- Email: [email protected]
Influence of residual stress on stress concentration factor for high strength steel welded joints
Journal of Constructional Steel Research 72 (2012) 20–28
Contents lists available at SciVerse ScienceDirect
Journal of Constructional Steel Research
Influence of residual stress on stress concentration factor for high strength steel welded joints Jin Jiang ⁎, Mingshan Zhao School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798
a r t i c l e
i n f o
Article history: Received 20 June 2011 Accepted 5 September 2011 Available online 14 October 2011 Keywords: Residual stress Stress concentration factor High strength steel Hole-drilling method
a b s t r a c t In this study, a set of plate-to-plate T and Y joints specimens made from high strength steel plates with yield stress equal to 690 MPa is investigated. The joints are fabricated by SMAW welding procedure. Two groups of specimens with different welding procedures are included: one group is composed by the joints with welding completed at ambient temperature and the other group is composed by the joints with welding completed at a preheating temperature of 100 °C. The residual stress near the weld toe is investigated for both groups. Hole-drilling method is applied to investigate the residual stress distribution and variation in joints. Sequentially coupled thermal-stress analysis is then conducted with finite element package ABAQUS to investigate the residual stress distribution in the joints. Finally, the effects of residual stress on the stress concentration factor distributions of the joints are evaluated. A new parameter is put forward in stress concentration factor evaluation to combine the residual stress effect. © 2011 Elsevier Ltd. All rights reserved.
1. Introduction High strength steels, defined as steel with minimum yield strength of greater than 460 MPa, are increasingly used in industries. High strength steel was firstly used in Japan in the 1960s for construction and after that more and more countries show great enthusiasm on this area. Landmark Tower in central Yokohama is the first project using high strength steel in the construction of building in Japan [1]. The first UK project using high strength steel is the Hutton Field floating structure, 200 km off the coast of Norway. The tension-leg platform was fabricated from steel with minimum yield strength of 795 Mpa [2]. Another application of high strength steel with minimum tensile strength 600 MPa is the Shimizu super high rise building which is 550 m high comprising 127 levels for reducing the column section size [1]. In America, a research programme was initiated to develop a better performance steel specifically for use in bridges to improve the weldability, toughness and weathering resistance of existing steel grade. So, it can be concluded that HSS is applied broadly in the world for its merits such as higher strength to cost ratio and artistic. However it is still short of understanding some issues of high strength steel. Firstly, the stress–strain behaviour of high strength steel is different from mild steel in respect that high strength steel generally exhibits reduced capacity for strain hardening after yielding. There is restriction in structural design that yield ratio is not allowed to have
⁎ Corresponding author. E-mail address: [email protected] (J. Jiang). 0143-974X/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcsr.2011.09.007
a value greater than 0.85 in design equations to ensure a adequate ductility in the member to develop plastic failure behaviour as a defence against brittle fracture. Secondly, it is regarded that fatigue is one of the major problem causing the degradation of offshore structures in long term integrity. Thirdly, the application of high strength steels is limited by codes and standards involving application of design formulae are restricted to steels with yield strength smaller than 500 MPa in most cases. Finally, the residual stresses due to welding in high strength steel may do harm to the integrity of structures. Residual stress not only affects the initiation and onset of the propagation of surface cracks but also changes the path of a crack as it grows below the surface. It is mentioned that the effect of residual welding stresses on the performance of welded structure is particularly significant when low stresses are applied [3]. For high strength steel structures, how serious will the residual stresses impose influence on the behaviour of those structures is still obscure. Therefore, accurately understanding of the distribution of stress concentration around welded intersection and the effects of residual stress on the stress concentration factors are of great importance since it plays a key role in the prediction of fatigue life. In this paper, the residual stress distribution a set of plate-to-plate T and Y joints specimens made from high strength steel plates with yield stress equal to 690 MPa is investigated. To investigate the influence of preheating on the residual stress, two groups of specimens with different welding procedures are investigated for the residual stress with hole-drilling method: one group is composed by the joints with welding completed at ambient temperature and the other group is composed by the joints with welding completed at a preheating temperature of 100 °C. The plate-to-plate T and Y joints are chosen in this investigation
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
3. Experimental specimens
Table 1 Mechanical properties of RQT701 steel plate and LB-70L electrode. Items
Minimum Yield Tensile Strength Strength (MPa) (MPa)
RQT701 690 LB-70L 685
Minimum Average Minimum Impact Energy elongation (%) 27 J@ −45 °C 108 J@ −60 °C
790–930 755
21
18 27
in respect that this kind of joints is broad application in structural engineering and it is also basic and key building block for box sections and joints. Additionally, little residual stress data can be found for this new material. Sequentially coupled thermal-stress analysis is then conducted with finite element package ABAQUS to investigate the residual stress distribution in the joints. After that, the stress concentration factor under the action of static axial tensile loading applied on the brace plate is studied. Finally, the effects of residual stress on the stress concentration factor distributions of the joints are evaluated.
2. Historical background Extensive research into steel with tensile strength of 600 MPa are carried out since 1988 in Japan for looking for benefits in the prevention of damage caused by earthquakes. Stacy [4] listed some important issues for application of high strength steel in offshore engineering. Agerskov and Petersen [5–6] carried out the comparative investigation on fatigue performance of welded joints made of high strength steel and mild steel in constant amplitude and variable amplitude by full scale offshore tubular test and a fracture mechanics analysis. Sharp [7] summarized guidance and code topics relevant to high strength steel. Miki [8–9] reviewed the histories of the development and use of high strength steel for bridge structures in Japan. It was mentioned that the biggest problem of application for high strength steel is to achieve a balance between tensile strength and fatigue performance without losing good weldability. Additionally, fatigue performance of typical welded joints including longitudinal, cruciform and out-of-plane gusset welded joints were discussed with respect to the dependence on steel strength and joint size. HSE Research Report 105[10] gave a detailed review of the performance of high strength steel used in offshore. Bjorhovde [11] gave a brief review of mechanical properties of high strength steels, including the properties of tensile strength and stiffness, ductility, toughness, weldability and corrosion resistance. The effect of residual stress on fatigue behavior of high strength steel components has also spurred great interesting recently. Sonsino [12– 13] evaluated effect of residual stresses on the fatigue behavior of high strength steel weld joints under different kinds of loading. Stacey reviewed residual stresses distributions in several kinds of welded joints [14]. Transverse and longitudinal through-thickness residual stress distributions in range of geometries of welded CHS joints are summarized. Experimental work to determine residual stress distribution at the chord weld toe of three T circular section joints was carried out by Payne and Porter-Goff [15]. Three tubular joints with yield stress ranging from 365 to 491 MPa were investigated. Block sectioning was applied to determine through-thickness distribution of residual stresses in the weld toe region at crown and saddle positions.
In the present experimental investigation, a number of plate-toplate T and Y joints, made of high strength steel with minimum yielding stress of 690 MPa, were fabricated by welding. This high strength steel, RQT701, which was supplied by Corus Group, is quenched and tempered structural steel with improved forming and welding performance by substituting some alloying element with carbon. In the process of welding, an ultra low hydrogen and moisture resistant type covered electrode for 690 MPa high tensile strength steel for low temperature service, LB-70L, which is equivalent to the class ASME/AWS A5.5 E10016-G and supplied by Kobelco of Japan, was employed [16]. The diffusible hydrogen content of this electrode is 4 ml/100 mg. The welding procedure was carried out according to AWS D1.1 2008 [17]. Table 1 gives the mechanical properties of the RQT701 plate and LB-70L electrode. Table 2 lists the welding parameters for fabrication. Two series of specimens were included to compare the influence of preheating on residual stress distribution near the weld toe. The first group consisted of joints with welding completed at ambient temperature while the second group consisted of joints with welding completed at a preheating temperature of 100 °C. For each group, there are 12 different geometries, consisting of 3 different plate thicknesses and 4 joint angles. They were employed to explore the variation of the residual stress near the weld toe when geometry is different. For the 135° joints, the welding profile is shown in Fig. 1. Fig. 2 is applied for the welding profile of 60°, 75° and 90° joints (Table 3). 4. Experimental investigation for residual stress 4.1. Theorical basis The localized stress is relaxed by introduction a hole into a residually stressed structure and correspondingly the local strain on the surface of the testing structure is changed. It is the principle of the hole-drilling method of measuring of residual stress firstly proposed by Mather [18]. The initial stress state at point P(R, α) in a thin plate subjected to a uniform residual stress, as shown in Fig. 3, can be written as follow in polar coordinates: 8 0 σx > > > σr ¼ 2 ð1 þ cos2αÞ > < 0 σ σθ ¼ x ð1− cos2αÞ > 2 > 0 > > : τrθ ¼ − σx ð sin2αÞ 2
ð1Þ
When a hole is drilled through in the center of a plate, as shown in Fig. 4, the stresses in the vicinity of the hole was obtained by Kirsch and can be expressed as [19]: 8 00 σx 1 σ 3 4 > > 1− 2 þ x 1 þ 4 − 2 cos2α > σr ¼ > 2 > r 2 r r > < 00 σ 1 σ 3 σθ ¼ x 1 þ 2 − x 1 þ 4 cos2α > 2 2 r r > > > 00 σ 3 2 > > : τrθ ¼ − x 1− 4 þ 2 cos2α 2 r r
ð2Þ
where: r is the ratio of R to R0; R and R0 denote hole radius and arbitrary radius from hole center respectively. By comparing the stress in
Table 2 Welding parameters for the specimens. Welding position
Base Metal
Current
Flat
RQT701
DCEP
Heat Treatment Preheat
Non-preheat
100 °C
30 °C
Interpass Temp.
