Fatigue Behaviour of Orthotopic Steel Decks of Road Bridges

Commission of the European Communities

technical steel research

Properties and service performance

Measurement and interpretation of dynamic loads in bridges

Phase 3: Fatigue behaviour of orthotopic steel decks of road bridges

Commission of the European Communities

technical steel research

Properties and service performance

Measurement and interpretation of dynamic loads in bridges

Phase 3: Fatigue behaviour of orthotopic steel decks of road bridges

C. Beales Transport and Road Research Laboratory

Old Workingham Road Crowthome, Berkshire RG11 6AU

United Kingdom

Contract No 7210 - KD/807 (1.7.1986-31.12.1988)

Final report

Directorate-General Science, Research and Development

1990 EUR 12792 EN

Published by the

COMMISSION OF THE EUROPEAN COMMUNITIES

DirectorateGeneral

Telecommunications, Information Industries and Innovation

L2920 Luxembourg

LEGAL NOTICE

Neither the Commission of the European Communities nor any person acting on behalf of the Commission is responsible for the use which might be made of

the following information

Cataloguing data can be found at the end of this publication

Luxembourg: Office for Official Publications of the European Communities, 1990

ISBN 92-826-1505-7 Catalogue number: CD-NA-12792-EN-C

© ECSC-EEC-EAEC, Brussels ■ Luxembourg, 1990

Printed in Belgium

C O N T E N T S

Page

1. INTRODUCTION 1 2. ORTHOTROPIC STEEL BRIDGE DECKS - GENERAL CONSIDERATIONS 2 3. TRRL PROPOSALS 5 4. STATIC LOAD TESTS ON A DECK PANEL 5

4.1. Test panel 5 4.2. Details of test rig and test procedure 6 4.3. Instrumentation 7 4.4. Test results 7

4.4.1. Stress distributions 7 4.4.2. Influence Unes of stress 8

4.5. Provisional fatigue assessment 8 5. CONSTANT AMPLITUDE FATIGUE TESTS ON TYPE 'B' CONNECTION 9

5.1. Fabrication of test specimens 9 5.2. Test rig 10 5.3. Instrumentation 10 5.4. Loading arrangement 10 5.5. Test procedure 11 5.6. Test results 12

5.6.1. Crack development 12 5.6.2. Residual stresses 13 5.6.3. Weld classification : S-N curves 14 5.6.4. Calculated fatigue Ufe 14

6. CONSTANT AMPLITUDE FATIGUE TESTS ON TYPE 'A' CONNECTION 14 6.1. Fabrication of test specimens 14 6.2. Test rig 15 6.3. Instrumentation 15 6.4. Loading arrangement 15 6.5. Test procedure 15 6.6. Test results 15

6.6.1. Crack development 15 6.6.2. Weld classification : S-N curves 16 6.6.3. Calculated fatigue Uves 17

7. MEASUREMENTS ON A BRIDGE UNDER TRAFFIC LOADING 17 7.1. Bridge inspection 17

III

7.2. Strain measurements 18 7.2.1. Instrumentation 18 7.2.2. Results 18

7.3. Conclusions for the bridge 19 8. VARIABLE AMPLITUDE FATIGUE TESTS ON TYPE 'B' CONNECTION 20

8.1. Derivation of loading spectrum 20 8.2. Test specimens, instrumentation and test rig 21 8.3. Loading arrangement 22 8.4. Test procedure 22 8.5. Test results 22

8.5.1. Crack development 22 8.5.2. Fatigue endurance 22

9. SUMMARY OF TEST RESULTS AND FURTHER DISCUSSION 23 10. CONCLUSIONS 26 11. REFERENCES 27 12. ACKNOWLEDGEMENTS 29 APPENDICES 31

Appendix 1 - UK bridges with orthotropic steel decks 33 Appendix 2 - TRRL proposals - Technical annex 43 Appendix 3 - Manufacture of test specimens 47 Appendix 4 - Test specimens - Welding details 57 Appendix 5 - Inspection of test specimens 63 Appendix 6 - Test specimens - Weld sizes 77

TABLES AND FIGURES 81

IV

TABLES

1. Commercial vehicle types for fatigue assessment (from BS5400: part 10)

2. Provisional fatigue assessment 3. Stresses in type 'B' fatigue specimens (1st series) 4. Fatigue test results - type 'B' specimens 5. Residual strains 6. Stresses from tests on deck panel 7. Stresses in type 'A' fatigue specimens 8. Fatigue test results - type 'A' specimens 9. Stresses from measurements on bridge 10. Original Rheden spectrum 11. Rheden spectrum χ 1.5 12. Spectrum selected for the study 13. Stresses in type 'B' fatigue specimens (2nd series)

FIGURES

1. Types of longitudinal / transverse stiffener connection 2. Typical influence lines - crossbeam to deck plate connection 3. Typical influence lines - trough to crossbeam connection 4. Trough to crossbeam connection -

typical influence lines under free flowing traffic 5. Plan view of deck panel 6. Deck panel - details of test crossbeam 7. Deck panel - test crossbeam weld details 8. Deck panel - test crossbeam support conditions 9. Static load test rig 10. Loading grid for static tests on deck panel 11. Strain gauge installation - connection *A' 12. Distribution of stress around connection 'A' 13. Strain gauge installation - connection 'B' 14. Distribution of stress around connection 'B' 15. Strain gauge installation - connection 'C' 16. Distribution of stress around connection 'C' 17. Influence lines - gauge 10 18. Influence lines - gauge 13 19. Influence lines - gauge 19 20. Influence lines - gauge 49 21. Influence lines - gauge 52 22. Influence lines - gauge 67 23. Influence lines - gauge 70 24. Influence lines - gauge 82 25. Influence lines - gauge 88 26. Influence lines - gauge 89 27. Influence lines - gauge 90 28. Influence lines - gauge 91 29. Distribution of vehicles across deck 30. Variation of fatigue life with position of vehicles

- V

31. Fatigue test rig 32. Strain gauge positions - type 'B' fatigue specimens 33. Section at apex of trough - type 'B' specimen 34. Crack development - specimen 3B 35. Crack development - specimen 4B 36. Crack development - specimen 6B 37. Crack development - specimen 8B 38. Effect of cracking 39. BS5400 and Eurocode S-N curves 40. Fatigue tests at constant amplitude - type 'B' specimens 41. Strain gauge positions - type 'A' fatigue specimens 42. Strain gauge positions - fatigue specimen 8A 43a. End of test cracks - specimen 1A 43b. Crack development - specimen 1A 44a. End of test cracks - specimen 2A 44b. Crack development - specimen 2A 45a. End of test cracks - specimen 3A 45b. Crack development - specimen 3A 46a. End of test cracks - specimen 5A 46b. Crack development - specimen 5A 47a. End of test cracks - specimen 8A 47b. Crack development - specimen 8A 48. Fatigue tests at constant amplitude - type 'A' specimens 49. Strain gauge positions on bridge 50. Distribution of damage and cycles (from Table 12) 51. Typical output from chart recorder 52. Crack development - specimen 10B 53. Crack development - specimen 12B 54. Crack development - specimen 13B 55. Crack development - specimen 14B 56. Tests on type 'B' specimens

VI -

1. INTRODUCTION The work described in this report is the TRRL

contribution to a six nation collaborative research programme on the "Measurement and interpretation of dynamic loads in bridges". This is the third phase of research under this title. The British contribution to the first phase concerned the development of measurement techniques and the collection of traffic load and stress data at three UK bridges. This work was reported in July 1979'**. The second phase investigated the factors which relate the load spectra to the stresses generated by them. The British contribution, reported in 1983<2>, highlighted the important part played by the asphalt surfacing in reducing stresses and hence increasing fatigue life. However, it was accepted that the variable properties of the surfacing make it difficult to take account of in design.

The objectives of the third phase of research were to improve the understanding of the fatigue behaviour of the orthotropic decks of steel bridges. At TRRL the research has concentrated on the welded connection between the longitudinal and transverse stiffeners (otherwise referred to as the trough to crossbeam connection). The work was divided into five main stages, each reported in half-yearly progress reports :

Progress Report no.1 for the period 1.07.86 to 31.12.86: general assessment of orthotropic decks and detailing of test programme.

Progress Report no.2 for the period 1.01.87 to 30.06.87: static load testing of a deck panel in the laboratory and preliminary fatigue assessment.

Progress Report no.3 for the period 1.07.87 to 31.12.87: constant amplitude fatigue tests on specimens representing type 'B' trough to crossbeam connection (see Fig 1).

Progress Report no.4 for the period 1.01.88 to 30.06.88: constant amplitude fatigue tests on specimens representing type 'A' connection.

Progress Report no.5 for the period 1.07.88 to 31.12.88: variable amplitude fatigue tests on type 'B' specimens.

This final report brings together all of the work described above. Results from tests on a UK bridge are included which were an addition to the original test programme. A fatigue assessment of the type 'A' and 'B' connections is given and suggestions are made for modifications to the designs to improve the fatigue behaviour of the connection. Further testing is recommended to verify these design changes.

1 -

2. ORTHOTROPIC STEEL BRIDGE DECKS - GENERAL CONSIDERATIONS

Eight major bridges in the UK with orthotropic steel decks are listed, together with the main deck dimensions, in Appendix 1. Although few in number, they represent a significant financial investment and form major road links.

The Severn and Wye bridges and the Beachley viaduct together link South-West England to the industrial heartland of South Wales. They are amongst the most heavily trafficked bridges in the country carrying around eight hundred thousand heavy goods vehicles per year in the slow lanes of each carriageway. The crossing is at present undergoing a thirty-five million pound strengthening operation. Some work involves strengthening the deck, particularly under the slow lane wheel tracks. One area requiring routine inspection and repair is the welded connection between the longitudinal and transverse stiffeners.

Bridges of this type, with large box sections fabricated from thin steel plates, stiffened with welded plates or trough sections, were first used in the UK in the mid 1960s. The Severn, Wye and Forth bridges used trapezoidal shaped longitudinal deck stiffeners, about 3 to Ai metres long. These were butted up - to the transverse stiffeners (crossbeams) and fillet welded all round. The remaining bridges listed in Appendix 1 were built after 1970 and used 'V* shaped longitudinal stiffeners about 14 metres long. These passed through cut-outs in the crossbeams, the 'V' stiffeners themselves being joined end to end some distance away from the crossbeams.

Fatigue failures of the early type of connection have been reported <=»·■*> . Nunn

<e> had earlier found fatigue cracks

in an identical connection in an experimental bridge deck panel installed in a trunk road at Denham. These cracks occurred after only four years of normal trafficking. Nunn concluded that poor fabrication standards had accelerated the onset of fatigue and that fabrication tolerances and welding sequence also influenced the fatigue life. It was recommended that the connection be assessed for fatigue using the data for class G welds, BS 153**» (since superseded by BS5400*"

:r>).

Much work has been carried out at TRRL in recent years on the assessment of orthotropic decks. The work has involved the measurement of strains on bridges and on full-scale deck panels tested in the laboratory. Fatigue tests on specimens representing sections of the bridge deck have also been carried out. Much of this work has been published as TRRL laboratory reports«»-··»·»»-*****·*»». Two important factors have emerged from this work. Firstly, the calculated fatigue lives for most of the welded connections in the deck which occur under the wheel tracks are less than the 120 year design life required by the British Standard Β55400

<-τ>. These

calculations are based on measured strains on the unsurfaced deck and use the table of vehicle types and transverse distribution of vehicles detailed in the British Standard.

Secondly, the effectiveness of the bridge deck surfacing in reducing stresses in most of the welded connections has been realised. Beales'**' showed that there was a significant increase in the calculated fatigue life for most connections if the measured strains from the surfaced deck were used.

Some indication of the magnitude of the stiffening effect of the surfacing can be seen in Fig 2. In Fig 2a longitudinal influence lines of stress are shown for the crossbeam to deck plate connection. Stresses were calculated from strains measured by a gauge installed close to, and with its axis perpendicular to, the weld and mid-way between troughs. Measurements were made in the laboratory on a full scale bridge deck panel loaded by a single wheel to 20kN. The wheel was located mid-way between troughs and strains were recorded as the wheel was moved, in increments, along the panel. The effective influence line is seen to be very short, rising from zero through the negative and positive peak values and returning to zero in little over a metre. Influence lines for a multi-axled vehicle can be calculated by superposition from the single wheel data. Since, in this case, the influence line for the single wheel is shorter than the distance between axles on any heavy goods vehicle (HGV), the vehicle influence line is simply a series (one per axle) of magnified cycles of the single wheel stress. The calculated influence line for a 32kN wheel load is also shown in Fig 2a.

Fig 2b shows influence lines for an identical gauge location from measurements on a bridge. In this case, loading was from a two-axle HGV; the influence line for the front wheel (32kN) is shown. The full line shows the stresses recorded with the bridge deck surfacing removed. Although the longitudinal scale of the graph has been expanded, it can be seen that the peak values are in good agreement with the laboratory measurements for the 32kN loading.

The broken line in Fig 2b represents the stresses recorded with the bridge deck surfaced with 38mm of mastic asphalt. The surfacing temperature during the test was 14CC. The stress range is seen to reduce from 101N/mma (-68 to +33) to 14N/mm* (-13 to +1). If, as a first estimate, fatigue life is considered to be inversely proportional to the cube of the stress range, then the effect of the surfacing in this case is to increase life by a factor of 375 over that calculated for the unsurfaced deck. This is, however, an optimistic assessment and the following factors must also be considered:

(i) The stiffness of the surfacing is highly dependent on temperature. In very hot conditions the effectiveness of the surfacing in reducing stresses in the deck can be almost totally lost.

(ii) Asphalt displays visco-elastic properties and hence its effectiveness in reducing stresses in the deck is dependent on vehicle speed.

- 3

(iii) Cracking of the asphalt has been shown to lead to a loss of composite action of the surfacing'

1*·

1β>. Debonding

of the surfacing from the steel deck plate would be expected to have a similar effect.

(iv) The stress ranges given here were based on measurements from a bridge surfaced with 38mm of mastic asphalt. Thinner surfacings or those with a lower modulus would be expected to reduce stresses to a lesser degree.

These variables make it difficult to assess the overall 'surfacing factor'. Consequently, fatigue assessments are normally based on measured or calculated stresses for the unsurfaced deck. Nevertheless, the surfacing factor is almost certainly the reason why significantly more fatigue failures have not occurred in heavily trafficked bridges.

The characteristics of the crossbeam to deck plate connection (short longitudinal influence line, large reduction of stress from surfacing) may be considered typical of those connections close to the deck plate. In contrast, the influence line for the trough to crossbeam connection is long and the surfacing factor small.

Influence lines for this connection are shown in Fig 3. The influence line for the single wheel extends to adjacent crossbeams, consequently there is interaction between the multiple axles of an HGV. A typical influence line for a two-axle vehicle is shown in Fig 3a, calculated from the superposition of the single wheel stresses. In Fig 3b, influence lines are shown from measurements on a bridge for a vehicle of similar wheel loads and axle spacing. In this case the maximum stress range is only reduced by a factor of 2.4, from 45N/mm

Ä to 19N/mm

z. In an assessment of this connection

on the Wye bridge, an overall 'surfacing factor' of 3 was considered appropriate, increasing the calculated fatigue life from 4.3 years for the unsurfaced deck to 12.9 years for the surfaced deck'

1**.

Typical influence lines under traffic loading for the trough to crossbeam connection are shown in Fig 4. Sharp peaks, corresponding to the passage of 5 HGVs close to the instrumented trough, are shown in Fig 4a. Stresses below 5N/mm

z were caused by cars or heavy vehicles in the adjacent

lane. A more detailed record of vehicles VI and V2 is given in Fig 4b. These are characteristic traces of four-axle HGVs. It is possible to establish from the trace that the two vehicles were travelling in close convoy, approximately 32 metres apart at 64kra/hr.

Much information exists related to the early design of trough to crossbeam connection. Influence lines have been obtained for the surfaced and unsurfaced decks, the weld classification has been established and the fatigue life calculated. Cuninghame

<xe> has investigated repairs to welds

which have cracked in service. Little information has been found relating to the later designs of connection. There is

- 4

no reason to believe that the surfacing factor for the later design is greater than that for the early design but it is likely that the Joint classification is improved by continuing the trough through the crossbeam. It is the aim of this research to investigate the later designs of connection in detail, to assess the fatigue life of the type of connections in service today and to establish whether current designs are suitable for future use in bridges.

3. TRRL PROPOSALS The TRRL proposals, submitted to the Commission in April

1986, are given in the Technical Annex in Appendix 2. The tasks to be carried out may be summarised as follows:

(i) Fabricate a trial deck panel incorporating different designs of trough to crossbeam connection.

(ii) Dynamically load the panel and study the stress distribution around each connection using equipment designed to sense thermal emissions produced by the thermoelastic behaviour of the material.

(iii) Measure strains on the panel under the action of a single static wheel load.

(iv) Carry out constant amplitude fatigue tests on specimens representing two of the designs of connection.

(v) Carry out fatigue tests at variable amplitude on one design of connection and assess the application of the Palmgren-Miner summation;

The second of these tasks has not been completed. Technical difficulties with the hydraulic control equipment and problems associated with the measuring equipment could not be resolved in the time available.

Additional work not specified in the original proposals was, however, undertaken. Following unexpected fatigue failures on some specimens at very short endurances, strains were measured for a continuous two week period on a bridge under traffic loading. The important 'surfacing factor' was thereby assessed.

4. STATIC LOAD TESTS ON A DECK PANEL 4.1 Test panel

An orthotropic steel deck panel, shown in plan view in Fig 5, was fabricated and tested at TRRL in the mid 1970s. The panel comprises five 'V' shaped longitudinal stiffeners passing through cut-outs in the three intermediate crossbeams. The panel had been used to determine stresses around the central crossbeam and in the two adjacent 4.57m bays either side of the crossbeam. The outer 3.05m bays provided longitudinal continuity.

5 -

The panel was modified for use in this project. The central crossbeam was removed from the panel by flame cutting and grinding and a replacement crossbeam, shown in Fig 6, fitted in its place. Welding details are given in Fig 7.

The central crossbeam thus incorporated five trough to crossbeam connections. The outer two were not used as test connections because of possible edge effects from the side of the panel. The three remaining connections are shown in perspective in Fig 1 and have been identified as types Ά', 'B' and 'C'.

Connections type 'A' and 'C' are typical of those to be found in the later UK bridges and examples of these types can be seen in Appendix 1. Cope holes are cut out of the crossbeam around the apex of the troughs. It is understood that this is to ease fabrication of the panel, particularly the fit-up between the trough and crossbeam around the tight bend at the apex of the trough. Additional cut-outs are incorporated at the top corners of the crossbeam around the continuous trough to deck plate weld.

No cut-outs are incorporated in the type 'B' connection and the longitudinal stiffener is welded all round, on both sides of the crossbeam plate. No examples of this type of connection exist in the UK but a similar design was used in two Korean bridges (from a 1980s British design«1"'* ) . This type of connection is believed to be commonly used on the continent of Europe, though trapezoidal shaped longitudinal stiffeners are often used in preference to 'V' shaped stiffeners.

The crossbeams of the panel were bolted to girders which were clamped to the main longitudinal members of a reaction frame. Special attention was paid to the design of the central girder and to the connection between this girder and the test crossbeam. The design and detailing of this connection, shown in Fig 8, was based'on that found on the Avonmouth bridge. 4.2 Details of test rig and test procedure

The test rig is shown in Fig 9. It consists of two lower longitudinal beams supporting the panel through cross-girders and crossbeams. Vertical columns at the corners of the frame support two upper longitudinal beams. A transverse beam, suspended from and reacting against the upper longitudinal members, carries a single wheel and axle assembly. The wheel is loaded hydraulically and applies load to the test panel through a 1000 χ 20 radial ply tyre, typical of those used on HGVs. Under a test load of 20kN, the 'footprint' of the tyre measures approximately 235mm long by 185mra wide.

The wheel and axle assembly can be moved, by means of a handwheel, along the transverse supporting beam. Similarly, the transverse beam can be moved longitudinally along the panel. The test procedure consists of monitoring strain

gauges while a static load is applied through the wheel which is systematically positioned at a number of locations on the panel. These locations are marked out as a grid, shown in Fig 10, the loading positions being at the Intersection of the longitudinal and transverse lines. The full test used 702 measurement positions. The panel was tested without surfacing. 4.3 Instrumentation

Strain gauges were installed on the test crossbeam and on the webs of the three troughs as shown in Figs 11, 13 and 15. The gauges were normally installed as bi-axial pairs with one element parallel to and one element perpendicular to the line of the weld. The grid length of the gauges was 6.35mm and the centreline of the grid was located 15mm from the root of the weld. At key points, notably at the end of the weld near the cope-hole in connections 'A' and 'C' and at the apex of the trough in connection 'B', rosette gauges were installed.

Single element gauges 88-91 were installed on connection 'A' at a later stage, following unexpected failures of fatigue test specimens of this type. A reduced testing programme was carried out to determine the influence surface of stress for these gauges.

Strain gauge signals and outputs from load cells located either side of the axle to accurately determine the applied load, were recorded onto magnetic tape. Strains were subsequently adjusted to match a load of exactly 20kN and resolved stresses were computed. 4.4 Test results

The basic strain data for each of the 91 strain gauge elements and for each of the 702 loading positions have been printed out but are not presented in this report because of their bulk. The data are, however, available by personal application to the project officer at TRRL. Selected data for the important gauges and wheel positions are presented in the following two sections. Rosette gauges will hereafter be referred to by the gauge number of the centre element and stresses quoted will be those resolved perpendicular to the weld unless otherwise stated. 4.4.1 Stress distributions

Stress distributions around the three connections are shown in Figs 12, 14 and 16. For each connection two positions of the wheel are considered; (a) wheel over centre of trough and 1220mm from crossbeam and (b) wheel over centre of trough and directly over the crossbeam. Position (a) was found to produce high stress at the apex of the trough in connection 'B' (gauge 49) and high stress in the web of the trough at the bottom of the weld in connections 'A' (gauge 13) and 'C' (gauge 70). Very high stresses were later

7 -

measured in connection 'A' at the top of the weld near the crossbeam cut-out (gauges 88 and 89) with the wheel positioned over the crossbeam. 4.4.2 Influence lines of stress

Influence lines of stress for the gauges located at the highest stressed locations are given in Figs 17 to 28. In each case the principal longitudinal and transverse influence lines are shown, that is, the line on which the maximum stress was recorded. The longitudinal influence line for the gauge on the soffit of trough 'B' (Gauge 49 - Fig 20) is seen to be very similar to that measured in earlier tests on a deck panel and on the Wye bridge (Fig 3). 4.5 Provisional fatigue assessment

A provisional fatigue assessment of the three connections was carried out using a program developed at TRRL and broadly based on the procedures described in the British Standard BS5400 partlO«"". The vehicle types used in the assessment are given in Table 1. Vehicles were transversely distributed across the deck as, for example, in Fig 29 where the centre of the distribution is located directly over trough 'B'.

