BS 5447 and ASTM E399 govern the
not done then the effects on non-ambient temperature on experimental determination of Klc. The latter covers two
clip gauge calibration will have to be taken into account, types of specimen geometry, a three point bend SENB4
and more obviously the reliability of the clip gauge will with W = 2B and a compact tension specimen CT (Figure
be compromised. B.3.1).
If testing is done with the crack at the top and especially Both specimens contain a fatigue crack of the depth
when it is done in a bath, this causes difficulties in specimen indicated in Figure B.1.

Figure B.3.1 Fracture touchness test piece geometries.

 WTIA – TECHNICAL NOTE 10 APPENDIX B PA G E 4 7

The specimens are loaded in a testing machine and a Klc measurements, and at the crack tip calculated from
plot is made of force against the amount of opening of the the surface measurement. The standard COD specimen is
crack mouth. The latter is measured by the clip gauge. a 3 point notch bend specimen SENB4 with W = 2 (see
Figure B.3.1). This specimen has been chosen because
The load at the onset of crack growth, PQ, is measured it is compact and has a simple deformation mode during
from the load/displacement graph, and this is used in one general yielding when the two arms rotate as rigid bodies
of the following equations to determine Klc. The method (apart from change in elastic deformation) about a hinge
of determining Pj from the graph is rigidly specified in point in the ligament (see Figure B.4.1).
the standard.
According to slip line theory, for a rigid-plastic
The equations are essentially the same as the basic
material, the centre of rotation is 0.5 (W-a) below the
equation (1) with geometry correction and an expression
notch tip. BS 5762 gives an empirical relationship which
for the specimen compliance.
includes an elastic deformation term as a function of a
The initially calculated value of K (normally termed modified stress intensity factory K that depends on the
KQ), must satisfy certain conditions before it can be classed ratio a/W
as a ‘valid’ Klc.
δ= +
These are
1) Pmax/PQ < 1.10
Where V P is the plastic component of the clip
2) The stress intensities during fatigue cracking are limited
gauge displacement. A shorter crack length has been
to preclude the danger of plastic deformation at the
proposed.16
crack tip.
The force displacement record obtained from a standard
B.4 Determination of δc, the Critical test is usually one of the five types shown in Fig. B.4.2.
COD In type (i) fast unstable fracture occurs without any prior
slow stable crack growth and the COD obtained at fracture
The COD is sensitive to the root radius of the notch is denoted as δc.
and it is necessary to use either a fatigue crack of a very
sharp machined notch to obtain valid results. Pop-in occurs in type (ii) at a force of Pc. That is, a
sudden but small advancement of the crack front occurs
The limiting radius is of the order 0.02 mm which, in the mid-thickness and not at the edges. The unstable
although small, is much larger than the maximum root fracture initiates under plane strain conditions, but is
radius required for valid Klc tests in high strength materials immediately arrested. The reasons for this arrest may be
(Figure 2.9). BS 5762 only permits a fatigue crack for the relaxation in triaxiality as the fracture nears the plate
COD testing. surface and/or the increase in radius of the crack tip due to
the toughness of the material. In homogeneous material the
The COD should be measured indirectly at the surface
arrest subsequent to Pop-in may be due to loss of driving
of the specimen with the same clip gauge as is used for
force in insufficiently soft loading systems.

Z When testing a weldment, Pop-in is generally

Knife edge associated with initiation of unstable fracture in the HAZ
Vg and subsequent arrest in tough base metal. After Pop-in,
further straining is required to reinitiate the fracture which
now develops shear lips at the plate surface. Material which
Machined slot shows this ductile behaviour subsequent to pop-in gives
a
an ambiguous COD and using pop-in as the δc gives an
Fatigue crack extremely conservative figure. With type (iii) to (v) slow
δ stable growth occurs before instability. In this case the
COD that corresponds to initiation is denoted by δj.
Plastic Determination of the point of initiation on the curve
hinge 1 (W-a)
M n M is not always easy and is discussed below. Fast unstable
fracture occurs after slow crack growth, but before the
C W maximum load is reached in type (iii). Pop-in occurs in
θ type (iv) after initial slow crack growth. The behaviour
of the ductile specimens is represented by type (v)
where the maximum load P is attained before final
fracture which may or may not be unstable. Although
θ in the standard(the COD at maximum load δm is one of
the COD’s listed to be recorded) it does not have any
Figure B.4.1 Relationship between δ and Vg, for rigid
fundamental significance and is not a material constant.
rotation about C.12
As the size of the specimen is reduced so the COD at
 PA G E 4 8 APPENDIX B WTIA – TECHNICAL NOTE 10

maximum load, δm, decreases, and since δj is constant, B.5 The Determination of Fracture
for plane strain fracture, initiation occurs closer to the Toughness of Welds
maximum load. In very small specimens, initiation can
occur after maximum load. It is unlikely, in this case, B.5.1 General
that the value of δj obtained can be applied to a full scale The designer needs to know the fracture toughness
structure. Thus in very thin plate, it may be necessary to of weld metal, HAZ and parent plate as affected by the
use a specimen wider than that specified in the standard. heat of welding. It is hoped that, in the future, tables of
However, the material would be so ductile that in a full such values for Australian steels will be readily available.
scale structure the critical flaw size would probably be In the mean time, however, engineers or others may well
measured in metres. have to arrange or actually carry out tests to determine
these values. Some guidance is therefore given here on
Detection of fracture initiation is difficult when there
the fracture toughness testing of weld metal. Fracture
is slow stable crack growth. The standard follows the
toughness testing of HAZ is even more difficult than that
technique which suggests that several specimens should
of weld metal and is best done in co-operation with an
be tested, stopping each test after a different amount of
experienced testing organisation.
slow crack growth. The COD at the end of the slow crack
growth is determined for each specimen and the results B.5.2 Forming a Straight Fatigue Pre-crack
extrapolated to zero crack growth to determine δj.
The residual stress pattern in a weld zone may cause
The slow crack growth can be obtained in steel a W-shaped fatigue crack to grow, or more than one.
specimens by breaking the remaining ligament at liquid Fatigue cracks which are not straight give spurious results
nitrogen temperature. With non. ferrous specimens, the in fracture toughness tests, so precautions must be taken
initial crack growth can be stained either with recorder ink, to ensure that a single straight crack grows.
or by heating, before breaking them open. Unfortunately,
Precompression
multiple specimen or it-Curve techniques for determining
δj are not applicable to weldments because of the scatter Precompression relieves residual stresses in the
in results. standard through-thickness notched B x 2B weld metal
test specimen. The shape of the fatigue crack obtained
Alternative methods are: passing a constant AC or DC on square-notched specimens with and without the use of
current through the specimen and measuring the change precompression is shown in Fig. B.5.1. Fig. B.5.2 shows
in potential across the notch22 or detection of acoustic the effect of the improper crack front shape upon the
emission from cracking with piezo-electric crystals.23 observed fracture toughness.

