Experimental Research of Welding Residual Stress of Butt Welded Joint of Thick Steel Plate

Abstract

The thickness of steel plates used in the structure has gradually increased to meet the load-bearing capacity requirements of long-span steel bridges. Thick steel plate welded by arc welding process will result in considerable welding residual stress with complex distribution. Large welding residual stress will significantly impact the performance of steel bridges. At present; residual welding stress of thick steel plates is not considered enough in bridge engineering; which could lead to potential dangers in safety. In this paper; a butt welded joint with a thickness of 80 mm was designed; the residual welding stress was measured by X-ray diffraction method; and the distribution of residual stress in the direction perpendicular to the weld seam and along the direction of the weld seam is analyzed; and the distribution pattern of welding residual stress in thick steel plate is systematically studied. It is found that in the area near the weld; the stress in the direction along the weld seam is more significant than that in the perpendicular direction; the peak stress in the direction perpendicular to the weld seam usually appears in the weld seam and the heat-affected zone; and the maximum value tends to appear close to the last weld bead on the surface; on the path perpendicular to the weld seam; the stress in the direction perpendicular to the weld seam is distributed in a “Π” shape; the stress in the direction along the weld seam is distributed in an “M” shape; and the stresses in the direction along and perpendicular to the weld seam are symmetrically distributed; however the axis of symmetry may appear anywhere in and around the weld seam. The results could serve as a reference for welding residual stress analysis and guide optimization design of steel bridges made of thick steel plates.

1. Introduction

With the continuous development of transportation and the increasing demand for traffic construction, the application of long-span steel bridge is gradually increasing. The increase in the span of the steel bridge inevitably leads to a rise in the internal force and the thickness of the steel plate. In recent years, the maximum plate thickness of steel plate of long-span steel bridges has been more than 40 mm, among which the maximum plate thickness of the main beam reaches 80 mm, the maximum plate thickness of the superstructure reaches 115 mm, and the maximum plate thickness of the bridge tower could reach 250 mm.

In the connection of steel structure, welding is one of the primary forms of thick steel plate joining, but welding will lead to residual welding stress, welding defects and stress concentrations in welded parts, which could result in deterioration of fatigue performance, yield strength, tensile strength, elongation at fracture, plastic deformation and other mechanical properties [1,2,3,4]. The increase in steel plate thickness increases the possibility of welding defects (including holes, inclusions, under penetration, cracks) and complicates welding residual stress distribution. The adverse effects are especially obvious in welded thick plates, and this is more evident for arc welding processes than in laser or electron beam welding [5,6].

Welding residual stress refers to the residual self-balancing stress in the weldment after welding and cooling [7,8,9]. The research on residual welding stress mainly focuses on three aspects: (1) Factors affecting the distribution of residual welding stress; (2) The distribution pattern of welding residual stress; (3) The influence of residual welding stress on the performance of steel structures.

(1)

Factors affecting the distribution of welding residual stress

Klassen [10] used the X-ray diffraction method to measure the residual welding stress and carried out finite element analysis to study the residual stress distribution of the I-beam welded with a 15 mm thick web and a 30 mm thick flange under the influence of ambient temperature. Hwang et al. [11] used X-ray diffraction and numerical simulation to study 80 mm EH40 steel V-shaped butt welded thick plates, and proposed a new heat source model for numerical simulation. Balakrishna et al. [12] used neutron diffraction to study influence of different welding processes on residual stress distributions. Ibrahim et al. [13] studied the influence of the SAW process on the residual stress of steel plates with thicknesses of 50 mm, 75 mm, and 100 mm using numerical simulation. Genchev et al. [14] studied the 20 mm thick S335J+N steel butt welded specimens and found that the history of welding heat impacts the mechanical properties of the base metal in the welded area. Wang et al. [15] studied the welding deformation of tube-sheet joints using the three-dimensional thermo-elastoplastic finite element method and compared the analysis results of fixed heat source, moving heat source, single-pass welding, and multi-pass welding. Yang [16] took the welded thick plates of the Chongqing Daning River Bridge as the research project, studied the influence of ambient temperature on the residual welding stress of 48 mm thick Q370qE steel butt welded plates. Fu [17] used numerical simulation to study the influence of ambient temperature and welding layers on residual stress and residual deformation. Chi et al. [18] studied the effect of post-weld heat treatment on the residual stress distribution of 100 mm thick 20MnNiMo steel butt welded thick plates. Wang et al. [19] conducted thermo-elastoplastic finite element analysis on 390 mm thick 20MnMoNb butt welded plates, and found that post-heat treatment can significantly ease the stress distribution and reduce the peak value of residual stress. Zou et al. [20] studied the residual stress distribution on the surface and inside of 45 mm thick Q345qD steel butt welded thick plates, and found that the latter welded area has a greater influence on the residual stress distribution than the earlier welded area. He et al. [21] analyzed the residual stress distribution of 40 mm thick plate fillet welds by elastic-plastic finite element analysis, and found that the initial loading amplitude has a significant effect on the residual stress relaxation. Liu et al. [22] detected the section of the 50 mm Q345D steel butt welded joints, and found that the local repair has a great impact on the peak residual stress. Zhou et al. [23] studied the influence of the welding process on the residual stress distribution of EH47 steel 70 mm multi-layer multi-pass welding and found that the final residual stress distribution has a good correlation with the welding through numerical simulation and experimental research.