Each vehicle in turn was 'run' along the longitudinal influence lines of measured stress (stress calculated from measured strain) and the influence line due to the vehicle calculated by superposition of the single wheel stresses. The single wheel influence lines were extrapolated out to the adjacent crossbeams and'to the outer troughs as shown, for example, in Fig 24. The 'Rainflow' cycle counting method was used to determine histograms of stress ranges.

For this provisional assessment, S-N data for Class F welds (BS5400, part 10) were used. This was an estimate of the likely classification of the connections based on guidance from the Standard (orthotropic decks are not within the scope of this Standard) and from past experience. The centreline of the distribution of vehicles was positioned at a number of points to determine the worst (lowest fatigue life) location for each connection and measurement position. Fatigue lives were calculated for a 2.37. probability of failure (using mean minus two standard deviation S-N data) and for a vehicle flow of one million HGVs per year. The results are summarised in Table 2.

For connection 'B* the lowest fatigue life was calculated for gauge 49 on the soffit of the trough adjacent to the crossbeam. A life of 6.3 years was calculated for the vehicles centred directly over the trough (line 15). the lowest life (6.1 years) was calculated for the vehicles just off the centreline of the trough on line 14.

For connections 'A' and 'C* the lowest fatigue life was calculated for the gauge on the web of the trough at the

bottom of the weld, gauges 13 and 70 respectively. Again the worst case was with the vehicles centred directly over, or adjacent to, the centreline of the trough.

The effect on fatigue life of the position of the vehicles is seen in Fig 30. Fatigue lives increase sharply as the loading moves away from the centre of the instrumented trough but this is clearly to the detriment of the adjacent trough. The optimum position would therefore appear to be to to locate wheel tracks mid-way between troughs by suitable arrangement of the lane markers. Such an arrangement may not be feasible for such frequently occurring connections but there is merit in avoiding wheel tracks coinciding with infrequently occurring connections such as the web of box to deck plate connection and the longitudinal splice joints between deck panels.

Connections 'A' and 'B' were selected for the second phase of testing to establish their constant amplitude fatigue performance. The very short life calculated for connection 'B' was of particular concern. If the weld classification F (assumed in the provisional assessment) was confirmed, failures of this type of connection could be expected on heavily trafficked bridges even allowing for the surfacing factor. Connections 'A' and 'C' might be expected to have similar weld classifications. 'A' was selected, having the better provisional fatigue assessment. Testing began with specimens representing the type 'B' connection. 5. CONSTANT AMPLITUDE FATIGUE TESTS ON TYPE 'B' CONNECTION 5.1 Fabrication of test- specimens

Eight test specimens were manufactured; they are illustrated in Fig 31. The specimens are full scale and comprise a 1500mm length of deck plate and trough with a central crossbeam and an end plate. An internal gusset plate is fitted 100mm in from the open end of the trough to stiffen the specimen under the loading plate.

Details of the specification for the manufacture of the specimens together with the fabrication drawings of the specimens (for types 'B' and 'A') and other items are given in Appendix 3.

The component parts of the specimen were clamped together on a flat, rigid bed, deck plate down - troughs uppermost, in the same way that a full size deck panel would be fabricated. 6mm, single pass, Manual Metal Arc (MMA) fillet welds were specified for all the main welds - trough to deck plate, crossbeam to deck plate and trough to crossbeam. The welding procedure adopted by the manufacturer is given in Appendix 4.

After the specimens were manufactured and prior to their delivery to TRRL, independent engineers were employed to inspect the specimens to ensure that they conformed to the

- 9

specification. In particular they were asked to carry out a magnetic particle flaw detection inspection on the important trough to crossbeam welds. The specification for the inspection of the specimens and the inspection report, together with the material certificates, are given in Appendix 5.

A detailed survey of weld sizes for the trough to crossbeam weld was carried out at TRRL after the specimens had been tested. The results are summarised in Appendix 6. Both leg lengths and the throat thickness were measured; (i) all round the trough to within 50mm of the deck plate, (ii) around the apex of the trough where cracking occurred and (iii) at the crack initiation point.

From the inspections, it was concluded that the quality of the welds was acceptable and representative of those that could be found on a typical bridge. 5.2 Test rig

The specimens were loaded in a reaction frame, as shown in Fig 31. The test rig was an adaptation of a design used by Cuninghame in tests on similar connections'1«*'.

The central crossbeam was bolted to an inverted 'T' section 'cross-girder' which, in turn, was bolted to the test bed. The detailing of the connection between the crossbeam and cross-girder (with respect to the overlap of the plates, the web thickness of the cross-girder and the bolt sizes and spacings) was identical to that detail on the test panel which was itself modelled on the Avonmouth bridge.

One end of the specimen was fixed and the free end was loaded by a hydraulic actuator through a steel plate. An adjustable steel strap was fixed between the test bed and a bracket bolted to the internal gusset plate. As the vertical load was applied horizontal forces were produced which gave bending in the crossbeam. The amount of bending could be altered by adjusting the angle of the strap. 5.3 Instrumentation

Four 45° rosette strain gauges were installed at the apex of the trough, on the trough and crossbeam plates, as shown in Figs 32a and 32b. The gauge positions on the specimens were identical to those on the panel at this location.

All gauges were positioned on the centreline of the trough with the centre of the middle element of the strain gauge 15mm from the root of the weld. 5.4 Loading arrangement

Data from the static load tests on the panel indicated that the likely fatigue failure mode would be a weld toe

10

failure, through the trough plate, at the apex of the trough. Maximum stress (-20.5 N/ram2) was recorded by gauge 49 at this location with the wheel positioned directly over the trough and 1220mm from the crossbeam. Corresponding stresses in the crossbeam were -9.7 N/mm2 (gauge 46) and -3.0 N/mm2 (gauge 52). By symmetry, a stress of -19.0 N/mm2 was calculated for the apex of the trough on the opposite side of the crossbeam to gauge 49 with the wheel in this position.

The fatigue specimens would be loaded so that the distribution of stress around the apex of the trough was the same as (or as close as possible to) that measured on the panel with the wheel load in this 'worst' position.

Strain gauges were balanced to zero with the specimen freely suspended from a crane at points close to the crossbeam. In this position the specimen balanced horizontally. Great care was needed when bolting the specimen to the centre and end supports to avoid producing very large strains at the gauge positions. Occasionally it was necessary to shim the end support. These problems were caused by the (unavoidable) distortions in the specimens caused by the welding process.

Adjustments were made to the load and the strap to obtain the desired (nominal) stress range at specimen gauge l and the correct proportion of stress at the other three gauge locations. Stresses at the four gauge positions at the minimum and maximum actuator loads are given in Table 3a. A compression negative sign convention has been adopted throughout this report for stresses and for downward acting actuator loads which produce compressive stresses in the apex of the trough. In practice it was not possible to achieve precisely the correct distribution of stress around the trough. Stress ranges actually achieved, compared to the 'target' panel stresses, are given in Table 3b. They have been normalised so that the stress at gauge 1 has a value of 1.0. 5.5 Test procedure

Each specimen was tested continuously. Visual inspections were made and strain gauge readings were recorded at approximately 8 hour intervals. The normal loading frequency was 6Hz during the early stages of the test, reducing to 3Hz when cracking was indicated. At the higher loading rate, ten million cycles could be achieved in about 3 weeks, assuming no interruptions. In practice trips, set to safeguard the specimens from overloads, occasionally operated resulting in lost time.

Strains from the centre element of each of the four strain gauges were displayed on a chart recorder, set to operate for a few seconds every half hour. Drift of the mean strain from gauge l was observed Just before cracks became visible. This served as an early indication of cracking at

- 11

which time visual inspection of the specimens became more rigorous.

Visual inspection consisted of an examination of the weld using a xlO magnifying glass. Normally cracks were spotted when they were between 10 and 15mm long. Each end of the crack was marked and crack growth subsequently charted. Accurate measurements of the distance between marks were made at the end of the test with the specimen removed from the rig. Dye penetrant tests were carried out on specimens which did not fail to confirm that cracking had not occurred. 5.6 Test results 5.6.1 Crack development

Of the eight specimens originally manufactured. one specimen (5B) was badly distorted and had to be rejected. Specimen IB suffered an accidental overload during set-up and was also rejected. Results for the remaining specimens are summarised in Table 4a.

Specimens 2B and 7B, which were tested at the lowest stresses (95 and 100 N/mm2 respectively) were uncracked after 11.7 and 13.2 million cycles respectively.

The remaining specimens (3B, 4B, 6B and 8B) all suffered weld toe failures through the trough plate as expected. Cracks initiated near the apex of the trough and grew around the webs of the trough. One specimen was sectioned as illustrated in Fig 33. It confirmed that the crack had penetrated through to the inside of the trough plate.

A second crack, in the toe of the weld at the crossbeam plate, also developed in specimen 6B.

The crack development is shown in Figs 34 to 37 for specimens 3B, 4B, 6B and 8B respectively. The figures show the length of crack and the number of cycles when the crack was first observed and at the end of the test. They also show the number of cycles to failure which, in this report, is defined as a crack 25mm long.

Fatigue failure may be defined in a number of ways, four possible approaches being:

(i) first visible crack - requires very frequent and detailed inspection in order that this state is spotted as early as possible.

(ii) crack penetration through parent plate - difficult to detect in closed section specimens.

(iii) total failure - not practical for some specimens or loading arrangements.

(iv) crack of specified length.

- 12

Criterion (iv) was adopted for this work. For specimen 3B the crack growth was rapid and there is

little difference in the number of cycles between a 25mm crack (0.9 million cycles) and a 100mm crack (1.1 million cycles). However, for specimen 4B crack growth slowed down after about 3 million cycles (34mm crack length). Defined failure at 25mm crack length was therefore a conservative approach. A 25mm long crack could be easily seen during the routine inspections. 5.6.2 Residual stresses

Strain gauge readings taken at the end of each test o"n the failed specimens give an indication of the original residual stress produced by the fabrication process. These residual stresses are (partly) relieved by the fatigue cracking and stress, of opposite sign to the initial residual stress, remains at the end of the test.

These 'end of test' strains are summarised in Table 5a. The sequence of welding (see Appendix 4) suggests that tensile residual stresses would be produced at the apex of the troughs. This was indicated in most cases though tensile strains were present in the trough on one side of the crossbeam in specimen 3B suggesting that compressive residual stresses were initially present in this case.

Residual strains were also measured on the rejected specimen 5B. Strain gauges were balanced to zero before a sawcut was progressively made through the trough at the toe of the weld. The results are given in Table 5b. With the sawcut extending approximately 100mm either side of the apex of the trough, large tensile strains were recorded by the gauge adjacent to the sawcut with larger compressive stresses on the opposite side of the crossbeam. Further sawcutting to completely remove a section of trough containing gauge 1 did not result in any further significant change in recorded stress. These results suggest that compressive residual stresses were present on the sawcut side of the crossbeam with larger tensile residual stresses on the other side of the crossbeam, a result similar to that obtained in specimen 3B.

As noted earlier, the onset of cracking was usually preceded by a change in the mean stress at gauge 1. The behaviour of the readings from this gauge varied somewhat between the four specimens but some common features are shown in Fig 38. During the initial change in mean stress the stress range remained constant. There then followed a period of rapid change of both mean stress and stress range as residual stresses were relieved and the stress in the trough due to the applied load was reduced by the presence of the crack. Finally the stress range was reduced to zero and the mean stress indicated the magnitude (but opposite sign) of the relieved residual stress.

- 13

5.6.3 Weld classification: S-N curves In view of the small number of test reults obtained, the

data have been compared with established S-N design curves. Consideration has been given to both Eurocode«1β> and British Standard (BS5400'"**' ) design curves though the data are only presented graphically in this report against the former.

Eurocode classifications 50, 80 and 125 and BS5400 classifications G, E and C are considered in this report. The corresponding classifications (50, G), (80, E) and (125, C) are similar and may be regarded as equivalents. The mean minus two standard deviation curves are shown in Fig 39. A constant amplitude non-propagating stress range (σ0) is assumed in both codes, at 5 χ 10* cycles in Eurocode and at IO-7 cycles in BS5400. In addition, stresses below 0.55σο (corresponding to 10e cycles) are assumed to be non-damaging in Eurocode. For classes other than C, the S-N curves change slope from m=3 to m=5 at σο. For class C, the corresponding slopes are m=3.5 and m=5.5.

The results of the tests on the type Β specimens are summarised in Table 4a and shown in Fig 40. Data are for the stress at gauge l and the cycles for a 25mm crack in the trough plate at the weld toe. The four failures are close to the mean line of Eurocode class 80 and well within the 95% confidence limits. Specimens 2B and 7B which did not fail are shown as 'run-outs' in the figure. 5.6.4 Calculated fatigue life

From the data obtained in the panel tests, histograms of stress ranges were obtained for gauge 49 (on the trough plate at the apex of the trough) for the vehicle loading shown in Table 1. The histogram produced with the centreline of the distribution of vehicles centred directly over the trough is given in Table 6. Using the equations of the mean minus two standard deviation S-N curves for Eurocode class 80 and BS5400 class E, fatigue lives were calculated for a 2.3% probability of failure. The Palmgren-Miner method of damage summation was used and failure assumed at E(n/N) = 1. The lives are 12.7 years (Eurocode) and 10.8 years (BS5400) for one million HGVs. 6. CONSTANT AMPLITUDE FATIGUE TESTS ON TYPE 'A' CONNECTION 6.1 Fabrication of test specimens

The manufacture of these specimens was similar to that already described for the type 'B' specimens. The fabrication drawings of the specimen are given in Appendix 3, the welding details in Appendix 4 and the inspection report in Appendix 5 Weld sizes of the specimens tested are given in Appendix 6.

14

6.2 Test rig This was identical to that used for the type 'B'

specimens and shown in Fig 31. 6.3 Instrumentation

Analysis of the static load test data from the deck panel suggested that weld toe failure through the trough plate at the bottom of the weld would occur. Strain gauges were installed to determine the stress at this point in the test specimens. Other gauges were installed around the bottom of the weld, on both sides of the trough, to ensure that the distribution of stress in this area was similar to that measured on the panel. After fatigue cracks unexpectedly occurred at the top of the welds in the first specimen tested (specimen 1A), additional gauges were installed at the top of the weld on subsequent specimens. Strain gauge positions are shown in Figs 4la and 41b.

An alternative strain gauge layout was adopted for the last specimen tested (specimen 8A, see Fig 42) corresponding to gauges installed on a bridge to record stresses under traffic loading. The bridge tests are discussed in section 7. 6.4 Loading arrangement

Similar procedures to those discussed in section 5.4 for the type 'B' specimens were adopted. Stresses recorded by the gauges at the minimum and maximum actuator loads are given in Table 7a. Normalised stress ranges are compared with the 'target' panel stresses in Table 7b. The stress at the top of the weld in specimen 1A was estimated from measurements made during the setting up of specimen 2A (gauges had not been installed in this location on the first specimen). 6.5 Test procedure

This was similar to that described in Section 5.5 above. The loading frequency for the type 'A' specimens was 3Hz throughout. 6.6 Test results 6.6.1 Crack development

Crack development is shown in Figs 43 to 47 for specimens 1A, 2A, 3A, 5A and 8A respectively. Two sets of results were obtained from each specimen, on the northside and southside of the trough at end 2 (see, for example, Fig 43a). End of test cracks are illustrated in Fig 43a and crack growth is charted in Fig 43b. Crack length and cycles are tabulated for the three stages; crack first observed, defined failure and end of test. As with the type 'B' specimens, failure is defined as a crack 25mm long. In specimen 5A, the growth of crack c was extrapolated to determine the number of cycles to failure (25mm), the test having been halted at 2.15

15

million cycles when these cracks were only 18mm and 20mm long, crack a (northside) having by this time grown almost the full length of the weld.

Four different cracks developed, though not all in the same specimen:

Crack a was a weld toe failure through the trough plate at the bottom of the weld. This crack was expected from the panel test results. It occurred in all but the most lightly loaded specimen (1A) but not always on both sides of the trough. In all cases the crack initiated within 25mm of the bottom end of the weld.

Crack b was a weld end failure through the crossbeam plate at the bottom of the weld. It occurred in a total of 5 locations in specimens 2A, 3A and 5A.

Crack c was a weld end failure through the trough plate at the top of the weld. It occurred in all 5 specimens, on both sides of the trough and on both sides of the crossbeam. In all cases cracks initiated at very low endurances.

Crack d was a weld end failure through the crossbeam plate at the top of the weld. It occurred in specimens IA, 2A and 8A but at much longer endurances than the corresponding crack c. For this reason it is regarded as a secondary crack and has not been assessed in the following classifications. 6.6.2 Weld classification: S-N curves

The test results are summarised in Table 8. The stress at gauge 1 or 5 has been used to represent

the fatigue stress at crack a (Northside or Southside respectively).

Measurements on both the deck panel and the fatigue specimens show that the stress at the top of the weld is similar at gauge positions 9 and 10 (or 11 and 12). Fatigue specimen gauges 9 and 11 (corresponding to panel gauge 88) have been used to represent the stress at crack c.

Gauges 13 and 15 on specimen 8A (corresponding to panel gauge 90) has been used to represent the stress at crack b, estimated values being used for the first four specimens. It is accepted that these estimated stresses are only an indication of the actual stress and will be treated with due caution.

The results are shown in Fig 48 against Eurocode S-N design curves.

It is concluded that Eurocode class 50 or BS5400 class G is appropriate for the weld end failure through the trough plate at the top of the weld (crack c). For the weld toe failure through the trough plate at the bottom of the weld

- 16

(crack a), the appropriate classes are class 125 (Eurocode) or class C (BS5400).

A high classification is indicated for crack b from the estimated stresses. 6.6.3 Calculated fatigue lives

Histograms of stress ranges were obtained for panel gauges 13, 88 and 90 from the static load test data and the vehicle loading of Table 1. These are shown in Table 6 for the case of the centreline of the vehicle distribution directly over the trough.

Using mean minus two standard deviation S-N data for Eurocode class 50 and the histogram for gauge 88, the calculated fatigue life for weld end failure at the top of the weld (crack c) is 4.8 years (4.3 years for BS5400 class G). This life is for one million vehicles per year.

For the weld toe failure (crack a), lives very much greater than 120 years (the design life for UK bridges) are calculated using either the Eurocode or BS5400 classifications (125 or C) and the histogram for gauge 13.

The maximum stress range, calculated for gauge 90 from the panel tests, is not more than 40 N/mm2 (see Table 6). Given the high classification indicated in Fig 48, and allowing for a reasonable margin of error in the estimated stresses, fatigue cracks are unlikely to develop at this point in bridges in service.

7. MEASUREMENTS ON A BRIDGE UNDER TRAFFIC LOADING The calculated fatigue life of less than 5 years for the

failure at the top of the weld was sufficiently short to be of immediate concern for existing bridges in service. It was decided to carry out a limited inspection of a heavily trafficked bridge to see if cracking had occurred at this point and to measure strains under traffic loading for a short period. Being close to the deck plate, surfacing might be expected to reduce the stresses in this area. A limited strain measurement trial would help to establish the surfacing factor. 7.1 Bridge inspection

The most suitable bridge for the trials had trough to crossbeam connections of the type 'C' design. Other bridges with type 'A' connections were lightly trafficked. Weld classifications established for the type 'A' connection cannot with certainty be applied to the connections on the bridge but measurements on the deck panel suggest that stresses around the connections are similar.

Visual inspection was made of five trough to crossbeam connections under the slow lane wheel tracks. Both sides of

17

the trough and crossbeam were inspected, giving twenty potential crack locations. No cracks were found.

This was a very small sample and the method of inspection less than rigorous. It was agreed that ultrasonic inspection of this detail would be included in future routine inspections of the bridge. 7.2 Strain measurements 7.2.1 Instrumentation

Strain gauges were installed in similar positions to those on the type 'A' fatigue specimens (Fig 49). The number of channels available on the recording equipment was limited and so single element strain gauges were used throughout. The instrumentation was installed around one trough, calculated to be directly under the slow lane nearside wheel track. The temperature of the steel deck plate was measured by a single thermocouple.

Recording equipment was installed to monitor the gauges and store the strain ranges produced by the passing vehicles. Vehicle flow was not recorded. Data were recorded continuously for a two week period in November 1988. 7.2.2 Results

Stress ranges for the eight strain gauges are given in Table 9. The data suggest that the centreline of the distribution of vehicles is slightly biased towards the side of the trough containing gauges 5-7. Data from these gauges are therefore used in the following analyses.

Data from the gauges on the bridge may be compared with data from gauges in similar locations on the deck panel with the vehicle loading of Table 1. The period of recording on the bridge was short and it cannot therefore be assumed that the full spectrum (types and relative frequencies) of vehicles represented in Table 1 crossed the bridge during this time. It is therefore not appropriate to compare the full spectra of Tables 6 and 9. Instead, the maximum stress ranges only are compared on the assumption that at least one vehicle of the most damaging type will have crossed the bridge during the recording period.

The comparisons are summarised below: (i) Stress at top of weld

Panel gauge 88 - maximum stress range 80 N/mm2 Bridge gauge 8 - maximum stress range 40 N/ram2

(ii) Stress in trough at bottom of weld Panel gauge 13 - maximum stress range 70 N/mm2 Bridge gauge 7 - maximum stress range 48 N/mm2

18

(iii) Stress in crossbeam at bottom of weld Panel gauge 90 - maximum stress range 40 N/mm2 Bridge gauge 5/6 - maximum stress range 56 N/mm2

The results confirm that the surfacing effect is greatest for the gauges close to the deck plate. There is no obvious explanation for the stresses in the crossbeam at the bottom of the weld being higher in the surfaced deck than in the unsurfaced deck panel.

