The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328

https://doi.org/10.1007/s00170-022-08679-2

ORIGINAL ARTICLE

Multi‑parameter optimization of PLA/Coconut wood compound

for Fused Filament Fabrication using Robust Design
John D. Kechagias1   · Stephanos P. Zaoutsos2 · Dimitrios Chaidas1 · Nectarios Vidakis3

Received: 12 November 2021 / Accepted: 4 January 2022 / Published online: 11 January 2022
© The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2022

Abstract
This study investigates the effects of four variables during fused filament fabrication of organic biocompatible composite
material, PLA with coconut flour, at the ultimate tensile strength (UTS), and elasticity module (E) of the printed parts. The
parameter optimization uses Taguchi L18 design and regression models. The examined deposition variables are the layer
thickness, the nozzle temperature, the raster deposition angle, and filament printing speed. The effects of the above variables
on the strength of the parts are essential to enhance the mechanical response of the printed parts. The experimental outcomes
are investigated using the ANOM and ANOVA and modeled utilizing linear regression models. In addition, an independent
experiment was repeated three times at optimum parameters’ levels to evaluate the methodology, giving predictions errors
less than 3%. The observed results showed that the raster deposition angle dominates among the other variables in the studied
experimental area.

Keywords  Taguchi · Optimization · Additive manufacturing · 3D printing · FFF · FDM · Mechanical strength · PLA wood ·
Eco-friendly material · Sustainability

1 Introduction techniques [1–3]. Fused Filament Fabrication (FFF), also

Additive manufacturing (AM) is a sub-group of manufac-
turing processes that uses metallic, plastic, ceramic, and
composite materials in solid-state (powders, filaments,
sheets, bars, etc.) or fluids (liquid or droplets) that are depos-
ited using plenty of different deposition mechanisms and
Article Highlights
• Organic PLA wood material for eco-friendly additive
manufacturing
• Robust design with applying the L18 Taguchi orthogonal array
• Mechanical response optimized considerably after Taguchi
methodology
• Three evaluation experiments at optimum conditions with errors
less than 3%
* John D. Kechagias
jkechag@uth.gr
1
Department of Forestry, Wood Science and Design,
University of Thessaly, Karditsa 43100, Greece
2
Department of Energy Systems, University of Thessaly,
Larissa 41500, Greece
3
Mechanical Engineering Department, Hellenic
Mediterranean University, Heraklion 71410, Greece
known as Fused Deposition Modeling (FDM) or 3D printing
(3DP), is the most popular among AM processes due to the
simple deposition mechanism and the variety of available
materials that exist in the market. Nowadays, more innova-
tive materials and colors are extensively used in an eco-
friendly way in applications, including the furniture industry,
design, and fashion, among others [4–9].
The strength of the FFF parts is one of the main apprehen-
sions of the end-users when the material deposition param-
eters are selected before the printing process starts [10–12].
The researchers have studied the strength characteristics of
FFF parts for over two decades [1, 13–16]. Part orientation
(PO) inside the build space and the deposition parameters
affect the interlaminar strength between the deposited mate-
rial [17, 18] and, consequently, the FFF parts' mechanical
response and surface quality [19–24]. The raster Deposi-
tion Angle (DA) also affects the tensile stress of polylactic
acid (PLA) FFF parts; 0° raster DA results in higher tensile
stress than 45° and 90° [25]. Finally, the Layer Thickness
(LT) interacts with the part orientation and affects the tensile
strength of the ABS-FFF parts [26].
Many other studies investigated the effects of the layer
thickness, part orientation, raster deposition angle, raster

13
Vol.:(0123456789)
4318 The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328
width, and air gap (infill density) on the strength character-
istics of the FFF test specimen [14, 27, 28]. They concluded
that the strength of the FFF parts varies according to the
quality of the interlayer bond formation [27, 29]. The ther-
mal properties of the polymeric materials and the deposition
rates need optimization to achieve a robust bond interface
and better strength [30]. All the investigations conclude that
different printing parameters are appropriate for each other
material for optimizing the quality of FFF parts, such as
mechanical response and shape accuracy [31].
In the literature, the organic biocompatible composite,
PLA with wood compounds (PLA/W) investigated in dif-
ferent material synthesis [12, 32–35] and FFF 3D printing
conditions [36–38]. Faludi et al. [12] reported dependable
adhesion between PLA and wood flours, and they proved
that interfacial adhesion is strong in PLA/wood composites.
The use of para-wood powder derived from the furniture
industry was tested as an infilled material in the PLA matrix
in Chansoda et al. [35]. The authors showed that wood pow-
ders influence composite PLA/wood filament quality and
mechanical properties. Zandi et  al.[38] investigated the
fatigue behavior of a PLA/W (Timberfill, 8% wood fibers)
processed through FFF process parameters, i.e., the layer
thickness (LT), the nozzle diameter, the infill density (ID),
and the printing speed (PS). They used the Taguchi L27
orthogonal array and found that the LT was the most critical
parameter. In Kain et al. [19], two wood fiber contents (15
and 25%; Wood fibers, ARBOCEL® C100) mixed with PLA
(Ingeo™ 3251D) tested in different raster deposition angles
(DA) during FFF (0–90°; step 15°). The wood content of
25% gave better strength than 15%, and that the raster DA
affected FFF parts' mechanical response. The mechanical
performance of PLA/W parts, changing the nozzle tempera-
ture (NT) between 210 and 250 °C, was also investigated in
[39]. The increase in the NT from 210 to 230 °C improves
the strength slightly, which agrees with other studies con-
cerning FFF composites with PLA as a base material (see
[18, 31]. However, above 230 °C, the wood particle’s degra-
dation influences the strength and is not suggested. Ayrilmis
et al. [37] investigated the impact of LT on the water absorp-
tion and strength of FFF-PLA/W specimens. Increasing the
Fig. 1  Dogbone: (a) 2D views,
and (b) 3DP of the specimens
13
LT increases the water absorption and the cross-sectional
porosity, and the durability decreases. More, the density
of the printed PLA/W part increased as the printing speed
decreased, and the tensile and flexural strength altered sig-
nificantly by the deposition rates [36].
The above researchers investigated the influence of the
FFF parameters on the strength of different PLA wood flour
contents. In conclusion, to the authors' knowledge, it was not
reported any research in the literature where the influence of
four parameters, i.e., the LT, NT, raster DA, and PS, on the
mechanical response of the 3D printed PLA/W is studied.
Moreover, the proposed experimental area of the selected
parameters is investigated for the first time.
Therefore, this work is an experimental investigation of
how these four FFF parameters, the LT, NT, raster DA, and
PS, influence the ultimate tensile strength (UTS) and the
elasticity module (E) of the PLA mixed with coconut wood
flour FFF parts. The results were analyzed and presented
using statistical tools such as main effect plots (MEP), anal-
ysis of variances (ANOVA), interaction charts, and linear
regression models. So, the optimum parameters proposed
to optimize the strength of the organic biocompatible com-
posite FFF-PLA/W material.
2 Materials and methods
2.1 Preparation of experiment
According to ASTM D638 standards (see Fig. 1a), a dog-
bone was designed having a thickness of 4 mm and trans-
lated in STL format at the SolidWorks CAD program. Then,
the Craftbot Plus 3D printer (Fig. 1b) fabricated the speci-
mens. Its platform is made of aluminum, thus giving the
capability for printing specimens using different types of
material. The maximum volume of the vat was 250 X 200
X 200 mm, and the printing speed was up to 200 mm/s.
The commercially available material NEEMA3D™ WOOD-
PLUS consisted of 30% wood fibers of coconut and addi-
tives, and 70% pure PLA polymer was utilized. The filament
diameter was 1.75 mm. The specific gravity was 1.2 g/cc
The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328 4319