Welding Condition
150 °C
170 A
26 V
2.2 kJ/mm
22
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28 Table 3 Specification of the specimens.
t2
F
y
θ (°)
β (°)
L1 (mm)
L2 (mm)
W (mm)
t1 (mm)
t2 (mm)
R (mm)
L (mm)
135
N.A.
440
440
420
60
30
440
440
420
75
60
440
440
420
90
60
440
440
420
8 12 16 8 12 16 8 12 16 8 12 16
8 12 16 8 12 16 8 12 16 8 12 16
5 5 5 6 6 6 6 6 6 6 6 6
16 22 28 22 31 41 21 30 39 20 27 35
l2
E
θ
G J I
H
D
tw
C
R
B
t1 x
l
A
l1 Fig. 1. Welding Profile for 135° joints.
this point before and after hole-drilling, stress relaxation can be determined due to drilling the hole. On the assumption that the material of the plate is homogenous and isotropic and its stress–strain curve is linear, the relieved strain in the point P(R, α) can be obtained by substituting the stress relaxation into biaxial Hooke's law: 8 i ε1 þ ε3 1 h 2 2 0:5 > > > σ max ¼ 4A − 4B ðε3 −ε1 Þ þ ðε3 þ ε1 −2ε2 Þ > > < i ε þ ε3 1 h 2 2 0:5 þ σ min ¼ 1 ðε3 −ε1 Þ þ ðε3 þ ε1 −2ε2 Þ > 4B 4A > > > ε þ ε3 −2ε2 > : tan2α ¼ 1 ε3 −ε1 where; A ¼ −
ð3Þ
1þυ 1 1þυ 4 1 3 − ; B ¼ − ; 2E 2E 1 þ r r2 r4 r2
pads as shown in Fig. 5. The milling cutter should be guided carefully so as to make the cutter progress in a straight line devoid of side pressure on the hole and friction at the non-cutting edge. A high-speed air turbine was employed to form good hole shape and adaptability to incremental drilling as shown in Fig. 6. To make the measured points close to the weld toe of the joint, a special supporting set was designed as shown in Figs. 5 and 6. In the process of test, the ASTM 837-08 was followed [20]. Since there is physical limitation for the test rig that it cannot be very close to the weld toe, the 75° and 90° joints are cut for their brace plates.
4.3. Strain gauge scheme
4.2. Test setup The RS-200 Milling Guide, a high-precision instrument for analyzing residual stress by the hole-drilling method through positioning and drilling of a hole in the center of a special strain gauge rosette, was applied for measuring the residual stress in the specimens. Its ruggedness and flexibility make it equally suitable for laboratory or field application. Since positioning precision of the milling guide has great influence on the accuracy of measurement, the RS-200 milling guide with a microscope installed was applied herein and it is secured to samples with quick-setting and frangible adhesive to bond its foot
A special type of strain rosette, FRAS-2, which is designed by TML to facilitate positioning three grids on one side of the measurement point, was used to measure the released strain of the specimen during drilling. For the 135° and 60° joints, in the longitudinal direction (the x axis in Fig. 7), three strain gauges, denoted by A, B and C, were positioned at points where the x coordinate are respectively equal to 25 mm, 75 mm and 125 mm with the 5 mm away from the weld toe. In the transverse direction (the y axis in Fig. 7), in the middle of plate width, where the x coordinate is 75 mm, another two strain gauges denoted as B1 and B2 were positioned at the points where the y coordinate are respectively equal to 20 mm and 35 mm. For the 75° and 90° joints the strain gauge A and C were removed. In the middle of plate width, another strain gauge B3 was added at the 50 mm point. Table 4 lists residual stress testing results.
F E
y
Y
l2
σ max σ '' r
R0
θ
σ '' θ R
R D C
β J I
σ min
θ
G H
l
x
X
B
t1 A
σy
σx
l1 Fig. 2. Welding Profile for 120°, 105° and 90° joints.
Fig. 3. Stress state at point P(R, α) before drilling a hole.
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
23
5. Numerical investigation for residual stress 5.1. Modelling procedure In this paper, the commercial finite element modelling package ABAQUS was used in process of welding simulation. A sequentially coupled thermal-stress analysis was conducted by assuming that the stress/displacement solution is dependent on a temperature field but there is no inverse dependency. Sequentially coupled thermal-stress analysis was performed by first solving the pure heat transfer problem, then reading the temperature solution into a stress analysis as a predefined field. Firstly, a non-linear transient thermal analysis was conducted to determinate the temperature distribution during the welding. Secondly, the temperature data was used as a thermal loading input for the mechanical analysis. Before running the modelling, several assumptions for the modelling were put forward. First, the stress due to the welding has negligible influence on the temperature field. Second, the latent heat, the amount of energy released or absorbed by a substance during a change of state that occurs without changing its temperature, is ignored in the modelling. In the paper, only the 135° joint with 16 mm base plate is chosen for numerical modelling to investigate the residual stress distribution. 5.2. Thermal analysis In the thermal analysis, transient non-linear analysis was conducted for 2D plate model to determine the temperature history throughout the plate. Element ‘birth and death’ technique was adopted to simulate the weld filler variation with time. All the elements, including the base plate and weld filler (regions IJGH in Figs. 1 and 2), were created firstly. However, the elements in the weld filler were deactivated before analysis. The weld filler was divided into four lumps. The first lump was activated at 0.001 s. The welding speed was 2.6 mm/s so that the time to finish the first run of 150 mm was 57.7 s. Then, the second lump was activated within 0.001 s and the heat transfer time was again set as 57.7 s. Such procedure was then repeated for the third and forth lumps. Thermal conductivity, k, which is the property of a material reflecting its ability to conduct heat, is obtained from the standard EC3 (1–2) [7]. The melt droplets were modelled by a volumetric heat source with uniform density. The move of heat source was achieved by defining the amplitude of heat source density curve. The heat source density was defined by the equation:
q¼
0:8UI le hl v
High speed turbine assembly
Supporting set
Fig. 5. High speed drilling setup for the RS-200 milling guide.
In Eq. (4), 0.8 is the arc efficiency factor, I is the arc current, U is the arc voltage, le is the depth of each lump and hl is the height of each lump. To consider the heat losses from the surface of welded plate, both convection and radiation were assumed, as shown in Eq. (5). The convection coefficient h was assumed as 15 W/ (m 2K) and the emissivity ε is defined as 0.2. 5.3. Mechanical analysis Transient non-linear stress analysis was carried out for recognizing the distribution of residual stress. In the mechanical analysis, the temperature history obtained from the thermal analysis was input as a thermal loading into the structural model. Temperature dependent mechanical properties, including the thermal expansion coefficients, Young modulus, Poisson's ratio and yield strength were calculated by applying the reduction factors listed in the EC3(1–2) [7]. The mesh and analysis steps were as the same used in the thermal analysis. 5.4. Modelling results Since only the part near the weldment in base plate is interested, the results for temperature variation and residual stress distribution in this part, in the range of 50 mm from the weld filler contour, are shown.
ð4Þ
Brace Plate
Microscope 15 15 15 5 25
C
Added in 75 and 90 specimens
50
weld toe
B3 B2 B1 B 50
A
Supporting set
Fig. 4. Alignment setup for the RS-200 milling guide.
25
y
x Chord Plate
Fig. 6. Strain gauges scheme for residual stresses measurement.
24
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
Table 4 SCF testing results. Joint geometry
Ambient temperature
Preheating
Degree
Thickness(mm)
A
B
C
A
B
C
135
8 12 16 8 12 16 8 12 16 8 12 16
1.43 1.61 1.70 1.61 1.00 1.63 1.46 1.68 1.52 1.62 1.53 1.78
1.53 1.82 1.95 1.41 1.15 1.68 1.57 1.74 1.55 1.83 1.84 2.15
1.48 1.74 1.72 1.54 2.44 1.45 1.51 1.52 1.43 1.76 1.68 1.65
1.54 1.48 1.76 1.60 1.67 2.10 1.41 1.57 1.46 1.54 1.72 1.62
1.62 1.53 1.96 3.10 1.73 3.25 1.69 1.83 1.59 1.94 1.88 1.88
1.60 1.30 1.88 1.92 1.58 4.84 1.54 1.62 1.52 1.62 1.76 1.80
60
75
90
Fig. 8. The temperature distribution at t = 58.7 s (2nd lump was added).
