Sie sind auf Seite 1von 10

Applied Surface Science 257 (2011) 4113–4122

Contents lists available at ScienceDirect

Applied Surface Science


journal homepage: www.elsevier.com/locate/apsusc

Analysis of a fibre laser welding case study, utilising a matrix flow chart
J. Karlsson ∗ , A.F.H. Kaplan
Department of Applied Physics and Mechanical Engineering, Luleå University of Technology, SE-971 87 Luleå, Sweden

a r t i c l e i n f o a b s t r a c t

Article history: For fibre laser welding of an eccentric corner joint, the quality of the resulting weld cross section was
Received 26 March 2010 studied with respect to the dependence on process parameters like lateral laser beam alignment, beam
Received in revised form inclination, focal plane position or welding speed. The complex load situation of the support beamer
27 September 2010
was simplified to bending of one corner. Due to fatigue load, the weld properties causing the peak stress
Accepted 30 November 2010
are essential, in particular the top and root shape of the weld cross section. For the parameters varied,
Available online 17 December 2010
the resulting shapes were categorized into different top and root classes, determined by certain key
dimensions, considering also welding defects like undercuts. The shapes are boundary conditions for
Keywords:
Laser welding
Finite Element Analysis of the joint under load for quantitative comparative analysis of the maximum
Parameters stress. As two high strength steel grades were joined, the hardness transition across the weld was of
Geometry interest, too. High speed imaging of the weld pool surface shape provided additional information on the
Load relation between the parameter input and quality output. The different trends identified were discussed
Shape and guidelines were derived. As the systematic documentation of results is unsatisfactory in welding,
Defect a new method was developed and applied for the first time, called the Matrix Flow Chart. It enables
an illustrative view on the resulting welding trends in a combined manner and is extendable by other
researchers.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction ing creates a narrow HAZ, often even without softening, or at least
of little contribution as a narrow part of the overall weld material
The quality of a weld is regarded from two points of view. Dur- compound under mechanical load. Laser hybrid welding of ultra
ing manufacturing the process parameters chosen determine the high strength steel up to a tensile strength of 1700 MPa has proven
welding process and in turn the weld quality. When the welded to maintain its properties, while during tensile testing the cracking
product is in service, the weld tries to resist the load conditions took place in the base material [1].
(beside other criteria) for which poor weld quality can cause frac- Besides the metallurgical material properties, essential for the
ture. mechanical behaviour of a weld under load is its geometry, particu-
Laser welding is characterized by superior abilities compared to larly the top and root surface shapes, moreover welding defects. In
conventional welding (e.g. resistance spot or electric arc welding), Fig. 1 common laser welding defects in metals are shown, accord-
like enabling high precision and high quality, but is held back by ing to the standards [2], accompanied by possible explanations of
high investment costs and difficulties in achieving the high quality their causes.
in a robust manner. The process window is often narrow and the Different welding defects have to be suppressed to achieve a
physical mechanisms of the process are complicated and not fully quality that can handle the loading situations that the final product
understood. One application of high potential as an enabler tech- will have to endure. Due to their good beam quality even at high
nology is laser welding of high strength steel of different grades, as power the recently emerging fibre and disc lasers have made it
traditional methods soften the steel in the heat affected zone (HAZ). possible to increase the welding performance. Unfortunately those
Due to its high speed and the narrow interaction zone, laser weld- laser welds are often accompanied by heavy spatter ejection, a
defect which has to be suppressed [3]. For monitoring of welding
defects a promising method are photodiodes [4,5] or cameras [5].
Advanced methods to analyze the melt pool and vapour flow during
∗ Corresponding author at: Division of Manufacturing Systems Engineering, Luleå laser welding are high speed imaging [3,6,7] or Finite Element Anal-
University of Technology, SE-971 87 Luleå, Norrbotten, Sweden. ysis (FEA) [8]. Recently, it was shown that combinations of analysis
Tel.: +46 920 49 1675; fax: +46 920 49 2228; mobile: +46 70 520 47 50. methods can be a powerful tool, e.g. FEA and high speed imag-
E-mail addresses: jan.karlsson@ltu.se (J. Karlsson), alexander.kaplan@ltu.se
(A.F.H. Kaplan).
ing [8], X-ray imaging of pore formation followed by modelling [6]
URL: http://www.ltu.se/tfm/produktion (J. Karlsson). or analysis of photodiode monitoring by continuing thermal emis-

