当前位置:首页 >> 电力/水利 >>

55、BS 7608-1993 Fatigue design and assessment of steel structures


BRITISH STANDARD

BS 7608:1993
Incorporating Amendment No. 1

Code of practice for

ww

w.

bz

fx

w.

co

m

Fatigue design and assessment of steel structures

BS 7608:1993

Committees responsible for this British Standard
The preparation of this British Standard was entrusted by the Welding Standards Policy Committee (WEE/-) to Technical Committee WEE/44, upon which the following bodies were represented: AEA Technology Association of Consulting Engineers British Constructional Steelwork Association Ltd. British Railways Board Department of Transport Department of Transport (Transport Research Laboratory) ESDU International Ltd. Electricity Association Federation of Manufacturers of Construction Equipment and Cranes Institution of Civil Engineers Institution of Structural Engineers Ministry of Defence Process Plant Association Railway Industry Association of Great Britain Society of British Aerospace Companies Limited United Kingdom Offshore Operators Association Welding Institute

This British Standard, having been prepared under the direction of the Welding Standards Policy Committee, was published under the authority of the Standards Board and comes into effect on 15 April 1993 ? BSI 03-1999

ww
Amendments issued since publication Amd. No. 8337 Date Comments February 1995 Indicated by a sideline in the margin

The following BSI references relate to the work on this standard: Committee reference WEE/44 Draft for comment 87/77328 DC ISBN 0 580 21281 5

w.

bz

fx

w.

co

m

BS 7608:1993

Contents
Committees responsible Foreword Section 1. General 1.1 Scope 1.2 References 1.3 Definitions 1.4 Symbols and units 1.5 Assessment 1.6 Design life 1.7 Fatigue loading 1.8 Basis of fatigue analysis 1.9 Factors on fatigue life 1.10 Features influencing fatigue behaviour 1.11 Fracture mechanics Section 2. Classification of details 2.1 General 2.2 Classification of details 2.3 Unclassified details 2.4 Workmanship and inspection 2.5 Welded steel decks Page Inside front cover iv 1 1 1 3 3 3 3 4 4 5 5 6 6 24 24 25 27 27 27 27 27 28 29 30 30 31 31 31 33 34 34 34 46 47 51 61 69 70 71

ww

w.

Section 3. Stress calculations 3.1 General 3.2 Stress range in parent material 3.3 Stress range for welds 3.4 Effective stress range for details in unwelded members in which the whole or part of the stress is compressive 3.5 Calculation of stresses 3.6 Geometrical stress concentrations 3.7 Stresses in welds attaching shear connectors 3.8 Axial stresses in bolts 3.9 Derivation of stress spectra Section 4. Allowable fatigue stresses 4.1 Tensile stress limitations 4.2 S-N curves 4.3 Modifications to basic S-N curves 4.4 Treatment of low stress cycles 4.5 Treatment of high stress cycles 4.6 Joints subjected to single stress range 4.7 Joints subjected to a stress spectrum Annex A (normative) Fatigue design philosophy Annex B (normative) Explanatory notes on detail classification Annex C (normative) Guidance on the calculation of stress concentration factors Annex D (normative) Guidance on the use of fracture mechanics Annex E (normative) Fatigue testing and the use of test data to define design stresses Annex F (normative) Cycle counting by the reservoir method Annex G (informative) Background notes on sources of data

bz

fx

w.

co

m

? BSI 03-1999

i

BS 7608:1993

Annex H (informative) Bibliography Figure 1 — Weld quality category determined by adjacent detail Figure 2 — Reference stress in parent metal Figure 3 — Reference stress in weld throat Figure 4 — Typical example of stress concentrations due to geometrical discontinuity Figure 5 — Typical example of stress concentration caused by a geometrical hard spot Figure 6 — Stress concentration factors Figure 7 — Example of hot spot stresses in a nodal joint Figure 8 — Summary of mean-line Sr-N curves Figure 9 — Summary of standard basic design Sr-N curves Figure 10 — Sr/UTS-N curves for bolts with cut or, ground or rolled threads under axial loading (class X) Figure 11 — Toe grinding to improve fatigue strength Figure 12 — Grinding of weld at tubular nodal joint Figure 13 — Typical Sr-N relationship Figure B.1 — Edge distance Figure B.2 — Failure modes at weld ends Figure B.3 — Failure modes in cruciform and T-joints Figure B.4 — Failure modes in transverse butt welds Figure B.5 — T-junction of two flange plates Figure B.6 — Cruciform junction between flange plates Figure B.7 — Alternative method of joining two flange plates Figure B.8 — Local grinding adjacent to cope hole in type 7.1 joint Figure B.9 — Use of continuity plating to reduce stress concentrations in type 8.1 and 8.2 joints Figure B.10 — Example of type 8.3 or 8.4 joint Figure B.11 — Single fillet corner weld in bending Figure B.12 — Example of a third member slotted through a main member Figure C.1 — Types of misalignment and distortion Figure D.1 — Flaw dimensions Figure D.2 — Transverse load-carrying cruciform joint Figure F.1 — Example of cycle counting by reservoir method Figure G.1 — Comparison of proposed S-N curves for bolts in axial loading with curves from ESDU data sheets

Page 80 35 36 36 37 38 39 40 41 42 43 44 45 45 52 52 53 53 53 54 54 55 55 56 56 57 61 63 64 70 79 7 8 9 10 12 13 15

ww

Table 1 — Classification of details: plain material free from welding Table 2 — Classification of details: bolted or rivetted, spliced or lapped connections Table 3 — Classification of details: fasteners and shear connectors Table 4 — Classification of details: continuous welded attachments essentially parallel to the direction of applied stress Table 5 — Classification of details: welded attachments on the surface or edge of a stressed member Table 6 — Classification of details: full penetration butt welds between co-planar plates Table 7 — Classification of details: transverse butt welds in sections and tubes

w.

bz

fx

w.

co

m

ii

? BSI 03-1999

BS 7608:1993

Page Table 8 — Classification of details: load carrying fillet and T-butt joints between plates in different planes Table 9 — Classification of details: slotted connections and penetration through stressed members Table 10 — Classification of details: circular tubular members Table 11 — Classification of details: seam welds Table 12 — Classification of details: branch connections Table 13 — Upper limits of flaw area of planar flaws and slag inclusions in transverse butt welds Table 14 — Details of basic S-N curves Table 15 — Nominal probability factors Table D.1 — Values of M1, M2 and M3 Table D.2 — Values of g and f? Table E.1 — Fatigue test factor F Table G.1 — Sources of the content Table G.2 — Experimentally determined fatigue strengths of various types of threaded connection expressed in terms of stress range/UTS 17 20 21 22 23 26 32 33 66 67 70 72 77 Inside back cover

List of references

ww

w.

bz

fx

w.

co

m

? BSI 03-1999

iii

BS 7608:1993

Foreword
This British Standard has been prepared under the direction of the Welding Standards Policy Committee. It has been produced with the objective of covering the general aspects of fatigue design of fabricated steel elements within the subject area regardless of the type of structure in which they are situated. It is anticipated that application standards for particular types of structure will in due course be modified to omit their own fatigue design rules and to cross refer to this code of practice, thereby achieving a co-ordinated approach to the subject. Such application standards will, however, still include fatigue loading and safety requirements for the structures to which they relate. This code is based directly upon the recommendations of existing codes and other reference documents, such as the Department of Energy Publication Offshore installations: Guidance on design, construction and certification. At this stage no attempt has been made to update existing recommendations except in cases where they were found to be in need of clarification. Background notes on the sources of the data given in this code are provided in Annex G. As a code of practice, this British Standard takes the form of guidance and recommendations. It should not be quoted as if it were a specification and particular care should be taken to ensure that claims of compliance are not misleading. It has been assumed in the drafting of this British Standard that the execution of its provisions is entrusted to appropriately qualified and experienced people.
NOTE The numbers in square brackets used throughout the text of this British Standard relate to the bibliographic references given in Annex H.

A British Standard does not purport to include all the necessary provisions of a contract. Users of British Standards are responsible for their correct application.

ww
Summary of pages This document comprises a front cover, an inside front cover, pages i to iv, pages 1 to 80, an inside back cover and a back cover. This standard has been updated (see copyright date) and may have had amendments incorporated. This will be indicated in the amendment table on the inside front cover. iv
? BSI 03-1999

w.

bz

Compliance with a British Standard does not of itself confer immunity from legal obligations.

fx

w.

co

m

BS 7608:1993

Section 1. General
1.1 Scope
This British Standard gives recommendations for methods for the fatigue design and assessment of parts of steel structures which are subject to repeated fluctuations of stress. It is concerned with wrought structural steel with a specified minimum yield strength of up to 700 N/mm2 operating in the sub-creep regime. This British Standard is applicable to the following: a) parent material remote from joints; b) welded joints (in air or sea water) in such material1); c) bolted or rivetted joints in such material; d) shear connectors between concrete slabs and steel girders acting compositely in flexure. Guidance on general fatigue design philosophy is given in Annex A, which also contains a brief description of the method of using this document. The relevant application standard or other specification for the particular structure under consideration should specify the following: 1) the loading to be assumed for design purposes, including its magnitude and frequency; 2) the required life of the structure; 3) the environmental conditions; 4) the required nominal probability of failure (see 4.2). This British Standard takes no account of the possible onset of unstable fracture from a fatigue crack. This possibility should be considered and guarded against by appropriate material selection. This British Standard does not apply to the following: — orthotropic decks; — wire ropes; — bonded connections; — steel reinforcement in concrete; — out of plane joints between hot rolled rectangular or square hollow sections; — pressure vessels; — castings; — peening.

1.2 References
1.2.1 Normative references This British Standard incorporates, by reference, provisions from specific editions of other publications. These normative references are cited at the appropriate points in the text and the publications are listed on the inside back cover. Subsequent amendments to, or revisions of, any of these publications apply to this British Standard only when incorporated in it by updating or revision. 1.2.2 Informative references This British Standard refers to other publications that provide information or guidance. Editions of these publications current at the time of issue of this standard are listed on the inside back cover, but reference should be made to the latest editions.

1.3 Definitions

bz

w.

ww

1)

For welded joints fatigue strength is same for all steels.

fx

w.
1.3.1 fatigue 1.3.2 stress cycle

For the purposes of this British Standard the following definitions apply.

the damage of a structural part by the initiation and gradual propagation of a crack or cracks caused by repeated applications of stress

a pattern of variation of stress at a point defined by the cycle counting method and consisting of a change in stress between defined minimum (trough) and maximum (peak) values and back again
NOTE 1 Also known as cycle of stress. NOTE 2 One loading event may produce one or more stress cycles at any particular point.

1.3.3 stress range Sr the algebraic difference between the two extremes (reversals) of a stress cycle
NOTE 1 Also known as range of stress. NOTE 2 See 3.2 and 3.3 for parent metal and weld metal respectively.

1.3.4 initial non-propagating stress range So the constant amplitude stress range below which (in the absence of any previous loading) a crack is assumed not to propagate
NOTE Its magnitude depends on the joint class and, for joints in air or adequately protected against corrosion, it is assumed to be the stress corresponding to a life of 107 cycles on the design S-N curve. For unprotected joints in a corrosive environment it should be assumed that So = 0.

co

m

? BSI 03-1999

1

BS 7608:1993

1.3.5 detail class the rating given to a particular structural detail to indicate which of the fatigue strength (S-N) curves should be used in the fatigue assessment
NOTE 1 Also known as joint class. NOTE 2 The class is denoted by one of the following letters: A, B, C, D, E, F, F2, G, S, T, W or X. The categorization takes into consideration the local stress concentration at the detail, the size and shape of the maximum acceptable discontinuity, the stress direction, metallurgical effects, residual stresses, fatigue crack shape, and in some cases the welding process and a post-weld improvement method.

NOTE It is derived from the relevant basic S-N curve modified, if necessary, to allow for the influence of material thickness, environment, fatigue strength improvement techniques (additional to those given in Table 1 to Table 12), stress relief (see 4.3.5) or workmanship (see 2.4).

1.3.13 fatigue loading the loading on a structure which is liable to cause fatigue cracking
NOTE It may be composed of several different types and magnitudes of loading events (see 1.7).

1.3.6 service life the intended operating life, which is the period in which a structure or component is required to perform safely with an acceptable probability that it will not require repair or withdrawal from service as a result of fatigue 1.3.7 design life the period within which there is a defined nominal probability that failure by fatigue cracking will not occur
NOTE This may be longer or shorter than the service life (see 1.6).

the constant amplitude stress range Sr causing failure in the appropriate number of cycles (N) 1.3.9 S-N curve

1.3.10 basic S-N curve

the S-N curve, for the required probability of failure, for a detail of basic thickness (see 4.3.2) tested in air without the application of any fatigue strength improvement method apart from those given in Table 1 to Table 12 1.3.11 standard basic S-N curve the basic S-N curve for two standard deviations below the mean (d = 2) assuming that the test data are represented by a normal distribution of log life 1.3.12 design S-N curve the S-N curve adopted for design purposes for the detail under consideration

ww

the quantitative relationship between the fatigue strength S and the number of cycles N corresponding to a specific probability of failure for a detail, derived from test data

w.

bz

1.3.8 fatigue strength

fx

1.3.15 load spectrum

a tabulation showing the relative number of occurrences of all the loading events of different types and intensities expected to be experienced by the structure in its design life 1.3.16 cycle counting method a particular method of counting the numbers of stress cycles of different magnitudes which occur in a service stress history
NOTE The loads applied to the structure, considered in sequence, generate a particular stress history at each detail of interest. This stress history can be broken down into equivalent numbers of stress ranges of different magnitudes by the operation of cycle counting. The particular method of counting recommended in this code is the rainflow or reservoir method (see 3.9).

1.3.17 design spectrum a tabulation of the number of occurrences of all the stress ranges Sr of different magnitudes produced by the load spectrum in the design life of the structure or component, to be used in the fatigue assessment
NOTE 1 Also known as stress spectrum. NOTE 2 Different components of a structure may have different design spectra.

1.3.18 Miner’s summation a cumulative damage summation based on the rule devised by Palmgren and Miner

w.

NOTE In the case of a crane this could involve the lifting, movement and depositing of a load. With respect to a bridge it could be the approach, passage and departure of a train or, for short lengths, a bogie or axle, over a railway bridge or one vehicle over a highway bridge. For design purposes each loading event is assumed to repeat a given number of times in the design life of the structure.

co

a defined loading sequence on the structure which may be characterized by its relative frequency of occurrence as well as its magnitude and geometrical arrangement

m

1.3.14 loading event

2

? BSI 03-1999

BS 7608:1993

1.4 Symbols and units
For the purposes of this British Standard, the following symbols and units are used. A Co C2 Net area of cross section (in mm ) Parameter defining the mean line Sr-N relationship Parameter defining the Sr-N relationship for two standard deviations below the mean line Parameter defining the Sr-N relationship for d standard deviations below the mean line Stress concentration factor under fatigue loading Applied bending moments (in N·mm)
2

∑ ---N
? fL ? ?Y

n

Miner’s summation Load factor for fatigue Standard deviation of log N Nominal tensile yield stress (in N/mm2)

1.5 Assessment
Any structure which is exposed to loads which may cause cyclic stress variations in its structural members, and is therefore liable to suffer fatigue cracking in those members, should be assessed in accordance with this British Standard. A summary of the fatigue assessment procedure is given in A.5.

Cd

Kf M, M1

1.6 Design life

S

SB So

Fatigue strength of the joint using the basic Sr-N curve (in N/mm2)

Sr

Sr1, Sr2 . . . Individual stress ranges (Sr) in a design spectrum (in N/mm2) Z d m n1, n2 . . . Elastic modulus of section (in mm3) Number of standard deviations below the mean Sr-N curve
m

ww

Stress range in any one cycle (in N/mm2)

w.

Constant amplitude initial non-propagating stress range (in air So = Sr at N = 107 cycles) (in N/mm2)

bz

Fatigue strength of the joint under consideration, including any required thickness correction (in N/mm2)

Inverse slope of log Sr log N curve (i.e. Sr ·N = constant) Number of cyles of damaging stress ranges Sr1, Sr2 . . ., etc in a design spectrum in cycles) Plate thickness of the member under consideration (in mm) Thickness relevant to the basic Sr-N curve for the joint (in mm) Weld throat thickness (in mm)

t tB w

fx

P, P1

Applied axial forces (in N)

w.

N1, N2 . . . Number of cycles to failure under constant amplitude loading with stress ranges Sr1, Sr2 . . ., etc., corresponding to n1, n2 etc. (in cycles)

The design life is the period for which it is assumed, for purposes of calculation, that a structure or component will be required to perform safely, and within which there is a defined probability that failure will not occur by fatigue cracking. The structure should, whenever possible, be designed to be damage tolerant (see A.3). In some circumstances it may be decided to make the design life either longer or shorter than the service life. If the design life is made longer (see A.4), then the result is the introduction of an arbitrary factor of safety on life, but if it is shorter there is a greater probability that failure may occur during the service life. The latter approach should only be considered if the structure, or the relevant part of it, can safely and economically be inspected and any cracks be detected and repaired before any crack reaches a length which causes failure under static loading; this requires a conscious decision with the strength of the structure, the need for inspection and the feasibility of repair in mind. It is also necessary that the strength of the structure in the presence of any cracks should have been evaluated at the time of the design.

1.7 Fatigue loading
In assessing fatigue performance a realistic estimate of the fatigue loading is crucial to the calculation of life, and all types of cyclic loading should be considered. Cyclic loading from different sources may be significant at different phases of the life of a structure, e.g. construction, transport, installation, in-service, and may involve different frequencies.

co

m

? BSI 03-1999

3

BS 7608:1993

The design load spectrum should be selected on the basis that it is an upper bound estimate of the accumulated service conditions over the full design life of the structure. Account should be taken of all likely operational and environmental effects arising from the foreseeable usage of the structure during that period. The loading should not, however, be factored (i.e. ?fL = 1.0). The following are some important sources of cyclic loading that should be considered, any or all of which may be relevant in particular applications: a) fluctuating loads; b) acceleration forces in moving structures; c) pressure changes; d) temperature fluctuations; e) mechanical vibrations; f) environmental loading (wind, currents and waves, especially when vortex shedding is induced, e.g. on slender members). Particular care should be taken to assess dynamic magnification effects where loading frequencies are close to one of the natural frequencies of the structure. In some instances the loading to be assumed for fatigue design purposes will be specified in the relevant application standard. Where such information is not available reasonable assumptions as to the loading to be expected in service should be made, and it may be useful to obtain data from existing structures subjected to similar effects. In particular, in assessing an existing structure, a design spectrum may be able to be compiled from strain readings or loading records obtained from continuous monitoring. In any case the objective is to define the spectrum in terms of the numbers of cycles of each of the individual stress ranges expected in the life of the structure. Where a non-standard loading is used the design spectrum should be divided into a sufficient number of intervals of stress and the use of stress ranges in each interval should be such as to represent adequately the damage effect over the whole spectrum.

