8/2/2019 Effect of Welding Sequences on Residual Stresses
1/14
Effect of welding sequences on residual stresses
Tso-Liang Teng a, Peng-Hsiang Chang b,*, Wen-Cheng Tseng c
a University of National Defense, Chung Cheng Institute of Technology, Ta-shi, Tao-Yuan 335, Taiwan, ROCb Department of Mechanical Engineering, Da-Yeh University, 112 Shan-Jiau Road, Da-Tsuen, Changhua 515, Taiwan, ROC
c 455 Wing Air Force, Taiwan, ROC
Received 26 March 2002; accepted 2 November 2002
Abstract
Accurately predicting welding residual stresses and developing an available welding sequence for a weld system are
pertinent tasks since welding residual stress is inevitably produced in a welded structure. This study analyzes the
thermomechanical behavior and evaluates the residual stresses with various types of welding sequence in single-pass,
multi-pass butt-welded plates and circular patch welds. This is achieved by performing thermal elasto-plastic analysis
using finite element techniques. Furthermore, this investigation provides an available welding sequence to enhance the
fabrication process of welded structures.
2003 Elsevier Science Ltd. All rights reserved.
Keywords: Welding sequences; Residual stresses; Butt-welded; Circular patch welds
1. Introduction
Metallurgical welding joints are extensively used in
the fabrication industry, including ships, offshore struc-
tures, steel bridges and pressure vessels. Among the
merits of such welded structures include a high joint
efficiency, water and air tightness, and low fabrication
cost. However, residual stresses and distortions can oc-
cur near the weld bead due to localized heating by the
welding process and subsequent rapid cooling. High
residual stresses in regions near the weld may promote
brittle fractures, fatigue, or stress corrosion cracking.Meanwhile, residual stresses in the base plate may re-
duce the buckling strength of the structure members.
Therefore, welding residual stresses must be minimized
to control them according to the respective require-
ments. Previous investigators have developed several
methods, including heat treatment, hammering, pre-
heating, vibration stress relieving, and weld sequencing
to reduce the residual stresses attributed to welding. In
these methods, to choose an available welding sequence
is more simple and efficient for reduction welding re-
sidual stresses. Because many welded structures which
cannot be post-weld manufacturing measures after
welding contain residual stresses of varying degree.
Therefore, developing an available welding sequence
and accurately predicting welding residual stresses for
welds system are necessary for achieving the safest de-
sign.
For a investigation of reducing weld residual stress,Jonassen et al. [1] described the effect of welding pro-
cedures on reducing the residual stresses for butt-welded
steel plates. Rybicki et al. [2,3] developed a method for
reducing tensile stresses on the inner surfaces of the girth
welded pipes. The process entails inductively heating the
outside of a welded pipe while cooling the inner surface
with flowing water. Josefson [4,5] calculated the welding
residual stresses that were numerically analyzed for a
girth-butt welded thin-walled pipe during different post-
weld treatments. Brust and Rybick [6] developed a
method called backlay welding that can be effective in
producing compressive residual stresses on the pipes
* Corresponding author. Tel.: +886-4-8511221; fax: +886-4-
8511224.
E-mail addresses: [email protected], g910404@ccit.
edu.tw (T.-L. Teng).
0045-7949/03/$ - see front matter 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0045-7949(02)00447-9
Computers and Structures 81 (2003) 273286
www.elsevier.com/locate/compstruc
http://mail%20to:%[email protected]%2C/http://mail%20to:%[email protected]%2C/8/2/2019 Effect of Welding Sequences on Residual Stresses
2/14
inner surface. Ueda et al. [7] investigated the effective-
ness of the heat-sink welding to improve the residual
stresses of a pipes circumferential joint. Chou and Lin
[8] reduced residual stresses by parallel heat welding type
304 stainless steel specimens. For the effect of welding
sequence, Weck [9] and Watanabe et al. [10] have stud-
ied how welding sequence affect residual stress build-up.
Kihara et al. [11,12] investigated how welding sequence
affects residual stress and shrinkage in slit-type welds
and circular-path welds. Jonassen et al. [1] describes the
effects of certain block and other special welding pro-
cedures on the magnitude of residual stresses in butt-
welded steel plates of 1-in. thickness.
When welding a long butt-welded joint, multi-pass
butt-welded joint or a path joint, various types of
welding sequences are used in order to reduce residual
stress and distortion. The selection of a proper welding
sequence is an important practical problem. However,
accurately predicting the residual stresses of weldingsequences is extremely difficult because the thermal and
mechanical behaviors in welding include a local high
temperature, temperature dependence of material prop-
erties, and a moving heat source. Therefore, this investi-
gation performs a thermal elasto-plastic analysis using
finite element techniques to analyze the thermomechan-
ical behaviour and evaluate the residual stresses with
various types of welding sequence in single-pass, multi-
pass butt-welded plates and circular patch welded plates.
