Journal of Materials Processing Technology 164–165 (2005) 986–989

Optimizing the laser-welded butt joints of

medium carbon steel using RSM
K.Y. Benyounis ∗ , A.G. Olabi, M.S.J. Hashmi
School of Mechanical and Manufacturing Engineering, Dublin City University, Dublin 9, Ireland

Abstract

The optimization capabilities in design-expert software were used to optimize the keyhole parameters (i.e. maximize penetration (P), and
minimize the heat input, width of welded zone (W) and width of heat-affected zone (WHAZ )) in CW CO2 laser butt-welding of medium carbon
steel. The previous developed mathematical models to predict the keyhole parameters in terms of the process factors, namely laser power (LP),
welding speed (S) and focused position (F) were used to optimize the welding process. The goal was to set the process factors at optimum
values to reach the desirable weld bead quality and to increase the production rate. Numerical and graphical optimization techniques were
used. In fact, two optimization criteria were taken into account. In this investigation, optimal solutions were found that would improve the
weld quality, increase the productivity and minimize the total operation cost. In addition to that, superimposing the contours for the various
response surfaces produced overlay plots.
© 2005 Elsevier B.V. All rights reserved.

Keywords: Laser welding; RSM; Optimization; Keyhole parameters

1. Introduction heat input, bead width of fusion zone as well as bead width
of HAZ should be minimized.
Laser welding with high power density, high degree of au- Minimizing the heat input would result in reducing the
tomation and high production rate is extremely advantageous welding cost through reduced energy consumption and in-
in automotive application [1]. But the point is how to express creased welding productivity through high welding speed. By
the weld bead parameters in terms of process input factors utilizing the above advantages, the weld bead profile could be
to determine the optimum welding conditions. Considering optimized. The objective of this study is to optimize the auto-
that welding is usually done with the aim of producing a good genous laser-welded joints subjected to maximize penetration
joint at low cost. However it is impossible to achieve low-cost and minimize both the fusion zone width and HAZ width. In
welding and good junction without optimization. Trial and er- order to achieve these objectives, mathematical models were
ror methods were previously used to determine the optimal developed to relate the important weld bead parameters and
process conditions for the required weld joint quality [2,3]. the laser welding input variables [6].
Optimization of the weld bead volume ‘minimize’ in SAW The mathematical models developed and optimized for
was studied [4]. Also, optimization of the impact strength of the weld bead profile are very useful to identify the correct
spiral-welded pipes in SAW at different serving temperatures and optimal combination of the laser welding input variables,
was investigated [5]. Despite the different optimization tech- in order to obtain superior weld quality at relatively low cost.
niques used in the previous studies, the goals were reached
and optimal welding conditions were identified to achieve
the desirable weld quality with minimum cost [4,5]. For a 2. Experimental
strong weld, bead penetration should be maximized and the
Medium carbon steel with chemical composition in weight
∗ Corresponding author. percent of 0.46% C, 0.2% Si, 0.7% Mn and Fe, balance
E-mail address: khaled.benyounis2@mail.dcu.ie (K.Y. Benyounis). was used as work piece material. The size of each plate was

0924-0136/$ – see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.jmatprotec.2005.02.067
 K.Y. Benyounis et al. / Journal of Materials Processing Technology 164–165 (2005) 986–989 987

Table 1
Independent process variables and experimental design levels used
Variables Code Unit −1 0 +1
Laser power LP kW 1.2 1.3125 1.425
Welding speed S cm/min 30 50 70
Focused position F mm −2.5 −1.25 0