1.8 Basis of fatigue analysis
It should be recognized that uncertainties exist in assessing both the stresses resulting from applied loads and the response of a particular joint, which together control fatigue performance. The basis of the fatigue analysis should be to use a best estimate of these stresses, noting the uncertainties involved, with S-N curves derived from experimental data. In this way, uncertainties associated with the life of a particular joint, e.g. size, weld detail, local environment, can be separated from those associated with applied stress. In order to evaluate fatigue lives it is necessary to establish the long term distribution of stress range taking into consideration all stress variations which can reasonably be expected during the life of the structure which have magnitudes and numbers large enough to cause fatigue effects. Using the calculated long term distribution of stress range, the method of assessment described in this code is based on the Palmgren-Miner rule for damage calculation (see 4.7). The resulting calculated fatigue lives, which enable critical parts to be ranked in terms of fatigue sensitivity, should be used, in conjunction with an assessment of the consequences of failure of specific members, to establish a priority basis for developing a selective inspection programme to be followed during the service life of the structure.

ww

w.

bz

fx

1.9 Factors on fatigue life
The standard basic S-N curves (see 4.2) are based on the mean minus two standard deviation curves for relevant experimental data. Their use therefore implies a finite probability of failure for the calculated life. Thus, an additional factor on life, i.e. the use of S-N curves based on the mean minus more than two standard deviations, should be considered for cases of inadequate structural redundancy. In defining this factor on life, account should be taken of the accessibility of the joint and the proposed degree of inspection as well as the consequence of failure. As a crack grows in one part of a structure, the load may be shed to other members and lead to further fatigue cracks in those members. Because of the sensitivity of calculated life to the accuracy of estimates of stress, particular care should be taken to ensure that stresses are not underestimated.

w.

co

m

4

? BSI 03-1999

BS 7608:1993

1.10 Features influencing fatigue behaviour
For both welded and bolted steel structures it has been established that the fatigue life is normally governed by the fatigue behaviour of the joints, including both main and secondary joints. The best fatigue behaviour will be obtained by ensuring that the structure is so detailed and constructed that stress concentrations are kept to a minimum and that where possible the elements are able to deform in their intended ways without introducing secondary deformations and stresses due to local restraints. Stresses may also be reduced by increasing the thickness of parent metal which should improve fatigue life although with some types of joint fatigue strength tends to decrease with increasing thickness (see 4.3.2). The best joint performance is achieved by avoiding joint eccentricity and misalignment and welds near free edges and by other controls over the quality of the joints. Performance is adversely affected by concentrations of stress at holes, openings and re-entrant corners. Guidance in these aspects is given in Table 1 to Table 12 and Annex B and Annex C.

The magnitude and nature of stresses that will cause propagation of a crack, and consequently reduce the number of stress repetitions to cause failure, are affected by the presence of residual stress, inherent flaws in welds and adjacent parent metal, surface flaws and any other stress raisers interfering with the flow of stress. These are taken into account in the classifications given in Table 1 to Table 12.

1.11 Fracture mechanics
In some situations the normal fatigue assessment procedures may be inappropriate, but fracture mechanics methods may be helpful. Guidance on the use of fracture mechanics is given in Annex D.

ww

w.

bz

fx

w.
5

? BSI 03-1999

co

m

BS 7608:1993

Section 2. Classification of details
2.1 General
For the purposes of fatigue design, joints are divided into several classes, each with a corresponding design S-N curve (see 4.2). This classification depends upon the following: a) the geometrical arrangement and proportions of the detail; b) the direction of the fluctuating stress relative to the detail; c) the location of possible crack initiation at the detail; d) the methods of manufacture and inspection. In any structure or component, which is liable to be subjected to repeated applications of stress, every welded joint or other form of stress concentration, such as a bolt hole, is potentially a source of fatigue cracking. Hence, each part of each constructional detail should be considered individually and should, where possible, be placed in its relevant joint class in accordance with the criteria given in Table 1 to Table 12. Otherwise the detail should be dealt with in accordance with 2.3. In bolted or rivetted joints fatigue cracks will normally initiate from the bolt or rivet hole. However, in welded details there are several locations at which fatigue cracks may, potentially, initiate; these are as follows: 1) in the parent metal of either part joined adjacent to: i) the end of the weld; ii) a weld toe; iii) a change of direction of the weld; 2) in the weld metal starting from: i) the weld root; ii) the weld surface; iii) an internal flaw. In the case of members or elements connected at their ends by fillet welds or partial penetration butt welds and flanges with shear connectors, the crack initiation may occur either in the parent metal or in the weld throat; both possibilities should be checked by taking into account the appropriate classification and stress range. For other details, the classifications given in Table 1 to Table 12 cover crack initiation at any possible location in the detail. Notes on the potential modes of failure for each detail are given in Annex B.

2.2 Classification of details
Table 1 to Table 12 correspond to the following basic types of details: 1) plain material; 2) lapped or spliced, rivetted or bolted joints; 3) fasteners; 4) continuous longitudinal welded attachments; 5) other welded attachments; 6) transverse butt welds in plates; 7) transverse butt welds in sections and tubes; 8) load-carrying fillet and T-butt joints; 9) slotted connections and penetrations through stressed members; 10) details relating to tubular members; 11) seam welds in vessels; 12) branch connections to vessels. Each classified detail is illustrated and given a type number. Table 1 to Table 12 also give various associated criteria and the diagrams illustrate the geometrical features and potential crack locations which determine the class of each detail. This information should be used to assist with initial selection of the appropriate type number. A detail should only be designated a particular classification if it conforms in every respect to the tabulated criteria appropriate to its type number. Class A is generally inappropriate for structural work and the special inspection standards relevant to classes B and C cannot normally be achieved in the vicinity of welds in structural work.

ww

w.

bz

fx

w.

co

m

6

? BSI 03-1999

Table 1 — Classification of details: plain material free from welding
Design Type Class Dimensional Manufacturing Special Product Location of stress number inspection form potential crack requirements requirements requirements area initiation Rolled steel structural plates and sections Away from all welds or structural connections Member of constant or smoothly varying cross section with no holes or re-entrant corners All surfaces fully machined and polished. No flame cutting Edges as rolled or machined smooth. No flame cutting 1.2 B 1.1 A See B.2.2 Notes Sketch

See also note for type 1.4. Net cross section 1.4 C

BSI

Any cutting of edges by planing or machine flame cutting with controlled procedure

The controlled flame cutting procedure should ensure that the resulting surface hardness is not sufficient to cause cracking. Types 1.3 and 1.4. The presence of an aperture, re-entrant corner or other discontinuity implies the existence of a stress concentration and the design stress should be the stress on the net section multiplied by the relevant stress concentration factor (see Annex C)

At a small hole (may contain bolt for minor fixtures)

Hole drilled or reamed

Net cross section

w. b
1.5 D

This type may be deemed to include bolt holes for attaching light bracing members where there is negligible transference of stress from the main member in the direction Sr The classification includes allowance for the stress concentration created by the hole.

ww

zf xw .
. .

At any external or Member with internal edge or without apertures, re-entrant corners or other discontinuities

Any flame cut edges subsequently machined or ground smooth

Net cross section

1.3

B

All visible signs of drag lines should be removed from the flame cut edge by grinding or machining.

co m

? BSI 03-1999

BS 7608:1993

7

Table 2 — Classification of details: bolted or rivetted, spliced or lapped connections
Product form Location of potential crack initiation Dimensional requirements Manufacturing requirements Special Design Type Class inspection stress number requirements area Gross 2.1 C Notes Sketch

See B.2.3 regarding the tightening of bolt groups. 2.3 D

At joint made with rivets or precision bolts, at the hole At joint made with high strength friction grip bolts, at the hole Joint made with rivets, at the hole Joint made with Close tolerance Torque to at least 0.7 × proof load precision bolts, hole or use lock nuts at the hole Joint made with black bolts, at the hole Net section

The classifications include allowance for the stress concentrations created by the bolt hole.

2.4

D

2.5

D

2.6

E

2.7

G

ww

w. b

zf xw .

BSI

co m

8
Rolled steel structural plates and sections At joint made with high strength friction grip bolts, at the hole

BS 7608:1993

Double covered Holes drilled or At joint made symmetrical reamed with high strength friction joint grip bolts, away from the hole

This covers connections designed for slip resistance at the ultimate limit state and where secondary out-of-plane bending of the joint is restrained or does not occur i.e. double-covered symmetrical joints. Failure initiates by fretting in front of the hole. Failure initiates at edge of hole.

Bolts tightened in accordance with BS 4604-1:1970 or BS 4604-2:1970

2.2

C

? BSI 03-1999

Table 3 — Classification of details: fasteners and shear connectors
Product form Location of potential crack initiation At thread root. Fastener in a butt joint with fastener axis parallel to Sr Dimensional Manufacturing requirements requirements Type Design Special stress number inspection area requirements Tensile stress area 3.1 Class Notes Sketch

Screw threads conforming to BS 3643-2:1981

For the use of black bolts conforming to BS 4190:1967 and subjected to fluctuating tensile loads, see 3.8.

In weld throat Shear connectors between encased in concrete connector and any member

ww

w. b

All regions stressed in through thickness direction to be free from lamellar flaws and tears

Effective 3.2 weld throat area (see 3.7)

S

zf xw .

Where bolts or screwed rods are pre-tensioned to a value in excess of an applied external load, including any prying forces, stress fluctuations will be governed by the elasticity of the pre-compressed elements. The increase in tension will rarely exceed 10 % of an external load applied concentrically with the bolt axis, but where the load is eccentric, e.g. as in the sketch, a further increase will result from prying action.

BSI

This type covers embedded shear connectors at any position along a girder. The fatigue strength of the member adjacent to the shear connector is covered by type 5.1.

co m

? BSI 03-1999

Threaded fasteners

Bolts conforming to BS 3692:1967, BS 4395-1:1969 BS 4395-2:1969

X (see 3.8 and Figure 10)

This classification applies to failure at the root of the thread in normal commercial quality threaded components. Attention should be paid to the details of head fillets, waisted shanks and thread run-out in components, not covered by an appropriate British Standard, to ensure that they have satisfactory fatigue resistance. A higher fatigue resistance (see 3.8) can be obtained with a rolled thread on material which has previously been fully heat-treated, but such components should be subject to special test and inspection procedures.

BS 7608:1993

9

Table 4 — Classification of details: continuous welded attachments essentially parallel to the direction of applied stress
Product form Location of potential crack initiation At a long welded attachment (in the direction of Sr) away from the weld end Dimensional Manufacturing requirements requirements Special inspection requirements Proved free of all flaws which are likely to degrade the joint below its stated classification (see 2.4.3) Design stress area Type Class number Notes Sketch

Butt or fillet welded web or attachment with width not greater than 50 mm

Automatic weld with no stop/starts

4.2

C

Accidental stop/starts are not uncommon in automatic processes. Repair to the standard of a C classification should be the subject of specialist advice and inspection and as a result, the use of this type is not recommended. For situation at the ends of flange cover plates see joint type 5.4. Backing strips, if used, need to be continuous and either not attached or attached by continuous fillet welds.

Welds with stop/starts

4.3

D

ww

w. b

At an intermediate gap in a longitudinal weld

Intermittent fillet weld with g --# 2.5 h

4.4

E

The limiting gap ratio g/h applies even though adjacent welds may be on opposite sides of a narrow attachment (as in the case of a longitudinal stiffener with staggered fillet welds). Long gaps between intermittent fillet welds are not recommended as they increase the risk of corrosion and, in the case of compression members, may cause local buckling. If intermediate gaps longer than 2.5h are required the class should be reduced to F.

zf xw .

If the backing strip is attached by discontinuous fillet welds see type 4.6.

BSI

co m

10
? BSI 03-1999

BS 7608:1993

Rolled steel structural plates, sections and built-up members

,

Butt weld with Weld reinforcement full dressed flush penetration and no backing strip

Minimum 4.1 transverse cross section of member at location of potential crack initiation

B

Finish machining should be in the direction of Sr The significance of flaws should be determined with the aid of specialist advice and/or by the use of a fracture mechanics analysis. The non-destructive testing (NDT) technique should be selected with a view to ensuring the detection of such significant flaws. This type is only recommended for use in exceptional circumstances.

.

Table 4 — Classification of details: continuous welded attachments essentially parallel to the direction of applied stress
Product form Location of potential crack initiation At a cope hole Dimensional requirements Manufacturing requirements Special inspection requirements Design stress area Type number Class Notes Sketch

At the ends of a discontinuous fillet weld attaching a backing strip

The backing strip to be continuous

4.6

ww

w. b

zf xw .
E

BSI

This type includes tack welds to the edges of continuous longitudinal backing strips irrespective of spacing, provided that the welds conform in all respects to the workmanship requirements for permanent welds and that any undercut on the backing strip is ground smooth. The effects of tack welds which are subsequently fully ground out or incorporated into the butt weld by fusion, need not be considered. If the backing strip is either not attached or is attached by a continuous fillet weld, see type 4.3.

co m

? BSI 03-1999

Rolled steel structural plates, sections and built-up members

Weld may or may not continue round plate ends

Minimum 4.5 transverse cross section of member at location of potential crack initiation

F

The existence of the cope hole is allowed for in the joint classification: It should not be regarded as an additional stress concentration in relation to cracking in the flange. With regarded to the web see notes for type 7.1.

BS 7608:1993

11

Table 5 — Classification of details: welded attachments on the surface or edge of a stressed member
Product form Location of Dimensional Manufacturing Special potential crack requirements requirements inspection initiation requirements Design stress area Type Class number F Notes Sketch

At weld toe or end l # 150 mm of a long, narrow w # 50 mm attachment Weld not within 10 mm of member edge

5.3

F2

The decrease in fatigue strength with increasing attachment length is because more load is transferred into the longer gusset, giving an increase in stress concentration.

BSI

At weld toe or end l > 150 mm w > 50 mm of a long, wide attachment

5.4

G

At weld toe or end Weld within of any attachment 10 mm of close to edge of member edge stressed member

Grind out undercut

ww

w. b

5.5

G

zf xw .

The classification may be deemed to include stress concentrations arising from normal eccentricities in the thickness direction. This type includes parent metal adjacent to the ends of flange cover plates regardless of the shape of the ends.

This type applies regardless of the shape of the end of the attachment. In all cases, care should be taken to avoid undercut on element corners or to grind it out to a smooth profile should it occur. In particular, weld returns across a corner should be avoided and the use of cover plates wider than the flange, to which they are attached, is not recommended. The classification applies to all sizes of attachment. It would therefore include, for example, the junction of two flanges at right angles. In such situations a low fatigue classification can often be avoided by the use of a transition plate (see also joint type 6.5).

co m

12
? BSI 03-1999

BS 7608:1993

At weld toe of stud Rolled shear connector steel structural plates, sections, tubes, and built-up members

Minimum 5.1 transverse cross section of member at location of potential crack initiation 5.2

At weld toe or end Weld length (parallel to Sr) of a short l # 150 mm attachment (in direction of Sr)

F

Butt welded joints should be made with an additional reinforcing fillet so as to provide a similar toe profile to that which would exist in a fillet welded joint.

Table 6 — Classification of details: full penetration butt welds between co-planar platesa
Dimensional Manufacturing Special Product Location of inspection form potential crack requirements requirements requirements initiation Rolled steel plates only At transverse butt weld joining two single plates end to end Plates of equal Misalignment slope # 1 in 4 width and thickness Longitudinal axes in line Dress flush reinforcement Proved free of all flaws which are likely to degrade the joint below its stated classification (see 4.3) Design stress area Type Class number C Notes Sketch

BSI

Any width or thickness change # 1 in 4 slope Other weld processes and positions than those in type 6.2. Grind smooth any undercut

6.3

E

This category can be raised to class D provided that the reinforcement is ground smooth until flush with the plate surface on both sides, in addition to grinding smooth any undercut. Any NDT should be carried out after grinding.

On permanent backing strip No permanent tack welds within 10 mm of edge. Grind smooth any undercut

ww

w. b
6.4 F

If the backing strip is fillet or tack welded to the plate (type 5.2) the detail class will not be reduced below class F unless permanently tacked within 10 mm of the member edge, in which case it will be class G (type 5.5).

zf xw .

Longitudinal axes in line

Flat position shop welds, not submerged arc. Grind smooth any undercut

6.2

D

Shop welds made entirely in the flat position, either manually or by an automatic process other than submerged arc, tend to have a better reinforcement shape from the point of view of fatigue than positional, site or submerged arc welds, i.e. larger re-entrant angles at the toes.

co m

? BSI 03-1999

Minimum 6.1 transverse cross section of member at location of crack initiation

The significance of flaws should be determined with the aid of specialist advice and/or by the use of a fracture mechanics analysis. The NDT technique should be selected with a view to ensuring the detection of such significant flaws. This class should not normally be used in structural work. (See B.2.2 and B.6.2.6.)

BS 7608:1993

a

The joints covered by this table may also fail from internal weld flaws if they are more severe than the external geometrical discontinuity. Weld quality is therefore pertinent to the various classifications (see 2.4.3).

13

Table 6 — Classification of details: full penetration butt welds between co-planar platesa
Product form Location of potential crack initiation At transverse butt weld joining two single plates end to end Dimensional requirements Manufacturing requirements Special inspection requirements Proved free of all flaws which are likely to degrade the joint below its stated classification (see 4.3) Design stress area Minimum transverse cross section of member at location of crack initiation Type number Class Notes Sketch

Where the end of one plate is butt welded to the surface of another, refer types 8.1 and 8.2.

a

The joints covered by this table may also fail from internal weld flaws if they are more severe than the external geometrical discontinuity. Weld quality is therefore pertinent to the various classifications (see 2.4.3).

ww

w. b

zf xw .

BSI

co m

14
Rolled steel plates only

BS 7608:1993

Longitudinal axes in line

Build up corners to radius # 0.15W Corner reinforcement ground flush for 2t Grind smooth any undercut

6.5

F2

Abrupt width change

The effect of the stress concentration at the corner of the joint between two individual plates of different widths in line is included in the classification. Where the end of one plate is butt welded to the surface of another refer to types 8.1 and 8.3. Stress concentrations due to abrupt changes of width can often be avoided by tapering the wider plate (see types 6.2, 6.3 and 6.4).

.

? BSI 03-1999

Table 7 — Classification of details: transverse butt welds in sections and tubes
Product form Rolled steel sections and built-up members Location of potential crack initiation At transverse butt weld joining two sections of similar profile end to end. Dimensional requirements Manufacturing requirements Special inspection requirements Design stress area Type number Class Notes Sketch

Steel tubes

At the toes of circumferential butt welds in tubes

Weld made from both sides Weld made from one side on permanent backing strip

Weld made from Weld to be one side with no checked for full backing strip penetration

ww

w. b
7.3 7.4 E F 7.5 F2

Weld made from both sides with overfill dressed flush

Proved free of all flaws which are likely to degrade the joint below its stated classification (see 2.4.3)

7.2

zf xw .
C

BSI

This joint is frequently made using a cope hole. This gives improved access to the flange butt welds when webs or longitudinal stiffeners have already been attached. The end of the web butt weld at the cope hole can be considered to be equivalent to class D with the appropriate stress concentration factor provided that the end of the butt weld and the reinforcement within a distance equal to the radius r are ground flush (see Figure B.8). If there is no grinding class E with the appropriate stress concentration factor should be used. Cope holes of triangular shape are not recommended.

The significance of flaws should be determined with the aid of specialist advice and/or by the use of a fracture mechanics analysis. The NDT technique should be selected with a view to ensuring the detection of such significant flaws. Use of this type is not recommended.

co m

? BSI 03-1999

Full penetration Misalignment slope # 1 in 4. butt weld with longitudinal axes in line.

7.1 Minimum transverse cross section of member at location of crack initiation.

F2

Butt welds between rolled sections or between built-up sections are prone to weld flaws, which are difficult to detect, in the region of the web/flange junction. Special preparations, procedures and inspection may be undertaken in exceptional circumstances and type 6.3 may then be applied unless the weld is made on a permanent backing (type 6.4). Dressing of the weld reinforcement is advised to overcome poor reinforcement shape resulting from the greater misalignments which may occur in the jointing of sections.

BS 7608:1993

15

Table 7 — Classification of details: transverse butt welds in sections and tubes
Product form Steel tubes Location of potential crack initiation At the toes of circumferential butt welds between tubes and conical sections Dimensional requirements Manufacturing requirements See types 7.2 to 7.5, as appropriate Special inspection requirements See types 7.2 to 7.5 as appropriate Design stress area Type number 7.6 Class Notes Sketch

ww

w. b

zf xw .

