10

09 layout final:Layout 1

9/8/09

4:37 PM

Page 193

SUPPLEMENT TO THE WELDING JOURNAL, OCTOBER 2009

Sponsored by the American Welding Society and the Welding Research Council

Transient High-Frequency Welding

Simulations of Dual-Phase Steels
Numerical and experimental simulations were used to investigate
high-frequency welding of advanced high-strength steels
BY R. BAUMER AND Y. ADONYI

ABSTRACT
Continued development of advanced high-strength steel (AHSS) requires a corresponding improvement
in joining technology. One promising
joining method is high-frequency butt
joint welding. Seeking to validate the
utility of this process for joining AHSS
flat sheet specimens for steel mill processing lines, high-frequency butt joint
welding of flat sheet steel was investigated through a combined numerical
and experimental simulation methodology. Simulated welds were produced
and pre-Curie and post-Curie temperature heating rate differences were observed with infrared radiation (IR)
imaging. Good correlations were
found between numerical predictions
and actual heating rates. Final metallographic analysis revealed complete
coalescence of faying surface, with
only minor hardening at the weld interface. It was concluded that high-frequency welding shows good potential
for coil joining in steel processing
lines.

Introduction
Modern steel coil processing lines
(such as pickling and galvanizing) benefit
greatly from a continuous feed of steel
strip, a process that requires coil end joining (Ref. 1). As is shown in Fig. 1, continuous processing is achieved through the
combined use of an accumulator (Fig. 1B)
(Ref. 2) and a coil end welding machine
R. BAUMER is a Graduate Student at Massachusetts Institute of Technology, Cambridge,
Mass. Y. ADONYI is the Omer Blodgett Professor
of Welding and Materials Joining Engineering at
LeTourneau University, Longview, Tex.

(Fig. 1C). Due to the finite capacity of the

accumulator, one main constraint on welding process selection is welding speed. Additional constraints, such as material
thickness, result in a variety of welding
processes being used for coil end joining,
including gas tungsten arc welding
(GTAW), gas metal arc welding (GMAW),
flash welding (FW), resistance (mash)
seam welding (RSEW-MS), and laser
beam welding (LBW) (Ref. 3). Note that
the last two are mostly used on tin coating
and recoiling lines, where the sheet thickness is less than 1 mm.
Joining of advanced high-strength
steel (AHSS) coil ends brings additional
challenges to coil end joining due to inherent high strength, prompting difficulties in end shearing, and faying surface
alignment, and a propensity for localized
hardening in welding microstructures. Indeed, excessive hardness in the fusion and
heat-affected zones has been reported in
the case of laser beam welding of dualphase (DP) and transformation-induced
plasticity (TRIP) steels (Ref. 4). The presence of such excessive hardness can provide a metallurgical notch, precipitating
joint failure during subsequent mill operations. To eliminate potential problems
stemming from the as-cast microstructures found in fusion welding, a solid-state
welding process is desired for joining
AHSS coil ends. While resistance welding
(RW) and flash welding (both solid-state
processes) are currently widely employed
in coil joining, RW is slow and is limited to
small thicknesses (Ref. 1) while FW can
be difficult to control because it is susceptible to irregular arcing and incom-

plete fusion on the strip edges. Even resultant strip breaks of 0.2% are not acceptable, as equipment is damaged and
production lost. Therefore, an improved
solid-state joining process is desired for
joining AHSS coil ends.
Previous work has demonstrated that a
coupled high-frequency induction heating/pressure welding (termed hyper-interfacial bonding) operation can produce
faying surface coalescence in butt joint configurations and minimize thermally induced
changes in grain size of ultrafine-grained
steel (Ref. 5). Heating times for 5 5 30mm specimens were shown to be very rapid
(0.2 s to 1600C at 1 MHz and 5059 kW)
(Ref. 5), indicating that high-frequency
welding can satisfy the time constraints associated with coil end joining.
Additionally,
previous
research
demonstrated that high-frequency welding
could produce good welds in AHSS specimens (Fig. 2), as evidenced by successful
limited dome height formability testing
(Ref. 6). The long history of successful
high-frequency induction welding (HFIW)
of joints in tubular products and structural
shapes (Ref. 7) also suggests the usefulness of high-frequency welding for coil end
joining.
This present work builds on this foundation by developing numerical and experimental techniques for simulating
high-frequency welding of dual-phase
steel coil ends, thereby 1) providing insight
into fundamental high-frequency heating/material interactions, 2) establishing
operating parameters, and 3) demonstrating the feasibility of joining DP steel coil
ends with high-frequency welding.

