materials

Article
Discrete-Event Simulation Thermal Model for
Extrusion-Based Additive Manufacturing of PLA
and ABS
Sunil Bhandari 1,2, * and Roberto A. Lopez-Anido 1,2
1 Advanced Structures and Composites Center, University of Maine, Orono, ME 04469, USA; rla@maine.edu
2 Department of Civil and Environmental Engineering, University of Maine, Orono, ME 04469, USA
* Correspondence: sunil.bhandari@maine.edu; Tel.:+1-207-581-2402




Received: 10 October 2020; Accepted: 3 November 2020; Published: 5 November 2020


Abstract: The material properties of thermoplastic polymer parts manufactured by the extrusion-based
additive manufacturing process are highly dependent on the thermal history. Different numerical
models have been proposed to simulate the thermal history of a 3D-printed part. However, they are
limited due to limited geometric applicability; low accuracy; or high computational demand. Can the
time–temperature history of a 3D-printed part be simulated by a computationally less demanding,
fast numerical model without losing accuracy? This paper describes the numerical implementation
of a simplified discrete-event simulation model that offers accuracy comparable to a finite element
model but is faster by two orders of magnitude. Two polymer systems with distinct thermal
properties were selected to highlight differences in the simulation of the orthotropic response and
the temperature-dependent material properties. The time–temperature histories from the numerical
model were compared to the time–temperature histories from a conventional finite element model
and were found to match closely. The proposed highly parallel numerical model was approximately
300–500 times faster in simulating thermal history compared to the conventional finite element model.
The model would enable designers to compare the effects of several printing parameters for specific
3D-printed parts and select the most suitable parameters for the part.

Keywords: fast thermal simulation; temperature-dependent properties; thermal history

1. Introduction
Extrusion-based additive manufacturing of thermoplastic polymers is a thermally driven process.
Thermal history affects viscoelastic deformation [1–3], bonding [4–8], and residual stresses [9,10].
Consequently, dimensional accuracy and the strength of the manufactured part are driven by the
thermal history of the part.
Wang et al. [11] analyzed the strengths and weaknesses of different polymer-based additive
processes including extrusion-based additive manufacturing. Tan et al. [12] reviewed the state of
the art of commodity, engineering, and high-performance polymers used in additive manufacturing.
Polymers undergo melting, thermal transitions, and solidification during the extrusion-based additive
manufacturing process. Tan et al. [12] highlighted the different mechanical, rheological, and thermal
properties that affect the additive manufacturing processability and the 3D-printed part properties.
Several polymers have been successfully used in the extrusion-based additive manufacturing process.
Initially amorphous polymers were preferred for additive manufacturing including acrylonitrile
butadiene styrene (ABS) [13–16], polyethylene terephthalate glycol (PETG) [8,17], and polycarbonate
(PC) [18–20]. The thermal and rheological properties of semi-crystalline polymers caused difficulties in
filament production and showed higher shrinkage and warping during the 3D printing process [21,22].

Materials 2020, 13, 4985; doi:10.3390/ma13214985 www.mdpi.com/journal/materials

Materials 2020, 13, 4985 2 of 21

Polylactic acid (PLA) [23–26], despite being a semi-crystalline polymer, is widely used in extrusion-based
additive manufacturing, as the PLA crystallizes slowly and consequently shrinks and warps less
than other typical semi-crystalline 3D-printed polymers. Recently, however, several semi-crystalline
polymers such as polyamides (PA) [27,28], polyether ether ketone (PEEK) [29–31], and polypropylene
(PP) [21,32,33] have been used for the extrusion-based 3D printing process. Suitable numerical models
enable designers to account for the different properties of the polymers during the 3D printing process
and optimize the design of 3D-printed parts [12,34–36].
Different numerical models have been proposed by researchers to simulate the thermal history
of 3D-printed parts. Sun et al. [5] used a mathematical model to approximate the temperature of
3D-printed parts to model bond forming between acrylonitrile butadiene styrene (ABS) filaments.
Wang et al. [37] used a simplified mathematical model to simulate the temperature and the subsequent
warp deformation in extrusion-based 3D-printed ABS parts. These simplified layer-by-layer models
have low accuracy and the results can only be interpreted qualitatively.
Compton et al. [38] discussed a 1D transient thermal model to describe a build process and
analyze warping and cracking in thin-walled structures. The thermal model was solved by using a
finite difference method that calculated the temperature at the nodes at each time step. Zhang et al. [39]
used an adaptable, boundary adjusting finite difference method to simulate the thermal history of a
3D-printed polylactic acid (PLA) part. Stockman et al. [40] presented a thermal model tailored for
additive manufacturing that was based on the 3D finite difference method. The researchers used
coarse meshing in time and space along with simplifying assumptions about the solidification process.
The finite difference scheme-based models work well for simple geometries such as thin-layered walls
and rectangular cuboids. However, for more complex geometry parts that extrusion-based 3D printing
usually produces, a method that can account for changes in geometry is necessary.
Finite element analysis (FEA) modeling has been used to simulate the thermal history of 3D-printed
parts with complex geometry. Ji and Zhou [41] used a finite element model that accounted for
temperature-dependent material properties. D’Amico and Peterson [42] described an adaptable
FEA model capable of simulating heat transfer in 3D and at sufficiently small time scales to capture
rapid cooling. El Moumen et al. [43] discussed a 3D thermomechanical model that simulates the 3D
printing process using FEA. Zhou et al. [44] described a finite element based on element activation to
model the thermal history of a 3D-printed part. Zhou et al. [45] described a voxelization-based finite
element simulation to simulate the thermal history of 3D-printed parts. Brenken et al. [3] used FEA
modeling to simulate the thermal history, final deformed shape, and residual stresses in the 3D-printed
short-carbon-fiber-reinforced ABS polymer. Finite difference methods and finite element methods
solve systems of linear equations for each time step during the period of simulation. Such models slow
down non-linearly as the size of the part to be simulated grows.
Heat flow through a body in the form of a partial differential equation is shown in Equation (1).
Finite element methods and finite difference methods approximate a solution to the heat equations at
discrete locations by solving a system of linear equations.

∂T ∂ ∂ ∂ ∂ ∂ ∂
!
ρcp − kx + ky + kz T=Q (1)
∂t ∂x ∂x ∂y ∂y ∂z ∂z

where,

ρ = density;
cp = the specific heat capacity;
T = temperature;
kx , k y , and kz = conductivity in the x, y, and z directions, respectively;
Q = heat flow.

During 3D printing, the boundary of the part changes with each new deposition of beads.
As a result, finite element methods and finite difference methods need to update the heat capacity
Materials 2020, 13, 4985 3 of 21

matrix, the conductivity matrix, and the boundary conditions at each time step. Furthermore,
non-linearities such as temperature-dependent conductivities and specific heat capacities require
iterative solution of the system of linear equations at each time step. As a result, a system of
linear equations-based solutions to the heat equation is slow and computationally expensive for
simulating heat transfer during the 3D printing process. Some simplifications have been proposed
for speeding up the simulations. McMillian et al. [46] reduced the 3D model to a 1D model by
using geometric simplifications to create a computationally efficient finite difference method for
metal additive manufacturing. Zhang and Shapiro [47] proposed a linear-time thermal simulation of
as-manufactured fused deposition modeling components. They used the concept of “active body” to
update the temperatures only in the elements that have been recently activated and using analytical
equations to solve the temperature of “inactive” elements with “lazy updates”. The research work
considered the thermal properties of ABS as constants for the simulation.
A phenomenological approach has been adopted in this paper to develop the thermal model for
extrusion-based additive manufacturing. The phenomenological observations, are:
The geometry of the part changes rapidly during 3D printing. For desktop-scale printing on
extrusion-based 3D printing, a bead deposition speed of 30–60 mm/s is commonly used [48]. It amounts
to a material deposition rate of 30–60 cm3 /h. For large-scale 3D printers, the deposition rates can reach
230 cm3 /h [49]. The high rate of deposition continuously changes the geometry of the 3D-printed parts.
Heat flow from one region of a part to another region of the same part is not instantaneous, i.e.,
heat transfer takes time. Polymers have low thermal diffusivity, which is a measure of the rapidity of
heat propagation through a material [50].
The deposited element cools down quickly to reach the temperature of the surrounding
environment (approximately 8 s for ABS 400 material [5,51]) and deposition in a distant layer
has a negligible effect on the thermal history of an element.
Based on the aforementioned observations, a solution to the heat transfer problem that examines
local exchange of heat between neighboring regions at small time steps can be formulated. This research
work treats deposition of a small bead of material as a discrete event. At small time steps, the deposited
section of bead exchanges heat with its immediate neighbor via conduction, convection, or radiation.
Temperature is considered static for the elements in a layer that are far away from the layer being
actively deposited.
The objectives of the research presented herein are listed below. A model that meets these
objectives would allow designers to quickly compare the effects of different printing parameters on the
thermal history of a 3D-printed part and allow selecting the most suitable printing parameter for a
given part.
Develop an efficient numerical model for thermal discrete-event simulation (DES) that generates
the part geometry and mesh by interrogating the G-code, and uses parallel computation for fast
analysis of extrusion-based additive manufacturing.
Simulate material orthotropy in 3D-printed parts by using information on the orientation of
deposited beads from the G-code.
Capture the temperature-dependent response of the material in the simulation of the thermal
behavior of 3D-printed polymers.
Verify the accuracy of the proposed numerical model by comparing it with results from a finite
element model using a commercial FEA package (Abaqus).

