Content uploaded by Julian Jauk
Author content
All content in this area was uploaded by Julian Jauk on Jan 07, 2025
Content may be subject to copyright.
463 | ASCAAD 2024
INTEGRATION OF WEAVING AND SEWING MOTIONS INTO THE
3D PRINTING OF CLAY FOR LIGHTWEIGHT CERAMIC
STRUCTURES
KILIAN KLEIBEL-MCGEE, LUKAS GOSCH, JULIAN JAUK,
KRISTIJAN RISTOSKI AND MILENA STAVRIC
Graz University of Technology, Institute of Architecture and Media,
Graz, Austria
kiliankleibel@gmail.com
lukas.gosch@tugraz.at
julian.jauk@tugraz.at
ristoski@tugraz.at
mstavric@tugraz.at
Abstract. This paper introduces a novel method for fabricating
lightweight ceramic structures by incorporating traditional weaving
patterns and sewing techniques into a 3D printing process. Traditional
clay forming techniques like slip casting or pressing are less economical
for these complex patterns. Instead, the study demonstrates the
feasibility of using additive manufacturing, particularly paste-based
extrusion of clay, to create intricate woven-like structures. The research
focuses on developing an efficient fabrication process that allows for a
range of weaving patterns with varying densities and sizes. By
integrating sewing and weaving motions into the print path, the study
expands the design vocabulary of additive manufacturing. Results were
achieved by using the developed method, enabling the production of
ceramic panels with variable pattern-densities ranging from 73% to
100%. This innovative approach paves the way for advanced
applications in architecture, interior design, and industrial
manufacturing, including enhanced cooling systems, acoustic
optimization, and light or air filtration.
Keywords: 3D printing, clay, weaving, lightweight structures, patterns
K. KLEIBEL-MCGEE, L. GOSCH, J. JAUK, K. RISTOSKI, M. STAVRIC
464 | ASCAAD 2024
1. Introduction
Clay has played a pivotal role in architecture throughout history, serving as a
fundamental material in construction (Baker and Ashby 2018). Over time,
various shaping techniques have evolved, ranging from hand molding and
wheel throwing (Rahaman 2017) to more industrialized processes like
extrusion, slip casting, and pressing, each offering unique advantages and
limitations (Händle 2007). However, these shaping techniques are limited in
mass customization and in producing complex woven-like structures due to
the intricate nature of these designs.
In contrast to traditional shaping techniques, rapid prototyping methods such
as 3D printing offer significant advantages, particularly in the ability to swiftly
modify and adjust parameters (Wolf, Rosendahl, and Knaack 2022). This
flexibility promotes a more sustainable production process, as it eliminates the
need for individual molds required in methods like slip casting. These features
enable the efficient testing and production of a wide range of patterns,
facilitating iterative design and accelerating the overall development process
(Jauk 2024). Additionally, 3D printing offers significant advancements in the
fabrication of these complex woven-like clay structures by providing precise
control over clay deposition, which is essential for creating intricate patterns.
This approach enables varying heights (z-values) within individual layers of
printing paths, allowing the material to arch over previous extrusions, thereby
facilitating the creation of these structures. Furthermore, the extrusion process
in 3D printing closely resembles the flow of threads, which naturally aligns
with the creation of woven structures, enabling the production of complex
patterns. The integration of digital design tools enhances this process by
allowing for precise modeling and simulation before fabrication, reducing the
time and effort needed to achieve the desired outcomes.
The aim of this research is to expand the vocabulary of additive
manufacturing by adding weaving and sewing motions to the 3D printing of
clay. The long-term goal of this research is to advance the creation of
lightweight ceramic facade panels or screens that offer beneficial properties
such as light, sound, and air filtration, as well as cooling and humidity control
to positively influence the environmental climate. Traditionally, mashrabiya
screens in Middle Eastern architecture have achieved similar goals by using
intricate wooden lattice work to regulate light, enhance ventilation, and
provide cooling effects through natural airflow (Bagasi, Calautit, and Karban
2021). This work represents the initial step towards these broader goals. The
research objectives of this paper are to develop a controlled fabrication
process for 3D printing lightweight woven-like ceramic structures and to
elaborate their properties.
