Progress in Additive Manufacturing (2023) 8:541–559

https://doi.org/10.1007/s40964-022-00347-x

FULL RESEARCH ARTICLE

Stress‑adapted fiber orientation along the principal stress directions

for continuous fiber‑reinforced material extrusion
Tim Heitkamp1 · Sebastian Kuschmitz2 · Simon Girnth1 · Justin‑Dean Marx1 · Günter Klawitter1 · Nils Waldt1 ·
Thomas Vietor2

Received: 17 March 2022 / Accepted: 15 September 2022 / Published online: 25 September 2022
© The Author(s) 2022

Abstract
A proven method to enhance the mechanical properties of additively manufactured plastic parts is the embedding of continu-
ous fibers. Due to its great flexibility, continuous fiber-reinforced material extrusion allows fiber strands to be deposited along
optimized paths. Nevertheless, the fibers have so far been embedded in the parts contour-based or on the basis of regular
patterns. The outstanding strength and stiffness properties of the fibers in the longitudinal direction cannot be optimally
utilized. Therefore, a method is proposed which allows to embed fibers along the principal stresses into the parts in a load-
oriented manner. A G-code is generated from the calculated principal stress trajectories and the part geometry, which also
takes into account the specific restrictions of the manufacturing technology used. A distinction is made between fiber paths
and the matrix so that the average fiber volume content can be set in a defined way. To determine the mechanical properties,
tensile and flexural tests are carried out on specimens consisting of carbon fiber-reinforced polyamide. In order to increase
the influence of the principal stress-based fiber orientation, open-hole plates are used for the tensile tests, as this leads to
variable stresses across the cross section. In addition, a digital image correlation system is used to determine the deformations
during the mechanical tests. It was found that the peak load of the optimized open-hole plates was greater by a factor of 3
and the optimized flexural specimens by a factor of 1.9 than the comparison specimens with unidirectional fiber alignment.

Keywords Fiber reinforced additive manufacturing · Design for additive manufacturing · Material extrusion · Principal
stress · Curvilinear fiber · Continuous fiber

1 Introduction economical production of arbitrarily complex components

directly from 3D CAD data without significant additional
Material extrusion (MEX) is characterized by the production effort. However, this leads to anisotropic material proper-
of parts by continuous layer-by-layer application of a molten ties with poorer mechanical properties in the build-up direc-
thermoplastic strand. The initial material is usually in the tion [2, 3]. Thus, the MEX process is often ruled out when
form of a thermoplastic filament, which can be continuously selecting suitable manufacturing processes for production
unwound from a filament spool [1, 2]. In the past, the MEX in industrial practice, since the mechanical properties of the
process was primarily used for the fast and cost-effective manufactured parts are rarely sufficient for the load case at
production of prototypes. However, further developments hand [3]. In addition to pure thermoplastics, filaments with
of the process and the materials used have led to an increas- added fillers can also be processed to improve the strength
ing number of end products being manufactured using the of parts. In this context, fillers in the form of reinforcing
process [2]. The layer-by-layer process principle enables the fibers are primarily suitable [3–5]. A comparatively simple
approach is the use of short fiber reinforced filaments [6].
Fibers with a length of 0.1–1 mm are added to the thermo-
* Tim Heitkamp plastic filaments. Thus, the filaments can be processed with
tim.heitkamp@hs-hannover.de conventional extruders. In general, short fiber reinforcement
1 improves the tensile and flexural strength and stiffness of the
Hochschule Hannover, 30459 Hannover, Germany
parts, while it reduces the toughness, yield strength and duc-
2
Institute for Engineering Design, Technische Universität tility [7–10]. Due to the significantly improved mechanical
Braunschweig, 38108 Braunschweig, Germany

13
Vol.:(0123456789)
542 Progress in Additive Manufacturing (2023) 8:541–559

properties compared to short fiber reinforcements, continu- stiffness properties in the fiber direction than across the
ous fibers are increasingly used for additive manufacturing fiber direction. Due to this extreme anisotropy in the fiber
of fiber-reinforced thermoplastic parts [11, 12]. Due to their direction, fiber-reinforced plastics offer enormous light-
high strength and stiffness values, carbon, glass or aramid weight design potential when the fibers are aligned in line
fibers are preferably used as reinforcing fibers. Two different with the load.
approaches have been established for embedding continuous For manufacturing using MEX, a G-code (a machine
fibers in parts manufactured by MEX, which differ in the code readable by the 3D printers [29, 30]) is required.
timing of fiber impregnation (see Fig. 1). One approach is to The G-code is generated during the slicing process. In the
impregnate the fibers during the process within the extruder. slicing process, the 3D part geometry is first loaded into
Accordingly, the fibers are initially in the form of rovings. the slicing software and the build orientation is defined.
In the print head, the supplied fiber strand bonds with the Then the part is sliced into layers and divided into outer
molten filament and is deposited together on the part (see and inner contours (perimeters) and an inner regular infill
Fig. 1a). This approach has already been studied and demon- pattern. The perimeters and infill patterns correspond to
strated in various research papers [13–18]. Another concept the tool paths to be traced by the print head [2]. Manufac-
involves fiber impregnation prior to the process. For this turers of fiber printing systems, such as Markforged [31]
purpose, semi-finished products consisting of a thermo- and Anisoprint [32], also offer corresponding slicing soft-
plastic matrix and continuous fibers embedded therein are ware (Markforged–Eiger [33], Anisoprint–Aura [34]) for
used, which are applied to the print bed or the part with an their respective systems, which allows slicing of parts and
extruder (see Fig. 1b) [13, 19]. creation of tool paths for the pure thermoplastic and for the
The use of continuous fiber reinforcements leads to continuous fiber strands. These are user-friendly slicing
a significant increase in the strength of the parts. The programs that specify optimized process parameters that
mechanical properties of continuous fiber reinforced can be varied only to a very limited extent. Various options
parts fabricated using different approaches and materials are available in the slicing software for fiber alignment.
have been described in detail in numerous studies [14–17, These are, for example, contour-based, unidirectional or
20–24]. The mechanical properties obtained depend on a rhombic infill patterns (see Fig. 2), which are compara-
number of factors. In addition to the basic process param- ble to the infill patterns of conventional MEX. However,
eters, the fiber and matrix materials, fiber volume fraction, the embedded fiber strands are of particular importance
interfacial adhesion and the number of defects in the part with regard to the mechanical properties. Due to the ani-
also influence the mechanical properties [25]. Delamina- sotropic properties, fiber orientation adapted to the load
tion and fiber-matrix debonding have been shown to be the is essential. With existing slicing programs, however, the
main failure mechanisms in various studies [16, 26–28]. fibers are placed in the part on the basis of regular pat-
Another decisive factor for the mechanical properties is terns. Load-adapted fiber placement based on numerical
the fiber orientation in the parts. Parts reinforced with simulation methods is not possible with conventional slic-
continuous fibers exhibit many times higher strength and ing programs.