2. Materials and Methods

2.1. Materials
PLA and ABS (3DXTech, Grand Rapids, MI, USA) were selected for this study. The polymers were
chosen because they are widely used polymers used in the extrusion-based 3D printing process and
represent two classes of polymers used in additive manufacturing, i.e., semi-crystalline and amorphous.
Materials 2020, 13, 4985 4 of 21

PLA is a semi-crystalline material with a specific heat capacity that is dependent on temperature,
with sharp changes at melting, cold crystallization, and glass transition temperatures [52,53]. ABS is an
amorphous polymer
Materials 2020, with
13, x FOR a fairly
PEER REVIEWconstant specific heat capacity [54]. Differential scanning calorimetry
4 of 21

carried out during a previous study [8] was used for the specific heat capacity of PLA.
scanning calorimetry carried out during a previous study [8] was used for the specific heat capacity
of PLA. Model
2.2. Numerical
TheNumerical
2.2. followingModel
assumptions were used for the numerical model:

• HeatThe following
exchange assumptions
occurs betweenwere used for the
neighboring numerical
elements model:
only, for a small time step.
• •
The effect of radiation
Heat exchange is not
occurs considered.
between neighboring elements only, for a small time step.
• • The resistance
Contact effect of radiation is not considered.
is not considered in calculations involving thermal conductivity.
• Contact resistance is not considered in calculations involving thermal conductivity.
Figure 1 shows the flowchart of the numerical model developed for this study.
Figure 1 shows the flowchart of the numerical model developed for this study.

Figure
Figure 1. 1. Flowchartofofthe
Flowchart thenumerical
numerical model
modelused
usedinin
this study.
this study.

The model takes the G-code from the slicing software as input. The G-code is the sequence of
machine instructions that directs the actions of the 3D printers. The G-code is generated by a slicing
Materials 2020, 13, 4985 5 of 21

Materials 2020, 13, x FOR PEER REVIEW 5 of 21

The model takes the G-code from the slicing software as input. The G-code is the sequence
of machine
software usinginstructions that directs
the geometrical the actions
model and theof printing
the 3D printers.
parameters TheasG-code
input.is For
generated by a
this study,
slicing software
Simplify3D using
version the
4.1.2 geometrical
was modelthe
used for slicing and the printing
geometrical parameters
model. as input.
Simplify3D For this
was used study,
because it
Simplify3D
is version
a widely used 4.1.2software,
slicing was usedenables
for slicing the geometrical
modification model.printing
of different Simplify3D was usedand
parameters, because
allowsit
is a widely used slicing software, enables modification of different printing parameters,
for the visualization of the printing process. The model reads the G-code and generates the movement and allows for
the visualization
segments, of theofprinting
the speed movement process. The model
for each segment,readsandthewhether
G-code theand extrusion
generates istheonmovement
for each
segments, the speed of movement for each segment, and whether the extrusion
segment. Table 1 shows the conversion of a section of the G-code output to a movement segment. is on for each segment.
X,
Table
Y and1Zshows
are thethe conversion
G-codes of absolute
for the a section position
of the G-code
of theoutput
X, Y, andto aZmovement
axes. G1 issegment.
the G-codeX, Yinstruction
and Z are
the linear
for G-codes for the absolute
interpolation. F is position
the G-code of the
forX, Y, andextruder
setting Z axes. G1 is the
speed G-code
[55]. Eachinstruction
movementfor linear
segment
interpolation. F is the G-code for setting extruder speed [55]. Each movement
consists of an initial point, a final point, the movement speed, and information about whether segment consists of an
initial point,
extrusion a final
occurs forpoint, the movement speed, and information about whether extrusion occurs for
the segment.
the segment.
Table 1. Conversion from a G-code to a movement segment.
Table 1. Conversion from a G-code to a movement segment.
Movement Segment
G-code X1 Y1 Z1 X2 MovementY2 Segment
Z2 Speed
G-code Extrusion
(mm) (mm) (mm) (mm) (mm) (mm) (mm/min)
X1 Y1 Z1 X2 Y2 Z2 Speed
G1 X0 Y0 F4800 Extrusion
(mm) (mm) (mm) (mm) (mm) (mm) (mm/min)
G1 Z0.400 F1002 0.000 0.000 0.000 0.000 0.000 0.400 1002 off
G1Y89.347
G1 X89.347 X0 Y0 F4800
F4800 0.000 0.000 0.400 89.347 89.347 0.400 4800 off
G1Y89.347
G1 X110.653 Z0.400 E2.8347
F1002 0.000 0.000 0.000 0.000 0.000 0.400 1002 off
G1 X89.347 89.347 0.00089.3470.0000.4000.400 110.653
89.347 89.347
89.347 0.400 1300 on
F1300 Y89.347 F4800 0.400 4800 off
G1 X110.653 Y89.347 E2.8347
G1 X110.653 Y110.653 E2.8347
F1300
110.653
89.347
89.347
89.347
0.400
0.400 110.653
110.653
89.347
110.653
0.400
0.400
1300
1300
on
on
G1 X110.653 Y110.653 E2.8347 110.653 89.347 0.400 110.653 110.653 0.400 1300 on

In Table 1, X1, Y1, and Z1 are the coordinates of the initial point of the movement segment. X2,
In Table
Y2, and 1, X1,
Z2 are theY1, and Z1 areof
coordinates thethe
coordinates
final point of of
thethe
initial point of the
movement movement
segment. The segment. X2, are
coordinates Y2,
and Z2 are the coordinates of the final point of the movement segment. The coordinates
absolute coordinates with reference to the machine axes of the 3D printer. Figure 2 shows the are absolute
coordinatessegment
movement with reference to the machine
with reference axis toaxes of the 3Daxes.
the machine printer.
TheFigure
initial2point
showsofthe
themovement
segment issegment
marked
with reference axis to the machine axes. The initial point of the segment is marked as (X1,
as (X1, Y1, Z1) and the final point of the segment, where the printer tool head is located, is marked Y1, Z1) and
as
the final point
(X2, Y2, Z2). of the segment, where the printer tool head is located, is marked as (X2, Y2, Z2).

Figure 2. Movement
Movement segment
segment coordinates
coordinates with reference to machine axes.

The mesh for the DES model described in this paper includes additional information that is not
available in
in typical
typicalmeshes
meshesused
usedininfinite
finiteelement
elementmodels.
models.In In
finite element
finite models,
element the the
models, mesh usedused
mesh has
nodes
has defining
nodes the coordinates,
defining and connectivity
the coordinates, defining
and connectivity the elements.
defining The mesh
the elements. used for
The mesh this
used study
for this
has, inhas,
study addition to nodes
in addition toand connectivity,
nodes the information
and connectivity, about adjoining
the information elements,elements,
about adjoining called neighbor
called
neighbor information for this study. For an element, the information about whether the faces are
exposed to the boundary or connected to another element is the neighbor information for that
element. For this study, hexahedral elements (brick elements) were considered with six faces. Figure
Materials 2020, 13, 4985 6 of 21

Materials 2020, 13, x FOR PEER REVIEW 6 of 21

information for this study. For an element, the information about whether the faces are exposed to the
3boundary
shows a configuration
or connected toofanother
elements in relation
element is thetoneighbor
their neighbors.
information Element 0 iselement.
for that hidden in ForFigure 3 on
this study,
the left. Figure 3 on the right is an exploded view of Figure 3 on the left so
hexahedral elements (brick elements) were considered with six faces. Figure 3 shows a configuration that each element in the
mesh can be in
of elements viewed.
relationFortoelement 0, which isElement
their neighbors. at the center, the neighbors
0 is hidden in Figure at the
3 onfront and the
the left. back3are
Figure on
elements 1 and
the right is 2, respectively.
an exploded view ofSimilarly,
Figure 3 onthetheneighbors at the
left so that eachright, left, top,
element in theand
meshbottom
can beareviewed.
4, 3, 5,
and 6, respectively.
For element 0, which Element
is at the0 exchanges
center, the heat with elements
neighbors at the front1, 2,and
3, 4,the
5, back
and 6are viaelements
conduction (and
1 and 2,
radiation if radiation
respectively. Similarly, is the
considered).
neighborsElement 0 does
at the right, not
left, have
top, andany face exposed
bottom are 4, 3, 5,toand
the boundary and
6, respectively.
does not0 exchange
Element exchangesheat heat with the environment
with elements 1, 2, 3, 4, via convection.
5, and Considering
6 via conduction element if2,radiation
(and radiation its only
neighbor is element
is considered). Element0 at0the
does front
notface.
haveElement
any face2exposed
exchanges heat
to the with element
boundary 0 vianot
and does conduction.
exchange
Element
heat with2the loses heat to thevia
environment environment
convection.via convectionelement
Considering (and radiation
2, its onlyif radiation
neighbor is is element
considered) 0 at
through the back, left, right, top, and bottom surfaces.
the front face. Element 2 exchanges heat with element 0 via conduction. Element 2 loses heat to the
The neighbor
environment information
via convection (and changes
radiationwith each newiselement
if radiation deposited
considered) through andtheneeds
back, to beright,
left, updatedtop,
once an element
and bottom is deposited.
surfaces.

(a) (b)
Figure 3. AA configuration
configuration of
of elements
elements showing
showing neighbors. (a)
(a) elements
elements on
on DES
DES model
model and
and the
neighbors; (b) exploded view of the elements and neighbors.