0169-4332/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.apsusc.2010.11.185
4114 J. Karlsson, A.F.H. Kaplan / Applied Surface Science 257 (2011) 4113–4122

Fig. 1. Different types of defects occurring during laser welding.

sion modelling from high speed imaging evidence of the weld pool a low C-steel fillet joint with different gaps and chamfers, where
dynamics [7]. undercut was likely to take place [13]. For laser hybrid welding of
Also the limits of acceptable weld surface properties are defined an eccentric stainless steel butt joint a detailed investigation of the
by the standards, but are still unsatisfactory as the knowledge for top surface was carried out [14]. Little toe radius was the primary
load situations is so far not enough generalized. The shape of laser criterion (being also a new welding standard guideline) for fatigue
welds was often studied [9], but not in a systematic and generalised crack initiation, but surface ripples by resolidification could prevail
context between publications. Typical for laser welding are nail- and shift the cracking location, guiding the cracks along the ripples.
shaped cross sections, often generated by Marangoni convection For fatigue bending load either the top or the root often is under
depending on the weld pool geometry and also on the shielding gas compression and thus uncritical, while the other side governs the
type [10]. For the new generation of high brightness lasers (Yb:fibre fatigue crack initiation and propagation essentially by its shape. An
laser, disc laser), spiking can disturb the root shape [11]. Numeri- overview of fatigue analysis of welds is given in [15], and of fatigue
cal simulation of the heat and fluid flow during laser welding has properties and cracking behaviour in [16,17].
been studied by several authors [12] to understand the weld sur- Despite a large number of experiments, case studies and appli-
face motion and its resolidification, governing the weld shape. Laser cations of welding (including laser welding) carried out in the past,
hybrid welding is a complex process owing to its combination with the transfer of knowledge to other applications is difficult. Even
a gas metal arc, but it enables to shape the root (by the laser) and large welding databases were reported to be unsatisfactory since
the top (by the electric arc) separately, as studied systematically for each parameter situation is different, i.e. it is difficult that one point
J. Karlsson, A.F.H. Kaplan / Applied Surface Science 257 (2011) 4113–4122 4115

2. Methodology

2.1. Methodological approach

A certain joint and material situation was under investigation,


to be fibre laser welded. For a preferred laser beam power, several
(mainly geometrical) process parameters were varied in a selective
manner, to keep the number of experiments limited, but system-
atic. The resulting weld cross sections were commonly arranged in
a comparable manner, suitable for predicting trends of shapes in
the parameter vicinity. As fatigue bending across the weld is the
load criterion applied, the top and root shapes of the weld cross
sections are of key importance as possible stress raisers. There-
fore the shapes were judged and categorized, including definition
of the key dimensions. From the parameter variations, trends of
the weld shapes and of possible defects can be derived. The weld-
ing process was recorded by high speed imaging. Accompanying
information was achieved by comparing the observed weld pool
shape for some of the varied parameters. As two different high
strength steel grades were joined, trends in the hardness distribu-
tion across the weld were derived, too. The weld quality achieved
is the essential input determining stress raisers and possible frac-
ture under load. Comprehensive FE stress analysis was conducted
for different joints and weld shapes, to be published in another
paper. Selected results from the stress analysis are presented to
illustrate the stress behaviour, the impact of the top and root
shapes and differences for different joint types. This is of partic-
ular importance to transfer the main findings to similar geometries
of relevance. The findings are discussed and conclusions are
drawn.
Fig. 2. Work piece joint setup. (a) cross-section of the support beam and positioning While the research normally would end at this stage, the article
of the laser towards the joint, (b) joint and dimensions, (c) cross section of the weld.
continues by applying the Matrix Flow Chart as a new documenta-
(d) 3D-illustration of the melt pool in laser keyhole welding (here: butt joint) [i)
Weldox 960, ii) Domex 700, iii) weld joint, iv) laser beam path and direction, v) tion method for each of the above findings presented, thus creating
beam inclination angle, vi) laser cut surfaces, vii) root and viii) top side of weld cross a systematic nucleus of results that hopefully is easier to track, to
section, respectively.]. interpret, to discuss, to refer to and to extend.