(a) thermoplastic thermoplastic (b) thermoplastic

prepreg filament
filament (matrix) filament filament

fiber

z
x dual extruder
y

previous layer

Fig. 1  Schematic representation of the approaches for the insertion of continuous fibers in the MEX process

13
Progress in Additive Manufacturing (2023) 8:541–559 543

(a) fiber perimeter + undirectional infill (b) fiber perimeter + „isogrid“ infill

fiber tool path thermoplast tool path

(c) fiber perimeter + „rhombic“ infill (d) fiber perimeter

Fig. 2  Exemplary infill patterns with continuous fibers generated with the slicing software Aura (Anisoprint [34]), fiber perimeter with unidirec-
tional infill in (a), with “isogrid” infill in (b), with “rhombic” infill in (c) and without additional fiber infill in (d)

1.1 Stress‑adapted fiber placement Li et al. [39] generated load-path appropriate tool paths
for fiber insertion in printed composite parts. The load paths
In recent years, various approaches have been pursued to are based on the tensile and compressive stresses occurring
introduce fibers into the part in a way that is appropri- in the part. The mechanical tests were carried out on contin-
ate to the load. By means of numerical methods, it could uous fiber-reinforced open-hole plates. The optimized infill
be shown that a load-appropriate alignment of the fibers patterns showed 67.5% and 62.4% higher tensile and flex-
can lead to a significant improvement of the mechani- ural strengths, respectively, than the conventional straight
cal properties. Zhang et al. [35] used FE analyses and infill patterns. The strength-to-weight ratio of the optimized
streamtracer plots to generate fiber orientations based on tensile and flexural specimens exceeded the conventionally
principal stress tensors. In the numerical calculations, the specimens by 55.1% and 35.2%, respectively.
optimized orientation resulted in a significant reduction in The mechanical properties of open-hole tensile specimens
stress peaks compared to 90°/0° and 45°/−45° multilayer made of continuous fiber-reinforced thermoplastic were
composites. investigated by Hou et al. [40]. Optimized fiber orientation
Malakhov and Polilov [36] also used numerical methods was based on stress gradients. Comparisons with unidirec-
to calculate load-oriented fiber orientations in order to be tional fiber orientation increased the ultimate strength by
able to adapt fiber paths to geometric discontinuities, such 42% and reduced the stress concentration by 36%.
as notches or holes, in an optimized manner. A principal
stress-oriented infill pattern resulted in a 3.2-fold reduction 1.2 Aims and scope
in stress peaks compared to the straight fiber orientations in
the numerical model. The aim of this work is to develop a method for the stress-
Malakhov et al. [37] presented an iterative simulation- adapted design and manufacturing of continuous fiber-
based method to calculate the fiber paths in order to manu- reinforced components using MEX in order to improve the
facture parts with variable stiffnesses. In the simulations, the mechanical properties of the components. In contrast to
optimized fiber orientation resulted in a 6 times improve- existing approaches, computer aided internal optimization
ment in stiffness compared to unidirectional fiber orienta- (CAIO) is used for fiber reinforced MEX for the first time
tion. The iterative simulation method for adjusting variable to the best of the authors' knowledge. With the aid of this
stiffnesses was practically investigated by Sugiyama et al. method, it is possible to precisely determine the principal
[38]. The effects of the optimized fiber alignment on the stress trajectories with any number of flexibly selectable
mechanical properties of a fiber paths uniformly surround- starting points. The CAIO method allows effortless adap-
ing the hole on one side, the stiffnesses could be increased tation of any number of trajectories even for complex part
by factors of 4.5 and 9.4, respectively. The maximum load geometries. In addition, the method allows the generation
capacity of the open-hole plates could be increased by a fac- and selection of defined fiber trajectories without recalcu-
tor of 1.2 and 1.6, respectively. lations, which is a disadvantage in other simulation-based

13
544 Progress in Additive Manufacturing (2023) 8:541–559

methods [41]. The plotted principal stress trajectories can The directions of the principal stresses can be determined
be used directly as fiber toolpaths. Accordingly, the fibers according to the following equation [44].
strictly follow the principal stress directions.
2𝜏xy
To validate the method, solid specimens with embedded tan(2𝜑∗ ) = (4)
continuous fiber strands are fabricated. On the one hand, 𝜎xx − 𝜎yy
open-hole plates are used as specimens, for which only the
first principal stress direction is taken into account. In addi- The tangent function is periodic with π. Therefore, Eq. (4)
tion, flexural specimens with layerwise alternating fiber with φ* and 𝜑* = φ* + π/2 provides the two mutually per-
paths along the first and second principal stress directions pendicular principal directions [44]. According to Neuber
were printed and tested. Mechanical tests on optimized and [45], the directions of the first principal stress, described by
unidirectional reinforced specimens will show the impact on the angle φ1, can be determined directly with respect to the
the mechanical properties. In addition, it will be investigated underlying coordinate system using Eq. (5) for plane stress
whether the principal stress-based fiber orientations differ problems.
from unidirectional fiber orientations in terms of failure 𝜏xy
mechanisms. tan 𝜑1 =
𝜎xx − 𝜎2 (5)

The use of Eq. (5) has the advantage that the angle φ1,
2 Principal stress‑based fiber alignment which is to be assigned to the first principal stress direction,
can be calculated directly. Thus, the ambiguity is avoided
Compared to conventional isotropic materials, the aniso- when calculating tan(2φ) from Eq. (4).
tropic properties and high specific strengths of the reinforc- The principal stresses can be represented in the form of
ing fibers offer considerable potential for lightweight con- so-called trajectories, which are tangential to the principal
struction applications, especially when aligned to withstand stress directions at each point. In composite parts, these tra-
loads. Fibers exhibit outstanding strength and stiffness prop- jectories can be considered as load-adapted fiber paths that
erties in the longitudinal direction. Transverse to the fiber, take into account the anisotropic fiber properties (see Fig. 3).
the strengths and stiffnesses are many times lower. Accord-
ingly, the principal stress directions can be used for optimal
fiber alignment. The basic idea of this approach is that the 3 Materials and methods
shear stresses τ become zero, while the normal stresses σ
take extreme values, so that the longitudinal strength prop- In this section, the study design and methods used are pre-
erties of the fibers can be used optimally [42]. In an infini- sented. The approach used to generate the G-code for the
tesimal material section, the shear stresses are tangential to principal stress-based fiber alignment is discussed, as well
the sides of the material section, while the normal stresses as the fabrication setup. In addition, the specimen design and
are normal to them. When the material section is rotated by the mechanical testing procedure are described.
angle φ, the shear stresses can be completely eliminated,
while the normal stresses become maximum. In this case, 3.1 Principal stress‑based toolpath planning
σ1 ≥ σ2 as well as σ1 ⊥ σ2. The calculated stress states can be
summarized in a matrix S [43]: A suitable method, which is also used in this work for the
[ ] visualization and generation of the optimized fiber align-
S=
𝜎xx 𝜏xy
(1) ments, is the CAIO method. Moldenhauer presents in [41] a
𝜏xy 𝜎yy CAIO approach to visualize the principal stress trajectories
in conventional finite element analysis (FEA) software. His
According to [43] solving the eigenvalue problem with
mathematically proofed method is based on the following
unit matrix I in Eq. (2) yields eigenvalues σ1,2 corresponding
hypothesis [41]:
to Eq. (3), which correspond to the first and second principal
stresses. The isotherms of an orthotropic steady-state 2D ther-
mal conduction problem with the thermal conductivi-
det (S − 𝜎I) = 0 (2) ties ­k1 and k­ 2 are tangential to an arbitrarily prescribed
√ direction field y′ = f(x, y) provided that the local ori-
𝜎xx + 𝜎yy
(
𝜎xx − 𝜎yy
)2 entation of k­ 1 follows the direction field y′, and perfect
𝜎1,2 = ± 2
+ 𝜏xy (3) insulation exists perpendicular to this ­(k2 = 0).
2 2
The principal stress directions are direction fields which
can be computed statically. The orthotropic thermal analysis