(i) (ii) (iii) (iv) (v)

Pm
Pc Pc Pi
Pc Pc Pi Pi Vi
Vi Vi

Vc Vc Vc Vc Vm
Force

c – critical
i – initiation
m – maximum

Vp Vp Vp Vp Vp
Clip Gauge Displacement

Figure B.4.2 Types of force displacement records.

 WTIA – TECHNICAL NOTE 10 APPENDIX B PA G E 4 9

The chevron notch (Figure B.5.3) ensures that a single

crack is nucleated at the point of the chevron and guided Fatigue Machined As welded
by the grooves until it reaches the full thickness of the As welded and pre-
crack notch
material. Figure B.5.4 shows the alternative precompression compressed
treatments13 Precompression is thus an essential element
of the COD testing of weld metals. Very large forces are
required for large specimens. A 2500 kN (250 tonne) short
stroke hydraulic jack in a dedicated load frame can be
used for this task.7 effective ε/εy is increased to 1.45. Fracture
The non dimensional COD design curve is = 1 2 which surface
corresponds to a critical half crack length of (a) (b)
Figure B.5.1 The fatigue pre-crack front shape obtained
ac = 0.133 with through thickness notched, as welded, weld metal
COD specimens (a) without precompression and (b) with
taking εy as 1.5 x 10-3, the estimate of critical half precompression.13
crack length is:

The half length of a through crack equivalent to the There is a factor of over 7 between the equivalent
surface crack 1 mm deep and 40 mm long, from Figure flaw size and the calculated maximum flaw size for
4.2.6.1. using fracture initiation. There would also be a considerable
margin between the strain necessary to initiate a fracture
a/2c = 1/40 and a/B = 1/16, is found to be a/B = 0.076 and that necessary to propagate the fracture to a critical
or a = 1 2 mm
length at which catastrophic fracture would occur. Thus
it is concluded that the vessel is safe to operate for a
limited period. Since the existing crack may grow by
 Fatigue cracked & stress relieved stress corrosion it would not be safe to continue operation
O Locally compressed, fatigue
Critical Stress Intensity Factor Nmm-3/2

cracked and stress relieved. without planned periodic inspections.

4,000   WORKED EXAMPLES

O
USE OF BASIC FRACTURE MECHANICS
3,000
EQUATIONS

O B.1.3 Case Studies
2,000 O


O O
O Example 1
Valid K1C 
O
O

O A 2 m wide, 100 mm thick aluminium alloy 5083-0
O
1,000 O O plate component has a through-thickness planar crack 20
O
 O  mm long in a weld aligned perpendicular to the applied
stress. The plate supports a load of 1,400 tonnes. The
material fracture toughness is 25 Mpam1⁄2. Is it safe?
-200 -160 -120 -80 -40 0
– Converting to Sl units
Temperature °C
Width = 2m
Figure B.5.2 The influence of fatigue crack shape on Thickness = 0 1 m
fracture toughness.13 Force = 1.4 x 103 tonnes
= 1.4 x 106 kg
= 1.37 x 107 N
Straight fatigue crack
Stress = 1.37 x 107 /(2 x 0 1) = 69 MPa
– Using equation (1)
ac = Klc2/(πσ2) = 252/(π x 692) = .0424m.
ie. the total length of a critical crack is 85 mm, so
clearly a 20 mm crack is not critical.
Before declaring the structure safe, however, the
Fracture surface engineer should ascertain how the crack grew during
service to 20 mm. Processes not covered by the elementary
Figure B.5.3 The chevron notch ensures nucleation of a fracture mechanics covered so far eg. fatigue or stress
single fatigue crack which is then guided by the grooves corrosion, might cause the crack to continue to grow and
until it reaches full thickness of the material.
reach the critical size.
 PA G E 5 0 APPENDIX B WTIA – TECHNICAL NOTE 10

Example 2 λ = 0.12 and it may be verified that Y = 1.79 in comparison

with √π or 1.77 implicit in equation (1).
A 1.2" transverse crack is discovered in the 10" wide
lower flange of a crane rail which sustains a maximum EXAMPLE OF THE CALCULATION OF THE
tensile stress of 25 ksi. The fracture toughness of the steel
BULGING FACTOR
is approximately 150 ksi√in. Is the rail safe?
Method 1: K = B.2.5.1 Example:
= 25√π x 0.6 A 9ft. diameter 3" thick sphere contains a 16"long
34.3 ksi√in through wall crack. The material properties are as follows:
Method 2:a = Kc2/(πσ2) Yield stress σy = 30.7 ksi
= 15O2/(π x 252) Tensile strength σu = 59.4 ksi
= 11 5" Fracture toughness Klc = 102 ksi√in
Method 2a = 23" What is the burst pressure? (This example was taken
Method 1: Conclusion from real life – the burst pressure was actually 1245 psi).

The applied stress intensity K (≈30 ksi√in) is clearly Working

much less than the critical value (Kc = 150 ksi√in) and a = 8"
the girder is quite safe. r = 1.8"
Method 2: Conclusion R = 54"
The 1.2" crack needs to grow to 23" to become critical, t = 3"
so the girder is quite safe. M = 1.28
The use of the elementary basic fracture toughness σ = Klc/1.28 √π x 9.8 = 14 4 ksi
equation (2) can result in a critical crack size that exceeds The pressure corresponding to this hoop stress is;
the size of the component, and the answer may seem P = 2tσ/R = 1596 psi
ridiculous. Nevertheless, the important conclusion that
the component is safe remains unaltered. The application of this model to partial-thickness
flaws is discussed below.
CALCULATION OF THE CALIBRATION FACTOR
Example: Consider the crane rail described in example USE OF BASIC FATIGUE (PARIS) EQUATION
B.3.1.2 with the following modifications: B.4.1 Example
(i) the lower flange is cracked at the edge rather than
somewhere within it. A transverse crack extending 8 mm in from the surface
(ii) due to building alterations, the rail is in the open is found in the 300 mm steel tie rod of an extrusion
air overnight and immediately before use the next press. The press load of approximately 1,850 tonnes is
morning is at 0°C, at which temperature its fracture shared equally by four such rods. Between press cycles,
toughness has fallen to 55 ksi√in. which occur 4,100 times per month, there is no load on
the tie rods. A fracture mechanics evaluation yields a
Calculation of Calibration Factor. The flange may be shape factor Y = 2.1 in equation (14) for this crack and
represented by an edge notched tensile member, in which: concludes that when the crack is 60 mm deep the tie rod
For this value of A, the appropriate formula for Y is will be in danger of brittle fracture. The fatigue behaviour
Y = 1.99 - 0 41 (0 12) + 18.70 (0.12)2 - 38 48 of the steel is known from experiment to be.
(0 12)3 + 53.85 (0.12)4 It is desired to keep the press operating. How long
= 1.99 - 0.049 + 0.269 - 0.066 + 0.011 will it be possible to put off replacing the tie rod?
= 2.15 The following are the alternatives available to the
and equation 4 can be used. plant engineer:
(i) 24 months, when the tie rod was due to be written
Method I
off.
K=Y (ii) 7 months, when the next scheduled maintenance
Method II period is due.
Conclusions: Both calculations show that the crane (iii) 21⁄2 months, which is the soonest a new rod of
rail is on the point of catastrophic failure. (Note: This identical specification could be delivered.
is an illustrative example. The steel specified for crane (iv) 2 weeks, which is the soonest a makeshift tie rod
rails would in fact be required to have adequate fracture could be installed.
toughness 20°C below its minimum possible service (v) Operation is unsafe, and must be suspended until
temperature. See 46 and TN 11). the tie rod is replaced.
In the centre-cracked version of this example, Data
 WTIA – TECHNICAL NOTE 10 APPENDIX B PA G E 5 1