(2)

Distribution pattern of residual welding stress

Chang et al. [24] studied the welding residual stress distribution characteristics of butt welded joints of 30 mm thick plate of bridge steel POSTEN60 and POSTEN80 and the effect of high strength steel welding cooling process on the residual stress relaxation of weld fusion zone (FZ) and heat affected zone (HAZ). Bang et al. [25] analyzed the residual stress field of 13 mm thick STS304 L steel under different welding methods, and found that the residual stress caused by laser welding was 13%~15% smaller than that caused by SAW. Acevedo et al. [26] used neutron diffraction to measure the residual stress distribution of K-shaped welded joints made of 30 mm thick S355J2H steel, and found that the transverse residual stress of K-shaped welded joint is greater than the longitudinal residual stress and is close to the yield strength of the material. Bank et al. [27] used the blind hole method to measure the residual stress, and studied the residual stress distribution of butt-welded austenitic stainless steel plates with a thickness of 20 mm. Peng et al. [28] studied the distribution of residual stress in the welding of 44 mm thick box-shaped compression bars of the Wuhu Yangtze River Bridge. Huang et al. [29] carried out experimental research on the magnitude and distribution of welding residual stress at the joints of the Dongjiang Bridge, found that the peak value of residual welding stress can reach 0.9 times the yield strength of the material. Liu et al. [30] carried out a numerical simulation on the butt welded and fillet welded joints of railway steel bridges, and analyzed the residual welding stress of the overall joint with a plate thickness of 28–40 mm. Qiang [31] studied the spatial distribution of residual stress in the welding of 16–100 mm thick Q345qD steel plates, and used numerical simulation to study the effect of shot peening on the relaxation of residual stress. Deng and Murakawa [32] used three-dimensional, thermo-elastic–plastic, large deformation finite element method to simulate welding distortion in a low carbon steel butt-welded joint with 1 mm thickness, and found that the inherent strain method can effectively predict the welding deformation in the thin plate butt-welded joint with 1 mm thickness. Dai et al. [33] numerically investigated the residual stress distributions in a medium thickness SUS304 steel pipe butt-welded joint before and after repair welding, found that that the peak value of axial residual stress after repair welding was significantly increased, while the variation of the maximum value of hoop residual stress was limited.

(3)

Influence of welding residual stress on the performance of steel structures

Ahmad [34] studied the effect of residual stress on the fatigue performance of 12 mm thick dissimilar steel butt welded plates. Gyubae et al. [35] used 50 mm and 80 mm welded thick plates of EH36 steel to study the influence of different welding processes on welding residual stress and analyzed the effect of residual welding stress on crack propagation and fracture toughness of thick plates. Li et al. [36] studied the mechanical properties of the 56 mm thick 15MnVNq steel plate of Jiujiang Yangtze River Bridge, and expounded the mechanism of preheating to prevent hydrogen-induced cracks. Gu et al. [37] studied the longitudinal residual stress distribution of steel column made of 40 mm thick welded plate steel, found the asymmetric distribution of the residual stress along the symmetry axis of the box section leads to different effects of the residual stress on the ultimate bearing capacity around the two axes. Peng et al. [28] studied the distribution of residual stress in the welding of 44 mm thick box-shaped compression bars of the Wuhu Yangtze River Bridge, found that residual stress has a significant impact on the ultimate bearing capacity of components, and will also reduce the stability coefficient of compression bars. Yang et al. [38] studied the distribution of longitudinal and transverse residual stress in thick plates with the thicknesses of 25 mm, 40 mm, 60 mm, and 80 mm, and the effect of residual stress on distribution of damage under low cycle reciprocating load was studied.

To sum up, the research on the influencing factors of residual welding stress mainly focuses on steel plates thinner than 50 mm. In terms of residual stress distribution, due to the constraints of the large thickness of the thick steel plate and many factors affecting the residual welding stress, the distribution of residual stress in the plate is complicated, making the relevant experimental research difficult.

Aiming at the deficiencies in the current research on thick plates of steel bridges, a welding residual stress test for an 80 mm thick plate is conducted in this paper. Distribution patterns of longitudinal and transverse residual stress were explored by point-taking tests on several concerning paths on the top and bottom surfaces of the welded thick plate. The results could serve as a reference for welding residual stress analysis and guide optimization design of steel bridges made of thick steel plates.

2. Experimental Details

2.1. Dimension and Material Properties

The dimension of the specimen is 402 mm × 300 mm × 80 mm, which is butt welded of two 200 mm × 300 mm × 80 mm base metals, as shown in Figure 1 and Figure 2.

The base metal is S420ML steel that meets the requirements of “Hot rolled products of structural steels—Part 4: Technical delivery conditions for thermomechanical rolled weldable fine grain structural steels” (BS EN 10025–4:2004). The nominal chemical composition of S420ML steel is shown in

Table 1. S420ML chemical composition.

Grade	Chemical Composition (%)
	C	Si	Mn	P	S	Nb	V	Ti	Cr	Ni	Mo	Cu	N
S420ML	≤0.18	≤0.55	≤1.8	≤0.03	≤0.025	≤0.06	≤0.14	≤0.06	≤0.35	≤0.85	≤0.13	≤0.6	≤0.027

, and the nominal mechanical properties at room temperature are shown in

Table 2. S420ML minimum yield strength at room temperature.