INTEGRATION OF WEAVING AND SEWING MOTIONS INTO THE 3D
PRINTING OF CLAY FOR LIGHTWEIGHT CERAMIC STRUCTURES
465 | ASCAAD 2024
2. Background and Related Work
3D printing with clay has emerged as a transformative technology in both
artistic and architectural fields, expanding the possibilities of clay-based
design and construction. The artist Piotr Waśniowski exemplifies this
innovation with his detailed, modular ceramic small-scale objects
(Waśniowski, 2023). This is particularly relevant as it demonstrates the
potential for high-precision fabrication with clay, even at a small scale. ETH
Zurich focuses on advanced digital fabrication techniques, including large-
scale clay 3D printing, to develop innovative construction methods and
materials (Anton et al., 2020; Jenny et al., 2022). The Italian company WASP
contributes to the field on an industrial scale with a range of commercially
available clay 3D printers, highlighting the growing impact and potential of
3D printing in ceramics. Despite these advancements, most 3D clay printing
projects rely on layer-by-layer approach characteristic for fused deposition
modeling (FDP).
Traditional weaving is the process of interlacing two distinct sets of yarn
or threads to form a fabric or cloth. The longitudinal threads are called the
warp, and the lateral threads are the weft. The basic principles of weaving
involve the use of a loom, which holds the warp threads in place while the
weft threads are woven through them. Similar to the process of weaving,
which interlaces threads to produce fabric, 3D printing can employ analogous
motions to construct layered clay structures. By incorporating the techniques
of traditional weaving and sewing, the 3D printing process can unlock new
capabilities in the fabrication of complex and durable clay structures with
unique aesthetics.
Friedman et al. (2014) explored integrating 3D printing, clay, and weaving
through robotic fabrication. Other studies demonstrate advancements like
spatial print trajectories for non-planar printing and complex geometries
(Alothman et al. 2019). The research presented in this paper advances existing
studies by incorporating material properties into the analysis. Further research
is needed to optimize 3D printing due material properties, improve design
software accessibility, and understand the efficient and reliable production as
well as the mechanical capabilities of woven-like ceramic elements.
3. Methodology and Equipment
This research employed a mixed-methods approach, integrating digital
fabrication techniques, 3D printing, with traditional clay crafting methods. For
the implementation of the weaving motions in the utilized printer, the weft
thread was arched over the warp thread during instances of weft thread
overlap, while the warp thread was pressed against the weft thread during
K. KLEIBEL-MCGEE, L. GOSCH, J. JAUK, K. RISTOSKI, M. STAVRIC
466 | ASCAAD 2024
instances of warp thread overlap. The objectives were to; a) Develop an
efficient 3D printing process. b) Investigate material properties and behavior
during drying and firing.
Conducted in two phases, the first phase investigated how 3D printing
processes can be adapted to create woven-like elements by analyzing various
patterns for their printability and conducting bridging tests. The second phase
involved material experiments and analysis of deformation during drying and
firing and the light transmission properties of these elements.
3.1. HARDWARE SETUP
For the fabrication of the samples presented in this research, clay was used as
the base material, alongside technical equipment including a mixer, a 3D
printer, and a kiln.
3.1.1. Material
The clay powder used for the experiments was type 208 (Georg Schneider,
Boden, Germany). As per the manufacturer’s specifications, type 208 exhibits
a water absorption rate of 10% when fired at a temperature of 1070 °C, which
decreases to 7% at a higher firing temperature of 1140 °C. The chemical
composition of this clay type is: SiO2 - 75.0%, Al2O3 - 19.8%, TiO2 - 1.4%,
Fe2O3 - 0.9%, CaO - 0.2%, MgO - 0.3%, K2O - 2.2%, Na2O - 0.2%.