The neighbor
Based information
on the deposition changes
path, with eachtime
the deposition newiselement
generated deposited
for eachand needsFigure
element. to be updated
4 shows
once an element is deposited.
the process for calculation of the element deposition time and the element orientation. Each
Based on
movement the deposition
segment is divided path,
intothe deposition
points. time is
The points generated
mark for each
the physical element.
location of Figure 4 shows
the deposition
the process for calculation of the element deposition time and the element orientation.
head during the 3D printing process. The points are created in a way that ensures that the spacing Each movement
segment two
between is divided into points.
consecutive pointsTheis a points
length mark the physical
less than the length location of the deposition
of an element. The time at head during
which the
the 3D printing
movement headprocess.
is at eachThe points
point are created
is calculated as in
tp. aInway that4d,
Figure ensures that
the red dotsthemark
spacing between two
the movement of
consecutive points is a length less than the length of an element. The time at which
the deposition head. A square box with a width equal to the bead width, as shown in Figure 4d, is the movement head
is at each
used point
to find is calculated
elements that as
aretpdeposited.
. In Figure 4d, thecenter
If the red dotsof mark the movement
the element is inside of the
the deposition
square box,head.
the
A squareisbox
element with a width
considered equaland
deposited to the
thebead width, time
deposition as shown
for the inelement
Figure 4d,
is tis usedorientation
p. The to find elements
of the
that are is
element deposited.
calculatedIfasthe thecenter of the representing
unit vector element is inside the squaresegment.
the movement box, theElement
element 0isisconsidered
deposited
deposited
at and element
time t0, and the deposition time for at
1 is deposited thetime
element
t3. is tp . The orientation of the element is calculated as
the unit
Thevector
choicerepresenting
of the element the movement
dimensionssegment.
is made Element
such that0 isthedeposited at time
process of beadtdeposition
0 , and element
can1beis
deposited at
simulated time t3 . Each element has a length and width equal to the bead width of the 3D-printed
correctly.
part. The height of the element equals the bead height of the 3D-printed part.
Materials 2020, 13, x FOR PEER REVIEW 7 of 21
Materials 2020, 13, 4985 7 of 21

Figure 4. Calculation of the element deposition time and orientation. (a) movement segment generated
from
FigureG-code; (b) top-view
4. Calculation of theof element
hexahedral elements;
deposition (c) and
time movement segment
orientation. (a) divided
movementintosegment
points;
(d) elementfrom
generated deposition.
G-code; (b) top-view of hexahedral elements; (c) movement segment divided into
points; (d) element deposition.
The choice of the element dimensions is made such that the process of bead deposition can be
simulated
After correctly. Each time
the deposition element hasthe
for all a length andhave
elements width equal
been to the bead
calculated, the width of the 3D-printed
first element is added to
part. The height
the numerical of theThe
model. element equals the
first element hasbead height ofboundary
a conductive the 3D-printed part. face and a convective
at the bottom
After the
boundary deposition
at the time right,
front, back, for allleft,
the elements have been
and top faces. Untilcalculated,
the time whenthe first
theelement is added
next element is
to the numerical model. The first element has a conductive boundary at the bottom
deposited, for a time increment dt, heat losses to the environment due to convection are calculated face and a
convective boundary
using Equation (2). at the front, back, right, left, and top faces. Until the time when the next element
is deposited, for a time increment dt, heat losses to the environment due to convection are calculated
using Equation (2). 𝑑𝑄 = ℎ · 𝐴 · (𝑇 − 𝑇 ) · 𝑑𝑡 (2)
where, dQ = h·A·(T − Tenv )·dt (2)
𝑑𝑄 = heat loss to environment via convection,
where,
ℎ = convective heat transfer coefficient,
dQ =𝐴heat lossof
= area tothe
environment via convection,
surface exposed to the environment,
𝑇 = temperature
h = convective at the coefficient,
heat transfer centroid of the element,
𝑇 of= the
A = area temperature of the environment,
surface exposed and
to the environment,
𝑑𝑡 = time increment.
T = temperature at the centroid of the element,
Tenv =After
temperature of the
adding the environment,
next element to and
the model, the neighbor information is updated for all the
dt = time increment.
elements in the model. Conduction of heat from one element to the other is calculated using Equation
(3).
After adding the next element to the model, the neighbor information is updated for all the elements
in the model. Conduction of heat from one (𝑇 −to𝑇the
) other is calculated using Equation (3).
𝑑𝑄 = 𝑘 ·element
𝐴· · 𝑑𝑡 (3)
𝑑𝑥
where, (T1 − T2 )
dQ = k·A· ·dt (3)
𝑑𝑄 = heat conducted to another element, dx
Materials 2020, 13, 4985 8 of 21

where,

dQ = heat conducted to another element,

k = conductivity,
A = area of the conducting surface,
T = temperature at the centroid of the element,
Tenv = temperature of the environment,
dt = time increment.

The sum of heat exchanges through all surfaces and the new temperature of the element are
calculated using Equations (3)–(6).
X
dQtotal = dQ (4)

Qold = Told ·ρ·Cold ·V (5)

Qnew = Qold − dQtotal (6)

Qnew
Tnew = (7)
ρ·Cnew ·V
where,

dQtotal = total heat change from all surfaces,

Qold = total heat in the element during previous time step,
Told = temperature at the centroid of the element during the previous time step,
ρ = density of the element material,
C = the specific heat capacity of the element material at the old temperature,
V = volume of the element,
dt = time increment,
Qnew = total heat in the element during this time step,
Tnew = temperature at the centroid of the element after current time step, and
Cnew = the specific heat capacity of the element after this time step.

Equation (7) is iterative and uses specific heat capacity versus temperature data for the element.
Mass (volume times density) is assumed to be constant during the process.
Equation (8) shows the heat transferred through six different faces. The heat exchanged by an
element with its neighbors is calculated using Equation (8). Equation (8) accounts for heat transferred
via conduction. If any of the faces of the element is exposed to the environment, Equation (2) is used to
account for heat exchange with the environment via convection for that face.

T−T f
 
dx A f
    

 dQ f 
  kxx 0 0 0 0 0 





   T−Tl 
dy Al
    



 dQl 


 
 0 k yy 0 0 0 0 







   T−Tb 
dQb 0 0 kxx 0 0 0
dx Ab
   
  
  
 =  dt (8)
  

dQr

0 0 0 k yy 0 0
 T−Tr 
dy Ar

   

 
 
   


 dQt  0 0 0 0 kzz 0 
T−Tt


 
dz At

 
   
  
 

dQd   0 0 0 0 0 kzz
   
  T−Td 
dz Ad

 


where,

dQ f , dQl , dQb , dQr , dQt , and dQd = heat exchange with the element connected to the front, left, back,
right, top, and down faces, respectively;
T f , Tl , Tb , Tr , Tt , and Td = temperature of the element connected to the front, left, back, right, top,
and down faces, respectively;
Materials 2020, 13, 4985 9 of 21

dx, dy, and dz = distance between the centroids of connected elements in the x, y, and z directions;
A f , Al , Ab , Ar , At , Ad = contact area with the element connected to the front, left, back, right, top,
and down faces, respectively;
dt = time step.

Considering the front and back faces, the faces in the global x direction, and Ab = Af = Ax ,
Al = Ar = Ay , At = Ad = Az for hexahedral elements used in this study, Equation (9) can be formed.

dQ = dQ f + dQb + dQr + dQl + dQt + dQd

 
T−T f T−T T−T T−Tt T−Td
dQ = kxx dx Ax + kxx dx b Ax + k yy dy l y + k yy T−Tdy
r
A y + k zz dz A z + k zz dx A z dt (9)
     
−Tb +2T−T f −Tl +2T−Tr
 −T +2T−T 
t d
dQ = kxx dx Ax + k yy dy A y + kzz dz Az dt

Substituting Equation (9) in Equation (7), Equation (10) can be formed, which is equivalent to the
forward time center space (FTCS) finite difference scheme for the Fourier heat transfer equation [56].
      −T +2T−T  
−Tb +2T−T f −Tl +2T−Tr
ρ·Cnew ·V·Tnew = ρ·Cold ·V·Told − kxx dx A x + k yy dy A y + k zz
t
dz
d
Az dt (10)

The temperature of each active element in the model is updated using the procedure described
in Equations (2)–(7). When an element is added to the model, the neighbor information is updated
to define the new boundary conditions. If an element is in a layer far away from the element just
deposited, the far-away element is made inactive. Temperature is not updated for inactive elements in
the model. For this study, a 20-layer distance was considered far, and elements more than 20 layers
away from the element currently being deposited were made inactive. Figure 5 shows the changing
neighbors with the addition of elements to the model. At time t = t1 , element 1 has all faces exposed to
the environment. At time t = t2 , element 2 is added with neighbor element 1 at its back. The neighbor
information of element 1 is also updated such that element 2 is at the front of the element 1.
A stable time increment for a uniform grid in finite difference methods with forward Euler
integration is given by Equation (11) [56].

∆x2
!
∆t < (11)
2α

where,

α= K
ρ·C and
∆x = distance between the nodes.