13
Progress in Additive Manufacturing (2023) 8:541–559 545

shear stress normal stress principal stresses

2
1

2 2
1
1
tensile pressure

Fig. 3  Principal stress trajectories of a beam subjected to bending with an infinitesimal material section

is independent from the static analysis and uses the principal were first read in as coordinates and connected in the form
stress directions as local systems for the conductivities ­k1 of a curve. A specific path width was assigned to the fiber
and ­k2. Extreme orthotropic ratios (k1/k2 for the first, k2/k1 paths. The layer was then subdivided into the fiber paths
for the second principal stress direction, at least greater than with specific width and the intermediate spaces. Separate
1/1e−4) force the isotherms to follow the direction fields. process parameters, such as extruder temperature, printing
According to the described procedure, a FE stress analysis speed and print bed temperature, were defined for fiber paths
is first performed in ABAQUS. The principal stress direc- and the intermediate infill. Furthermore, cutting commands
tions in the integration points of the finite elements of a 2D were implemented for the fiber paths, which are located
or 3D shell part are determined. The mutually perpendicular about 40 mm before the end points of the fiber path due to
principal stress directions in the elements are transferred as the distance between extruder and cutting unit. In addition,
local systems to a subsequent thermal FEA. The structural the G-code for unreinforced layers was generated, which was
elements of the stress analysis are replaced by thermal ele- used for the first and last layers. The G-codes were then
ments without changing the node and element numbering. merged and sent to the printer.
For common problems, two thermal boundary conditions in
the form of heat sources with different temperatures in two
nodes of the mesh are sufficient to visualize the isotherms 3.2 Setup and materials
over the entire mesh. They represent the principal stress
trajectories. The placement of the two nodes with thermal For the printing of the specimens, in this work a modified
boundary conditions is discussed in [41]. Renkforce RF2000 (Conrad Electronic SE, Hirschau, Ger-
The result of this method is therefore a continuous tem- many) with a Duet3D control board with RepRap firmware
perature field in which the isotherms follow the principal (Duet3D Ltd., Peterborough, Great Britain) was used, which
stress directions. No iterations are necessary for the deter- can be accessed by a browser interface (see Fig. 6). The
mination of the principal stress trajectories. With the help printer's conventional printhead was replaced with a dual-
of post-processing software, any number of fiber trajectories extruder system from Markforged (Markforged Inc.), which
can be generated from the continuous field at regular inter- is also used in its MarkTwo printer. Polyamide 6 (Nylon
vals or at defined points without any further calculation steps White) was used as the matrix filament and carbon fiber fila-
and output in the form of coordinates. The process flow is ment CFF from Markforged (Markforged, Inc., Watertown,
shown in Fig. 4. MA, USA) was used as the reinforcing fiber. To prevent
Conventional slicing programs are not able to generate moisture absorption of the polyamide, the filament spools
toolpaths for parts with complex fiber paths. Therefore, in were stored in the airtight drying box PolyBox™ (Polymaker
this work, the 3D modeling software Rhinoceros3D version B.V., Shanghai, China) with a controlled relative humid-
7.0 with the visual programming environment Grasshopper ity content of < 25%. The diameter of the nylon filament is
was used to generate the toolpath. Since solid specimens 1.75 mm and the diameter of the continuous fiber filament
were to be used for the investigations, infill patterns were is 0.4 mm. Approximately 1000 fibers are embedded in the
generated for the inter-fiber spaces (see Fig. 5). A written continuous fiber filament. The fiber volume fraction has been
program distinguished between fiber paths and intermediate measured in various investigations and varies between 32.8
areas within the layers. In the program used, the fiber paths and 34.5% [46, 47].

13
546 Progress in Additive Manufacturing (2023) 8:541–559

CAD model 1.

1.
FEA with defined load
Additional FEA
for other load
cases
One load
No case
Yes
2. Calculation of the principal
2.
stress directions in the
elements

Replacing the structural

elements with thermal
elements

Assignment of orthotropic
thermal conductivities in
local systems according to
the principal stress directions 3.

Repeat FEA with thermal

elements

3.
Principal stress directions as
contour lines

Fig. 4  Flowchart of the CAIO method using orthotropic heat conduction as a numerical integrator with curved open-hole plate under tensile
loading, according to Moldenhauer [41]

fiber tool paths intermediate infills

bottom and top layer fiber reinforced layer

20 mm

Fig. 5  Exemplary infill pattern of the bottom and top layers and the fiber-reinforced layers

