Available online at www.sciencedirect.

com
Available online at www.sciencedirect.com

Available ScienceDirect
ScienceDirect
online atonline
Available www.sciencedirect.com
at www.sciencedirect.com

ScienceDirect
Procedia CIRP 00 (2019) 000–000
Procedia CIRP 00 (2019) 000–000
ScienceDirect www.elsevier.com/locate/procedia
www.elsevier.com/locate/procedia
Procedia CIRP 00 (2017)
Procedia 000–000
CIRP 84 (2019) 491–495
www.elsevier.com/locate/procedia

29th
29th CIRP
CIRP Design
Design 2019
2019 (CIRP
(CIRP Design
Design 2019)
2019)

Design
Design for
for AM: Contributions
AM:28th
Contributions from
from surface
surface
CIRP Design Conference,
finish,
finish,
May 2018,
part
part
Nantes,
geometry
geometry and
France and part
part
positioning
positioning
A new methodology
A. Castelão, B.
to
A.
analyze
R. Soares a the functional a and physical
a*, C. M. Machadoa, M. Leitea, A.J.M. Mourãoa
a architecture
a of
A. Castelão, B. A. R. Soares *, C. M. Machado , M. Leite , A.J.M. Mourão
existingUNIDEMI,
products
a
a for an assembly oriented product family identification
Departamento de Engenharia Mecânica e Industrial, Faculdade de Ciências e Tecnologia, Universidade NOVA de Lisboa
UNIDEMI, Departamento de Engenharia Mecânica e Industrial, Faculdade de Ciências e Tecnologia, Universidade NOVA de Lisboa
* Corresponding author. Tel.: +351962930535; E-mail address: ba.soares@fct.unl.pt
* Corresponding author. Tel.: +351962930535; E-mail address: ba.soares@fct.unl.pt
Paul Stief *, Jean-Yves Dantan, Alain Etienne, Ali Siadat
École Nationale Supérieure d’Arts et Métiers, Arts et Métiers ParisTech, LCFC EA 4495, 4 Rue Augustin Fresnel, Metz 57078, France
Abstract
Abstract
* Corresponding author. Tel.:
With the increase +33 3 87 37final
in production 54 30;parts
E-mail Additivepaul.stief@ensam.eu
by address: Manufacturing, the study of the surface finish and dimensional accuracy become
With the increase in production final parts by Additive Manufacturing, the study of the surface finish and dimensional accuracy become
critical, in order to determine if it is possible to avoid costly and time-consuming post-production, especially when using Fused deposition
critical, in order to determine if it is possible to avoid costly and time-consuming post-production, especially when using Fused deposition
Modelling of Polymers. This is especially true when the goal is functional parts, where the surface quality and dimensional accuracy may
Modelling of Polymers. This is especially true when the goal is functional parts, where the surface quality and dimensional accuracy may
preclude the use of FDM. With that in mind a study was performed to evaluate the surface finish and dimensional deviations of FDM parts,
preclude the use of FDM. With that in mind a study was performed to evaluate the surface finish and dimensional deviations of FDM parts,
Abstract
when considering multiple processing parameters, by way of a design of experiments coupled with response surface methodology. The results
when considering multiple processing parameters, by way of a design of experiments coupled with response surface methodology. The results
showed that the critical parameters were face orientation, layer height and printing speed, and equations predicting the surface finish and
showed that the critical parameters were face orientation, layer height and printing speed, and equations predicting the surface finish and
dimensional
In today’s businessaccuracy were obtained,
environment, with
the trend further more
towards experimental
productresults
variety validating the equations.
and customization With these equations
is unbroken. wedevelopment,
Due to this propose guidelines for of
the need
dimensional accuracy were obtained, with further experimental results validating the equations. With these equations we propose guidelines for
manufacturing
agile and part by
reconfigurable FDM whensystems
production surface emerged
finish and/or
to dimensional
cope with accuracy
various is of importance,
products and productso families.