Grade	Nominal Thickness (mm), Minimum Yield Strength (MPa)
	≤16	>16~40	>40~63	>63~80	>80~100	>100~120
S420ML	420	400	390	380	370	365

and

Table 3. S420ML tensile yield strength at room temperature.

Grade	Nominal Thickness (mm), Tensile Strength (MPa)
	≤40	>40~63	>63~80	>80~100	>100~120
S420ML	520~680	500~660	480~640	470~630	460~620

.

2.2. Welding Process Parameters

The design of the welding followed the provisions specified in the “Code for Welding of Steel Structures” (GB/T 50661–2011), which is suitable for welding of thick steel plates. The groove form adopted a double-sided U-shaped groove, the groove angle was 30°, the edge was 3 mm, the root arc radius was 6 mm, and the root gap before welding was less than 2 mm. The joint details are shown in Figure 3.

The welding adopted CO₂ gas-shielded arc welding, the ambient temperature was 25 °C, and the humidity was 53%. The welding wire used for welding was HTW-58 (ϕ1.2), and the nominal mechanical properties and chemical composition of the welding wire are shown in

Table 4. Mechanical properties of the deposited metal.

Product	Mechanical Properties of the Deposited Metal
	Yield Strength (MPa)	Tensile Strength (MPa)	Elongation (%)	Impact Energy(J) (°C)
HTW-58	472	562	31	96(−29)

and

Table 5. Chemical composition of the deposited metal.

Product	Chemical Composition of the Deposited Metal (%)
	C	Mn	Si	P	S
HTW−58	0.088	1.21	0.6	0.012	0.01

. The mechanical properties of the deposited metal of the welding wire are slightly higher than those of the base metal, and the chemical composition is the same as that of the base metal. It conforms to the principle of equal strength and the exact composition for welding wire selection.

The welding of the specimen was done by manual welding. The polarity of the power supply was direct current electrode positive (DCEP). The butt welded joint was multi-layer multi-pass welded. Before welding, the surface inside and on both sides of the weld seam are heated with flame spray gun at temperatures within 120–150 ℃. As specified by GB/T 50661–2011, the maximum thickness of a single filler is 6 mm for CO₂ gas-shielded arc welding, therefore there are at least 14 layers along the thickness direction. After calibration with the manufacturing factory, it was decided to have 23 layers and 60 weld passes. The distribution of weld passes is shown in Figure 4 (side A is the earlier welded side). The welding direction is shown in Figure 5. Weld pass 1 is root welding, weld passes 2~39, 44~56 are filler welds, and weld passes 40~43, 67~60 are cover welds.

In the welding of thick steel plate, the welding path is divided into bottom welding, filling welding and cover welding. Bottom welding gives priority to heat input control to avoid defects such as breakdown and incomplete welding. Generally, welding current and voltage are slightly smaller. Filler welding can improve production efficiency and melt speed of welding wire as much as possible on the premise of ensuring welding quality. The welding current and voltage are slightly larger than that of bottom welding. The main purpose of cover welding is to cover the groove to ensure the quality. The determination of welding current and voltage is based on experience of the manufacturing factory. In multi-layer and multi-pass welding, increasing welding parameters are helpful to improve the welding rate when the heat input is not exceeded. The specific welding parameters used in this paper are provided by the manufacturer. The welding process parameters are shown in

Table 6. Welding process parameters.

Weld Bead	Interpass Temperature (°C)	Electric Current (A)	Voltage (V)	Power Polarity	Welding Speed (cm/min)
1	100	200~220	28	D.C.E.P	50
2~6	120	240~260	30~32	D.C.E.P	45
7	140	240~260	30~32	D.C.E.P	40
8~18	150~220	260~280	30~32	D.C.E.P	45~53
19~39	150~250	260~300	30~34	D.C.E.P	45~53
40~43	200~250	280~300	32~34	D.C.E.P	50~53
44~56	150~250	260~300	30~34	D.C.E.P	45~53
57~60	200~250	280~300	32~34	D.C.E.P	50~53

.

2.3. Ultrasonic Defect Detection of the Specimen

The original defects or welding defects inside the weld seam and the base metal, such as holes, inclusions, incomplete penetration, cracks, and others, will generate local constraint weak points and stress concentration in the weldment, which will lead to local welding stress changes. On the other hand, when the residual welding stress reaches a certain level, new cracks will be generated, or the initial defects will deteriorate, causing local residual stress redistribution.

According to the scanning method specified in “Thicker steel plates—Method for ultrasonic inspection” (GB/T 2970–2016), this paper formulated an ultrasonic testing scheme aiming at detecting the type and location of defects inside the weld seam and the base metal, which provided a reference for residual stress analysis.

The model of the ultrasonic flaw detector used in this test was CT-1000H, equipped with an 8 mm × 12 mm single crystal angled probe and a 20 mm diameter single crystal straight probe, as shown in Figure 6.

Through the comprehensive and meticulous inspection of the steel plate and the base metal, no welding defects were found inside the weld seam or any initial defects inside the base metal.