BSI

co m

16
? BSI 03-1999

BS 7608:1993

C E F F2

Class and stress should be those corresponding to the joint type as indicated in types 7.2 to 7.5, but the stress has also to include the stress concentration factor due to overall form of the joint. If a stiffener or diaphragm is situated adjacent to the joint, see also type 10.3

Table 8 — Classification of details: load carrying fillet and T-butt joints between plates in different planes
Product form Rolled steel structural plates, sections and built-up members Location of potential crack initiation At toe of weld joining two members end to end with third member transverse through joint Dimensional requirements Full penetration butt weld with longitudinal axes in line Manufacturing requirements Any undercut should be ground smooth particularly on the corners of member X Special inspection requirements All regions of plate Y stressed in the through thickness direction to be free from lamellar defects and tears. Design stress area Cross section of member X Type number 8.1 Class Notes Sketch

Member Y can be regarded as one with a non-load-carrying weld (see joint types 5.2 and 5.5). In this instance the edge distance limitation applies. Misalignment between the two plates X shall be kept to be a minimum.

Partial penetration butt or fillet weld

8.2

ww

w. b
8.4 F2

At toe of weld joining the end of one member to the surface of another

Full penetration butt weld

8.3

Partial penetration butt or fillet weld

zf xw .
F2 In this type of joint failure is likely to occur in the weld throat unless the weld is made sufficiently large. (See joint type 8.5). Misalignment between the two plates X shall be kept to be a minimum. It may be necessary to include a stress concentration factor in the design calculation (see B.8.2). F See joint type 8.1. See joint type 8.2

If width permits make weld continuous around the joint; otherwise grind ends flush with edge of member X.

It may be necessary to include a stress concentration factor in the design calculation (see B.8.2).

BSI

.

co m

? BSI 03-1999

F

Weld metal failure will not govern with full penetration welds.

BS 7608:1993

NOTE Butt welded joints should be made with an additional reinforcing fillet so as to provide a similar toe profile to that which would exist in a fillet welded joint (see B.5.2.2).

17

Table 8 — Classification of details: load carrying fillet and T-butt joints between plates in different planes
Product form Rolled steel structural plates, sections and built-up members Location of potential crack initiation In weld throat of any incompletely fused joint Dimensional requirements Manufacturing requirements Special inspection requirements Design stress area Effective weld throat area Type number 8.5 Class Notes Sketch

At toe of transverse load carrying fillet weld in lap joint on both sides of plate symmetrically

Weld toe not less than 10 mm from edge of member

See notes

8.6

F2

The classification may be deemed to include stress concentrations arising from normal eccentricities in the thickness direction. Where a narrow attachment, Y, is transferring the entire load out of a wide member, X, as in the case of a welded lap type connection between, for example, a cross brace and a gusset, the stress in the gusset at the end of the cross brace will vary substantially across the section. For assessing the stress in the gusset X the effective width should be taken as shown below: Minimum weld length 150 mm

At toe of transverse load carrying fillet weld in lap joint on one surface only

8.7

ww

NOTE Butt welded joints should be made with an additional reinforcing fillet so as to provide a similar toe profile to that which would exist in a fillet welded joint (see B.5.2.2).

w. b

zf xw .

G

BSI

For failure in the cross brace at Z the cross brace, Y, is the “member” and the gusset is the “attachment”. (See type 8.8.) Minimum weld length 150 mm

co m

18
? BSI 03-1999

BS 7608:1993

W

This includes joints in which a pulsating load may be carried in bearing, such as the connection of bearing stiffeners to flanges. In such examples the welds should be designed on the assumption that none of the load is carried in bearing.

Table 8 — Classification of details: load carrying fillet and T-butt joints between plates in different planes
Product form Location of potential crack initiation At the ends of longitudinal load carrying fillet welds, with the weld end on a plate edge Dimensional requirements Manufacturing requirements Special inspection requirements Design stress area Cross section of member Y Type number Class Notes Sketch

Avoid weld returns round laps

NOTE Butt welded joints should be made with an additional reinforcing fillet so as to provide a similar toe profile to that which would exist in a fillet welded joint (see B.5.2.2).

ww

w. b

zf xw .

BSI

co m

? BSI 03-1999

Rolled steel structural plates, sections and built-up members

Weld toe within 10 mm of edge of member

Grind any undercut

8.8

G

This type applies regardless of the shape of the end of the attachment. In all cases, care should be taken to avoid undercut on element corners or to grind it out to a smooth profile should it occur. In particular, weld returns across a corner should be avoided and the use of cover plates wider than the flange, to which they are attached, is not recommended.

BS 7608:1993

19

Table 9 — Classification of details: slotted connections and penetration through stressed members
Product Location of Dimensional Manufacturing Special Design Type Class form potential crack requirements requirements inspection stress number initiation requirements area Steel plates, sections or tubes In stressed member at toe of butt weld connecting slotted through member Length of slotted through member, parallel to Sr # 150 mm and edge distance $10 mm 9.1 F Notes Sketch

Any length and edge distance < 10 mm In parent or weld metal adjacent to a penetration, on a plane essentially perpendicular to Sr

9.3

G

9.4

D

In this situation the relevant stress should include the stress concentration factor due to the overall geometry of the detail. Full penetration welds are normally required in this situation. If they are not used, the possibility of failure through the weld has to be considered (see type 9.6)

9.5

F

For failure at weld toe (see sketch).

ww

w. b

In weld metal of fillet or partial penetration welds around a penetration, on a plane essentially parallel to Sr

9.6

W

The stress in the weld should include an appropriate stress concentration factor to allow for the overall joint geometry. This type of failure could subsequently lead to failure of type 9.4. Weld metal failure does not govern with full penetration welds.

zf xw .

BSI

co m

20
? BSI 03-1999

BS 7608:1993

,

Length > 150 mm and edge distance > 10 mm

9.2

F2

This classification does not apply to fillet welded joints (see joint type 8.2). However it does apply to loading in either direction (L or T in the sketch).

.

Table 10 — Classification of details: circular tubular members
Product form Steel tubes Location of potential crack initiation Adjacent to the toes of full penetration welds in nodal joints At toe of weld attaching diaphragm or stiffener to tubular member Dimensional Manufacturing requirements requirements Special inspection requirements Design stress area Type number 10.1 10.2 T F Class Notes Sketch

In the stressed member at the toes of bevel butt or fillet welded attachment in a region of stress concentration

10.3

In the tube at the toe of full penetration or fillet welded gusseted connections Weld throat failure in fillet welded gusseted connections

w. b
10.4 F Effective weld throat 10.5 W

ww

zf xw .
F or F2 or G

Class depends on attachment length (see types 5.2 to 5.4) but stress should include the stress concentration factor due to overall . shape of adjoining structure

BSI

The design stress has to include any local bending stress adjacent to the weld end. For failure in the weld throat of fillet welded joints.

co m
.

? BSI 03-1999

In this situation design should be based on the hot spot stress as given in 3.6.2. Stress should include the stress concentration factor due to overall shape of adjoining structure.

BS 7608:1993

21

Table 11 — Classification of details: seam welds
Product form Location of potential crack initiation From the root or surface of a joggle joint, transverse to the weld Dimensional requirements Manufacturing requirements Special inspection requirements Design stress area Type number Class Notes Sketch

From the weld root of a joggle joint

11.2

F

From the root or surface of a continuous fillet weld forming a lap joint, transverse to the weld.

11.3

In the weld throat of a lap joint

ww

From the weld toe of a lap joint

w. b

11.4

Weld throat

11.5

zf xw .

D

BSI

F2

Allowance has to be made for bending due to the non-alignment of forces.

W

co m

22
Steel plates

BS 7608:1993

11.1

D

? BSI 03-1999

Table 12 — Classification of details: branch connections
Product form Location of potential crack initiation At crotch corner Dimensional requirements Manufacturing requirements Special inspection requirements Design stress area Type number Class Notes Sketch

At weld toe in shell

12.2

At weld toe in branch

ww

w. b
12.3 F

zf xw .
F Stress has to include the relevant stress concentration factor. Stress has to include the relevant stress concentration factor.

BSI

co m

? BSI 03-1999

Steel plates

12.1

D

Stress has to include the relevant stress concentration factor.

BS 7608:1993

23

BS 7608:1993

2.3 Unclassified details
2.3.1 General Details which are not expressly classified should be treated as class G, or class W for load-carrying weld metal, unless a higher classification can be justified either by reference to published experimental work or by carrying out special tests. Such tests should be sufficiently extensive to allow the basic or design S-N curve to be determined in the manner used for the standard classes (see Annex E). 2.3.2 Post-welding treatments Where the classification of Table 1 to Table 12 does not give adequate fatigue resistance, the performance of weld details may be improved by post-welding treatments such as controlled machining, grinding or peening. When this is required, and the proposed improvement method is not covered by 4.3, or the improvement stipulated therein is insufficient, the detail should be classified by testing (see 2.3.1). It is recommended that no advantage of improvement techniques should be taken at the initial design stage, but they can represent a useful option when the need to increase fatigue life is discovered, for example, at a late stage of fabrication or when the structure is already in service.

2.4 Workmanship and inspection
2.4.1 General

2.4.2 Detrimental effects

The following occurrences can result in a detail exhibiting a lower performance than its classification would indicate: a) weld spatter; b) accidental arc strikes; c) unauthorized attachments (temporary or permanent); d) corrosion pitting; e) weld flaws, particularly in transverse butt welds (see 2.4.3); f) poor fit-up in bolted connections; g) eccentricity and misalignment.

ww

In general, welding workmanship should be in accordance with BS 5135:1984. Where the classification of a detail is dependent upon particular manufacturing or inspection requirements, the necessary standards of workmanship and inspection should be indicated on the relevant drawings; this applies equally to welded and bolted connections. In addition, the class of any critical detail which is classified as class D or better should also be indicated on the drawings.

Transverse butt welded joints (see Table 6) usually fail in fatigue by cracking from the weld toe and the classifications given in Table 6 relate to this mode of failure. However, these welds are also potentially vulnerable to failure from internal flaws. If, therefore, the fatigue strengths indicated by the classifications given in Table 6 are to be achieved, it is essential that the internal quality of the welds is also of a sufficiently high standard. In the case of stresses corresponding to class C, for practical purposes a “flaw-free” weld is required. For class D the upper limits of flaw should be determined in accordance with PD 6493:1991. The maximum permissible flaw sizes of planar flaws or slag inclusions corresponding to classes E to G are shown in Table 13. For other types of flaw, e.g. porosity, and for joints in classes other than those given in Table 13, the significance of any flaws should be assessed by reference to PD 6493:1991. Similarly for all joint classifications, guidance relating to the interaction of flaws in close proximity to each other is also given in PD 6493:1991. It is not always necessary to achieve the classification given in Table 6, and in such cases weld quality may be relaxed. For example, suppose that in a tension member there is a type 6.2 butt weld adjacent to a transverse non-load-carrying fillet weld (type 5.2, see Figure 1). In this case, assuming that no other details of lower classification are present, the design stress will be limited to class F by the fillet weld. There is then no need for the adjacent butt weld to have a fatigue strength above class F, and the maintenance of internal quality to the standard required for class F will be adequate.

w.

bz

fx

w.

co

2.4.3 Quality categories for transverse butt welds

m

It is important, therefore, that in fatigue loaded structures fabrication should adequately be controlled and that they should subsequently be inspected with these matters in mind. It is also necessary to ensure that regular inspections are carried out during the service life to check that no unauthorized attachments have been made to the structure in regions where such attachments would reduce the design class. Any attachments found during such inspections should be removed, and the area ground out and checked by non-destructive testing to ensure that no unacceptable flaws remain.

24

? BSI 03-1999

BS 7608:1993

2.5 Welded steel decks
The classifications given in Table 1 to Table 12 should not be applied to welded joints in orthotropic steel decks of, for example, highway bridges. Complex stress patterns usually occur in such situations and specialist advice should be sought for identifying the stress range and joint classification.

ww

w.

bz

fx

w.
? BSI 03-1999

co
25

m

Table 13 — Upper limits of flaw area of planar flaws and slag inclusions in transverse butt welds (see note 1)
Stresses corresponding to class: Flaws more than 5 mm from the near surface: Depth between centre of flaw and the nearer surface: 7.5 mm mm2 10 mm mm2 12.5 mm mm2 25 mm mm2 50 mm mm2 15 mm mm2 20 mm mm2 Flaws within 5 mm of a surface: Plate thickness: 25 mm mm2 50 mm mm2 100 mm mm2

G

112

168

260





112

co m

ww

w. b

zf xw .

BSI

26 E F F2
? BSI 03-1999

BS 7608:1993

28 54 80

40 80 110

54 110 160

110 260 440

180 560 —

13 33 60

17 45 80 160

22 60 112 240

38 116 220 —

60 200 — —

NOTE 1 The flaw areas given in this table relate to the area of the circumscribing rectangle on a plane perpendicular to the direction of stressing. Acceptable areas for intermediate depths or thicknesses may be obtained by interpolation. Where areas are not quoted (for lower classes and large depths) they are so large that the flaws would certainly be repaired in any event. NOTE 2 Specialist advice should be sought on the interpretation of the limits in this table in terms of ultrasonic test signal characteristics.

NOTE 3 Clustered flaws, other than porosity, are not acceptable; porosity should be limited to a level where the detection of other flaws is not prejudiced. NOTE 4 It may be necessary to impose more stringent limits than those shown in this table in order to ensure that adequate weld quality is maintained and/or to avoid other modes of failure.

BS 7608:1993

Section 3. Stress calculations
3.1 General
The procedure for the fatigue analysis for welded structures is based on the assumption that it is only necessary to consider ranges of cyclic stress in determining the fatigue life, i.e. mean stresses are neglected. However, for non-welded details subjected to stress cycles which are partly or wholly in compression the effective stress range is modified (see 3.4). Unless otherwise noted in Table 1 to Table 12, the stress should be based on the net section. Where appropriate, allowance should be made for geometrical stress concentrations (see 3.6). In estimating the maximum principal stress, shear and torsional effects may be neglected where they are small, i.e. less than 15 % of a coexistent direct stress. Stress ranges for post-weld heat-treated joints should be assumed to be the same as for as-welded joints (but see 4.3).

NOTE

For tubular nodal joints see 3.6.2.

If the direction of principal stress rotates during the stress cycle, e.g. at the end of a web stiffener on a crane gantry girder where the shear stress changes as the crane wheels pass over it, the cyclic stress should be derived from the stresses calculated at the two extremes, i.e. the peak and trough, of the stress cycle. The peak and trough values of principal stress should be those on principal planes which are not more than 45° apart. Thus, if ?x, ?y and ? are the coexistent values (with appropriate signs) of the two orthogonal direct stresses and the shear stresses at the point under consideration, the relevant principal stress will be selected if either: a) ?x – ?y is at least double the corresponding shear stress ? at both peak and trough; or b) the signs of ?x – ?y and T both reverse or both remain the same at the peak and the trough. In either a) or b), provided that ?x2 > ?y2 at both peak and trough, the required stress range will be the algebraic difference between the numerically greater peak principal stress and the numerically greater trough principal stress.

bz

ww

w.

fx

In most situations the potential fatigue crack will be located in parent material adjacent to some form of stress concentration, e.g. at a weld toe or bolt hole. Provided that the direction of the principal stress does not change significantly in the course of a stress cycle, the relevant cyclic stress for fatigue assessment should then be taken as the maximum range through which any principal stress passes in the parent metal adjacent to the potential crack location, as shown in Figure 2. Tension stresses are considered positive and compression stresses negative. In practice the through-thickness component of stress is rarely relevant and can usually be ignored.

3.3 Stress range for welds
In load-carrying partial penetration or fillet-welded joints, where cracking could occur in the weld throat, the relevant stress range is the maximum range of shear stress in the weld metal. This should be taken, for each stress cycle, as the algebraic or vector difference between the greatest and least vector sum of the shear stress, based upon the effective dimensions of the weld throat (see Figure 3). It should be assumed that none of the load is carried in bearing between parent materials. In situations where the length of the weld is parallel to the direction of stress, fatigue of the weld may be ignored.

w.
3.5.1 General

3.4 Effective stress range for details in unwelded members in which the whole or part of the stress is compressive
In unwelded details where the stress range, allowing for dead load and residual stresses (due to fabrication), is entirely compressive, the effects of fatigue loading may be ignored. However, when the resultant stress range involves stress reversals through zero, the effective stress range to be used in the fatigue assessment should be obtained by adding 60 % of the range from zero stress to maximum compressive stress to that part of the range from zero stress to maximum tensile stress.

3.5 Calculation of stresses
Stresses should be calculated using elastic theory and taking account of all axial, bending and shearing stresses occurring under the design loading. No redistribution of loads or stresses, such as may be allowed for checking static strength at ultimate limit state, including implicit allowance for redistribution in simplified elastic design rules, or for plastic design procedures, should be made.

co

m

3.2 Stress range in parent material

? BSI 03-1999

27

BS 7608:1993

3.5.2 Effects to be ignored The effects of the following need not be included in stress calculations: a) residual stresses in welded details; b) stress concentrations due to: 1) a weld detail itself; 2) bolt, rivet or small drilled holes (but excluding types 1.3 and 1.4 in Table 1). 3.5.3 Effects to be included All stresses arising from sources other than those specifically excluded in 3.5.2 should be included in the analysis. For example, the following stresses should be included together with any others that influence the stress applied to the joint: a) stress concentrations due to the overall joint geometry (see 3.6); b) eccentricities or misalignment occurring in a joint detail, except where otherwise indicated in this code; c) shear lag, restrained torsion and distortion, transverse stresses and flange curvature; d) stress distribution in wide plates; e) cracking of concrete in composite elements; f) stresses in triangulated skeletal structures due to load applications away from joints, member eccentricities at joints and rigidity of joints; g) residual stresses (but for non-welded details under, nominally, fully or partly compressive loading; see 3.4); h) fabrication tolerances, except as covered in B.6 for misalignment.

3.6 Geometrical stress concentrations
3.6.1 General

Unless otherwise indicated in Table 1 to Table 12 the stress concentrations inherent in the make-up of a welded joint (arising for example from the general joint geometry and the weld shape) have been taken into account in the classification of the details. However, where there is an additional geometrical discontinuity, such as an aperture or a change of cross section (see Figure 4 and Figure 5), which is not a natural characteristic of the standard detail category itself, the resulting stress concentration relevant to fatigue design should be determined either by special analysis or, where appropriate, by the use of predefined stress concentration factors, such as those given in Figure 6.

ww

w.

bz

Figure 4 shows a typical example of a joint which could be analysed using the stress concentration factors given in Figure 6(a) while Figure 5 shows an example of a joint which would require special analysis because the geometrical layout of the joint obviously creates a “hard spot” and hence a non-uniform stress distribution. In general form the joint is obviously type 8.4 (class F2). Conversely, Figure B.10 illustrates a case where flexibility produces a non-uniform stress distribution. In either case the corresponding applied stress for design purposes should be the geometrical stress at the weld toe (obtained by extrapolation) but excluding the local stress concentration caused by the fillet weld (see Figure 5). Other typical examples are tubular nodal joints (see 3.6.2) and the joints shown in Figure B.3, Figure B.5, Figure B.6, Figure B.8 and Figure B.9. For the purposes of this code, these geometrical stresses are regarded as being equivalent to the fatigue stress concentration factor (Kf) multiplied by the nominal stress (see Figure 5). The derivation of any stress concentration factor used should be validated taking account of the guidance given in Annex C. In general, for joints for which standard stress concentration factors are not applicable and which therefore require special analysis, the fatigue stress at the weld toe should be obtained by extrapolation to the weld toe position from results calculated at distances of 0.4t and greater from the weld toe.
NOTE This is to avoid taking into account the stress concentration effect of the weld itself.

fx

If the fatigue stress concentration factor, Kf, is obtained from stress analysis or strain measurements on prototype or actual structures, the aim should be to determine the stress close to the detail., e.g. crotch corner, weld toe, but excluding the stress concentration due to the detail itself. In general, a suitable stress is that which would be measured over a gauge length of 3 mm to 5 mm starting 0.1t (but not more than 5 mm) from the detail, where t is the plate thickness. A similar criterion should be applied if use is made of published data obtained for geometrically similar structures.

w.

co

m

28

? BSI 03-1999

BS 7608:1993

3.6.2 Tubular nodal joints For nodal joints, the stress range to be used in the fatigue analysis 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 set-on connection at least four equally spaced points around the joint periphery will need to be considered. For any particular type of loading, e.g. axial loading this “hot spot” stress range is the product of the nominal stress range in the brace and the appropriate stress concentration factor (SCF). The “hot spot” stress is defined as the greatest value of the direct stress around the brace/chord intersection of the extrapolation to the weld toe of the geometric stress distribution near the weld toe. This hot spot stress incorporates the effects of overall joint geometry, i.e. the relative sizes of brace and chord, but omits the stress concentrating influence of the weld itself, which results in a local stress distribution (see Figure 7). Hence, the “hot spot” stress is considerably lower than the peak stress but provides a consistent definition of stress range for application with the design S-N curve (curve T shown in Figure 8 and Figure 9). Stress ranges on both the brace and chord sides should be considered in any fatigue assessment. The calculation of “hot spot” stress may be undertaken in a variety of ways, e.g. by physical model studies, finite element analysis, or by use of semi-empirical parametric formulae. The position of the “hot spot” in relation to the crown and saddle can be determined by the first two methods but not in all cases by parametric equations. When physical models are used, care should be taken in obtaining the geometric stress extrapolated to the weld toe as described above. When 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 brace/chord 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.