2.2. Experimental set-up and process parameters


in the N-dimensional parameter space (e.g. N = 20–30 parameters)
learns from another. Design of Experiments is a systematic method The joint under investigation is an eccentric corner joint with
for solving N-dimensional problems, but was successfully applied 6 mm interface, see Fig. 2(b), which aims at making a support beam
in laser welding only in limited process parameter domains and consisting of four equal welds, see Fig. 2(a). Two different grades
when supported by an expert (additional intuitive skills), as the of high strength steel are welded together, namely Domex 700
process is too complex and trends are not monotonic. Despite these (6 mm thick) and Weldox 960 (15 mm thick). The chemical com-
limitations in transferring knowledge and the desire for improve- position and important material properties of the two materials
ments, new approaches and solutions are hardly addressed. (steel maker: SSAB) are shown in Tables 1 and 2. Both plates (each:
Inspiration from other disciplines is appreciated. For example in sample size 100 mm × 200 mm) were laser cut, enabling zero gap
machining a management information system was developed [18]. before welding. The resulting oxide layer was not removed. The
While many solutions have been developed, like for geometric welding technique applied is deep penetration laser welding, as
properties such as shape identification, the treatment of complex it is particularly suitable for maintaining the high steel strength,
and abstract data such as parameter combination in a multidi- but also due to other advantages like high precision, high welding
mensional space is rare, and strategies are not obvious. Within speed and little distortion. Fig. 2(c) shows a typical laser weld cross
genetics developed artificial neural networks have been devel- section of the joint. Fig. 2(d) illustrates the long-section of the melt
oped, connecting input-properties with selected output-criteria, pool and of the vapour capillary (keyhole) that is induced by and
weighed and marked [19]. Increased treatment and visualization accompanying the continuous wave (cw) laser beam [21]. Several
of data depends on the human cognitive ability and disability [20], melt flow phenomena can occur in the weld pool, which essen-
e.g. the meaning of colours for our perception. Regarding welding tially influence the resulting weld quality at the resolidification
and usage of knowledge transfer visualization, perception and front.
cognition are barely mentioned but can be the key for unravelling a As will be discussed later for the FE-analysis, three different
new way of how to transfer knowledge in a new and more efficient joint types are under comparison, all suitable for laser welding. The
manner. here studied eccentric joint with 6 mm interface is properly dimen-
The present article shows, discusses and categorizes the results sioned with respect to the thinner plate and enables in the case of
of the weld quality for systematic variation of the parameters, full penetration shaping of both the top and the root geometry of
during dissimilar fibre laser welding of two high strength steel the weld, which is promising for influencing the fatigue cracking
grades. In addition, a method was developed to present and docu- behaviour and therefore studied in the present article.
ment the results in a new, extendable manner, hopefully easier to The process parameters varied during the experiments are
survey. welding speed v, inclination ˛ of the laser beam as well as its focal
4116 J. Karlsson, A.F.H. Kaplan / Applied Surface Science 257 (2011) 4113–4122

Table 1
Chemical composition [wt%] of the materials used for the plates.

Steel type/element C Si Mn Cr Ni P S N Al Fe

Domex 700 0.12 0.10 2.10 – – 0.025 0.01 0.015 – Balance


Weldox 960 0.20 0.5 1.60 0.70 1.5 0.02 0.010 0.0 0.018 Balance

Table 2
Mechanical properties of the materials used.

