See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/321225714

Thermal Modeling of 304L Stainless Steel for Selective Laser Melting: Laser
Power Input Evaluation

Conference Paper · November 2017

DOI: 10.1115/IMECE2017-72224

CITATIONS READS
13 6,091

2 authors:

Diego de Moraes Aleksander Czekanski

York University York University
5 PUBLICATIONS 46 CITATIONS 142 PUBLICATIONS 1,501 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Diego de Moraes on 24 April 2018.

The user has requested enhancement of the downloaded file.

 Proceedings of ASME 2017 International Mechanical Engineering Congress and Exposition
IMECE2017
November 3-9, 2017, Tampa, Florida, USA

IMECE2017-72224

THERMAL MODELING OF 304L STAINLESS STEEL FOR SELECTIVE LASER

MELTING: LASER POWER INPUT EVALUATION

Diego Augusto de Moraes Aleksander Czekanski

York University York University
Toronto, Ontario, Canada Toronto, Ontario, Canada
moraes@yorku.ca alex.czekanski@lassonde.yorku.ca

ABSTRACT technologies brought into the market different types of

Selective Laser Melting (SLM) process is a Powder Bed techniques using several powder alloys. The first developments
Fusion (PBF) technique, which has shown significantly growth came with Direct Metal Deposition (DMD), Laser Engineering
in the recent years. The demand for this process is justified by Net Shaping (LENS) and Laser Cladding, similar processes that
the versatility and ease in manufacturing the parts from 3D are very analogous to Thermal Spraying coating. It consists in a
models as well for the increased complexity of engineered parts powder nozzle that successive supply a selected metal powder
generated from topology or shape optimization. Automotive, through a substrate according to the desired geometry. A high
aerospace, medical and aviation industries are taking great power laser is responsible for melting the powder that is injected
advantage of this process due the unique geometry and the process repeats layer-by-layer until the final height is
characteristics found in the components. To enhance the benefits reached. These are the first AM processes known to have utilized
of SLM, a vital task is to analyze the laser power input impact on metallic alloys as raw material. [2].
the temperature distribution through the powder bed, important After the first developments successfully started to be
for posterior residual stresses analysis. The Finite Element employed in the industry, the second technology wave ascended,
Method proposed in this study is a transient thermal model, able introducing a new concept, which is mostly known as powder
to predict temperature distribution through different sections of bed technique. The first revolutionary concept named Selective
the powder bed when performing a single track of the laser Laser Sintering (SLS) came in the mid-1980s. It was patented by
scanning. Furthermore, the impact of the laser power input is Deekard C. in Texas, U.S. and it is the first to be classified as
carried out utilizing SS 304L, a low cost Stainless Steel alloy that Powder Bed Fusion (PBF) process [3]. Subsequentially, Direct
can be employed in the SLM process, in order to determine the Melting Laser Sintering (DMLS) and Selective Laser Melting
influence on the temperature distribution along the different (SLM) were patented in Germany and Finland in the 1990’s and
cross sections. Electric Beam Melting in Sweden in the 2000’s to supply the
increased demand for the technology [4,6].
Keywords: Additive Manufacturing, Selective Laser Melting, Nowadays it is common to classify SLS, DMLS, SLM and
Moving heat source, Finite Element Method, Transient thermal EBM as Powder Bed Fusion (PBF) process. The reason is the
analysis, SS 304L fact that all of them share common features, such as a bed filled
with metallic powder, a roller to distribute new layer of material
1 INTRODUCTION and a high powered electron beam or laser as heat source to
Additive manufacturing (AM) using fine metal powder achieve solidification by full or partial melting phenomena [2].
plays an important role in industrial advanced manufacturing, Many other variants of AM using metallic alloys are available in
and is considered by many the fourth industrial revolution, the market, but they are very similar to the ones described in this
extensively studied and analyzed by academia and industrial section and will not be cited in this work [5]. Since SLM is the
institutions [1]. In recent years, the rapid development of these focus of this work, the parameters and the environmental

