offshore technology report 2000/078 · 2019. 12. 13. · fatigue behaviour of the joints under...

HSE Health & Safety

Executive

Static strength of cracked high strength steel tubular joints

Prepared by University College London for the Health and Safety Executive

OFFSHORE TECHNOLOGY REPORT

2000/078

HSE

Health & Safety

Executive

Static strength of cracked high strength steel tubular joints

B Talie-Faz, W D Dover and F P Brennan UCL NDE Centre

Department of Mechanical Engineering Torrington Place

London WC1E 7JE

HSE BOOKS

© Crown copyright 2002

Applications for reproduction should be made in writing to:

Copyright Unit, Her Majesty’s Stationery Office,

St Clements House, 2-16 Colegate, Norwich NR3 1BQ

First published 2002

ISBN 0 7176 2307 6

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means (electronic, mechanical, photocopying, recording or otherwise) without the prior written permission of the copyright owner.

This report is made available by the Health and Safety Executive as part of a series of reports of work which has been supported by funds provided by the Executive. Neither the Executive, nor the contractors concerned assume any liability for the reports nor do they necessarily reflect the views or policy of the Executive.

ii

SUMMARY

A series of nine static strength tests were performed on full-scale pre-cracked tubular welded joints. Six T-joints and three Y-joints were loaded to failure, in axial and out-of-plane-bending loading respectively, using purpose designed testing rigs.

The specimens were available from two previous fatigue test programmes conducted on a high strength jack-up steel (SE702). All specimens had at least one through thickness fatigue crack at the weld toe.

The aim of this work was to support studies concerned with the assessment of the static strength of cracked tubular joints. Very little published information is available concerning the residual strength of members containing cracks and fabricated from high strength steel.

The results from this study will support the development and verification of procedures for the prediction of the residual static strength in cracked tubulars.

iii

TABLE OF CONTENTS

SUMMARY........................................................................................................ iii TABLE OF CONTENTS ...................................................................................... iv 1. INTRODUCTION............................................................................................. 1 2. EXPERIMENTAL DETAILS ............................................................................. 2

2.1. Specimens ...............................................................................................2 2.2. Materials.................................................................................................2

2.2.1. Fracture Toughness Tests ........................................................................ 2 2.3. Joint Fabrication ....................................................................................3 2.4. Test Arrangement ................................................................................... 3

2.4.1. Y-joints ............................................................................................... 3 2.4.2. T-joints ............................................................................................... 4

2.5. Instrumentation ...................................................................................... 4 2.5.1. SCXI Data Acquisition System ................................................................. 4 2.5.2. Strain Gauges ....................................................................................... 4 2.5.3. Linear Variable Displacement Transducers.................................................. 4 2.5.4. Load Cell............................................................................................. 5 2.5.5. Measurements Taken During Test ............................................................. 5 2.5.6. Chord Ovalisation.................................................................................. 5 2.5.7. Crack Opening ...................................................................................... 6 2.5.8. Video Footage .......................................................................................7 2.5.9. Crack Growth Measurement ..................................................................... 7

2.6. Experimental Procedure..........................................................................7 2.6.1. Fatigue Precracking ................................................................................7 2.6.2. Elastic Loading ......................................................................................7 2.6.3. Static Strength Test.................................................................................7

3. RESULTS AND DISCUSSION ............................................................................9 3.1. Y-joints ...................................................................................................9

3.1.1. Specimen Y1 ........................................................................................ 9 3.1.2. Specimen Y4 .......................................................................................10 3.1.3. Specimen Y3 .......................................................................................10

3.2. T-joints .................................................................................................10 3.2.1. Specimen T6 .......................................................................................10 3.2.2. Specimen T1 .......................................................................................11 3.2.3. Specimen T4 .......................................................................................11 3.2.4. Specimen T2 .......................................................................................12 3.2.5. Specimen T5 .......................................................................................12 3.2.6. Specimen T3 .......................................................................................12

3.3. Failure Assessment ...............................................................................13 3.3.1. Evaluation of FAD Parameters ................................................................14 3.3.2. Failure Assessment Diagrams ..................................................................16

3.4. General Discussion ...............................................................................17 4. CONCLUSIONS .............................................................................................18 5. REFERENCES ...............................................................................................19 TABLES AND FIGURES .....................................................................................21

v

1. INTRODUCTION

Most of the static strength tests on large-scale tubular joints of the type used in fixed platforms have failed in a ductile manner in the laboratory [1-4]. This is because the material has to pass an acceptance test based on the CTOD, which ensures ductile behaviour. HSE have used this information as the basis for guidance in 50D steel [5-7]. More recently high strength steels have been used in the construction of fixed platforms and for jack-ups used for production rather than exploration. In addition, the level of in-service inspection is reducing as more emphasis is placed on the use of flooded member detection to indicate through-thickness cracks. This means that tubular joints in a structure may contain a large crack at the point when the maximum design load is experienced. A typical high strength steel, in welded form may not be as ductile as BS7191 355D. It is also likely to have a higher yield to ultimate strength ratio. This research is aimed towards developing guidance on the residual strength of cracked tubular joints and studying the influence of yield to ultimate ratio on failure capacity for tubular joints.

The availability of pre-cracked joints from a previous study [8] provided an opportunity for data comparison with joints made from conventional structural steels such as BS7191 355D.

1

2. EXPERIMENTAL DETAILS

2.1. Specimens A total of nine cracked tubular joints were available from previous projects.

Three Y-joints were part of a variable amplitude corrosion fatigue study [8]. The project involved out-of-plane-bending fatigue loading on SE702 Y-joints in two different testing environments. Its main objective was to examine the corrosion fatigue behaviour of the joints under simulated service loading and environmental conditions. All joints were fatigue loaded until a through crack was detected [8-10].

Six T-joints were also part of the full-scale fatigue-testing programme. These axially loaded joints were also tested under two different environmental conditions. The effects of cathodic protection on the corrosion fatigue behaviour of these joints under constant amplitude loading was analysed. All T-joints were cracked on at least one side [8,10].

The joint parameters are shown in Table 1. Figure 1 shows the dimensions of each tubular joint. Table 2 shows the history of each of the nine joints. Figures 2 and 3 show the percentage cracked area and the position of each crack for Y-joints and T-joints respectively.

2.2. Materials All nine tubular joints were fabricated from SE702 which is a member of the Super Elso family of steels. It is Creusot Loire Industrie’s (CLi) equivalent of the A517GrQ standard [11]. The quoted chemical composition is given in Table 3.

The specified mechanical properties are given in Table 4.

Independent tensile tests were carried out at Cranfield University on the same batch of SE702 used to fabricate the tubular joints. The specimen used were 5.5mm in diameter. Data from the Charpy impact tests performed at Cranfield are presented in Table 5 [9,10]

Both Creusot Loire and Cranfield University also conducted hardness tests on the parent plate, the heat affected zone and the weld metal of a T-Butt weld specimen. The data produced are presented as Vickers Hardness number in Tables 6 and 7.

2.2.1 Fracture Toughness Tests KIc tests were conducted on an Instron servohydraulic test machine according to BS7448. Specimens were cut from the tubular Y-joints after the full-scale tubular fracture tests. Three types of specimen were produced as described below. The specimens were cut as compact tension specimens with a thickness of 12mm. All of the material was cut from undamaged parts of the Tubular Y-joints.

2

P specimens cut from parent plate of the chord remote from the weld.

T specimens cut from the welded region of the chord so that one surface had HAZ still present.

W specimens cut as P specimens which were subsequently given a weld deposit which was machined off leaving an HAZ region on the specimen surface.