Fig. 2  (a) A view of a PLA/

Coconut organic composite
specimen mounted in the tensile
grips of the INSTRON 3382
universal testing machine,
and (b) broken specimens of
0.1 mm LT

(ASTM D1505), the melting point between 140–150°, and
the nozzle's diameter was 0.4 mm. The minimum contended
tensile strength and modulus for the NEEMA3D™ PLA
series filaments is 70 MPa 1900 MPa, respectively (ASTM
D882) as measured by the manufacturer.
The static tensile testing is performed on a strain-controlled
Instron 3382 universal testing machine with a load capacity of
100 kN, as per ISO, equipped with a gripping fixture as shown
in Fig. 2. The cross-head speed rate was 1 mm/min. At the
same time, force and displacement data for each test coupon
were recorded through a data acquisition system and stored
through BlueHill software for further treatment.
Figures  3a–f presents all graphs of eighteen experi-
ments of the L18 Taguchi orthogonal array. At first glance,
Figs. 3a–f shows that the raster DA most affects the UTS and
E. The 0° DA presents the higher UTS and E in all graphs,

Fig. 3  Stress–Strain curves
for all specimens: (a-c) with
LT 0.1 mm, and (d-f) with LT
0.3 mm

13
4320 The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328

Table 1  Parameters with levels strands of the woven pattern. Finally, the printing speed
Parameters Units Levels determines the time between layers and affects the inter-
laminar quality [30]. All other FFF process parameters were
1 2 3
kept constant in this study to minimize the noise error.
Variable parameters This study utilizes the L18 Taguchi Orthogonal Array
Layer Thickness (LT) (mm) 0.1 0.3 - (OA). This OA consists of eighteen arrays and eight columns
Nozzle Temperature (NT) (°C) 180 200 220 [44, 45]. The eighteen arrays show the number of the experi-
Deposition Angle (DA) (o) 0 45 90 ments, 18 experiments, and the columns show the possible
Printing Speed (PS) mm/s 30 40 50 variables [41]. We have used only 4 out of 8 columns, i.e.,
Constant parameters four parameters (LT, NT, DA, and PS). The process error
Travel speed mm/s 40 (e) is estimated by considering the empty columns. So, it is
Infill density % 100 unnecessary to repeat each combination of the 18 experi-
Bed temperature o
C 60 ments. Table 2 lists the variable parameters (LT, NT, DA and
Room temperature o
C 20 PS), the measured attributes (UTS and E), and the eighteen
Room humidity % 50 (18) fractional combinatorial experimental design accord-
Outline/perimeter shells/top Number 1/1/1 ing to the Taguchi L18 ­(21 × ­37) OA. A qualified individual
layers/bottom layers carefully did the eighteen mechanical response experiments
to eliminate noise errors.
Table  2 shows substantial max–min values between
and the 90° DA is the lowest. However, the effects of the eighteen experiments that were DA, LT, NT, and PS values
other processing parameters are not apparent, and statistical changing.
modeling should be applied to optimize the levels of the
parameters.
3 Results and discussion
2.2 Design of experiment
Table 2 tabulates all the results from the experiments con-
In the current study, the Taguchi L18 (2 1 × 3 7) experi- ducted in this work. The UTS is the higher tension a mate-
mental approach was employed [40–42]. This methodol- rial can sustain while being stressed before breaking. At the
ogy adopts balanced experiments according to Taguchi's same time, the elasticity module (E) is the linear proportion
proposed Orthogonal Arrays (OA). Although the number of between the stress and the strain in the elastic deformation
the executed experiments is a fraction of the full combina- zone of the stress–strain curve. Both UTS and E characterize
torial design, the methodology can find the best parameter the materials' withstand properties. Table 2 also summarizes
levels and construct predictive mathematical models [20, the basic statistics of the mechanical performance (mean,
43]. Finally, three validation experiments against the best min, max, spread). The spread values (max–min) of the UTS
conditions are applied to verify the experimental design and and E are considerable concerning the data set of the 18
evaluate the experiment's spread at the optimum conditions. samples. This variability is about 100% for the UTS and 69%
The first step in the above approach is selecting the vari- for E. Therefore, modeling the process is highly significant,
able parameters that affect the aspired attributes: the UTS reducing variability considerably. Finally, process optimiza-
and the elasticity module (E). This task is critical and has tion is vital for sustainable FFF printings [46].
two main concerns: (i) to select the appropriate variable It is worth here noting that the deposited beds' cooling
parameters and levels and (ii) all the experiments of the process affects the adhesion and fusion between sequential
decided orthogonal array (OA) to be achievable (should do layers, which, in turn, are affected by the LT, NT, and PS
all experiments contained in the OA). [10, 47–50]. Therefore, controlling the FFF process PS, NT,
Table 1 summarizes the selected variables and the con- and LT is conclusively significant [38, 51–54]. In Fig. 4,
stant parameters after the literature review in the introduc- side views and broken surfaces pictures are presented by
tion section and a 'trial-and-error' procedure. According to SEM for 0.3 mm LT, showing good interlaminar adhesion
the literature review, layer thickness and nozzle temperature and bonding. Scanning electron microscopy enabled detailed
(LT and NT) influence the dimensional accuracy, surface examination of the area of failure and the failure mode of
roughness, and mechanical response of the FFF parts [10]. the fabricated specimens. The morphological shape of the
In addition, raster deposition angle (DA) is also an essen- fractured cross-sectional surface implies a brittle mode of
tial parameter [18]. Furthermore, it affects the strength of failure (Fig. 4a-f). This fact verifies the already observed
the FFF parts, as this parameter defines the direction of the failure mode optically exhibited in tensile testing with no