3.7 Stresses in welds attaching shear connectors
3.7.1 General For shear connectors conforming to the dimensional recommendations of BS 5400-5:1979, the design stresses for fatigue in the weld metal should be calculated in accordance with 3.7.2 and 3.7.3. Where the dimensions of the shear connectors and/or the concrete haunches are not in accordance with BS 5400-5:1979, the fatigue strength should be determined by test (see Annex E).
NOTE The assessment of the effects of local wheel loads on shear connectors between concrete slabs and steel plates is beyond the scope of this code. This effect may, however, be ignored if the concrete slab alone is designed for the entire local loading.

3.7.2 Stud connectors The stresses in the weld attaching stud shear connectors should be calculated from the following expression (from BS 5400-5:1979): stress in weld = 2 shear load on stud ------------------------------------------------------------------------- × 425 N/mm appropriate nominal static strength of the connector 3.7.3 Channel and bar connectors The stresses in the weld metal attaching channel and bar shear connectors should be calculated from the effective throat area of weld, transverse to the shear flow, when the concrete is of normal density and from 0.85 × throat area when lightweight concrete is used. For the purposes of this clause the throat area should be based on a weld leg length which is the least of the dimensions tabulated below:
Channel connector Bar connector

w.

bz

fx

w.

ww

Actual leg length Thickness of channel web Half the thickness of beam flange

co

NOTE In normal density concrete, where the thickness of the beam flange is at least twice the actual weld leg length and the weld dimensions conform to BS 5400-5, the effective weld areas are: 50 × 40 bar connectors × 200 mm long, 1697 mm2; 25 × 25 bar connectors × 200 mm long, 1018 mm2 127 and 102 channel connectors × 150 mm long, 1272 mm2; 76 channel connectors × 150 mm long, 1 081 mm2.

m

0.6 × actual leg length 0.125 × (height + breadth of bar) Half the thickness of beam flange

? BSI 03-1999

29

BS 7608:1993

3.8 Axial stresses in bolts
This clause applies only to bolts with threads conforming to BS 3643-2:1981, manufactured in accordance with BS 3692:1967, BS 4395-1:1969, BS 4395-2:1969 or to BS 4190:1967. Black bolts conforming to BS 4190:1967 should only be used under fluctuating axial load if they are faced under the head and turned on the shank in accordance with BS 4190:1967. This clause is not applicable if the bolts are welded. The stress range should be calculated on the tensile stress area of the bolt and should include the effects of axial and bending loads, including any effect of prying. It should take into account the pre-load in the bolt and the compressibility and specified fit-up of the connected parts. If reliance is to be placed on this pre-load, it should be at least 1.5 times the design tension including the effects of prying.
NOTE 1 Permissible stresses in bolts are covered in 4.2.2.

3.9 Derivation of stress spectra
In situations in which the loading spectrum is not specified, e.g. in a relevant application standard, it will be necessary to derive the expected spectrum. In general the total loading on a structure will be composed of several different loading events, each with different magnitudes, geometrical arrangements and frequencies of occurrence. To derive the design spectrum for any particular detail an appropriate mix of loading events, representative of the loading to be expected in a given time interval, should be considered. It should be applied to the relevant influence line(s) in order to obtain the pattern of stress fluctuations to which the detail will be subjected in that time interval. Account should be taken of the possibility of two or more loading events occurring simultaneously, or following each other in particular orders, such that higher stress ranges may be caused. This pattern should be broken down into a convenient spectrum of cycles, expressed in terms of stress ranges Sri and numbers of applications ni, by the reservoir counting method (see Annex F). The numbers of cycles so counted should be combined with the appropriate total numbers of occurrences of the various loading events in the design life of the structure to compile the overall design spectrum. In most computer programs used for the derivation of cycle counts the results are accumulated into a relatively small number of stress range intervals. In most instances about 40 intervals will be sufficient.

ww

NOTE 2 If the bolt is not pre-tensioned to the minimum proof load, the stress range in the bolt will be a higher proportion of the applied load.

w.

Due to the form of the relationship between the increase in stress in the bolt and the externally applied load, the stress range should be based on the full applied load rather than the fluctuating portion of the load. For a bolt which has been pre-tensioned to the minimum proof load, the stress range in the bolt is up to approximately 20 % of the load applied to the bolt including the additional force due to prying action. Where the design is based on a simple elastic or plastic analysis of the connection, the stress range should be taken as 20 % of the applied load including the prying force. Where the design procedure includes suitable checks to limit the extension of the bolt at working load, the stress range may be taken as 10 % of the applied load including the additional force due to prying action.

bz

fx

w.

co

m

30

? BSI 03-1999

BS 7608:1993

Section 4. Allowable fatigue stresses
4.1 Tensile stress limitations
The procedures given in this code for deriving fatigue stresses should only be considered as valid provided that the calculated maximum fibre stress on the net area of a member, remote from geometric stress concentrations and excluding self-regulating stresses (such as residual or thermal stresses) does not exceed 60 % of yield stress under “normal operating conditions” and 80 % of yield stress under “extreme loading conditions”. In this context the predicted number of cycles exceeding “normal operating conditions” in the anticipated life of the structure should be not greater than 100. All the basic Sr-N curves are applicable, for the relevant value of d, to joints which are either: a) in air; or b) exposed to sea water, but adequately protected from corrosion.
NOTE The effectiveness of cathodic protection in relation to fatigue has not yet been proven for structural steels with yield stress, ?Y > 400 N/mm2.

For design purposes, however, the curves may have to be modified in order to allow for the factors given in 4.3. 4.2.2 Axially loaded bolts The S-N curves for axially loaded bolts with cut, ground or rolled threads up to 25 mm diameter (class X) are shown in Figure 10 and are expressed in terms of the following ratio:

4.2.1 Plain material and welded or bolted joints For each class of joint the relationship between the applied stress range, Sr, and the number of cycles to failure, N, under constant amplitude loading conditions is of the following form: log N = log Co – d/? – m log Sr where Co is a constant relating to the mean Sr-N curve; d ?

fx
(2) (3)

is the number of standard deviations below the mean; is the standard deviation of log N;

bz

log Cd = log Co – d?

then equation (1) can be written as: SrmN = Cd

Thus, from equations (1), (2) and (3), the required basic S-N curve can be derived for any desired value of d. The nominal probability of failure corresponding to various possible values of d, based upon an assumed normal distribution, is shown in Table 15. The standard basic Sr-N curves are to be taken to represent two standard deviations below the mean lines i.e. d = 2. They are shown in graphical form in Figure 9, and the corresponding values of C2 are included in Table 14.

ww

The relevant values of these terms for joints in air are shown in Table 14, and the mean line relationships are plotted in Figure 8. They are applicable to all steels covered by this code, using the equation:

w.

m is the inverse slope of the log Sr versus log N curve [see equation (3)].

w.
(1)
NOTE

For bolts with diameter greater than 25 mm the Sr ? fatigue strengths defined by the class X ? ------------- –N ? UTS? curves should be reduced in accordance with 4.3.2.
UTS is the ultimate tensile strength.

Regardless of the actual tensile strength of the bolt material, it should never be assumed to be greater than 785 N/mm2.

4.3 Modifications to basic S-N curves
4.3.1 General In order to derive the design Sr-N curves, the basic Sr-N curves should be modified, as appropriate, to allow for the factors given in 4.3.2 to 4.3.5. 4.3.2 Effect of material thickness The fatigue strength of welded joints and of bolts is to some extent dependent on material thickness, strength decreasing with increasing thickness. The basic Sr-N curves relate to the following thicknesses and bolt diameters: a) nodal joints (class T): 16 mm; b) non-nodal joints (classes B to G): up to 16 mm; c) bolts (class X): up to 25 mm diameter.

co

4.2 S-N curves

stress range on the tensile stress area ---------------------------------------------------------------------------------------------------------------------------nominal tensile strength of the bolt material

m

? BSI 03-1999

31

BS 7608:1993

Table 14 — Details of basic S-N curves
Class Co Log10 Co Loge m Standard deviation, ? Log10 Loge C2 So (N = 107 cycles) N/mm2

B C D E F F2 G W S T
NOTE

2.343 × 1015 1.082 × 1014 3.988 × 1012 3.289 × 1012 1.726 × 1012 1.231 × 1012 0.566 × 1012 0.368 × 1012 2.13 × 1023 4.577 × 1012

15.3697 35.3900 4.0 14.0342 32.3153 3.5 12.6007 29.0144 3.0 12.5169 28.8216 3.0 12.2370 28.1770 3.0 12.0900 27.8387 3.0 11.7525 27.0614 3.0 11.5662 26.6324 3.0 23.3284 53.7156 8.0 12.6606 29.1520 3.0

0.1821 0.2041 0.2095 0.2509 0.2183 0.2279 0.1793 0.1846 0.5045 0.2484

0.4194 0.4700 0.4824 0.5777 0.5027 0.5248 0.4129 0.4251 1.1617 0.5720

1.01 × 1015 4.23 × 1013 1.52 × 1012 1.04 × 1012 0.63 × 1012 0.43 × 1012 0.25 × 1012 0.16 × 1012 2.08 × 1022 1.46 × 1012

100 78 53 47 40 35 29 25 82 53a

For details of the S-N curve for class X see 4.2.2.

For example the T curve expressed in terms of log10 is: log10(N) = 12.6606 – 0.2484d – 3log10(Sr)
a Idealized

hot spot stress.

ww

t B?1 ? 4 S = S B ? ----? t? where S SB t

w.
(4)

For joints of other thicknesses, apart from those involving longitudinal welds (types 4.1 to 4.4, 11.1, 11.3 and 12.1) or bolts of other diameters, or where there is substantiated evidence that thickness does not influence the fatigue strength of the type of joint under consideration, correction factors on life or stress should be applied to produce a relevant Sr-N curve. The correction on stress range is of the following form:

is the fatigue strength of the joint under consideration; is the fatigue strength of the joint using the basic Sr-N curve; is the greater of 16 mm or the actual thickness of the member or bolt diameter under consideration; is the maximum thickness relevant to the basic Sr-N curve (i.e. tB = 16 mm for welded joints or 25 mm for bolt diameters).

tB

No thickness correction need be applied in the case of butt welds with the weld reinforcement machined flush, e.g. joint types 4.1, 6.1 and 7.1.

bz

fx

4.3.3 Effect of sea water in relation to unprotected joints For unprotected joints exposed to sea water the basic Sr-N curves should be reduced by a factor of 2 on life for all joint classes. In addition, the correction relating to the number of small stress cycles (see 4.4) is not applicable. For high strength steels (?Y > 400 N/mm2) these penalties may not be adequate and the use of such materials under corrosion fatigue conditions should be approached with caution. 4.3.4 Effect of weld improvement by toe grinding For welded joints involving potential fatigue cracking from the weld toe, which are given a remedial treatment by controlled local machining or grinding of the weld toe, the S-N curve can be assumed to be improved in strength by 30 % (but see 2.3.2). This is equivalent to a factor of 2.2 on life.

w.

co

m

32

? BSI 03-1999

BS 7608:1993

Nominal probability of failure %

31 16 2.3 0.14

a Mean-line b

curve. The standard design curve.

NOTE 1 The benefit of grinding may be claimed only for welded joints which are adequately protected from corrosion, since the presence of a corrosive environment can cause pitting in the dressed region. NOTE 2 Also in the case of partial penetration and fillet welds, where failure may occur from the weld root, grinding of the weld toe cannot be relied upon to give an increase in strength.

ww

w.
0.5 1.0 3.0 2.0b

50

0a

bz
d

Toe grinding is normally carried out either with a rotary burr or by disc grinding, but the treatment should be approached with caution. It should produce a smooth concave profile at the weld toe with the depth of the depression penetrating into the plate surface to at least 0.5 mm below the bottom of any undercut or flaw at the weld toe which can be located by magnetic particle or liquid penetrant inspection (see Figure 11). The maximum depth of local machining or grinding should not exceed 2 mm. If the reduction in plate thickness exceeds 5 % this should be taken into account in the stress calculation. Particular care should be taken if disc grinding is used since that can easily result in an excessive reduction in thickness. In the case of a multi-run weld more than one weld toe may need to be dressed. Where toe grinding is used to improve the fatigue life of fillet welded connections, care should be taken to ensure that the required throat thickness is maintained. In some situations, e.g. a nodal joint between tubes of nearly equal diameter under out-of-plane bending stresses, the stress distribution over the weld surface may be fairly uniform. In that case grinding the weld toe will merely transfer failure to an adjacent “toe” between surface beads of the weld, so that the whole weld surface would need to be ground (see Figure 12). Table 15 — Nominal probability factors

4.3.5 Effect of stress relief for welded details If the applied stress range is fully tensile, stress relief will have no effect on fatigue strength. However, if it can be demonstrated that a compressive component of the combined applied and residual stress exists, and can be quantified, it would be permissible to assume that the relevant value of Sr is the sum of the tensile component and 60 % of the compressive component. It is because this recommendation is unlikely to be satisfied very often that the benefit is generally ignored (see 3.2). In assessing the potential influence of stress relief account should be taken of the following facts that: a) the extent to which stress relief actually reduces the residual stress in large complex joints is largely unknown; b) if joints are locally stress-relieved and then welded into a complete structure long range residual stresses will still exist.
NOTE Such long range stresses will then be magnified by the inherent stress concentration present around the joint.

fx

w.

4.4 Treatment of low stress cycles
Under fluctuating constant amplitude stresses there is a certain stress range, which varies both with the environment and with the size of any initial flaws, below which an indefinitely large number of cycles can be sustained. In air, and in sea water with adequate protection against corrosion, it is assumed that this non-propagating stress range, So, is the stress corresponding to N = 107 cycles to failure as predicted by the design S-N curve (relevant values of So for the standard basic curves are shown in Table 14). In all cases So is the stress derived after allowing for any required modifications to the S-N curves (see 4.3). For unprotected joints in sea water it should be assumed that So = 0. When the applied fluctuating stress has varying amplitude, so that some of the stress ranges are greater than and some less than So, the larger stress will cause growth of the flaw, thereby reducing the value of the non-propagating stress range below So. Thus, as time goes on, an increasing number of stress ranges below So can themselves contribute to crack growth. The final result is an earlier fatigue failure than could be predicted by assuming that all stress ranges below So are ineffective. When considering variable amplitude loading, it should be assumed that the S-N curve has a change of inverse slope from m to (m + 2) at N = 107 cycles (see Figure 13). This is equivalent to assuming that the number of repetitions of each stress range Sr less than So is reduced in the proportion (Sr/So)2.
NOTE This correction does not apply in the case of unprotected joints in sea water.

co

m

? BSI 03-1999

33

BS 7608:1993

It may assist in calculations to note the following:

nS r ?m n- = -------------n ? Sr for ( S r $ S o, ) ----- = ----------- ? -------? N C 2 107 ? S o? for
(m + 2) ? ?( m + 2 )? n ? Sr n- = nS r ? S # S , ----? ------------------------------ = ----------- ? -------? r o ? ? 2 N C2 So 107 ? S o? ? ?

m

4.7 Joints subjected to a stress spectrum
4.7.1 General For a joint subjected to a number of repetitions ni of each of several stress ranges, Sri, the value of ni corresponding to each Sri should be determined from standard loading rules (if applicable), from stress spectra measured on a similar structural member, or by making reasonable assumptions as to the expected service history, as appropriate (see 1.7). The number of cycles to failure Ni at each stress range, Sri, should then be determined from the basic S-N curves, modified as necessary in accordance with 4.3, for the relevant joint class at the selected probability of failure. The design should then be modified so that the cumulative damage (Miner’s) summation is as follows:

4.5 Treatment of high stress cycles
For high stress cycles the design Sr-N curves for a) nodal joints (class T); b) non-nodal joints (classes B to G) where local bending or other structural stress concentrating features are involved and the relevant stress range includes the stress concentration; may be extrapolated back linearly (on the basis of log SR versus log N) up to a limit of a stress range equal to twice the material nominal tensile yield stress (2?Y). However, for joints in a region of simple membrane stress the design Sr-N curves should only be extrapolated back to a stress range given by twice the tensile stress limitations given in 4.1. For the class W curve, extrapolation may be made back as for non-nodal joints but up to a limit of stress range defined by half the values given above in this clause (i.e. with reference to shear instead of tensile stress).

ww

For a joint subjected to a number of repetitions of a single stress range, the range should be not greater than that defined by equation (1) (see 4.2.1) and Table 14 for the relevant joint class, number of cycles and required probability of failure. If the stress range is lower than the initial non-propagating stress range So (see 4.4) fatigue need not be considered.

w.

4.6 Joints subjected to a single stress range

bz

If all values of Sri are less than So fatigue need not be considered. However, if any values of Sri exceed So, all the stress ranges, including those below So, need to be included in the summation. This is because the higher stresses in the spectrum are capable of propagating cracks which may then be propagated further by the lower stresses. Some of the lower stresses in the spectrum can, however, be assumed to be less damaging than the higher stresses, as indicated in 4.4.

fx

w.

n1 n2 n3 ------+ ------- + ------- + . . . = N1 N2 N3

co
∑ ----N

m

n i is not greater than unity i

(5)

34

? BSI 03-1999

BS 7608:1993

4.7.2 Miner’s summation greater than unity If the condition defined by equation (5) (see 4.7.1) is not met n > 1.0 ) the following alternative actions ∑ ---N should be considered. a) Strengthen the detail to reduce the value of Sr. The strengthened detail will be satisfactory if the reduced values of stresses are less than the original values divided by the factor (i.e. if n? 1 ? m ? ---∑ ? N? If, however, they are greater than the original values divided by the factor

n ? 1 ? (m + 2) ? ---∑ ? N? the strengthened detail will not be satisfactory. Intermediate values should be reassessed by the procedure given in 4.7.1. b) Redesign the detail to a higher class. As a guide, for upgrading to any class having an S-N curve with slope m = 3, the minimum required value of So should be between

Figure 1 — Weld quality category determined by adjacent detail

ww

w.

bz

fx

w.
? BSI 03-1999

co

times the value of So of the original class of the detail.

m

n ? 0.33and ? n ? 0.2 ? ---? ∑ N? ? ∑ ---N?