that a designer
To knowsand
design what changesproduction
optimize can be
manufacturing part by FDM when surface finish and/or dimensional accuracy is of importance, so that a designer knows what changes can be
made
systems as to a part
well as toand what precautions
choose the optimal must be taken
product whenproduct
matches, printing said part, in orderare
to comply with a specified
most ofsurface finish/dimensional
made to a part and what precautions must be taken when printinganalysis methods
said part, in order to needed.
comply withIndeed,
a specified the known
surface methods aim to
finish/dimensional
tolerance.
analyze a product or one product family on the physical level. Different product families, however, may differ largely in terms of the number and
tolerance.
nature of components. This fact impedes an efficient comparison and choice of appropriate product family combinations for the production
© 2019
©
system. 2019
A new The
The Authors.
Authors.Published
methodology Published byby
is proposed Elsevier
ElsevierB.V.
to analyzeB.V.existing products in view of their functional and physical architecture. The aim is to cluster
© 2019 The Authors. Published by Elsevier B.V.
Peer-review under responsibility ofof
thethe
scientific committee of the
of CIRP Design Conference 2019 2019.
these Peer-review
Peer-review under
products in new
under responsibility
assembly oriented
responsibility of the scientific
product
scientific committee
families for the
committee the CIRP
of optimization
the CIRP Design
of
Design Conference
existing
Conferenceassembly
2019 lines and the creation of future reconfigurable
assembly systems. Based on Datum Flow Chain, the physical structure of the products is analyzed. Functional subassemblies are identified, and
Keywords: Surface Finish; Dimensional Tolerances; Fused Deposition Modelling (FDM); Design of Experiments (DoE); Design for Additive Manufacturing
Keywords: Surface Finish; Dimensional Tolerances; Fused Deposition Modelling (FDM); Design of Experiments (DoE); Design for Additive Manufacturing
a functional
(DfAM).analysis is performed. Moreover, a hybrid functional and physical architecture graph (HyFPAG) is the output which depicts the
(DfAM).
similarity between product families by providing design support to both, production system planners and product designers. An illustrative
example of a nail-clipper is used to explain the proposed methodology. An industrial case study on two product families of steering columns of
thyssenkrupp Presta France is then carried out to give a first industrial evaluation of the proposed approach.
© 20171. The
Introduction
Authors. Published by Elsevier B.V. applied so that the semi molten filament can then be extruded
1. Introduction applied so that the semi molten filament can then be extruded
Peer-review under responsibility of the scientific committee of the 28th CIRP to form
Design a part layer by2018.
Conference layer [8].
to form a part layer by layer [8].
Under the umbrella term of Additive Manufacturing (AM),
Under the umbrella term of Additive Manufacturing (AM),
thereAssembly;
Keywords: are 7 families
Design of technologies
method; that are used for turning
Family identification The manufacturing of end use parts introduces several new
there are 7 families of technologies that are used for turning The manufacturing of end use parts introduces several new
virtual 3d models to physical parts [1]. constraints to the production by AM technologies. These
virtual 3d models to physical parts [1]. constraints to the production by AM technologies. These
Although it started as a rapid prototyping technology, used constraints are:
Although it started as a rapid prototyping technology, used constraints are:
in conceptual design and manufacturing checks among others  Surface roughness - A typical end use part has to adhere
in conceptual
1. Introduction design and manufacturing checks among others  Surface roughness - A characteristics
typical end use manufactured
part has to adhere