13
The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328 4321

Table 2  Parameter Variable parameters, ­L18 ­(12 × ­73) Attributes measured

combinations and experimental
measurements 1 2 3 4 5 6 7 8
o o
No exp LT (mm) NT ( C) DA ( ) PS (mm/s) Empty Columns UTS (Mpa) E (MPa)
1 0.1 180 0 30 1 1 1 1 17.420 692.62

2 0.1 180 45 40 2 2 2 2 7.043 410.59

3 0.1 180 90 50 3 3 3 3 5.480 347.95
4 0.1 200 0 30 2 2 3 3 15.410 627.77
5 0.1 200 45 40 3 3 1 1 10.240 448.88
6 0.1 200 90 50 1 1 2 2 7.745 375.92
7 0.1 220 0 40 1 3 2 3 16.920 651.88
8 0.1 220 45 50 2 1 3 1 12.580 478.50
9 0.1 220 90 30 3 2 1 2 10.010 445.97
10 0.3 180 0 50 3 2 2 1 16.920 678.84
11 0.3 180 45 30 1 3 3 2 12.145 425.52
12 0.3 180 90 40 2 1 1 3 8.970 374.89
13 0.3 200 0 40 3 1 3 2 15.420 624.65
14 0.3 200 45 50 1 2 1 3 12.080 468.83
15 0.3 200 90 30 2 3 2 1 10.010 402.12
16 0.3 220 0 50 2 3 1 2 15.210 611.23
17 0.3 220 45 30 3 1 2 3 12.007 483.17
18 0.3 220 90 40 1 2 3 1 10.150 431.88
Average 11.987 498.96
Min 5.480 347.95
Max 17.420 692.62
Spread 11.94 344.67

necking phenomena. Figure 4a, b presents an SEM magnifi-
cation of the side view of a specimen where illustrated both
the number and length of PLA sequenced layers. Moreover,
in Fig. 4c, d, the brittle nature of the failure is also confirmed
by the resulting surface texture.
3.1 Effects of the variable parameters on UTS and E
The diagrams that show the effects of the variable param-
eters on an attribute are known as ANOM diagrams (analy-
sis of means) or main effect plots (MEP). Such diagrams
explain the influence of each variable graphically in consid-
eration of the attribute measure. So, we quickly observe the
optimum parameter levels and the trend lines. For example,
by employing these plots, the effects of the four process
parameters (LT, NT, DA, and PS) on both the UTS and E
can be extracted (see: Fig. 5).
The MEP plots showed that the dominant parameter in
the utilized experimental space for the UTS and E measures
is the raster DA. The three other variables are significant at
a lower level. According to the MEP plots, the optimized
values for all the variables are zero raster DA, 30 mm/s PS,
220 °C, and 0.3 mm LT; UTS and E are edified as 'the maxi-
mum the best' attributes.
A way for identifying how one parameter interacts with
another is by utilizing the interaction plots [44, 53, 55–57].
For example, Fig. 6a, b shows the LT, NT, DA, and PS inter-
actions concerning the UTS and E accordingly. In addition,
these two plots show a strong interaction between the domi-
nant parameter raster DA and the others, PS, NT, and LT.
Analytically, the interaction plots showed that: (i) LT
interacts synergistically with NT and DA, i.e., the trend lines
have similar directions, (ii) LT interacts anti-synergistically
with PS, i.e., the trend lines have complex directions, (iii)
DA interacts synergistically with NT and PS, and (iv) NT
and PS interact anti-synergistically.
After the MEP and interaction plots, the ANOVA is
the qualitative tool to investigate the importance of each
variable parameter on the quality attributes (here, the UTS
and E). Analysis of variances (ANOVA) decomposes the
errors of each variable on the total error. ANOVA fits
linear or quadratic mathematical models and studies the
least-squares differences. The 'MEP' plots showed that
the raster DA is the dominant parameter in our case by
utilizing linear mathematical models for the 'UTS' and 'E'
analysis of variances (see: Tables 3 and 4, respectively).
The ANOVA showed that the raster DA variable affects
80.1% and 92.6% on UTS and E, respectively. The three