35

BS 7608:1993

Figure 3 — Reference stress in weld throat

ww

w.

bz
36
? BSI 03-1999

fx

w.

co

Figure 2 — Reference stress in parent metal

m

BS 7608:1993

Figure 4 — Typical example of stress concentrations due to geometrical discontinuity

ww

w.

bz

fx

w.
? BSI 03-1999

co
37

m

BS 7608:1993

Figure 5 — Typical example of stress concentration caused by a geometrical hard spot

ww

w.

bz
38
? BSI 03-1999

fx

w.

co

m

BS 7608:1993

ww
Figure 6 — Stress concentration factors

w.

bz

fx

w.
? BSI 03-1999

co
39

m

BS 7608:1993

ww
Figure 7 — Example of hot spot stresses in a nodal joint

w.

bz
40
? BSI 03-1999

fx

w.

co

m

BS 7608:1993

Figure 8 — Summary of mean-line Sr-N curves

ww

w.

bz

fx

w.
? BSI 03-1999

co
41

m

Figure 9 — Summary of standard basic design Sr-N curves (mean minus two standard deviations)

ww

w. b

zf xw .

BSI

co m

42
? BSI 03-1999

BS 7608:1993

ww

w. b

zf xw .

BSI

co m

? BSI 03-1999

BS 7608:1993

Sr Figure 10 — -----------– N curves for bolts with cut, ground or rolled threads under axial loading (class X) UTS

43

BS 7608:1993

ww
Figure 11 — Toe grinding to improve fatigue strength 44
? BSI 03-1999

w.

bz

fx

w.

co

m

BS 7608:1993

Figure 12 — Grinding to weld at tubular nodal joint

ww
Figure 13 — Typical Sr-N relationship

w.

bz

fx

w.
? BSI 03-1999

co

m
45

BS 7608:1993

Annex A (normative) Fatigue design philosophy
A.1 S-N relationships The Sr-N relationships for the various joint classes have been based on statistical analyses of available experimental data obtained under tensile loading. The analyses involved linear regression analyses of log Sr and log N with the slopes of the curves predetermined. In addition some minor empirical adjustments were made to ensure compatibility of results between the various classes. The change of slope in the curves from m to (m + 2) (see Figure 13) is a mathematical device to avoid difficulties in cumulative damage calculations using Miner’s rule. The bent S-N curve should not be assumed to represent the results which would be obtained in tests under constant amplitude loading. As far as welded joints are concerned, it has been shown experimentally that, when high tensile residual stresses are present, fatigue strength is a function of stress range alone; mean stress and stress ratio have no significant effect. In general it is impossible to predict what residual stresses may be present in any particular structure and therefore the design rules are based upon the assumption that high tensile residual stresses will be present. For simplicity the same assumption has been made with regard to non-welded details. A.2 Fatigue life for various failure probabilities The standard basic Sr-N curves in Figure 9 are based on two standard deviations below the mean line assuming a log normal distribution, with a nominal probability of failure of 2.3 %. In certain cases a higher probability of failure could be acceptable, for example where fatigue cracking would not have serious consequences or where a crack could be easily located and repaired. On the other hand, in critical situations it may be desired to use a lower probability of failure. The nominal probabilities of failure associated with various numbers of standard deviations below the mean curve are given in Table 15. The Sr-N curves appropriate to other numbers of standard deviations below the mean curve can be derived from equation (1) (see 4.2).

A.3 Damage tolerant design In general it is recommended that structures and components should be designed to be “damage tolerant”. Because of the scatter in fatigue performance, and the possibility of use beyond the required minimum life, there is a risk that a structure or component will fail in service. Damage-tolerant design should ensure that when fatigue cracking occurs in service the remaining structure can sustain the maximum working load without failure until the damage is detected. It is recommended that the following design features should be used to help achieve damage tolerance: a) selection of materials and stress levels to provide low rates of crack propagation and long critical crack lengths; b) provision of multiple load paths; c) provision of crack-arresting details; d) provision of readily inspectable details. Damage tolerance depends on the level of inspection the user is prepared to apply to the structure and is not automatically ensured by replaceable components. Inspection of structures should be planned to ensure adequate detection and monitoring of damage and to allow repair or replacement of components. The following factors should be taken into account: 1) location and mode of failure; 2) remaining structural strength; 3) detectability and associated inspection technique. This should be based on the largest flaw not likely to be detected rather than the smallest it is possible to find; 4) inspection frequency; 5) expected propagation rate allowing for stress redistribution; 6) critical crack length before repair or replacement is required. A.4 Safe life design In situations where regular inspection is not possible, or is otherwise impractical, safety from possible catastrophic failure should be achieved by ensuring that the calculated life is many times longer than the life required in service.

ww

w.

bz

fx

w.

co

m

46

? BSI 03-1999

BS 7608:1993

∑ ---N
(see 4.7).

n

n If the value of ∑ ---is unsatisfactory modify the N peak stress range (and hence all the other stress ranges) or the joint detail so as to give a satisfactory value which is equal to or less than 1.0 (see 4.7.2).

Annex B (normative) Explanatory notes on detail classification
B.1 Introduction This annex gives background information on the detail classifications given in Table 1 to Table 12, including notes on the potential modes of failure, important factors influencing the class of each detail type and some guidance on selection for design.

ww

w.

A.5 Fatigue assessment procedure A structural member may contain a number of potential fatigue crack initiation sites. Although all need to be checked (see 2.1), the regions of the structure subjected to the highest stress fluctuations and/or containing the severest stress concentrations would normally be checked first. The basic recommended procedure can be summarized as follows. a) Select the required design life of the structure, which will be influenced by whether the design basis is to be “damage tolerant” or “safe life”. b) Make the best estimate of the loading expected in the life of the structure (see 1.7). c) Estimate the resulting stress history at the detail under consideration (see section 3). d) Reduce the stress history to an equivalent number of cycles ni of different stress ranges Sri using a cycle counting technique (see 3.9). e) Rank the cycles in descending order of stress range to form a stress spectrum. f) Classify the detail in accordance with section 2 and Table 1 to Table 12. g) Use this classification to define the basic design Sr-N curve (see 4.2). h) Modify the S-N curve, if necessary, to allow for variables such as material thickness, corrosion, weld improvement methods, (see 4.3). i) Select the peak stress range in the spectrum, and hence for all the other stress ranges, and calculate the value of

B.2 Non-welded details (see Table 1 and Table 2) B.2.1 Potential modes of failure In unwelded steel, fatigue cracks normally initiate either at surface irregularities, at corners of the cross sections, or at holes and re-entrant corners. In steel, which is connected with rivets or bolts, failure generally initiates at the edge of the hole and propagates across the net section. However, in double covered joints made with high strength friction grip bolts this mode of failure is eliminated by the pre-tensioning, providing joint slip is avoided, and failure initiates on the surface near the boundary of the compression ring due to fretting under repeated strain. B.2.2 General comments In welded construction, fatigue failure will rarely occur in a region of unwelded material since the fatigue strength of the welded joints will usually be much lower. Class A requires special manufacturing procedures which generally render it inappropriate for structural work. Hence assessment of fatigue strength for this class is not included in this code. Classes B and C should be applied with caution. B.2.3 Fit-up and pre-tensioning of bolted connections It is important to ensure that the specified fit-up of bolted connections is achieved in practice. Otherwise the stress ranges applied to the bolts may be much higher than those assumed in design and hence lead to premature failure. When a group of bolts is being tightened, it is important to check the torque on all the bolts after all have been tightened. This is because it is possible for some pre-tension to be lost as later ones are tightened, even if they were originally tightened with a torque spanner. B.3 Fasteners and shear connectors (see Table 3) In threaded fasteners fatigue cracks 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. In welded shear connectors fatigue cracking tends to occur in the weld throat initiating from the root. B.4 Continuous welded attachments essentially parallel to the direction of stress (see Table 4) B.4.1 Potential modes of failure

bz

fx

w.

co

m

? BSI 03-1999

47

BS 7608:1993

ww

w.

Away from weld ends, fatigue cracks normally initiate at stop-start positions or, if these are not present, at weld surface ripples. With the weld reinforcement dressed flush, failure tends to be associated with weld flaws. However, in the case of discontinuous welds (types 4.4 and 4.5) fatigue cracks will occur at the weld ends. B.4.2 General comments B.4.2.1 Edge distance (see Figure B.1) No edge distance criterion exists for continuous or regularly intermittent welds away from the ends of an attachment (see types 4.1 to 4.5). However it is important to limit the possibility of local stress concentrations occurring at unwelded corners as a result of, for example, undercut, weld spatter and excessive leg length at stop-start positions or accidental overweave in manual welding. Although this criterion can be specified only for the “width” direction of an element, it is equally important to avoid undercutting on the unwelded corners of, for example, cover plates or box members [see Figure B.1(b) and Figure B.1(c)]. If it does occur, it should subsequently be ground out to a smooth profile. B.4.2.2 Attachment of permanent backing strips If a permanent backing strip is used in making longitudinal butt welded joints it should be continuous or made continuous by welding. These welds, and those attaching the backing strip, should also conform to the relevant class requirements. The classification will reduce to class E or F (type 6.3 or 6.4) at any butt welds in the backing strip or class E at any permanent tack weld (see type 4.4). Transverse butt welds on backing strips may be downgraded by tack welds close to their ends (see type 6.4).

B.4.2.3 Tack welds Tack welds, unless carefully ground out or buried in a subsequent run, will provide potential crack locations similar to any other weld end. Their use in the fabrication process should be strictly controlled. B.5 Welded attachments on the surface or edge of a stressed member (see Table 5) B.5.1 Potential modes of failure (see Figure B.2). When the weld is parallel to the direction of the applied stress fatigue cracks 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 cracks may also initiate at the weld root. The cracks then propagate into the stressed member. B.5.2 General comments B.5.2.1 Stress concentrations Stress concentrations are increased, and hence the fatigue strength or joint classification is reduced, where the following apply: a) the weld ends or toes are on, or near, an unwelded corner of the element. This is the reason why an “edge distance” is specified for some of the joints; b) 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 will occur through welds adjacent to its ends. B.5.2.2 Weld forms Full or partial penetration butt welded joints of T form, including cruciform joints, (such as would connect attachments to the surface of a stressed element) should be completed by fillet welds of leg length at least equal to 25 % of the thickness of the attachment (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 thus, in considering the effects on the stressed element, it is immaterial whether the attachment is fillet or butt welded to the surface, since a similar toe profile results in both cases.

bz

fx

w.

co

m

48

? BSI 03-1999

BS 7608:1993

B.6 Transverse butt welds (see Table 6) B.6.1 Potential modes of failure (see Figure B.4) With the ends of butt welds machined flush with the plate edges, fatigue cracks 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, failure is more likely to occur in the weld material from embedded flaws or from minor weld flaws which become exposed on the surface, e.g. surface porosity in the dressing area (see type 6.3). In the case of butt welds made on a permanent backing strip, fatigue cracks initiate at the weld metal strip junction and then propagate into the weld metal. B.6.2 General comments B.6.2.1 Misalignment The classifications may be deemed to allow for the effects of any accidental axial or centreline misalignment up to the lesser of 0.15 times the thickness of the thinner part or 3 mm, provided that the root sides of joints with single-sided preparations i.e. single bevel -J, -U or-V forms are back-gouged to a total width at least equal to half the thickness of the thinner element. 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 thickness direction may be neglected. However, where such support is not provided, e.g. tension links, and where the amount of misalignment exceeds the limits stated above, the design stress should include an allowance for the bending effects of any intentional misalignment, i.e. the nominal distance between the centres of thickness of the two abutting components. For components tapered in thickness, the mid-plane of the untapered section should be used. The nominal stress should be multiplied by the following factor: t 13 e 1 + 6 ---- × ---------------t 1 t 13 + t 23 where t1 is the thickness of the thinner plate; t2 is the thickness of the thicker plate; e is the offset between the plate centre lines.

Thus, when t1 = t2, the stress concentration factor e becomes 1 + 3 t For other cases, including angular misalignment, reference should be made to PD 6493:1991. B.6.2.2 Element edges In all cases, failures tend to be associated with plate edges and care should be taken to avoid undercut at the weld toes on the corners of the cross section of the stressed element (or on the edge at the toes of any return welds). Should it occur, any undercut should be ground out to a smooth profile. B.6.2.3 Part width welds Butt type welds may also occur within the length of a member or individual plate as, for example, in the case of: a) a plug weld to fill a small hole; b) a weld closing a temporary access hole with an infill plate. Although such geometries have not been given specific categories in Table 3, types 6.3 and 6.4 may be deemed to cover plug and infill plate welds. B.6.2.4 Joints welded from one side only Unless made on a permanent backing (type 6.4) welds made entirely from one side are not classified since the root condition will be dependent upon the welding procedures adopted. Accordingly, their use is not recommended unless subject to special tests and strict procedural control. B.6.2.5 Penetration of butt welds It is recommended that butt welds transmitting stress between plates, sections or built-up members connected end-to-end should be full penetration welds. B.6.2.6 Dressing of butt welds (types 6.1 and 7.2) Fatigue tests on butt welds with the weld overfill dressed flush have shown that class C is a realistic design classification provided that the weld is free of flaws. However, in the great majority of situations, a fatigue strength higher than class D cannot be justified because of the possible presence of flaws which are too small for reliable detection using current non-destructive inspection methods, but which are of sufficient size to reduce the fatigue strength of the joint. Previously buried flaws revealed by flush-grinding of a weld can result in a severe reduction in the fatigue strength of the weld and they should be assessed.

ww

w.

bz

fx

w.

co

m

? BSI 03-1999

49

BS 7608:1993

ww

B.6.2.7 Cruciform and T-joints between plates in the same plane (type 6.5) Often the load is transmitted from a member to a transverse member primarily via flange plates in the same plane (see Figure B.5). This can occur in the case of a junction between cross girders and main girders, diagonals and truss chords, or in Vierendeel frames. If a full penetration butt weld is used and the joint geometry is in accordance with the requirements of type 6.5, and if in addition the joint is either of the cruciform type (see Figure B.6) or if the transverse member is relatively stiff, i.e. its width is at least three times the width of the stressed member, then the classification can be considered to be effectively class F2 with a stress concentration factor of unity (see type 6.5). Otherwise the classification should be assumed to be class E (i.e. type 6.3) with the appropriate stress concentration factor. In the case of trusses, secondary stresses due to joint fixity should be taken into account. The fatigue strength of both flange plates may be improved by the insertion of a smoothly radiused gusset plate in the transverse member so that all butt welds are well away from re-entrant corners (see Figure B.7). B.7 Transverse butt welds in sections (see Table 7) Butt welds between rolled sections or between built-up sections are frequently made using cope holes to provide access for making the weld in the flange. The end of the web butt weld at the cope hole can be considered to be equivalent to class D provided that the end of the butt weld, and the weld reinforcement within a distance equal to the radius of the cope hole, are ground flush (see Figure B.8). The relevant stress should include a stress concentration factor to allow for the presence of the cope hole. Mitred cope holes of triangular shape are not recommended.

B.8 Load-carrying fillet and T-butt joints (see Table 8) B.8.1 Potential modes of failure (see Figure B.3) Failure in cruciform or T-joints with full penetration welds will normally initiate at the weld toe, but in joints made with load-carrying fillet or partial penetration butt welds cracking may 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; cracks normally initiate at the weld end and propagate into the plate perpendicular to the direction of applied stress. The stress concentration is increased, and the fatigue strength is therefore reduced, if the weld end is located on or near to the edge of a stressed member rather than away from the edge. B.8.2 Cruciform joints (types 8.1 and 8.2) Where the third member is a plate it may be assumed that plane sections remain plane in the main members and that axial and bending stress distribution in the Sr 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 will occur. In this case the stress parameter should be the peak stress concentration at the joint. In the absence of published data on a particular joint configuration, the stress concentration factor may have 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 in line with the main member components (see Figure B.9). In this type of connection it is important that the joint regions of the third member are checked before welding for lamellar rolling flaws and after welding for lamellar tears. B.8.3 T-joints (types 8.3 and 8.4) T-joints are distinguished from types 8.1 and 8.2 by the absence of a similar member in line on the far side of the joint. In this case an axial load in member X will induce bending moment, and hence curvature, in member Y. Unless the latter is very stiff in bending an uneven stress distribution will result. 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 will increase the peak stress further (as in the case of types 8.1 and 8.2).

w.

bz

fx

w.

co

m

50

? BSI 03-1999

BS 7608:1993

As far as fatigue failure of the transverse member Y is concerned, member X is treated as an attachment (types 5.2 to 5.4) 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. B.8.4 Load-carrying fillet welds failing in the weld (type 8.5) Class W 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. In certain cases difficulty of access may only allow welding 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, may cause failure of the corner weld in bending (see Figure B.11). Where axial stress is also present, the stress range at the face of the weld may be different from that at the root. Failure from ripples or stop-start positions on the face may give a higher strength than class W, but expert advice should be sought if a higher strength is required. In most cases failure from stress fluctuation in the root will be critical and this should always be classified as class W. B.9 Slotted connections and penetrations through stressed members (see Table 9) Slotted connections exist where a narrow member is slotted through a single main member away from an end connection (see also Figure B.12). 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 5.2 to 5.5. This detail should generally 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 risk from lamellar tears. However, with respect to stress fluctuation in member B, the detail shown in Figure B.12 is type 5.5 (class G) at point Y. If member B is critical and member A is not, circular cut-outs at the corners of member B will improve the class to class F (type 5.2). Particular attention is drawn to the making of this type of joint with fillet welds instead of full penetration welds. In that situation the joint becomes type 8.2 and type 8.5.

bz

B.10 Branch connections (see Table 12) B.10.1 Potential modes of failure There are three main sites for fatigue cracking in branch connections, the weld toes in the shell and the branch and the crotch corner. In every case account should be taken of the stress concentration in the region of potential fatigue cracking due to the gross structural discontinuity introduced by the nozzle. B.10.2 Stress concentrations adjacent to branch connections Three possible stress concentrations due to structural discontinuities in nozzles should be considered when calculating Sr. a) Crotch corner. The class D fatigue design curve is used, based on nominal hoop stress range multiplied by Kf at the crotch corner where Kf is as defined in 3.6.1. b) Weld toe in shell. The appropriate fatigue design curve is normally used with nominal hoop stress range multiplied by Kf at the weld toe, where Kf is as defined in 3.6.1. The possibility of stresses arising in the shell as a result of mechanical loading on the nozzle as well as pressure loading should be taken into account. c) Weld toe in branch. This region should be treated as in item b). Again, the possibility of mechanical as well as pressure loading should be taken into account.

w.

ww

fx

w.

