BS '.76D3:2014
Guide to fatigue design and
assessment of steel products
...meking excellence a habit
B5 ?EflB:2D1Il
BRITISH STAHDAHD
Publishing and copyright information
The El5t copyright notice displayed in this document indicates when the document
was last issued.
@ The British Standards Institution 2014
Published by E5t Standards Lirnited 21114
l5EI'-l Eli-‘E El SEE 5405"] 5
lE5 51 ,lJEl].1D
The following H5! references relate to the worl: on this document:
Committee reference WEEJET
Draft for comment IBIEEHEIIEIEZ DE
Publication history
First edition April T993
Second lpresentl edition March EH11‘-l
Amendments issued since publication
Date
Tent affected
BRITISH STAN DAHD
B5 TEDHIZDIII
Contents
Foreword iv
Scope
i‘
Normative references E
Terms and definitions 3
Symbols and units E
Fatigue assessment procedure 3
Design life H
Fatigue loading 5
flI*'-il 'Ili|'-li |Il"i-|l -
El'l1v'l|'Cii"llTiEI‘iII-HI CDl"iSICI'E'l'iEl‘lZIDl’iS
ID
SI
‘ID
Factors on fatigue life l'l‘
Features influencing fatigue behaviour
11
12
13
Fracture mechanics LE
Classification of details 1'2
Unclassified details 3.?
11
workmanship and Inspection
‘IS
‘IE
Stress calculations #2
Ailowabie fatigue stresses S3
fl‘
32
.|ll|.l1I'lE!1t-ES
Anne:-t A [normative] Fatigue design
EB
Anneii B {normative} Errpianatory notes on detail classification
Annex C {normative} Guidance on stress analysis B2
PD
Anne:-i D [normative] Guidance on the use of fracture mechaniu
Hit?
Anne: E {normative} Fatigue testing and the use of test data to define design
S‘lZI"E5-SES
Ill-5
Anne:-i F {normative} Weld toe improvement techniques
‘HS
Anne:-i G [normative] Assessment of tubular node joints ‘i3-El
Anne:-i H [normative] Cycie counting by the reservoir method
Bibliography
T35
BB
List of figures
Figure ‘i -— Definition of length l. for use in thiciinesebending correction B
Figure 2 - Reference stress in parent metal 43
Flgure 3 - Reference stress on weld throat 45
Figure 4 - Typical example of stress concentrations due to geometrical
discontinuity 4?
Figure S — Typical example of stress concentration caused by a geometric hard
spot dd
Figure E - Fatigue stress concentration factors =15
Figure T — Comparison of nominal, structural and hot-spot stresses in a beam
with a welded cover plate SD
Figure B - Fieiative stiffness effects on the fluctuating load in a bolt in a
concentrically clamped and concentrically loaded bolted joint 5.?
Figure El e Iviean S,-N cunres 5-if
Figure ‘iii — Standard basic design S,-H curves SE
Figure 11 — Sr-Al curves for bolts with threads under direct loading {class it-‘ll
SB
Figure 12 — Modifications made to S;l‘-l curves for welded joints in sea water E3
Figure 13 — Typical S;N relationship E5
Figure B.1 - Welds at plate edges It
Figure B.2 - Failure modes at weld ends and weld toes of welded
attachments 5'2
Figure B.3 — Failure modes In cruciform and T-joints for joint types indicated 5'3
Figure B.-ii — Failure modes in transverse butt welds for joint types indicated F3
Figure B-S — T-junction of two flange plates F5
Figure B.E — Cruciform junction between flange plates FE
Figure BJ - Alternative method for joining two flange plates ?E
Flgure B.B - Local grinding adjacent to cope hole in type 5.2 joint F?
ID The British Standards Institution ZD14
e
i
BS TBUBIZB14
BRITISH STANDARD
Figure BB - Lise of continuity plating to reduce stress concentrations in type 2.1
and 2.2 joints 29
Figure B.‘iiIi — Example of type 'i".5i or 2.4 T-joint B-El
Figure B.il — Single fillet corner weld in bending {type ?.Ei}
B-El
Figure B.l2 — Example of a third member slotted through a main member Bf
Figure C.‘i — l beam with cover plate showing distribution of structural stress and
definition of hot—spot stress B-it
Figure C.2 - Types of hot—spot BE
Figure C3 - Possible brick element model of an I beam with a cover plate B2
Figure C.-ii - Node numbers superimposed upon the weld mesh section in Figure
C.3cj BB
Figure C.S -- Node numbers superimposed upon the weld mesh section in Figure
C.3ej BB
Figure C.E — Element numbers superimposed upon the weid mesh section in
Figure C-3cl B5‘
Figure C.i' - Element numbers superimposed upon the weld mesh section in
Figure C.3ei BS
Figure CB - Calculation of the SSE stress at node n2 of the brick mesh shown in
Figure C.3 to Figure C.S BE
Figure C.'.ii — Possible shell element model of an I beam with a cover plate Ell‘
Figure C.t{i — Node numbers superimposed upon the shell element weld mesh
section in Figure C.Elc} .92
Figure C.'i1 — Node numbers superimposed upon the shell element weld mesh
section in Figure C.Elfi S2
Figure C.'i2 - Element numbers superimposed upon the shell element weld mesh
section in Figure C.9c} and di S3
Figure C.'i3 - Element numbers superimposed upon the shell element weld mesh
section in Figure C.Sfj SL5‘
Figure C.t-'-l — Calculation of the SSE stress at node n2 of the shell mesh shown in
Figures CE to Figure C.1.?. Sid
Figure C.tS — Stress distributions across sections of an I-beam with a cover
plate
E35
Figure C.tEi - Region of Til or HF integration for an edge attachment SE
Figure C.'i? - Stress distribution through l beam flange underneath node n2 for
solid mesh shown in Figure CA, Figure C.5 and Figure C.? St-IS
Figure C.‘lB - Distribution of correctly averaged stresses plotted against distance
y Si?’
Figure C.iSi — Distribution of correctly averaged nodal forces plotted against
distance y lttfl
Figure El-2D ~ Brick element mesh with definition of weld toe element size {fl
and plate thickness {ti ldi‘
Figure C.21 - Brick element model of an I beam showing the region of
connection between the connectivity representing the joining surface {shaded}
to a cover plate when the weid overfill is not modelled ltif
Figure C22 — Dimensions used for inclined element representation of a fillet
weld H12
Figure C.23 — Dimensions used for thicker element representation of a fillet
weld
l'l.'.i.i-‘
Figure C24 - Example of symmetrical welded joint for which hot-spot stress is
underestimated using methods in this annex IDS
Figure C.25 - Types of misalignment and distortion
Figure D.1 — Flaw dimensions lEl.fi
Figure D.2 — Transverse load—carrying cruciform joint
Figure D.3 — Crack opening modes
fdd
lid
iii
Figure F,1 ~— lvlulti-run weld in tubular nodal joint requiring improvement of
every weld toe f2d
Figure F.2 - Recommendations for weld toe grinding
I22
Figure F.3 - Toe grinding to improve fatigue strength I22
Figure F.4 — Effect of TIE or plasma torch position on resulting weld profile
li
I
ID The British Standards Institution 2014
1'23
BRITISH STAIlI DARD
BS TEDBIZDIII
Flgure F.S — Modification to design S-N cunre for untreated weld resulting from
weld toe dressing l.2rl
Figure F.Ei — ‘Weld toe peening methods
Figure F.Ir' — ‘Weld toe peening l'.2ti
f.?.'.i
Figure EB — Modification to design S-iii curve for untreated weld resulting from
weld toe peening 1Ei'lIl
Figure G-1 — Example of hot—spot stresses in a tubular node joint T31‘
Figure G.2 - Locations A and B of stresses used for linear extrapolation to weld
toes to determine hot-spot stresses in tubular joints is-ll
Figure H.1 - Example of cycle counting by reservoir method I25
List of tables
Table 1 — Classification of details: plain material free from welding
f-ll
Table 2 - Classification of details: bolted and riveted connections 15
Table 3 — Classification of details: butt welds and continuous welded
attachments essentially parallel to the direction of applied stress T5
Table 4 - Classification of details: welded attachments on the surface or edge of
a stressed member TB
Table S - Classification of details: full penetration butt welds between co-planar
plates 2!?
Table E — Classification of details: transverse butt welds in sections, tubes and
pipes .22
Table if — Classification of details: load carrying fillet and T-butt joints between
plates .25
Table B - Classification of details: slotted connections and penetrations through
stressed members 23
Table B — Classification of details: circular tubular members 25
Table ll] — Classification of details: branch connections in pressurized
containers Ell
Table 11 — Guidance on non—destructive testing of planar imperfections in
welds 39
Table 12 — Fatigue based acceptance levels for embedded non-planar
imperfections in butt welds -iii
Table 13 - Fatigue based acceptance levels for undercut in transversely stressed
welds ‘ii fl rl{.l
Table 14 - Effect of misalignment on the fatigue strength of transverse butt
welded joints -ill
Table 1S — Effect of misalignment on the fatigue strength of cruciform welded
joints -til
Table IE — Effect of misalignment on the fatigue strength of transverse butt or
cruciform welded joints being assessed in terms of hot-spot stress rll
Table 1? - Stresses used in fatigue assessments involving applied shear
stresses #2
Table 1B - Details of basic S-Al curves SH
Table lfl — ltlominal probability factors Elli
Table 2|] — Details of design S-ill curves for steel in sea water
El
Table C.l - The performance of structural stress calculation procedures SSE, Tl'l
and NF for assessing hot—spot type "a" weld toes or ends iii?‘
Table D.‘i s- Use of stress intensity corrections with nominal or hot-spot
Stf-ESS
H5
Table E.1 — Fatigue test factor, F
H5
Table F.1 — Summary of weld toe peening methods I23
Table F.2 -— improvement in fatigue strength due to weld toe peening
1.29
Summary of pages
This document comprises a front cover, an inside front cover, pages i to vi,
pages 1 to lilo, an inside back cover and a back cover.
Q The British Standards Institution 2B1-ll
e
ill
BS TEUBIZBIII
BRITISH STANDARD
Foreword
Publishing information
This British Standard is published by BSI Standards Limited, under licence from
The British Standards institution, and came into effect on 31 March 2i.'l1-ii. lt was
prepared by Technical Committee WEEBT. Acceptance levels for flaws in welds.
A list of organizations represented on this committee can be obtained on
request to its secretary.
Supersession
This British Standard supersedes BS Tr'BDB:1'.?rSl3, which is withdrawn.
Information about this document
Guidance on general fatigue design philosophy is given in Annex A, which also
contains a brief description of the method of using this British standard. A more
general method for assessing welded joints using the hot-spot stress, only
included previously for assessing tubular joints, is also introduced.
The relevant application standard or specification for the particular product
being assessed specifies the following:
al
the loading to be assumed for design purposes, including its magnitude and
frequency:
bi the required life of the structure:
cl
the environmental conditions;
dj
the required nominal probability of failure.
This is a full revision of the standard, and introduces the following principal
changes [1]:
I
Introduction of the hot-spot stress method with guidance on finite element
stress analysis {FEAl.
I
New correction for both plate thickness and applied bending with
allowance for welded joint proportions.
I
Additional weld details; some have been reclassified.
I
Weld Eluality requirements based on fitness for purpose.
I
Revised sea water corrosion fatigue data.
I
New rules for bolts.
I
Design data to resist shear fatigue failure.
I
Guidance on stress calculation for combined loading.
I
Revised cumulative damage rules.
I
Comprehensive guidance on use of weld toe improvement methods.
I
New guidance on acceptance fatigue testing and statistical analysis of
results.
European standards containing fatigue rules for steel structures and pressure
vessels have been published since the 1993 edition of this British Standard. it is
therefore not applicable to product areas covered by them. it is applicable to a
wide range of other steel product areas that do not have specific fatigue rules.
lv
I D The British Standards institution 2D14
BRITISH STANDARD
BS ?ElJB:2l]"lII
Use of this document
As a guide, this British Standard takes the form of guidance and
recommendations. It should not be quoted as if it were a specification or a code
of practice and claims of compliance cannot be made to it.
Presentational conventions
The guidance in this standard is presented in roman {i.e. upright] type. Any
recommendations are expressed in sentences in which the principal auxiliary
verb is "should".
Commentary. e.xplanatlon and general informative material is presented in
smaller italic type, and does not constitute a normative element.
Contractual and legal considerations
This publication does not purport to include all the necessary provisions of a
contract. Users are responsible for its correct application.
Compliance with a British Standard cannot confer immunity from legal
obfigafions.
El The British Standards Institution 211111-l
I
v
BS ?EUB:2lIl1II
vi
I
El The British Standards Institution 2Cl‘|4
BRITISH STANDARD
ibis page deliberately left blank
BRITISH STANDARD
BS TEDBIZDIII
1 Scope
1.1
General
This British Standard gives methods for assessing the fatigue life of parts of steel
products that are subject to repeated fluctuations of stress. it is applicable to all
areas of industrial application that are not covered by other British Standards
containing fatigue assessment rules.
lllDTl'E Some British Standards have specific product acceptance tests for fatigue lite,
but do not have assessment rules. in such cases the guidance in this British Standard
might be applicable for product development purposes.
1.:
Applications not covered
This British Standard is not applicable to the following application areas:
al
lighting columns lsee BS EH Alli:
bl
concrete building and civil engineering structures {see BS Elli 1592};
cl
steel building and civil engineering structures [see BS Elli lBB3 {all partslj;
dj
composite steel and concrete building and civil engineering structures [see
BS EH 1994 {all partslj;
ei unfired pressure vessels {see BS EN 13--145i: and
fl
1.3
fixed offshore structures isee BS EH ISD 1991112].
Illlaterials
This British Standard covers:
aj
wrought steel material products;
bl
welds in fully machined areas of steel casting;
cl
ferritic alloy and low alloy steels;
dl
austenitic and duplex stainless steels;
el
unprotected weathering steels; and
fl
threaded fasteners.
It is applicable to yield strengths in the range 2-Bl] llllmmi to BED lillmmf and
ultimate tensile strengths in the range are to 1 2-iltl llllmmi for material
thicknesses 3 mm and greater.
This British Standard is not applicable to the following:
1.-is
‘ll
proprietary fasteners;
2j
steel castings;
3}
cold drawn products;
ill
wire ropes: and
S}
steel for reinforcement in concrete.
Manufacturing processes
This British Standard is applicable to machined products with the following
exceptions:
al
rough sawn surfaces:
bl surfaces requiring high quality surface finish {e.g. lapping, polishing, honing.
fine grindingl: and
ID The British Standards Institution 2lJ14 I
1
BS TBUBIZBIII
BRITISH STANDARD
cl
machined details with sharp corners le.g. key ways. un-radiused shouldersi.
The following manufacturing processes are also covered:
1}
cold formed wrought products;
2}
weld toe improvement methods;
3}
arc welded joints. with the exclusion of joints between rectangular and
square hollow sections:
it}
in-line butt welds made by power beam and friction welding:
S_l
tensioned and un-tensioned bolted joints and hot-driven riveted lap joints
loaded in shear; and
Bl
thermal cutting.
This British Standard is not applicable to the following manufacturing processes:
il
resistance welding processes and brazing;
ii,l-
contact joints under pressure where fretting occurs;
iii] adhesively bonded joints;
iv} shearing and punching; and
v‘,l
surface hardening-
Environment
The fatigue design data in this British Standard are applicable to internal and
external air environments. They are applicable to structural steel products
exposed to sea water.
They are not applicable to unprotected stainless or weathering steel products in
sea water or aggressive corroding environments leg. chloride, sulphide, strong
acid or alkali}.
The data are applicable to products operating at temperatures below the creep
range of the steel.
This British Standard is not applicable to products operating in the creep regime.
Normative references
The following documents, in whole or in part, are normatively referenced in this
document and are indispensable for its application. For dated references, only
the edition cited applies. For undated references, the latest edition of the
referenced document {including any amendments] applies.
BS 35113-2, lSCl metric screw threads — Part .2: Specification for selected limits of
size
BS 3592. lSCl metric precision hexagon bolts. screws and nuts — Specification
BS iliiilli. rso metric blaclr hexagon bolts. screws and nuts - specification
BS 4395 {all parts}, Specification for high strength friction grip bolts and
associated nuts and washers for structural engineering
BS TlEl1[l, Guide to methods for assessing the acceptability of flaws in metallic
structures
BS Elli 1cn1.1, irlr'elo'.lng — Recommendations for welding of metallic materials —
Fart l‘: General guidance for arc welding
BS Etl 1ti‘i1-2, Welding - Recommendations for welding of metallic materials Fart .2: Arc welding of ferrltlc steels
2
I
ID The British Standards Institution 2B1-ll
BRITISH STANDARD
BS TEDBIZDIII
BS Elil ltl11~3. Welding - Recommendations for welding of metallic materials Part 3: Arc welding of stainless steels
BS Elli IDIE3 {all parts}, Delivery requirements for surface condition of hot-rolled
steel plates, wide flats and sections
BS Elsi IBB3-1-B, Eurocode 3 — Design of steel structures — Design ofjoint
BS Eill ISD 3505, Mechanical properties of corrosion-resistant stainless steel
fasteners - Bolts, screws
BS Elil ISD llIJ‘i4, Hexagon head bolts - Product grades A and B
BS Elli lSCl 1'-lCi‘l?, Hexagon head screws - Product grades A and B
BS Elsi ISD IITB2, Hexagon socket head cap screws
BS EH ISD SD13, Thermal cutting — Classification of thermal cuts — Geometrical
product specification and duality tolerance
BS ISD 12ltiB, llrletallic materials - Fatigue testing - Fatigue cracl: growth
method
Terms and definitions
For the purposes of this British Standard, the following terms and definitions
apply.
cycle counting method
method of counting the numbers of stress cycles of different magnitudes which
occur in a senrice stress history
lllDTE The loads applied to the structure, considered in sequence, generate a
particular stress history at ea ch detail of interest. Th ls stress history can be broken
down into equivalent stress ranges by the operation of cycle counting-
detail class
rating given to a particular structural detail to indicate which of the fatigue
strength {S-Nj curves should be used in the fatigue assessment
lllDTE l Also known as joint class.
lllClTE 2
The class ls denoted by one of the following letters: A, B, C, D, E, E F2, G,
G2, 5,, S, TJ, Wt or x. The categorisation taxes into account the stress being used in
the assessment l'e.g. nominal, hot~spot or shear stress}, the local stress concentration
at the detail, the siae and shape of the maximum acceptable discontinuity; the stress
direction. metallurgical effects. residual stresses and a post-weld improvement
method.
deflgnlde
period within which there is a defined nominal probability that failure by
fatigue cracking is unlikely to occur
NDTE This can be longer or shorter than the service life {see Annex .A,l.
design spectrum
tabulation of the number of occurrences of all the stress ranges. 5,; of different
magnitudes produced by the load spectrum in the design life of the structure or
component, to be used in the fatigue assessment
NDTE l‘
Also known as stress spectrum.
ll|'ClTE 2
Dil'ferent components of a product can have different design spectra.
ID The British Standards Institution 2lJ14 I
3
B5 TEUB 21314
BRITISH STANDARD
3.5
fatigue
damage ef a structural part by the initiatien and gradual prepagatien ef a cracls
er cracics caused by repeated appiicatiens ef stress
3.5
fatigue failure
achievement ef a threugh-sectien fatigue craclc er a sufficiently large fatigue
cracit te cause static failure er excessive determatien
NOTE
3.?
See Annex A.
fatigue life. N
number ef stress cycles that preduce a given prebabllity ef fatigue failure
3.5
fatigue leading
leading en a structure which is liable te cause fatigue cracking
NOTE it can he cempesed ef severai different types and magnitudes ef ieading
events isee Ciause .T".i.
3.9
fatigue strength
censtant amplitude stress range. 5,, causing failure in a specified number ef
cycles ihl)
3.1I'J
het-spet stress. SH
structural stress at a weld tee er weld end
3.11
initial nen-prepagating stress range. SM
censtant amplitude stress range belew which {in the absence ef any previeus
leading} a craclc is assumed net te prepagate
MUTE i
Aise irnewn as censtant ampiitucie fatigue iimit; CAFL.
HEITE .? its magnitude depends en the structurai detaii being assessed. Fer parts in
air er adecruateiy pretected against cerresiurr. it is assumed te he the stress range
cerrespending te a iife ef iii? cycies en the design 5-H curve fer aii detail‘ ciasses
except 5, and 5, . fer which it cerrespends te ii? cycies. Fer unpretecteci ieints in a
cerresive envirenment it sheuid be assumed that SM = ti fer aii ciasses.
3.11
lead spectrum
tabulatien shewing the relative number ef eccurrences ef all the leading events
ef different types and magnitudes expected te he experienced by the structure
in its design life
3.13
leading event
defined leading sequence en the structure
NOTE i This can be characterized try its reiative frequency ef eccurrence as weii as
its magnitude and geerrietricai arrangement.
MUTE 2 Fer errarnpie. in the case ef handiing equipment this ceuid invieive the
iifting, mevement and de_eesiting ef a iead. Fer design purpeses each ieading event
is assumed te repeat a given numher ef times in the design iife ef the structure.
3.11‘-1
i'iiiiner‘s summatien
linear cumulative damage summatien based en the rule devised by Falnrigren
and Miner
3.15
rieminal stress
structural stress that weuld EJti51. in the absence ef the structural discentinuity
being censidered
4
I
Iii The British Standards lnstitutien 2014
BRITISH 5TAH DARD
55 3505:2514
NOTE iiieminai stress is a reference stress that can he caicuiated using eiementar_v
theery ef structures. it exciudes the effects ef structurai discentinuities ie.g. vveids.
epenings, thicirness changes) and secendary bending due te a iecai vveid detaii.
3.15
-5-hf EUHFE
quantitative relatienship between the fatigue strength 5 and the number ef
cycles iv cerrespending te a specific prebability ef failure fer a detail, derived
frem test data
3.15.1
basic S-hi‘ curve
5-N curve. fer the reeuired prebability ef failure. fer a detail ef basic thicltness
isee 15.3.2) eperating in air witheut the appllcatien ef any fatigue strength
imprevement methed apart frem these given in Table 3 te Table ‘lb
3.15.2
design S-hi curve
S-iii curve adepted fer design purpeses fer the detail being assessed
NDTE it is derived frem the relevant basic S—l‘1i curve medified, if necessary, te aiievv
fer the influence ef materiai thicirriess, bending, envirenrnent, fatigue strength
imprevement techniques iadditienai te these given in Tabie 5' te Table iti,l_. stress
reiief isee 15.3.5.1 er vverirmanship isee Clause ‘Hi.
3.15.3
standard basic S-N curve
basic 5-N curve fer EiT.T% prebability ef suniival [twe standard deviatiens ef
leghi belew the mean {d = 2]} assuming that the test data are represented by a
nermal distributien ef leg life
3.1?
service life
peried in which a structure er cempenent is required te perferm safely with an
acceptable prebabiiity [see A1} that it is unlikely te require repair er withdrawal
frem senrice as a result ef fatigue craclcing
3.15
slepe transitien peint
peint en the 5-hl cunre beyend which it is eictrapelated at a shallevver slepe fer
use in cumulative damage calculatiens ef fatigue under variable amplitude
leading
3.15
stress cycle
pattern ef variatien ef stress at a peint defined by the cycle ceunting methed
and censisting ef a change in stress between defined minimum itreugh} and
maximum {peak} values and bacic again
NOTE t Aise irnesvn as cycie ef stressNDTE 2
peint.
3.1-5
fine leading event can preduce ene er mere stress cycles at any particular
stress range 5,
algebraic difference between the twe extremes ireversalsi ef a stress cycle
NDTE ‘l'
Aise irnewn as range ef stress.
NDTE 2
See 75.2 fer parent metai.
NOTE 3
See 15.3 fer vveid metal‘.
fii The British Standards lnstitutien 2514 I
5
55 ?5=U5:2[l14
BRITISH STANDARD
3.21
structural stress
surface vaiue ef the linearly distributed stress acress the sectien thickness arising
frem applied leads (ferces, mements, pressure, etc.) and the cerrespending
reactien ferces en the particular structural part
NOTE The linear stress distrihutien includes the effects ef gress structural
discentinuities {’e.g. presence ef an attachment, aperture, change ef cress-sectien,
misalignment; intersectien ef members] and clistertien-inducecl eencling mements.
l-levvevei; it excludes the netch effects ef lecal structural discentinuities i'e.g. vveld
tee, vveid endl which give rise te nen-linear stress distributiens acress the sectien
thickness.
4 Symbels and units
Unless etherwise indicated, fer the purpeses ef this British Standard, the
fellewing symbeis and units apply.
A
l'-let area ef cress sectien {in mm‘)
a
Weld threat thickness [in mm,'i
ti
Expenent in cerrectien term fer thickness and bending
in
Parameter defining the mean line 5,-lil relatienship
E_-,
Parameter defining the 5,-lll relatienship fer twe standard
deviatiens ef leg lil belew the mean line
Cd
Parameter defining the 5,-N relatienship fer d standard deviatiens
ef leg N frem the mean line
D
_
P
n
lv1iner's surnm atien EH
d
t-lumber ef standard deviatiens ef leg lil frem the mean 5,-N curve
E,
Elastic medulus at temperature T {in Himmfl
e
Axial misalignment leccentricity er centrealine mismatch} [in mm}
F
Fatigue test facter
g
Gap between welds in intermittently welded jeint lin mmi
h
Weld length fin mm)
it',,
Belt stiffness {in Hlmml
lily
Summatien ef clamped cempenents stiffness {in lllmm}
it,
Stress cencentratien facter under fatigue leading
km
Stress magnificatlen facter due te misalignment
km
Eerrectien facter fer plate thickness and bending
l.
Overall attachment length, including welds [see Figure ‘ii [in mm]
i
Attachment length parallel te directien ef leading censidered fin
mm}
lvl, lirl'
Applied bending mements {in hi-mm}
m
Inverse slepe ef leg 5,-leg lll curve ii.e. 5,“-lil = censtantl
lil,, lill, . . .
Number ef cycles te failure under censtant amplitude leading with
stress ranges 5,1. 5,, . . .. etc-. cerrespending te n,. ny etc. fin cycles}
5
I
him
Eenstant amplitude endurance cerrespending te Sm {in cycles}
NU,
Endurance at slepe transitien peint. 5,,,. en 5-hi curve fin cyclesl
n
llfumber ef fatigue test results
IE The British Standards lnstitutien 2514
BRITISH STAH DARD
BS 3555:2514
n,, n; . . .
l'-lumber ef cycles ef damaging stress ranges 5,1. 5,2 . . ., etc. in a
design spectrum
F, F", PH, PL
Applied axial ferces iin fill
P,
Applied ferce range fin hi)
Fl‘
Stress ratie {ratie ef minimum te maximum algebraic value ef
applied stress] er cherd radius in rnm {see Annex G]
r
Radius {in mm]
5
Fatigue strength ef the structural detail under censideratien,
including any required thickness cerrectien iin lllimmfl
5,,
Fatigue strength ebtalned frem the basic 5,-lil curve fin hiimmfl
5,,
Het-spet stress lin Nlmrnfl
.5“,
Het-spet stress range fin hllmml}
5“,
Censtant amplitude initial nen-prepagating stress range iin air 5,, =
5, at iii = ‘iii’ cycles fer all design classes except 5, and 51 when it
ceincides with if = 15“ cyclesl fin hllmmfl
NOTE Als-e lines-vn as censtant amplitude fatigue limit i'C.4Fl.,l.
5,“
5tress range at slepe transitien peint l,in Hlmmfl
5,
Stress range in any ene cycle fin lillmmfl
5,,. 5,; . . .
individual stress ranges l5,} in a design spectrum fin llllmmfl
SW
Resultant stress range en weld threat {in hilmmfl
5.5
Standard deviatien ef leg lll
irii
Plate width er thickness ef lengitudinal attachment [in mm)
T
Applied terque [in l'~l.rnml
t
Plate thickness ef the member under censideratien {in mm}
t,_.,
Thiclcness relevant te the basic 5,-lll curve fer the detail {in mm}
t,
Cever plate thickness lin mml
t,,,.,.
Effective plate thickness [in mm}
w
tlembined site ef effective weld threats fin mm
2'
Sectien medulus {in mmfl
rrl, .1},
Direct stresses acting in the x- and y-directiens fin hllmm-*-l
duh
Applied bending stress range fin lillmmfl
.»l.-rm
Applied membrane stress range iin l"Iln'|n'i1}
..-in,
Direct stress range en weld threat {in lllirnmfl
.:1-tr“,
Engineering shear stress range en weld threat fin hilmmfl
rr.,
l‘-leminal tensile yield strength (in lillmmf}
r
Shear stress lln hiimmfl
_-fr
Shear stress range [in l'~llmm*"l
Jr
Transverse shear stress range en weld threat {in fllmmfl
--lr,,
Lengitudinai shear stress range en weld threat {in hilmmfl
sll
Degree ef bending la.-iglieem + dab}
I5 The British Standards lnstitutien 2514 I
T
BS ?555:2514
BRITISH STANDARD
Figure l
Definitien ef length l. fer use in thickness-bending cerrectien
l.
-‘I
F'-
.
‘ll-I
5 Fatigue assessment precedure
A stressed element can centain a number ef petentiai fatigue crack initiatien
sites. All ef these sheuld be checked isee ‘l2.1l. The regiens subiected te the
highest stress fluctuatiens andier centaining the severest stress cencentratiens
sheuld nermaily be given the highest prierity. The design precedure in this
British Standard invelves calculatien ef the fatigue damage accumulatien during
the design life er cemparisen ef the maximum applied stress range with the
relevant censtant amplitude fatigue limit. it is primarily intended fer use in
“safe life" design {see A.3.1l. It might alse be suitable fer "damage—telerant"
design [see A.3.2l.
The steps that sheuld be feilewed are:
a}
establish the required design life ef the preduct isee Clause El;
bl
establish a censervative estimate ef the leading expected in the life ef the
preduct fsee Clause T];
cl
estimate the resulting stress histery at the detail under censideratien isee
Clause 15};
dl
reduce the stress histery te an equivalent number ef cycles n, ef different
stress ranges Srl using a cycle ceunting technique fsee 15.9};
el
classify the detail in accerdance with Table 1 te Table ‘I5;
fl
use this classificatien te define the basic design 5,-lil curve {see 15.1};
gl
calculate the resulting fatigue life en the basis ef cemparisen ef the
maximum applied stress range and the relevant EAFL {see 15.5} er a
cumulative damage calculatien [see 15.7];
hl
if all applied stress ranges are belew the relevant CAFL [see 15.5}, the detail
can be assumed te have a life exceeding the specified design life in a} {see
Clause El;
ll
if seme ef the applied stress ranges exceed the relevant EAFL and that the
life calculated using the fuil design stress spectrum and the cumulative
damage methed isee 1s.ri exceeds the specified design life in al. the
requirement fer safe life design is met; and
ll
if the precedure in il results in a calculated life less than the specified design
life in a}, the requirement fer safe life design is net met. The fellewing
measures sheuld be taken:
B
I
ll
adjust the detail design se that a higher 5,-hi curve can be used;
ll
if 1} is net adequate, increase the cress-sectien at the petentiai
fatigue crack initiatien site te reduce the stress ranges.
I5 The British Standards lnstitutien 2514
BRITISH STANDARD
BS 3555:2514
3}
if measures ll and iii result in severe ecenemic censequences, use ef
a damage-telerant appreach invelving periedic in-service
nen-destructive testing [MDT] fer the detectien ef fatigue cracking
may be empleyed te ensure that the everall prebability ef failure
witheut warning during the design life is ne less than that assumed
fer safe life design. Fer further guidance see A.3.2.
5
Design life
The design life ef a preduct is usually pre-determined by facters such as market
expectatiens, planned service life er ebselescence, centract er warranty
requirements and uncentrellable deterieratien by mechanisms ether than
fatigue, such as wear and tear er cerresien. Where a design life has net already
been specified fer a preduct in which fatigue is a petentiai failure mede, it
sheuld be selected eri the basis ef the peried ef service ever which the
prebability ef failure by fatigue is required te be lew. This is achieved in design
by the use ef lewer-beund fatigue strength data lsee Clause flil and
upper-beund leading data [see Clause 3].
3
Fatigue leading
when assessing fatigue perlermance a realistic estimate ef the fatigue leading is
crucial te the calculatien ef life, and all types ef cyclic leading sheuld be taken
inte acceunt. Cyclic leading frem different seurces might be significant at
different phases ef the life ef a structure, e.g. manufacture, transpert, sterage,
installatien service, and can invelve different leading medes and frequencies.
Llncertainties exist in assessing beth the stresses resulting frem applied leads and
the respense ef a particular jeint, which centrel fatigue perlermance. The basis
ef the fatigue analysis sheuld be the use ef an upper beund estimate ef these
stresses, recerding the uncertainties invelved, cembined with 5-N curves derived
frem experimental data. In this way, uncertainties asseciated with the life ef a
particular jeint. e.g. site. weld detail. iecai envirenment. can be separated frem
these asseciated with applied stress.
Llncertaintles can alse exist in the number ef applicatiens ef the lead expected
te eccur during the preduct’s design life. The design lead spectrum sheuld be
selected en the basis that it is an upper beund estimate ef the accumulated
service cenditiens, including beth leading and number ef cycles, ever the full
design life ef the preduct. The adeptien ef mean plus twe standard deviatiens
data fer applied leads levels er an upper beund estimate based en knewledge
ef the actual er predicted leading envlrenment and applied numbers ef cycles,
when used with the design {i.e. mean minus twe standard deviatiens ef leg llll
5,-hi data in Clause 15, usually results in an acceptably lew prebability ef failure
during the design life, cemmensurate with safe-life design principles.
Because ef the sensitivity ef calculated life te the accuracy ef estimates ef stress,
stress ranges sheuld net be underestimated. Acceunt sheuld be taken ef all
likely eperatienai and envirenmental leads arising frem the fereseeable usage
ef the preduct during that peried. Use ef this appreach cempensates fer the use
ef design 5-lil cunres that cerrespend te a finite prebability bf failure and means
that lead facters are net required.
The fellewing are seme lmpertant seurces ef cyclic leading that sheuld be taken
inte acceunt. any er all ef which can be relevant in particular applicatiens:
al
fluctuating leads:
bl acceleratien ferces in meving structures;
cl
pressure changes;
ifii The British Standards lnstitutlen 2514 I
5
BS 3555:2514
BRITISH STANDARD
dl temperature fluctuatiens;
el
mechanical vlbratiens: and
f}
envirenmental leading lwind, currents and waves, especially when vertex
shedding is induced, e.g. en slender members}.
it is particularly impectant te assess dynamic magnificatlen effects where leading
frequencies are clese te ene ef the natural frequencies ef the cempenent er
structure. In seme instances the leading te be assumed fer fatigue design
purpeses is specified in the design specificatien. Where such infermatien is net
available, assumptiens as te the leading te be expected in seniice sheuld be
made, and it might be useful te ebtain data frem existing preducts subjected te
similar effects. in particular, in assessing an existing preduct, it might be pessible
te cempile a design spectrum frem strain readings er leading recerds ebtalned
frem centinueus menitering.
in all cases. the ebjective is te define the spectrum in terms ef the numbers ef
cycles ef each ef the individual stress ranges expected in the life ef the preduct.
if it is required te cenvert the spectrum inte a series ef censtant amplitude
blecks, the resulting simplified spectrum sheuld be equivalent in terms ef
fatigue damage te the actual spectrum. Te achieve this, a sufficient number ef
intenrals ef stress sheuld be selected te aveid discretien errers due te insufficient
reseiutien in the stress spectrum.
Envirenmental censideratiens
The fatigue assessment sheuld take inte acceunt the envirenmental cenditiens
that the preduct is expesed te during all phases ef its anticipated service life.
Fer example, preducts designed te eperate in seawater might be censtructed
and transperted in an air envirenment, installed and cemmissiened in a freely
cerreding marine envirenment, and eperated in seawater with cathedlc
pretectien. This British Standard prevides 5-til cunres fer three envirenmental
cenditiens and the rriest apprepriate curve sheuld be used fer each segment ef
the anticipated service life. Ferieds ef free cerresien sheuld be aveided, fer
example during temperary sterage ef preducts awaiting lnstallatien in seawater
er fellewing cemmissiening hydre-tests ef pressurized cempenents, as this can
cause pitting and an asseciated reductien in the expected fatigue life.
it might be necessary te take the eperating temperature inte acceunt lsee
15.3.3}. The fatigue strength ef steel preducts in air varies with temperature, in
accerdance with the cerrespending change in elastic medulus- Therefere it is
impreved at subaere temperatures. Hewever, in the case ef ferritic steels. as the
fracture teughness is reduced, it is pessible that everall fatigue life weuld be
reduced if the preduct suffers premature failure by brittle fracture frem a small
fatigue crack. Censequentiy, ne ailewance sheuld be made fer the beneficial
effect ef sub-aere temperature en fatigue strength but. where such eperatien is
expected, fer censistency with the fatigue design data previded in this British
Standard. the steel used sheuld be capable ef telerating the presence ef
threugh-sectien fatigue cracking at the minimum anticipated temperature. This
is achieved with steels that have minimum lew temperature impact preperties
specified in the material standard and are eperated at er abeve the stated
temperature.
There are ne particular recemmendatiens fer austenitic stainless steels eperating
at lew temperature, as they are net susceptible te brittle fracture. Fer elevated
temperatures, ne effect is assumed fer temperatures up te 155 “C, but at
temperatures higher than this fatigue strength decreases in line with the
decrease in elastic medulus fer temperatures belew the creep range ef the steel
cencerned. He guidance is previded fer cembined creep and fatigue.
T5
I E The British Standards institutien 2514
BRITISH STANDARD
BS 3555:2514
Facters en fatigue life
The reliability ef a preduct's fatigue life is dependent en the fellewing facters:
al
selectien ef a safe level ef fatigue leading isee Clause 3};
bl
cerrect calculatien ef stress ranges isee Clause 15};
cl
cerrect detail classificatien isee Clause 12l;
dl applicatien ef apprepriate centrels during manufacture lsee Clause 14l;
el
in seme cases, fatigue testing might be apprepriate {see Annex El.
The fatigue design data previded in this British Standard are in the ferm ef
mean S-lil cunres and the cerrespending standard deviatiens ef leg hi te enable
different prebabilities ef survival te be adepted. The standard basic 5-lll curves
isee 15.2l represent B‘.-'-T% prebability ef survival, as they are based en the mean
minus at least twe standard deviatiens ef legrv curves fer relevant experimental
data. Their use therefere indicates a finite prebability ef failure lup te 2.3%} fer
the calculated life. in seme circumstances, fer example in a failure investigatien
ef a part that has experienced fatigue cracking in service, it might be mere
apprepriate te assume a higher prebability ef failure and make use ef an S—l'il
curve less than twe standard deviatiens ef leg til belew the mean. Cenversely, it
might be apprepriate te use 5-lil curves based en the mean minus mere than
twe standard deviatiens ef leg N fer cempenents with inadequate structural
redundancy er difficult access fer inspectien. In selecting the number ef
standard deviatiens te be used te define the design 5-r-i curve, acceunt sheuld
be taken ef the accessibility ef the _ieint and the prepesed degree ef
manufacture and in-service inspectiens, as well as the censequences ef failure.
As a crack grews in ene part ef a structure, the lead might be shed te ether
members and lead te further fatigue cracks in these members.
Features influencing fatigue behavieur
Fer beth welded and belted steel preducts the fatigue life is nermaily geverned
by the fatigue behavieur ef the jeints, including beth main and secendary jeints
[2, 3]. Even fabricatien er handling aids, such as welded brackets er lifting lugs,
that remain in the cempleted preduct ceuid previde sites fer fatigue cracking
and sheuld therefere be assessed. Clptimum fatigue behavieur is ebtalned when
the preduct is detailed and censtructed such that stress cencentratiens are kept
te a minimum and. where pessible. the elements are able te deferm in their
intended ways witheut intreducing secendary defermatiens and stresses due te
iecai restraints. stresses can alse be reduced by increasing the thickned ef
parent metal er the weld threat, depending en the petentiai failure mede. in
the case ef the fermer, ailewance might need te be made fer the fact that
fatigue strength tends te decrease with increasing plate thickness fer seme types
ef jelnt when assessing the resulting benefit {see 15.3.2}.
lllptimurn jeint perfermance is achieved by aveiding jeint eccentricity and
misalignment. welds at free edges. and by ether centrels ever the quality ef the
jeints.
D The British Standards lnstitutien 2514
I
11
BS 3555:2514
BRITISH STANDARD
in the specific case ef pipe-te-pipe jeints made frem seamless pipe. the
accumulatien ef maximum allevvable manufacturing telerances fer thickness.
evality and diameter leads te a large petentiai fer girth weld misalignment and
asseciated penalty en fatigue life. Apart frem the resulting intreductien ef
secendary bending stresses, such misalignment can alse intensify the stress
cencentratien due tn the geemetry ef the weld reet bead in welds made frem
ene side. it is therefere advantageeus te recerd the actual range ef telerances
achieved fer each batch ef pipe supplied and use this infermatien te calculate
the maximum petentiai misalignment. Perfermance is alse adversely affected by
cencentratiens ef stress at heles. epenings and re-entrant cerners. Guidance in
these aspects is given in Table l te Table 15, Annex B, Annex C and Annex G.
The magnitude and nature ef stresses that cause prepagatien ef a crack and
therefere reduce the number ef stress repetitiens te cause failure are affected
by the presence ef residual stress. inherent flaws in welds and adjacent parent
metal, surface flaws and any ether stress raisers interfering with the flew ef
stress. These are taken inte acceunt in the classificatiens given in Table 1 te
Table ‘i5.
Fracture mechanics
in seme situatiens the nermai fatigue assessment precedures might be
inapprepriate. but fracture mechanics metheds might be helpful. Guidance en
the use ef fracture mechanics is given in Annex 5.
Classificatien ef details
12.1
General
Fer the purpeses ef fatigue design, jeints are divided inte several classes, each
with a cerrespending design 5-iii curve isee 15.2}. This classificatien depends
upen the fellewing:
a} the type ef stress being used te assess the detail lneminai er het-spet
stress};
bl
the geemetricai arrangement and prepertiens ef the detail;
c}
the directien ef the fluctuating stress relative te the detail,"
NCJTE Any reference te transverse er lengitudinal welds. welded jeints er
welded attachments refem te their erientatien with respect te the directien ef
stressing.
d}
the lecatien ef pessible fatigue crack initiatien at the detail; and
e}
the metheds ef manufacture and inspectien.
in any preduct er cempenent that is liable te be subjected te repeated
applicatiens ef stress. every welded jeint, including nen-structural attachments.
lifting lugs and temperary fabricatien aids left in place. er ether ferrns ef stress
cencentratien. such as a belt hele er cut edge, is petentially a sei.irce ef fatigue
cracking. Each part ef every censtructienai detail sheuld be assessed individually
and sheuld, where pessible, be placed in its relevant jeint class in accerdance
with Table i te Table 15. Where this is net pessible, the detail sheuld be
classified in accerdance with Clause 13.
in belted er riveted jeints, fatigue cracks nermaily initiate frem the belt er rivet
hele er, in the case ef frictlen-grip belted jeints. in the plate by fretting er in
the belt itself. in plate with cut edges er welded details fatigue cracks can
petentially initiate in the fellewing places:
1}
‘i2
frem any peint en the plate edge er plate surface;
I E The British Standards institutien 2514
BRITISH STANDARD
BS 3555:2514
2}
in the parent metal ef either part jeined adjacent te;
ll
the end ef the weld:
ii}
a weld tee;
iii} a change ef directien ef the weld;
3}
in the weld metal starting frem:
i}
the weld reet;
ii} the weld surface;
iii} an internal flaw.
in the case ef members er elements cennected at their ends by fillet welds er
partial penetratien butt welds. crack initiatien can eccur in the parent metal er
in the weld threat. Beth pessibiiities sheuld be assessed by talting inte acceunt
the apprepriate classificatien and stress range- similarly. fatigue crack initiatien
can eccur frem a weld tee en either the eutside er inside ef a full-penetratien
girth weld between pipes er tubes. The mest critical lecatien depends en the
relevant classificatien and stress range, beth ef which can differ between the
eutside and inside. Fer ether details, the classificatiens given in Table 1 te Table
15 cever crack initiatien at the lecatien indicated. hletes en the petentiai medes
ef failure fer each detail are given in Annex B.
12.2
Classificatien ef details
Table 1 te Table 15 cerrespend te the fellewing basic types ef details:
I
plain material {Table 1};
I
belted er riveted cennectiens {Table 2};
I
centinueus lengitudinal butt welds and welded attachments l_Table 3};
I
ether welded attachments lTabie rt};
I
transverse butt welds in plates lTable 5};
I
transverse butt welds in sectiens. tubes and pipes {Table 5};
I
lead~carr"ying fillet and T-butt jeints {Table 3};
I
sletted cennectiens and penetratiens threugh stressed members [Table B};
I
details relating te tubular members l.Table Bl: and
I
branch cennectiens te vessels {Table 15}.
Detail classificatiens are given fer assessments based en applied neminal stresses
and, where apprepriate, het~spet stresses. Where relevant, the need te apply a
cerrectien fer thickness and bending [see 15.3.2} and the applicability ef a weld
tee imprevement technique {see 15.3.5} are included. Each classified detail is
illustrated and given a type number. Table 1 te Table 15 alse give asseciated
criteria and diagrams that illustrate the geemetricai features and petentiai crack
lecatiens fer the directien ef leading shewn which determine the class ef each
detail. They sheuld be used te assist with initial selectien ef the apprepriate
type number.
A detail sheuld enly be designated a particular classificatien if it cenferms te all
criteria in Table 1 te Table 15 apprepriate te its type number er if a suitable
classificatien can be justified en the basis ef relevant published fatigue data .er
the results ef specific fatigue tests in accerdance with Annex E- Class A is
generally inapprepriate fer structural werk and therefere ne design data are
previded. The practical difficulty ef achieving the special inspectien standards
relevant te classes B and C might limit the feasibility ef adepting these
classificatiens in structural werk.
D The British Standards lnstitutien 2514
I
13
5S
3
BRm 5 H
5nu 5 E 5 1 I
I
5
H H" nu A In nu
_3mfl___"_
_“_ _m“h_____
m_____m__
"_ "_ _ _
___5-5_53__
iiE=!_ _i5thi
_T_ s
m_fink;
“H__ _____-_ _L_._-“H
_" _ m:_E_
_ ___F__
“_| _ _ " u_ _ __“__"_E35__"“__F_"_ __i "=_n_ _ "_
: _ _“li_ _ u_ E
_ _ _“ _ _ _“ _ _
___P__
______
.Em;
_ _H_ ._“ _ _:_E_IF
“E___m_rm_ m_ _ __Hn_u_“EH
__n_=I____ F“
_h____u____
_ "_ u_ “_h _ H_ "_ H‘_ _ _ _
___:__E
__n_______HOL
_ _ _" _ _' _ _ _ _ _" H_
BEE
B“
Hl___-ET
_H“_-_
_ “_ __u__
u_"“_____
_t n_ _m
F_“__
m_____ _ _ m_ m:_ _“_ _ n_ H_ _
____fl2;:_ n_ _ _ _ ___|_
EH__t _h_u¥_E_E_E
E
_ “_H__v_d _:u_"__
IE
___n_
H_ _" _u“_ P_ _ w_u _
fl_ m_ _ _ _ _ E_ _ __ u_E_i!_ _;
nfi_h_m___"_ _| _:_-HE__ ="_
fi_"___fl_ _H$_ _"_"_' _
finnH___=_
EEnE
E_m___
m__ _m_ _ _ i_:t_H
nE:Eli
_ _n "_ _ _! _
FEE
r_ "_=_ h_=5_ B_HE:
EHEU
EH_____"l!‘ _ _ m_“ !_“_
E
:_____“E
n_ H_ I|_ _ _ _ :I_“ E
I__“_
H_ _ :_ _ _ _ _|h_ _ _“_
E_if_E___“
I_ _ _ u _ _E"_“_ |-_"=“_H_h ‘_ "E__"_u_ _“_n“_ .