1 Copyright © 2017 by ASME

boundaries are taken from the selected manufacturers available component, which can lead to undesired warping and cracking.
in the market, such as SLM Solutions and Concept-Laser. Presently, academia and industry perform most tests of new
powder alloys on SLM by long trial and error experiments,
1.1 Selective Laser Melting (SLM) which generate a lot of material scrap, energy consumption and
Selective Laser Melting (SLM) is a metallurgic Additive require large resources allocation [4,8].
Manufacturing process that utilizes a high powered fiber laser to Several alloys are already validated to be commercialized
selectively scan a certain region in order to melt and by the machines manufacturers, including Nickel, Cobalt,
subsequentially fuse a powder alloy distributed through a Titanium and Aluminum alloys as well as several Stainless steel
substrate. After each slice is laser scanned, a roller or wiper is alloys [9,10]. However, one of the low cost Stainless Steel
responsible for distributing a new layer of powder on the top of available in the market, SS 304L, is not yet validated. Abd-
the previous iteration. The process builds the desired component Elghany and Bourell conducted an experimental study utilizing
with a layer-by-layer fashion until the full height of the part is Concept-Laser machine with different layer thickness and
attained. An inert gas, usually Argon, is utilized to control the scanning speeds while Brown conducted another experimental
powder bed environment. Figure (1) shows the basic setup in order to determine optimal parameters to build full dense
configuration of the machine [6]. parts [6,11]. The conclusion proved that it is feasible to use this
technology for low-cost alloys components as well as to achieve
up to 70% of the standard material properties.
In this research, the main objective is to present a numerical
procedure able to perform an effective measurement of the
thermal distribution through the powder bed utilizing different
laser power input values when scanning a single track, with
considerably low computational cost. The model is
conceptualized utilizing commercially available Finite Element
Method (FEM) software ABAQUS.

1.2 Finite Element Method developments

The rapid development of multi-physics Finite Element
software in the past 20 years, such as ANSYS and ABAQUS,
made it feasible to numerically study AM techniques and account
for most different behaviors that can be inferred by the process.
Since then, transient thermal analysis is developed with diverse
Figure 1 - General SLM machine configuration [7] procedures in order to account for temperature distributions,
thermal-kinetic models are used to predict final microstructure
The main applications are aerospace, medical, automotive and properties, and thermal-mechanical models are used to
and aviation engineering, but applications are not limited for account residual stresses and deformations. The state of art
those. Because of the improvements in mechanical design, research is performed via Discrete Element Method (DEM), on
brought especially by topology and shape optimization to which the particles are modeled as a single unit, usually called
fabricate lightweight components, the complexity of the Multiphysics analysis [12].
geometry increased substantially enough to make it impossible Early studies utilized concepts from Laser Welding and
or very costly to fabricate with conventional forming or Electrical Beam Welding (EBW) because of the lack of specific
subtractive manufacturing processes, such as casting or literature available in that time [13]. Some approaches for
machining. That it is one of the main reasons why SLM is highly metallurgic AM processes that have been developed through the
recommended for these type of applications [4]. years includes the model developed by Lino et al., Vastola et al.
During the last years, this technology has been improved and Rombouts. The first one captured the kinetics phenomena
drastically, allowing it to be implemented in high end and the microstructure into ABAQUS by using several
components for most diverse Engineering applications. It is subroutines, predicting the hardness and the volume fraction of
possible, with some types of alloys, to build fully dense its micro-constituents with the simulation for Laser Cladding
components that can offer similar or even slightly improved [14]. The second accounted the residual stresses by defining
material properties compared to those built in conventional phase dependent properties inside user subroutine UEXPAN on
manufacturing processes, however they tend to be less malleable ABAQUS to keep them on track and calculate the thermal strains
due the inherited rapid solidification of the process [4]. generated by the moving heat source [15]. Finally, UMATHT
The research challenge in material science field now is how and USDFLD, another user subroutines from ABAQUS, were
to effectively determine and optimize the process parameters for used by Rombouts to define all the thermal properties according
powder alloys, since conduction, convection, radiation, phase to its state, temperature, as well for the phase change [16].
transformation and evaporation are all happening at the same Although the strong effectiveness of Multiphysics analysis,
time. This behavior generates residual stresses in the built it has a very high computational cost since it incorporates several

