Upload
paikjg
View
44
Download
0
Embed Size (px)
DESCRIPTION
.
Citation preview
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
UK BRIDGES WITH ORTHOTROPIC STEEL DECKS
BRIDGE NAME Severn DATE OF COMPLETION 1966 BRIDGE TYPE Suspension MAIN SPAN LENGTH 988 metres SURFACING WATERPROOFING MEMBRANE Rubber bitumen
WEARING COURSE Hand laid mastic asphalt NOMINAL THICKNESS 38mm
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
LONGITUDINAL TO TRANSVERSE STIFFENER CONNECTION Crossbeam continuous. Troughs butted up to crossbeam and fillet welded all round with 6.5mm continuous welds.
114 305
381
102
6 4
\-m p. 89
NOTE all dimensions in millimetres
35
UK BRIDGES WITH ORTHOTROPIC STEEL DECKS
BRIDGE NAME Wye DATE OF COMPLETION 1966 BRIDGE TYPE Cable stayed MAIN SPAN LENGTH 235 metres SURFACING WATERPROOFING MEMBRANE Rubber bitumen
WEARING COURSE Hand laid mastic asphalt NOMINAL THICKNESS 38mm
DECK PLATE THICKNESS 11.4mm LONGITUPINAL STIFFENERS
TRANSVERSE STIFFENERS
SHAPE Trapezoidal PLATE THICKNESS 6.4mm WIDTH 305mm DEPTH 229mm SPACING (BETWEEN CENTRE LINES) PLATE THICKNESS 8mm DEPTH 394mm SPACING 4267mm
610mm
LONGITUDINAL TO TRANSVERSE STIFFENER CONNECTION Crossbeam continuous. Troughs butted up to crossbeam and fillet welded all round with 6.5mm continuous welds.
394
102
NOTE all dimensions in millimètres
75
- 36
UK BRIDGES WITH ORTHOTROPIC STEEL DECKS
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
TRANSVERSE 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
NOTE all dimensions in millimetres
- 37 -
UK BRIDGES WITH ORTHOTROPIC STEEL DECKS
BRIDGE NAME Avonmouth DATE OF COMPLETION 1974 BRIDGE TYPE Viaduct MAIN SPAN LENGTH 174 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 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 /
NOTE all dimensions in millimetres
38
UK BRIDGES WITH ORTHOTROPIC STEEL DECKS
BRIDGE NAME Cleddau DATE OF COMPLETION 1975 BRIDGE TYPE Viaduct MAIN SPAN LENGTH 213 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 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
NOTE all dimensions in millimetres
- 39
UK BRIDGES WITH ORTHOTROPIC STEEL DECKS
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
NOTE all dimensions in millimetres
40
UK BRIDGES WITH ORTHOTROPIC STEEL DECKS
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
TRANSVERSE 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
NOTE all dimensions in millimetres
- 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-"-^ii
ri
s τ« W
φ φ
r* "rl·--4»
u κ
κ nm, t^a ·" *>
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
THIRD ANGLE PROJECTION
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 * «
THIRD ANGLE PROJECTION
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
THIRD ANGLE PROJECTION
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«>c^a>.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 "
PiMtOOSie^i 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' ft^oJkí/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.¿^iAl\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
CONSULTING. INSPECTING
AND TESTING ENGINEERS
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
CONSULTING. INSPECTING
AND TESTING ENGINEERS
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 *
Stage Result Comments
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 rVO^aDriARt
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 ^AfLrVHtwjs» 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
CONSULTING. INSPECTING
AND TESTING ENGINEERS
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 ■v^A
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 ^OCTvo^ 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 -
CONSULTING. INSPECTING