H_ fl_ _h _ “__ _
_=_ _ _ _ _ _ _ V_ _ _ _
1‘-um
_" _ u_3ui_ _ h_
h_ _H_ fl_ _
G
_ u_ E_n_
_I__“____
___m
_ _ _ _ b_ :_
IE
EE_H__=_E_n_
-iii
-55______I
fii
FEE
E;
_ _U_d_“_Ufl
ti
_ _ _=_ E_ _ _ m
_ _“u_B_ _u EE5_ _"n_E_£__r__*_Hm_"H___n___2__ _ __ “_E"_m"__ _| _ "_
_:LE_
____
____ __" _u_h__l _H_fl__ _£n_ Eu|-_ 5
5_nu_:"-n_ _E_ _ _Hg“
_mt__
B
=u:H_m__fl_~EH__“_ u_H_gH_u_ _ _ _ _ _ H_ _ _H_ u_fl_____
u_nm___
E_E_l! _ i_ _u_-+_ _E- _ -i
_“i=___n__ _ :_" _ _: _ "_H_
_______
_“ _u _ __“_-H_ _ __ *_ '_‘ _ _ _ _
IE
-_ _"'_ _h_-_fl_ _ -_E_ _-£____
_ _ _ _ _ _fl_ _“ ___ =_“__:
n_ _
:_ -_ ____ _=E_‘EH
____
___"!_ -_ -_fl__!_ “_ ._:_
_" u_ ____"‘ _“ _ “_ _u 5_u 5_ _ “
_m_ E"_ _ _ “_L _" _:
_“un___m_“ _ _ _E _ _ _ 5E
E______“
_ _E_U_ _: _ m_ _ "_ _"_
E
n______ _E" _ "_ _“ H_
Eu
:3‘
3______
_-___i__‘
_ m_ _ _E:§_“_
_E
i _Eu_ _“"_ _ _ ‘_ _ H_
_"__i“En
E
p_ _ _-fl_"__ _
_£_“_E
u_ _="_ __|_
u_"__£_____h___
u_ __-_F"_
_U
3" _E
H“____
___“-_""“__!I__"_-____h _-___n_"_|_="‘__H“__::__
_ _ _ _ "_ _ _n_:_ _ _ _ _ _:____
_J lE_ _u _n"_ “"_n _:_
d:_ _ =_ _ _ _E
_“_ _ __”_H“ h_
gig
_Eam_HL_ _ _ _E! _ 5
_E _r:E_ _ _ l_E_E!_
-"_ _ In
_m__ D__ _u_ _|_ hH_
E
n_ "_ _ _ _ _ H_"_
__h__
_____n_
_ _I “_: _ _“ "_
“D
En_H_______U_
‘E_ E___w_
1:"_ _ "_
_E_u_~_a_“_: _m_:
______
_ _u _= _ E§_ ___"__ _ _ _ _ n_ _H h__“E
MP?E____m_E
_CE_ _ _ _ _
_“B_uaumu
_ _ fl_ m_ "_ _ _ _ _Bi
_I _ n___" _ "__m _ E"_h______
E_h
_ ____"_
MQ
_______
_____
h_"_ _“_E
"El:n_ _ _ _H_E___“_m_: _ _ _ _ _"_
_=_“_" _u_ _ gfi_ ___u_
:_ E_E_ _h _
_“ _ _ _ H_ _
__DE:Eu
_ _ I_ _ _ ‘|_F :_r_ m“_ ___"_ __ _ :_ “_
“_ _ "_ _ _ _n_ _ “_ _ un_n
HIE__ E“_ _‘:_ _
fi__E
_____t
: ______
ER
_'fi_! i_U _fl _: _ _fl__!_:_ H_ WED
:5‘:
in__E_3_ _i_ _
EE1e___
_if_5_1 _ _
E__ _ _ 2_ _ _ _ _H
_ _H_ E
_I_m___
_-_*=i_E_ __‘_ _ _ _n_E“mun
_____
Eh
_H__-i
_ “_ H_ _ _“ _ *
Em__
_ _ _“ _ H:_ n _ _ H_
_ _u 5_ _
_ _" _ _ _ _ ________
_ “E_ _ _ _ _
Efl_u_p_ _ _
___
_:“_F
______
_ __E__‘
_ _“_‘ __"_
_"_n___"_m_“1_“ _ _
H___
1
4
i
my
MM
% H“ ______W d B
____"_
______
Im
Umt M _m H _ _ _ _G M
5 _m
_m____
5R
m DA
_hE H _ _ _ _ T
L
RD
rm
i
5S
l
fig
__\____.I____“__\
_H_H____:__:___ EM
/
|_ _ _ _
35
5 3 _______5 1 4
M
A
,
_
1
__
E___\
____“_\.__"‘_____“___-fl
r_r_r______fi___
T“H
__ “H_______r___
____ _ _ _.-_j_-_
qI
III
j__ _
lid
I
___U1
l
"_ "_m_"_ ____"__ _- _ _ _ _ _fl _ _n_Eu ___"
_| _ _ ___"_E_ _ _ _* _ _ _ _ "_ E_ fi_
"uH5
E_ _“_m___=_“_H _"_£ L_E _| _ =_ n__q
_|__fl___
_ _. u_ _ _ “_"_ E.m_______“
__"_‘ _=_m_____I_ _ __H___E__
n_ w
_E
" E:
_- |E=_ -_“_J"“_‘ _fl _ _H _h _h _ __u__ _ h_“Eu_ “_ "_ _ n_“
m_E_E__E_ _
E_m!___:___m _ _:1= __“______iE35
_"___a_|_fi _H =._ _! _ :_ _E
:u____|___“=_fin__ _____H#5
___H
_ _n _ _
Eran“
S
:_ "_________
_ ":_ _ _: _ _" u_
E:HE______u_ _ Q:in
_:__l
_!._"__ _=__‘_ ____:_n_ _ _
_ _ u__fl__
=_H__ i"_|:_nH_m________
_ :E
_ __"_p___"
_ _“ _E_‘i_ _ _ _ _E .__E_ _ _ H_ _ _
E_u"_5 U_ :_ _HE:
___H__
__‘u ________
_ =_ _"_“u_ -_
E
_ __'_-_ __"_____F_____
_E_fl =_ _ _ u_ _ H_ _E
E_ _ n_ _
Eu:
2___5_3_5
_ :5d_ _
E_ 5_- _ _E_ _
_H_ d_ _ _
_ _ :_ "_ "_ _
"_ _ n_ _" _ _" _“ -_
E:
_“_ _ _;"_j
m_ "_ _ _ “_n _ _ _ _ _ _ _ E_
m:_h_ _
__“_____
___
In
__m“*5:
_i_- _U _ _:_E __’_ ____ _: =__
_ U _H
H__
H
;___
U
m_
H
1
_
m_H__'
:____m_flflgfi
E:
__ _ _ _ _ _ _ F
__“_
E
Lu_ _ _ "_ _“_
_I _m
“'"_!__"_
H
J__“rH__H_
__u_": |__=
"__E__“______|
u_n_ Iii"____u
:5
___u
HM:
_ _ D_“_ |_ _ _:
__q“__
____H__U_
_" _ -_"_HE__‘_“_" |n_-_mE_‘ n_
_E__n__
__““_u_" _._uHu____"_q _:_fl"_-_|“_ _
E
=____#
*_ _ _ _
fig
_ "_Eu_ H_ “_“FF§u__nflfl
_E“_
_HEEHUI
____
u_ED
‘_|_ _ _ _ _"______
H'_iE!_:F
_H
Lb
" _mm__u_£_
“ _ u£_m“_u_H E_=__"_D“:n__CH_=_:
_H__m:
u' un_
_E___ _ l_ _ _
__m=_E
_"n_ _FEm_“_-l ___"_n_ _
_“ _ ‘_ "_ I_ ___
_ _ _fl_ m“_ H_ _H_ _ _u
_____!__I_____m_______~__
__n__ _ n_ _ _l_H_
H_“ _In_ H_ _ “_Hu_“ _DH_ _ “_
:_=
_ _ _ _ _ "_ _ m_. E_ _
h_ _ _q_ _: F_1_ _ :_:__"
_=5
¥__“"__|_ ____n_
_H______ _ _ _ _EF_ L_.“ _E- _H|_ _ _m =
d__=_fi__
____m_
E_"____m:_u_Hh_ :
_F_ _ _ IE_ -_ E
_ _"_ _ "___j____1m_J ___
E_=_
_—_
‘______
_ :_ _:_"_
___q__
E
E_ _ _ _h__n_ "_H_ _E6%“E
n __§___
__" E_" _ _ _ "_._
t_" _____
_" _H" _ _ _ _
_ =_ _ _ "_ul _ "_ _ ‘
_ _“_ _ _m_ _ _um_
_u_E_ _n_ _______
_n_
___U_"
m___"__F___
5m__ _ _ E_ _ _ 5
_H___:__u____E
m____fl
_H_
_=_:_H____5=
_H____H_:____fi_E
!__H_
___H
=_:___:
_"_ _ _ _ _"_
_ _" _ "_m_ |_l "_ _ _
"_ "_ _ _"5_:“=_ H_H_
="_ _" =_ __ _"___ _
E5
___“__ _|=_ __ _‘ _ _
_ _n _ _ _ _ _t_ _ r
_"5_n5_ _ _ _5_ “_ _“
Em
_ _ _E_ _E _fl_ _E
_" ___= _ m_" u_L_ m_ _
_-““Ea
_Eh_F_ _ _ _ _“
-_J_____"
__m-_ _ _“_i________
___ _U _ _ __ _
_ _ _D_“_ _ _ _q _H_fl _ "_
E_H:
_i _H_m__
_m_‘n__ “_F_ _ _
Fl
:______n
_ _ _r_ u_m_ n_
U_n_ _ _H_ _ _ _ _ _ _H_ _
I_D____
_ ____
_ _ _m_ _ _“ __l __"__
=_ _ _ _ _ _"_
EE_E__"
_ _ _ |_ _ _ _ ___
_H_ _ _Eh_ _“"_ _
D____-__fl_:_____=_ “_ _ _ _ F:
“_"_n_J_u_" _ _“_
H_ _ "_ "_
_n_ _ “_ _
-.___-______
_ _ _ H_ _"_
llll
IE3:35
_____i3£
_ _ _ _ _u_“ _ _ n_
_ _ " _ _ ' _“ _ _
H__“____HEE2‘
E______ _ n_ _ _ _ _ _ _ u_____n_
£_2:ED:
w_ _“_ _F_-_-E:
H_ _“ _ _
______ _ _ _ _ _“ _
Bibs“
__"I_
_
__n__
_uH_____"_ _ _ _“__ ___E:_ _ _“_
_ _ &_: _ _ _ _ _ _ _ h_
_ _ _ “ _E _
1%
€H
_____H m___“
____H
“H
_m__H
_m____
h_____“
a
_ _" InH
T h E R_ r H E h _ _ _H
Im
___
HH H
M
M
__
H 2 __U1 4
I
__F 5_
B 5 T ED B 2 fl_ 1 l
.__"_T _ _ -_ _ -_ _ '_
fl¥¥
I
Ihm
%_rIt“%m__fl_ _ _ _ “_
"Wu"
__h_l_
kw
BHm 5 H
._ J_._-_ _ _ _ ~._*_ I
I!
_H___-_ _
§ _ _ £_ _¢
___
_.__._.
__
-
HE
__
1WJ “ac”
Q
fi__b
__§__“
Em
:m
_HE5
3
mH_“=:2“
_ _nflHum“_:uflE_U=_?m“E=_%E
H_E:u“m______
“Em_ _ _ ufl_ n_ ii:
_ Hi
EE_E
_ H_ _ _ _ _m_ _" _" _uH_ "_ “E.
_ _u___H|
_ uh:_
_ _ u_ "_u ‘HH“_ _‘_ fl_ _ ~_H- ‘:_ _ _ _ _H: _
__n_
l___“_“ _r _ u_|__m
"__"_
“_ P__n_____" n__r_"__ “:_"_________n_
_____“
E_u
_LI_!_ I ___n_
__fl____=
______
______fl_| _U ____
m_______"_
_ __ _"“_I _|__=I _- h_E"“__:n_____
_ u__fl_E‘_n- ____ _r “"_ -q_ “_ _u "_ _
E_
U
I
_!_ “__q____ _ $“_- U_q_n _ _ _ :_
E
__H
H
E____
"_ _ _ _ E_
E
“_ E_ __Umu_n_"_u_UE_ _ ____"uJ_--5_____F
___“__‘"______
_____=_h-_ _
EBEE
___-H___ _ _“ _ _ _ _" _ _ u
_h:E__"“___:
“_E H+_ _" _
_"_E____
H“____"_ _:"' n__ -_F _ “_ _
"____"_
uh
_HF:
'm_ _"____:
“__= _ _"_ _ "'_“__E___"_"_
__
E
Eu:
_H_
_?___fi__m_fl
_____fl_H=_____
____H
_“ _ ___n_
_" _ u_ _:"H_-_“ _"_. _“"
_:_D_
_:___n_____
_ _"_H_m__'
_ _ -_ n_
|“_ £H_ i _ _F _" u_E
H__:
-_-FBI
E515
'_ _ _ _ _ _ _I
n_ fi_ _ _
_ i_ fi_ _= _ “_
u_ "1:I_E_ _
-_ "_ -_ _- _ "_ "_-
=|H_u__
:____
“=-
E
n_H_i _ _ _"_ "_-
"_ fl_ _ |_ _ n_ n
_ _i _ _ fi_ _ _
__FE
u_ P_ _
_____
E:
n_ _ "_ !_ _ n_ “_*___H
____
_____
H_! _fi “-_ E_- E_ _
_ImH-U
IIIu__-_- :I_ P-_FE
fl_I _ _ _! _ =_ _ _E_
___mu
_ _ _ _ _ _ _ _ _ _ r_i
U
=_ "_ _ “_fl _ _- _ h_| IH_"
it__H_" EH_ "_ "_
___"_P_" __ h_u _
_ _d_ _ _ _ _
E
_ H_ u_ _ _ _ _ + In_BEE
3"!"
EHi
E=_"_ _E "_ _
=___ n_ "_ _ “_ _ _
“_ _ _ _ _E _
_ n_ _ _ H_ _
"_ “_ _ _ J_" _ _ _ :_U_
______fi___
_ __:_
_ _ _ _ _ _"_
_ "_ _“_ _$m_
_ _ _ _ _ ! _E “_ _
_‘_:_H_______ _
_“ _E_
_____
F_ _ _ _ £_~_ _
_"_-_ _q_ _ _
“_ _____
Ju_"_ _ ___“_F__fl__:H_ ___
_" E_I“__"_
_u “_ _“__m__n__
"_ _ "_
itu_____
5_" _ _ _ _ "_
_E___m_
r_ _ _ __ _
_HE_ _ _ _J_ “_ _ "_ H_ _ 5=3
fl_“ _ ._i _iE_ "‘_? _ n_
F_
-_l _ = _ _ “ |_ " _ :_ _ “_ H_ =_
“L-_
_ _U__“ _ _:_H “__H._
"M
E
fl_"_ _fi __ P_ _ _ _ "_
_:_ _“_"i_ _:
Fiiéi E-E1?
_" _ "_ F_ _|_" _1_E
_El
__"u_
_ _ ____“_
_ _ "_ “_- _ =_ _ _ "_
H_h_ _ _ _i _ _ _ _
_ _ _ :_ _ _ "_ n_ “
“_Eu_ u_ _r_ r __=__ _“_:_u_ '_ :_“_ _E_u_ _ _ _ mHE_ u_ I_ I_ FE
_ u_ _ _ _ h_! = ___“_ £“' _ _ “_ _ _ _ _
_HH__ h_ _E_ _“_ _ _ _ _ _"_q_ H_ _ h_ n _
_ _ fl__H_"_ _h
inn_ _ _ #|_ -_ "_
_ _ _ ! _ _m _
______
_n_ _ n_
___n_ _
E
H
E__u_ _ H_ H_ _
_u_ _¥_ _ _ _ t_ _
3:53
Ea_"_ _ _" _ _5
_" h_ m_“: _H _ _ "
_ }_ 5_ _ _
_“ _ un_ _: _L_ _ *
"_“_ __n _“____n_ _ _ "_ _ _ _ _ _" _ "_ n_
_ _n
5m_ _ “_ _ "_ _ E__=_____ _E_E;
_____
E:___u___E
EDD‘
____“__
E
__u_E_
______.=
__fl
E?
E
:______
__| E
i
IE
E
_E_ _ _ _ _
H_ H__n_
_E_ _ _
_"_______
___;
__H
Hm ___ M ____"_%_|
_5_ _ _ _“ EHEHE
fl_ _ E“_n_ _ _
H H d _ _ _M 5 m 5 t
-5
M
-M
H _ _ _ _D ____
I
_ _ _ _ H_ _ _ :_ _ _ _ 3“;
_.____-
_ "_ _ _ _
_“________
Us
n__n
_u_:E_ _ _ _ H_ _ H_ _ HH‘_“
5
HH DAHD
BH
m DA
_hE H _ _ _ _ T
HD
_£___b_EH__Ld:"H_____
___]_Er_ l_lh_ _ _ _ _
_____
_ _ ____n_
B 5 ?EDB
\
I-l_+‘_-___lli
_______D 1 4
-_IIIIIII
___
_1|r_
~~~~~~
_
i-I-I
I.
FE‘EHE
_ ___-n_“
n__u__“
_ _fl"__L. _ -._ _" _E _ _ "_nq
_ _ _‘\._\ _ \_ 1_\ \_ _\ _
In
__
I__H
J
_*______r__ _________
i_H_Egg
I____.
E__H
_ _mE_ __ __ _Eb?“
_ ”_ _________ _ _
_ _ _HiI_m_
E‘"_-_ _" _ "H_ "_ '_&
_ _ _ _ _ "_
_m
_u______fi
Lb
:"_H_______n_
u_Hm3_____1__
“DP
H=H“
“HEE
£_H_a_E_=_
u_nHu“_ifi“_nm_u5H=_fi_"EU=_nm:“_u“E_HflH
“u_H:_ _ .“
\
EE
__‘
___
___
_u“ __F____'
El_H_EU5
H_=I_H_U_H _- _ E
_l _" D_ “_____i
_ _mEi_ _________
i_"
_"_
H____
_____
_“_“E
____
_ _'
E_"__¥_n__:
_m _ _m_W_“_fi_E__ __ _
_El_=_____m_m_"h_ _:q_: _ _ _- _ _
fl_H_______
E
_ _5_ _"_ n_ _E
_"E:nE_Hi=_H_Ew"_ :E_u _ _
_ E__Tfi__E“__n__"_"__ .__Hn_ _ E
___‘_I__
D‘
m_E_ _m:_=_Et|_=“_‘ H"-_ __nH_ _Hu"_
u___D_
ti:
E
___
asu
Fu_|_ __ _i E_-_“_“_i_ _ _ _
_"EH_E
_m
-5____
__fim?
_"__"E_ _P_ _ "_
“u"n__" u_i"_=u"__=__mI_“:_*_
_ I__
r_______D_
_ _-_ _ .____"_" _D_ "_ U_‘ _' q _“ __B
_ _=___“_
fl___"_
BEL_
!____
____IL_ _ n____ _“ _H=_ _“E_ _D“_ _ _
Eh!!!
_E__-_ _“_u___"_ ___
___fl__inE:§_
_____
fl_ fl_ _ _ _ “_= _ “ H_ _u____£
_H_-___“
_____
_“_ _“ _H"___w_
"n_ _Hw_-__
_I
H
" _HFn___"
_"_ _-_"___-UE_n__ _
I_=_____
___“_
|_ _ fl_“____
__-__"=_ __n __u____
“_-_
_: _m___
__
""5
l__“L"-___"j"_“_'" _qL"____ _"_ _“_E
_E!
“H_I|£_____“
:"_n“_U_ ___"___"_"_
_1111:1]]11111:
______
_EH__3_H_ u_=" _ u_ fi_E “_ _n
:_flE
_"“EaEL"___
I-_ r_ _ C_I "_E_
_"_____!__D"_H|_ _“ u__mn__m_n"_ E___“__
_ ____“____U______
_" _ W_m _ !m_ _ _ n:__r ___fl_"u_mh __ _“_ H:
_‘ ___‘I_££______""__H___m_“ _n_-_“_U:_ _ _ _ _
m
____H
m__n_____E
_n-_£ _ _ _
_ _ j_ _ _:
'5
_“_ _ _
_m_ “_- _u _ _ I_ _
_ "_ _ _H“_ _ _ :
_ _ Efi'_ _-B_
_ &_ _ _ _
H______‘
E:
_ “_ ~_L _ _" _ _ _
_-_ "]_-_&W
qii
I
“H5
UM
_ _ "_ _ _
_ _ _J _ n _
_ _ _ E_ _ _
“FE
____
__"_
ii
E
_m _ _ _ _ _ _ :
_fl:_
u_F _ _= _ _ H_
Us
“_fl______“
ii__t _ _
EB
_"¥_ _ “_ _ _ _ _
_ "_ ._ _ "_ .
_ _“ _ _ _
Eu“
H
_ _ _“ _ _ _
__“___q_ -_" |_ _" _ __ “_ _ _ _! _
H_________
_ _ "_:“_
_
E
_!____1__“
fi_%E_
it
5“!
Eii§3fi_£_E
EEUS
_E_ 5__
_ _ =_ _“' _E|“_ - _ _
"___"_ E_fi_ _E
_ _ _ _ _ _ L_E
H‘
_ _n_ #_ "_ _ _ _ _ _ "____“_
_ _ _H_ u_ _ “_ I
_I n_ u_ u_ _ _ I
i-__"____ r_" _ _ _ _n _
_ _" _U _H _ _ _
____"_r _“_fi_"_ _ ______m_ n_ _fl =_ E
_ _E_ E_ _ _:
_____
n_m_ _ _ _ _ _ =_ _EH
_ q____
_ h_ "_ H_ "_ _i
_iF_=D_L_ _ _ _E__ _
_h______
u_ _ _ _ _ _ _ _"“__Bi
____________
m_______
_ u_ _: in_ _m_: _ __"_ E5P___h__“
b?_ _ _u_ _
_ _Eu
E_____"_ _ “_ __
_ _ E_"_- _ _ I _ _ _= _
_ _ “_"_ ="_: _" _" “_"_
_:fl_-_ _ F_Uh_ _ i _ _
______
_ _H _ _HI_ _ _ r _uH_
_ _ _ _ _ _ _ _E
um:__
E__ _ “_ _ _ _ U_ _
_: _ _ _ u_ E ___H____H_= :_ u_
Us__ “_ fi_" _ _ _____
_ LE|_ :_ _5_
__n_
__I?‘
H|EH
E“
F____£__
E_““EH5
u__n____
_H_EH_un=_"_“H=_"“_|_ hE"__“__n_"_ _ "_D_‘_
Hr_____
‘Ii1]
I
II
_-_HEEE51!‘
____"_ _ _ _"£_ _: E:
?fl_ _=El!_ =_ _
_“ _ _____l
£'_E_' _“|_E_r "_E_
E:
_Hu_IE
_uI_ _ _ _ _H_ _ "_ _ H_ _Hm
“H
E:
“________w_
“n_ -_“ H_nE_ u_ _ H_ "_
_U
H__-"_E
:_=h______ELn:
=__"____m
*_-"_i_"_flE_:_hHF_ _ _ _ “_ _ _._ _ _ ‘E
u_ _: _ _ _
_ _ _ _ _ g_ _ _"_____n_ _“_ _ flE3
__u_"__i_ _i _
EFE
HE
_ _“ E__ _ _
BE
-hn_ _-_ __" _
E5£_Ji
______“__EE‘£15533
_ _ "_ _ _ "_ _ "_u
iiup-__n
EF_ _ _
E_ n_m_ _ _ _ _ ;
__n______u__
_:_ _"_ "_
E35
_ “_ _ _ _
]I1__
_u_ _* _ _ _ "_ HH
_ _ _ _ “ -_ _
_u____!
_ _ _ _ _ _ L_"
_ _ _ E_ _ _
_"E
_"_n _ _m_ _
_ _ _"_ _ _ _ _"_
EE‘
pd
_"________fl
_"_ h____
____
___u_ _ _IE
_II_3?_
-1I_
_"_ _ _ _ _
____F_
Fin____
E_
$5
QM
H T h _ _ _ B r ___“ H h _ _ _H _" d_ H fi H _ "_ H
__
m
__
H 2 fl_
1
4
I
1
_____
B5
T
BHm 5 H
ED B 2 B 1 l
n
H
H?F______“_
II
EI
LI
mH__H__H__
___
i
D
v_ "_":_ _n_ :_
_______
iT
5
HH DAHD
_________"
__f___
in
___E____m_r
_ _ "_ |__
_ _ _ _ _ _ __u_______“_
I
m_ _ _E_I:_H_
_ _._
F_ _ _ "__
_ _ "_
ii:
1-iii;
“Hi
-Lr_
EE_i_ “__“
_E_ _ _ _" _ _
E_“_Efl__E_ _ E_p_
_n _ H5-_ “ _ "_ "_ __:_“E:
:_ _ _ _ _ _ HE=_u_E"_ __"_____"____E_ _
_=_" _ F_ _ =_E
_ _“_
_HI:
_ E_w_u_ _
:___
3'
BF
_ _ _ _ _=
_:!____n __“__H_u _ _: _I_=______
_m_u_m_ _ "i_ E_n______i§__fi_
"E_ _ _ _ _: Jug“
fl________
fl___ _u _ _ _ ~I_ i _:n
E_EH_ _ _D__“_
m_ _=I _ ._l "_
_m _ _h _E=_ _
_ _-_"_
_ _ _ _ _ _ n_ " n_ " _“ _= _" “_
_i _=_ mE_"_ _ _=_
“Iii
Du
n_ _ _ _ u_ _
m_ _" _ _I _ _| Uh_ _ "_H'Eh“-_. _Em:_ H_ ___EW
" _ "_
l_____B
_ _ _ -_ “ _" _ L_m__=_ HU__-_F_ _ _ _uH_ u_ _ _ "_
_ "H!E5_ _ _ _n_" _ _h_ fl_____5_ E__ __! _E_n _ _ _ _ _ _ _ _ ~_ “_
_- “E1
_"_ _ _H_ -_ _ "_ ___“inE
n_ _hm__ “_ ; E_ _ _ _m_g_:_ m_ _ _ _ _‘ _ "_n__"_
EIn_H5:_ q_ _F
_ - "_ _ _ _ _ ‘Efl______
|“_ Hfl_ _ u_ =_ _ _“ "_F
_ "_ I_: =fi_ _" _
__H_:_d
_ _ _E
“_ _ “_
‘E
*9
n__D
H____
m:HHmflfiLaH
HEEUE
H
D__fl__fi___
"“:ai_=! flfi_u____D_
£H_=nu“_uH__HH_u_ _W_ _ _ _ _ _ _ _m____u__ E
.i____H___
“E_ _ _
‘=1
n_ _ n-_ _ _n _ _
i
_ H_:_ _ _ _ h
m_ "_ : _ F_ u_ _
_ _ _ _ _m _ m_fl__
_ h_ _
_"__m:H_"_i__
_“ _ _ d_ _ _ *_ ‘_ ___I_ _1_ _ _
I______
_ _ "_ _
W
_u
S
‘EDI
D
E
__!__
H‘
E
___H_
E
I
inn
=-_ _ _ _
_ _ _ _ _ _‘ _ _ _ _ r_ _
_I _ _ _ _“ !_ “_
EEEE
___-_
E__ H_ _EI_
_ _ _ _ _: H_ _ _
_ _ “_ =_ _ _ _
J"_ -_ “_
IF_I
_“_BH_EE_ _ "
*3
__m_____“
_ _ _“ _ _“ _ _ _!m_ m_
_ED"
u_m_ _ fl“_ _ _ _ _ _
fl_ ____"=E_ ___ :_
_“_j:_ht__ I_
E_! u_ _ _ _ “ _ ‘_
_"_ m__-_"_
l_ E_
xi__ _ _ _ _ _ _
E_E_r _]_ _ _ _ _ _ _ _ _ _ _m_E
_ h_ u _ _ fl _ 5 _
r_ _ _ H_ :_ _ _ :
Em;
E:
H-EE
_u___1"_ _"_
n_F_ E_ _ _F _ _ _i _y_h“_nE_
Hi
_!_______ “_ :_U_ _ _u
E_ - _ '_ _ _ " _ _
r_“ _ _E_LE_ _ _:
_ __E__“m_ _ _
_u____
m__EI_ _ ’_______H_
__&_ _m_E_ _
____"__
Hun
£?_____
__ _ __ “_
E__I
n_ _ -_E -_ fl_ _ _-‘
EJEEE
“_u_ _ _ _"_ h‘
_EI_ _U_ “_u_ _ EFE
Eng
E_!_!_ _E
HIE
m_ _E _ E_u "_ _ _
iEn“"_ _m_ _q _ _ ___
_ uH_ _: E_ _H _ “_ _
EEUE
"_ _ _ _ _ “_ _r H_ _#_____ _
_“ _ _ _ _H_ w_E
_"_
i _ _ __E_ _ _
J
_"_:__“"_H_" _ _"|_
_i =___"
_ _m_ _ __
_£_ n_ __ _ =IE
E2" 2m__ §E E_‘_h _ _- _ "_
H1_:m_ _.H_ _EElEE
iE_ !_ _ _ _“
___"_
"_ _ _- _ _ _! _
E!_ _ u_ _" _ _
2_ __H_Em_ __Q
_ rE_ _h _r _ _E_ _ _ _ "_ fin
E_ 1_ _ "_
EEE_____
__“__TE_- _ _ __ lJ__ __ _ “ _ H_
_ !_ H_ _ _"
_ :_ _
__ I__
H_-___|_ _ !___
£_ Em:_ _ _ _HE____H
E_“F_ _55_ ______! =_ _ _ _ _ U_ E
_ _ “_" E__H_ _ _
E‘:_m_ _ "_|_
_ _ _ m_ _ _ :_H_fi
__
__H__"fi_ _ _ ____ _“_fl“_ __E_E
'_
___________E_ _
D:
:H_ _ _ g_ _ u_ __ _ q_ _ H_ _ _
nE____
__JH_ _ _ _ _ _
__15
it__ _ H_ _ _
EflagEB_ _ _ i:_5
EL____
|_ _ _L _ _
_Eufl_ _U_Em_ fl_ EE_ _“ H:_ _ "
5_____"_
EE
_“__m_F__ “_ _
EE:
______
EH:__:_
__5_ _E
_"_#1
_E
E
u_ _ _:
_______E___
E_
_!___
________
uH________
__n__
____H‘Du
Em?EI‘:_
HE I_ _H_"_
_ _£ _ _ _i
____I_H
_ “m_ _ _‘_ "_
_"_ _ _u
“FE
“E
___r_-_El“
_Eu_ _ _ _h_ “_ _" _ _ h_ _ _ _" J¥_ _ _ _ ~_ _
l
E
1B
i
H____
E
Hm ___ “F 1%
H H d _ _ _M 5 m 5 t
-5
M
-M
Md
H _ _ _ _G ____
I
_ _ H__.fl_ _-_ _ “_ -=-_
u__-
E
_
____'_
__u____
E _ _ _m_“ _:
E _ _ HY"_ _
BH
m DA
_hE H _ _ _ _ T
B 5 TEDB 1D 1 4
HD
I__ _ _E_ 5_ 1_ _
______fl_Imfi_“_ E_mT_
iii_.__£
“___
_ _“ _: H_fl_ _“_ _=“_ _u____"_
_ ___-_
_$2
E _ _:_ _ m
_"_!E_ __" _"H_ _
E“_flE_m_fiWHHEEHE
E
"___“
mma
ufl“__um__E
E=M£__L_
_fi_uEHh_L___"=_Eflfi_=u“gt__ fi‘_-__-_5::
E5
_ _ ._ _:_ _ _
‘FE
:5
_ ‘_ _E:
_H ___: _ E_
_“53
_____““_ _ _ _r_ _ _ _ _ _ _ _" _ _ n_ _“ _ _n
_“_______
E
_“__ ___ _
E
:_u__T__E
_ _m_ _ _ _ _ E_ _ _
___":
_-. _" _ "“_ _"“_ _H _" _ _ “_ _ _“ _Hn_ _" _ _im_ _ n_
“E_U__ _
EEflmb
_"t
_ _I“_ _ E_ E
H“_
_ _n _ _d_ _: "_n_ _ __m__ :__" _____-“U_fl__r_-_E
_ _ _ _"_ __ _ _“ "_ - _l ._“
£5?
______ _ m_ _ _
B-_E?
L "_#__"_ _ _:_
__E_
B
!‘_
.____I_____
m__"_
fl_ _" _E_3_ _ _: _ “_ _"
E_E-_‘_E_ _:_" _ _ __:_in“E:
fl_
H_ _ _ _
___n_____:_"_ E_ _
_ _H _ _ _ ' _in; E__ “_ H_3 fi_
_ _ H_ _ _ _ 2_
_ _ u_=_ _ E
H-.__-“__u
“J:
_H;
____“
m‘:=___:
J___-r:_"___:
n______
m_:|_h' __m*_"__"___"’_ _u___“__ _n_ “_ __ _“ "_ "_ |_ _ _ _
EEIEE
“__n__ _ -_ -_ _ _E _
d_ m_ "_ _ “_ _ _
as
_ "_ |_ u_ _ _ _ _: _n _“ -"_“ _ _U _ "_ _ _:_ _ _ _ _ _“_ _Eu________"_£5
fl_ l_ _H_ _ _ _
F_!_ _ H_ _ _ _
H_'“_|___ _“_ _
_E_-|_ _E_"__
_HED_
_ _H_ _ _-_ _| _ !_ _ _ E
_E_! _ E_: &_
___
ED"q_"__
_H_ _ _ _ _ _ _" _
_ D"
:E-_u_
“_u_m_ _
E_ _- _ _" _ *_ _ -_
_ ____
_LIE
_____
_ _u_ _
E_in
-__ _“ _ _ _n_u
_“_ _‘_ _ _ d_ _ _ _ _ _: fl___ “_ _ _ _ _ _ _ _ _
"_ _ _ _ _“E_ _ _Eu
_ “_ u_"_ _"_ _i_
___:
_i E_ _h _ _ _ _ _ _: ___-"____“_ _=____E_"_ "_ _ _ __“_E_ _E"_ __:_E"_ _ "l_!_HE:
_.___“_ _“
F
_53:ES
BEE
_ _i_ 3_
it__ _ _ _fl _ ___“"_ _ Ev
in_
HE__ _ E_ fi_
=_“fi_ _ :£
F_s _ _ ‘ _ _
5
_ _" In____
T h E H_ r H E h _ _ _H
Im
___:
HH H
M
M
__
_H 2 __U1 4
I
1
fl_
B5
T
BHm 5 H
ED B 2 B 1 4
11
“___
“__
:5
_ _ _" :-_ =_ u
____
EH3
_______
____"_ _____
_ _ !_fl _h
_“_FE
___rr_ “_ _ _“ _H_ “___“_L_
__:__ _ __Eiv
__ __Hn1_=_ _ _ '_H_m_ _ “___;Fig
_m_ _En_fl _= _E-_
E _ _ _u _ _ m_
___"
_ F_ _ _ _ "-_fl ___: _E_____
:3:
_"_n"__"_ "_ _" _ "_| _
:5"
_ "_ _ m_ _ _ _ _ _ _
5
__E___ “H_| _u E_ _ I
____“u_ !_ _“_E_I“ _ m
_ _ _! _ _ _ _H_ 1_
Ci__
E
iiii _:3
:_ _ _
unfit!"
U35.
E
_ H_ _ _ _ u
_“_ _ "E_ _ _H_ _
S
flag‘
£5
H_"_ "_ _
_" u_ _ _u_u________
F_ _ fl_ n_
m__ _ _ _ £"_ _ “_ “_n=_ _
n_ “_ “_ “_ _
_- _" l_“ _ _ _D_ :_
n_
"_ LH_ _ _ _ _ E_ _"________
u____ _ _
“___I_:__“
"_ _ _ £ n_ _
_“ _ _ :_ "_ _E
E:
___H
___
ThM
n_ _ _‘__E _ __
_E____
uH__"_
5:
"___
_ _ E_ "_ %
i_=_ _I_ "_3:_
‘__ ___ _ _
U_ _H_ _ _
______
+5
M
“w___fi___
Ifi$U#H_fi__
fl=":_____H__
“_=fifl_fl_H_
“£g=“_;“UHCHfl_u _ _ M_ _H_ "_|Du_ _ $;_ _ m_ fl_“_“_ N_ _ _ _ _
DD
E__E
m_m_
_=_ -_ _:
3
i inmi
"__“_
_“ H_ “_ _ :_“H|_
EU:_ m_ _“ _ _ _ _
“___.__ _ h_ _ _5 _= _ _ “ _
_ “_ _ _ ___“__E_ _ _ _ u_ _
__U__:E_ E_ _ u_E
U__.Ti:
_m______
_____U_
______
"E
_" _ "_ _ |_ _ _
HI
_ "_ "_ _ _ h_ _ _
|__"___ " _ " _ n_
_i _ _ _ t_ _
_n__:__
n_n_ _ _ _
___”
__.
BEE
H-_fl__Z_
__n_
EE__u_u_
E_n_ _n_ _!
_____
_u p“__E __;_ _ _U_E__ _H _ _ _ _; E__“ _T _ m_ u_fl _5n__
E__ _h_ _ _ _
Ii__u _ _ _ _ EH53
n_ i_ _ _ i_ _E
__.E
E:
"_ "_ _ _
_ !_ ~ _ _ _ E _ _ _ _: _ _
El“
_ _“ _ fi_ = F_ jU =__:!_ ____ H_
n_m_E_
_n_ _ _I _fl_H _ _ E
_____“Hfi__ _ _ _ m:____ “_
_?_ _________l
“___:
"_ H_ _
E
_n____
_ _" 5_ _t _ _- _flh_ _Dn_ _ _ _J _ _ _ __u_ :_
_u_"_n_ _ _
_ _ I!_ _ -_ _H _ “_ L_
Q:
"_H_ _m_ _ _
E__“ _ t_ “_ !_ _:
_ H_ _ "_ _ _ :_ _ _ _| _ _ _ E_
_ _E_ _ _ _m_ _
_ _ _m_______
____
_“_fi_ _ _
Egg
EEH
_
fl_ _n_ _ _E_I_:" _L -_
_ "_ _ _F _m _H=_ _*
“m_ _ d_ H_ _ _ _r.-_‘
_‘"__u____ _ -_d _i _ _ _"_ H_n __EH_ __U_ _
_“u_ _ _ _" __"_ _
_____"_ “_ H_ ___ fl_n
_"_ _____L__
_ _ _|u"_“_ 5
______"_ _D_H_ _“_H_ ufl_
m__ _ _ _1“_ _Ei__"_'_n_ m___:_ _EE_ n_ _
ED__:
_H___=___ _"_ _ _ _"_"“_ _ _
_ _ _H _. _ _ _fi_H_
_“ _n____ _- __“_
“_5___h_
u_ _u_ H_ _ _ _
IE_ _ _ _ _
H‘:
=__n“__ _ _ _
-"_ _I H_ '_J _ _ -F_
"_ _ u_ u_ "_ _ _
_ _ _fl
‘__
_“_ _ _ _ _
___i__
_“_ _D_ _ _
_H_ ~‘_ _ _ _H _ _ _ _ _ _=_H__:_
_ _ _ _ _u _ _r
_ _@_d_m_ _
:____u
u_fl‘_ _ _ _ _- _ _
_u_"
_ _ _ “_ ___=__ _ “_ _
#2#_ _n_ _ _ _
____
_ '__=_=_
_ h_
_:m_ _ _ _lI_ _
F
__=____ _ _-_
m_ _ _m_u fl"_ h_ _ _ _ _ _ _ “_ H_ U_ “
_ _ H_Em_ fl_ Iu_ _m_ _ _ ___
=____"_
:_ u_ _ _ _ E_ _
'_
“W
__‘___
n_ _ _-_ _"
__m_i__ _ _
_u_ E_m_ u_m_ _
!_ n__fl"__E__u____
_ _ _ _ _ _ _u
_I _ _ " _ J U_ _ _
1-H
_ _- _ fl_ U_ _i #_ :_ _ _ _ : Em__ _ _ i_
_ _ _U F|_ _ _
_m _ _ _ _ _
_ _ n_E_ _
E
HE______
EHE
_" _ m_ _ _3u_ _ _n ____‘
I_ "_ "_: E:m
33:51
____E___
#3‘“_ _ _~_ _ _
E_"_ :_ H_iU
i_!E:
______" _“ ‘_ _ _ _
__m____
3‘
EDE:
_"_UH”#3
t_ _ _ _
_u_____
H__h
_u"2
_ _ _F
E55
_ _ _“ _- _# _ _
_ _ _"_E _
£_fl_ _ _ _
_ h_ _H_ _"‘_n_
fi__i_HI
___F_u__
:______
_ _ _ D_
i
IE
E
_uh_ _ _“ _"
ii:
___H_“
in
Hm ___ W___|_ %_|
H H d _ _ _M 5 m 5 t
-5
M
-M
H _ _ _ _D ____
I
u_____ D_ _ -_ __I _ _u _
E_u___E_“_ “_:_ _
_ _"F|:_H _m=_|“_
___“
_ _ r_u _n _ _
__=______
_‘ _" I_h_ _ n_ "_ _" _ "
_ n_ "_ _ “_ _" _
___":_ _ _ m_ _ _
E__H_ _n_ _ _ _
_ I_ _ _“ _ _" _L“_ !_
E__“"_h|_ h|_ -_
5
HH DAHD
nu H
H DA
_hE H" _ _ _ _ *|
%¥
HD
B 5 ?E
fin]
_ _“___E_
D B _______D 1 flq
flif
_E
____u"_H____m_E
_H________
___"
Eu
___
____"_
DE
Ej‘_uE_|Hdn_I _"hE_|n’nuE_"_mI__F“_!_"mE“Hfl_"_'_n~“flh____n____=
."_n_="_ .“_ _ _“ _ !_" _ _ _ "_
_ _"_E_5"_"_ _ I
Eh:
_ _“ _u I_ _ H_ _ “_H
__"_
E_-__|E_n "___-____ _h_____
u_ _ _ _ "_ _ _
E_uh_ _ __n Ln_ _"_ ! _Ema
_E_" “_.=_| D_H _E. “u_
_ _ _H_-__
_ _ _H_=!_ _ _ _
“Q
Hit“I
H_______|_|
_5
“ _‘_“"=:_“_ _ -_ fl_ _ _____H
fl_#3
_u._“_ _ _"_
___fl__
“________F
- ___;"_ _ _ _ _ _
__H___ _ _ '_ _ H
_ _ n_ _. _ _
EEng:
___
_ _ _ _ _ _I _ _u____H__
_n"___“
_ _ _ “_.___?
_ _ E_S:_ E_ _ _
_ _:
E_!_-_h_“ _ _ _ __ I fl_ _E _€"___?
_____‘
_ _ _ _ _ ____“
“_ _"_ _ _ _H_
_ h_u_ _ _r E
___"
___-___“_ 5
_H
m
_m_ “_ _ #
_:
_E _ fl_ _ _
iiii _"_‘"__m___ _
_ _“ _ _“ -_ .H_
_" _ _t__
_ _£
_ndfl_fl
u_ _“ _ H_ _ _“ _fl
E__ _ u _ _ _ E____:_ _ _ m_F_ __ _ _mn_ F_
n_ _ ___ :_ _
n_ _ u_ “_ fl _ _
_"___
__IH_ _-_fl
_E_:H
Nd
i_ _ _:h__ _ _"_ _
‘_=__-__-bh
___E
___:
m
fig
E_E__"
_“u _fim"_Hn_Hmi 5_:=E_n“"_ua_£ _ E _ _ _ “ _m "_“ "_ E_ _ _
_H_E
1"_
_
_ _ _ “_
-_-___:
EFEEUHE
E__ _ E§ H__" _ “_ |_ _i _ _
_ _ _ _ _ _ _ _E
_ n_ _ -_E_‘H_ q_ _ _ "_ .
_n_E__ _F _ iE
E
‘_ _ q_ H_ _ =_
__fl_
_"_“_fi_____:
_ =u_ _“ _
:5:u_ _ _ _ _fl _ _ _
i-"__‘_"__|_._ _ __
_ "_ _ _ r _
_“____"_=
-_ h_ _ _ _ _
E5EE iii
_ _ r_ _ fl_ _h _ _F _
__h__ _ H_uE_“;_ _
E5:_ =_ _ _ E_
_:_____U_ _
E__ "_ F_ n_
_ _ F_ _ _ _
H‘
EDEHE
_ _n _ _H _ B_
_ ______
n_ _ E_ _ _ _EH_" _H_ _ _ _ r E_" HF_ h“_ _n_“ E
FE
h__ _ _u _
__"____L_ _ "_ _ _ _ _:
“__mm_"_E _U|_ _
"_____ _“ -_H_ _ _ "_ _U=
_ _ "_ _: “_ _ “:_ u_
_"_ u_ _ __ _u_ fl_
Em#5
____
_H3
_ _:_ _
___n___ fl_ _! _ E_ "_ n
_ “_ _m_ _“_ _u _ H_ _
5*
I_ "_ _ _ _ _ _
3_ H_u_ _ _
_ _ _ _ -_ "_ h
_ _ _ “_ _ _ u_-
n_ “_ m_ _- _ _"
DI“_____
:____fl_ _ ____fl"_ _ _u_ _ _ _ u_ _ _ H_____|
_fin
"_u_ _ _ _ _“_ H_H_n_ _
_"EI__H
_ _- _ _ _ _" H _____m___D _“_ E___
_ _ ____“
P_ _ _ _ |_ _ _
______"_ _ “_ =_ __ _ _ g_
___|_ _ _“_"u._= _H _ fl_H
__I‘______
E:
_$I_1 _|£_fl. _“_
_‘EE__E
_ n_ _ _ _
_ _ m_ “_ _ _ _
_ _ u_ _ _ _ E_
_ _ _ 5“_:_“_ __-__n ______ _. _ _ _
_"___"
n____!
“_ _ _
_ _ fl_ _ _§_______ _m_ _uE_ _m__ _?__n__
d‘BE
_ _u _u __£_E
_“
_|_ _ _ “_d|_ _ _H _ _ __ . ____ _t _ _ _ _:
_ m_ _ _ _ _ &_ _ _ _ _n:-_ _ _ : _ _ _ _ _"_ _ _ _
Eii
_n _ _ fl_
_u__q
_ m“_ _fl__
_____
“__¥___+
"__"___fl___=__
“+:_u¢_?:"_.i! _E“ m_5_ _ E_ _fl =_ _ _
"_ _ HE_ "_ Du_r n_
in-H__EE
_ ___“_ _ n_I-EGE -__1
_;_ _=_ _ _
_“_u_3-u_ _“ _E_
___3_-i
_ __:__“__h._n_
_ =_ _ _ _
H‘
Th _ _ _
B r ___" H“ h _ _ _H _" d H M 5 H _flH
m
r
H
I U __|__ 4
I
2 ___.
B5
T
BHm 5 H
ED B 2 fl_ 1 l
_. _l \I_lH!\_ I\l_I _ r
fin
_ _ _ _______
__ _ _
“
___
Eu_ _I_‘£_“
‘if
i
_'_
ii:
ii
_ fi___ ____
m_ _n_“ _ _"
_ _ "__H“_ _w_ u_ih___________
_{_£_____ _H_ _ _ _
E_i _“ _m"_ _ _E_ _ E:E‘__ _ _ _"_ _Hu_"_ “_u_
N_ _ _ _ _E_ _ E_ E_ g_
E
_ 5_ _j _ _U _‘ _ _ i_m_____ _"_ _m_ _ _
_ " _fi “_ _" _ _ | _ h _| _":
‘FE“EH
_n_“_ ____n_ “_ J__: __\ flJ___: _ _
A:
-_ _ __=__i _§_“
_!_$:_fl_ _ _ _ ___
_u_ q_ m_ _ _fi n_H_r _
£_ £_n_n_ _
Pi:
“Elia:
m__H_n__ _ _ _ $HE
J
_I_“_ £_ _ _E _= “_ _ _
_ “_:_ _n
_ _ “ _ "_ _ _n _
D
G
HW
__
5
JEE-
E_-___-
_m
“___”
=2“
_m_Ifl““H35”
D_fl___:
"U“_: n_"_mH_HHU_H
_“£_=________
M“m_n!fi=_D_UmH_ “_fl_ _ _
H"-
‘E
“___:
_ _ HI=_| _ _"_
@_ _l __ _E_ _. _U_
E53
__"___ _ _ _ H_ _
_ _ _ _ _ _ _FE
u_ _u _ _ E_-E_u_
:_ u_ _ ______
=_ _
_ _ _ _ _ _I
_ _H_ :_ _ _ _
_"EH
__!______ _ _ "_
3;_H_ _ _ _ E_
__“_
_ _ u_ =_ |_ _ r
_______
-E
__1______
“_ d_ _ _ "_ _ _" _ _
PE
_ _mu_ _ _ _ _r _
E
“E-_ -_ -_ -_H_ BEE!!
:_=_w_ _“_ “ii
€1_
"_ E_ _ _
_ H_ _ "_ _U=_|
EH:__ _
_ _¥_ _h
_:___"_U_-_|q_:_
_ _ n_ n_ _"n_ ?U_ iE
m_ _ _ _ _- _ _ _
__J"_
_ _ “_ =_ H_fl_ _ nfi
“DmJ-_ _______ _ _ _ H_
_€E£_ _ i
_ _" _ _u_ _h_ "_ n_ _ _ F
_£_:__u_m_=
_ fl_ fl_ _"_h_E5_ _ !____n____n_
Hm_ _ _ _ _ _ =_ _ :_ _ _ E5_ :_ _ _
_ _ ____!_| _H_ “____:
_"_ =____=:_ "_ _ _n:
_"_=u
E
_"_ _ _ _ _ E
H__: _ “ _ E_ _
E___"_ _ _ __ =_
fl_‘ _ '_ H“_ |_ _ _
__
_
W_____
_“___U_____
_ _ E_u_ _
_“_u"_ _ _ _ _ __£ __; _"_i______En_
_|mEHN
_ -_ “_ -_ _ fi_ _ “_
___“.
E
_ _ _ _fi_ _" _ H_:
_H___u“______ _ _
m______ _ _
_IS
E=_E-_“_ 5
_ _ _i _H _ _
_i_F_R_
BE?“
fl______
m_ _ _ _ _ _
En
n_u_ _ _F
_"_“I
"EH
____
[_1_
‘
_ _?_ _
EEE
_ "_ U_3_ _§
_ _ _ _m__E_ “_ u_ _
_n_:_ _ n_
'
E
Hm ____ m |___"_%
_“ _n _n _r _ nE
H __H
-5
In___:M _ _ _
_E
fl_____:_ _ _ _ _fl_ "_"_
m___tHUt
_H _B:_ 1_‘“ _ w
_ "HE
_ n_ _ _1_ _ _ _ “_ E
_n_"___:
_ “_ _ _ _: _ _ _" _
_"__h_H__
_!____E
_ _ __ _ _
‘HDH
__"_U_
__!__‘__"____
_ “_ _“H__“ _ _“ _ _n _ _P_“_ _ _is
__l
__________
._1“_ _ _ -_ _ _
-M
H I flu I 4
£_H____
U_H___q_
_ _ ____-_m_"__u
_ _" _H -_ _ _ "_| _ n_“
5
HH DAHD
BH
m DA
_hE H _ _ _ _ T
%_ rlkui
HD
_______H___
km3___?_ _If_
_h _
B 5 ? E D B _______D 1 4
_ _ _ :’_ _1_ :
*
*__ _ “_ _ _ q_
___‘
___"__ _ _F _ _ _"
_k_
E_ _F"_ _ _ _ i_ _
Eu:
__E__-H
__n_UE_ _fih _: fl_ _ fl_
_‘“____:H_“_u
_"_ E_____"_
______
_ni________ u_ _ _ HEE
EH:
Eng“-
_“ H_ _ _ _H_ _ .
:_ _ ____
n_ _
=_ H_n|‘
E
_ _ _ :_ _ _
___:_:_ _i _ _ _
m__ _ :_ _ “E_ __“_____
n_m_ H_ “_ _
_H_ _ _ _ _ _!
_H_n_“_ _E :!__"_ _ _ _
:n_ _H_ =_EH_ _ E
_ r_" _H_. _h_ " _ "_ "_.
“___;
:5_'_H_:
G
m_:_ _____uu_ _ '_ _ __g__
“ _r _D_ _“ _ “_H _"_ _“_ _t: _h =_ _ _ _BEE?