13
4322 The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328

Fig. 4  SEM images of the

longitudinal layered side of
specimens (a, b) and cross-sec-
tional fracture surface (c, d, e, f)
in SPC13 0.3–200-0–40

other variables (PS, NT, and LT) are insignificant for UTS
and E, respectively. The F values are below two, and the P
values are higher than 0.05.
The outcomes occurring from Figs. 5 and 6 as well as of
Tables 3 and 4 conclude in the following:
The raster DA parameter dominates both the UTS and
E values; 80.1% and 92.6%. The F values are higher than
four (F > 4) and P values smaller than 0.05 (P < 0.05), which
means this conclusion is statistically significant. The best
value for the raster DA is 0°, and the worst the 90°. The
interactions of raster DA with the other parameters are
slightly synergistic (Fig. 6). The mechanical response (UTS
and E) increases by decreasing the raster deposition angle.
It means physically that the 0° aligned filament shows the
highest UTS and E values.
The other three parameters, NT, PS, and LT, are less
important for the utilized experimental space. The F values
are smaller than 2, and P values are higher than 0.05 (F < 2
and P > 0.05). The F value for the LT concerning the UTS
attribute is between 2 and 4 (2 < F < 4), which statistically
means that it is more significant than the NT and PS. These
statistical values show that the PS, NT, and LT affect the
optimal thermal bonding condition in an anti-synergistic
way (see interaction charts, Fig. 6).

Fig. 5  Main effects plot for (a)

UTS and (b) E

13
The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328 4323

Fig. 6  Interaction plots between DA, PS, NT, and LT according to the: (a) UTS and (b) E

Considering previous experimental work and knowing
that the interlaminar bonding conditions are affected ini-
tially by the PS and NT, and secondly, by the LT [58], we
decided to construct two 'MEP' plots for 0.1 and 0.3 mm LT
(see Fig. 7). These MEP plots decompose better the influ-
ence of raster DA, NT, and PS according to each LT level.
Figure 7a–d shows that the NT is a significant parameter for
0.1 mm LT and minor for 0.3 mm LT, denoting that the inter-
laminar conditions are different in the case of 0.3 and 0.1 LT.
Specifically, the deposited strands cooling process is
affected more by the PS than the NT for 0.3 LT, while for 0.1
LT is affected evenly by NT and PS. Furthermore, Coogan
and Kazmer [29] proved that the longer the material stays at
a higher temperature than its glass transition level, the better
the bond becomes. Therefore, increasing NT or decreasing
PS leads to a rise in UTS and E for 0.1 mm LT, which fol-
lows the finding of a similar investigation for the same range
of PS values, 30–50 mm/s [59].
The NT has different trend lines for 0.1 and 0.3 LT. In
the case of 0.1 LT, when the NT increases, the UTS and E
increase (Fig. 7a, b). On the other hand, for 0.3 LT, the NT
is insignificant for both UTS and E (Fig. 7c, d).
Finally, concerning the PS parameter, Fig. 7a–d shows a
significant influence for both UTS and E attributes for each
of the two LT (0.1 and 0.3 correspondingly). By increasing
the PS, the UTS and E decrease for the 0.1 mm LT while
rising for the 0.3 mm LT. In the total experimental space
(Fig. 5), the trend line of the PS is similar to that of the
0.1 LT (Fig. 7a, b). The ANOVA shows that the PS is an
insignificant parameter (F < 2 and P > 0.05; Tables 3 and

Table 3  ANOVA: UTS versus Source DF Adj SS Adj MS F-Value P-Value %

DA; PS; NT; LT
DA 2 176.746 88.373 35.68 0.000 80.1%
PS 2 6.592 3.296 1.33 0.307 3.0%
NT 2 6.857 3.428 1.38 0.295 3.1%
LT 1 5.627 5.627 2.27 0.163 2.6%
Error 10 24.766 2.477 11.2%
Total 17 220.587 100.0%
R2 88.77%
R2 adj 80.91%
R2 pred 63.62%

13
4324 The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328

Table 4  ANOVA: E versus DA; Source DF Adj SS Adj MS F-Value P-Value %

PS; NT; LT
DA 2 208,926 104,463 88.24 0.000 92.6%
PS 2 1769 884 0.75 0.498 0.8%
NT 2 2991 1495 1.26 0.324 1.3%
LT 1 25 25 0.02 0.888 0.0%
Error 10 11,838 1184 5.2%
Total 17 225,548 100.0%
R2 94.75%
R2 adj 91.08%
R2 pred 82.99%

4) following Zandi et al. [38]. Remark that Fig.  7a–d is
employed to analyze MEP plots for the two different LTs
as explained in the text to construct more accurate linear
regression models and not replace Fig. 5a, b. Process opti-
mization employs the L18 OA and not Fig. 7a–d.
3.2 Modeling and validation
In the case of our optimization problem, it is evident that
the raster DA affects the most in the proposed experimental
area and that the other three parameters have minor effects.
ANOVA (see Tables 3 and 4) validates these results, where
the F-values of LT, PS, and NT are all below the number 4,
which statistically means that they have minor significance.
The only significant parameter according to Tables 3 and
4 is the raster DA having F-values higher than four and
P-values lower than 0.05.
Considering the ANOVA (Tables 3 and 4) and the MEP
plots (Figs.  5 and 7) as well as the interactions charts
(Fig. 6), linear regression models were developed for both
the UTS and E for the 0.1 and 0.3 mm LT, respectively
(Eqs. 1–4).
UTSLT=0.1mm =2.84 − 0.0889 ∗ DA − 0.0839 ∗ PS
(1)
+ 0.0797 ∗ NT ± e
ELT=0.1mm = 549 − 2.570 ∗ DA − 3.62 ∗ PS + 1.04 ∗ NT ± e
(2)
UTSLT=0.3mm =15.89 − 0.06615 ∗ DA + 0.0186 ∗ PS
(3)
− 0.0056 ∗ NT ± e

Fig. 7  MEP plots: (a) DA, NT

and PS vs UTS for 0.1 LT, (b)
DA, NT and PS vs E for 0.1 LT,
(c) DA, NT and PS vs UTS for
0.3 LT and (d) DA, NT and PS
vs E for 0.3 LT