2 Copyright © 2017 by ASME

physics components compared to conventional thermal or beam that follow a Gaussian profile. Figure (2) shows the
thermal-mechanical analysis. This is a thermal study and there different profiles of TEMxx for lasers [18,19].
are some features not captured by it, such as the evaporation
phenomena, Marangoni effect caused by fluid dynamic of the
molten pool and kinetics of the phase changes. These cited
effects demand a Multiphysics analysis of the process, which is
not the scope of this study.

2 THERMAL MODELING
The temperature distribution of the process, described as
T(x,y,z,t), is found by solving the three dimensional heat
conduction equation (1) with suitable boundary conditions.

dT
 (T )C p (T )    q (r , t )  Q (r , t ) (1)
dt
q ( r , t )   kT (2)
Figure 2 – Different Transverse electromagnetic mode for
r  x, y, z (3) lasers [19]

There are several formulations in the literature for a moving

Where the ( ) and ( ) are the density and specific heat with heat source. The first developments in numerical simulation to
temperature T dependent values, ( , ) is the heat flux and r account for the phenomena came with Goldak et al in 1984 [20].
represent the three dimensional Cartesian coordinates, in this His model for welding processes has been widely used since
study. By substituting equation (3) into (2) and decomposing the then, with some modifications depending on the process studied.
Laplace operator, we have: Figure (3) and equation (7) and (8) express the formulation.

T   T    T    T 
C p  k   k    k Q (4)
t x  x  y  y  z  z  Heat Flux

The boundary conditions for the thermal analysis depends

on the environment inside of the chamber. In this case, besides
the conduction that is happening in the powder bed, the thermal
conduction and radiation are accounted considering the powder
emissivity, the type of inert gas and the initial temperature inside
the chamber [17]. The following equations (5-6) are utilized in
the model.

qconv  h (T  T0 ) (5)
qrad   p (T 4  T0 4 ) (6) Figure 3 - Double ellipsoid heat source model (heat source
moving through X-axis) [20]
Where ℎ is the convective heat transfer coefficient, is the  3x 2 3 y 2 3z 2 
emissivity of the surface emissivity, is the Stefan-Boltzmann  2  2  2 
6 3 f f  abs P a b c 
constant and T and T₀ are the surface and ambient temperature Q f ( x, y, z )  e  f 
(7)
[17]. Radiation and convection are important to predict the a f bc 
cooling rates inside the powder bed. In the following sections,
the proper emissivity of the powder bed is calculated. The  3x 2 3 y 2 3z 2 
 2  2  2 
implementation is done via user-defined subroutine FILM. 6 3 f r abs P a 
Q r ( x , y , z ) 
b c
e  r  (8)
a r bc 
2.1 Laser modeling as moving heat source
The laser utilized in the process is modeled following a The heat flux ̇ ( , , ) in this formulation is divided into
Gaussian profile. For the SLM process the laser used is an two parts, the rear and the front part. The representation of the
Ytterbium fiber laser, which has a TEM00 profile. Transverse ellipsoids can be visualized at figure (3), on which the heat
electromagnetic mode (TEM) is the classification used for laser source moves in the X-axis, scanning the bed with steady speed.
The front ellipsoid has parameter, and the rear has , both

3 Copyright © 2017 by ASME

expressed in meters. The parameter is the correspondent front as porosity-dependent, since each machine manufacturer
heat fraction and the rear part is represented by , both 0.6 and provides its own powder porosity. The implementation follows
1.4 respectively [20]. Parameters b and c are used to determine Sih and Barlow, which validated the formulation and it is
the width the depth of the profile, is the powder absorption expressed in the equations (10-12).
coefficient, and finally x, y and z are the coordinates from the
model. Figure (4) below show the distribution in a stationary  p  AH  H  (1  AH ) s ( 10 )
point.
0.908 2
AH  ( 11 )
1.908 2  2  1