__#_
_E_
_u 3___:___“_
:_ _ _ E
“__i_ _ __"_£__ ‘_ _ _E_"_ fl_"_ _
_n _ "___
_ _ _‘ _ _ _ _
_“_
nE"__E_ _=_ _
E___:__Fah
_“_
_w____ _ _“ _
H___“
‘__H_______
__:_
_ “_ :_ &_
h_ "1_: _ _“ _“ _ _
?__“_
E___ _ _ _ _
E_FHH_ ‘______fl_____“‘E_"_ ___ _ U:=__E________ _ _
______“
"_ _ _mh__“ _ _
___fl_:_
H__ H_ _ “_ “_fi
H_ _ _ |_ "_ _
ii
=_____
DD
5
EE
H
h_“_
E
‘___
“EH
"__D
W
=3:
“N_=__"_fl“u5__w___DE_fi
Elli
Cu“
m_EiHEUBHEHH
_=mfi_"“___"_fi__ur _"_ _ _fi_ _ 5
W_
H_
_ _ _ E_ _ _ _
5+
______
_‘_"_
_ _ _ _ _ _ I"_ _ _ _
_ E_ _ _ EH_
_ _“ H_5r_ _H__“_ _ u_ “ _ H_ m_ _ fl_H_“ _ _
‘____ _|_i_U _=_|H_ "_
‘_ _P_ _ _ ‘_ _ _uH_H_
E"_ii
_ _ m_ !_ _
___“E_“_____;_:_ n_U__ _=_ _HE _ “_ _ _H
_"__=____H
_ :|_
_ “ _ " _ ___“
_ - "_ ._
_ _ _i _ _ _ _ _ _
_-_._
__‘___?
HE
“I:
H:H-i :
ffi__fl_
______Iu_ _ _ _ _ _
____
_ ___;
._-___
_ "_‘_ W_I_“_
fl_ I_ _ _- l_ E!_
§__!_ ____E1
____i
_ _!._ _ __: ___=_.
_| _h_ _:_n_ _§______
m_ _ _ _ F_I _ _E‘_
_____"
m____m
b_"_ "_____n__
_fi _ |"_‘ _ _ a
Eu_"_ _ _ _ n_
Eu_ _ _ "_ _ _ m_ n_ _ ___fl_u
_ _ __ _I_n__:_"_ £3
_ _ __"_ _ _H_
_"H:3;
_n“__fl _E__”__fl_E
H_;_ I_ _ _
_ _:-_ ___|_ _ ___
E
m_n_ “_ _ _ _
_ _ _ ! _ _U _ H_ lU_ _ _
_ _ _ §_ _:H_ _:___“m|__*
_ "r_ !_ "_ "_ _“ E“H_ "_
_ I_
_ IE_E__
_ ‘_ _E_E is__ _EIE
_ _H_I _ _: Pu_ _ _ "_
_ =_ _ h_ 5_ fl_E
n_ _ n_ _ Eu_ _E _
_u_ “E
u_ _ _ _
_D_H____ _ _ _ _
H55“
_______
fi_i _ _ _ _H_ _
_ i_ _u _ _ “_ _ _ _i _E_ _ _“ _ q
___:r_ “_ “_ E_- _ _- "_
H
_
____“
_ __
_“_ :_
_ "_ u_ _ _ _
_: “___“_ _ _ _ _
E53
_um_____
u_H_ _ _ _ _ _ _ H_
_- _ n_ £_ _ _“ _ u
_I -_ ‘_ h_ _ _ ‘ _|“ _ H_
_" _"H__ "_ _ _ _h_
_‘ _ _n =_ _ _t u_ _“_ "h_ _“ _ _ "_ _ _
fl_u"_ ____ "_ _ _ _“_ _
_ "_ _P _ _E_
___"_
"ma
_ _E__
E
E;
H_ _ _ _ _ _ _ EEE‘
FEE
___H_
" m__m_ fi___ _
:_I:_u_ _ _ _
_-fl_ _"-_ "_ "_
’§E____ "_ _ HE
_B____:__"_ £_Hu_ _ F_
uM_ _ fl_fi"__
_: E
“_____" __ _ H_
“_ i__“u_ _r__: _"_
_" _ _m__"m_"u_u_ _
‘___"___:-
_D_P_
___“
“I:
“_"_!_ _ _ _fi_______‘D
____
E_____
_ _1_ _ _ _
_"______
_“_‘_" _“_ _ _"____Dufi
:_“_ _“__
E
E:
ii
__ i_ _:-_ _:
_ "_ H_ "_ u_ _ !_ n_ " _1 :_" _ _H _ _
_ _ _ _ _H _ "_
_!fl
EEE
u_____“
_ “_ _ _ _" _u__‘“_____ _ _". r_§ _ _ H_ “_ _H_
E5
IE“
___”__ m_ “_ _H_
_=__“____H_
___;
i__m_
_"_ _ fl_
mi
____m
___“
G T h _ _ _ B r ___“ E h _ _ _H _" d I M 5 H H H
m
__
__H _ _ _ _ D 4
1
I
23
HUD =UlD
iiBi
in__¥¥F
___
E_"F_:H____:_n _L :_“ _
_E_E__M
_ __mE_ _w
_ _ H_h_!_ “_ P__ "__m_‘_£_
__ _H_ H_ _
_5__UE______H
m=____
__ifi_ ___H-__ _ _
=_ _ _ _ _ _i|:_ _u_ _ E
ti
5:“
_ _ _ _ :_ _ _ _ _ "_ H_ _ _ _ in
E
_ _ _H__ _ _ _
m___=
_ _ _ _ _ _ __ =__ ___
_fl__ _H_ _
E___:_ _ _h__£_D__
E
_w _ _“ _ _ _n _ _
"___
___D__ -_ _ _ '_ _ _" “_ _ _I
‘:___"_ =___"_E__I_ _ _
hEH.=_ _ _"_ _“_ _ _-“__=_fi___u___'
_ :_ I_ _____ _
“___
H__ _ _ -_ _5-Hi_i"_ _ _ _-_ _ _ _n_ _ _ _ _ _ _
_“I:_tim_ _ _ _ _“ :_
E_ _ "_E P_ E
I__ _! _ _"_ _ _
‘___
_____"_ __ _
Fl
:5_
a__I___“_ _H _ H_- _ H_
£1_m_ _ “ _ _ _
_" _1:
_ _ “H_ "_ _ _ _ -_L__E_="_ _ _ _ E"E
_HEEu
_" _ n_ _“ "_ _ _ _ _"_‘
gig_h_ m_ _ _ _ “_
fl_flI___:“_
£_"_ _ _ _ _
H¢
_ _ H_ _ _ _
_u_u" __ H__E_“_ _ “
U
Du
HEEm
_E____u
Eu#
_ EB:
_ _ _ _ '_
E___
‘H__|_:
H _n_I“
H
E
m_
E
__
¥_ fi_ “_H fl_ _
_ _ _“ _ H_ _ "_ :
ii“:
_:
____=___fi
n__D
m
=____H_
m_Cflfl“Hflnflfl
_fl___“
"U_“:=__"HH_u_H
_H_“E£“________
=m“_fl=__uDm__Ufl_“
_"_
"__
=__
_ _n _" £_ n|_ _ a_ H_
ii5:3
*____
_ _fl _F _ E_ E_ E_H_
£_ _ -_ _“_. :_ _ _! _ H_ _n_
_ _-_"_ IE__ -_ _ fl_. "_
_____ iii
_________n__
_ iH_ _ _u_____
_ _ _ __"_
_ _ _" _ _. _ _“ _
_ _E_ _E_E_ _ _ n_E _ _ _ - _
___»
m_
E
_ _ _ _ _E_£E
_ _ _i=_ _ _ _ _ _
_H_"_h _ _" _hi _ _m
H_£_____
_D_fl_ _H
u_ _u _ _ “ _ _ _
fi_P____
E-
"_m=_____:
__‘_m____t
___!
H_ __“|_______"_____m_i
_“__EH_=__"
__ _H_E_______i
_ m___:___H_h_ _ HE_: _ _" _ _- H_ "_ _ _UE_ _m_ _ _ _ _ _u=_ _
55::
____“
H_ _ _- _ :_ _ _" —_ Ea?
_1“
_u____’
fl“_E
‘_ _£____I
-_ _‘ |_" _u _ _gm
_“E
________£2
_ _ _ "_ _“ -_ _ _. _El_ EEE
_"_C__
_ _" _ !_U_h‘_’ _“ E_! _=L_E
-I_ _u_ _H"_ _ _
_ _“"_P _ "_JHh_ ="_ &_ _"_Enfl_ :
_Eu;
fl“_____
=_E
‘_______
_w
HE
HE
__"_“m_ "_E
_ _____
nw
dd
_"E
E
_HE_"E____"_ F_ _E
_ "_
33_!i_ _ E_i§ _“ _
=_i _"H_ __ _- _ _: _
‘_ _ “_ _§
_u_____“E_ _ _ ;_
E__ _ _- “_ “_ _
EDE_u“__F_ _
_ _ _ _ E_ _
_ -E_ #n_ _“
G
~__"_____-
___“
nag_ E_ n_ _
_ _ _4
i
_"___H__
EE_ _ _ _ r
Eh_“_ _“ _ _ "_
_:;
_ _ _"_ _ _
E
MM
% H ___“ d _ _ _M
___“
5
m Umt M _m H _ _ _ _G M
_H_ H
m DA
_hE H _ _ _ _ T
B 5 TED3
HD
AH:
_______D 1 4
1%
_'m_1_ ‘_ _ _ h_|
___“
IE
W
HMk
EW
IM
___| _____
WI
_____
| I|‘_____\_____‘_‘__|| l
“___:
"___?I
_l|
1
1
I_E_n__
Em“___:_ _ _
I___
______"
h_ "_______
1“
_ __
-
E‘:
__
_uE_F_:
_____
"u_§________
_n_ __ :_£_“_
_nH_ :_" _"__ _m_ _ _
5__“_ _: __" _n_“ _
H_£EU5
Em_ “_ _ _
__IH_isE_
_.___I"_ _E_ _ _ _ _I"_
_t _ u"_ ’_E _ L_m_ "
_Er_" _E_ _m
i__=__"
___|u_"___n___
£2
_E
Eh‘E_H_m_Hm=_m___m=__:_Enm5_E_FE_d§u"_Efi d_r _ fl_| _
En
-In__ H_ “_ -_ _ =
1:‘
Elli
___"=_‘ _#__ _ _
_u_ "_F_ _ H_" _5_
m_ _ _" _ _ _ _ _
_"H‘
_ H_ _ __"r_H"_
_m__-_________
E_ 5_
HI“:
_ _ _fl_ _u_
i
_
:5_E_ _: _
!__Ei"i
_ _ _E=
_U
___
Efi _mEl_ _ "__-_____ HE
E
___
EJUE
“__fl__
"_ _m_"_=__ _ ___"_
_ =_ _ _a _ m_E _ _
E
EH“
_-___Em_
=5‘__i_E__
E5
m_ _ _ "_n_ H_:
“___
_ _u"__E_"__"
!_=
_ _ _ £__ h_ _ _ fi_ _ “
_ E_H_“n_ _ _
_nu _ “_ F_" _" : _
fl_____
__u_
____
’__r_ _ _i_ _
_ _in;
_m_ _ _
E
"hi
_ H_ |H_ _ "_-
HEB____
u__“_“ _ _ :“_ _ _
_m__‘H_m__H__-
_ _n_ _ _ _"H_ _“ H_
5__ '“_ _-“ _- _
E______
_ _E_ _ '
___ _H_ _ _u _H
ED__“ _ _ ___"__r HEh______
_ "_
_=u_ _ _ _ _ _ _ _ _ 1_ "_ _ _ _
_ H_ _E_ n_n“_ _ _|
"_ _.E_ _ _ _ _ _: _
_ u_ _H_ _ _:__n_E _ “ :_ _ n_ _ " _u
I_"_ "_u _ _ H_
_“E
_ _ _n_ _ _
_u _ '_ ____‘ __" “__ _"_ _
_u_ _u______
_ _ _En
5?_um?_ _ _
fi_fl______m__“
_ _ D_ _ _____
___“
_b H_ _ u_I
_ _ ____
_______ _ _n _ _ m_
E
_ _ “_ _ _ _
_"__ _" _ =_“_ _
_____
_m_ _ _“ _ _ _" "_ _ _ _
_F n_ _: _ _“
_l_ _ _ _ _“
E
_ “_=
_ H_ H_ u_ _ H_
__“_fir;
_=_Dhu_
_- _ _ _ _ _D_-L_ _ _
_ “_ _m _ _“ :_
E3
_ u_ _n_ _ _ _
_ _ _ _ E“_
E:
B_ ___"
_ IE_- _. ___
_“____“
=_ "_ “_ _ "_I+h_ "_
_ _:E_F_ _h_iE_
_____E____:__
_"_____.___:_ ___"___‘E"
__£__
=5uH_“_ d_ n_ u_ EHE
E:
_ _" _ _" _ L_I "_ "_H
Etna
fin _ _u _ n_ _ _u _ n_ " fl_ '_H_ E_ _ _I _
_h_ _m_ _ _nfl_ _ |_ _H=__"“_ ____“fl_________ _ _
EEng
_ _H_ _ _ _
___‘
i____"
_=___H
_"_____r
_F_ _ _ _! _" _n_
I.
_u_ _Hr _ _EH“_ _ E1_ _H_:_
_ ___" _ =_“_h _ u_n _" _ "un_ "_ -__ _“_
_ _=_" _ _“"_|_ H_Lnfl_ “_ h_ _ _ _ ______I
__
_ _ _ _ _ _ “ -_ _ -_ _I_ _ '_-_
_ _ _ _ _ "_
_‘____ _ _"
_ _ “ _ _n _ n_ u_ _ E:
_H_ _ :_ _ _
_" _ _r_ _ _ _ _
_ E_“ _ _m_ "_
___“_ _E_ m_I_ _“_
5:
______“______
:E
______ d_ “_"_ _m_ _ _ _
_ _ "_ _ E_ h_ _ -_ E
_ _mE_fi_ _fl_
_ _n _ _i
_ _ ?_ _H_ E_ _______:_ _ _ _ _ _ fig!_H_ _ _
_“ _m"_ _L _ _ _ "_ _ E_ "_ _E_E _ _ _
____:_ _ _ _
_l _" _ _ _ _ _
_“_ _ _ _ _ _
_i n_ _ _ u_E
____
n_ _ _ U_ _E_ _ _ |umDT‘D:
h_ _ _E _ _ _n _
2I___“E_fi___
_"_“ _ _ _ _ H_
___“
_"_“___:
u___u_ _ E_
_" _hm_ _ _E_ "_
_“___n_
H_=_ :_ m_ _
_____m_ "__H_ m_ _ _I
5
hH5
U_“ _ _ _ u_ _ _ _
‘D
u=_ _ _ _ _ _
E_____ _ :_ _ _ |_" _ _“ _' _ u_ _
_=_ _E_ _‘
_:______
___
=53
:u=_=I:
_. _ _ _ _- _ :_
______
I"
_ U_nF_¥_ _ _
_“ _ u__F___ _
_____
E
u_ “_ H_
____
E2
‘___:
E
_n_ _
_" _u_ _" _ ’fl_
-|____
_n_ u_fl
_i_ _ _ _E_" _ |_ _ _
_" _ _| _ _ _ H_
£_ fi_ E_ _ _
“En
"___"
m_____
_'____“_
a
_ _" InH
T h E H_ r H E h _ _ _H
Im
___
HH H
M
M
__
H 2 __U1 4
I
_______5
B5
T EM E B 1 4
BHm 5 H
5
HH DAHD
___-I
_WrL|T
II_________
_H_ _ _ _h_H_ _
_:__b_____ _"E_ _“ _ _ _
E“___:
_~___E__i
EE___ _ E_ _ ___
_ _ - _____ _ _ - _ r
I
______ __
fiflfl
_3
mi
"E:
i‘‘FE
_H£_-_ij53:'_ _ m_
*
'-
ET
I_IIEEII
_U_it
“_ _E _uhE_ _h 5___P_ ‘“_ n_r=u_U _ -I_ "_E_""__E3“I______iH_ _ ______ _ __=____H
"_ H_ _ _="“_ _n _ E‘
_ _| _ _i ‘ "_ "_ _ _ _"_____ _gum
=n________
_E_I
___" _ _
_ _ _u_“=E_H_ _" _5
_I _: -_ L_"_ m_L I_Hr_
_ uE_n_fl _" _: _u _ _ _ _
"_!E_ ____" _ _
3
____-E_ -_“ ___EH_ "
_=E_“ _ _ _uF_ _ E_fl
E___tam
_ _ _H _“ _ m_ _
"___
Hfiflfl“
Ifi
____
_ fl_ _ _ ___: _ _ _n"_ _ E="_£H_ _ E_"___"
_m_L_ _ _ _ "_ _H_“ _ :_ !$‘______E____
____" _
_u____"
___EH“
_ E_ _ __ _ _ H
__"_£_ _ _"_ £_ _5m__ _ _
“_H:__._E _ _“ EEH“
L__ithF_ _ _i f_.
HI
E_:Eh“h_____
__h ""__ _H_ n_IF
_H
:__n:_n_ _ “_
_ =_ _u_
E5!________u_ _ _ _ _ _ ___:
G
___flH__:_ _ __u____|
___h___
_ FE
_ _ fl_ _ J_" _| _ _ “_
_= "_ _“_ _i “_m_ _fl__u_
____“
_ _“_ :_"_ _ m_
E
E:
‘E
‘E5
_n___E
_H:_ _E_: _ _
‘_E
“_£__-_
_ _ _u" _ _ _ _
_" _ i_____H
_u_ _ _ _ _ H_= _ - _ I_ _ _ “_u -_EEK__ _ _=_ _ u_
“___
___“
E
____:u &_ _ _ |:_ _ _ _
_55
______
_ m_$____
_ “_ _ _ :_ _ m_!___
__"“__:_r___ __m"__ "__ “__ __-"___F-____
_____
_ "_ I_ _“ =_ _-
____m__ _ E‘h_ _ _ _ _ "_ _ _| _ _ HI
I
_ _ _ _ _ H_
“E_ _ E_ _ __F____ _ u _ _ _H _E ‘
_" _ “_ _-_ n_ _ H_ ‘_
E
*_ _ _ _ _ _H_
H
j
_w___fl
i
‘_____E_ _
_“_
_H_ _d_ "_
I
!F.
_ i_mI _h___.
Eu_n _"_=1 _ _ “_E_H__ _uE_____
flE
_ _fl_ _ _nfi_ _E _ “_iu_s_H
H1
n_E512:
E:
_E _i_ __ 3_ _
“_____
_Em_H_E:Hm___fl_ _wfimflE_r“"_u“:d_F_N“_HE_ B
_E__
_E
u§£3
E_"_H
_Um____D5““_E"=_ _Hu_ n“_ “_-
_EE
___Eu___E§_ §_ "_ u
|_ni_nHIE§_i
__u__:-__i__i_£ __i‘_I_§__
____|:_
_ _ _ :_
“HE
“___:
m_ _ _H"_: _ _h "_u fl_u
_L___"_EH _ _n_n"____" _ _
“_ _b"________
“_ ___ _ _
FIE5R_ _E_
_H=___I
_ _ _a_ _ _ E_u“_ _:h_ _ _ _
E__ _ _ _ _h_ "_§_
_ _!u____
_ _"_ “_ _
___‘:
_T|_ _E_"_ _
______“
_____
_i_hm_H__
nE‘_ E_ _ _1_ _i _ _ ElE_
_“_ H_“_ _
_E_= _u_ _ _“_:
______n
_ _ _ _ _ _ _ _ “_ _
EHn__u____“
__ _u __“ ="___“___ _ E_m_ ‘_ E
EHHi
EE
E:__= _ “_ _ _ "_ N_
“:-_ "_ _ _ _"_
_ _ _ n_u _ H_ _
=__
_m_ _ _ _|_“‘_ n_n"_ _- _r_J_
__EB:_§_ _ _ _
F
_H_ _U_ _
_
_u=_
__fi_5
___“!_ _
_m L_ E_ _ : _ _ ___
__u_"_"_u_ _ _ _ _
_E
_Ea
£_:_ _
_"_"_"__"_h__"_ _E_"-_
_____E_n_
_" _ _ _ _ r_
n_ _ _ _ H_ :_E_
hi
_ u_ _
fl_ _:E_ _ __“ __"_
HE
_ _fi_ _ _n_ EE
_ _ _"_ _ “
_HE
_ _ _ r_ _
_"_ _ _ __ _E_m _"_
_"_ _ "_ _ _I
‘IE
&_ _H_ _I _ _ l_ _
EEFE“
_ _ _ _H5_ _ _ U;n_ _ _ _ _ _ "_ E_ _" _ __ _ __ _q_‘_d_____ _
____:_____H_n
_ _ _ E ‘D
_E_______1_
fig: _ _ H_ _ n_ _ _ _
_" __“E_ _n_I_.___
___E
_ _‘ _u_ _ U_
_“M_ _ _m_ _ h_
m_=_ _ _ =_ _n _
E“:
_ _ _ _ _: _ _ _ _
H_LE"_ _ | _H F_
H
_i _ "-_ _ _I _ _
fl_ _ _" _ i_ i_
_"_H_____
_-_ _ _
flit
_=_H_ _
‘_ _“ _ _ H_ _
u_____ "_ "_
_ _ u_ _ _
EEH
Eh-in
_ _E H_ H_ _
_" _ _| _ _ _ _ _"
_ _ _' "_ _| u_ _
ECU
[
IE
i
E
__m ____ Wm M H
I
E
___“d _ _ _M 5 m 5 fi t M _m H _ _ _ _G M
P
HEEWE
_ _ _fl__ __
“_____
__;
_ _ E_ _ _
BH
m DA
_hE H _ _ _ _ T
B 5 TEDE QD 1 4
HD
I
all
I-_____
I
___‘"Ff
‘F
E rt
T
I
$+
_"____€F
I
i
1
ii:
‘E:
_" :_ “_m__ImF_ "_E
____________n_“
_“n_"_ _ _ _ _ _
_" _ E_ _ _____"
_ I_H_ _=_ _ I_F
_ "h____u_
_“_ _“ fi_E_ _E_ _"“_"_
I
K
I
iIIfl.|_
E_ ____:
dE_ _Lrn_m: _ _ _ _H_
__"
_:E
iHi
__E:
“-___”
.E:
______
___"_“"__t“_m
_!q"__m_:_u“ t_ "&__E
_: _
_"_F_“=FaEfl_| _n"_“Efl-_=_ _ -_I_ _HEBii
_H_"1_n_H__H_H_mh_U_ £_ _
‘2_
__
nED_" _ h"_ _I _ _ H_
___
_"____:_ _"§_1_ g_ __ ___
HIEIHE
5_H_h_P_5_
Dczmun
___E
Ehfi___=
_En
EmH5m__m___
u_nEn__5uE_Ffl:um_Ed_r _ _fl |_
E:_ _ “_=_ &_
SEE”_ _ _ _ _ _
“E-_
___
u_Jfl_‘
EH“
Eel
Em
___I
‘HIE
F____‘
___‘:-_
__-fl___
PHHFm_
in
_I__
____"
_I_H__
_=_E__I
-____ ._=_D€__ I_ |_ _ _"_ _I_ _‘l_'_ _ _fl_ _ _ "_ “_"E‘
_n _ H_- H_ E_ n_ _
n_ _ _ fl_ _ -_ _
fl _L _ fi_ _ H
H_ H___‘
_ _ _ _ :_ ___ U5__H__
_":_ _E_ __“_
_:____
anH
§_ _ E_ _
_ H_ _ H_ “_ _
_ _u_ _ "_ “_ _
m__ _ _ _ =
____
Eu:
_ _ _ !_:
in
i
E
E:______ H__ _ _ _ _H___ _ _ _ _
__:_
__u_____
“EH :5
‘E
_____m HL;Q___: P_____ ___H__
U
__b_E
___m_anHE
_“H“_"_:J_H_H_m_
E_.H_F;"_nm_i"_Eflm_i:__:_E____ _ _"_ _
H_ _“m_ _ _ :_ _ _
_ _ _ _H_m_ M_
EE
____m__“_-H
"____"_ _E___?_q_E n_ _E_ H_ n__ _"_
tHi
“_ _“:_ ___“___ _H
IKE
__ __ _"_‘
E.E__ -“_ “_ _
_“ _ _ _ _ _ _ ~
_ "_ _ _ “__ _: Eh‘
H_ _=___ _ _ _ _m_ _
IE
__E_H_____:
_ h_“I___
_ _ _ _U_
_ __H_m_ ‘_H
=__!_! _ ______ :_u_ _ _| _
___
"___:
Lu“
fl___n__- H_- h
fl_ I_ _ !_ 1_ I_
U_ _ ‘_ _ ‘_ _H _ _Hu_ _‘
h_ _" _ "_ _n _
iq_E__nH_ _:_
_.“E
_“_fl _-_ _"‘_
HE“?
FIE
_n____:1_“_ _L_| "_ _:
_ _ ““E
____
_._ T_H_“
m_ _= _ : _ _ i_ _ _
__:_
Em“
E
_ _i _ ~_ _h |_I _H_
_UE_Hi
___“H_m_ _ _E_:_H_i ____"
E_
_ _ _ “______
_H____ _ -_“ =_m“_ _E _ m_E"_ 2
_"_ u_‘fi_ I_ I_ _
"h__“_____E__-___”;
____
_“ _____U____ E
_:__
_ _ _ _ "_ _| _ :H_
_ _ _ H_ _ _ _ E
=_ _ _-_ _
E_E_u_3
_5t5E
_ _$_ 5_5
_" _ "_ _ _ _ _ _ _
i_____
____H___
E
__
EE
:5
E
_ _n_ fl_=
_m_ _ _
_Eq1
__‘_H_“
_ u_ _ _ _ _
_ L_" _" h_ "_ _
-
_P_ _=_ _ _ _I _ _
_“____q
_ m_ _ _
E ThE
___m_"
B r H E h _____
InH
Im
___
HH H
M
m
__
_H 2 __"_1 4
I
_______ 1:
_%‘H%
II"____
1:
_ _ _______I _ _ _fl
%RA
=______
_L
___ _ r_
______\__
LT
I[IE
@:_____
P
__E__-#
Efl_‘BE
_ fl_ _ _ _iH
__"=___=
DH
______“
_____E
_____
_"_ _-__!_- ____, _ _“_
_flu
I_ _ _! w_ _ _“ _I __u_
_E|L_" _H__n____
: "_ =_ _ _ _
________
_H _: “_ _ _
E
_____n_ "_ "_ _ _E_D_ ____r__'____|_|_H_t_ _ ___E_
_U_ _ _ §
Egg“_ “_ ¥_ “_ _ _ _
__"_fi_ _ _H ¥_" ‘_
_ .h_ _ _E _ "_] u_ _
_ 5H____:i_ _ _i
fifit#EE
__"E__ _ _fi _~I _ u_1 _ u_
iiE-__: _- _ _ _ !_ _ _
E!______ | _h _ _ H_ m_
_ _ _ _" _“_:_ uE_ "_
E_E_ _ _ _! _H
______";-
__________
___
lb__"_E___
__:___
_uHE
_ E_ _" _ _ _ !___.HE_ I_ _n_ "_ _“ _ _.
E_“___
_"_‘ “____:_ _ "_ "___
E
_"_U I_ _ ______
I_
_“‘_____:
‘_ _ _E_fl_ I_ h_
_ _ “___H__n
_ _“_£__
‘EH
_ _ “_ _ _ _" _ u_ _
_ _H_h_n_ _ E_" _I_ LH_ _"H
n______
ii“E
____
U________
___:__ _ _“_ _ Ifl____
___.iii
i_____‘i tflu
Fi_l
_ _ _" _! _ _“ -m
__2_E
_ Ti_ _I___:
_F_h_n_=T
_. _=_. _an
E_hI___DE
___
u_ _ _ _ _ _ _ i_:__“_=
HE
H“_
E
u_ fl_ “_ _ "_ "_ _
“Rh___d_
u_n _ u_ h_n I
‘_ _“ fi_ _ _ _ fi_- _ _ _ _L_ _
in_ _ E_ _
____T
_ ___":
___
E
E
__J
_;_
_fl_G
:_-_ _ _ _
5
F
_ _" n_ _“ _ _Eh
_ _ _ E_ _ _ :
H‘
____u_
_ "_ _n_ _ _ i_E
fl___'_____
“___”
_LD=_!I“I
_m_ “_ U_ :_ -_ _
_fi _ _ _ |_ _ _ _ _
_ _ “ _ i “_ _ _ _
fi_=_.
_ _‘-H_ _ _ _
_____
E
_u_ _“ _ n_ _ n_
Ii“_n-_ _ -_El_ H_
___“_ n_ _ _ _- _ _ "_ - _
HEEEEE
_E _ _ _r _ _ Kh_
EE
_ _h_ _ _ _ _ _“ _“ _ _ _" _E_“ "_ _ _
FEEFEEE
u__“_
_ _ _ _ "_ _ _E_ l
_ _" _ _fl _ _
i_ _ #_ “_ _ _H_ -_
_uI_5*
__
-_-_ _ _m_‘ _-u‘_ _r_ _
HE___:
_ _ H_ _ _
:|fl__
I___“_E_ _____H_ _
EH
_ _ _ !_
"_ §_ _ _ “_q
______u
:3
_ "_ _ _ H_. _. "_
___“
‘_:_“_ _" _ H_ :_ _ _ n_ m_"H_ _m _* _
_ _ _ E_ _! _ fliinif
:_i_L_m____
_ _ _ _‘_
_Ei_ _L1“
_:_
_ _ _" _flnb_" _. 1
_ _m_ "_ _-_E“_ I_ _ fl_
_____"_ _" _i _ !_ :_
_m“E
_;_:__"_ "_
_fl_E_
_ _ _ fl_ “_ _ _ _ “_
___!
N
EE
_m_ _ _ _ _
_" _" _ _ _ _
_____“_ _“ u_I _ !_ _n
H____"_ _ n_ l_ _ H_ u_“ u_ __ _ *_ _m_ :_ _ _i__
_ _a_ u_ iE_n_
_" n_ _ _uE_ _ _ _ _ _
_"H__m_ __ “_
-_ED
_ _h_"_E_ __ _“ i__=_ "_
_n _ _D___
"_ “_ _m__:_ “_Efl__“"____"
___ _
fl____I
_ _ "_ _ u_fl __Eu“
__|__I
_m_ _ _ _! _ !_
_ _: _ _ m_ _
=53
_"_: 1_ _ _ _
FIE;
_:__“
“___-_
__‘_ _u_____"_ _
_“
JE
all
in
i_ _“ h_Hfl_H__"_HflI_ _-_E_ _
_!_“_“___:
_____
___“_ _ _ E
_ _I _ =I_ _ _
H
:_ _“ _ t_ _
_"____n_
NE
m_ _ _ E__=“_ u_“___n_" _=_ !_ H_
E_m__”__fl__:
___ ___"__EEEE__ H_: _=_ _
n_"______
E
i
E
__m ____ W WM H
H
H d _ _ _M 5 M 5 fi t M _m H _ _ _ _G M
1
____fi_
E ‘H:
"_ _ _ _ _ __m_n__
_£€
_Eu_ _ _ p_ _n _w_
at_:
_
D___
_______
E___£__E_ _r#_ !
_i_!_ _
E
E:
___“:
i_ _ _ _=_ _H:_ _h
m_ E_u _ "_ n_ m_ _ _ _ _ _ _ :_l
_ _ _:_I_=n_ “_ _
23
“___“
_E
__-H_n_:___ -_
_ _ !_m_ _
2
_" _ _ _ _ _
_ _ED__
“_ I_ _ _ '
E_ _ m_ _ “_
Emmhun
‘EM
___D
m
_ u_“W_fi_fl"_“uD_H:_flm__“
D___“HE_flH£_=Hm“_uEH“fl____ u_ _m_ _
__“_
HE
_ _ _________
_ _ _ _£ _"_ _ _" _ i "_ _“ -_' "_ _
BH
m DA
_hE H _ _ _ _ T
B 5 TEDB 1D 1 4
HD
MMHH
%
I-
i
_r_______"_
IH
_____1___
_1
H _1 F
E5
__E__I_m
5__
h_ _ H
___fi_
_ _E
E_ _u _ _ _: +__E___
EflFE_ _n _ _
____"in"_______
_I _ _ _ E_ _ u_= _ fi_ t_ _ g_ _ H_
BEEE“___:
_ _H__:I_“_ _"_m_H_._
____
_ _ ____
~_ H_ "_ _m_"
_ i_ _ _ _ ____‘:IHEE
‘_hWE
_ _ “_ h_:_ _
5::_fl_ "_ m_“ _ I_ _
E__ m_H
_E
____"!H_.____i _ ___n_
EE__u___=
2"
__“
FE‘
_ I__:_ :__._____
_ “_
_ _ _ _ _ _:"_l_: _ _ “
_|n H_n _ H_ _" “ "___“___!___“__:_
_" _ u_ _I _ ufl_h _ _
m_ _ _ _ _ _ _ _ _: _ F_ _“ _ . _ =
_ _ “_ _ _ _
E
Ehfi
E
ti
_=_ _ _ _ _ _ _ _ _
-_ H_ H_ I_ _ H_
E___!__i
_n_ _H_ _
Hi!
Ii
m_r _ ______
_‘ _ _ hn
_ _"I_ _-_ _ _
_Hu_|_H_h__ J“_ _ _ _
_ “_ _ “_n J_
E1_n_ w_ _
fl__H______ _ _ :
_"__
g_F_ _ _ _ _ _ _“
_ _ n_ _ _ "_ "_ _
=__ _ _ _ _ _
_b_H_ _H_
Ed
-l
Z
D
_ _ _E_f" _ "_ _
duh“__iEI
_ _ "_ m_ _ _ "_ _ _ _" i
H“___“
"___"_
inn
=5
=_r_= _
mH___fl
___;
_ fl_ _ _
EEIE.
_hi
:_ _u_ _
_ "_ _" D_ :E_
________5
_____
H_fl__
m
EHJEHE_
m_n_E__
"__h=fi“_u_u _ "_ _ n_ :__i___;__u
IE5
___fi__
_ _ =_ _ L_u _ _
_____H
__I
_‘__“_____
_"EH
_ h_ _ _
"_ =_ En_ _En_ _= _
_n _ _- _ _i _ _-" _-
_=_ _ _ fl_"_"_ _ “_ _ _ _ _“ _ _" _ _=
_" _____E_
_____
_ _ _ _!_“
_Hi5:_ _H3i_ _fi3u__ _!___;_lEFEE
$__"_ ____u_ . _ _ i _ _ _ "'_-
_l_fl‘_“F_m_
_-__I“___:
:'______"
"__l -____‘
n____D"_-_______"_' _T. _ _ “n
“___“_ _ "fl_ "_ _"u_ _"“_
_ " _H _ _ _ _ _ _ _ _ _E_ _ _
_ _ m_ _ _| _ _ E__ n_: _ “ _ _
“_ .E
!_ _ "_ _ _ . ___n_
EH_ _ “_ _m_!
_3_ _ E_ "_ _: "_ h_ _ _ EE
E__ _ _ _=_fi_
E_P _ E_ ! u_
_
_"_______
2
_ EH
___
_ _H _ _- _ _ _
fl_
E
m__ "_ _ "
E_ _ _ _ "_
H
{
E
5 T h E B r H E h _ _ _H _" d I M 5 H H H M _H
m
__
2U 1
4
I
29
Mm W ____HH DAHD
E 5 T ED B 2 B 1 4
L
Li:
___:
:_
:___“___
T_:_._______m
____D_
“_ _F_ "_ t__A
"_ _ ____
_ !_fi _ _" _ _
ii
“___:
flan:Eli
m_ _ “_ _“ _ _ _u_‘!______________"E___ _ __ _
____‘
__F_U____:__F_______"_
H_“E
_ _ F_ _ _ _ _ _
_‘G
"H_h_H‘___ _ _ _
“_ _u_ _J ____
_ _ _ _ _ "_ _
[
_ _ _" _ _n
_“_fin___‘ _ _ _ "_
_ _ _____ _ _£3
m_| _ _ __________
=_ _ _ _“"
_ _ _fl_ =
_ _ fi_I _ _5 _“H-___:
_n_ fi_ _ h_ _ "_ H_ _ H_
_ _m"_ u_ _H_ _" :_“ _m_
___
_“_
‘H
"
H_ H_ _ ____“:
IJ
-_ -_-_“_ -j_E_n_=_ _ "_ _
:5E:
Dm
____
_ ____£_
___
fi___ _ _
2_ H_ l_ _
E__j
3
in
“Bi
_ H_ _ _ _u"=_
_:__
_"_ _HE_"_ _ _ “
h__"___"____ _ _
_ _ " _ _ _ _ _ _ _ _ _- _ _ _ _| _LW_ _
m__n_“___
_ _rE:_HE“m_:"E_“ h_m_Ln_ _nwm|_q“_E __amfi_ _: m_Hm:_’Wung“_m_m_‘ _EmE___E_uu‘ _ U_ H_ _H _ m_ u_ _" _ E_ _H m_ H
*8_"___h
N___h___
_"“I:_ _ F_ _ _
m__u__ _H_ :_
_" _ _ _h _ ~_ _u
___!-_
_ fl_HE;_ _n _"_ _“-n_J _ _ in
_ H_ =_ _ _' _ _ _ i_
______
*5
m
_m_Hi_"“_!=£“fl_#H_H_fl_“_ _“_n=W_u_“_u _ _m_=_“_ _ _
_ r"_ _ "_ h_ u_ _
:_ _ "_ _ _ “_u _
_H_n__E_
____
_E_H_ i:_fii_E_“:_ E_
IEE
‘_ _ _
E"__ u_ n_ _ _!_" ______H_ _ _ _ _E_______‘
_". |_"n_ "_ _u_ _h
_ _“ _ _ m_ “E
_E
u_DEE
_ _ _ _ m_
'_ "n_ _" -_"_E "_ _
n_ _H_:_n5_ _ _ _ _
_"E
_ _ _u_" q_
"DC_n__ "_ _ _ _ “_" u_
_ _ _ 3_ _“ _" q
_H“ _EH
:
_~_1_ _ _:_E_"_ _ _ _F__n_
_‘“H_USLIP
U____H__"
H____u
EH
___"
__F___FH____
__u___
_ m_____m_
_ _u_ _ &_" _m_ _ n_ _ _
_"_n_fl' In_"I_“_| ____"_I ___“ __“ _
E
__
1_
_ _ =flU_ _
r_-
3D
i
E
T h ___ B W
r
H H d E m ___m
-5
5
5t
H
U
t-
H
I flu
I
____
HI H
m DA
_hE H" _ _ _ _ T_
n_ 5 T E D B 1 D 1 fl_
HD
EH7 T1
F
Hugh
"__n_~fi__H_______
___I_ _ _ H_ _ fl_ nE_“_“_ r____"___+
‘D
HE:
“Ci_
_ H_ _ _H_ _I _# ._"_ _ “
H
H_H_ "__I“__"|_ ___fl_hn_ __fi_t_ _
__n_
L
-_
"___“
Q-__""__
__q_“‘___“__" _
Ii:=__ _ _
_* _ _E__-H"_ P‘_;
“n_ _=___
_ _ _H_ _ _-_ _:
E“___:Hfl_ _ _“ _Eand
_ _ _ _ _ " _ n_ _ _“ _= _ _fi _ _“ _
E
"___:_ “_* _ fl_mt:H
E_ _ _ _ _ _ "_Iu__ "_ H_ _ _" |_ _H
E_E_ "_ '_ H_
fl___I_ _____-H____=______
_ _ _-_ D|_ _ _ _
_E__|_m _"_E_:_ _ “_"_ "__u_______
_‘ E
_ m:_ _ n___fi____
Eta“h_i_ _ _ _
iii‘
___"___E
q____
______"
_ _ “_ ___
u_n____EE_
__
_H_ _“___
_" _ _ h
_ H_ _ _ _ "_
_h_“___n_:_
-_-__U___ "_ _ :_ _ _H
E_:_ iu1_ _
m_____________
_E_ _ fi_ _ :__ _ _ “_ _ _ H
“___:
_'_n_"_ |- _ _ H_":
"_ _ _ _ "_ _
ii
m_-__ _ ______fl
_ _ _:
“_ "_ n_ _ _ _“u_ _h ‘
_ “_ "_ _"u_ "_n _ _
_h _ _ _ _ _“
_ H_ _ |-F
_fl_-IHimi
___;
E=_“ l_-?i_
_"“_‘=_ _"_
D“:1
_m_ _m_ _ fi_ _ _
4‘
mfi
_“_I___!
mgI
_Ju_
_
_H_u___H_
_h"_fl=_ _ n_ "_.
_ _EH:
I_ _m_i _
!_£_u_ _ n_ n_ _
_i___fiH!_ _ _ _
U
E____
._H______
___u
_HH¢E
_ _ _ _ ~_
_Eh
_" uH_H=£G_"
E
“ E_u_ _HmE=_:_mnflH_u_ H=_ E_ _ _H
____“
_ _ _“ _Gfi_fl _E_ _
_ ________
_“_
_u _ _H_ _ _
_ _ _ _ _= _- _
___“
H
_ _ H_ _“ _E_
nE"=_ _ _ _l _H_
Hfl_
_ _H'_m_ mL_F_ l_h“
:_ _E
._ _ _ _. _ _ _E_- _
_" _r _ h_ H_ _ _ _-
:n_ _ _ _ E _u
_ _ _E _ _ _
_n _ _ u_ _“ _ n_
_ _1 _ _ _ _
m_ E_ _“ _ m_ “_
__u
___;
_H_I_ _._ _
_ H__‘_H=_H_fi_ _
_ _ _| D"_ _u _ "_ m 5
_ _ u_Eu_ Eh_ _ _
_ _u _ _ n _ “_ _ _ _
_ W_ _ u_ _ _ _ I _! _ _ _ "_-“_ _ "_ "_n
_“_ _“E
u_ _ _
__u_
in_fi _ _ H_ “=_ :_ _:
E_L_“_ H_ n_
1_U_“E_ _ _
__D_L__ _ _ _ "_ _ £u:
u_ _" _ "_“ _ "_ “_‘
_"_ "_ _ "_ _E _ _ -_ "_
I;
__"___-_ H__| -_E_ "I__" u|_ _ l
EEEEEE
_ _ _ _m_ _ _
E_'_HH_“_: _ "_ _ n_
_"_______
_flu“_ _ _
_ _H _ “_ _
1_ _ _ _ _ -Hi
E_ u_ _" _ _
E-E
_HE_ _ _
_ _ "_ _=_ _" _____
_E
“_ _ _
|_ u_ _ _ "_ "_
Jim
_ _ -_-_ _ _ h_n_
_ _ _ _ u_ u_ _
E
UEWF
1%_ u_ _
:_U____
_H'fi__
_m_"___
_____E_
5 T h E B r H E h _ _ _H _" d I M 5 H H H M _H
m
__
2U 1
4
I
______ ___.
BRITISH STANDARD
BS ?EUB:2i]14
13
13.1
Unclassified details
General
Details that are nbt expressly classified shbuld be treated as class G2, br class W1
fbr lbad-carrying weld metal, unless a higher classificatibn can be justified either
by reference ta published experimental wbrit pr by carrying but special tests.
Such tests shbuid be sufficiently extensive tb allew the basic Ell‘ design 5~N curve
tc: be determined in the manner used fbr the standard classes isee Annex E}.
13.1
Pbst-weiding treatments
Where the classificatibn bf Table ti tb Table ‘Iii dbes nbt giye adequate fatigue
resistance, the peribrmance bf weld details may be imprbyed by pbst-welding
treatments, as detailed in 15.3.5 and Annex F. The detail types that are suitable
fbr applicatien pf the techniques described in Annex F are indicated in Table 4
tb Table iii- when imprclyeci fatigue perfermance is required. and the prepesed
imprbyement methed is net cbyered by Annex F, cur the imprciyement stipulated
therein is insufficient, the detail shbuld be classified by testing [see 13.1}.
14
Wbrkmanship and lnspectibn
14.1
General
The classificatiens fbr structural details in Table 1 tb Table Tl] refer tb specific
mbdes bf fatigue failure. Hewever, the details cbncerned can cbntain features
that cbuld prbyide aiternatiye sites fbr fatigue cracit initiatibn br increase the
seyerity bf thbse cbyered by the classificatibn scheme, in bbth cases pbssibly
resulting in lbwer fatigue perfbrrnance. Therefbre, in fatigue~lbaded steel
prbducts there shbuld be adequate cbntrbl and inspectibn bf manufacturing
quality, tbgether with acceptance limits related specifically tb fatigue resistance.
‘Where the classificatien bf a detail is dependent bn particuiar manufacturing c-r
inspectien requirements, the necessary standards bf wbriimanship ancl inspectien
sheuld be indicated bn the reieyant drawings.
in practice, the byerail quality requirements fbr a prbduct are determined by a
number bf perfbrmance requirements, depending bn its applicatibn. These
might br might nbt be sufficient tb cbyer the quality requirements needed fbr
fatigue perfermance.
Typical perfermance requirements bther than fatigue can include ene br mbre
bf the fellewing:
ai
static strength [resistance tb fatigue br buclciing};
bi
resistance tb defbrmatibn br cieflectibn;
cl
energy absbrptibn {ductility};
dl
geemetricai precisibn idimensibnal tblerancesi;
ei
clearance tb mbying parts:
f}
fit-up ta adjacent parts;
g]
resistance tb cbrrbsibn;
hi
laclt bf bbstructibn tb fluid flbw;
ii
aesthetic appearance; and
ji
hygiene {eliminatibn bf creyicesi.
Sbrne bf the recbmmendatibns in the sub-clause can impbse yery high
requirements fbr specific aspects bf quality.
31
1 E The British Standards lnstitutibn IBM
BRITISH 5TA|lI DARD
B5 2503:2014
The existing prbduct specificatlbn requirements shbuld therefbre be adequate
fbr fatigue perlbrrnance purpbses. if nbt. then the nbn-cbnfbrming aspects
shbuld be enhanced.
It shbuld nbt be assumed that the quality requirements needed fbr fatigue
perfermance albne adequately cbver thbse required fbr bther perfbrmance
requirements. In sbme cases they might nbt.
Eluality aspects which are detrimental tb fatigue perfermance are described
in 14.2.
Guidance bn quality levels needed tb ensure that the varibus detail classes are
achieved is given in 14.3.
14.2
14.2.1
Quality aspects detrimental tb fatigue
General
The presence bf unspecified nbtches in a new prbduct can have the same effect
an the remaining fatigue life as a fatigue crack pf the same site. such nbtches
can therefere significantly shbrten the fatigue life bi the prbduct, by eliminating
the peric-d bf early crack grbwth.
The mbst severe nbtches are thbse that exhibit the fbllbwing characteristics:
ai
planar with a sharp tip;
bi
brientated nbrmal tb the directibn bf fluctuating stress;
cl
lbcated at br clcise tb a surface; and
dl
lbcated at pr clese tb bne bf the classified initiatien sites.
Either quality aspects which can adversely affect fatigue life are thbse which
result in an increase in cyclic stress at bne bf the classified lnltiatibn sites. These
are usually caused by unspecified variatlbns in prbduct gebrnetry.
The mbst cbmmbn types bf quality aspects irnpbrtant tb fatigue are listed in
14.2.2 tb 14.2.4.
sbme quality restrictibns are given in Table 1 tb Table ill. Guidance bn bthers is
given in 14.3.
14.2.2
Parent materials
Particular material quality aspects that shbuld be cbntrblled fbr fatigue purpbses
are:
ai
surface imperfectibns irblling flaws and their repairs, cbrrbsibn pitting,
accidental damage such as arc strike and weld spatter];
bi
internal laminatibns and inclusians;
cl
incbrrect geemetry {e.g. deviatiens in shape, flatness br thickness at the
limits bf telerances. see hibtei.
dl
cracks due tb inadequate rembval bf hydrbgen le.g. heat treatment br
plating prbcesses, bbits bf Grade 111.9 and abbve}.
ei
liquid metal assisted cracking iLMACl in heavy galvanized prbducts.
.-'-.lL'l'TE Design stresses calculated using nbrrrinal thickness are applicable tb plates
manufactured tb the thickness tblerance requirements bf BE Ell.-' ltit'.'l51 and
E5 EN l'lJli2£-il:2ii1lil Classes A, E and E.
14.2.3
Bblted jbints
Particular quality aspects that shbuld be cbntrplled far fatigue purpeses are:
al
cbrrect alignment bf hbles {in un-tensibned lap jbints in particular];
E The British Standards lnstitutibn 21114
I
33
BS TEUBIZB14
BRITISH STANDARD
bl
cerrect fit-up between laying surfaces prier te tightening {mest impertant
with pre-tensibned tensile jeints};
cl
cbrrect bblt tightening prbcedure and sequence; and, in cases where a
tbrque measure methbd is applied;
d]l
cbrrect frictibn cbefficients in threads and under washer faces and the
cerrect lubricatien required by the tbrque tightening specificatien.
14.2.4
Welded jeints
Particular quality aspects that sheuld be centreilecl fer fatigue purpeses are:
al
cracks lusually a sign ef seribus material er precess preblemsl;
bl
lack bf fusibn br penetratien {surface and embedded};
cl
undercut at weld tbes (transverse tb stress directibn];
d]|
vblumetric flaws {e.g. pbrbslty, slag};
e}
excess weld cap er reet prefiles;
fl
less ef weld threat lbutt and fillet welds}; and
g]
linear br angular misalignment in crucifbrm br in-line butt welds.
Nb unauthbrlzed attachments, intrbduced as fabricatien aids br later in service,
shbuld be welded tb the prbduct in regiens where such attachments wbuld
reduce the classificatien assumed fbr design. Unless allbwance is made fbr them
at the design stage, they shbuld be rembved and the area greund but and
checked by MDT te ensure that he unacceptable flaws remain.
14.3
14.3.1
Cbntrbi bf quality fbr fatigue
General
Table ‘l tb Table ID prbvide limited infbrmatibn bn manufacturing requirements
fbr certain type numbers. In general, mbst bf the necessary cbntrbls bn
fatigue-related quality aspect: described in 14.2 are nbt previded in the tables.
Guidance en cbntrbl bf these quality aspects is given in 14.3.2 tb 14.3.4.