Furthermore, this study provides an available welding
sequence to improve the fabrication process of welded
structures.
2. Analysis model
2.1. Thermomechanical model
Welding residual stress distributions are calculated
by a finite element method. Fig. 1 presents the analysis
procedures.
2.1.1. Thermal model
In the thermal analysis, a total of 160 load steps in-
crease from 0.001 to 10 s were required to complete the
heating cycle. Only 30 load steps increment were typi-
cally required for the weldment to return its initial
(room) temperature. The time increments were auto-
matically optimized for each time step by the computer
program. The modified NewtonRaphson method was
used in each time step for the heat balance iteration.
This study simulates weld thermal cycles for SAE 1020
steel are shown in Fig. 2. The convective heat coefficientson the surfaces were estimated (using engineering for-
mulae for natural convection) to be 15 W/m2 K.
2.1.2. Mechanical model
In the mechanical analysis, the temperature history
obtained from the thermal analysis was input as a
thermal loading into the structural model. The thermal
strains and stresses can be calculated at each time in-
crement. Also, the final state of residual stresses will be
accumulated by the thermal strains and stresses. During
each weld pass, thermal stresses are calculated from the
temperature distributions determined by the thermal
Nomenclature
q density
C specific heat
T temperature
t timefqg heat fluxQ the rate of internal heat generation
g unit outward normal vector
hf film coefficient
TB bulk temperature of the adjacent fluid
TA temperature at the surface of the model
N element shape functionsfTeg nodal temperature vectorC q
RVCNTN dV
KRV
BTDB dVRAhfNN
TdA
fFegRVQN dV
RAhfTBN dA
fPg surface force vectorffg body force vectorfug displacement vectorfeg strain vector
frg stress vectorB straindisplacement matrixL differential operator matrix
fRgRANTfPg dA
RVNTffg dV
fDreg nodal stress increment matrixfDepg fDeg fDpgfDeg elastic stiffness matrixfDpg plastic stiffness matrixfUeg nodal displacement vectorfDTg temperature increment matrixfCthg thermal stiffness matrixfDTeg nodal temperature increment matrixM temperature shape functionm1fK1g
RV
BTfDepgB dVm1fK2g
RV
BTfCthgM dV
rY longitudinal residual stressrX transverse residual stress
r/ circumferential stress
rr radial stresses
274 T.-L. Teng et al. / Computers and Structures 81 (2003) 273286
8/2/2019 Effect of Welding Sequences on Residual Stresses
3/14
model. The residual stresses from each temperature in-
crement are added to the nodal point location to de-
termine the updated behavior of the model before the
next temperature increment. The material was assumed
to follow the von Mises yield criterion and the associ-
ated flow rules. Phase transformation effects were not
considered in the current analysis due to lack of material
information, especially at high temperatures, such as the
near-melting state.
2.1.3. Element birth and death
The model in this study adopts the technique of ele-
ment birth and death to simulate the weld filler variation
with time in single-pass, multi-pass butt-welded joints
and circular patch welds. All elements must be created,
including those weld fillers to be born in later stages of
18
200
650
16 00
700
350
200
7018
0.00 0.01 0.10 1.00 10.00 100.00 1000.00time (sec)
0.00
400.00
800.00
1200.00
1600.00
Temperature(C)
Fig. 2. Simulated weld thermal cycles for SAE 1020 steel.
Governing Equation Boundary Condition
Governing Equation of the Finite
Element Model
Temperature Field
Compatibility EquationEquilibrium
Equation
Thermal Elasto-Plastic
Equation
Basic Equation of the
Finite Element Model
Displacement Field,
Stress Field
Thermal Model
Analysis
Mechanical Model
Analysis
{ } { }
C
T
tL q Q
T+ = { } { } ( )q h T T
T
f B A =
[ ] [ ]{ } { }C T K T Fe e e
+ =
ki,ljij,klkl,ij +
0lj,ki
=
{ }Te
{ }eU { }e
m
e
m
eK U K T R+ + =1
1
1
2{ }{ } { }{ } { }
}U]{B}[D{}{ eep
e =
}T]{M}[C{ eth
0fij,ij =+
Fig. 1. Flow diagram of the analysis procedure.