13
Progress in Additive Manufacturing (2023) 8:541–559 547

Fig. 6  Renkforce RF2000 with

Markforged Dual Extruder and
Polymaker Polybox

filament feeder
fiber feeder
dual extruder

printer bed

fiber filament

nylon filament

dry box

Table 1  Process parameters for the fabrication of the specimens from The simulation and calculation of the principal stress
nylon and carbon fiber trajectories was performed according to the procedure
Parameter Value Unit described in Sect. 3.1. An element number of 38,400 was
used for the flexural specimens and an element number
Extruder temperature 270 °C of 47,460 was used for the open-hole plates. CPS4R ele-
Print bed temperature 90 °C ments are used as element type for the structural analysis,
Nylon filament diameter 1.75 Mm
and DC2D4 elements are used for the thermal analysis. The
CF filament diameter 0.4 mm
G-code was generated using the developed program in Rhi-
Layer height fiber/nylon 0.125 mm
noceros3D/Grasshopper. An exemplary process flow for a
Bottom and top layers (nylon) 0.2 mm
flexural specimen is shown in Fig. 8.
Strand width (nylon) 0.4 mm
The generation of fiber paths for the open-hole plates was
Fiber width 0.9 mm
performed at constant distances in the direction of the first
Fill density 100 %
principal stress for both specimen types. In order to achieve
Nylon printing speed 40 mm/s
a comparable fiber volume fraction Vf, this resulted in 12
Fiber printing speed 8 mm/s
fiber tracks per layer with a spacing of approx. 1.9 mm in the
Infill angle +45/−45 °
straight-line direction for the principal stress-based orienta-
tion and in 14 fiber tracks with a spacing of approx. 1.7 mm
in the straight-line direction for the unidirectional fiber ori-
The specimens were fabricated from both fiber-reinforced entation (see Figs. 9, 10). The second principal stress direc-
and pure nylon layers. An overview of the process param- tion was not considered for the open-hole plates.
eters used is shown in Table 1. The fiber orientation for the flexural specimens, on the
other hand, is more complex due to the high relevance of
both principal stress directions. Therefore, a distinction was
3.3 Specimen design and fabrication made in the layers between the first and second principal
stress directions. The complete filling of each layer allowed
Open-hole plates and flexural specimens were used as speci- an alternating layer structure with alternating fiber orien-
mens to investigate the influence of optimized fiber orienta- tations, each following the first or second principal stress
tions. The dimensions of the specimens are shown in Fig. 7. direction (see Fig. 11a). The spacing between the fiber paths
As a reference to the optimized specimens, fiber-reinforced is subject to certain restrictions. On the one hand, the width
specimens with unidirectional fiber orientation and compa- of the deposited fiber strand has to be taken into account for
rable fiber volume fraction were additionally used. With the the fiber orientation. Thus, the generated paths must have a
exception of the fiber orientation, the same process param- distance of at least 0.9 mm to each other. Secondly, a mini-
eters as listed in Table 1 were used. mum radius of curvature must not be undershot, otherwise

13
548 Progress in Additive Manufacturing (2023) 8:541–559

(a) (b)

z
x

8 mm
Fig. 7  Dimensions of the open-hole plate in (a) and the flexural specimen in (b)

Setup of the FE model

Definition of
• geometry,
(a) • loads,
• boundary,
• element type & size

Calculation of the principal

stresses
• readout and calculation of the
(b) principal stress directions in
the nodes
• red – first principal stress,
blue – second principal stress

Visualization & selection of

(c) principal stress trajectories
• reading of principal stress
trajectories in the post-
processing software
• generation of trajectories in
defined adjustable intervals
and any number
(d) • export of trajectories as
coordinates

G-code generation
• specifying the process
(e) parameters
• generation of G-code for
fibers and intermediate fills
y • output as .gcode file

Fig. 8  Exemplary process flow for the generation of a stress-adapted in (b), visualized principal stress trajectories in the post-processing
infill pattern using the example of a flexural specimen; FE model in software in (c, d), visualization of the G-code of the specimen in (e)
(a), principal stresses at right angles to each other in the nodal points

13
Progress in Additive Manufacturing (2023) 8:541–559 549

(a)

(b)

y
15 mm Starting point Ending point Cutting point

Fig. 9  Comparison of fiber orientations for the open-hole plate along the principal stress trajectories in (a) and with unidirectional orientation in
(b) with the starting, ending, and cutting positions

Fig. 10  Optimized open-hole

plate in (a) and open-hole plate
(a)
with unidirectional fiber rein-
forcement in (b)

(b)

20 mm

fiber breakage may occur. The exact minimum radius of of an open-hole plate is about 30 min and the fabrication
curvature for the carbon fiber filament is not known. As an time of a flexural specimen is about 160 min.
approximation, a radius of curvature corresponding to the For better comparability, the specimens were prepared
width of the deposited fiber strand (0.9 mm) was used in this with a fiber volume fraction as identical as possible. The
work. Taking these restrictions into account, the fiber curves fiber volume fraction describes the proportion of fibers in
were generated, which are shown in Figs. 11 and 12. the total volume of the part and can be determined with
The underlying standard prescribes a minimum number the following formula [48]:
of five specimens. This minimum number was taken into vf
account for the present work. In order to reduce the pro- Vf = × 100 % (6)
vc
duction time of the flexural specimens, the middle part of
the specimen was made of pure unreinforced thermoplastic where Vf is the fiber volume fraction, vf is the total volume
with a height of 3.2 mm and layer thicknesses of 0.2 mm. of the fibers and vc is the total volume of the composite
The height of the fiber-reinforced layers was thus 4 mm in part. Due to the complex fiber characteristics in principal
total with a layer height of 0.125 mm. The fabrication time

13
550 Progress in Additive Manufacturing (2023) 8:541–559

(a) (b) 2 1

z
y
(c)
x

y
15 mm
x

Fig. 11  Schematic layered structure in (a) and comparison of fiber orientations for the flexural specimen along the principal stress direction in
(b) and with unidirectional orientation in (c)

(a)

(b)

20 mm

Fig. 12  Fabricated optimized flexural specimen in (a) and unidirectional reinforced flexural specimen in (b)

Table 2  Theoretically determined fiber volume fractions, subdivided determined fiber volume fractions of the specimens with
by specimens and layers optimized and unidirectional fiber orientation. Furthermore,
Specimen/layer Fiber vol- discrepancies between the different layer layouts of the
ume fraction flexural specimen were found. The fiber orientations corre-
(%) sponding to the first and second principal stresses or to a 0°
Open-hole plate (optimized) 9.5 and 90° infill pattern resulted in fiber volume fractions in the
Open-hole plate (unidirectional) 9.8 respective layers that could not be precisely adjusted. There-
Flexural specimen (optimized) 5.4 fore, Table 2 also lists the different fiber volume fractions,
Flexural specimen (optimized—σ1) 2.6 subdivided by layers. For the calculation, a fiber volume
Flexural specimen (optimized—σ2) 2.8 fractions of the fiber filament of 34% was assumed.
Flexural specimen (unidirectional) 5.8
Flexural specimen (unidirectional—0°) 2.8 3.4 Tensile and flexural tests
Flexural specimen (unidirectional—90°) 3.0
The tensile tests were performed on the Zwick/Roell Zmart.
Pro BZ1 universal testing machine (Zwick Roell, Ulm, Ger-
stress-based alignment compared to unidirectional fiber many). The mean values of the measured values are given for
alignment, it was not possible to set an identical fiber vol- the results. The specimen deformation was measured over
ume fraction. Table 2 shows the respective theoretically the recorded crosshead travel. The test speed was 10 mm/

13
Progress in Additive Manufacturing (2023) 8:541–559 551

min. The 3-point bending test was performed to determine record the displacements. The images were taken at intervals
the characteristic values under bending load. The tests were of one second. The experimental setup is shown in Fig. 13.
performed on the same testing machine with identical testing The microstructure of the fracture surfaces of the flex-
speed. The machine force was recorded by load cell and the ural and tensile specimens were examined using a Zeiss
deformation by the crosshead travel of the testing machine. Leo 1455VP scanning electron microscope (Carl Zeiss AG,
All mechanical tests were performed under the same envi- Oberkochen, Germany).
ronmental conditions.
A digital image correlation (DIC) system was used, which
allows non-contact detection of deformations. The Aramis 4 Results
5 M measuring system from GOM (Carl Zeiss GOM Metrol-
ogy GmbH, Braunschweig, Germany) was used for this The results of the mechanical tests are described below.
purpose. The system has two cameras and thus enables the Figure 14 shows the force–displacement diagrams for the
measurement of two- and three-dimensional deformations. tensile and flexural tests for the optimized and unidirectional
In order to measure the deformations of the specimens, sto- reinforced specimens. The average maximum forces with
chastic patterns were applied to the surfaces of the speci- the corresponding standard deviations for the tensile and
mens, which allowed the measurement system to optically flexural tests are shown in Fig. 15.