 2
1  
 s 2  3.082  
Heat Flux (W/m²)

    
H 
2
( 12 )
 1   
 s 1  3.082    1
    

Where is the effective powder emissivity, is the bulk

emissivity of the material, is the gap emissivity, is the area
fraction of the powder surface and is the powder porosity [23].
Below is the emissivity calculated from 0% (full bulk material)
to 100% (full porous material) porosity values of SS 304L alloy.
Note that for 0% porosity the correspondent emissivity is ,
Figure 4 – Static heat flux from the fiber laser modeled as which is 0,44 in this case. Since the porosity adopted is 40%, the
Gaussian profile using Goldak formulation equivalent powder bed emissivity is ɛ=0,5652.

2.2 Phase transformation

The phase transformation has to be considered in order to
accurately capture the temperature distribution during phase
transition. The implementation of Goldak’s equivalent heat
capacity method is applied in this study, in order to capture the
phenomena. The need is justified by the intrinsic characteristics
of alloys, on which the phase change is not an isothermal
transformation, but within a range from the solidus and liquidus
temperature ( and ) taking the latent heat as consideration.
With this, the problem can be solved with a single region
problem, instead of a two region problem[21,22].

 C p (T ), T  TS
 Lf Figure 5 - Porosity-dependent emissivity for SS 304L alloy

Ceq (T )  C p (T )  , TS  T  TL (9)
 TL  TS
The emissivity of the powder it is an important factor when
 CL , T  TL
performing these type of analysis, since the radiation and thermal
conductivity of the powder bed are heavily relied on the
Where ( ) is the temperature-dependent specific heat for emissivity value.
solid state, is the specific heat for liquid state, is the latent
heat of fusion, and are the liquidus and solidus temperature 2.4 Temperature-dependent properties
respectively. The definition of proper temperature-dependent properties
is important to achieve accuracy in the results. For this study, the
2.3 Emissivity of the powder bed properties are taken from the Choong K. report on the
The emissivity of the powder surface is an essential thermophysical properties of Stainless Steel [24]. For the thermal
parameter to account for the radiation inside the chamber during analysis, the specific heat, thermal conductivity, coefficient of
the process. The surface emissivity is being applied in this work expansion, density, latent heat of fusion, solidus and liquidus
temperature are necessary to simulate the temperature behavior

4 Copyright © 2017 by ASME

during scanning. The effective thermal conductivity will be contact area between them is minimal. For this reason, a relation
further calculated according to its actual powder state at section gives the effective thermal conductivity from the solid thermal
2.5, following formulations of Wakao and Kagou [25]. However, conductivity and the powder state properties. Wakao and Kaguei
for the liquid state the thermal conductivity will remain as formulated this relationship, on which was reinforced by Sih and
determined by Choong. Figure (6-8) shows the temperature- Barlow with experimental data [23,25].
dependent properties including Specific heat, coefficient of Equation (13) gives the effective thermal conductivity for
expansion and the thermal conductivity. In further sections, the the powder bed. This will be the thermal conductivity for
remaining properties will be explored. temperatures ranging from 300 to 1703K in the model. For liquid
phase, the original values found on Choong K. publication,
ranging from 1733 to 2800K, are implemented.

  kR 
k eff  k g  (1  1   ) 1   1   (1   ).
  kg 
 
 
 
 2  B  k   k s 
.  1  g  ln 
 Bk g   2  ks   Bk 
 Bk g     g 
  1     1  
 k s  ks
   
Figure 6 - Specific heat of SS 304L versus Temperature
  
  
B 1 B 1  kR  k contact  ( 13 )
     
2  Bk g   kg  kg
 1   
ks   
    

Where is the powder bed effective thermal

conductivity, is the inert gas thermal conductivity, is the
contact area fraction between the particles in the powder bed and
represents the bulk thermal conductivity of the material. The
deformation parameter , contact thermal conductivity
and the conductivity by thermal radiation are expressed
Figure 7 - Coefficient of expansion of SS 304L versus below:
Temperature 10
1  9
B  1.25  ( 14 )
  