Physical Simulation Overview

KEYWORDS
High-Frequency Welding
Dual-Phase Steels
FEA Modeling
Heating Rates

High-frequency induction heating/

pressure welding of dual-phase steels was
performed at small scale through induction
heating using a solid-state-controlled,
state-of-the-art 100-kW variable-frequency
(250400 kHz) induction welding power
supply. After heating specimens to forging
WELDING JOURNAL 193-s

WELDING RESEARCH

Baumer

Baumer

10

09 layout final:Layout 1

9/8/09

4:38 PM

Page 194

C
A

Fig. 1 Schematic representation of accumulator and coil end welding machine utilized in continuous steel strip processing mills. A Strip to mill; B
accumulator moving vertical rollers allow for a varying amount of strip
to be stored, enabling the stored strip to be fed to the mill line while keeping
the coil end stationary; C welding machine utilized to join the end of the
coil in the mill and the lead end of the next coil. (Fig. 1B reproduced after
Ref. 2 with permission from MetalForming/PMA Services, Inc.)

WELDING RESEARCH

C
Fig. 2 High-frequency welds made in advanced high-strength steels (different composition than the DP600 but still possessing near 600 MPa ultimate tensile strength). A As-welded specimens; B transverse micrograph. Previously published in Ref. 6. Used with permission from US Steel
Research Europe, Kosice, Slovak Republic.

temperatures and turning off the induction

coil power supply, controlled deformation
was delivered by the hydraulic ram system
of a connected Gleeble 1500 thermomechanical simulator. All heating and deformation timing was precisely controlled via
a LabVIEW control program connecting
the two systems. Small-scale welding specimens (1.5 44 89 mm) were rigidly constrained in aluminum jaws and heated by a
water-cooled, copper induction coil Fig.
3A. Atmospheric shielding was accomplished by flooding the welding chamber
with argon.
As compared to the industry standard
high-frequency
resistance
welding
(HFRW) often employed in tube welding

194-s OCTOBER 2009, VOL. 88

Fig. 3 A Physical simulation setup. Note results reported in this present

investigation utilized a single-turn induction coil and did not use the two-turn
coil shown above; B electromagnetic model simulation geometry with representative mesh density; C thermal model simulation geometry with representative mesh density.

(Ref. 7), the experimental simulation technique reported here is distinguished by 1) a

transient process unable to reach the
steady-state operating condition characteristic of HFRW, 2) controlled relative motion between faying surfaces (i.e., not dependent upon the speed or V angle of the
advancing pipe), and 3) variable frequency
between 250 and 400 kHz using the latest
solid-state power control technology.
Fundamentals of the Heating Mechanisms

Heating of welding specimen edges

prior to forging is the consequence of both
eddy current heating, accentuated by the
skin effect, and hysteresis heating (ferro-

magnetic materials only). Eddy current

heating occurs due to resistive heating
losses accompanying induced current flow
in a material. By consequence of the skin
effect, current distributions are restricted
to shallow penetration depths on faying
surfaces, leading to rapid heating and high
efficiencies (Ref. 8). For ferromagnetic
materials, hysteresis heating also occurs, a
direct consequence of inelastic magnetization/demagnetization (represented by
the area enclosed by the BH curve of a
material) (Ref. 9). Naturally, hysteresis
heating ceases when the Curie temperature (approximately 1033 K for ferromagnetic materials) (Ref. 10) is reached and
materials become paramagnetic.

10

09 layout final:Layout 1

9/8/09

4:38 PM

Page 195

Fig. 4 A Comparison of numerical and experiment temperatures; B heating rates vs. time at the faying surface (weld interface center). Numerical simulation results for DP600 heated with a 1750-A induction coil current input (315 kHz). Physical experimental results shown for DP600 welded at 315 kHz and
40 kW.

Table 1 Material Input Parameters for Numerical Simulations of High-Frequency Induction Heating of DP600 Sheet Steel
Temp.
K

273
373
473
573
673
773
873
973
1073
1273

Thermal
Conductivity
Wm11

Temp.

Specific
Heat
Jkg11

Temp.
K

59.5
57.8
53.2
49.4
45.6
41
36.8
33.1
28.5
27.6

323
473
573
623
723
823
973
1023
1073
1123

450
520
565
590
650
730
825
1100
875
846

273
373
473
573
673
773
873
973
1073
1173

Resistivity
m

0.3
0.38
0.44
0.42
0.65
0.78
0.92
1.11
1.34
1.55

Magnetic
Loss
Wkg1
0.00
5512
22050
49612
88200
137812
198450
270112
352800
446512