Annex C (normative) Guidance on the calculation of stress concentration factors
C.1 General The stress to be used with the Sr-N curves in Figure 8 and Figure 9 is, except in the case of class T, the nominal stress range at the potential cracking site shown in the sketches in Table 1 to Table 12. It is the stress which would be calculated by conventional engineering methods, which does not take 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 it is necessary to calculate the stress 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. In the case of tubular nodal joints, which have to be designed using the class T Sr-N curve, the relevant stress is the so-called “hot spot stress” range. This is considered in more detail in C.2. 51

? BSI 03-1999

co

m

BS 7608:1993

Figure B.1 — Edge distance

ww

Figure B.2 — Failure modes at weld ends

w.

bz
52
? BSI 03-1999

fx

w.

co

m

BS 7608:1993

Figure B.3 — Failure modes in cruciform and T-joints

Figure B.4 — Failure modes in transverse butt welds

ww
Figure B.5 — T-junction of two flange plates

w.

bz

fx

w.
? BSI 03-1999

co

m
53

BS 7608:1993

Figure B.6 — Cruciform junction between flange plates

ww

Figure B.7 — Alternative method of joining two flange plates

w.

bz
54
? BSI 03-1999

fx

w.

co

m

BS 7608:1993

Figure B.8 — Local grinding adjacent to cope hole in type 7.1 joint

ww
Figure B.9 — Use of continuity plating to reduce stress concentrations in type 8.1 and 8.2 joints

w.

bz

fx

w.

co
? BSI 03-1999

m
55

BS 7608:1993

Figure B.10 — Example of type 8.3 or 8.4 joint

Figure B.11 — Single fillet corner weld in bending

ww

w.

bz
56
? BSI 03-1999

fx

w.

co

m

BS 7608:1993

Figure B.12 — Example of a third member slotted through a main member

ww

w.

bz

fx

w.
? BSI 03-1999

co
57

m

BS 7608:1993

C.2 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.10. It is this magnified stress which has to be used as the input to the Sr-N curve for that particular detail. The effect on the fatigue life will be quite marked because life is proportional to the third or the fifth power of the stress depending on the position on the Sr-N curve. More complex shapes can be dealt with in the same way using published solutions [l, 2 and 3] for most of the regular shapes of cut-out; two common details are shown in Figure 6. For shapes not included in the standard texts stresses can be measured on a photoelastic model [4] or by strain measurements on an acrylic [5] or metal model, or the full scale item. In the case of a metal model, or the full scale metal structure, it is important to apply load cycles of sufficient magnitude and quantity to ensure that a shaken down state is reached. The design of the models and the positions at which strains are measured have to recognize the basis of the Sr-N curves so that the relevant stress is used. In areas of low stress gradient this is not too difficult but there is a need for careful assessment in regions of steep stress gradients. A commonly used and powerful tool for stress prediction is the finite element analysis 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 will give results which are insufficiently accurate. Too fine a mesh will pick up the influence of the weld shape, if that is modelled, or the false influence of the idealized joint shape. Although they may be expensive in computer time, convergence checks will 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 work rather than the immediate needs of the designer. These include techniques such as thermal imaging and X-rays.

C.3 Three dimensional stress systems It will be apparent that for most of the welded joint details shown in Table 1 to Table 12 the stress required to calculate the fatigue life is at the surface of the members. The stress analysis methods described in C.2 will all 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.10) or by brace loads on the walls of hollow section chords (Figure 7), the standard analytically based solutions are not sufficient. It is then necessary to use three dimensional methods such as photoelasticity and finite elements for a full understanding of the stress pattern. Strain gauges will give the surface stresses but if there is a need to know the stresses through the thickness of the material, for crack growth calculations for example, they will not be of much help. Finite element methods can be used to calculate the stresses in all types of joint. The selection of element types and sizes has to reflect the geometry and the stress gradient so that a sufficiently reliable prediction of the stress at the weld detail is obtained. The modelling of the joint has to reflect the overall geometry and the stiffness without introducing the details of the weld profile. C.4 Tubular joints It was mentioned in C.1 that the design of tubular joints, using the class T Sr-N curve, is based upon the range of “hot spot stress”. The stresses in tubular nodal joints arise from the following three main causes: a) the basic structural response of the joint to the applied loads (nominal stresses); b) the need to maintain compatibility between the tubes (geometric stresses); c) the highly localized deformations in part of the tube wall near to the brace-chord intersection (local stresses). Nominal stresses arise due to the tubes behaving as beam-columns, and may be calculated by frame analysis of the structure. Geometric 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 weld. They can also cause a maldistribution of the nominal membrane stresses around the brace circumference.

ww

w.

bz

fx

w.

co

m

58

? BSI 03-1999

BS 7608:1993

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 thickness, 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. Under cyclic loading fatigue cracks will appear first at the point or region of highest stress range; this is called the “hot spot” and the stress there is the “hot spot stress”. In formal terms it is the greatest value of the range of principal stress extrapolated to the weld toe through the local stress region around the brace-chord periphery (see Figure 7). It incorporates the effects of overall tube geometry, i.e. the relative sizes of the brace and chord, but omits the concentrating influence of the weld geometry which changes around the periphery of any brace/chord intersection. Welded plate specimens loaded axially or in bending exhibit a linear stress distribution on the outer surface which is easily calculable. This is disturbed by the presence of the weld which induces a region of rapidly rising local stress over a small distance from the weld toe. The continuation of the linear stress distribution to the weld toe by extrapolation will then give a stress at the toe which omits the concentrating effect of the weld profile. Again this is easily calculated. In the case of class T and class X tubular joints, however, such a linear (or near linear) stress distribution is found only within a small region some distance from the weld toe2), dependent on joint size. However, not all tubular joints will exhibit a region of stress linearity near the weld toe. Class K and class Y joints in particular can, in certain locations, exhibit geometric stress distributions which are non-linear, and to maintain the general definition of “hot spot stress”, i.e. that found at the weld toe but omitting the concentration caused by the weld geometry, non- linear extrapolation through the region of local stress is necessary. It can be seen, therefore, that a more general definition of “hot spot stress” can be given as the peak value of geometric stress found at the weld toe.

A further difficulty that can arise in certain locations on certain tubular joints is that the geometrical stress perpendicular to the weld toe may not be the maximum principal stress (rosette strain gauges or finite element analysis have to be used to discern this phenomenon). In these cases, the value of stress quoted is taken as the (conservative) value of extrapolated maximum principal stress. This practice follows that of other fatigue design codes. It has been demonstrated that the region of local strain increases with joint size. This is quite reasonable as weld dimensions are dependent on brace wall thickness. The extent of the linear stress region approaching the “hot spot” weld toe is given in Figure 7 (see also reference [6]) which defines the maximum extent of the local region, for class T and class X joints, as the greater of 4 mm3) or 0.2?rt where r and t are the brace outer radius and wall thickness respectively. This expression has been obtained empirically, though the dependence on the parameter ?rt was originally drawn from a similar dependence on the wavelength of bending stress in tubes (see reference [7]). Where the “hot spot” lies near, but not on, an axis of joint symmetry it is suggested that the extrapolation procedure follows that given for the adjacent axis. Thus it can be seen that, as local region size scales with tubular connection size, “hot spot stress”, as defined by linear extrapolation to the weld toe, is dependent only on the overall geometry of the tubular connection. This should also be the case when “hot spot stress” is predicted by analytical means. The stress within the local region increases rapidly towards the weld toe according to the overall weld geometry; the end bead may locally further intensify stresses over a few millimetres distance. Such local stresses control fatigue crack initiation and early crack growth, and decay rapidly as crack growth continues. Their measurement is inherently of high uncertainty and low reproducibility due to weld profile variation and potential inaccuracy in strain gauge positioning. Thus a characteristic stress which controls the complete fatigue life of a tubular joint has to be evaluated some distance from the weld toe. The “hot spot stress” obtained by linear extrapolation to the weld toe is used as an expression of such a stress.

2) 3)

Except for class X joints where ? = 1.

The alternative value of 4 mm is only used in practice for very small tubular joints where disproportionately large weldments are produced. The local region is in these cases determined by the geometry of the weld toe bead (typically a few millimetres in size) as the stresses rise into the weld which being disproportionately large forms a ring of very stiff and inflexible material joining the brace to the chord.

ww

w.

bz

fx

w.

co

m

? BSI 03-1999

59

BS 7608:1993

mean radius of brace cylinder β = ---------------------------------------------------------------------------------mean radius of chord cylinder mean radius of chord cylinder γ = --------------------------------------------------------------------------------------wall thickness of chord cylinder wall thickness of brace cylinder τ = --------------------------------------------------------------------------------------wall thickness of chord cylinder The form of these equations assumes that some dependence of the parameters can be deduced logically (by, for example, using simple beam theory to predict a component of stress at the chord crown); otherwise a polynomial or power law dependence is assumed. Each set of parametric equations is limited in application in the following three ways: 1) Restrictions in types of joint geometry. 2) Restrictions on parametric validity range. 3) Restrictions on loading cases.

ww

w.

bz

fx

w.

co

length of chord cylinder α = ---------------------------------------------------------------------------------mean radius of chord cylinder

m

Apart from the use of analytical methods (e.g. finite element analysis) or experimental stress analysis methods on model joints, the main method of defining the “hot spot stress” is by means of empirical parametric equations (see references [9, 10, 11, 12, 13]). In general they are only applicable to some simple types of joint. They should only be used within established ranges of validity and when the equations have been shown to be reliable within that range. The parametric equations were developed by fitting analytical or model stress concentration factor data to equations which are functions of non-dimensional tubular joint geometric parameters, as follows:

The parametric equations deal exclusively with co-planar tubular connections; thus, to use these parametric equations to assess peak stresses at connections of greater complexity implies that stress distributions around individual braces do not interact. However, this is not the case when bracing members are nearer to each other than two or three chord diameters: brace interaction can result in underestimation of “hot spot stresses” when braces are examined separately. For example, consider the case of a connection where two perpendicular T-joints A and B connect the chord at the same point on the chord axis but in different planes i.e. a V-joint. The load on brace A causes a “hot spot stress” at the chord intersection of brace A which can be reduced by the stiffening effect of brace B. However, the load on brace A also causes some stress concentration at the chord intersection of brace B and vice versa. Thus, the total “hot spot stress” at the chord intersection with brace A is due to brace A behaving as a loaded T-joint, modified by the presence of brace B, plus the load on brace B producing some stress concentration at brace A. Hence, the presence of other nearby bracing will alter the values of the stress concentration factor from that obtained from the equations treating the connection as two separate joints. The magnitude of the discrepancy depends on the relative direction and value of the loads applied to the bracing members as well as to the distance between the planar braces. Stress values obtained in these circumstances can be as easily underpredicted as overpredicted. None of the equations in question covers joints with internal or external stiffeners. Each set of equations assumes freely supported chord end conditions with a minimum chord length to radius ratio so as to avoid interaction of chord end loading stress concentration and tubular connection stress concentration, and incorporate chord beam bending stresses. In practice there is little significance in this latter effect, except where axial brace loading causes peak stresses at the chord crown. This only occurs either for very long chords (? > 32 according to reference [10]) or for ? > 0.9.

60

? BSI 03-1999

BS 7608:1993

There are also restrictions on the range of the other three variables: ?, ?, ?. This is due to limitations on the range of original data obtained, or the methods used to obtain these data, but also that marked changes in the nature of the stress distribution occur at extreme values of these parameters. To exemplify this, high ? (? > 0.9) connections show a radically different peak stress distribution from similar lower ? connections, with the hot spot moving from the normal saddle position to the crown, and for low ? connections thin shell idealization used for finite element calculation is not valid. Non-compliance with parametric validity range can result in a considerable increase in the variation of performance, both conservative and non-conservative, of these equations. In addition the stress concentration factor (SCF) values obtained from the parametric formulae are in terms of maximum principal stress. Research into the stress distribution in nodal tubular joints is a continuing activity. A source of the latest assessed data is the Department of Energy Publication Offshore Installations: Guidance on design, construction and certification and its background document: Background to new fatigue design guidance for steel welded joints in offshore structures. C.5 Rectangular hollow sections There is relatively little published research or service experience on the subject of the fatigue performance of welded joints in structures using rectangular hollow sections but see reference [13]. C.6 Distortion and misalignment A source of stress concentration in all types of structures can be linear or angular misalignment at joints (see Figure C.1 and B.6.2.1). This can arise from poor attention to assembly or erection procedures or from welding distortion. Stresses can be set up by the local bending effects induced by this misalignment and these have to be taken into account. Particularly sensitive are circular section pipes and vessels where offset and angular distortion in the longitudinal seams and ovality can individually or in combination magnify the nominal stress by an amount which can seriously degrade the fatigue performance under pulsating pressure.

bz

NOTE detail.

ww

References [14, 15 and. 16] deal with this matter in some

w.

fx

w.

Annex D (normative) Guidance on the use of fracture mechanics
D.1 Introduction The objective of this annex is to provide guidance on the use of fracture mechanics methods for the fatigue assessment of structures or components subjected to high cycle fatigue conditions, in situations where the normal fatigue strength assessment methods outlined in this code may be unreliable or inappropriate. Nevertheless, it is strongly recommended that fracture mechanics should be used with considerable care. In general it is not a suitable method by which to try to define precise fatigue strengths or lives, since the results will be dependent, to a very marked extent, upon the assumptions which are made regarding, for example, the values of the constants in the crack growth equation, the size of the initial flaw(s) and the shape of the resulting fatigue crack, e.g. for a crack at a weld toe, whether it is semi-elliptical or straight-fronted. It would be unusual for all this information to be available at the design stage. Thus, if the objective is to define a particular fatigue strength or life, care should be taken to make the most pessimistic assumptions.

co
Figure C.1 — Types of misalignment and distortion 61

? BSI 03-1999

m

BS 7608:1993

H1, H2 Kmax l

NOTE This problem is outside the scope of this code. Reference should be made to PD 6493.

b) When the effects of relatively minor variations in the geometrical or stress parameters for a given detail are being studied. c) When the joint detail under consideration is unusual and is not adequately represented by one of the standard joint classifications, or when a joint is subjected to the influence of another stress concentration. d) When defining the frequency of in-service inspections. e) When assessing the remaining fatigue life of a structure in which fatigue cracks already exist. In the case of item e) the structure will contain cracks whose sizes will have to be determined by measurement and the sizes assumed will have to allow for possible errors in such measurements. In other situations it has to be assumed that small, but unmeasurable, flaws exist at points of stress concentration (e.g. at weld toes) and that it is from them that fatigue cracking may originate. The size of such flaws will therefore have to be assumed taking into account the warnings stated above. The procedure recommended in D.3 to D.8 is based upon the principles of linear elastic fracture mechanics. D.2 Symbols and units For the purposes of this annex only the following symbols and units are used.

w.
a

bz
m Mk M1, M2, M3 M4, M5 N Ni R %K %Kth %?a, %?b ?o
a Length

ww

a af 62

Measure of current crack length (in mm) Final value of crack lengtha (in mm)

fx

w.

Fracture mechanics can, however, be a very useful method for carrying out parametric studies, where the objective is to define the relative influence of a particular set of variables. In that situation all the variables, except the one under consideration, can be held constant and thereby enable its influence to be evaluated. These recommendations are NOT intended to replace, in any way, the normal fatigue assessment procedures outlined in this code when such procedures are applicable. For example, they are not intended to be used as a method to circumvent the normal requirements for good workmanship. Some typical situations in which the normal procedures may be inappropriate and in which the use of fracture mechanics may be helpful are as follows. a) When assessing the fitness for purpose of a structure known to contain flaws whose size, shape and distribution are outside normally acceptable limits (see 2.4.3) but which would be difficult to repair.

ai A B c f? fw Fa, Fb

Initial value of crack lengtha (in mm) Constant in the crack propagation equation Plate thickness (in mm) Half the surface length of a semi-elliptical surface crack of depth a (see Figure D.1) (in mm) Function dependent on the position around the crack circumference A finite width correction factor Functions of crack size and shape and the proximity of the crack tip to free surfaces for axial and bending stresses respectively A curve fitting function Functions of crack shape for semi-elliptical surface cracks Weld leg length (see Figure D.2) (in mm) = Fb/Fa

h H

Values of H at the surface and at the deepest point on the crack front, respectively [see Figure D.1(a)] Maximum value of stress intensity factor in the cycle (N·mm–3/2) Overall length from weld toe to weld toe of an attachment, measured in the direction of the applied stress (in mm) Constant in the crack propagation equation Magnification factor on stress intensity to allow for the presence of a stress concentration, such as weld toe Functions of crack shape for semi-elliptical surface cracks Function of crack shape for elliptical buried cracks Number of cycles (in cycles) Number of cycles to crack initiation (in cycles) Stress ratio ( = Smin/Smax) Range of stress intensity factor at the tip of the crack (in N·mm–3/2) Threshold value of %K for crack propagation (in N·mm–3/2) Range of axial and outer fibre bending stresses respectively at the location of the crack, based on gross section properties (in N·mm–3/2) Complete elliptic integral of the second kind

being measured in the direction of propagation.

co

g G1, G2

m

? BSI 03-1999

BS 7608:1993

ww
Figure D.1 — Flaw dimensions

w.

bz

fx

w.
? BSI 03-1999

co
63

m

BS 7608:1993

Figure D.2 — Transverse load-carrying cruciform joint D.3 General background As outlined in 2.1, fatigue cracks 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. Cracks originating from the weld toe normally initiate at small flaws while cracks originating from the root often start from areas of deliberate lack of penetration; in both cases the initiating feature can therefore be regarded as a planar discontinuity. In essence, therefore, most fatigue cracks in welded joints can 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 predict the life by integrating the relevant crack growth law. In doing so it is assumed that the real flaws can be idealized as sharp-tipped cracks, which propagate at a rate, da/dN which is a function of the range of the stress intensity factor, %K. The overall relationship between da/dN and %K is normally observed to be a sigmoidal curve in a log da/dN versus log %K plot. There is a central linear portion. At low values of %K the rate of growth falls off rapidly to a threshold stress intensity factor, %Kth, below which no significant crack growth will occur. At high values of %K, when the maximum stress intensity factor in the cycle approaches the critical stress intensity factor for failure under static load, the rate of crack growth accelerates rapidly. However, for practical purposes, it is usually 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 %K up to failure.

m da -------- = A ( ? K ) dN

w.
a ai

The relevant equation for the rate of crack propogation da/dN (in mm/cycle) is given by: (D1)

w.

ww

bz

fx

where A and m are constants which depend on the material and the applied conditions, including environment and cyclic frequency; %K is the range of stress intensity factor corresponding to the applied stress cycle and instantaneous fatigue crack dimensions. Integration of equation (D.1) gives the number of cycles, N, required to propagate a crack from an initial size, ai, to a final size, af, as: 1 f 1 N = N i + --- ------------------da A ∫ ( ? K )m

where Ni is the life to crack initiation which is usually assumed to be zero. D.4 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 waveform. Whenever possible, data relevant to the particular material, product form and service conditions should be used and where any doubt exists concerning the influence of the environment such data should be obtained. 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 da/dN.

co

m
64
? BSI 03-1999

BS 7608:1993

If this relationship is used, it is recommended that the calculated value of A should be multiplied by 2.0 to allow for scatter. Alternatively, it has been found that the values: m=3 A = 3 × 10–13 correspond to the upper bound to many published data for ferritic steels [plain plate, weld metals and heat affected zones (HAZ)] in which fatigue crack growth was by the striation mechanism. Higher growth rates have been observed in some weld metals and HAZ, particularly at high %K as %Kmax approaches the critical value for fracture, when the normal crack growth mechanism is accompanied by cleavage or microvoid coalescence. If such behaviour is thought to be relevant, the recommended value of A = 6 × 10–13. In the absence of specific corrosion fatigue data, it is recommended that, for structural ferritic steels operating in a marine environment at temperatures up to 20 °C, the crack growth constants should be assumed to be: m=3 A = 2.3 × 10–12. D.5 Initial flaw size ai It will usually be found that the calculated life is very sensitive to the assumed value of ai. It is therefore important that ai should not be underestimated. For nominally flaw-free welded joints failing from the weld toe it is recommended that ai should be assumed to lie within the range 0.1 mm to 0.25 mm unless a larger size is known to be relevant. The correct value to be used will be able to be deduced from the calibration calculations (see D.8).

ww

w.

bz

fx

w.
where Fa and Fb

co

1.315 × 10 A = --------------------------------m 895

–4

ai ----- = 0.1. 2c

D.6 Limit to fatigue crack propagation af In the fatigue assessment an upper limit should be set to the size af to which a crack may be allowed to grow without failure occurring during operation by any of the following modes, as appropriate: a) unstable fracture; b) yielding of the remaining section; c) leakage (in containment vessels); d) stress corrosion; e) instability (buckling); f) creep. D.7 Range of stress intensity factor, %K D.7.1 General Application of the crack propagation equation (see D.3) requires a knowledge of the range of stress intensity factor, %K, at the crack tip. Hence, for the actual or assumed dimensions and position of the idealized flaw, the value of %K corresponding to the range of stress (see 3.2 and 3.3) should be estimated either from relevant published solutions or from specific stress analyses of the structure. D.7.2 Semi-elliptical surface cracks (see reference [17]) In general the value of %K for a partial thickness flaw is of the form: πa ? K = ( M ka F a ?σ a + M kb F b ?σ b ) ---------Φo %?a and %?b are the ranges of the nominal axial and bending stresses at the crack tip; are functions of crack size and shape and the proximity of the crack tip to free surfaces (suffix a refers to axial loading and suffix b to bending);

m

In the absence of sufficient data to define those values, it is recommended that the value of A obtained from a mean regression fit to the crack propagation data should be multiplied by a factor of a least 4 to allow for the statistical nature of fatigue behaviour. When such data are not available, suitable values of A and m should be determined from other relevant published data. For structural steels tested in air at R = 0, m has usually been found to lie within the range 2.4 to 3.6. The mean value of A (in N·mm) has been found to be related to m by the equation:

In the absence of definite information about the shape of the initial flaws, it is also recommended that, for joints with welds transverse to the direction of stress, it should be assumed that the flaw at the weld toe is long and continuous, i.e. ai ----- = 0. 2c At the ends of longitudinal welds, however, it would be realistic to assume that the initial flaw was a semi-elliptical in shape with

? BSI 03-1999

65

BS 7608:1993

M k and Mk
a

b

are magnification factors to allow for the presence of a stress concentration, such as a weld toe; is the complete elliptic integral of the second kind.