13
The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328 4325

Table 5  Validation tests No LT NT DA PS UTS E UTS E UTS E

Actual Predicted (Eqs. 1 & 2) error

1 0.1 180 0 30 17.42 692.6 14.66 627.6 -15.8% -9.4%

2 0.1 180 45 40 7.043 410.6 9.82 475.7 39.6% 15.9%
3 0.1 180 90 50 5.48 348.0 4.99 323.9 -8.9% -6.9%
4 0.1 200 0 30 15.41 627.8 16.26 648.4 5.5% 3.3%
5 0.1 200 45 40 10.24 448.9 11.42 496.5 11.6% 10.6%
6 0.1 200 90 50 7.745 375.9 6.58 344.7 -15.0% -8.3%
7 0.1 220 0 40 16.92 651.9 17.01 633 0.6% -2.9%
8 0.1 220 45 50 12.58 478.5 12.17 481.1 -3.2% 0.6%
9 0.1 220 90 30 10.01 446.0 9.85 437.9 -1.5% -1.8%
Predicted (Eqs. 3 & 4)
10 0.3 180 0 50 16.92 678.8 15.81 620.5 -6.5% -8.6%
11 0.3 180 45 30 12.145 425.5 12.465 471.1 2.6% 10.7%
12 0.3 180 90 40 8.97 374.9 9.67 385.2 7.8% 2.8%
13 0.3 200 0 40 15.42 624.7 15.51 607.2 0.6% -2.8%
14 0.3 200 45 50 12.08 468.8 12.725 521.3 5.3% 11.2%
15 0.3 200 90 30 10.01 402.1 9.37 371.8 -6.3% -7.5%
16 0.3 220 0 50 15.21 611.2 15.58 636.2 2.5% 4.1%
17 0.3 220 45 30 12.007 483.2 12.235 486.7 1.9% 0.7%
18 0.3 220 90 40 10.15 431.9 9.44 400.9 -6.9% -7.2%
Evaluation experiments Predicted (Eqs. 3 & 4)
19 0.3 220 0 30 14.93 591.1 15.23 593.8 2.0% 0.5%
20 0.3 220 0 30 14.79 579.4 15.23 593.8 3.0% 2.5%
21 0.3 220 0 30 15.28 578.8 15.23 593.8 -0.3% 2.6%
Average 15.23 593.8 15.23 593.8 1.5% 1.8%

ELT=0.3mm = 444 − 2.379 ∗ DA + 2.12 ∗ PS + 0.392 ∗ NT ± e that the models accurately predict both the UTS and the E
(4) attributes.
Note that the validation experiment is repeated three
Table 5 summarizes the model's predictions (Eqs. 1–4)
times at the optimum levels for the UTS and E according to
and error percentages. Finally, three evaluation experiments
the MEP plot for all 18 experiments (Fig. 5; raster DA = 0°,
are executed to validate the developed regression models
PS = 30 mm/s, NT = 220°, and LT = 0.3 mm). Therefore,
(Table 5; Fig. 8). The specimens for the validation experi-
Fig. 5 should be considered for the process optimization as
ments were built with the optimized parameters according
it has the orthogonality property and not Fig. 7a–d. Perceive
to the MEP in Fig. 5 (zero raster DA, 0.3 mm LT, 30 mm/s
that the orthogonality property is described in ref [44] and
PS and 220 °C NT). All three experiments give accurate
means that all combinations between two columns exist in
predictions with an accuracy lower than ± 5%, which means
an orthogonal array and appear evenly.

Fig. 8  Validation experiments
and strain–stress curves for the
optimized parameter values
(0.3mm LT, 220 °C NT, 0­ o DA,
and 30 mm/s PS)