6. Test for stress concentration factor Fig. 8 illustrates the temperatures at 1.0 s when the 1st lump is activated. At this moment, the heat is localized in the weld filler and the temperature in base steel does not raise much. Also, the 2nd, 3 rd and 4th lumps are still deactivated at this moment, the elements in those areas can be seen as ‘blank’ and they do not have function of heat conduction. When the 2nd lumped of the weld filler is activated, as shown in Fig. 9, the maximum temperature is shifted to this new lump. However, it can be seen that the maximum temperature in the joint do not change much. It just enlarges the area that heat affected. It is similar for the 3 rd lump and 4th lump, as shown in Figs. 10 and 11. After that, the cooling procedure comes in for the joint. Figs. 12 and 13 give the temperature distribution at 800 s and 2500 s. When the time went to 2500 s, the temperatures in the joint reduce to near 30 °C. Figs. 14 and 15 show the transverse stress when the joint completely cooled down. Fig. 14 illustrates the stress distribution when the 100° preheating is carried out. It can be found that the transverse stress shows a layered distribution in the chord plate. The stress in the weld is particularly high when compared with the stress in the base plates. High transverse residual stress is localized at weld toe and weld root position. Another place with high transverse stress need to be noticed is the lower surface of the chord plate rightly below the weld. From Figs. 14 and 15, it can be found that the residual stress in the plate surface only localized in the vicinity of the weld toe and the weld root. When the distance from the weld toe is 15 mm, the magnitude of the residual stress has dropped down quickly. Figs. 16 and 17 give the comparison for the modelling results with the corresponding values from the testing. The modelling results seem a litter higher than the values obtained from the test at the point of 5 mm from the weld toe. In the 2D modeling, the plane strain elements are used so that the structural deformation along the plate width is ignored.
Fig. 7. The temperature distribution at t = 1.0 s (1st lump was added).
6.1. Test setup The Instron Model 8506 Dynamic Materials Testing System is introduced when tensile loading is to be applied in the specimen. With maximum tensile loading capability of 2000 kN, it is an advanced multiprocessor-based control console which provides full digital control of a testing system. It consists of a closed load and four columns frame with movable crosshead, a hydraulic actuator to apply a force, gripping mechanisms to hold the mechanical test specimen, and a load cell to measure the force. The position of the actuator, under closed loop control by controlling the hydraulic fluid flowing through a servo-valve supplying the actuator, is measured by a displacement transducer. 6.2. Specimen assembly To fix the specimen in the grip of testing machine, a set of supporting joints, made of mild steel S355 with thickness of 45 mm, is designed. It is designed in such a way that have nearly triple thickness when comparing with specimens to make sure the failure will turn out in specimens rather than the supporting joints. The specimen and supporting joint is connected by 12 high strength hexagon bolts of grade 10.9HR. In each end of connection, 6 bolts are positioned in two lines. Fig. 18 shows the profile of the specimen and supporting joint after assembly. Fig. 19 is the full view of the testing machine when the specimen and the supporting set are fixed in the grips. 6.3. Strain gauge scheme FLA-2 of TML strain gauges, which has only one grid, were adhered on the surface of the specimen. In the direction of width of specimen, three strain gauges in a line were used. In the direction of length of specimen, two lines strain gauges, which were away from weld toe 5 mm and 15 mm respectively, were applied, shown in Fig. 20.
Fig. 9. The temperature distribution at t = 116.4 s (3rd lump was added).
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
Fig. 12. The temperature distribution at t = 2500 s.
Fig. 10. The temperature distribution at t = 230.1 s (4th lump was added).
When stress concentration factor and hot spot stress in weld toe of the samples were focused, the stress distribution along direction of weld toe was investigated by monitoring the strains near the weld toe shown subjected to tensile loading in brace plate. Table 5 is the SCF testing results for all the joints.
In the experimental investigation, a carefully planned testing was organized to analyze the residual stresses and SCF values in various high strength steel plate-to-plate joints. Linear interpolation method is employed to estimate the residual stress at the weld toe. In addition, linear interpolation method is used for hot spot stress at the weld toe and SCF values without considering the residual stress are illustrated. By comparing the SCF values without residual stress effect and another group SCF values incorporating residual stress, the influence of residual stress on SCF is analyzed. At any point in the specimen, the total stress which is assumed to be smaller than the yielding stress σy can be expressed as: ð5Þ
In Eq. (5), σrs is the residual stress at a certain point in the specimen; σf is the elastic stress caused by the applied load in the specimen. According to the definition of SCF, SCFf can be written as: SCFf ¼
σf σn
ð6Þ
In Eq. (6), SCFf is the stress concentration factor without considering residual stress; σn is the nominal stress caused by the applied load in the specimen. When residual stress is considered, the equation for the SCF should be expressed as: SCFtotal ¼
σtotal σ rs þ σf σ rs þ SCFf ⋅σn σ rs ¼ ¼ ¼ þ SCFf σn σn σn σn
Fig. 11. The temperature distribution at t = 800 s.
In Eq. (7), SCFtotal is the stress concentration factor including residual stress and stress caused by the applied load. By defining SCFrs ¼ σσrsn , the equation can be written as: SCFtotal ¼ SCFrs þ SCFf
ð8Þ
Therefore, the effect of residual stress can be measured by the residual stress factor (RSF), which is defined as:
7. Analysis and discussion
σtotal ¼ σrs þ σf
25
ð7Þ
RSF ¼
SCFrs SCFtotal
ð9Þ
The specimen nomenclature for Figs. 21, 22, 23 and 24 is that the first term in the legend denotes the intersection angle (135°, 60°, 75° and 90°) of the plate, the second term denotes the plate thickness and the final team identifies whether preheating was applied. The character ‘P’ herein denotes joints with preheating and ‘A’ indicates the specimens welded in ambient temperature. From Eqs. (8) and (9), it can be seen that when tensile residual stress exists SCFrs will be smaller than SCFtotal and hence RSF shall within the range [0, 1]. In addition, in general, with the increase of the applied stress, the RSFs will be reduced gradually from 1. This fact also implies that whenever the RSF value is outside the range of [0, 1], compressive residual stress exists. Fig. 21 illustrates the RSF values variations of 135° joints in different applied stress. It can be seen that when that the applied stress is equal to 50Mpa, at the 5 mm point, the RSFs for the 8 mm, 12 mm and 16 mm joints without preheating are equal to 0.65, 0.51 and 0.69, respectively. The correspondingly values for the preheating joints with 8 mm, 12 mm and 16 mm plate thickness are equal to 0.61, 0.41 and 0.58, respectively. Hence, it can be seen that at 5 mm point, the RSF for the preheating joints was slight smaller than values without preheating. When the applied stress increases to 250Mpa, RSFs for the 8 mm, 12 mm and 16 mm joints without preheating are equal to 0.27, 0.17 and 0.30, respectively. The correspondingly RSFs for the preheating joints with 8 mm, 12 mm and 16 mm plate thickness are equal to 0.24, 0.12 and 0.20, respectively. Therefore, when
Fig. 13. The transverse residual stress for the 135°joint welded in ambient temperature.
26
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
t
420 Specimen θ Fig. 14. The transverse residual stress for the 135°joint welded with 100 °C preheating.
Tranverse residual stress (MPa)
200
t 45
Testing Modelling
150
Supporting Joint 360
100
50
0
45 -50
0
10
20
30
40
50
60
Fig. 17. Assembly of the specimen and supporting joint.
Distance from the weld toe (mm) Fig. 15. Comparison of transverse residual stress between testing and modelling for 135° joint with 16 mm base plate welded at ambient temperature.
shown in Fig. 22. It is concluded that the influence of residual stress is more serious when a low applied stress exists. Fig. 23 shows the relationship between RSFs and the applied stress for 105° joints. For the 75°–8 mm-P joint and the 75°–8 mm-A joint,
the applied stress increases from 50 MPa to 250 MPa, the RSFs for the 8 mm, 12 mm and 16 mm joints without preheating were reduced by 58.5%, 66.7% and 56.5%, respectively. For the 8 mm, 12 mm and 16 mm joints with preheating, the correponding RSFs reduction are equal to 60.6%, 70.7% and 65.5%, respectively. Thus, even though the residual stresses in joints without preheating are higher than the joint with preheating, the RSFs reduction for the joints without preheating is faster than the preheated joint. When the applied stress increases from 250 MPa to 400 MPa, RSFs reduce 30% approximately for all the specimens. Similar conclusion can be drawn from 60° joint, as
Tranverse residual stress (MPa)
200 Testing Modelling
150
100
50
0
-50
0
10
20
30
40
50
60
Distance from the weld toe (mm) Fig. 16. Comparison of transverse residual stress between testing and modelling for 135° joint with 16 mm base plate welded at 100 °C preheating.
Fig. 18. Full view of testing machine.
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
27
RSF
0.8
15 15 15 5
0.7
8mm-P
8mm-A
0.6
12mm-P
12mm-A
0.5
16mm-P
16mm-A
0.4 0.3 0.2
30
0.1 45
0.0 weld toe
0
50
100
150
200
250
300
350
400
Nominal stress (MPa)
45
Fig. 20. RSF of 135° joints under different nominal stresses.
30
the 90° joints. Similar to the case shown in Fig. 21, negative RSFs were resulted when the applied stress is 50 MPa which indicated the existence of compressive residual stress.
Fig. 19. Strain gauge scheme for SCF measurement.