?o

The value of ?o may be obtained from standard tables but an approximate solution, which is more convenient for calculation purposes, is as follows: ? Φ ο = ? 1.0 + 1.464 ? ? = ? 1.0 + 1.464 ? ?a --? ? c? c ? --? ? a?
1.65 ?0.5

a a For ----- > 1.0 the solution for ----- = 1.0 should be 2c 2c used. The values of g and f? depend upon the position on the crack boundary and on the crack shape. Their values are as given in Table D.2 . As a a before, for ----- > 1.0 use the solution for ----- = 1.0. 2c 2c The value of fw is given by
0.5

1.65 ?0.5

a 0 # ----- # 1.0 2c 2c ----- < 0.5 W a a- # 0.1 and --- < 1.25 ? a -- + 0.6? for 0 # ----?c ? B 2c a a --- < 1.0 for 0.1 ≤ ----- ≤ 1.0 B 2c a) Axial loading Fa

w.

? a? 2 a? 3? = ? M 1 + M 2 ? --- + M 3 ? --gf Φ f W ? B? ? B? ? ? ?

The values of M1, M2 and M3 depend upon the a crack shape, ----- , and are as given in Table D.1. 2c

a- # 0.5 ----2c M1

ww

Table D.1 — Values of M1, M2 and M3 a- #1.0 0.5 < ----2c

a? 1.13 – 0.09 ? -? c? 0.89 ----------------------- – 0.54 a? ? 0.2 + -? c? 24 ? 1 a? ? - + 14 ? 1 – ? -- ? 0.5 – -------------------------? c? a? ? ? 0.65 + ? -? c? a? 4 0.2 ? -? c? c? 4 – 0.11 ? -? a?

M2

M3

bz

fx

The value of H2 is given by the equation: a? a ?2 H 2 = 1 + G 1 ? --- + G 2 ? --? B? ? B?

The values of H1, G1 and G2 depend upon the crack shape and are as follows: a (As before, if ----- > 1.0 the value 2c a for ----- = 1.0 should be used.) 2c

w.

Subject to the following conditions, the values of F at the deepest point on the crack front [point Q, Figure D.1(a)] and at the ends of the crack (point S) are as listed in a) and b) as follows. Conditions:

b) Bending Fb = HFa

where Fa is as defined in a) for axial loading; H = H2 at the deepest point on the crack front; H = H1 at the ends of the crack.

co

? ?

for 1 < a -- #2 c

c If W $ 20c, it can be assumed that ---- = 0 , so W that fw = 1.0.

m

? ?

a ? for ? -- #1 ?c ?

? π c? ? a ? 0.5? f W = ? Sec ? ----- --? W ? ? B? ? ? ?

66

? BSI 03-1999

BS 7608:1993

Table D.2 — Values of g and f?
At deepest point [point Q, Figure D.1(a)] At crack ends [point S, Figure D.1(a)]

a ----- # 0.5 2c g ff 1.0 1.0
a? 2 1.1 – 0.35 ? --? B?
0.5 ?a --? ? c?

a 0.5 < ----- #1.0 2c c? ? a? 2 1.1 – 0.35 ? -- --? a ? ? B? 1.0

a for ----- # 0.5: 2c

. K for crack in plate with stress concentration M k = ------------------------------------------------------------------------------------------------------------------------------K for same crack in plate without stress concentration

Thus, Mk 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 30 % of plate thickness. At crack depths greater than that corresponding to Mk = 1.0 it should be assumed that Mk = 1.0.

ww

When a flaw or crack is situated in a region of stress concentration, such as a weld toe, and the value of K does not already incorporate the influence of the stress concentration, it is necessary to include a correction factor, Mk, which is a function of crack size, geometry and loading, in general as follows:

w.

bz

fx

a For 0.5 < ----- ≤ 1.0 : 2c

w.

For transverse butt welds, K (cruciform) butt welds and members with transverse non-load-carrying fillet welds, Mk has been found to be a function of crack depth a, plate thickness B and the overall length l of the “attachment” measured from weld toe to weld toe [see Figure D.1(b)]. The relationships are as follows (see reference [18]): 1) Tension

co

m

? BSI 03-1999

67

BS 7608:1993

2) Bending

2a b) 0 # ------ # 0.7. W The range of stress intensity factor (see reference [19]) is given by the equation: ? π a? ? ? K = M k ?σ a ? π a Sec ? -----? W? ? ? ? where
0.5

The nature of the finite element model used to calculate Mk is such that the solutions produced are not applicable for a = 0. Thus, for a semi-elliptical crack they do not apply to the ends of the crack, at the surface. Experience indicates that it is reasonable to assume Mk is equal to the elastic stress concentration factor at the weld toe, or to adopt the value of Mk corresponding to a very small crack, such as a = 0.15 mm. The latter approach is compatible with the fact that in steels there are inherent crack-like flaws of this order of depth at all weld toes. D.7.3 Weld root crack in cruciform joint (see Figure D.2) The solution for a weld root crack in a cruciform joint was derived for tension loading but it is conservative to apply it for bending. It applies for the following: h a) 0.2 # --- #1.2; B

ww

w.

However, experience indicates that they can also be a applied to semi-elliptical cracks ? - # 0.5? ? 0 < ----?. 2c

bz

The values of Mk given in 1) and 2) were calculated by 2D finite element analysis for profiles representing sections of the welded joint geometry. Thus, they are directly applicable to the case of a a straight-fronted weld toe crack i.e. ----- = 0 2c

fx

and dimensions are as indicated in Figure D.2. In this case Mk is always less than unity and its value may increase or decrease with increase in · h. crack size, depending on the value of --B

w.
? BSI 03-1999

68

co

m

BS 7608:1993

E.1 Introduction Potentially there may be a need to carry out fatigue testing for two different purposes, as follows: a) to establish the relevant S-N curve, or joint classification, for the design of some detail which is not adequately covered by the classifications given in 2.2; b) to establish whether some prototype structure is capable of carrying the fatigue loading expected during the service life of the structure. In a) the objective is to obtain a design S-N curve, under constant amplitude loading, in a similar manner to that used for the standard classes (see 2.3.1). In fatigue acceptance testing, the objective would normally be to apply to the component or structure loading simulating that to be expected in service. Any fatigue tests which may be carried out should be performed using equipment and/or testing machines with certified calibration. Care has to be taken to ensure that specimens are not overloaded prior to fatigue testing. E.2 Fatigue tests to establish joint classification Fatigue testing for joint classification purposes involves carrying out constant amplitude tests under tensile loading (usually with lower limit stress = 0) with the stresses selected so as to give endurances to failure reasonably evenly distributed over the linear part of the relationship between log (stress range) and log (endurance), (i.e. typically over the range 105 to 2 × 106 cycles). It is usually appropriate to test not less than eight nominally identical specimens representative of the detail under consideration.

ww

w.

bz

fx

w.

co

Annex E (normative) Fatigue testing and the use of test data to define design stresses

m

D.8 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: a) the fatigue strength of a joint class with a detail similar to that under consideration; or b) 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.

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. Having derived the relevant experimental fatigue test results it is necessary to calculate the corresponding mean and mean – 2 standard deviation S-N curves, and to compare them with the corresponding curves for the standard joint classes (see 4.2 and Figure 8 and Figure 9). The detail can be considered to fall within the highest joint class for which both the mean and mean – 2 standard deviations experimentally determined S-N curves are above the corresponding design S-N curves. E.3 Fatigue acceptance test The objective of the fatigue acceptance test is to establish whether some prototype component or structure is capable of carrying the fatigue loading expected during its service life. Where the service loads vary in a random manner between limits, they should be represented by an equivalent series of loads agreed between the designer and the acceptance engineer. Care should be taken to ensure that occasional high service loads in the test spectrum are neither too large nor too numerous since, if they are, the fatigue lives which are obtained may not be representative (due to the fact that occasional high stresses may retard fatigue crack growth). Alternatively, the test load should be the maximum imposed service load, and the number of repetitions representing the total applications of all service loads should be agreed between the designer and the acceptance engineer. The latter should supervise the tests. 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. Owing to the great increase in scatter for tests carried out near the fatigue limit, it is recommended that stresses should be selected to give specimen lives not exceeding 2 × 106 cycles.

? BSI 03-1999

69

BS 7608:1993

Table E.1— Fatigue test factor F
Effective number of specimens Fatigue test factor F

1 2 3 4 10

5 4.2 3.9 3.75 3.5

ww
Figure F.1 — Example of cycle counting by reservoir method

w.

bz

F.1 General The purpose of cycle counting is to reduce an irregular series of stress fluctuations to a simple list of stress ranges. The method given in F.2, and shown in Figure F.1, is suitable when dealing with short stress histories, such as those produced by individual loading events. 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, will be the same as that obtainable by the rainflow method.

fx

w.

co

m

Annex F (normative) Cycle counting by the reservoir method

F.2 Method F.2.1 Derive the peak and trough values of the stress history, due to one loading event. Sketch the history due to two successive occurrences of this loading event. The calculated values of peak and trough stresses may be joined with straight lines if desired. Mark the highest peak of stress in each occurrence. If there are two or more equal highest peaks in one history, mark only the first such peak in each occurrence. F.2.2 Join the two marked points and consider only that part of the plot which falls below this line, like the section of a full reservoir. F.2.3 Drain the reservoir from the lowest point leaving the water that cannot escape. If there are two or more equal lowest points the drainage may be from any one of them. List one cycle having a stress range Sr1 equal to the vertical height of water drained. F.2.4 Repeat F.2.3 successively with each remaining body of water until the whole reservoir is emptied, listing one cycle at each draining operation. F.2.5 Compile the final list which contains all the individual stress ranges in descending order of magnitude Sr1, Sr2, etc. Where two or more cycles of equal stress range are recorded, list them separately. F.2.6 For non-welded details only, a horizontal line representing zero stress should be plotted and those parts of the stress ranges in the compression zone modified as in 3.4.

70

? BSI 03-1999

BS 7608:1993

Annex G (informative) Background notes on the sources of data
G.1 Introduction In general, the objective was to produce a code which was identical with what was considered to be the best practice represented in the set of currently available codes. At this stage substantial amendments were only made when the relevant source text was considered to be in need of improvement or when it would obviously be helpful to include some explanatory matter. It was therefore known, and accepted, that the code was, in some respects, out of date and would need to be updated in the light of new knowledge. A summary of the main sources used in the compilation of the code is given in G.2. Details of the derivation of those new clauses which are not either obvious or self-explanatory are given in G.3. G.2 Sources of the content Table G.1 shows the primary sources of the various clauses forming the code. In some instances the text is not identical with that in the source since some editing was carried out but, except where otherwise noted, there is no substantial change in meaning between the clause and the relevant source. In Table G.1 which follows, the following abbreviations have been used:

GN

Dept of Energy Publication Offshore Installations: Guidance on design, construction and certification (1990)

BS 8118 BS 8118-1:1991 followed by the relevant clause number ECCS Recommendations for the fatigue design of steel structures, European Convention for Constructional Steelwork (ICOM I5O) 1985

In many instances the clauses relating to a particular subject in the various source documents were almost identical with each other even though only one source is quoted. G.3 Comments on new clauses G.3.1 General Some comments on the derivation of those new clauses which are not self-explanatory are given in G.3.2, G.3.3 and G.3.4.

ww

w.

BS 5400 BS 5400-10:1980 followed by the relevant clause number

G.3.2 2.4.3 Quality categories for transverse butt welds In general concept this clause, although new, is self-explanatory and follows directly from the application of PD 6493. Indeed, the quoted flaw sizes for slag inclusions and porosity are copied from Table 5 of PD 6493:1991. The areas of planar flaws corresponding to the various classes were derived (see reference [25]) by fracture mechanics calculations, taking account of experimental fatigue test results for transverse butt welds with lack of penetration flaws. G.3.3 3.2 paragraph 2 It was considered that 6.2.3 of BS 5400-10:1980 was in need of improvement and clarification and a new version was created by the committee. The basic problem is that there are few relevant data to use in defining a sensible rule. G.3.4 3.8 and 4.2.2 Fatigue strength of bolts in tension In the reference standards under consideration design rules for bolts are contained both in BS 5400-10:1980 and the ECCS Publication Recommendations for the fatigue design of steel structures 1985. The rule in the latter is simple in that it specified the use of an S-N curve with slope 3.0 and a strength of 36 N/mm2 at 2 × 106 cycles; clearly this is an approximation to make it fit with the S-N curves for welded joints. Unfortunately the rule in 6.5 of BS 5400-10:1980 is not easy to interpret. However, it is believed to mean that, for fatigue design purposes, the stress range in the bolt has to be multiplied by 1 700/?u (where ?u is the UTS of the bolt material) and the resulting figure then has to be compared with the class B design curve (see Table 17(a), type 1.12 of BS 5400-10:1980). However, the experimental justification for this rule was not known. In view of this situation it was necessary to rewrite this clause and it was decided to base the rule on the experimental data published in ESDU data sheets 67020, 68045 and 69001. Expressed in terms of stress range/UTS the results may be summarized as shown in Table G.2.

bz

fx

w.

co

m

? BSI 03-1999

71

BS 7608:1993

Table G.1 — Sources of the content
Clause in this British Standard Relevant source document

1.5 1.6

Background Design life Paragraph 1 Paragraph 2

w.

1.7

Fatigue loading Paragraph 1 Paragraph 2 Paragraph 3 Paragraph 4 Paragraph 5 Paragraph 6

ww

1.8

Basis of fatigue analysis Paragraph 1 ? Paragraph 2 ? ? Paragraph 3 ? ?

bz

fx

Section 1 General 1.3 Definitions 1.3.1 1.3.2 1.3.3 1.3.4 1.3.5 1.3.6 1.3.7 1.3.8 1.3.9 1.3.10 1.3.11 1.3.12 1.3.13 1.3.14 1.3.15 1.3.16 1.3.17 1.3.18

3.1.1 of BS 5400 and 1.2.11 of BS 8118 New (self-explanatory) A.2 of BS 5400 A1.05 of ECCS New (self-explanatory) 1.2.2 of BS 8118

3.1.12 of BS 5400 New (self-explanatory) New (self-explanatory) 3.1.13 of BS 5400 A.1.11 of ECCS Generalization of 3.1.2 of BS 5400 1.2.20 of BS 8118 A.1.17 of ECCS 3.1.10 of BS 5400 3.1.16 of BS 5400 New (self-explanatory)

paragraph 1 of 4.1 of BS 5400 and 1.2.2 of BS 8118

Section 1 of TWI Bull. 17 May 1976 and
7.2.1 of BS 8118

4.2.1.10 c) of GN 7.4 (final paragraph) of BS 8118 7.4 of BS 8118 and 4.2.1.10 c) of GN 4.2.1.10 d) of GN and 7.4 of BS 8118 9.3.3.2 of BS 5400 9.3.4 of BS 5400

4.2.1.10 d) of GN 4.2.1.10 e) of GN

w.

co

m
? BSI 03-1999

72

BS 7608:1993

Table G.1 — Sources of the content
Clause in this British Standard Relevant source document

1.9

Factors on fatigue life

4.2.1.10 f) of GN 4.5 of BS 5400 amended to take account of thickness effect (4.3 in this code) New (self-explanatory)

1.10 Features influencing fatigue behaviour

1.11 Fracture mechanics Section 2 classification of details 2.1 General Paragraph 1 Paragraph 2 Paragraph 3 Paragraph 4 2.2
? ? ?

5.11.3 of BS 5400

2.3

2.4

ww
Welded steel decks General

Workmanship and inspection 2.4.1 2.4.2 Paragraph 1 Paragraph 2 2.4.3

w.

Unclassified details 13.1 13.2

2.5

Section 3 Stress calculations 3.1 3.2 4.2.1.11 (paragraph 1) of GN

Stress range in parent material Paragraph 1 Paragraph 2 Paragraph 3 Paragraph 4

bz

Classification of details Paragraph 1 Paragraph 2 Paragraph 3 Note

fx

w.
5.2.1 of BS 5400 5.2.2 of BS 5400 5.3.1 of BS 5400 5.4 of BS 5400

New (no comment needed) 5.1.2.2 of BS 5400 5.1.2.3 and 5.1.2.4 of BS 5400 New (no comment required)

5.3.2 of BS 5400, with addition of “weld flaws” ECCS 8.01 and 8.02 plus some new material New

6.2.1 of BS 5400 New (see comment) 6.2.1 and 6.2.2 of BS 5400 4.2.1.11 (paragraph 3) of GN

co

5.1.1.2 of BS 5400 5.1.1.1 of BS 5400 and 4.2.1.10 b) of GN

m

? BSI 03-1999

73

BS 7608:1993

Table G.1 — Sources of the content
Clause in this British Standard Relevant source document

3.3

Stress range for welds

4.2.1.11 b) (paragraph 2) of GN and 3.1.8 b) and 6.3 of BS 5400 6.1.3 of BS 5400

3.4

Effective stress range for details in unwelded members in which the whole or part of the stress is compressive Calculation of stresses 3.5.1 3.5.2 3.5.3

3.5

3.6

New, but self-explanatory

3.7

3.7.3 3.8 3.9 Axial stresses in bolts

Derivation of stress spectra

w.

4.1 4.2

Tensile stress limitations S-N curves

ww

Section 4 Allowable fatigue stresses

4.3

Modification to basic S-N curves 4.3.1 4.3.2 4.3.3 4.3.4 4.3.5

bz

Stress on welds attaching shear connectors 3.7.1 3.7.2

fx

3.6.2

4.2.1.11 a) of GN

6.4.1 of BS 5400 6.4.2 of BS 5400, but taking account of all shear loads rather than just longitudinal 6.4.3 of BS 5400 New, see comment in G.3 New, but no comment required

Based upon 4.2.1.13 a) of GN and Table 10 of BS 5400 (but extensively amended)

New, but no comment needed 4.2.1.13 b) ii) of GN 4.2.1.13 b) i) and 4.2.1.13 c) of GN 4.2.1.13 b) iii) of GN New, but no comment required

w.
? BSI 03-1999

3.6.1 Paragraph 1 Paragraph 2 Paragraph 3

Based on 4.1.2 of BS 5400
? ? ?