The detail classes given in Table 1 te Table tlli denete the highest recemmended
5-hi curves fer each detail type. Any manufacturing recemmendatiens in the
Table 1 te Table re assume that the detail is te be stressed up te the limit
allbwed by that S-N witheut a shbrtfall in fatigue life. In practice there might be
a number bf details in a prbduct that are stressed tb similar levels, but have
different detail classes. Dn the assumptibn that all details have been assessed tb
have a satisfactbry fatigue life, details with the higher classes are nbt stressed tb
the maximum limit permitted by the classificatien. Fer example if a cempenent
has a lbng wide welded attachment en its surface (detail type 4.5, neminal stress
class G1 and there is a transverse deuble sided butt weld {detail type 5.3, class D}
clese by and stressed te the same level then the butt weld cannet be stressed te
its maximum permissible limit le.g. Hp iilimmif at 2 x ibii because the class G
detail weuld be limited tb SE1 Himmi at 2 x Till“ cycles. Thus, in the case bf the
butt weld the required class fbr this prbduct is G, l.e. 4S"l-"ii lbwer in stress terms
than the maximum permitted detail class ill accbrding tb the table. The required
class is therefere a functien bf the stress spectrum enly and is the lbwest S—l'ii
curve which just prevides an acceptable fatigue life-
34
l E The British Standards lnstitutibn 2lI.‘ri4
BRITISH STANDARD
BS 2503:2014
It is therefere pessible te select the fatigue quality centrel requirements fer a
particular prbduct bn a fitness-fbr purpese basis taking intb acceunt the actual
levels bf stressing, l.e. the required class. Fbr preducts with medium br lbw cyclic
stressing this is likely te result in significant quality cbntrbl savings cbmpared tb
thbse required fer meeting the maximum permitted class fer all the details in
the prbduct. The required class at any lecatien in the preduct can be readily
determined by repeating the calculatien precedures in Clause 15 and Clause 15
fer varieus Hil curves until the lewest acceptable curve is feund fer that
lecatien. The required class, as in the case ef the detail class in Table 1 te
Table 15, depends en the directien ef stress fluctuatien and is likely te be
different in the tvvb brthegbnal dlrectibns.
14.3.2
14.3.2.1
Parent material
lillaterial quality
Where material is te be welded, selectien ef a suitable carben equivalent value
liIE‘v'l is impertant fer preventien ef cracks.
Where material is tb be used fbr transverse elements in welded crucifbrm br
T-jbints {see element '1'“ in Table 2, types 2.1 te 2.4} and there is the risk bf
lamellar tearing, the fellewing testing shbuld be carried eut:
ai
threugh-thickness defennatien {2} testing; and
bl ultrasenic inspectien te ensure that it is free frem laminatiens and inclusien
cluster.
14.3.2.2
Surface cenditien
The fellewing sheuld be carried eut te ensure the surface is in an acceptable
cendltien.
al
visual inspectien sheuld be carried eut te ensure that the surfaces are free
frem damage fer required classes D and abeve {cerresien pits, arc strikes
and spatter, mechanical damage}.
bl
MDT bf any areas where temperary welding has been carried but fer
required classes F and abeve {repair bf relling flaws, rembved temperary
attachmentsl sheuld be undertaken te ensure ireedem frem cracksci
Excessive irregularities isee Type 1.4. Table 11 in thermally cut edges sheuld
be greund smeeth fer required classes G and abeve.
dl
Ebld fbrmed radiused edges shbuld be checked fbr cracks by MDT where the
class requirement is G er abeve in the transverse directien and E er abbve in
the lengitudinal directien.
el
14.3.3
Belts bf Grade 'lti.5 and abeve which are te be ceated er galvanized sheuld
preferably net be subject te acid er electreaziepesitien precesses. if this is
unaveldable, heat treatment te remeve hydregen sheuld be carried eut
effectively lethenrvise stress cerresien cracking might eccur}.
Belted jeints
Greups ef heles in lap jeints, {e.g. detail types 2.5 and 2.1}, with belts carrying
lbad in bearing shbuld be jig, Eemputer Numerically Ebntrbiled {CHE} br match
drilled te ensure uniferm bearing between all belts in the grbup. Apprbpriate
telerances bn diameter, pesitien and alignment bf parts shbuld be specified.
ID The British Standards lnstitutibn 2014
I
3S
BS TBUBIZB14
BRITISH STANDARD
when the fatigue design ef a belted jeint is based en belts being pre-tensibned
tb a nbminated er minimum value, it shbuld be ensured that the required
pre-tensibn is achieved. Bblt tightening, (and therefere bblt pre-tensibn}, can be
ebtalned using tbrque, angle br bblt-stretch measurements. The belt pre~lbading
methbd and hence pre-lead accuracy shbuld be selected after assessing the
criticaiity bf the jbint. Fer all pre-leading metheds, the cerrect measurements,
terque sequences and terque—checking data sheuld be defined- Calibratien tests
sheuld be carried eut te cenfirm that the required pre-tensien is achieved.
Established data en similar fasteners may be used if ailewance is made fer
variatien between batches and preducts. A system ef menitering and recerding
shbuld be implemented tb ensure that precedures are feilewed cbrrectly.
Fit-up between parts with flat cbntact surfaces shbuld be designed tb ensure
uniferm cbntact at the initial hand-tight stage prier tb final tensibning.
The required belt pre-lead sheuld be ebtalned by the fbllbwing prbcedure.
all
Using a terque centrel methed apply, in a defined sequence, sufficient
tbrque te bring the cempenents inte full face te face cbntact.
bi
Using a tbrque measurement, bblt br nut part-turn br bblt-stretch methbd,
that is apprepriate fbr the criticaiity bf the jbint, apply the final pre-lead.
This prbcedure dees net preclude the use bf a final stage bf tightening using
ether metheds.
Where a hellew jack might be used te tensien leng studs er belts, tests sheuld
be carried eut te measure the pre-lead less, due te the applied dynamic leads.
This sheuld be taken inte acceunt when determining the minimum pre-tensibn
tb be used as the basis bf design.
14.3.4
Welded jeints
Welding is a prbcess where netch-like imperfectibns and deviatiens in geemetry
are difficult tb eliminate entirely, particularly in arc welds. The degree bf severity
depends primarily en the level bf cbntrbl bf the prbcess, but alse en bther
facters such as the type bf prbcess, the materials, the jeint geemetry and access.
The size ef imperiectien that can be telerated witheut cempremising the
fatigue life based en the details classified in Table 3 te Table 1-ti becemes smaller
as the required class {defined in 14.3.1} increases and the chbice bf viable details
becemes mere limited. Fbr example, if the required class is C and a transverse
butt weld in a plate is required, enly detail type 5.1 is acceptable. This detail
invelves rembval bf all lbcal geemetricai stress raisers such as weld tees, welding
access frbm beth sides and a high level bf l'~l[1T. If the cempenent crbss~sectibn is
a mere cemplex shape than a flat plate then the access fer welding is reduced
and the likeliheed ef preducing larger imperfectibns and alse net detecting
them by HDT increases significantly. Fer this reasen the maximum detail class fer
a butt weld in this case is reduced tb F2 [see detail type l;'i.1l, a reductien bf fbur
classes.
in cbntrast, if the required class is bnly W1 er less, a very wide chbice bf jeints is
available. This includes transverse lbad—carrying fillet welded jeints where the
design allews majer areas bf the leint te remain unfused with severe netches at
the weld reets isee detail types 2.2 and T.-'-‘ll Whilst these cruciferm and T-jeints
can be stressed in the parent metal te the level ef neminal stress class F2 as far
as failure frem the tee is cencerned, failure frem the reet threugh the weld
metal is limited te W1 {detail type ?.B]l.
35
l E The British Standards lnstitutibn 2lI.‘ri4
BRITISH STANDARD
BS 2503:2014
Dlrectien ef stress fluctuatien with respect te weld axis is a further impertant
factbr. Weld imperfectibns with mere netch-like features lcracks, lack bf fusibn.
lack bf penetratien. undercut, slag inclusibns} are generally brientated parallel
tb the weld axis and are therefere much mbre severe in terms bf fatigue
perfermance when the stress directibn is transverse tb the weld axis. l_bcal stress
raisers such as weld tbes are similarly mbre unfaveurably brientated with respect
te transverse stresses- This is reflected in the detail classificatiens where, fer
example. a centinueusly fillet welded T-ieint stressed lengltudlnally is acceptable
fer a required class bf E ler El if it centains step-starts, see detail types 3.2 and
3.3 respectivelyl; whereas if stressed transversely threugh the base ef the Tee it
weuld be restricted typically tb neminal stress class E br F {detail type 4.2 br 4.3
respectively, within certain dimensibnal limits}. If stressed transversely threugh
the stalk bf the Tee it weuld be restricted tb W1 with respect tb weld threat
stresses.
The cbntrbl ef welding in preductien is generally achieved by use ef the
fellewing measures:
al
relevant dbcumented weld precedures qualified by test;
b}
welders apprbpriately qualified by test fbr the range bf precedures used bn
the prbduct;
c}
apprepriate welding quality management system;
d} apprepriate inspectien and testing precedures and acceptance criteria fer
the preductien welds; and
e}
gbbd practice fbr sterage and handling bf welding cbnsumables tb prevent
dampness and cbntaminatibn.
Where fatigue perfermance is an impertant requirement fbr the prbduct [l.e.
the requirement is class W1 er higher} all these cbntrbl measures shbuld be
adbpted in accerdance with ES EH 1b11-1, BS EM 15111-2 {fer ferritic steels} and
BS Elli 1l}11~3 lfer stainless steelsl.
Fer small mass-preduced initial preducts it might be practicable te inspect the
quality bn a randbm sampling basis by destructive tests [e.g. sectibning bf
welds}. lilbutine perfermance testing {by leading} is usually impracticable fbr
verifying fatigue life because bf the time it takes. in the maibrity bf cases
verificatibn that the required quality has been achieved in prbductibn is by ll-IDT.
If high fatigue classes are required the limitatiens bf the varibus testing
techniques in detectien, characteriaatien and sizing bf the varieus welding
imperfectibns shbuld be taken inte acceuntIn seme instances it might be current practice te use the same acceptance
criteria in prbductibn as used fbr prbcedure er weld qualificatibn {see
BS EH ISD 'lS5lZlEI-1 and BS Eltl 2B2-1} br these in BS EH ISD SE12. Such
specificatibns are based bn arbitrary quality criteria and might have aspects that
are nbt adequate fer higher fatigue class requirements. They are alsb likely te
centain requirements that incur additibnal cbst fer lbwer fatigue class
requirements. Fer these reasens it ls preferable te specify quality requirements
that are based mere en fitness-fer-purpese principles. Guidance is given belew
en apprepriate limits bf these imperfectien types which need te be centreiled
fbr fatigue purpeses in additibn tb thbse specified in Table 3 tb Table ll]. This
guidance is expressed in terms bf the “required class", described in 14.3.1, sb
that an ecbnbmic level bf quality can be specified fbr any given prbduct.
ID The British Standards lnstitutibn 2014
I
32
BS 2EOB:2014
BRITISH STANDARD
Where an existing preductien weld quality specificatien exists fer the preduct, it
sheuld be checked te ensure that it cevers the types and limiting sizes bf
imperfectien recemmended belew fbr cbntrbliing fatigue perfermance. Where it
dees nbt fully cever the fatigue quality requirements the specificatien shbuld be
amended accerdlngly and it shbuld be made clear that if any fatigue
imperfectien limit is exceeded in preductien it is cause fbr rejectlbn. The scbpe
ef inspectien might need te be enhanced in accerdance with the
recemmendatiens that iellbw.
if ne preductien weld quality specificatien already exists fer the preduct the
specificatien sheuld be based en the recemmendatiens in this clause, previded
that any bther relevant perfermance requirements {see 14.1} db nbt require
enhancement abeve thbse required fbr fatigue.
‘Where the fatigue stressing requirements in a preduct are mederate, e.g. a
maximum class requirement ne higher than F2 anywhere, a general
fitness-fer~purpese requirement fer the preduct quality sheuld be based en the
maximum class required, subject te any enhancement fer ether perfermance
requirements {see 14.1}. if enly a small minerity bf specific welds in the prbduct
experience high fatigue leading and have a higher class requirement than F2, it
might be mere ecbnbmical tb identify these by their class requirement sb that
extra inspectien and higher quality requirements can be targeted tb thbse welds
that are mere critical. The welds in questien sheuld be indicated en the
drawings with their class requirement as defined in 14.3.1 {nbt their detail class
as given in Table 3 te Table 1-ti}.
Table 12 gives guidance en the fatigue acceptance ef vblumetric embedded
imperfectibns te meet Class C er lbwer {Class I3 is excluded because ne
embedded imperfectibns are permitted}.
Table 13 gives acceptance levels fbr undercut at transverse weld tbes. These
limits might net be small eneugh fer ether perlermance requirements {see 14.1}
and might be larger than in seme general quality cbntrbl specificatibns.
The welds sheuld net be smaller than the specified size. This applies particularly
te transversely leaded partial penetratien butt welds and beth the leg length
and threat dimensibn bf transversely leaded fillet welds, where the fatigue life
is significantly reduced by small shbrtfalls in these dimensibns.
With regard tb deviatiens frem the neminal geemetry, particular attentibn is
drawn te jeint misalignment, as the acceptance levels in many quality
centrel-based specificatibns are tee large fer fatigue-leaded jeints. it is usually
necessary te allew fer the effect ef even rieminally acceptable misalignment as a
seurce ef additienal stress. This effect can be expressed in terms ef the stress
magnificatlen facter, k,,,, which is used either te increase the estimated applied
stress range er te reduce the fatigue strength ebtalned frem the relevant S-iii
curve {see B.S.2.1}. Assessment bf any misalignment is required fbr transverse
butt welds as Class D refers tb perfectly aligned jeints. This is alse true fbr any
assessment bf petentiai weld tee failure using the het-spet stress, altheugh that
effect ceuid be included in the calculatien ef the het-spet stress if the relevant
type and extent ef misalignment was included in the finite element medei er
the het-spet stress was ebtalned by surface stress extrapelatien frem strains
measured en the actual welded detail cencerned. Hewever, linear misalignment
eit up tb l1l.{lS, cerrespending te km: 1.15, is included in the data behind Classes
F and F2 fbr cruciferm jeints. Therefbre, bn a fitness-fbr-purpbse basis, bnly
misalignment abeve this level needs tb be censidered when assessing cruciferm
jeints, Table 14 te Table 15 illustrate the effect ef misalignment en fatigue
strength fer these three situatiens.
313
I E The British Standards lnstitutibn 2lI.‘r14
m DARD
BRHE H 5T
BS 2E0B ED
__113I.H:-_
:%“HG
"_flan
“___:
Em
_uHm_fl"_ _u_m_ _ _
H_ _ _ _ _EmH__ _ _ _ _ _ _ m_UU__ _ _
1IEm
DH“___i
P
“___:U"_ _ _ _
___1
___: _ Hx_
Z
U
_fl_______ "___
4
_E£m___m_E
E
U
n_
Hu_wmEHcmfiw
_HuE_____
_=mn"_mHu__“mHE"n_mu_HE_“Hmfl_H_____n__
Dn _ n
__W
__g
_ _ _ “_ _ 2
HHHIE2
HI
_ HgH__“
_E+__
“mD
H
M
_my
gm
mHu_W_H_nHfi“m_H"_wu|fi:_m“"HEu_mH_EH_Uu_m u_E_Lm_ " E_ _ _m_ _ _E_“ u_m_ _E_ _ "_
_H
_‘_
I
“___
Q_m_"H_H_n____u
_uH“:EHm_:DHfhmfi_|H“mE$_wU=mE:w_u_J fi_nmUH_“ hw__En3_m__E_
H_ u_E
I
E_fl_uH"____
___:
wCHfl_uEH|_
i_2ufiGP_H“m_m_+H_mM_u“HmFE"_m_U"flEU_._m“_ H_En_ m_H_ m"_
_flm_
_“_
H__w_:JP“m"_m"wEH€_m_“hnH_1m£Uu_t:Lfl_wE_u:HUfl_Ewu_HEm_Eu_=hUm_ E_U u_ _
_
__
_ __E_____
E _=“EU;
“u_n"H_uy_| _m_“Hfl_Hun
I_Ufl H_=EU_ £hm_ “_fl fi_ _“UW=_flE_H_fln_M"l_flUu
1
_“_I
HH
“_____n_
m_HE
_____“
_§___
“Emt_m___E5E_Hm"flE“w_H:um_mfl_m___fl_E_
Uuzw_‘ mH“m_Efium_“m__ “_=m_ flmH_ _ n_
E
E____H_E
E
U:
gm
E:
Z
_hfi____mH_:umn_H_wm£E“_mH_mn_Hd_HE"P*_fim_Ewm_H_ E_mfi_mm_ _:_Em_ _g
“___:flit
uD"_ _ _:
H U_____
_“H_fl_“E_m“_w_D"=__mU“E_“_m____:
fl_H"U_$_fl_“ __fl _fl=mn“EH____ENU__=“
U_um H_ _ _________Wu_fi_ _ _ _ _ _ hH_n_ _ _ “_U_H_ _ _1
H
mg!“ mpg:
L_’__
_ !_ _ _ mH_ _ H_ _ _ _
____f__
__1
E_u_____
_U__
__u_
=_mE_E:
___“_m_E§_
fi_:H_mwr~_ _H:Dm_“n_m_uHE_%_d _F m_ _fiY _ _Efi_ _ _U _ _ “
_n__"H
_L
_H__ ____E
"_ _ _ _ _ _"__
i
_mE___“_m_fl u_:mm_"Hnu_pEw:r_n_flmufi“_$H:um_n_Mtm_ "_flE_ “_
cm
L
U U U U 1
H W"
___
:1: _H_ _ _H_
_ “3“
___?_ "_+__" __:
__“_
___ ___
__|_ _1_
___
H_____ZDH_
____
_ "_z____E“
“ _ _ _ ____
_u"_J_
___: _ _?_ H_ _J_ _ _
_$_ __“_ _2_ ____]___"
_ _ ] _ _ _ _ “___
D___
‘U___;_H+_ “3_"_3
EEF
u__H____ EEI
H
_"______H__
___H_H_____cm
E#_ m_"_ _ "_u__
_ _ _____U
“__D
:_
“U__“m___fi___H_H_
Em
_ _u_fl“_5fl“_|U_“fi___U_fi _ _EN_ "_E“_U:_ :fi“_ _ _"; _CDfi_ _ _ w_u_ _ _ _ U
:|__
m_“_ _E_ _
“___”
_MED:
“ _HmD_" _h _m_fi_fi_ m_ _"m“_"m_r _m_ "_ _uh_ _“ n
‘Emu
Z___
]1-1
H_ _ _ L_
m_H_m_
_V
_
H_H
m
_
um___
ED
__ J"
u_ U_H
Mfl_ Dm__n_hI_____n______ __u__H_ _m_ _
Bd_ _;
____
_ _ _ _ Nw__*En__ __ _ _ m_r _ _ _ _ _fi_*_ _ _
E__ __ _ _ "_~ _ _ _H_ _1:5m
l
H__ _ _ _
m_ _"_ _ _ _
____
U" E
m_
“__ H1M
__H3_ Fwfl
NU _ _ _ _ _ _ _
Em
mm
Q
gm ____
H__ _n_
E T H_ E h
________t _____
H d_ ___ __-fl_ 5 m H
M m H 2 fl_ 4
1
_-
39
B5 TEDBIZD14
BRITISH STANDARD
Table 12
Fatigue based acceptance levels fbr embedded ncm-planar imperfectibns in butt
welds -‘"
Required
class
Maximum length crf
sc-lid inclusien. mm
Maximum ‘iii
prejected area bf
As-
eeresinr an
Stress
welded
' relieved by
Individual pare
diameter. mm it
I'==|"J'Eli'iP|1 "" '3
I FWHT
I
E
1.5
Di“
5"]
3
'EI.25t, 5 3 mm
maximum
5 2-4
We E is
4
"1 Lse
51
_ ‘lil
-35 -
letter
-
-r r -il;]
E csr
'
I crest. E mm
-
I"-lb limit
- fl-:1 limit
-- -
i maximum
..
-
-
t.nu1 .n'l.iu|.iu
ti Applicable tc: imperfectibns which are nbt lbcated within 5 mm bf the surface er an
adiacent imperfectien iclr within a distance equal tcl the maximum permitted length.
whichever is the smaller]'1‘ Fur assessing pcircssitv, the area cif radicigraph used shcruld be the length cxf the weld
affected bv purcrsltv multiplied bv the maximum width bf weld.
'5‘ Any pbrcisitv level which cibstructs an ultrasenic inspectien sheuld be rejected.
‘"1 Univ measured by radicigraphv.
El Accurate measurement clil"ficult by MDT.
Table 13
Fatigue based acceptance levels fbr undercut in transversely stressed welds “l “l
_ Depth crf undercut cl
Required class
i
_
_
Butt welds
Fillet welds
hlcrt permitted
—
lIl.t]25t, £1 mm
llilfit -:1 mm
I[l.il‘it, 5 l].5 mm
lJ_iJ25t -F11 rnrn
r
i
i
Ii
i
-|1|-nur-|
a.cn'si, s 1 rnrn
e.as_t_, s 1 mm
F2
El.ll]t, S 'l mm
lll.Cl?5t, 5 1 mm
E
lLl.ll}t.s'l mm
Ill-.ll3lt.-c l mm
*i Applicable unlv td undercut with a clearlv visible reet radius. Sharp urdercut shbuld net be accepted as the depth
cannat be measured and can ccrnceal craclcs c:|r laclc bf sidewall fusitan.
E“ Applicable fur 10 s t s Al] mm; undercut in thinner car thicker material sheuld be assessed as a planar flaw in
accerdance with E5 I-'El‘llJ.
‘ii Llndercut in centinueus welds stressed in the lengitudinal directibn crnlv dees nbt aftect the fatigue lite cn‘ a jeint.
Therefcire, there are nci limits tn their sixes tram the fitness-fcir-purpcise vievvpciint-
Al]
e
El The British Standards lnstitutibn 2014
BRITISH STAN DAFID
Table 14
B5 ?EDB:2D14
Effect pf misalignment cm the fatigue strength cit transverse butt welded jeints
class
km fl
cerrespending linear misalignment. eit ‘I
D
E
F
F2
G
1
1.14
1.34
1.52
1.34
ill
El-I35
B-11
D.1?
El-23
ea
_ 2-1?
u-as
""" ism is the stress magnificatidn facter due td any farm df misalignment (linear, angular
nr a cnmbinati-en] that can be assumed tn be included if the butt weld is designed cm
the basis bf the neminal applied stress range in cenjunctibn with the 5-l'l.l curve fer the
indicated class"l The cdrrespcinding values far the extent cit angular misalignment depend an tn-d many
factcirs tci tabulate them {see 5.5.2.11}-
Table 15
Effect ef misalignment en the fatigue strength crf cruciferm welded jeints
class
i Full-penetratien butt welds
' Effective
cerrespending
km fl
linear
misalignment.
I By; er
F
F2
Ll]
1-12
ii-U5
li-lllti
G
1.35
ii-1?
E3 3 .- 1-E3
Fillet pr partial-penetratien welds
Effective
cerrespending linear
lrm fi
misalignment. eft f"
F
-—
i-l]
1-2i'.'r
D-D5
i112
-- -.9-ii. -3- . -.3 _._- i-43 - ..- *5‘-if-'
‘ii Effective values bf km allew fer the fact that the neminal stress class refers tb jeints in
which eit can be up tcr lJ.tl'3. Hewever, they refer tb the extent bf any farm bf
misalignment ilinear, angular ur a cci-ml:|-inaticin} that can be assumed tcr be included it
the crucifcirm jeint is designed cm the basis cif the neminal applied stress range used in
cenjunctinn with the s-iii curve fer the indicated class.
“l The cerrespending values fbr the extent nf angular misalignment depend cm tub many
facters td tabulate them isee B-5-2-fl.
Table 1E-
Effect bf misalignment an the fatigue strength cl-f transverse butt cir crucifcrrm
welded jeints being assessed in terms bf hcrt~sp-bt stress
glass
____fr_m "ti
D
E
F
F2
G
1
1.14
1-34
1-52
1-E4
ea
2.1?
"__l:i;ll'l"E.'!'!|2'|l‘.1lI1-Elll'llH linear misalignment, aft '1
El
l].fi5
l]-11
i]-1?
El-23
_ e-as
“’ irm is the stress magnificatlen facter due te any term ef misalignment ilinear. angular
er a cembinatibni that can be assumed ta be included if the weld jeint is designed an
the ha-sis df the het-spet stress range calculated assuming perfect alignment used in
ccrniuncticin with the 5-hi curve fcir the indicated class.
Fl The cerrespending values fcir the extent crf angular misalignment depend cm ten many
facters ta tabulate them isee fl-5-2.tl-
E The British Standards lnstittrtic-n 2014
l
41
BRITISH STANDARD
BS TBBBIZB14
15 Stress calculatiens
1s.1
General
The prbcedure fer the fatigue analysis fer welded structures, including flame-cut
edges, is based en the assumptibn that it is enly necessary te censider ranges ef
cyclic stress in determining the fatigue life. i-e- mean stresses are net talren inte
acceunt- Hewever, fer nen-welded details subjected te stress cycles which are
partly er whelly in cempressien the effective stress range might be lewer
[see t5.4l.
The stress used tu assess the structural details included in Table 1 ta Table iii
depends an the class and the mede ef petentiai fatigue craclcing censidered. In
mest cases the latter is by direct {er nermai] stress fatigue failure under craclc
epening mede I isee Annex [Ill and the fatigue assessments are perfermed using
either neminal er het-spet structural principal stress ranges. The cerrespending
appreaches are referred te as the neminal stress-based isee 15.5} er hetspet
stress-based fatigue assessment methed isee 15.2}. Hewever. classes S1 and 5,
refer tcs petentiai shear fatigue failure under craclt cr-pening medes ll crr lll
isee Annex D]. In such cases the stress used in the fatigue assessment depends
en the lecatien bf cracking, the ratie ef applied shear te direct stress and
whether cembined applied shear and direct stresses act in-phase [i-e- principal
stress directiens remain censtant during lead cycling} er net. as summarized in
Table 1?‘. As indicated, in scrrne cases a direct stresebased class might be
apprepriate. Further details are previded in 15.2. 15.3 and 15.4.
Table ii’
stresses usecl in fatigue assessments invelving applied shear stresses
Ease
Fatigue craclting mede
Fa rent metal
failure Efrem
Shear rnedes ll er Ill
weld tee er
Nermal mede I
Weld threat
s,
cembined in-phase
direct and shear
Di‘ GI, B5
stresses
apprepriate
cembined medes I and
cembined
ll c1r lll
but-crf-phase direct
5, but
endurances
and shear stresses
halved
Pure shear
5'2
Shear medes ll er Ill
.
f-lermal mede I
leaded fillet
er partial
penetratien
weld}
Pure shear
class
Stress
'
hleminal shear
stress range
belt surface}
failure lfrcim
rciet uf
directly
E Leading
cembined in-phase
direct and shear
D. E. F. F2, G
W1
Maximum neminal
er het-spet
structural principal
stress range
I“-ieminai shear
stress range
Heminal resultant
stress ra nge
stresses with rlne'ib.3
cembined medes l and
cembined in-phase
ll ur lll
direct and shear
5-r
stresses with n'e:-{I-3
Ecimbined mcides I and
cembined
ll er ill
crut-crf-phase direct
and shear stresses
Si. but
endurances
halved
Heminal stresses sheuld be used with neminal stress s-rr curves. The neminal
stress sheuld include the effects ef lucal stress cencentratiens if the detail being
assessed is net fully represented by the descriptiun and figures in Table 1 td
Table iii. Alternatively, in assessments ef petentiai failure frem a weld tee,
het-spet stresses sheuld be used isee 15.2]. Stress ranges fer pest-weld
heat-treated jeints sheuld be assumed te be the same as fer as-welded jeints
lbut see ‘le.3l-
42
l
El The British Standards lnstitutibn 2014
BRITISH STANDARD
15.2
BS TEBBIIDIA
Stress range in parent material
In mest cases the pcitential fatigue cracic is lecated in parent material adjacent
tu seme fcirm cif stress cencentratien, e.g. at a weld tee er belt hele. Frci-vided
that the directien ef the principal stress dees net change significantly in the
cuurse ef a stress cycle {c 2i]"}, the relevant cyclic stress fer fatigue assessment
sheuld be talcen as the maximum range threugh which any principal stress
passes in the parent metal adjacent te the petentiai crack lecatien, as shewn in
Figure 2al. Tensien stresses are censidered pesitive and cempressien stresses
negative. In practice. the threugh-thickness cempenent ef stress is rarely
relevant and may usually be lgnured.
MUTE Fer tubular ncrdai jeints see Annex rs-
ln estimating the maximum principal stress. shear stresses less than 15% ef a
ceexistent direct stress may be neglected. in the case ef pure shear. the relevant
stress is the neminal shear stress range en the sectien pf petentiai fatigue
failure, as shewn fur applied tcirsien ef a tube in Figure 2bl.
Figure 2
Fteference stress in parent metal
4----""_"l
ii‘,/""
if/1
I
F ii
i
2
-
P-P‘
iifri
"— i
H H
m
. :
:|
F,
r —'*
at
3
KL
H
al Applied direct stress
llley
1
Welded attachment
3
Fetential cracic lecatien
2
Neminal stress = iF'r'.-4. + MEI]
4
Stress cllstributien
I A
_r’
j
\_
_./"T ‘Tas-
_
/
\.
gr
1
l
F
2
bl Applied shear stress
itey
1
Ndminal stress = Tfriir
2
Fetential craclt lecatien
E The British Standards lnstitutibn 21314
ll
43
BS TBBBIZB14
BRITISH STANDARD
if the directien ef principal stress changes during the stress cycle. e.g. as a result
ef a phase difference between the twe fluctuating lead seurces. the cyclic stress
range sheuld be derived frem the principal stresses calculated at the twe
extremes, l.e. the pealc and treugh. ef the cembined leading cycle.
The pealc and treugh values ef principal stress sheuld be these en principal
planes which are net mere than 45“ apart. Therefere, if cl“, rs}, and r are the
ceexistent values iwith apprepriate signsl ef the twe erthegcinal direct stresses
and the shear stresses at the peint under censideratien. the relevant principal
stress sheuld be selected if either:
a)
cr, -s tr, is at least deuble the cerrespending shear stress r at beth pealc and
treugh; er
bl
the signs ef rs; — rig and r beth reverse er beth remain the same at the pealc
and the treugh-
In either all er bi. erevided that rr,1 e cl,’ at beth pealc and treugh. the required
stress range is the algebraic difference between the numerically greater pealc
principal stress and the numerically greater treugh principal stress.
i.-'v"here cycling is ef such a cemplex nature that it is net clear which twe lead
cenditiens weuld result in the greatest value ef principal stress range. they
sheuld be established by calculating principal stresses fer all pairs ef lead
cenditiens- Alternatively, it may be assumed that the required stress range is the
difference between the algebraically greatest and smallest principal stresses
eccurrlng during the whcile leading cycie regardless ef their dlrectibns.
The classificatien fer welded jeints Table 3 te Table 1lJ depends en the directien
ef leading and the classificatien scheme is generally arranged en the basis that
the welded jeint under censideratien is either parallel {lengitudinal jeints} er
nermai [transverse jeints} te the stress directien indicated. Applied stresses that
are net acting in either ef these directiens might be effectively less damaging
than assumed. but there might be the rlslr ef fatigue failure frem a different
lecatien where the classificatien might be lewer- Fer example. the class ef a type
3.2 detail dreps frem E te E er F with respect te any stress acting nermai te the
weld tee. Therefcire. fer stress directiens that are greater than =15‘ frem that
shewn in the figures the welded jeint sheuld be assessed with respect te every
petentiai fatigue cracic initiatien site. The assessment sheuld be based en the
maximum principal stress range if it acts within 145* ef nermai te the petentiai
fatigue cracic while fer greater angles it can be based en the nermai stress range
acting en the plane ef petentiai fatigue craclting censidered. as determined
frem the principal stress rangesLlnless etherwise stated in Table 1 te Table itl. the stress sheuld be based en the
net sectien. Where apprepriate. ailewance sheuld be made fer geemetricai
stress cencentratiens isee 15.5.4].
15.3
5tress range in fillet welds
in lead—carrying partial penetratien er flllet—weided jeints. where craclring ceuid
eccur in the weld threat. the relevant reference stress 5,,_, is the resultant stress
range in the weld metal. Frevided that the directien ef the principal stress dc:-es
net change significantly in the ceurse ef a stress cycle ic2lJ“}. this sheuld be
talcen. fer each stress cycle. as the vecter difference between the greatest and
least vecter sum ef the applied nerrnal and shear stresses, based upen the
effective dimensiens ef the weld threat. as detailed in Figure 3- It sheuld be
assumed that nene ef the lead is carried in bearing between parent materials.
The tin value that influences the cheice ef detail class isee Table 1?} is the ratie
lengitudinal shear stresslenglneering shear stress er rtryr new in Figure 3.
Hewever. the lengitudinal shear stress range er., may be neglected if
.et,.,}.ecr,,, E lIl.1S.
44
1
El The British Standards lnstitutien 2lJ14
BRITISH STANDARD
BS TEBBIIBIA
Where stress cycling is due te rnere than ene lead seurce but the dlrectiens ef
the stresses remain fixed. the resultant stress S“, is based en the maximum ranges
ef the leads {e.g. PM and P1 in Figure 3} en the weld. Hewever, if the directien ef
the stress vecter en the weld threat changes by mere than 213" during a cycle
between twe extreme lead cenditiens. the resultant stress range is the
magnitude ef the vecter difference between the bvve stress vecters. Where
cycling is ef such a cemplex nature that it is net clear which twe lead cenditiens
weuld result in the greatest value ef 5,, the vecter difference sheuld be feund
fer all pairs ef extreme lead cenditiens- Alternatively. it may be assumed that:
fll
Stir‘ = {IIF-I-,.I1'i.a.i' _ 4'7.-_min}2 +IT.max T f1|'l1rlfi)I+II,rJi'.|'i|.E|I _ Tainan}: IE5
Figure 3
Fteference stress en weld threat
lpi \‘
1
bf
l
l
e
H
1
.
.4...
l
F,
I
-2,"...
l
E
-
g_h';,
.
-
Rey
1
Transverse direct stress
.1
NDTE
Ifw
PM
PH. E 1' flrf
vvh
wbfrb
= —-- + »---——-
Alse linevvn as engineering shear stress range en weld threat
2
P
Lengitudinai shear stress av. = —i—
vvh
3
4
Resultant stress range 5... = lerrf.,.+erfl"-5
w = cembined siee ef effective weld threats
ts.4
Effective stress range fer details in unwelded members in
which the whele er part ef the stress is cempressive
In unwelded details where the stress range. allewing fer dead lead and residual
stresses {due te fabricatien]. is entirely cempressive, the effects ef fatigue
leading may be ignered. Hewever, when the resultant stress range invelves
stress reversals threugh aere. the effective stress range te be used in the fatigue
assessment sheuld be ebtalned by adding Efillrir ef the range frem ;-:ere stress te
maximum cempressive stress te that part ef the range frem zere stress te
maximum tensile stress-
El The British Standards lnstittrtien 21314
ll
4S
BRITISH STANDARD
BS TBBBIZB14
1S.S
calculatien ef stresses
Stresses sheuld be calculated using elastic theery and talcing acceunt ef all axial.
bending and shearing stresses eccurring under the design leading. llie
redistributien ef leads er stresses, [e.g. as might be allewed fer checlcing static
strength at ultimate limit state. including implicit ailewance fer redistributien in
simplified elastic design rules. er fer plastic design precedures}. sheuld be made.
1S.-E
15.5.1
Calculatien ef neminal stresses
General
The neminal stress. 5,... at any lecatien is the summatien ef the membrane and
bending stresses at that lecatien. Examples ef neminal stresses in parent
material and a fillet weld threat are shewn in Figure 2 and Figure 3 respectively.
'lS.E.-I
Effects te be ignered
The effects ef the fellewing sheuld net be included in neminal stress
calculatiens:
al
residual stresses in welded details [see 1'.i.l3.3fl];
bl stress cencentratiens due tez
'lS.E.3
1}
a weld detail itself;
2}
belt. rivet er small drilled heles {but excluding types 1.3 and 1.4 in
Table 1}-
Effects te be included
All stresses arising frem seurces ether than these specifically excluded in ‘IS-5.2
sheuld be included in the analysis. The fellewing stresses sheuld be included
tegether with any ethers that influence the stress applied te the jeint:
al
stress cencentratiens due te the everall jeint geemetry isee 15.5.4}:
bl eccentricities er misalignment eccurring in a jeint detail, except where
etherwise indicated in this cede;
cl
shear lag. restrained tersien and distertien. transverse stresses and flange
curvature;
dl stress distributien in wide plates:
15.5.4
el
stresses in triangulated slreletai structures due te lead applicatiens away
frem jeints. member eccentricities at jeints and rigidity ef jeints;
fl
residual stresses ibut enly fer nen-welded details er stress relieved welds
under. neminally. fully er partly cempressive leading; see 15.4 and 15.3.5};
and
gl
fabricatien telerances. except as cevered in 5.2.2 fer misalignment.
Eeemetrical stress cencentratiens
Unless etherwise indicated in Table 1 te Table iii the stress cencentratiens
inherent in the malte-up ef the structural detail cencerned le.g. arising frem the
general geemetry ef a welded jeint and the weld shape} have been talren inte
acceunt in the classificatien ef the details. Hewever, where there is an additienal
geemetricai discentinuity, such as an aperture er a change ef cress sectien {see
Figure 4 and Figure S}. which is net a natural characteristic ef the standard
detail categery itself, the resulting stress cencentratien relevant te fatigue
design sheuld be determined either by specific analysis er, where apprepriate. by
the use ef predefined stress cencentratien facters [4, 5. E]. such as these given in
Figure ti.
45
l
El The British Standards lnstitutien 21314
BRITISH STANDARD
BS TEI3B:2l314~
Figure 3 shews a typical example ef a member which ceuid be analysed using a
stress cencentratien facter, such as given in Figure filal. and the resulting stress
used in cenjunctien with the neminal stress-based assessment methed.
Figure 4
Typical example ef stress cencentratiens due te geemetricai discentinuity
1
s.
2
i
|
I-5
I
I
i
I‘.
|
1
|
3
I
-L
qt llllllllllllllll!
"- Isl‘-
_- II-
‘I: 1-
I- _
llllllfi Eli
6
4
5
ice
Fetential craclr. lecatiens
welded attachment
Typical stress distributien
The design stress is applied te the apprepriate plain material classificatien
l. "l. :i-l.I 'l"-.l—-=
lillanhele er re-entrant cerner. The design stress fer lecatien 4 er E sheuld be talten as the
stress en the net sectien multiplied by the fatigue stress cencentratien facter (= lc,.S,]l
l':i
At the attachment the design stress is applied te the apprepriate neminal stress jeint
classificatien
In centrast, Figure 5 shews an example ef a welded jeint which weuld resluire
special analysis because the geemetricai layeut ef the jeint ebvieusly creates a
hard spet and hence a nen-uniferm stress distributien. in general ferm the jeint
is type 2.4.
El The British Standards lnstlttrtien 21314
ll
43'
BS 25133221314
Figure 5
BRITISH STANDARD
Typical example ef stress cencentratien caused by a geemetrlc hard spet
_-n|I""
ar
/1
iifl
2
A
ax
Q77
I-it
gan_'_it
—l'_*l'
L
_
'
—|-I‘
—I
if-."
i-in
'-
.--I
i:
in
in
-in in
—|—
in
inin
in
in
id-|.l_
A-|
imin
in
ii
i—-I
inin
in
id
Z.1.11-r
$--r
-1:
ltey
5, at weld tee taking inte acceunt scr due te weld tee
SI at weld tee [by extrapelatien] = Design stress = ieffectivelyjl 141.3, {neminal}
SI {neminal}
t.|Jf'|-. l|—I
4B
l
El The British Standards lnstitutien 21314
BRITISH STANDARD
Figure ii
BS TE13B:2D14~
Fatigue stress cencentratien facters
ft
3.15
1
it,‘
3-it
H
iii. ‘L 5. %.
3.13 Illt1.‘_'i\c __ 15
I"~
I
3.2
2.5
fir
I
as
,
1 r.
1-B
s
-—
___
___
4.1
__,
4-2
"Ir.
_
E__
I
I
_
'--I
,-,
cs
-I
Sis
"I
.-r Ll
,.
lit ___
E
2.13
El
B
"-
I
E
S“-..
u
2.2
‘I-5
E
'
I"
___
r
r’
1|
B
Q
I
2’its
Ir--_
4-4
“I
I”
r
s.
11.5
53
NDTE
lit’, values are censervative when E c 13.12
all Fatigue stress cencentratien facter fer unreinferced apertures it’. {based en net stress at ill-iii
lfey
1
Free edge
2-ti
2
5tress fluctuatien
um
it
2-4
2.2
3
r/
iii1.ll1ll4$1
ms
v|
I 2
E
1.4
I-_ ‘Q-'|-5
1.2
1 ti
K_r;_|
'l
‘S
1.-B
it
1
2-4
iffy ‘I-B
I r
ll
ll--2
ll-.4
'll.1l
B-5
Sr
‘iii
1-tl
fir
bl Fatigue stress cencentratien facter fer resentrant cerners if. [based en net stress at it-ill}
itey
1
Length ef straight a 2r
2
5tress fluctuatien
Figure B-11] illustrates a case where flexibility preduces a nen-uniferm stress
distributien. in all cases the cerrespending applied stress fer design purpeses
sheuld be the structural stress at the weld tee. Either typical examples are
tubular nedal jeints isee Annex El and the jeints shewn in Figure B-3.
Figure B-5. Figure 5.5. Flgure B-3 and Figure B-5.
Fer such cases and these fer which standard stress cencentratien facters are net
applicable. the alternative het-spet stress-based fatigue assessment methed
sheuld be adepted with stresses determined in accerdance with 15.3.
il The British Standards lnstitutien 21314
ll
45'
BS 25133221314
BRITISH STANDARD
15.?
15.1.1
Ealculatien ef het-spet stresses
General
The het-spet stress 5,. is the structural stress at the weld tee. This might vary
aleng a weld tee. as illustrated in Figure T. in which case the largest value is
generally used fer design.
Three metheds fer calculating the het-spet stress are described in Annex E while
Annex 43 prevides rules fer the specific case ef tubular jeints-
Figure If
cemparisen ef neminal. structural and het-spet stresses in a beam with a welded cever
plate
4
Er
T
lte
Structural stress. defined ever flange thiclcness. 1
5
Ferce. P
Neminal stress = l'v'l.l'Z+F.fA
E
fvlement. fill
Het-spet stress
if
sectien medulus, 2 and area. A
Flange thiclcness. t
.P| 1. r.il"-. l—e
1s.;'..'t
Effects te be ignered
The effects ef the fellewing sheuld net be included in het-spet stress
calculatiens:
al
residual stresses in welded details [fer exceptiens. see 1S.2.3f}];
bl stress cencentratien clue te the weld tee-
1.s.?.3
Effects te be included
All stresses arising frem seurces ether than these specifically excluded in 15.3.2
sheuld be included in the analysis. The fellewing stresses sheuld be included
tegether with any ethers that influence the stress applied te the jeint:
al
stress cencentratiens due te the everall jeint geemetry;
bl
eccentricities er misalignment eccurring in a jeint detail, except where
etherwise indicated in this cede;
S13
l
El The British Standards lnstitutien 21314
BS T1':i0B:2014~
BRITISH STANDARD
cl
shear lag. restrained tersien and distertien. transverse stresses and flange
curvature:
d}
stress distributien in wide plates;
e}
stresses in triangulated slceletal structures due te lead applicatiens away
frem jeints. member eccentricities at jeints and rigidity ef jeints;
fl
residual stresses in stress-relieved welded jeints in which the whele er part
ef the applied stress cempenent is cempressive. see 15.4 and 15.3.5:
gl fabricatien telerances.
‘fS.B
Axial stresses in belts
This sub-clause applies enly te ISD metric threads [B5 3543-2, B5 EN I50 4014.
BS El‘-l 1517.3 401?, B5 Ehl ISD 4252]. It is applicable te hexagen belts, screws and
nuts {E15 35512] high strength frictien grip belts lBS 4335-2] and stainless steel
belts IE5 EH I513 3505]. Fer blaclc belts isee I35 4150i it sheuld enly be used fer
fluctuating leads if the belt head underside has a machined face. This sub-clause
is net applicable te welded belts.
The stress range sheuld be calculated en the tensile stress area. BS 3532. Table
10. ef the belt and sheuld include the effects ef axial and bending leads.
including any effect ef prying. It sheuld talce inte acceunt the pre-lead in the
belt and the cempressibility and specified fit-up ef the cennected parts. Where
the fatigue design ef a jeint relies en the pre~lead ef the belts te limit the stress
range in the belts. the pre-lead sheuld be at least 1.5 times the calculated
applied belt leadThe cempressibility effect in a belted jeint is determined frem the relative
stiffness ef the belt and the clamped cempenents.
The pertien ef the applied lead that is carried by the belt can be determined
frem the relative stiffness ef the belt and clamped cempenentsFigure ti illustrates the relative stiffness effects en the fluctuating lead en the
beltWhere the applied lead includes beth a permanent tensien cempenent and a
fluctuating tensien cempenent, the tetal applied lead sheuld be assumed te be
a fluctuating lead unless a full jeint calculatien is carried eut which acceunts fer
beth the permanent and fluctuating cempenents ef applied leadNDTE l
Permissible stresses in bells are cevered in 15.2.2.
NDTE 2
l.r'.'-Ill .2231‘-'.l:2ti03 I;-'jl gives derailed instructiens en the calculatien precedures
fer belted jeints.
El The British Standards lnstitutien 2014
ll
S1
BS 3"Bi3B:213‘l 4
Figure El
BRITISH STANDARD
lielative stiffness effects en the fluctuating lead in a belt in a cencentrlcally clamped and
cencentrlcally leaded belted jeint
A
B
If I
C
T i.
|:|_
-------isTi5------
—"Lc
_'i.
F‘
M
-..
-is
Fr
i=
T it'll:
,
Fe
I-
.l 4" l.
__le__
ltey
en:
3., =
displacement ef belt due te belt
tightening
displacement ef jeint interface due te
F,=
fluctuating lead in belt
FL, =
prepertien ef Fa that relieves jeint
belt tightening
=
=
=
_}|cf,'_ |'1 ==
l_'l n_'If
lead applied te jeint
belt pre-lead
clamping effect
Fm =
=
clamping actien at jeint interface
Eiispacement
1
Belt
Belt fluctuating lead.
=
=
1*--r;.I__:=.:I'I
residual clamping lead
residual bearing lead under belt head
displacement due te F,
Lead
clamped plates
Ilia
F. = —vr..
Illfis-fffcl
where: lib = belt stiffness and it‘, = summatien ef clamped cempenents stiffness.
A
initial state
E
Leaded state
1s.s
E
Assembled state
Derivatien ef stress spectrum
in situatiens in which the leading spectrum is net specified. e.g. in a relevant
applicatien standard. the expected spectrum sheuld be derived.
S2
4
El The British Standards lnstitutien 2014
BRITISH STANDARD
BS 2513522014
In general the tetal leading en a structure ls cempesed ef several different
leading events. each with different magnitudes. geemetricai arrangements and
frequencies ef eccurrence. Te derive the design spectrum fer any particular
detail. a mix ef leading events that is representative ef the leading te be
expected in a given time interval sheuld be established. it sheuld be applied te
the relevant influence linelsl in erder te ebtain the pattern ef stress fluctuatlens
te which the detail is te be subjected in that time interval. Acceunt sheuld be
talcen ef the pessibiiity ef twe er mere leading events eccurring simultaneeusly.
er fellewing each ether in particular erders. such that higher stress ranges might
be caused. In selecting lead cycles fer inclusien in the spectrum lew stress
ranges. belew S hli'mmf_. and eccasienal high stress ranges that centribute less
than 155 ef the tetal damage may be ignered [but alse see 15.1].
This pattern sheuld be brelren dewn inte a cenvenient spectrum ef cycles.
expressed in terms ef stress ranges 5,. and numbers ef applicatiens Hi. by a
suitable cycle ceunting methed {see Annex H]- The numbers ef cycles ceunted
sheuld be cembined with the apprepriate tetal numbers ef eccurrences ef the
varieus leading events in the design life ef the structure te cempile the everall
design spectrum.
ln mest cemputer pregrams used fer the derivatien ef cycle ceunts the results
are accumulated inte a relatively small number ef stress range intervals
(semetimes called bins]. In mest instances abeut 40 intervals is sufficient-
15
Allewable fatigue stresses
15.1
Tensile stress limitatiens
The precedures given in this British Standard fer deriving fatigue stresses sheuld
enly be deemed te be valid if the calculated maximum fibre stress en the net
area ef a member. remete frem geemetric stress cencentratiens and excluding
self-regulating stresses {such as residual er thermal stressesl. dees net exceed
50‘!-ii ef yield stress under nermai eperating cenditiens and B055 ef yield stress
under extreme leading cenditiens. in this centext the expected number ef cycles
exceeding nermai eperating cenditiens in the anticipated life ef the structure
sheuld be net greater than 100.
15.2
15.2.1
$-lil curves
Plain material and welded. riveted er belted jeints
Fer each design class the relatienship between the applied stress range. 5,, and
the number ef cycles te failure. N. under censtant amplitude leading cenditiens
is ef the fellewing ferm:
legl'v'=leg i'.',,-dSD-mlegS,
i2l
The values ef these terms fer jeints in air are shewn in Table 1B. and the mean
line relatienships are pletted in Figure El. They are applicable te all steels
cevered by this British Standard, using the equatien:
leg Cd = leg CD — c:iSD
l3l
Then Equatien 2 can be written as:
5,"“lll' = Ed
ll-Ill
E The British Standards lnstitutien 2014
ll
S3
BS 2505:2014
BRITISH STANDARD
Thus. frem Equatien 2 te Equatien 4. the required basic 5-N curve can be
derived fer any desired value ef cl‘. The neminal prebability ef failure
cerrespending te varieus pessible values ef d. based upen an assumed nermai
distributien, is shewn in Table 151. The standard basic 5,-N cunres sheuld be talcen
te represent twe standard deviatiens belew the mean lines i-e- cf = 2- They are
shewn in graphical ferm in Figure 10, and the cerrespending values ef [2 are
included in Table 1B. if a curve fer a welded jeint cresses the class B curve fer
plain steel, the class B curve geverns isee Figure llici and 15.51.