8. Conclusions when the applied stress is 50 MPa, the residual stresses in the weld toe are compressive so that RSFs are negative and outside the range [0, 1]. For the 75°–12 mm-P joint, the RSF at the 5 mm point was equal to 5.78. It is greatly outside the range [0, 1] as that the compressive residual stress at the 5 mm point is − 105.2 MPa which is larger than the applied stress. When the applied stress was increased to100MPa, the RSF values was −1.53 because SCFtotal becomes positive. With further increase of the applied stress, the RSF will approach into 0. Note that for a given applied stress, there exists a critical compressive residual stress which will lead to zero total stress (or SCFtotal = 0) and therefere the RSF will approach infinity. Furthermore, when the applied stress is slightly smaller than this critical value, the RSF approaches positive infinity quickly while when the applied stress is larger than this critical value, the RSF reduces quickly to minus infinity value. For the 105°–12 mm-P joint, this critical value is roughly equal to 61.8 MPa. It alos can be seen that when the applied stress is larger than 250 MPa, the RSF is comparatively small when comparing with low stress is applied. Fig. 24 gives an illustration for
Based on the analysis in the paper, some conclusions are drawn: (1) Frequently, the maximum tensile residual stress perpendicular to weld toe appears in the middle of plate. In the transverse direction, the magnitude of residual stress reduces greatly from the point B to point B2 for 135° joints; however, it may increase in a smaller magnitude from B1 to B2. It means that the change of residual stress with the distance away from weld toe is non-linear. When the distance from the weld toe is 50 mm, the residual stress fluctuates around 0. (2) Preheating can effectively reduce the magnitude of residual stress of near the weld toe. Therefore, it is significant that evenly high-quality preheating of the steel plate should be applied during the welding of high strength steel joints. (3) When the applied stress ranges from 0 to 200 MPa, the influence of residual stress on SCF is obvious. When the applied stress is beyond 200 MPa, this influence is much relieved. The
Table 5 Residual stress testing results. Specimen cases θ (°) 135
Stress in the measuring points (MPa) t (mm) 8 12 16
60
8 12 16
75
8 12 16
90
8 12 16
Preheating treatment
Brace–cutting treatment
A
B
Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No
No No No No No No No No No No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
− 78.8 23.2 − 82.4 − 15.6 47.7 12 − 29.5 76.6 88.1 − 75.3 − 11.4 − 18.5 – – – – – – – – – – – –
119.1 151 62.8 102.6 132.8 213.6 118.5 240.4 108.7 239.7 87.1 278.4 − 32 − 46.5 –105.2 35.2 27.3 139.2 − 66.2 − 70.3 − 32.5 − 8.8 51.1 24.9
B1
B2
99.2 96 48.5 79.7 114.9 103.9 56 63.3 22.5 127.6 90.5 − 17.6 137 36.7 45.6 96.9 45.6 243.6 76.4 129.0 34.6 67.5 91.9 52.4
118 90.9 36.9 32.3 53.4 68.1 65.3 93.7 43.5 60 79.8 62.2 36 22.3 12.4 24.4 − 7.7 58.1 25.2 28.2 11.2 26.9 25.2 38.2
B3
C
– – – – – – – – – – – – − 10.5 − 12.6 − 13.5 − 22.7 15.6 19.1 8.0 28.5 6.2 5.3 5.2 –26.4
42.6 –25.6 –68.2 –42.7 16 –68.5 –27.8 –29.4 –47.6 81.8 –76.4 –45.8 – – – – – – – – – – – –
28
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
0.9
References
0.8
8mm-p
8mm-A
0.7
12mm -P
12mm -A
0.6
16mm -P
16mm -A
RSF
0.5 0.4 0.3 0.2 0.1 0.0
0
50
100
150
200
250
300
350
400
Nominal stress (MPa) Fig. 21. RSF of 60° joints under different nominal stresses.
20 15
8mm-P
8mm-A
10
12mm-P
12mm-A
16mm-P
16mm-A
RSF
5 0 -5 -10 -15 -20
0
50
100
150
200
250
300
350
400
Nominal stress (MPa) Fig. 22. RSF of 75° joints under different nominal stresses.
influence of residual stress on stress concentration effect is evident when low stress is applied. (4) RSF is an effective parameter that can indicate the effect caused by the residual stress. When the tensile residual stress exists, the RSF should be in the range (0, 1). When the compressive residual stress exists, the RSF will jump to a high value (which is plus in algebra) firstly and with the increase of the applied stress, the parameter will drop down to a minus value. When the applied stress increases continually, the RSF gradually approaches to 0. (5) Residual stress does not always harmful to the integrity of structures. When the compressive residual stress exists, residual stress could be is beneficial as it could relieve the stress concentration effect caused by the applied load.
1 0
RSF
-1 -2
8mm-P
8mm-A
12mm-P
12mm-A
16mm-P
16mm-A
-3 -4
0
50
100
150
200
250
300
350
Nominal stress (MPa) Fig. 23. RSF of 90° joints under different nominal stresses.
400
[1] Pocock Graham. High strength steel use in Australia, Japan and the US. The Structural Engineer Nov 2006:27–30. [2] Salama MM, Tetlow JH. Selection and Evaluation of High Strength Steel for Hutton TLP Tension Leg Elements. Offshore Technology Conference, Houston, Texas; May 1983. [3] Clarin Mattias. High strength steel local buckling and residual stresses. Lulea, Lulea University of Technology: Department of Civil and Environmental Engineering; 2004. Licentiate. [4] Stacey A, Sharp JV, King RN. High strength steels used in offshore installations. International Conference of OMAE; 1996. p. 417–33. Volume III. [5] Agerskov, H., Petersen R.I., et al. An investigation on fatigue in high-strength steel offshore structures. Welding in the World 1997;41(4):328–42. [6] Petersen, R.I., Agerskov, H., and Lopez, M.L. Fatigue life of high strength steel offshore tubular joints, Report of Technical University of Denmark, Lyngby, Denmark. ISBN: 87-7740-167-0. [7] Sharp JV, Billingham J, Stacey A. Performance of high strength steel used in jackups. Marine Structures 1999;12:349–70. [8] Miki C, Homma K, Tominaga T. High strength and high performance steels and their use in bridge structures. Journal of Constructional Steel Research 2002;58:3–20. [9] Anami K, Miki C. Fatigue strength of welded joints made of high-strength steels. Progress in Structural Engineering and Materials 2001;3:86–94. [10] Billingham J, Sharp JV, et al. Review of the performance of high strength steels used offshore. Bedfordshire, UK; 2003. [11] Bjorhovde R. Development and use of high performance steel. Journal of Constructional Steel Research 2004;60:393–400. [12] Sonsino C.M., Lagoda T., et al. Damage accumulation under variable amplitude loading of welded medium and high strength steels. International Journal of Fatigue 2004;25(5):487–95. [13] Sonsino C.M. Effect of residual stresses on the fatigue behaviour of welded joints depending on loading conditions and weld geometry. International Journal of Fatigue 31(1): 88–101. [14] Stacey, A., Barthelemy,J.Y., et al. Incorporation of residual stresses into the SINTAP defect assessment procedure. Engineering Fracture Mechanics 2000;67(6):573–611. [15] Payne, J.G., Porter-Goff, R.F.D. Experimental residual stress distributions in welded tubular T-nodes. I Mech E Conference Publications Institution of Mechanical Engineers 1986:109–116. [16] AWS. ANSI/AWS A5.5. Specification for Low-Alloy Steel Electrodes for Shield Mtetal Arc Welding. Miami, USA: American Welding Society; 2006. [17] AWS. ANSI/AWS D1.1. Structural Welding Code-Steel. Miami, USA: American Welding Society; 2008. [18] Mather J. Determination of Initial Stresses by Measuring the Deformation Around Drilled Holed. Transactions of the ASME 1934;56(4):249–54. [19] Kabiri M. Measurement of Residual Stress by the Hole-Drilling Method:Influence of Transverse Sensitivity of the Gages and Relieved Strain Coefficients. Experimental Mechanics 1984;25:252–6. [20] ASTM. E837-08. Standard Test Method for Determining Residual Stresses by HoleDrilling Strain-Gage Method. West Conshohocken, PA 19428–2959, United States: ASTM International; 2008.