T.-L. Teng et al. / Computers and Structures 81 (2003) 273286 275
8/2/2019 Effect of Welding Sequences on Residual Stresses
4/14
the analysis. The method proposed herein does not re-
move elements to achieve the element death effect.
Instead, the method deactivates them by multiplying
their stiffness by a severe reduction factor. Although
zeroed out of the load vector, element loads associated
with deactivated elements still appear in element-load
lists. Similarly, mass, damping, specific heat, and othersuch effects are set to zero for deactivated elements. The
mass and energy of deactivated elements are excluded
from the summations of the model. An elements strain
is also set to zero as soon as that element is killed.
Similarly, when elements are born, they are not ac-
tually added to the model, but are simply reactivated.
When an element is reactivated, its stiffness, mass, ele-
ment loads, etc. return to their full original values.
Thermal strains are computed for newly activated ele-
ments according to the current load step temperature.
2.2. Verification
To confirm the accuracy of the present method, a
specimen was constructed using multi-pass butt-welding,
with a length, width and thickness of L 1000 mm,W 400 mm, t 25:4 mm, respectively, as shown inFig. 3. The welding was done using the submerged arc
technique. Pass sequences and welding parameters are
shown in Table 1 [13]. The material was assumed to
follow the von Mises yield criterion and the associated
flow rules. Linear kinematic hardening was assumed.
Furthermore, Refs. [13,14] specifies the mechanical
properties and stressstrain curves of base metal, weld-
metal, and the heat-affected zone (HAZ) for weldments
of ASTM A36 carbon steel. Therefore, these data are
used here for the residual stress analysis of the butt-
welded joints. The symmetric finite element model has
572 elements and 640 nodes after meshing.
The size of the finite element mesh has a great effecton the accuracy of the results and computational cost.
To examine the adequacy of element size, effect of mesh
refinement in the weld area was studied. The new model
with refined meshes consists of 696 elements and 771
nodes. Very little difference in the results between these
two different mesh models was found. Therefore, the
original FEM model without mesh refinement in the
butt-welded joints can be worked for this verification
study.
Figs. 4 and 5 portray the distribution of the trans-
verse and longitudinal residual stress on the thick plate.
Shim et al. [13] presented experimental results for the
same problem. Additionally, the ABAQUS finite ele-
ment package is applied as a comparison. As the Fig. 4
indicate, the ABAQUS package result showed slightly
lower tensile transverse stress near the weld centerline.
X
YZ
25.4 mm
400 mm
weld bead
A
200 mm
A
weld bead
25.4 mm
1000 mm
12 34 5
678 9
10 11
Fig. 3. Geometry of multi-pass butt-weld.
Table 1
Schematic of pass sequences along with welding parameters for
each pass
Pass no. (111) Voltage (V) Current (A) Speed (mm/s)
1 25 190 3.34
25 26 215 4.70
6 25 190 3.34
79 26 220 4.70
1011 27 250 4.70
0.00 0.05 0.10 0.15 0.20
X (m)
-1.00E+8
0.00E+0
1.00E+8
2.00E+8
3.00E+8
ResidualStress
(Pa)
Present Method
Experiment [14]
ABAQUS [14]
Fig. 4. Transverse residual stress at the top surface of the plate.
276 T.-L. Teng et al. / Computers and Structures 81 (2003) 273286
8/2/2019 Effect of Welding Sequences on Residual Stresses
5/14
Present method results showed good tendency to the
experimental results. As Fig. 5 indicate, both analysis
results show tensile stress near the weld centerline.
Therefore, the procedure presented here is suitable for
analysis of residual stresses and distortions due to welds.
3. Analysis model
3.1. Analysis of the single-pass butt-weld
3.1.1. Specimen and material properties
Fig. 6 illustrates two thin-wall plate sections that are
jointed by a single-pass butt-weld. The length, the width
and thickness of the plate are assumed to be 300, 100
and 5 mm, respectively. The plate material is SAE
1020, and the mechanical properties are dependent on
the temperature history as illustrated in Fig. 7. As Fig. 7
indicates, mechanical properties of metals change under
various conditions when temperature increases, the
modulus of elasticity, yield stress and thermal conduc-tivity decrease while the thermal expansion, specific heat
and Poisson ratio increase. Furthermore, the width of
weld zone was assumed as that of the heat source.
Autogeneous weldment was assumed. These means that
weld metal, HAZ, and base metal share the same me-
chanical properties.
3.1.2. Welding conditions
The welding parameters chosen for this analysis were
as follows: welding method, gas tungsten-arc welding;
welding current, I 110 A; welding voltage, V 20 V;and welding speed, v 5 mm/s, respectively. The heatsources are applied along the weld path for practical
welds.