B
Field
T
0
0.015
0.03
0.045
0.06
0.075
0.09
0.105
0.12
0.135

H
Ampm-1

293 K
r

523 K
r

773 K
r

1023 K
r

0
100
200
300
400
500
1000
2000
3000
4000

0.0
11.0
16.6
20.3
23.2
25.5
35.2
44.8
50.7
55.2

0.0
9.2
13.8
17.0
19.4
21.3
29.5
37.7
42.9
46.8

0.0
6.7
10.0
12.3
14.1
15.5
21.6
27.9
31.9
35.2

0.0
1.9
2.9
3.6
4.2
4.7
6.8
9.4
11.4
13.2

Note: Thermal conductivity values were taken from SAE 1008 carbon steel (Ref. 18); specific heat capacity values were taken from SAE 1010 (3231023 K) and SAE 1008 (1073
1123 K) (Ref. 16). Resistivity values were determined using DP980 (Ref. 17). Magnetic loss was determined at 280 kHz using Equation 9. Relative magnetic permeability was
determined using Equation 10.

vious research indicates that control

of
deformation
input parameters
of total faying surface upset, deformation strain rate,
and forging operation temperatures
can improve coalescence of faying
surfaces
during
high-frequency
welding of highperformance steels
(HPS) (Ref. 11).
Fig. 5 Comparison of numerical and experiment temperature distributions. While the significance of these
Experimental data obtained through static heating trials of DP600.
mechanisms is certainly recognized
in this present reFundamentals of the Deformation
search,
neither
Mechanisms
total upset nor deformation rates were
considered as primary variables in this
Weld quality in high-frequency weldinitial process development, as fundaing is strongly dependent upon the nature
mental heating parameters were of priof deformation at the faying surface. Premary concern.

Objectives
The purpose of this numerical and
physical simulation study was to understand and characterize the thermal phenomena governing the bond quality in
transient high-frequency joining of DP
steels. The study was also intended to validate the concept of designing a steel coil
joining prototype and provide means to
link parametric effects found at small scale
to future full-scale implementation.

Methodology
Induction Heating Simulation Overview

Induction heating is modeled by solving both the electromagnetic response to

current flow through the heating coil and
the thermal response of the specimen to
eddy current and hysteresis heating. The
electromagnetic regime is described by
Maxwells equations and associated constitutive relations (as given by Ref. 15)

 
B = 0

(1)

WELDING JOURNAL 195-s

WELDING RESEARCH

Baumer

Baumer

10

09 layout final:Layout 1

9/8/09

4:38 PM

Page 196

WELDING RESEARCH

Fig. 6 Sample 18: DP 600, 42 kW, 315 kHz, 4 s heating. A Welded sample; B base metal; C grain-refined zone; D weld interface.

 
D = 0


 
B
E =
t



 D
H = J +
t


D = E


B = H


J = E

(2)
(3)
(4)
(5)
(6)
(7)

where B is magnetic induction, D is electric flux density, E is electric field strength,

H is magnetic field strength, is magnetic
permeability, is dielectric constant, and
is conductivity.
Solutions in the electromagnetic
regime yield eddy current and hysteresis
power losses, which comprise the thermal
generation term in the generalized heat
equation (Refs. 12, 15)


T
(k T ) + Q = C
t
196-s OCTOBER 2009, VOL. 88

(8)

where Q is internal heat generation (in this

problem, coming from hysteresis and eddy
current power losses), k is thermal conductivity, is density, and C is specific
heat.
Solving the heat equation yields the resulting temperature distribution in the specimen. Due to the temperature dependence
of all material properties, coupling between
the governing equations is highly nonlinear,
requiring a numerical solution method.
(Refer to Refs. 9 and 1215 for detailed explanations of the governing equations and
associated numerical solutions.)
In this present work, a systematic numerical simulation of induction heating of
flat sheet DP steels was conducted using
the commercial finite element analysis
(FEA) software MagNet and ThermNet
from Infolytica Corp. MagNet enables a
user to model components and currentcarrying conductors (using a CAD-type interface) and solve for the resulting electromagnetic field interaction between the
two via an FEA-based solution method.
Solutions include static, time-harmonic,
and transient responses in two or three dimensions. Similarly, ThermNet provides

an FEA-based solution to the transient

thermal response of a component subject
to heating or cooling phenomenon. Solutions include static and transient responses
in two or three dimensions. Induction
heating is modeled by solving electromagnetic and thermal systems together.
Numerical Simulation: Specimen
Geometry, Mesh, and Boundary
Conditions

Model geometry was defined in two dimensions only, operating under the assumption that edge heating effects were
negligible. Dual-phase steel welding specimens were each said to be 1.6 mm thick
and 43.5 mm in length, with a 1-mm root
opening between faying surfaces. The induction coil was modeled as a solid copper
conductor of 6.35 mm diameter with a
liftoff of 4 mm. Aluminum jaws utilized to
contain specimens during physical trials
were also included in the model Fig.
3B. Simulations were conducted in a simulation cell (referred to as the AirBox) of
dimensions 176 176 mm.
Following the definition of specimen

10

09 layout final:Layout 1

9/8/09

4:39 PM

Page 197

Fig. 7 A Temperature evolution at the weld interface for DP600; B temperature evolution at the weld interface for SAE 1018; C temperature distribution at time of maximum temperature (immediately prior to upset) of DP600; D temperature distribution at time of maximum temperature (immediately
prior to upset) of SAE 1018.