All the specimens were fatigue tested to give a fatigue crack for the fracture tests. The fracture tests were conducted at laboratory temperature (similar to the large-scale tubular fracture tests).

Six specimens each (a total of 18 specimens) from three different regions from the tubular joint designated, W, T and P were tested to failure. In general the fracture surfaces were slant which was probably due to the specimen thickness. A typical test result is shown in Figure 4 and this shows that a KQ analysis would be necessary.

The nominal toughness values have been calculated using 5% offset of the initial slope of the Compact Tension specimens Load-Displacement Curves. The toughness results are tabulated in Table 8 for each of the specimen for nominal toughness KQ and Maximum Toughness (Kmax) and the results are also shown as an average for different kinds of specimens. The units used for the determination of the toughness are MPa m1/2.

2.3. Joint Fabrication Both the brace and the chord were rolled in two halves and subsequently seam welded together to give 16mm thick tubes. Details of the seam welding are given in Table 9. Post weld heat treatment was applied to the seam welds [8]. The position of the seam welds is illustrated in Figure 5.

It was important that the welding between the brace and the chord was representative of that used in the construction of offshore structures as the weld detail is known to have a significant effect on the joint’s fatigue life. Details of the intersection welds are given in Table 10. It should also be noted that no post weld heat treatment was specified for the completed joint [8].

2.4. Test Arrangement 2.4.1. Y-joints The Y-joints were tested in a purpose built rig, which consisted of a 250kN Instron actuator with a 100mm stroke able to apply out-of-plane-bending loading to the brace of a Y or K-joint. The two ends of the chord were restrained from movement in any direction. This setup was also used for fatigue precracking of Y1. The failure load predicted from the API RP2A Load and Resistance Factor Design [12] and the HSE guidance [13] was approximately 500kN. This meant that while the setup provided

3

enough capacity to test the first Y-joint, there was a need for an actuator with at least as much capacity as the value calculated by using the API guidance. After the initial Y-joint test, the 250kN actuator was replaced with a 500kN actuator with a similar stroke length (see Figure 6). This involved modifying the rig. The resulting format did not alter the angle or the direction of the load being applied.

2.4.2. T-joints The uncracked joint capacity was calculated by using the API RP2A LRFD parametric equations and the HSE equations. The API RP2A Load and Resistance Factor Design was used rather than the Working Stress Design (WSD), as it uses reliability based calibration which was considered to be more suitable for this type of testing.

The rig shown in Figure 7 was constructed for the T-joint tests. A system of two jacks pushing against a reinforced box section was designed. The box section was bolted to the brace via high strength steel studdings that passed through the box as shown in Figure 8. The jacks were powered via a single hydraulic pump which could reach an oil pressure of up to 700 bar, meaning that the capacity of each jack was about 2500kN.

2.5. Instrumentation 2.5.1. SCXI Data Acquisition System An SCXI (Signal Conditioning Extensions for Installations) data acquisition system made by National Instruments was utilised to record measurements from strain gauges, LVDTs and any other voltage outputting device. It consisted of two SCXI-1322 boards used for measuring strain and an SCXI-1100 board for monitoring changes in potential difference. The strain measuring boards were connected to 32 external terminals each capable of recording from one linear gauge. The LVDTs were linked to the SCXI-1100 board via a panel comprising of 32 BNC sockets.

2.5.2. Strain Gauges Stacked rosette strain gauges were bonded to the surface of the joint around the crack area and on the brace. For the first two tests involving the Y-joints (Y1 & Y4), gauges were placed 50mm away from the crack tip. Two more gauges were bonded to the brace in line with the load application point and the saddle, parallel to the weld seam as shown in Figure 9. The same approach was employed for the T-joint tests concerning the gauges around the crack area (top face). The brace gauges were placed at 180˚ to each other for the first two T-joint tests (T6 & T1) and from then on at right angles to each other so that any conceivable bending in the system could be detected and possibly quantified. The gauges were then connected to the SCXI-1322 board via a series of terminals and electric cables.

2.5.3. Linear Variable Displacement Transducers Two types of LVDTs were used. The transducer with the larger stroke length (DC25) was used to measure the brace displacement in the direction of the applied

4

load. The smaller LVDT was used for chord ovalisation measurement. Both types of LVDTs were linked to the SCXI-1100 board via BNC connectors.

2.5.4. Load Cell A load cell was used to provide confirmation of the load output from the jacks. An excitation voltage was supplied through the same power supply used for the LVDTs. The readings from the load cell were initially recorded via the SCXI unit. It was found that the voltage measurement feature of the SCXI unit could not quantify the readings to the accuracy required. A digital voltmeter capable of recording potential difference to 10µV was used instead. 2. 5.5. Measurements Taken During Test a) Y-Joints Load versus displacement data were obtained via the actuator LVDT and the load cell of the servo control system. Both readings were recorded at 10-second intervals using two digital voltmeters connected to the mini-controller. A PSI function generator was used to supply a steady rise in voltage, and subsequently, load to the actuator. The minimum ramp rate of one cycle for every 1000 seconds was used with the testing mode in load control.

b) T-Joints The LVDT with a stroke of 50mm was calibrated and attached to the back of the box section, in line with the centreline of the brace and monitored as a function of time during the test.

For the first two tests, the load was measured using the strain gauges attached to the brace. For tests T4, T2, T5 and T3 the load cell was used. In all tests, the jack pressure was monitored to give confirmation of the load readings. The values from the load cell were consistently about 8% lower than those of the pressure gauge. This may be attributed to the fact that there is a certain amount of friction loss in the jacking system which could not be accurately quantified.

2. 5.6. Chord Ovalisation As the brace was axially loaded, load was transferred through the weld and to the chord. The shape of the cross section of the chord changed from circular to elliptical away from the clamped ends. The value of deformation was recorded. The ovalisation of the chord for the T-joints was monitored by the DC15 LVDT (stroke length 30mm). The values were recorded by the SCXI DAQ system at regular time intervals. Figure 10 shows the setup for the chord ovalisation measurement.

Due to the lack of access to the inside of the chord, chord ovalisation was not measured for the Y-joints.

5

2.5.7. Crack Opening Photogrammetry Photogrammetry is a technique which uses high resolution, high quality imaging. This is the first time it has been used in conjunction with tubular joint experiments. With the help of the Geomatics Centre at UCL, this technique was employed on all T-joint experiments and the final Y-joint (Y3).

Vision metrology is a sub-topic of photogrammetry. This technique is used to complement conventional instrumentation in structural testing and monitoring. In its most basic form, the technique is based on following the movements of points on a specimen using a number of cameras and a computer.

The technique proved the most effective and accurate available means for measuring crack opening displacements of the magnitude and range required for this project. Its major advantages include the fact that it is a non-contacting measuring system and is very accurate, as it has been known to measure with sub millimetre accuracy depending on the number of cameras being used.

Initially a suitable three dimensional axis system had to be chosen so that the data recorded could be interpreted in cartesian co-ordinates. For ease of camera calibration and presentation of the data in the required form, the brace of each tubular joint was aligned with the y-axis, while the x-axis was positioned along the chord and the z-axis perpendicular to both the chord and brace.

Retro-reflective targets with an adhesive side were placed on the specimen in a relatively accurate grid. A more dense cluster of targets was placed on the area around the weld toe. Figure 11 shows the arrangements of the targets on a T-joint. As a reference, larger targets were placed in the camera shots, where they were deemed to be still and not effected by the movements during the test. The distance between a few of the targets was measured manually for scaling purposes and to confirm the results from the images taken. A simple illustration of the system is shown in Figure 12.