3. Measurement of Residual Stress

In this paper, the non-destructive X-ray diffraction method was used to measure the residual stress on the surface of the specimen. X-ray is an electromagnetic wave with a wavelength between 0.001 and 10 nm. It has the characteristics of extremely short wavelength, high energy, strong penetrating ability, and is invisible to the naked eye. Due to the limited penetration ability of X-ray to steel, the X-ray diffraction method can only measure the residual stress value of the surface and sub-surface of the specimen, but it has a complete set of mature detection methods and rigorous theoretical analysis. The main advantage is that the measurement process is not complicated and does not change the original stress state inside the specimen.

3.1. Basic Principles

The X-ray diffraction method assumes that under a certain stress, the spacing of specific crystal planes of the crystal grains of the material will change. This lattice strain is consistent with the macroscopic strain. The X-ray diffraction technique satisfying the Bragg equation is used to measure the material’s lattice strain. The relationship between the macroscopic strain and the stress is established according to the elastic mechanics. The macroscopic stress is obtained by fitting the measurement results with the least square method.

3.2. Definition of Stress Direction

Welding residual stress is generally divided into three categories: the residual stress along the weld direction is called longitudinal residual stress, the residual stress perpendicular to the weld direction is called transverse residual stress, and the residual stress along the thickness direction.

Since bridge weldments are mostly slender rods, the direction along the length of the rods is also called the longitudinal direction. There may be welds in multiple directions on the same weldment, thus, the classification of longitudinal and transverse residual stresses determined by welding direction can easily cause misunderstanding.

In this paper, the three-dimensional Cartesian coordinate system was used as a reference to define the stress in each direction. Taking the central position of the weld seam at the finishing end along the thickness direction as the origin, the X-axis along the length direction of the plate, the Y-axis along the width direction of the specimen (the direction of the weld seam), and the Z-axis along the thickness direction of the specimen, a right-handed coordinate system was established as shown in Figure 7. σ_x, σ_y, and σ_z were used to denote the residual stresses parallel to the X, Y and Z coordinates, respectively. That is, according to the actual direction of the residual stress, it was divided into residual stress in the X direction, residual stress in the Y direction, and residual stress in the Z direction.

3.3. Measuring Point Layout

The upper surface was denoted as surface A; the lower surface was represented as surface B. A total of 70 measuring points were arranged on both surfaces of the specimen, including 45 measuring points on surface A and 25 measuring points on surface B. The arrangement of the measuring points is shown in Figure 8 and Figure 9. Each measuring point measured the residual stress σ_x in the X direction and the residual stress σ_y in the Y direction. The measuring points are plotted on the boundaries in the figure, however, the centers of the measuring points move 3~5 mm inward in the actual measurement.

3.4. Measurement Process

The model of the X-ray stress measuring instrument used in this test is PROTO iXRD, as shown in Figure 10. The main components of the instrument include an X-ray diffraction stress analyzer, XRDWin2.0 analysis software, X-ray tube, linear array detector, goniometer, mechanical arm, and fixture.

The test was carried out with reference to the “Non-destructive Testing-Practice for residual stress measurement by X-ray” (GB/T 7704–2017), the specific process is as follows:

(1)

Surface pretreatment. Grind the surface of the specimen to make the surface smooth and flat, and prepare for point marking.

(2)

Marking of measuring points. According to the arrangement of points, mark the points to be measured on both surfaces of the specimen.

(3)

Electrolytic polishing. The points to be measured were mechanically polished to ensure that the X-ray irradiation points were bright and smooth. Then, an electropolishing instrument and NaCl solution were to measure the points to eliminate surface processing stress and improve detection accuracy.

(4)

Setup of parameters. Corresponding parameter configurations were performed on the operating system according to the type of specimen before the test. The standard material type was ferrite, the target type was Cr target, the X-ray type was CrKα ray, the diffraction crystal plane was {211}, the standard value of diffraction angle 20 was 156°, the elastic modulus E₂₁₁ was 206.612 GPa, Poisson’s ratio υ₂₁₁ was 0.247934. The peak search method was the Gaussian function fitting method.

(5)

White noise removal. In X-ray diffraction, the superposition of various incoherent scattering constitutes the background of the diffraction peak. Scattering causes the diffraction curve not to be an isolated diffraction peak, which causes large errors in test results. Therefore, the background curve was determined by adjusting the background voltage, and the original curve was subtracted from the background curve point by point to obtain a pure Bragg diffraction curve and improve the measurement accuracy.

(6)

Instrument verification. After adjusting the stress tester, the zero stress block (the average stress value should be within ±14 MPa, and the standard deviation should not be greater than 7 MPa) and the standard stress block (the average stress value should be within −492 ± 35 MPa) were measured respectively [39]. Five consecutive measurements were carried out with the strain gauge. The actual measurement results showed that the stress of the zero stress block was 13.19 MPa, and the stress of the standard stress block was −489.13 MPa, both of which aligned with the actual condition. The standard deviations of the two tests were ±4.77 MPa and ±4.75 MPa respectively, both within ±7 MPa, and the test accuracy met the requirements.

(7)

Testing. Laser aiming was used to determine the correct position of the measuring point. The actual measurement was done automatically by the X-ray strain gauge. First, the azimuth ${\psi }_0$ was anchored at 0°, and the diffraction angle 20 was scanned continuously from a high angle of 167° to a low angle of 143° with a scanning step of 0.05°. After the detector received the X-ray signal, it displayed the diffraction curve on the operating system. The diffraction curve was an approximate single-peak continuous curve, the Gaussian function fitting method was used to determine the peak position and thus to determine the direction of the diffraction line. According to the same-tilt fixed ${\psi }_0$ method, two sets of data (14 in total) were collected for each measuring point, the least square method was used to fit a straight line to each data set, and the average stress value was calculated.