Table 2 Representative Physical Simulation Input Parameters for DP600 and SAE 1018
Material

Frequency
kHz

Power
kW

Gap
mm

Liftoff
mm

Upset
mm

Disp.
Rate
mm s1

Trial Time
s

SAE 1018
DP 600
SAE 1018
DP 600

316
315
403
398

42
42
42
42

1
1
1
1

4
4
4
4

1
2
2
2

10
100
20
20

4
4
4
4

Note: See Appendix 1 for all data.

geometry, the second step in FEA implementation was spatial discretization. In the
electromagnetic model, due to the rapidly
changing electromagnetic fields, the maximum element size (MES) in the sample was
to be no larger than 0.25 mm. No MES was
specified for the AirBox or the coil, and the
computer generated one appropriate to the
geometry Fig. 3C. For the thermal
model, no maximum element size was set
for any part of the model, and an appropriate mesh was automatically generated
Fig. 3C. A polynomial order of two was set
for all element equations.
The electromagnetic boundary condi-

tions were set to have flux lines tangential

to the AirBox perimeter and at the intersection of the AirBox with the sheet specimen. This facilitated the approximation
that the sheet specimen was being inductively heated in an infinitely large volume.
In the thermal regime, the adiabatic
boundary condition was applied to all surfaces, assuming that radiative and convective heat losses were small compared to
conduction of heat within the specimen.
Numerical Simulation: Material Properties

Material properties requiring defini-

tion for successful FEA simulation of induction heating are mass density, thermal
conductivity, specific heat, electrical resistivity (or conductivity), magnetic loss (hysteresis loss), magnetic permeability, and
electric permittivity. Mass density was
taken to be 7600 kg.m3. Thermal conductivity and specific heat were determined
through handbook data for low-carbon
steel (Ref. 16), with the values defined as
shown in Table 1. Electrical resistivity as a
function of temperature was measured
through a modified form of ASTM B-193
(Ref. 17) (Table 1). While simulations
were based on resistivity data collected on
WELDING JOURNAL 197-s

WELDING RESEARCH

Baumer

Baumer

10

09 layout final:Layout 1

9/8/09

4:39 PM

Page 198

Fig. 8 Resistivity vs. temperature profiles for DP600, DP980, and SAE
1025. Resistivity data for SAE 1025 taken from Ref. 16. Resistivity of DP
steels determined through a modified form of ASTM B193-02 (Ref. 17).

WELDING RESEARCH

DP980, additional resistivity testing of

DP600 demonstrated the values to be
nearly identical Fig. 8. Magnetic loss
was determined through the empirically
based Steinmetz equation (Ref. 18):

< P (t ) >= C f B

(9)

where <P v(t)> is the average power loss

per volume; C, , and are constants (set
equal to 87500, 1, and 2, respectively); f is
frequency; and B is the peak magnetic induction. Magnetic permeability was found
through the following equation
r (H ,T ) = 1 +

(T

T ) k1

k2
k2 + H

(10)

where r is the relative magnetic permeability, k1 and k2 are constants (set equal
to 10 and 100, respectively); and Tc is the
Curie temperature. Magnetic induction
(B-field) was subsequently determined by
treating as the first derivative of the B-H
curve and performing numerical integration (Ref. 15).
In the initial stages of material property
definition, experimentally obtained heating profiles of DP steel at low powers (1
kW) were utilized to optimize the correlation between numerical and experimental
simulation heating profiles. Assuming that
hysteresis heating dominates the heating
mechanism before the Curie temperature,
hysteresis loss was treated as a correction
factor and the constants of Equation 1
were iteratively adjusted until an optimal
correlation was obtained with experimental heating curves. Following material
property definition, high-power simulations were conducted (10001750 Arms coil

198-s OCTOBER 2009, VOL. 88

Fig. 9 Hardness data on DP600 high-frequency welding simulation

specimens.