The images were captured by dimming the surrounding lights and shinning a high intensity white light onto the targets which reflect the light back. The images were recorded by the two cameras at a previously specified rate and were transferred to the computer. These images were then combined so that they could be displayed as an animation of the test in 2-D or 3-D. The data from each target also provided displacement values in all three co-ordinates as well as a combined value for movement in three dimensions. Figure 13 shows the difference between two images from the initial and the later stages of a test [14,15], while Figures 14a and 14b show the image of the same T-joint taken from both cameras.

Figure 15 demonstrates one of the features of this technique. It shows the general shape of the T-joint in 3-D by only recording the target positions [16].

6

2.5.8. Video Footage A video account of the last three tests involving a T-joint (T2, T5 and T3) was recorded. The joints were positioned so that the crack (or the larger of the two cracks) was facing towards the camera. Crack propagation could be observed, particularly during the latter stages of each test. The tests were recorded on 8mm video-camera tapes and were later digitised and recorded on compact disc.

2.5.9. Crack Growth Measurement The procurement of the images from photogrammetry allowed the measurement of the crack length. Each image was viewed carefully and a decision was made as to how far the crack had progressed using the retro-reflective targets around the crack as reference points. Although this method of measurement is not very accurate, the information obtained was both valuable and necessary. Tables 11 (i-vi) show the crack growth values at various time increments and load values during each T-joint test.

2.6. Experimental Procedure 2.6.1. Fatigue Precracking The original rigs with modifications, which were used to grow the fatigue cracks in all nine joints, were used to grow the cracks in two of the specimens, Y1 and T6. It was decided that the remaining cracks in both T and Y-joints covered a reasonable range of percentage cracked areas and thus precracking for the rest of the specimen was not required. Table 12 shows the original crack size and the precracking load range applied to Y1 and T6 along with the final crack size.

2.6.2. Elastic Loading Elastic loading was carried out for the first T-joint test (before the rig was modified) to check the operation of the various gauges and the outputs of the actuator. The specimen was loaded to approximately 5% of the total capacity and then unloaded to ensure that the instrumentation readings returned to zero.

2.6.3. Static Strength Test a) Y-Joints The function generator was used to generate a slow ramp from zero up to the maximum capacity of the actuator. It was known that the 100mm stroke length was not enough for this test. To overcome this, the base of the actuator was packed with spacers after the stroke had been fully extended. The test was then restarted.

The test was stopped when the crack reached the crown positions on the joints or more accurately, when the rate of brace displacement increased rapidly relative to the constant load increase.

7

b) T-Joints The oil pressure regulator of the hydraulic pump was used as the manual load controller for the duration of the test. Tearing of the crack could be heard throughout most tests.

The tests were deemed complete when one side of the joint failed completely, which resulted in the load being taken by the joint to drop substantially. Alternatively, in most cases there was total separation between the brace and the chord (T6, T1, T5 and T3). There was normally a warning in that the pressure reading from the gauge would drop off by a few percent before failure occurred. Figure 16 shows the separated brace from the chord for test T1.

8

3. RESULTS AND DISCUSSION

Static Strength Capacity The main test results are summarised in Table 13 which shows load values at three different stages during each test and the original percentage cracked area in each case.

The peak load in each case was used with the HSE characteristic loads to give the failure load ratio and plotted as a function of crack size. The data are shown in Figure 17 together with the mean line obtained from Figure 4, reference 17. This line includes data from large scale tests, small scale tests and numerical analysis. In most cases, the failure data for SE702 are below this line. A mean line for the high strength steel data was plotted to further establish the relationship between the two materials. There is only a small difference between the mean lines for the high strength steel and the mild steel, with the HSS mean line being about 5% lower in relation to the intercept at the ordinate.

Detailed Results 3.1. Y-Joints 3.1.1. Specimen Y1 As explained in section 2.6.3. all Y-joint specimens had to be tested in several stages. Specimen Y1 was precracked and thus the percentage cracked area had been increased from 45% to about 70%. The test involved three large cycles.

Figure 18 illustrates the load-displacement curves for the static strength testing of the first Y-joint. The first test showed that the load was linearly related to the displacement up to a load of 100kN. Subsequently, the slope of the load-displacement curve changed to give an approximately linear region up to 170kN. After this point, the crack began to tear at a rapid rate until the maximum stroke length had been reached and the joint had to be unloaded. Spacers were added to the base of the actuator to allow an increase in the overall brace displacement.

As the specimen was loaded for the second test, a linear slope was observed, until the last two points where more tearing resulted in a rapid increase of displacement.

The same process was undertaken for the third and final try. The results were quite similar to the second test, as only the last three points were not in line with the previous loading line. It should be noted that crack extension occurred in each test. By the last run, the cracked area was almost 100% (of half the total welded area) and initial branching of the crack into the chord was observed.

The specimen was loaded one last time, however, this time the maximum load taken was about 20kN less than the previous tests. The results from this test were omitted as they provide no useful data. This type of last test was performed for all the other Y specimens in order to ensure that the joint had failed.

9

3.1.2. Specimen Y4 Specimen Y4 was tested in the same manner as the previous Y. Figure 19 illustrates the two tests performed on this specimen. The first two tests were enough for the joint to reach its peak capacity of 270kN. Again both curves were linear up until the last few points. No spacers were required for the first two tests. Spacers were added for the confirmation test, which saw the load rise up till 250kN. The crack had started to branch off into the chord indicating that the test was over.

3.1.3. Specimen Y3 Test Y3 was expected to be very similar to Y4. The percentage cracked areas were almost identical.

Figure 20 displays the three tests that were performed on Y3. The first curve was not perfectly linear past the early part of the test. The displacement was growing more rapidly with respect to the load applied. A new test was performed as the stroke had run out on the previous test.

In order to load the joint up again, spacers had to be added to the base of the actuator. The second test was linear until the last two points which were in line with the final few points of the previous test. The results of the final test were very similar to the second test indicating that the final two points were in line with the other two curves. Thus a complete load-displacement graph could be plotted by joining the last few points of the second and third test to the final stages of the first test. The overall brace displacement was 190mm.

3.2. T-Joints 3.2.1. Specimen T6 The test was attempted three times, twice with the 1000kN servo hydraulic actuator in the original T-Joint rig used to pre-crack the specimen and once with the static jacks and box beam setup (Figure 7). The first servo hydraulic actuator test reached a maximum of 930kN and appeared to be close to a peak value. The second servo hydraulic actuator test (using a higher pressure) reached a maximum of 950kN with clearer evidence of reaching a peak value. The third test confirmed the two earlier tests and showed a definite peak in the load-displacement curve at 960kN.

The brace displacement recorded for these tests was different, as the first two used the servo hydraulic actuator LVDT, whilst the third test used the LVDT at the back of the box beam. The displacement measured in the third test is likely to be more representative of the brace displacement.

Figure 21 shows the load-displacement curves for specimen T6 using the two methods of brace displacement, the chord ovalisation and the maximum crack opening displacement data. It can be seen that the chord ovality curve tends to a constant value of displacement in the latter stages of the test. The general shape of

10

the other three curves was similar in that all three had a distinctive peak at the maximum load. From the similarity of the load-Maximum COD curve to the load-brace displacement curve, it may be assumed that the top crack was the dominant defect.

The fracture surface indicated ductile tearing throughout, excluding a small region at the crown points of the chord.

Figure 22 is a plot of mode I crack opening displacement along the crack at various loads throughout the test. The curves show that the crack was relatively symmetrical about the hot spot stress point.

The change in crack length as a function of load was measured and recorded in Table 11(i).