Max load = 1850 tonnes = 18.1 MN

Area of tie rods = 4π (0.15)2 = 0.283m2
Max stress in
tie rods 18.1/0.283 = 64 MPa
Min stress = 0
∴ Stress range
Δσ = 64 MPa
ao = 0.008m
ΔK = YΔσ√a = 12 MPam1/2
Working
Conclusion:
The tie rod can be replaced after 7 months as scheduled,
but cannot safely be left 24 as per financial forecast.
Notes
(i) By substituting in equation (16), it can be calculated
that the time it would take the crack to grow to
60 mm would be 82,000 cycles, or 20 months.
(ii) The approximation that Y is constant is acceptable
as a/d grows from .027 to 2.
(iii) If the approximation (equation 17) is made that the
crack continues at the initial da/dN of 1 5 x 109m/
cycle, and increase of only 12 mm from 8 to 20 mm
is calculated after 20 months and to 23 mm after
24 months. This approximation is unacceptable.
PA G E 5 2 WTIA – TECHNICAL NOTE 10
 WTIA – TECHNICAL NOTE 10 APPENDIX C PA G E 5 3

OTHER MEASURES OF FRACTURE TOUGHNESS

OTHER MEASURES OF FRACTURE parameter. Like δc it is potentially more useful in ductile

materials than is Kic; however as yet there is no extensive
TOUGHNESS data on the values of Jlc for various materials, and it is
not likely to be of great use to practicing engineers for
C.1 ‘J’ Integral some time.
The J integral, like the COD, shown in Fig. C.1.1
Note that Jlc =
extends Fracture Mechanics well into the elastic-plastic
regime. The most simple explanation of a J Integral is
graphical. Consider two test specimens, identical except
C.2 ‘R’ Curves
that one has a crack of length a, the other has a crack of C.2.1. Variation of Toughness with Crack Length
length a + δa.
In normal circumstances many structural applications
The superimposed load/displacement records of these operate in generalised plane stress so that final failure,
two are shown in Figure C.1.1. whether of a ductile or brittle nature, can be preceded
The area under each curve represents energy (work by the significant amounts of slow stable crack growth
= force x distance) and the difference in area, the energy from a pre-existing flaw. This apparent increase in
change due to a crack increment ‘δa’. The area is Jδa fracture toughness after initiation is due principally to
for a unit thickness specimen or JBδa for a specimen of crack blunting and is analogous to work hardening. A
thickness B. Thus ‘J’ is defined: unique plane stress equivalent of Klc, which can define
the stable growth process, does not exist. The YFM
Where U/B is the potential energy per unit thickness initiation parameters δi and Ji are material parameters
and ‘a’ is crack length. largely independent of test piece geometry but since ductile
The critical value of J, ie. ‘Jlc’ is a material toughness materials usually display considerable resistance to crack
growth after initiation, Fig. C.2.1(a) the use of initiation
values is normally unnecessarily conservative. Because
of this, the resistance (R) curve based on KR δR or JR is
Linear elastic
attracting increased attention. GR in the ductile curve is
Crack length a a general fracture mechanics parameter representing KR
δR or JR. The R Curve will predict when cracks in some
materials and structures will stop automatically as the
material properties change.
Shaded area
Load, P

is JBδa In order to make use of the cracking resistance or

Crack
R Curve of a component it is necessary to examine the
length structure and determine the corresponding variation in its
a + δa compliance or crack driving force with crack length.
C.2.2 Variation of load with crack length
C.2.2.1 Crack path geometry
The structure properties must also be considered.
Displacement, u The driving force under a constant load will change with
crack length. As a crack traverses a reinforced panel the
Figure C.1.1 Derivation of J integral from load-
driving force will change drastically at each stiffener.
displacement curves.8
 PA G E 5 4 APPENDIX C WTIA – TECHNICAL NOTE 10

KR

δR
Ductile
GR

Klc

Brittle

GO
aO
(a) a (b) a

Figure C.2.1 Variation of fracture behaviour with crack length (a) ductile material (b) brittle material.

The driving force will initially increase due to the lower and hard systems. A system is soft if the load is applied by
section remaining to support the load L, but drops in the springs or very long and thin tension members, because
reinforced section as the crack runs into the thicker section. the load varies only slightly with deformation. If the load
Different structures will have different driving force vs is applied by gravity the system is infinitely soft. If the
a curves, ie. they will have different compliances, and load is applied by short thick tension or compression
there will of course be a different curve for each applied members the system is hard, and any deformation, crack
stress level. opening or crack extension rapidly reduces the load. If the
load is applied by an ideal screw or wedge the system is
C.2.2.2 The concept of system compliance or stiffness infinitely hard. The variation of load with crack extension
Structures vary widely in compliance, ie. the level is shown in Figure C.2.3.2(a). The resultant stress induced
of constraint with which they elastically oppose or in a beam is shown schematically in (b) and the resultant
encourage deformation of an element of the structure. crack driving force is shown in (c).
Figure C.2.3.1 illustrates the concepts of soft, intermediate Note that in the type of structure shown in Figure
C.2.2. crack driving force is not a simple function of load.
P Compliance has an important effect on crack behaviour,
P3
which may be understood by considering the elastic or
potential energy stored in a stressed system.
P2 This energy available up to a certain crack extension
G a., is analogous the area under the curve in Figure C.2
3.2 (a). It can be seen, then, how much more energy is
P1 stored in a soft system than in a hard system This energy is
immediately available to perpetuate an energy-absorbing
(ie. tough) fracture which would be stable in a hard system.
As will be seen below, a certain material may give slow,
stable crack growth in a hard system, but unstable fracture
in a soft system.
a > >
P3 P2 P1 In terms of crack driving force, a soft system has a
steeply rising curve, while in a hard system the crack
P driving force may actually reach a maximum and start
a decreasing as crack length increases. These differences
also occur among types of testing machines (See B.2).
Figure C.2.2 Effect of specimen geometry on relationship The crack driving force curve (representing the system)
between toughness and crack length.
may now be superimposed on the cracking resistance R
 WTIA – TECHNICAL NOTE 10 APPENDIX C PA G E 5 5

Soft Soft

Force Stress
P σ Intermediate
Intermediate

Hard
Hard

(a) (b)
ao a1 Crack length a

Soft Soft

Stress
Intermediate Intensity
Factor
K Intermediate

Hard Hard

(c)
a a
Figure C.2.3.1 (a) (b) (c) Effect of compliance on fracture behaviour. (a) Force vs crack length.
(b) Average stress vs crack length. (c) Stress intensity factor vs crack length.