current input, 250400 kHz, and 4.5 s heating time). A complete list of material input
properties for DP600 is provided in Table
1. Electromagnetic models were solved
using the time-harmonic approximation
(material nonlinearities were still considered) while the thermal mode was solved
for the transient solution. A time step of
0.25 s was used in the electromagnetic
model, while a 0.10 s time step was utilized
in the thermal model.
Experimental Simulation Methodology

Dual-phase steel and SAE 1018 carbon

steel specimens (1.5 mm thick) were water
cut to 1.75 3.5 in. (44.45 88.9 mm), sandblasted, and painted with hightemperature HiE-Coat 840-M paint (Ref.
19), designed to ensure consistent material
emissivity and corresponding infrared radiation (IR) camera accuracy. Welds were
made on both DP and carbons steel, with
the primary operating parameters being 4mm coil liftoff, 42 kW, 315 kHz, and a 1-mm
root opening (Table 2). During heating, the
transient thermal response was collected
via two digital IR cameras, which provided
(when operated together) an accurate
measure of a range of temperatures between 298 and 1733 K. Data collected with
these two cameras enabled temperature
evolution at the weld interface and temperature distributions transverse to faying
surfaces to be measured. Following welding, metallurgical analysis was performed
using optical microscopy and Vickers microhardness indentation (1000-g load and
12-s dwell time). Selected specimens were
cross sectioned transverse to the weld interface, mounted, prepared to a 1.0 m
final polish, and chemically etched using
2% Nital etch.

Results
Numerical Simulation

Temperature evolution profiles for the

faying surface of DP600 specimens (FEA
simulation at 1750 Arms input current and
315 kHz) clearly indicated the distinct
heating mechanisms occurring before and
after the Curie temperature Fig. 4A.
Before the Curie temperature (~1033 K),
heating rates are as high as 1000 K s1.
After the Curie temperature, heating rates
drop to approximately 225 K s1, a decrease of approximately 75% Fig. 4B.
Comparison with physical simulation results indicates that the numerical values
fall within approximately 10% of the experimental data, emphasizing that the dramatic decrease in heating rate after the
Curie temperature observed in our numerical simulations is indeed a real result.
While heating profiles at the faying surface correspond well between the numerical and physical simulations, temperature
distribution comparisons (sampled at the
peak temperature in static heating trials)
reveal a growing divergence (with increasing distance from the weld interface) between the numerical and experimental
temperature distributions Fig. 5. Divergence between numerical and experimental simulations in the transverse
heating profiles suggests that our material
property values could be further refined.
We consider Equations 9 and 10 to be the
most likely source of this divergence, especially since Equation 10 deviates somewhat from formulations given in the
literature (Refs. 13, 15). However, the accurate temperature evolution profiles indicate that our general methodology is
sound, and we anticipate that future re-

10

09 layout final:Layout 1

9/8/09

4:40 PM

Page 199

Table 3 Summary of Statistical Analysis

Pre-Curie Temp
Heating Rate
Ks1
Mean
Mean
St. Dev.
Observations
df
Hypo. Dif.
t Stat
t Stat-Critical
1 Tail P Value
DP 600>SAE 1018 by
with Confidence

DP 600
904.9
62.8
8

Post-Curie Temp
Heating Rate
Ks1

SAE 1018
786.0
68.0
9

DP 600
227.6
25.1
25.1

15
60
1.856
1.753
0.04
7.6%
96%

Max. Temp.
K

SAE 1018
225.6
42.2
42.2

DP 600
1697.6
8.9
7

13
0
0.122
1.77
0.45
0%

SAE 1018
1692.5
27.2
3
2
0
0.317
2.920
0.390
0%

Table 4 Composition Information

1.52 mm DP 600
GN (Ref. 23)
AISI 1018 (Ref. 24)

Mn

Si

Cr

Mo

Al

Ti

Other

0.084

1.50

0.009

0.0073

0.010

0.032

0.315

0.054

0.017

<0.002

0.150.20

0.600.90

0.04

0.05

0.150.30

V = 0.003
Cu = 0.037
Cu 0.20

finement of model material properties

(particularly magnetic loss and permeability) could improve results.
Physical Simulation

Both SAE 1018 and DP600 steel sheet

specimens (1.6 44 89 mm) were welded
at the parameters in Table 2, successful
welds being produced in all cases Fig.
6. During welding, IR thermal imaging of
specimens was obtained, enabling the determination of temperature evolution of
the weld interface and temperature distributions transverse to the weld interface
(measured right before the weld plunge).
Curie Temperature Effect on Heating Rate