4 p T 3 xR
kR  ( 15 )
1  0.132 p

18k s , P  1 Pa
kcontact   ( 16 )
 ks , P  1 Pa

The deformation parameter B is calculated here as a

function of the powder bed porosity . The thermal conductivity
term to account radiation , is calculated according to the
Figure 8 – Bulk thermal conductivity of SS 304L versus emissivity of the powder , the Stefan Boltzmann constant,
Temperature
which is = 5.67 10 and the powder particle
2.5 Effective thermal conductivity of the powder bed diameter .
The powder thermal conductivity considerably differs from According to Wakao and Kaguei, the thermal conductivity
the bulk material, since there is gas filling the particles and the due contact ( ), varies according to the chamber pressure.

5 Copyright © 2017 by ASME

For low pressures the term is calculated as dependent on the increment [27]. The moving heat source is implemented using
particle diameter and the contact area, expressed as the fractional DFLUX, a user defined subroutine for ABAQUS. In order to
contact area , and the bulk conductivity [25]. However, in execute and debug the subroutine that is written in Fortran
SLM systems available in the market, it is common to utilize language, the software is linked with Intel Parallel Compiler 17.0
atmospheric pressure, and for this reason, = , and Visual Studio 2012 [28]. The main parameters utilized in the
following Equation (16) formulation [24, 25]. numerical procedure, besides the material properties previously
The contact area between particles is used to calculate the presented, are listed on the table (1).
contact fraction term , as shown by the formula (17) at figure
(9). Table 1 - Parameters for SLM simulation
Parameters Values
Latent heat of fusion, [J/Kg] 273790
Solidus temperature, [K] 1703
Liquidus temperature, [K] 1733
( 17 ) Porosity of the bed, [%] 0.4
Solid emissivity, [ ] 0.44 [29]
Powder emissivity, [] 0.5652
Preheat Temperature, [K] 300 [9]
Figure 9 - Fractional contact area as function of the contact area Absorption coefficient , [W] 0.40
and the particle diameter [25] Laser spot diameter, [µm] 50 [9,10]
Layer thickness, [µm] 75 [9,10]
Laser Power input, [W] 100, 200, 400 [9,10]
Scanning speed, v [m/s] 1.0
Average powder diameter, [µm] 50 [6]
Convection coefficient, h [W/m²K] 10 [17]
Gas thermal conduct., [W/m*K] 0.016 [17]
Fractional contact area, 0.01023 [25]
Deformation parameter, B 1.9614
Time step, t [ms] 1.5

The domain of the model is limited to 1,5mm by 1,5mm.

For this type of analysis, a very small area is sufficient to obtain
the temperature distributions and it will save computational costs
as well. The layer thickness was set to 75μm, following the
general layer thickness used. The ABAQUS model utilized a
Figure 10 - Thermal conductivity of the powder bed DC3D8 hexahedron element mesh with 25μm cube size. The
element size adopted is not considered very fine for this case of
2.6 Finite element modeling and parameters finite element analysis, however the refinement of the mesh did
The software used for the present work is ABAQUS. The not show high impact on the results, hence the presented
main feature from the software needed to perform the work is the dimensions will be adopted. In consequence of the element size,
flexibility to utilize user-defined subroutines, allowing a proper the thickness of the powder bed corresponds exactly to three
implementation of the moving heat source, and ABAQUS is elements size. The total number of elements is 10800 with
ideal for that. Some considerations, however, has to be made in correspondent 14884 nodes, proving to be a very low
order to simulate the process. For the three dimensional model computational cost to solve it numerically.
using transient thermal analysis, the distributor of ABAQUS
recommends to use the following equation (17) to determine the 3 RESULTS AND DISCUSSION
minimum time increment [27]. The simulation was performed with three set-ups, utilizing
the power input of 100, 200 and 400W and capturing the
C p temperatures through the bed. A longitudinal and a perpendicular
t  l 2 ( 18 ) cross section along with the laser scanning direction were chosen
6 k eff
in order to perform the nodal measurements.
Where , , ∆ are respectively the specific heat, density,
thermal conductivity and the element size from meshing. This 3.1 Model validation
avoid false oscillations that may appear in the boundary Our model was validated from the work performed by Parry
vicinities as well for inaccurate solutions for very small time et al., on which it was utilized Ti6Al4V for the analysis. The
comparison is necessary to match the shape of the comet by