3.1.2. Equipment
A 20-liter MS 12 CR-KIP mixer (Rauch, Graz, Austria) was used to prepare
the clay mass by mixing clay powder with water, resulting in a suitable
consistency for 3D printing. The WASP Delta 40100 Clay 3D printer (WASP,
Massa Lombarda, Italy) was used for additive manufacturing, supported by a
Continuous Feeding System (CFS) to ensure uninterrupted printing. For
firing, a Top 60 R top-loading kiln (Nabertherm, Lilienthal, Germany) with a
maximum temperature of 1320 °C and a 60-liter volume was employed, using
firing programs with peak temperatures from 900 °C to 1250 °C and a
maximum temperature increase of 110 °C per hour.
3.2. SOFTWARE
Existing slicing software, primarily designed for thermoplastic printing with
small nozzles, lacks the precise control required for large-scale clay printing
(1-30 mm nozzles). The unique properties of clay, like flow behavior and
shrinkage, necessitate strategic printing path arrangements. Standard printing
software was inadequate due to the unique logic of the printing paths required
for the woven-like structure. Therefore, Grasshopper, an algorithm editor
INTEGRATION OF WEAVING AND SEWING MOTIONS INTO THE 3D
PRINTING OF CLAY FOR LIGHTWEIGHT CERAMIC STRUCTURES
467 | ASCAAD 2024
within Rhinoceros 3D (McNeel Europe S.L., Barcelona, Spain), was used for
precise path design. It allows intricate customization and manipulation of
printing paths, crucial for achieving the desired complexity.
The Grasshopper script was developed and iteratively refined during
experimentation, incorporating material properties and post-processing
measurements for precise control of the structure's geometry. For the
generation of the G-Code, the software plugin Termite (Julian Jauk, 2023) was
utilized.
3.2.1. Input parameters:
Hex radius specifies the radius of the hexagon u p to 20 cm as this corresponds
to the maximum printable size of the printer.
Density determines the number of warp threads, which in turn influences the
number of weft threads. The higher the number of warp threads, the smaller
the openings in the pattern, which influences pattern-density and therefore the
light and air transmission.
Arc specifies the height of the weft arcs. It is used to determine how much
material is extruded per stitch. It is essential that the arc of the weft thread 2
is sufficiently high to clear the extrusion of the weft thread 1.
Nozzle size (diameter) influences the extrusion width as well as the height of
the various layers, with the latter needing to be at least half the diameter of the
nozzle.
Frame below specifies the height of the substructure, which consists of a
frame and the warps and shifts the pattern by the same ratio in the z-direction.
Frame above specifies the height of the frame on top of the pattern.
Figure 1. Grasshopper script with inputs displayed on the left and outputs on the right.
3.2.2. Output
Travel contains the travel paths as curves connecting the print paths.
Medium contains the weft threads as print paths to control their extrusion
separately.
K. KLEIBEL-MCGEE, L. GOSCH, J. JAUK, K. RISTOSKI, M. STAVRIC
468 | ASCAAD 2024
Slow contains the remaining print paths, including the frame and warp threads.
To enhance clarity, an information window is provided for each of these
outputs, displaying the number of print paths.
Flat creates the print path for a hexagonal flat element that has the same radius
and volume as the woven-like element, and it is used for the comparison of
the elements.
Density calculates the percentage of the woven-like structure that would
remain open if it were projected onto a planar surface.
DensGeo contains the corresponding silhouette of the pattern for a graphical
representation of the pattern-density.
Trim allows excess material to be removed with a syringe needle following
the outermost edge of the element.
4. Method Development
At the beginning of the research, it was decided to restrict the overall geometry
to a hexagonal form, serving as the foundational shape for the weaving
patterns. This choice stemmed from the radial printing area of the clay 3D
printer, where the hexagonal shape allowed for more efficient use of the entire
space, enabling the creation of adjacent samples in a broader range of scales.
The three directional possibilities inherent in the hexagonal geometry
further enable the creation of multiaxial woven fabrics and 2D structures with
at least three yarn systems arranged at non-right angles to one another (Cherif
2016), broadening the design possibilities. Initially, patterns were printed on
a horizontal print bed. Due to the extruder being fixed orthogonally to the print
bed, the nozzle obstructed the print path during negative z-axis movements.
To avoid this obstruction, later patterns were printed on an inclined surface.