K
K K σ = σ4
σ = σ3 KR
KR KR KR
a b c d

K
σ = σ2 KR = σ4 πa 4
a2
a1 a4

σ = σ1 σ = σ1

ao ∆a ao ∆a ao ∆a ao ∆a

Figure C.2.4 (a), (b), (c), (d) Comparison of crack driving force and cracking resistance (R) curves.
 PA G E 5 6 APPENDIX C WTIA – TECHNICAL NOTE 10

Curve (representing the material of a given thickness) • No success has been achieved in correlating CVN
to determine whether or at what load or stress unstable energy with COD in Weld metals.
fracture will occur.
The Charpy test is also used to characterise the
C.2.4 Use of R-Curves brittle-ductile transition temperature in terms of fracture
appearance or energy absorbed in several ways.
The driving force in the loaded system may be stated
• Nil ductility transition temperature (NDTT) is the
in terms of K, J or δ for comparison with the cracking
highest temperature at which the whole fracture
resistance curve KR JR or δR of the material plotted against
surface is crystalline in appearance.
crack length as in Figure C.2.4.
• 20J or other energy level transition temperature is that
Therefore, if the driving force exceeds the energy at which a smooth curve, drawn through a plot of CVN
absorbed in creating the new surfaces unstable crack test results, intersects that energy.
growth will occur or continue. If the driving force is less
• 50% fracture appearance transition temperature (FATT)
than the cracking resistance, fracture will not initiate and
is that at which the fracture surface is 50% fibrous and
if already occurring, will stop.
50% crystalline on a macroscopic scale.
In a) the crack will be stable if it is less than ‘a1’.
Empirical correlations exist that relate a Charpy impact
In b) the crack will grow until it reaches ‘a2’ and value to the Klc, however they should be used with caution
then stop. until a data base is built up. Eg. for a ‘bridge steel’
In c) the crack will grow continuously ie., the crack
is unstable. An example is a soft testing machine. K2lc/E = A (CVN),
In d) the two curves are just tangent, and the stress where A = 4-8
level, σ4 defines the load at which unstable crack growth Another suggested relationship is
will occur.
K2lc/E = 2 (CVN)3/2
At present R Curves are used only for thin sheet
structures such as in aircraft, and so are not a feature of (N.B. Both equations are in British units)
general structural design. For the low accuracy required of a fracture mechanics
evaluation (within a factor of 2) fracture toughness data
C.3 Charpy Impact Values can often be obtained from the empirical correlation with
The Charpy test is probably the most widely used Charpy data. However great care should be taken, because
measure of the fracture properties of materials. The value of the number of ways already mentioned in which the
yielded by the test is called “Notch Toughness”. Charpy test fails to represent the real situation.
It has the advantages: Better correlation is found between Klc values and
CVN values measured on an instrumented specimen at
• It is a cheap and simple test and is used universally.
the same (slow) strain rate as that at which Klc values
• It uses small samples. are determined.4 Such slow bend CVN test results
• It is useful to compare materials or treatments. conservatively estimate Klc using equation (C) with A
However it has the disadvantages: = 5 for low-strength structural steels, below and in the
transition temperature range. CVN impact data similarly
• It does not meet all of the criteria mentioned in Section give estimates of KID
2.2 for useful fracture toughness tests.
• It produces only empirical values that although useful The following correlation has been found for high
for comparison, are of little use in design unless an strength steels (σy ≥ 100 ksi) on their upper shelf:
extensive body of empirical experience is available, eg. (also in British Units)
it cannot be used to evaluate the danger of a known flaw. Charpy values are used by designers of pressure
• The Transition Temperature determined from a vessels to select their materials taking into account
series of Charpy values is likewise not the transition tensile properties, thickness, stress relieving treatments
temperatures found from the fundamental fracture and operational temperatures, but the relationships are
toughness measured on full-thickness test pieces. based almost entirely on experience.
 WTIA – TECHNICAL NOTE 10 APPENDIX D PA G E 5 7

THEORY BEHIND GENERAL YIELDING

FRACTURE MECHANICS
YIELDING FRACTURE MECHANICS –
CRACK-OPENING DISPLACEMENT

D.1 Yielding at the Crack Tip an experiment is described in 7.4. The degree of safety
A crack tip is usually considered to have a profile of the structure of interest may then be estimated from
similar to that shown in Figure D.1 (a). However the the empirical COD design curve. For the purposes of
zone at the crack tip stretches before breaking as shown the COD design curve, it is convenient to define a new,
in (b). This crack opening displacement (COD) or amount non-dimensional parameter φ where
of stretch, measured directly in inches or metres, is the φ = δ/2π ε y a (in general)
parameter which is useful in materials too tough for the φ = (4/π 2) In sec (πδ/2σ y ) for σ ≤ σ y
use of LEFM. φ c = δ c /2π ε y a max
Tough and ductile material stretch more ie. have greater
for a particular material. As a means of estimating φc
COD values than do brittle and hard materials. is given below, this will define a critical crack dimension
It should be recognised however, that COD is usually amax. φc varies with δc/amax which means the larger it is,
dependent on thickness, and only data relevant to the the tougher is the material. It is related to σ/σy, and the
application should be used. COD Design Curve37 allows the safe φ to be calculated
for any stress a (knowing σy). Once the safe φ is known
D.2 The Concept of Critical COD calculation of δc/a shows whether the situation is safe.
For applied stresses below the yield strength the crack
opening displacement is given by: D.3 The COD Design Curve
Up to σ/σy = 0.5, the safe region is bounded by the
curve = (σ/σy)2. That is, a crack is safe if φ ≥ (σ/σy)2.
Where σy and εy are the yield strength and the yield For σ/σy = 0.5, the safe region is bounded by the line
strain respectively. φ = (ε/εy) – 0.25.
Again a crack is safe if φ ≥ (ε/εy) – 0.25.
For small stresses (In sec θ = 1⁄2θ2 when θ is small)
this equation reduces to: Using COD to determine the significance of a known
crack in a vessel (ie. whether the vessel will fail) requires
several steps encompassing the various relationships
implied in the above inequalities.
The critical COD, δc at which a material of a certain These steps are:
thickness breaks may be determined experimentally. Such 1) Determine δc from eg. a laboratory experiment on the
material. Other material properties needed are σy and
E or εy;
2) Estimate the service stress in the material.
3) Compare the service stress with the yield stress, and
decide which of the above equations apply.
4) Use the applicable equation to derive
5) Knowing (from experiment), to give the maximum
safe (ie. critical) crack half length amax
a b 6) The crack existing in the pressure vessel wall is generally
not through-thickness. By reference to Figure 4.2.6.1
we can find the half length a of a through-thickness
crack of equal severity.
Figure D.1.1 Conventional (a) and actual (b) crack tip 7) Compare amax (5) with a (6): if a < amax the structure is
configurations.
judged to be safe
PA G E 5 8 WTIA – TECHNICAL NOTE 10
 WTIA – TECHNICAL NOTE 10 APPENDIX E PA G E 5 9