6 Copyright © 2017 by ASME

determining properly parameters for the double ellipsoidal the temperature distribution is represented from 0 to 750µm
profile previously presented. Although this is a thermal study, (0.75mm) of the powder bed.
accounting for a proper profile helps to predict accurate values The temperature distribution is presented below for
of temperature distribution through the bed, even there are no different power inputs at the 0,75ms of the total time in the
fluidic considerations prior included. The Figure (11) shows the analysis. Three different sections were captured as well, the
comparison between both molten pool profiles. surface one represents the top of the powder bed (0 µm), -25µm
and -75µm represents the total distance in Z-axis, as shown in
the figure (16).
(1)

P=400W
v=1m/s
(2) t=0.75ms
Figure 11 - Comparison between our model molten pool profile
(1) and Parry et al. model (2)
Figure 13 - Temperature distribution for P=400W
The grey color represented on both molten pools are the liquid
material, and it will be adopted for further representation of
liquid material, in this study it is referred to the region that reach
1733K or more (Liquidus temperature).

3.2 Temperature distribution along different sections

The numerical procedure described in the previously
sections is now utilized to simulate one single track scanning P=200W
through the powder bed, utilizing power input of 100, 200 and
v=1m/s
400 Watts and speed of 1m/s. The powder bed and the coordinate
system used is showed in the figure (12) below. t=0,75ms

Figure 14 - Temperature distribution for P=200W

P=100W
v=1m/s
t=0,75ms

Figure 15 - Temperature distribution for P=100W

Figure 12 - Powder bed coordinate system of the analysis
The power input of 400 Watts induced very high
The longitudinal cross section follow X-axis and the temperatures into the model compared to the two other values.
perpendicular follows the Y-axis. Both cross sections are done Figure (16) shows a very large molten pool from that goes from
in the middle of the powder bed. Since the Y-section is simetric, the top to the bottom of the powder bed. In that region, the

7 Copyright © 2017 by ASME

temperature reached 2700K in the top surface and 1712K in the Power input of 200 Watts, showed previously at figure (17), with
bottom. the temperature distribution at figure (14), would be the optimal
parameter for this alloy, since it reached the melting point at the
NODAL
TEMPERATURE (K)
top and at -25µm of the layer. For -75µm (bottom), the maximum
temperature attained was around 1200K, which is not a problem,
but the component density would have to be evaluated to see if
it reaches desired values.
The final temperature distribution obtained is along X-axis.
In this case a specific node at the center of the powder bed is
picked and the variance of the temperature is captured through
the whole laser scan, which takes 1,5ms (time step). In this case,
in the figure (18), it is evident the large thermal gradient for
power input of 400 Watts, compared to 200 or 100 Watts. This is
an important indication to track residual stresses in this process,
since large the thermal gradients induce large thermal strains that
could lead to defects in the component.
It is also possible to identify the moment when the material
is melting, represented in the green curve (P=400W), the
temperature changing its slope between 1703 and 1733K. At
Figure 16 – Temperature profile utilizing power input of power input of 100 Watts, since it did not reach 1703K, the same
P=400W does not occur. For 200 Watts, the temperature stays almost
In the other hand, utilizing power input of 100 watts steady from 0.8ms until the end of the simulation at 1.5ms. One
of the reason is the internal heat generation during solidification
induced temperatures below the melting point of the material,
and because the small interval analyzed.
around 1200K in the top surface against 770K at the bottom. The
melting temperature is 1703K, however we cannot assume that
the material would not melt in this case, since the material is in
powder state and some of the small particles eventually would
melt. This can be verified by the study of Abd-Elghany and
Bourrel, on which they managed to produce coupons for testing
utilizing SS 304L in a 100 Watts SLM machine. One important
finding from this experimental study is when using 70µm layer
thickness, the coupons showed lower density, yield and ultimate
tensile strengh when compared to 30µm and 50µm tickeness
[11]. This is approximetally the layer thickness that we utilized
in this study (75µm), showing that the numerical procedure
match with the experimental data.