Using the stress ranges from the tests on the bridge and the weld classification established in the fatigue tests, the fatigue life for cracking at the top of the weld is calculated to be 280 years for Eurocode class 50 and 106 years for BS5400 class G. The lives are for a 2.3% probability of failure and the calculation assumes that the vehicle flow across the bridge during the two week recording period was completely representative of the flow throughout the year.

The temperature of the bridge deck during the recording period ranged from -1.6° to 18.0°C with a mean of 7.6°C. This is below the annual mean temperature and the surfacing effect will therefore, on average, be less than that indicated above.

The large difference in the calculated lives from the British and European codes can be explained by the fact that a large proportion of the stress ranges are below the constant amplitude stress range limit σσ. This point is different in the two codes. A life of 49 years is obtained if a slope of m=3 is assumed for all stresses. 7.3 Conclusions for the bridge

From the stress ranges obtained from the unsurfaced deck panel and the weld classifications established in the fatigue tests, calculated fatigue lives in excess of 120 years (the design life of the bridge) were obtained for failure at the bottom of the weld. Any surfacing effect may therefore be regarded as an added bonus.

There appears to be a considerable 'surfacing factor' at the top of the weld. Although the data from the deck panel and the bridge are not directly comparable, there is at least an order of magnitude difference between the calculated fatigue lives of the surfaced and unsurfaced decks. This is not inconsistent with the increases in fatigue life from surfacing obtained for other connections near the deck'1*'.

For the bridge on which measurements were made, there is a 2.3% probability of fatigue failures occurring at the top of the weld within 106 years (using the British Standard weld classification). It is therefore concluded that failures in service are unlikely if current levels of traffic flow are maintained. The bridge is believed to be carrying around half a million HGVs per year. However, if it were carrying the

19

full BS5400 traffic flow of two million HGVs per year, the corresponding life would be reduced to only 26.5 years, well below the required design life. 8. VARIABLE AMPLITUDE FATIGUE TESTS ON TYPE 'B* CONNECTION 8.1 Derivation of loading spectrum

Measured strains from the static load tests on the deck panel for strain gauge 49 (on the apex of the trough, adjacent to the weld) provided the basic data for the spectrum. Highest stresses were measured at this position during the static testing. Fatigue tests at constant amplitude were also been related to the stress at this location. A stress spectrum was calculated from these data for the simulated traffic flow of the Rheden Bridge using the computer assessment program of the University of Liege'1". It was agreed that all participants in this research would use this common approach in order that the test results would have a basis for comparison. The spectrum is shown in Table 10.

Using the parameters of the mean-line S-N curve for Eurocode Class 80 (the classification established in the constant amplitude tests), the stress spectrum of Table 1 was assessed. The Palmgren-Miner cumulative damage method was used and endurances calculated assuming failure to occur when E(n4/Ni) = 1.

Number of cycles Ση± = 85212 Cumulative damage £(n±/Ni) = 0.628 χ IO-3

Endurance = {1/£(η*/Νι)} χ Ση±

= rni/£(nt/Nt) = 135.6 χ 10e· cycles

An endurance of 135.6 χ 10e cycles is calculated for a 50% probability of failure. Assuming a loading frequency of around 6 Hz, failure for a specimen achieving the mean-life would take around 262 days. This is an unacceptably long duration and it is therefore necessary to modify the spectrum to achieve a more practical mean-life.

Increasing stresses throughout by a factor of 1.5 would reduce the endurance by approximately 1/1.5s* (ignoring the change of slope of the S-N curve at low stresses), that is, to about 40 χ 10* cycles or 78 days. Such a spectrum, recompiled into 10 N/mm2 classes, is given in Table 11. For the mean-line S-N curve of Eurocode Class 80, this spectrum is assessed as follows.-

20

Number of cycles En» = 85212

Cumulative damage E(ni/Nâ) = 2.672 χ IO-3

Endurance = Enâ/E(n»/Nâ) = 31.9 χ 10e cycles

Further analysis of this spectrum is shown in Fig 50. Calculations show that stresses below 50 N/mm

2 contribute

only 2.1% to the total damage of the spectrum while making up 88.3% of the total number of cycles. A practical spectrum in which stresses below 50 N/mm

2 are deleted, is given in Table

12.

Three values of cumulative damage for this spectrum

have been calculated:

(i) assuming a change in slope of the mean-line S-N curve at 5 χ 10«* cycles - Eurocode Class 80

E(ni/Ni) = 26.166 χ 10"3

(ii) assuming a change in slope of the mean-line S-N

curve at 10"" cycles - BS5400 Class E

E(n»/N») = 27.451 χ 10~3

(iii) assuming a slope of m=3 of the mean-line S-N curve

for all stresses

E(n»/N») = 28.040 χ 10"3

The three values are similar, differing by around seven per cent. In the latter case (iii), the mean-life endurance is:

Εηι/Γ(ηι/Νι) = 99820 / (28.040 χ 10"3)= 3.6 χ 10* cycles

This represents approximately 7 days testing at 6Hz. If the Palmgren-Miner assumption proves to be conservative, longer endurances would be expected.

An equivalent stress range oe of a spectrum can be defined as follows:

σβ = [(l/En.) χ Είη^σι3)]1'3

The equivalent stress range is that stress range which, if applied to all cycles in the spectrum, would give the same calculated damage as the cumulative sum of the damage caused by the individual stresses and their corresponding numbers of cycles. For the spectrum of Table 12, σβ = 97.4 N/mm2. 8.2 Test specimens, instrumentation and test rig

These were identical to those used for the tests on the Type 'B' specimens at constant amplitude and described in the respective parts of Section 5. The reports on the inspection

21 -

of the specimens and on the weld sizes are given in Appendices 5 and 6. 8.3 Loading arrangement

The same procedure was used for setting up the specimens as that described for the tests at constant amplitude. Stresses, at the minimum and maximum actuator loads, are given in Table 13a and normalised stress ranges, compared with the target panel stresses, in Table 13b.

After each specimen was set up in the test rig, a static load test was carried out to determine the relationship between the applied load and the stress at gauge 1. From these data, the loads required to produce stress ranges at 10 N/mm2 intervals between 55 and 245 N/mm2 (from Table 12), from a constant small pre-load of -2kN, were calculated. 8.4 Test procedure

The hydraulic actuator used to load the specimens was driven by a controller programmed to apply the spectrum of stress ranges given in Table 12. Stress ranges were selected at random until all the cycles in the 99820 cycle block had been applied; the block was then repeated. A typical example of the load signal, and the corresponding output from the middle element of strain gauges 1 and 2, is shown on the chart record in Fig 51. As before, drift of the mean strain from gauge 1 was observed just before cracks became visible.

The loading rate was constant, consequently the cycling rate depended on the magnitude of the applied stress range. On average it was about 4.5Hz. 8.5 Test results 8.5.1 Crack development

Crack development in the specimens tested at variable amplitude was similar to that for the specimens tested at constant amplitude. Each of the four specimens cracked through the trough plate at the toe of the weld, with the crack initiation point close to the apex of the trough. In one case, specimen 14B, a second crack developed at the weld toe in the crossbeam plate. Crack development is shown in Figs 52 to 55. 8.5.2 Fatigue endurance

The test results are summarised in Table 4b (for specimen 14B, the endurance for the crack in the trough plate is given). The mean endurance of the four specimens is 1.88 χ 10* cycles.

The results are shown on an S-N curve (Fig 56), together with the results from the constant amplitude tests on similar specimens. The variable amplitude test results are plotted at

22

the equivalent stress range level. At this stress range, the S-N curves suggest a 50 per cent probability of failure (a mean endurance) of around 3.6 κ 10* cycles. The actual endurances under variable amplitude loading are therefore much lower than expected. None of the four specimens achieved the expected endurance.

In the Palmgren-Miner cumulative damage summation method, failure is assumed to occur when the damage sum E(n»/N») equals unity. Three values of £(n»/N») were calculated for the applied spectrum of Table 12. Given that the mean endurance of 1.88 χ 10* cycles represents 18.83 repetitions (1880000/99820) of the spectrum, the corresponding values of E(n«./N4) to failure are:

(i) For mean-line S-N curve of Eurocode Class 80 E(n*/N») = 18.83 χ (26.166 χ IO"3) = 0.493

(ii) For mean-line S-N curve of BS5400 Class E Ε(ηι/Νι) = 18.83 χ (27.451 χ IO"3) = 0.517

(iii) For mean-line S-N curve with m=3 for all stresses E(n»/N») = 18.83 χ (28.040 χ IO"3) = 0.528

These values represent about half the Palmgren-Miner predicted value.

9. SUMMARY OF TEST RESULTS AND FURTHER DISCUSSION This research set out to rigorously examine the

performance of types of trough to crossbeam connection used in bridges today. Early designs of connection, with the trough butted up to the crossbeam, had been shown elsewhere to have a poor fatigue performance. Failures of this type of connection have occurred in service and the design is no longer used. Two more recent designs have been examined here, both with the trough passing through cut-outs in the crossbeam.

The type 'B' connection was welded all round. Influence surfaces of stress were obtained for a number of points around the connection from static load tests on a full-scale deck panel. Highest stresses were recorded in the trough plate at the apex of the trough.

Full-scale specimens were tested under constant amplitude loading to determine the weld classification. Fatigue cracks developed through the trough plate at the toe of the weld, as expected, with endurances consistent with weld class 80 (Eurocode) and class E (BS5400). Under BS5400 traffic loading centred directly over the trough (the worst case condition), fatigue lives of between 11 and 13 years (depending on the code used) were calculated. The calculations were based on a 2.3% probability of failure and

23 -

a vehicle flow of one million HGVs per year. The calculations also used the Palmgren-Miner method of damage summation and assumed failure at E(n»/N») = 1.0.

Tests on identical specimens at variable amplitude produced similar failures. The loading spectrum was based on stress data from the static load tests and the traffic flow recorded on the Rheden bridge. The spectrum was modified to increase the damage factor in order that fatigue failures would be produced in a practical timescale. The mean endurance from the four specimens tested was approximately half that predicted by Miner.

Other work to assess the effect of the bridge deck surfacing on trough to crossbeam connections of the early design suggested that a life improvement factor of 3 could be applied to the calculated life for the unsurfaced deck. In both designs failures occur at the apex of the trough and the surfacing factor applies to stresses at this point. It is therefore reasonable to assume that a similar surfacing factor could be applied to the Type 'B' design considered here.

The implication for bridges in service with this type of connection is that there is a 2.3% probability of fatigue cracks developing in this connection within 18 years assuming :

(i) weld class 80 (Eurocode) or class E (BS5400) (ii) a life improvement factor of 3 from the surfacing (iii) failure at E(n±/N*) = 0.5 (iv) a traffic flow of one million HGVs per year (v) the centreline of the distribution of vehicles

centred directly over the trough (vi) vehicles of the type and mix given in Table 1

The design clearly fails to meet the 120 year fatigue life requirement of the British Standard.

The type 'A' connection differed from the type 'B' connection by having cut-outs in the crossbeam at the apex of the trough and around the trough to deck plate weld. Thus there was no welded connection between the trough and the crossbeam at the apex of the trough where the type 'B' specimens failed.

Analysis of the static stresses from the load tests on the deck panel initially suggested that failure would occur through the trough plate at the weld toe, initiating at the bottom of the weld and progressing up the side of the trough. These cracks did occur, but only after cracks had developed in the trough plate at the top of the weld at the weld end. Further measurements on the deck panel confirmed that high stresses were present at this point. Cracks also developed in the crossbeam plate at both ends of the weld.

- 24

Tests at constant amplitude on specimens representing the type 'A' connection suggested that the cracks in the trough plate at the top end of the weld be classified Eurocode class 50 or BS5400 class G. The weld toe failure through the trough plate at the bottom of the weld was consistent with weld class 125 (Eurocode) or class C (BS5400).

A fatigue assessment of the connection suggested that failures at the bottom of the weld, either in the trough plate or in the crossbeam, were unlikely to occur for well in excess of 120 years, the design life of UK bridges. However, a fatigue life of less than 5 years was calculated for the failure in the trough plate at the top of the weld.

Measurements under traffic loading on a bridge surfaced with 38mm of mastic asphalt showed that the maximum stresses at the top of the weld were only half of those calculated from the data from the tests on the unsurfaced deck panel. Insufficient data were obtained from this short trial for a general reassessment of this detail under surfacing conditions. In particular, further measurements would be necessary at a range of surfacing temperatures. However, there appears to be a significant reduction in stress at this point from the surfacing which would be consistent with measurements on other connections close to the deck plate.

For the bridge on which the measurements were made, calculations suggest that cracking is unlikely to occur in less than 100 years. However, it is known that the traffic flow across the bridge is only about half a million HGVs per year, about one quarter of the design requirements of BS5400. Furthermore, measurements were made over a two week period in November when the mean surfacing temperature was 7.6eC. This is lower than the annual mean temperature; consequently, the 'average' surfacing factor would be less than that assumed in these calculations. While fatigue cracking at this point may not present a problem for this structure, it is clearly a detail to be avoided in future designs.

The type 'C' connection is similar in design to the type 'A' connection, differing mainly in the shape of the cut-outs around the apex of the trough and around the trough to deck plate welds. Results from the static load tests were similar though higher stresses were recorded in connection 'C* which consequently had an inferior provisional fatigue assessment. Fatigue tests were not carried out on type 'C' specimens but it is likely that cracking would have developed in the same way and weld classifications would have been similar to those of the type 'A' connection. The same general conclusions are therefore likely to apply to the type 'C' connection.

The results of the tests suggest that an 'optimised' design might be one which incorporated the cut-out in the crossbeam at the apex of the trough but not the cut-outs around the trough to deck plate welds. Further testing would be needed to assess this design. In particular, the following

2^

points would need to be investigated: (i) cracks initiating at the top of the trough where the

three welds (trough to deck plate, crossbeam to deck plate and trough to crossbeam) meet. This is not a good design detail. Cracks did not develop here during the fatigue tests on the type 'B' specimens but the loading was not arranged to produce highest stresses at this point.

(ii) cracks initiating at the weld end, through the trough plate, at the bottom of the weld. Cracks did not occur at this point in the tests on the type 'A' specimens but, again, the loading was not arranged to produce highest stresses at this point. A poor classification was obtained for the weld end failure at the top of the weld. If a similar classification were obtained at the bottom of the weld, this could become the critical fatigue point.

(iii) fatigue tests at variable amplitude to establish whether the Palmgren-Miner method of damage summation is appropriate for this design.

10. CONCLUSIONS 1. Fatigue failures occurred in specimens representing

the type 'B' connection (trough passing through crossbeam and welded all round) under constant amplitude loading. Cracks developed at the weld toe, through the trough plate at the apex of the trough, consistent with classifications 80 (Eurocode) and E (BS5400).

2. Fatigue lives were calculated using data from static load tests on an unsurfaced deck panel. For the vehicle loading of BS5400 (part 10) centred directly over the trough (the worst case condition), there is a 2.3% probability of failure within 11 to 13 years (depending on the code used). The calculation is based on a vehicle flow of one million heavy goods vehicles per year and uses the Palmgren-Miner method of damage summation.

3. Tests on identical specimens at variable amplitude suggest that calculations based on Palmgren-Miner (failure occurring at E(n/N) = 1) are optimistic. The mean endurance of four specimens tested using a loading spectrum based on traffic measurements on the Rheden bridge was approximately half that predicted by Palmgren-Miner.

4. Other work on a similar connection suggests that fatigue lives may be increased by the bridge deck surfacing. A factor of three has been suggested for surfacings such as 38mm of mastic asphalt.

5. The type 'B' connection fails to meet the 120 year design life required for UK bridges when assessed by the BS5400 code of practice for fatigue.

26

6. Fatigue failures occurred under constant amplitude loading on specimens representing the type 'A' connection. This connection was similar to the type 'B' connection but with cut-outs in the crossbeam at the apex of the trough and around the trough to deck plate welds. Cracks developed at the top and bottom of the weld.

7. Weld end failure through the trough plate at the top of the weld was consistent with classifications 50 (Eurocode) and G (BS5400). A fatigue life of less than 5 years was calculated using the same methods and assumptions as described in 2 above.

8. Failure at the toe of the weld, through the trough plate at the bottom of the weld, was consistent with classifications 125 (Eurocode) and C (BS5400). Calculated fatigue lives were well in excess of 120 years.

9. Measurements under traffic loading on a bridge surfaced with 38mm of mastic asphalt suggested that stresses at the top of the weld were considerably reduced by the composite action of the surfacing. Insufficient data were available for a general reassessment of the connection under surfaced conditions. For the bridge on which the measurements were made, cracks are unlikely to develop for 100 years. However, traffic flow on this bridge is considerably below normal design requirements.

10. The detailing at the top of the weld in the type 'A' connection is not recommended in future designs.

11. The type 'C' connection was similar in design to type 'A'. Results from the static load tests were similar though higher stresses were recorded in the type 'C' connection. Fatigue tests were not carried out on type 'C' specimens but it is thought that the results would have been similar to those obtained for the type 'A' connection. The same general conclusions may be applied to both the types 'A' and 'C' connections.

12. An 'optimised' design with a cut-out in the crossbeam at the apex of the trough but not around the trough to deck plate weld has been suggested. Fatigue tests, at constant and variable amplitude, on specimens representing such a connection would be required to assess its performance.

11. REFERENCES 1. PAGE, J. 1979. Measurement and interpretation of dynamic loads in bridges: Phase 1. Final report on contract no. 7210-SA/8/809. Crowthorne: Transport and Road Research Laboratory.

27 -

2. PAGE, J. 1983. Measurement and interpretation of dynamic loads in bridges: Phase 2. Final report on contract no. 7210-KD/804. Crowthorne: Transport and Road Research Laboratory. 3. STREAMS, B. 1987. Prediction of future incidence of fatigue cracking from observed rates on a structure. International conference on fatigue of welded constructions. Paper 36. The Welding Institute. Brighton UK, April 1987. 4. CUNINGHAME, J R. 1987. Strengthening fatigue prone details in a steel bridge deck. International conference on fatigue of welded constructions. Paper 38. The Welding Institute. Brighton UK, April 1987. 5. NUNN, D E. 1974. An investigation into the fatigue of welds in an experimental orthotropic bridge deck panel. TRRL Laboratory Report 629. Crowthorne: Transport and Road Research Laboratory. 6. BSI. 1972. Steel girder bridges. Part 3B, Stresses. BS153: 1972. London: British Standards Institution. 7. BSI. 1980. Steel, concrete and composite bridges. Part 10 Code of practice for fatigue. BS5400: 1980. London: British Standards Institution. 8. NUNN, D E and J R CUNINGHAME. 1974. Stresses under wheel loading in a steel orthotropic deck with trapezoidal stiffeners. TRRL Supplementary Report 53UC. Crowthorne: Transport and Road Research Laboratory. 9. NUNN, D E and J R CUNINGHAME. 1974. Stresses under wheel loading in a steel orthotropic deck with V-stiffeners. TRRL Supplementary Report 59UC. Crowthorne: Transport and Road Research Laboratory. 10. MORRIS, S A H . 1976. Stresses under dynamic wheel loading in a surfaced steel orthotropic deck with V-stiffeners. TRRL Supplementary Report 237. Crowthorne: Transport and Road Research Laboratory. 11. NUNN, D E and S A H MORRIS. 1974. Trials of experimental orthotropic bridge deck panels under traffic loading. TRRL Laboratory Report 627. Crowthorne: Transport and Road Research Laboratory. 12. MORRIS, S A H and H HOWELLS. 1974. Derivation of stress spectra from measurements on orthotropic bridge decks during normal trafficking. TRRL Supplementary Report 44UC. Crowthorne: Transport and Road Research Laboratory. 13. CUNINGHAME, J R. 1982. Steel bridge decks: fatigue performance of joints between longitudinal stiffeners. TRRL Laboratory Report 1066. Crowthorne: Transport and Road Research Laboratory.

28 -

IA. BEALES, C. 1985. Assessment of the fatigue lives of the welds in the decks of the Severn and Wye bridges. TRRL Working Paper WP/B/101/85. Crowthorne: Transport and Road Research Laboratory. (Unpublished paper available on direct personal application only). 15. BEALES, C. 1988. The effect of cracks in the asphalt on stresses in an orthotropic steel bridge deck. TRRL Working Paper WP/B/155/88. Crowthorne: Transport and Road Research Laboratory. (Unpublished paper available on direct personal application only). 16. CUNINGHAME, J R. 1978. Interim report on strengthening of stiffener to crossbeam joints in steel decks. TRRL Working Paper WP/BD 36. Crowthorne: Transport and Road Research Laboratory. (Unpublished paper available on direct personal application only). 17. TAPPIN, R G R and Ρ J CLARK. 1985. Jindo and Dolsan bridges: design. Proceedings of the Institute of Civil Engineers. Vol 78, Part 1, pp 1281-1300. December 1985. 18. COMMISSION OF THE EUROPEAN COMMUNITIES. 1984. Eurocode no.3 : Common unified rules for steel structures. 19. BRULS, A. 1987. Measurement and interpretation of dynamic loads in bridges: Phase 3. Progress report no.2: Sept 1987. University of Liege.

12. ACKNOWLEDGEMENTS The work described in this report was carried out in the

Bridges Division (Head of Division Dr K A Gallagher) of the Structures Group of the Transport and Road Research Laboratory. The author would like to thank Mr Ball for his practical assistance and Mr Cuninghame for his helpful advice.

