Pernet Nagel Zhang
Pernet Nagel Zhang
Pernet Nagel Zhang
net/publication/359437775
CITATIONS READS
0 729
3 authors:
Hao Zhang
James Madison University
21 PUBLICATIONS 540 CITATIONS
SEE PROFILE
All content following this page was uploaded by Benoît Pernet on 24 March 2022.
ScienceDirect
Procedia CIRP 00 (2022) 000–000
www.elsevier.com/locate/procedia
Abstract
With the increasing popularity of 3D printed products, material consumption becomes a concern in additive manufacturing in recent years.
Various lightweight structures (e.g., infill patterns) have been designed and used to reduce material use while supporting product functionality.
Product designers and engineers often experience difficulties in choosing the infill structure when solid material is not necessary in an application,
thus losing the potential of material and cost saving opportunities. This is due to a lack of comprehensive study on mechanical properties and
sustainability performances of these infill patterns. The objective of this study is to understand the mechanical behaviors, economic and
environmental benefits through compression test, life cycle cost assessment, and life cycle assessment. Fourteen common infill patterns including
Grid, Lines, Triangle, Cubic, Tetrahedral, Concentric, Concentric 3D, Zigzag, Gyroid, Octet, Cross, Cross 3D, Tri-hexagonal and Quarter Cubic
have been examined. The samples were printed with ASTM D695 standard cylinders made of Polylactic Acid using fused deposition modeling.
Compression tests were conducted on an Instron test platform and the data was collected to compare the load and weight relations of all structures.
The research results will provide a comprehensive mechanical inventory for product design and manufacturing. In addition, this study identifies
opportunities for more robust infill pattern design for additive manufacturing.
what pattern would suit their needs the most, resulting in time, Additive Manufacturing still being in development implies a
material and performance loss. multitude of research has been and is being and are conducted
to improve the technique. Finite Element Analysis (FEA) is a
Most of the time, the principal goal of the infill pattern is to good asset to help the research about infill patterns. Using it to
emulate or even ameliorate the mechanical properties such as derive infill patterns from the predicted stress profiles within
stiffness and rigidity, of a full part, while saving material and parts can increase substantially the stiffness of a part while
manufacturing time, and, thus, money. This is the case in the maintaining a similar loading value [14]. The same way, an
automotive and aeronautical industries for example, where both increase of 3.5 times the loading capacity of a beam filled with
government regulations and consumer demand impact the the honeycomb infill pattern has been observed with an
seeking of lightweight parts [4]. The strength-to-weight ratio is adaptation of the pattern to the predicted stress profiles [15].
the most important factor that affects the choice of a particular Although the Fused Deposition Modeling (FDM) process
infill pattern. Printing lightweight elements implies the use of does not permit it, using Computer Aided Design (CAD) to
less material and the reduction of the print time. Baich et al. design infill patterns might be the future of this field. It can
showed that print cost has more impact on the overall allow multiple different infill patterns in the same part to obtain
production cost than material cost [5]. Manufacturing time better mechanical properties [16]. CAD also helps creating new
being too long is one explanation for the fact that AM still has infill patterns such as bioinspired ones [13]. A new approach
not replaced subtractive manufacturing for many series- combining structural analysis and optimization based on human
produced pieces, although some companies tend to use it bones structures has been presented by Wu et al. to “assist users
increasingly [6]. Some infill patterns generate longer print in designing lightweight and mechanically strong prints” [17].
times because of their design (ex: when the angles are too sharp, Lattice structures are also a solution that is being researched
the print nozzle has to reduce its speed, therefore making the [18]. All the aforementioned aspects are outside of the scope of
print time longer) [7]. The mechanical solicitations that the part this study but are important since they focus on infill patterns
will have to withstand also plays a major role. A component and how to improve them.
that is being loaded in multiple directions should be Some tests have already been conducted to try to understand
manufactured with a different pattern than one that is solely the influence and properties of infill patterns better. Fernandez-
constrained in one direction. A pattern that, for example, is not Vicente et al. proved that while the infill density mainly
aligned with the principal stress field will not be effective determines the tensile strength, the chosen infill pattern also
enough and make the part less strong and stiff. [8] A designer plays a role in this strength [19]. Racz also studied the tensile
should also consider the infill percentage or density, as it is a properties of 3D printed specimens considering different infill
criterion with a high influence on the mechanical abilities of the rates, infill patterns and printing orientation [20]. While testing
part. Most of the time, a completely full print will mechanically the tensile strength of different infill patterns, Yeoh et al. also
perform better than one with a density below 100%, but the tested the hardness of parts filled with different patterns [21].