4. Results and Analysis

Compressive residual stress is generally considered as desirable for mechanical properties of welded structures, while tensile residual stress is detrimental to static, dynamic and fatigue properties of the structure, therefore the concern of the investigation is more focused on tensile residual stress.

4.1. Stress along the Path Perpendicular to the Weld Seam

From the perspective of practical engineering, the stress perpendicular to the weld seam is of more concern, since the stress in this direction in truss and steel box girder is more often the direction in which the stress is generated and transmitted by other loads.

4.1.1. Stress on the Transverse Direction through the Center of the Weld Seam

The residual stress on the path L_{yA = 150} with the Y coordinate of 150 mm on surface A of the specimen and the path L_{yB = 150} with the Y coordinate of 150 mm on surface B of the specimen is the concern, as shown in Figure 11. The stress on this path is least affected by the run-on and run-off plates, which can well reflect the stress distribution on the path perpendicular to the weld seam in the welded thick plate in actual engineering. The residual stress distribution after welding is shown in Figure 12 and Figure 13.

(1)

Stress characteristics near the weld seam

As can be seen from Figure 12 that on surface A of the specimen, the peak tensile stress in the Y direction is 283.79 MPa, which is about 0.75 times the yield strength of the steel. The peak tensile stress appears at the position of 12.5 mm away from the center of the weld seam, which is located at the top of the last weld bead on surface A, as shown in Figure 14. The valley in the Y direction is compressive stress with a magnitude of −6.85 MPa that is 12.5 mm away from the right side of the weld center, which is located at the junction of the first weld bead and the second weld bead on surface A. It shows that the residual stress fluctuates significantly on the surface of the weld seam and is closely related to the specific position. In general, the stress value at the surface of the weld bead is more significant than at the fusion line.

It can be seen from

Table 7. Statistical information of peak residual tensile stress near the weld seam.

Surface	Surface A				Surface B
Direction of stress	Residual stress in the X direction		Residual stress in the Y direction		Residual stress in the X direction		Residual stress in the Y direction
Sequence of peak	1st	2nd	1st	2nd	1st	2nd	1st	2nd
Stress value (MPa)	163.73	145.41	283.79	180.04	35.47	17.61	143.69	128.25
Stress to yield strength ratio	0.43	0.38	0.75	0.47	0.09	0.05	0.38	0.34
Position(mm)	−50	80	−12.5	−50	−25	50	12.5	−50
Location feature	BM of the last WB on A	BM of the 1st WB on A	TS of the last WB on A	BM of the last WB on A	BM of the 1st WB on B	BM of the last WB on B	TS of the last WB on B	BM of the 1st WB on B

BM indicates base metal, TS indicates top surface, WB indicates weld bead, A indicates surface A, B indicates surface B.

that the peak tensile residual stress in the X direction on surface A can reach up to 163.73 MPa, which is about 0.43 times the yield strength, and it is located at a position of about 30 mm near the fusion line of the last weld bead on the surface. The peak tensile residual stress in the X direction on surface B is relatively small, only 35.47 MPa, and it is located near the fusion line between the first weld bead and the base metal on the surface. The peak tensile residual stress in the Y direction is 143.69 MPa, which is about 0.38 times the yield strength, and it is located on the top of the last weld bead on the surface. It can be seen that the peak tensile residual stress in the Y direction is larger than that in the X direction on the same welding surface, and the peak tensile residual stress in the Y direction is more likely to appear on the weld seam, while the peak tensile residual stress in the X direction is more likely to occur in the base metal area.

Combined with the analysis of the second peak, it can be seen that the peak tensile residual stress in the X direction is relatively stable, its maximum value can reach 0.38–0.43 times the yield strength, and the position often appears near the weld seam or within 50–80 mm away from the weld seam. The peak tensile residual stress in the Y direction fluctuates greatly, its value can reach 0.34~0.47 times the yield strength, the maximum value can reach 0.75 times the yield strength, and the position often appears near the last weld bead or about 50 mm away from the weld seam within the area.

It can be seen from Figure 12 and Figure 13 that the residual tensile stress in the Y direction is often greater than the stress in the X direction near the weld seam. The reason might be that the longitudinal length of the weld seam is greater than the transverse width, and the total longitudinal compressive plastic deformation is greater than the transverse compressive plastic deformation, but the transverse direction is affected by the internal tensile stress to form a compressive stress zone on the surface. On both sides of the weld seam in surfaces A and B, the stresses in X and Y directions show that the stress around the last weld bead is greater than the stress around the first bead. This is caused by the fact that the last weld bead is more constrained than the earlier weld beads.

(2)

Overall pattern of residual stress change

The larger stress fluctuations on the weld seam have some influence on the research on the pattern of residual stress change. It can be seen from Figure 15 that analyzing the pattern of residual stress change from an overall perspective, considering the drastic stress change in the Y direction near the weld seam and the sparse arrangement of measuring points, it is very likely that some high-value stresses have been missed. The stress at the weld seam in the Y direction is generally greater than the stress in the Y direction away from the weld seam, and the stress value shows a downward trend from the weld seam to the direction away from the weld seam, like a “Π” shape. At the position away from the weld seam, the stress might be tensile or compressive.