3.1.3. A finite element model for the single-pass butt-welds
This work develops a two-dimensional symmetrical
plane stress model to estimate the residual stresses of the
single-pass butt-weld using the finite element method.
The model employs two-dimensional four-node plane
elements, including the finite element meshes for the
butt-welded joint. Fig. 8 demonstrates the finite elementmeshes for the butt-weld, along with the refined meshes
used in the weld area. The symmetric model has 500
elements and 561 nodes after meshing.
0.00 0.05 0.10 0.15 0.20
X (m)
-2.00E+8
0.00E+0
2.00E+8
4.00E+8
ResidualStress(
Pa)
Present Method
Experiment [14]
ABAQUS [14]
Fig. 5. Longitudinal residual stress at the top surface of theplate.
Fig. 6. Geometry of single-pass butt-welds.
T.-L. Teng et al. / Computers and Structures 81 (2003) 273286 277
8/2/2019 Effect of Welding Sequences on Residual Stresses
6/14
3.1.4. Mesh sensitivity study
To examine the adequacy of element sizes, the effect
of mesh refinement in the weld area was studied. The
new model with refine meshes consists of 600 elementsand 671 nodes. Results from two mesh densities with the
same material model and geometry. Figs. 9 and 10 both
display the distributions of the longitudinal residual
stress rY along the X-direction (at Y 150 mm) andY-direction with 500 and 600 finite elements mesh
model. Very litter difference in the results between
these two different mesh models was found. It ap-
pears that the main characteristics of the residual
stress results from the different meshes are almost the
same. Therefore, the original FEM model without mesh
refinement in the weld joint can be worked for this
study.
3.2. Analysis of the multi-pass butt-weld
3.2.1. Specimen and material properties
Fig. 11 presents two thick-wall plate sections that arejoined by a multi-pass butt-weld. The length, width and
thickness of the plate are assumed to be 1000, 200 and
12.7 mm, respectively. The mechanical properties are the
same as illustrated in Fig. 7.
3.2.2. Welding conditions
Four passes were involved in the model. Table 2 lists
the welding parameters chosen for this analysis.
3.2.3. A Finite element model for the multi-pass butt-weld
This investigation develops a two-dimensional sym-
metrical plane strain model to calculate the residual
0 400 800 1200 1600 2000
Temp (oC )
0.00
2.00
4.00
6.00
MaterialProperties
Material Properties
Yield Stress
Young's Modulus
Poisson's Ratio
Expansion
Conductivity
Specific He
Symbol Material Properties Unit
--y Yield Stress 108
Pa
---E Youngs Modulus 1011 Pa--- Poissons Ratio 10 1
--- Expansion 10 5 m m K
---k Conductivity 10 2W K m,---c Specific Heat 102 J K Kg,
Fig. 7. The mechanical properties of weldments.
278 T.-L. Teng et al. / Computers and Structures 81 (2003) 273286
8/2/2019 Effect of Welding Sequences on Residual Stresses
7/14
stresses of the thick-wall butt-weld using the finite ele-
ment method. The model employs two-dimensional
four-node plane elements, including the finite element
meshes for the multi-pass butt-weld. Fig. 11 displays the
finite element meshes for the butt-weld, along with the
refined meshes used in the weld area. The symmetric
model has 192 elements and 231 nodes after meshing.
3.3. Circular patch welds analysis
3.3.1. Specimen and material properties
Fig. 12 depicts two thin-wall circular plate sections
that are joined by circular patch welds. The radius of the
inner plate (R1), outer plate (R2), and thickness (t) of the
plate are assumed to be 600, 1500 and 5 mm, respec-
tively. The mechanical properties are the same as dis-
played in Fig. 7.
3.3.2. Welding conditions
The welding parameters chosen for this analysis were
as follows: welding method, gas tungsten-arc welding;
welding current, I 110 A; welding voltage, V 20 V;and welding speed, v 5 mm/s, respectively.
3.3.3. A finite element model for the circular patch welds
This study develops a two-dimensional plane stress
model to calculate the residual stresses of the circular
patch weld using the finite element method. Fig. 13 de-
picts the finite element meshes for the circular patch
welds, along with the refined meshes used in the weld
area. The model has 733 elements and 743 nodes after
meshing.