3.2.2. Specimen T1 Figure 23 shows the plot of load versus brace displacement, chord ovality and maximum COD values. The plot has a distinctive peak at 3000kN. As the maximum load was reached, continuous tearing occurred and the pressure gauge indicated a drop in load (displayed by the unloading section). Figure 16 shows the separated brace from the chord for test T1.

The fracture surface showed major ductile tearing at the top. The bottom surface was mostly light reflective and crystallographic implying brittle fracture.

Figure 24 shows part of the fracture surface of the second test. The original crack can be clearly seen.

Figure 25 illustrates the COD readings. The crack growth is relatively equal from both sides around the saddle and it has a symmetric shape. This figure also shows the rapid tearing stage as the end of the test took place.

The change in crack length as a function of load was measured and recorded in Table 11(ii).

3.2.3. Specimen T4 The load-displacement plot results are shown in Figure 26. Some data were lost for this test.

From Figure 26, an elastic region can be observed, followed by a peak at 3050kN.

The change in crack length as a function of load was measured and recorded in Table 11(iii).

11

3.2.4. Specimen T2 Specimen T2 was the only specimen that had equal size cracks on both sides (about 17% each).

Figure 27 is the load-displacement plot using various methods of measuring displacement. It has a clear and distinctive peak at about 2800kN. The much smaller value of maximum COD compared with the other displacements, indicates the fact that the monitored crack (at the top) was not the dominant defect. The test ended as there was a sudden drop in the load. About 20% of the brace remained attached to the chord.

There was comparatively little tearing during this test, as indicated by the fracture surface. There was more tearing at the bottom which meant that the bottom crack had become the dominant defect. The fracture surface indicated more brittle fracture during this test than any of the others. Figure 28 illustrates the COD values at different loads during the test. The crack did not extensively increase in length.

The change in crack length as a function of load was measured and recorded in Table 11(iv).

3.2.5. Specimen T5 The load-displacement curves are shown in Figure 29. It can be seen that there was a plateau once the maximum force was reached. From 20mm until about 42mm the crack was tearing under constant load. All four curves closely resembled each other. It should be noted that the chord was permanently deformed after the test.

The fracture surface was mostly dull and fibrous around the original crack, thus suggesting ductile tearing. There was more evidence of brittle fracture on the opposite side of the weld.

From Figure 30 it can be seen that the crack grew symmetrically and at a relatively constant rate. The crack opening displacement for this test is the largest of the whole project. This can partially be explained by the large deformation in the chord indicated by the chord ovalisation readings.

The change in crack length as a function of load was measured and recorded in Table 11(v).

3.2.6. Specimen T3 The load-displacement curves are shown in Figure 31. The curve shows a plateau after reaching the highest load of 3100kN. Visible deformation was observed in testing of this specimen, although not to the extent of the previous test.

The fracture surfaces indicated ductile tearing on the top side and to some extent on the bottom fracture face.

12

The crack opening displacement curves (Figure 32) indicate that the crack did not grow evenly.

The change in crack length as a function of load was measured and recorded in Table 11(vi).

3.3. Failure Assessment In order to analyse the data obtained, the use of Failure Assessment Diagrams was employed. FADs are used to consider failure by linear elastic fracture as one limiting criteria and failure by plastic collapse as the second. When performing a structural integrity assessment of a flaw in a stressed structure an assessment point is derived from two different calculations and plotted on the FAD. The structure is deemed unsafe if the point calculated lies on the curve or falls outside it, and safe, if the point is within the curve.

The first version of the failure assessment curve defined by Harrison et. al. is shown in Figure 33 [17]. The equation determining the shape of the curve is as follows:

Where the fracture parameter, Kr, is expressed in terms of the plastic collapse parameter, Sr.

BS 7910 The PD 6493 procedure for the assessment of defects in welded components, originally published in 1980 and subsequently revised in 1991, has been used with some success in industry and is now applied to offshore structures [18]. Its main applications are the fitness-for-purpose assessment of fabrication and in-service defects, inspection scheduling and the determination of whether post-weld heat treatment is required. PD 6493 was further revised as a result of The Welding Institute’s work with the British Standard Institution as BS 7910 which refines the approaches of the 1991 edition but retains the same principles [19].

There are three levels of assessment depending on the input data available, the level of conservatism and the degree of accuracy required.

Level 1 Level 1 is the screening level introduced into the 1991 PD 6493. It provides a conservative estimate from its use of a simplified FAD with in-built safety factors and requires conservative estimates of the applied stress, the residual stress and the fracture toughness input. The curve for Level 1 assessment is shown in Figure 33.

This level of assessment is consistent with the CTOD design curve in the 1980 version of PD 6493.

13

Level 2 This is considered to be the normal assessment route applicable for general structural steel application and makes use of FAD with no additional safety factors. The assessment equation is identical to Equation (1). The Level 2 assessment utilises the strip yield model and allows for the interaction of fracture and collapse. Figure 33 shows a Level 2 FAD in comparison to the conservative Level 1 curve.

Level 3 This level employs a full tearing instability approach and provides a more accurate description of ductile materials. For this reason, Level 3 assessment was used for the assessment in this study.

In a Level 3 assessment, the stress ratio, Sr, is replaced by a load ratio, Lr.

The plastic collapse axis is normalised to yield stress and not to flow stress. Thus the assessment curves extend beyond unity and represent material strain hardening as seen in Figure 34. The load ratio, however, cannot exceed cFLOW / cYIELD. This limiting load ratio is defined as Lr MAX. Different values of Lr MAX are used for various steels due to the alternative strain hardening characteristic of the materials. If the stress-strain curve for the material is not available, as would be the case if the heat affected zone of a weld is analysed, the following FAD equation can be applied:

This expression also has a cut-off at Lr = cFLOW / cYIELD. It requires only knowledge of the yield and tensile strengths of the material.

As with previous FADs any data point falling under the curve (Figure 34) is considered safe while the point lying on or falling outside the boundaries of the curve are deemed unsafe.

3.3.1 Evaluation of FAD Parameters Brittle fracture is predicted when the applied stress intensity factor, KI, equals the critical stress intensity factor, KIc (the fracture toughness). The fracture axis is defined by the stress intensity ratio, Kr as:

KK Ir (3) K IC

KI has to be calculated in terms of the relevant geometry and in the present case two complications exist. Firstly, the KI expression for a part through crack in a tubular joint would have to take into account the local bending and this means that one would use the hot spot stress. Once the crack has penetrated the chord wall, the calculation for KI

14

can be performed by using available solutions for through-thickness cracks in tubes [20]. For situations where two cracks existed, the corrections for two unequal cracks in a flat plate can also be used [21]. The other complication is the crack orientation for a chord crack in a tubular joint. This will vary with the p ratio (ratio of brace and chord diameters) so that for high op ratios, the crack orientation may be greater than 45 to the loading direction. In effect, this changes the fracture mode from mode I to a mixture of Modes I and III. It should be noted that the failure of a tubular joint may well be different to that of a closure weld on a tube for this reason. For the purpose of this analysis, the calculation of KI for both loading cases was as shown below:

Axial Tension Case

The definitions for P, M, 8, R and t are shown in Figures 35a and 35b [20].

F is a correction factor for various crack sizes. As mentioned in the previous paragraph, an extra correction factor, FI, was taken into consideration for joints with two cracks. This value was read off the Figure on page 199 of reference 21.