Near the weld seam, on the path away from the weld seam, the residual stress in the X direction on surfaces A and B has experienced a rise-fall process, roughly showing an “M” shaped distribution and specific characteristics of symmetry about the center of the weld seam.

The pattern of residual stress change in the X and Y directions is considered mainly caused by the combination of four factors:

(1) Regardless of the influence of the weld seam residual height and under ideal welding conditions, on the path through the center of the weld seam and is perpendicular to the direction of the weld seam, the stress in the Y direction shows an ideal “Π” shape distribution, and the stress in X direction shows a perfect “M” shape distribution on the surface of the butt welded thick plate specimen.

(2) The use and cutting of run-on and run-off plates impact residual stress. Tan et al. [40] recorded the phenomenon of stress increase in the area away from the weld seam when studying the residual welding stress of the ultra-thick steel plate with 120 mm constraints at both ends of the weld seam. This explains the phenomenon that run-on and run-off plates will affect residual stress distribution and the stress increase near the edge of the specimen in the test.

(3) The edge of the welded specimen is affected by the processing technology, the stress near the edge will rise by a certain amount [41]. The cutting of the specimen is completed by gas cutting and mechanical grinding. When the edge of the specimen is processed by gas cutting, the large thermal processing stress will cause uneven deformation or plastic deformation of the cutting surface and the surrounding metal, however, the shrinkage of the material during the cooling process is restrained, resulting in tensile stress.

(4) Measurement errors cause stress fluctuations. The main metallographic structure of S420ML steel is ferrite and a small amount of pearlite, and the grains are fine and uniform. The X-ray diffraction method is used to measure the residual stress in the base metal with high accuracy. After welding, some coarse-grained structures are produced on the surface of the weld seam, which makes the fused metal in a multi-phase state, which will have a certain impact on the accuracy of the residual stress. The X-ray diffraction method is based on the idea of statistics, and the measurement results can meet the requirements of engineering applications after peak search and white noise removal. Stress fluctuations caused by large measurement errors at the weld seam than in the base metal region should be considered in the analysis.

(3)

Symmetrical characteristics of residual stress

It can be seen from Figure 15 that on the paths of both surfaces of the specimen, from the perspective of numerical value, the stress symmetry in the X-direction is better, and the stress symmetry in the Y-direction is poorer. The wandering of the welding arc and its energy impact on the arc edge at the moment of welding causes asymmetry of residual stress about the weld seam. From the perspective of the trend of stress change, the stress change trends in the X and Y directions demonstrate a good symmetry, however, the symmetrical distribution is not completely symmetrical about the center of the weld seam. Affected by the surface welding sequence, the welding center may be located somewhere near the weld seam.

(4)

Stress difference between different welding surfaces

It can be seen from Figure 16 and Figure 17 that the residual stress in the X and Y directions of welding surface A is generally larger than that of welding surface B. The peak tensile stress of surface A is 163.73 MPa, which is 128.26 MPa higher than that of surface B (35.47 MPa); the second peak tensile stress of surface A is 156.1 MPa, which is 138.49 MPa higher than that of surface B (17.61 MPa). Notably, the residual stresses on surfaces A and B in the X direction are in opposite state, namely the residual stress in the X direction of surface A is mainly tensile, and the residual stress in the X direction of surface B is mainly compressive, the peak tensile stress (163.73 MPa) and compressive stress (−239.5 MPa) occur at the same X coordinate of respective surfaces. The residual stress in the Y direction of surface A is dominated by tensile stress, the maximum tensile stress is 283.79 MPa; the residual stress in the Y direction of surface B is in an alternating state of tension and compression, the maximum tensile stress is 143.69 MPa, the maximum compressive stress is −139.08 MPa.

There are two reasons for the above difference. First is that in the three passes near the surface, there are 12 welding passes on surface A and 10 welding passes on surface B. The energy received by surface A is larger than that of surface B, and the influence of welding heat is greater than that of surface B, thus the area of plastic deformation in and around the weld seam is larger than that of surface B. When cooling and shrinking, the tensile stress and its influence range are also larger than those of surface B. The second is the self-balancing of the residual stress along the thickness direction, the resultant force of the tensile stress area and the compressive stress area in the section along the thickness direction are in a state of balance. The tensile stress of surface A and its area is large, which leads to an increase in compressive stress and its area of surface B.

4.1.2. Stress on the Transverse Direction through the Quintile Points of the Weld Seam

The paths L_{yA = 60} and L_{yA = 240} with Y coordinates of 60 mm and 240 mm on surface A of the specimen are two transverse paths symmetrical about the center of the weld seam, as shown in Figure 18.

As concluded from Section 4.1.1, from the perspective of the trend of stress change, the stress change trends in the X and Y directions demonstrate a good symmetry, although the symmetrical distribution is not completely symmetrical about the center of the weld seam. However, for simplification while not greatly deviating from the actual situation, in the actual stress measurement, only the stress value on the right side of the transverse centerline of the weld seam was measured, and the residual stress value on the left side of the weld seam was estimated assuming the symmetrical distribution of the residual stress to study the pattern of residual stress distribution.