Fig. 13  Test setup of the tensile

test with clamped tensile speci-
men with stochastic pattern and
the digital image correlation
system (DIC)

(a) (b)
12 5
optimized optimized
10 unidirectional 4 unidirectional
8
3
Load [kN]
Load [kN]

6
2
4
1
2

0 0
0 1 2 3 4 5 0 1 2 3 4
Displacement [mm] Displacement [mm]

Fig. 14  Force–displacement diagrams of the tensile tests in (a) and the flexural tests in (b)

13
552 Progress in Additive Manufacturing (2023) 8:541–559

Fig. 15  Overview of the average 12

maximum force of the tensile
and flexural tests, subdivided
10
into optimized and unidirec-
tional reinforced specimens

Maximum load [kN]

with the corresponding standard 8
deviations
6

0
optimized unidirectional optimized unidirectional
Flexural tests Tensile tests

Considering the average values obtained (see Fig. 15), half, as well as several smaller broken pieces. The clearly
it becomes apparent that the optimized specimens absorb visible damage around the hole is conspicuous, indicating
significantly higher peak loads in both test series. Compared that areas of high stress occur not only on the sides of the
with the unidirectional reinforced open-hole plates, the opti- hole transverse to the load. This is also supported by the
mized open-hole plates with an average measured maximum measurements using the digital image correlation system
force of approx. 10 kN have a peak load that is greater by a (see Fig. 18).
factor of 3. The optimized flexural specimens, on the other All fractured flexural specimens with unidirectional fiber
hand, withstand a maximum load of approx. 3.7 kN, a factor orientation show an expected central fracture. In contrast,
of 1.9 greater than the unidirectional reinforced specimens. the optimized specimens show a completely different frac-
As expected, all specimens exhibit brittle fracture. Three ture pattern. All specimens show a fracture right of center
of the six unidirectional reinforced open-hole plates exhibit with an approximately diagonal course through the speci-
failure with multiple fractures (see Fig. 14a). This is due to men. The fracture occurs on the lower half along the second
the failure and fracture of individual fiber strands at short principal stress direction σ2 (compression) and on the upper
time intervals. In all three specimens, the first break is fol- half along the first principal stress direction σ1 (tension) (see
lowed by a clearly visible drop in force with subsequent also Fig. 20).
complete failure of the specimen. The results of the measurement using the digital image
The greater effects of principal stress-based fiber orienta- correlation system are shown in Fig. 18. The measurements
tion on the mechanical properties of the tensile specimens were performed on each specimen. However, very similar
compared with the flexural specimens are attributed to part deformations are observed in all measurements. Therefore,
weakening due to the hole. A significant amount of unidi- the measurements of the first specimen, respectively, at 50%
rectional fiber strands before and after the hole does not and at 90% of the applied maximum load are presented as
result in any increase in strength, which is also a primary representative. Based on the measured deformations of the
criticism of conventional fiber orientation. In contrast, a open-hole plates, the optimized specimens (Fig. 18a) show
fiber alignment based on the principal stress leads to a more a significantly more homogeneous stress distribution at 50%
uniform force distribution. It also results in simultaneous of the maximum load compared with the unidirectional rein-
loading of all fiber strands and also reduces stress peaks at forced specimens (Fig. 18b). At 50% peak load, significant
the weakened cross section. This is also evident from the stress peaks transverse to the applied load can already be
measurements made using the digital image correlation sys- observed at the hole edges with unidirectional fiber ori-
tem (Fig. 18). entation. These deformations increase progressively with
Figures 16 and 17 show all broken tensile and flexural increasing load. Above and below the hole, only very slight
specimens. With the exception of one optimized specimen, deformations are observed. At 90% load, the optimized
which shows a deviating fracture pattern (see Fig. 16b, specimens show a more homogeneous deformation over the
Specimen 1), all open-hole plates are fractured at the surface, but exhibit increased deformations at the upper and
point with the smallest cross section. All unidirectional lower sides of the hole. No compression is observed on the
reinforced specimens have two broken pieces, while the sides of the hole. The comparatively low strains at the weak-
optimized specimens are broken into an upper and lower est part cross section are remarkable. Nevertheless, even the

13
Progress in Additive Manufacturing (2023) 8:541–559 553

(a) 1 (b) 1

2 2

3 3

4 4

5 5

6 6

25 mm

Fig. 16  Fractured tensile specimens with unidirectional fiber orientation in (a) and with optimized fiber orientation in (b)

(a) (b)
1 1

2 2

3 3

4 4

5
5

25 mm

Fig. 17  Fractured flexural specimens with unidirectional fiber orientation in (a) and with optimized fiber orientation in (b)

13
554 Progress in Additive Manufacturing (2023) 8:541–559

optimized unidirectional
(a) 50 % FMax 90 % FMax (b) 50 % FMax 90 % FMax

(c) 50 % FMax (d) 50 % FMax

90 % FMax 90 % FMax

Fig. 18  Measured deformations using DIC for optimized tensile specimens in (a), unidirectional reinforced tensile specimens in (b), optimized
flexural specimens in (c), and unidirectional reinforced flexural specimens in (d)

optimized open-hole plates show a dominant fracture pattern

◂
Fig. 19  Microstructure of fractured cross sections of optimized ten-
at the hole sides transverse to the load. sile specimens in (a) and (b), unidirectional reinforced tensile speci-
mens in (c, d), optimized flexural specimen in (e, f), and unidirec-
When comparing the measured deformations of the flex-
tional reinforced flexural specimen in (g, h)
ural specimens, very different deformations can be observed.
While the flexural specimens with unidirectional fiber ori-
entation (Fig. 18d) show deformations corresponding to the shift to the edges of the specimen with a lower relative fiber
maximum bending moment, the optimized specimens show volume fraction.
hardly any deformations in the middle area of the speci- The microstructure of the fracture surfaces was inves-
men (Fig. 18c). In the optimized specimens, the deformation tigated using SEM images to determine the main failure
shifts to the sides of the specimen, which is also expected mechanisms and to investigate possible differences between
from the off-center fractures of the specimens. The high the fiber orientations. For this purpose, images were taken of
fiber content in the most stressed areas in the center of the both specimen types with optimized and unidirectional fiber
specimen prevents large deformation changes and leads to a orientations. The images are shown in Fig. 19.