Nomenclature A,B: Calibration constant for hole-drilling method E: Young's modulus h: Convection coefficient hl: The height of each lump of weld k: Thermal conductivity I: The arc current l: Weld filler length in the chord plate surface le: Depth of each lump l1: The length of the chord plate l2: The length of the brace plate q: Heat source density R: Gauge circle diameter or weld root length R0: Diameter of the drilled hole t1: The thickness of the chord plate t2: The thickness of the brace plate U: The arc voltage W: Plate width β: End preparation angle for the specimens ε: Relieved strain due to the hole-drilling εe: The emissivity coefficient θ: Joint angle or the angle between the uniform normal stress in r direction and x ν: Poisson's ratio σc: Calibration stress σr: Uniform normal stress in r direction σθ: Uniform normal stress in θ direction τrθ: Uniform shear stress σmax: Maximum principle stress σmin: Minimum principle stress AWS: American Welding Society HSS: High strength steel RSF: Residual stress factor SCF: Stress concentration factor
Contents lists available at SciVerse ScienceDirect
Journal of Constructional Steel Research
Influence of residual stress on stress concentration factor for high strength steel welded joints Jin Jiang ⁎, Mingshan Zhao School of Civil and Environmental Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798
a r t i c l e
i n f o
Article history: Received 20 June 2011 Accepted 5 September 2011 Available online 14 October 2011 Keywords: Residual stress Stress concentration factor High strength steel Hole-drilling method
a b s t r a c t In this study, a set of plate-to-plate T and Y joints specimens made from high strength steel plates with yield stress equal to 690 MPa is investigated. The joints are fabricated by SMAW welding procedure. Two groups of specimens with different welding procedures are included: one group is composed by the joints with welding completed at ambient temperature and the other group is composed by the joints with welding completed at a preheating temperature of 100 °C. The residual stress near the weld toe is investigated for both groups. Hole-drilling method is applied to investigate the residual stress distribution and variation in joints. Sequentially coupled thermal-stress analysis is then conducted with finite element package ABAQUS to investigate the residual stress distribution in the joints. Finally, the effects of residual stress on the stress concentration factor distributions of the joints are evaluated. A new parameter is put forward in stress concentration factor evaluation to combine the residual stress effect. © 2011 Elsevier Ltd. All rights reserved.
1. Introduction High strength steels, defined as steel with minimum yield strength of greater than 460 MPa, are increasingly used in industries. High strength steel was firstly used in Japan in the 1960s for construction and after that more and more countries show great enthusiasm on this area. Landmark Tower in central Yokohama is the first project using high strength steel in the construction of building in Japan [1]. The first UK project using high strength steel is the Hutton Field floating structure, 200 km off the coast of Norway. The tension-leg platform was fabricated from steel with minimum yield strength of 795 Mpa [2]. Another application of high strength steel with minimum tensile strength 600 MPa is the Shimizu super high rise building which is 550 m high comprising 127 levels for reducing the column section size [1]. In America, a research programme was initiated to develop a better performance steel specifically for use in bridges to improve the weldability, toughness and weathering resistance of existing steel grade. So, it can be concluded that HSS is applied broadly in the world for its merits such as higher strength to cost ratio and artistic. However it is still short of understanding some issues of high strength steel. Firstly, the stress–strain behaviour of high strength steel is different from mild steel in respect that high strength steel generally exhibits reduced capacity for strain hardening after yielding. There is restriction in structural design that yield ratio is not allowed to have
⁎ Corresponding author. E-mail address: [email protected] (J. Jiang). 0143-974X/$ – see front matter © 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.jcsr.2011.09.007
a value greater than 0.85 in design equations to ensure a adequate ductility in the member to develop plastic failure behaviour as a defence against brittle fracture. Secondly, it is regarded that fatigue is one of the major problem causing the degradation of offshore structures in long term integrity. Thirdly, the application of high strength steels is limited by codes and standards involving application of design formulae are restricted to steels with yield strength smaller than 500 MPa in most cases. Finally, the residual stresses due to welding in high strength steel may do harm to the integrity of structures. Residual stress not only affects the initiation and onset of the propagation of surface cracks but also changes the path of a crack as it grows below the surface. It is mentioned that the effect of residual welding stresses on the performance of welded structure is particularly significant when low stresses are applied [3]. For high strength steel structures, how serious will the residual stresses impose influence on the behaviour of those structures is still obscure. Therefore, accurately understanding of the distribution of stress concentration around welded intersection and the effects of residual stress on the stress concentration factors are of great importance since it plays a key role in the prediction of fatigue life. In this paper, the residual stress distribution a set of plate-to-plate T and Y joints specimens made from high strength steel plates with yield stress equal to 690 MPa is investigated. To investigate the influence of preheating on the residual stress, two groups of specimens with different welding procedures are investigated for the residual stress with hole-drilling method: one group is composed by the joints with welding completed at ambient temperature and the other group is composed by the joints with welding completed at a preheating temperature of 100 °C. The plate-to-plate T and Y joints are chosen in this investigation
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
3. Experimental specimens
Table 1 Mechanical properties of RQT701 steel plate and LB-70L electrode. Items
Minimum Yield Tensile Strength Strength (MPa) (MPa)
RQT701 690 LB-70L 685
Minimum Average Minimum Impact Energy elongation (%) 27 J@ −45 °C 108 J@ −60 °C
790–930 755
21
18 27
in respect that this kind of joints is broad application in structural engineering and it is also basic and key building block for box sections and joints. Additionally, little residual stress data can be found for this new material. Sequentially coupled thermal-stress analysis is then conducted with finite element package ABAQUS to investigate the residual stress distribution in the joints. After that, the stress concentration factor under the action of static axial tensile loading applied on the brace plate is studied. Finally, the effects of residual stress on the stress concentration factor distributions of the joints are evaluated.
2. Historical background Extensive research into steel with tensile strength of 600 MPa are carried out since 1988 in Japan for looking for benefits in the prevention of damage caused by earthquakes. Stacy [4] listed some important issues for application of high strength steel in offshore engineering. Agerskov and Petersen [5–6] carried out the comparative investigation on fatigue performance of welded joints made of high strength steel and mild steel in constant amplitude and variable amplitude by full scale offshore tubular test and a fracture mechanics analysis. Sharp [7] summarized guidance and code topics relevant to high strength steel. Miki [8–9] reviewed the histories of the development and use of high strength steel for bridge structures in Japan. It was mentioned that the biggest problem of application for high strength steel is to achieve a balance between tensile strength and fatigue performance without losing good weldability. Additionally, fatigue performance of typical welded joints including longitudinal, cruciform and out-of-plane gusset welded joints were discussed with respect to the dependence on steel strength and joint size. HSE Research Report 105[10] gave a detailed review of the performance of high strength steel used in offshore. Bjorhovde [11] gave a brief review of mechanical properties of high strength steels, including the properties of tensile strength and stiffness, ductility, toughness, weldability and corrosion resistance. The effect of residual stress on fatigue behavior of high strength steel components has also spurred great interesting recently. Sonsino [12– 13] evaluated effect of residual stresses on the fatigue behavior of high strength steel weld joints under different kinds of loading. Stacey reviewed residual stresses distributions in several kinds of welded joints [14]. Transverse and longitudinal through-thickness residual stress distributions in range of geometries of welded CHS joints are summarized. Experimental work to determine residual stress distribution at the chord weld toe of three T circular section joints was carried out by Payne and Porter-Goff [15]. Three tubular joints with yield stress ranging from 365 to 491 MPa were investigated. Block sectioning was applied to determine through-thickness distribution of residual stresses in the weld toe region at crown and saddle positions.
In the present experimental investigation, a number of plate-toplate T and Y joints, made of high strength steel with minimum yielding stress of 690 MPa, were fabricated by welding. This high strength steel, RQT701, which was supplied by Corus Group, is quenched and tempered structural steel with improved forming and welding performance by substituting some alloying element with carbon. In the process of welding, an ultra low hydrogen and moisture resistant type covered electrode for 690 MPa high tensile strength steel for low temperature service, LB-70L, which is equivalent to the class ASME/AWS A5.5 E10016-G and supplied by Kobelco of Japan, was employed [16]. The diffusible hydrogen content of this electrode is 4 ml/100 mg. The welding procedure was carried out according to AWS D1.1 2008 [17]. Table 1 gives the mechanical properties of the RQT701 plate and LB-70L electrode. Table 2 lists the welding parameters for fabrication. Two series of specimens were included to compare the influence of preheating on residual stress distribution near the weld toe. The first group consisted of joints with welding completed at ambient temperature while the second group consisted of joints with welding completed at a preheating temperature of 100 °C. For each group, there are 12 different geometries, consisting of 3 different plate thicknesses and 4 joint angles. They were employed to explore the variation of the residual stress near the weld toe when geometry is different. For the 135° joints, the welding profile is shown in Fig. 1. Fig. 2 is applied for the welding profile of 60°, 75° and 90° joints (Table 3). 4. Experimental investigation for residual stress 4.1. Theorical basis The localized stress is relaxed by introduction a hole into a residually stressed structure and correspondingly the local strain on the surface of the testing structure is changed. It is the principle of the hole-drilling method of measuring of residual stress firstly proposed by Mather [18]. The initial stress state at point P(R, α) in a thin plate subjected to a uniform residual stress, as shown in Fig. 3, can be written as follow in polar coordinates: 8 0 σx > > > σr ¼ 2 ð1 þ cos2αÞ > < 0 σ σθ ¼ x ð1− cos2αÞ > 2 > 0 > > : τrθ ¼ − σx ð sin2αÞ 2
ð1Þ
When a hole is drilled through in the center of a plate, as shown in Fig. 4, the stresses in the vicinity of the hole was obtained by Kirsch and can be expressed as [19]: 8 00 σx 1 σ 3 4 > > 1− 2 þ x 1 þ 4 − 2 cos2α > σr ¼ > 2 > r 2 r r > < 00 σ 1 σ 3 σθ ¼ x 1 þ 2 − x 1 þ 4 cos2α > 2 2 r r > > > 00 σ 3 2 > > : τrθ ¼ − x 1− 4 þ 2 cos2α 2 r r
ð2Þ
where: r is the ratio of R to R0; R and R0 denote hole radius and arbitrary radius from hole center respectively. By comparing the stress in
Table 2 Welding parameters for the specimens. Welding position
Base Metal
Current
Flat
RQT701
DCEP
Heat Treatment Preheat
Non-preheat
100 °C
30 °C
Interpass Temp.