The KIC test results show that the SE702 material was not brittle in the test thickness used and produced data with relatively small scatter. The Kmax values were about 40% more than the KQ values with overall averages of 142 and 104 MPa m1/2. The differences between each type of specimen were also small, at most 6%, and confirmed that the T specimens, cut with the original HAZ intact, were appropriate for the purpose of the Failure Analysis Diagram. The values

15

chosen for the purpose of illustrating the analysis were a KQ of100.8 MPa m1/2 for the HAZ and a KQ of 103.8 MPa m1/2 for the parent plate. These values were chosen even though Kmax/KQ was greater than 1.1 because the specimens were cut from thin plate.

For CTOD data, Kr is replaced by į r .

where 8I is the applied CTOD obtained from a modified form of the CTOD design curve [23]:

The plastic collapse criterion is defined as the ratio of the net stress to the flow stress. Plastic collapse occurs when this ratio is equal to unity. The Stress ratio, Sr, can be defined as:

while the load ratio is represented by:

3.3.2 Failure Assessment Diagrams Figures 36 to 41 show the failure assessment diagrams for the specimens using BS 7910 Level 3 analysis. The calculated values for Kr and Lr are given in Tables 14, 15 and 16.

Figures 36-38 show the data for a parent plate KIC of 103.7MPa.m1/2. Figures 39-41 show the same data for a KIC value appropriate to the HAZ of 100.8MPa.m1/2. In each case, the initial tearing load, the maximum load and the twice compliance load was used to give the nominal stresses in the brace.

16

It can be seen that these three approaches have varying degrees of conservatism with the peak load calculation being the least conservative but being more consistent with the approach shown in Figure 17, the normalised ultimate static strength capacity.

From Figure 36, the initial tearing load results, it can be seen that only one T-joint result was inside the envelope All the other results were outside.

Figure 37 shows the maximum load data. All data lie outside the envelope.

Figure 38, for the twice compliance load data, shows a similar behaviour to the maximum load data in that all data lie outside the envelope.

Figures 39-41 show a similar trend, but this time all the Kr values are slightly larger. For these graphs only one experimental point fell within the FAD safe region. For the maximum load data this was test T2, which was a specimen with a double crack.

Figures 42-47 show the data using Kmax instead of KQ. It can be seen that the results are less conservative and for say, the initial tearing load four data points fall within the safe zone.

3.4. General Discussion The normalised static strength capacity approach appears to be simple to operate and provides an appropriate method for the offshore industry. The mean curve for the high strength steels is slightly lower than that for BS7191 355D by about 5%. In both cases the mean curves are above the (0,1) – (1,0) line. The slight difference between the SE702 and the BS7191 355D material could be related to the crack path. It was noted that for the SE702 tests, the crack remained close to the weld toe at all times. For BS7191 355D tests the crack often branches away from the weld toe and grows into the parent plate.

In comparing the mean failure line of Figure 17 with the FAD data, the biggest discrepancy appears to be for the T-joints with two cracks. Thus, for T4 and T2, the data is above the line in Figure 17 but for the FAD graphs, T4 and T2 appear to give low values. This may well be due to the inadequacy of the multiple crack correction used (this solution was originally developed for flat plates). The solution seemed to be adequate for T6, but inadequate for T4 and T2, giving only a 3% increase in the stress intensity factor for the presence of two cracks.

Figures 36-47 highlight the importance of the fracture toughness value used. It would seem that for the SE702 data that the KQ toughness value needs to be used as Kmax gives more data points within the safe zone.

17

4. CONCLUSIONS

1. Nine static strength tests were successfully completed on high strength steel tubular welded joints (T and Y) made from SE702.

2. Normalised ultimate static strength results were compared with typical data produced using BS4360 50D steel. It was found that the reduction in the static strength of cracked high strength steel joints relative to the uncracked strength was greater than that for joints fabricated from 50D steel by 5% at the ordinate intercept.

3. An FAD analysis was attempted and showed that failure could be predicted if the lower bound value of fracture toughness , the KQ value, was used. It would appear that the Kmax value of toughness is unconservative. It is also necessary to have the KIC values for the material in the condition that it is to be used.

4. Solutions for multiple cracks in tubulars were not available and there is a need for this information.

18

5. REFERENCES [1] Hadley, I et al., “Static Strength of Cracked Tubular Joints: New Data and Models”,

International Conference on Offshore Mechanics and Arctic Engineering, 1998. [2] Cheaitani, M. J., “FAD Based Tearing Analysis of New TWI Data on the Static Strength

of Cracked Tubular Double-T Joints”, International Conference on Offshore Mechanics and Arctic Engineering, 1998.

[3] Lalani, M., Nichols, N. W., and Sharp, J. V., “New Developments in Tubular Joint Static Strength Technology”, BOSS Conference Paper, Volume III, Boston, 1994.

[4] Klasen, B., and Wastberg, S., “Failure Assessment Diagram Method for Tubular Joints”, ISOPE Conference Paper, Honolulu, 1997.

[5] Department of Energy, “Static Strength of Large Scale Tubular Joints, Engineering Assessment”, OTH 89 297.

[6] Department of Energy, “Background to the New Static Strength Guidance for Tubular Joints in Steel Offshore Structures”, OTH 89 308.

[7] Department of Energy, “Static Strength of Large Scale Tubular Joints, Test Programme and Results”, OTH 89 543.

[8] Etube, L. S., Myers, P. T., Brennan, F. P., Dover, W. D., Stacey, A., “Constant and Variable Amplitude Corrosion Fatigue Performance of a High Strength Jack-up Steel”, ISOPE Conference Paper, Montreal, 1998.

[9] Etube, L. S., Brennan, F. P., Dover, W. D., “Variable Amplitude Corrosion Fatigue (VACF) of High Strength Steels Used in Jack-Up Structures”, Final Report, UCL NDE Centre, January 1998.

[10] Myers, P. T., Brennan, F. P., Dover, W. D., “Corrosion Fatigue Fracture Mechanics of High Strength Jack Up Steels”, Final Report, UCL NDE Centre, January 1997.

[11] Coudert, E., Renaudin, C., “Variable Amplitude Corrosion Fatigue Behaviour and Hydrogen Embrittlement of High Strength Steels for Offshore Applications”, ISOPE Conference Paper, Montreal, 1998.

[12] “Recommended Practice for Planning, Designing and Constructing Fixed Offshore Platforms”, API RP2A 20th Edition, American Petroleum Institute, Washington, 1993.

[13] “Offshore Installations: Guidance on Design, Construction and Certification”, Fourth Edition, UK Health and Safety Executive, HMSO, London, 1990.

[14] Robson, S., Woodhouse, N. G., “Monitoring Concrete Columns Using Digital Photogrammetric Techniques”, Department of Civil Engineering, City University.

[15] Robson, S., Cooper, M. A. R., “Digital Photogrammetric Monitoring of Small Scale Structural Deformation”, 3rd Symposium on Surveillance and Monitoring Surveys, Melbourne, 1995.

[16] Woodhouse, N. G., Robson, S., Eyre, J., “Some Roles of Vision Metrology and Associated Visualisation Tools in Structural Testing and Engineering Experimentation”, Department of Geomatic Engineering, University College London, 1999.

[17] Stacey, A., Sharp, J. V., and Nichols, N. W., “Static Strength Assessment of Cracked Tubular Joints”, OMAE Conference Paper, Florence, 1996.

[18] Stacey, A., Burdekin, F. M., Maddox, S. J., “The Revised PD 6493 Assessment Procedure – Application to Offshore Structures”, OMAE Conference Paper, Florence, 1996.

[19] BS 7910:1997, “Guide on Methods for Assessing the Acceptability of Flaws in Structures”, British Standard Institution, 1997.