13
Progress in Additive Manufacturing (2023) 8:541–559 555

(a) (b)

Top layer breakage

Debonding

Fiber filament
matrix
Delamination

500 µm 50 µm

(c) (d)
Debonding
Fiber breakage

Delamination

Fiber filament
matrix Fiber pullout
Large fiber
Fiber filament
impregnation
matrix

250 µm 250 µm

(e) (f)
Matrix fracture

Delamination Fiber pullout

Fiber breakage
First principal stress
(tensile) fibers

Fiber breakage under

transverse tension
Second principal Bottom layer
stress (pressure) fibers breakage

25 µm 25 µm

(g) (h)

Fiber pullout
Delamination
Fiber pullout

Intermediate infill Fiber debonding

Fiber breakage

25 µm 50 µm

13
556 Progress in Additive Manufacturing (2023) 8:541–559

All fracture surfaces show comparable failure mecha- occurs in the direction of the second principal stress direc-
nisms. The optimized tensile specimens are characterized by tion (compression). On the compressive side, however, the
the significantly higher fiber content in the fracture surfaces fracture occurs along the first principal stress direction (ten-
of broken fiber strands. The unidirectionally reinforced ten- sion). The curved fiber strands therefore lead to a slanted
sile specimens show predominantly protruding or pulled-out fracture progression. The fracture path is illustrated in
fiber strands in the surrounding matrix. Based on the frac- Fig. 20. The debonding, which can be observed in the tensile
ture surfaces of the optimized and unidirectional reinforced specimens even under tensile loading, is intensified by the
tensile specimens (Fig. 19a, c), delamination between the transverse loading of the fiber strands. The inhomogeneity
fiber layers is clearly visible. This indicates a low level of of the fiber strands promotes cracking and fracture of the
layer adhesion between the fiber strands. Within the lay- specimen under transverse tension. Therefore, hardly any
ers, mainly exposed or vacancies caused by pulled out fibers matrix around the adjacent transversely embedded fibers is
can be seen, indicating a low interfacial adhesion between visible at the fracture surface of the flexural specimens (see
the matrix and the fibers embedded therein. In some cases, Fig. 19f, h).
entire fiber bundles are present without surrounding matrix
(see Fig. 19b). Debonding and delamination have also been
identified by, e.g., in [16, 26–28] as major failure mecha- 5 Conclusion
nisms in printed continuous fiber reinforced components.
In some areas (e.g., Fig. 19d, f), non-uniform fiber embed- In this work, a method was presented that allows the con-
ding is also observed, with matrix-dominated areas and areas tinuous fiber-reinforced additive manufacturing of parts
of high fiber density. This was also described in [49]. The with load-oriented fiber alignment. The fibers are aligned
layered structure is clearly observed in the flexural speci- along the first and, if necessary, the second principal stress
mens (see Fig. 19e, g). The fracture surfaces of the flexural direction. A CAIO approach was chosen to calculate the
specimens also show delamination between the layers. The principal stress trajectories. A developed program generates
fracture surfaces of the unidirectionally reinforced flexural the G-code required for manufacturing based on the part
specimens (Fig. 19g, h) show that the fracture is at the level geometry and the calculated trajectories. The use of a dual
of the transversely embedded fiber strands. In this context, extruder allows the production of parts made of unreinforced
the transversely embedded fibers act as a weak point. The thermoplastic with the possibility of local integration of con-
same behavior is also seen in the optimized bending speci- tinuous fiber strands.
mens, where the fibers are aligned along the principal stress Based on the presented method, tensile tests on carbon
directions. On the tensile side of the specimens, the fracture fiber reinforced open-hole plates and flexural tests were

(tensile) (pressure)
F 1 2

Specimen fracture

Fig. 20  Fracture progression along the first principal stress direction for the lower half and along the second principal stress direction for the
upper half of the flexural specimen

13
Progress in Additive Manufacturing (2023) 8:541–559 557

performed. The specimens were fabricated and tested with Author contributions Conceptualization was contributed by T.H.;
both principal stress-based and unidirectional fiber align- methodology was contributed by T.H.; formal analysis was contrib-
uted by T.H., S.K., and S.G.; T.H. and J.D.M contributed to design and
ment. For the tensile specimens, only the first principal manufacturing of the test specimens; T.H. and J.-D. M. contributed to
stress direction was considered, while for the flexural speci- investigation; T.H. contributed to writing—original draft preparation;
mens both the first and second principal stress directions S.K., S.G. and N.W. contributed to writing—review and editing; T.H.
were considered. In this case, the slicing program allows an contributed to visualization; G.K., N.W. and T.V. contributed to super-
vision; G.K., N.W. and T.V. contributed to project administration; G.K.,
alternating layer structure with alternating fiber orientations. N.W. and T.V. contributed to funding acquisition. All authors have read
In order to better compare the tensile and flexural specimens and agreed to the published version of the manuscript.
with different fiber orientations, similar fiber volume frac-
tion was used. During the mechanical tests, it was found that Funding Open Access funding enabled and organized by Projekt
the principal stress-based orientation of the fibers led to a DEAL. Funded by the Ministry for Science and Culture of Lower
Saxony (MWK), School for Additive Manufacturing SAM.
significant increase in the strength of the parts. The peak
load of the optimized tensile specimens exceeded that of the
Declarations
tensile specimens with unidirectional fiber orientation by a
factor of about 3. The optimized flexural specimens showed Conflict of interest The authors declare that they have no conflict of
a peak load that was higher by a factor of 1.9 compared interest.
with flexural specimens and unidirectional fiber orientation.
Based on measurements using digital image correlation, it Open Access This article is licensed under a Creative Commons Attri-
was shown that a principal stress-based fiber orientation bution 4.0 International License, which permits use, sharing, adapta-
tion, distribution and reproduction in any medium or format, as long
leads to a more homogeneous stress distribution. This results as you give appropriate credit to the original author(s) and the source,
in a reduction in stress peaks, especially in weakened part provide a link to the Creative Commons licence, and indicate if changes
cross sections. were made. The images or other third party material in this article are
The proposed process therefore makes it possible to included in the article's Creative Commons licence, unless indicated
otherwise in a credit line to the material. If material is not included in
significantly increase the strength of continuous fiber-rein- the article's Creative Commons licence and your intended use is not
forced and additively manufactured parts or to reduce the permitted by statutory regulation or exceeds the permitted use, you will
weight or fiber content while maintaining the same mechani- need to obtain permission directly from the copyright holder. To view a
cal properties. In principle, the method is also useful for copy of this licence, visit http://​creat​iveco​mmons.​org/​licen​ses/​by/4.​0/.
additively manufactured parts without fiber reinforcement,
for example for additively manufactured lattice structures.
In the context of fiber-reinforced parts, however, the stress-
based fiber alignment is of particular importance. Since the References
fibers exhibit outstanding strength and stiffness properties in
1. Gebhardt A (2016) Additive Fertigungsverfahren: Additive Manu-
the longitudinal direction, but only very low strengths and facturing und 3D-Drucken für Prototyping—Tooling—Produk-
stiffnesses transverse to the fiber orientation, minimization tion, 5th edn. Hanser, München
of the shear stresses is a target for improving the mechani- 2. Gibson I, Rosen D, Stucker B (2015) Additive manufacturing
cal properties. This is made possible with principal stress- technologies: 3D printing, rapid prototyping and direct digital
manufacturing, 2nd edn. Springer, New York
based fiber orientation. In future, principal stress-based 3. Mazurchevici A., Nedelcu D, Popa R (eds) (2020) Additive manu-
fiber alignment using simulation-based calculations in slic- facturing of composite materials by FDM technology: a review.
ing programs could lead to an increasing use of continuous Indian J Eng Mater Sci 27:179–192
fiber-reinforced material in the industry. It should be noted, 4. Ngo TD, Kashani A, Imbalzano G et al (2018) Additive manufac-
turing (3D printing): a review of materials, methods, applications
however, that larger loads can only occur in one direction and challenges. Compos B Eng 143:172–196. https://​doi.​org/​10.​
parallel to the build plane. In this context, processes must 1016/j.​compo​sitesb.​2018.​02.​012
be developed in future that enable additive manufacturing 5. Mohan N, Senthil P, Vinodh S et al (2017) A review on compos-
of continuous fiber-reinforced parts with a fiber orientation ite materials and process parameters optimisation for the fused
deposition modelling process. Virtual Phys Prototyping 12:47–59.
in multiple spatial directions. Also, the improvement of https://​doi.​org/​10.​1080/​17452​759.​2016.​12744​90
the interfacial adhesion between fiber and matrix and the 6. Bryll K, Piesowicz E, Szymański P et al (2018) Polymer compos-
adhesion between the layers should be given special atten- ite manufacturing by FDM 3D printing technology. MATEC Web
tion, since in this work it was shown that delamination and Conf 237:2006. https://d​ oi.o​ rg/1​ 0.1​ 051/m
​ atecc​ onf/2​ 01823​ 70200​ 6
7. Ning F, Cong W, Qiu J et al (2015) Additive manufacturing of
debonding are the main failure mechanisms. carbon fiber reinforced thermoplastic composites using fused
deposition modeling. Compos B Eng 80:369–378. https://​doi.​
Supplementary Information The online version contains supplemen- org/​10.​1016/j.​compo​sitesb.​2015.​06.​013
tary material available at https://d​ oi.o​ rg/1​ 0.1​ 007/s​ 40964-0​ 22-0​ 0347-x.