CASE STUDIES

E1 EARLY CASE HISTORY WELD E1.2 Detection of Flaws

FLAWS IN NEW BOILER During final erection of the boiler and attachment of
40 tubes and down-comers to the imported boiler drum,
E1.1 Main Problem opportunity was taken to ultrasonically examine a number
This case involved the acceptability of a major boiler of the nozzle-to-shell welds which did not appear to have
drum in a power station where many drum nozzles the same quality as required for the attachment welds to the
contained flaws significantly outside the limits proposed tubes. This resulted in detection of flaws in many nozzles
by the purchaser. It involved contractual problems involving as shown in Fig. E.1.1. This boiler drum was one of two
hundreds of thousands of dollars and the availability of which contained over 200 nozzles in each drum.
the power station at a critical time.
E1.3 Required Weld Quality
The drum was required to comply with BS1113 (Water
Tube Boilers) which specified that:
“no butt weld in a drum . . . shall be acceptable if it
40 mm has . . .
100 mm φ
1) cracks or areas of lack of fusion or incomplete
Nozzle penetration . . .
2) any elongated inclusion having a width exceeding 1 5
s mm or a length exceeding . . . 13 mm for any thickness
exceeding 25 mm . . . ”
d
The drum fabricator also agreed that the nozzles be
inspected ultrasonically.
100 mm Defect
thick shell
E1.4 Check on Weld Quality
C
L At the time, ultrasonic testing (UT) in Australia had
not been widely developed to a high level particularly
of for complex nozzle welds; and naturally the overseas
Nozzle
fabricator suspected that UT had been incorrect. However,
their re-examination on site of the nozzle welds indicated
that the Australian tests had been very thorough and the
flaws were reported accurately.

E1.5 Confusion over Weld Quality

A major aspect of this problem related to terminology.
The BS1113 committee intended that a “butt weld in a
drum” only applied to the drum and not to nozzle welds
which at the time were not covered in detail in any code.
However, the BS definition of “butt welds” indicated that
nozzle-to-drum welds were in fact butt welds and therefore
Figure E1.1 Early Case History – Flaw in Boiler Nozzle.
were literally required to meet the above weld quality.
 PA G E 6 0 APPENDIX E WTIA – TECHNICAL NOTE 10

The fabricator’s prior experience had suggested that the Depth (d) 2.4 mm 2.4 mm 2.7 mm 3.0 mm
level of quality used in these welds had been successful in Length any 60 mm 14 mm 11 mm
boiler drums where no UT had been the norm. Hence his
UT at a level to indicate gross flaws had been considered Note: If flaw is near a surface and s 7 d (see Figure
adequate by him. It should be stressed that at this time UT E1), the flaw depth is taken as s + d in depth, and the
of nozzles were relatively new – previously only visual and allowable depth is 1⁄2 the listed values.
possibly magnetic particle inspection had been used. Overseas special research using fatigue testing
All the authorities involved were concerned with indicated independently that slightly larger flaw depths
flaws which greatly exceeded the limits thought to be might be acceptable.
covered in specifications, particularly as nozzles are a
E1.8 Re-examination of Each Nozzle
critical area in any boiler. They were naturally worried
about the future safety and reliability of the boiler over Using the above criteria and allowing for the small
a long life, keeping in mind one shutdown of a power errors which can occur with the measurement of these
station boiler may cost more than $100,000. flaws, the acceptance of each nozzle was reviewed. Three
border-line cases arose and it was decided to repair these
E1.6 Resolution of Acceptable Weld Quality in-situ and to include another safety valve nozzle which
could be conveniently repaired in the same area where
All parties including the power and boiler authorities post weld heat treatment was required.
concerned appreciated that:
a) very large cost and delay would result from extensive E1.9 Service Experience
repair to the boiler drum which had most tube-to-nozzle No problems have been reported with the boiler and
welds already made; and it is understood that periodic ultrasonic checks are made
b) weld acceptance standards for these welds might to monitor any flaw developments.
justifiably differ from that for main butt welds. E1.10 Observations from this Early
As a result, assistance was sought in judging what Experience
was an acceptable standard. This was done using past These are:
experience with cracked boiler nozzles, FM and knowledge a) FM or full scale fatigue and brittle fracture tests were
that the weld quality was typical of the highly experienced the only means of giving a quantitative estimate of
drum fabricator. acceptable flaws.
b) FM used in a most conservative manner gave results
E1.7 FM Analysis
which were extremely useful and economic.
At the time no standards were available relating to FM c) The FM results were tempered to some extent by
and an analysis from first principles taking approximately 50 previous experience and “engineering judgement”.
hours was carried out using linear elastic fracture mechanics d) The authorities concerned accepted the recommendations
(LEFM). This was considered most appropriate because: on the basis of a sound case, despite no standard
a) low cycle fatigue was expected to be the main mode guidelines being available.
of crack extension; and e) This case is typical of a number which arise in
engineering in that flaws are detected at an embarrassing
b) for conservatism, very low fracture toughness was
stage when repairs are difficult. A multi-million dollar
assumed (a lower bound value of 40 MPa x/m or 36
ksi √in). Other conservative assumptions included: case is the Trans Alaska Oil Line where welds which had
not been inspected required untimely inspection. Weld
(i) the values of stresses (pressures, thermal, nozzle acceptance standards for these welds were changed as
bending, and residual) and stress ranges in the area a result of FM analysis to avoid further huge losses,
of the flaw (use was made of US photo-elastic without reducing the reliability of the pipeline (Ref. 1).
studies on nuclear vessel nozzles), f) Non-destructive testing of known accuracy was an
(ii) allowance for thickness and shape departures, essential requirement for safe application of FM.
(iii) slag inclusions were treated as planar flaws,
E2 RECENT CASE HISTORY WELD
(iv) the maximum number of pressure and temperature
cycles in the life of the boiler was used,
FLAWS IN NEW PRESSURE
VESSEL
(v) a safety factor of 1.5 on Klc or 2.3 on flaw size
was used. E2.1 Main Problem
The following was recommended and at the time was Recently one of the first carbon steel pressure vessels
probably the first recommendation on the acceptance of fabricated in Australia with ultrasonic inspection replacing
continuous slag in boilers or pressure vessels: radiographic examination was found to have mid-thickness
 WTIA – TECHNICAL NOTE 10 APPENDIX E PA G E 6 1