As can be seen in Figure 19 and Figure 20, on the path L_{yA = 60}, the peak tensile residual stress in the X direction is 37.97 MPa, and the peak tensile residual stress in the Y direction is 152.91 MPa, which is about 0.40 times the yield strength, both appearing at a position 80 mm away from the center of the weld seam. The peak tensile residual stress in the X direction on the path L_{yA = 240} is 173.91 MPa, which is about 0.46 times the yield strength, and it appears at a position 50 mm away from the center of the weld seam. The peak tensile residual stress in the Y direction is 361.43 MPa, which is about 0.95 times, appearing at the fusion line between the weld seam and the base metal at a distance of 25 mm from the center of the weld seam.

The residual stress varies greatly around the weld seam, the X-direction stress on the path L_{yA = 60} and L_{yA = 240} basically conforms to the “M” shaped distribution, and the Y-direction stress does not conform to the “Π” shaped distribution, however, it seems more like an “M” shaped distribution, which is related to the relatively sparse arrangement of measuring points.

It can be seen from Figure 21 and Figure 22 that the stresses in the X and Y directions of the path L_{yA = 240} near the weld seam is greater than the stress of the path L_{yA = 60}, and the residual stress in the X direction is greater than the residual stress in the Y direction in the farther area. From the perspective of the pattern of residual stress distribution, the stress in the vicinity of the weld seam has experienced a process of initially increasing and then decreasing. The earlier welded area (L_{yA = 240}) has higher stress values in the X and Y directions than the latter welding area (L_{yA = 60}) and has narrower coverage. The occurrence of this phenomenon is related to the following two factors: one is the influence of the welding process. The specimen is multi-layer multi-pass welded with a total of 60 weld passes. The welding direction of the 60 weld passes all starts from the run-on plate with a Y coordinate of 300 mm and finishes at the run-off plate with a Y coordinate of 0 mm. The welding heat source always moves from one end to the other, and the heating and cooling of the weld seam and its surrounding metals also start from the earlier welding area. When welding several layers of welds on the surface, the metal temperature in the earlier welding area is always lower than in the latter. From the perspective of preheating, the preheating of the metal in the latter welding area is more uniform and the temperature is higher than that of the earlier welding area, therefore, the peak stress value is smaller and the stress change is relatively smooth. The second is the restraint effect of run-on and run-off plates. For each pass, the metal connecting run-on and run-off plates and the base metal cools to form a new constraint, which increases the constraint of the starting end of the specimen. However, when the weld bead enters the runoff plate, the temperature of the weld bead in the runoff plate is higher than that of the base metal, the elastic modulus and yield strength of the steel are low, and the plasticity is large, therefore the weld bead at the runoff plate is less constrained in the base metal region. The constraint on the starting end is larger than on the finishing end. Therefore, the peak value of residual welding stress at the starting end is larger, and the stress change is more severe. This effect cannot be eliminated after cutting off run-on and run-off plates.

4.1.3. Stress on the Transverse Direction through Weld Ends

The paths L_{yA = 297} and L_{yA = 3} on surface A of the specimen through the weld ends are two transverse paths symmetrical about the center of the weld seam, as shown in Figure 23. Similar to Section 4.1.2, for simplification while not greatly deviating from the actual situation, in the actual stress measurement, only the stress value on the right side of the transverse centerline of the weld seam was measured, and the residual stress value on the left side of the weld seam was estimated assuming the symmetrical distribution of the residual stress.

It can be seen from Figure 24 and Figure 25 that the peak tensile residual stress in the X direction on the path L_{yA = 3} is 131.75 MPa, which is about 0.35 times the yield strength, and appears in the center of the weld seam. The peak tensile residual stress in the Y direction is 250.85 MPa, which is about 0.66 times the yield strength, and it appears near the fusion line between the weld seam and the base metal. The peak tensile residual stress in the X direction on the path L_{yA = 297} is 117.29 MPa, which is about 0.31 times the yield strength, and it appears at a position 50 mm away from the center of the weld seam. The peak tensile residual stress in the Y direction is 115.1 MPa, which is about 0.30 times, and appears in the center of the weld seam. The patterns of residual stress distribution in X and Y directions are consistent with the results discovered in the previous section.

Two of the four residual stress distribution lines have double peaks, and two have single peaks. The residual stress varies greatly around the weld seam. Considering the sparseness of the measuring point layout, it can be assumed that the stress in the Y direction on the path L_{yA = 3} and L_{yA = 297} basically conforms to the “Π” shape distribution, and the stress in the X direction conforms to the “M” shape distribution.

From the above analysis, it can be known that the distribution pattern of residual stress at the weld ends of the specimen is slightly different from the distribution pattern of residual stress near the center of the specimen. There are three main reasons for this phenomenon: 1. The run-on and run-off plates’ restraint significantly influence the residual stress at the edge during welding, and the local stress redistributes after run-on and run-off plates are removed. 2. The arrangement of measuring points near the edge is less and the spacing is larger, there is only one measuring point in and around the weld where peak stress is most likely to occur, resulting in the possibility of not capturing the critical points of stress changes. 3. The center of the measuring point is only 3 mm away from the specimen’s edge, and the specimen’s manufacturing may also affect the stress in this area.