13
558 Progress in Additive Manufacturing (2023) 8:541–559

8. Ferreira RTL, Amatte IC, Dutra TA et al (2017) Experimental mechanism, materials and properties. Compos Struct 232:111476.
characterization and micrography of 3D printed PLA and PLA https://​doi.​org/​10.​1016/j.​comps​truct.​2019.​111476
reinforced with short carbon fibers. Compos B Eng 124:88–100. 26. Luo M, Tian X, Shang J et al (2019) Impregnation and interlayer
https://​doi.​org/​10.​1016/j.​compo​sitesb.​2017.​05.​013 bonding behaviours of 3D-printed continuous carbon-fiber-rein-
9. Torrado Perez AR, Roberson DA, Wicker RB (2014) Fracture forced poly-ether-ether-ketone composites. Compos A Appl Sci
surface analysis of 3D-printed tensile specimens of novel ABS- Manuf 121:130–138. https://​doi.​org/​10.​1016/j.​compo​sitesa.​2019.​
based materials. J Fail Anal Prev 14:343–353. https://​doi.​org/​10.​ 03.​020
1007/​s11668-​014-​9803-9 27. Liu T, Tian X, Zhang M et al (2018) Interfacial performance
10. Ning F, Cong W, Hu Z et al (2017) Additive manufacturing of and fracture patterns of 3D printed continuous carbon fiber with
thermoplastic matrix composites using fused deposition mod- sizing reinforced PA6 composites. Compos A Appl Sci Manuf
eling: a comparison of two reinforcements. J Compos Mater 114:368–376. https://​doi.​org/​10.​1016/j.​compo​sitesa.​2018.​09.​001
51:3733–3742. https://​doi.​org/​10.​1177/​00219​98317​692659 28. Caminero MA, Chacón JM, García-Moreno I et al (2018) Inter-
11. Naranjo-Lozada J, Ahuett-Garza H, Orta-Castañón P et al (2019) laminar bonding performance of 3D printed continuous fibre rein-
Tensile properties and failure behavior of chopped and continu- forced thermoplastic composites using fused deposition model-
ous carbon fiber composites produced by additive manufacturing. ling. Polym Test 68:415–423. https://d​ oi.o​ rg/1​ 0.1​ 016/j.p​ olyme​ rtes​
Addit Manuf 26:227–241. https://​doi.​org/​10.​1016/j.​addma.​2018.​ ting.​2018.​04.​038
12.​020 29. Deutsches Institut für Normung e. V (1983) DIN 66025-1: Pro-
12. Blok LG, Longana ML, Yu H et al (2018) An investigation into grammaufbau für numerisch gesteuerte Arbeitsmaschinen. Beuth
3D printing of fibre reinforced thermoplastic composites. Addit Verlag GmbH, Berlin
Manuf 22:176–186. https://​doi.​org/​10.​1016/j.​addma.​2018.​04.​039 30. Girnth S, Koopmann J, Klawitter G et al (2019) 3D hybrid-mate-
13. Prüß H, Vietor T (2015) Design for fiber-reinforced additive rial processing in selective laser melting: implementation of a
manufacturing. J Mech Des. https://​doi.​org/​10.​1115/1.​40309​93 selective coating system. Prog Addit Manuf 4:399–409. https://​
14. Kuschmitz S, Schirp A, Busse J et al (2021) Development and doi.​org/​10.​1007/​s40964-​019-​00082-w
processing of continuous flax and carbon fiber-reinforced ther- 31. Markforged Inc. (2022) Carbon Fiber Composite 3D Printer:
moplastic composites by a modified material extrusion process. Markforged Mark Two. https://m ​ arkfo​ rged.c​ om/3​ d-p​ rinte​ rs/m
​ ark-​
Materials (Basel) 14:2332. https://​doi.​org/​10.​3390/​ma140​92332 two. Accessed 16 Mar 2022
15. Matsuzaki R, Ueda M, Namiki M et al (2016) Three-dimensional 32. Anisoprint LLC (2022) Desktop Anisoprinting. https://a​ nisop​ rint.​
printing of continuous-fiber composites by in-nozzle impregna- com/. Accessed 08 Feb 2022
tion. Sci Rep 6:23058. https://​doi.​org/​10.​1038/​srep2​3058 33. Markforged Inc. (2022) Eiger 3D Printing Software | Markforged.
16. Tian X, Liu T, Yang C et al (2016) Interface and performance of https://​markf​orged.​com/​softw​are. Accessed 16 Mar 2022
3D printed continuous carbon fiber reinforced PLA composites. 34. Anisoprint (2022) Aura | Anisoprint. https://a​ nisop​ rint.c​ om/a​ ura/.
Compos A Appl Sci Manuf 88:198–205. https://d​ oi.o​ rg/1​ 0.1​ 016/j.​ Accessed 10 Feb 2022
compo​sitesa.​2016.​05.​032 35. Zhang H, Yang D, Sheng Y (2018) Performance-driven 3D print-
17. Tian X, Liu T, Wang Q et al (2017) Recycling and remanufactur- ing of continuous curved carbon fibre reinforced polymer compos-
ing of 3D printed continuous carbon fiber reinforced PLA com- ites: a preliminary numerical study. Compos B Eng 151:256–264.
posites. J Clean Prod 142:1609–1618. https://​doi.​org/​10.​1016/j.​ https://​doi.​org/​10.​1016/j.​compo​sitesb.​2018.​06.​017
jclep​ro.​2016.​11.​139 36. Malakhov AV, Polilov AN (2016) Design of composite struc-
18. Fischer A, Rommel S, Bauernhansl T (2013) New fiber matrix tures reinforced curvilinear fibres using FEM. Compos A Appl
process with 3D fiber printer—a strategic in-process integration Sci Manuf 87:23–28. https://​doi.​org/​10.​1016/j.​compo​sitesa.​2016.​