[2], AM, with its increase in quality and decrease in cost, of theto product a maximum range and
value(s) of roughness, measured in (µm). and/or
[2], AM, with its increase in quality and decrease in cost, to a maximum value(s) of roughness, measured in (µm).
quickly disseminated throughout various industry clusters [3]. assembled
There in arethis system.
several wayInofthis context, the
measuring mainroughness,
surface challenge in
quickly disseminated throughout various industry clusters [3]. There are several way of measuring surface roughness,
Due AM to isthe
nowfast development
a technology used for in production
the domain of
of small modelling
with the andmostanalysis is nowused
commonly not being
only toRacope
whichwith single
is the
AM is now a technology used for production of small with the most commonly used being Ra which is the
communication
batches, or oneand ofana kind
ongoing
parts,trend
producingof digitization
end use partsand
batches, or one of a kind parts, producing end use parts for
for products, a
arithmeticlimited
meanproduct range
deviation.
arithmetic mean deviation.
or existing product families,
multiple products
digitalization, [4]–[7]. enterprises are facing important
manufacturing  also
but Dimensional
to be ableTolerances
to analyze and– Maximum
to compare deviation
productsallowed
to define
multiple products [4]–[7].  Dimensional Tolerances – Maximum deviation allowed
One in
challenges of such
today’stechnologies
market is Fused Deposition
environments: a Modelling new product
continuing for a given element.
families. It can be observed that classical existing
One of such technologies is Fused Deposition Modelling for a given element.
(FDM),
tendency widely
towards used in of
reduction theproduct
prosumer market duetimes
development to itsand  Geometrical
low product families Tolerances
are regrouped – in
Maximum
function deviation
of clients allowed
or features.
(FDM), widely used in the prosumer market due to its low  Geometrical Tolerances – Maximum deviation allowed
acquisition
shortened cost,
product as well asIninaddition,
lifecycles. companies with
there is low production However,
an increasing for a assembly
given feature geometry,
oriented productoften in relation
families with other
are hardly to find.
acquisition cost, as well as in companies with low production for a given feature geometry, often in relation with other
demandvolume and/or high
of customization, complexity
being at the parts.
same FDM is part
timeisinpart of
a global the features.
Onfeatures.
the product family level, products differ mainly in two
volume and/or high complexity parts. FDM of the
material extrusion family, in which a continuous filament of
competition
material with competitors
extrusion family, in allwhich
over the world. This
a continuous trend,of main characteristics: (i) the number of components and (ii) the
filament
thermoplastic material is fed into a printhead, where heat is
whichthermoplastic
is inducingmaterial is fed into a from
the development printhead,
macro where heat is type of components (e.g. mechanical, electrical, electronical).
to micro
markets, results
2212-8271 in The
© 2019 diminished lot sizes
Authors. Published due B.V.
by Elsevier to augmenting Classical methodologies considering mainly single products
2212-8271 © 2019 The Authors. Published by Elsevier B.V.
product varieties
Peer-review (high-volume
under responsibility of to
thelow-volume production)
scientific committee of the CIRP[1]. or solitary,
Design Conference 2019 already existing product families analyze the
Peer-review under responsibility of the scientific committee of the CIRP Design Conference 2019
To cope with this augmenting variety as well as to be able to product structure on a physical level (components level) which
identify possible optimization potentials in the existing causes difficulties regarding an efficient definition and
production system, it is important to have a precise knowledge comparison of different product families. Addressing this