NODAL
TEMPERATURE (K)

Figure 18 – Temperature distribution at one node located on the

center of the powder bed in the transient analysis

4 CONCLUSION
The transient thermal analysis presented in this study was
able to predict temperature distribution through the powder bed
of Selective Laser Melting (SLM), a desirable information when
determining optimal power input to build components using SS
304L. Although this study is not experimentally checked to
evaluate its accuracy, it showed very close behavior on what was
found in the literature, in this case the work of Abd-Elghany and
Bourell was used, proving that the formulations adopts are ideal
for this case.
From the results presented, the power input of 200 Watts for
SS 304L proved to be optimal in order to build a full dense
Figure 17 - Temperature profile utilizing power input of component with considerably good mechanical properties.
P=200W Utilizing 100 Watts as power input could be considered an option

8 Copyright © 2017 by ASME

as well, since the temperature distribution reaching around https://slm-solutions.com/. [Accessed: 01-Apr-2017].
1200K in the surface can melt significantly amount of powder, [11] Abd‐Elghany, K., and Bourell, D. L., 2012, “Property
however it will present low density and poor mechanical Evaluation of 304L Stainless Steel Fabricated by
properties, as proved by the work of Abd-Elghany and Bourell Selective Laser Melting,” Rapid Prototyp. J., 18(5), pp.
[11]. 420–428.
For 400 Watts, the temperature distribution thorough the bed [12] Seufzer, W. J., 2014, “Additive Manufacturing Modeling
is very high, reaching values of 2700K at the surface and around and Simulation A Literature Review for Electron Beam
1700K at the bottom. This can induce high thermal stresses in Free Form Fabrication,” NASA Tech. Memo., (April), p.
the part, which can lead to cracks and warping, a factor highly 218245.
undesirable in several applications. Another phenomenon that [13] Shen, N., and Chou, K., 2012, “Thermal Modeling of
could prejudice the component when utilizing high power input Electron Beam Additve Manufacting Process - Powder
is the evaporation of the materials, which can result in losing Sintering Effects,” Proc. ASME 2012 Int. Manuf. Sci.
material and consequentially affecting the final geometry. Eng. Conf., (June 2012), pp. 287–295.
Future work include a thorough study using thermal- [14] Costa, L., Vilar, R., Reti, T., and Deus, A. M., 2005,
mechanical model to perform cooling rates and residual stresses “Rapid Tooling by Laser Powder Deposition: Process
analyses as well as the experimental work utilizing SS 304L with Simulation Using Finite Element Analysis,” Acta Mater.,
different building parameters and its effects in the final 53(14), pp. 3987–3999.
mechanical properties of the component. [15] Vastola, G., Zhang, G., Pei, Q. X., and Zhang, Y. W.,
2015, “Controlling of Residual Stress in Additive
ACKNOWLEDGMENTS Manufacturing of Ti6Al4V by Finite Element
This research is supported by Lassonde School of Modeling,” Addit. Manuf.
Engineering and NSERC/Quanser Chair in Design Engineering [16] Rombouts, M., 2006, “Selective Laser Sintering/melting
program. of Iron-Based Powders,” (september 2006).
[17] Denlinger, E. R., Jagdale, V., Srinivasan, G. V., El-
REFERENCES Wardany, T., and Michaleris, P., 2016, “Thermal