[20] Zahoor, A., “Closed form Expressions for Fracture Mechanics Analysis of Cracked Pipes”, The Journal of Pressure Vessel Technology, Vol. 107, pp. 203-205, May 1985.

[21] The Society of Materials Science, “Stress Intensity Factors Handbook”, Edited by Y. Murakami, Vol. 1, pp 198-199, Japan, 1987.

[22] Coudreuse, L., Renaudin, C., Bocquet, P., Cadiou, L., “Evaluation of Hydrogen Assisted Cracking Resistance of High Strength Jack-up Steels”, Marine Structures, vol. 10, pp 85106, 1997.

[23] Anderson, T. L., “Fracture Mechanics – Fundamentals and Applications- 2nd Edition” CRC Press New York, 1995.

19

Table 1: Specimen Parameters

Y Joint Parameters T Joint Parameters T 16 a 10.85 T 16 a 7.26 D 457 p 0.71 D 457 p 0.71 L 2480 y 14.28 L 1660 y 14.28 t 16 ' 1.00 t 16 ' 1.00 d 324 e 35.00 d 324 e 90.00 l 1390 l 900

Table 2: Testing History of Specimen

Test in order tested Tested Environment C.P. %Cracked Area Y1 Air - 45.0% Y3 Sea Water -800mV Ag/AgCl 45.6% Y4 Sea Water -1000mV Ag/AgCl 43.3% T1 Air - 16.3% T2 Air - 18.2%

17.2% T3 Sea Water -1000mV Ag/AgCl 16.3% T4 Sea Water -800mV Ag/AgCl 7.8%

18.0% T5 Sea Water -1000mV Ag/AgCl 20.0% T6 Sea Water -800mV Ag/AgCl 18.0%

Table 3: Chemical Composition of SE702

Element SE702 Quoted Chemical Composition % C 0.12

Mn 1.05 Si 0.25 S

Table 4: Mechanical Properties of SE702

cy (MPa) UTS (MPa) Elongation (%) Carbon Eqv. (IIW) Paris Constant 'C' Paris Constant 'm' 700 790-940 16 0.599 2.72 x 10-12 3.532

Table 5: SE702 Charpy Test Data

Temperature (oC) Specimen Charpy Energy, J Room Temp. S1 149

S2 152 -40 S3 127

S4 112 -60 S5 92

S6 79

Table 6: SE702 Vicker’s Hardness Data

Location Hardness (Hv) Parent Plate ~250

HAZ 272 - 374 Weld Metal 252 - 297

Table 7: SE702 Hardness Data

Weld Metal CGHAZ Parent Specimen CAP ROOT CAP ROOT -

Average 1 305 249 392 311 253 2 305 253 389 300 260

Range 1 276 - 336 245 - 251 373 - 409 262 - 363 242 - 268 2 360 - 336 247 - 258 363 - 401 272 - 345 243 - 274

Sample Size 1 10 4 8 10 11 2 10 4 9 10 11

Table 8: KQ and Kmax Fracture Toughness values for SE702SE702

Specimen W Specimen T Specimen P Specimen No. KQ 5%off Kmax KQ 5%off Kmax KQ 5%off Kmax

1 104.0 147.0 100.0 148.8 102.0 142.2 2 106.0 147.4 102.0 150.9 104.0 141.4 3 106.0 144.1 99.0 148.8 104.0 140.3 4 101.0 149.8 102.0 156.2 102.0 143.2 5 118.0 145.5 102.0 151.5 106.0 142.4 6 106.0 149.6 100.0 152.4 104.0 143.6

Average 106.8 147.2 100.8 151.4 103.7 142.2

22

Table 9: Seam Welding Details

Seam Welding Details Weld Process Semi Automatic MAG

Weld Consumable Oerlikon Fluxocord 42 / OP1TT Pre / Postheat 125oC for 90 Minutes

Heat Input

Table 11(iii): Crack Growth Data Collected During the Third Test

Table 11(iv): Crack Growth Data Collected During the Fourth Test

The Fourth Test (T2) Epoch Time Load (kN) Max. COD Crack Length Alignment

0 0 0 0.00 240 -110 , 130 16 82 76 0.10 240 -110 , 130 32 165 250 0.20 240 -110 , 130 48 247 1254 0.77 240 -110 , 130 64 329 1790 1.77 240 -110 , 130 80 412 2296 3.42 250 -115 , 135 96 494 2520 4.54 265 -125 , 140

112 576 2790 6.54 285 -140 , 145 124 638 2618 11.61 400 -215 , 185

Table 11(v): Crack Growth Data Collected During the Fifth Test

The Fifth Test (T5) Epoch Time Load (kN) Max. COD Crack Length Alignment

0 0 0 0.00 260 -130 , 130 20 102 0 0.33 260 -130 , 130 40 204 400 0.78 260 -130 , 130 60 306 1180 1.95 260 -130 , 130 80 408 1820 3.71 260 -130 , 130

100 510 2150 5.55 280 -140 , 140 120 612 2550 8.79 345 -165 , 180 140 714 2730 10.62 375 -185 , 190 160 816 2840 16.85 465 -245 , 220 180 918 2650 28.55 584 Crown , Crown 196 1000 2680 36.61 584 Crown , Crown

24

Table 11 (vi): Crack Growth Data Collected During the Sixth Test

The Sixth Test (T3) Epoch Time Load (kN) Max. COD Crack Length Alignment

0 0 0 0.00 260 -140 , 120 20 100.6 0 0.19 260 -140 , 120 40 201.2 366 0.60 260 -140 , 120 60 301.8 870 1.31 260 -140 , 120 80 402.4 1276 1.71 260 -140 , 120

100 503.1 1900 3.28 260 -140 , 120 120 603.7 2350 5.22 290 -140 , 150 140 704.3 2790 7.13 380 -180 , 200 160 804.9 2860 8.61 415 -200 , 215 180 905.5 3092 14.33 470 -240 , 230 196 986 2894 28.16 584 Crown , Crown

Table 12: Precracking History

Joint Type Original Cracked Area Type of Loading Loading Range Frequency Final Cracked Area Y-Joint (Y1) 45% Cyclic 75 kN 0.8 70.5% T-Joint (T6) Top - 17%

Bottom - 0% Cyclic 260 kN 0.8 25.8%

45.6%

Table 13: Final Results Table

Tests Initial Tearing Load (kN) Max. Load (kN) Twice Comp. Load %Cracked Area Comments Y1 115 185 185 70.5% Crack extended from about 45% - 510mm long

1st 500kN actuator test - 400mm long 400mm long Side A 350mm long / 150mm through Side B 520mm long / 505mm through One side cracked 240mm length / 100mm through Side A 130mm long / 14mm deep Side B 275mm long / 100mm through Side A 240mm long / 120mm through Side B 210mm long / 100mm through One side cracked 260mm length / 105mm through One side cracked 260mm length / 65mm through

Y4 250 270 270 43.3% Y3 225 264 264 45.6% T6 615 961 785 25.8%

46.3% T1 2100 2977 2530 16.3% T4 - 3056 * - 7.8%

18.0% T2 2296 2869 2740 18.2%

17.2% T5 2150 2881 2845 20.0% T3 2350 3131 3065 16.3%

25

Table 14: FAD Parameters for the Initial Tearing Load Case using KQ

Initial Tearing Load KQ = 103.7MPa.m1/2 KQ (HAZ) = 100.8MPa.m1/2

Tests (kN) Kr Lr Kr Lr Y1 115 1.936 0.230 1.992 0.230 Y4 250 3.172 0.500 3.264 0.500 Y3 225 2.855 0.450 2.937 0.450 T6 615 1.338 0.253 1.377 0.253 T1 2100 1.191 0.224 1.226 0.224 T4 - - - -T2 2296 0.858 0.325 0.883 0.325 T5 2150 1.335 0.235 1.374 0.235 T3 2350 1.460 0.256 1.502 0.256