4.1. HORIZONTAL PRINT BED
The initial experiments aimed to assess the behavior of clay extrusions when
printed in a single layer on a flat surface with integrated weaving and sewing
motions. Additionally, the experiments sought to optimize the setup for
printing primarily flat objects.
To facilitate sample removal after printing, a 40 cm by 40 cm wooden plate
was used as the print bed and covered with plastic foil. The elastic nature of
the plastic foil allowed it to shrink alongside the samples as they dried,
preventing cracking or warping. To ensure the adhesion of the foil to the
wooden plate, a thin layer of water was applied, securing the foil during
printing; subsequently, the wood absorbed the water after approximately 30
minutes, facilitating the foil to distort along with the samples’ shrinkage.
INTEGRATION OF WEAVING AND SEWING MOTIONS INTO THE 3D
PRINTING OF CLAY FOR LIGHTWEIGHT CERAMIC STRUCTURES
469 | ASCAAD 2024
The mimicking of weaving and sewing motions was achieved by
integrating the z-axis into the print path and executing movements in an arc
from point to point over a flat printed substructure. Upon analysis, it was
found that the adhesion between the clay and the foil is weaker than the
adhesion between two clay extrusions. Consequently, optimal results are
achieved by first printing a substructure followed by a weaving pattern on top.
Additionally, continuous extrusion without interruptions is beneficial, as each
interruption in the extrusion process presents a potential weak point in the
final structure. These results were observed during subsequent experiments,
in which various patterns were tested for their printability. The findings were
gradually integrated into later attempts.
4.1.1. Kumiko Pattern (KP)
KP was built upon the eponymous design, printed in two phases (Figure 2, a,
b). The first layer consisted of flat straight lines in three directions, forming
triangles. The second layer mimicked weaving motions with arcs, connecting
the midpoints of the triangles and moving over the clay extrusions of the first
layer.
4.1.2. Plain Dutch Weave (PDW)
This pattern utilizes the Plain Dutch Weave technique, where the weft thread
passes over every other warp thread. Initially, a substructure is printed to form
the warp threads (Figure 2, c, d). Instead of going under the warp threads, the
weft threads always start their arc on a warp thread, ensuring close contact and
adhesion between the layers.
Figure 2. a) Linear representation of KP; b) the print paths of KP in the second layer were
not continuous, resulting in a disjointed appearance due to the printer starting and stopping at
each arc.; c) linear representation of PDW; d) tiny cracks (~ 0.1 mm) were noticeable at the
start of each arc of PDW due to the tensile forces exerted on the extrusion during printing.
4.1.3. Twilled Dutch Weave 1 (TDW
1
)
TDW
1
pattern is a variation of the PDW pattern, in which the weft threads
pass over two warp threads (Figure 3, a, b). The weft threads starting on every
second warp thread prevented continuous extrusion, causing prominent start
K. KLEIBEL-MCGEE, L. GOSCH, J. JAUK, K. RISTOSKI, M. STAVRIC
470 | ASCAAD 2024
and end points in the printed arcs. This disrupted the intended twill appearance
by breaking the continuous flow characteristic of twilled weaves.
4.1.4. Twilled Dutch Weave 2 (TDW2)
TDW2 builds upon the TDW1 but introduces a spiral arrangement to the arcs,
creating a distinctive twilled aesthetic (Figure 3, c, d).
Figure 3. a) Linear representation of TDW1; b) prominent start and end points disrupted the
intended twill appearance of TDW1; c) linear representation of TDW2; d) material squeezing
at TDW2 impacting the coherent extrusion.
This spiral arrangement strategically conceals the endpoints of the previous
arcs. Unlike the original TDW1, TDW2 allows for shorter arcs while
maintaining the desired visual effect. The steepness of the print path in TDW2
led to the nozzle being in the path of material extrusion during negative z-
movements. This scenario resulted in material squeezing, impacting the
coherent extrusion.
4.1.5. Stockinette Stitch 1 (SS1)
SS1 is an adaptation of the stockinette stitch, which is recognized as one of the
fundamental stitches in knitting. In this adaptation, loops are created starting
on one warp thread and extending over the next one, linking together like a
chainmail (Figure 4, a, b). The print path of the loops was positioned at a
distance above the substructure. Due to the weight of the clay material, the
loops settled over and between the warp threads. Similar to TDW2, downward
movements squeezed the material at the loop endpoints, causing visual
inconsistencies with distinct start and end points.