4. Results and discussion
4.1. Single-pass butt-weld
4.1.1. Longitudinal residual stresses
A stress acting parallel to the direction of the weld
bead is termed a longitudinal residual stress, as denoted
by the letter rY. Fig. 14 illustrates the distributions of the
residual stress rY along the X-direction (at Y 150
mm). High tensile stresses arise in regions near the welddue to a resistance contraction of the material as cooling
commences. Compressive stresses occur in regions re-
moved from the weld for self-equilibriating purposes.
The maximum stress value is as high as the materials
yield stress.
4.1.2. Transverse residual stresses
A stress acting vertical to the direction of the weld
bead is known as an transverse residual stress, denoted
by the letter rX. Fig. 15 represents the distributions of the
residual stress rX along the Y-direction. As this figure
reveal, the stress distributions are symmetrical at the
Fig. 8. The finite element mesh for the single-pass butt-weld.
0.02 0. 06 0.100.00 0.04 0.08
-5.00E+7
5.00E+7
-1.00E+8
0.00E+0
1.00E+8500 elements
600 elements
Fig. 9. Longitudinal residual stress distribution along the X-
direction for different finite element mesh.
T.-L. Teng et al. / Computers and Structures 81 (2003) 273286 279
8/2/2019 Effect of Welding Sequences on Residual Stresses
8/14
middle of the plate, while the tensile stresses occur at the
middle of the plate, and the compressive stresses occur
at the end of the weld.
4.1.3. Effect of welding sequences for the single-pass butt-
weld
Reducing the residual stresses in weld structures
during an early stage of design and fabrication is of
priority concern. For this reason, the effects of weldingsequence on the residual stresses are characterized in the
following. This research investigates the effect of pro-
gressive welding, backstep welding and symmetric
welding on residual stresses for a thin-wall butt-weld as
revealed in Fig. 16. Figs. 17 and 18 both display the
distributions of the longitudinal residual stress rY along
the X-direction (at Y 150 mm) and Y-direction withprogressive welding, backstep welding and symmetric
welding. These figures reveal that the longitudinal re-
sidual stresses of symmetric welding are smaller than
those of the other welding sequences. Since the sym-
metric welding reduced restrained force of the weldment
for why the magnitude of the residual stresses are
smaller than those of the other welding sequences.
4.2. The multi-pass butt-weld
4.2.1. Longitudinal residual stresses
A stress acting parallel to the direction of the weld
bead is termed a longitudinal residual stress as denotedby the letter rY. Fig. 19 depicts the distributions of the
residual stress rY along the X-direction. The longitudi-
nal residual stress develops from longitudinal expansion
and contraction during the welding sequence. A high
tensile residual stress arises near the weld bead along the
weld line, and then decreases to zero, ultimately be-
coming compressive as distance from the weld line. The
residual stress value (110 MPa) approaches the yield
stress of the material. The tensile and compressive re-
sidual stress exist at the weld bead and away from the
welding line on the plate due to the self-equilibrium of
the weldment.
0.05 0.15 0.250.00 0.10 0.20 0.30
Y-direction (m)
-2.00E+7
2.00E+7
6.00E+7
1.00E+8
-4.00E+7
0.00E+0
4.00E+7
8.00E+7
1.20E+8
LongitudinalResidualStressy
(Pa)
500 elements
600 elements
Fig. 10. Longitudinal residual stress distribution along the Y-direction for different finite element mesh.
280 T.-L. Teng et al. / Computers and Structures 81 (2003) 273286
8/2/2019 Effect of Welding Sequences on Residual Stresses
9/14
4.2.2. Transverse residual stresses
A stress acting vertical to the direction of the weld
bead is known as a transverse residual stress as denoted
by the letter rX. Fig. 20 presents the transverse residual
stress variations at the center of the weld bead (X 0,
Y 500 mm) through the plates thickness. The tensilestress occurred at the upper surface of the plate and was
gradually transferred to compressive stress because of
the local bending upwards. Furthermore, the stress dis-
tributions at both surface areas showed a similar mag-
nitude.
Table 2
Welding pass number and parameters for each pass
Pass
no.
Welding parameters
Current(A) Voltage(V) Speed(mm/s) Weldingmethod
1 190 25 3.34 Gas tung-
sten-arc
welding
2 190 25 3.34
3 215 26 5.04
4 215 26 5.04
X
YZ
12.7 mm
400 mm
1000 mm
Weld Bead
AA
Weld Bead
1
2
3
4
Weld passes
Fig. 11. The geometry and finite element mesh for the multi-
pass butt-welds.
R1
R2
Inner Plate
Outer plate
Weld Bead
r
Fig. 12. Geometry of circular patch welds.
Fig. 13. Finite element mesh for the circular path welds.