Table 15: FAD Parameters for the Maximum Load Case using KQ

Maximum Load KQ = 103.7MPa.m1/2 KQ (HAZ) = 100.8MPa.m1/2

Tests (kN) Kr Lr Kr Lr Y1 185 3.115 0.370 3.204 0.370 Y4 270 3.426 0.540 3.525 0.540 Y3 264 3.299 0.520 3.394 0.520 T6 961 2.091 0.396 2.151 0.396 T1 2977 1.689 0.317 1.738 0.317 T4 3056 * 1.433 0.404 1.474 0.404 T2 2869 1.072 0.406 1.103 0.406 T5 2881 1.789 0.314 1.841 0.314 T3 3131 1.945 0.342 2.001 0.342

Table 16: FAD Parameters for the Twice Compliance Load Case using KQ

Twice Compliance Load KQ = 103.7MPa.m1/2 KQ (HAZ) = 100.8MPa.m1/2

Tests (kN) Kr Lr Kr Lr Y1 185 3.115 0.370 3.204 0.370 Y4 270 3.426 0.540 3.525 0.540 Y3 264 3.299 0.520 3.394 0.520 T6 785 1.780 0.323 1.757 0.323 T1 2530 1.435 0.270 1.477 0.270 T4 - - - - -T2 2740 1.024 0.388 1.054 0.388 T5 2845 1.767 0.310 1.818 0.310 T3 3065 1.904 0.334 1.959 0.334

26

Table 17: FAD Parameters for the Initial Tearin g Load Case using Kmax

Initial Tearing Load Kmax = 142.2MPa.m1/2 Kmax (HAZ) = 151.4MPa.m1/2

Tests (kN) Kr Lr Kr Lr Y1 115 1.412 0.230 1.326 0.230 Y4 250 2.314 0.500 2.173 0.500 Y3 225 2.082 0.450 1.956 0.450 T6 615 0.976 0.253 0.917 0.253 T1 2100 0.869 0.224 0.826 0.224 T4 - - - -T2 2296 0.626 0.325 0.588 0.325 T5 2150 0.974 0.235 0.915 0.235 T3 2350 1.064 0.256 0.999 0.256

Table 18: FAD Parameters for the Maximum Load Case using Kmax

Maximum Load Kmax = 142.2MPa.m1/2 Kmax (HAZ) = 151.4MPa.m1/2

Tests (kN) Kr Lr Kr Lr Y1 185 2.272 0.370 2.134 0.370 Y4 270 2.499 0.540 2.347 0.540 Y3 264 2.406 0.520 2.260 0.520 T6 961 1.525 0.396 1.432 0.396 T1 2977 1.232 0.317 1.157 0.317 T4 3056 * 1.045 0.404 0.982 0.404 T2 2869 0.782 0.406 0.735 0.406 T5 2881 1.305 0.314 1.226 0.314 T3 3131 1.418 0.342 1.332 0.342

Table 19: FAD Parameters for the Twice Com pliance Load Case using Kmax

Twice Compliance Load Kmax = 142.2MPa.m1/2 Kmax (HAZ) = 151.4MPa.m1/2

Tests (kN) Kr Lr Kr Lr Y1 185 2.272 0.370 2.134 0.370 Y4 270 2.499 0.540 2.347 0.540 Y3 264 2.406 0.520 2.600 0.520 T6 785 1.246 0.323 1.700 0.323 T1 2530 1.047 0.270 0.983 0.270 T4 - - - - -T2 2740 0.747 0.388 0.701 0.388 T5 2845 1.289 0.310 1.210 0.310 T3 3065 1.388 0.334 1.304 0.334

27

900mm

457mm

1660mm

16mm

16mm 324mm

35 °

16mm

324mm

457mm

00mm 990mm 9

1390mm

2480mm

Figure 1: Nominal Specimen Dimensions for Tubular Welded Joints

28

70.5%

Y1

43.3%

Y4

45.6%

Y3

Figure 2: Percentage Cracked Area and Position of Each Crack for All Y-Joint Specimen

29

T6 T1

25.8%

46.3%

16.3% Side A

Side B

7.8% Side A

Side B 17.2% 18.0%

T4 T2

18.2%

T5 T3

20.0% 16.3%

Side B

Side A

Figure 3: Percentage Cracked Area and Position of Each Crack for All T-Joint Specimen

30

KQ M

Pam

1/2

160

140

120

100

80

60

40

20

0 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5

Clip Gauge Opening (mm)

Figure 4: KQ agains t Clip Gauge Opening

Figure 5: Position of the Seam Welds

31

Figure 6:Testing Set-up for the Y-Joint Specimen

Figure 7:Static Strength Testing Setup for the T-Joint Specimen

32

Studding

Nuts

Brace Supporting Backing Box

Plate Welded Ring Sections Locking Ring Ring

Figure 8: Side View of the Box Section Used in the T-Joint Setup

Figure 9: Placement of Strain Gauges on the Brace of a Y-Joint

33

Figure 10: Chord Ovalisation Measurement Setup

Figure 11: Arrangement of the Retro-Reflective Targets on a T-Joint

34

Stable Structure Specimen

Reference Target

Camera

Computer

Figure 12: The Basic Setup for Data Collection Using Photogrammetry

Figure 13: The Difference Between Two Images From the Initial and the Latter stages of a test

35

Camera a

Camera b

Figures 14a and 14b: Photogrammetry Images taken Instantaneously by Cameras a and b

36

Figure 15: General Shape Obtained by Using Two Cameras 500mm Apart

Figure 16: Separated Brace from the Chord Following the Completion of the Test

37

Exp

erim

enta

l Fai

lure

Loa

d H

SE C

hara

cter

istic

Loa

d

2

1.5

1

0.5

0

HSS Data Mean

HSE Data Mean

Y1 Y4 Y3 T6 T1 T4 T2 T5 T3

0.0% 10.0% 20.0% 30.0% 40.0% 50.0% 60.0% 70.0% 80.0% 90.0% 100.0%

% Cracked Area

2.5

Figure 17: Normalised Ultimate Static Strength Capacity Using HSE Equations for Characteristic

Strength of Tubular Joints Compared With Previous Results Mean Line

Load

(kN

)

200

180

160

140

120

100

80

60

40

20

0 0 20 40 60 80 100 120 140

Displacement (mm)

1st Test 2nd Test 3rd Test

Figure 18: Load-Displacement Graph for Y1

38

0

50

100

150

200

250

300

Load

(kN

)

1st Test 2nd Test

0 50 100 150 200 250

Displacement (mm)


0

50

100

150

200

250

300

Load

(kN

)

1st Test 2nd Test 3rd Test

0 50 100 150 200 250

Displacement (mm)


39

1200

1000

800

600

400

200

0

Brace Disp. (LVDT) Chord Ovality Brace Disp. (PhotoG) Maximum COD L

oad

(kN

)

0 5 10 15 20 25 30 35 40 45 50

Displacement (mm)

Figure 21: Load-Displacement Curves for T6 Using Various Displacements

Mod

e I C

rack

Ope

ning

Dis

plac

emen

t (m

m)

0

2

4

6

8

10

12

14

16

18

20

-250 -200 -150 -100 -50 0 50 100 150 200 250

Position Around the Saddle (mm)

Original Crack Length 350mm (-190mm , +160mm)