4.2. Stress on the Path along the Weld Seam

This section studies the distribution of welding residual stress along the three longitudinal paths along the direction of the weld seam. As shown in Figure 26, the distribution of the three paths is as follows: the path L_{yA = 0} along the weld seam through the center of the weld seam on surface A, the path L_{yA = 25} along the weld seam which is about 3 mm away from the fusion line on surface A, the path L_{yA = 50} along the weld seam which is about 50 mm away from the centerline of the weld seam on surface B.

As can be seen from Figure 27 that the path L_{yA = 0} is the weld seam centerline on surface A of the specimen. The tensile residual stresses in the Y direction of the three highest value measuring points on the centerline of the weld seam are 195.91 MPa, 127.78 MPa, and 115.10 MPa, which are 0.52, 0.34, and 0.30 times the yield strength, respectively. The tensile stress on the top surface of the weld bead has been at a relatively high level. The maximum tensile residual stress in the X direction is 131.75 MPa, and the remaining stresses in the X direction of the other measuring points are all compressive stresses. The stress in the Y direction is always greater than the stress in the X direction, the stress in the Y direction is basically in a state of tensile stress, and the stress in the X direction is basically in a state of compressive stress. The main reason for this phenomenon is that manual welding is difficult to control the welding direction of the weld bead accurately, the weld bead is not strictly in a straight line. Among the five measuring points on the path, the two measuring points with the lowest stress are located at the weld bead junction, and the remaining measuring points are all on the top surface of the weld bead.

It can be seen from Figure 28 that path L_{yA = 25} is located in the heat-affected zone of welding on surface A, the maximum tensile residual stress in the Y direction in the heat-affected zone can reach 361.43 MPa, and the average stress is 185.94 MPa, which is about 0.49 times the yield strength, which illustrates that the residual stress in the Y direction is always at a high stress level near the fusion line. The highest tensile residual stress in the X direction is 83.25 MPa. From the perspective of the pattern of residual stress distribution, the residual stress generally presents a trend of rising from the starting end, then falling, and then rising to the finishing end. As shown in Figure 29, the measurement points with Y coordinates between 50 and 150 mm are on the fusion line since the weld bead is not straight, while the rest are about 3 mm away from the fusion line.

As can be seen from Figure 30 that on path L_{yB = 50}, the maximum tensile residual stress in the Y direction is 153.22 MPa, which is about 0.40 times the yield strength, and occurs at the edge of the specimen. In addition, the residual stress in the Y direction produces a peak value, the stress is 98.23 MPa in tension, which is about 0.26 times the yield strength. The residual stress in the Y direction changes gradually from the compressive stress to the peak value of the tensile stress, then decreases to the compressive stress, and finally rises to the tensile stress from the starting end to the finishing end. This asymmetry is caused by the slightly stronger restraint effect of the run-on plate at the starting end on the longitudinal expansion and contraction of the weld seam than at the finishing end, and the difference in the position of the measuring point relative to the edge of the weld seam. The maximum tensile residual stress in the X direction is 155.8 MPa, which is about 0.41 times the yield strength, and occurs at the starting end’s edge. It can be seen from the figure that the residual stress in the X direction presents a pattern of first falling, then rising, then falling, and then rising. The distribution of residual stress has a certain symmetry. At the same time, affected by the welding direction, the stress value is slightly higher on the right and slightly lower on the left.

5. Conclusions

This paper introduces the 80 mm thick plate butt welded specimen manufacturing process. It eliminates the influence of welding defects of the weld seam and the base metal’s internal defects on the specimen’s residual stress distribution by ultrasonic flaw detection technology. The basic principles of stress measurement by the X-ray diffraction method are briefly described, the residual stress of the specimen is measured, and the stress distributions in X and Y directions on several representative paths are studied. The following conclusions can be drawn:

(1)

In the area near the weld seam, the tensile stress in the Y direction is greater than in the X direction. The peak tensile stress in the Y direction ranges from 128 to 361 MPa, and the maximum value is about 0.95 times the yield strength of the material. The peak tensile stress in the X direction ranges from 17 to 174 MPa, and the maximum value is approximately 0.46 times the yield strength of the material.

(2)

The peak tensile stress in the Y direction usually appears in the weld seam and the heat-affected zone, and the maximum value tends to occur close to the last weld bead on the surface. The peak tensile stress in the X direction usually occurs within the 50–80 mm range from the center of the weld seam, and the peak stress on both sides has little difference.

(3)

The tensile stress in the Y direction near the weld seam is generally high and fluctuates greatly, and the stress on both sides shows an overall downward trend. The stress level is low in the area far away from the weld seam, which might be compressive or tensile. On the path perpendicular to the weld seam, the stress in the Y direction is distributed in a “Π” shape.

(4)

The stress in the X direction near the weld seam might be tensile or compressive, but the stress on both sides shows an upward trend. However after reaching the peak stress, the stress shows a downward trend, and the stress may be tensile or compressive. The stress in the X direction is distributed in an “M” shape.

(5)

The stresses in X and Y directions are basically symmetrically distributed, however the axis of symmetry may appear anywhere in and around the weld seam. In actual welding, the symmetry of stress value and change trend may degenerate into the symmetry of change trend.

(6)

Measures can be taken to reduce and mitigate residual welding stress, including post weld heat treatment, shot blasting, vibration treatment, stretch processing, explosive treatment and pulsed magnetic treatment, however these measures’ effect on reducing residual welding stress of thick steel plate is awaiting investigation.