13
4326 The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328
4 Conclusions
In this experimental research, four variable parameters (LT,
NT, DA, and PS) are examined as far as their effect on the
ultimate tensile strength (UTS) and the elasticity module
(E) of FFF specimens with PLA wood flour (70–30%) as
filament material is concerned.
The observed UTS and E values were between
5.48–17.42 MPa and 347–692 MPa, accordingly for solid
PLA/Coconut (70%PLA/ 30%Coconut) parts. In Afrose
et al. [25], FFF of solid specimens of pure PLA remarked
values between 31–39 MPa and 1246–1538 MPa for the
UTS and E, respectively. The above indicates that UTS and
E of the composite PLA/C are considerably smaller than
that of pure PLA. In Afrose et al. [25], zero raster DA also
optimizes the UTS for pure PLA, following the observations
of the current research for the PLA/C composite material.
Additionally, it is notable that the specifications of the used
PLA/C material in filament form have UTS and E minimum
values of about 70 and 1900 MPa, respectively, denoting that
the PLA/C FFF parts have the 25% and 36% UTS and E of
the filament specifications correspondingly.
The experimental results were analyzed using the main
effect plots (MEP), interaction diagrams, and ANOVA and
conclude the following:
• The raster DA parameter dominates on both the UTS and
E values, while when this parameter increases, both UTS
and E attribute decrease. The DA influences the UTS
and E about 80% and 90%. Physically the zero-oriented
filament shows higher UTS and E values like in fiber
composites. This conclusion follows the literature [25].
• The optimal thermal bonding condition is affected by
the PS, NT, and LT in an anti-synergistic way (see
interaction charts, Fig. 6) but trivially.
• The PS parameter affects the UTS and E slightly. When
decreases, it improves UTS and E for the 0.1 LT and
decreases the UTS and E for the 0.3 LT.
• The NT parameter insignificantly affects both the UTS
and E for the 0.3 LT. Concerning the 0.1 LT, both the
UTS and E increase slightly when the NT increases.
• The LT is a significant parameter but is highly depend-
ent on the interlaminar bonding condition (PS and NT)
and should be studied individually.
• Finally, the developed linear regression models proved
adequate, having an accuracy better than 5% by three
independent validation experiments.
The three repetitions in optimum parameters’ levels
prove that: (i) the spread values are very close, so the
optimization is validated (in optimum conditions, the
spread values minimized), and (ii) the regression model
13
gives accurate predictions, so the Taguchi methodology
is validated. In case that we had higher spreads or inac-
curate predictions, the optimization would have flaws. In
this case, a new design would be needed.
As future work, the authors propose a multi-parame-
ter multi-objective optimization of the PLA/wood parts,
including more wood flour types and wt. %.
Authors contribution  Conceptualization was contributed by JK;
Methodology was contributed by JK, SZ and DC; Formal analysis
and investigation were contributed by JK, SZ, DC and NV; Writ-
ing—original draft preparation, was contributed by JK; Writing—
review and editing, was contributed by JK, SZ; and NV; Resources
were contributed by JK and SZ; Supervision was contributed by JK;
All authors have given approval to the final version of the manuscript.
Availability of data and material (data transparency)  All data generated
or analyzed during this study are included in this published article.
Declarations 
Ethics approval  Not applicable.
Consent to participate  Not applicable.
Consent for publication  Not applicable.
Conflict of interest  The authors have no competing interests to declare
that are relevant to the content of this article.
Authors declaration  The submitted work is original and has not been
published elsewhere in any form or language.
References
1. Wong KV, Hernandez A (2012) A Review of Additive Manufac-
turing. ISRN Mech Eng 2012:1–10. https://d​ oi.o​ rg/1​ 0.5​ 402/2​ 012/​
208760
2. Oliveira JP, Santos TG, Miranda RM (2020) Revisiting fundamen-
tal welding concepts to improve additive manufacturing: From
theory to practice. Prog Mater Sci 107:100590. https://​doi.​org/​
10.​1016/j.​pmats​ci.​2019.​100590
3. Oliveira JP, LaLonde AD, Ma J (2020) Processing parameters in
laser powder bed fusion metal additive manufacturing. Mater Des
193:108762. https://​doi.​org/​10.​1016/j.​matdes.​2020.​108762
4. Pringle AM, Rudnicki M, Pearce JM (2018) Wood Furniture
Waste–Based Recycled 3-D Printing Filament. For Prod J 68:86–
95. https://​doi.​org/​10.​13073/​FPJ-D-​17-​00042
5. Yap YL, Yeong WY (2014) Additive manufacture of fashion and
jewellery products: a mini review. Virtual Phys Prototyp 9:195–
201. https://​doi.​org/​10.​1080/​17452​759.​2014.​938993
6. Goh GD, Yap YL, Tan HKJ et al (2020) Process–structure–properties
in polymer additive manufacturing via material extrusion: A review.
Crit Rev Solid State Mater Sci 45:113–133. https://​doi.​org/​10.​1080/​
10408​436.​2018.​15499​77
7. Aida HJ, Nadlene R, Mastura MT et al (2021) Natural fibre fila-
ment for Fused Deposition Modelling (FDM): a review. Int J Sus-
tain Eng. https://​doi.​org/​10.​1080/​19397​038.​2021.​19624​26
The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328
8. Vidakis N, Petousis M, Savvakis K et al (2019) A comprehensive
investigation of the mechanical behavior and the dielectrics of
pure polylactic acid (PLA) and PLA with graphene (GnP) in fused
deposition modeling (FDM). Int J Plast Technol 23:195–206.
https://​doi.​org/​10.​1007/​s12588-​019-​09248-1
9. Vidakis N, Petousis M, Velidakis E et al (2020) Three-dimensional
printed antimicrobial objects of polylactic acid (PLA)-Silver nano-
particle nanocomposite filaments produced by an in-situ reduction