180 mm long × 80 mm width with thickness of 5 mm. Trial

samples of butt joints were performed by varying one of the
process variables to determine the working range of each
variable. Absent of visible welding defects and at least half-
depth penetration were the criteria of choosing the working
ranges. The experiment was carried out according to the de-
sign matrix in a random order to avoid any systematic error in
the experiment using a CW 1.5 kW CO2 Rofin laser provided
by Mechtronic Industries Ltd. Argon gas was used as shield-
ing gas with constant flow rate of 5 l/min. Two transverse
specimens were cut from each weldment. Standard metallo-
graphic procedures were made for each transverse specimen.
The bead profile parameter ‘responses’ were measured using Fig. 1. Optimization steps.
an optical microscope with digital micrometers attached to it
with accuracy of 0.001 mm, which allow to measure in x and matical models in terms of coded factors are listed below:
y axes. The average of two measured weld profile parameters
was recorded for each response. Heat input = 1260 + 118.29 × LP − 600 × S + 240
× S 2 − 51.43 × LP × S (1)
3. Optimization
P = 3.68 + 0.46 × LP − 0.53 × S + 0.54 × F (2)
The optimization module in design-expert searches for a
combination of factor levels that simultaneously satisfy the
requirements placed (i.e. optimization criteria) on each of W = 2.42 + 0.26 × LP − 0.56 × S − 0.38 × F − 0.31
the responses and process factors (i.e. multiple response op-
timization). Numerical and graphical optimization methods × S 2 + 0.30 × F 2 + 0.23 × S × F (3)
were used in this work by choosing the desired goals for
each factor and response. The optimization process involved
combining the goals into an overall desirability function. The WHAZ = 0.53 + 0.06 × LP − 0.16 × S + 0.03 × F
numerical optimization finds a point or more that maximize − 0.08 × LP × S (4)
this function. While, in the graphical optimization with multi-
ple responses you need to define regions where requirements
simultaneously meet the proposed criteria. Superimposing or 3.2. Optimization method
overlaying critical response contours on a contour plot. Then,
visual search for the best compromise becomes possible. In 3.2.1. Numerical optimization
case of dealing with many responses, it is recommended to Two criteria were introduced in this numerical optimiza-
do numerical optimization first, otherwise you may find it tion. The first criterion is to reach full-depth penetration and
impossible to uncover a feasible region. The graphical opti- to minimize the following: heat input, width of the fusion
mization displays the area of feasible response values in the zone and width of HAZ. In other words, to reach full-depth
factor space. Regions that do not fit the optimization criteria penetration at relatively low welding cost by reducing the
are shaded [7]. Fig. 1 shows the flow chart of the optimization laser power and increasing the welding speed as well as to
steps. obtain excellent joints. While, in the second criterion the goal
was to reach half-depth penetration (i.e. P ≈ 2.5 mm) and to
3.1. Development of mathematical models minimize the following: heat input, width of the fusion zone
and width of HAZ. However, the joint type will be double-
The mathematical models were developed and presented sided butt joint to obtain superior junction. Tables 2 and 3
previously [6]. Box-Behnken design was used to develop the illustrate the goal, lower and upper limits as well as the im-
models. Table 1 presents the selected process control param- portance for each response and factor in the first and second
eters with their limits and units. The adequate final mathe- criteria, respectively.
988 K.Y. Benyounis et al. / Journal of Materials Processing Technology 164–165 (2005) 986–989

Table 2 respectively. While for the second criterion the limits are
The first criterion of numerical optimization 1.2–1.24 kW, 69.77–70 cm/min and −2.03 to −1.71 mm,
Name Goal Lower limit Upper limit Importance respectively.
Laser power Minimize 1.2 1.425 4
Welding speed Maximize 30 70 4
Focused position Is in range −2.5 0 3
Heat input Minimize 822.857 2280 5 4. Results and discussion
Penetration Is target = 5 4.99 5 5
Width Minimize 1.342 3.681 2 Table 4 shows the welding conditions, which lead to full-
Width HAZ Minimize 0.375 0.872 2 depth penetration at relatively low welding cost. It is evi-
dent that to achieve full-depth penetration, the optimal work-
Table 3 ing range for the laser power has to be between 1.38 and
The second criterion of numerical optimization 1.42 kW and the welding speed has to be between 30.48
Name Goal Lower limit Upper limit Importance and 35.55 cm/min using a focused position spanning from
Laser power Minimize 1.2 1.425 4 −0.43 to 0 mm. However, the full-depth penetration achieve-
Welding speed Maximize 30 70 4 ment has a negative effect on both the bead width of WZ
Focused position Is in range −2.5 0 3 and HAZ, due to the high laser power and slow welding
Heat input Minimize 822.857 2280 5 speed used. Table 5 presents the results of the second crite-
Penetration Is target = 2.5 2.499 2.5 5
Width Minimize 1.342 3.681 2
rion in the numerical optimization. It is clear that to achieve
Width HAZ Minimize 0.375 0.872 2 half-depth penetration, the optimal laser power ranges be-
tween 1.2 and 1.24 kW, the welding speed ranges between
69.77 and 70 cm/min and the focused position spanning from
3.2.2. Graphical method −1.71 to −2.03, because of half-depth penetration the weld-
For each response, the limits lower and/or upper have ing has to be double-sided butt-welding to obtain excellent
been chosen according to the numerical optimization welded joints. In this case, the heat was introduced twice,
results. The same two criteria, which are proposed in the which would make the total heat input for the two passes
numerical optimization, were introduced in the graphical to be around 1700 J/cm but it is still less than the minimum
optimization. In the first criterion the lower and upper limits heat input of 1960 J/cm in the first optimization criterion. The
for the laser power, welding speed and focused position reduction in the heat input results in less distortion and im-
are 1.38–1.41 kW, 30.48–35.21 cm/min and −0.43 to 0, prove the weld quality. As the welding speed was doubled,