- 29 -

A P P E N D I C E S

APPENDIX 1

UK bridfies with orthotropic steel decks

Forth Severn Wye Erskine Avonmouth Cleddau Humber Kessock

33

UK BRIDGES WITH ORTHOTROPIC STEEL DECKS

BRIDGE NAME Forth DATE OF COMPLETION 1964 BRIDGE TYPE Suspended truss MAIN SPAN LENGTH 1006 metres SURFACING WATERPROOFING MEMBRANE Rubber bitumen

WEARING COURSE Hand laid mastic asphalt NOMINAL THICKNESS 38mm

DECK PLATE THICKNESS 12.7mm LONGITUDINAL STIFFENERS

TRANSVERSE STIFFENERS

SHAPE Trapezoidal PLATE THICKNESS 6.4mm WIDTH 343mm DEPTH 191mm SPACING (BETWEEN CENTRE LINES) PLATE THICKNESS 9.5mm DEPTH 479mm SPACING 2997mm

686mm

LONGITUDINAL TO TRANSVERSE STIFFENER CONNECTION Crossbeam continuous. Troughs butted up to crossbeam and fillet welded all round with 6.5mm continuous welds.

343

203

479

NOTE a l l dimensions in œ i l l i e e t r e s

9-5

102

- 34


BRIDGE NAME Severn DATE OF COMPLETION 1966 BRIDGE TYPE Suspension MAIN SPAN LENGTH 988 metres SURFACING WATERPROOFING MEMBRANE Rubber bitumen


DECK PLATE THICKNESS 11.4mm

LONGITUDINAL STIFFENERS SHAPE Trapezoidal PLATE THICKNESS 6.4mm WIDTH 305mm DEPTH 229mm SPACING (BETWEEN CENTRE LINES) 610mm

TRANSVERSE STIFFENERS PLATE THICKNESS 6.4mm DEPTH 381mm SPACING 4572mm


114 305

381

102

6 4

\-m p. 89

NOTE all dimensions in millimetres

35


BRIDGE NAME Wye DATE OF COMPLETION 1966 BRIDGE TYPE Cable stayed MAIN SPAN LENGTH 235 metres SURFACING WATERPROOFING MEMBRANE Rubber bitumen


DECK PLATE THICKNESS 11.4mm LONGITUPINAL STIFFENERS


SHAPE Trapezoidal PLATE THICKNESS 6.4mm WIDTH 305mm DEPTH 229mm SPACING (BETWEEN CENTRE LINES) PLATE THICKNESS 8mm DEPTH 394mm SPACING 4267mm

610mm


394

102

NOTE all dimensions in millimètres

75

- 36


BRIDGE NAME Erskine

DATE OF COMPLETION 1971

BRIDGE TYPE Cable stayed

MAIN SPAN LENGTH 305 metres

SURFACING WATERPROOFING MEMBRANE Rubber bitumen WEARING COURSE Hand laid mastic asphalt NOMINAL THICKNESS 38mm

DECK PLATE THICKNESS 12.7mm

LONGITUDINAL STIFFENERS


SHAPE 'V' PLATE THICKNESS 6.4mm WIDTH 307mm DEPTH 262mm SPACING (BETWEEN CENTRE LINES)

PLATE THICKNESS 9.5mm DEPTH 381mm SPACING 4267mm

610mm

LONGITUDINAL TO TRANSVERSE STIFFENER CONNECTION

Trough continuous through crossbeam. 6.4mm fillet welds both sides of crossbeam plate.

12-7 307

381

9-5


- 37 -


BRIDGE NAME Avonmouth DATE OF COMPLETION 1974 BRIDGE TYPE Viaduct MAIN SPAN LENGTH 174 metres SURFACING WATERPROOFING MEMBRANE Rubber bitumen


DECK PLATE THICKNESS 12.7mm LONGITUDINAL STIFFENERS SHAPE 'V'

PLATE THICKNESS 6.4mm WIDTH 305mm DEPTH 257mm SPACING (BETWEEN CENTRE LINES) 572mm

TRANSVERSE STIFFENERS PLATE THICKNESS 14mm DEPTH 381mm SPACING 3658mm

LONGITUDINAL TO TRANSVERSE STIFFENER CONNECTION Trough continuous through crossbeam. 6.5mm fillet welds both sides of crossbeam plate.

257

381

14 /


38


BRIDGE NAME Cleddau DATE OF COMPLETION 1975 BRIDGE TYPE Viaduct MAIN SPAN LENGTH 213 metres SURFACING WATERPROOFING MEMBRANE Rubber bitumen


DECK PLATE THICKNESS 12.7mm LONGITUDINAL STIFFENERS SHAPE 'V'

PLATE THICKNESS 6.4mm WIDTH 305mm DEPTH 262mm SPACING (BETWEEN CENTRE LINES) 610mm

TRANSVERSE STIFFENERS · PLATE THICKNESS 9.5mm DEPTH 381mm SPACING 4267mm

LONGITUDINAL TO TRANSVERSE STIFFENER CONNECTION Trough continuous through crossbeam. 6.5mm fillet welds both sides of crossbeam plate.

305

381

/ 9-5


- 39


BRIDGE NAME Humber DATE OF COMPLETION 1981 BRIDGE TYPE Suspension MAIN SPAN LENGTH 1410 metres SURFACING WATERPROOFING MEMBRANE Rubber bitumen

WEARING COURSE Machine laid mastic asphalt NOMINAL THICKNESS 38mm

DECK PLATE THICKNESS 12mm LONGITUDINAL STIFFENERS SHAPE 'V'

PLATE THICKNESS 6mm WIDTH 286mm DEPTH 266mm SPACING (BETWEEN CENTRE LINES)

TRANSVERSE STIFFENERS PLATE THICKNESS 12mm DEPTH 375mm SPACING 4525mm

LONGITUDINAL TO TRANSVERSE STIFFENER CONNECTION Trough continuous through crossbeam. 8mm fillet welds both sides of crossbeam plate.

12 286

266

375

12


40


BRIDGE NAME Kessock DATE OF COMPLETION 1982 BRIDGE TYPE Cable stayed MAIN SPAN LENGTH 240 metres SURFACING WATERPROOFING MEMBRANE Rubber bitumen

WEARING COURSE Machine laid mastic asphalt NOMINAL THICKNESS 38ram

DECK PLATE THICKNESS 14mm (slow lanes)

LONGITUDINAL STIFFENERS


SHAPE Trapezoidal PLATE THICKNESS 8mm WIDTH 300mm DEPTH 275mm SPACING (BETWEEN CENTRE LINES) PLATE THICKNESS DEPTH max 1500mm SPACING 4000mm

600mm 12 to 16mm (slow lanes)

LONGITUDINAL TO TRANSVERSE STIFFENER CONNECTION Trough continuous through crossbeam. 6mm fillet welds both sides of crossbeam plate.

14 1_

300

135


- 41

APPENDIX 2

TRRL proposals - technical annex

Objectives Description of the work

Methods Summary

43 -

Measurement and Interpretation of Dynamic Loads in Bridges Phase 3:

Fatigue Behaviour of Orthotropic Steel Decks of Road Bridges

OBJECTIVES Approximatly 15 per cent of the total number of road

bridges in the UK are constructed wholly or partly of steel. In the latter case, steel/concrete composite structures are commonly used in bridges up to 200m span. Wholly steel bridges are rarley economical for short spans (except for moving bridges) but are used almost exclusivly for spans in excess of 200m. Concrete and steel compete for the middle range of spans between 100-200m. However, concrete is becomming more competitive and a bridge having a main span of 440m was constructed in Northen Spain in 1983.

The longer span steel bridges are normally of box girder construction with orthotropic decks. High stresses induced in the welded connections in the deck plate by wheel loading from passing vehicles have resulted in premature fatigue failures in a number of European bridges of this type. This in turn has led to expensive strengthening, repair and inspection work and some inevitable loss of favour for steel bridges.

The objectives of this third phase of ECSC sponsored research is to increase the understanding of the fatigue behaviour of orthotropic decks and to improve the design of certain welded connections to enhance the fatigue life of these bridges. An alternative is to reduce the wheel load stresses by increasing the steel content of the bridges. However, this will lead to increased material costs and less competive designs.

DESCRIPTION OF THE WORK Recent work at the Transport and Road Research

Laboratory (TRRL) has investigated the effect of asphaltic road surfacings on the wheel load stresses induced in the welded connections of orthotropic steel bridge decks. Fatigue life predictions indicate that for unsurfaced decks, most of the welded connections are likely to suffer fatigue failures before the design life of the bridge. It has been further established that stresses induced, by wheel loading, in the welded joints close to the deck plate are significantly reduced by the composite action of the surfacing material on the steel deck plate. Consequently, the fatigue lives of these joints are increased and they meet the design requirements. However, the joint between the longitudinal stiffener and the transverse crossbeam has been found to be less affected by the surfacing and has also been shown to have low fatigue life. Improvements to the design of this detail are therefore required.

44

Most European bridges of this type (long span, steel orthotropic deck) have trapezoidal or 'V' section longitudinal stiffeners passing through the crossbeams. Commonly, a cope hole is cut out of the crossbeam at the apex of the longitudinal stiffener. It is the objective of this programme of research to study, in detail, variations on the design of this joint. In particular, to determine the stress distribution around the joints for various positions of wheel loading, to obtain data relating to the fatigue performance of the joints and, if possible, to improve the design of the connection.

METHODS Fabricate a trial deck panel, approximately 15.3m long

by 3.4m wide, comprising 5 longitudinal 'V' stiffeners and 5 transverse crossbeams (including the two end crossbeams), dividing the panel into two 4.6m central bays and two 3m end bays. Incorporate different designs of longitudinal stiffener to crossbeam connection into the central crossbeam.

Dynamically load the panel and study the stress distribution around each detail in turn using equipment designed to sense thermal emissions produced by the thermoelastic behaviour of the material.

Instrument the panel with strain gauges in identified areas of high stress. Determine stress distributions from strain gauge readings under static single-wheel loading. Select two favourable designs of connection for fatigue testing.

Manufacture a number of test-pieces, composed of a 1.5m length (approximately) of 'V' stiffener and associated deck plate, with a central crossbeam incorporating one of the selected designs of connection.

Fatigue test up to eight test-pieces of each type at constant stress amplitude to obtain S-N (Wohler) fatigue curves, placing emphasis on long endurances. Select the most favourable design for variable amplitude fatigue testing.

Fatigue test a minimum of six test-pieces using a stress spectrum derived from information collected in phases l and 2 of the ECSE Bridges programme, again placing emphasis on long endurances. Compare the results with the variable amplitude tests with endurances predicted from the constant amplitude tests using appropriate representation of the constant amplitude S-N (Wohler) fatigue curves and the Palmgren-Miner summatation.

- 45 -

SUMMARY Approximatly 15 per cent of the total number of road

bridges in the UK are constructed wholly or partly of steel. Steel bridges have been used almost exclusively for long span structures (in excess of 200m), though concrete bridges are becomming more competitive.

Premature fatigue failures in the orthotropic steel decks of some bridges has led to expensive strengthening, repair and inspection work and some inevitable loss of favour for steel bridges. It is the objective of this third phase of the ECSC sponsored research to restore confidence in and retain the competitiveness of steel bridges. This will be achieved by research directed towards increasing the understanding of the fatigue performance of these structures.

The programme of research by the Transport and Road Research Laboratory will involve static load tests on a full scale bridge deck panel incorporating different designs of longitudinal stiffener to transverse crossbeam connection. THis has been identified as a particularly fatigue prone detail. Fatigue tests on test-pieces will be carried out at constant stress amplitude and at variable amplitude simulating conditions under normal trafficking. This will enable the fatigue lives of these details to be compared and hopefully lead to an improved design of this connection.

46

APPENDIX 3

Manufacture of test specimens

Specification: Materials Preparation of edges and ends Welding Fabrication

Fabrication drawings: Items list Fatigue specimen 'A' Fatigue specimen 'B' End support Centre support Fatigue specimen assy

47

SPECIFICATION This specification describes the requirements for steel test specimens representing sections from a Full details of the specimens are 1071.07/G01.004,.../G01.005 and .../G01, assembly procedures are described below.

steel bridge deck. given in drawings 006. Materials and

1. Materials All steel to BS4360 grade 50B. Steel plates must be flat and free from defects. Each specimen should be identifiable with the plates from which it was made and a sample of each plate provided (600 χ 100mm) together with a copy of the mill certificate. 2. Preparation of edges and ends The edges and ends of the deck plate and the transverse stiffeners and the ends of the troughs may be finished by planing, sawing or machine gas-cutting. The edges of the troughs are to be planed to the bevel shown on the drawing. 3. Welding

(i) MMA welding is to be carried out in accordance with BS5135 using low hydrogen electrodes to BS639 (1976) classification E51 B120H. (ii) A written record is required of the electrodes, O.C. voltage, current and order of welding (including tack welds). (iii) All 6mm fillet welds, including tack welds, to be made full size in a single run. (iv) Before commencing assembly of the specimens the contractor shall make such trial welds as may be required to demonstrate the soundness of the welding procedure and the competence of the operator. The welding of all the specimens shall be carried out by the same operator.

4. Fabrication (i) Trough sections are plate.

to be cold formed from flat

(ii) No heating components.

is to be applied to any of the

(iii) Each specimen is to be assembled and securely clamped, deck plate down, to a substantial frame or table before welding is started, such that all parts are held in close contact. The maximum permitted clearence between the central transverse stiffener (the crossbeam) and the deck plate and between the crossbeam and trough is lmm.

48

(iv) The deck is to be clamped to the frame or table at each corner to minimise distortion. (v) Materials are to be free from rust, oil, grease etc before welding is started. (vi) Tack welds are permitted between all component parts EXCEPT between the crossbeam and trough. Particular care and attention should be paid to the quality and uniformity (between specimens) of these welds which are to be the subject of fatigue endurance testing. For the specimens shown in drawings .../G01.004 and .../G01.006 these welds should not be continued round the end of the crossbeam inside the cut out. (vii) After welding is completed, the ends of the deck plate shall be flattened, without heating or hammering, to within 2mm from a straight edge placed across the top of the deck plate 50mm from the ends of the specimen. (viii) Each specimen is to be identified by a number of the form FS20.nL where L is the letter Α, Β or C depending on the type of specimen (see drawing titles) and η is a number, following the sequence of manufacture of each type of specimen, starting at 1 for each type of specimen. This identification mark is to be painted in white durable paint on the side of the trough at one end in figures approximately 25ram high. (ix) No remedial treatment is permitted except after agreement with TRRL.

- 49

TRANSPORT and ROAD

RESEARCH LABORATORY Department of the Environment

Crowthorne Berkshire

haa Re.

G01

001

002

003

004

005

006

007

008

009

-.010

011

Q12

013

OU

015

016

017

Ha. eft

Τ

1

1

1

1

1 '

16

12

2

28

32

2

24

ITEMS LIST

Daaerlatiea af han

FATIGUE SPECIMEN ASSY

END SUPPORT

CENTRE SUBPORT

BRACKET

FATIGUE SPECIMEN 'A'

FATIGUE SPECIMEN 'Bc

FATIGUE SPECIMEN'C'

BOLT?eUNCx4«00inLG HEX STEEL GR S

BOLT ^ UNCx 2-50 in LG HEX STEEL GR S

BOLT M16χ60 LG STEEL GR 8.8

NUT Zft UNC ORD HEX STEEL GR 5

WASHER ^ ¡ n PLAIN STEEL ZINC PL

WASHER M16 PLAIN STEEL ZINC PL

WASHER 7/8 in TAPER STEEL ZINC PL

Enauwtr.·

Dram

J. ADORE Y.

Bwriaa! Aaaravai

TWe

FATIGUE SPECIMEN ASSY

1

km *.i.rr

Da* ■art Mo. baat Oati

Category ,, Sèrto* Ê

.1071 ·07/(

TRRL Star« Rai Ma.

02BH973

02 WS 255

02WS916

MMM«. IBM·

Til

Fa»

Ρ

Ρ

Ρ

Ρ

Bat* 1

Sì Sa A2

A2

A3

A2

A1

A1

A1

ND

ND

ND

ND

ND

ND ND

.

M Re.

Addna lam

300 -GO!

50

tima »«n raojfcnw

ui

S

«Λ

Si

: mr' , rau t '■ ■ 'f n

* ' ■tlCNC au'^ "OVI VM dO.*

ι >u

* J l ! > _

* T O C P

PART SECTON W W

as ? wT

p~ «τη / / « Γ Of I»H »

SECTION Ζ Ζ

ί M01ES DRiu4?S POSH TCI ¿ΰ 1

•ΤτΤτί

V Φ-"-îi

ri

s τ« W

φ φ

r* "rl·--4»

u κ

κ nm, tâ ·" *>

laaofBuy > •X*Tt « t i * * ■β··"»* a* .

DETAIL V - t . i r ï .

T»»arât ■» ï ' W r t

?

SECTION X X

I HOLES D H H l ^ ? S /

POS* T t X p 0 . 9 _

7 EOUâl PITCHES OF )«0>«ΙΤ%7ΓΤ.Ρ

31OO.MSI l l l TP

SECTION YY

NOTES -I WCLO STMWIS TO I S i W . P å f t T 1

ÛT: äii /GOI -004 :

KHír:n IAC:CATOTT —« M TO

■son U M FATIGUE SPECIMEN A 1071 -07

Ui ro

α. < rζ

o ! ¡Λ

Ζ ui α. co

c/>

THIRD ANGLE PROJECTION

¿ Χ ( Ι Ζ - Γ Ν

? HOLES O m i t ¿ 1β BEFORE

WELOING Ml/TS

POSN TOL | Í 0 . 9

t MOLES OWIL aí 77

POSN TOL ¿ 0 9

STOCK

PART SECTION W W

-ï

0.

Is I

\

i w'

Μ βθ.0 Τ Ρ

φ f

> — Τ

Ι « , W

φ φ

51 7*%Ν | / « Ε DET. ILV

TTÕCT

THIS PICE TO BE MACHINED TO FIT UNOCOSIOE OP L.FCT PLATE » CLAMPED

6 J 7 BEFORE WELDING

DETAIL V SOLE 1:1

TTP1CAL « 7 POSHS

©

SECTION X X

SECTION Ζ Ζ

NOTES

I WELD SYMBOLS TO BS199>PAPT 2

1 LL

— · H S TO

os neo CA sos

j «ron« A t i A .UM.·'

FATIGUE SPECIMEN Β 1071 ·07/θΟΟ -005


S

LU

O . c

ζ

o (Λ

ζ LU

Q.

(Λ Z3 CO

MAKE FROM 6 1 0 · 2 29 . t:S kg/tn

UNIVERSAL BEAM

7S

4

t S 00 T.P.

ι r

«1 +

i t .

ΠΓ

Φ~""γ Φ"'"^τ

— Φ — ^ φΙ_φ

Κ 0 900 Γ.Ρ 900 Τ.ρ

I HOLES DRILL p 1%

POS. TOL. β" 09

β HOLES DRILL fl 25

„POS. TOL.0 09

O

o

! 0

ι

I s

¡l

<g>

BEL •ATI H H · , ιΤΛυι »ATI

TRANSPORT tmi ROAD RESEARCH LABORATOR*

lvi·· ' — -ι at I*. | a....!

■' *■ . . mi

·«·» M.S. TO

BS. . 3 6 0 GRX3A

'—PAINT GUARD

PRIMER/UCOAT.

END SUPPORT 071 -07/G0I -00

LU Q. <

O

m LU CL CO

ω

MAKE FROM 305 «305 «97kg/m UNIVERSAL COLUMN

TRANSPORT in. ROAO

RESEARCH LABORATORY DeoeMrtanl « * ■ * · t new aajaaaaaj

Omeatman« of T amcw i

tkVoea1fTpofa>a) Ρ Α * % · ™ Τ ^

• » n m n m ·

TOLERANCE!

■ . •Arc . · )«» ·

0·» ·*■»··*» Τ « . ρ ( « α » ρ 1 * «


600

333

/

7 EQUAL PITCHES OF 3·00ίη (76·2 ) T.P.

121-00 ¡η (533A.1T.P.

'^""ΦΦ""Φ~"Φ

_Φ~^

_ Φ"

8 HOLES DRILL ff 25 POS.TOL0O3

/

t ι— t tfc— t atas

·«**· M.S.TO

Et5. ¿360. CR. ¿3 A

'—PAINT GUARD

PRIMER/UCOAT

t ba··« md tfwjap «dfn

ELE.L.

Tr».

CENTRE SUPPORT

HV87 DATI

C D

CTI

θ

Kau 1:5

1071 »a··» i

07/(

DATI « D O « .

G O I 0 0 2 argaraaT Eat* an a a a a i lm AI

I IT .OI

s


8 HOLES DRILL 0* 25 IN ITEM 00 2 POSN FROM TEST RIO

7 'τώι' | ~ — T E S T RIG

TRANSPORT airi ROAD RESEARCH UCORATORV

1.II . I I I . I I I . ' Til . 1 1

t .— - . M . — .1 Μ Ι , Μ 1\>..ιι ι I". ■ · . MjWI

SEE ITEM LIST

JADOREY. /»<?..«?

FATIGUE SPECIMEN

ASSY

■Oui DAM aeaaj. row pail ape»» mm IMj |ajot».