Analysis of the continuous heating profiles (Fig. 7) indicates a striking difference

in heating rate before and after the Curie
temperature (inflection point in the heating curve), similar to the results obtained
through numerical simulations. This difference was the expected result due to the
sudden change in heating mechanism
upon reaching the Curie temperature. The
magnitude of the heating rate differences
was significant, with the post-Curie temperature heating rate being nearly 75%
less than the pre-Curie temperature heating rates (~225 K s1 vs. ~900 K s1).
Heating Rate Comparisons between DP600
and SAE 1018

In addition to a difference in heating

rate before and after the Curie temperature, comparison of continuous heating
profiles between DP600 and SAE 1018
welding specimens revealed a noticeable
difference (also observed qualitatively
during experiments) in induction heating
response. Statistical analysis performed
using a one-sided t-test (data assumed to
be normally distributed and variances assumed to be unequal) indicated a statistically significant difference between the
mean heating rates for DP600 and SAE
1018. For mean pre-Curie temperature
heating rates, DP600 was shown to be
7.6% greater than SAE 1018 at a 96% confidence level (Table 3). However, postCurie temperature heating rates and
maximum temperatures were not shown
to have any statistically significant
differences.
The lack of a statistically significant difference in heating rate after the Curie temperature indicates that heating arising from
resistive losses is essentially the same in the
two materials, prompting the supposition
that the difference in heating rate between
DP600 and SAE 1018 must arise from a difference in the magnetic response of each
material. Support for this conjecture is supported first by the similarity in resistivity vs.
temperature profiles for dual-phase and
carbon steel (Fig. 8) (Refs. 16, 17). Secondly, such an explanation is bolstered by
the reported sensitivity of the magnetization
response of materials to their composition
and microstructural differences (Refs.
2022). Considering that compositions
(Table 4) (Refs. 23, 24) and microstructures

(by definition of a dual-phase vs. plain carbon steel) are different between DP600 and
SAE 1018, the magnetization response
ought to be different between the two materials, as was observed experimentally.
Metallography

Microstructure analysis of the DP600

specimens revealed grain coarsening at
the weld interface, indicating that upset
was not sufficient to expel the thermally
affected base metal completely from the
weld interface Fig. 6. Microhardness
profiles across the welding zone indicated
a slight hardening at the weld interface
Fig. 9. However, in an extreme case of failure to upset the cast microstructure, hardness at the weld interface was double
those values found in other specimens, underscoring the importance of sufficient
upset to displace unwanted material from
the faying surfaces.
Deformation Response

A maximum force of approximately

275 kN was measured at the faying surfaces, indicating that a maximum pressure
of 20 MPa ensured faying surface coalescence and upset.

Conclusions
1) Fundamental induction heating
mechanisms were revealed through a combined numerical and physical simulation
effort. Specifically, in both DP600 and

WELDING JOURNAL 199-s

WELDING RESEARCH

Baumer

Baumer

10

09 layout final:Layout 1

9/8/09

SAE 1018 (ferromagnetic) materials, heating rates were found to be significantly different before and after the Curie
temperature. Additionally, induction heating rates were found to be dependent on
the ferromagnetic material type.
2) Satisfactory simulated welds can be
produced in sheet steel specimens at 42
kW, 315 kHz, and heating for 4 s with 1mm root opening (gap) between faying
surfaces. This result demonstrates that
high-frequency induction heating/pressure
welding could serve as an excellent solidstate joining process for use in joining steel
coil ends in continuous coil processing
mills.
3) The usefulness of the coupled numerical and experimental simulation tech-

4:40 PM

Page 200

nique was clearly evidenced by the insights

gained into the high-frequency induction
heating process.
Additional study of this new process is
certainly warranted and successful realization of a full-scale prototype of the joining
process is optimistically anticipated.
Acknowledgments
The financial support of this work by
the 2007 AISI/FeMETGrant and US Steel
Europe R&D sponsorship is gratefully acknowledged. We also thank Prof. Robert
Warke for his invaluable contributions to
the project, as well as the following LeTourneau University students: Caleb Mel-

bom, Jordan Smith, Josh Swenson,

Stevenson Jian, Jack Dunaway, Jerica
Cadman, Jody Carter, Mitch Plant, and
Steve Wolbert.
References
1. Halmos, G. T., ed. 2006. Roll Forming
Handbook. Boca Raton, Fla.: CRC Press.
2. Steinmeyer, L. 1999. Accumulate uptime
by joining coils. Metalforming magazine.
http://archive.metalforming magazine.com/1999/
01/coilkent/coils.htm.
3. Shaffer, S. 2003. Coil joining criteria for
tube and pipe mills. Tube & Pipe Journal, www.thefabricator.com/TubePipe Production/TubePipeProduction_Article.cfm?ID=260.
4. Kang, C.-Y., Han, T.-K., Lee, B.-K., and
Kim, J.-K. 2007. Characteristics of Nd:YAG

Appendix 1 Data Table

Table 5 Summary of Data Collected during April 10, 2008, Trial
Spec.
Num

WELDING RESEARCH

Freq.