Welding Condition
150 °C
170 A
26 V
2.2 kJ/mm
22
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28 Table 3 Specification of the specimens.
t2
F
y
θ (°)
β (°)
L1 (mm)
L2 (mm)
W (mm)
t1 (mm)
t2 (mm)
R (mm)
L (mm)
135
N.A.
440
440
420
60
30
440
440
420
75
60
440
440
420
90
60
440
440
420
8 12 16 8 12 16 8 12 16 8 12 16
8 12 16 8 12 16 8 12 16 8 12 16
5 5 5 6 6 6 6 6 6 6 6 6
16 22 28 22 31 41 21 30 39 20 27 35
l2
E
θ
G J I
H
D
tw
C
R
B
t1 x
l
A
l1 Fig. 1. Welding Profile for 135° joints.
this point before and after hole-drilling, stress relaxation can be determined due to drilling the hole. On the assumption that the material of the plate is homogenous and isotropic and its stress–strain curve is linear, the relieved strain in the point P(R, α) can be obtained by substituting the stress relaxation into biaxial Hooke's law: 8 i ε1 þ ε3 1 h 2 2 0:5 > > > σ max ¼ 4A − 4B ðε3 −ε1 Þ þ ðε3 þ ε1 −2ε2 Þ > > < i ε þ ε3 1 h 2 2 0:5 þ σ min ¼ 1 ðε3 −ε1 Þ þ ðε3 þ ε1 −2ε2 Þ > 4B 4A > > > ε þ ε3 −2ε2 > : tan2α ¼ 1 ε3 −ε1 where; A ¼ −
ð3Þ
1þυ 1 1þυ 4 1 3 − ; B ¼ − ; 2E 2E 1 þ r r2 r4 r2
pads as shown in Fig. 5. The milling cutter should be guided carefully so as to make the cutter progress in a straight line devoid of side pressure on the hole and friction at the non-cutting edge. A high-speed air turbine was employed to form good hole shape and adaptability to incremental drilling as shown in Fig. 6. To make the measured points close to the weld toe of the joint, a special supporting set was designed as shown in Figs. 5 and 6. In the process of test, the ASTM 837-08 was followed [20]. Since there is physical limitation for the test rig that it cannot be very close to the weld toe, the 75° and 90° joints are cut for their brace plates.
4.3. Strain gauge scheme
4.2. Test setup The RS-200 Milling Guide, a high-precision instrument for analyzing residual stress by the hole-drilling method through positioning and drilling of a hole in the center of a special strain gauge rosette, was applied for measuring the residual stress in the specimens. Its ruggedness and flexibility make it equally suitable for laboratory or field application. Since positioning precision of the milling guide has great influence on the accuracy of measurement, the RS-200 milling guide with a microscope installed was applied herein and it is secured to samples with quick-setting and frangible adhesive to bond its foot
A special type of strain rosette, FRAS-2, which is designed by TML to facilitate positioning three grids on one side of the measurement point, was used to measure the released strain of the specimen during drilling. For the 135° and 60° joints, in the longitudinal direction (the x axis in Fig. 7), three strain gauges, denoted by A, B and C, were positioned at points where the x coordinate are respectively equal to 25 mm, 75 mm and 125 mm with the 5 mm away from the weld toe. In the transverse direction (the y axis in Fig. 7), in the middle of plate width, where the x coordinate is 75 mm, another two strain gauges denoted as B1 and B2 were positioned at the points where the y coordinate are respectively equal to 20 mm and 35 mm. For the 75° and 90° joints the strain gauge A and C were removed. In the middle of plate width, another strain gauge B3 was added at the 50 mm point. Table 4 lists residual stress testing results.
F E
y
Y
l2
σ max σ '' r
R0
θ
σ '' θ R
R D C
β J I
σ min
θ
G H
l
x
X
B
t1 A
σy
σx
l1 Fig. 2. Welding Profile for 120°, 105° and 90° joints.
Fig. 3. Stress state at point P(R, α) before drilling a hole.
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
23
5. Numerical investigation for residual stress 5.1. Modelling procedure In this paper, the commercial finite element modelling package ABAQUS was used in process of welding simulation. A sequentially coupled thermal-stress analysis was conducted by assuming that the stress/displacement solution is dependent on a temperature field but there is no inverse dependency. Sequentially coupled thermal-stress analysis was performed by first solving the pure heat transfer problem, then reading the temperature solution into a stress analysis as a predefined field. Firstly, a non-linear transient thermal analysis was conducted to determinate the temperature distribution during the welding. Secondly, the temperature data was used as a thermal loading input for the mechanical analysis. Before running the modelling, several assumptions for the modelling were put forward. First, the stress due to the welding has negligible influence on the temperature field. Second, the latent heat, the amount of energy released or absorbed by a substance during a change of state that occurs without changing its temperature, is ignored in the modelling. In the paper, only the 135° joint with 16 mm base plate is chosen for numerical modelling to investigate the residual stress distribution. 5.2. Thermal analysis In the thermal analysis, transient non-linear analysis was conducted for 2D plate model to determine the temperature history throughout the plate. Element ‘birth and death’ technique was adopted to simulate the weld filler variation with time. All the elements, including the base plate and weld filler (regions IJGH in Figs. 1 and 2), were created firstly. However, the elements in the weld filler were deactivated before analysis. The weld filler was divided into four lumps. The first lump was activated at 0.001 s. The welding speed was 2.6 mm/s so that the time to finish the first run of 150 mm was 57.7 s. Then, the second lump was activated within 0.001 s and the heat transfer time was again set as 57.7 s. Such procedure was then repeated for the third and forth lumps. Thermal conductivity, k, which is the property of a material reflecting its ability to conduct heat, is obtained from the standard EC3 (1–2) [7]. The melt droplets were modelled by a volumetric heat source with uniform density. The move of heat source was achieved by defining the amplitude of heat source density curve. The heat source density was defined by the equation:
q¼
0:8UI le hl v
High speed turbine assembly
Supporting set
Fig. 5. High speed drilling setup for the RS-200 milling guide.
In Eq. (4), 0.8 is the arc efficiency factor, I is the arc current, U is the arc voltage, le is the depth of each lump and hl is the height of each lump. To consider the heat losses from the surface of welded plate, both convection and radiation were assumed, as shown in Eq. (5). The convection coefficient h was assumed as 15 W/ (m 2K) and the emissivity ε is defined as 0.2. 5.3. Mechanical analysis Transient non-linear stress analysis was carried out for recognizing the distribution of residual stress. In the mechanical analysis, the temperature history obtained from the thermal analysis was input as a thermal loading into the structural model. Temperature dependent mechanical properties, including the thermal expansion coefficients, Young modulus, Poisson's ratio and yield strength were calculated by applying the reduction factors listed in the EC3(1–2) [7]. The mesh and analysis steps were as the same used in the thermal analysis. 5.4. Modelling results Since only the part near the weldment in base plate is interested, the results for temperature variation and residual stress distribution in this part, in the range of 50 mm from the weld filler contour, are shown.
ð4Þ
Brace Plate
Microscope 15 15 15 5 25
C
Added in 75 and 90 specimens
50
weld toe
B3 B2 B1 B 50
A
Supporting set
Fig. 4. Alignment setup for the RS-200 milling guide.
25
y
x Chord Plate
Fig. 6. Strain gauges scheme for residual stresses measurement.
24
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
Table 4 SCF testing results. Joint geometry
Ambient temperature
Preheating
Degree
Thickness(mm)
A
B
C
A
B
C
135
8 12 16 8 12 16 8 12 16 8 12 16
1.43 1.61 1.70 1.61 1.00 1.63 1.46 1.68 1.52 1.62 1.53 1.78
1.53 1.82 1.95 1.41 1.15 1.68 1.57 1.74 1.55 1.83 1.84 2.15
1.48 1.74 1.72 1.54 2.44 1.45 1.51 1.52 1.43 1.76 1.68 1.65
1.54 1.48 1.76 1.60 1.67 2.10 1.41 1.57 1.46 1.54 1.72 1.62
1.62 1.53 1.96 3.10 1.73 3.25 1.69 1.83 1.59 1.94 1.88 1.88
1.60 1.30 1.88 1.92 1.58 4.84 1.54 1.62 1.52 1.62 1.76 1.80
60
75
90
Fig. 8. The temperature distribution at t = 58.7 s (2nd lump was added).