1 O

! O O

O

ι— O

i"

Θ

1071 -07/000-GOI

APPENDIX A

Test specimens - welding details

Type 'A' specimens Type 'B' specimens

57 -

WELDING DETAILS - TYPE 'A' FATIGUE SPECIMENS

Electrodes : ESAB OK UNITRODE 48.00 DC (70 OCV) 3.25 mm FOR TACKING 95-110 AMPS

1. WELD M16 NUTS TO GUSSET PLATE 2. TACK GUSSET PLATE TO TROUGH

2 NO. 25 mm LONG TACKS

3. WELD GUSSET PLATE TO TROUGH

A mm ELECTRODES 115 AMPS

A. CLAMP DECK PLATE TO BED. BOLT BACKING PLATE TO END PLATE. TACK END PLATE TO TROUGH WITH 2 NO. 25 mm TACKS AT OUTER ENDS. TACK TROUGH TO DECK PLATE WITH A NO. 25mm TACKS.

l· A I 300 I

GRIND OUT TACKS PRIOR TO WELDING 5. CLAMP CENTRE CROSSBEAM IN PLACE - NO TACKS

58

6. WELD SEQUENCE FOR TROUGH TO DECK PLATE, END PLATE TO DECK PLATE AND CROSSBEAM TO DECK PLATE

1» -4 h 4 8 12 lê *

3 7 </ '5 _ J* Ί

04 2 é /O /4

-

A mm ELECTRODES - 140 AMPS

7. WELD SEQUENCE FOR TROUGH TO CROSSBEAM (BOTH SIDES)

'J/ f\

59 -

WELDING DETAILS - TYPE 'B' FATIGUE SPECIMENS

Electrodes : ESAB OK UNITRODE 48.00 DC (70 OCV) 3.25 mm FOR TACKING 95-110 AMPS

1. WELD M16 NUTS TO GUSSET PLATE 2. TACK GUSSET PLATE TO TROUGH

4 NO 25 mm LONG TACKS

3. WELD GUSSET PLATE TO TROUGH

4 MM ELECTRODES 115 AMPS

4. CLAMP DECK PLATE TO BED. BOLT BACKING PLATE TO END PLATE. TACK END PLATE TO TROUGH WITH 2 NO. 25 MM TACKS AT OUTER ENDS. TACK TROUGH TO DECK PLATE WITH 8 NO 25 MM TACKS.

i

1

Γ 3 0 0 *~ 1

m B J 3 0 0 ' Γ 300

GRIND OUT TACKS PRIOR TO WELDING

5. CLAMP CENTRE CROSSBEAM IN PLACE - NO TACKS

60 -

6. WELD SEQUENCE FOR TROUGH TO DECK PLATE, END PLATE TO DECK PLATE AND CROSSBEAM TO DECK PLATE

4 . 4 n —

2 "/ / / IS _

r ' 2. 6> IO 1 +

.Γ

4 MM ELECTRODES - 140 AMPS TYPICAL WELD SEQUENCE FOR TROUGH TO END PLATE AND TROUGH TO CROSSBEAM (BOTH SIDES)

WELDS 1 AND 2 WELDED BEFORE TROUGH TO DECK PLATE WELDS

- 61

APPENDIX 5

Inspection of test specimens

Specification Inspection reports: Type 'B' specimens (1st series)

Type 'A' specimens Type 'B' specimens (2nd series) Material certificates

- 63

SPECIFICATION

The following inspections are to be carried out on the completed specimens : (i) Check the overall dimensions, tolerances and detailing against the drawings and specification. (ii) Visually inspect all the welds and make spot checks on the weld sizes. (iii) Carry out a dye penetrant or magnetic particle inspection of the trough to crossbeam welds. In addition, measure and record the (two) leg lengths and throat thickness of these welds. The inspector should report any unsatisfactory workmanship or practices or departures from the specification or drawings immediately to the TRRL project officer. A brief written report should be produced on completion of the job detailing the checks carried out, any departures from the specification and the results of the weld inspections described in (iii) above.

64 -

CONSULTING. INSPECTING

AND TESTING ENGINEERS

TELEPHONE: 0 1 7 3 0 3 4 6 1

TELEX: β 10518 SANBER G

MESSRS. SANDBERG

Inspection Report

Repon No. {

Sheet 1 of: ^

Clicni >gftwSrla~ Λ Η Ο 2 Ο Λ Ρ Kc«>câ>.ví UtecTtoVaft igS Job No. 3 / 4 * ( 0 « &

Contract STÏCICL. lÄTSreuMd**«» .FS>£o.

W o K S ^ u M S f t f t E H C INI)UÍ>Í?ÍLÃ b p . .vJføfcZD U f t l . L W I L C / . Tel .No.o3_qV3.St3 , «

Inspector(s) ■■l.£t"h.Gnr Person Contacted u a y ^ ¿ v ^ .bufo/ffcfl

Visit Datc(s) % / l j ^ " |

* Inspection Stages Key overleaf. Resul ts. A » Accept. R = Reject.

Drawing No.

ίΟΊ.-ΟΉύο. ΟΛ'

Item mark (No. off)

*·

IB.13.3A

Inspec

Stage

.C A N O l b . „

l e

·Λ.£Ε.£ο IC

ae

tion*

Resul)

A

.A

A fl

rríT.Gutí S^t .uoJS. h ?at Fslo.T?>.

" D I N N O M K D l&Te-Si 'E-

ϊ ) .Μ,ΐ>Λ,ΰΜΑι_ C\kt.\L<L, OiburtL EJCIMA. UOUcrJ

3Γ~ΤΤϋ_*ί?Γ »si,£:L.öS, F/Vr\i T>CW£.C.TtoO TèYT\

ßF" hlOO.rlCr C A H . I Í D ÄWT ΙΓΛ Prti_ü2DAríai YV-ιΐτ.

CLI-C*?"«» S?tÄir^d\T»oM TUu<»:—

SuKñvcir 0«οΓί.ΤνοιΛ e r Ρ*ΙΛΓΙ£ Arto IføuOrtS.

° / Α L O J Î . T U \ ί. ^ , Ο Τ \ Α o r ÙctL R A T ^ É A I Ù

P L Ä T Ä A »si D I M . D Ì K J I SefcT.iV«! Y\J*£ .

Vi.DTtt ÔF TßoaCU

U t P T r V "

PiMtOOSieî k¿Tr>i_a£*J £.»ΛΟ Γ_ΑΓ_Λ Prviò MVODLÉ'

S><DLT,O«Í Ρι-ΑΓ-Τ

ALV. IFAAS Τ ϋ υ ^ Ο T o 'L·? UvTU ira Sa>P.r. o "

L>ÖM/\h* C. ,

AS A ¡ W ¿

/w A L J ¿ w.rw TIW «£*or?r,-w *>- T U C0

I A

LervSWtéF "LVck. PLAUS. U d i i . t L i M i T Ú "EC

U-MM ûF TÌWrtfc' ftôJkí/StBr"ïffTuu W.TWMS

lU¿ TbUca-A^r^fT. __f TUir M o O . I N A O Ä O ^

Gti/ttvi ùhx TVVt Per tK t t ^ i ^ ^ o í UT¿ftir>tí.¿îAl\i

f M o T e S ^ .

Enclosures: ¡A.PfQ ?*?A\ . M A TL C t I ÎS '. U l t ^ p ^ ä C e . * ^

Further visits required? Yes^Hf Inspector's Signature Date φ/?ι Inspection Authority Signature Date

65



TELEPHONE: 0 1 7 3 0 3461

TELEX: 9 1 9 5 1 8 SANBER G

MESSRS. SANDBERG

Inspection Report

Continuation sheet

Repon No. £

Sheet 2. °f= S Job No. H*.o2.

Drawing No. Ilem mark

(No. off)

Inspection *

Stage Result Comments

SB, L LCrX... Ù i

22>

iSTùlg,

ÏC A.

i c A

A

ΡΐΛ~£ A r i o YW\£ É.rC»D T ^ A T Í Ο Γ Α Τ Ι Λ Λ .ΠϋΛ UtVí —3>ΜΜ or" DfeAWjtjT, .TUdk Ate* fco

ïktmrtisictas o#\i T W Î)i2*.«4j<cr*, Ttvfi TU»·*

DcMOÄvorsJ LSiLC Leuen ι»» Gzu.KfcTE^^

¡ΑΛΤ> CLA.M?«-£>, ΙΓΛ (vcLcajoAt^tií. ν*<Τ\\

ÎAf iD V^CXD. I ÎC v>£«\ »»A pEeA&iS*, TU.s.

ÎGAJC KC "fite fifto»MÍTH T¡» <U ε Tlk: 1-LD¿£,

■ilr-in.j-^ otO TUiC ΤάΰΓ3\υχ.\(λΰΔ P»fiD T L '6 >s»d2Jt¿: TU¿ CLAUP.ICÛ e r e TC» ^receuTfeisr

¡ ß c D . T l t a tri »Λ T ^ U K Í O T Û &.£ SAÏÏ'MACTÛIIA»

iT lk É.MÙ P L A T Í S , ULTttödCU UA<JvM\ Λ

¡ U I Z A J U J pUÖTt* "β,ΰΐΧκΟ TO Τ Γ + Ι £ "Ê>ftcJC »νΐΡτ<Β ifLürOO Te» ΈχΓ S^LiOÍIL-l ΐ>Λ'ΐΰ2ΐΐΐ> ¿1 iSûHEiO

¡Affea » r »*Λ*5 CûK?La-ei> Μ>Ι> R A C T

[Π4Λ Wff> fvca¿Ht> "Q 2>¿: ftrx¿fTí>il?u£ i l Söf

jCfOStáLAL. K a T t S ^ ,

[ffle D t t £ ?L_AT?W ,UÜ«£-r "îUe M i ù û u r W I Í N *

¡ P L A T ¿ & WtCftøl wi£LOtCO Wè&r T ^ v ^ û T k

!&£ ' S L C U T L M TàSTaCTieD OUe Te, TU;»: kloPuA

i l F Û J T Τ \ \ ^ (Ans ΙΑ( \Α \ )Β^*»9Ι_Ε. .

¿ M M M L K L I J C T W Q J Í M te tsu i l t i íD ι r* fttíaíliX :tf

W.Tft S?cÆ>ricfruc*J."iTfeSc' K t ^ Sue.uu_~l

U r 4 D c f i . ^ i 1 l í UVA "Πλι£ SlaST&ÏW. LjCtv Ü I M T I + £

D f C f i P L A T É T Û ÊAlD^MiDDüfTSltSlTvu^ PUKTS

Z.M fKQic t iD Ti» á e o í P T £s«ar Cr£Kiii\l(:Te5.\

WZ3L Inspectora Signature Date tør Inspection Authority Signature Date

66



TELEPHONE: 0 1 7 3 0 3 4 6 1

TELEX: Β 1951 θ SANBER G

MESSRS. SANDBERG

Inspection Report

Continuation sheet

Report

Sheet

Job No

No.

?>

1 of:

H S O I

?>

Drawing No. Iiemmark (No. off)

Inspection *


4BT**& ift Α. Μ·λΓ> Gdf f t <¿¿5i¿ QrVti.vl£t?TeR f ^ Y i i c A L

â CKe-KtlCAri> PßoPeRTTer.

M»L1S CiäfS" A»40 SAFA^L¿9 or ÇUr-a*. rt »ûoX lûorAM*) I A J » U L T » C S,r£MT TÛ^fA Ir4cfi.<5»

t»f fÁTTCtiK Tc*ìT iTeMS rV% i«M rVOâDriARt

WÍTI4 S?t2nr^cftTÍD/¿H. rtAr«rA.Ai_s" i x rmu-S.

ôl= ^ûcerOuexifdirç» T U A £*47ft.f*3".I »rJfVs

;i/oTbßM£Ei £ U Μ2.?«£Α·-4τ- TlUT TUlSollfb»

¡M¿r 2í&u,££O.

fc. jjlk^WélJCbt |V?2OJAL! CrtT.fíCMUf SuiVoiiÛ

ψΑ M^ M¿yis *£Æ " i lk UI¿ID¿9 £k<£S N¿7*

>CW_/ +V»iUi ι^Λ .

WIÍLDÍPVÍC»

£o«4>uNAl\j^>

C e n d a l . N M ¿ V

ÍN CutaJBdLiar.crA tNiTH H i l ¡Zervix o f T* í?ZL

bv/tiû. Tlf*: Ti i¿:?fWc: Ι Ι £ > ? Ι _ Λ . ^ « Λ T ¿ I \ I M

|Λ2»^ΑΤ T W r ÂfLrVHtwjs» AKjftM TÍ2cMTl4ir

PZfty<,BÍ^- L F S ^ U N I Í or 714 é" D . M O K I Û ^ S Avi D

T l k : SiitSr+r Dc.T¿c[TvUFí5f A K > Ù i k Aú/2£ti>

TW AT TVlÄ- K d t e : A<xc:rTA"Jli_£:.

l i k ÛUkiTio»>i uFTVl¿ IAKU¿£SÍ2Í£· FTLUT l4clbS _

I exRAiBit^.TMAT, KlHeßflUs. N Û U L D A I ^ M A I I X

!£¿ A 2ά^τΤνΔΓ45 ΐΤ\νΰ«ΛωτΤΛ,νΤ^να>,»Γς ^ r f f i r U a i * uP rAAH CAUS¿ U^.T¿: ft\_ùTcv"

D,Vra3Tûryl,rVb ÍrAa»¿ L EUILD.OJC IA?" Tl^S£ UtiLDl

l i k i irlwkLù u tOû^TcûLY &.D WÙU- &JtA TU-r

'&ÛU»2»r£> Β?ΓΗ.Μ.Γ\£ AG¿£t?To U K O Ê rkU-ftUrJx

j.i>i"ïlU \hfJítFÍc£:/'Lar Té. rAAtìt W t í l k fttwr ¡S.2J3 Hii£ APPU/DTÛ f l k IN^T'SAÎÛA.V^ HoldS WûS, ι Ν Γ ύ Ζ Μ Λ . ¡m-tarroÄUi·^ uSiíD I N {HLA¿ohHce: ΜιΤΊ* ^Pcc. ]CLAü^f£L.ιJ"HFΛI>¡^^C.

,

ES«. üNiTfautr <A .ao fe Llfi : IS7.V. Es ι u i & tfe W » 0

Inspector's Signature #/A¿# Dale V/7A7 Inspection Authority Signature Date

67



TELEPHONE: 0 1 7 3 0 3 4 6 1

TELEX: 9 1 9 5 1 8 SANBER G

MESSRS. SANDBERG Report No.

Sheet 1 of:

Magnetic Particle/Dye Penetrant

Flaw Detection Report

Client tlìAttFcST to42> tCAO töeTftßcW U&ÆATURIA 4 Job No. ? U-^tfL

Contract STÆV-TeT S^tUcfisTS fíía

Works ' J t y lWfog E N C . U t X . L T D . k ^ f & f r M L^JJrHSürface Condition A5 v>lfru>,sä>

Inspector(s) il.T>ta L\kT Person Conlacte&MllM¿XÍ>S t¿?¿Z Dale: S l l J ' f O

Specification for Testing Method fcSdcTl'. 1 ^ ^ \ Acceptance Criteria^ S ^ 1 5 l l ' rK J ^ f i L A T r A g ^ . r\.

EQUIPMENT/CONSUMABLES

DYE PENETRANT MAGNETIC PARTICLE

Remover _£ν^β_ ¡Method rAAüst&^lxiX M M * £ ¿ N o K f ^

Penetrant ■vÂ

Contrast Aid »t ucP*

Developer M[ft Delecting

Medium 1 U F

H ¿ M . TEST RESULTS

7SÍO

fTVPtl

SB.

"UmxCKl Tc» LJZcSS&líAiVA £rAvî>OUr ÔCTvo^ T v ^ U Í ^

t û û 7 o M P T C°AC2>£Ò ΟΙΛΤ Ο »Λ TU Λ S Ì O C T W ^ ι » ο ivccûCûftKs:^

»rJïTU CL\e^lTí, SSeXtrTcAuoiO C L A I / S É ¿Li I 0 ΙΐΛΡςχΠονΛ ùY Ci¿l?L TtrST SF¿OMeTKÍ% .

Au. VNÍiLibS Htdi£ t eu tão Tí 2>¿: h2e0£ ciF iVvM S*.erv-.vr.u\fõr

USi rbcL Z&A

Signed:

68 -



TELEPHONE: 0 1 7 3 0 3 4 6 1

TELEX: 91951 θ SANBER G

MESSRS. SANDBERG

Inspection Report Report No. 1 ¡ΖετΓ

Sheet l of:

Cliegansport and Road Research Labora tor ies Job No. J/4902

Contract Fatigue Test Specimens.

Works M* Dünbar Engineering Ltd Rutherford Way Crawley Tel. No. 0295-25437

Inspcctor(N) J.E.Thompson Person ContactedMr Mews

Visit Daic(s) 3rd December 19&7

* Inspection Stages Kc> otcrlcaf. Results. A = Accept. R« Reject.

Drawing No. , , Inspection*

Item mark

(No. off) Stage Result Comments

1071.07/001.004

FS29A

A1 t o A8

i n c l u s i v e . '

11C

! 2C

A

A

3D

3E

A

A

Surface condit ion of p l a t e and Trough

mate r ia l y Overall l eng ths and widths y

hole p o s i t i o n s and other dimensions

whithin t o l e r ances s e t during the V i s i t

of Mr. Wright.

All welds acceptable but with the following

d e v i a t i o n s ; - Trough t o Crossbeam welds

Specimen 4A0ne f i l l e t weld i r r e g u l a r

p r o f i l e .

Specimen fcft. One f i l l e t weld some undercut

t o trough and of poor p r o f i l e .

Specimen 8A One f i l l e t weld I r r e g u l a r

p r o f i l e .

Trough t o Deck p l a t e welds

Specimen 8,4

.Areas of f i l l e t welds below s i z e noted

as 5m/m with i n t e r m i t t e n t u n d e r c u t .

These welds onSpecimens 1A, 2k, 3A, and 7A

a l so contained " S t o p / s t a r t " C ra t e r s with

a c en t r a l Gas ho l e .

All the above areas were l o c a t e d and Marked

with a ï a i n t s t i ck for easy i d e n t i f i c a t i o n s

by the c l i e n t .

Enclosures tJotJ£

Further visits required? Ycs'frw " Inspector's SignatureVSP (% *e /"S / I . . ^ 7

Inspection Authority Signature I ■3/ . i |vlt>&_. t3c>LO

_63ihjsz

69

CONSULTING. INSPECTING AND TESTING ENGINEERS TELEPHONE: 0 1 - 7 3 0 3 4 6 1 TELEX: 919518 SANBER G

MESSRS. SANDBERG

Inspection Report

Continuation sheet

Report No.

Sheet 2 of:

Job No. J /4902

Drawing No. Item mark (No. off)

Inspection *


4A M .P.p.Β exfcmintions of the welds revealed no further significant defects.

Flatness tolerance . It was noted that the flatness of the deck plate where required was \*ithin the specified limits,but in the overall length

a Bow was present on all specimens,but this would seem unavoidable when considering the amount of weld metal deposited.

The items have yet to have the Full Number marked on them and a number of areas of sharp edges have yet to beyfcmoved . I understand that these matters will be attended tc before despatch.

All items examined were accepted after telephone discussions with Mr. Beale . During this discussion I observed that the deviations in the welds e? noted,

were typical of the flaws that would occur during actual fabrication of this type of constuction.

70



TELEPHONE: 0 1 7 3 0 3461

TELEX: 919518 SANBER G

MESSRS. SANDBERG

Inspection Report Report No. 2 / HJW

Sheet 1 of: /

Client Tart»*Aar ΛΗΟ ?e>/So E r t t ì t t a t LAÍC.*WTOA *£<* Job No. J/ V*ol

Contract FrtTiûuC. TiLsT SítciMErí^ F S ^ O .7WP¿1?» .

Works'JOK D U M 3 « A t^C.îrÔuSraALrO.ûnteAfîiîûlJrrl CffrtWUÌl Tel. No.

Inspector^) H.J.WRIGHT Person Contacted

&^νί.5·νΐ»7

WX £ΓΑ»Α Κ ί * ϊ 2 £

Visit Date(s) l o / * S / - a i

φ Inspection Stages Key overleaf. Results. A »Accept. R — Reject.

l t í7t .u7kioo.

Item mark (No. off)

eo i"

{ft To t>B

42» I S . 4-E,

Aus S a

3&

¿3

Inspection *

Stage

it ic

It

a.a

^ D

"bî>

ID

Result

ft

ft

t «

A

ft

K. *

e. *

ΡΑΓ.Εαύ.ΤϊώΓ Sp£jCiV»e»JS "ϊγλεΒ> £oFF » Λ Ρ * * £ Ο 12. To £ B

D I M £ N S Æ I M K \ _ CjAítlk^.ViûrAL· £j>.ft»A.^»Auôr\i

AAö R i M D¿T¿2ÍAIftfo "ΤΪΛΛ" CtVöß»£X> Û A T i t *

fVcLo9.ûtV«tE. VsVTM C U e r * T 3 *?cC*a\û.E»£>aTS.

Ç,ua??tti£. CarOTarAS fir Hi£ VÂTÉ'S. wo"ï twv.s .

% U J Í I H I V S . or frrtX. ?U>.7¿S

WA M.ùTtA^ *.f* "tinvcIrC. P L A T « ^ ± 1 » . M

ù u o r v o i s i "FfccM α ο ο Ϊ \ Α Γ Α T ^ MIOUVE. VLP.rcx

Op£#J C U D T k tA.ODt_»I R-rtT£- " ± i - Λ . ^

"&Pé>. £>»0 To N I O D L É . PLATÍ. ΛΤ PdftVC e>r"

TîlouCvA jV/Aa»K7Tes.t>ïs TfcarA —'S Τα — 'fc ~ ~

S»oU¿. e.r 7rlc5>£. Tfeo^H's. K id í . LvO" SrkZT Αώο ¿¡las lUòTWc u t t o u L FLATCC PaLLûi u P D^c

72». TrVvt. VJEATJiisSif.

HOUL Lbx*vr.,e»*\ fttiû DiftM¿r±r.

¿ H A U ç L Ö Ä T H ftÜ«3" klûE? lCÔuuic& ftU. t?öAMtb.

UBIDBÄCJJIT Tõ. V^ÄTX cri er jc S i o i oFTÜoacr·, ÔBJO\

WDCJLOiT Το ?U>T£. «OTU ^ . 0 * 3 Ο Γ ΤδυαςΗ

Dûitt C«dTei7o«3 A4»£.T» e «OD PLAWÄ e»oLS.

Enclosures \A?fQ 2abfÃ l\* * *>T»4 —

Further visits required? j l ^ N o Inspector's Signalure

¿Æ t €>/&/%*

Inspection Authority Signature

71 -



TELEPHONE: 0 1 7 3 0 3461

TELEX: 9 195 18 SANBER G

M E S S R S . S A N D B E R G

Inspection Report

Continuation sheei

Report No. u HJ.ÏÏ

Sheet 2. o f : 7.

Job No. J / 4 3 o ï .