Orthotropy in the conductivity of the 3D-printed material due to different conductivities along
the bead and across the beads is considered. The conductivity matrix transformation from the local
system to the global system is shown by Equation (12).
    
s 0   kxx c2 + k yys2
  
 c −s 0  kxx 0 0  c cs kx − k y 0 
[K]G = [T ]T [K]l [T ] =  s
     

c 0  0

k yy 0  −s
 c 0  =  cs kx − k y
  k yy c2 + kxx s2 0  (12)
0 0 1 0 0 kzz 0 0 1
 
0 0 1

where,

[K]G = the conductivity matrix in the global system (machine axes),

[K]l = the conductivity matrix in the local system,
[T ] = the transformation matrix,
c = cosine of bead deposition orientation to the machine X axis,
orthotropic material.
𝑇−𝑇
⎧ 𝐴 ⎫
1 1 𝑑𝑥
⎡ 𝑘 𝑐 +𝑘 𝑠 𝑐𝑠(𝑘 − 𝑘 ) 0 𝑐𝑠(𝑘 − 𝑘 )
0 0 ⎤⎪𝑇 − 𝑇 ⎪
𝑑𝑄 ⎢ 2 2 ⎥⎪⎪
𝐴 ⎪
⎪
1 1 ⎥ ⎪ 𝑑𝑦
Materials 2020, 13, ⎫ ⎢ 𝑐𝑠(𝑘 − 𝑘 ) 𝑘 𝑐 +𝑘 𝑠
⎧ 4985 𝑐𝑠(𝑘 − 𝑘 ) 0 0 0 ⎪ 10 of 21
𝑑𝑄
⎪ ⎢2 2 ⎥ ⎪𝑇 − 𝑇
⎪
⎢ 1 1 ⎥ 𝐴 ⎪
𝑑𝑄 𝑑𝑥
=⎢ 0 𝑐𝑠(𝑘 − 𝑘 ) 𝑘 𝑐 +𝑘 𝑠 𝑐𝑠(𝑘 − 𝑘 ) 0 0 ⎥ 𝑇−𝑇 𝑑𝑡 (13)
⎨ 𝑑𝑄 ⎬ ⎢ 2 2
⎥⎨ 𝐴 ⎬
s = sine of bead 𝑑𝑄 1
⎪ deposition
⎪ ⎢ 1
𝑐𝑠(𝑘 orientation
−𝑘 ) 0to the machine
𝑐𝑠(𝑘 − 𝑘X)axis,
𝑘 𝑐 +𝑘 𝑠 0 0 ⎥ ⎪ 𝑑𝑦 ⎪
⎩𝑑𝑄 ⎭ ⎢2 2 ⎥⎪𝑇 − 𝑇 ⎪
kxx = in-plane conductivity
⎢ 0 along the deposition
0 of0the bead, 0 𝑘 0 ⎥⎪ 𝐴 ⎪
⎪ 𝑑𝑧 ⎪
k = in-plane conductivity⎣ 0 perpendicular
0 0
to the deposition 0 bead,0and
of the 𝑘 ⎦ ⎪𝑇 − 𝑇 ⎪
yy 𝐴 ⎭
⎩ 𝑑𝑧
kzz = out-of-plane conductivity in the machine Z axis.

(a) (b)

(c) (d)
Figure 5. Changing
Changing neighbor
neighbor information
information with
with addition
addition of
of elements (a) tt == tt11,,(b)
elements to the model. (a) (b) tt == tt22,,
(c) tt == tt33, and (d)t t==t4.t4.
and(d)

Equation
The neighbor (13)information
shows the heat transferred
is updated similarlyfrom forthe faces of
addition an element
of element viaelement
3 and conduction for an
4, as shown
orthotropic
in Table 2. material.
 T−T f A 
    
  2 + k s2 1 1
k c cs k − k 0 cs k − k 0 0
 
 dQ f  xx yy x y x y f

 
  2 2   dx

 

T−T
     
   1 cs k − k k c 2 + k s2 1
cs k − k 0 0 0
 l
A

dQ
   


 l


  
 2 x y xx yy 2 x y 



 dy l 


     T−T
1 2 2 1
   
 dQb  0 cs k − k k c + k s cs k − k 0 0 b

A
    
x y xx yy x y
  
=
  2 2


 dx b 
dt (13)
T−Tr
   
dQ 1 1
   
2 2
 
 2 cs kx − k y 0 2 cs kx − k y k yy c + kxx s 0 0  dy Ar 
r 

 
 
 

 


   
dQt 
   
T−T
 
 dz t At 
  
0 0 0 0 kzz 0 
   

  
  
 dQ 
  
0 0 0 0 0 kzz   T−Td Ad 
   
d
 
dz

The neighbor information is updated similarly for addition of element 3 and element 4, as shown
in Table 2.
Materials 2020, 13, x FOR PEER REVIEW 11 of 21

Materials 2020, 13, 4985 11 of 21

Table 2. Neighbor information for cells with addition of new element at each time step.

Time Step Elements Front Back Right Left Top Bottom

Table 2. Neighbor information for cells with addition of new element at each time step.
t = t1 1 env env env env env env
Time 1 2 env env env env env
t =Step
t2 Elements Front Back Right Left Top Bottom
t = t1 1
2 env
env env
1 env env
env
env env
env
env
env
11 2
2 env
env env
env env
env env
env env
env
t = t2
t = t3 22 env
env 11 env env
env 3 env
env env
env
13 2
env env
env env
2 env
env env
env env
env
t = t3 21 env 1 env 34 env env
2 env env env env
3 env env 2 env env env
1
2 env
2 env
1 env
env 4
3 env
env
env
env
t = t4
23 env
env 14 2
env env
3 env
env env
env
t = t4
34 3
env env
4 21 env
env env
env env
env
4 3 env 1 env env env
Once all the elements have been deposited, the model is allowed to run for a specified period of
time Once
to simulate the cooling
all the elements havedown of the 3D-printed
been deposited, the modelpart. The time–temperature
is allowed data period
to run for a specified for each
of
element
time are exported
to simulate to adown
the cooling file. of
ThetheRust programming
3D-printed part. Thelanguage was used
time–temperature forfordiscrete-event
data each element
simulation
are exporteddue
to atofile.
its speed, automatic
The Rust memory
programming management
language was usedwithout use of a garbage
for discrete-event collector,
simulation and
due to
safe concurrency. Abaqus 2020 and its 3D printing module was used for thermal
its speed, automatic memory management without use of a garbage collector, and safe concurrency. finite element
modeling.
Abaqus 2020 and its 3D printing module was used for thermal finite element modeling.

2.3. Part Geometry, Printing Parameters, G-Code

G-code Generation,
Generation,and
and Inputs
Inputs for
for the
the DES
DES Model
Model
solid modeling
A solid modelingcomputer-aided
computer-aideddesign
designsoftware
software(SolidWorks
(SolidWorks version
version 2016
2016 SP5.0)
SP5.0) waswas used
used for
for generating the geometry of the 3D-printed part. A miniature geometry ashtray
generating the geometry of the 3D-printed part. A miniature geometry ashtray model, asmodel, as used by
Zhou et al. [45], was used for this study. The geometry was considered complex enough to showcase
the effects of 3D printing. The STL file was created using SolidWorks. Simplify3D software was used
for slicing and generating G-code tool path.
generating the G-code path. Figure 6a shows the 3D model of the STL file imported
into Simplify3D software. TheThe part
part has
has a diameter
diameter of 48.8 mm and a height of 12.7
12.7 mm.
mm. Figure 6b
shows the deposition path for the 3D-printed part.

(a) (b)
Figure 6. 3D model and printing path for the 3D-printed part. (a) 3D model of the part, (b) Printing
path for the part.

The printing parameters used for this study are listed in

in Table
Table 3.
3.
Materials 2020,13,
Materials2020, 13,4985
x FOR PEER REVIEW 12
12of
of21
21

Table 3. Printing parameters used for this study.

Table 3. Printing parameters used for this study.
Parameter Value
Parameter Value
Bead width 0.8 mm
Bead
Beadwidth
height 0.4 0.8
mmmm
Bead height 0.4 mm
Move speed 90 mm/s
Move speed 90 mm/s
Deposition speed
Deposition speed 60 mm/s
60 mm/s
Infillangle
Infill angle 0° 0◦
◦
Deposition temperature
Deposition temperature 200forC PLA,
200 °C for PLA, °C◦for
210210 C for ABS
ABS
Infill percentage
Infill percentage 100%100%
First layer speed 12 mm/s
First layer speed 12 mm/s

A temperature of 25 °C was used as the environmental temperature. A convective heat transfer

A temperature of 25 ◦ C was used as the environmental temperature. A convective heat transfer
coefficient of 100 W·m−2−2·K−1
−1 was assumed for this study. The thermal properties of PLA used for this
coefficient of 100 W·m ·K was assumed for this study. The thermal properties of PLA used for this
study are listed in Table 4.
study are listed in Table 4.
Table 4. Thermal properties used for the PLA material during finite element analysis.
Table 4. Thermal properties used for the PLA material during finite element analysis.
Property Value
Property
Value
density, ρ 1240 kg/m3 [57]
density, ρ 1240 kg/m 3
conductivity in the x direction, kxx 0.25 W·m −1·K[57]
−1 [58]
conductivity in the x direction, kxx 0.25 W·m ·K−1 [58]
−1
conductivity
conductivity in the y in the y direction,
direction, kyy kyy 0.2W·m
0.2 W·m −1·K−1
−1 ·K −1
conductivity
conductivity in the z in the z direction,
direction, kzz kzz 0.2W·m
0.2 W·m ·K−1−1
−1−1·K

Specific heat capacities for PLA as a function of temperature are shown by the peaks and valleys
Specific heat capacities for PLA as a function of temperature are shown by the peaks and valleys
in Figure 7. The specific heat capacity of PLA changes rapidly at the glass transition temperature at
in Figure 7. The specific heat capacity of PLA changes rapidly at the glass transition temperature
approximately 55 °C, at the cold crystallization temperature at approximately 120 °C, and at the
at approximately 55 ◦ C, at the cold crystallization temperature at approximately 120 ◦ C, and at the
melting point at approximately 150 °C.
melting point at approximately 150 ◦ C.
specific heat capacity (W . kg-1 . K-1 )

Figure 7. Specific heat capacity of PLA vs. temperature.

Figure 7. Specific heat capacity of PLA vs. temperature.
Table 5 Thermal properties used for the ABS material during finite element analysis. The thermal
Table 5 Thermal properties used for the ABS material during finite element analysis. The thermal
properties are taken from the publication by Zhou et al. [45].
properties are taken from the publication by Zhou et al. [45].
Materials 2020, 13, 4985 13 of 21

Table 5. Thermal properties used for the ABS material during finite element analysis [45].