reactive melt mixing process. Biomimetics 5:42. https://​doi.​org/​10.​
3390/​biomi​metic​s5030​042
10. Chaidas D, Kechagias JD (2021) An investigation of PLA/W parts
quality fabricated by FFF. Mater Manuf Process. https://​doi.​org/​
10.​1080/​10426​914.​2021.​19441​93
11. Vidakis N, Petousis M, Maniadi A et al (2020) Sustainable addi-
tive manufacturing: mechanical response of acrylonitrile-butadiene-
styrene over multiple recycling processes. Sustainability 12:3568.
https://​doi.​org/​10.​3390/​su120​93568
12. Faludi G, Dora G, Renner K et al (2013) Improving interfacial
adhesion in pla/wood biocomposites. Compos Sci Technol 89:77–
82. https://​doi.​org/​10.​1016/j.​comps​citech.​2013.​09.​009
13. Ćwikła G, Grabowik C, Kalinowski K et al (2017) The influence
of printing parameters on selected mechanical properties of FDM/
FFF 3D-printed parts. IOP Conf Ser: Mater Sci Eng 227:012033.
https://​doi.​org/​10.​1088/​1757-​899X/​227/1/​012033
14. Mohan N, Senthil P, Vinodh S, Jayanth N (2017) A review on
composite materials and process parameters optimisation for
the fused deposition modelling process. Virtual Phys Prototyp
12:47–59. https://​doi.​org/​10.​1080/​17452​759.​2016.​12744​90
15. Bikas H, Stavropoulos P, Chryssolouris G (2016) Additive manu-
facturing methods and modelling approaches: a critical review.
Int J Adv Manuf Technol 83:389–405. https://​doi.​org/​10.​1007/​
s00170-​015-​7576-2
16. Vidakis N, Petousis M, Velidakis E et al (2020) On the strain rate
sensitivity of fused filament fabrication (FFF) processed PLA,
ABS, PETG, PA6, and PP thermoplastic polymers. Polymers
12:2924. https://​doi.​org/​10.​3390/​polym​12122​924
17. Durgun I, Ertan R (2014) Experimental investigation of FDM
process for improvement of mechanical properties and produc-
tion cost. Rapid Prototyp J 20:228–235. https://​doi.​org/​10.​1108/​
RPJ-​10-​2012-​0091
18. Kechagias JD, Vidakis N, Petousis M (2021) Parameter effects
and process modeling of FFF-TPU mechanical response. Mater
Manuf Process. https://​doi.​org/​10.​1080/​10426​914.​2021.​20015​23
19. Kain S, Ecker JV, Haider A et al (2020) Effects of the infill pat-
tern on mechanical properties of fused layer modeling (FLM) 3D
printed wood/polylactic acid (PLA) composites. Eur J Wood Wood
Prod 78:65–74. https://​doi.​org/​10.​1007/​s00107-​019-​01473-0
20. Fountas NA, Kechagias JD, Manolakos DE, Vaxevanidis NM
(2020) Single and multi-objective optimization of FDM-based
additive manufacturing using metaheuristic algorithms. Procedia
Manuf 51:740–747. https://d​ oi.o​ rg/1​ 0.1​ 016/j.p​ romfg.2​ 020.1​ 0.1​ 04
21. Fountas NA, Kitsakis K, Aslani K-E et al (2021) An experimental
investigation of surface roughness in 3D-printed PLA items using
design of experiments. Proc Inst Mech Eng J: J Eng Tribol. https://​
doi.​org/​10.​1177/​13506​50121​10593​06
22. Kechagias J, Kitsakis K, Zacharias A et al (2021) Direct 3D Printing
of a hand splint using Reverse Engineering. IOP Conf Ser: Mater Sci
Eng 1037:012019. https://​doi.​org/​10.​1088/​1757-​899X/​1037/1/​012019
23. Chaidas D, Kitsakis K, Kechagias J, Maropoulos S (2016) The
impact of temperature changing on surface roughness of FFF pro-
cess. IOP Conf Ser: Mater Sci Eng 161:012033. https://​doi.​org/​
10.​1088/​1757-​899X/​161/1/​012033
24. Aslani K-E, Chaidas D, Kechagias J et al (2020) Quality per-
formance evaluation of thin walled PLA 3D printed parts using
the taguchi method and grey relational analysis. J Manuf Mater
Process. https://​doi.​org/​10.​3390/​jmmp4​020047
4327
25. Afrose MF, Masood SH, Iovenitti P et al (2016) Effects of part
build orientations on fatigue behaviour of FDM-processed PLA
material. Prog Addit Manuf 1:21–28. https://​doi.​org/​10.​1007/​
s40964-​015-​0002-3
26. Vidakis N, Vairis A, Petousis M et al (2016) Fused deposition
modelling parts tensile strength characterisation. Academic J
Manuf Eng 14
27. Gurrala PK, Regalla SP (2014) Multi-objective optimisation of
strength and volumetric shrinkage of FDM parts. Virtual Phys Pro-
totyp 9:127–138. https://​doi.​org/​10.​1080/​17452​759.​2014.​898851
28. Srivastava M, Rathee S (2018) Optimisation of FDM process
parameters by Taguchi method for imparting customised proper-
ties to components. Virtual Phys Prototyp 13:203–210. https://d​ oi.​
org/​10.​1080/​17452​759.​2018.​14407​22
29. Coogan TJ, Kazmer DO (2017) Bond and part strength in fused
deposition modeling. Rapid Prototyp J 23:414–422. https://​doi.​
org/​10.​1108/​RPJ-​03-​2016-​0050
30. Vanaei HR, Shirinbayan M, Deligant M et al (2021) In-Process
Monitoring of Temperature Evolution during Fused Filament Fab-
rication: A Journey from Numerical to Experimental Approaches.
Thermo 1:332–360. https://​doi.​org/​10.​3390/​therm​o1030​021
31. Chen K, Yu L, Cui Y et al (2021) Optimization of printing param-
eters of 3D-printed continuous glass fiber reinforced polylactic
acid composites. Thin-Walled Struct 164:107717. https://​doi.​org/​
10.​1016/j.​tws.​2021.​107717
32. Motas JG, Gorji NE, Nedelcu D et al (2021) XPS, SEM, DSC and
Nanoindentation Characterization of Silver Nanoparticle-Coated
Biopolymer Pellets. Appl Sci 11:7706. https://​doi.​org/​10.​3390/​
app11​167706
33. Bhagia S, Lowden RR, Erdman D et al (2020) Tensile properties of
3D-printed wood-filled PLA materials using poplar trees. Appl Mater
Today 21:100832. https://​doi.​org/​10.​1016/j.​apmt.​2020.​100832
34. Bulanda K, Oleksy M, Oliwa R et al (2020) Biodegradable poly-
mer composites based on polylactide used in selected 3D tech-
nologies. Polimery 65:557–562. https://d​ oi.o​ rg/1​ 0.1​ 4314/​polim​ery.​
2020.7.8
35. Chansoda K, Suwanjamrat C, Chookaew W (2020) Study on
processability and mechanical properties of parawood-powder
filled PLA for 3D printing material. IOP Conf Ser: Mater Sci Eng
773:012053. https://​doi.​org/​10.​1088/​1757-​899X/​773/1/​012053