[1] Schwab, K., 2016, “The Fourth Industrial Revolution: Modeling of Inconel 718 Processed with Powder Bed
What It Means, How to Respond,” 2016 [Online]. Fusion and Experimental Validation Using in Situ
Available: Measurements,” Addit. Manuf., 11, pp. 7–15.
https://www.weforum.org/agenda/2016/01/the-fourth- [18] Andani, M. T., 2015, “Modeling , Simulation , Additive
industrial-revolution-what-it-means-and-how-to- Manufacturing , and Experimental Evaluation of Solid
respond/. and Porous NiTi.”
[2] Frazier, W. E., 2014, “Metal Additive Manufacturing: A [19] Roberts, I. A., Wang, C. J., Esterlein, R., Stanford, M.,
Review,” J. Mater. Eng. Perform., 23(6), pp. 1917–1928. and Mynors, D. J., 2009, “A Three-Dimensional Finite
[3] Deekard, C. R., 1989, “Method Abd Apparatus for Element Analysis of the Temperature Field during Laser
Producing Parts by Selective Sintering.” Melting of Metal Powders in Additive Layer
[4] Yap, C. Y., Chua, C. K., Dong, Z. L., Liu, Z. H., Zhang, Manufacturing,” Int. J. Mach. Tools Manuf., 49(12–13),
D. Q., Loh, L. E., and Sing, S. L., 2015, “Review of pp. 916–923.
Selective Laser Melting: Materials and Applications,” [20] Goldak, J., Chakravarti, A., and Bibby, M., 1984, “A
Appl. Phys. Rev., 2(4). New Finite Element Model for Welding Heat Sources,”
[5] ASTM International, 2013, “F2792-12a - Standard Metall. Trans. B, 15(2), pp. 299–305.
Terminology for Additive Manufacturing [21] Goldak, J., 1989, “Modeling Thermal Stresses and
Technologies,” Rapid Manuf. Assoc., pp. 10–12. Distortions in Welds,” Second International Conference
[6] Brown, B., “Characterization of 304L Stainless Steel by on Trends in Welding Research, Gatlinburg, p. 11.
Means of Minimum Input Energy on the Selective Laser [22] Escobar, L. E. C., 2016, “Predictive Modeling,
Melting Platform.” Simulation, and Optimization of Laser Processing
[7] Sidambe, A. T., 2014, “Biocompatibility of Advanced Techniques : Uv Nanosecond-Pulsed Laser Micro-
Manufactured Titanium Implants-A Review,” Materials Machining of Polymers and Selective Laser Melting of
(Basel)., 7(12), pp. 8168–8188. Powder Metals,” Rutgers, The State University of New
[8] Arman Ahmadi, 2016, “A Computational-Based Jersey.
Framework for Modeling the Mechanical Properties of [23] Sumin Sih, S., and Barlow, J. W., 2004, “The Prediction
the Metallic Parts Fabricated by Selective Laser of the Emissivity and Thermal Conductivity of Powder
Melting,” University of Toledo. Beds,” Part. Sci. Technol., 22(3), pp. 291–304.
[9] “Concept-Laser” [Online]. Available: [24] Kim, C. S., 1975, Thermophysical Properties of
https://www.concept-laser.de/home.html. [Accessed: Stainless Steels, Argonne National Laboratory.
03-May-2017]. [25] Wakao, N., and Kagei, S., 1982, Heat and Mass Transfer
[10] “SLM Solutions Group AG” [Online]. Available: in Packed Beds, Taylor & Francis.

9 Copyright © 2017 by ASME

[26] Kruth, J., Mercelis, P., and Vaerenbergh, J. Van, 2007, Intel Fortran Compiler And Microsoft Visual Studio (64-
“Feedback Control of Selective Laser Melting,” Proc. Bit Architecture Systems Only),” (December), pp. 1–8.
[27] “Abaqus Documentation” [Online]. Available: [29] G. G. Gubareff, J. E. Janssen, R. H. T., 1960, Thermal
http://50.16.225.63/. Radiation Properties Survey: A Review of the Literature.
[28] Nikolaos, V., 2014, “Linking Abaqus 6.13 / 6.12 With

10 Copyright © 2017 by ASME

View publication stats