6. Test for stress concentration factor Fig. 8 illustrates the temperatures at 1.0 s when the 1st lump is activated. At this moment, the heat is localized in the weld filler and the temperature in base steel does not raise much. Also, the 2nd, 3 rd and 4th lumps are still deactivated at this moment, the elements in those areas can be seen as ‘blank’ and they do not have function of heat conduction. When the 2nd lumped of the weld filler is activated, as shown in Fig. 9, the maximum temperature is shifted to this new lump. However, it can be seen that the maximum temperature in the joint do not change much. It just enlarges the area that heat affected. It is similar for the 3 rd lump and 4th lump, as shown in Figs. 10 and 11. After that, the cooling procedure comes in for the joint. Figs. 12 and 13 give the temperature distribution at 800 s and 2500 s. When the time went to 2500 s, the temperatures in the joint reduce to near 30 °C. Figs. 14 and 15 show the transverse stress when the joint completely cooled down. Fig. 14 illustrates the stress distribution when the 100° preheating is carried out. It can be found that the transverse stress shows a layered distribution in the chord plate. The stress in the weld is particularly high when compared with the stress in the base plates. High transverse residual stress is localized at weld toe and weld root position. Another place with high transverse stress need to be noticed is the lower surface of the chord plate rightly below the weld. From Figs. 14 and 15, it can be found that the residual stress in the plate surface only localized in the vicinity of the weld toe and the weld root. When the distance from the weld toe is 15 mm, the magnitude of the residual stress has dropped down quickly. Figs. 16 and 17 give the comparison for the modelling results with the corresponding values from the testing. The modelling results seem a litter higher than the values obtained from the test at the point of 5 mm from the weld toe. In the 2D modeling, the plane strain elements are used so that the structural deformation along the plate width is ignored.
Fig. 7. The temperature distribution at t = 1.0 s (1st lump was added).
6.1. Test setup The Instron Model 8506 Dynamic Materials Testing System is introduced when tensile loading is to be applied in the specimen. With maximum tensile loading capability of 2000 kN, it is an advanced multiprocessor-based control console which provides full digital control of a testing system. It consists of a closed load and four columns frame with movable crosshead, a hydraulic actuator to apply a force, gripping mechanisms to hold the mechanical test specimen, and a load cell to measure the force. The position of the actuator, under closed loop control by controlling the hydraulic fluid flowing through a servo-valve supplying the actuator, is measured by a displacement transducer. 6.2. Specimen assembly To fix the specimen in the grip of testing machine, a set of supporting joints, made of mild steel S355 with thickness of 45 mm, is designed. It is designed in such a way that have nearly triple thickness when comparing with specimens to make sure the failure will turn out in specimens rather than the supporting joints. The specimen and supporting joint is connected by 12 high strength hexagon bolts of grade 10.9HR. In each end of connection, 6 bolts are positioned in two lines. Fig. 18 shows the profile of the specimen and supporting joint after assembly. Fig. 19 is the full view of the testing machine when the specimen and the supporting set are fixed in the grips. 6.3. Strain gauge scheme FLA-2 of TML strain gauges, which has only one grid, were adhered on the surface of the specimen. In the direction of width of specimen, three strain gauges in a line were used. In the direction of length of specimen, two lines strain gauges, which were away from weld toe 5 mm and 15 mm respectively, were applied, shown in Fig. 20.
Fig. 9. The temperature distribution at t = 116.4 s (3rd lump was added).
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
Fig. 12. The temperature distribution at t = 2500 s.
Fig. 10. The temperature distribution at t = 230.1 s (4th lump was added).
When stress concentration factor and hot spot stress in weld toe of the samples were focused, the stress distribution along direction of weld toe was investigated by monitoring the strains near the weld toe shown subjected to tensile loading in brace plate. Table 5 is the SCF testing results for all the joints.
In the experimental investigation, a carefully planned testing was organized to analyze the residual stresses and SCF values in various high strength steel plate-to-plate joints. Linear interpolation method is employed to estimate the residual stress at the weld toe. In addition, linear interpolation method is used for hot spot stress at the weld toe and SCF values without considering the residual stress are illustrated. By comparing the SCF values without residual stress effect and another group SCF values incorporating residual stress, the influence of residual stress on SCF is analyzed. At any point in the specimen, the total stress which is assumed to be smaller than the yielding stress σy can be expressed as: ð5Þ
In Eq. (5), σrs is the residual stress at a certain point in the specimen; σf is the elastic stress caused by the applied load in the specimen. According to the definition of SCF, SCFf can be written as: SCFf ¼
σf σn
ð6Þ
In Eq. (6), SCFf is the stress concentration factor without considering residual stress; σn is the nominal stress caused by the applied load in the specimen. When residual stress is considered, the equation for the SCF should be expressed as: SCFtotal ¼
σtotal σ rs þ σf σ rs þ SCFf ⋅σn σ rs ¼ ¼ ¼ þ SCFf σn σn σn σn
Fig. 11. The temperature distribution at t = 800 s.
In Eq. (7), SCFtotal is the stress concentration factor including residual stress and stress caused by the applied load. By defining SCFrs ¼ σσrsn , the equation can be written as: SCFtotal ¼ SCFrs þ SCFf
ð8Þ
Therefore, the effect of residual stress can be measured by the residual stress factor (RSF), which is defined as:
7. Analysis and discussion
σtotal ¼ σrs þ σf
25
ð7Þ
RSF ¼
SCFrs SCFtotal
ð9Þ
The specimen nomenclature for Figs. 21, 22, 23 and 24 is that the first term in the legend denotes the intersection angle (135°, 60°, 75° and 90°) of the plate, the second term denotes the plate thickness and the final team identifies whether preheating was applied. The character ‘P’ herein denotes joints with preheating and ‘A’ indicates the specimens welded in ambient temperature. From Eqs. (8) and (9), it can be seen that when tensile residual stress exists SCFrs will be smaller than SCFtotal and hence RSF shall within the range [0, 1]. In addition, in general, with the increase of the applied stress, the RSFs will be reduced gradually from 1. This fact also implies that whenever the RSF value is outside the range of [0, 1], compressive residual stress exists. Fig. 21 illustrates the RSF values variations of 135° joints in different applied stress. It can be seen that when that the applied stress is equal to 50Mpa, at the 5 mm point, the RSFs for the 8 mm, 12 mm and 16 mm joints without preheating are equal to 0.65, 0.51 and 0.69, respectively. The correspondingly values for the preheating joints with 8 mm, 12 mm and 16 mm plate thickness are equal to 0.61, 0.41 and 0.58, respectively. Hence, it can be seen that at 5 mm point, the RSF for the preheating joints was slight smaller than values without preheating. When the applied stress increases to 250Mpa, RSFs for the 8 mm, 12 mm and 16 mm joints without preheating are equal to 0.27, 0.17 and 0.30, respectively. The correspondingly RSFs for the preheating joints with 8 mm, 12 mm and 16 mm plate thickness are equal to 0.24, 0.12 and 0.20, respectively. Therefore, when
Fig. 13. The transverse residual stress for the 135°joint welded in ambient temperature.
26
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
t
420 Specimen θ Fig. 14. The transverse residual stress for the 135°joint welded with 100 °C preheating.
Tranverse residual stress (MPa)
200
t 45
Testing Modelling
150
Supporting Joint 360
100
50
0
45 -50
0
10
20
30
40
50
60
Fig. 17. Assembly of the specimen and supporting joint.
Distance from the weld toe (mm) Fig. 15. Comparison of transverse residual stress between testing and modelling for 135° joint with 16 mm base plate welded at ambient temperature.
shown in Fig. 22. It is concluded that the influence of residual stress is more serious when a low applied stress exists. Fig. 23 shows the relationship between RSFs and the applied stress for 105° joints. For the 75°–8 mm-P joint and the 75°–8 mm-A joint,
the applied stress increases from 50 MPa to 250 MPa, the RSFs for the 8 mm, 12 mm and 16 mm joints without preheating were reduced by 58.5%, 66.7% and 56.5%, respectively. For the 8 mm, 12 mm and 16 mm joints with preheating, the correponding RSFs reduction are equal to 60.6%, 70.7% and 65.5%, respectively. Thus, even though the residual stresses in joints without preheating are higher than the joint with preheating, the RSFs reduction for the joints without preheating is faster than the preheated joint. When the applied stress increases from 250 MPa to 400 MPa, RSFs reduce 30% approximately for all the specimens. Similar conclusion can be drawn from 60° joint, as
Tranverse residual stress (MPa)
200 Testing Modelling
150
100
50
0
-50
0
10
20
30
40
50
60
Distance from the weld toe (mm) Fig. 16. Comparison of transverse residual stress between testing and modelling for 135° joint with 16 mm base plate welded at 100 °C preheating.
Fig. 18. Full view of testing machine.
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
27
RSF
0.8
15 15 15 5
0.7
8mm-P
8mm-A
0.6
12mm-P
12mm-A
0.5
16mm-P
16mm-A
0.4 0.3 0.2
30
0.1 45
0.0 weld toe
0
50
100
150
200
250
300
350
400
Nominal stress (MPa)
45
Fig. 20. RSF of 135° joints under different nominal stresses.
30
the 90° joints. Similar to the case shown in Fig. 21, negative RSFs were resulted when the applied stress is 50 MPa which indicated the existence of compressive residual stress.
Fig. 19. Strain gauge scheme for SCF measurement.