Power

kHz

kW

Mat.

No.

Root
Upset
Opening
Size
mm

mm

Disp. Forge
Rate Delay

Set
Trial
Time

Liftoff

Time to
Max.
Temp.

mm

K s-1

K s-1

1.0

744

388

mm s-1

Max. Pre-Curie Temp. Post-CurieTemp.

Temp.
Heating
Heating
Rate
Rate

1018

310

300 set

1018

317

42

0.5

2.0

2.4

929

265

1018

317

42

0.5

2.5

2.8

1505

840

310

1018

317

42

2.5

2.7

1434

696

301

1018

317

42

1.5

2.5

2.5

1401

794

273

1018

317

42

3.0

3.0

1539

837

286

1018

317

42

3.5

3.9

1628

724

232

1018

317

42

4.0

4.2

1661

787

205

1018

318

42

4.0

4.0

1706

807

212

10

1018

316

42

10

4.0

4.7

1710

694

196

11

1018

316

42

10

0.25

4.8

3.8

1707

889

226

12

1018

317

42

10

0.25

4.8

4.4

1714

806

200

13

DP 600

315

42

20

0.25

4.75

4.0

1731

969

250

14

1018

316

42

100

0.25

4.5

4.7

1706

836

172

15

DP 600

315

42

20

0.25

4.5

1688

240

16

DP 600

313

42

20

0.25

3.9

1698

831

219

17

DP 600

309

42

20

0.25

1715

242

18

DP 600

308

42

100

0.25

4.2

1701

965

207

19

DP 600

315

42

100

0.25

4.2

1679

915

185

20

DP 600

319

42

20

0.25

3.7

1697

923

216

21

DP 600

319

42

20

0.25

3.6

1697

959

236

22

DP 600

318

42

100

0.25

3.3

1703

872

259

23

DP 600

318

42

100

0.25

3.7

1707

805

248

24

1018

403

42

20

0.25

4.6

1695

692

224

25

DP 600

398

42

20

0.25

3.6

1670

755

226

Frequency, root opening size, upset, displacement rate, forge delay, set trial time, and liftoff were input parameters. Time to maximum temperature, maximum temperature,
pre-Curie temperature heating rate, and post-Curie temperature heating rate were determined via analysis of IR thermal imaging.

200-s OCTOBER 2009, VOL. 88

10

09 layout final:Layout 1

9/8/09

laser welded 600 MPa grade TRIP and DP

steels. Proc. 5th International Conference on Processing and Manufacturing of Advanced Materials THERMEC' 2006. Vancouver, Canada,
pp. 39673972.
5. Saida, K., Jeong, B., and Nishimoto, K.
2004. Development of hyper-interfacial bonding technique for ultra-fine grained steels and
microstructural analysis of bonded joints. Science and Technology of Welding and Joining 9
(6): 548554.
6. Adonyi, Y., Jurko, V., and Lesko, A. 2007.
Modeling and characterization of highfrequency induction welds in advanced high
strength steels. Proc. AIST steel properties and
applications conference, Combined with Ms and
T07, Materials Science and Technology. Detroit, Mich., pp. 195206.
7. Smith, W., and Roberts, J. 1993. Highfrequency welding. Welding, brazing and soldering. ASM Handbook Vol. 6, pp. 252, 253.
Materials Park, Ohio, ASM International.
8. Choi, J.-H., Chang, Y.-S., Kim, C.-M.,
Oh, J.-S., and Kim, Y.-S. 2004. Penetrator formation mechanisms during high-frequency electric resistance welding. Welding Journal 83 (1):
27-s to 31-s.
9. Garbulsky, G. D., Marino, P., and Pignotti, A. 1997. Numerical modeling of induction
heating of steel-tube ends. IEEE Transactions
on Magnetics 33(1): 74652.
10. Lindgren, Art. 1989. Magnetic particle
inspection. Nondestructive evaluation and qual-