effective printing time has to be taken into account [9]. Saniman et al. investigated the flexural properties of various
Engineers and designers can choose among an infinity of types of infill patterns and concluded that the introduction of
infill patterns for their part, and when it comes to actual an infill pattern reduced the flexural strength by at least 50%
workpieces, infill patterns play a major role, hence it is and the flexural strain by not more than 40% [22]. Other studies
important to select the one that fits the most. There are two showed that printing parameters, i.e. build orientation, layer
major categories of infill patterns. 2D Infill Patterns are thickness, feed rate and printing path had tangible effects on
identical on every layer and provide the majority of their strength, stiffness, tensile strength and tensile modulus
support in one dimension. 3D infill patterns, on the other hand, [23&24].
form 3D infill structures that add strength to the part in a more
uniform way. Most of the 3D infill patterns are labelled as Despite the already conducted research, there is still a gap
isotropic, meaning that they are equally strong in all directions which this paper’s research is trying to fill in: how do the
[10]. Grid, Lines, Triangle, Concentric, Zigzag, Cross and Tri- mainstream infill patterns compare to each other under
hexagonal are all 2D infill patterns, and they are patterns that compressive tests. The running of such tests should provide
will be investigated in this paper. Their design repeats itself some information about the behaviour of each of the fourteen
layer after layer, although sometimes it rotates on itself from patterns as well as about the importance of infill density.
layer to layer to create some different shapes. Concentric 3D, Consequently, the goal of this study is to propose a
Cubic, Gyroid, Octet, Tetrahedral, Cross 3D and Quarter Cubic classification of fourteen of the most used infill patterns based
are all examples of 3D infill patterns and complete the list of on their strength-to-weight ratio.
infill patterns investigated in this paper. This study will benefit the engineering community by
Different slicer software have different infill patterns providing designers, engineers and even private individuals a
available. This study only tests 14 of them, but other patterns ranking of the infill patterns to choose from when they want to
and designs are provided by different slicer software [12]. design a piece that should be lightweight and will be loaded in
Research always goes forward as new patterns are being compression.
created, mostly inspired from nature [13]. Those bioinspired This paper will first detail the methodology of the testing
designs are however not available in classical slicer software then display and analyse the results, and finally conclude with
infill patterns library and are, like other slicer software’s a comparison of the infill patterns and discuss what could be
patterns, thus outside the scope of this study. done in the future.
Author name / Procedia CIRP 00 (2022) 000–000 3
A 3D model was designed using Solidworks software. To be Before proceeding with the test, all the specimens were
consistent in testing method, all specimens were printed using weighted. A scientific balance with a precision of 0.01 g was
the same parameters with a Dremel Digilab 3D45 machine. The used for the weight measurements.
printer has a build volume of 255x155x170 mm. The 3D model The compression tests were performed on an Instron 5567
was sliced using two different software, Dremel DigiLab 3D Machine. Figure 3 represents the actual setup of the
Slicer and Ultimaker Cura. Table 1 presents the printing compression testing machine used in this study. Test cylinders
parameters used to manufacture the test specimens. were carefully placed in the exact same location for every
iteration of the test. Similarly, sing the machine’s integrated
Table 1. Printing parameters of the testing specimens ruler, the starting height for each test was always the same.