All the basic 5,-ill curves are applicable. fer the relevant value ef cl, te details in
air er ether nen-cerresive envirenments. er fer details with adequate cerresien
pretectien in the ferm ef paint er ether ceating.
Fer design purpeses. hewever. the cunres might have te be medified in erder te
allew fer the facters given in 15.3 and 15.4-
Figure 5
Mean .5,-ill curves
1I
S00
are
1511
100
30
Bil
Tu
AR
Sil-
..
5|‘
’
"'1
/Ii?
/
_
-
I’/
1B
E
—_I'./1 .1 1?
Q5;._-,2
'3ls. .1_.,.*'
-|
\—
‘Ii
II
I-.
IMVII mfg-m
lg,
:55?
I
I
_!!E!
20
Ill‘
I
1°‘
2315-
W!
2
S11-S
W‘
2
311-S
my
I
3-ES
W‘
it
These curves sheuld net be used fer calculatien purpeses [see Table 1B}.
ltey
it Endurance I'll. cycles
al Mean 5,-N cunres fer direct stress failure
S4
I
El The British Standards lnstitutien 2014
"r'
Stress range 5,. lilimmf
BRITISH STANDARD
Figure 5
BS 2505:2014
Mean 5,-iii curves
,|,s11_
1.1111
ill
1'
r
e
Ill
if
rs-11
'l'-.il
I-00
‘P0
B0
‘ill
—
-
1-allIl gspe
BIB '
sr
40
34
211
_ ,,_
QI/:'III
~
Infill
2!
111-S
W5
'--will‘
-‘
I
112-I I
23115
1°‘
2
345
*-
-“F
2'
3115:.-I“.
These curves sheuld net be used fer calculatien purpeses {see Table 1Bl.
ltey
it
Endurance l'v', cycles
bl Mean 5,~l'il cunres fer shear stress failure
If
F
-5.:
400
-
..
‘r
5hear stress range. 5,. lilfmmf
.-
"'
Ellllll 2-llll
15-ll
1-1111
ss
an
111
5-ll
Sill
fill
-.
lllll Il l l II
IIIII
iiaigr
311 -
i
——- "--
flltéfll I Z
24 '
“jug
I
Ill
mi
II-ES
If
23115
mg.
2'
111-Sim‘
These curves sheuld net be used fer calculatien purpeses isee Table 1Bl.
ltey
it Endurance N. cycles
‘r’
Het-spet stress range. 5H,. llllmmf
cl Mean 5,-—hl curves fer use with het—spet stress
1l'.l The British Standards lnstitutien 2014
I
SS
BS 2505:2014
Figure 10
BRITISH STANDARD
standard basic design 5,-ill curves
Sllll
_""
If 4110
.-J.1-l
'
'
I
"'-.
Er
I-I|I' IE Ill
1511
1411
311'
lltl
‘I'll50
S0
l
A
illi-
2'll-
14‘
'
"I-f
I-f
WI”- 1'
EmsBflhl
lb
"
23115
11f
"rs
l-I
l
W1’lfilfi1%1555“
-ll-0
10
1
1
1
hi-|.
‘E
P
-
ll-c
I
q
I
I-l'1
F
IHF
‘I
2311-S
1-if
231-5
'I|I|-_-T:
I|I|"1_
111
These curves sheuld net be used fer calculatien purpeses {see Table 1B}
1
2345-
H 111"
-I
Adjustments sheuld be made. where apprepriate in accerdance with 15.3 and 15.4.
ltey
it Endurance N. cycles
‘r’ 5tress range 5,. Hlmmf
al Standard basic design 5,-l'1l cunres lmean minus twe standard deviatiens ef leg fl] fer direct stress
failure
StillIIII iifl-ll
300
Iif
l ill’
5...
2411
154
In.
-
ll-1:
1"
11
1‘-
I
_—
‘I00
-
in
flll-
_.
-:l-E
l,_:
3'0
50
sri
IE,
.1rll l
:
--1"-
+-
I1
411'
=
rt‘
1
l
30
rs
-.-...
111'
14"
II
'|.I
ll-
1.-.
hill
11f‘
—|-I
lil-
lull
.
-I'- 1-Iii
'
r..
til‘
I
ssts
'
..
._.i......
ill
,
.
it
... ._. f..-.-Li
211-S
E 111‘
These cunres sheuld net be used fer calculatien purpeses {see Table 15}
ltey
.11
Endurance iii, cycles
"1"
Shear stress range 5,. llilmmf
bl Standard basic design 5, -.-‘ii cunres [mean minus twe standard deviatiens ef leg lill fer shear stress
failure
S5
I 1'0 The British Standards lnstitutien 2014
BRITISH STANDARD
Figure 10
BS 2505:2014
standard basic design 5,-iii curves
S-lilll ._
-W
If 1.1111 -
\
-/’
_-Ia
31111
200
tsn -
~
1-lI'lI
till
!!!
Bill
"ill
511
Sil-
III
.1 _
[I fill?
p1:.|iSI'“——-
ii-Ll-il
I
1103“ -
..-.
.-.
I—I
1
I'I
20
Illl
ill"
2311-'5
'
is‘
231'-5
11:‘
231-ll
rrf
2111-'5
x iii‘
These curves sheuld net be used fer calculatien purpeses {see Table 1Bl.
Adjustments sheuld be made, where apprepriate in accerdance with 15.3 and 15.4.
lt Endurance ill. cycles
"r' Het-spet stress range. 5,.,. flrrnmf
cl standard basic design 5,-ill curves {mean minus twe standard deviatiens ef leg lill fer use with
het-spet stress
15-1-1
Axially leaded belts
The mean and design {mean — 2.513} S-I'll curves fer axially leaded belts up te
2S mm diameter {class it} are shewn in Figure 11- These are expressed in terms
ef the stress range en the tensile stress area. calculated in accerdance with 15.5.
Fer belts with diameter :.-2S mm. the fatigue strength defined by the class it
curve sheuld be reduced in accerdance with 15.3.2. in additien. the fellewing
cenditiens apply.
al
Class it is directly applicable te belts with cut er greund threads and te
relied threads that have then been heat treated.
bj
if thread relling is carried eut after any heat treatment the fatigue strength
defined by class it may be increased by 25%-
cl
If the belt is electre plated the fatigue strength defined by class it sheuld be
reduced by 2054.
il The British Standards lnstitutien 2014
I
S2
BS 2505:2014
Figure 11
BRITISH STANDARD
5,-iii curves fer belts with threads under direct leading tclass lltl
S00
T i1-00
V 'Iq‘!,- .- , ,
-4ifasII
.3e iiiE
"~
200
11-
-
-11-
I-
200
150-
2
1.! "-
I-.4
-1:: e-en1:-e
l.l'lTl-I-JI'|fl=
-ll-ii
‘I
311
20
I'll
1-r
234!-
re
E
E
.
.
l
r
1
‘- LI-
2311-1'1
1-if
23115
111‘
III!-
These curves sheuld net be used fer calculatien purpeses isee Table 15}‘E3’
r-.1-3-4.
SB
Endurance ill. cycles
'1'
stress range 5,. 1'-l.-lmmf
Class ill. mean
3
Eenstant amplitude
class it. design
4
variable amplitude
I Ill The British Standards lnstitutien 2014
it
111‘
BRITISH STANDARD
Table 15
class
BS 2505:2014
Details ef basic 5-l'1l cunres
11,,
Leg...c,,
m
standard
c.
deviatien ef
leg.,.,f-1. 50
2.34-3 it 1015
1.052 x 10“
3.5-B5 it 10"!
3.255 ls '0'"
-1m-.:1r' 1:|=
1.225 it 313"
F2
G
152
W1
J4
1.231 it '12“
5.15515 ls '12"
3.502 11'. '13"
2.5130 ill 10"
5.255 it 10'"
5-
5.502 x 10"‘
5:.-
3.5451-1 10“
15.3552
14.0342
12.50132
12.5155
12.2320
12.0500
11.2525
11.5515
11-3525
11.5554
15.2210
15.5555
4.0
3.5
3.0
3.0
3.0
3.0
3-0
3.0
3-0
3.0
5.0
5.0
0.1521
0-2041
0.2055
0.2505
0.2153
0.2225‘
0-1253
0-1552
0.2140
0.2134
0.2350
0.3500
3.0
0.2330
5.250 x 10" 12.542
Fer example the Tl curve is leg l"-.l = 12.542 - 0.233d -— 3leg 5,
Tl
1.01
4-23
1.52
1-04
5.33
4.32
2.50
1.45
5.33
3.51
2.00
x
x
x
x
x
x
x
x
x
x
x
I015
1013
10"
10"‘
10"
10"
I0"
10"
10"’
10“
10"
5.55 x 1015
3.02 x 10”
s,,,r1-1=1e=
s,,.,r1-1-=s--111'
cycles}
ill.-‘mm’
cycles} 1'-l-‘mm’
100
25
S3
42
40
35
25
25
21
33
52
50
31
25
23
21
12
14
12
15
45 lat 15"
45 lat 10“
cycles} El
cycles} "1
32 lat 10“
cycles} fl
32 lat 10“
52 fl
cycles} fl
35 fl
‘ll idealized het-spet stress-
fl Curve extrapelated te f-l=l0f cycles witheut slepe change.
ll.ltIlTE The held figures are re be ralren as exact and definitive; figures that are net held are calculated exactly
frem the exact definitive figures and then re-uhelecl te the displayed number ef significant figures.
15.3
15.3.1
Medificatiens te basic S-hi curves
General
In erder te derive the design 5,-N curves. the basic S,—l"1l curves sheuld be
medified in accerdance with Clause 5 and. as apprepriate. te allew fer the
facters given in 15.3.2 te 15-3.5.
15.3.2
Effect ef material thickness and bending
in the case ef petentiai fatigue failure frem the tee er end ef a weld er frem a
thread reet in a belt. the fatigue strength is te seme extent dependent en
material thicltness. istrength decreases with increasing thiclrness}. The basic S,—ltl
curves relate te the fellewing thiclrnesses and belt diameters:
al tubular nedal jeints lciass Til: up te ta = 15 mm:
b}
neh-nedal jeints {classes 5 te G2}: up te ta = 25 mm:
cl
belts lciass it}: up te 1,, = 25 mm diameter.
Similarly. in the case ef petentiai fatigue failure frem either the tee er end ef a
weld threugh the thiclcness ef the leaded member er frem the reet threugh the
threat ef a transversely-leaded fillet er butt weld. the fatigue strength depends
en the degree ef threugh—thicl-zness bending. The fatigue strength increases with
increasing bending cempenent fer a decreasing stress range gradient threugh
the thickness. Hewever: the design 5,-N curves relate te applied leading
cenditiens that preduce predeminantly membrane stresses- The petentially
detrimental effect ef increased thicltness but beneficial effect frem applied
bending are cembined by the applicatien ef the cerrectien facter lc,_., en stress
ranges ebtalned frem the relevant S,-hi curve. such that:
E The British Standards lnstitutien 2014
I
S5
BS 2505:2014
BRITISH STANDARD
5 = fits-5a
1'52
where,
fer t 5- 25 mm:
1-...=(r...= 1....-jl'**[1 +c1. 1er:r1-'1]
rs}
fer-4mmsts2Smm:
1--...=[1-.e11{{r,..-"tit-1}]-[1+e.1ee1-1]
rh
and
b
is 0.25 er 0.2 as indicated in Table 4 te Table 10 fer as-welded jeints
er 15.3.5 fer impreved weld tees er 0.25 fer belts:
t,,,.,.
is the greater ef tg er the actual thiclcness t in mm ef the member er
belt diameter under censideratien. Hewever, if in a neminal
stress-based assessment the weld detail can be described in terms ef
the dimensiens L and t isee Figure 1} and Lft s 2. then t,,,., is the
greater ef 0-5L er ti.Fer het-spet stress-based assessments {see 15.2}, ta,-1. = tFer cases ef increasing threugh-thiclchess stress range gradient, s1 sheuld be
assumed te be sere-
lf the bending stress cempenent is due te misalignment. fl sheuld be assumed te
be aere.
Fer assessments ef tubular nede jeints. ti sheuld be assumed te be aere as the
class already includes ailewance fer bending.
NDTE The increase in fatigue strength fer thicknesses less than 25 mm resulting
frem the use ef Equatien 2 enly arises in the presence ef bending.
Table 15
Iilerhinal prebability facters
I-iemihal prebability ef failure
d
‘lt50
31
15
2.3
0-14
‘ll Mean-line curve.
'“ The standard design curve.
50
I
El The British Standards lnstitutien 2014
0 fl
0-5
1-0
2.0 ‘fl
3.0
B m TE H
m
____T
B5 ?EDBED
DAHD
E___u__
_“H“___:
_H_ _ _
___:
_:1.“
H_F_h _ _ __h_qfi
___“;5___“: H____H
____H__
_"_H
F
:2
“___;
_"H__m _Mg”
5
"_______"=
J
tE
___“
_LE
UnuJ_mE_
;“H
“___
_u___;
fl__mfl____
:"U%_nH“E_m:EM|"_“Dm_Efiu‘H“|;_"muwE“_fl_“ u_L"_H _“fl_m _ _
_nHED:
_E
EFm_canE
fl_fiHm_Uu"m_fi=____“;
h_IhnE
EmUfl_dh"_Em‘_" __m___?m_ _" _E _m_“m_E_u_____m
_ _ “_‘g_ E_ _mnE_ _En_E "W_ _fl ____F_"
t“ __
_E
HE
Mm
mfl_HE_E___E___
mm_E=__E
_HE
= w_m_uE‘"_dm“_wm_E
Hm_fi _Em_dE_um“_E_h _m? r_flU_m_* _H
E
mE:_ =_ _ _E_ E_ _ H_m_ E_ “_
_"___:_E
___‘
E
_ mE__E_fi_E_E_h__
m_H_ __Em_u_
_u _ fl_ _ _ _E_ w_ _ _ _ _ "mm____
1
:_"___;_
_H_"_ _ _____m
______m_
___ u_ _ "_ _ E:2
_ _ _P_ _ _ _____"_qm
5_I_ ___
Nuh“
N_“
m_u__“_ _win___“
Em
_
_ _Em_ _ “3“3“
_ _ _ _ _m_ _ =u_ _E_m_ "_ _ “ _ _ _ m _ “_ w_ E_
LE;
I
ufi
E
__N
m
_Hx__:
__
m____fi3__n_
HE___“
nu
H
EH6
:E
E
_ _=_ r_
J“
um E
5 _"___m E HE Q“ E
_j_“__x
:5I _
_“_m___u
1
____H___
i
E: _ EH
____EH
EH;in _
m“fi_H___w
in__u_
_
Eu"
_fi__
___
_mF_Hh_ _
___"__m] _u_ _ _
_m_ _ _m E;
F
E m_£_ _!
$___
_LE
E;
“___
Hm_
hfl EH_fi=3_
_JE
ME
fifluuq
u_nmm_E_u_EM_‘uI_ _Eu1M_’Em_H_Eh_ _ __
_: m_ _ _:
I
___-_fl
Hh_E_
M
mWEu___:____u_ _“_u_$_____u“_ _h_1:
______
E _ _ _=
____"_
EH3?
E
HE
M_ Tw_EH|_JHFm_E‘mTb_ mHu_w "_“‘ _MH‘h_ hE_ :_“u E_ _ _ _maH2m_ _ v_H “_ _ H_ _ m_ _ _ _ m_ _kY_ : _ m_ _ m_
m:_____
_W__ EH
__G _w_____
R __H
%fi _“_ _ _H ___: q__
___:___ Wm
mm H;
ad “___” ad ___; H; E H;
1
_H_D___
_ fi_ _ _ _"
“___:H
“Hi
:w__w___
____n
_____ I:
~_fi______
___:i :2
mmd
Hmé___ ___:
2:___
___:
H "___:
:5" ___:'_E_ _ _ _____“m____H___m EH
____“
“E
nhu“_ m_m_£
_ _3'_"mflhE_nmP‘_" _E _U §_“ m_ _
"EEK:
_fiH“_h
____D
_?_____“
UN
_=___m_:
U_“____“
_N“!u_ _mfl_E__“_fin_5_ ___"_uN_D“___:
_
H_U_ _ __
_E“__ E
m_ h_____
EH “H MN
m___ :____H;_u__an___
mm! _$
___“ _
E Q“
___: H;
HE ___“ _“___|fl
“___:
I
E:
mH_I_ _ fl____fl
En____
HEM“ EH_:K_“__
_____H__
Ed
“___
E E
inan mm
_ _H“_ _ “___:R___H___
Em
___u__H__
H '_mg.“
___H=_
I :H__"__m_
______
Hh
I
1
_"__é_
“M
m__W _“___M
QM E HR E
______H__
K_“___:__
*__“___E_knE___
_ _“_ ]_ flm_ E__H__
___h___
_u____ M_
___“______
m_H
QM qm
gm
_"_ _ _H_ _"___HNEH H___q___
_H_____i
fl___H____
_____u_____H ____"
:5?
up :2
_m__m
E___ _____
INN
H= _ _ rm
_=
EH “___:
M
mm
U D
mm mm
im
M
3 mi
E U flu
I?Mm E
Q
M _:
Q
_m E
_t h
H_ ___“ H
________t ___
H ___"a r __"5_ m___“
M mH 2D
-I1
4
-
E
I
B5 TEUBIZU14
15.3.3
BRITISH STANDARD
Effect cif temperature
There are nd restrictidns an the use df the fatigue design cunres far cempenents
er structures that eperate at sub-aerd temperatures, previded that the steel
threugh which a fatigue craclt might prepagate is shewn tn be sufficiently
tdugh tb ensure that fracture dpes nclt initiate frem a fatigue craclt.
With regard tn elevated temperature, the design curves are directly applicable
fbr temperatures up tn 15D ‘E. Fcir higher temperatures T, but belpw the creep
range fer the steel cencerned. the fatigue strength dbtained frem the 5-H curve
sheuld be multiplied by E, i2.i]Ei:-ilill‘-" far structural steels pr ET l'2.i1l-i1e‘ii15 far
austenitic steels. where E. is the elastic mddulus pf the steel in tllmmf at
temperature T.
15.3.11
Effect pf sea vvater
The fatigue strength pf steel welded jeints might be reduced in the presence pf
sea water, even if cathbdic prcitectibn is applied. Eurrnsicin prc-itecti-an in the
fcirm bf paint pr bther cdatings is effective; this shbuld remain intact thrnughput
the service life pf the jnint.
Eathddlc prdtectidn prdducing vciltages in the range -E51] mv td -1 1i]lIl rri‘v'
dfiers same resistance tp cerresien fatigue. but cinly in the highicycle regime
[transitien at censtant amplitude endurance pf apprcn-iimately 1D“ cycles.
depending an the class}. In particular. a penalty pf 2.5 en life, pr 2.tl in the case
pf tubular jeints. is applied befpre the translticin, after which the penalty
reduces until inaalr perfermance is achieved fbr I"-l 2 1|]? cycles, as illustrated in
Figure 13. The resulting design 5,-N curves are detailed in Table El].
Fdr unprntected jeints eitpc.-sed ta sea water the basic 5,-til curves sheuld be
reduced by a facter pf 3 an life. in the case cit class is fer plain steel, the part at
the ccirnpdnent pr structure cencerned is first cln-wn—graded td class E tn allew
fdr surface rdughening pr pitting frem cdrrdsidn {see Table 1. Table 3 and
Table 51. Far all classes the cerrectien relating ta the number pf small stress
cycles lfsee 15.11} is nbt applicable. Details pf the resulting design 5,—l'il curves are
included in Table 15 and an eitampie is included in Figure 14.
NOTE The reccintrnendatlcins in this sub-clause are based largely dn studies bf
cdrrdsipn fatigue under the cpnditldns e.=rperi'enced by dffshdre structures pr vessels
ciperatlng in the hlcirth Sea {wave lciadlng frequency in the range t1. 15 tn lJ.5H.".' and
sea water temperature in the range 5 tn Ed“ fl. The detrimental effect pf sea water;
with pr witheut cathpdi.-: prdsectlpn, ceuid tie even greater at i'tIi'l.-ti'El' frequencies and
higher temperatures.
Fur high strength steels {c.r.|. 2- TDD hllmmf] these penalties might nc-t be adequate
because pf their greater susceptibility tp craclting frem hydrbgen embrittlement
and the use pf such materials under cbrrbsibn fatigue ccinditipns shbuld be
apprdached with cautidn.
He guidance is previded an the cerresien fatigue strength pf austenitic dr
duplex stainless steels.
Hp guidance is currently available fpr cpnsidering cbrrbsibn fatigue failure by
shear [craclt dpening mbcies ll pr lll]i.
E2
l
El The British Standards lnstitutibn 20111
BRITISH STANDARD
Figure 12
BS TEIIIBIIDIA
lillpdificatipns made tp S,-N curves fer welded jeints in sea water
'|ll'
Eli-II
l1I""l'l"'|'
i.
1
I"1I'|1ll'|
l
11"'Ill1|"[
I‘
I
l'1l"I'l1"'
ll-l.'l'l1 —
—|
Elllll —-.__E
1
-"5-. "1-.
Eli-II -
"1-i.
II
'|'|.|_
3
"H.
15-II ——1il'll' .95
I1
lit
lili-
—-
2",-"'i
"1-i HE
I-I
"'I-i.
'-.
‘-.
"1-t
‘H.
"'I-
S
"'I-.
—I
“I-\__H‘
rf
5"
il
"-. 1--.
fl-
-I
"\p_\‘-
_"1.
i
i
i
i
i i i
D
i
ii
Ii
i
L
D
i
i
A
—
'--.
iii-
an — 5w
"-.__H
_
I
_
"-~.___fi|
l‘“~.
J
i
I
I
1-
IL!
I
5"“-.
illi-
1-H‘
‘L.-|'l
1l.'l
ill‘
il
""-nah
H
-P‘
§
H_
D‘
||—
F1
H1
qil
Tili-
.,. _
.r'i'
"1-._____:I
1I
“_
.E:_1L-.-.
€_
:ye;
ltey
i.-.:r-.i-5-t
Endurance ill, cycles
"i"
Stress range S,, lillmmi
ln air
-=1
Ebnstant amplitude
Freely cdrrcidlng in sea water
in sea water- with cathedlc Fl rcitectipn
5
variable amplitude
15.3.5
Effect pf weld tpe imprevement
Fpr welded ldints invplving pbtential fatigue cracltlng frcim the weld tde, the
fatigue strength can be increased by the appiicatlnn qf a weld the imprbvement
technique. Tetihniques based pn imprqvement pf the weld tpe gecimetry tn
reduce the lcical stress ccincentraticin by grinding pr re-melting qr pn the
intreductien pf beneficial cdrnpressive residual stress by peening are described in
Al'1l"lElIl F. Frevided that the chcisen technique is applied in accerdance with
Anne: F and acceunt is tal-zen pf petentiai alternative sites fur fatigue craclr
initiaticin, the imprcivement in fatigue strength specified in Annex F may be
assumed. In general this amdunts tn a 55% increase in fatigue strength at
hi = 1[li cycles and rdtatidn pf the S-hi curve tp a sinpe pf rn = 3.5, but the
benefit might be less under sbme lciading cpnditicins in the case bf the peening
techniques {e.g. after tensile pre-leading dr fcir ciperaticin under high tensile
mean stressl. Higher imprcivements might be achievable but these shduid be
ccinfirmed by special fatigue testing in accerdance with hnneit E.
In the assessment pf imprcived weld tcie with respect td pdtential fatigue failure
frcim the tpe the currecticins fpr thicltness and bending, Equatibn E and
Equatien 1", shciuld be applied assuming b = El..'ll'J fpr dressed weld tpes nr £1.25
fcir peened weld tbes.
NOTE 1 The benefit cif dressing may be claimed cinly fdr welded jeints which are
adequately prqtected frem ccirrcisibn, as the presence pf a cdrrdsiire envlrcinrrrent can
cause pitting in the dressed regidn.
NOTE 2 in the case pf partial penetration and fillet welds, where failure can eccur
frpm the weld rpcit, imprevement cif the weld tde cannpt be relied updn tci give an
increase in fatigue strength.
El The British Standards lnstittrtipn 2514
l
53
BS ?5BB:Il.B14
BRITISH STANDARD
15.3.5
Effect cif stress relief fdr welded details
lf the applied stress range is fully tensile. stress relief has l"lD significant effect an
fatigue strength. Hewever. if it can be derndnstrated that a cempressive
cempenent c-f the cembined applied and residual stress exists, and can be
quantified, it may be assumed that the relevant value qf S, is the sum pf the
tensile cempenent and l5lJ'il'li c:-f the cempressive cempenent.
NOTE 1' As this unliliely id be satisfied very ciften this benefit is generally ignered
isee 'lS.E.3,l.
in assessing the pqtential influence bf stress relief, acceunt sheuld be taken cif
the fcillciwing:
ai
the extent td which stress relief actually reduces the residual stress in large
cemplex jeints is largely unltnewn; and
bl
if jeints are lcically stress-relieved and then welded intci a cdmplete structure
lpng range tensile residual stresses still exist.
NOTE ii Such icing range stresses can be as high, and have the same effect. as the
residual stresses induced by welding.
15.4
Treatment pf lew stress cycles
Under fluctuating cci-nstant amplitude stresses there is a stress range, which
varies bath with the envirciriment ancl with the size pf any initial flaws, belew
which an indefinitely large number pf cycles can be sustained. This is referred tci
in this British Standard as the nan-prdpagating stress range but a cqmmcinly
used alternative is CAFL. In air and sea water with adequate prptectici-n against
cbrrbsibn, it is assumed that this nqn-prbpagating stress range, Sm, {given the
suffix "dc" td denpte censtant amplitude leading], is the stress cerrespending tcr
ill = 1l]' cycles tn failure liar 1iIl"i' cycles in the case cif classes 5, and S2} as
ebtalned frem the design S-ill curve [relevant values qf Sm fbr the standard basic
curves are shewn in Table 1B and Table Elli- In all cases Sm is the stress derived
after allewing fpr any required mndificatipns tci the S-ill curves lsee 15.31. Far
unprcitectecl jeints in sea water it sheuld be assumed that sg, = 1.1.
When the applied fluctuating stress has varying amplitude, sq that sqme ci-f the
stress ranges are greater than and seme less than SM, the larger stresses cause
grdwth -pf the flaw, thereby reducing the value bf the nbn-prqpagating stress
range belew SM. Therefcire. as time gcies qn, an increasing number cif stress
ranges belew sg can themselves centribute tn cracl: grdwth. The final result is
an earlier fatigue failure than weuld be calculated by assuming that all stress
ranges belew SD, are ineffective. Therefc:-re, when cqnsidering variable amplitude
leading acceunt sheuld be talten pf the fatigue damage due tp stress ranges
belew Sm, but exceeding 5 hlimmf isee 15.51}, using a medified cunie. This is
prdduced by assuming that the S-ill curve is extrapelated beybnd ill = ll]? cycles
until it has a change qf inverse slcipe frem m tn lm + 2} at lllm, l,see Figure 1-ill,
where NM = 5 x 1l]i cycles fcir classes E-l tq W1 and ls‘. cir 151“ cycles fcir classes 5,
and 51. The stress range at this slepe transitien peint is 5,,,,, suffix "av" deneting
variable amplitude leading: values are included in Table lls and Table Eli. This
prbcedure is equivalent tp assuming that the number pf repetitiens pf each
stress range 5, less than SD, is reduced in the prepertien l5,iS,,,,l-ii.
NDTE l
water.
This cerrectien tides net apply in the case pf unprc.-tected jciints in sea
NDTE .1‘ The fellewing might assist in calculatiens.‘
fer 5, a Sm.
n
H5?’
n
( 5, T"
hi _ cl ' sintii s,,.,
E4
1
El The British Standards lnstitutibn 25111
BRITISH STANDARD
BS TEBBIIBIA
fr-iJr S, 5 Sm,
n
i'1SI""+fI"
n
S. “"“'f
.f'll_ cysg, _sir1eiI_s,,,)
NOTE Si This appreach might be tee censervative fer seme stress spectra. including
these that censlst predeminantly ef stress ranges belew SM with enly eccasienal
stress ranges abeve it. A different precedure ceuid be justified en the basis ef
specific tests in accerdance with Annex E. Alternatively; the damaging effect ef stress
ranges belew Sm can be calculated using fracture mechanics isee Annex Oi.
Figure 13
Typical S,-ill relatienship
1
T
.
.
‘
2
__/_________
Sec
II I lIl
Ear
.
II I I IIII
...__H_4
I I I II
1..-ul
7%
it
I
T
-I -'I-—l-I -I -I -I —|-I -I -I L
Hat
NOTE
Hill‘
x
Only the pertien ef this figure shewn as a full line is based en experimental data.
ICE
I‘-.l-—I-3'5
Endurance N, cycles — leg scale
"r'
Stress range 5,, leg scale
Static limitatiens
3
Effective curve used fer calculatiens
censtant amplitude leading in clean air
Ill‘-'l1fI‘"'l"*il ‘-iiltlilllllt i"_1'1PIl‘il-lilit Iilillllflll.
er, with details pretected against
'l'1"il'-l""'ilIEl'lt til‘ '5hf"iBl"B 5|l7'PE ‘ill 5?”
cerresien, in sea water
F‘-""""E *'='h“""E New
Llnpretected details in sea water.
=1
15.5
Treatment ef high stress cycles
Where Iecal bending er ether structural stress cencentrating features are
invelved, and the relevant stress range includes ailewance fer the stress
cencentratien [including by the use ef the het-spet stress}, the class B te G2, W1
and Tl design S,-N curves fer welded jeints may be extrapelated baclr. linearly
len the basis ef leg 5,, versus leg Ni up te a limit ef a stress range equal te twice
the material neminal tensile yield stress flag].
Hewever. fer jeints in a reglen ef simple membrane stress the design 5,-ill curves
sheuld enly be extrapelated baclt te a stress range given by twice the applicable
tensile stress limitatien given in 'lE.1.
ill-5 all the S-N cunres fer welded jeints cress the class B curve fer plain steel in
the high stressllew endurance regime te give higher fatigue lives fer a given
stress range. the class B, er class E fer unpretected jeints in a cerresive
envirenment, curve sheuld be used if it indicates a lewer fatigue strength than
the welded jeint curve.
Fer the class S, and S, cunres, extrapelatien may be made baclt as fer the ether
design classes but up te a limit ef stress range defined by half the values given
in this sub-clause {l.e. with reference te shear instead ef tensile stress}.
El The British Standards lnstitutien 2514
l
E-S
BS TBBBIIIB14
BRITISH STANDARD
15.5
leints subjected te a single stress range
Fer a jeint subjected te a number ef repetitiens cif a single stress range l[i.e.
censtant amplitude leading], the range sheuld be net greater than that defined
by Equatien It {see 15.1.1} and Table 1B er Table 25 fer the relevant
envirenment, jeint class, number ef cycles and required prebability ef failure.
Apart frem the case ef unpretected jeints in sea water, if the stress range is
lewer than 5,,, isee ‘lti.ill it may be assumed that fatigue failure will net eccur.
This might alse be the case fer seme variable amplitude leading isee 1B.1'}.
15.?
Jeints subjected te a stress spectrum
Fer a jeint subjected te a number ef repetitiens n, ef each ef several stress
ranges, S,,, [l.e. variable amplitude leading} the value ef n, cerrespending te
each 5,, sheuld be determined frem standard leading rules {if applicable}, frem
stress spectra measured en a similar structural member, er by malting reasenable
assumptiens as te the expected service histery. as apprepriate isee Clause 1‘ and
t5.Bl. The number ef cycles te failure, N, at each stress range. S,,. sheuld then
be determined frem the basic S-ill curve, medified as necessary in accerdance
with 15.3, fer the relevant jeint class at the selected prebability ef failure. The
fatigue damage sum O due te the stress spectrum te be endured fer the
required life ef the jeint is then calculated using f-.iliner's linear cumulative
damage rule. as fellews:
n
n
n
nO=—I'+-j+—3-+..etc=F,—I
llli llls -‘Ills
I'll.
(Bl
If all values ef 5,, are less than Sn, it may be assumed that fatigue failure will net
eccur. Experience le.g. [B.B]l indicates that this can be assumed te be the case
with a statienary narrew band variable amplitude leading spectrum, fer which
the histegram ef stress cycles {including all increases due te misalignment,
thicltness cerrectien etc.) fellews a Rayleigh distributien, if the stress range
cerrespending te B times the reet mean square cif the stress amplitude fer the
stress—time histery is ne greater than the initial nen-prepagating stress range Sm,
tsee 15.1}.
Hewever, if any values ef 5,, exceed 5D,, all the stress ranges, including these
belew Sm, sheuld be included in the summatien. This is because the higher
stresses in the spectrum are capable ef prepagating craclcs which might then be
prepagated further by the lewer stresses. Seme ef the lewer stresses in the
spectrum may be assumed te be less damaging than indicated by the basic S-N
curve {see 15.-ill.
In general, the required fatigue life is achieved if El s-.1 l.e. Fatigue testing
[e.g. [1B. 11, 12]} has cenfirmed that this assumptien is safe if the leading
preduces a spectrum with gradual variatien in stress range, netably narrew-band
randbm leading cenferming te a Rayleigh distributien ef pealts. Hewever, under
seme stress spectra, including these invelving fully-tensile stress cycling abeut a
high tensile mean stress er where there is little variatien in the maximum
applied tensile stress, tests have shewn that fatigue failure can eccur when D -c
l, typically when D - {1.5 but semetimes even lewer I1-ill. Therefere, if there is
any uncertainty abeut the nature ef the service stress spectrum, er fer
particularly critical cases. it is advisable te limit D te 5.5. Alternatively. El can be
established fer the particular stress spectrum and welded jeint type cencerned
by reference te relevant published data er by special testing {see Annex El.
55
1
O The British Standards lnstitutien 25111
BRITISH STANDARD
BS TEBBIIBIA
If the damage sum D ebtalned frem Equatien B is greater than the limiting
value selected fer design {e.g. 5.5}, measures that can be taken te impreve the
fatigue strength ef the jeint include strengthening the detail te reduce the
applied stresses, re-design ef the detail te a higher class er the applicatien ef a
weld tee imprevement technique (see Annex F}.
El The British Standards lnstitutien 2514
l
5T
BS ?5BB:Il.lIl14
BRITISH STANDARD
Annex A
[rierrnative]
A.1
Fatigue design
5-N relatienships
The S,-N relatienships fer the varieus structural detail classes have been based
en statistical analyses ef available experimental data ebtalned under tensile er
shear leading. The analyses invelved linear regressien analyses ef leg 5, and leg
N with the slepes ef the curves predetermined. in additibn seme miner empirical
adjustments were made te ensure cempatlbility ef results between the varieus
classes [1*3].
The change ef slepe in the curves frem m te lm + 1] {see Figure 111} is a
mathematical device te aveid difficulties in cumulative damage calculatiens
using ii.-1iner's rule. The bent 5-ill curve sheuld net be assumed te represent the
results that weuld be ebtalned in tests under censtant amplitude leading.
As far as welded jeints are cencerned, it has been shewn experimentally that,
when high tensile residual stresses are present, fatigue strength is a functien ef
stress range alene; mean stress and stress ratie have ne significant effect. in
general it ls impessible te predict what residual stresses might be present in any
particular structure and therefere the design rules are based upen the
assumptien that high tensile residual stresses are lil-tely te be present. Fer
simplicity the same assumptien has been made with regard te nen-welded
details. This is a reasenable assumptien te make fer beth welded tas-welded er
stress relieved} and nen-welded details if leng range tensile residual stresses are
intreduced, fer example as a result ef imperfect fit-up during subsequent
assembly invelving the parts cencerned.
Altheugh the fatigue test data used te determine the design S,-N relatienships
were ebtalned frem arc welded jeints it has been feund that they are alse
suitable fer similar jeints made with either pewer beam lelectren beam er laserl
er frictien welding [15].
A.I
Fatigue life fer varieus failure prebabilities
The standard basic S,-N cunres in Figure 1lI'l are based en twe standard deviatiens
belew the mean line assuming a leg nermai distributien, with a neminal
prebability ef failure ef 2.3‘l-ll. in certain cases, a higher prebability ef failure
ceuid be acceptable. fer example where fatigue craci-ting weuld net have serieus
censequences er where a cracic ceuid be easily lecated and repaired. In
situatiens where, fer example, there is ne structural redundancy er where the
jeint is uninspectable it might be desirable te design against a lewer prebability
ef failure.
The neminal prebabilities ef failure fer a lcnewn stress spectrum asseciated with
varieus numbers cif standard deviatiens belew the mean cunre are given in
Table ls. The Sr-N curves apprepriate te ether numbers ef standard deviatiens
belew the mean cunie can be derived frem Equatien it isee 15.1}.
NOTE The everall prebability ef failure during the design life ef a typical preduct is
substantially lew'er than the values in Table 19 lwhich are enly applicable te the
fatigue strength distributien} when the upper beund values pf leading are assumed
lsee clause iii.
EB
1
O The British Standards lnstitutien 25111
BRITISH STANDARD
A.3
A.3.1
BS TEBBIIBIA
Fatigue design philesephy
Safe life design
The precedures and design data previded in this British Standard are primarily
intended fer use in safe life design. The safe life design appreach is based en
the use ef standard lewer-beund fatigue endurance data and an upper beund
estimate ef the fatigue leading. it therefere prevides a censervative estimate ef
fatigue strength and dees net depend en in~service inspectien fer fatigue
damage.
A.3.1
Damage-telerant design
The damage-telerant design methed sheuld enly be used if the preduct, er the
relevant part ef it, can be safely and ecenemically inspected by apprepriate NOT,
and any craclts detected and repaired befere they reach a length that ceuid
cause failure under static leading. The fellewing sheuld be talten inte acceunt
and evaluated at the design stage when deciding te talce a damage-telerant
appreach:
al the strength ef the structure;
bl
the censequences ef failure; and
cl
the need fer inspectien and the feasibility ef repair.
NOTE Oamage-teierant design might be suitable fer applicatien where a safe life
assessment shews that fatigue has a significant effect en design ecenemy er where a
higher risl: ef fatigue cracking during the design life may be justified than is
permitted using safe life design principles.
Damage-telerant design sheuld ensure that when fatigue craclting eccurs in
service the remaining intact material can sustain the maximum werlcing lead
witheut failure until the damage is detected. Therefere, a prescribed inspectien
and maintenance pregramme fer detecting and cerrecting any fatigue sheuld be
put in place and feilewed threugheut the life
The fellewing design features sheuld be used te help achieve damage telerance:
1]
selectien ef materials and stress levels te previde lew rates ef cracic
prepagatien and leng critical cracic lengths;
2}
previsien ef multiple lead paths;
3} previsien et cracic-arresting details; and
it}
previsien ef readily inspectable details.
Damage telerance depends en the level ef inspectien te be applied te the
preduct and is net autematically ensured by replaceable cempenents. inspectien
sheuld be planned te ensure adequate detectien and menltering qf damage
and te allew repair er replacement ef cempenents. The fellewing facters sheuld
be talten inte acceunt:
ll
lecatien and mede ef failure;
ii)
remaining structural strength;
iill detectability and asseciated inspectien technique, which sheuld be based en
the largest flaw net liltely te be detected rather than the smallest it is
pessible te find;
iv] inspectien frequency;
vi
expected prepagatien rate allewing fer stress redistributien; and
vi] critical craclr. length befere repair er replacement is required.
El The British Standards lnstitutien 2514
1
59
BRITISH STANDARD
BS T5BB:1B1l'l-
The calculated fatigue lives enable critical parts te be ranked in terms ef fatigue
sensitivity. when in-service inspectien is an eptien. this can be used, in
cenjunctien with an assessment ef the censequences pf failure ef specific
members, te establish a prierity basis fer develeping a selective inspectien
pregramme tn be feilewed during the service life ef the preduct. Fer
nen-structurally redundant parts where the censequence ef failure is high, er fer
which in-service inspectien weuld be difficult te achieve, it is eften preferable te
re-design critical parts te a higher fatigue classificatien te reduce the level at
future inspectien required.
Annex B
lnermative]
B.1
Explanatery netes en detail classificatien
General
This annex gives bacltgreund infermatien en the detail classificatiens given in
Table 1 te Table 15. including the petentiai medes ef failure. impertant facters
influencing the class ef each detail type and seme guidance en selectien fer
dedgn.
B.1
B.Z.1
Hen-weldecl details {see Table ‘l and Table 2}
Fetential medes ef failure
in unwelded steel, fatigue craclts nermaily initiate at surface irregularities,
cerners ef the cress sectiens er heles and re-entrant cerners. in steel which is
cennected with rivets er belts. failure generally initiates at the edge ef the hele
and prepagates acress the net sectien. Hewever, in deuble cevered jeints made
with high strength frictien grip belts this mede ef failure is eliminated by the
pre-tensiening, previding jeint slip is aveided, and failure initiates en the surface
near the beundary ef the cempressien ring due te fretting under repeated
strain.
B.1!
B.1.I.‘l
General cemments
Classes A te C
ln welded censtructinn. fatigue failure rarely initiates in regiens ef unwelded
material as the fatigue strength ef the welded jeints is usually much lewer.
Class A requires special manufacturing precedures which generally render it
inapprepriate fer structural werlt. Hence assessment ef fatigue strength fer this
class is net included in this British Standard.
Classes B and C sheuld be applied with cautien in cases where the high class is
dependent en surface finish as this might degrade in service, fer example frem
cerresien er abrasien. Hewever, with due attentien te the detrimental features
mentiened. there might be scepe fer adepting a higher design curve than these
specified. Such alternatives sheuld be established in accerdance with Annex E.
B.1.2.2
Fit-up and pre~tensiening ef belted cennectiens
The specified fit-up -ef belted cennectiens sheuld be achieved in practice- If net,
the stress ranges applied te the belts might be much higher than these assumed
in design. which ceuid lead te premature failure.
Alter a greup ef belts has been tightened, the tergue en all the belts sheuld be
checlted. This is because it is pessible fer seme pre-tensien te be lest as later
enes are tightened, even if they were eriginally tightened using tensien er
tbrque centrel. If net, bending ceuid be intreduced with the result that stress
ranges applied te the belts weuld be higher than these assumed in design-
ill]
1
El The British Standards lnstitutien 25111
BRITISH STANDARD
B.2.I.3
B5 TEDBIIDI4
Fasteners
In threaded fasteners fatigue craclcs normally initiate at the root of the thread,
particularly at the first load-carrying thread in the joint. Alternatively, failure is
sometimes located immediately under the head of the bolt, particularly in bolts
with rolled threads and in joints subjected to bending loads.
B3
Continuous butt welds and welded attachments essentially
parallel to the direction of stress (see Table 3)
B.3.1
Dressing of butt welds {type 3.1}
Fatigue tests on butt welds with the weid overfill dressed flush have shown that
class C is a realistic design classification provided that the weld is free of fiaws.
However, this classification is reliant on the detection of surface and embedded
flaws that could reduce the fatigue strength below class E. This has been shown
to be possible E14] but it does require the provision of l'-IIIJT techniques and
operators capable of both detecting and sizing flaws, a requirement that is
generally beyond the scope of routine workmanship inspection. Guidance is
provided in Clause 14 and E5 i"I-‘rill. Acceptance levels for welding flaws to meet
class E are given in 14.3.4, or they can be determined using fracture mechanics
{see Anne: Di.
B.3.2
Potential modes of failure
with the reinforcement dressed flush ltype 3-1}. failure in longitudinal butt
welds tends to be associated with embedded flaws. In continuous butt welds or
in butt or fillet welded continuous attachments {types 3.2 and 3.3], away from
weld ends, fatigue craclcs normally initiate at ripples or lurnps at stoplstart
positions on the weld surface. However, in the case of discontinuous welds
ltypes 3.4 and 3.5] fatigue craclcs occur in the parent metal at the weld ends.
B-3-.3
B.3.3.1
General comments
Welds near plate edges [see Figure B.1}
Although welding attachments to plate edges can result in a very iow
classification ie.g. types 1'i.i' and i'.Tr'l there is no downgrading for welds close to
a plate edge or ones that accidentally overlap an edge. However, as with edge
attachment welds, in such cases the possibility of local stress concentrations
occurring at unwelded corners as a result of, for ercarnple, undercut, weld
spatter and eiccessive leg length at stop-start positions or accidental overweave
in manual welding should be limited. Similarly, the unwelded corners of, for
esrample, cover plates or bo:-r members [see Figure E.'la]i and Figure £-l.1bl] should
not be undercut. lf this does occur. it should subsequently be ground out to a
smooth profile.
Figure El.1
Welds at plate edges
‘l
ll-TE
1
l
Avoid or grind out to a smooth profile any undercut to these exposed corners
E The British Standards Institution 2014
l
T1
BRITISH STANDARD
B5 TEDBIZU14
Attachment of permanent baclcing strips
5.3.3.2
if a permanent baclcing strip is used in rnaicing a longitudinal butt welded joint
it should be continuous or made continuous by welding. Any welds in the
baclcing strip, and those attaching it, should also conform to the relevant class
requirements- The classification might reduce to class E or lower {type 5.3 or 5-4}
at any transverse butt welds in the baclcing strip that have not been dressed
flush or nominal stress class E at any permanent tacit weld {see type 3.Ei.
Tacit welds
H.333
Tacl: welds. unless carefully ground out or buried in a subsequent run. provide
potential crack locations similar to any other weld end. Their use in the
fabrication process should be strictly controlled.
Welded attachments on the surface or edge of a stressed
member [see Table 4]
H.41-
Potential modes of failure {see Figure 5.2]
B.A.1
When the weld is parallel to the direction of the applied stress fatigue craclcs
normally initiate at the weld ends, but when it is transverse to the direction of
stressing they usually initiate at the weld toe; for attachments involving a single
fillet weld, as opposed to a double, weld craclcs might also initiate at the weld
root. In each case the craclcs then propagate into the stressed member.
Figure B.2
Failure modes at weld ends and weld toes of welded attachments
l
r
-l-—I-
2
""
*1.
_
I Illlllrllllli In
.r
""
H.
lIIIl
-I-—-lh
s-._,_,_-r
It
5
For classification purposes, an "attachment" should be talten as the adjacent structural element connected by welding to the
stressed element under consideration. Apart from the particular dimensional requirements given for each type in Table A the
relative size of the "stressed element’ and the "attachment" influences the correction for thickness and bending isee 15-3.1].
lliey
Long attachment
Short attachments
Weld failure craclcs ltype, Tr'.B and I-'.1lJ}
Joint types 4.2 to 4.5 lor 4.? at edge}
.h|L|I. I' -.l—I
Joint types 4.1 and 11.2 lor 4.? at edgell
LIT
B.4.I
B.ill.I.1
General comments
STIESS EDFI CE l"lTl*EI'l"i'iJ I15
Stress concentrations are increased, and the fatigue strength or joint
classification is therefore reduced, where the following apply:
al
the weld ends or toes are on an unwelded corner of the element; and
bl the attachment is long in the direction of stressing and, as a result, transfer
of a part of the load in the element to and from the attachment is liicely to
occur through welds adjacent to its ends. This effect is further intensified
with thiclc attachments.
T2
l
El The lilritish Standards Institution Itlillll
BRITISH STANDARD
B-11.2.2
BS TBBBIIBIA
Weld forms
Full or partial penetration butt welded joints of T form, including cruciform
joints, [e.g. those that would connect attachments to the surface of a stressed
element} should be completed by fillet welds of sufficient leg length to produce
nominal weld toe angles no greater than Bill‘ {see Figure B-3}. The fillets exclude
the possibility of an increase in stress concentration arising at an acute
re-entrant angle between the element surface and the toe of the weld, and it is
therefore unimportant whether the attachment is fillet or butt welded to the
surface when assessing the effects on the stressed element, as a similar toe
profile results in both cases-
Figure B-3
Failure modes in cruciform and T-joints for joint types indicated
mi-
'lS
rs
sEEl*'
'l.5
q
IJ.
41
ii
B *—"
-L5
Er
75
Iii-I-P
NOTE T-joint shown with minimum weld angle to be achieved when adding
fillet weld to full or partial penetration butt welded joint.
B-S
B-S-1
Transverse butt welds [see Table 5 and Table El
Potential modes of failure (see Figure B-4}
With the ends of butt welds machined flush with the plate edges, fatigue craclcs
normally initiate at the weld toe and propagate into the parent metal, so that
the fatigue strength depends largely upon the toe profile of the weld. if the
reinforcement of a butt weld is dressed flush isee type S-1i, failure is more lilcely
to occur in the weld material from embedded flaws or from minor weld flaws
which become e:-tposed on the surface, e.g. surface porosity in the dressing area.
In the case of butt welds made on a permanent baclcing strip, fatigue craclcs
initiate at the weld metal baclcing strip junction and then propagate into the
weld metal-
Figure B-4
Failure modes in transverse butt welds for joint types indicated
5.2. S-3
5.2. 5.3
-e ngmmxuxe 1-e
E The British Standards Institution 2014
l
F3
BS TBBBIZB14
BRITISH STANDARD
B.S.Z
B.S.It.1
General
liiiisalignment
Butt welded joints between plates or tubes are susceptible to misalignment {see
E.liil and therefore transverse joints might experience secondary bending under
applied axial loading. No allowance is made for this in the classifications. For
elements where out-of-plane bending is resisted by contiguous construction,
e.g. beam flanges supported by webs, wide plates supported by effectively
continuous stiffeners, eccentricities due to axial misalignments in the thiclcness
direction may be neglected.
However. where such support is not provided. e.g. tension iinlis, the design stress
should include an allowance for the bending effects of any misalignment ll. l.e.
the nominal distance between the centres of thiclcness of the two abutting
components, e. For components tapered in thiclcness, the mid~plane of the
untapered section should be used. The nominal stress should be multiplied by
the following stress magnification factor;
—Ii
lr = 1 + GE
TI
'"
ti ti‘ rd
rs. ll
where
t,
is the thiclcness of the thinner plate
ti.
is the thiclcness of the thicicer plate
e
Thus. when t, = ta, the stress concentration factor becomes 1+3? For other cases. including angular misalignment, see B5 T910. For additional
solutions specific to girth welded joints in pipes, reference should be made to
Dl~lv~FlP-GIGS.
B.S.I-I
Element edges
Fatigue failures in butt welded plates tend to be associated with plate edges
and undercut at the weld toes on the corners of the cross section of the stressed
element for on the edge at the toes of any return welds} should be avoided. If ft
does occur, any undercut should be ground out to a smooth profile.