8. Conclusions when the applied stress is 50 MPa, the residual stresses in the weld toe are compressive so that RSFs are negative and outside the range [0, 1]. For the 75°–12 mm-P joint, the RSF at the 5 mm point was equal to 5.78. It is greatly outside the range [0, 1] as that the compressive residual stress at the 5 mm point is − 105.2 MPa which is larger than the applied stress. When the applied stress was increased to100MPa, the RSF values was −1.53 because SCFtotal becomes positive. With further increase of the applied stress, the RSF will approach into 0. Note that for a given applied stress, there exists a critical compressive residual stress which will lead to zero total stress (or SCFtotal = 0) and therefere the RSF will approach infinity. Furthermore, when the applied stress is slightly smaller than this critical value, the RSF approaches positive infinity quickly while when the applied stress is larger than this critical value, the RSF reduces quickly to minus infinity value. For the 105°–12 mm-P joint, this critical value is roughly equal to 61.8 MPa. It alos can be seen that when the applied stress is larger than 250 MPa, the RSF is comparatively small when comparing with low stress is applied. Fig. 24 gives an illustration for
Based on the analysis in the paper, some conclusions are drawn: (1) Frequently, the maximum tensile residual stress perpendicular to weld toe appears in the middle of plate. In the transverse direction, the magnitude of residual stress reduces greatly from the point B to point B2 for 135° joints; however, it may increase in a smaller magnitude from B1 to B2. It means that the change of residual stress with the distance away from weld toe is non-linear. When the distance from the weld toe is 50 mm, the residual stress fluctuates around 0. (2) Preheating can effectively reduce the magnitude of residual stress of near the weld toe. Therefore, it is significant that evenly high-quality preheating of the steel plate should be applied during the welding of high strength steel joints. (3) When the applied stress ranges from 0 to 200 MPa, the influence of residual stress on SCF is obvious. When the applied stress is beyond 200 MPa, this influence is much relieved. The
Table 5 Residual stress testing results. Specimen cases θ (°) 135
Stress in the measuring points (MPa) t (mm) 8 12 16
60
8 12 16
75
8 12 16
90
8 12 16
Preheating treatment
Brace–cutting treatment
A
B
Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No Yes No
No No No No No No No No No No No No Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes
− 78.8 23.2 − 82.4 − 15.6 47.7 12 − 29.5 76.6 88.1 − 75.3 − 11.4 − 18.5 – – – – – – – – – – – –
119.1 151 62.8 102.6 132.8 213.6 118.5 240.4 108.7 239.7 87.1 278.4 − 32 − 46.5 –105.2 35.2 27.3 139.2 − 66.2 − 70.3 − 32.5 − 8.8 51.1 24.9
B1
B2
99.2 96 48.5 79.7 114.9 103.9 56 63.3 22.5 127.6 90.5 − 17.6 137 36.7 45.6 96.9 45.6 243.6 76.4 129.0 34.6 67.5 91.9 52.4
118 90.9 36.9 32.3 53.4 68.1 65.3 93.7 43.5 60 79.8 62.2 36 22.3 12.4 24.4 − 7.7 58.1 25.2 28.2 11.2 26.9 25.2 38.2
B3
C
– – – – – – – – – – – – − 10.5 − 12.6 − 13.5 − 22.7 15.6 19.1 8.0 28.5 6.2 5.3 5.2 –26.4
42.6 –25.6 –68.2 –42.7 16 –68.5 –27.8 –29.4 –47.6 81.8 –76.4 –45.8 – – – – – – – – – – – –
28
J. Jiang, M. Zhao / Journal of Constructional Steel Research 72 (2012) 20–28
0.9
References
0.8
8mm-p
8mm-A
0.7
12mm -P
12mm -A
0.6
16mm -P
16mm -A
RSF
0.5 0.4 0.3 0.2 0.1 0.0
0
50
100
150
200
250
300
350
400
Nominal stress (MPa) Fig. 21. RSF of 60° joints under different nominal stresses.
20 15
8mm-P
8mm-A
10
12mm-P
12mm-A
16mm-P
16mm-A
RSF
5 0 -5 -10 -15 -20
0
50
100
150
200
250
300
350
400
Nominal stress (MPa) Fig. 22. RSF of 75° joints under different nominal stresses.
influence of residual stress on stress concentration effect is evident when low stress is applied. (4) RSF is an effective parameter that can indicate the effect caused by the residual stress. When the tensile residual stress exists, the RSF should be in the range (0, 1). When the compressive residual stress exists, the RSF will jump to a high value (which is plus in algebra) firstly and with the increase of the applied stress, the parameter will drop down to a minus value. When the applied stress increases continually, the RSF gradually approaches to 0. (5) Residual stress does not always harmful to the integrity of structures. When the compressive residual stress exists, residual stress could be is beneficial as it could relieve the stress concentration effect caused by the applied load.
1 0
RSF
-1 -2
8mm-P
8mm-A
12mm-P
12mm-A
16mm-P
16mm-A
-3 -4
0
50
100
150
200
250
300
350
Nominal stress (MPa) Fig. 23. RSF of 90° joints under different nominal stresses.
400
[1] Pocock Graham. High strength steel use in Australia, Japan and the US. The Structural Engineer Nov 2006:27–30. [2] Salama MM, Tetlow JH. Selection and Evaluation of High Strength Steel for Hutton TLP Tension Leg Elements. Offshore Technology Conference, Houston, Texas; May 1983. [3] Clarin Mattias. High strength steel local buckling and residual stresses. Lulea, Lulea University of Technology: Department of Civil and Environmental Engineering; 2004. Licentiate. [4] Stacey A, Sharp JV, King RN. High strength steels used in offshore installations. International Conference of OMAE; 1996. p. 417–33. Volume III. [5] Agerskov, H., Petersen R.I., et al. An investigation on fatigue in high-strength steel offshore structures. Welding in the World 1997;41(4):328–42. [6] Petersen, R.I., Agerskov, H., and Lopez, M.L. Fatigue life of high strength steel offshore tubular joints, Report of Technical University of Denmark, Lyngby, Denmark. ISBN: 87-7740-167-0. [7] Sharp JV, Billingham J, Stacey A. Performance of high strength steel used in jackups. Marine Structures 1999;12:349–70. [8] Miki C, Homma K, Tominaga T. High strength and high performance steels and their use in bridge structures. Journal of Constructional Steel Research 2002;58:3–20. [9] Anami K, Miki C. Fatigue strength of welded joints made of high-strength steels. Progress in Structural Engineering and Materials 2001;3:86–94. [10] Billingham J, Sharp JV, et al. Review of the performance of high strength steels used offshore. Bedfordshire, UK; 2003. [11] Bjorhovde R. Development and use of high performance steel. Journal of Constructional Steel Research 2004;60:393–400. [12] Sonsino C.M., Lagoda T., et al. Damage accumulation under variable amplitude loading of welded medium and high strength steels. International Journal of Fatigue 2004;25(5):487–95. [13] Sonsino C.M. Effect of residual stresses on the fatigue behaviour of welded joints depending on loading conditions and weld geometry. International Journal of Fatigue 31(1): 88–101. [14] Stacey, A., Barthelemy,J.Y., et al. Incorporation of residual stresses into the SINTAP defect assessment procedure. Engineering Fracture Mechanics 2000;67(6):573–611. [15] Payne, J.G., Porter-Goff, R.F.D. Experimental residual stress distributions in welded tubular T-nodes. I Mech E Conference Publications Institution of Mechanical Engineers 1986:109–116. [16] AWS. ANSI/AWS A5.5. Specification for Low-Alloy Steel Electrodes for Shield Mtetal Arc Welding. Miami, USA: American Welding Society; 2006. [17] AWS. ANSI/AWS D1.1. Structural Welding Code-Steel. Miami, USA: American Welding Society; 2008. [18] Mather J. Determination of Initial Stresses by Measuring the Deformation Around Drilled Holed. Transactions of the ASME 1934;56(4):249–54. [19] Kabiri M. Measurement of Residual Stress by the Hole-Drilling Method:Influence of Transverse Sensitivity of the Gages and Relieved Strain Coefficients. Experimental Mechanics 1984;25:252–6. [20] ASTM. E837-08. Standard Test Method for Determining Residual Stresses by HoleDrilling Strain-Gage Method. West Conshohocken, PA 19428–2959, United States: ASTM International; 2008.
Nomenclature A,B: Calibration constant for hole-drilling method E: Young's modulus h: Convection coefficient hl: The height of each lump of weld k: Thermal conductivity I: The arc current l: Weld filler length in the chord plate surface le: Depth of each lump l1: The length of the chord plate l2: The length of the brace plate q: Heat source density R: Gauge circle diameter or weld root length R0: Diameter of the drilled hole t1: The thickness of the chord plate t2: The thickness of the brace plate U: The arc voltage W: Plate width β: End preparation angle for the specimens ε: Relieved strain due to the hole-drilling εe: The emissivity coefficient θ: Joint angle or the angle between the uniform normal stress in r direction and x ν: Poisson's ratio σc: Calibration stress σr: Uniform normal stress in r direction σθ: Uniform normal stress in θ direction τrθ: Uniform shear stress σmax: Maximum principle stress σmin: Minimum principle stress AWS: American Welding Society HSS: High strength steel RSF: Residual stress factor SCF: Stress concentration factor