4:41 PM

Page 201

ity control. ASM Handbook Vol. 17, p. 89. Materials Park, Ohio: ASM International.
11. Swenson, J., Adonyi, Y., and Warke, R.
2008. Deformation effects during variable highfrequency welding of modern steels. Proc. 8th
International Conference on Trends in Welding
Research, Pine Mountain, Ga. (in-press).
12. Skoczkowski, T. P., and Kalus, M. F.
1989. Mathematical model of induction heating
of ferromagnetic pipes. IEEE Transactions on
Magnetics 25(3): 27452750.
13. Chaboudez, C., Clain, S., Glardon, R.,
Rappaz, J., Swierkosz, M., and Touzani, R.
1994. Numerical modeling of induction heating
of long workpieces. IEEE Transactions on Magnetics 30(6): 50285037.
14. Kim, H.-J., and Youn, S.-K. 2008. Three
dimensional analysis of high-frequency induction welding of steel pipes with impeder. Journal
of Manufacturing Science and Engineering,
Transactions of the ASME 130(3): 031005-1 to
031005-7.
15. Favennec, Y., Labbe, V., Tillier, Y., and
Bay, F. 2002. Identification of magnetic parameters by inverse analysis coupled with finiteelement modeling. IEEE Transactions on Magnetics 38(6): 36073619.
16. 2002. Physical properties of carbon and
low-alloy steels. Properties and selection: Irons,
steels, and high performance alloys. 10th ed.
ASM Handbook Vol. 1, pp. 195199. Materials
Park, Ohio: ASM International.
17. ASTM B193-02. Standard Test Method

for Resistivity of Electrical Conductor Materials.

West Conshohocken, Pa.: ASTM International.
18. Li, J., Abdallah, T., and Sullivan, C. R.
2001. Improved calculation of core loss with
nonsinusoidal waveforms. Proc. Industry Applications Conference, 36th IAS annual meeting.
Chicago, Ill., pp. 22032210.
19. Aremco Products, Inc. 2008. High Temperature Specialty Coatings, technical bulletin
A5. Valley Cottage, N.Y. www.aremco.com/
PDFs/A5_08.pdf.
20. Ryu, K. S., Nahm, S. H., Kim,Y. B., Yu,
K. M., and Son, D. 2000. Dependence of magnetic properties on isothermal heat treatment
time for 1Cr-1Mo-0.25V steel. Journal of Magnetism and Magnetic Materials 222(1-2): 12832.
21. Martinez-de-Guerenu, A., Gurruchaga,
K., and Arizti, F. 2007. Nondestructive characterization of recovery and recrystallization in
cold rolled low carbon steel by magnetic hysteresis loops. Journal of Magnetism and Magnetic Materials 316(2): 842845.
22. Kupferling, M., Appino, C., Basso,V.,
Bertotti, G., Fiorillo, F., and Meilland, P. 2007.
Magnetic hysteresis in plastically deformed lowcarbon steel laminations. Journal of Magnetism
and Magnetic Materials 316(2): e842e845.
23. Auto/Steel Partnership. 2007. Advanced
high strength steel (AHSS) weld performance
study for autobody structural components.
www.a-sp.org/publications.htm.
24. AISI 1018. Alloy Digest. ASM International, Materials Park, Ohio, p. 2002.

Preparation of Manuscripts for Submission

to the Welding Journal Research Supplement
All authors should address themselves to the following questions when writing papers for submission to the
Welding Research Supplement:
Why was the work done?
What was done?
What was found?
What is the signicance of your results?
What are your most important conclusions?
With those questions in mind, most authors can logically organize their material along the following lines,
using suitable headings and subheadings to divide the
paper.
1) Abstract. A concise summary of the major elements of the presentation, not exceeding 200 words, to
help the reader decide if the information is for him or her.
2) Introduction. A short statement giving relevant
background, purpose, and scope to help orient the
reader. Do not duplicate the abstract.
3) Experimental Procedure, Materials, Equipment.
4) Results, Discussion. The facts or data obtained
and their evaluation.
5) Conclusion. An evaluation and interpretation of
your results. Most often, this is what the readers remember.

6) Acknowledgment, References and Appendix.

Keep in mind that proper use of terms, abbreviations,
and symbols are important considerations in processing
a manuscript for publication. For welding terminology, the
Welding Journal adheres to AWS A3.0:2001, Standard
Welding Terms and Denitions.
Papers submitted for consideration in the Welding
Research Supplement are required to undergo Peer Review before acceptance for publication. Submit an original
and one copy (double-spaced, with 1-in. margins on 8 12
x 11-in. or A4 paper) of the manuscript. A manuscript submission form should accompany the manuscript.
Tables and gures should be separate from the manuscript copy and only high-quality gures will be published. Figures should be original line art or glossy
photos. Special instructions are required if gures are
submitted by electronic means. To receive complete instructions and the manuscript submission form, please
contact the Peer Review Coordinator, Erin Adams, at
(305) 443-9353, ext. 275; FAX 305-443-7404; or write to
the American Welding Society, 550 NW LeJeune Rd.,
Miami, FL 33126.

WELDING JOURNAL 201-s

WELDING RESEARCH

Baumer