Drawing No. Ilcm mark (No. off)

Inspection *


^ S ft

1A

lfc

Jfc» U TêLCfMûfsiA. G^j£A^t?tt»utNÍ,K!ÍC.ScwL¿s

WlAS ttAfoTìMCù Of TWii. DiiC.7.C PBOJCWÄ

If4 Ttt£ DiMOA.ur . fHU C*CCKS,; ftrjO Ü£ V4fA

S A r ^ r l i i C s T o L C A V / ¿ " M c ü . ΡΛ TUcH AfttJ^

Ori TWc CiutCïr.oM O f TUí t?£r«CcJtr» IrslxiD*

UZ &£»<_■££ kÌNS .Hr*ÆMCO T W N T »Τ rioAUo

M£Ar4. Λ COMPLUTLÍ . H£j_ö P.V5A ftUsMQ Tire.

SnC>£. C F TrVt CslrllC.tfftL. W Ì C O O *MlÜr\ Wt.*\±>

Cu/£. UMCQaftL VJCC L__s>A"riA a2 TUc V4(\ol£.

Vj£i_o UAt/iM<f To rtAísTAo PVf«Ä α*λ.._Α

jViexuC TVlaA ff,\j£ A í f o t ¿ M K L£trL£>it.T>»j

!ÍÔK. Íu»6»csr4 û F ΤΒ=\ΓΙ*4<, Ori TMcS¿

|Sr«fX\r»4E»Ä,,ir»4«^ Τ Λ Ο Τ Τ Η Κ Τ UvfcCdl fT__T

i Nt_LÙ_ vizShr l^tX Ο ι Α Λ Λ W O H.frSi

At%0 >N Tei l_£ftj_ TNtAC AS TkefvJefie..

WÊUJÚR AppßuJAA. CûiTvRcfirtr: _F rtd.rAfxS

NfrS ftc<JoTftSu£ Prt» PoultUS, A M O tVftù S¿EM

Ι_Λ\ΟΟΑ. _ _ ù ΑττΛΟ SÄÄUUHT αΡ Ti. P i \T_ . .

(ΑΡΙΓΕΛΙΑ\_ CdrnFiCAÍéS i4£R_ FS*MÛ ΑΕ£4ΤΓΓΛΪ_<_-IN ACCGtlùPirrïCc Huí* t & ^ ' & c . r W l C PrrtÄ^TftUPiE)

AC£c&0.rA»4LM.

S« FTPBCD f i o u c μ Í L A T É S (Γ#ΛΤ Ν1β.S8 S t t T S ΐ « ϊ > £

FLAT i -r iO p»t>4ö tL_C*l R r i f ë i i ~7f iS4_L K2.IÍB.» _ G ( I A O Í _ . __í2.

I

IMtODuí. "FIAT Η Λ » é¿S__ LSH G A O S

Κ)\__>.ΤΛ C v s s a u w i u Ä ' . ' i£LerJ(ZADtS

OK.E&ftB.. UA>TT!_ÛÉ. 4 * . o o . .&«<<: lS" I£

t S v <V1S I lo. l L & \

72

CONSULTING. INSPECTING M E S S R S . S A N D B E R G


TELEPHONE 0 1 - 7 3 0 3 4 6 1

TELEX: 919518 SANBER G

Report No. ¡Arw

Sheet 1 of:

Magnetic Particle/j Flaw Detection Report

Client |e»u>?fl¿r FVTMP 2 o f t _ Q.&.£PP.O\ Lft2o«LAToft.»=g Job No. j / 4 3 o L .

Contract "FÃTt uE. TsJT S?-r tM-H\ FS a^, TV F-. S. ■

Works/ÄtxLVwßaÄ ENtVlN0vfc.T2._ LjD.^ínfôí^lÂiiiaHljdSurfaceCond mon AS t»}..—=»-x>

Inspcclor(s) h.J.Y/RíUíT PersonContacted: M f l Λ IM__V-Jt_i_ Dale: t _ J _ > J f c V

Specification for Testing Method BS 6 û ? 2 : 1 9_1

BS 5 l 3 5 : Ϊ 9 5 * , TABLE 1 9 , Acceptance Criteria (¿UALITY CATEGORY . E_ .

Remover

Penetrant

Developer

DYE PENETRANT

N/A

K/A

I S / A

EQUIPMENT/CONSUMABLES

MAGNETIC PARTICLE

¡ ¡Method Ü A G N A F L I K PERMANENT YOKE YM5

I Contrast

¡Aid UAGNAFLUX WCPi

Detecting

Medium MAGNAFLUX ΤΗ}1

DRAV/ING

lull, ell

DOO.OOÍ'

ITEiá

'I EST RESULTS

LEVEL OP

INSPECTION RESULTS

fÓr.6i_a.T«_.r ÇA_Cir4_j_S

F£_ö_TH?_.i_.

t*AQ.<c¿> Í S T â - B

Tío-c »4 PtAT_.s To Ntùouc At4û __r_D PvATTS.

ι OÛ7O Kt S.CrafncAKiT _uP.HÌ__,

I A _ F _ _ T _ . _ Ac___P77>.<u_.

f s^ . . . . ß t fVU "ÍUS O E T IK L ^ T U C T . O M Witti lNSr\_j_>rA »2_?__Γ t_C l ^ T V J , __00_ίΛ_->

Signed: /M¿y ^

73 -

^ J

. * > .

British ¡Steel Corporation Scunthorpe Works aSilTÆIåf Telephone Scunthorpe (0724) 843411 TO Box No. 1. Scunthorpe. Pruefbescheinigung _ . _._.Λ. South Humberside. DN16 IBP. Telex 5260;

103

DM Cam o,.,_

Ί / c / o c

;< JAMES 8 R I 0 G E STEEL STOCKH I!'*' ' 1 0 6 - 1 1 6 ASTON CHURCH ROAD

: BIRMINGHAM B7 3RX

OLOERS « h - i r - f - i * i i f l

eeVST J - M O r w n 3 - - T 5 7 4 - 50mm

M . y t / C - n - K l l-_r*_r Hap-η d» aammantJw/H-. 8» <gn»-t Aa/lr-o**-v_aÂ--_gar-

&74 CAST NO

7 1 / 532 froauq paaonown

BV574 . / i / o

BS 4 3 6 0 - 7 9 5ÜB C R . b S C / F .

Γ™" PLATES CÚNTR0L RJLLfcu Cm

l.v

MMS MMS IMS

I 1 0 0 0 0 250Ú 1 2 . 5 0

JAMES

CUSTOIA

IRDth:

AOVIU

ITEMS

SUPPÜ

6 6 5 0 8 ^

BRIOGF

,VIA:.X¿O

D AGAIiMS'

LG

•!RMIX5:'"M

73- 2 oo«73

MVrj,

THE

N/MM2 %onOL

E

«♦89

y

TR

STEEL sroæHOirøsLO

ι «OVE [ROER

ANSVER

567

f-ÛtiuTAAHL·

SE

22

ι F W » T » Ka__BKl^)aBSá>B_l

CHARPY V 1 0 x 10x 2MM JOULES 0 C A»

LONGITUDINAL

108 100 108 105

\r.\ DCÎ£ tf;

ÁNDATE 5#/??.%^

<7/

2>cöji tV

COTWQ. F _

ΓΧ Λ λ Α Α Λ \ Α Λ

CP

•»ta*.«*

13 .34 1.3a . 0 1 ' ; . t i l / .¡JCV .>.·Ιν. . υ ^ . . , / > ALVO 2 3 C L K U * . ^ . * . ' X c

I Γ * l

n K« «I!

»Il k· TM

rs f

Varaa tsuwd by lha BSC w M

from the H a p p i » that il i t ·

■ n m * o u i d wm*. Any icapnnt of ■ copy of ■ BSC tasi carn íca ta w * h o u l the ι

and OCurvM reprodUCIon OÍ Ihn ongtnel

XAJUUUWUU.XX XX X X " < .· ' ' s. JUJUUUXXrXAJt Χ A A A A .· · ·. Λ · Χ^νΧΛΧΙΙΧΛΛτΧ.Α,λΛ.' A » .·. A · " M * . by ih« BOS · ν κ · Μ

M*»»e*V >F»>i»i aS· . 7 _ i _^ f , «ι its* Ρ· Τ**Λ Ι Γ> j i * ·

On behftff o* i h * FV.i.ih Sî*»<^ C c p o P t "

T e « H r t u i * M a m i g « ,·

F· _ν^«Ί Coor.o-ΐ.' C h / tUV A b n a h f f V ^

/ * * ,*

Teesside Works P.O. Box 29. Redcar.Cleveland TS 10 SRO Telephone.0642474111 Tele« 587401

CERTIFICATE OF TES'

Cwlloma.

tot paction

• S C O. t» . No

3 4 1 4 3 4 enlomara O.dBt Ma.

o. . . 3 0 / 0 4 / 8 7

C a r t No 187574

Adv tea .No 66965

4 Y\h ' R S L es 19

Spacilktiio. 50A A .R . BS4360 1986 50A AS ROLLED

TEST CERTIFICATES AND ANALYSIS 3 COPIES OF TEST CERTIFICATE

Goduti 3.fc.iBiM» HOT ROLLED CO IL

S

l l l i t V M i f l

Numb« rt

ITEM 3

1Z0\ ?-

Or, .

1

1

O i m i n . w i . .

MM

1350« 6 .00

Car t /Ha i l No. flaca No.

Spirii .c «itou __"" M M

56(15375 843S5

/OA

Tl . ld

S l ra t ·

a •

N/mm'

355

474

J

■ J*»

s k r f ·· —

TaaaUa

StraaglK

a at

N/mm'

490

640

556

SC

t long

% A

I . 0

30C

IE

16

y

Λ

Λ » ·.

ν / / '

ν

CBarpy Impaci Valúas

Un i i . Av.

B A S I C O X Y G E N S T E E L

A N A L Y S I S %

C.

.00

.33

.12

l i t ;

Si.

.000

.500

.OIS

Mn.

.00

1 .60

P.

.000

.050

s. .000

.050

NB

.000

.100

1.15 .0 18 .019 .025

X x / «/"

ÔLA/Q/m e vo

li S SI.

&ol<tSO ι

Page No 1 and

V

.000

.100

.005

1 ♦ (·:υ :.';""

l a s t

Ú .1 .ι o .

.û.\'SV.

Ht.' 1 > ν Λ

ι ' « / i

V l__f

All Tasi Canille i l a t I f luad by |ha B S C will

contain iM i t m b o t f t d taal Any rtclplani

oi ■ COPY ol · ns C. Teet Canille ate without

lha laai fhou l ·η»υ ι · ('om tn* f u p p l i · '

that it t i a u u t and acculala rapto duc i lon

ol lha ongin.) .

T fiato fatuità afa canutad by i h · n,ui,h Staal

Co» po» ai «· n and comply « u h iho loqul i tmtnt i

ol i h · Produci Ootcilption.

On bah»,l ol lha Bninh Staal Co'poraiion.

TaattiQai W o t U

s.u lûvJL I

APPENDIX 6

Test speciaens - weld sizes

Weld measurements Weld sizes: Type *A' specimens

Type 'B' specimens

77

SURVEY AREAS: © and © full length of weld © and © crack a initiation point

Type 'A' speel»ens

Measurements in millimetres to nearest £ mm

TROUGH

•9* X7 o 50 O Β OD tu

SURVEY AREAS: All round trough to within 50mm of deck plate Around crack zone, 75mm either side of centreline At crack initiation point

Type 'B' speci„ens

WELD MEASUREMENTS

78

SPECIMEN

NUMBER

IA

2A

3A

5A

8A

SURVEY

AREA

1

2

3

A

1

2

3

A

1

2

3

A

1

2

3 A

1

2

3

4

min

6y

6*

7

6*

7.

67

5_

7

6_

7

la.

max

B. -

9

—

Bi 7 9

—

9Ì

Bi Bi —

e 7.

e. 7

■ 67

H Bi

H

av

7i

Β

lì

li

Bi

li

li

H

H

H

min

7

6

6*

6

6

e i

5¿

6

67

ài

l«r

max

e -

li —

7Ì

67

8

—

H ei H —

7

t i

e 6

7i

6* 7* 7

av

7i

Bi

ei

1

Bi

1

6

7

67

7

min

ti

Si

Si

6

6

6

6

ε τ

6

6

t

max

e -

7

—

6

6

7

—

7

6

7

—

67

6»

7

57

7

67

67

67

av

7

6

6

67

67

65

e 7

€i-

67

t-7

WELD SIZES - TYPE 'A' SPECIMENS

79

SPECIMEN NUMBER

2B

3B

AB

6B

7B

8B

ÌOB

12B

13B

1AB

SURVEY AREA

1 2 3 .

1 2 3

1 2 3

1 2 3

1 2 3

1 2 3

1 2 3

1 2 3

1 2 3

1 2 3

1_ min max av

5 9 7 5 9 7

6 9 7 6i 9 7

8i -

5i 8 6i 5i 8 6i

6. -

5i 8i 7 6 8i 7

6

5 8i 6i 5 7i 6i

5 8i 6i 5 8i 6i

8i -

5 8i 6i 5 8i 7

8i -

5 9 7 6 7i 6i

7

6 8i 7 67 87 7i

7i -

6i 9i 7 67 9i 7i

7

1«= min max av

Ai 6 5 Ai 6 5 no crack

5 67 6 5i 67 6

6

5 67 6 5 67 6

5

Ai 8 6 Ai 8 6

Ai -

Ai 7 Si Ai 6 5 no crack

Ai 67 5i Ai 5i S

5

Ai 8 6 Ai 8 67

67 -

A 7 6 A 6i 5i

5i -

Ai 8 7 6 8 7

7

Ai 7i 6 67 7i 7 - 7i -

t min max av

Ai 6 5 Ai 6 5

A 6 5 A 5i 5

5i -

Ai 6 5i Ai 5i 5

Ai -

A 67 5i A 67 5i

A

5 7 6 5 67 6

A 67 S A 6 5

Ai -

3i 7 5i A 7 Si

Ai -

Ai 7i 5i Ai 7 6

Ai -

5 8 6 5 8 6

5i -

A 6i 5 A 67 5i

6

HELD SIZES - TYPE 'B' SPECIMENS

80 -

CD

ro

Total axle»

5

4

3

2

ChaaBia typa

Articulated

Articulated

Rigid

Articulated

Rigid

Rigid

Average axla spacing., m

@ ®^> ® ® 10 1.5 5.0 1.5

®—M® @<S> io as is

@(2) <·)(·) 1.5 3.5 1.5

®-kg) <S> 15 4.5

@ WW 4.0 1.5

@ <S> (.0

Loading Group

M

L

H

M

L

H

M

L

H

M

L

H

M

L

H

M

L

Total weight, k N

360

250

335

260

145

280

240

120

215

140

90

240

195

120

135

65

30

Aale loada. kN

60 70 70

40 45 45

55 100

45 85

35 50

50 50

40 40

20 20

45 85

30 55

20 35

60

55

40

50

30

15

80

60

90

65

30

90

80

40

90

70

40

80

60

90

65

30

90

80

40

85

55

35

90

70

40

85

35

15

No. in each group par million commercial vehielaa

14 500

15000

90000

90000

90000

15000

15000

15000

30000

30000

30000

15000

15 000

15000

170000

170000

180000

Vehicle designation

5 AM

5AL

4AH

4AM

4AL

4RH

4RM

4RL

3AH

3AM

3AL

3RH

3RM

3RL

2RH

2RM

2RL

(θ) denotes 2 single tyres / axle

O r-

2.01

( φ ) denotes 2 double t y r e s / a x l e

.Q-çpair--

1.8m

CZ_ ç p a i r

TABLE 1 COMMERCIAL VEHICLE TYPES FOR FATIGUE ASSESSMENT (FROM BS5A00: PART 10?

CONNECTION TYPE

Ά'

'Β'

■

'C'

GAUGE

NUMBER

IO

13

19

49

52

67

70

82

CENTRELINE OF VEHICLE DISTRIBUTION

LONGITUDINAL

LINE t

NUMBER

3

5

7

3

5

7

3

5

7

2

11

13

15

14

11

13

15

14

19

21

23

19

21

23

24

19

21

23

18

a b c

'

1 \ *

V V V 1 ►

INSTRUMENTED TROUGH

d denotes aost damaging position

a b = d c

a b

c = d

a b c d

a b c d

a b c d

a b

c = d

a b c d

a b c d

FATIGUE LIFE ##

(YEARS)

1048 859 972

43 29 27

40 54 79 38

9 6 6 6

40 31 33 31

1530 990 877

28 17 14 14

18 24 43 17

• SEE FIG 10

· · BASED ON 1 , 0 0 0 , 0 0 0 HGVs PER ANNUM OF THE BS 5400 VEHICLES SPECTRUM. ASSUMED WELD CLASS F .

TABLE 2 PROVISIONAL FATIGUE ASSESSMENT

- 83

SPECIMEN

NUMBER

2B

3B

AB

6B

7B

8B

NOMINAL

STRESS

RANGE

σ»

(N/mm*)

9 5

1 5 0

1 2 5

1 1 5

1 0 0

2 0 0

LOAD

kN

P m ir-» — 2

ι" m a* χ ""._?__.

Pan ir-» — 2

í'max —

—' _.

¿ m ir» ~" ·

Pm.K - A 2

Γπ> ir-» — _-·

P m . x - A 2

Pmi n — 2

Pmaax 34

4 m ir» """ ·

Fm*»x — 6 2

STRESS (N/mm2)

( v e COMPRESSION)

Ο ι Os; O's σ.»

α-, *n - Ι Α + 7 - 2 - 9

σ Μ Κ - 1 0 9 - 5 1 - 7A + A

O R 9 5 5 8 7 2 1 3

Om»r, - 1A + 1 5 - 3A - 3 3

Omaax - 1 6 4 - 8 2 - 1 A 6 - 8

O R 1 5 0 9 7 1 1 2 2 5

Omln - 10 + 2 6 - 1 5 - 2A

σ „ · χ - 1 3 A - A0 - 1 1 1 - 1 5

σ η 12Α 6 6 9 6 9

Omin - 1 3 - 1 2 - 7 + 7

Oma.« - 1 2 8 - 7 6 - l O A + i e

O R 1 1 5 5A 9 7 11

Omi,-, - 1 1 - 9 - 1 1 + 3

Ornaax -111 65 90 + 1 3

O R 1 0 0 5 6 7 9 10

Omlr, 0 + 1 5 - 2 2 - AO

σ_.«χ - 2 0 0 - 9 7 - 1 7 3 - 17

O R 2 0 0 1 1 2 1 5 1 2 3

.

a ) STRESSES AT MINIMUM AND MAXIMUM ACTUATOR LOADS

DECK PANEL

2 0 kN SINGLE WHEEL STRESS

(N/mm2)

NORMALISED (TARGET) S T R E S S

RANGE

FATIGUE SPECIMENS

NORMALISED STRESS RANGE

2B

3B

AB

6B

7B

8B

σ_.«»

- 2 0 . 5

1 . 0 0

Οι

1 . 0 0

1 . 0 0

1 . 0 0

1 . 0 0

1 . 0 0

1 . 0 0

o^«s.

- 9 . 7

0 . Α 7

Ö 2

0 . 6 1

0 . 6 5

0 . 5 3

0 . Α 7

0 . 5 6

0 . 5 6

(α»»)

- 1 9 . 0

0 . 9 3

Os

0 . 7 6

0 . 7 5

0 . 7 7

0 . 8 Α

0 . 7 9

0 . 7 6

Osi

- 3 . 0

0 . 1 5

σ_>

0 . 1 3

0 . 1 7

0 . 0 7

0 . 10

0 . 1 0

0 . 1 2

b l NORMALISED STRESS RANGES

TABLE 3 STKgSSES IN TYPE ' B ' FATIGUE SPECIMENS ( 1 s t SERIES)

- 84

SPECIMEN NUMBER

2B 3B AB 6B 7B 8B

STRESS RANGE

Οι (N/mm2)

95 150 125 115 100 200

CYCLES TO

FAILURE X10 &

>11.70 0.90 2.80 1.67

>13.20 0. A7

aj. SPECIMENS TESTED AT CONSTANT AMPLITUDE

SPECIMEN NUMBER

10B 12B 13B 1AB

MAXIMUM STRESS RANGE

Oi (N/mm2)

2A5 2A5 2A5 2A5

CYCLES TO

FAILURE xioe-

2. 12 0.78 1 .86 2.75

b_l SPECIMENS TESTED AT VARIABLE AMPLITUDE

TABLE 4 FATIGUE TEST RESULTS - TYPE 'B' SPECIMENS

85

SPECIMEN

NUMBER

3B

4B

6B

ΘΒ

1.1

+ 314

-54Α

-605

-846

1.2

+ 662

- 29

- 75

-108

1.3

- 45

-341

-324

-125

2.1

+ 190

-301

-425

-357

GAUGE NUMBER

2.2 2.3

-AAI -217

-681 -159

-92A +210

-7A2 - BO

3.1 3.2 3.3

-717 -605 +122

-231 -AA7 411

- 24 -549 -908

-583-1130 -714

4.1 4.2 Α.3

+463 + 13 304

+ 81 +178 +250

+134 +120 +281

+ 7 - 7 +1A9

a) STRAINS AT END OF TEST

NOTE all strains in microstrain, tension positive

^V

3 3 3-2

gnu ) ) ) ) ) )

¿ <·

( ( ( ( ( ( ! ) ) ) . ) ) )

Γ

■ ■<%,. »1st cut

3rd cut

s»

~\

~7

^

\ >

S e c t A - A '

CONDITION

BEFORE CUTTING

AFTER 1st CUT

AFTER Ath CUT

1.1 1.2 1.3

0 0 0

+A69 +A77 -233

+A97 +A68 -218

GAUGE

2.2

0

-36A

-350

NUMBER

3.1 3.2 3.3

0 0 0

-72A -11A6 -515

-737 -1160 -513

A.2

0

-92

-85

bl RESIDUAL STRAINS SPECIMEN 5B

TABLE 5 RESIDUAL STRAINS

86

STRESS RANGE O R (N/ram*)

0-10 10-20 20-30 30-AO 40-50 50-60 60-70 70-80 80-90 90-100

CONNECTION 'A' GAUGE 13

6A772A5 A81807 355687 255613 83593 22933 A063

0 0 0

NUMBER OF CYCLES-

CONNECTION 'A' GAUGE 8β

6118059 718A70 279685 317A2A 175765 61A90 38862 3899

0 0

CONNECTION 'A' GAUGE 90

8A56SS5 5206AS 286755 100389

0 0 0 0 0 0

CONNECTIOK 'B' GAUGE A9

3119612 125A02C

~> C __. /-_._. 3027ΑΞ 14663·:. 205SS7 7821A 21212 11A0-. 1969

For 1.000.000 HGVs described in Table 1 Centre of distribution of vehicles over centreline c f trough

TABLE 6 STRESSES FROM TESTS ON DECK PANEL

- 87

œ OD

SPECIMEN

NUMBER

IA

2A

3A

5A

8A

NOMINAL

STRESS

RANGE

Ox / On

(N/mm2)

1 2 5

1 5 0

2 0 0

1 7 5

1 6 0

LOAD

kN

Lft* i r - » -~ a£_

P m . x 6 1

i m ir» ~~ c.