36. Yang T-C, Yeh C-H (2020) Morphology and Mechanical Properties
of 3D Printed Wood Fiber/Polylactic Acid Composite Parts Using
Fused Deposition Modeling (FDM): The Effects of Printing Speed.
Polymers 12:1334. https://​doi.​org/​10.​3390/​polym​12061​334
37. Ayrilmis N, Kariz M, Kwon JH, Kitek Kuzman M (2019) Effect
of printing layer thickness on water absorption and mechanical
properties of 3D-printed wood/PLA composite materials. Int J
Adv Manuf Technol 102:2195–2200. https://​doi.​org/​10.​1007/​
s00170-​019-​03299-9
38. Zandi MD, Jerez-Mesa R, Lluma-Fuentes J et al (2020) Experi-
mental analysis of manufacturing parameters’ effect on the flex-
ural properties of wood-PLA composite parts built through FFF.
Int J Adv Manuf Technol 106:3985–3998. https://d​ oi.o​ rg/1​ 0.1​ 007/​
s00170-​019-​04907-4
39. Guessasma S, Belhabib S, Nouri H (2019) Microstructure and
mechanical performance of 3D printed wood-PLA/PHA using
fused deposition modelling: Effect of printing temperature. Poly-
mers 11:1778. https://​doi.​org/​10.​3390/​polym​11111​778
40. Tsiolikas A, Mikrou T, Vakouftsi F et al (2019) Robust design
application for optimizing ABS fused filament fabrication process:
A case study. IOP Conf Ser: Mater Sci Eng 564:012021. https://​
doi.​org/​10.​1088/​1757-​899X/​564/1/​012021
41. Kechagias J, Petropoulos G, Vaxevanidis N (2012) Application
of Taguchi design for quality characterization of abrasive water
jet machining of TRIP sheet steels. Int J Adv Manuf Technol
62:635–643. https://​doi.​org/​10.​1007/​s00170-​011-​3815-3
13
4328 The International Journal of Advanced Manufacturing Technology (2022) 119:4317–4328
42. Zhang JZ, Chen JC, Kirby ED (2007) Surface roughness optimiza-
tion in an end-milling operation using the Taguchi design method.
J Mater Process Technol 184:233–239. https://​doi.​org/​10.​1016/j.​
jmatp​rotec.​2006.​11.​029
43. Kechagias JD, Aslani K-E, Fountas NA et al (2020) A comparative
investigation of Taguchi and full factorial design for machina-
bility prediction in turning of a titanium alloy. Measurement
151:107213. https://d​ oi.​org/1​ 0.1​ 016/j.m
​ easu​rement.​2019.1​ 07213
44. Phadke MS (1989) Quality engineering using robust design. Pren-
tice Hall, Englewood Cliffs, New Jersey, p 07632
45. Gopinath C, Lakshmanan P, Palani S (2021) Fiber laser microcut-
ting on duplex steel: parameter optimization by TOPSIS. Mater
Manuf Process. https://​doi.​org/​10.​1080/​10426​914.​2021.​19819​39
46. Saxena P, Stavropoulos P, Kechagias J, Salonitis K (2020) Sus-
tainability assessment for manufacturing operations. Energies.
https://​doi.​org/​10.​3390/​en131​12730
47. Harris M, Potgieter J, Archer R, Arif KM (2019) In-process thermal
treatment of polylactic acid in fused deposition modelling. Mater
Manuf Process 34:701–713. https://​doi.​org/​10.​1080/​10426​914.​
2019.​15666​11
48. Tamburrino F, Graziosi S, Bordegoni M (2019) The influence of
slicing parameters on the multi-material adhesion mechanisms of
FDM printed parts: an exploratory study. Virtual Phys Prototyp
14:316–332. https://​doi.​org/​10.​1080/​17452​759.​2019.​16077​58
49. Hashemi Sanatgar R, Campagne C, Nierstrasz V (2017) Investiga-
tion of the adhesion properties of direct 3D printing of polymers
and nanocomposites on textiles: Effect of FDM printing process
parameters. Appl Surf Sci 403:551–563. https://​doi.​org/​10.​1016/j.​
apsusc.​2017.​01.​112
50. Deisenroth DC, Moradi R, Shooshtari AH et al (2018) Review
of heat exchangers enabled by polymer and polymer composite
additive manufacturing. Heat Transf Eng 39:1648–1664. https://​
doi.​org/​10.​1080/​01457​632.​2017.​13842​80
51. Melenka GW, Schofield JS, Dawson MR, Carey JP (2015) Evalu-
ation of dimensional accuracy and material properties of the Mak-
erBot 3D desktop printer. Rapid Prototyp J 21:618–627. https://​
doi.​org/​10.​1108/​RPJ-​09-​2013-​0093
13
52. Vyavahare S, Kumar S, Panghal D (2020) Experimental study of
surface roughness, dimensional accuracy and time of fabrication
of parts produced by fused deposition modelling. Rapid Prototyp
J 26:1535–1554. https://​doi.​org/​10.​1108/​RPJ-​12-​2019-​0315
53. Fountas NA, Papantoniou I, Kechagias JD et al (2021) Experimen-
tal investigation on flexural properties of FDM-processed PET-G
specimen using response surface methodology. MATEC Web Conf
349:01008. https://​doi.​org/​10.​1051/​matec​conf/​20213​49010​08
54. Vinoth Babu N, Venkateshwaran N, Rajini N et al (2021) Influ-
ence of slicing parameters on surface quality and mechanical
properties of 3D-printed CF/PLA composites fabricated by FDM
technique. Mater Technol. https://​doi.​org/​10.​1080/​10667​857.​
2021.​19150​56
55. Kechagias JD, Ninikas K, Petousis M et al (2021) An investigation
of surface quality characteristics of 3D printed PLA plates cut
by CO2 laser using experimental design. Mater Manuf Process
36:1544–1553. https://​doi.​org/​10.​1080/​10426​914.​2021.​19068​92
56. Kechagias JD, Fountas NA, Ninikas K et al (2021) Surface char-
acteristics investigation of 3D-printed PET-G plates during CO2
laser cutting. Mater Manuf Process. https://​doi.​org/​10.​1080/​
10426​914.​2021.​19819​33
57. Kechagias J, Ninikas K, Petousis M, Vidakis N (2021) Laser
cutting of 3D printed acrylonitrile butadiene styrene plates for
dimensional and surface roughness optimization. Int J Adv Manuf
Technol
58. Savvakis K, Petousis M, Vairis A et al (2014) Experimental deter-
mination of the tensile strength of fused deposition modeling
parts. In: IMECE2014. ASME, Montreal, Quebec, Canada
59. Christiyan KGJ, Chandrasekhar U, Venkateswarlu K (2016) A study
on the influence of process parameters on the Mechanical Proper-
ties of 3D printed ABS composite. IOP Conf Ser: Mater Sci Eng
114:012109. https://​doi.​org/​10.​1088/​1757-​899X/​114/1/​012109
Publisher's Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.