Table 4
Optimal welding condition based on the first criterion
No. Laser power Welding speed Focused position Heat input Penetration Width Width HAZ Desirability
1 1.41 35.21 −0.00 1967.11 4.99984 2.62583 0.778076 0.243
2 1.41 35.08 −0.00 1972.52 4.99972 2.62288 0.778613 0.243
3 1.40 33.57 −0.00 2035.54 4.99999 2.58837 0.784574 0.234
4 1.41 34.41 −0.04 2006.33 4.99999 2.61728 0.78523 0.230
5 1.39 32.87 −0.00 2065.7 5.00000 2.57212 0.787292 0.224
6 1.42 35.55 −0.05 1960.78 4.99999 2.64371 0.781039 0.223
7 1.39 32.28 −0.00 2090.83 4.99999 2.55602 0.788761 0.213
8 1.38 30.52 −0.00 2166.54 4.99441 2.50473 0.79157 0.128
9 1.41 30.73 −0.32 2210.56 4.99999 2.61339 0.827088 0.106
10 1.42 30.48 −0.43 2238.7 4.99682 2.64875 0.838873 0.059

Table 5
Optimal welding condition based on the second criterion
No. Laser power Welding speed Focused position Heat input Penetration Width Width HAZ Desirability
1 1.20 69.77 −1.71 833.96 2.49964 1.40749 0.383966 0.991
2 1.20 70.00 −1.71 834.653 2.49988 1.40172 0.382422 0.990
3 1.20 69.31 −1.73 835.792 2.499 1.44525 0.385042 0.986
4 1.21 70.00 −1.75 836.94 2.5 1.42156 0.380906 0.985
5 1.20 68.48 −1.78 839.744 2.5 1.51241 0.387102 0.976
6 1.22 70.00 −1.84 842.912 2.5 1.47593 0.37693 0.971
7 1.22 70.00 −1.89 845.724 2.49903 1.50353 0.375005 0.964
8 1.22 70.00 −1.92 847.305 2.499 1.51897 0.373951 0.960
9 1.23 70.00 −1.94 848.642 2.49922 1.532 0.373073 0.957
10 1.24 70.00 −2.03 855.035 2.49971 1.59706 0.368844 0.940
 K.Y. Benyounis et al. / Journal of Materials Processing Technology 164–165 (2005) 986–989 989

the welding cost will be less resulting in improving the pro-

cess productivity. Also, the bead width of welded zone and
HAZ are significantly less in the second criterion. The graph-
ical optimization results allow visual inspection to choose the
optimum welding condition. The shaded areas on the overlay
plots in Figs. 2 and 3 are the regions that do not meet the
proposed criteria.

5. Conclusions

The following points were concluded from this investiga-

tion among the factors limits considered:
1. Design-expert software can be used for optimizing the
weld bead parameters and finding the corresponding op-
timum process factors.
2. Full-depth penetration has a strong effect on the other bead
parameters investigated.
3. Strong, efficient and low-cost weld joints could be
achieved using the optimum welding conditions.
Fig. 2. Overlay plot shows the reign of the optimal working condition based
on the first criterion at LP = 1.42 kW.

Acknowledgements

The authors gratefully thank the Libyan Government for

providing the financial support through the cultural affairs of-
fice in London. The authors also thank the technical support
from the workshop in the School of Mechanical and Manu-
facturing Engineering, Dublin City University.

References

[1] C. Dawes, Laser welding, Abington Publishing, New York, 1992.

[2] S.A. Jasbir, Introduction to Optimum Design, McGraw-Hill, 1989.
[3] M. Aoki, Introduction to Optimization Techniques, Macmillan, New
York, 1971.
[4] V. Gunaraj, N. Murugan, Prediction and optimization of weld bead
volume for the submerged arc process, Part 2, Welding J. (2000)
331s–338s.
[5] K.Y. Benyounis, A.H. Bettamer, A.G. Olabi, M.S.J. Hashmi, Predict-
ing the impact strength of spiral-welded pipe joints in SAW of low
carbon steel, in: Proceedings of IMC21, Limerick, Ireland, 2004.
[6] K.Y. Benyounis, A.G. Olabi, M.S.J. Hashmi, Effect of laser-welding
parameters on the heat input and weld-bead profile, J. Mater. Process.
Technol. 164–165 (2005) 978–985.
Fig. 3. Overlay plot shows the reign of the optimal working condition based [7] Design-Expert software, v6, User’s Guide, Technical Manual, Stat-
on the second criterion at LP = 1.2 kW. Ease, Inc., Minneapolis, MN, 2000.