Thermal Specific Heat

Density
Temperature (◦ C) Conductivity Capacity
(kg·m−3 )
(W·m−1 ·K−1 ) (J·kg−1 ·K−1 )
0 0.23 780 1050
50 0.25 1040 1050
100 0.28 1490 1050
150 0.29 1710 1050
200 0.31 1865 1050
250 0.33 2020 1050

For a grid of 0.8 mm × 0.8 mm × 0.4 mm, with K = 0.20 W·m−1 ·K−1 , ρ = 1240 kg·m−3 ,
and C = 1400 J·kg−1 ·K−1 , a stable time increment for PLA was calculated as 0.6944 s.
For a grid of 0.8 mm × 0.8 mm × 0.4 mm, with K = 0.33 W·m−1 ·K−1 , ρ = 1050 kg·m−3 ,
and C = 780 J·kg−1 ·K−1 , a stable time increment for ABS was calculated as 0.1985 s.
For both cases, a time step of 0.01 s was chosen for the DES model. The maximum time increment
that can be used for the DES model is the time required to deposit an element. The effect of smaller
time increments on the results was studied by using four different time steps of 0.01, 0.005, 0.0025,
and 0.00125 s.

2.4. Finite Element Modeling Using Abaqus with the AM Modeler Plugin
Abaqus 2020 with the AM modeler plugin was used for the finite element analysis of the part.
A cuboid with a length of 50.4 mm, a width of 50.4 mm, and a height of 13.2 mm was created. An 8-node
hexahedral mesh (DC3D8) was created, where each element had a length and a width of 0.8 mm and a
height of 0.4 mm. The width and height of the elements matched the width and height of the deposited
beads. A total of 34071 elements were deposited in the final part during the analysis. A homogenous
solid section property and material orientation directions matching the 3D printing direction were
assigned to the part. The material properties used for the finite element analysis of PLA are listed
in Table 4. A temperature of 25 ◦ C was used as the environmental temperature. A convective heat
transfer coefficient of 100 W·m−2 ·K−1 was assumed for this study. The thermal properties of PLA used
for this study are listed in Table 4.
A transient thermal analysis was carried out with a time step of 1 s. Using a small time step
would give a better resolution of temperature, capturing the peak temperatures better. However,
the time required to complete the finite element simulation and the space required to store the output
results increase with a decreasing time step. Increasing the time step would improve the speed of
the simulation but miss capturing the peak and valley temperatures. A time step of 1 s was used
as a compromise to obtain a good time–temperature resolution within a reasonable time period.
A predefined temperature of 200 ◦ C was used for PLA and a predefined temperature of 210 ◦ C was
used for ABS.
The AM modeler plugin was used to define the progressive element activation and cooling
interactions. A G-code to event series converter was written in the Rust programming language.
The event series file generated was used as an input to the AM modeler plugin. An analysis job was
set up with 6 processors enabled.
Two elements, one deposited at the beginning of the simulation and another midway through
the simulation, were chosen to compare the results from the FEA model to the results from the
discrete-event simulation model.

3. Results
Figure 8 shows the nodal temperature distribution in the part at 766 s from the beginning of the
print for the finite element simulation. The finite element simulation ran until 778 s from the beginning
Materials 2020, 13, 4985 14 of 21

of the print. The elements that are deposited later are hotter compared to the elements that were
Materials 2020, 13, x FOR PEER REVIEW 14 of 21
deposited earlier.
Materials 2020, 13, x FOR PEER REVIEW 14 of 21

Figure 8. Temperature distribution in the model from FEA results.

Figure 8. Temperature distribution in the model
model from
from FEA
FEA results.
results.
Figure 9 shows the temperature versus time history of the 3D-printed part at two different
Figure
Figure 9 shows the temperature versus time
time history of the
the 3D-printed part
part at two
two different
locations for9theshows
PLA the temperature
material. versushistory
The thermal history of from
obtained 3D-printed
the DES model at
closely different
matches
locations
locations for
for the
the PLA
PLA material.
material. The
The thermal
thermal history
history obtained
obtained from
from the
the DES
DES model
model closely
closely matches
matches
that obtained from the FEA model. The DES model has a better resolution as it shows results from
that obtained
that obtained from the
from the FEA
FEA model.
model. The DES model has a better resolution as it shows results from
every 0.01 s. Hence, it can record theThe
sharpDES model
peaks thathas a better
arise when resolution
a new elementas it is
shows results
deposited from
next to
every
every 0.01
0.01 s.
s. Hence, it
Hence, it can
can record
record the
the sharp
sharp peaks
peaks that
that arise
arise when
when aa new
new element
element is
is deposited
deposited next
next to
to
another element.
another
another element.
element.
(°C) ( C)

C) C)
°

°
temperature(
temperature

°
temperature(
temperature

(a) (b)
(a) (b)
Figure 9. Comparison of the time–temperature history calculated using discrete element simulation
Figure
and FEM9. for
Comparison of the
the PLA part. (a)time–temperature
The thermal history history
of thecalculated using element
34th deposited discrete with
element simulation
the centroid at
and
(x = FEMyfor
26.0,
= 26.0, == 3.4,
the and
3.4, PLA ==0.2
and zzpart.
0.2(a) Theand
mm),
mm), thermal
and history
(b) the
(b) the of the
thermal
thermal 34th of
history
history deposited element
of the 12473th with the
deposited centroid
element at
with
(x =centroid
the 26.0, y =at3.4, =
(x and
31.6,
= 31.6, ==28.2,
z =yy0.2 mm),and
28.2, and
and =1.8
zz=(b) the
1.8 thermal history of the 12473th deposited element with
mm).
mm).
the centroid at (x = 31.6, y = 28.2, and z = 1.8 mm).
Figure 10
Figure shows the
10 shows the temperature
temperature vs. time history
vs. time history of
of the
the 3D-printed
3D-printed part
part at
at two
two different
different locations
locations
for Figure
the ABS 10 shows
material. the
The temperature
thermal vs.
history time history
obtained of
from the
the3D-printed
DES model part at
closelytwo different
matches
for the ABS material. The thermal history obtained from the DES model closely matches the thermalthelocations
thermal
for the ABS
history
history material.
obtained
obtained fromThe
from the thermal
the FEA history obtained from the DES model closely matches the thermal
FEA model.
model.
history obtained from the FEA model.
Materials 2020, 13, 4985 15 of 21
Materials
Materials2020,
2020,13,
13,xxFOR
FORPEER
PEERREVIEW
REVIEW 15ofof2121
15

temperature ( C)

C)
temperature (°°C)

temperature ((°C)
°
temperature
(a)
(a) (b)
(b)
Figure10.
Figure 10.Comparison
Comparisonof
Comparison ofthe
of thetime–temperature
the time–temperature history
time–temperature history calculated
calculated using
using discrete
discreteelement
elementsimulation
simulation
and
and FEMFEM for
for ABS
ABS part.
part. (a)
(a) The
The thermal
thermal history
history of
of the
the 34
34
th deposited element with the centroid at (x =
th deposited element with the centroid at (xat
=
FEM for ABS part. (a) The thermal history of the 34th deposited element with the centroid
26.0, y = 3.4, and z = 0.2 mm), and (b) the thermal history of the 12473th deposited element with the
(x = 26.0,
26.0, y =and
y = 3.4, 3.4, zand z =mm),
= 0.2 and (b)
0.2 mm), andthe
(b)thermal history
the thermal of the
history of 12473th deposited
the 12473th element
deposited withwith
element the
centroid at (x = 31.6, y = 28.2, and z = 1.8 mm).
the centroid
centroid at (xat= (x = 31.6,
31.6, y = 28.2,
y = 28.2, and zand z =mm).
= 1.8 1.8 mm).

Forboth
For bothPLA PLA andand ABS,
ABS, thethe time–temperature
time–temperature history history starts
startsrecording
recordingafter afterthe thefirst
firsttime
timestep,
step,
For both PLA and ABS, the time–temperature history starts recording after the first time step,
which
which is is
1s 1s
forfor
the the Abaqus
Abaqus FEA FEA
model model
and and
0.01 s 0.01
for thes for
DES the DES
model. Asmodel.
a result,As thea result,
initial the initial
temperature
which is 1s for the Abaqus FEA model and 0.01 s for the DES model. As a result, the initial
temperature
shown shown
in theshown
graphs in the graphs is the
lower than the initial temperature
of 200 ◦ C of 200 °C and
for PLA 210 and 210ABS.
◦ C for °C
temperature inisthe
lower
graphs thanis lower initial
thantemperature
the initial temperature for
of 200PLA °C for PLA and 210 °C
for ABS.
Figure Figure 11 shows the effect of a decreasing time step on the results of the DES model. Figure
for ABS.11Figure
shows11the effect
shows theofeffect
a decreasing time step
of a decreasing time onstep
theonresults of theofDES
the results the model.
DES model. Figure 11a
Figure
11a shows
shows the thermal
the thermal history history
of theofof
34ththe 34th deposited element
for the for the 3D-printed partthewith
ABSthe ABS
11a shows the thermal history thedeposited
34th depositedelement element 3D-printed
for part with
the 3D-printed part with material.
the ABS
material.
Four different Four different time steps of 0.01, 0.005, 0.0025, and 0.00125 s are shown. However, the
material. Four time stepstime
different of 0.01,
steps0.005,
of 0.01,0.0025,
0.005,and 0.00125
0.0025, ands 0.00125
are shown.
s are However,
shown. However, the thermal the
thermal histories obtained by using the four time steps are too close to see any difference in the line
histories obtained by using the four time steps are too close to see any
thermal histories obtained by using the four time steps are too close to see any difference in the line difference in the line plots.
plots. The first reheating peak is marked with a red ellipse. The marked area is zoomed in and shown
The first
plots. Thereheating peak peak
first reheating is marked
is markedwith with
a redaellipse. TheThe
red ellipse. marked
marked areaarea
is zoomed
is zoomed in and
in andshown
shown in
in Figure 11b to view the difference in results for models using the different time steps. The peak
Figure
in Figure11b11b to view the the
to view difference
differencein results for models
in results usingusing
for models the different time steps.
the different time The peak
steps. Thevalues
peak
values seem to be converging with smaller time steps. With each halving of the time step, the
seem to be converging with smaller time steps. With each halving
values seem to be converging with smaller time steps. With each halving of the time step, the of the time step, the difference in
difference in peak temperature becomes smaller. However, the difference between the peaks for
peak temperature
difference in peakbecomes
temperature smaller. However,
becomes smaller. the difference
However, between the peaks
the difference betweenfor models
the peaksusing fora
models using a 0.00125 s time step and a 0.01 s time step is approximately ◦ C, which is very0.3 °C, which is very small
0.00125 using
models s time astep and as 0.01
0.00125 time sstep
timeandstepa is approximately
0.01 s time step is0.3
approximately 0.3 small
°C, which compared
is very to the
small
compared to the temperature of the elements in the model.
temperature
compared to of
thethe elements inofthe
temperature themodel.
elements in the model.

220
220 0.01 s 0.01 s
200 0.005 s 98.25 0.005 s
0.01 s 0.01 s
200 0.0025 s 98.25 0.0025 s
0.005 s 0.005 s
180 0.00125 s 98.2 0.00125 s
0.0025 s 0.0025 s
180 0.00125 s 98.2 0.00125 s
160 98.15
(°C)(°C)

160 98.15
140 98.1
temperature

140 98.1
120 98.05
temperature

120 98.05
100 98
100 98
80
97.95
80
60 97.95
97.9
6040
97.9
97.85
4020
97.85
0 100 200 300 400 500 600 700 800 900 146.65 146.7 146.75 146.8 146.85 146.9 146.95 147 147.05
20 time (s) time (s)
0 100 200 300 400 500 600 700 800 900 146.65 146.7 146.75 146.8 146.85 146.9 146.95 147 147.05
time (s) time (s)
(a) (b)
(a) (b)
Figure11.
Figure 11.Effect
Effectof
ofaadecreasing
decreasingtime
timestep
stepon
onthe
theDES
DESmodel
modelresults
resultsfor
forABS.
ABS.(a)
(a)The
Thethermal
thermalhistory
historyof
of the
34th34th
Figure
the 11. deposited
Effect
deposited element
ofelement with
a decreasing
with time
the the centroid
step
centroid at
onatthe = =26.0,
(x
(x DES 26.0,y y==3.4,
model 3.4,
results
and zz =
andfor = ABS.
0.2
0.2 mm), and
(a) The
mm), (b)
(b)the
thethermal
andthermal history
thermal
history
of ofthe
the 34th
history of the34th
34thdeposited
deposited deposited element
element element
with zoomed
thezoomed
centroidinatat(x the
= 26.0, = 3.4, and peak.
firsty reheating z = 0.2 mm), and (b) the thermal
history of the 34th deposited element zoomed in at the first reheating peak.
Materials 2020, 13, 4985 16 of 21

Materials 2020, 13, x FOR PEER REVIEW 16 of 21

Figure 12 shows the effect of making the 34th deposited element inactive when it is 10, 15, 20,
Figure 12 shows the effect of making the 34th deposited element inactive when it is 10, 15, 20,
25, and 30 layers away from the most currently deposited layer. Figure 12 also shows the thermal
25, and 30 layers away from the most currently deposited layer. Figure 12 also shows the thermal
history of the 34th deposited element when it is active throughout the analysis. The reheating peaks
history of the 34th deposited element when it is active throughout the analysis. The reheating peaks
are approximately 4 ◦ C high after 10 layers of deposition. If the element is made inactive after 10 layers
are approximately 4 °C high after 10 layers of deposition. If the element is made inactive after 10
of deposition, the reheating peaks that are lower than 4 ◦ C will not be recorded by the model. Similarly,
layers of deposition, the reheating peaks that are lower than 4 °C will not be recorded by the model.
the reheating peaks after 15, 20, 25, and 30 layers of deposition are 0.6, 0.3, 0.08, and 0.01 ◦ C high.
Similarly, the reheating peaks after 15, 20, 25, and 30 layers of deposition are 0.6, 0.3, 0.08, and 0.01
The choice for distance to the currently deposited layer should be based on the accuracy desired for
°C high. The choice for distance to the currently deposited layer should be based on the accuracy
the model. Bonding between thermoplastic beads only happens at a temperature higher than the
desired for the model. Bonding between thermoplastic beads only happens at a temperature higher
glass transition temperature [5,51]. Thermal models used to predict such bonding can use a 10-layer
than the glass transition temperature [5,51]. Thermal models used to predict such bonding can use a
distance or lower for inactive elements. Young’s modulus for thermoplastic materials increases rapidly
10-layer distance or lower for inactive elements. Young’s modulus for thermoplastic materials
at the glass transition temperature with decreasing temperature. Models predicting residual stress
increases rapidly at the glass transition temperature with decreasing temperature. Models predicting
are more sensitive to temperature changes at a lower temperature, as Young’s modulus for materials
residual stress are more sensitive to temperature changes at a lower temperature, as Young’s
increases with decreasing temperature [43]. For such models that predict residual stresses, inactive
modulus for materials increases with decreasing temperature [43]. For such models that predict
elements that are 20 layers away might be necessary.
residual stresses, inactive elements that are 20 layers away might be necessary.
220
inactive elements 10 layers away
200 inactive elements 15 layers away
inactive elements 20 layers away
180 inactive elements 25 layers away
inactive elements 30 layers away
160 all elements active

140

120

100

80

60

40

0 100 200 300 400 500 600 700 800

time (s)

Figure 12. Effect of the distance of the inactive element from the most recently deposited layer.
Figure 12. Effect of the distance of the inactive element from the most recently deposited layer.
The FEA modeling-based simulation for the PLA material was completed in 1 h 36 min (5766 s)
with 6The FEA
cores. Themodeling-based
simulation using simulation for thewas
the DES model PLAcompleted
material was completed
in 11.3 s with 6 in 1 h 36
cores. min (5766
A speed up ofs)
with 6 cores. The simulation using the DES model was completed
510 times was obtained using discrete-event simulation in this case. in 11.3 s with 6 cores. A speed up
of 510 times was obtained using discrete-event simulation in this case.
The FEA modeling-based simulation for the ABS material was completed in 1 h 30 min (5420 s)
with 6The FEA
cores. Themodeling-based
simulation using simulation for thewas
the DES model ABScompleted
material was completed
in 11.1 s with 6 in 1 h 30
cores. min (5420
A speed up ofs)
with 6 cores. The simulation using the DES model was completed
488 times was obtained using discrete-event simulation in this case. in 11.1 s with 6 cores. A speed up
of 488 times was obtained using discrete-event simulation in this case.
Figure 13 shows the speed up in the DES model by using a higher number of cores. The core of
Figure 13used
the algorithm shows
for the speed up
calculating intemperatures
the the DES model by using
is highly a higher
parallel. number
The of cores.
temperature The
of an core of
element
the algorithm used for calculating the temperatures is highly parallel. The temperature
is dependent only on its own properties and the properties of its immediate neighbors. The simulation of an element
is dependent
can be sped up only on aitshigher
by using own number
properties and cores.
of CPU the properties of its immediate neighbors. The
simulation can be sped up by using a higher number of CPU cores.
Materials 2020, 13, 4985 17 of 21
Materials 2020, 13, x FOR PEER REVIEW 17 of 21

35

30

25

20

15

10
1 2 3 4 5 6
number of CPU cores

Figure 13. Time to finish the simulation vs. number of CPU cores used.
Figure 13. Time to finish the simulation vs. number of CPU cores used.
4. Discussion
4. Discussion
The experimental observations form different research works show certain features in the
thermal Thehistory
experimental
that couldobservations
also be found form different
in the thermal research
history works
presentedshow certain
by this features
model. It hasinbeenthe
thermal history
observed that could
in various researchalsoworks
be found
that inthethe thermal history
temperature of thepresented
deposited bybead
this model.
decreases It has been
rapidly
observed in various research works that the temperature of the deposited
after deposition [5,51]. The rapid cool down of the deposited element is due to the loss of heat to the bead decreases rapidly
after deposition
neighboring bead[5,51].
due toThe rapid cool
conduction anddown
the lossof of
theheat
deposited element is due
to the environment dueto tothe loss of heat
convection. Thetoheat
the
neighboring bead due to conduction and the loss of heat to the environment
loss is characterized by an exponentially decreasing temperature with time. The time–temperature due to convection. The
heat loss
history graph is predicted
characterizedby the bymodel
an exponentially
shows a similar decreasing
temperature temperature
decrease inwith time. The
the deposited time–
element.
temperature
Another notablehistory
featuregraph
in thepredicted by the model
time–temperature historyshows
graph aofsimilar temperature
the deposited beads in decrease
3D printing in theis
deposited element. Another notable feature in the time–temperature history
the rapid reheating by conduction of heat from the beads deposited above it [39,48,59,60]. The reheating graph of the deposited
beads
of in 3D printing
the elements is evenismore
the rapid
rapid reheating by conduction
than the cooling down of of theheat from with
element, the beads deposited
an almost above it
instantaneous
[39,48,59,60].
rise to the peak. TheThereheating of the elements
model presented is even more
in this manuscript rapidsimilar
shows than the coolingof
reheating down of theelements
deposited element,
with an almost instantaneous rise to the peak. The model presented in
due to freshly deposited beads above it. A common observation in the thermal history of 3D-printed this manuscript shows similar
reheating
parts is theofexponential
deposited elements
decrease due to freshlypeaks
in reheating deposited beads
with the above in
increase it. the
A common
distance observation
to the newest in
the thermal history of 3D-printed parts is the exponential decrease
layer deposited [47,51,59]. The model presented here shows a similar decrease in the reheating peaks,in reheating peaks with the
increase
with in the distance
depositions in layersto20the newest layer
depositions away deposited
having a[47,51,59].
negligibleThe model
effect on the presented here shows
thermal history of thea
similar decrease
3D-printed part. in the reheating peaks, with depositions in layers 20 depositions away having a
negligible effect
The DES model on the
hasthermal
a betterhistory
resolutionof theas 3D-printed
it stores resultspart.from every 0.01 s. Hence, it can record
the sharp peaks that arise when a new element is deposited nextfrom
The DES model has a better resolution as it stores results every 0.01
to another s. Hence,
element. it can record
The difference in
the sharp
peaks peaksthe
between that arise
FEA whenand
model a new
the element
DES model is deposited
is higher next
for theto another
PLA partelement.
than that The fordifference
the ABS
in peaks
part. PLAbetween
exhibits themuchFEA modelchanges
sharper and the in DESthemodel
specific is heat
higher for thecompared
capacity PLA part to than
ABS thatandfora the ABS
smaller
part. PLA exhibits much sharper changes in the specific heat capacity compared to ABS and a smaller
time step is necessary to capture the changes in specific heat capacity of the PLA parts. However,
time step is necessary to capture the changes in specific heat capacity of the PLA parts. However, the
the FEA model with a largertime step can only capture the changes in the specific heat capacity with a
FEA model with a largertime step can only capture the changes in the specific heat capacity with a
coarser resolution.
coarser resolution.
The speed up results mainly due to two reasons. Firstly, the discrete element model only considers
The speed up results mainly due to two reasons. Firstly, the discrete element model only
immediate neighbors for heat transfer for a small time step. As a result, solving a large system of
considers immediate neighbors for heat transfer for a small time step. As a result, solving a large
linear equations is avoided. Furthermore, since only information about immediate neighbors is used
system of linear equations is avoided. Furthermore, since only information about immediate
for each element in a given time step, each element can be updated individually, independent of
neighbors is used for each element in a given time step, each element can be updated individually,
other elements that are not its immediate neighbors. As a result of this locality, parallel threads
independent of other elements that are not its immediate neighbors. As a result of this locality,
running on different cores can be used to update the temperature of each element for a given time step.
parallel threads running on different cores can be used to update the temperature of each element for
Secondly, the model considers the elements in far-away layers (at least 20 layers away from layer being
a given time step. Secondly, the model considers the elements in far-away layers (at least 20 layers
currently deposited for this study) to be inactive and does not update the temperature of such elements.
away from layer being currently deposited for this study) to be inactive and does not update the
temperature of such elements. Consequently, the number of elements that need to be updated at each
Materials 2020, 13, 4985 18 of 21

Consequently, the number of elements that need to be updated at each time step is limited. As a result,
the model can simulate the thermal history of 3D-printed material in linear time, i.e., increasing the
model size increases the simulation time proportionally. In contrast, in conventional finite element
models, increasing the size of the model increases the simulation time non-linearly, making the thermal
simulation of 3D-printed large parts prohibitively long.
The numerical model presented is significantly faster than the equivalent finite element model.
The DES model developed would allow designers to quickly compare the effects of different printing
parameters on the thermal history of a 3D-printed part and allow the designers to select the most
suitable printing parameters for a given part.

5. Conclusions
The following conclusions were drawn from the research work presented:

(1) A fast and accurate numerical model can be formulated based on heat exchange between
deposited elements to generate the thermal history of a polymer extrusion-based 3D-printed
part. The formulated DES numerical model was shown to be fast and accurate for the thermal
simulation of extrusion-based additive manufacturing.
(2) Orthotropy in 3D-printed parts can be accounted for in the simplified numerical model by using
the information on the orientation of deposited beads from the G-code. The effect of orthotropic
conductivity in the 3D-printed PLA part was simulated using the DES model. The DES model was
able to account for the orthotropic conductivity of the 3D-printed PLA part used for this study.
(3) The temperature-dependent material response can be captured in the DES model simulation of
the thermal behavior of 3D-printed polymers. The temperature-dependent specific heat capacity
was used for PLA and ABS. Temperature-dependent conductivities were used for ABS. The DES
model can incorporate the effects of the temperature-dependent material properties of the 3D
printing polymers.
(4) The accuracy of the DES model is comparable to the accuracy offered by finite element models.
The results from the DES model were found to be comparable with the results from the FEA model.

Author Contributions: Conceptualization, S.B. and R.A.L.-A.; methodology, S.B.; software, S.B.; supervision,
R.A.L.-A.; writing—original draft, S.B.; writing—review and editing, R.A.L.-A. All authors have read and agreed
to the published version of the manuscript.
Funding: This research was provided by the Transportation Infrastructure Durability Center at the University
of Maine under grant 69A3551847101 from the U.S. Department of Transportation’s University Transportation
Centers Program, and the Malcolm G. Long ‘32 Professorship in Civil Engineering.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Cattenone, A.; Morganti, S.; Alaimo, G.; Auricchio, F. Finite Element Analysis of Additive Manufacturing
Based on Fused Deposition Modeling: Distortions Prediction and Comparison With Experimental Data.
J. Manuf. Sci. Eng. 2019, 141. [CrossRef]
2. McIlroy, C.; Olmsted, P.D. Deformation of an amorphous polymer during the fused-filament-fabrication
method for additive manufacturing. J. Rheol. 2017, 61, 379–397. [CrossRef]
3. Brenken, B.; Barocio, E.; Favaloro, A.; Kunc, V.; Pipes, R.B. Development and validation of extrusion
deposition additive manufacturing process simulations. Addit. Manuf. 2019, 25, 218–226. [CrossRef]
4. Yin, J.; Lu, C.; Fu, J.; Huang, Y.; Zheng, Y. Interfacial bonding during multi-material fused deposition
modeling (FDM) process due to inter-molecular diffusion. Mater. Des. 2018, 150, 104–112. [CrossRef]
5. Sun, Q.; Rizvi, G.M.; Bellehumeur, C.T.; Gu, P. Effect of processing conditions on the bonding quality of FDM
polymer filaments. Rapid Prototyp. J. 2008, 14, 72–80. [CrossRef]
Materials 2020, 13, 4985 19 of 21

6. Caminero, M.A.; Chacón, J.M.; García-Moreno, I.; Reverte, J.M. Interlaminar bonding performance of 3D
printed continuous fibre reinforced thermoplastic composites using fused deposition modelling. Polym. Test.
2018, 68, 415–423. [CrossRef]
7. Coogan, T.J.; Kazmer, D.O. Prediction of interlayer strength in material extrusion additive manufacturing.
Addit. Manuf. 2020, 35, 101368. [CrossRef]
8. Bhandari, S.; Lopez-Anido, R.A.; Gardner, D.J. Enhancing the interlayer tensile strength of 3D printed short
carbon fiber reinforced PETG and PLA composites via annealing. Addit. Manuf. 2019, 30, 100922. [CrossRef]
9. Casavola, C.; Cazzato, A.; Moramarco, V.; Pappalettera, G. Residual stress measurement in Fused Deposition
Modelling parts. Polym. Test. 2017, 58, 249–255. [CrossRef]
10. Casavola, C.; Cazzato, A.; Karalekas, D.; Moramarco, V.; Pappalettera, G. The Effect of Chamber Temperature
on Residual Stresses of FDM Parts. 2019, 87–92. [CrossRef]
11. Wang, Y.; Zhou, Y.; Lin, L.; Corker, J.; Fan, M. Overview of 3D additive manufacturing (AM) and corresponding
AM composites. Compos. Part A Appl. Sci. Manuf. 2020, 139, 106114. [CrossRef]
12. Tan, L.J.; Zhu, W.; Zhou, K. Recent Progress on Polymer Materials for Additive Manufacturing.
Adv. Funct. Mater. 2020, 30, 2003062. [CrossRef]
13. Mazzanti, V.; Malagutti, L.; Mollica, F. FDM 3D Printing of Polymers Containing Natural Fillers: A Review
of their Mechanical Properties. Polymers 2019, 11, 1094. [CrossRef] [PubMed]
14. Wang, X.; Jiang, M.; Zhou, Z.; Gou, J.; Hui, D. 3D printing of polymer matrix composites: A review and
prospective. Compos. B. Eng. 2017, 110, 442–458. [CrossRef]
15. Rodriguez-Panes, A.; Claver, J.; Camacho, A.M. The Influence of Manufacturing Parameters on the Mechanical
Behaviour of PLA and ABS Pieces Manufactured by FDM: A Comparative Analysis. Materials 2018, 11, 1333.
[CrossRef]
16. Peterson, A.M. Review of acrylonitrile butadiene styrene in fused filament fabrication: A plastics
engineering-focused perspective. Addit. Manuf. 2019, 27, 363–371. [CrossRef]
17. Barrios, J.M.; Romero, P.E. Improvement of Surface Roughness and Hydrophobicity in PETG Parts
Manufactured via Fused Deposition Modeling (FDM): An Application in 3D Printed Self-Cleaning Parts.
Materials 2019, 12, 2499. [CrossRef] [PubMed]
18. Pranzo, D.; Larizza, P.; Filippini, D.; Percoco, G. Extrusion-Based 3D Printing of Microfluidic Devices for
Chemical and Biomedical Applications: A Topical Review. Micromachines 2018, 9, 374. [CrossRef]
19. Harris, M.; Potgieter, J.; Archer, R.; Arif, K.M. Effect of Material and Process Specific Factors on the Strength
of Printed Parts in Fused Filament Fabrication: A Review of Recent Developments. Materials 2019, 12, 1664.
[CrossRef]
20. Tanikella, N.G.; Wittbrodt, B.; Pearce, J.M. Tensile strength of commercial polymer materials for fused
filament fabrication 3D printing. Addit. Manuf. 2017, 15, 40–47. [CrossRef]
21. Carneiro, O.S.; Silva, A.F.; Gomes, R. Fused deposition modeling with polypropylene. Mater. Des. 2015, 83,
768–776. [CrossRef]
22. Wang, L.; Gardner, D.J.; Bousfield, D.W. Cellulose nanofibril-reinforced polypropylene composites for
material extrusion: Rheological properties. Polym. Eng. Sci. 2018, 58, 793–801. [CrossRef]
23. Popescu, D.; Zapciu, A.; Amza, C.; Baciu, F.; Marinescu, R. FDM process parameters influence over the
mechanical properties of polymer specimens: A review. Polym. Test. 2018, 69, 157–166. [CrossRef]
24. Ferreira, R.T.L.; Amatte, I.C.; Dutra, T.A.; Bürger, D. Experimental characterization and micrography of 3D
printed PLA and PLA reinforced with short carbon fibers. Compos. Part B Eng. 2017, 124, 88–100. [CrossRef]
25. Tao, Y.; Wang, H.; Li, Z.; Li, P.; Shi, S.Q. Development and Application of Wood Flour-Filled Polylactic Acid
Composite Filament for 3D Printing. Materials 2017, 10, 339. [CrossRef]
26. Chiulan, I.; Frone, A.N.; Brandabur, C.; Panaitescu, D.M. Recent Advances in 3D Printing of Aliphatic
Polyesters. Bioengineering 2017, 5, 2. [CrossRef]
27. Rebaioli, L.; Fassi, I. A review on benchmark artifacts for evaluating the geometrical performance of additive
manufacturing processes. Int. J. Adv. Manuf. 2017, 93, 2571–2598. [CrossRef]
28. Ferreira, I.; Machado, M.; Alves, F.; Torres Marques, A. A review on fibre reinforced composite printing via
FFF. Rapid Prototyp. J. 2019, 25, 972–988. [CrossRef]
29. Gordelier, T.J.; Thies, P.R.; Turner, L.; Johanning, L. Optimising the FDM additive manufacturing process to
achieve maximum tensile strength: A state-of-the-art review. Rapid Prototyp. J. 2019, 25, 953–971. [CrossRef]
Materials 2020, 13, 4985 20 of 21