T.-L. Teng et al. / Computers and Structures 81 (2003) 273286 281
8/2/2019 Effect of Welding Sequences on Residual Stresses
10/14
4.2.3. Effect of welding sequences for the multi-pass butt-weld
A butt-welded plate joint of thick plate is considered
as a model for analysis under the various welding se-
quences in Fig. 21. Fig. 22 presents the distributions of
the longitudinal residual stress rY with various types of
welding sequences. Longitudinal residual stresses be-
tween various weld sequences do not appear to signifi-
cantly differ. Fig. 23 illustrates the distributions of the
transverse residual stress rX with various types of
welding sequences. The transverse residual stress rX of
case (A) welding procedure is smaller than the other
weld sequences. This difference might be attributed to
0.02 0.06 0.100.00 0.04 0.08
X-direction (m)
-5.00E+7
5.00E+7
-1.00E+8
0.00E+0
1.00E+8
LongitudinalResidualSt
ressy
(Pa)
Fig. 14. Longitudinal residual stress distribution along the X-
direction.
0.05 0.15 0.250.00 0.10 0.20 0.30
Y-direction (m)
-1.50E+8
-5.00E+7
5.00E+7
-1.00E+8
0.00E+0
1.00E+8
Transver
seResidualStressx
(Pa)
Fig. 15. Transverse residual stress distribution along the Y-
direction.
Progressive Welding
Backstep Welding
Symmetric Welding
300 mm
200 mm
300 mm
300 mm
1 2 3 4
X
Y
200 mm
1 23
X
Y
200 mm
Y
X
4
1
Fig. 16. The different welding sequence for thin-wall butt-
welds.
0.02 0.06 0.100.00 0.04 0.08
-5.00E+7
5.00E+7
-1.00E+8
0.00E+0
1.00E+8Progressive Welding
Symmetric Welding
Backstep Welding
X-direc tion (m)
ResidualSt
ress
(Pa)
Longitudinal
Y
Fig. 17. Longitudinal residual stress distribution along the X-
direction for different weld sequences.
282 T.-L. Teng et al. / Computers and Structures 81 (2003) 273286
8/2/2019 Effect of Welding Sequences on Residual Stresses
11/14
two reasons: (a) the symmetric welding sequences can
reduce the residual shrinkage of the plate or (b) the
symmetric welding sequences have pre-heating and post-
heating effects.
4.3. Circular patch welds
4.3.1. Circumferential residual stress
Stresses along the welding tangent direction are
called circumferential stress, as denoted by the letter r/.
Fig. 24 depicts the distribution of circumferential re-
sidual stress and reveals that the patch centers residual
stress value is 67 MPa, while the inner minimum value is
52 MPa. The vicinity of weld beads HAZ stress value
rapidly changes from tensile quickly to compressive
stress. This is attributed to equilibrium. Because the
outside plates temperature effect is less than the inner
one, the material is less compressed and elongated than
they. The stress value slowly approaches zero.
4.3.2. Radial residual stressStresses along the welding normal direction are called
radial stresses as denoted by the letter rr. Fig. 25 depicts
the distribution of radial residual stress. The residual
stresses do not markedly differ from each other because
the weld line and patchs radial shrinkage values are
almost equivalent. The weldment has a uniform stress
field in the patchs central region and the residual
stresses are nearly the same according to Fig. 25. The
weldment has a uniform stress field in the patchs central
region and the residual stresses are nearly the same ac-
cording to Figs. 24 and 25.
4.3.3. Effect of welding sequences for the circular patch
welds
This work investigates how progressive welding,
backstep welding and jump welding affect residual
stresses for circular patch welds as illustrated in Fig. 26.
Fig. 27 illustrates the distributions of the residual cir-
cumferential stress r/, with various types of welding
sequences. The circumferential residual stresses between
various weld sequences do not appear to significantly
differ. Fig. 28 depicts the distributions of the radial re-
sidual stress rr, with various types of circular patch
welding sequences. The radial residual stress rr, of the
0.05 0.15 0.250.00 0.10 0.20 0.30
Y-direction (m)
-2.00E+7
2.00E +7
6.00E+7
1.00E+8
-4.00E+7
0.00E+0
4.00E+7
8.00E+7
1.20E+8
LongitudinalResidualStressy
(Pa)
Progressive Welding
Backstep Welding
Symmetric Welding
Fig. 18. Longitudinal residual stress distribution along the Y-
direction for different weld sequences.