2212-8271©©2017
2212-8271 2019The
The Authors.
Authors. Published
Published by Elsevier
by Elsevier B.V. B.V.
Peer-review
Peer-review under
under responsibility
responsibility of scientific
of the the scientific committee
committee of the of theCIRP
28th CIRP Design
Design Conference
Conference 2019.
2018.
10.1016/j.procir.2019.04.247
492 A. Castelão et al. / Procedia CIRP 84 (2019) 491–495
2 A. Castelão, B. A. R. Soares, C. M. Machado, M. Leite, A.J.M. Mourão / Procedia CIRP 00 (2019) 000–000

Although post processing is used in most cases, and several

techniques have been studied [9]–[11] the study of the
roughness and dimensional tolerances of net shape parts is of Table 1 – Blockstec Blocks one properties
interest for applications where such constraints are more
Printing Properties Blocks One
relaxed, i.e., when the requirements are less stringent.
Furthermore, given that the technology has seen a rapid Technology FDM
evolution new studies are required. Extruder 0.4 mm @ 180-260 ºC
Print Volume 200x200x200 mm
As such this work focuses on determining the surface Layer Thickness 50 to 300 microns
roughness, using Ra, and dimensional tolerances, (outputs), of
Print speed Up to 100 mm/s
an FDM machine using PLA as feedstock, and the
Material 1,75 mm PLA
manufacturing parameters (inputs) that influence both
tolerances. This is accomplished by performing a design of
experiments (DoE) to determine the influential parameters, The material used was Filkemp’s Poly Lactic Acid (PLA)
coupled with the response surface methodology (RSM), to PLA-N.
obtain predicting equations that correlate the influential The specimen shape for the roughness tests were that of a
parameters with the outputs. cube with 10 mm per side, which allowed for the
measurement of 3 faces per test (figure 2).
A DoE is a technique used to analyses several system
inputs and the response of the output(s) to the variations in
inputs [12]. By testing multiple possible inputs at once, it
avoids one factor at a time (OFAT) techniques, that are costly,
time consuming, and as important, cannot test for interactions
between inputs, which might be of relevance to a particular
system [13]. RSM is a statistical technique used to obtain the
relationship between the system inputs and outputs, for a
given range of input parameters, in order to understand the
system or in optimization studies, improving the system
response (outputs) by optimizing the inputs. Developed by
Figure 2 - Specimens used in tests
Box and Wilson [14], it uses a second-degree polynomial to
model the system response, that although an approximation, it
For the dimensional deviation tests a 3-axis test specimen
is of great help to understanding complex systems or when
was created. In this case since the goal was to measure
there is little information. It has seen use in the AM field, with
dimensional accuracy among the axis no rotation was
Taufik [11], using RSM to model the roughness of an ABS
performed (Figure 3).
part finished by laser; Mohamed [15], used RSM to study the
influence of several parameters on build cost and dimensional
errors; Mohamed [16], used RSM to optimize FDM
parameters to increase flexural modulus; and Rathee [17] used
FDM to study build time requirements.

2. Materials and methods

2.1. Printer, Material and Specimen

The printer used was a Blockstec Blocks One (Figure 1), a

CoreXY FDM printer with the specifications shown in Table Figure 3 – test specimens used in the measurements of dimensional
1. deviations. a) 5mm b) 20mm c)35mm d) 50mm e) 65mm.

2.2. Experimental procedure

The experimental procedure was as follows

 Identification of the likely important parameters
 Create a full factorial DoE.
 Measure the surface roughness/dimensional
variation from the nominal value
 Perform ANOVA to determine the important
Figure 1 – Blockstec Blocks One
Perform a RSM using the Central Composite
Design (CCD) to ensure rotatability, in order to
. obtain the response equation.
 A. Castelão, B. A. R. Soares, C. M. Machado, M. Leite, A.J.M. Mourão / Procedia CIRP 00 (2019) 000–000 3
A. Castelão et al. / Procedia CIRP 84 (2019) 491–495 493

3. Results and Discussion The tables with the results of the RSM are shown in the
appendix.
3.1. Screening experiments
3.2. RSM Roughness
The input variables considered to be of influence on the
surface finish and dimensional deviations are: The roughness results of the RSM are shown in Equation 1
𝑅𝑅𝑅𝑅𝑅𝑅𝑅𝑅ℎ𝑛𝑛𝑛𝑛𝑛𝑛𝑛𝑛 = 12.417 + 3.956𝐿𝐿𝐿𝐿 − 4.404𝛳𝛳 − 2.219 𝐿𝐿𝐿𝐿𝐿𝐿 (1)
 Infill Percentage (I%) With an Adjusted R2 of 77.25%.
 Layer Thickness (Lt)
 Print Speed (v) Equation 1 shows that the surface roughness decreases as
 Extrusion temperature (T) Lt decreases and ϴ increases. The reason for the first one is
 Size (Sz) (Dimensional Only) that with lower Layer thickness the smaller the difference
 Orientation (ϴ) relative to the print bed (roughness between peaks and valleys. Orientation wise, for very small
only) angles in the XoY plane the staircase effect of the layer by
layer deposition of AM is very pronounced which increases
With these parameters both DoE require 25 specimens with surface roughness.
5 specimens for each variable combination, resulting in a full The adjusted R2 is quite good, although there is a caveat. In
factorial DoE. With the DoE complete an Analysis of AM there are two orientations where the values of roughness
Variance (ANOVA) was performed to determine the variables are significantly smaller than for any given direction,
that have the most impact on the responses (Table 2). specifically 0º and 90º. At these orientations, the deposition
has no staircase effect which significantly decreases
Table 2 – Important variables after screening
roughness. Figure 4 shows this difference in roughness.
Responses
Roughness Dimensional
Lt Lt
Influencing
variables

ϴ v
v T
Sz

With the most important variables determined, the RSM

using a CCD was performed to obtain the equations predicting
the response behavior to changes in variables. The relation
between RSM Levels and the values of the variables are
shown in tables 3 and 4. Figure 4 – Roughness along ϴ with Lt at Level 0 (0.1 mm)