30. Vaezi, M.; Yang, S. Extrusion-based additive manufacturing of PEEK for biomedical applications.
Virtual Phys. Prototyp. 2015, 10, 123–135. [CrossRef]
31. Arif, M.F.; Kumar, S.; Varadarajan, K.M.; Cantwell, W.J. Performance of biocompatible PEEK processed by
fused deposition additive manufacturing. Mater. Des. 2018, 146, 249–259. [CrossRef]
32. Spoerk, M.; Sapkota, J.; Weingrill, G.; Fischinger, T.; Arbeiter, F.; Holzer, C. Shrinkage and
Warpage Optimization of Expanded-Perlite-Filled Polypropylene Composites in Extrusion-Based Additive
Manufacturing. Macromol. Mater. Eng. 2017, 302, 1700143. [CrossRef]
33. Sodeifian, G.; Ghaseminejad, S.; Yousefi, A.A. Preparation of polypropylene/short glass fiber composite as
Fused Deposition Modeling (FDM) filament. Results Phys. 2019, 12, 205–222. [CrossRef]
34. Provaggi, E.; Capelli, C.; Rahmani, B.; Burriesci, G.; Kalaskar, D.M. 3D printing assisted finite element
analysis for optimising the manufacturing parameters of a lumbar fusion cage. Mater. Des. 2019, 163, 107540.
[CrossRef]
35. Somireddy, M.; Singh, C.V.; Czekanski, A. Analysis of the Material Behavior of 3D Printed Laminates Via
FFF. Exp. Mech. 2019, 59, 871–881. [CrossRef]
36. Bhandari, S.; Lopez-Anido, R.A.; Wang, L.; Gardner, D.J. Elasto-Plastic Finite Element Modeling of Short
Carbon Fiber Reinforced 3D Printed Acrylonitrile Butadiene Styrene Composites. JOM 2019, 72, 475–484.
[CrossRef]
37. Wang, T.-M.; Xi, J.-T.; Jin, Y. A model research for prototype warp deformation in the FDM process. Int. J.
Adv. Manuf. 2006, 33, 1087–1096. [CrossRef]
38. Compton, B.G.; Post, B.K.; Duty, C.E.; Love, L.; Kunc, V. Thermal analysis of additive manufacturing of
large-scale thermoplastic polymer composites. Addit. Manuf. 2017, 17, 77–86. [CrossRef]
39. Zhang, J.; Wang, X.Z.; Yu, W.W.; Deng, Y.H. Numerical investigation of the influence of process conditions
on the temperature variation in fused deposition modeling. Mater. Des. 2017, 130, 59–68. [CrossRef]
40. Stockman, T.; Schneider, J.A.; Walker, B.; Carpenter, J.S. A 3D Finite Difference Thermal Model Tailored for
Additive Manufacturing. JOM 2019, 71, 1117–1126. [CrossRef]
41. Ji, L.B.; Zhou, T.R. Finite Element Simulation of Temperature Field in Fused Deposition Modeling.
Adv. Mat. Res. 2010, 97, 2585–2588. [CrossRef]
42. D’Amico, A.; Peterson, A.M. An adaptable FEA simulation of material extrusion additive manufacturing
heat transfer in 3D. Addit. Manuf. 2018, 21, 422–430. [CrossRef]
43. El Moumen, A.; Tarfaoui, M.; Lafdi, K. Modelling of the temperature and residual stress fields during 3D
printing of polymer composites. Int. J. Adv. Manuf. 2019, 104, 1661–1676. [CrossRef]
44. Zhou, Y.; Nyberg, T.; Xiong, G.; Liu, D. Temperature Analysis in the Fused Deposition Modeling Process.
2016, 678–682. [CrossRef]
45. Zhou, Y.; Lu, H.; Wang, G.; Wang, J.; Li, W. Voxelization modelling based finite element simulation and
process parameter optimization for Fused Filament Fabrication. Mater. Des. 2020, 187, 108409. [CrossRef]
46. McMillan, M.; Leary, M.; Brandt, M. Computationally efficient finite difference method for metal additive
manufacturing: A reduced-order DFAM tool applied to SLM. Mater. Des. 2017, 132, 226–243. [CrossRef]
47. Zhang, Y.; Shapiro, V. Linear-Time Thermal Simulation of As-Manufactured Fused Deposition Modeling
Components. J. Manuf. Sci. Eng. 2018, 140. [CrossRef]
48. Geng, P.; Zhao, J.; Wu, W.; Ye, W.; Wang, Y.; Wang, S.; Zhang, S. Effects of extrusion speed and printing speed
on the 3D printing stability of extruded PEEK filament. J. Manuf. Process 2019, 37, 266–273. [CrossRef]
49. Love, L.J.; Duty, C.E.; Post, B.K.; Lind, R.F.; Lloyd, P.D.; Kunc, V.; Peter, W.H.; Blue, C.A. Breaking Barriers in
Polymer Additive Manufacturing; Oak Ridge National Lab (ORNL): Oak Ridge, TN, USA, 2015.
50. Nunes dos Santos, W.; Mummery, P.; Wallwork, A. Thermal diffusivity of polymers by the laser flash
technique. Polym. Test. 2005, 24, 628–634. [CrossRef]
51. Bellehumeur, C.; Li, L.; Sun, Q.; Gu, P. Modeling of Bond Formation Between Polymer Filaments in the Fused
Deposition Modeling Process. J. Manuf. Process 2004, 6, 170–178. [CrossRef]
52. Wang, L.; Gardner, D.J. Contribution of printing parameters to the interfacial strength of polylactic acid
(PLA) in material extrusion additive manufacturing. Prog. Addit. Manuf. 2018, 3, 165–171. [CrossRef]
53. Solarski, S.; Ferreira, M.; Devaux, E. Characterization of the thermal properties of PLA fibers by modulated
differential scanning calorimetry. Polymer 2005, 46, 11187–11192. [CrossRef]
54. Khun, N.W.; Liu, E. Thermal, mechanical and tribological properties of
polycarbonate/acrylonitrile-butadiene-styrene blends. Polym. Eng. Sci. 2013, 33, 535–543. [CrossRef]
Materials 2020, 13, 4985 21 of 21

55. Brown, A.C.; de Beer, D. Development of a Stereolithography (STL) Slicing and G-Code Generation
Algorithm for an Entry Level 3-D Printer. In Proceedings of the AFRICON, Pointe-Aux-Piments, Mauritius,
9–12 September 2013; pp. 1–5. [CrossRef]
56. Recktenwald, G.W. Finite-difference approximations to the heat equation. Mech. Eng. 2004, 10, 1–27.
57. Lee, S.T.; Kareko, L.; Jun, J. Study of Thermoplastic PLA Foam Extrusion. J. Cell. Plast. 2008, 44, 293–305.
[CrossRef]
58. Lule, Z.; Kim, J. Thermally conductive and highly rigid polylactic acid (PLA) hybrid composite filled with
surface treated alumina/nano-sized aluminum nitride. Compos. Part A Appl. Sci. Manuf. 2019, 124, 105506.
[CrossRef]
59. Kousiatza, C.; Karalekas, D. In-situ monitoring of strain and temperature distributions during fused
deposition modeling process. Mater. Des. 2016, 97, 400–406. [CrossRef]
60. Costa, S.F.; Duarte, F.M.; Covas, J.A. Towards modelling of Free Form Extrusion: Analytical solution of
transient heat transfer. Int. J. Mater. Form. 2008, 1, 703–706. [CrossRef]

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional
affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).