weld flaws slightly greater than specified in the code. E3 CASE HISTORY – WELD FLAWS
Shell thickness of 100 mm necessitated heavy repairs if
the flaws were to be removed. Acceptance of such flaws
IN EXISTING VESSELS
was requested. E3.1 Main Problem
E2.2 Detection of Flaws A number of carbon steel transportable vessels were
found to not comply with Statutory Regulations and on
The flaws which consisted of three slag inclusions up examination the welds (and materials) did not meet pressure
to 25 mm length were located along the circumferential vessel standards. In view of the huge costs involved, a
weld using standard ultrasonic testing (UT) prior to heat critical assessment of their safety was required to establish
treatment. Using special UT techniques the flaws were whether the vessels should be replaced.
sized and located accurately. The height of the flaws did
not exceed 2 mm and appeared to be fine slag inclusions. E3.2 Detection of Flaws
The flaws were remote from structural discontinuities, eg.
Flaws first detected by UT and RT when the vessels
openings and supports.
which had been in service for a number of years were being
E2.3 Fracture Mechanics Assessment assessed for acceptance by the Statutory Authority. These
flaws consisted primarily of continuous incomplete root
A Very simple LEFM assessment was made in a penetration at mid-thickness of double-welded butt joints.
very conservative manner and this showed that such
flaws could be readily accepted for this particular high E3.3 Assessment of Flaws
pressure application. This was tackled in two ways:
E2.4 Acceptance of Flaws a) full scale mechanical tests; and
These flaws were accepted because: b) fracture mechanics analysis.

a) FM coupled with wide experience and judgement In this case it was practicable to use existing vessels to
indicated that the flaws should not affect safety or supply test samples which showed clearly that the welds
performance; were able to safely withstand the internal pressure. This was
largely due to the relatively low nominal pressure stresses
b) Accurate flaw sizing by UT indicated the height of in the shell during discharge, which resulted from relatively
the flaw and its orientation with respect to maximum thick material and because of weld reinforcement.
stress. The height of the flaw is vital in FM but is not
measured in radiographic examination. A check using FM indicated that through-thickness
flaws up to approximately 100 mm could be tolerated.
c) Following postweld heat treatment and hydrostatic
testing, further inspection showed no growth in flaw size. E3.4 Observations
d) Radiographic examination was used to re-examine the a) Full scale mechanical testing and fracture mechanics
flaws but there were no signs on the radiograph – ie. were used together to provide clear evidence for the
the flaws would have been accepted if the standard acceptability of the welds.
radiographic examination was used.
b) To ensure that no excessively long flaws were present,
E2.5 Observations a pressure test of each vessel was carried out to check
for leaks, cracks and distortion.
a) FM was able to give a valid basis for acceptance of
the flaws for this particular application. c) To ensure that the vessels are not discharged unsafely
each vessel should be visually examined for severe dents.
b) UT can be far more searching than radiographic testing
and will be more likely to reject thick welds if the weld d) This recent and important example is typical of a
acceptance standards for radiographic examination are number of cases where the costs of testing and FM
used. It provides data on the height of the flaw, which analysis are justified.
is not obtained by RT.
c) Authorities and owners are loath to accept flaws E4 CASE HISTORY – PNEUMATIC
outside the code values accepted in the Contract as TESTING OF LARGE VESSEL
this encourages a general reduction in the present
E4.1 Main Problem
quality which can be readily achieved in Australia
and throughout the world. Extensive modifications were to be made to a very
large and old pressure vessel constructed of carbon steel. It
d) However, where special conditions exist and a sound was required to pressure test the vessel after modification.
case is made, eg. as above, authorities and owners Because of the internal refractory lining and probable
may accept flaws outside the code limits. foundation damage, hydrostatic testing was to be replaced
e) The weld acceptance values in SAA Codes are being by pneumatic testing. This test is dangerous and it was
reviewed to ensure they are equally applicable to UT required to determine the conditions under which the
and RT. pneumatic test should be carried out.
 PA G E 6 2 APPENDIX E WTIA – TECHNICAL NOTE 10

E4.2 Detection of Flaws measure of security against subsequent failure up to the

It was not practicable to assess all flaws of this vessel design pressure provided there is no significant increase
which was over 40 m high. However, a good idea of the in flaw size or decrease in material toughness.
general quality of welds was known from experience of
construction at that time and by assessment of the large E5 CASE HISTORY – FATIGUE
sections removed from the vessel for modification. Vessel CRACKING IN ROD MILL
construction had been to spot radiographic standards in
the 50’s but replacement nozzles, etc. were to be subject to E5.1 Main Problem
100% RT and MT. Hence an estimate only was available
of most flaws, but a very good idea of the quality was A number of large horizontal ball mills were found
known for the areas of high residual stress where welded to be cracked at a time when maximum output from
modifications were made. the plant was required. It was clear that the mills had
operated safely with cracks, but it was required to know
Weld imperfections included linear and angular what was the best action to take to minimise costs and
misalignment as well as shape departures due to local avoid catastrophic failures, particularly as two cracks had
high temperatures in the shell during service. each exceeded 10 m in length.
E4.3 Material Properties E5.2 Detection of Flaws
From technical references and review of plate These were first noted visually and checked by
certificates for the vessel, it was clear that the steel had magnetic particle inspection, local grinding and UT. Cracks
relatively poor fracture toughness. Brittle fracture could were located at the toes of the fillet welds attaching the
be a real risk if high residual stresses were present in two central flanges to the shell. Cracks had propagated
areas with large undetected flaws and when stressed at around the shell up to a depth of 12 mm.
low temperatures. Service temperatures were high.
E5.3 Assessment of Cracking
Hence Charpy V and COD tests were conducted at
different temperatures on steel samples taken from the Stress analysis indicated that the bending stress range
cut-outs for nozzles, etc. was almost sufficient to account for the cracking. This
and the position of cracking suggested fatigue cracking.
E4.4 FM Analysis The stress analysis was complicated by the high axial
load developed by rod wedging.
Using LEFM and YFM with:
a) A conservatively estimated flaw size; Approximate FM analysis indicated that cracking
would continue and lead to complete failure within a
b) The maximum stress during the test (including certain number of mill revolutions.
100% allowance for residual stress and for estimated
misalignment); and This indicated that the mill with the largest cracking
c) The measured material properties, it was possible should be remedied as soon as possible. Crack growth
to find a suitable testing temperature to avoid brittle rates increase rapidly with increased crack size and ΔK.
fracture. E5.4 Observations
This required a slight lowering of the test pressure to a) As a result of the analysis, repairs were made by
avoid unnecessary overstressing and cost of preheating, extensive welding in-situ in a way to avoid unequal
yet at the same time be able to detect a flaw which could stresses. The repair method could not completely
become critical in later service. relieve residual stresses but improvements were made
by reducing the stress concentrations.
E4.5 Observations
b) This case is a very common example of fatigue cracking
a) FM greatly aided in determining the test pressure (1 in a wide range of machines and structures used in
2-1 35 x design pressure) and the test temperature materials handling and mining.
(over 70°C) to provide a safe yet suitable test.
c) The use of FM to estimate the progression of cracks
b) FM enabled residual stresses to be catered for and the critical crack size is invaluable in deciding the
realistically. optimum timing of repairs to avoid excessive loss of
c) The greatest hazard occurs with this test during local production, failure, and excessive repair.
first yielding of the steel with reshaping of the vessel d) FM aids in providing a sound basis for the initial
and redistribution of yield point residual stresses. commissioning of critical equipment which may be
d) This test is a “warm overstressing” and provides a good subject to distortion (or high thermal stresses).
WTIA – TECHNICAL NOTE 10 APPENDIX F PA G E 6 3

PERMISSIBLE SERVICE TEMPERATURES

ACCORDING TO STEEL TYPE AND THICKNESS

Table F1 Permissible Service Temperatures According to Steel Type and Thickness.