Table 3 – RSM levels and relation to variables (roughness tests)

As the figure shows the roughness is ever decreasing from

Input
Level 0º to 90º, but the real values at those points are smaller. Since
-1.682 -1 0 1 1.682 these points do not appear in any level of the RSM, this
0.05 0.7 0.1 0.13 0.15
discontinuity in surface roughness does not appear in the
Lt [mm]
equation, and as such the equation is not valid at 0º and 90º.
v [mm/s] 8 25 50 75 92

ϴ [°] -5 15 45 75 95 3.3. RSM dimensional deviation

The dimensional deviations results of the RSM are shown

Table 4 - RSM levels and relation to variables (dimensional variation tests) in Equation 2, 3 and 4.

Level 𝐷𝐷𝐷𝐷𝐷𝐷� = −0.024 − 0.079𝑆𝑆𝑆𝑆 + 0.042𝐿𝐿𝐿𝐿 − 0.018𝐿𝐿𝐿𝐿 � − 0.031𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆

Input
-2 -1 0 1 2 (2)
With an Adjusted R2 of 83.8%
Sz [mm] 5 20 35 50 65

Lt [mm] 0.05 0.1 0.15 0.2 0.25 𝐷𝐷𝐷𝐷𝐷𝐷� = 0.01 − 0.053𝑆𝑆𝑆𝑆 + 0.058𝐿𝐿𝐿𝐿 − 0.016𝐿𝐿𝐿𝐿 � + 0.013𝑣𝑣 −
v [mm/s] 10 30 50 70 90 0.0313𝑆𝑆𝑆𝑆𝑆𝑆𝑆𝑆 (3)
200 205 210 215 220 With an Adjusted R2 of 86.9%
T [°C]
494 A. Castelão et al. / Procedia CIRP 84 (2019) 491–495
4 A. Castelão, B. A. R. Soares, C. M. Machado, M. Leite, A.J.M. Mourão / Procedia CIRP 00 (2019) 000–000

𝐷𝐷𝐷𝐷𝐷𝐷� = 0.152 + 0.042𝐿𝐿𝐿𝐿 (4)

With an Adjusted R2 of 2.8% Regarding dimensional variation:

Analyzing the first two equations, the Adjusted R2 is high, For the least variation, important surfaces should be
with the main difference being that the response along the Y perpendicular to the XoY plane.
axis is influenced by printing speed which does not happen to For a more controlled variation part dimensions should be
the X axis, possibly due to a quirk in either the RSM or the somewhat large.
printer itself. Given this difference in equations, the difference If accuracy is important the Z axis should be avoided for
in response is quite low, although the real deviations in Y are those accurate measurements.
closer to the predicted values. The geometric deviations in Z Parts to be printed should be designed to be printed larger
have an Adjusted R2 of 2.8%, which means that the equation than the nominal dimensions due to the fact that most
is not a good predictor of the response. The causes are mainly variations are negative, which also allows that, if dimensional
due to how 3d printers work. Namely, since the printer has to accuracy would have to be improved, below the IT9
find 0, where the print bed is, a small change in this value is measured, post processing is a necessity.
bound to change the deviations. Furthermore, other effects
such as layer compression, non-orthogonality between the 4. Conclusions
print bed and the Z axis, and non-planar print bed, can also
cause problems. Finally, in cases where the tuning of the 0 This study aimed to obtain predicting equations for surface
point is done manually, as is the case with the printer used, is roughness and dimensional tolerances for parts produced by
possibly the biggest contributor to Z variations, which can be FDM, and to develop guidelines for Design for Additive
seen in table 6 for the Center point repetitions. Manufacturing. The results show that only a few of the
Finally, in all tested specimens there is no clear pattern to printing parameters have a statistically significant influence
the deviations, with some being positive, others negative, on the response, with the prediction equations show good
sometimes in the same batch of specimens, with the exception adjusted R2 values, except for the dimensional variation in the
of larger parts where the deviations are always negative. The Z axis.
tolerance falls between IT9 and IT12 in the X and Y This also allowed for the determination of the surface
directions, and over IT14 for the Z direction. These tolerances roughness and dimensional variation, for any given value of
are within the expected values for pre-processed polymer the influencing parameters.
parts, which has an impact in the guidelines. The predicting equations confirmed that the best
orientations for surface roughness are 0º or 90º although it
3.4. Guidelines was somewhat surprising to discover that the Ra at 90º is
smaller than at 0º. At low angles of orientation, the staircase
With these tests Guidelines for FDM regarding surface effect is much more prevalent than at higher angles, resulting
roughness and geometric deviations can be given. in higher roughness. Dimensionally the study showed that in
the Z axis, a correct bed leveling and setup is of paramount
Regarding improving surface roughness: importance, and that the 3d printer was relatively similar
values of dimensional variation in the XoY plane.
Parts should ideally be a collection of rectangular prisms
and should be orientated such as the biggest surface area is on The development of the predicting equations also allowed
the print bed, since it provides the smallest value of Ra. the development of a number of guidelines to improve surface
Chamfers should be used at 45º whenever possible, instead roughness and Dimensional tolerances, furthering the
of fillets, for a constant Ra, since the equations predict an knowledge of DfAM.
equal value in all directions of the chamfer.
Parts with curved faces should have those faces printed Acknowledgements
vertically (90º) to ensure constant surface roughness.
Cylinders should be printed vertical for the best possible This work was supported by FCT, through
values of surface roughness. UNIDEMI, project UID/EMS/00667/2019.
Layer height should be as small as possible.
If any face is angled with any other face, the part should be Appendix A. Screening experiment
positioned, insuring the angled surface is at least at 45º or
higher for the best results. Table 5 – DoE results for RSM (surface Roughness)
Figure 5 shows a part with all of the above considerations.
Inputs Responses