74

co

6.1.4.1 of BS 5400 6.1.6 of BS 5400, but not including plate buckling 6.1.5 of BS 5400, but with addition of a), f) and g)

m

BS 7608:1993

Table G.1 — Sources of the content
Clause in this British Standard Relevant source document

4.4 4.5 4.6

Treatment of low stress cycles Treatment of high stress cycles Joints subjected to a single stress range Joints subjected to a stress spectrum 4.7.1 Paragraph 1 Paragraph 2 4.7.2

4.2.1.13 c) of GN and 11.3 of BS 5400 4.2.1.13 d) of GN 4 b) of TWI Bull. Vol 17 May 1976

Paragraph 2

A.3 Damage tolerant design A.4 Safe life design

bz

A.5 Fatigue assessment procedure Annex B B.1 Scope

B.2 No-welded details B.2.1 B.2.2

ww

w.
H.2.1 of BS 5400 H.2.2 of BS 5400 H.3.1 (part) of BS 5400 H.3.2.1 of BS 5400 H.3.2.2 of BS 5400 H.3.2.5 of BS 5400

B.3 Fasteners and shear connectors B.4 Continuous welded attachments B.4.1 B.4.2.1 B.4.2.2 B.4.2.3

fx
New New

A.2 Fatigue life for various failure probabilities A.3 of BS 5400

w.

Annex A A.1 General Paragraph 1

A.1 of BS 5400 Based on A.2 of BS 5400

Based on 7.3 of BS 8118

co

11.1 of BS 5400 and 4.2.1.13 e) of GN TWI Bull. May 1976 11.5 of BS 5400

m

4.7

? BSI 03-1999

75

BS 7608:1993

Table G.1 — Sources of the content
Clause in this British Standard Relevant source document

B.5

Welded attachments on the surface or edge B.5.1 B.5.2.1 B.5.2.2 H.3.1 (part) of BS 5400 H.3.2.3 of BS 5400 H.3.2.4 of BS 5400

B.6

Transverse butt welds B.6.2.1 B.6.2.2 B.6.2.3 B.6.2.4 B.6.2.5 B.6.2.6 B.6.2.7 H.4.2.1 and H.4.2.2 of BS 5400 with new formula H.4.2.3 of BS 5400 H.4.2.4 of BS 5400 H.4.2.5 of BS 5400 H.4.2.6 of BS 5400 New

H.4.3 of BS 5400 (type 3.9 and 3.10)

B.7 B.8

Tranverse butt welds in sections or tubes Load-carrying fillet and T butt joints B.8.1 B.8.2 B.8.3 B.8.4

B.9

Slotted connections

B.10

Branch connections to vessels

ww

Annex C. Guidance on the calculation of stress concentration factors

w.

bz

fx

H.4.3 of BS 5400 (note to type 3.6)

TWI Bull. Vol 17 May 1976, Table 3, type 5

H.4.3 of BS 5400 (types 3.7 and 3.8) H.4.3 of BS 5400 (type 3.9 and 3.10) H.4.3 of BS 5400 (type 3.11) H.4.3 of BS 5400 (types 3.7 and 3.8, final paragraph) Appendix C of BS 5500 Although completely new, the text of this annex appears to be non-controversial and to require no special comment.

w.

co
? BSI 03-1999

76

m

B.6.1

H.4.1 of BS 5400

BS 7608:1993

Table G.1 — Sources of the content
Clause in this British Standard Relevant source document

Annex D. Guidance on the use of fracture mechanics

Based upon appendix D in the ECCS Publication (TC6) Recommendations for the fatigue design of steel structures 1985, with some additional material as follows: a) the equation of A in D.4 is taken from reference [20];

c) D.6 is based upon 5.1 of PD 6493:1991; d) in D.7 the value of K for semi-elliptical surface cracks is taken from reference [23]; e) the values of Mk in D.7 are taken from reference [24].

Annex F. Cycle counting by the reservoir method

Table G.2 — Experimentally determined fatigue strengths of various types of threaded connection expressed in terms of stress range/UTS

mm

bz
N/mm2

Sheet no.

Bolt diameter

UTS

fx
105

w.
Cut or ground 106

Annex E. Fatigue testing and the use of test data to define design stresses

co
105

The text of this annex requires no particular comment. Appendix B of BS 5400-10

m
Form of threads 106

b) the suggested values of ai in D.5 are based upon the findings in references [21] and [22];

Rolled, then heat treated

Heat treated then rolled 105 106

67020 67020 68045

6 to 19 6 to 19 50 to 75

w.
1 040 to 11 110 1 030 to 2 250 405 to 703 593 to 1 186 558 to 1 675

0.37 0.30

0.24 0.178 0.46 0.38 0.38 0.32 0.40 0.34

ww

0.38 0.322 0.25 0.188

0.20 0.162 0.18 0.134 0.26 0.22 0.144 0.102

0.48 0.42

69001 69001
NOTE

10 to 19 6 to 25

0.44 0.33

0.34 0.27

The upper of each pair of figures is the mean value and the lower is approximately mean – 2 standard deviations.

? BSI 03-1999

77

BS 7608:1993

1+d? FS = C ? ---------------? 1 + 3 d? where d is the diameter in inches C is a constant representing the fatigue strength for the limit of zero diameter bolts. The typical value of C was quoted to be ± 14 ksi (= ± 96.5 N/mm2).

ww

w.

bz
78
? BSI 03-1999

fx

w.

co

The results for bolts with cut or ground threads are summarized in Figure G.1, together with the proposed mean and mean – 2 standard deviations curves for use in this standard. The separation between the mean and mean – 2 standard deviations curves in terms of life is approximately the same as that in ESDU data sheet 69001, and that is how the (mythical) mean curve was derived. An analysis by Heywood (see reference [26]) of fatigue test results for bolts, both with machine cut and rolled threads, showed that fatigue strength was linearly related to the tensile strength of the bolt material. Heywood also showed that fatigue strength was related to size by the following empirical relationship:

This implies that, compared with a 25 mm diameter bolt, the fatigue strength of 38 mm and 50 mm diameter bolts are reduced by factors of 0.908 and 0.858. However, it is clear that a size correction of the form represented by equation (4) (see 4.3) is a more logical approach and that gives, for the same dimensions, reduction factors of 0.90 and 0.84. These are virtually identical with the empirical factors derived from Heywood’s relationship and it was therefore decided to make equation (4) applicable to bolts.

m

ww

w. b

zf xw .

BSI

co m

? BSI 03-1999

BS 7608:1993

Figure G.1 — Comparison of proposed S-N curves for bolts in axial loading with curves from ESDU data sheets

79

BS 7608:1993

Annex H (informative) Bibliography
1. PETERSON, R.E. Stress concentration factors, John Wiley and Sons Inc., 1974. 2. ROARK, J.R. and YOUNG, W.C. Formulas for stress and strain, McGraw-Hill, 1975. 3. Engineering Sciences Data Unit. Data sheets, Vol. 3. 4. HEYWOOD, R.B. Photoelasticity for designers, Pergamon Press, 1969. 5. WORDSWORTH, A.C. Experimental determination of stresses at tubular T and X joints, Joint Australasian Welding and Testing Conference, Perth, 1977. 6. IRVINE, N.M. Review of stress analysis techniques used in UKOSRP conference on Fatigue in Offshore Structural Steels, Institute of Civil Engineers, London, February 1981. 7. CLAYTON, A.M. and IRVINE, N.M. Stress analysis method for tubular connections, Paper 30, European Offshore Steels Research Seminar, The Welding Institute, 1978. 8. GIBSTEIN, M.B. Parametric stress analysis of T-joints, Paper 26, European Offshore Steels Research Seminar, The Welding Institute, Cambridge, 1978. 9. KUANG, J.G. et al. Stress concentration in tubular joints, Society of Petroleum Engineers Journal, August 1977. 10. WORDSWORTH, A.C. and SMEDLEY, G.P. Stress concentrations at unstiffened tubular joints, Paper 31, European Offshore Steels Research Seminar, The Welding Institute, Cambridge, 1978. 11. WORDSWORTH, A.C. Stress concentration factors at K and KT tubular joints, Fatigue in Offshore Structural Steels Conference, ICE, London, February 1981. 12. EFTHYMIOU, M. and DURKIN, S. Stress concentrations in T/Y and gap/overlap K joints, Paper Bill, BOSS Conference, 1985. 13. WARDENIER, J. Hollow section joints, Delft University Press, 1982. 14. WYLDE, J.G. and MADDOX, S.J. Effect of misalignment on the fatigue strength of transverse butt welded joints, in Significance of deviations from design shapes, I.Mech.E. Conference Publications, 1979. 15. BURDEKIN, F.M. The effects of deviations from intended shapes on fracture and failure, in Significance of deviations from design shapes, I.Mech.E. Conference Publications, 1979-2.

16. MADDOX, F.J. Fitness for purpose assessments for misalignment in transverse butt welds subject to fatigue loading, International Institute of Welding, IIW-XIII-1080-85, 1985. 17. NEWMAN J.C. and RAJU W.S. Stress intensity factor equations for cracks in three-dimensional finite bodies subjected to tension and bending loads. NASA Technical Memorandum 85793, National Aeronautics and Space Administration, Langley Research Centre, Virginia, April 1984. 18. MADDOX S.J. and ANDREWS, R.M. Stress intensity factors for weld toe cracks, Proc. Conf. Computer aided assessment and controls of localized damage, Springer Verlag, Berlin, 1990. 19. FRANK K.H. and FISHER J.W. Fatigue strength of fillet welded cruciform joints. Journal of Structural Division, Vol. 105, No. ST9, 1979, pp 1727-1740. 20. GURNEY T.R. Fatigue of Welded Structures, Cambridge University, Press, 2nd Edition 1979, p.61. 21. SIGNES E.G., BAKER R.G., HARRISON J.D. and BURDEKIN F.M. Factors affecting the fatigue strength of welded high strength steels. Br. Weld. J, 14 (3) 1967. 22. WATKINSON F., BODGER P.H. and HARRISON J.D. The fatigue strength of welded joints in high strength steels and methods for its improvement, Weld. Inst. Conf. Fatigue of Welded Structures, Brighton, July 1970. 23. NEWMAN J.C. and RAJU I.S. Stress intensity factor equations for cracks in three dimensional finite bodies subjected to tension and bending loads, NASA Tech Memo 85793, April 1984. 24. MADDOX S.J., LECHOCKI J.P. and ANDREWS R.M., Fatigue analysis for the revision of PD 6493:1980, Weld,. Inst. Report 3873/1/86 (for Department of Trade and Industry). 25. GURNEY T.R. The derivation of an acceptance standard for small planar defects in transverse butt welds, Proc. Conf. EIS 90, Engineering Integrity Society, Coventry, March 1990. 26. HEYWOOD R.B. Designing against fatigue, Chapman and Hall, 1962.

ww

w.

bz

fx

w.

co

m

80

? BSI 03-1999

BS 7608:1993

List of references (see 1.2)
Normative references
BSI standard publications
BRITISH STANDARDS INSTITUTION, London

Informative references
BSI standards publications

BRITISH STANDARDS INSTITUTION, London

Other references

Offshore installations: Guidance on design, construction and certification, 4th edition, HMSO, 1990. Background to new fatigue design guidance for steel welded joints in offshore structures, HMSO, 1984. Recommendations for the fatigue design of steel structures, European Convention for Constructional Steelwork, (ICOM15O), 1985. GURNEY T.R. Fatigue design rules for welded steel joints, Welding Institute Research Bulletin, Vol. 17, (Report R124/5/76) May 1976. Fatigue strength of steel screw threads with large root radii under axial loading, ESDU Data Sheet 67070, Engineering Sciences Data Unit International Ltd., London, 1967. Fatigue strength of large steel screw threads under axial loading, ESDU Data Sheet 68045, Engineering Sciences Data Unit International Ltd., London, 1968. Fatigue strength of steel screw threads under axial loading (not greater than 1 inch diameter), ESDU Data Sheet 69001, Engineering Sciences Data Unit International Ltd., London, 1969.

? BSI 03-1999

ww

w.

BS 5400, Steel, concrete and composite bridges. BS 5400-10:1980, Code of pratice for fatigue. BS 5500:1991, Specification for unfired fusion welded pressure vessels. BS 8118, Structural use of aluminium. BS 8118-1:1991, Code of practice for design.

bz

fx

w.

BS 3643, ISO metric screw threads. BS 3643-2:1981, Specification for selected limits of size. BS 3692:1967, Specification for ISO metric precision hexagon bolts, screws and nuts. Metric units. BS 4190:1967, Specification for ISO metric black hexagon bolts, screws and nuts. BS 4395, Specification for high strength friction grip bolts and associated nuts and washers for structural engineering. BS 4395-1:1969, General grade. BS 4395-2:1969, Higher grade bolts and nuts and general grade washers. BS 4604, Specification for the use of high strength friction grip bolts in structural steelwork. Metric series. BS 4604-1:1970, General grade. BS 4604-2:1970, Higher grade (parallel shank). BS 5135:1984, Specification for arc welding of carbon and carbon manganese steels. BS 5400, Steel, concrete and composite bridges. BS 5400-5:1979, Code of practice for design of composite bridges. PD 6493:1991, Guidance on methods for assessing the acceptability of flaws in fusion welded structures.

co

m

BS 7608:1993

BSI — British Standards Institution
BSI is the independent national body responsible for preparing British Standards. It presents the UK view on standards in Europe and at the international level. It is incorporated by Royal Charter. Revisions British Standards are updated by amendment or revision. Users of British Standards should make sure that they possess the latest amendments or editions. It is the constant aim of BSI to improve the quality of our products and services. We would be grateful if anyone finding an inaccuracy or ambiguity while using this British Standard would inform the Secretary of the technical committee responsible, the identity of which can be found on the inside front cover. Tel: 020 8996 9000. Fax: 020 8996 7400. BSI offers members an individual updating service called PLUS which ensures that subscribers automatically receive the latest editions of standards. Buying standards

Orders for all BSI, international and foreign standards publications should be addressed to Customer Services. Tel: 020 8996 9001. Fax: 020 8996 7001.

BSI 389 Chiswick High Road London W4 4AL

ww

Copyright

Copyright subsists in all BSI publications. BSI also holds the copyright, in the UK, of the publications of the internationalstandardization bodies. Except as permitted under the Copyright, Designs and Patents Act 1988 no extract may be reproduced, stored in a retrieval system or transmitted in any form or by any means – electronic, photocopying, recording or otherwise – without prior written permission from BSI. This does not preclude the free use, in the course of implementing the standard, of necessary details such as symbols, and size, type or grade designations. If these details are to be used for any other purpose than implementation then the prior written permission of BSI must be obtained. If permission is granted, the terms may include royalty payments or a licensing agreement. Details and advice can be obtained from the Copyright Manager. Tel: 020 8996 7070.

w.

Subscribing members of BSI are kept up to date with standards developments and receive substantial discounts on the purchase price of standards. For details of these and other benefits contact Membership Administration. Tel: 020 8996 7002. Fax: 020 8996 7001.

bz

BSI provides a wide range of information on national, European and international standards through its Library and its Technical Help to Exporters Service. Various BSI electronic information services are also available which give details on all its products and services. Contact the Information Centre. Tel: 020 8996 7111. Fax: 020 8996 7048.

fx

Information on standards

w.

In response to orders for international standards, it is BSI policy to supply the BSI implementation of those that have been published as British Standards, unless otherwise requested.

co

m


相关文章:
BS_7608_1993_图文.pdf
BRITISH STANDARD BS 7608:1993 Incorporating Amendment No. 1 Code of practice for Fatigue design and assessment of steel structures BS 7608:1993 Committees ...
疲劳讲义(BS & IIW--OK).ppt
[5]BS7608:1993 CODE OF PRACTICEFOR FATIGUE DESIGN AND ASSESSMENT OF STEEL STRUCTURES 1993 (参见它的电子读本---大连交通大学编译 2005) 参见它的电子读本-...
大连交通大学疲劳讲义.pdf
例如:英国标准(BS7608:1993)(Code of practice for fatigue design and assessment of steel structures)、国际焊接学会标准(IIW)(Fatigue design of welded joints...
疲劳讲义_图文.ppt
例如:英国标准(BS7608:1993)(Code of practice for fatigue design and assessment of steel structures)、国际焊接学会标准(IIW)(Fatigue design of welded joints...
EN15085-3培训_图文.pdf
通常该曲线都适用于特定类型的接头(对接焊缝、 角焊缝等) 英国标准(BS7608:1993)(Code of practice for fatigue design and assessment of steel structures) 国际...
基于名义应力法的焊接结构疲劳寿命评估概述_图文.ppt
BS标准概述 Fatigue design and assessment of steel structures 钢结构疲劳设计与...BS7608标准(1993)的应用流程 大连交通大学 载运工具先进技术研究中心 怎样使用BS...
焊接接头的设计与工艺2015.3_图文.pdf
Page 30 3、焊接结构设计英国:钢结构疲劳设计与评估实用标准 (Code of practice for fatigue design and assessment of steel structures) BS(7608:1993) Page 31...
焊接结构虚拟疲劳试验技术研究_图文.pdf
1996.Fatigue design of welded joints and components 1996 3.BS 7608.Code of practice for fatigue design and assessment of steel structures 1993 4.刘德刚...
疲劳讲义(BS & IIW--class)_图文.ppt
[5]BS7608:1993 CODE OF PRACTICEFOR FATIGUE DESIGN AND ASSESSMENT OF STEEL STRUCTURES 1993 (参见它的电子读本---大连交通大学编译 2005) 参见它的电子读本-...
调车机车的车体设计_图文.pdf
根据英国标准BS7608:1993(codepracticeforFatiguedesignandassessmentofsteelof 个COMBINl4弹簧单元、5152个MASS21质量单元。 分别对静强度、疲劳强度和车体振动模态进行...
抓斗卸船机金属结构疲劳寿命估算.pdf
Key words: Fatigue life Nominal stress method ...design and assessment of steel structures# 的规定...北京: 国防工业出版社 5 BS 7608: 1993 Code of...
平行流换热器疲劳故障分析及解决方案_图文.pdf
[4]BS7608 Codeofpractice forFatiguedesignand as 1)在设计平行流换热器的...sessmentofsteelstructures.BritishStandard,1993. [5]Stephens IRalph,FatemiAli...
焊接构件虚拟疲劳仿真分析_图文.pdf
[ 4 ] BS7608. Code of p ractice for fatigue design and as2 sessment of steel structures [ S ]. London, B ritish Standards Insution, 1993. [5...
太阳能电池片导电胶串焊技术及电池片受力分析_图文.pdf
BS7608 - 1993 Fatigue Design and As[3] IIW Joint Working Group. XIII...sessment of Steel Structures[ S] . British:BSI, 1993. Document Recommen-...
基于焊缝疲劳寿命预测的后桥设计改进.pdf
参考文献 图 11 后桥纵向力物理台架试验示意图[ 1] British Standard BS7608. Code of practice for fatigue design an assessment of steel structures[M] . ...
690MPa级海洋平台用钢手工电弧焊条熔敷金属性能分析_王....pdf
BS7608 B S Code of practice for fatigue design and assessment of 李 向伟...Fatigue prediction of spot welded joints using equivasteel structures[S]. ...
BS8100-1 铁塔及桅杆-荷载.pdf
BS8100-1 铁塔及桅杆-荷载_电力/水利_工程科技_专业...(vortex shedding) 3.4 Fatigue life assessment 3...and a guide covering the loading and design of ...
AS 4100-1990 Steel Structures.pdf
the limit state steel structures Standards of other...Methods of analysis Brittle fracture Fatigue Fire ...113 13.3 DESIGN AND DETAILING REQUIREMENTS . . ...
疲劳讲义(BS & IIW--OK)_图文.ppt
[5]BS7608:1993 CODE OF PRACTICEFOR FATIGUE DESIGN AND ASSESSMENT OF STEEL STRUCTURES 1993 (参见它的电子读本---大连交通大学编译 2005) 参见它的电子读本-...
references.pdf
Code of Practice for Fatigue Design and Assessment of Steel Structures. BS 7608, 1993. 35. Radaj, D. “Local Fatigue Strength Characteristic Values for ...
更多相关标签: