INVESTIGATION a platform and a laser or electron

 

INVESTIGATION INTO THE EFFECT OF
BUILD PARAMETERS AND BUILD ANGLE ON THE DOWN SKIN SURFACES FOR STAINLESS STEEL
LASER MELTED PARTS

Aurelia Pana – 1723161

Faculty of Science and Engineering, University of Wolverhampton,
Telford, TF2 9NT, UK

 

1. INTRODUCTION INTO LASER MELTING

In the Additive Manufacturing (AM)
industry several technologies
have been developed, powder bed fusion being the leading technology for
obtaining metal parts.
Some
of powder bed fusion processes are alternately known as selective laser
sintering, selective laser melting, direct metal laser sintering, direct metal
laser melting, and electron beam melting. All powder bed fusion processes have
a similar basic operating principle.
The main differences between processes are in the way layers are deposited to
create parts and in the materials that are used. (W.E. King et al., 2015) The design
stage of the process starts by
creating 3D CAD model of the desired object. Using a pre-defined slicing
program, the 3D CAD file is converted into a series of thin parallel layers that
fully describe the geometry of the desired object creating a 2D image of each layer (Wikipedia,2015).
The data obtained is transferred to a computer controlled laser device.
Depending on the equipment and the method used, thin layers of powder are
spread with a recoating system onto a platform and a laser or electron beam is
used to fuse the powder at locations specified by the model of desired
geometry. When one layer is completed, a new layer of powder is applied and the
process is repeated until a 3D part is produced (S. Bose et al.,2017). The
applications of rapid prototyping are vast. Over the last years, AM
technologies are continuously expanding in industrial sectors like
architectural, medical, dental, automotive, furniture and jewellery, and many
others. (Metal AM, 2014)

2. THEORY AND BACKGROUND

In AM processes,
laser sintering is a technique that uses the directed energy generated by a
laser to melt particles of metal powder, layer by layer forming a solid
structure (Wikipedia, 2014).
For the experiment discussed in this paper, the laser sintering method applied
is Selective Laser Melting (SLM). SLM is a process based on powder bed fusion where metal powder is
completely melted in order to obtain dense parts (J. Jhabvala, 2010). SLM method involves a number of
steps that move from the virtual CAD description to the physical resultant
part. The first step is to create a 3D CAD and then convert the CAD file onto an
STL file format because nearly every SLM machine accepts this type of file format.
After completing these steps, using computer software the STL file will be mathematically
sliced it in 2D cross sections, representing the layers that will form the
solid structure of the part. The data obtained is converted into a readable
file for the SLM equipment and transferred to the machine that will be used to
create the physical object. The machine must be properly set up prior to the
build process. Such settings would relate to the build parameters. The primary
process parameters for SLM that are related to the quality of the part in terms
of surface finish are the scanning speed, laser power, layer thickness, hatch
distance and beam offset (Y.Tian,2017),
shown in Figure 1.

Figure 1- Process parameters

The build parameters can be standard
depending on the machine used, or either modified by entering them manually by
a trained operator. (I. Gibson
et. al.,2015) Laser power is the energy brought by the laser beam to the
powder bed and influences the melting temperature. (P.
Hanzl et. Al., 2015) Scanning
speed (speed of beam over the powder bed) has an important role in laser
sintering because it can affect the mechanical proprieties of the final part.
When increasing the scanning speed, a better quality of the surface finish can
be obtained (L.
Taimisto, 2009). Layer thickness usually influences the building time and
represents the depth of each successive addition of metal powder to the
building platform. A lower thickness of layers can decrease surface roughness (S.
Dadbakhsh, L. Hao, 2014). Therefore, all build parameters
need to be carefully chosen to avoid creating a defective part.

 

3.EXPERIMENTATION

In the following experiment two sets
of parts was manufactured and an investigation into the effects of the build
angle and downskin condition on laser melting Stainless Steel powdered
material, has been carried out. The SLM equipment used in this work is an EOS
M270 machine which has a Yb-fibre laser with a variable focus diameter 100
?m – 500 ?m and a maximum power output of 200 W (3RSystems,2014). Following
the generic steps for the SLM process, first of all, the 3D CAD model for both
sets of test samples was created using Solidworks
2017 software. In Figure 2 is illustrated the 3D CAD model and its dimensions. 

Figure 2- 3D CAD model of test samples

In Table 1 the angle dimensions (?) for
all 11 A parts and B parts are presented.

Table 1 – Built Angle parts “A” and “B”

Built
Angle (?) for parts “A” and “B”

1

2

3

4

5

6

7

8

9

10

11

18

21

24

27

30

33

36

37

39

42

45

After designing the 3D CAD models
for both sets of test samples, the CAD files ware converted onto STL files.
Using Magics Software both sets of
parts ware fixed onto a virtual platform, without the aid of support structures,
although for overhanging structures inclined from 0 to 45 degrees support
structures are needed (Y. Kajima et.al., 2017). For
the next step, the STL files for all test samples and the support files, ware sliced
and converted using EOS RP TOOLS
software in the SLI format, which in the EOS language represents the part layer
by layer. The thickness of each layer was set at 0.02 mm. After completing this
step, the files were transferred to the machine, where settings for the build
parameters and exposure strategies ware made as it follows in Table 2:

Table 2 – Built parameters test samples

Build parameters for set “A”

Build parameters for set “B”

Pre-contour

Post-contour

Pre-contour 1

Pre-contour 2

Post-contour1

Post-contour 2

Power = 40W

Power = 40W

Power = 40W

Power = 40W

Power = 40W

Power = 40W

Scan
speed = 700 mm/s

Scan speed = 700 mm/s

Scan speed = 800 mm/s

Scan speed = 1200 mm/s

Scan speed = 1600 mm/s

Scan speed = 1800 mm/s

Offset = 0.020 µm

Offset = 0.00 µm

Offset = 0.03 µm

Offset = 0.02 µm

Offset = 0.01 µm

Offset = 0.00 µm

In the exposure strategy, pre-contours and post-contours can
be given by the standard settings of the machine. Pre-contours are used to
define sections to be melted and post-contours are used to define the final
size of the component. Contours are an important aspect because the quality of surface
finish is influenced by them. (F. Caligano et. Al. 2012) In Figure 3 below, the
pre-contour is illustrated as “contour without beam offset”, and post-contour
is presented as “contour with beam offset”

Figure 3- Contour strategy

In this experiment, the contouring values for group “B” of
test samples ware modified in order to investigate if the surface finish of the
downskin surfaces will be improved when using two pre-contours and two
post-contours with different scanning speed. Once the build contours and other parameters
ware set on the EOS M270 machine, the build platform was fixed and orientated
inside the build chamber. The build chamber was pre-filled with argon gas to
protect the parts from the effects of oxidation. After the completion of the
SLM process, the test samples were cut off from the build platform using
WireEDM process. In Figure 4 both sets of test samples obtained by SLM are presented.

Figure 4 – test samples obtained by SLM

Using Guyson Euroblast 4SF dry blasting
machine, all parts were blasted with aluminium oxide and after this process,
for each test sample, the downward facing surface was analysed using the
confocal laser microscope Olympus LEXT TS 150 (Figure 5).

 

 

 

Figure 5- Analysing
surface finish

 

3.RESULTS AND ANALYSIS

Surface roughness it is characterised by the deviations in the direction
of the normal vector of a real surface from
its ideal form (G. Strano et. al., 2013). In Table 3 the average values obtained by analysing the
surface roughness using the confocal microscope are presented, where Ra is the arithmetical mean deviation
of the roughness profile. Ra can be calculated by the formula shown in Figure 6,
where Z(x)
is the deviation of surface height at x from the mean height
over the profile.

Figure 6 – Ra representation

For this experiment, all values for
the average surface roughness have been measured by the confocal laser
microscope. Knowing that the parts were designed with varying
angles of the overhang surface to determine the quality of the downward-facing
surface when changing the build parameters, and knowing that for Stainless
Steel the angle of 45? is used as the minimum build angle for the process
without support structures, the surface roughness values obtained for set A and
set B will be compared and discussed.

 

 

 

 

 

Table 3 – Ra average values
for downskin surfaces.

Part
Number

 
Overhanging
Angle (?)

 Ra for Set A

Ra for Set B

1

18

30.9338

29.8409

2

21

21.8308

22.4608

3

24

24.5165

27.3188

4

27

36.7436

25.0643

5

30

24.5165

29.126

6

33

22.1805

30.0908

7

36

14.2019

18.4023

8

37

30.1682

10.0956

9

39

9.2447

13.4115

10

42

6.4263

13.7339

11

45

14.9983

11.5549

 

For part number A1 and B1, Ra values are approximately the
same and the aspect of the overhang surface built at 18? is poorly. See images
from Figure 7 below:

Figure 7– Downskin surface Part A1 and B1

Part A1

Part B1

 

In the chart
below, Figure 8 a profile of the surface roughness for part A1 and B1 is
drowned.

Figure 8- Profile of surface roughness for
part A1 and B1

Looking at Figure 4, a slight
improvement can be observed regarding the quality of surface roughness when
increasing the overhang angle. However, even when modifying the build
parameters for contours, for the overhang angles that are lower than 30 degrees
the quality of surface finish is still unsatisfying. An image of parts A5 and
B5 with the overhanging angle of 30 degrees is presented in Figure 9.

Figure 9 – Surface finish part A5 and B5

In the chart from Figure 10, a
profile of the surface roughness for part A5 and B5 is showing that there is a
there is a small difference between the profiles created. For set B, even if
the average Ra values are higher (see Table 3), the quality of the surface
seems to improve.

Figure 10-
Profile of surface finish for part
A5 and B5

Looking at Table 3, the Ra average
value for part A8 is considerably increased compared to Ra average value for B8
with the overhanging build angle at 37 degrees. In figure 11 the profile of the
surface for both parts is highlighting the fact that for the prescribed dimensions
of sample A8 and B8, set B has a better quality of the downskin surface where
two pre-contours and two post contours ware applied.

 

 

Figure 11-Profile of surface finish for part
A8 and B8

In Table 4, the quality of the
surface finish can be compared by simply visualising the images showed. The
differences between set A and set B are clearly visible.

Table
4 – Images of set A8 to A 11 and B 8
to B 11

Comparing the Ra average values for
sample A11 and B11 (Table 3), the difference between them is insignificant. In
Figure 12 the profile of surface finish shows that the quality of downskin
surface for both sets is almost similar with small differences of Ra values.

Figure 12- Profile defining surface finish for part A11 and B11

Therefore, it
is revealed that surface roughness by SLM can be varied by modifying the
processing parameters.

4.DISCUSSION

Analysis
of Ra average values for both set of test samples in relation with build angle
(?) of the downward facing surfaces, can be seen in Figure 13, which shows a
clear dependence of surface angle on Ra. As ? increases, the value of Ra
decreases.

Figure13
– Ra in relation with ?

The variable parameters for contouring as the offset of the beam,
and the scanning speed function of the laser ware modified for set B, and the
standard settings ware used for set A. Looking at Figure 13, for set B when the
SLM process was optimised by changing the contouring parameters, the average
values for Ra are variating between 10µm-30µm, and for set A Ra values are
between 7µm-37µm. Even though the difference between both sets of parts it is
not enormous in terms of surface finish, modifying the exposure type by adding
one more pre-contour and a post-contour, seems favourable when increasing the
scanning speed and laser beam offset. A study made by F.Caligano (2014) reveals
the importance of using support structures for overhanging surfaces in order to
avoid  “staircase effect” of angled walls
and surfaces. It was discovered that for surfaces with an angle smaller than 30?,
the staircase effect tended to increase. The minimum orientation of the
overhang surface that provided an acceptable quality of surface finish, was the
orientation at 45?.The experiment was carried out on two types of material AlSi10Mg
and Ti6Al4V alloys. Referring to Figure 13, the minimum value of Ra is obtained
at the 36? angle for set B, and for set A at 39?. For both sets of test
samples, surface roughness seems to remain stabilized. A good quality of the
downward-facing surfaces is difficult to obtain by SLM process without the aid
of supports. Introducing support structures for the overhanging surfaces will
make the process less cost-effective, however the surface that has the lowest requirement for surface
finish can be considered as the bottom surface when orientating and fixing the
part in the design stage of the process, since the surface attached to the support
structure will have an increased roughness after removal of the support
structure (K. Zeng 2015).

 

5.CONCLUSIONS

An investigation into the effect of
build angle and downskin condition obtained in SLM using Stainless Steel has
been presented to further the understanding of the relationship between surface
roughness and process parameters. It was found that for Stainless Steel,
considering ideal the Ra average value obtained at 45?, and comparing it with
all the other values, at the standard parameters settings, the minimum angle values
for overhanging surfaces that can be considered acceptable for SLM process
without the use of supports are ranging between 39?-45?. When the process
parameters ware modified, the minimum angle values for overhanging surfaces are
ranging between 36?-45?.To conclude, for overhanging surfaces with an angle
below 36? support structures are needed. The orientation of the
part must be considered in the design stage in order to reduce the number of
supports and avoid damaging the quality of the surface after their removal.

 

 

 

 

6. REFERENCESS

1. F.
Calignano, D. Manfredi, E. P. Ambrosio, L. Iuliano, P. Fino (2012) Influence
of process parameters on surface roughness of aluminium parts produced by DMLS,
Accessed
on 26th November 2017,Available at < http://porto.polito.it/2505584/2/2505584.pdf>

2.
F.Caligano (2014), Materials
& Design, Design optimization of supports for overhanging
structures in aluminum and titanium alloys by selective laser melting, 64, pp
203-213

3. G.Strano, L. Hao, R.M. Everson, K.E. Evans
(2013), Journal of Materials Processing Technology, Surface roughness analysis, modelling and prediction in selective laser
melting, 213, pp. 589-597

4. I. Gibson, D. Rosen, B. Stucker (2015),Additive Manufacturing
Technologies,3D Printing, Rapid
Prototyping, and Direct Digital Manufacturing, Second Edition, ISBN 978-1-4939-2113-3,pp
3-5

5. J. Jhabvala, E. Boillat, T. Antignac, R. Glardon (2010)
Virtual and Physical Prototyping, On the
effect off scanning strategies in selective laser melting process,5,pp. 99-109

6. K.
Zeng (2015) Optimization of support
structures for selective laser melting, Accessed on 28th
November 2017, Available at < https://ir.library.louisville.edu/cgi/viewcontent.cgi?article=3250&context=etd  >

7. L.Taimisto
(2009), Bachelor’s thesis, Process
parameters in laser sintering process, Accessed on 24th November
2017, Available at < https://www.doria.fi/bitstream/handle/10024/70831/nbnfi-fe201109082361.pdf?sequence=3 >

8. Metal AM (2014), Applications for metal
Additive Manufacturing technology, Accessed on 15th November 2017, Available at < http://www.metal-am.com/introduction-to-metal-additive-manufacturing-and-3d-printing/applications-for-additive-manufacturing-technology/ >

9. P. Hanzl,
M. Zetek,
T. Bakša,
T. Kroupa (2015),Procedia Engineering, The Influence of Processing Parameters on
the Mechanical Properties of SLM Parts,100, pp.1405-1413

10. S. Bose,
D. Ke,
H. Sahasrabudhe,
A. Bandyopadhyay (2017), Progress in Materials Science, Additive Manufacturing of Biomaterials, 92, pp 45-111

11. S. Dadbakhsh, L. Hao (2014),The
Scientific World Journal ,Effect of Layer
Thickness in Selective Laser Melting on Microstructure of Al/5?wt. % Fe2O3
 Powder Consolidated Parts, 2014.

12. W. E. King, A. T.
Anderson, R. M. Ferencz , N. E. Hodge, 
C. Kamath, S. A. Khairallah and A. M. Rubenchik (2015),Applied Physics
Reviews, Laser
powder bed fusion additive manufacturing of metals; physics, computational, and
materials challenges,2,pp
1-3

13. Wikipedia (2015) Selective
laser melting, Accessed on 12 November 2017, Available at < https://en.wikipedia.org/wiki/Selective_laser_melting  >

14. Wikipedia
(2013) Sintering, Accessed on 21st
November 2017, Available at < https://en.wikipedia.org/wiki/Sintering >

15. Y.Tian, D.Tomus,P. Rometsch, X. Wu (2017), Additive Manufacturing, Influences of processing parameters on
surface roughness of Hastelloy X produced by selective laser melting,13,pp
103-112

16.Y. Kajima,
A. Takaichi,
T. Nakamoto,
T. Kimura,
N. Kittikundecha,Y. Tsutsumi,
N. Nomura,
A. Kawasaki,
H. Takahashi,
T. Hanawa (2017), Journal of the Mechanical
Behaviour of Biomedical Materials, Effect of adding support structures
for overhanging part on fatigue strength in selective laser melting,78,pp. 1-9

17. 3R Systems 2014, EOSINT
M270 – metal,
Accessed on 24th November 2017, Available at < http://www.3rsystems.cz/en/3d-print/3d-print-dental/eosint-m270-metal >