Permissible service temperature, °C
Steel type
Thickness, mm
(see Table F.2)
≤6 >6 ≤12 >12 ≤20 >20 ≤32 >32 ≤70 >70
1 –20 –10 0 0 0 5
2 –30 –20 –10 –10 0 0
3 –40 –30 –20 –15 –15 –10
4 –10 0 0 0 0 5
5 –30 –20 –10 0 0 0
6 –40 –30 –20 –15 –15 –10
7A –10 0 0 0 0 —
7B –30 –20 –10 0 0 —
7C –40 –30 –20 –15 –15 —
8 –40 –30 — — — —
Note: This table is based on available statistical data on notch toughness characteristics of
steels currently made in Australia or New Zealand. Care should be taken in applying this
table to imported steels as verification tests may be required. For a further explanation,
see WTIA Technical Note 11

Tables F.1 and F.2 are reproduced with permission from AS/NZS 1554.1:2000 Structural
Steel Welding Part 1: Welding of Steel Structures, pages 73 and 32.
 PA G E 6 4 APPENDIX F WTIA – TECHNICAL NOTE 10

Table F2 Association of Steel Type Numbers to Australian and New Zealand Steels.
Specification and grade of parent metal
Steel
type AS/NZS 3678 AS/NZS 3679.1 NZS 3415
AS 1163 AS 1397 AS 1450 AS 1548 AS/NZS 1594 AS/NZS 1595 AS 2074
AS/NZS 3679.2
1 C250 G250 C200 7-430 HA1 All Grades C2 200 250 Fe 430A
G300 H200 7-460 HA3 C3 250 300
C250 HA4N C7A-1 300
H250 HA200 A1006
HA250 XK1016
HA250/1
HU250
HA300
HA300/1
HU300
HA1006
HA1010
HA1016
HXA1016
2 C250 L0 — — 7-430L0 — — — — 250L0 Fe 430C
7-460L0 300L0
3 — — — 7-430L20 XF300 — — 250L15 250L15 Fe 430D
7-430L40 300L15 300L15
7-430L50
7-460L20
7-460L40
7-460L50
4 C350 G350 C350 5-490 HA350 — C1 350 350 Fe 510A
H350 7-490 HA400 C4-1 WR350 400 Fe 510B
HW350 C4-2 400
C7A-2
5 C350 L0 — — 7-490L0 — — — WR350L0 350L0 Fe 510C
400L0
6 — — — 5-490L20 XF400 — — 350L15 350L15 Fe 510D
5-490L40 400L15 400L15
5-490L50
7-490L20
7-490L40
7-490L50
7A C450 G450 C450 — — — — 450 — —
7B C450L0 — — — — — — — — —
7C — — — — — — — 450L15 — —
8 — — — — XF500 — — — — —
 WTIA – TECHNICAL NOTE 10 APPENDIX F PA G E 6 5

LODMAT ISOTHERMS °C
(Lowest One Day Mean Ambient Temperature)
Based on records 1957 – 1971
kindly supplied by:
Australian Bureau of Meteorology.
PA G E 6 6 WTIA – TECHNICAL NOTE 10
WTIA – TECHNICAL NOTE 10 APPENDIX G PA G E 6 7

MEMBERS OF WTIA TECHNICAL PANEL 1

PRESSURE EQUIPMENT

Mr Randeep Agarwal, Woodside Energy Ltd

Mr Stan Ambrose, Consultant
Mr Greg Bartley, GWB Engineering Pty Ltd
Mr Harry Better, ETRS Pty Ltd
Mr Alan Beveridge, Loy Yang Power Management Pty Ltd
Mr Brennan Borserio, Incitec Ltd
Mr Joe Boutros, Pressure Vessel Engineering
Mr Bruce Cannon, BHP FPD
Dr Colin Chipperfield, CRC for Welded Structures
Mr Robert Crnkovic, HSB (Australia) Pty Ltd
Mr Andrew Crocker, ANSTO
Mr Richard Eager, Consultant
Mr Alistair Forbes, Welding Technology Institute of Australia
Dr Peter Ford, FCA Australia Pty Ltd
Mr Mark Gobell, ASC-Engineering Pty Ltd
Mr Roger Griffiths, BP Refinery (Bulwer Island) Ltd (Chairman)
Mr Brian Jacobs, Caltex Refineries (NSW) Pty Ltd
Mr Lician Kent, Retired
Mr Michael Liddell, Esso Australia Ltd
Prof Valerie Linton, Adelaide University
Dr Stuart Mann, HRL Technology Pty Ltd
Mr Alan McClintock, HERA NZ
Mr Mike Pitrun, Wollongong University
A Prof John W Price, Monash University
Mr Bob Small, Pacific Power
Mr Garry Smith, Edison Mission Operations and Maintenance
Dr Hugh Stark, University of NSW / Stark Pty Ltd
Mr Zoran Sterjovski, University of Wollongong
Mr Robert Svensk, HSB (Australia) Pty Ltd
Mr Joe Vetter, Stanwell Corporation Ltd
Mr Tim Vukomanovic, BHP Direct Reduced Iron Pty Ltd
PA G E 6 8 WTIA – TECHNICAL NOTE 10
 WTIA – TECHNICAL NOTE 10 PA G E 6 9

EXPERT TECHNOLOGY TOOLS

These Technical Note, Management System and other Expert Technology Tools may be obtained from the WTIA.
Technical advice, training, consultancy and assistance with the implementation of Management Systems is also
available through the WTIA’s OzWeld Technology Support Centres Network and School of Welding Technology.
WTIA – PO Box 6165
Silverwater NSW 1811 Australia
Phone: +61 (0) 2 9748 4443
Fax: +61 (0) 2 9748 2858
Email: [email protected]
Visit our Internet site at http://www.wtia.com.au