Test nº Lt [mm] V [mm/s] ϴ [º] Ra [µm]

1 0.07 25 15 7.101
2 0.13 25 15 11.117
3 0.07 75 15 6.557
4 0.13 75 15 9.233

Figure 5 – Example of a part using the guidelines

 A. Castelão et al. / Procedia CIRP 84 (2019) 491–495 495
A. Castelão, B. A. R. Soares, C. M. Machado, M. Leite, A.J.M. Mourão / Procedia CIRP 00 (2019) 000–000 5

5 0.07 25 75 14.224 References

6 0.13 25 75 21.383
[1] ASTM International, “ISO/ASTM 52900: 2015 Additive
7 0.07 75 75 15.262
manufacturing-General principles-terminology,” Geneva,
8 0.13 75 75 22.934 Switzerland, 2015.
9 0.05 50 45 9.452 [2] I. Gibson, D. Rosen, and B. (Brent) Stucker, Additive
10 0.15 50 45 19.715 manufacturing technologies : 3D printing, rapid prototyping, and
direct digital manufacturing. .
11 0.1 8 45 14.440
[3] M. K. Thompson et al., “Design for Additive Manufacturing:
12 0.1 92 45 14.029 Trends, opportunities, considerations, and constraints,” CIRP Ann.,
13 0.1 50 -5 14.485 vol. 65, no. 2, pp. 737–760, Jan. 2016.
14 0.1 50 95 7.317 [4] Y. Jin, Y. He, and A. Shih, “Process Planning for the Fuse
Deposition Modeling of Ankle-Foot-Othoses,” Procedia CIRP, vol.
15 0.1 50 45 16.199
42, pp. 760–765, 2016.
[5] S. Santos, B. Soares, M. Leite, and J. Jacinto, “Design and
Table 6 – DoE results for RSM (Deviations)
development of a customised knee positioning orthosis using low
cost 3D printers,” Virtual Phys. Prototyp., pp. 1–11, Aug. 2017.
Inputs Responses [6] D. Singh, R. Singh, K. S. Boparai, I. Farina, L. Feo, and A. K.
Test Sz Lt v T Dim X Dim Dim Verma, “In-vitro studies of SS 316 L biomedical implants prepared
nº [mm] [mm] [mm/s] [ºC] [mm] Y[mm] Z[mm]
by FDM, vapor smoothing and investment casting,” Compos. Part
1 20 0.2 70 215 0.18 0.17 0.29 B Eng., vol. 132, pp. 107–114, Jan. 2018.
2 50 0.2 70 215 0.07 0.03 0.35 [7] F. Kabirian, B. Ditkowski, A. Zamanian, R. Heying, and M.
Mozafari, “An innovative approach towards 3D-printed scaffolds
3 20 0.1 70 215 0.04 0.01 0.14
for the next generation of tissue-engineered vascular grafts,” Mater.
4 50 0.1 70 215 0.16 0.11 0.16 Today Proc., vol. 5, no. 7, pp. 15586–15594, 2018.
5 20 0.2 30 215 0.11 0.14 0.34 [8] C. K. Chua and K. F. Leong, 3D Printing and Additive
6 50 0.2 30 215 0.14 0.11 0.38 Manufacturing: Principles and Applications, 4th Ed. Singapore:
World Scientific, 2014.
7 20 0.1 30 215 0.01 0.04 0.12