B.S.2.3
Part width welds
Butt type welds might also occur within the length of a member or individual
plate, e.g. in the case of:
al
a plug weld to fill a small hole;
bi
a weld closing a temporary access hole with an infill plate.
Aithough such geometries have not been given specific categories, the relevant
type in Table S or Table E. may be deemed to cover plug and infill plate welds.
ll This includes unintentional misalignment to the extent of the acceptance limit
specified by the manufacturer or fabrication standard being used. "lightening that
acceptance limit isee 1'4;-ll would be beneficial from the fatigue viewpointTld
Ir
El The British Standards Institution Itllllll
BRITISH STANDARD
B-S-2.4
BS TEBBIIDIA
Joints butt welded from one side only
Butt welds should be produced with full-penetration welds. The most reliable
way to achieve this in joints welded from one side is by the use of permanent
baclcing ltypes S-S and E-El. However, the design class is relatively low lciass Fl
and the alternative use of temporary non-fusible baclcing {types S3, Sal, ti-E and
E-Fl allows some improvement [15]. in this case, inspection of the weld root to
assess whether or not there is full penetration is feasible and should be carried
out to justify class E. ln the case of butt welds in pipes ltypes l:'i.E- and SJ} the
surfaces of the pipe and baclcing should be aligned to within 1:1 mm to avoid
the seepage of weld metal through the gap and the subsequent formation of a
cold lap on the inside of the pipe [ll-Ir]. in the absence of any baclcing {types S-3,
s.a, ii-E and G-Tl the main requirement is full penetration and to justify class E,
this should be assessed by inspection. if this is only possible from one side of the
joint, as in pipe girth welds, specialist MDT such as automated ultrasonic testing
lALlTl should be carried out, further guidance is given in BS 1'Ei1lZi- Experience
indicates that the techniques used are capable of detecting root flaws due to
incomplete penetration down to approximately if-I-5 mm in depth [liliB.S.I.S
Penetration of butt welds
Butt welds transmitting stress between plates, sections or built-up members
connected end~to*end should be ‘full penetration welds.
B.S.I.=E
Dressing of butt welded types S-1 and ii-Ii
The information given in 5.3.1 for longitudinal butt welds is equally applicable
to transverse welds, with the additional recommendation that the dressing is
sufficient to ensure that all the traces of the weld toe are removed.
B.S.3
Cruciform and T-joints between plates in the same plane [type 5-5]
Gften, the load is transmitted from a member to a transverse member primarily
via flange plates in the same plane isee Figure El-S}. This can occur in the case of
a junction between cross girders and main girders, diagonals and truss chords, or
in "v'ierendeel frames.
Figure B-S
T-junction of two flange plates
i
‘\
oo*/
o-‘T
l-
'3'
‘Ir
\ ‘*1
2
,,o""'*
/ If '\ 4*‘1
tr‘ \_5
t
llley
I
Transverse member
3
Type 4.?
2
Full penetration butt weld
ti
Type S-Ei lwith stress concentration
factor}
E The British Standards Institution ..'t[l'l1l
l
IFS
BS TBBBIZB14
BRITISH STANDARD
if a full penetration butt weld is used and the joint geometry is in accordance
with the requirements of type S-5. and if in addition the joint is either of the
cruciform type {see Figure El.E]i or if the transverse member is relatively stiff, i-e.
it-s width is at least three times the width of the stressed member, the nominal
stress-based classification should be as given for type S.-El with a stress
concentration factor of unity.
Figure lil.E
¢ ,
Eruclform junction between flange plates
/
/
Z’!
.
A
‘IIII:
$4‘
1
A
iii.
\e***‘*
Type S-G {with stress concentration
3
Typical stress pattern
factor}
2
Full penetration butt weld
Eltherwise the classification should be assumed to be that of type S-2 with the
appropriate stress concentration factor.
ln the case of trusses, secondary stresses due to joint fixity should be taicen into
account. The fatigue strength of both flange plates can be improved by the
insertion of a smoothly radiused gusset plate in the transverse member so that
all butt welds are further from re-entrant corners lsee Figure E.i'l.
Figure B.?
Alternative method for joining two flange plates
l‘
‘Rmno-,‘.‘i*’
Ill
e,‘-‘_
*5
is
-*5"
!'*~o
’—
,.--at
all \ll"'
T""'-r
ii
**’
\#***
Q
|
Sr
2
lcey
1
Types 1-3 or 1--'-I lwith stress
concentration factorl
if-IS
e
El The British Standards Institution Itllllll
2 Types S-2, S-S or E-1
BRITISH STANDARD
B-E
BS TBBBIIBIA
Transverse butt welds in sections (see Table G)
Butt welds between rolled or built up sections [type E-.1] are prone to weld flaws
in the region of the weblflange junction. The low classification reflects the fact
that these are very difficult to detect. The normal classifications for transverse
butt welds in Table S should only be adopted if special preparations, procedures
and inspection have been undertalcen that show the weld is free from
significant flaws.
The same types of joints are frequently made using cope holes to provide access
for malting the weld in the flange. The end of the web butt weld at the cope
hole {type 5.2] is equivalent to class D if the end of the butt weld, and the weld
reinforcement within a distance equal to the radius of the cope hole, are
ground flush isee Figure B.B]i. Gltherwise, class E is applicable {type ti.3]l.The
relevant stress should include allowance for the stress concentration effect of
the of the cope hole isee Dhl‘v-ltP-E2lZl3i- fvlitred cope holes of triangular shape
should not be usedFigure B-3
Local grinding adjacent to cope hole in type 5.2 joint
2
-|--l-l-
l
/1
if /
-ri
Y
/
lg
-rrr _.f~fi§'r-:-.:".-;:. -
icey
B-if
B-if-1
1
Distance = r
I
Grind
3
liadius r
Load-carrying fillet and T-butt joints [see Table 7)
Potential modes of failure {see Figure B-3]
Failure in cruciform or T—joints with full penetration welds normally initiates at
the weld toe, but in joints made with load-carrying fillet or partial penetration
butt welds craciting might initiate either at the weld toe and propagate into the
plate or at the weld root and propagate through the weld. In welds parallel to
the direction of the applied stress, however, weld failure is uncommon; craclcs
normally initiate at the weld end and propagate into the plate perpendicular to
the direction of applied stress. lilevertheless, provision is made for possible shear
failure through the weld throat {see type ‘.l'.1lfi}.
E The British Standards Institution ..'t[l11l
l
Til’
BS TBBBIZIII14
BRITISH STANDARD
B.'l'.2
Gruciform joints [types if-1, 7.2 and i".B}
Eruciform joints between plates are susceptible to misalignment
[see Figure B.Sl a]] and thus secondary bending stresses are lilcely to be
introduced under applied axial loading. However‘, the classifications allow for
the effects of any accidental axial or centreline misalignment up to the lesser of
El-GS times the thicltness of the thinner part or 2 mm. The secondary bending
stress due to any misalignment that exceeds these limits should be allowed for
in the calculation of the design stress. Equation B-1 is applicable for axial
misalignment but other formulae for calculating the relevant fr," for assessing
potential fatigue failure from the weld toe or weld root are given in BS Tull].
vvhere the third member is a plate it can be assumed that plane sections remain
plane in the main members and that the axial and bending stress distributions in
the 5, direction are unaffected. Where the third member is an open shape, for
example, an I section or a hollow tube, particularly if different in width, a
discontinuity in the main member stress pattern is iiltely to occur. in this case the
stress parameter should include the stress concentration at the joint, for
example by using the hot-spot stress- ln the absence of published data on a
particular joint configuration, the stress concentration factor for use with the
nominal stress might need to be determined by finite element or model analysis.
Plane sections may be assumed to remain plane where the main member stress
can be continued through the transverse member by additional continuity
plating of comparable cross-sectional area, which is ln line with the rnain
member components [see Figure B.El bl]. in this type of connection the joint
regions of the third member should be checiced before welding for lamellar
rolling flaws and after welding for lamellar tears.
TB
Ir
El The British Standards Institution Itllllll
BRITISH STANDARD
Figure ti-El
BS TBBBIIBIA
Use of continuity plating to reduce stress concentrations in type 1'-1 and 1‘-2
joints
—j1
2
al Cruciform joint between plates
icey
1 Axial misalignment or eccentricity. e
2
,_,
Q.
‘F.
Srw-int
2
,,
‘ill’
=‘ill’ril ere‘-re ‘
-rs‘
ti,-lfr‘
'-.'~e~.'e-"'~.
is
Through-wall direction
q_
"-il,I" §.|I "'rt" r-."'r-."-."'ri."'ri-
'1-[I I
1
fig
‘N
B-if-S
-I-in
*
W -iris
bl Gruciform joint between sections
icey
H
—
d
u
"I5
1
Typical stress pattern with plating
3 Eontinuity plating
2
Typical stress pattern without
plating
4
Eletail it show in Figure B-Ba}
T-joints [types 3'-3 and if-ill
T-joints [Figure B.1lIli are distinguished from cruciform joints by the absence of a
similar member in line on the far side of the joint. In this case an axial load in
member it would induce a bending moment and cunrature in member "r". Unless
member ‘i’ is very stiff in bending an uneven stress distribution results. Members
with bolted end connections via transversely welded end plates are particularly
susceptible to local increase of stress {see Figure B-10}. If the transverse member
is an open or hollow section, local bending increases the pealc stress further {as
in the case of types Ii’-1 and 1-2). These effects are included in the hot-spot stress
but appropriate stress concentration factors should be applied to the nominal
SITESS.
As far as fatigue failure of the transverse member 1*‘ is concerned. member it is
treated as an attachment {types A-1 to ~l'i.E,i and the stress parameter is the stress
in the transverse member without the application of a stress concentration
factor. In hollow or open transverse members this is often magnified by local
bending of the walls.
E The British Standards Institution -.'t[l11l
l
TB
BS TBBBIZIII14
BRITISH STANDARD
Figure E‘-.1El
Example of type T-3 or 1'-4 T-joint
I
L
l!..._..
5r""i"";.i"ll
l
Rey
1
B-14
i
L
Stress distribution due to flexibility of member "'r"
i-oad-carrying fillet welds failing in the weld [types 18. ?.1lIl and
9.5}
Provision is made for assessing load~carrying fillet welds with respect to
potential failure from applied direct and shear stresses. Depending on the
proportion of shear. either class W1 or class 5,, is applicable isee 15-3 and Table
Ii’). Apart from cases of pure shear, class W1 is the most lilcely relevant class in
practice- Glass W1 is primarily intended to apply to all fillet or partial
penetration butt weld joints where bending action across the throat does not
occur. Where lapped joints are welded on two or more sides, or T- or cruciform
joints are welded from both sides, such bending action is normally prevented.
However. if significant bending does occur and it can be quantified, it can be
included in the calculation of the resultant stress range and hence used in
conjunction with class 'v'v'1.
e.:r.s
T-joints failing by bending in the weld {type 1-'9]
in certain cases difficulty of access might only allow welding of T-joints to be
done on one side of the joint. This applies particularly to small hollow members
with welded corners which, if subject to loading that distorts the cross section.
might cause failure of the corner weld in bending {see Figure B-ll and type Tr'-EllWhere axial stress is also present, the stress range at the face of the weld might
be different from that at the root. Failure from ripples or stop-start positions on
the face can give a higher strength than class G, but expert advice should be
sought if a higher strength is required. in most cases, failure from stress
fluctuation in the root is critical and should be classified as class G with
application of the thicirness and bending correction ir,,,Figure B-11
Single fillet comer weld in bending {type I-"-Bl
I —'—'JI-q,
-seq,-_
'15-
I“-
"- —'-I "-I ".-r
I-. -\-I
.-.|-.—‘I
I-,-r-I‘-I-I. .
I
l
R lr|I_,|-'.vI'I§"
'I '.|-I-
I
icey
Bl]
e
1
Bending stress at root
2
Bending stress on weld face
El The British Standards Institution Itlfllll
3
Corner detail
BRITISH STANDARD
BS TBBBIIBIA
Slotted connections and penetrations through stressed
members {see Table Bl
Slotted connections exist where a narrow member is slotted through a single
main member away from an end connection {see also Figure B.11l- In this case,
the narrow member should be assumed to transmit the stress which the parent
material would have carried before the slot was cut. The part of the narrow
member projecting out of the plane of the stressed plate then becomes,
effectively, a welded attachment, so that the classification becomes the same as
for types 4.1 to 4.5. This detail should be avoided where possible, as slots are
difficult to cut accurately and fit-up for welding is often poor. Where member B
is called upon to carry high tensile stress, a slot in member A avoids any rislc
from lamellar tears.
However, with respect to stress fluctuation in member B, the detail shown in
Figure B12 is type 4-? {nominal stress class El at point ‘r’. if member B is critical
and member A is not, circular cut-outs at the corners of member is improve the
class to that of type 4-3 lnominal stress class Fl.
Particular attention is drawn to the maiclng of this type of joint with fillet welds
instead of full penetration welds. In that case, the joint becomes type Tr‘.-T.-1 and
type i‘.B-
Figure rs.t-11
Exarnple of a third member slotted through a main mernh-er
=-lSll
if‘
ill‘
ililiiifliiiiie
‘II
jil
-.-.l
rrrparinc
5t'*___“"
\I.-\"rI.*\.-1‘IrIrI\‘\\-"\\fl,\‘
_,_,,_
A
IIIIIII%I
Ia
"'\.-‘\-"i. "\.-‘L
IIII
\IeI. \ .-
A
rafl
'\-\-‘\-'\
i. l. I.|I_I l.
Iiiai
“T '-nr'""s
Iii‘!-1
- "?. "'9-
NFJTE i
GZ
—-
I-
"'Ir .-"\-'Ia .
3
I—
N~F“
--
L\-"'\-“|I-"\.-
-
i -
Dimensions are in rnm.
NGTE -1 llrlain member A is slotted while narrow member B is continuous and is
welded all around the slor in ANDTE -i‘
Classes refer to nominal stress+I:rased design," hot-spot stress class is El in both
EBSE5.
liley
1 Class F2 {type B-El
3 Gut out
2
4
Class F ftype B-11
Alternative detail if member "E" is critical
E The British Standards Institution -.'t[l11l
I
B1
BRITISH STANDARD
BS TBBBIZB14
B-S
B-5-‘I
Branch connections (see Table 1tl)
Potential modes of failure
There are four main sites for fatigue craciting in branch connections, the weld
toes in the shell and branch. the weld root, leading to cracicing through the
weld throat, and the crotch corner. In every case account should be talcen of the
stress concentration in the region of potential fatigue craciting due to the gross
structural discontinuity introduced by the nozzle.
B-S-2
Stress concentrations adjacent to branch connections
Four possible stress concentrations due to structural discontinuities in noaales
should be talcen into account when calculating 5,.
al
Crotch corner. The class El fatigue design curve is used, based on nominal
hoop stress range multiplied by it’, at the crotch corner where it’, is as
defined in 15--E-4.
bi
Weld toe in shell- The appropriate fatigue design curve is normally used
with nominal hoop stress range multiplied by ft’, at the weld toe. where ic, is
as defined in 15.5.4. Alternatively, the hot-spot stress isee 15.?) could be
used directly. The possibility of stresses arising in the shell as a result of
mechanical loading on the nozzle as well as pressure loading should be
talcen into account-
cl
Weld toe in branch- This region should be treated as in item b‘,i. The
possibility of mechanical as well as pressure loading should be talcen into
BlZiZIIl'l.-ll"llI-
d}
Wald root. The class Er or F fatigue design curve is used, based on the
nominal hoop stress range on the minimum transverse cross-section of the
weld multiplied by it’, at the weld root, where it’, is as defined in 15-5-4. The
possibility of mechanical as well as pressure loading should be talcen into
account-
NDTE See B-2.2 for potential failure from weld root under stresses acting normal to
weld length.
Annex C[normative]
C-1
Guidance on stress analysis
General
The stress to be used with the S,-hi curves in Figure ‘El and Figure ll} depends on
the choice of method. In all the cases included in Table 1 to Table ill one option
is the nominal stress range at the potential cracltlng site shown in the slcetches
in the tables. it ls the stress that would be calculated by conventional
engineering methods, which does not talce into account the effect of the local
shape of the detail or the weld on the stress field. This stress is easily calculated
in the case of simple axially loaded members or simple beams in bending. In
more complex details the stress should be calculated adjacent to the detail
which is analogous to the stress in the simple member. This means calculating
the stress which is developed by reason of the shape of the structure, but
without the perturbation in the stress field caused by the detailed shape of the
weld itself. This annex provides guidance lG-2 and G-ll on the determination of
the resulting increase in nominal stress by means of experimental or numerical
stress analysis or the use of appropriate stress concentration factors.
B2
e
El The British Standards Institution 2lIl11l
BRITISH STANDARD
BS TEBBIIBIA
In the case of details in which the site of potential fatigue cracking is the weld
toe {Table 3 to Table 1B} an alternative approach is to use the specified S,-N
curve. usually class D for plate structures or class TI for tubular joints, in
conjunction with the hot-spot stress range. This is intended to include any stress
concentration effect arising from the nature of the structure containing the
weld detail as well as that due to the weld detail itself, but to exclude the stress
concentration due to the weld toe, lvlore detail is given in f-l.e.
Two dimensional stress systems
An example of a simple two dimensional stress system is a butt weld in a plate
in a direction at right angles to the direction of stressing, such as a transverse
weld in the flange of a box girder. In this situation the nominal stress can be
calculated from simple bending theory. If a hole is made in the plate on the line
of the weld the stress at the ends of the transverse diameter of the hole is
magnified so that the weld at those points sustains a higher stress than at a
distance from the hole. Figure 4 shows a typical stress distribution in such a case;
further examples are shown in Figure B-5 to Figure B-ll]. it is this magnified
stress which should be used as the input to the S,-iii curve for that particular
detail. The effect on the fatigue life is quite marlced because life is proportional
to the third or the fifth power of the stress depending on the position on the
S,-N curvelvlore complex shapes can be dealt with in the same way using published
solutions [4, S, til for most of the regular shapes of cut-out; two common details
are shown in Figure E. For shapes not included in the standard texts, stresses can
be obtained from measured strains on a model or the full scale item. in such
cases, the load cycles applied should be of sufficient magnitude and quantity to
ensure that a shaiten down state is reached by the strain gauge.
The designs of the models and the positions at which strains are measured
should recognize the basis of the S,-N curves to ensure that the relevant stress is
used. This is more difficult in regions of steep stress gradients than in areas of
low stress gradient.
Alternatively. the stresses can be calculated using the finite element stress
analysis {FEA} method. This requires careful mesh generation to model the joint
so that the relevant location for stress computation is selected. Too coarse a
mesh in relation to the local stress gradient gives results which are insufficiently
accurate. Too fine a mesh piclcs up the influence of the weld shape, if that is
modelled, or the false influence of the idealized joint shape. Convergence
checlcs provide a guide to the optimum mesh size.
Some other methods of stress measurement and visualization exist but tend to
be more suited to experimental woric rather than the immediate needs of the
designer. These include techniques such as thermal imaging and Ill-rays.
Three dimensional stress systems
For most of the welded joint details shown in Table 3 to Table ill the stress
required to calculate the fatigue life is at the surface of the members. The stress
analysis methods described in C-2 give solutions where the stress field is two
dimensional or axisymmetric- in the case of more complex stress fields, such as
those produced by out~of~plane bending moments in plates {Figure B-15} or by
brace loads on the walls of hollow section chords {Figure G-ll, the standard
analytically based solutions are not sufficient. in these cases, three dimensional
methods such as photo-elasticity and FEA should be used for a fuii
understanding of the stress pattern. Strain gauges give the surface stresses but if
there is a need to itnow the stresses through the thicitness of the material, {e.g.
to separate the membrane and bending stress componentsi, they are not
suitable.
El The British Standards Institution 211111
I
B3
BRITISH STANDARD
BS ?BBB:2lIl'll'l-
Finite element methods can be used to calculate the stresses in all types of joint.
The selection of element types and sizes should reflect the geometry and the
stress gradient so that a sufficiently reliable estimate of the stress at the weld
detail is obtained. The modelling of the joint should reflect the overall geometry
and the stiffness without introducing the details of the weld profileC-ti
E-4.1
Calculation of hot-spot stress
introduction
The hot-spot stress is widely used for the fatigue design of tubular nodal joints.
The same approach offers many advantages for application more generally in
the fatigue design of welded joints between plates and other structural sections
when considering potential fatigue failure from a weld toe or end.
An example of the difference between the nominal and structural stress is
shown for the simple case of an l-beam with a welded cover plate in Figure E-1-
This detail is classified in ‘l2-2 as class G for assessment based on nominal
stresses- The class G design S-N curve was obtained from fatigue test results
obtained from similar beams with cover plates and so the stress concentration
effect of the cover plate welded to the flange is included in the S-N curve-
Therefore, the relevant nominal stress that should be used with the class G S-N
curve is that at the weld toe position based on the beam section, applied force
and bending moment at that location, excluding the stress concentration effect
of the cover plate, i-e. it is the structural stress at the weld toe that would be
calculated using standard formulae for the beam in the absence of the cover
plateFigure C-1
I beam with cover plate showing distribution of stnrctural stress and definition
of hot-s pot stress
l
E
FIIIB
II
i
Rey
Structural stress
Nominal stress
'l-I|.lI"r-. I-l
B-ll
e
Hot-spot stress
El The British Standards Institution 2lfl11l
Ifi
‘- xei-.
Iwnn Innis
'¥-‘P-
Inll tl lo
Imil wu
1IlI AIl l c
I
E
I
L
III
l uliurl
x ul il s IHI lI *
.-
IIIIII
k ||||||||||i|I|"""
'ILI lI h 'IJI lI l\
I‘
BRITISH STANDARD
BS TEBBIIBIA
In contrast, the actual structural stress at the weld toe. as would be obtained
from stress analysis of the complete beam with a cover plate, is higher due to
the stress concentrating effect of the attached cover plate and the extra stress
concentration where the weld crosses over the web of the beam. This can be
seen in the stress distributions across the width of the beam and approaching
the cover plate {Figure C-1].
The structural stress is useful for assessment of potential fatigue craclcing from a
weld toe or the end of a short or discontinuous longitudinal weld. in such cases,
the assessment is based on the maximum structural stress range at the weld toe,
which is also referred to as the hot-spot stress range rf5,,,lApplication of the hot-spot stress approach to tubular joints {see Annex GI was
aided by the provision of parametric formulae relating the hot-spot stress
concentration factor to tube form, dimensions, joint type and mode of loading
for a wide range of joint configurations. However, such formulae do not exist
for the much wider range of welded joints that need to be assessed in
non-tubular products. Therefore. it is necessary to determine hot-spot stresses by
detailed stress analysis, typically by FEA or from strain measurements. This annex
presents details of procedures for calculating hot-spot stresses that are
consistent with the design S~l'll curves in Clause 1'5.
E-4-2
Hot-spot stress calculation from surface stresses
As the through-thicltness stress distribution remote from a weld is lilcely to be
linear, the stresses on free surfaces remote from the weld are equal to the
structural stress. Therefore, the structural stress at the weld toe, the hot-spot
stress, can be estimated by extrapolation from the surface stresses at locations
near the weld. This technique is called surface stress extrapolation {SSE}. Stresses
are used directly to calculate the weld toe structural stressThe hot-spot stress is the maximum principal stress at the weld toe if its
direction is within :45‘ of the normal to the weld toe or. outside this range, the
larger of the minimum principal stress or the stress component acting normal to
the weld toe. Strictly spealcing, the extrapolation should be performed for each
stress component at the extrapolation point locations and the extrapolated
values at the weld toe used to calculate the corresponding principal stressesHowever, in practice it is sufficient to use either a principal stress or the stress
component acting normal to the weld toe, whichever is dominant in accordance
with the criteria in this sub-clause.
‘v'arious extrapolation methods have been proposed and investigated [lB],
dependent on the location of the weld toe {weld toe on plate surface or edge,
weld toes in tubular joints] and the nature of the stress analysis-
Where the stress distribution depends upon the plate thiclrness, type “a”
hot-spot regions in Flgure E-2, the most common SSE method ls linear
extrapolation from stresses on the surface at distances tun and 1.llt from the
weld toe. The corresponding hot-spot structural stress is given by:
E-I
SH = I-5?-l'll'|j'_||,g_|j _ I:I-EliI|t'.F'"|_[];|-
It
J
This was the method used to determine hot-spot stresses in the fatigue test
database used to validate the choice of class D as the hot-spot stress-based
design curve [19]. Alternatives include linear extrapolation from stresses ll.St and
‘i.St from the toe and quadratic extrapolation from three locations, fix-It, fl-9t
and 1--fit from the weld toe.
El The British Standards Institution 29111
I
BS
BS TBBBIZIII14
BRITISH STANDARD
in the case of weld toes on plate edges, type "b" hot-spot region in Figure G-2.
the stress distribution approaching the weld toe does not depend on the plate
thiclcness. Extrapolation methods have therefore been developed that use
stresses located at absolute distances from the weld toe rather than proportions
of plate thiclcness.
Figure E-2
Types of hot-spot
/'
Q
I
53"-J ta;
-,yyj..
-'1i-
allr
la
\/ L.
icey
Type "a"
1
2
Type "b"
For stress analysis based on measured strains or FEA of a relatively fine mesh
model, quadratic extrapolation from stresses on the plate edge at distances
4 mm. E mm and 12 mm from the weld toe should be used. with regard to
mesh site, in order to produce an accurate value of the stress at the 4 mm
location the weld toe element should have a node at 4 mm. Either 2 mm linear
or 4 mm quadratic elements should be usedThe corresponding hot-spot stress is given by:
_
5i-i - 3'-1-riririi " 3'1s.mra f tlrzmm
ll:-El
Alternatively, for stress analysis based on FEA with relatively coarse Ill mm x
ll} mm quadratic elements, linear extrapolation from stresses on the plate edge
at distances of S mm and lS mm is suitable [15]. The corresponding hot-spot
stress is given by:
SH = 1-Sogmm — B-So15mm
BE
I
El The British Standards Institution 29111
lfl.3l
BRITISH STANDARD
BS ir'BDBI2D14
Guidance on the procedure for applying the SSE method on the basis of FEA is
illustrated through a typical l-beam mesh of bricit elements. as shown in Figure
C-3. it shows a complete mesh and the two parts when it is separated along the
plane through the weld. The two resulting separate mesh regions are shown in
Figure C-4 and Figure C-S where the nodes on the common face are identified.
Additional nodes on the top face of the l-beam are shown in Figure C.S- The
corresponding element numbers are shown in Figure C-E and Flgure C.i'- These
are suitable for the SSE calculations. The weld toe is represented by nodes n2 to
n1l.1- The SSE stress at node n2 is calculated from the stresses at nodes n2, n41.
nS2, n-S3 and ni-'4 {Figure C-S1. The stress distribution using the values at these
nodes is shown in Figure C.S- The interpolated stresses at iltlt and 1.Elt from the
weld toe are used in the extrapolation, as shown.
The stress at fl-r-it from the weld toe should be determined from the stresses in
the element adjacent to the weld toe {e-g. elements eT1, ei'2 of Figure C-Til that
are not in or under the weld- FEA software normally calculates the stress at
node n2 as the average of the value at the node from all the elements it is
attached to lel, e2, e21 of Figure C-5 and ei-'1, ei'2 of Figure C-31. The stresses in
elements el, e2 and e21 tend to be lower than those in e21, e'l1 because there
is more material on that side of the weld toe. The stresses at node n2 should
therefore only be talcen from elements ei'1 and e22, where they are
representative of those in the main plate-
Frgure C 3
Possible briclr. element model of an l beam with a cover plate
tel
-II ll
II
\‘ ivi
-?_-"5-+“_|'i-I ,,.
‘RI
-i'.I -II‘I
"-
l-":I."=_|r-l‘
III:1
II]j,III;lg
I[lg
I‘
-I
1+-. .|.
-I -I-I-I
-—-
hill:
'I -I.-hl-I -a,I| . r: .*;ll
{iI"I;:i1
"',t‘,li;'.s=‘ -1 l:F- .
grllll.I*';.I*';:'r,+
I_'l....
:1-I .|-Ir-I--_
-I:2:-:-I‘
Ill
I:-I
iii '‘IIIxi"III"I'II-1:1:
Ir‘:-rtI‘ ‘III:
file: "lI'IIll'
tylf
l'I I ,|'I'.|'
I I I| - ‘I-‘I-?',|f,f I',j'I
-h.
I——-
‘frrf
III’
tiII,-,
'I-‘Ii‘Ir
1|,
II
"'|"II"-fir
‘-'-hr
\-qtIv-.
“raw-s
ft}
IIIII;IlI’I'I-r , €%"€l%fl=‘-€i-ff,r’ \%$‘
I
sf..SE!
§' II.I!lI' Ilselit-ifil ‘dlrljl
rrliri
+'5'fil.'+'Il'I
1"
‘—-
l
bl
I]
-I
‘TI.
‘-‘-:_l,|
II]:If
Weld region and adjacent elements
Elements underneath the weld
II II
‘I--___.--T"
H
q
'_",_-|:*,,_. 1-"
IIII
dl
ei
lei
SS'
~
.
!
.
I
2-fl’
Elements in beam adjacent to the weld
Close up of elements beyond the weld
Close up of elements underneath the weld
E The British Standards lnstittrtion 2lIl14
I
El?
“mu”?
%%MMW%%
-~.‘-~“”__%__§_
W%@%~mM"
"“fl§“
“
‘
Q
fig
_
F
g”
~
E
_3_" _H?l_ ':__Jlr _ “ '
“
Q?
fl
M
gQQ
%M_%% W”~fl4__‘Q”
‘mafiaflg‘ H?@g.EM
H D d_ E
U
F
E
__L5
E
_ _HW_ _M_ T_ J
fig
W%Q~@m
.M .
@
H‘_m‘
MM\E WEI’
‘I
dI
"_i _‘F
UW
E
__m__
ND
M
H
lfl ____|D
_
t
h U _____E D H H E
MH H W m
t
hE
M
rm __.____ _ _
r_____
_flU W _" _m F ___m____U __-E
__L
*h_
BS TEDBIZU14
Figure EB
BRITISH STANDARD
Ealculaticm crf the 55E stress at nude n2 cut the brick mesh shewn in Figure E3 tcr Figure
3*
sti
.—=
-II‘
'l'
1:--— -—~
nlii
lte
l
I
Stress
Distance tram weld tbe bn surface
n52
nf-3
3
I!"-J
n‘i'L
Structural stress at nude n2
A typical shell mcrdel cit the same detail is shewn in Figure EB. This mesh
illustrates the use at inclined weld elements. Mc-re details c-n the use crl shell
elements fcrr calculatiens UT structural stresses are given in E.5. The weld tee is
represented by the rrcrde line n2 tu n1I.'Ii {Figures C.'lD and iI.'l1l. The elements
representing the weld metal and the material beneath the weld are shewn in
Flgure [.12. Elements adjacent td these In the main sectibn are given in Figure
C.13.
Surface stress extrapelatien can be used te calculate the structural stress at ncrde
nit. The stress distributien aicrng the line ef n2 ta ni"'£l in Figure E.1i is used tu
estimate the stresses at l.'i.-ilt and 'i.Dt frem the weld tee. The stress shbuld be
-El"l1Zl5El"| cunsistently an the side bf the shell that represents the tap face at the
beam sectien. The erttrapcalatibn frem these twe values shbuld be made tb the
lcrcatibn bl nude n2 {Figure lI.‘l4l.
The SEE methed assumes that a stress at apprcntimately i].4t fr-am the weld tbe is
net influenced by the stress cencentratien effect cit the welded jeint. A
definitldn df thiclcness. t. fer the I beam was made in Figure -‘.'I.1. It shbuld be
assumed that the flange thicltness is used where the ccwer plate is welded tb the
flange. It considerably lbwer SSE stress weuld be estimated if the whele beam
depth was assumed fur lbcatictns where the ccnrer plate weld passes cwer the
beam web [nudes n5, n6 and n? cri Figure {.4}. The structural stress cwer the
web is liicel1.r tb be higher {Figure L1}; the flange thiclcness, t, shcruld be used fur
SSE calculatiens here.
The structural stress ctn the side crf the shell elements asseciated with the weld
tc-e being assessed sheuld be used fer 55E.
Ell]
l
El The British Standards lnstitutibn 2014
BRITISH STANDARD
BS TEDBIIDHI
If the surface stresses are cibtained frem strain measurements. using electrical
resistance strain gauges. the gauges shciuld ideally be lcicated at the
eictrapelaticin pci-siticiris. Alternatively. a series bf strain measurements might be
made ccwering the range that includes the eictrapcilaticin lcicaticrns, fbr eicample
using multi-element strip strain gauges. In such cases, sufficient gauges shbuld
be used tci enable a curve tn be fitted tci the results tcr establish the required
strains at the eittrilpblatinn Pusitiens by interpcilatinn- Depending cm the
extrapelatien prcrcedure adcrpted. strains icir strain ranges if they are measured
under cyclic leading} shcruld be substituted ‘I'Cll' stresses in Equatien 12.1 tn
Equatien 12.3. If the stress state is clese tn uni-aitial. the resulting structural strain
can be ccrntrerted tn stress simply by multiplying by the elastic medulus bf the
material. Hciweirer, fbr greater accuracy, acceunt shbuld be talcen bf any
bi-aiciality and principal stresses established using strain gauge rcisettes.
Figure {.9
Pcrssihle shell element mnctel nf an I beam with a ccnrer plate
/
ibl
El
l
‘I-1.-
.-‘it-.
| I - . - _- _. -__-__-_:__:__::__:___ _ _
nil:-ii.
I s || J
'5iiil‘l'SI\I‘lI.*ti- /’
.;"ir,||lt_
1 '| 'I'_ |_
P‘
In
1' I-' ‘III:-IL.‘ —l-| I
-I —.—.r'I' '.'--I
'F|-I _ 15+
1:.1.‘..-+1‘ -iiJ
1"
-F
-_-._-_:s-__-_.
1" — '- - 1--I
—-1-
_l_—:-.__ |
IP-‘‘Sr-"'
I‘ 11|—'_-- *‘i_1l-
—;.—._£ $--L-It
I I-tl-
r
S
T;
If}
ii
Hr
stir
‘L--",,1I-
iieI=.€a.-sifriisi. ‘“"’
/1'I "\\$-“=*
31%.?3*!-. ..:.:~‘§‘i“i'i;i;._
dis
it
IISe.-"=“""'*
{fill
rItI'IltI'i
Sffieis-‘:-fitfiryti
ryr
silly
rI*F.
rii;
;~rle-i .:~ii,tr,
.firfirI‘.
c#'-a."~'.“i'‘ly“"“"~"l'‘rife;
lgfit
i
i
'0
M
""'-Q-“'-"-ii
I till‘-ri»""’-Ir "~:i'Ill
*1t.._____,.r
%-.*3-'=- ;,.";}=nmt-
all--‘cc-‘Es-w
0—#;§#
'0I"
%;'?III
#1-=1
'0
we‘i
--rig“
Ii. .i". :;'IiI_l
lte
all
Weld regictn and adjacent elements
bl
Element; underneath ti-IE w,_=_i|g
cl
Elcise up bf elements underneath the
weld
dl
Elcise up bf elements underneath the
weld with seme bf the inclined "weld"
e}
elements l'E-l'i‘|t'_‘i'ti'EtIl
Elements beybnd the weld
f]
Clcise up -bf elements beycind the weld
E The British Standards lnstitutibn I-!l]11I
II
‘Ell
RS i"El]'B:2[l14
Figure t'I.ifl
BRITISH STAN DARD
I
II...
Nude numbers superimposed upen the shell element weld mesh section in Figure E Elcl
$~
1-Lianne.» fsapgty
l."l
Tl
s.
NDTE
The nodes correspond to those on elements with the same face shown in Figure if. ii
Figure E.‘l1
Node numbers superimposed upon the shell element weld mesh section in Flgure E Elf]
I133
nay
-IZIE5
iiii; iiiiii
#35555 Q?’
9Q4%,?
5%;
‘iii-i§is
~‘"=sea7~‘~T‘i§-“fie
-l'.l
il'.l
pg;
l'} Ti
NOTE
SI-1
999$
Ill
5.?
-'l'fl'l'
i@%
The nodes correspond to those on elements with the same face shown in Figure E I'll}
El The British Standards Institution I-!l]‘l=!l
-5‘
F
“___
E
___
1
2
EH
EmM fl
MD md
H
m _D E H 5 U D__ E ‘Hm P D
M
u
P D H _1__h E _ _ _ _h H H E m E M W E M m E _ _ _ ___"
_
%_‘W
_“%WW
%_uM__ gflfififink
$0
§_0‘fi$
P___"
‘_ _L _ I0
_ . ___R_hp‘! “_ guafl
1‘ti
“
‘l&_1kl___l__
____E H D H
n
F
H M
C
Sd
E
W
&£¢_£_ _ _ m afil
$
é
_
%I:‘
H:‘
§ll“
‘
___
“II
W‘
II
N U E _____m
F B
__|___ 1
d
3
m
fl
HPDnd
% @%®M&MrWW %M$w%W
g
M
___|_D _ _ _L" _U_ _ _E
®% Qwfi
%
_____F
D n _ _ _m m E H H M m m E
5Em E
_____-I
L" W
___ H
F9
h
H
____ M
$
NUE
__m
d
m
f
m PDH
Int D
M
___|__
E DH EM
______
M _“
_____-5_
Hm m E
E _m
m f
E
t
h St
F9 m C H
h W
d
d
M
ID 1 A
93
BS TBBBIZB14
Figure E.li!l
BRITISH STANDARD
Calculation of the 55E stress at node n2 of the shell mesh shown in Figures E3 tci Figure
112.12
l
il
3 *--‘-
l
L
n2
n13
l.‘l.lil
i-i:ir
.
..
nliii
.,.. 2
n35
nil-I.'i
Rey
l
2
Stress
Distance from weld toe through thicltness
c.-ca
12.4.3.1
3
structural stress at node n2
Hot-spot stress calculation using through-thickness stresses
General
Finite element analysis can be used to calculate the through-thicltness
distribution of stress as shown in Figure E.1. Therefore a force per uriit length of
weld and a moment per unit length of weld can be calculated. These values can
be used with the section area per unit length and the section modulus per unit
length to calculate the structural stress. This technique is called through
thiclcness integration {'l'|'l]l because the stress distribution is integrated to
calculate forces and moments.
Surface stress eictrapolation and TTI both use stresses from FEA models to
calculate the structural stresses. FEA also calculates forces. These are more
fundamental quantities which of-fer an alternative method of calculating section
forces and moments- Additional processing is needed to convert forces to loads
per unit length of weld. The methods associated with this are described in this
sub-clause and the technique is called nodal force ll'ilFl.
B4
Ii
El The British Standards Institution 2014
BRITISH STANDARD
BS TEBBIIBIA
The Tl'l and HF methods estimate the structural stress from the distributions of
forces and moments underneath the weld toe. For the ii-lF method the nodes in
the finite element model should lie on the plane normal to the surface. It is
therefore important to decide the section thicltness over which the calculation
of force and moment is to be made. This decision is similar to the choice of
thicltness that is needed for SSE. Figure C.1S shows that the distribution of stress
across the beam section is linear remote from the weld toe. The same structural
stress is therefore calculated irrespective of the chosen region of integration. ht
the weld toe. there is a small region of stress concentration. This region is a
significant proportion of the flange thiclcness. However. it is only a small
proportion of the whole beam depth at the web. Integration across the whole
web might not show a stress concentration because much of the stress
distribution here is the baclcground linear stress. The Til and NF methods should
therefore use an integration depth equal to the flange thicltness even over the
region of the web- For eicample the TTI or HF stress for node nS [Figure E.-‘Ill
should be calculated from nodal values at locations nS. ntis and n2? lthe
stressinodal force for node n2?‘ should be from elements in the flange {elements
el:'iB, eEE+ e'iB, eTi"EIl- and riot from element eB4 which is in the web, Figure E.'i'l.
A similar judgement should be made for other geometries, including the edge
attachment shown in Figure lI.'lE~ [see E.Ti'l.
Figure C 1S
Stress distributions across sections of an l-beam with a cover plate
""_‘-—-n.-
il The British Standards Institution ..'t[l'l1l
l
SIS
BS TBBBIZIII14
Figure E.lS
BRITISH STANDARD
Flegion of '|"I'I or HF integration for an edge attachment
1
I
. y /'\
I
Loading
2
'lTl and NF region
Figure E.1?
Stress distribution through I beam flange unclerrieatlt node n2 for solid mesh shown in
Figure L4, Figure ELS and Figure CJ
I
S
j e'll
-1‘ "'-.
‘B
§ eit
;
’
;
‘ edl
“Te
Q ellt
i
-I--I-H
"H.
1-LL‘-L‘
"-I-i.
-
n2
NDTE
-2
i
n13
ntli
The legend shows which element the unaveraged stresses were from
lliey
1
2
BE
Stress
Distance from the weld toe through
thicltness
Ii
D The British Standards Institution IIIJM
3
Structural stress at node n2
BRITISH STANDARD
Figure 12.13
BS TEBBIIBIA
Distribution of correctly averaged stresses plotted against distance y
"fl
rt in crlnlliil
I’: ifl crlniilll
ii if aiiisii
in
. crlrillll
if
Distance y is given in Figures t'1.E and (LT
Rey
it
"i"
Distance y
Stress
15.4.3.1
Hot-spot stress calculation using Tl'|
The TTI method is illustrated through a typical l beam mesh of briclc elements as
shown in Figure E3. it shows a mesh and sections of mesh separated along the
plane through the weld. The separation is shown in Figure [ill and Figure CS
where the nodes on the common face are identified. The weld toe is
represented by nodes nil to ntti. The TTI stress at node n1 is calculated from the
stresses at nodes n2. n'.i3 and n24 {Figure E.Sl. The stress distribution from these
nodes is shown in Figure C.1'.i'. There are multiple results because each ncide is
connected to more than one element. The elements used for the Tl'l calculation
are shown in the icey. These are the elements in the flange at the weld toe but
not underneath it {see Figure lI.i",l. Elements under the weld eitperience less
stress and should therefore be ignored.
Figure l'I.'l‘i" shows the calculated stress distribution using the correct averaging
of stress at each node {continuous llnel. The equivalent linear distribution is also
shown iclottedl line. The value of stress at node nit from the linear distribution
is equal to the ‘l'T| structural stress at this location.
The TTI structural stress has membrane and bending components. Lising the
nomenclature below and shown in Figure E.‘lB,, they can be calculated as
follows:
I
Correctly averaged stress at node nlil is iilnlill.
I
Height from the underside of the plate. y {Figure C.S. Figure 12.6} of node
nil] is yi-
I
Thiclcness of the part at the location is r.
-I
Membrane stress Elm is calculated using Equation I:.II-.
__r{~[-ii-*ii+~[ii i'+1i1}.i.i».-+1-ii
fl'"_
at
rl'.'.1l,l
The bending stress ah is calculated using Equation E.S.
El The British Standards Institution ..'t[l'i1l
l
ST
BS TBBBIZIII14
BRITISH STANDARD
TE“ =Z,(fl"~‘l"Il'lifl“l-"lit .I))l'..TIrfiI.Ff"I.})—u“£-
{E15}
where
“I'
Ilrllllrl
I 1]
1
. +ri.i il +rrini'.i+
- 1iiI—rii -sir.-i
.. +1:-'i.i)]
—E[rriniiii{—E-irf
+1-ire.--i
ms;
The Tl'i hot-spot structural stress is then given by Equation ET.
51-i= iitatdiiil
It-5"]
Shell elements generally provide structural stresses as output- The structural
stress on the side of the shell associated with the weld toe should be used.
l2.-1.3.3
Hot-spot stress calculation using HF
The HF method is illustrated through a typical I beam mesh of briclc elements as
shown in Figure C-3. The diagram shows a mesh and sections of mesh separated
along the plane through the weld- The separation is shown in Figure i‘-2.4 and
Figure t‘.S where the nodes on the common face are identified. The weld toe is
represented by nodes n2 to n'liEl. The NF stress at node n2 is calculated from the
nodal forces at nodes n2. n13 and n2-ll {Figure C.S}. The nodal force at each
node is summed from elements that are at the weld toe but not underneath it.
For eicample, the nodal force at node n2 should the summation of the nodal
forces at that node from elements eilt and e?.'-I {Figure t2.T]|.
A manipulation of the forces over a length of weld is needed to convert the
nodal forces into structural stresses. The nodal forces at each section {for
eirample at is = h2. Figure CZ-' and Figure C-3} are converted into a section force
Fl and a section moment Mi at that section as follows [Figure E.‘lSlt
Fl = .='INFi
i|"t'f.B,l
i'v'li = rivri til lj-'i — i_rr:rl ji
i'c.s.I
Where the nodal force at node nlii is HFI.
The section forces and section moments should then be converted to a
distribution of line forces fi {or force per unit length of weldl and a distribution
of line moments mi for moment per unit length of weld}. The membrane and
bending stress can be calculated as follows:
r:rm.=
rib --
'fi
—
(I T)
fC.‘l'i'll
(“mi i-i IF
E
,
The l'ilF hot-spot structural stress is then calculated using Equation C-1
SIB
Ii
El The British Standards Institution Itllllll
fill ll
BRITISH STANDARD
BS TEBBIIBIA
The relation between section and line forces and moments is given in Equation
C-12 for linear elements and in Equation C-13 for quadratic elements:
Fl
‘iii
‘T ll
'
ri
ii-ii
F2 : ii -_
hi“: -- ,r1'
_
H
and
fi-_
_
1
H"
E‘
H
F.
.
ilhl
hl
is ,f'.3
M
2M
- if 3
3"“
i_ji-;_i_:g
.l'HI
l --i"'-I = ii ftl K iii:
_
oiii iii I-ll!
Fl =i iii
I5
F3
—iil
‘iii
H]
1'
l'i_,
_1ii'I iii ll-ii’
I
.-I--fl =—I- iil
I5
iii-I3
—iil
iiiii
m
I icisi
i':iI
is ml
——iii -.1M
1
it-:3
where
F1
is the section force at corner node 1
M1
is the section moment at corner node l
F2
is the section force at mid-side node 3
M2
is the section moment at mid-side node 2
F3
is the section force at corner node 3
M3
is the section moment at corner node 3
f1
is the associated line force at corner node 1
mt
is the associated moment at corner node 1
fl
is the associated line force at mid-side node I
m2
is the associated moment at mid-side node I
f3
is the associated line force at corner node 3
ir.-3
is the associated moment at corner node 3
ht
is the element side length between the corner nodes
Typical h values are shown in Figure C.E and Figure CJ.
A similar set of matriic equations should be created for each element along the
weld toe- A master matriii should then be assembled. the coefficients being
added where two elements contribute to one node. The master matriii is then
inverted to obtain the line forces and moments along the weld toe.
C-5
C.S.1
Accuracy and limitations of hot-spot stress analysis
Use of brick element models for hot-spot stress calculations
Continuum elements tend to have only linear degrees of freedom.
Two-dimensional {ID} models and three-dimensional {SDI models have two and
three degrees of freedom per ricide respectively. Three-dimensional models are
constructed from volumetric elements that are generally called briclts.
El The British Standards Institution ..'tl311l
l
S13
BS TBBBIZB14
BRITISH STANDARD
The assumed distributions of displacement in continuum elements are generally
either linear or quadratic. Elements are therefore called linear or quadratic
respectively. Bending modes of deformation tend to produce curved deformed
shapes which quadratic elements are able to represent. However, linear elements
are generally unable to represent bending accurately. The stress distributions at
fatigue cracic initiation sites often have a bending component, so linear
elements might not be suitable for accurate hot—spot stress
calculations isee C-S-3l.
The use of quadratic 3D briclc element models provides a relatively detailed
indication of the distribution of stresses at structural details. Briclc models can
accurately represent the geometry of a detail lilte a welded joint, including the
weld profile. This contrasts with shell models which represent the centre-plane
of thin sections. its a result, accurate calculation of the stress distribution at a
potential fatigue cracic initiation site, lilte the weld toe, is possible using a very
fine briclc model, The need for fine meshes, however, means that briclc models
tend to be larger and more time consuming to prepare and run than equivalent
shell element models [C-S-2}.
The size of briclt models can be reduced and the compleicity of the assumed
geometry simplified, but these changes affect the accuracy of the model.
Modelling accuracy is discussed in greater detail in C.S.L'l. Some possible
modelling techniques are described below. The dimension of the elements at the
fatigue initiation site under consideration relative to the underlying plate
thicltness is important. Figure C.2Ei shows the dimension. f, of an element normal
to a weld toe. The recommendations in C-S-3 are based on the condition f E 2t.
The 3D briclc model shown in Figure C-3 to Figure C-3‘ has elements that
represent the overfill of the fillet weld around the cover plate. Figure C-31
shows an alternative mesh where the overfill has not been modelled. Instead,
the cover plate is connected to the flange via a ring of elements around the
cover plate perimeter as shown by the shading.
Figure C-19
Distribution of correctly averaged nodal forces plotted against distance y
. iiiFs
Q liiF3
Q iirt
Ii,
I
Distance y is given in Figures C-1’ and C-13.
lliey
‘l"
1[li§l
Distance y
l
D The British Standards Institution EDI-4
it
l'-lodai force
m
BH W E H ____T
H
__H_U r E
___“ ________ __U_
_B
r
M
E M m E H _'_____ m E m W M d__ w H M D H
i___#r_#w____h“"flhw_mF_“m _ W" h h*_F_F_’_-_#‘___#"__§fl""_h“__h_h:i‘I‘-__h_&__ "_fl:__I--__+___I_fi.____-"_“."_hI__-_I_ _£_“fl-N%_ _ flm'mwvmvm w
_ #H__" m_ "_ _m$“h_‘"“____
_fi
‘
h
_
fi
"
_
l
I
.
_
*
_
______mm
_ _ _ _%_ _ _ _ _ _1h_ _ _q_
:
‘___-_ _ Jl_. ___.. _. _ __“.__" !_I ‘
*i_I___I_
f'_-_l_Il _l_ l .
__""
__m_
B 5 T E D B "___ D 1 4
DAHD
_mw E H
t
__"__E
E M m ___H t
*I_I_=
ll
it
__
*
__“
___
Ill...
__
_ _“fink.
_z
I
a n ___“P H 1 E m _“H H E 5 5
I tI
gmmwkg%
___+_+____"
_ _ _ _‘_ _‘a
*_ _"I_ fl
_+H"_m
_
0'
!._“r_ _fi
___“I E H
__Wé§aw
gI-“_ _ _ §n_ _ _“_ n_“".‘Q
‘‘"8
:."_I___..‘_“_ii.‘
‘__
I.mK
-I.