Printing parameters Values R6.35
Printing temperature 230 °C
Build plate temperature 60 °C
Layer height 0.1 mm
25.40
Wall thickness 0.8 mm
Printing speed 50 mm/s
Figure 5. Grid infill pattern with infill density from 20% (left) to 100% (right)
in Dremel DigiLab 3D Slicer
3500 12000.00
3000
2500
2000 10000.00
1500
1000
8000.00
500
Maximum load (N)
0
1 2 3 4 5 6 6000.00
Cylinders
Figure 4. Maximum loads (in N) for each of the 6 cylinders tested 4000.00
For each of the fourteen infill patterns, five different infill 20% infill 40% infill 60% infill 80% infill 100% infill
densities were tested: 20%, 40%, 60%, 80% and 100%. They
were selected as 20% incremental to 100%. Figure 5 shows a
Figure 6. All compression test results
screenshot of five models created with the Dremel DigiLab 3D
Author name / Procedia CIRP 00 (2022) 000–000 5
Max Load/Weight ratio α for all patterns and all infill % 20% 40% 60% 80% 100%
2900.00
2700.00
2500.00
α (N/g)
2300.00
2100.00
1900.00
1700.00
1500.00
Figure 7. Strength-to-weight ratio α value for all 14 infill patterns with density from 20 to 100%
possible to draw any link between the density being higher than
The maximum load weight ratio (α) is calculated as 20 or 40% and the ratio being higher accordingly possible to
maximum load divided by infill pattern weight. In this study draw any link between the density being higher than 20 or 40%
metric units were used. and the ratio being higher accordingly. But this is a tendency
and some values do not follow this rule, as for example the α
𝑚𝑎𝑥𝑖𝑚𝑢𝑚 𝑙𝑜𝑎𝑑 of “Grid 60%” being higher than “Cubic 100%” or the α of
α=
𝑤𝑒𝑖𝑔ℎ𝑡 “Cross 20%” being higher than “Octet 100%”.
Table 4 below shows the average values of the maximum
In Figure 6, the circles represent all the specimens with 20% load-to-weight ratio α for all infill patterns.
infill density, the squares are for 40% infill density, the
triangles for 60%, the lines for 80% and the diamond represent Table 4. Ranking of the tested infill pattern by their average α value.
the specimens with 100% density. Each colour represents an
infill pattern: Grid (red), Lines (dark red), Cubic (orange), Infill Pattern Average α
Triangle (light green), Tetrahedral (green), Concentric (light Cross 2531.92
blue), Concentric 3D (blue), Zigzag (dark blue), Gyroid (gold), Grid 2505.04
Triangles 2392.60
Octet (grey), Cross (purple), Cross 3D (white), Tri-hexagonal Concentric 2335.80
(yellow) and Quarter Cubic (pink). Tri-hexagonal 2310.20
Figure 6 shows that the maximum load of the infill patterns Lines 2303.82
Cross 3D 2286.24
can tolerate increases with the increase of weight. However, Zigzag 2207.22
some patterns outperform lower weight patterns. For example, Octet 2117.22
for the same 6000 N maximum load, three infill patters can be Concentric 3D 2108.83
chosen, grid 40% infill density with 2.42g, zigzag 60% infill Cubic 2022.91
Tetrahedral 2004.01
density with 3g, and quarter cubic 80% with 3.31g. In this case, Gyroid 1986.82
the designer can choose grid 40% infill density to reduce Quarter Cubic 1943.71
material use in the product and still achieve the same
mechanical strength of the product. For designers, the testing Unsurprisingly, the best strength-to-weight ratio α is
results shown in Figure 6 can be used to assist structural design. presented by specimens with 2D patterns (i.e. Grid, Cross,
Figure 7 shows the different values of α for each pattern and Lines, Triangles, Concentric, Tri-hexagonal, and Zigzag). Only
each infill percentage. Each bar represents one pattern. This Cross 3D performs slightly better than Zigzag. This result can
graph illustrates the fact that infill patterns indeed have an be explained by the fact that 2D infill patterns have their
influence on the mechanical properties of a part. Some, like material always deposited in the same direction, thus they can
Grid or Tri-hexagonal, have rather high α (>2100 N/g for all handle stress from that one direction (here compression) better
five infill densities), while others tend to be generally lower. than in other directions. On the other hand, 3D infill patterns
Although an infill density of 100% does not mean a tend to use more material to be isotropic and support stress
completely full part, almost all patterns with 100% infill equally in all three directions. The same conclusion can be
present a higher ratio than with density from 20 to 80%. This drawn from Figure 7 that 2D infill patterns seem to perform
tendency is also true for 80% infill density having a higher ratio better than 3D infill patterns.
than 20 to 60% infill density, although it diminishes. It
decreases even more for the 60% density, as it is no more
6 Author name / Procedia CIRP 00 (2022) 000–000