WTIA Technical Notes TN 6-85 – Control of Lamellar Tearing

Describes the features and mechanisms of this
TN 1-96 – The Weldability of Steels important mode of failure and the means of controlling
Gives guidance on the preheat and heat input conditions tearing through suitable design, material selection,
(run size, current, voltage) required for acceptable welds fabrication and inspection. Acceptance standards, repair
and to avoid cold cracking in a wide variety of steels. methods, specification requirements and methods of
The Note is applicable to a wide range of welding investigation are proposed. Four appendices give details
processes. on the mechanism, material factors, tests for susceptibility
and the important question of restraint.
TN 2-97 – Successful Welding of Aluminium
This note covers the major welding processes as they TN 7-98 – Health and Safety in Welding
are used for the welding and repair of aluminium and its Provides information on all aspects of health and
alloys. Information is given on the processes, equipment, safety in welding and cutting. Designed to provide
consumables and techniques. It also provides information this information in such a way that it is readily useable
on the range of alloys available and briefly covers safety, for instruction in the shop and to provide guidance to
quality assurance, inspection and testing, costing and management. Recommendations are given for safe
alternative joining processes. procedures to be adopted in a wide variety of situations
found in welding fabrication.
TN 3-94 – Care and Conditioning of Arc
Welding Consumables TN 8-79 – Economic Design of Weldments
Gives the basis and details for the correct care, storage Principles and guidance are given on methods and
and conditioning of welding consumables to control procedures for optimising design of weldments and
hydrogen and to ensure high quality welding. welded joints and connections to maximise economy in
welding fabrication. Factors influencing the overall cost
TN 4-96 – The Industry Guide to Hardfacing for of weldments which need to be considered at the design
the Control of Wear stage are discussed.
Describes wear mechanisms and gives guidance on
TN 9-79 – Welding Rate in Arc Welding
the selection of hardfacing consumables and processes
Processes: Part 1 MMAW
for a wide range of applications. Includes Australian
hardfacing Suppliers Compendium 1998. Gives practical guidance and information on the
selection of welding conditions to improve productivity
TN 5-94 – Flame Cutting of Steels during manual metal arc welding (MMAW). Graphs are
Gives a wealth of practical guidance on flame cutting provided showing rates as a function of weld size. The
including detailed procedures for efficient cutting, selection graphs enable a direct comparison of different types of
of equipment and gases, practices for identifying and curing welding electrodes when used for butt and fillet welds in
defective cutting, methods of maximising economy and various welding positions.
other important guidance on the use of steels with flame
TN10-02 – Fracture Mechanics
cut surfaces.
Provides theory and gives practical guidance for
Flame Cut Surface Replicas the design and fabrication of structures, planning of
These have been developed to complement Technical maintenance and assessment of the likelihood of brittle
Note Number 5 by defining three qualities of flame cut or ductile initiation from flaws in ferrous and non-ferrous
surface. Each set of three is contained in a convenient alloys. Engineering critical assessment case histories are
holder with a summary sheet of main flame cutting data. discussed.
 PA G E 7 0 WTIA – TECHNICAL NOTE 10

TN 11-98 – Commentary on the Structural Steel TN 18-87 – Welding of Castings

Welding Standard AS/NZS 1554
Provides basic information on welding procedures for
(A joint publication with AISC). the welding processes used to weld and repair ferrous and
The Note complements AS/NZS 1554 parts 1 to 5, by non-ferrous castings. It also provides information on the
presenting background information which could not be range of alloys available and briefly covers non-destructive
included in the Standard. It discusses the requirements of inspection, on-site heating methods and safety.
the Standard with particular emphasis on new or revised
clauses. In explaining the application of the Standard to TN 19-95 – Cost Effective Quality Management
welding in steel construction, the commentary emphasises for Welding
the need to rely on the provisions of the Standard to achieve Provides guidelines on the application of the AS/NZS
satisfactory weld quality. ISO 9000 series of Quality Standards within the welding
and fabrication industries. Guidance on the writing,
TN 12-96 – Minimising Corrosion in Welded development and control of Welding Procedures is also
Steel Structures given.
Designed to provide practical guidance and information
TN 20-94 – Repair of Steel Pipelines
on corrosion problems associated with the welding of
steel structures, together with possible solutions for Provides an outline of methods of assessment and repair
minimising corrosion. to a pipeline whilst allowing continuity of supply.
TN 13-00 – Stainless Steels for Corrosive TN 21-99 – Submerged Arc Welding
Environments Provides an introduction to submerged arc welding
(A Joint publication with ACA) equipment, process variables, consumables, procedures
Provides guidance on the selection of stainless steels for and techniques, characteristic weld defects, applications
different environments. Austenitic, ferritic and martensitic and limitations. Describes exercises to explore the range
stainless steels are described together with the various of procedures and techniques with the use of solid wire
types of corrosive attack. Aspects of welding procedure, (single and multiple arcs) and provides welding practice
design, cleaning and maintenance to minimise corrosion sheets, which may be used by trainees as instruction
are covered. sheets to supplement demonstrations and class work, or
as self-instruction units.
TN 14-84 – Design and Construction of Welded
Steel Bins TN 22-02 – Welding Electrical Safety
Written because of the widely expressed need for Provides information and guidance on welding
guidance on the design and fabrication of welded steel electrical safety issues: welding equipment, the human
bulk solids containers, this Technical Note gathers body and the workplace.
relevant information on functional design, wall loads,
stress analysis, design of welded joints and the fabrication, TN 23-02 (Draft) – Environmental Improvement
erection and inspection of steel bins. It also contains a Guidelines
very comprehensive reference list to assist in a further Provides information and guidance on how to reduce
understanding of this very broad subject. consumption in the Welding and Fabrication industry,
while reducing the impact on the environment at the
TN 15-96 – Welding and Fabrication of
same time.
Quenched and Tempered Steel
Provides information on quenched and tempered steels Binder (holds approximately ten Technical Notes).
generally available in Australia and gives guidance on
welding processes, consumables and procedures and WTIA Management Systems
on the properties and performance of welded joints. MS01-TWM-01 Total Welding Management System
Information is also provided on other important fabrication Interactive CD-ROM
operations such as flame cutting, plasma cutting, shearing Welding Occupational Health, Safety &
and forming. Rehabilitation Management System
TN 16-85 – Welding Stainless Steel MS02-OHS-01 OHS&R Managers Handbook
This Technical Note complements Technical Note MS03-OHS-01 OHS&R Procedures
Number 13 by detailing valuable information on the MS04-OHS-01 OHS&R Work Instructions
welding of most types of stainless steels commonly
MS05-OHS-01 OHS&R Forms and Records
used in industry.
Four Expert Technology Tools incorporated into
TN 17-86 – Automation in Arc Welding one Interactive CD-ROM
Provides information and guidance on all the issues MS06-ENV-01 Welding Environmental
involved with automation in arc welding. The general Management System (Draft)
principles are applicable to automation in any field. Interactive CD-ROM
 WTIA – TECHNICAL NOTE 10 PA G E 7 1

WTIA Pocket Guides

These handy sized Pocket Guides are designed to
be used on a practical day-to-day basis by welding and
other personnel.

PG01-WD-01 Weld Defects

Will assist Welders, Welding Supervisors and others in
the identification and detection of defects, their common
causes, methods of prevention and in their repair.

PG02-SS-01 Welding of Stainless Steel

A concise guide for Welders, Welding Supervisors
to welding processes and procedures for the fabrication
of stainless steel including Codes, Standards and speci-
fications, cleaning and surface finishing, good welding
practice and precautions.
PA G E 7 2 WTIA – TECHNICAL NOTE 10