‘IH
T h E d E fl H Hm D H _ _ _ _._
m E 5 h_
F
_m_ U E F___ H
r
E ddE
IE t D
t
h E _m ___:E t W h E r E E M m E H ______5
__H B V E b E E fl
mumW? {MM
O 5%
H lm‘DEEhmm Em_fl_H“___ Wmifl
Bmflr kflfiL_ "fl_ _" ___Emhm‘mflsEwm
Esh
H3gt
fu
_____
h{
r
E m GV E
_fit fi
M H _r E _ _ -_ E
t __H E
WE H
Emmi EdM“H} umf _ _ _ _D_ HmUVmm hm HE hmE3 EEEH
fit
t
DE
‘Mmw
M
____
h_____
___
_"‘|_mh1_____1_ p'_____
_ I"q_‘ ______l__L__.____
___
_ mH__E__?HHHm
____“
H H _ih_H._H ___“__
_I i.______“
fl_____flfl___.
____-____"fl+
__
____. _.F‘_.- _l
*_**i_ __-_________I
______
__“‘-_-h___
‘%_.__-__
fl. _'_q___|l__'
_ __
_DM|%‘m"r_ '*.__
illill
_l_
___I_r.||n‘I
__
_|_|__*_.
__
l1_.-_._________"__"_"__._____1|
_T__hT__J
__h_h|
r
__II__
+____H|___J_|
| |_|“|
P&||_|
_|_
____
_.
J_
_______
H___
.__._-___;
_l-_
__
_ m“_ _ _ _‘I-_
____
“fin”
H-‘fin
J___-fir
I__ -_I _- _ “_- __
*
__-_=___
_I ._
___?"
_+_ _:_l___!“_-_-_
_
I‘
_
ha‘
%X ‘I'M‘&_q
Q
%;_I__I‘
W1 %
:§‘:_-I
.‘.-________
ix‘I
‘fifl
.
_ ii
II
E T h E B ___. ___" h
-E
___§__-___E
___" d H d ______ ___“ ___“___H_ U __U__u H 1___ U 4
__|
1
_'
-___
_"_
1.
F QU
E
__-I
__LA _______
D
.
E H 5 _m H 5
#_ “H‘Wm%q_ “ $m
U5
E d__ __iD_ I _m d M E d__ E E m E H t
Mum__“_ _h _ _fi _ _ “ _ "1_
“I_flmn“‘-'_“ h_“.m_H“._ §*m_“ ___+___fi____
£__:___
_ 1_l _fl“F_l m_" I____'__f_fl_
E‘
I_______mfi_
¢
L
_______-‘
_. _ ____+_
_ _._ -_ -_
gun!
___*
__
__ _flu_’ -__“_ _m_“g___$-Ifl‘
I__iI‘____.
E
__
__| E P H 5 E H _ _ _H t _m H D
____ B fi
E t WE M
Q
g___‘5I‘__$““___fl
Efii‘n.‘W
____“$_$ “_& $_d__§:_“§
__I’’
M
g
‘__$_
l
Th E
1 DI
_m d _m E
_-
E
Ifl 5 h H
E E mEH t5
5
_h D U M h B U E
H
Ih
T h _ _ _ B Int __h h HW d E _|d 5 M 5 t ___“ U t _m _" I D 4
1
____"_H__ H
E U. U flu
t
D
WE H
HL
BRITISH STANDARD
Figure C13
BS TEDBIIDI4
Dimensions used far thicicer element representation pf a fillet weld
H1
"I-|.
-
II-
1-+5
___--‘H;
Ilr""'
-'-
Hl
91 .""-Q-"'9'
.1-""""r
The thickness cif the elements is based upen a weld threat thicltness cif w.
llle if
Thiclcness = t,
| |,.I_|.l.
U-J
Thiclcness = it, + w}
Thickness = lit; + w}
E.5.I
Use pf shell element mpdels fer het-spet stress calculatiens
Shell elements are part pf a greup pf elements called structural elements.
Meshing ef structures with shell elements dees nut require an eicplicit
representatien bf the thicltness bf the plates they represent. The thicltness is a
preperty which determines the stiffness bf the shell. The threugh-thicltness
distributien c-if stresses is estimated can the basis bf beth linear and rcitatienal
degrees pf freednrn. The HF methed uses the nedal mements bf retatienal
degrees ef freeclem ta estimate the distributien bf the bending cempenent ef
structural stresses. It is net pessible ta represent the geemetry and stiffness ef
relatively shert structural details such as weids explicitly using shell eiements.
The stiffness is that bf a sectien pf plate bf the same size as the shell element.
Fillet weld details may be represented in twe ways. fine methed is based upen
inclined weld elements meant tn represent the centre-plane bf the fillet weld.
E1IlEll!'l1|:|-E5 are shewn in Figure [Lil] and Figure C12. Figure {L12 shews seme cif
the elements representing the fillet {e-g. elements ell ta eefll rerneyed se that
the unfused land underneath the weld can be seen. Figure L22 shews a
suggested geemetry ta be used fer this sart af shell medel fer the I beam with a
ccwer plate.
Ann-ther pessible representaticsn ef fillet welds is based upen a thiclcening bf the
elements in the plate underneath the fillet weld as shewn in Figure E13.
fit The British Standards lnstituticin EH14
1
‘llflil
BS 1503:1514
BRITISH STANDARD
The accuracy bf these metheds has been assessed and is presented in 12.5.3
tegether with a plane mesh. witheut explicit representaticrn bf the weld. The
plane mesh is similar tci the geemetry shewn in Figure 12.23, except that the
sectien thiclcnesses are nbt medified tn represent the stiffening bf the weld.
The assumed distributibns bf displacement in structural elements in FEA are
generally either linear er quadratic. Elements are therefere called linear cir
quadratic.
C.S.3
Accuracy bf medelling and pest-precessing metheds
The accuracy nf the yaricius metheds fcir the calculatien pf het-spet stresses has
been determined threugh the analysis bf seme test cases [ED]. These all
censidered weld tcres er ends {referred ta as type "a" in Figure {.2}. The
accuracy cif each methed was determined by cemparisen with the results
ebtained frem benchmark reference cases in which the structural het-spet
stresses were determined using SSE frem fine mesh brick mudels [fbur elements
threugh the plate thickness]. The results bf this benchmark analysis led tn the
recemmendatiens detailed in Table C.1. These are enly valid fbr element sizes, f
[Figure CED), up tn twice the underlying plate thickness, t. Fer brick mcidels it is
assumed that there are at least twe elements threugh the plate thiclcness.
The benchmark analysis upen which the recemmendatiens in Table {.1 are
based did net ccwer every circumstance. Therefere a benchmark study sheuld be
cpnducted in any case where there is a need fcir additibnal ccrnfidence in the
chcisen methbd bf caiculating the but-spet stress.
C.S.4
Llmitatinns pf the het-spet stress calculatien metheds
Attentien is clrawn tc:- the preblem bf using het-spet stress calculatien metheds
that are related ta the plate thickness when cnnsiderlng the tee pr encl pf an
edge attachment weld [type “b” in Figure 121.2}. At this stage enly SSE shbuld be
used fbr determining the hbt-spet stress in such cases. As this is essentially a
I-dimensibnal prciblem. any bf the FE mbdelling metheds is suitable, prbyicled
the element sizes in the reginn being assessed meet the cenditiens described
in 1.11.4.2.
There are alse sbme welded jeint geemetries fer which the hot—spot stress
determined by the metheds described in this annex is under-estimated. such that
it is yirtually the same as the neminal stress. Cpnseguently, use bf the het-spet
stress design class {D} rather than the neminal stress class weuld be
nun-cunseryatiye. Particular pr-bblems arise with simple cruciferm er T-jbints
between plates fbr which a cress-sectien nbrmal tu the weld tee being
censidered is symmetrical, as in Figure L24. The het-spet stress at the weld tee
is lcnewn ta increase with increase in attachment length. but this is net detected
using any pf the three calculatien metheds in this annex.
1l]4
l
E The British Standards lnstitutibn 2014
BRITISH STANDARD
Figure C 24
BS TEDBIIDIA
Example at symmetrical welded jeint fur which het-spet stress is underestimated using
metheds in this annex
it
_
l
./"I
X
\
A new methed that has been shc:-wn tp include the stress cencentrating effect cif
the attachment site is tn equate the het-spet stress te the stress 1 mm belew
the weld tee [21]. A disadvantage cif the methbd is the need fcir scilid FE mciciels
with very small elements, altheugh it is claimed that elements dewn tci 1 mm in
size are sufficient. An extensive database frem a range bf welded jeints has
been shewn T.-Ci be ccinsistent with class D an the basis crf the ‘I mm stress [21].
An alternative appreach is ta assume that the piate thickness is effectively less
than its actual value sa that greater weight is given ta stresses iecai tb the weld
tee. This appreach is recemmended by Deng et al [22] fcir assessing any
symmetrical {as described abeve} jeint assuming that the effective plate
thicltness is half the actual. Thus, SSE by linear extrapcilatibn vvcruld use stresses
lecated i].2t and iJ.St frem the weld tbe, while calculatien bf the equivalent
membrane and bending stresses fbr the TTI and HF metheds weuld be based an
half the plate thickness.
A similar prablem can arise with shell element FE medeis if members that are
large eneugh ta attract leading cannet de this because they are mci-delleci with
twp-dimensibnal elements.
Scime bf these deficiencies in the metheds bf calculating the l'il‘J‘lI-S|';lItIllI stress are
the subject bf research and imprqvements might appear in future. Meanwhile,
unless the neminal stress-based appreach can be used, expert advice shbuld be
sciught fbr assistance in areas crf dnubt pr ccincern. As a guide te the validity bf
a calculatien nf the het-spet stress fer a fillet weld. it sheuld be at least 15%
higher than the neminal stress at the weld tee.
Misalignment and distcirtibn
A seurce bf stress cencentratien in all types bf structures can be linear er
angular misalignment at jeints isee Figure LL25 and B.S.2.1}. This can arise frem
pbcir attentibn tn assembly cir erecticin precedures er frem welding -EilS‘lIDl"lIlDl"|.
Stresses can be set up by the iecai bending effects induced by this misalignment
and these sheuld be taken inte acceunt.
NOTE References i'.?3—.2'5,i' cever this in sbme detail.
Er The British Standards lnstituticin EH14
1
‘ilIlS
BS TBBBIZB14
BRITISH STANDARD
Circular sectien pipes and vessels are particularly sensitive as ciffset and angular
distnrtien in the lengitudinal seams and evality can individually er in
cbmbinaticin magnify the neminal qr hbt~spcit stress by an amciunt which can
sericiusly degrade the fatigue perfermance under pulsating pressure. Similarly,
sbme degree bf linear andibr angular misalignment is virtually unaveidabie in
butt welded jeints between cu-planar plates and cruciferm jeints. Unless the
extent ef misalignment is already allewed far in the 5-H curve. it sheuld be
included in the calculatien at the het-spet stress. It is rarely practical te include
misalignment in a FE meclel. Therefbre, if relevant, the calculated value pf the
het-spet stress sheuld be multiplied by km , the stress magnificatlen facter due
tci misalignment isee B.S.2.1l, tci determine the actual vaiue.
Figure [.25
Types pf misalignment and distcirtiun
E
is
ll
all Linear misalignment {plate centre-line bffseti
bl Angular distcirtibn
I
‘i
c} Clvaiity
1l]E
l
E The British Standards lnstitutibn Ellie
)
(
i
I
BRITISH STANDARD
Table -E.l
BS TEBBIIBIA
The perfermance bf structural stress calculatien prbcedures SSE. Tl'l and HF fbr assessing
hbt-spbt type "a" welcl tbes br ends“
Element type
Weld mbdelling
metlibd
Hbt-spet stress
cleterminatibn methbd
Eamment
Quadratic brick
Eluadratic brick
Weld mbdelled
SSE br NF
'l'Tl if benchmark study
Weld nbt mbdelied
Til if benchmark study
Linear brick
Linear brick
Weld mbdelled
SSE br l'~lF
NF
Weld nbt mbdelled
NF
Benchmark study required
fbr assessing end bf
lengitudinal weld
Quadratic shell
Weld mbdelied using
inclined weld elements
Benchmark study
recemmended fbr assessing
SSE, TTI br HF,
weld tees by TTI er HF
Quadratic shell
Weld mbclelled using
SSE, TTI br NF
Benchmark study
recemmended fbr assessing
weld tbes by TTI er NF
thicker elements
Weld nbt mbdelled
SSE, TTI br NF
Weld mbdelied using
inclined weld elements
HF
Linear shell
Weld mbdelled using
thicicer elements
Hbt recbmmended
Linear shell
lnr'eld_nbt_n1bdelied I
NF
Quadratic shell
Linear shell
I
l—I
Benchmark study
recbmmended fbr assessing
weld tbes
TTI if beri_chHm_a|_'k_stud_y
IaJ—I
fl Applicable lbr elements sixes fit s 2. Reference case was FE mbdel with lJ.25t quadratic brick mesh and weld
included. Structural het-spet stress was calculated using SSE.
Annex D
{nbrmative}
Guidance bn the use bf fracture mechanics
Bacitgrbund
This annex prbvides guidance bn the use bf fracture mechanics methbds fbr the
fatigue assessment bf structures br cbmpbnents subjected tb high cycle fatigue
cbnditibns. in situatibns where the nbrmal fatigue strength assessment metheds
in this British Standard might be unreliable br inapprbpriate.
in general, fracture mechanics is nbt suitable fbr calculating precise fatigue
strengths br lives as the results are largely dependent upbn the assumptibns
made, {e.g., the values bf the cbnstants in the cracic grbwth equatibn, the size bf
the initial flawfsl and the shape bf the resulting fatigue crack, fbr example, fbr a
crack at a weld tbe, whether it is semi-elliptical br straight-frbnted}. l‘~lbt all bf
this infbrmatibn is available at the design stage. Therefbre, if the bbiective is tb
define a particular fatigue strength br life, assumptiens made shbuld be very
censervative.
Fracture mechanics can, hbwever, be a useful methbd fbr carrying but
parametric studies, where the bbjective is tb define the relative influence bf a
particular set bf variables. in that situatibn all the variables, except the bne
under cbnsideratibn, can be held cbnstant and its influence can be evaluated.
The guidance in this annex dbes nbt replace the nbrmal fatigue assessment
precedures butllned in this British Standard when such precedures are
applicable. Far example. they are nbt intended tb be used as a methbd ta
circumvent the nbrmal requirements fbr gbbd wbrkmanship.
Sbme typical situatibns in which the nbrmal prbcedures might be inapprbpriate
and in which the use bf fracture mechaniu might be helpful are as fbllbws.
El The British Standards lnstitutibn Efll-It
1
‘ill?
BS TBBBIZB14
BRITISH STANDARD
al
When assessing the fitness fer purpbse bf a structure knbwn tb centain
flaws whbse size, shape and distributien are butside nbrmally acceptable
limits {see 14.3.4} but which wbuld be difficult tb repair.
NOTE in this case see as rare.
bl when the effects bf relatively miner variatiens in the geemetricai er stress
parameters fer a given detail are being studied.
cl
When the jeint detail under censideratien is unusual and is net adequately
represented by bne bf the standard jbint classificatiens, br when a jeint is
subjected tb the influence bf anbther stress cencentratien.
dl When defining the frequency ef in-service inspectiens.
el
When assessing the remaining fatigue life bf a structure in which fatigue
cracks already exist.
fl
When determining the damaging effect bf stress ranges belew the cbnstant
amplitude fatigue limit in the calculatien bf fatigue life under spectrum
lbading.
in the case bf item ej the structure wbuld centain cracks whbse sizes have tb be
determined by measurement and the sizes assumed have tb allew fbr pessible
errers in such measurements- In ether situatiens it has te be assumed that small,
but unmeasurabie. flaws exist at pbints bf stress cencentratien {e.g. at weld
tbes} and that it is frbm them that fatigue cracking can brlginate. The size bf
such flaws shbuld therefere be assumed.
in general this annex cevers the applicatien bf fracture mechanics tb fatigue
cracking frbm a weld tee threugh the stressed member {Figure ill] br frbm a
weld rbbt threugh the weld threat {Figure I12]- lteference shbuld be made tb
BS i'S+'l{l fbr censideratien ef fatigue cracking frem embedded flaws. The
prbcedure recemmended in I13 te D.E is based upen the principles bf linear
elastic fracture mechanics. similar tb that in BS ?SilEl, which alse centains mere
detailed guidance and relevant design data.
b.z
Symbels and units
Fer the purpeses bf this annex enly, the fbllbwing symbeis and units, {which are
censistent with HS it-lllll. are used.
a
Measure ef current crack size {length er depthl fl {in mml
a,
Final vaiue bf cracic size“ {in mm}
a,
Initial vaiue bf crack size" {in mml
A
{Ibnstant in the crack prepagatien equatibn
E
Plate thickness {in mml
c
Half the surface length bf a semi-elliptical surface crack bf depth a
{see Figure D.1} [in mm]
dardlll
Flate bf fatigue crack prepagatien lmmicyclel
Fm, F,,
Functibns ef crack size and shape and the preximity ef the crack
tip tb free surfaces fer membrane and bending stresses
respectively
h
Weld leg length {see Figure ill} {in mml
L
Overall length frem weld tee te weld tee bf an attachment,
measured in the directibn bf the applied stress {in mml
P" Length and depth are measured in the directien ef prepagatien.
1[lB
1
E The British Standards lnstitutibn Ellie
BS ?l':i{lB:1lIl11'-'s
BRITISH STANDARD
rn
tienstant in the crack prepagatien eguatien
Mg
{Ibrrectibn factbr bn stress intensity factbr dependent bn crack
shape and size fbr bending stress
Mm
Cbrrectibn facter en stress intensity facter dependent en crack
shape and size fbr membrane stress
M,‘
lvlagnificatibn factbr bn stress intensity facter tb allew fbr the
presence bf a stress cencentratien, such as weld tee {suffices it,“ er
it], used te indicate membrane er bending}
Figure D.1
N
I"-lumber bf cycles {in cycles]
N,
Number bf cycles te crack initiatien {in cyciesl
arc
Range bf stress intensity facter at the tip bf the crack {in iii-mm"?-“*"l
alt',,,
Thresheid value bf bit fer crack prepagatien {in lil-mm'l"l
qig
Cbmplete elliptic integral bf the secbnd kind
Flaw dimensiens
r_
__
I
ea-—|'
1;
l
lTl
2‘
1.
i
lcflt
'
l
a} Semi-eliptical surface crack
l.
E
ll
E
{
E
bl Dimensiens ef jeints with cracks at the weld tee
E The British Standards lnstitutibn 2{li-It
1
il]El
B5 TEDB EH14
Figure D I
BRITISH STANDARD
Transverse load-carrying cruciform joint
-‘I-I
l
2
i
l
HE
_._.
g
If
if
r
]
|..._
so
l
'|
General baclcground
As outlined in 11.1, fatigue craclcs in welded joints can originate either from the
toe or root of the weld, depending on the type of joint, or from planar or
non-planar flaws in the weld. Eraclcs originating from the weld toe normally
initiate at small flaws while craclcs originating from the root often start from
areas of deliberate laclr of penetration; in both cases the initiating feature can
therefore be regarded as a planar discontinuity. Most fatigue craclcs in welded
joints can therefore be regarded as starting from a pre-existing planar flaw and
their behaviour can be described by the use of fracture mechanics analysis.
The objective of such analysis is to calculate the life by integrating the relevant
cracic growth law. In doing so it is assumed that the real flaws can be idealized
as sharp-tipped cracits. which propagate at a rate, dardhi. which is a function of
the range of the stress intensity factor, alt.
Fracture mechanics is applicable to the behaviour of cracks under loading that
can cause them to propagate, including fatigue loading. There are three
possible modes of cracic opening, depending on the type of loading, as
illustrated in Figure I13. This annex is applicable only to craclt opening mode l,
generally the most severe, although the principles described would apply for any
mode. When more than one mode is relevant, the corresponding stress intensity
factor is normally referred to as it,, but the suffix is not included in this annex
for clarity.
The overall relationship between asrarv and nit’ is normally observed to be a
sigmoidal curve in a log dafdfll versus log .£i.iii plot. There is a central linear
portion. At low values of tilt’ the rate of growth falls off rapidly to a threshold
stress intensity factor, .sir,,,, below which no significant craclr growth ls lilcely to
occur. At high values of tilt, when the maximum stress intensity factor in the
cycle approaches the critical stress intensity factor for failure under static load,
the rate of cracli growth accelerates rapidly. However, for practical purposes, it is
often sufficiently accurate to ignore the existence both of the threshold and of
the failure regions and to assume that the central linear portion applies for all
values of tilt’ up to failure. An important exception is item fl in D.1 when I5-H"-l is
used to identify the applied load cycles that can be neglected as being
effectively non-damaging.
1'lIEi
l
E The British Standards Institution 2014
BRITISH STAN DARD
Figure I13
BS ?ElJBIID14
Eracli opening modes
ai lvlode l {Opening}
bl lvlode ll {In-plane shearl
cl lvlode lll liIli.rt-of-plane shear]!
The relevant equation for the rate of cracic propagation daldlv lin mmlcyclel is
given by:
da
dill
.
lllll
— = .a{.-_ixl"‘
where
A
is a constant that depends of the material and applied conditions
including environment and cyclic frequency:
rn
is a constant that depends of the material and applied conditions
including environment and cyclic frequency;
ax
is the range of stress intensity factor corresponding to the applied
stress cycle and instantaneous fatigue cracic dimensions.
Integration of Equation D.1 gives the number of cycles. N. required to
propagate a cracic from an initial size. a_, to a final size. a,. as:
l"v'-
1"‘
1
. a
*“'".ll""i'l'" H
real
If the case being assessed is one in which N, cycles are required to initiate a
fatigue cracic, the total fatigue life is l‘il' + l'l.l,. However, N, is usually assumed to
be zero for fatigue failure from stress concentration features in welded joints.
The application of Equation [$1.2 for calculating the fatigue life under spectrum
loading is equivalent to the use of |v|iner's rule isee 15.i'l assuming El = 1 [3].
However, in contrast to the application of Miner's rule using a constant
amplitude 5-N curve. direct allowance can be made for the damaging effect of
stress ranges below the constant amplitude fatigue limit on the basis that only
combinations of craclc length and stress range that produce Mt’ values greater
than the threshold value all-It,__ are damaging (l.e. cause cracic growth}. However,
again in contrast to the use of lvliner's rule, the order of application of the stress
spectrum should be talcen into account to ensure that stress ranges that actually
occur when the cracic is long enough for alt to exceed .-tit,“ are not neglected.
Therefore, unless the order of application of stress cycles is known. cycles should
be applied in the most damaging plausible order, established if necessary by
performing the calculations for various orders to determine the worst case.
El The British Standards institution 21314
1
l'l1
RS TEDBIZD14
BRITISH STANDARD
if there is uncertainty about the validity of IvIiner’s rule for the type of load
spectrum being assessed, isee 15.7}, the life Iv calculated using Equation lJ.2
should be reduced accordingly {e.g. halved if D = IIl.S is appropriate}.
Values of A and m
The values of A and m depend upon the material and the applied conditions,
such as stress ratio, environment. test frequency and waveformWhenever possible, data relevant to the particular material, product form and
service conditions should be used and where any uncertainty exists concerning
the influence of the environment such data should be obtained in accordance
with BS ISD 12ilJB. Provided that sufficient data are available to enable them to
be defined, the chosen values should correspond to the mean plus two standard
deviations of log daldill, representing the same 92.2% probability of survival as
the design 5-N cunres-
Comprehensive guidance on A and m values relevant to welded materials.
including allowance for applied stress ratio and various environments, is given in
BS Ftilfl. in the case of structural steels under conditions comparable with those
experienced by welded joints containing high tensile residual stress, the
following conservative fatigue craclc growth parameters are recommended:
I
in air, m = 3, A = 5.21 x Iii“ and tilt,“ = E3 l~l.mm'i"'f:
I
in a marine environment at temperatures up to 2!] “C. m = 3. A = 2-3 x ill“:
and arr,,, = zero.
Initial flaw size a,
The calculated life is usually very sensitive to the assumed value of a,. Therefore
a, should not be underestimated.
For nominally flaw—free welded joints failing from the weld toe a, should be
assumed to lie within the range D.1 mm to I125 mm l2E, 221 unless a larger size
is known to be relevant. The correct value to be used is calculated from the
calibration calculations isee D.Bi.
In the absence of definite information about the shape of the initial flaws, for
joints with welds transverse to the direction of stress it should be assumed that
an
the flaw at the weld toe is long and continuous, i.e. Fl = ill At the ends of
c
longitudinally loaded welds, however, it would be realistic to assume that the
-Eli
initial flaw was a semi-elliptical in shape with 2- = ill .
-C
Limit to fatigue craclc propagation a,.
in the fatigue assessment, an upper limit should be set to the size a, to which a
cracic could grow without failure occurring during operation by any of the
following modes, as appropriate:
al
unstable fracture;
bl yielding of the remaining section:
112
l
cl
lealiage {in containment vessels};
di
stress corrosion;
e)
instability {buckling}; or
fl
creep.
E The British Standards Institution 2&1-4
BRITISH STANDARD
D.1
D.?.1
BS TEBBIIBIA
Range of stress intensity factor, AK
tiieneral
Application of the crack propagation equation isee DJ] requires knowledge of
the range of stress intensity factor, alt. at the cracic tip- Hence, for the actual or
assumed dimensions and position of the idealized flaw. the value of alt
corresponding to the range of stress isee 15.2 and 15.3} should tie estimated
either from relevant published solutions, notably those in BS TE-lllli, or from
specific stress analysis of the structure.
D.?.1
Stress intensity factor correction factor
in general the value of iilii is of the form:
|—
{DJ}
"|"lTH
.-Jilif = lhrfymfl-'fmJdm + ill-I'f|.;i;.lI-"l|;,,.rlrI|;.l?
U
NOTE gslu is only relevant in the assessment of elliptical embedded or semi‘-elliptical
surface cracks. it is a function of air,‘ for straight-fronted cracks l'ai'c=t'l,l it is l.lJ
BS '.lFl1{l provides an extensive range of solutions for the correction factors Mm,
Mg and llrl,,.. in cases that apply specifically to welded joints the following apply!
a}
Correction factors that do not distinguish between membrane and bending
stresses are used in conjunction with the total stress range dam + nah.
bl
If the distinction between membrane and bending stresses is not known.
the correction factors for membrane stress should be used in conjunction
with the total stress range .£irr,,, + deg.
In the cases shown in Figure D.1 and Figure D.2, the solutions in BS TBIB allow
arc to be calculated for semi-elliptical or straight-fronted lair = ill cracks at weld
toes but only straight-fronted cracks propagating from the weld root in
cruciform joints. In a fatigue crack propagation analysis the procedure is to
calculate arr at the crack tip El isee Figure D.1} and, in the case of an elliptical
flaw, 5 and use Equation D.1 to determine the incrementlsl of crack growth for
the relevant .’!ill'.' valuelsl and the number of applied load cycles. The process is
then repeated for the new crack size until a = aj. Depending on the purpose of
the analysis this might correspond, for example, to complete failure, leakage of
a container, the attainment of a detectable crack or the end of the required life.
D.?.l.ll
D.?.3.1
Stress range
Gen erai
BS 291111 provides detailed guidance on the determination of the stress range on
to be used in Equation D.1 However, this annex has some additional guidance
to allow a choice to be made between the use of the nominal or hot-spot stress.
Table D.1 summarizes the resulting recommendations.
D.T.El.2
Iilominal stress range
The guidance in ES 2910 leads to the determination of the same stress range as
that described in 15.5 for use in nominal stress-based assessments. Thus. do in
equation I13 should include allowance for the stress concentration effects of
any gross structural discontinuities and misalignment. The additional stress
concentration effect of the weld detail, which is not required for nominal
stress-based assessments using S-N curves, also needs to be considered. If this is
not already included in the it solution it is introduced using the correction
factor, Mk, which is a function of crack size, geometry and loading, as follows:
El The British Standards institution 2tl1-I1
1
113
BS TBBBIZB14
BRITISH STANDARD
My
-
It’ for crack in plate
it
for same
crack
in
plate
with stress concentration
without stress
lillil
concentration
The solution in B5 rare for weld toe surface cracks is a function of crack depth
a, plate thickness .El and the overall length l. of the attachment measured from
weld toe to weld toe {Figure D.1}. l'vl,, normally decreases with increase in crack
depth, from a value equal to the stress concentration factor in the absence of a
crack down to unity at crack depths of typically Eollt of plate thicltness. at crack
depths greater than that corresponding to Mg = 1.l] it should be assumed that
M, = 1.tl.
In the corresponding B5 'i'EillI1 solution for weld root cracks in cruciform joints
{Figure D.2I l'v'l,,, is a function of crack length a, weld size h and plate thickness B.
It is always less than unity and its value might increase or decrease with increase
in crack size, depending on the value of hill‘.
In all the cases shown in Figure D.1 and Figure [Ii.2 the stress range is that in the
loaded plate at the weld toe.
D.?.3.3
Hot-spot stress range
The potential use of the hot-spot stress range is confined to the assessment of
fatigue cracking from a weld toe. It is not applicable to the case of fatigue
failure from the root of a cruciform joint [Figure I12). The hot-spot stress
includes the stress concentration effects of gross structural discontinuities.
misalignment {which in practice is likely to be allowed for by applying l:,,, to the
hot-spot stress determined for an aligned joint, see B.5.2.1l and of the weld
detail, but excludes the stress concentration effect of the weld toe isee 1S.?j.
Therefore, the only further correction needed before using it in Equation 13.3 is
My. However, a complication with the Mk solutions in BS Tfllil is that they
already include the stress concentration effect of the weld detail, as reflected in
the UH value. As that effect is included in the hot-spot stress lexcept for those
cases highlighted in 12.5.4} the ll-l,, solution for i.ltl = 13.5 should be used to reduce
its influence. However. the actual titl value should be used if there is any
uncertainty that the hot-spot stress has been determined correctly [see 12.5.11},
which is a consenrative assumption for any case.
In practice, if the hot-spot stress is determined on the basis of calculation of the
through-thickness stress distribution and corresponding membrane and bending
stress components i,i.e- by the TTI or NF method, see t'I.lI.3]l. these can be used
directly in Equation I13. M, values would also be required, as described above.
D.B
Calibration
When the problem under consideration is one in which fatigue cracking from a
weld toe is involved, the fracture mechanics formulation which is used for the
fatigue assessment should be shown to predict, with acceptable accuracy, either:
al
the fatigue strength of a joint class with a detail similar to that under
consideration: or
bl test data for joints which are similar to those requiring assessment.
Such calibration checks should be based upon realistic estimates of the mean
values of the various parameters.
114
l
E The British Standards Institution 2i]1-4
M D A in D
B S _____ E D B _"____D
B R" H E H _ _ _ _ T_
1
A_
E
_"ufll
“fi__E
HU_B
E__-fl_mD_“__
HEB“HE;U
_=u_DEfi_“ £__="fiium-_w___“:MHmu_¢ _fl H_“ “W_ _“_ _ “ _ H_fi=N_ _ _fi
=_ _‘ENE
mcicmu
U-____fi___
_H_fl_m““_ _HEW
_m _ m_“m_ _ _
“fl_m _ _ “m_ _+" _DHhmfl_" ___
______________
________w
______
"_ _ _n_ _ _ _ _ _ _fl_ _D_ _ _ _ Hga
DE
_ _flEmmi
_ _Y
E_:E
u_E_"_w “_m_EDmH_“________
fi_ _ _ _ _
_________“__
_m_#H_ _ "_ _ _ fi_{
_ m_ _ “_ H_d
_m_EI_ _ _ H_
1--I-_-_-I -‘I1 .-1
Eu Eu
_n_ D_‘_U_ _ _W_E_ _ mF_
_____5_m]_m&_m_§mH5
______u
E
Eng:
E_whH:
HE:
E_E_
Rpm:
_ _£E“_u=E_u_ _Er_ _
_u"___
“__fi__
“__D
_ 2fimtm
H“_u___:
_wmmhi
_UuW_Hn_H:m_E=“m_M|_Hm_um=_mMu“_=_ _m_
DED_____
_E_Efi_Em_ _ _ E_um_ _h_“m_mnu
_____D_WEM_HD_EN_ED_ __H_______‘
_EEda
"E&_E___
_ "nu
____“___
:E_a
H____________
__ t
_E"_
_____
_______i
E______r
_mE:_ ED_2_ E_
mfi______i
_Wmm?
_wH_ Im“W_I1HH_m__u
_%
E
E T h E E r ___“ _“ h 5t E H d H M S H H ___H_
_m H __J_U 4
1
-
1 1
S
BRITISH STANDARD
BS TBBBIZIII14
Annex E
[normative]
E.1
Fatigue testing and the use of test data to define
design stresses
introduction
Fatigue testing might need to be carried out for two different purposes:
ai
to establish the relevant S-ill curve, or joint classification, for the design of
some detail which is not adequately covered by the classifications given in
Clause 12; or
b]i
to establish whether some prototype structure is capable of carrying the
fatigue loading expected during the service life of the structure.
in al. the objective is to obtain a design 5-ill curve, under constant amplitude
loading, in a similar manner to that used for the standard classes lsee hut}.
Statistical methods that are consistent with BS ISD ‘i2'llIlI-' are provided in this
annex. In fatigue acceptance testing, the objective would normally be to apply
to the component or structure loading simulating that to be expected in service.
Guidance on the design and production of welded fatigue test specimens is
provided in PD lSDiTft 14345.
Any fatigue tests should be performed using equipment andlor testing machines
with known calibration le.g. using equipment that conforms to
B5 EH ISD Title-ll. The specimens should not be overloaded prior to fatigue
testing.
E.2
I5.1.1
Fatigue tests to establish joint classification
Fatigue test procedure
Fatigue testing for joint classification purposes involves carrying out constant
amplitude tests under tensile cyclic stresses. In the case of welded specimens,
ideally these should be full-scale structural components but valid data can be
obtained from specimens incorporating the weld detail of particular interest
provided plate and weld sizes are full-scale and allowance is made for the
likelihood that they would not embody the high tensile residual stress that is
likely to exist in the actual structure. Ways of doing this include local spot
heating, the introduction of additional weld beads normal to the weld of
interest or performing the test at a high positive stress ratio or under conditions
of high tensile mean or maximum stress. Tests may be performed at various
stresses, with repeat tests at some of them, selected so as to give endurances to
failure reasonably evenly distributed over the linear part of the relationship
between log lstress rangel and log {endurance}, {l.e. typically over the range rte
to 2 x lot" cycles}. Alternatively, all tests may be performed at the same stress
level.
It is usually appropriate to test not less than eight nominally identical specimens
representative of the detail under consideration.
If the detail is subsequently to be used in an environment other than air at
normal ambient temperatures, then the service environment [e-g- corrosion
conditions, temperature] should be simulated in the fatigue tests. In those
circumstances it is also important that the loading frequency should be similar to
that expected in service.
11E
l
E The British Standards Institution 2iIi14
BRITISH STANDARD
E.1.Z
E.2.2.1
BS TEBBIIBIA
Tests performed to validate a design 5-ill curve
Approach
Use is made of the mean 5-iii curve for the class selected, and its standard
deviation of log f-l. Sflg. initially the assumption is made that the new test results
form part of the same population as that used to determine the design S~l'il
curve {this is called the null hypothesis]. Then, hypothesis testing is used to show
that, under this assumption, it is very unlikely {at a specified significance levelj
that the new results would be as high as they are. This is the basis for regarding
the null hypothesis as implausible. and for accepting the alternative hypothesis
that the new results actually belong to a population having longer fatigue lives
than the main database. Thus. use of the selected class would be valid.
Assuming a Sill significance level, the condition for accepting the alternative
hypothesis is:
i
i.scs.so,,.
IDQ N[Egt Z Il'J'glll|l,g' +
'i' n
l'E. ll‘
where
logNy.,.,y
stress;
is the mean logarithm of the fatigue life from the tests at a particular
logl'il,, is the logarithm of the corresponding fatigue life from the mean S-lil
curve for the design class;
n
is the number of fatigue test results.
NOTE I
The value l.iErt_'i .|'s obtained from standard normal probability tables for a
probability of 0.95.
NOTE 2 The 5% level of significance is commonly considered to give a sufficiently
low probability of concluding that the populations are different in the case where
they are actually the same. Alternatives include l..2B5 at the lfili-t level of
significance, l.S'ElIl at 2.5% and 2.33 at lit.
Equation E.1 can also be expressed:
—-
l.-Ii-4S sol
llliliestalllfdif lllll
I
{E.I',l
where:
S‘j".l'il= C,,x IDIITE-fl
-.|-1
-
is the geometric mean of the test fatigue lives.
lsasic assumptions are that:
al the slope of the mean s-rv for the test results is the same as the slope rn of
the design curve, and
bl
the standard deviation of log I'll about that mean 5-ill cunre {assuming that
its slope is ml is the same as that for the database that produced the design
curve. {l.e. Stlgji.
Tests are available to check these assumptions if there is any uncertainty
isee 2tll-
El The British Standards institution 2[l14
1
112
BS TBBBIZB14
BRITISH STANDARD
E.2.I.I
Tests performed at the same stress level
Each specimen is tested to failure- Equation E.2 is then satisfied, where filgls
the geometric mean fatigue life obtained at the stress level used and n is the
total number of tests. Unless the new results lie on an 5-ill curve with the same
slope as the design curve. this approach only validates the selected class at the
stress level used for the tests.
E.2.2.3
Ftepeat tests at a number of stress levels
Each specimen is tested to failure. Equation E.2 is then applied in turn for each
stress level. This approach validates the design curve over the range of stress
levels used, even if the s-rv curve for the new test results does not have the
same slope m as the design curve for the selected class.
E.2.2.4
Tests performed to produce an S-lll‘ curve
The specimens are tested to failure at various stress levels and an s-rv curve is
fitted by regression analysis {see A.1}. This should be of the same form as the
selected design curve and have the same slope m {see E.2.2.1} such that
Equation E.2 can be modified to compare this curve and the mean S-ill curve for
the design class being validated to give the condition:
c.,.,.,, e Cj;,xllIll%lm5'
lE-31‘
Consequently, the mean curve fitted to the test results is expected to be on or
above the following 5-lil curve:
.
,
s;".rv=c,,a1crl%_i
rice
ltather than testing every specimen to failure, Equation E-1‘-i could be used as a
target curve to be achieved in every test. In this case the specimens fatigue
tested do not need to fail, simply to achieve lives on or above the target curve.
E.3
Fatigue acceptance test
The objective of the fatigue acceptance test is to establish if some prototype
component or structure is capable of carrying the fatigue loading expected
during its service life-
iilihere the service loads vary in a random manner between limits, they should
be represented by an equivalent series of variable amplitude load cycles.
Ciccasional high service loads in the test spectrum should not be
unrepresentatively large or too numerous as, if they are, the fatigue lives which
are obtained might not be representative {due to the fact that occasional high
stresses can retard fatigue crack growth}.
Alternatively, the test could be performed under constant amplitude loading at
the maximum imposed service load. with the required number of repetitions
being estimated to produce fatigue damage equivalent to that produced by the
actual senrice loading spectrum. allowing for fatigue crack growth retardation. if
relevant.
As in the case of tests to establish the classification of a joint {see E.2}, if the
structure is to be used in an abnormal environment then the environment and
loading frequency should be simulated in the acceptance tests-
The geometric mean life obtained from the effective number of specimens
should be at least equal to the design life multiplied by the factor. F. from
Table E.1.
11B
l
E The British Standards Institution 2i]14
BRITISH STANDARD
BS TEBBIIDIA
Dwing to the great increase in scatter for tests carried out near the fatigue
limit, stresses should be selected to give specimen lives not exceeding
2 x ID‘ cycles.
Table E.1
Fatigue test factor, F
lrlumber of test result. n
Fatigue test factor. F
I
S-4
2
3
4
‘ill
4.3
3-S
3.?
3.2
MUTE
Test factor obtained from Equation E.2 modified to compare filmy, assuming
SD = il.2.
Annex F
{nonnative}
F.1
Weld toe improvement techniques
Background
The weld toe is a primary source of fatigue cracking because of the severity of
the stress concentration it produces. Apart from a relatively sharp transition
from the plate surface to the weld, dependent on the weld profile, the stress
concentration effect is enhanced by the presence of minute crack-like flaws,
extending to depths {below any undercut} of a few tenths of a mm [25, 2?].
These flaws are an inherent feature of fusion welds and are not regarded as
defects- Fatigue cracks readily initiate at these flawsThe weld toe improvement methods described in this annex rely on two main
principles:
a}
Reduction of the severity of the weld toe stress concentration. Two methods
are given, burr grinding and re-melting by TIE or plasma dressing. The
primary aim is to remove or reduce the size of the weld toe flaws and thus
extend the crack initiation part of the fatigue life- A secondary aim is to
reduce the local stress concentration due to the weld profile by achieving a
smooth blend at the transition between the plate and the weld facebl
introduction of beneficial compressive residual stress. An alternative. or
additional, approach is to introduce beneficial compressive residual stresses
in the weld toe region. These have the effect of clamping the weld toe in
compression, with the result that an applied tensile stress would first
overcome the residual stress before it became damaging. The applied stress
range is therefore less damaging. This annex covers techniques that all
achieve this aim by plastic deformation of the weld toe region, namely
hammer, needle. high-frequency and shot peening. Compressive residual
stresses are then produced as a result of the constraint imposed by the
surrounding elastic material.
An important practical limitation on the use of improvement techniques that
rely on the presence of compressive residual stresses is that the fatigue
performance of the treated weld is strongly dependent on the applied mean
stress of the subsequent fatigue loading. In particular, their beneficial effect
decreases as the maximum applied stress approaches tensile yield. disappearing
altogether at maximum stresses above yield. Thus. in general the techniques are
not suitable for structures operating at applied stress ratios {R} of more than DA
or maximum applied stresses above around Bl]‘lirr yield. The occasional
application of high stresses, in tension or compression, can also be detrimental
in terms of relaxing the compressive residual stress.
Er The British Standards institution 21314
1
115'
BS TBBBIZB14
BRITISH STANDARD
The techniques in this annex are particularly suitable for treating transverse fillet
weld toes. Although in principle they could equally be used to treat butt weld
toes, in practice it is usually more convenient simply to grind the butt weld
overfill flush to improve its fatigue performance. The information provided in
this annex refers specifically to transverse fillet weld toes, including the toes of
external fillets in full or partial penetration joints and the toes of fillet welds
that pass around the ends of longitudinal attachments. and the specified extents
of improvement in fatigue performance resulting from the use of the
improvement techniques apply only to such cases.
This annex provides only general guidance on the application of the
improvement techniques. For more detailed guidance reference can be made to
llllll Recommendations I29]. The extent of the improvement in fatigue
performance resulting from the use of the techniques is also defined in this
annex-
The improvement techniques covered by this annex are intended to increase the
fatigue performance of the welded joint treated with respect to potential
fatigue failure from the weld toe. In practice, the possibility of failure initiating
at some other location with little improvement in fatigue life should be taken
into account. This is especially relevant to joints with load-carrying fillet or
partial penetration welds where fatigue cracking might still propagate from the
weld root. Even nominally non-load-carrying fillet welds can fail from the root
when the toe has been improved. Consequently, when weld improvement is
planned, full penetration welds or fillet welds with extra~large throats should be
used where possible. particularly for welds at the ends of cover plates and
longitudinal stiffeners.
in the case of a multi—run weld, more than one weld toe might need to be
treated. This is particularly the case in situations where the stress distribution
over the weld surface is fairly uniform, such as a nodal joint between tubes of
nearly equal diameter under out-of-plane bending {Figure F.1}; treatment of the
weld toe on the member surface would merely transfer failure to an adjacent
toe between surface beads of the weld. Consequently, some or all other weld
toes would need to be treated, in which case the most practical approach would
be to grind the whole weld surface rather than to treat individual weld toes
{see Figure F.1}.
Figure F.1
liliulti-run weld in tubular nodal joint requiring improvement of every weld ‘toe
//*1»-22
l
A
If the stress over the weld surface is fairly uniform grinding the toes A merely transfers failure to the
inter-run toes ii and C- In such situations the whole weld needs to be groundThis annex applies to the treatment of weld toes on members at least E mm
thick in any arc welded steel structure that is subjected to fatigue loading. Due
to lack of experimental data for extra high strength steels, the specified
improvements in fatigue performance apply only to structural steel and stainless
steel grades up to maximum specified yield strength of E-lliti lrlimmf. However. it
is reasonable to expect that. in principle, the methods would also improve the
fatigue performance of welded higher strength steels, and welds on members
less than E mm in thickness. In the absence of relevant published data, such
benefit should be quantified by special testing {see Annex E}.
12[l
1
E The British Standards Institution 2i]14
BRITISH STANDARD
BS TEDBIIDI4
with regard to service loading conditions. the tensile stress limitations ln 15.1
still applv. The specified benefits of burr grinding and TIE dressing ezctend to
low-cvcle fatigue strain cvcling conditions in regions of geometric stress
concentration. However, special restrictions are imposed regarding applied pealc
stresses and stress ratios in the case of the peening techniques.
F.1
Preparation
The ease of application and benefit from anv of the improvement techniques
covered by this annex: is improved if the weld to be treated has a favourable
profile with a well-defined weld toe. A favourable profile has a low weld toe
contact angle, 45 or less, a generous weld toe radius, ‘i mm or more, and
freedom from weld toe undercut or cold laps. Therefore. if it is planned to use
an improvement technique, it is advisable to trv to produce fillet welds with
such features. If the weld to be treated has a poor profile, it is advisable to
grind a shallow groove along the toe to help guide the improvement technique
tool l_T|G or plasma torch, peening hammer. etc.l. Such preparation might not be
required before shot peening, but light grinding maires it easier to checlt that
the treatment has been applied correctfv and, reduce the stress concentration. In
the case of the treatment of welds that pass around the ends of longitudinal
attachments, including corner gussets, it is also advisable to use full-penetration
welds, at least for the first 3|] mm, to reduce the rislr. of fatigue failure from the
weld root. Further benefit cornes from tapering the attachment end to provide
a smoother transition to the main plateIn all cases the weld to be treated should be de~siagged and cleaned bv
wire-brushing to remove scale, rust or paint. It is also advisable to remove traces
of oil or other surface contaminants before TIE or plasma dressing.
F.3
F3
Weld toe dressing techniques
Burr grinding
The prlmarv aim of grinding is to remove or reduce the siae of the weld toe
flaws from which fatigue craclcs propagate and to reduce the local stress
concentration effect of the weid profile bv srnoothlv blending the transition
between the plate and the weld face.
Toe grinding is normallv carried out with a high speed pneumatic, hvdraulic or
electric rotarv burr grinder with rotational speed from 15 lZllJlZl to so U-Ell] rpmThe tool bit is normally a tungsten carbide burr ior rotating filel with a
hemispherical end. To avoid a notch effect from too small a groove radius fr}, it
should be scaled to the plate thickness ft) at the weld toe being ground and the
grinding depth id} such that the root radius of the groove is no less than lIl.25t
or 4d {see Figure F2}. If a suitable burr for achieving such a radius directlv is not
available, as is liicelv to be the case with thiclc sections, the final groove might
need to be produced bv blending the sides of the sharp groove bv additional
grinding.
Er The British Standards institution EH14
1
TIT
BRITISH STANDARD
B5 ?El]'B:1lII'll'l-
Figure F..."-I
Recommendations for weld toe grinding
d = min. I15 mm
below undercut
l‘l'I ::- I125
1:
,..,
I
‘l
rid 2 -‘-I
1
-H
2
€,..s"\
3
'5
1.
1
E
uor To scats
lfey
Ground profile
4
Minimum throat to be maintained
Original profile
'5
E
Grinding depth
Weld root
Original toe
LI.lI"-i.I—I
The tool is centred over the weld toe and pushed or pulled along the toe to
produce a smooth concave profile at the weld toe. The depth of the depression
penetrating into the plate surface should be at least I15 mm below the bottom
of any undercut or flaw at the weld toe, which can be located by magnetic
particle or liquid penetrant inspection {Figure F.3l. The maximum depth of local
machining or grinding should not exceed 2 mm. lf the reduction in plate
thiclcness exceeds 5% this should be talcen into account in the stress calculation.
The ends of fillet welds, such as those attaching longitudinal members, can only
be treated effectively if the weld can be carried around the end of the
attachment member to provide a distinct weld toe.
Where toe grinding is used to improve the fatigue life of fillet welded
connections. care should be talcen to ensure that the required throat thicltness is
maintained.
The final ground surface should be inspected by suitable l‘-i[.lT to confirm that
the weid toe has been removed completely and to ensure freedom from
surface—brealting flaws that might have been exposed as a result of the grinding.
Figure F.3
Toe grinding to improve fatigue strength
'F_
.-"'
al Stress in member I
T21
l
E The British Standards Institution EH14
FT
BRITISH STANDARD
Figure F3
BS TEBBIIBI4
Toe grinding to improve fatigue strength
-IIII-ii}
‘|_
2—-.
&\\‘“
3-o
NDTE Erinoling a weld toe tangentially top the plate surface as at st produces littfe improvement in strengthlSr.|'ncl'r'ng shouiol extend beiovv the surface as at B, in order to rernove toe fi'.a|.-vs-
bl Stress in member "r"
ltey
1 lvlember ‘r'
2 Flaw
3
4
|=.a.2
Member x
Depth of grinding should be at least f.l.5 mm but s it mm
below bottom of any flaw at the weld toe
I
TIE or plasma dressing
The objective of TlE or plasma dressing is to remove the weld toe flaws by
re-melting the material at the weld toe and to reduce the local stress
concentration effect of the local weld toe profile by providing a smooth
transition between the plate and the weld face.
Figure F.sl
Effect of TIE or plasma torch position on resulting weld profile
—-i—|—'—2
~+|-2
\
ii
liley
‘l
Elptimum shape
3
Toe :lIl.5 mm
2
tl.5 mm to 1.5 mm from weld toe
4
Unacceptable shape
Standard TIE or plasma welding equipment is used, without the addition of any
filler material. Argon is normally used as shielding gas. The addition of helium in
TIE dressing is beneficial as this gives a larger pool of melted metal due to a
higher heat input. However, one advantage of plasma dressing over TIE is that a
wider re-melted material pool is produced.