0.03 0.08 0.13 0.180.00 0.05 0.10 0.15 0.20
X-direction (m)
-8.00E+7
-2.00E+7
4.00E+7
1.00E+8
-1.10E+8
-5.00E+7
1.00E+7
7.00E+7
LongitudinalResidualStres
s
(Pa)
y
Fig. 19. Longitudinal residual stress distribution along the X-direction for thick plates.
0.002 0.0 06 0.010 .0140.000 0.004 0.0080 .012 0.016
Thickness t (m)
-7.50E+6
-2.50E+6
2.50E+6
7.50E+6
-1.00E+7
-5.00E+6
0.00E+0
5.00E+6
1.00E+7
TransverseResidualStressx
(Pa)
0
Fig. 20. Transverse residual stress variations at the weld line
through the thickness for thick plates.
T.-L. Teng et al. / Computers and Structures 81 (2003) 273286 283
8/2/2019 Effect of Welding Sequences on Residual Stresses
12/14
backstep welding is smaller than the other welding se-
quences because the post-weld heat treatment and pre-
heating effect in backstep welding are better than theother welding sequences.
5. Conclusion
The finite element method is employed herein to
evaluate residual stresses in single-pass, multi-pass butt-
welds and circular patch welds as well as to discuss how
welding sequences affect residual stresses. Based on the
results in this study, we conclude the following:
An extremely large tensile stress occurs near the weld
bead and a compressive stress appears away from the
weld bead in longitudinal residual stresses along the X-
direction for single-pass and multi-pass butt-welds.
The residual stresses are almost uniformly distributed
along the welding direction in longitudinal residual
stresses along the Y-direction for single-pass butt-welds,
except those near the two ends of the weld.
A tensile residual stress is produced at the center
region of the plates, and then suddenly becomes com-
pressive near the two ends of the weld in transverse re-
0.03 0.08 0.13 0.180.00 0.05 0.10 0.15 0.20
X-direction (m)
-8.50E+7
-3.50E+7
1.50E+7
6.50E+7
1.15E+8
-1.10E+8
-6.00E+7
-1.00E+7
4.00E+7
9.00E+7
LongitudinalResidualS
tress
z
(Pa)
UP-SURFACE RESIDUAL STRESS
Case : A
Case : C
Case : B
z
Fig. 22. Longitudinal residual stress distribution along the X-
direction in various weld sequences for thick-wall butt-welds.
0.03 0.08 0. 13 0.18
0.00 0.05 0.10 0.15 0.20
X-direction (m)
2.64E+6
7.64E+6
1.26E+7
1.76E+7
1.40E+5
5.14E+6
1.01E+7
1.51E+7
2.01E+7
TransverseResidualStressx
(P
a)
TRANSVERSE RESIDUAL STRESSx
Case: A
Case: C
Case: B
Fig. 23. Transverse residual stress distribution along the X-
direction in various weld sequences for thick-wall butt-welds.
X
YZ
12.7 mm
400 mm
1000 mm
Weld Bead
A
1
2
3
4
200 mm
2
1
3
4
200 mm
3
1
2
4
200 mm
12.7 mm
12.7 mm
12.7 mm
Case(A)
Case(B)
Case(C)
A
Weld Bead
Fig. 21. The different welding sequence for thick-wall butt-
welds.
284 T.-L. Teng et al. / Computers and Structures 81 (2003) 273286
8/2/2019 Effect of Welding Sequences on Residual Stresses
13/14
sidual stresses along the Y-direction for single-pass butt-welds.
Tensile stress occurs at the upper surface and is
gradually transferred to compressive stress at the bot-
tom surface in transverse residual stresses through the
plates thickness for multi-pass butt-welds.
In circumferential residual stress for circular patch
welds, the weldment has a uniform tensile stress field in
the patchs central region, and then decreases to com-
pressive, finally becoming zero as distance from the weld
bead.
In radial residual stress for circular patch welds, the
weldment has a uniform tensile stress field in the patchs
central region, and the residual stresses do not markedly
differ away from the weld bead.
Different welded geometrical configurations or wel-
ded joints have various available welding sequences that
0.15 0.45 0.75 1.05 1.350.00 0.30 0.60 0.90 1.20 1.50
Radius r (m)
-2.00E+7
2.00E+7
6.00E+7
1.00E+8
-4.00E+7
0.00E+0
4.00E+7
8.00E+7
1.20E+8
CircumferentialResidualS
tress
(Pa)
Fig. 24. Circumferential residual stress distribution for circular
patch welds.