0 kN 5 kN 20 kN 40 kN 68 kN 615 kN 825 kN 945 kN

Figure 22: Mode I Crack Opening Displacement at Various Loads for the First Test (T6)

40

Load

(kN

) 3500

3000

2500

2000

1500

1000

500

0

Chord Ovality Brace Disp. (PhotoG) Maximum COD

0 5 10 15 20 25 30 35 40 45 50

Displacement (mm)


Figure 24: Fracture Surface for Test T1

41

Mod

e I C

rack

Ope

ning

Dis

plac

emen

t (m

m)

0

5

10

15

20

25

30

35

50kN 718kN 1441kN 1930kN 1978kN 2100kN 2553kN 2601kN 2703 kN 2947 kN 1841kN


-200 -150 -100 -50 0 50 100 150 200


Figure 25: Mode I Crack Opening Displacement at Various Loads for the Second Test (T1)

Load

(kN

)

3500

3000

2500

2000

1500

1000

500

0 0 5 10 15 20 25 30 35 40 45 50

Displacement (mm)

Figure 26: Load-Displacement Curve for the Third Test (T4)

42

3500

3000

2500

2000

1500

1000

500

0

Brace Disp. (LVDT) Chord Ovality Maximum COD Lo

ad (k

N)

0 5 10 15 20 25 30 35 40 45 50

Displacement (mm)


Mod

e I C

rack

Ope

ning

Dis

plac

emen

t (m

m)

0

2

4

6

8

10

12

14 Original Top Crack Length 240mm (-110mm , +130mm)

200-200 -150 -100 -50 0 50 100 150


76 kN 250 kN 1254 kN 1790 kN 2296 kN 2520 kN 2790 kN 2618 kN

Figure 28: Mode I Crack Opening Displacement at Various Loads for the Fourth Test (T2)

43

3500

3000

2500

2000

1500

1000

500

0


oad

(kN

)

0 5 10 15 20 25 30 35 40 45 50

Displacement (mm)


Mod

e I C

rack

Ope

ning

Dis

plac

emen

t (m

m)

0

5

10

15

20

25

30

35

40

-200 -150 -100 -50 0 50 100 150 200



0 kN 400 kN 1180 kN 1820 kN 2150 kN 2550 kN 2730 kN 2840 kN 2650 kN 2680 kN

Figure 30: Mode I Crack Opening Displacement at Various Loads for the Fifth Test (T5)

44

3500

3000

2500

2000

1500

1000

500

0


oad

(kN

)

0 5 10 15 20 25 30 35 40 45 50

Displacement (mm)


Mod

e I C

rack

Ope

ning

Dis

plac

emen

t (m

m)

0

5

10

15

20

25

30

35 Original Crack Length 260mm (-140mm , +120mm)

250-250 -200 -150 -100 -50 0 50 100 150 200


0 kN 366 kN 870 kN 1276 kN 1900 kN 2350 kN 2790 kN 2860 kN 3092 kN 2894 kN

Figure 32: Mode I Crack Opening Displacement at Various Loads for the Sixth Test (T3)

45

Kr

1.0

0.8

0.6

0.4

0.2

Sr 1.0 0.8 0.6 0.4 0.2

0

SAFE

Level 2 Assessment Line

Level 1 Assessment Line

Collapse

Fracture

UNSAFE

Figure 33: Level 1 and Level 2 PD 6493 Failure Assessment Diagrams

1

0.8

0.6

Kr

0.4

0.2

0

Cut-off at 1.8 Typical of Austenitic Steels

Cut-off at 1.25 Typical of Mild Steels

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

Figure 34: Level 3 PD 6493 Failure Assessment Diagram

46

P

M

Figure 35a: Tube with a Through-Wall Flaw Subjected to Axial Tension (P) and Bending Moments

Figure 35b: Cross-Sectional Area of a Pipe Containing a Through-wall flaw

47

4

3.5

3

2.5

2

1.5

1

0.5

0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

T6 T1 T2 T5 T3 Y1 Y4 Y3

Kr

Figure 36: FAD Representing the Initial Tearing Load Point – K Q = 103.7MPa.m1/2

Kr

4

3.5

3

2.5

2

1.5

1

0.5

0

T6 T1

T4 T2

T5

T3 Y1

Y4 Y3

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

Lr 2

Figure 37: FAD Representing the Poin t of Maximum Load – KQ = 103.7MPa.m1/2

48

Kr

4

3.5

3

2.5

2

1.5

1

0.5

0

T6

T1 T2

T5

T3

Y1 Y4

Y3

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

Figure 38: FAD Plotted Using Loads Obtained Via the Twice Compliance Technique –

KQ = 103.7MPa.m1/2

Kr

4

3.5

3

2.5

2

1.5

1

0.5

0

T6 T1 T2 T5 T3 Y1 Y4 Y3

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

Figure 39: FAD Representing the Initial Tearing Load Point - HAZ KQ = 100.8MPa.m1/2

49

Kr

4

3.5

3

2.5

2

1.5

1

0.5

0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

T6 T1

T4 T2

T5

T3 Y1

Y4 Y3

Figure 40: FAD Representing the Point of Maximum Load - HAZ KQ = 100.8MPa.m1/2

Kr

4

3.5

3

2.5

2

1.5

1

0.5

0

T6 T1 T2 T5 T3 Y1 Y4 Y3

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

Figure 41: FAD Plotted Using Loads Obtained Via the Twice Compliance Technique HAZ KQ = 100.8MPa.m1/2

50

Kr

4

3.5

3

2.5

2

1.5

1

0.5

0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

T6 T1 T2 T5 T3 Y1 Y4 Y3

Figure 42: FAD Representing the Initial T earing Load Point – Kmax = 142.2MPa.m1/2

4

Kr

3.5

3

2.5

2

1.5

1

0.5

0 2 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8

Lr

T6 T1 T4 T2 T5 T3 Y1 Y4 Y3

Figure 43: FAD Representing the Poin t of Maximum Load – Kmax = 142.2MPa.m1/2

51

Kr

4

3.5

3

2.5

2

1.5

1

0.5

0

T6 T1 T2 T5 T3 Y1 Y4 Y3

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

Figure 44: FAD Plotted Using Loads Obtained Via the Twice Compliance Technique –

K max = 142.2MPa.m1/2

Kr

4

3.5

3

2.5

2

1.5

1

0.5

0

T6 T1 T2 T5 T3 Y1 Y4 Y3

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

Figure 45: FAD Representing the Initial Tearing Load Point - HAZ K max = 151.4MPa.m1/2

52

Kr

4

3.5

3

2.5

2

1.5

1

0.5

0

T6 T1 T4 T2 T5 T3 Y1 Y4 Y3

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

Figure 46: FAD Representing the Point of M aximum Load - HAZ Kmax = 151.4MPa.m1/2

Kr

4

3.5

3

2.5

2

1.5

1

0.5

0

T6 T1 T2 T5 T3 Y1 Y4 Y3

0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2

Lr

Figure 47: FAD Plotted Using Loads Obtained Via the Twice Compliance Technique

HAZ Kmax = 151.4MPa.m1/2

53

Printed and published by the Health and Safety ExecutiveC0.35 05/02

OTO 2000/078

£25.00

ISBN 0-7176-2307-6

9 780717 623075

http://www.hse.gov.uk

Static strength of cracked high strength steel tubular joints SUMMARY TABLE OF CONTENTS 1. INTRODUCTION 2. EXPERIMENTAL DETAILS 3. RESULTS AND DISCUSSION 4. CONCLUSIONS 5. REFERENCES

offshore technology report 2000/078 · 2019. 12. 13. · fatigue behaviour of the joints under...

Documents