Prnaax " 8 0

Prnlr, + 1 3

Prnaax " 8 0

P m I n + 2

Prnaax " 8 0

t m I r - » a__

P m . x 8 0

STRESS ( N / m m 2 )

( - v e COMPRESSION)

O i O-z Ο· σ = σβ. C e Ο τ» Ο ί ο O n 0 1 2

Omir, - 1 1 - 1 2 + 1 - 2 6 - 9 - 2 O m - x - 1 3 6 + 3 9 - 9 - 1 3 9 + A 1 - 1 0 O m - x - O m i r , - 1 2 5 + 5 1 - 1 0 - 1 1 3 + 5 0 - 8 - 3 9 * - A 7 * - 5 0 * - 4 7 *

Omir, + 1 0 - 6 + 8 + 1 1 + 7 + 1 0 - 1 - 1 - 2 0 Omaax ~ 1 A 0 + 5 3 - 1 1 - 1 3 1 + 6 6 - 1 - AS - 5 6 - 6 1 - 5 6 Omaax-Omtr, - 1 5 0 + 5 9 - 1 9 - 1 A 2 + 5 9 - 1 1 - A7 - 5 5 - 5 9 - 5 6

Omir-. - 1 5 - 7 1 + 5 - 2 - 7 3 + 1 2 + 1 0 + 3 + 3 O m - x - 2 1 5 + 1 - 1A - 1 9 A + 1 - 1 6 - 7 2 - 6 7 - 6 0 - 6 0 Omaax-Omtr, - 2 0 0 + 7 2 - 1 9 - 1 9 2 + 7 4 - 2 8 - 7 3 - 6 7 - 6 3 - 6 3

Omir, - 2 + 2 4 + 3 3 + A - A 2 + 3 8 - A - 6 - 3 - A O m - x - 1 A 6 + 9 8 + 2 2 - 1 7 1 + 1 5 + 2 0 - 6 0 - 6 1 - 5 7 - 5 7 O m - x - O m i . - , - 1 A A + 7 4 - 1 1 - 1 7 5 + 5 7 - 1 8 - 5 6 - 5 5 - 5 4 - 5 5

Omir, - 6 7 - 5 3 - 1 2 - 1 1 O m - x - 2 2 7 - 2 0 A - 7 5 - 7A O m - x - O m l r , - 1 6 0 - 1 5 1 - 6 3 - 6 3

TABLE 7a STRESSES AT MINIMUM AND MAXIMUM ACTUATOR LOADS

estimated stress

00 IO

DECK PANEL 20 kN SINGLE WHEEL STRESS (N/mm*) NORMALISED (TARGET) STRESS RANGE

FATIGUE SPECIMENS Om-x-Omir, (NORMALISED)

IA 2A 3Λ SA 8A

σΐ3»

Ι Α . 7

- 1 . 0 0

Οι-»

+ 3 . 7

+ 0 . 2 5

Ο ί ο

• Α . 2

- 0 . 2 9

O í a

Ι Α . 7

1 . 0 0

Ο ι »

+ 3 . 7

+ 0 . 2 5

Ο ί ο

- Α . 2

■ 0 . 2 9

Oeei

• 5 . Α

• 0 . 3 7

Ο ι

1 . 0 0

•1 . 0 0

1 . 0 0

■0 .82

•1 . 0 0

+ 0 . Α Ι

+ 0 . 3 9

+ 0 . 3 6

+ 0 . Α 2

■ 0 . 0 8

■0. 13

• 0 . 10

0 . 0 6

- 0 . 9 0

- 0 . 9 5

- 0 . 9 6

•1 . 0 0

- 0 . 9 Α

+ 0 . Α 0

+ 0 . 3 9

+ 0 . 3 7

+ 0 . 3 3

Ο«

0 . 0 6

• 0 . 0 7

0 .1Α

• 0 . 1 0

0 . 3 1

■0.31

■0 .37

■0 .32

• 0 . 3 9

- 6 . Α

- 0 . ΑΑ

Ο ί ο

- 0 . 3 8 • 0 . 3 7 - 0 . 3 Α • 0 . 3 1

C e s

- 5 . Α

- 0 . 3 7

Ο ι ι

- 0 . Α 0 - 0 . 3 9 - 0 . 3 2 - 0 . 3 1 - 0 . 3 9

■6.Α

- 0 . Α Α

Oiat

- 0 . 3 8

- 0 . 3 7

- 0 . 3 2

■0.31

TABLE 7b NORMALISED STRESS RANGES

TABLE 7 STRESSES IN TYPE 'A' FATIGUE SPECIMENS

CD O

SPECIMEN NUMBER

IA

2A

3A

5A

8A

CRACK a

NORTH SIDE

STRESS RANGE

Oi (N/mm2)

125

150

200

1AA

160

CYCLES TO

FAILURE X10«

>12.A

5.2

1 .8

1.4

6.3

SOUTH SIDE

STRESS RANGE

O-i (N/mm2)

113

1A2

192

175

151

CYCLES TO

FAILURE X10 e

>12.4

> 6.3

> 2.5

2. 1

> 7.0

CRACK b

NORTH SIDE

STRESS RANGE

Οχ-3 (N/mm2)

137*.

16A*

219*

158*

176

CYCLES TO

FAILURE X10 &

>12.4

> 6.3 ° 3

2. 1

> 7.0

SOUTH SIDE

STRESS RANGE

Oi-j (N/mm2)

12A*

156*

210*

192*

177

CYCLES TO

FAILURE X10*

>12. A

5.8

> 2.5

1.8

> 7.0

CRACK c

NORTH SIDE

STRESS RANGE

o«. (N/mm2)

39*

A7

73

56

63

CYCLES TO

FAILURE X10*

6.3

3.0

1 .8

2.7**

2.9

SOUTH SIDE

STRESS RANGE

O n (N/mm2)

50*

59

63

5A

63

CYCLES TO

FAILURE X10*

8.2

A.O

2.A

2.9**

5.2

denotes estimated stress denotes extrapolated cycles

TABLE 8 FATIGUE TEST RESULTS - TYPE 'A' SPECIMENS

STRESS RANGE On (N/ram»)

8-12 12-16 16-20 20-2A 2A-28 28-32 32-36 36-AO AO-44 44-A8 A8-52 52-56 56-60

1

15383 A61A 1790 876 A55 229 80 21 5 0 0 0 0

2

1A680 10A08 625A 3515 2116 1698 1082 A90 177 A2 A 1 1

NUMBER OF CYCLES

3

20851 11301 5613 2236 992 325 3A 3 0 0 2 0 0

GAUGE NUMBER* A

17866 11785 36A2 7A7 227 29 0 0 0 0 0 0 0

5

1AA5A 5038 2071 1153 682 A38 330 157 60 15 2 1 0

6

16098 10518 606A 3535 2569 1600 537 127 21 9 1 2 0

7

26099 12632 6787 2650 1A63 66A 186 28 2 1 0 0 0

8

20709 16227 10103 33A1 936 518 120 17 0 0 0 0 0

See Fig A9

TABLE 9 STRESSES FROM MEASUREMENTS ON BRIDGE

91

CD

STRESS

RANGE

N/mm2

0-5

5-10

10-15

15-20

20-25

25-30

30-35

35-A0

A0-A5

A5-50

50-55

55-60

60-65

65-70

70-75

75-80

80-85

85-90

90-95

95-100

100-105

105-110

110-115

115-120

120-125

125-130

130-135

135-140

1A0-1A5

1A5-150

150-155

155-160

160-165

CYCLES

Πι

0

33986

1509A

11723

7697

A857

2A97

17A7

1298

1115

786

896

671

593

A38

A20

222

306

2A7

172

AO

1A6

71

39

57

36

16

5

19

12

0

5

1

STRESS

RANGE

N/mm2

0-10

10-20

20-30

30-A0

AO-50

50-60

60-70

70-80

80-90

90-100

100-110

110-120

120-130

130-1A0

1A0-150

150-160

160-170

170-180

180-190

190-200

200-210

210-220

220-230

230-2A0

2A0-250

CYCLES

Πι

13 59 A

31713

15A96

96A0

A787

2371

17AA

1259

1092

908

685

529

3AA

369

23A

98

1A1

57

71

3 A

9

2A

7

5

1

TABLE 11

RHEDEN SPECTRUM X 1.5

STRESS

RANGE

N/mm2

50-60

60-70

70-80

80-90

90-100

100-110

110-120

120-130

130-140

140-150

150-160

160-170

170-180

180-190

190-200

200-210

210-220

220-230

230-240

240-250 ■

CLASS

. CENTRE

N/mm2

55

65

75

85

95

105

115

125

135

1A5

155

165

175

185

195

205

215

225

235

2A5

CYCLES

ni

23710

17AA0

12590

10920

9080

6850

5290

3AA0

3690

23A0

980

1A10

570

710

3A0

90

2A0

70

50

10

TABLE 12

SPECTRUM SELECTED FOR THE STUDY

TABLE 10

ORIGINAL RHEDEN SPECTRUM

SPECIMEN

NUMBER

ÌOB

1 2 B

13B

ÌAB

NOMINAL

STRESS

RANGE

Or>

N/ram 2

2A5

2AS

2A5

2 A 5

LOAC

KN

i m A r»

* rv»a__x

* m A r»

i ' m * «

* »n A n

» m « » χ

* m A t»

* m «h χ

1

- 2

- 7 3

o

— -_· - 8 9

- 2

- 8 A

- 2

- 8 9

Omir.

O m — χ

Om

Omir,

O m x

O «

O m l η

O m x

σ«

O m In

O m — χ

Om

(

O ,

- 1

- 2 4 6

2 4 5

- 3

- 2 A 8

2 4 5

+ 1 2

- 2 3 3

2 4 5

+ 4 7

- 1 9 8

2 4 5

STRESS ( N / m m Ä )

- v e COMPRESSION)

σ » σ 3

+ 1 8 - 2 2

- 1 3 0 - 1 9 4

1 4 8 1 7 2

+ 5 - 1 9

- 1 5 4 - 2 0 7

159 ìee

+ 2 4 - 2 8

- 1 3 2 - 2 1 0

1 5 6 1 8 2

+ Α7 - Α5

- 1 0 9 - 2 3 6

1 5 6 1 9 1

Oat

- 1 9

+ 1 9

3 8

+ 1

+ 3A

3 3

- 2A

+ 2 0

AA

- 1 7

+ 1 3

3 0

a l STRESSES AT MINIMUM AND MAXIMUM ACTUATOR LOADS

DECK PANEL

2 0 KN SINGLE WHEEL S T R E S S

( N / m m 2 )

NORMALISED (TARGET) S T R E S S

RANGE

FATIGUE SPECIMENS

NORMALISED STRESS RANGE

10B

1 2 B

1 3 B

1AB

σ*.»

- 2 0 . 5

1 . 0 0

σ ι

1 . 0 0

1 . 0 0

1 . 0 0

1 . 0 0

σ _ 6

- 9 . 7

0 . A 7

σ 2

0 . 6 0

0 . 6 5

0 . 6 Α

0 . 6 Α

(σ-,,)

- 1 9 . 0

0 . 9 3

σ ο

0 . 7 0

0 . 7 7

0 . 7 Α

0 . 7 8

σ β 2

- 3 . 0

0 . 1 5

ο-»

0 . 1 6

0 . 1 3

0 . 1 8

0 . 1 2

b i NORMALISED STRESS RANGES

TABLE 13 STRESSES IN TYPE ' Β ' FATIGUE SPECIMENS (2nd SERIES)

- 93

Type'A' Ό '

Deck plate V '

CD

Welded connection under test

Type Β « Λ »

Web of trough

Type'C

F i g i Types of longitudinal/transverse stiffener connection

40-1

20

:

os ti -2<H

-40-

-60-

Gauge position

Gauge position

K H L

Stress for 20kN single wheel load Calculated stress for 32kN load

a) Measurements on test panel

Influence lines for front wheel of test vehicle -wheel load = 32kN

Bridge deck surfacing removed — Bridge deck surfaced with 38nu_

of Bastie asphalt, surfacing temperature 14 *c

p) measurements on bridge

Fig 2 TYPICAL INFLUENCE LINES - CROSSBEAM TO PECK PLATE CONNECTION

95 -

20 Stress for 20 kN single wheel load Calculated stress for test vehicle (by superposition)

Gauge position

a) Measurements on test panel

Rear 2x25kN no

Front 32kN

4-5 ι

Offside wheels of test vehicle .Gauge position

Bridge deck surfacing removed Bridge deck surfaced with 38mm of mastic asphalt, surfacing temperature 12*C

Gauge position

Influence lines for test vehicle b) Measurements on bridge

Fig 3 TYPICAL INFLUENCE LINES - TROUGH TO CROSSBEAM CONNECTION

- 96 -

CD

10-

«Γ ° -a s_ -,<H

% -20-W OS H -30H en

-40-·

Γ 1 2 TIME (MINS) 3

a) Recording showing passage of 5 HGVs (V1-V5)

VI V2

Τ 1 2 TIME (SECS) 3 b) Recording of vehicles VI and V2 with expanded tlmebase

Direction of traffic flow

Gauge position

Fig 4 TROUGH TO CROSSBEAM CONNECTION -TYPICAL INFLUENCE LINES UNDER FREE FLOWING TRAFFIC

-15.24m (50 ft) -

-3.05m (10 f t ) - -4.57m (15 f t)- -4.57m (15 f t ) - 3.05m (10 f t ) -

T >

CD 00

ε o

Γ Γ' L

- ι ,

a

, r "j LT _Ζ"_Γ___, Z Z Γ " — _ \ — — "Ζ *Ι ~ JT~— . Γ '■ Ι _. — »— —. ■ — — .— — _— —Γ^— ■_, a— — , mu _— m i — BJWU.BBa_a»m»jBBtjB_m>_BB»>»a_^^ 1»»Β»_,^Β_Β_ΒΒΒ_Β1ΒΒΒΒΒΒ_ΒΒΒΒΙ

Deckplate thickness 12.7mm (O.Sinch) ' χ Test crossbeam

Fig 5 PLAN VIEW OF DECK PANEL

40mm

DETAIL'X'

50mm

^ W

R65

DETAIL 'Y'

Full R75

Plate thickness 12i

S • 3135mm

9 pitches of 305mm

See detail X

Γ -■»♦·» + +±+_±+± +·Τ+ 4-f--f + + + + TVf- + + + + + +^+ 4.4- + + + + + + ^ . ^ .

R32

-39 pitches of 76

E E o 00 η

Fig 6 DECK PANEL - DETAILS OF TEST CROSSBEAM

8

Crossbeam to deckplate 6mm Trough to crossbeam 6mm continuous fillet welds both sides continuous fillet welds both sides

Trough to deckplate 8mm continuous fillet weld

+ + - * · + + + Ί - Η - + + + + ·»- + · Ι - · ! - + + + + + + ■+ ■+·

Fig 7 DECK PANEL - TEST CROSSBEAM WELD DETAILS

. Trough thickness 6.35mm

40 bolts M22

_ .

1

Cross girder web thickness 10mm

4580mm·

5000mm

tin r il

1

Ì ε ε

CD CO CO

1

Fig 8 DECK PANEL - TEST CROSSBEAM SUPPORT CONDITIONS

o ro

Reaction frame

Cross girder support

2 Fig 9 STATIC LOAD TEST RIG

Single static wheel load

Extent of test

3353 mm 2286 mm

ι »*|«« lá

scale appro«.. 1:150

Γ Γ I 3 ^■~ Test crossbeam

Longitudinal section of panel

Lines at

304.8 mm spacing *4«

Lines at

152.4 mm spacing

Lines at

. 304.8 mm spacing .

8

I

en L U ro Q.

ε b

ΓΜ

<P Γ^

TO

C

_ l

'

i

f

F

TO α»

X l .Λ ΙΛ

O υ tf v t O)

1

X)

ε

XI

C o

i n r^ .— r~

r~ CO

η CD

cn t n

m i n ·— r^

t n r c n r ^ i n m r o i r ^ i n m r c n ar * n c o r o n n r M C M r s j r v i t N r

Transverse line numbers (T)

E

Ë

o

c o

Scale approx. 1:30

Fig 10 LOADING GRID FOR STATIC TESTS ON DECK PANEL

a) Gauge numbers

b) Gauge pos i t ions

F i g 11 STRAIN GAUGE INSTALLATION - CONNECTION 'A'

- 104

» „ χ

a) Wheel over trough and 1220am from crossbeam (position L7 TIS - see Fig 10)

Stress in N/_un= -ve compression

b) Wheel over trough and crossbeam (position L7 T31 - see Fig 10)

Fig 12 DISTRIBUTION OF STRESS AROUND CONNECTION 'A'

105 -

a) Gauge numbers

b) Gauge positions

Fig 13 STRAIN GAUGE INSTALLATION - CONNECTION 'B'

106 -

*<±r'ZÏ

a) Wheel over trough and 1220mm from crossbeam (position LIS T15 - see Fig 10)

Stress in N/mm= -ve compression


Fig 14 DISTRIBUTION OF STRESS AROUND CONNECTION 'B'

107

a) Gauge numbers

b) Gauge positions

Fig 15 STRAIN GAUGE INSTALLATION - CONNECTION 'C'

- 108

^ > ,

a) Wheel over trough and 1220mm from crossbeam (position L23 T15 - see Fig 10)

Stress In N/mm3 -ve compression

' « * -.*


Fig 16 DISTRIBUTION OF STRESS AROUND CONNECTION 'C'

109

TRANSVERSE INFLUENCE LINE 27

101

a \

z w VÌ w κ Η

W

0

LINE 6

10 I GAUGÍ POSITION

ΙΟi

1

w (X H

LONGITUDINAL INFLUENCE LINE 6

LINE 27

20 kN SINGLE WHEEL LOAD

TEST CROSSBEAM

TO J

Fig 17 INFLUENCE LINES GAUGE 10


-20-J 20 kN SINGLE WHEEL LOAD


LINE 15

Τ

-20 J

Fig 18 INFLUENCE LINES - GAUGE 13


ro LONGITUDINAL INFLUENCE LINE 5

20-

3 ,„ N 10-

κ H f 01 C

- , . -

TEST CROSSBEAM




10-, LINE 15

305r


ω LONGITUDINAL INFLUENCE LINE 15

10 -,

1 °

LINE 15

Ζ

w in 05-10-1

20 -J

TEST CROSSBEAM

Fig 20 INFLUENCE LINES - GAUGE A9

TRANSVERSE INFLUENCE LINE A3

-20 J 20 kN SINGLE WHEEL LOAD


-20



IO-,

Ζ

-io -i

LINE 22

GAUGE POSITION

305-nm al Β»

cr.



10-,

3 Ζ

ίο - i o -J

LINE 27

TEST CROSSBEAM


TRANSVERSE INFLUENCE LINE 15 lO-i

-20 J

LINE 24


σ> LONGITUDINAL INFLUENCE LINE 24 10 -i

1 ° Ζ § - io

-20 -J

lm LINE 15

I -*"Y .TEST

\ CROSSBEAM



Ζ

EXTRAPOLATED DATA

GAUGE POSITION

LONGITUDINAL INFLUENCE LINE 21 20 kN SINGLE WHEEL LOAD

TEST CROSSBEAM

-10 -J



20

10

Νβ 0 β

*10

æ w on

05 20

Η

w 30 40

50

LINE 6

GAUGE POSITION

20 kN SINGLE

WHEEL LOAD

20

10

\ °

H

α 20

30

40

50


LINE 31

TEST CROSSBEAM

Fig 25 INFLUENCE LINES GAUGE 88

TRANSVERSE INFLUENCE LINE 31 LINE 6


\ Ζ 8. «

20

10

0

-10

α! -20

-30

-40

-50


LINE 31

TEST CROSSBEAM



-20-J 20 kN SINGLE WHEEL LOAD

o τοΜπττίίηΤΝΑΙ. INFLUENCE LINE 5

10-,

"a ° ζ

κ - i o Η w -20 J

LINE 27

TEST CROSSBEAM



-20-1 20 kN SINGLE WHEEL LOAD

ro LONGITUDINAL INFLUENCE LINE 5

10 -,

a « ζ

06 - 1 0 Η

-20 -J

LINE 35 lm

TEST CROSSBEAM


15 1 Percentage of total number of vehicles

10-

N)

5-

ç rear wheels

Nearside wheels

Distribution of vehicles centred over trough 'Β'

(note: location of vehicles defined by position of front nearside wheels)

Fig 29 DISTRIBUTION OF VEHICLES ACROSS DECK

50·

40·

30

20·

10·

0·

Calculations based on: (i) mean-2s.d. S-N data for

class F welds (ii) 1,000.000 HGVs / year

(iii) vehicle types fro« BS5400 (see Table 1 of this report)

Gauge 13

in ce < w

ω [14

Η < Cu

50- ,

4 0 -

30·

J 20-M

10-

o-J

ς of distribution of nearside front wheels

Gauge 49

50-.