0.15 0.45 0.75 1.05 1.350.00 0.30 0.60 0.90 1.20 1.50
Radius r (m)
3.20E+7
4.20E+7
5.20E+7
6.20E+7
7.20E+7
2.70E+7
3.70E+7
4.70E+7
5.70E+7
6.70E+7
7.70E+7
RadialResidualStress
r
(Pa)
Fig. 25. Radial residual stress distribution for circular patch
welds.
1
12
34
1
2
3
4
progressive welding
backstep welding
jump welding
Fig. 26. The various welding sequences for circular patch
welds.
T.-L. Teng et al. / Computers and Structures 81 (2003) 273286 285
8/2/2019 Effect of Welding Sequences on Residual Stresses
14/14
depend on the restraints that happen during the welding
procedure. Symmetric welding is an available welding
sequence for single-pass welds. In multi-pass welds, case
(A) weld procedure is an available welding sequence.
Backstep welding is an available welding sequence for
circular patch welds.
More free space should be available for expansion
and shrinkage in the welding structure during the
welding procedure to prevent the rigid restraint in theweld bead, and, consequently, to decrease the residual
stress.
References
[1] Jonassen F, Meriam JL, Degarmo EP. Effect of certain
block and other special welding procedures on residual
welding stresses. Weld J 1946;25(9):492s6s.
[2] Rybicki EF, Mcguire PA. The effects of induction heating
conditions on controlling residual stresses in welded pipes.
Trans ASME J Eng Mater Technol 1982;104:26773.
[3] Koch RL, Rybicki EF, Strttan RD. A computational
temperature analysis for induction heating of welded pipes.
J Eng Mater Technol 1985;107:14853.
[4] Josefson BL. Residual stresses and their redistribution
during annealing of a girth-butt welded thin-walled pipe.
Trans ASME J Press Vessel Technol 1982;104:24550.
[5] Josefson BL. Stresses redistribution during annealing of a
multi-pass welded pipe. J Press Vessel Technol 1983;105:
16570.
[6] Brust FW, Rybick EF. A computational model of backlay
welding for controlling residual stresses in welded pipe.
J Press Vessel Technol 1981;103:22632.
[7] Ueda Y, Nakacho K, Shimizu. Improvement of residual
stresses of circumferential joint of pipe by heat-sinkwelding. J Press Vessel Technol 1986;108:1423.
[8] Chou CP, Lin YC. Reduction of residual stresses by
parallel heat welding in small specimen of Type 304
stainless steel. Mater Sci Technol 1992;8:17983.
[9] Weck R. Transverse contractions and residual stresses in
butt welded mild steel plates. Report No. R4, Admiralty
Ship Welding Committee, January 1947.
[10] Watanabe MS, Satoh K, Kimura K. Effect of welding
methods and sequences on the residual stress distribution
of welded joints. Japan Weld Society 1955;24(4):14653.
[11] Kihara H, Masubuchi K. Studies on the shrinkage and
residual welding stress of constrained fundamental joint.
Transportation Technical Research Institute, Report No.
20, 1956.[12] Kihara H, Masubuchi K, Mastuyama Y. Effect of welding
sequence on transverse shrinkage and residual stresses.
Transportation Technical Research Institute, Report No.
24, 1957.
[13] Shim Y, Feng Z, Lee S, Kim D, Jaeger J, Papritan JC, et al.
Determination of residual stresses in thick-section weld-
ments. Weld J 1992;(September):30512.
[14] Higashida Y, Burk JD, Lawrence FV. Strain-controlled
fatigue behavior of ASTM A36 and A514 grade F steels
and 5083-0 aluminum weld materials. Wel Res Suppl
1978;(November):334s44s.
0.15 0.45 0.75 1.05 1.350.00 0.30 0.60 0.90 1.20 1.50
Radius r (m)
-2.00E+7
2.00E+7
6.00E+7
1.00E+8
-4.00E+7
0.00E+0
4.00E+7
8.00E+7
1.20E+8
CircumferentialResidualStress
(Pa)
Circumferential Residual Stress
Progressive Welding
Backstep Welding
Jump Welding
Fig. 27. Circumferential residual stress distribution in various
weld sequences for circular patch welds.
0.20 0.60 1.00 1.400.00 0.40 0.80 1.20
Radius r(m)
3.20E+7
4.20E+7
5.20E+7
6.20E+7
7.20E+7
2.70E+7
3.70E+7
4.70E+7
5.70E+7
6.70E+7
7.70E+7
RadialResidualStressr
(Pa)
Radial Residual Stress r
Progressive Welding
Backstep Welding
JumpWelding
Fig. 28. Radial residual stress distribution in various weld se-
quences for circular patch welds.
286 T.-L. Teng et al. / Computers and Structures 81 (2003) 273286