of endless fibers using fused deposition modeling (FDM). Digit 04.​005
Prod Process Dev 2013:167–175 37. Malakhov AV, Polilov AN, Li D et al (2021) Increasing the bear-
19. Hu Q, Duan Y, Zhang H et al (2018) Manufacturing and 3D ing capacity of composite plates in the zone of bolted joints by
printing of continuous carbon fiber prepreg filament. J Mater Sci using curvilinear trajectories and a variable fiber volume frac-
53:1887–1898. https://​doi.​org/​10.​1007/​s10853-​017-​1624-2 tion. Mech Compos Mater 57:287–300. https://​doi.​org/​10.​1007/​
20. Goh GD, Dikshit V, Nagalingam AP et al (2018) Characterization s11029-​021-​09954-1
of mechanical properties and fracture mode of additively manu- 38. Sugiyama K, Matsuzaki R, Malakhov AV et al (2020) 3D printing
factured carbon fiber and glass fiber reinforced thermoplastics. of optimized composites with variable fiber volume fraction and
Mater Des 137:79–89. https://​doi.​org/​10.​1016/j.​matdes.​2017.​10.​ stiffness using continuous fiber. Compos Sci Technol 186:107905.
021 https://​doi.​org/​10.​1016/j.​comps​citech.​2019.​107905
21. Dickson AN, Barry JN, McDonnell KA et al (2017) Fabrication 39. Li N, Link G, Wang T et al (2020) Path-designed 3D printing for
of continuous carbon, glass and Kevlar fibre reinforced polymer topological optimized continuous carbon fibre reinforced com-
composites using additive manufacturing. Addit Manuf 16:146– posite structures. Compos B Eng 182:107612. https://​doi.​org/​10.​
152. https://​doi.​org/​10.​1016/j.​addma.​2017.​06.​004 1016/j.​compo​sitesb.​2019.​107612
22. Bettini P, Alitta G, Sala G et al (2017) Fused deposition technique 40. Hou Z, Tian X, Zhang J et al (2021) Optimization design and 3D
for continuous fiber reinforced thermoplastic. J Mater Eng Per- printing of curvilinear fiber reinforced variable stiffness compos-
form 26:843–848. https://​doi.​org/​10.​1007/​s11665-​016-​2459-8 ites. Compos Sci Technol 201:108502. https://​doi.​org/​10.​1016/j.​
23. Justo J, Távara L, García-Guzmán L et al (2018) Characterization comps​citech.​2020.​108502
of 3D printed long fibre reinforced composites. Compos Struct 41. Moldenhauer H (2018) Integration of direction fields with stand-
185:537–548. https://​doi.​org/​10.​1016/j.​comps​truct.​2017.​11.​052 ard options in finite element programs. MCA 23:24. https://​doi.​
24. Melenka GW, Cheung BK, Schofield JS et al (2016) Evaluation org/​10.​3390/​mca23​020024
and prediction of the tensile properties of continuous fiber-rein- 42. Henning F, Moeller E (2020) Handbuch Leichtbau: Methoden,
forced 3D printed structures. Compos Struct 153:866–875. https://​ Werkstoffe, Fertigung, 2nd edn. Hanser, München
doi.​org/​10.​1016/j.​comps​truct.​2016.​07.​018 43. Kessel S, Fröhling D (2012) Technische Mechanik: Zweispra-
25. Kabir SMF, Mathur K, Seyam A-FM (2020) A critical review chiges Lehrbuch zu Grundlagen der Mechanik fester Körper
on 3D printed continuous fiber-reinforced composites: history, : bilingual textbook on the fundamentals of solid mechanics/

13
Progress in Additive Manufacturing (2023) 8:541–559 559

Siegfried Kessel ; Dirk Fröhling = Engineering mechanics, 2., der Werkstofftechnik, 13, überarbeitete und, erweiterte. Springer
überarb. Aufl. Studium. Springer Vieweg, Wiesbaden Vieweg, Wiesbaden, Heidelberg
44. Schnell W, Gross D, Hauger W (1990) Elastostatik, 3. Aufl. 49. Chabaud G, Castro M, Denoual C et al (2019) Hygromechanical
Springer-Lehrbuch, Bd. 2. Springer, Berlin properties of 3D printed continuous carbon and glass fibre rein-
45. Neuber H (1971) Technische Mechanik Methodische Einführung: forced polyamide composite for outdoor structural applications.
Zweiter Teil Elastostatik und Festigkeitslehre. Springer, Berlin Addit Manuf 26:94–105. https://​doi.​org/​10.​1016/j.​addma.​2019.​
46. Van Der Klift F, Koga Y, Todoroki A et al (2016) 3D printing of 01.​005
continuous carbon fibre reinforced thermo-plastic (CFRTP) tensile
test specimens. OJCM 06:18–27. https://​doi.​org/​10.​4236/​ojcm.​ Publisher's Note Springer Nature remains neutral with regard to
2016.​61003 jurisdictional claims in published maps and institutional affiliations.
47. Dutra TA, Ferreira RTL, Resende HB et al (2019) Mechanical
characterization and asymptotic homogenization of 3D-printed
continuous carbon fiber-reinforced thermoplastic. J Braz Soc
Mech Sci Eng. https://​doi.​org/​10.​1007/​s40430-​019-​1630-1
48. Macherauch E, Zoch H-W (eds) (2019) Praktikum in Werkst-
offkunde: 100 ausführliche Versuche aus wichtigen Gebieten

13