AND TESTING ENGINEERS
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 *
Stage Result Comments
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
CONSULTING. INSPECTING
AND TESTING ENGINEERS
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.îr^OuSraALrO.^unteAfîiîûlJrrl 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^.Vi^urAL· £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£ V^ATÉ'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 2abfà l\* * *>T»4 —
Further visits required? j l ^ N o Inspector's Signalure
¿Æ t €>/&/%*
Inspection Authority Signature
71 -
CONSULTING. INSPECTING
AND TESTING ENGINEERS
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 *
Stage Result Comments
^ 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
AND TESTING ENGINEERS
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^l^AiiiaHljdSurfaceCond 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_a^A--_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-^UtiuTAAHL·
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.\ DCΣ 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
b) Wheel over trough and crossbeam (position L15 T31 - see Fig 10)
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
' « * -.*
b) Wheel over trough and crossbeam (position L23 T31 - see Fig 10)
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
TRANSVERSE INFLUENCE LINE 15
-20-J 20 kN SINGLE WHEEL LOAD
LONGITUDINAL INFLUENCE LINE 8
LINE 15
Τ
-20 J
Fig 18 INFLUENCE LINES - GAUGE 13
TRANSVERSE INFLUENCE LINE 27
ro LONGITUDINAL INFLUENCE LINE 5
20-
3 ,„ N 10-
κ H f 01 C
- , . -
TEST CROSSBEAM
20 kN SINGLE WHEEL LOAD
Fig 19 INFLUENCE LINES - GAUGE 19
TRANSVERSE INFLUENCE LINE 15
10-, LINE 15
305r
20 kN SINGLE WHEEL LOAD
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
LONGITUDINAL INFLUENCE LINE 15
-20
Fig 21 INFLUENCE LINES - GAUGE 52
TRANSVERSE INFLUENCE LINE 27
IO-,
Ζ
-io -i
LINE 22
GAUGE POSITION
305-nm al Β»
cr.
LONGITUDINAL INFLUENCE LINE 22
20 kN SINGLE WHEEL LOAD
10-,
3 Ζ
ίο - i o -J
LINE 27
TEST CROSSBEAM
Fig 22 INFLUENCE LINES - GAUGE 67
TRANSVERSE INFLUENCE LINE 15 lO-i
-20 J
LINE 24
20 kN SINGLE WHEEL LOAD
σ> LONGITUDINAL INFLUENCE LINE 24 10 -i
1 ° Ζ § - io
-20 -J
lm LINE 15
I -*"Y .TEST
\ CROSSBEAM
Fig 23 INFLUENCE LINES - GAUGE 70
TRANSVERSE INFLUENCE LINE 27
Ζ
EXTRAPOLATED DATA
GAUGE POSITION
LONGITUDINAL INFLUENCE LINE 21 20 kN SINGLE WHEEL LOAD
TEST CROSSBEAM
-10 -J
Fig 24 INFLUENCE LINES - GAUGE 82
TRANSVERSE INFLUENCE LINE 31
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
LONGITUDINAL INFLUENCE LINE 6
LINE 31
TEST CROSSBEAM
Fig 25 INFLUENCE LINES GAUGE 88
TRANSVERSE INFLUENCE LINE 31 LINE 6
20 kN SINGLE WHEEL LOAD
\ Ζ 8. «
20
10
0
-10
α! -20
-30
-40
-50
LONGITUDINAL INFLUENCE LINE 6
LINE 31
TEST CROSSBEAM
Fig 26 INFLUENCE LINES - GAUGE 89
TRANSVERSE INFLUENCE LINE 27
-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
Fig 27 INFLUENCE LINES - GAUGE 90
TRANSVERSE INFLUENCE LINE 35
-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
Fig 28 INFLUENCE LINES - GAUGE 91
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
Crack flrat observed ·
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
Crack flrat observed ·
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
Fig 45a End of test cracks - Specimen 3A
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
Fig 46a End of test cracks - Specimen 5A
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
Fig 52 CRACK DEVELOPMENT SPECIMEN 10B
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
Fatigue initiation point
Fig 53 CRACK DEVELOPMENT SPECIMEN 12B
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
Fatigue initiation point
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