[9] A. Boschetto, L. Bottini, and F. Veniali, “Finishing of Fused
8 50 0.1 30 215 0.15 0.12 0.1 Deposition Modeling parts by CNC machining,” Robot. Comput.
9 20 0.2 70 205 0.11 0.15 0.29 Integr. Manuf., vol. 41, pp. 92–101, Oct. 2016.
[10] R. Singh, S. Singh, I. P. Singh, F. Fabbrocino, and F. Fraternali,
10 50 0.2 70 205 0.16 0.04 0.24
“Investigation for surface finish improvement of FDM parts by
11 20 0.1 70 205 0.05 0.04 0.13 vapor smoothing process,” Compos. Part B Eng., vol. 111, pp. 228–
12 50 0.1 70 205 0.19 0.08 0.15 234, Feb. 2017.
[11] M. Taufik and P. K. Jain, “Laser assisted finishing process for
13 20 0.2 30 205 0.07 0.12 0.33
improved surface finish of fused deposition modelled parts,” J.
14 50 0.2 30 205 0.15 0.02 0.37 Manuf. Process., vol. 30, pp. 161–177, Dec. 2017.
15 20 0.1 30 205 0.07 0.07 0.12 [12] R. A. Fisher, “The Design of Experiments.,” Am. Math. Mon.,
1936.
16 50 0.1 30 205 0.17 0.11 0.14
[13] R. H. . A.-C. Myers, Response Surface Methodology : Process and
17 5 0.15 50 210 0.01 0.01 0.02 Product Optimization Using Designed Experiments. Wiley, 2016.
18 65 0.15 50 210 0.2 0.1 0.19 [14] G. E. P. Box and K. B. Wilson, “On the Experimental Attainment
19 35 0.05 50 210 0.16 0.17 0.15 of Optimum Conditions,” J. R. Stat. Soc. Ser. B, vol. 13, no. 1, pp.
1–45, 1951.
20 35 0.25 50 210 0.05 0.05 0.23
[15] O. A. Mohamed, S. H. Masood, and J. L. Bhowmik, “Parametric
21 35 0.15 10 210 0.07 0.02 0.15 Analysis of the Build Cost for FDM Additive Processed Parts
22 35 0.15 90 210 0.07 0.03 0.09 Using Response Surface Methodology,” Ref. Modul. Mater. Sci.
23 35 0.15 50 200 0.09 0.03 0.16 Mater. Eng., Jan. 2016.
[16] O. A. Mohamed, S. H. Masood, and J. L. Bhowmik, “Mathematical
24 35 0.15 50 220 0.07 0.04 0.15
modeling and FDM process parameters optimization using response
25* 35 0.15 50 210 0.01 0.02 0.06 surface methodology based on Q-optimal design,” Appl. Math.
26* 35 0.15 50 210 0.04 0.02 0.2 Model., vol. 40, no. 23–24, pp. 10052–10073, Dec. 2016.
27* 35 0.15 50 210 0.02 0 0.11 [17] S. Rathee, M. Srivastava, S. Maheshwari, and A. N. Siddiquee,
“Effect of varying spatial orientations on build time requirements
28* 35 0.15 50 210 0.05 0 0.3
for FDM process: A case study,” Def. Technol., vol. 13, no. 2, pp.
29* 35 0.15 50 210 0 0.01 0.09 92–100, Apr. 2017.
*Center point repetitions