Steel type/properties Rp0.2 (N/mm2 ) min Rm (N/mm2 ) min A5 (%) max Hardness (VH)

Domex 700 700 750 12 290


Weldox 960 960 980 12 340

plane position z0 and lateral offset y, all according to Table 3, result-


ing in 17 welded samples.
All other parameters are fixed, as follows: laser type: Yb:fibre
laser YLR-15000 (IPG Laser), maximum beam power: 15 kW (cw),
output fibre diameter: 200 ␮m, Beam Parameter Product (beam
quality for focusing) 10.4 mm mrad, wavelength: 1070 ± 5 nm.
The laser power applied was 5000 W (cw) corresponding to the
expected production capacities and for the welding speed that
achieves full penetration with this power. The average power den-
sity in the focus results then to 1.42 × 106 W/cm2 .
Optics (2 lenses): collimator length 150 mm, focal length:
500 mm, resulting magnification M = 3.3. Accordingly, the projected
fibre end results in a focal spot diameter of 0.67 mm and a Rayleigh
length (depth of focus) of ±8 mm. The longest among five focal
lengths was chosen, to provide a sufficiently long Rayleigh length
compared to the penetration depth (6 mm, but even when later
welding through the 15 mm plate).
Shielding gas: argon, gas flow 20 l/min, through a nozzle near and
aimed at the weld (top side).
High speed imaging (4000 frames per second) was applied
during the experiments (viewing at 45◦ from the side) to gain addi-
tional information. For this, a diode laser illuminated the weld,
enabling spectrally filtered imaging where most of the dominating
vapour emissions can be blocked [7,21].

2.3. Evaluation of the welds

As mentioned above, the resulting welds were etched, pho-


tographs were taken under an optical microscope and the shapes,
including possible defects, were judged and classified and criti-
cal dimensions were measured. The hardness (Vickers, HV) was
measured across the weld in a middle zone. Moreover, the weld
shape and HAZ shape were identified. In particular two HAZ can

Table 3
Variation of parameters.

Weld no. Weld speed (v) Focal position Inclination Lateral offseta
(m min−1 ) (z0 ) (mm) angle (˛) (◦ ) (y) (mm)

1 1.0 −1 5 0.25
2 1.0 −1 5 0.375
3 1.0 −1 5 0.50
4 1.0 −1 5 0.625
5 1.25 −1 5 0.375
6 1.5 −1 5 0.375
7 1.75 −1 5 0.375
8 2.0 −1 5 0.375
9 2.25 −1 5 0.375
10 2.5 −1 5 0.375
11 2.5 −1 10 0.5
12 2.0 −1 10 0.5
13 1.5 −1 10 0.5
14 1.5 −2 10 0.625
15 1.5 −3 10 0.75
16 1.5 −4 10 0.5
17 1.5 −5 10 0.5 Fig. 3. Weld cross section matrix of parameter changes. Slashed line: joint interface,
a
Offset from the left plate. dotted line: laser beam axis.
J. Karlsson, A.F.H. Kaplan / Applied Surface Science 257 (2011) 4113–4122 4117

Fig. 4. Weld cross sections and compared (upper) at four locations (indicated) along
the weld (top and root appearance, lower).

be found towards the Domex 700-plate. No fatigue testing was


carried out for this kind of joint, only for the partial penetration
L-joint shown later, as only for the latter fatigue testing equipment
for a whole beamer was available. For the resulting weld shapes
and for different joint types a comprehensive stress analysis was
carried out by Finite Element Analysis (ANSYS code, linear elas-
tic two-dimensional stress analysis), to be published in separate
papers [22]. A few results will be given just to illustrate the basic
trends.

2.4. The Matrix Flow Chart as a new documentation method

As mentioned before, a fundamental dilemma is the limited


transferability of research results in welding today. No guarantee
exists that the findings from one case study with its parameters are
applicable for another case and operating point. In order to transfer
welding information for use in a larger perspective, the creation of
a wide knowledge platform may be a solution.
Nevertheless, e.g. a solution to eliminate a welding defect can
be transferable. Guidelines and trend rules instead of numerical
parameter values can be a solution, but their range of validity is not
obvious. Moreover, standardized formats and categories are miss- Fig. 5. (a) Top and (b) root geometry classes identified for the eccentric corner joint:
ing through which knowledge could be combined and extended suggested dimensions of importance to measure.
[23]. A first approach for documentation of the physics of the laser
welding process for suppressing (or not) a certain welding defect
is the Bifurcation Flow Chart, BFC, recently applied [7,24]. The BFC
was immediately applicable to different findings that could be com- tially needs knowledge/experience from weld results, but aims to
bined, enabling tracking of knowledge as well as new conclusions be transferable to other welding cases and applications.
from the common view. However, the BFC-method is still not ideal An approach is to categorize different weld geometries as the
when the goals are different (from the original sequential mecha- quality output governing the durability of a weld when put under
nisms for a single output). load. This in combination with a system that contains changes in
In the present study the results are many different input- and parameters related to the results, all presented in a visually lucid
output-properties (parameter in, weld quality out), while the pro- manner, might help with knowledge transferability. Inspired by
cess is so far regarded as a black box (except the high speed the matrix arrangement from computational logics and electron-
imaging). Therefore a new approach was developed, the Matrix ics, the Matrix Flow Chart (MFC) was here developed, a kind of
Flow Chart. The creation of a table that is lucid, conceivable and black box variant of the BFC, but of parallel rather than sequen-
generalised may be needed. Instead of making the table consist of tial nature. The MFC method is presented and applied here for
pure numbers, it should contain rules for what can be expected the first time, to all essential results obtained, as will be shown in
to happen when different parameters are altered. Such a table ini- Section 5.
4118 J. Karlsson, A.F.H. Kaplan / Applied Surface Science 257 (2011) 4113–4122

3. Results and discussion

From the experiments made with the parameters in Table 3 the


following results where obtained:

3.1. Cross section geometry analysis

In Fig. 3 the weld cross sections of the experiments are shown.


They are placed in a manner so that all parameter changes are
visible in the same lucid manner, to visualize how the parameter
changes affect the shape of the welds.
In Fig. 4 one of the samples is cut into layers with 2 mm inter-
val. It can be seen that the weld does not vary notably in welding
direction. Hence it can be concluded that the chosen samples are
representative.

3.2. Top and root geometry classes

From the cross sections in Fig. 3, five different top and five
different root classes for an eccentric corner joint were distin-
guished, as shown in Fig. 5. The different top and root geometries
are defined by certain key dimensions that can be of importance for
the stress formation and crack initiation when the weld is under
load.
In Fig. 6 the top and root geometry classes are arranged in the
same manner as in Fig. 3, showing how they change with parameter
changes for the eccentric corner joint.
In Fig. 7 the trends of key geometry dimensions (ref. Fig. 5) by
parameter changes can be seen. The trends for the offset can be
expected, melting more material at the top when being left, but
reinforcing the root when shifted to right. When the focus position
changes, it can be seen that the width of the top side of the plate
increases. At the same time the width at the root side decreases
until the focus point reaches about half way into the plate. At this
point the width on the top and root starts to decrease/increase
respectively. The change in speed shows that the top weld width
increases with increasing speed, while root sagging decreases and
finally disappears beyond the penetration limit. Increasing speed
also results in decrease of topside undercut (less sagging, or heavy
vaporization). When the inclination angle changes from 5◦ to 10◦ ,
for different welding speeds root sagging and top reinforcement is
reduced. The weld width decreases with 5◦ and increasing speed,
while it increases with 10◦ .

3.3. High speed imaging geometry analysis

In Fig. 8 high speed images of the weld pool surface can be Fig. 6. Geometry and its class changes along with the welding parameter changes:
seen, including also the keyhole exit opening. In the series of (a) top, (b) root.
images the focal plane position z0 and the welding speed v is var-
ied, respectively. The width (note that the camera view is 45◦ the two partner metals, which is determined by the lateral off-
inclined sidewards) of the weld pool in the images remains rather set. However, this trend is accompanied by irregular local mixture
unchanged, while the weld pool length increases with speed and variations.
shows a distinct maximum when varying z0 .
3.5. Stress analysis
3.4. Hardness analysis
The resulting weld quality, particularly the top and root shape,
Fig. 9 shows the cross section of three samples with differ- is the crucial aspect. The shape is at the same time the result from
ent lateral laser beam offset and the corresponding (Vickers) the welding process, determined by the chosen parameters, and a
hardness distribution across the weld, at a vertical location in geometrical boundary condition for the load situation of the prod-
the middle of the weld. The Weldox 960 steel is located at the uct in service. Residual stress acts as a superimposing field, but
left side, followed by the outer (softened) and inner (hardened) was neglected for the first simulations, to enable a clear judge-
HAZ, the weld metal (mixing both steel grades), the softened ment of the isolated geometry impact. The formation of the stress
Domex 700 HAZ (it does not harden) and the Domex 700 field generated by the different weld shapes has been studied by
base metal. As can be seen, the transition in hardness basically Finite Element Analysis and will be published in additional papers
depends on the proportions of melting and in turn merging of in a comprehensive manner [22]. Fig. 10(a) shows the upper left
J. Karlsson, A.F.H. Kaplan / Applied Surface Science 257 (2011) 4113–4122 4119

Fig. 7. Geometrical property changes as a function of (a) offset, (b) focal position, (c) speed, (d) angle and speed.

Fig. 8. High speed images (45◦ inclined view from the side) of the weld pool shape Fig. 9. Hardness (HV) across the weld (left: Weldox 960, higher hardness, right:
(emphasized) when altering the focal position (upper) or the welding speed (lower). Domex 700) for three laser offset positions y.
4120 J. Karlsson, A.F.H. Kaplan / Applied Surface Science 257 (2011) 4113–4122

corner of a stress field computed for the here studied beamer. For
the sake of simplicity of the load situation, a slightly inclined force
acts at the upper left corner. The simulation was carried out two-
dimensionally across the beamer, thus creating two corners where
the 90◦ -angle increases and two where it decreases. Fig. 10(a)
shows the former case (with straight top and root shapes, thus
without the experimentally obtained shapes yet), thus the root
experiences tensile stress and its peak value and location are critical
for fatigue crack initiation, while the weld top is compressive and
uncritical. The different root shapes from above were studied and
show different stress raiser impact, as will be published separately
[22]. The lowest peak stress was computed for root Class ii.
Fig. 10. Calculated stress field (von Mises stress) for three joint types under load F:
(a) 6 mm weld eccentric corner joint, (b) 6 mm weld, flat corner joint (c) full 15 mm
penetration, flat corner joint.

Fig. 11. Matrix Flow Chart (MFC, “OR”-logic) for process parameter variation on (a) defects, (b) top/root classes, (c) top/root key dimensions, (d) high speed imaging–weld
pool size, (e) horizontal hardness transition, (f) stress analysis for three joint types.
J. Karlsson, A.F.H. Kaplan / Applied Surface Science 257 (2011) 4113–4122 4121

The stress field for the same load in a flat corner joint with raiser impact under load; from clear trends unexplored param-
the same 6 mm weld penetration from the top is calculated in eter regimes can be anticipated
Fig. 10(b). The design is advantageous, but here the root is more (iii) Turning the laser inclination angle slightly suppresses
difficult to control (e.g. via a gap) and it is a stress raiser for zero undercut and root dropout, but later leads to lack of
gap, while at the top side reinforcements have to be avoided due fusion/penetration; shifting the focal position into the joint
to interaction with another beamer. The same joint, but for full causes undercut and root sagging, which can be compensated
15 mm penetration is shown in Fig. 10(c). Here the root can be by increasing the welding speed (or the inclination angle); the
shaped properly and the deeper weld causes lower stress, but is MFC guides towards recommended laser beam parameters for
then overdimensioned, as the 6 mm plate governs the peak stress achieving a favoured weld shape and hardness transition
at the root. The comparison of the three joint types was carried (iv) Formulation of trends in the here developed MFC is a gen-
out extensively. The here presented experimental results were of eralized starting point for gradual identification/claim of
importance for transferring the findings to these other joint types parameter limits and also a potential guideline for eliminat-
to discuss their expected behaviour, before starting comprehensive ing welding defects; future work includes further trends, for
experiments. different joint types and materials, added to the MFC

4. Generalized knowledge documentation


Acknowledgements
From the experiments, trends were identified, as partially dis-
cussed above. These trends are documented in the new Matrix The authors are grateful for funding by VINNOVA – The Swedish
Flow Chart (MFC) developed, as shown in Fig. 11. In Fig. 11(a–c) Governmental Agency for Innovation Systems (project LOST, no.
the resulting weld quality according to Figs. 3, 6 and 7 is shown in 2006-00563) and by the Knut and Alice Wallenberg Foundation
dependence of the process parameters varied in three different out- (Fibre Laser, project no. KAW 2007-0119). The contributions from
put formulations (defects, top/root classes, top/root dimensions). the Swedish industrial and academic partners involved are highly
Note that the “OR”-logics is applied here. The method can also appreciated.
be applied to process-internal mechanisms, as shown in Fig. 11(d)
where the weld pool dimension resulting from high speed imaging References
is related to parameter changes, according to Fig. 8. The hardness
transition character as a function of the beam offset is shown in [1] K. Nilsson, A.F.H. Kaplan, O2 -Laser/Pulsed MIG Hybrid Welding of High Strength
Fig. 11(e), related to Fig. 9, while the FEA-findings from Fig. 10 are Steel, in: E. Halmøy (Ed.), Proc. 9th NOLAMP Conference, Trondheim, Norway,
2003.
documented in Fig. 11(f). [2] Swedish Standard, SS-EN ISO 13919-1 (1996) 27.
For example when the focal plane position is shifted into the [3] Y. Kawahito, M. Mizutani, S. Katayama, Elucidation of high-power fiber laser
work piece, the defects undercut and root sagging occur, as con- welding phenomena of stainless steel and effect of factors on weld geometry,
J. Phys. D: Appl. Phys. 40 (2007) 5854–5859.
nected in Fig. 11(a) by black circles (‘no connection’ is a white [4] S. Postma, R.G.K.M. Arts, Penetration control in laser welding of sheet metal, J.
circle). These trends might be valid in a more general manner, to Laser Appl. 14 (2002) 210–214.
be proven and extended by other experts and publications. When a [5] P. Norman, H. Engström, A.F.H. Kaplan, Proc. 11th NOLAMP Conference,
Lappeenranta, Finland, Lappeenranta University of Technology, 2007.
sample has the defects undercut and root sagging it can be seen, that
[6] A.F.H. Kaplan, M. Mizutani, S. Katayama, A. Matsunawa, State-of-the-art of
by altering the inclination angle or increasing speed (“OR”-logics), monitoring and imaging of laser welding defects, J. Phys. D: Appl. Phys. 35
each can eliminate the defects, since both of the parameters got the (2002) 1218–1228.
[7] P. Norman, H. Engström, A.F.H. Kaplan, Theoretical analysis of photodiode mon-
defects in one state and none in another state.
itoring of laser welding defects by imaging combined with modelling, J. Phys.
This kind of table can be evolved by adding more trends D: Appl. Phys. 41 (2008) 195502, pp. 9.
(connections) to already placed parameters, or by adding more [8] R. Fabbro, S. Slimani, I. Doudet, F. Coste, F. Briand, Experimental study of
information/states of the parameters and defects, or by adding the dynamical coupling between the induced vapour plume and the melt
pool for Nd–Yag CW laser welding, J. Phys. D: Appl. Phys. 39 (2006) 394–
new parameters and defects to the table. Numerical informa- 400.
tion about the parameter states cannot be included, since this [9] J.-H. Cho, D.F. Farson, M.J. Reiter, Analysis of penetration depth fluctuations in
is a generalization of trends meant to work for all applications. single-mode fibre laser welds, J. Phys. D: Appl. Phys. 42 (2009) 115501.
[10] P. Sathiya, M.Y. Abdul Jaleel, Measurement of the bead profile and microstruc-
The parameter numbers will vary from case to case. The MFC tural characterization of a CO2 laser welded AISI 904 L super austenitic stainless
only shows what could be expected when changing a parame- steel, Opt. Laser Technol. 42 (2010) 960–968.
ter value. The applicability for extension of the MFC has to be [11] M.H. Cho, D. Farson, C.Y. Lim, H.W. Choi, Hybrid laser/arc welding process for
controlling bead profile, Sci. Technol. Weld. Join. 12 (2007) 677–688.
proven and studied further in future publications, encouraging [12] K. Kazemi, J.A. Goldak, Numerical simulation of laser full penetration welding,
the scientific community in welding to combine their results with Comput. Mater. Sci. 44 (2009) 841–849.
the MFC. For the present article it is hoped that for the reader [13] Y. Yao, M. Wouters, J. Powell, K. Nilsson, A.F.H. Kaplan, Influence of joint geom-
etry and fit-up gaps on hybrid laser-metal active gas (MAG) welding, J. Laser
the survey on the results is supported by the new illustrative
Appl. 18 (2006) 283–288.
method and that the results can be tracked in a more systematic [14] M.M. Alam, Z. Barsoum, P. Jonsén, H.A. Häggblad, A. Kaplan, The influence
manner. of surface geometry and topography on the fatigue cracking behaviour of
laser hybrid welded eccentric fillet joints, Appl. Surf. Sci. 265 (2010) 1936–
1945.
5. Conclusions [15] W. Fricke, Fatigue analysis of welded joints: state of development, Mat. Struct.
16 (2003) 185–200.
The following conclusions can be drawn: [16] M. Ring, W. Dahl, Fatigue properties of laser-beam weldments on the high
strength Steels, Steel Res. 65 (1994) 505–510.
[17] L.L. Martinez, Fatigue behaviour of welded high-strength steels, Doctoral thesis,
(i) Improved documentation and generalization of knowledge is Royal Institute of Technology, Stockholm, Sweden, 1997.
desired, e.g. for transferring welding results systematically; [18] S. Chakrabarti, S. Mitra, B. Bhattacharyya, Management information system
multidimensional and graphical result arrangement facilitates as knowledge base model for machining process characterisation, Int. J. Adv.
Manuf. Technol. 34 (2007) 1088–1097.
the recognition of trends [19] T.M.D. Ebbels, B.F. Buxton, D.T. Jones, Visualisation of microarray and contex-
(ii) When varying 4 geometrical laser beam parameters for laser tual bioinformatic data using spring embedding and an ‘information landscape’,
welding of an eccentric corner joint, 5 top and 5 root shape Bioinformatics 22 (2006) e99–e107.
[20] G. Robinson-Riegler, B. Robinson-Riegler, Cognitive Psychology, Pearson Edu-
categories were distinguished, having different essential stress cation, 2004.
4122 J. Karlsson, A.F.H. Kaplan / Applied Surface Science 257 (2011) 4113–4122

[21] P. Norman, Process monitoring and analysis of laser welding and of milling, [23] J. Karlsson, T. Ilar, A.F.H. Kaplan, Knowledge platform approach for fiberlaser
Doctoral thesis, Luleå University of Technology, Luleå, Sweden, 2008. (http:// welding of high strength steel, in: Proc. 12th NOLAMP Conference, Copenhagen,
epubl.ltu.se/1402-1544/2008/35/index.html). Denmark, FORCE Technology, 2009.
[22] M. Alam, A study of the fatigue behaviour of laser and hybrid laser welds, [24] A.F.H. Kaplan, Advanced systematic theory of laser welding supported by high
Licentiate thesis, Luleå University of Technology, Luleå, Sweden, 2009. speed imaging, in: Proc. XVII GCL-HPL Conference, Lisbon, Portugal, SPIE, 2008.

Das könnte Ihnen auch gefallen