El The British Standards Institution 21314
1
IE3
BS TBBBIZIII14
BRITISH STANDARD
if pre-heating would be required for welding the steel concerned then
pre-heating to the lower end of the range of recommended temperatures
should be used before TlE or plasma dressing. In both cases the position of the
welding torch arc with respect to the original weld toe should be controlled.
Trials should be carried out before applying TIE or plasma dressing to establish
the correct conditions. The aim is to produce a profile that blends smoothly
from the weld face to the plate surface, as shown in- This should be possible
with the arc centred close to the weid toe on the plate side. ideally iIl.5 mm to
1.5 mm from the toe. An unacceptable result, Figure F.4, may be produced if the
arc ls too close to the weld toe, within lJ.5 mm, and the dressing conditions
should be modified to prevent this.
Further details of the techniques can be found in Iii‘, ES and Ellll.
Benefit
F.3.3
The classification of welds might, where indicated in Table ll to Table ll], be
raised when dressing of the weld toes is carried out in accordance with this
annex isee also 13.2}.
For assessments based on either nominal or hot-spot stresses. the S-I'll curve for
the untreated weld can be assumed to be increased in fatigue strength at ltli
cycles by a factor of 1.5 and the 5,-N curve rotated to a slope of m = 3.5, as
illustrated in Figure F.5.
Weld toe grinding should not be assumed to be effective in the presence of a
corrosive environment as this can cause pitting of the dressed surface. The full
potential of weld toe dressing might not be achieved if fatigue craclcing could
initiate at a location other than the treated weld toe. The design process should
still address all potential sites for fatigue cracic initiation and classify each one
accordingly.
Figure F.5
'
f-lodirficatlo n to design 5-N curve for untreated weld resulting from weld toe dressing
T
Slltl
I
fililfl
I
IIIII_l|
I
I
fI|lIl_|
I
I
IIII1I|
""1-.
""--.
I"rlltl—
""1--.
""1-.
-..H""'-i
iflfl -
--..___~_
""-I-.
T5-l'J —
ER
E
q
ri
~<
1
--..____h
‘H"""'-.
I-Iil -I ‘?EB-Z-IE
5l'=1'5'5'oc
*"""".,
III‘-Z
SII -
555::
Lil
ii
fltl
ill‘
IIIIl_lI
I
||-||r||.J
1ll5
I
I
||||||
til‘
ts‘
K
II‘-iey
it Endurance ili, cycles
1"
Stress range Sr Iilimmi‘
1 Toe sf-izrv = n.ss,¢..;i 3-5.11.1?
.1:
Untreated weld sfsv = .-t
Finally, the fatigue strength of toe dressed welds is limited by that of the parent
material. Thus, for applied stress ranges greater than that where the s-rv curves
for the treated weld and parent material [class B} intersect, the lower of the two
S-I'll‘ curves should be used.
12-II
l
El The British Stan dards Institution Iilll-1
BRITISH STANDARD
BS TEBBIIBI4
Weld toe peening techniques
F.4.1
Eene ral
Four weld toe peening techniques are covered by this annex:
all
hammer;
bl
needle;
cl
high-frequency; and
dl
shot peening.
Methods a} - d} aim to induce compressive residual stresses by plastic
deformation produced either by repeatedly hammering the weld toe region
with solid blunt-nosed tools or a bundle of small-diameter rods. both with
smooth, rounded tips {Figure E5}, or by directing high-velocity shot at the weld
toe. In use, the tips of the solid tools wear, causing flattening, and the surface
becomes rougher. Therefore, the end shape should be checlced after every use
and, if necessary, the tool tip re-ground and polished.
Figure F.E
Weld toe peening methods
rm
a]l Hammer peening
//
bl llleedle peening
cl High frequency peening
In the case of the hammering techniques, the tool is centred over the weld toe
and moved along it so that the repeated hammering produces an indentation at
the weld toe. Effective treatment requires reasonably accurate positioning of
the tip of the tool over the weld toe, which is facilitated by prior grinding, so
that metal on each side [both weld metal and parent plate} is deformed. This is
normally achieved by supporting the hammer firmly with the tool between 45
and Ell“ to the plate surface and approximately perpendicular {TF5 — ".-ltll to the
direction of travel {Figure F.i'} and lteeping the tool tip in close contact with the
weld toe as it is moved along. The peening operation is repeated until there is a
clear Indication of uniform plastic deformation in the form of a groove with a
uniform surface appearance. In the case of needle peening this is a surface
bright in appearance with a uniform distribution of small indentations.
El The British Standards Institution EH14
1
115
BS TBBBIZIII14
Flgure FJ
BRITISH STANDARD
Weld toe peening
‘I
In
. ___
_
2
/is 9
Er
.\\£5-///A
Itey
I
Bright surface
2
Feening depth Ill mm to I15 mm depending on
peening method
in order to ensure full coverage of the weld toe region. the following procedure
should be followed.
al
Apply needle peening until the relevant area is free of untreated spots,
checlt for ltlllbli coverage using a low power magnifying glass and record
the peening time talcen to achieve this.
bl Repeat the peening for the same length of time to achieve what is termed
ac-are coverage.
c}
A satisfactory result can normally be achieved with a total of four passes.
in the case of hammer or high-frequency peening the requirement is for a
surface that is smooth and free from obvious individual indentations. Again,
four passes are usually sufficient. Surface flaws should also be avoided. In this
respect. both peening methods. especially when applied to pealty or severely
convex weld profiles, can cause the piastically deformed metal to fold over the
original weld toe and leave a cracic-like lap. feature. The resulting fatigue
performance of the welded joint might actually be less than that of the original.
Therefore, it is advisable to grind the weld lightly before peening to improve its
shape and create a groove which facilitates a steady movement of the peening
tool. Any signs of folded laps or heavy flalcing of the surface should be removed
by local grinding, followed by re-peening.
Table F-1 summarixes the main characteristics of the equipment and objectives of
the three hammer or needle peening techniquesAn important feature of shot peening ls the need to ensure full coverage of the
surface to be treated. The usual technique for assessing adequate shot peening
coverage is to apply a fluorescent dye before peening and then inspect it with
the aid of a blaclt light. Checlcs can be made during shot peening so that
particular attention can be paid to regions where access by the shot is restricted.
its all the peening techniques rely on the retention of compressive residual stress
they cannot be relied upon to provide any benefit unless they are applied after
any operations that could introduce tensile residual stress ie.g- the assembly of
components when fit-up is imperfect].
125
l
El The British Standards Institution Iilll-1
BRITISH STANDARD
F.4.Z
l=.rfi.2.1
BS TEBBIIBIA
Practical comments
Hammer peening
The diameter of the tool tip influences the resulting appearance of a hammer
peened surface. In general, the smaller the diameter, the greater the lilrelihood
that the actual weld toe itself would be peened and eventual-ly disappear.
Peening with a large diameter tool {greater than 12 mm} does not usually reach
the weld toe but instead deforms material either side of it. Although in general
the desired effect is achieved with fewer passes using a large diameter tool, the
presence of the original weid toe is a disadvantage from the viewpoint of
inspection. in particular, it is not obvious that the toe has been correctly treated
ii.e. left in a state of compressive residual stress} and remnant traces of weld toe
confuse in-service inspection as it is difficult to distinguish between them and
fatigue craclcs. The use of a small diameter tool, or a combination of small and
large diameter tools, with the aim of deforming the actual weld toe offers the
best compromise. Inspection would then ensure that all traces of the original
weld toe had disappeared.
Hammer peening, even using modern silenced hammers, is a noisy operation;
the operator and others woriring in the vicinity should use ear protection.
‘vibration from peening equipment can cause physical discomfort or harm, and
the operator should wear vibration-damping gloves and not perform the
operation for extended periods of time.
El The British Standards Institution EH14
1
1.’-11-'
BS TBBB:ZlII1l'I-
Table F-1
BRITISH STANDARD
Summary of weld toe peening methods
Details
_
Tool
Machine
Method
Hammer peening
Needle peening
High frequency peening
3—5'.Irnm diameter
round-tipped chisel
Bundle of round-tipped
steel rods {e-g- around
Cine or more singie- or
double—radius ended tools,
thirty 2mm diameter
needles}
Standard weld
de-siagging needle gun
Pneumatic, hydraulic or
electrical impacting
hammer
typically 3—5mm in
, diameter
Purpose-built impacting
hammer activated by
vibrations produced by
ultrasonic transducer or
pneumatically
Typical
hammering
I 25-1l]tl blows per second
frequency
i
Typical travel
soo
4tl blows per second
till}-4tltl blows per second
sou
sop
speed,
m-i-elder-._
Number of
Typically 4
passes
_
_ _
_ I
Typical depth of I iIl.'lS—lIl.S {lI.'l.lSmm
plastic
Typically 4
I
_ _
D.1-i.'l.2
i
As required to produce
_ req_uired surface finish
fl.1S-B-4
minimum required]
deformation,
E15"
.=
Required
surface finish
I Uniform groove with
smooth surface free from
.-
.;
Bright surface with
uniform distribution of
Uniform groove with
smooth surface free from
obvious individual
small indentations over
obvious individual
indentations
complete surface
indentations
F.4.2.2
Iifeedle peening
lileedle peening is noisy and ear protection should be used. However, it is less
noisy than hammer peening and the equipment is lighter and generally easier to
use than hammer peening equipment.
F.4.2.3
High frequency peening
The relatively low frequencies of hammer and needle peening mean that their
effectiveness depends on the pressure on the tool against the treated surface
[typically at least Ill ltgfl. The result is that the vibrations of the tool are
transmitted directly to the hands of the operator and some effort might be
required to maintain alignment of the peening tool along the weld. In contrast,
the high-frequency peening methods are based on the generation and
utilization of impacts from very high frequency vibrations, with the result that
effective treatment is virtually independent of the pressure on the tool ii-3 ligfl.
lvoise and vibration are also considerably lower. In view of these factors, the
high frequency peening techniques offer advantages over hammer and needle
peening in terms of ease of use, health and safety and quality control of the
peening operation.
12B
1
El The British Standards Institution IBI4
BRITISH STANDARD
F.4.I.4
BS TEBBIIDIA
Shot peening
Shot peening entails impacting a surface with shot iround metallic, glass or
ceramic particles} varying in diameter from 5i] |.-tl'1'l to E mm with a force
sufficient to create plastic deformation. it is a non-contact process so that
particles are fired at the surface remotely and can be applied manually but
generally in an automatic or mechanized manner for consistency of the residual
stress profile. Checking for that consistency can be achieved using fluorescent
dyes sprayed or brushed onto the surface before shot peening and subsequently
checlired using a blaclt light during and after processing. Any residue of
fluorescence, which might be localized if it reflects the presence of a sharp
discontinuity lilte a steep weld profile or weld toe undercut, provides an
indication of inadequate coverage and the need for further peening. The
selection of shot size and type depends on the material and geometry of the
part and the strain sensitivity of that material.
F.r'-L3
Benefits
The classification of welds might, where indicated in Table 4 to Table til, be
raised by peening the weld toes isee also 13.21. However. in contrast to the
behaviour of toe dressed welds the improvement in fatigue strength depends on
the applied stress ratio, ii‘, and the maximum applied tensile stress, such that the
benefit decreases with increase in either. If practical, some compensation for this
problem arises if the peening operation is carried out while the welded joint is
subiected to a tensile stress. For example, such an approach might be possible if
part of the tensile stress experienced by the joint in service is due to the dead
load on the structure. However, in general the peening techniques are not
suitable for structures operating at applied H 1» ti-4 or maximum applied tensile
stresses above around sass yield. Similarly, their value is questionable under
loading conditions that include the occasional application of high stresses, in
tension or compression, as they can be detrimental in terms of relaxing the
compressive residual stress.
With the above limitations, on the basis of the application of weld toe peening
in accordance with this annex, for assessments based on either nominal or
hot-spot stresses the fatigue strength of the untreated weld at lit’ cycles may be
increased by a factor of up to 1.5 and the slope of the 5,-nl curve changed to
i"l"f = 3.5, depending on Fl and the maximum applied tensile stress l5,,,...l. as
indicated in Table F.2 and Illustrated in Figure F.B. As in the case of weld toe
dressing, the fatigue strength of peened welds is limited by that of the parent
material. For applied stress ranges greater than that where the S-N cunres for
the treated weld and class B intersect, the lower of the two 5-I'll cunres should
therefore be used. Similarly, if curve 2 in Figure F.tl crosses curve 3 then curve 3
should be used for higher applied stress ranges.
Table F...‘l
improvement in fatigue strength due to weld toe peening
Conditions
Improvement
it -c o
Increase fatigue strength at iii’ cycles by factor of 1.5.
fl -c fl‘ E. [l.2E.
Sum s EH55 yield
[l..?!B -=: if s i].4,
change slope of S,-N curve to m = 3.5 and treat detail
as stress-relieved in accordance with 15.3.6.
increase fatigue strength at ltli cycles by factor of 1.5.
change slope of 5,-iii curve to m = 3.5.
lncrease in fatigue strength by factor of 1.15 but no
Swine Biifl-"ii___§,iIEI-til
change in slope of S,-N curve.
Fl’ :=- l].4 or Sm“ :=- Blfllliil
yield
hlo benefit unless proved by fatigue testing
{see Annex El
El The British Standards Institution 21314
1
1.’-ISI-
BS TBBBIZB14
Flgure F.E
BRITISH STANDARD
Modification to design 5-Ill curve for untreated weld resulting from weld toe peening
‘I
I
L
If
5-D“
I
ll-l.'IlIl
""'-1.
3lIIl'-
I
I
I
I
I
L
-I
_|
I
|
I-
I
I
I
I I-|
I
I
I
|
|
I
I
I |
"'~=...""-:,‘__
""'-1.1-_
""-1.“-,,
"'|1i,.-In
illfll -
""1-.“-.
"'1-,“.,,.‘
IS-ll -
"'5...
5‘
IIII.‘
1-.‘
""1|.
_
_
‘III-‘-
“H
"'I-qh
-.,
“*-
1,‘-
‘-.."'I-.\....“'I||_‘
s=1.ss,..
—
|
.,___ ‘___‘.
si=1.1ss..
‘1-
_3_
s.-I 2
Sr=S|c
l"'
|J"I
D'I-I-I TZZ
I IDEIEEI-Z‘-3'5
an ._.
an
J. .|._....|._.|....s...s.z..l.
id
Lu
...J. as .r.....L.a._.|..sJ.....
1Ill5
.|.
.1.
tr‘
pl
if
Curve I: Basic design curve for untreated weld if it 5- tI.4 or Sm” 5 E-I:I'Il"ilI yield
Curve 2: SEN = l1.I5S,,,,[jf.1[li if {LIB -c Fl 5 I14
Curve 3: 5,3 if iii = lI.5S_HF'-".1i]i
al
H -c El, 5, _,,_,,.,,_,, = Sm" + lZl.l5S,,,,,,
bl
El ‘S Fl 5- lJ.El3 and Sm“ i’ Bflflb yield, S, ,,,.,,,,,,,, = 5,
ltey
it Endurance N, cycles
"r" Stress range 5,, lllimmf
Al'Il 1Eh‘. rs Assessment of tubular node joints
fnormativel
o.1
Fatigue of tubular joints
The stress range that should be used in the fatigue analysis of a tubular node
joint is the hot-spot stress range at the weld toe. This should be evaluated at
sufficient locations to characterize fully the fatigue performance of each joint.
For example, in the case of a tubular seteon connection at least four equally
spaced points around the joint periphery should be talcen into account. For any
particular type of loading, this hot-spot stress range can be defined as the
product of the nominal stress range in the brace and the appropriate stress
concentration factor l[5EFi, as indicated in Figure E.1, which also details the
nomenclature used to describe tubular joints.
13B
1
El The British Standards Institution IBI4
m DARD
BRHE H 5T
B S _____ E D B _"____B 1 4
_ __
%
m_H__EE_
minm$L_fi_____$
m_ _D__
_ _ H_
m_m_ _h_ _ _
E
m
numb“_ _ “ fi_fl _:“_ H_ mfi_ _ $_ _ )w
_E
mE_H“:_L"Hm_E:EMH_D_:|hU_L_
_HE
m
_m__E_ _ _u_m_“ W_m _ U_ L_ Ema“
pl
_u_E
mug
m_ _ “_E_ fi_E _H _
_nEDm_
_ _Hh_ _m“l_’1m
_jwNI
U_HE___w__
fl_ n_m_ u_ _P_fi_ _ __
$x2:§_ m_“_H_m
_H
E _ _ _ _E?fl_$_ _“ M
E:mumhn
F_:__
_ E_m _ _H m_ “_w
H_m_ _“_fi“_ "m_
M
u_ _ _ _ “_u_ H_ H4|
_q
Nm___
U_Em_
_M
H_m_ _ _ _
J_W
_ "m__ _ _H_ Em
_ 5_ _ _ u_ _D_ fi_m _ _ _ _E D_wm_
_“ _mEH
_ _ ___”
__
_H___‘
_ _m _ _ _ _
IT
EWg_R" _ _UR_
_ F_ _ _ _
hmfi
Egm
m____
E
___m__
_E_¢_
um__5m_I
m:HE:_
_“"_E_ __m_ _m_ _ _u _ _H
JR_
_
m
__
_m _m
é
__|
.
_l_ _|_l_ _
I_|_|1_|_
_m_“_M__: u_?m_H
_u" _ _
Jm
E T h E B r ___“ __h h
EH_ m_ _h ____5_H E _ _ U_“mag
5t
E H d ___" M _ _ _
H
_____H_____“
_m H _ _ _ _U
1
4
-
1
3
I
BS
B R _hE H
_____ B B B _____’B_ 1 4
E_mflmzfiu
fl_ u_ m_ U_h _ w_fiflH_m _m _ _ _“ _w _w
:E__ufi_
_ §_Efi____
_ m_"“Em_ u_ _h_m
_E
_wm_g_ _wm_ h_uh_mmflfi
_HE
_ mzflm
m_ _ _ m_ “_ Eu=_mmmhflm
DE;
E*_ n_£u_ _fiH_ _fi="_
“___"mmh:_ E_ _H_“_H_ m
H__ _ _r E L_u_m_fl_m _u_ mmmh
_ _E_ _u_uE
m
_ _ _E fl_ m_m m_E$_m_ _ m_
_ u_NEH
w_ _ _ _ _
_H_ w_ _ _ _ _
__§__
__
fi
r___ +
M ___
@
"___ _h @
m
_M
______
IP
I
_____ _-I
E
___m_
HHEDM_m_u_n=
:“_E:u___
E_ H2:_“m“_|_m"Em_" _a_H_m_
5
m
Wm
___. 3 I
_'
Q
___u_
“_ :_ _m__E"
“_U _ _ _uh_= E
"_ m_ u
T h _ _ _ B f H E In HW d E rd 5 M 5 t ___“ U t _m _" _ _ _ _ __“_ 4
I
H DAR D
5T
BRITISH STANDARD
BS TBBBIIBIA
In the fatigue assessment of tubular joints using the class Tl design curve
presented in Clause 15, the greatest values of the hot-spot stress ranges on both
the brace and chord sides should be talcen into account.
Stresses arising in tubular node joints
The stresses in tubular nodal joints arise from the following three main causes:
al the basic structural response of the joint to the applied loads {nominal
stressesl:
bl the need to maintain compatibility between the tubes {geometric stresses];
and
cl
the highly localized deformations in part of the tube wall near to the
brace-chord intersection {local stressesl.
Nominal stresses arise due to the tubes behaving as beam-columns, and can be
calculated by frame analysis of the structure.
Eeometric stresses result from the differences in deformation between the brace
and chord under load. For example, in a T-joint under axial tensile brace load,
the brace extends only very slightly, whereas the circular cross section of the
chord becomes significantly elongated to a pear—shape section. The differences
in deformation require the tube walls to bend so that the brace and chord
remain in contact at the welcl. They can also cause a maldistribution of the
nominal membrane stresses around the brace circumference.
Local stresses arise because of the geometric discontinuity of the tube walls at
the weld toes where an abrupt change of section occurs which increases until
the weld root is reached. Local stresses are not propagated far through the wall
thicltness, however, and therefore a local region of three-dimensional stress
occurs. The sharper the corner at the weld toe, and the greater the angle of the
overall weld profile to the tube wall, the higher the local stress,
Hot-spot stresses in tubular joints
Any of the methods described in Annex E may be used to establish the hot—spot
stress but the most widely used, including in the establishment of parametric
formulae for calculating the hot~spot stress concentration factor, is linear surface
stress extrapolation {SSE}. However, the locations of the stresses used to perform
the linear extrapolation for tubular joints are different from those defined in
Annex E for analysing plate joints, as shown in Figure E.2. These are based on
the assumption that the maximum extent of the local region where the stress is
influenced by the "notch" effect of the weld itself is the greater of 4 mm or
El..."-!.,rt, where r and is the brace outer radius and t is the wall thicltness.
MUTE The alternative value of rt mm is only used in practice for very small tubular
joints with disproportionately large welds. The local region in such cases is
determined by the geometry of the weld toe beacl {typically a few millimetres in
slze,l as the stresses rise into the welcl which, being clisproportionately large, forms a
ring of very stiff and inflexible material joining the brace to the chord.
This expression was obtained empirically. though the dependence on the
parameter .rt was originally drawn from a similar dependence on the
wavelength of bending stress in tubes isee references [31] and I32ll. The
hot-spot stress range is the maximum principal stress range at the weld toe,
based on the extrapolated values of the stress components at the weld toe.
Therefore, the maximum and minimum values of each stress component are
determined at each extrapolation point during a load cycle, the corresponding
ranges of each stress component at the weld toe are then determined by linear
extrapolation and finally these are used to calculate the principal stress range-
El The British Standards Institution 31314
1
133
BRITISH STANDARD
BS TBBB:.3B14
Figure E.2
Locations A and El of stresses used for linear extrapolation to weld toes to determine
hot-spot stresses in tubular joints
EH
[
'"--..
"ll-..
l
I
1Sit]
1
L
I
-
._ ah
ssrEl..-T
iii
A:
'
..--"
is wig
l. mi" _r’
I'IiI'I
s»
1.,
l f,-
\
.iei”"
_--"
R“
_--P‘
__.--F’
-"F.-'-_-
Te hillas
TE‘
-rf_g_. ..-W
-
.—.—I_-
.-|-i-I-—
_--"""
_--"'
.'—'-I-I-I
--I
.
'-I-F
_ .--"""
"_hI_
.--"
_._--'
2
'—l_|_._
I__|_-“-I—|
lcey
I
Brace
.3!
E.1I.
Eho rd
Calculation of the hot-spot stress
The calculation of hot-spot stress can be undertalcen in a variety of ways, e.g. by
physical model studies, finite element analysis or from hot-spot SEFs obtained by
use of semi—empirical parametric formulae. The first two allow the hot-spot
stress to be determined at any location whereas parametric equations tend to
concentrate on the positions of maximum hot—spot stress, usually at the crown
and saddle positions- When physical models are used, the geometric stress
extrapolated to the weld toe should be obtained as described in this sub-clause.
'lll.lhen finite element calculations do not allow for any effect of weld geometry.
the hot-spot stress at the weld toe can be estimated from the value obtained at
the bracelchord intersection. Parametric formulae should be used with caution
in view of their inherent limitations," in particular they should only be used
within the bounds of applicability relevant to the formula under consideration.
Extensive guidance on available parametric formulae for both co~planar and
multi-planar tubular connections and recommendations on the choice of
solution is given in B5 Elli ISD ISSUE. Euidance is also given on the effects of
ring stiffening. fvlore recent solutions are included in DI'»li.l~FiF'-I.C.3lI'l3. The hot~spot
SiIFs are defined as:
134
1
El The British Standards Institution 2IEl14
BRITISH STANDARD
BS TBBBIIBIA
SCF
Hot-spot stress
Ivominal
at weld toe considered
stress
in
loaded
if-ill
brace
with the hot-spot stress in terms of the maximum principal stress at the weld toe
obtained by extrapolation as described above.
The parametric formulae were developed by fitting analytical or model stress
concentration factor data to equations which are functions of non—dimenslonal
tubular joint geometric parameters. as follows:
length
H
I
gmean
_
I
radius
mean
mean
T
chord
of
radius
radius
mean
J]
of
chord
of
of
2
cylinder
E.2
fl
E.3
I
cylinder
cylinder
i
E.4
i
cylinder
wall thicltness of brace
wall
(
.
chord
_wall thicltness of chord
i
cylinder
brace
of chord
radius
cylinder
cylinder
I
E.5
1
thicltness of chord cylinder
The form of these equations assumes that some dependence of the parameters
can be deduced logically lby, for example, using simple beam theory to calculate
a component of stress at the chord crown); otherwise a polynomial or power
law dependence is assumed.
Each set of parametric formulae is limited in application in the following three
ways:
al types of joint geometry;
bl parametric validity range; and
c}
loading cases.
There are also restrictions on the range of the other four variables: ct, y, [_l, 1'.
These are due to limitations in the range of original data obtained, or the
methods used to obtain those data, but also because marlred changes in the
nature of the stress distribution occur at extreme values of these parameters. For
example, high Ii iii 5- ii.".ill connections show a radically different peali. stress
distribution from similar lower ,tl connections, with the hot-spot moving from
the normal saddle position to the crown, and for low y connections thin shell
idealization used for finite element calculation is not valid. lilon-compliance with
parametric validity range can result in a considerable increase in the variation of
performance, both conservative and non-conservative, of these equations.
El The British Standards Institution 31314
1
135
BRITISH STANDARD
BS TBBB:..'l.B1l'I-
E.S
Rectangular hollow sections
The class T.l design curve is based on fatigue test data obtained from joints
between relatively large circular section tubes, with wall thicltness generally
above ‘I5 mm. Alternative guidance directed more at thinner {down to 4 mm},
circular and rectangular section tubes is available in BS ISD 1434?.
Annex H
[normative]
H.1
Cycle counting by the reservoir method
Eeneral
The purpose of cycle counting is to reduce an irregular series of stress
fluctuations to a simple list of constant amplitude stress ranges that are lilcely to
produce the same fatigue damage. various methods exist but the most widely
used is the rainflow method, as described in ASTM E1tl4S. However, in practice it
is often necessary to deal with relatively short stress histories, such as those
produced by individual loading events. In such cases, the reservoir method given
in H.2 and shown in Figure H.1 is more convenient. It consists of imagining a
plot of the graph of each individual stress history as a cross section of a
reservoir, which is successively drained from each low point, counting one cycle
for each draining operation. The result, after many repetitions of the loading
event, is the same as that obtainable by the rainflow method.
Figure H.1
Example of cycle counting by reservoir method
i
.
E
r .
E
Q:be
E
'1-r'
I'|'I-
Ir
tr'I-'
T
4
Ur;
K
.-.11.
vf .
l
lte
x Time
l First occurance
I Second occurance
3 |-|ig|-15.51 pggy
H.2
tr’
4
Stress
Imaginary reservoir
5
Shaded areas are those parts of the reservoir which
successively became empty
Method
H.2.1 Derive the pealc and trough values of the stress history, due to one loading
event. Siretch the history due to two successive occurrences of this loading
event. The calculated values of pealc and trough stresses may be joined with
straight lines if desirable. Marlc the highest pealc of stress in each occurrence. If
there are two or more equal highest pealrs in one history, marl: only the first
peak-
H.2.I loin the two marlred points and asses only the part of the plot that falls
below this line, lilie the section of a full reservoir.
1315
1
B The British Standards Institution IEII4
BRITISH STANDARD
BS TEDBIIDI4
H.2.3 Drain the reserveir tram the levrest peint leaving the water that cannet
escape. If there are twp pr mpre equal lpvvest ppints the drainage can be from
any cine pf them. List t:-ne cvcle having a stress range 5,, equal tci the vertical
height pi vvater drained.
NOTE See 15.5 fer lcivv stress ranges that may he ignered.
H.2.4 Repeat H.I.3 successively with each remaining bpdv pf vvater until the
vvhple reservciir is emptied. listing cine cycle at each draining pperatipn.
H.2.5 Compile the final list which centains all the individual stress ranges in
descending erder pf magnitude SH, Sm, etc. Where tvvcs pr mere cvcles cit equal
stress range are recorded, list them separatelv.
H.2.fi Fer nan-welded details enly, a herizental line representing IEFU stress
sheuld he platted and these parts cal‘ the stress ranges in the cempressien zene
medified as in 15.4.
El The British Standards lnstituticsn EH14
1
131-'
B5 TEDBIZD14
BFIITISH STANDARD
Bibliography
Standards publications
For dated references, only the edition cited applies. For undated references, the
latest edition of the referenced document {including any amendments] applies.
Standards publications
ASTM Eton? - i.s5t2ti11le1, Standard practices for cycle counting in fatigue
analysis
EIS 43515-2, Specification for high strength friction grip holts and associated nuts
and washers for structural engineering - Part .2’: Higher grade holts and nuts and
general grade washers
B5 TEHD. Guide to methods of assessing the acceptability of flaws in metallic
structures
B5 EN ittl lall parts}. Lighting columns
BS EH 2ETl—1, Qualification test of welder — Fusion welding - Part 1': Steels
EIIS EH 1992 lell parts}, Eurocode .2 — Design of concrete structures
ElS El‘-I 1993 fell parts}, Eurocode Ii — Design of steel structures
B5 EH 1594 lall parts], Eurocode -ll — Design of composite steel and concrete
structures
BS Ehl 1l]lIl29':2D'lfl, Hot-rolled steel plates 3 mm thiclc or above - Tolerances on
dimensions and shape
ES Ehl 1l]DS1, Continuously hot-rolled strip and platelsheet cut from wide strip of
non—-alloy and alloy steels — Tolerances on dimensions and shape
as EN 131145 tall parts}. tlnii.-ed pressure vessels
B5 EH ISD 1SED5I-1, Specification and qualification of welding procedures for
metallic materials — Welding procedure specification - Part i: Arc welding
BS EH ISD 531?, Welding - Fusion-welded joints in steel, nicltel, titanium and
their alloys fhearn welding excluded} -s tjluality levels for imperfections
BS El‘-I ISD i‘El'Slfl2, Petroleum and natural gas industries — Flared steel offshore
structures
B5 EH ISD ‘i'5I!JD-1 Metallic materials - llerification of static uniairial testing
machines - Part l: Tensionicompression testing machines - Verification and
calibration of the force-measuring system
BS ISD 1210?, Metallic materials — Fatigue testing — Statistical planning and
analysis of data
BS ISD 1434?, Fatigue — Design procedure for welded hollow-section joints —
Recommendations
ouv~e|=~c2os. Fatigue design of offshore steel structures
PD SEED, Specification for unfired fusion welded pressure vessels
PD lSCllTFt:1434S, Fatigue » Fatigue testing of welded components — Euidance.
lSt‘JfTC lll.lli"lSC Jtlll
133
l
E The British Standards Institution 2014
BRITISH STANDARD
BS TEBBIIBI4
Dther publications
[1] MADDCDIL S.J. Background to the revision of BS ?E|'.lS."lS‘S3. TWI Ltd.
Cambridge, E0145"
[2] EURNET T.R. Fatigue of welded structures, Cambridge University Press, Ind
Edition, 1'3i"'EI
[3] MADDDR S..I- Fatigue strength of welded structures. Abington Publishing.
1331
[4] PETERSON R.E. Stress concentration factors, John Wiley and Sons |nc., 1334.
[5] RDARIC .i.R. and "i"CiLlNG W.C. Formulas for stress and strain, NlcGraw-Hiii,
ISITS.
[E] Engineering Sciences Data Unit. Data sheets, ‘tlol. 3.
[7] ‘v'D| 2230:2003, Systematic calculation of high duty bolted joints, Joints with
one cylindrical holt, Part 1, February 2003.
[3] BISHOP N '-iii Ivl and SHERR.4T'I' F: ‘Finite Element Based fatigue Analysis‘,
NAFEMS Ltd., East lcilhride, 2000.
['3] FRDST N E, MARSH it J and PDDR L P; ‘Metal Fatigue‘, ISBN 0193551143,
Clarendon Press, Cntford, 1334.
[I0] TUNNA J- Ivl. Random load fatigue: theory and experiment, Proc. lnstn rvtech
Engrs. Vol 133, No. C3, 1335, pp 243-EST[I1] IHANG 1"-H and MADDCHI-". S..i. investigation of fatigue damage to welded
joints under variable amplitude loading spectra, Int. I Fatigue, ‘ilol 31, No. I,
E003, pp 133-I52.
[12] GI..lRNE"il T.R. Cumulative damage of welded joints, Woodhead Publishing
Ltd., Cambridge. lSBl'~l~l3:333-1-BSSTS-333-3. 2005.
[I3] MADDDR. S..l. icey developments in the fatigue design of welded
constructions. llllIl' Fortevin Lecture. Proc. Ilw lnt. Conf. Welded Construction
for Urban Infrastructure, iSlIv1 fimisoara. ISBN 3?3-3353-1'i-1, 2003.
[141 IHANG '1’-H, MADDDR S i and Riid-EMJDD G R: ‘Re-evaluation of fatigue
cunres for flush ground girth welds’. Research Report 090, ISBN 0 ill TE 2134
I-', HSE Bool-ts, Sudhury, 2003.
[15] MADDDR S I: Fatigue of transverse butt welds made from one side. Welding
and Cutting. vol. 1', No.1. 2003. pp44—5.'l.
[IE] MADDD.‘-s‘. S .l and JDHNSTDN C: Factors affecting the fatigue strength of
girth welds: a re-evaluation of TWl’s resonance fatigue test database, Proc.
CJMAE 2011: 30th International Conference on Clffshore Mechanics and
Arctic Engineering, ASNIE, New ‘forlc, Paper no. ClMAE201‘l—4SiBl.
[13] IHANG ‘ti-H and MADDDN S J: ‘Fatigue testing of full scale girth welded
pipes under variable amplitude loading‘, Proc. of DMAE 2012: Eiist
international Conference on Dffshore Mechanics and .4rctic Engineering.
ASME. New ‘foils, Paper no. CIMl‘sE2012-03054.
[13] NlEIv'li, E, FRICRE, W and MADDEN, S. l. Fatigue analysis of welded
components, Designers guide to structural hot~spot stress approach. ISBN 133?B~1-B45513-124-0 ll'W, Woodhead Publishing Ltd., 20013.
[13] MADDEN SJ, Hot-spot stress design curves for fatigue assessment of welded
structures, Int. J Dffshore and Polar Engineering, 12112}, 2002. I10 134-141.
31
In preparation.
D The British Standards institution 2014
1
133'
BS 2BOB:201l'l-
BRITISH STANDARD
[20] SMITH, S, MADDDR, S. J, HE. W. 2HDl.i, D and SARASWAT. R. Computer
based fatigue analysis for welded joints, 'l'lilil industrial Members Report
9SSa, 2010.
[21] RIACI-, 2. G and "'i'AMADA, It. A method of determining geometric stress for
fatigue strength evaluation of steel welded joints, Int J Fatigue, 25, 2004, pp
1222-1293.
[22] DUNE. P. HDNE. H. It, DSAEE, D. A and PRAGER. M. Master curve method
for fatigue evaluation of welded components’, Welding Research Council
Bulletin 424, Welding Research Council, New ‘forlc, August, 2002.19.
[23] 'ii'ii“i"'LDE, J.Ci. and MADDDH, S.l. Effect of misalignment on the fatigue
strength of transverse butt welded joints, in Significance of deviations from
design shapes, l.Mech.E. Conference Publications, 1929.
[24] BURDERIN, F.M. The effects of deviations from intended shapes on fracture
and failure, in Significance of deviations from design shapes, l.Mech-EConference Publications, 1929-2.
[25] MADDCIIII, S.J. Fitness for purpose assessments for misalignment in
transverse butt welds subject to fatigue loading, international institute of
Illfelding, llW-Iillll-1000-BS, 1905.
[2E]SlC1NES, E.G., BAICER, R.G., HARRISCJN, J.D. and Bl..lRDElf.lN, F.M. Factors
affecting the fatigue strength of welded high strength steels. Br. Weld. J, I4
l3] 19152.
[22] WATRINSDN, F., BDDGER, P.H. and HARRISCJN, l.D. The fatigue strength of
welded joints in high strength steels and methods for its improvement, Proc.
Conf. Fatigue of Welded Structures, The Welding institute, Abington,
Cambridge, 1921.
[20] SCHNEIDER, C.R.A and MADDDI-‘ti. S. .|. Best practice guide on statistical
analysis of fatigue data, international institute of Welding, IlW— Hill-2130-0E.
2005.
[29] HAA-ISENSEN P J and MADDDR S J: iilrti Recommendations on methods for
improving the fatigue strength of welded joint, ISBN-13: 920-1-20242-015-4-4
Woodhead Publishing Ltd., Cambridge, 2013.
[30] BA}-tTE|=I. C F G and BDDTH G S, The fatigue strength improvement of fillet
welded joints by plasma dressing, in The Joining of Metals: Practice and
Performance. Proceedings, Spring Residential Conference, Coventry, Liis.
10-I2 Apr.19B1. Publ: Whetstone. London N20 BLW. UR: Institution of
Metallurgists. Publication 1401-BI-‘r'. NCi.1B. 'v'oI.2. ISBN
0901492-‘i4-4.Session 2. pp. 216-22E-; discussion pp. 253-9.
[31] lR'v"INE, N.M. Review of stress analysis techniques used in LlitIClSR‘F'
conference on Fatigue in Ciffshore Structural Steels, Institute of Civil
Engineers, London, February 19B‘i.
[32] CLA‘rTDH, A.M. and iRvlr~IE, N.M. Stress analysis method for tubular
connections. Paper 30. European Dffshore Steels Research Seminar, The
Welding institute. Cambridge.19'i'B.
[33] WDRDSWDRTH, A.C. and SMEDLET, t.S.P. Stress concentrations at unstiffened
tubular joints, Paper 31, European Clffshore Steels Research Seminar; The
Welding Institute, Cambridge, 1920.
140
l
E The British Standards Institution 2014
This page dellbera tely left blanl:
NU CCIF"'-‘ll’-lCi '."Ii| l1l-'.3Il_|T BS Pl;lif"'v1=E-S'C.'l‘-l Li'iCL'-"I AS Fi.Fll-ill? l [D l.l"l CCiF"r'FIllSl'll |_.-"-'-'lA'
Revisions
About us
I."-.'-' I'll‘-I'll‘; ‘I’:-‘;e".l'-.'| lII|:~. "||-=i’_ II'|lLl'..'i'II'y_ .;-'.I-.'|"-" |"lc"|I :-"|I'I'..JI"'-"I', ||"t"I':'i.'.'|=!I.':-"-.
.||'-:l -:-1' |': . !..
||'|I' !l||' I -'-..-‘r .': I"-|'I.I i':-.|||-I-|":-:1' .Iu-:l rs:-|"l '=.|- n'-‘-..- .".-ii Il.=i|'l'-
II.|'.|*:l -.|:lI.l -'||".
ll‘-=' if :'|-.'.'||'-:li'j|'- I-"||'-l||'||‘l|-'|1 ‘II :'I|iI '.1-=||'I1.'-|||'l'E. '-inf. I=|-i'I' :j.=||i-|..-"..' .-|".'I-='|'|.l'Ill-'!| "'
.| |li_'|||;-":l.|i|i|_- *|:r'||.-|l_ .||I|t '|_-'|||-'|] t||||_..||;-'1-:-.iI '.l|_lI_‘I -_-:'-||'..'.Jlf11c.r'- t_'-I:|:|_-.'.
l"l-‘rs ||'|,'.1II||:r", :|l .||l '. ,j|_". .I"-1 .Il rij:--.1. .|l| ;-s_'|_1-'r-. :|i|||_:'_.|_- '.l'.|'.|_11' :|- tr: IlI_'I|_I
l."-|""l -3.|'--II‘--!' Ilil-."l|' =1-.II." -
'.i'.." Hr-I-'.'| ".-l."""'t.'I'II5'=. .‘| "-:'J "-Il'|'-' :'.I..l'|l|:'.‘-'-:'II'- - Fur" ..|'.I:'t-‘|1e"IIl I|'|- .'|""-:"|'||1"|'|r'|"' "Ii ||r. '.|I"I"
-iv-= cor-'|r~ .--'ir'-.. -r||'-rri-.--1 thr- o,..-'|ln-. c-5 -".;' |'.'-'.-cl-|i':s .=in-': r.i'-'".-c-as to hr-'--'-’-1 -.-n .r
i||J=. "'.-t-'=.'I ll '-.'-..-|| ‘|I'||l .=.It: ||'.=IIf-:_'.Jl.SI.".' -.':-" .=||I I'||i.||||1y '.'.--Tl‘ ". .'s |1ll:T|'.!'- '11.-'|:'||1.=sl|l ||| |'|1I1r'I
li"i' |'.|I."-Ii: -':1|-"-"'- i"'P.=|:'-'=' ||'I‘I'II'|'| ll |' l'I"I-':'.'-.'|I"-c‘.-:'[--" F |'|'I'||'
Copynght
-"-.il the neta, :.o'|'.'-.-era sin-1 rs."-ru "‘:er=1at|r-*'- set out ='- e I E-r|".--an ':l|?"lI'.|ETl'.I'E| as-ti
-_-:l=e.* BS-I n-.ilr-'-cations are I".e .1-ro|:|e|:'_. or .=.s':l co-cvriglitec: :5: I!-SI_ -:-I so-me |sers.nr|
Information on standards
or t--'-t-1-,-tsat -.'.-i.-ns. con].-right r the -"re-'-'-tat:-'-r users -such .=i=. |l'-r- -reernat:-:-"a
y'-,'-- |.|.' ;:-::|i,-|ili_I '..';_|:, .~,'=l_l| '_l'|t' -r'r '.'.|I_'-':'-;_1|I 1I|._|!! 'i.'|_iI|| |_irr_| ||||,-'._|'_-:_=r| |"r-;--1:.
S1.|||I.1.||-:!|I".=sT|||.'| c.-.'J|J|r"t-I .=.-|‘I|l ii.=|'=. l0|'|‘|‘.=:Iiy l|I.-P|'=-;P'|'.! '.oi'l'| -'Il.'sll||.‘=l||||' 1-.:- !!"i' S-:‘.I|
I-'- *=.|-:cec-rl i nc out -c-re .=br-I2 lsr|"|'."i ‘i1-':r-rsarris ‘c-'.' '-'I=|1|r'g r.-.r '-v-‘I:-'. te at
ht--;'c-up -I-;-rr-"st.=.'I:r.-1-'t:: or c=:="|'.ert|r'o cl.’ C'..:-1c-"her Set-.-ices =ee"‘| or
.|.:-:1 F'.|1rI||1'_. ."l.rI liilrlil r'- I -l'lI'|_I- I || -_|'. I:-.' '|-;:.'-:|:l.|| r-il '-.':|||-|I ||' .| ||-!|||_--.-.Iil -:--.lrI||=
'5!"I.l'-1.'|I="I'.l'|" '.' I:'I"'Il'|.'
'-| ‘I._s'|_'.I1|.l'|'-'1 |r| ."|r:-, '-"-I"' -"-' hy .|-|'.-,- r|'||_'.'|r".
Buying standards
or o:l'er-'--Ise - o.'-1-toiit |'.'-' -:'-I -vrttl-in pt-'"' ss-or lI-:|rr- E'-S-- L!-rial s a"-cl etlvlce can
ne c-ota-ned rrc-"1 the C-:-o-.-r-gr-t -ii. Lc=.'r'-slrg L'Ieoartrnent
-‘c-_| :er E‘.-Lv =1 r-:I -:In-r."--.1-.sd i-‘DI '.-er-sinrit r-' ls‘-LI poo -cat--:-r:s |r-c-=_--2 'tg I:|r-1-sh
ar'c1 .=:-J-:-|1-:r-:l L-..'-:-pearl ariti ll"'I1i'IT'I3'illI.I"=|:- '-t.'1rirJ.:1rds, :.1r-:'.-ug|' -:.-.i' -.'-.-t-:'nn-a- -sl
n-..-:|iu.|:- r-.:-I".-'r.l'»o|s_ -.-.-'.t---='- h.'iri1 t-I:-,|.I-1"‘. ..s|'. :r'.r. he our-.:':|s.r-.1
l ',.'|I|. l'I'|'-'1 ='il|'Il'..-I‘ '-II.=|| .‘|'l-.5 l:'ili- ‘|l| '.l.||'-'!-.Ir-_'-. ‘|I'-"| II‘-ll -I‘-' '-In-’-|I.-=l|l'. l'.'|-'.'-.'- -I-1.-"1-i-.l'-:;- :'= .'.|l-..II|'-_ !|.u|l |-':: -t". I.r" I." -.-'.l|'-'-.'|| lr-:-"I -':-..-' i IJ'|IIIIl |" '-|'|-.'|| -'- I-'.:I"
Subscriptions
LIL.‘ '.-'IF=I§I:" ['lI' E.IEIEC.' E-I-.‘.I"' 'iI.="".IIIE"i E!'L'- ClE'F. "HE'S "'I TI|':I'.E' -J? "
:"EIE=-Pf iClf '|'I3|.I
I'.|3I'lI'lF!"Cl".
III-I 7ul'IIil-"I :l'I'Clffl|i!-I IIr"I i'J|'l -IIIJI f.-..IIILI." I'll Il|'| DfClCIl..-T'I1 CIIS 1|‘:
l'.|1:-._';"-.'.-|_'|T| r.':'I"I'-.'=..Jl'|".-.’l' t‘I!|i‘|I'I=.
:|I'I|f| I"-: I." |||:l|i -. .'|'!=:I'| .=|I|l |:'-.I- i:-t-.r'|:1 .|'. |I|-'I' -'l|-|'1 |||'-:11" l'.|' '. -:|||'.-'-'!:;lil'_ i"I'-'.|-:;'i'.
Useful Contacts:
Customer Services
‘tel: +44 3--1'; lJd-.-- '-uIJ'
Email [orders]: otcIr.'rsB"s ]'I.II.t: to-"'
Email {enq u|rles]: t si.-|'vi- -.'--.tilt:r-:---:_;'...|=:-.cu---Sul:|-scrlpllons
Tel: +4-‘I ii-1!: 0i:lrI- 900'
Email: snt:~'.cr pI-ii-'-1Ii'i|!nsi.;ir.':-_c- -‘on.
v'-.'-I' lE!rltls.h Standards Cl-nllne IIBSDLI -.1"-is I '-_'i-.i=- -' .t.'|n: .'1-I";-I'-".' i-.'- er.»-r ':--i ui.ii_'-
Iincrwl-edge Centre
'!I|!| J" .1I||l .-i|]|||.|1|-|| l_ ..':||-|'.||' .1|.:l i||'|-'I|.ll -:.-'|.: '.'.t"-:|.s'rl'. l||'||:| '|'I|lIl II-'-_~.!‘||1.:
1". .|'.'.|Ii.I:"-l|' ,._'-s1."" .I."-:l "|'|I|'-.l'|-:I '_1.| '|.' '.|| -,-|_||| ll .|l-.-'..|I,-'. I-|' |.|;: 1r:- '.l.1'i'
Tel: +-'--"l .-‘II EiE'|'_l'? ll:|U-1
EI'r!|.l-II: lr.|I|':'.'-.'l|'-:1|:T||'-:'|':I|t|rj'f;I t!-';--';||:'|.|t‘- tj-'-it:
"'-“'.-.= -.1.-III 'c-"-'|| ||I !'|'||:|"| Ir.-I" S’-L-'|'Il.I"'t". I:|'I.-|r-:||II| i'r'l- .'-I'=|| |I:-te|I.-|- -.=|l::-=.t.'-"Il'|.‘I
|_l|'.c-_s|_.'|['.. -,-" |l~:_' |_|I_.-'-" -|'.|_' pr~;t_- i_|'. ,.!._:r|r:..1r|_l'_._ I:-:_:-H‘. " '.-I|||l-_' :_|_||_:y ._I'||_l '.'.1s5rj'.|:_-I cu.
Copyright tli Licensing
'-=.t|r|'- || I‘-v lit-.1-"I-"to I BSI Subscribing lulernlaer
TIII -r-ill 2'-L‘ EIEIBE -'!:|3'C-'
PLUS 1. .1I| ||[|II.|1|I|I; ".1:-r'.'-: |I -"it: |||'.'.~|' tr: l-_l";.l '| |:::-I ' ;:- "'1 l'..1|".'|l:'t-|'. 'I'III "cl.-|"
|_' .'_I-' |':||1|-;_ [I"-;'-':j-|_-"-|I-,I"|c:_ r=_-|_r:-'::l||1r_|
El'l'!|bII: ---||yru|l.lliII'!.- .- :|I|||.:'I - -'-l'I
._||_.it;-|'|._|Ei.'.ji|!i,' ||_'.'_v!|I,-s;- !"|I l.;|.‘i_"_.' h.1|:_I .j.;|1I,- r|' yruir '.I._I"-','_|r-3'. |.',.t|i_'I|'i I|u_"-'.."'.'_I
'-To so: -.'-' re|'=.ace-:l
T-:- ll"-"l out "Ir."-ts |tD0u1 lsc-:.r.tI" '11] cl 03' S..o:cr=:.~r-'_1 lv1r.'"'t's'r st"-':l tric t'erIc-‘Its
-J" "'a|T-Lu.-r'.=1||_'- ole aso '.ss|t us--"gr-'_-up -;o|""shop
I.-‘v I" -.- Irlulti-Llser 1-Ietwort Licence IMUHLJ -.--- -. olo -sbl-. l.-'.- l‘--_l'.-! :-'.=|'-|.tc-'-'.t'.
punlir at or-'. -;'-r- -in-.|r -ltttnr-|'| !|rr.-|'=:|_=r. r-sr-1=.'--Ir" as lc--.-'-' or as I1'.=1."-..' ..'=e='. as '.--:--I
.'-.---.' '.-"i-tl'- on-.l.a'.-=-=.
:s|slii-s a=. -.-u-:.-'- as1l'-'--,- 'r- r1-.-r1 sole, '|l"..|; can he sI:'-'.- vou-
|lI'||‘I||"‘|i-I-T-‘|I -3-": ‘S |'|i'ii'|'I1 i-:'.-I‘ li.FIl'|i-" ||"l-'IlI1!.-'|1H.‘:~" I‘-'-1"-=||l l'I=-‘III.:".i'!ie=.-ii|.':'E-'I‘||'I‘II.:;'I I'|'|II
BS1 Group Headquarters
'-.F."I " -ir,-_.',|i']r '-|-|]". F-'-'_-.=|r| | -'~"r. - '. ',l.'s'. -'-"'.|_ - In
bsi.
.--rnak|rig excellence a habit.