40-

30-

20-

10-

0-

Gauge 70

Fig 30 VARIATION OF FATIGUE LIFE WITH POSITION OF VEHICLES

123 -

Hydraulic actuator

Reaction frame

F i g 3 1 FATIGUE TEST RIG

M1843

124 -

South side

North side

Fig 32a STRAIN GAUGE POSITIONS - TYPE 'B' FATIGUE SPECIMENS

125 -

ro

ï

ELEVATION

1^

4> --£ΞΕ]

^V-

-Λ/-

FT

Η

1> ©

Trough

VIEW ON A-A'

Flg_32b STRAIN GAUGE POSITIONS -

ΤΫΓΕ 'Β' FATIGUE SPECIMENS

VIEW ON B-B'

53

Fatigue crack — I — Crossbeam

Fig 33 SECTION AT APEX OF TROUGH - TYPE 'B' SPECIMEN

CONDITION

Crack first observed · Defined failure « End of test °

LENGTH mm

10 25 86

CYCLES X10*

0.78 0.90 1.09

150

00

crack Initiation point

100

I ΙΟ Ζ. w J !_ υ «: κ υ 50

10 _1_ _ι ι ι ι ι ι

10 CYCLES 10

Fig 34 CRACK DEVELOPMENT - SPECIMEN_3B

CONDITION

Crack first observed · Defined failure » End of test o

LENGTH ram

14 25 61

CYCLES X10*

2.56 2.80 4.43

150

Β

100

I ίο 2 ω y. υ < κ υ 50

10 crack

initiation point

_1 ■ > — I ι I _1 ι ι ι ι

10- CYCLES 10

Fig 35 CRACK DEVELOPMENT - SPECIMEN 4B

g

CRACK

a

b

CONDITION

Crack first observed Defined failure End of test

Crack first observed Defined failure End of test

• X

o

• X

O

LENGTH mm 11 25 79

14 25 43

CYCLES X10A

1.40 1.67 2.93

1 .40 2.05 2.93

Fatigue initiation point

150

100

X H O Ζ ω

υ < κ υ

50

• a

/ /b

10 10 CYCLES 10

Fatigue crack b initiation point

Fig 36 QRACJ^ DEVELOPMENT - SPECIMEN 6B

CONDITION

Crack f i r s t observed · Defined fai lure » End of t e s t o

LENGTH

■■

11

2 5

1 1 4

CYCLES

X 1 0 *

0 . 4 3

0 . 4 7

0 . 6 8

150

100

x H Ü Ζ W J

y. υ < κ υ

50

Ι

ι • I ■ • ■ I I I

■ · Ι ι . ι > ·

10 10 CYCLES 10

crack initiation point

Fig 37 CRACK DEVELOPMENT SPECIMEN 8B

ω ro

150

100

STRESS RANGE (N/ae2)

50

10"

Stress range

Mean stress

Crack first observed

ι I I I I I I I

Type *B' spec imens Typ ica l o u t p u t from gauge 1

y\

/ y

ι f l — l l I I I I

Load off (Residual stress)

10 10 ' M l 1-150

-50

MEAN STRESS (N/mm2)

-100

10' CYCLES

Fig 38 EFFECT OF CRACKING

1000

STRESS RANGE Ν/ι·«*

100

8

MEAN - 2 s.d. S-N

1 Eurocode class 2 BS5400 class C 3 Eurocode class 4 BS5400 class E 5 Eurocode class 6 DS5400 class G

CURVES:

125

00

50

CYCLES

Fig 39 DS5/.00 AND EUROCODE S-N CURVES

1000

ω

STRESS RANGE N/mm3

100

CYCLES

Fig 40 FATIGUE TESTS AT CONSTANT AMPLITUDE - TYPE 'Β' SPECIMENS

South side

North «ide

Fig 41a STRAIN GAUGE POSITIONS - TYPE 'A' FATIGUE SPECIMENS

- 135

Deck plate

Trough to deck plate weld

( ( ( ( ( ( ( ( T T T T

Web of trough

Crossbeam

Trough

Crossbeam

Crossbeam

Web of trough

Fig 41b STRAIN GAUGE POSITIONS - TYPE 'A' FATIGUE SPECIMENS

136

South side

Gauge numbers in brackets refer to gauges on other side of crossbeam

Noah side

Fig 42 STRAIN GAUGE POSITIONS - FATIGUE SPECIMEN 8A

- 137 -

South side

North side

Fig 43a END OF TEST CRACKS - SPECIMEN IA

138

NORTHSIDE SOUTHSIDE

150

ë

CRACK

C

d

COrtDITIOM

Crack first observed · Defined failure « Em! of teat o

Crack firat observed · Defined failure » Ετκ) of teat o

LEWCTH

26 25 SO

17 2S 39

CYCLES X10

6.64 6.30 12.39

11.11 11.90 12.39

100

¡5

ì υ

50

J I I Ι I I I I t t I I I I I I

150

10' IO 7

100

Si

50

CRACK

c

d

COBTDITIOrt

Crack firat observed · Defined failure » Ind of test o

Crack first observed · Defined failure χ End of test o

LENGTH

17 25 41

8

2S

13

CYCLES X10

6.64 6.20 12.39

11.11

12.39

J I I t l I I J t i l l

CYCLES 10' 1 0 ' 10 ' CYCLES 10'

F i g 43b CRACK DEVELOPMENT - SPECIMEN 1A

South side

North side

Fig 44a End of test cracks - Specimen 2A

140

NORTHSIDE SOUTHSIDE

150

CRACK

a

c

d

CONDITIO*

Crack firat observed ·

Defined failure χ

End of teat o

Crack flrat observed ·

Derined failure "

End of test r


Defined failure a

End of test O

L IMC ΤΗ

aa

10

2S

16S

S

25

SS

4

25

71

CYCLES

ΧΙΟ*

4.63

5.20

6.33

0.74

3.00

6.33

3.12

5.60

6.33

100

¡5

g G

50

165

10'

150

100

CRACK

b

c

d

CONDITION

Crack flrat observed Φ

Derined failure χ End or tast o


Defined failure χ

End of test o

Crack first observed ·

Defined rallure »

End of teat o

LEHCTH

as

4

25

68

6

2S

47

6

25

39

CYCLES

XI0·

3.12

5.80

6.33

0.74

4.00

6.33

3.38

6.20

6.33

Fi_L_.44b CRACK DEVELOPMENT _ SPECIMEN 2A

South side

North side


142

NORTHSIDE SOUTUSIDK 150

100

64 <

CRACK

a

b

c

CONDITIO«)

Crack flrat obear»ed a Derined fsllure χ End of test o

Crack flrat observed 8 Defined failure End of teat c

Creek first obsarved · Defined failure χ End of test o

LENGTH

7 25

168

3 25 65

10 25 44

CYCLES X10·

0.64 1.80 2.47

0.40 2.30 2.47

0.15 1.80 2.47

168 150

100

50

CRACK

b

c

CONDITION

Crack flrat observed · Derined failure " End or test o

Crack first observed · Defined failure χ End of test O

LENGTH

5 25 7

12 25 26

CYCLES XIO»

0.72

2.47

1.16 2.40 2.47

J L—I 1 l—L-l 10'

Fig 45b CRACK DEVELOPMENT - SPECIMEN 3A

South side

North side


144

NORTHSIDE SOUTHSIDE

150 CRACK

S

b

c

CONDITION

Crack firat observed 8 Defined failure χ End of teat o Crack flrat observed 8 Defined failure « End or teat Crack first observed 8 Defined failure « End of test o

LENGTH an 6 25 106 4 25 39 β 25 20

CYCLES ΧΙΟ· 0.84 1.3S 2.15 0.84 2.10 2.15 0.84 2.70 2.IS

100

9

5

b ol υ

50

_L_.I !_ io3

.1 J l_-I_I_l_l

150

100

CRACK

a

b

c

CONDITION

Crack flrat observed 8 Defined rallure χ End or test o Crack rtrst observed 8 Derined rallure χ End or test o Crack first observed · Defined failure χ End or test O

LENCTH

6 25 32 3

25 76 6 25 18

CYCLES ΧΙΟ* 0.84 2.10 2.15 0.55 1.80 2.15 0.82 2.90 2.15

CYCLES 107

Fig_4_6b CRACK_pEVELOPMEN_T_- SPECIMEN 5Λ

South side

North side

Fig 47a END OF TEST CRACKS - SPECIMEN 8A

146

NORTHSIDE SOUTHSIDE

150

CRACK

a

c

d

CONDITION

Crack first observed a Defined failure a End or test o

Crack first observed a Deflnerl rallure a End or test 0

Crack flrat observed s Defined failure a End of test o

LENGTH

5 25 99

8 25 67

15 25 148

CYCLES X10»

3.58 6.30 7.03

0.38 2.90 7.03

4.68 5.90 7.03

150

100 a β

ι Η Ü Ζ

w

a υ < cc υ

50

a

c

d

CONDITION

Crack flrat ohearved e Defined rallure a End of teat o

Crack first observed e Derined rallure χ End or test 0

Crac» flrat observed « Derined rallure a End or test o

LfBCTH

8 2S 13

3 25 40

5 25 10

CYCLES ΧΙΟ·

2.82

7.03

0.53 5.20 7.03

6.36

7.03

Fig 47b CRACK DEVELOPMENT SPECIMEN 8A

1000

è

STRESS RANGE N/mm

2

100

10

• " •a , " " " - » - ^ ^ . ^

^ ^ ■ " - ^ S , . ^ * 4 .

I 1 f i l i l i

^ " O ^ ^ O B

^. — ^

■ ^ — * .

^~~""^\*^ *"* ""* ^

"** ~v X

I I I I 1 I I 1

Mean 1 i ne _ "?5Λ c o n f i dence

1 i m i t s

o c r a c k a • c r a c k b χ c r a c k c

fc~ *

I 1 l i l i l í

. E u r o c o d e c l a s s 125

E u r o c o d e c l a s s 50

Ι Ι l i l i l í

IO5

10' io7

10' ioy

CYCLES

Fig 48 FATIGUE TESTS AT CONSTANT AMPLITUDE - TYPE 'A' SPECIMENS

<x

Gauge numbers in brackets refer to gauges on other side of crossbeam

Fig 49 STRAIN GAUGE POSITIONS ON BRIDGE

149

en o

4 8 - ,

Ρ Ε 3 8 -

R

C

E

Ν

Τ

Α

G 18-

Ε

Ζ β -

θ-^

■■m

jiff.;

m-.m

I "ri

if... ϋ 4 i

; * « ■

. · . - ■ "

v**; ¡frrç .¿ ;¡i

¡ «

ί : · *

!«t . « í

'??. ■VU!

%

.*< >'*<

;?<*.!

i?·"

5Í 4Í _.«__;aft_.!^__if«a3^]î ^DH-JblJ

damage :%ΐ eye les

jLñ.ÜLDLOL.o!.-i...DL..[._L_; 15 25 35 45 55 65 75 85 95 105 115 125 135 145 155 165 175 185 195 205 215 225 235 245

STRESS RANGE (CLASS CENTRE) N/W

Flg 50 DISTRIBUTION OF DAMAGE AND CYCLES (FROM TABLE 12)

ζ

w ai ae M

o

Τ" 10 Υ. TINE (SECONDS) -0

Fig 51 TYPICAL OUTPUT FROM CHART RECORDER

CONDITION

Crack first observed · Defined failure * End of test °

LENGTH mm

16 25 91

CYCLES XI0*

2.07 2.12 5.62

150

Ol

ro

100

I ΙΟ Ζ W

υ < υ

50

■ ι ι ι t l i l i 1 1

10 10° CYCLES 10

crack initiation point


CONDITION

Crack first observed · Defined failure a End of test °

LENGTH

mm

32 25 132

CYCLES ΧΙΟ*

0.92 0.78 2.94

150

S

100

I H

o

Ζ ω j s¿ υ < κ υ

50

■

/ /

/ / /

. Ι Ι ■ ι Ι ι ι Ι Ι l i l i l í

10 10 CYCLES 10



CONDITION

Crack first observed · Defined failure * End of test »

LENGTH mm 9

25 202

CYCLES X10* 1.44 1.86 5.95

150

2

100

X ίο ζ w -1 ii υ < κ υ 50

■

■

A 202

I

1

j

ι ■ ι 1 · ι ι ■

10 10 CYCLES 10


Fig 54 CRACK DEVELOPMENT - SPECIMEN 13B

s

CRACK

a

b

CONDITION

Crack f i r s t observed · Defined f a i l u r e * End of t e s t o

Crack f i r s t observed · Defined f a i l u r e * End of t e s t o

LENGTH ■■

3 25

149

34 25 8 3

CYCLES X10*

1 . 2 5 2 . 7 5 7 . 1 2

2 . 2 5 2 . 0 0 7 . 1 2

150

Fatigue crack a Initiation point

100

s X ΙΟ ζ ω j

s¿ υ < κ υ

50

•

■

• I I · B i l l

/ *

/ s^*

I /

' / / /

* * ^ 1 I I I · · · ·

10 10° CYCLES 10

Fatigue crack b Initiation point

Fig 55 CRACK DEVELOPMENT SPECIMEN 1AB

s

1000

500

STRESS RANGE (N/mm

3)

100

50·

10

SN CURVES FOR EUROCODE CLASS 80

_i 1 1 1 1

0 01 02

ηι/Ση·

APPLIED SPECTRUM

ï 1—ι—ι ι I I I

X

Mean line

95% confidence limits

Tests at constant amplitude Tests at variable amplitude

ï 1 1—ι—IIII τ 1—ι—I I I I I

10' ι ο ίο' 10e τ 1 — ι — I I I I I

10 CYCLES

Fig 56 TESTS ON TYPE 'Β' SPECIMENS

European Communities — Commission

EUR 12792 — Measurement and interpretation of dynamic loads in bridges Phase 3: Fatigue behaviour of orthotropic steel decks of road bridges

C. Beales

Luxembourg: Office for Official Publications of the European Communities

1990 — VI, 156 pp., tab., fig. — 21.0 x 29.7 cm

Technical steel research series

EN

ISBN 92-826-1505-7

Catalogue number: CD-NA-12792-EN-C

Price (excluding VAT) in Luxembourg: ECU 12.50

Fatigue failures have occurred in the orthotropic decks of bridges after less than 20 years in service. One welded connection to have suffered from premature fatigue cracking is the early design of connection between the longitudinal and transverse stiffeners. The objectives of this research were to assess the fatigue performance of more recent designs of connection and, if appropriate, to suggest design improvements. The research involved the static load testing of a full-scale bridge deck panel and the measurement of strains around three different designs of connection. Specimens representing two of the connections were tested under constant amplitude fatigue loading and one type was tested under variable amplitude loading. Measurements were also made on a bridge under traffic loading to assess the effect of the bridge deck surfacing. It is concluded that none of the three designs of connection assessed, and currently in use in major European bridges, meets the 120-year design life required for UK bridges when assessed by the BS5400 code of practice for fatigue. However, this assessment excludes the effect of the bridge deck surfacing which is expected to increase the life of the connection in service. An 'Roptimized' design is suggested but will require a similar testing programme to assess its performance.

Venta y suscripciones * Salg og abonnement · Verkauf und Abonnement * Πωλήσεις και συνδρομές Sales and subscriptions * Vente et abonnements · Vendita e abbonamenti

Verkoop en abonnementen · Venda e assinaturas

BELGIQUE / BELGIË

Moniteur balga /

Belgisch Staatsblad

Rue de Louvam 42 / Leuvenseweg 42

1000 Bruxelles/ 1000 Brussel Tel (02)512 00 28 Fax 51 1 01 84 CCP / Postrekening 000200550227

Autres distributeurs / Overige verkooppunten

Librairie •uropémnnm/ Europaaa Boekhandel

Avenue Albert Jonnart 50 / Alben Jonnartlaan 50 1200 Bruxelles / 1200 Brussel Tél. (02) 734 02 81 Fax 735 08 60

Jean De Lannoy

Avenue du Roi 202 /Koningslaan 202 1060 Bruxelles/ 1060 Brussel Tél. (02) 538 51 89 Télex 83220 UNBOOK Β

CREDOC

Rue de la Montagne 34 / Bergstraat 34

Ble 11 / Bus 11 1000 Bruxelles / 1000 Brussel

DANMARK

J. H. Schultz Information A/S

EFPubllkatloner

Ottillavei 18

2500 Valby TH. 38 44 22 88 Fax 36 44 01 41 GiFOkonto 6 00 08 86

BR DEUTSCHLAND

Bundesanzeiger Verlag

Breite SlraDe Postfach 10 80 06 5000 Köln 1

Tel (0221) 20 290 FernschreiberANZEIGER BONN 8 882 595 Fax 20 29 278

GREECE

O.C. Eleftheroudakla SA

International Bookslore

Nikis Street 4

10563 Athens

Tel (01) 322 63 23

Telex 219410 ELEF

Fax 323 98 21

ESPANA

Boletín Oficial dal Estado

Trafalgar. 27 28010 Madrid Tel (91) 448 80 00

MundlPrensa Libros. S.A.

Castello. 37 28001 Madrid Tel (91) 431 33 99 (Libros)

431 32 22 (Suscripciones) 435 36 37 (Dirección)

Télex 49370MPLIE Fax (91)275 39 98

Sucursal

Librarla Internacional AEDOS

Conseto de Ciento. 391

080O9 Barcelona Tal (93) 301 86 15 Fax (93) 317 01 41

Generalität de Catalunya

Llibreria Rambla dala aatudis

Rambla. 118 (Palau Moia)

08002 Barcelona

Tel. (93) 302 88 35

302 84 62

FRANCE

Journal officiel Service dea publications des Communautés européennes

26. rue Desaix 75727 Pans Cedex 15 Tél. (1)40 58 75 00 Fax (1) 40 58 75 74

PORTUGAL OSTERREICH

IRELAND

ITALIA

NEDERLAND

SDU uitgeverij

Christofe! Plantitnstraat 2 Postbus 20014

2500 EA 'sGravenhage Tel (070) 78 98 80 (bestellingen) Fax (070) 47 63 51

Imprensa Nacional

Casa da Moeda. EP Rua D. Francisco Manuel de Melo. 1092 Lisboa Codex Tel. (01) 69 34 14

Distribuidora de Livros

Bertrand, Ld."

Grupo Bertrand, SA

Rua das Terras dos Vales, 4A

Apartado 37 2700 Amadora Codex Tel. (01) 493 90 50 494 87 88 Telex 15798 BERDIS Fax 491 02 55

Government Publications Sales Office

Sun Alliance House Molesworth Street Dublin 2 Tel. 71 03 09

or by post

Government Stationery Office

EEC Section

6th floor Bishop Street Dublin 8 Tel. 78 18 66 Fax 78 06 45

UNITED KINGDOM

H M SO Booka (PC 18)

HMSO Publications Centre

51 Nine Elms Lane

London SW8 5DR

Tel. (071) 873 9090

Fax GP3 873 8463

Subagent:

Alan Armstrong Ltd

2 Arkwnght Road

Reading, Berks RG2 OSO

Tel. (0734) 75 18 55

Telex 849937 AAALTD G

Fax (0734) 75 51 64

Licosa Spa

Via Benedetto Fortini. 120/10

Casella postale 552 50125 Firenze Tel. (055) 64 54 15 Fax 64 12 57 Telex 570466 LICOSA I CCP 343 509

Subagente

Libreria scientifica Lucio de Blaaio AEIOU

Via Meravigli, 16 20123 Milano Tel. (02) 80 76 79

Herder Editrice e Libreria

Piazza Montecitorio, 117120

00186 Roma

Tel. (06) 679 46 28/679 53 04

Libreria giuridica

Via 12 Ottobre, 172/R 16121 Genova Tel (010) 59 56 93

GRANODUCHE DE LUXEMBOURG

Abonnements seulement Subscriptions only NUF fuF Abonnements

Messageries Paul Kraus

11. rue Christophe Plantin

2339 Luxembourg

Tei 499 88 88

Télex 2515

CCP 4924263

Reriouf Publishing Co. Ltd

Maif orders — Head Office:

1294 Algoma Road

Ottawa, Ontario K1B 3W8

Tel. (613) 741 43 33

Fax (613) 741 54 39

Telex 0534783

Ottawa Store:

61 Sparks Street

Tel. (613) 238 89 85

Toronto Store:

211 Yonge Street

Tel. (416) 363 31 71

JAPAN

Kinokuniya Company Ltd

177 Shiniuku 3Chome Shiniukuku Tokyo 16091 Tel. (03) 354 01 31

Journal Department

PO Box S5 Chitóse

Tokyo 156

Tel. (03) 439 01 24

MAGYAR

Agroinform

Kozpont

Budapest I Amia ut 93. H1012

Leveicim

Budapest. Pf 15H1253

Tel 36 ( I I 56 82 11 Telex (22) 4717 AGINF H61

Manz'ache Verlags

und Universitatsbuchhandlung

Kohlmarkt 16 1014 Wien Tel. (0222) 531 610 Telex 11 25 00 BOX A Fax (0222) 531 6181

SCHWEIZ / SUISSE / SVIZZERA

OSEC

Stampfenbachstraße 85 8035 Zürich Tel. (01) 365 51 51 Fax (01) 365 54 11

SVERIGE

BTJ

Box 200 22100 Lund Tel. (046) 18 00 00 Fax (046) 18 01 25

TURKIYE

Dünya super veb ofset A.S.

Narlibahçe Sokak No. 15 Cagaloglu Istanbul Tel. 512 01 90 Telex 23822 DSVOTR

UNITED STATES OF AMERICA

UNIPUB

4611F Assembly Drive Lanham. MD 207064391 Tel. Toll Free (800) 274 4888 Fax (301) 459 0056 Telex 7108260418

YUGOSLAVIA

Privrednlvjesnlk

Rooseveltov Trg 2

41000 Zagreb

Tel. 44 64 28

44 98 35

43 32 80

44 34 22

Teleks 21524 YU

AUTRES PAYS OTHER COUNTRIES ANDERE LÄNDER

Office des publications officielles

des Communautés européennes

2. rue Mercier

L2985 Luxembourg

Tél. 49 92 81 Téiex PUBOF LU 1324 b Fax 48 85 73 CC bancaire BIL 8109/6003/700

O)

NOTICE TO THE READER

All scientific and technical reports published by the Commission of the European Communities are announced in the monthly periodical 'euro abstracts'. For subscription (1 year: ECU 84) please write to the address below.

o

Price (excluding VAT) in Luxembourg: ECU 12.50 ISBN IS-f lSb- lSDS-?

* op * OFFICE FOR OFFICIAL PUBLICATIONS OF THE EUROPEAN COMMUNITIES

L-2985 Luxembourg 9 '789282"615058

Documents

Fatigue Behaviour of Orthotopic Steel Decks of Road Bridges