4.1.6. Stockinette Stitch 2 (SS2)
This pattern is developed from the SS1 involving the linking of start and end
points to ensure continuity (Figure 4, c, d). To address material cracks during
positive z-direction movements, an increased extrusion factor was
implemented at these points. Similar to the SS1, the loops settled between weft
threads due to their own weight. The impact of the nozzle squeezing on
material flow remained evident during negative z-direction movements,
highlighting areas where the start and endpoint were connected.
INTEGRATION OF WEAVING AND SEWING MOTIONS INTO THE 3D
PRINTING OF CLAY FOR LIGHTWEIGHT CERAMIC STRUCTURES
471 | ASCAAD 2024
Figure 4. a) Linear weave representation of SS1; b) clay mass post printing of SS1and visual
inconsistencies with distinct start and end points; c) linear weave representation of SS2; d)
clay mass post printing of SS2 and the impact of the nozzle squeezing on material flow
remained evident.
4.2. INCLINED PRINT BED
While the previously tested samples showed promising results, several
drawbacks were identified in the majority of them. As the extrusion moves in
an arc over an existing extrusion, small cracks develop as a result of the tensile
forces generated during the lifting process. This flaw can be addressed by
increasing the amount of material being extruded at these weak points. It has
also been noted that where the print head moves downward, the nozzle
obstructs the extrusion path, resulting in material compression.
To overcome this problem the print bed can be inclined at an angle α=21.5°
(Figure 5). This specific angle ensures that the print head maintains a
consistent upward or horizontal movement during printing, preventing any
downward movement that could cause compression of the extruded material.
In the next step, different patterns are printed on an inclined print bed and
compared with identical patterns printed horizontally.
Figure 5. Section drawing of the print setup with an inclined print bed. A – Clay extruder. B –
Inclined print bed with a tilt angle α= 21.5°.
4.2.1. Inclined SS2
First, as with the horizontally printed examples, the substructure was printed.
The angle of the print bed did not affect the adhesion of the material. In the
second phase the loops were printed. Starting with the lowest point, they were
printed horizontally so that only the beginning and end of each loop is attached
K. KLEIBEL-MCGEE, L. GOSCH, J. JAUK, K. RISTOSKI, M. STAVRIC
472 | ASCAAD 2024
to the substructure and the rest is printed in mid-air. Due to the weight of the
material, they naturally positioned themselves in the intended location.
Because they were printed beginning from the lowest point of the print bed,
working their way up, the loops were stacked on top of each other and not
interlinked (Figure 6, left). As the nozzle only moves in a horizontal direction,
there is less material compression.
Figure 6. a) SS2 printed inclined from the bottom to the top; b) SS2 printed inclined from the
top to the bottom. The line in black color shows the direction of the vector of steepest descent
of inclined printed bed.
This is particularly evident when compared with examples printed on a
horizontal print bed where the printhead has to move up and down to clear the
existing threads (Figure 4, c, d). The same example was printed a second time
using the same print path, except that the order of the printed loops was
reversed (Figure 6, right). Reversing the loop order interlinks the patterns
while positioning them within the print path of the subsequent layer. This
causes the nozzle to push material aside, creating a terraced structure that
deviates significantly from the original pattern.
4.2.2. Inclined TDW2
To avoid disrupting the material flow, the substructure was printed first
continuously from bottom to top in a zigzag motion. In the second phase, the
weave pattern was printed, starting again from the lowest point. This method
ensured that the print head moved only in the positive z-direction during
printing, preventing interference with the extrusion process. Consequently,
the extrusion maintained a consistent cross-section, resulting in a more
uniform pattern. This uniformity is particularly evident when compared to the
planar-printed sample. The ends of the arches show significant differences due
to the angle of the print bed (Figure 7).
INTEGRATION OF WEAVING AND SEWING MOTIONS INTO THE 3D
PRINTING OF CLAY FOR LIGHTWEIGHT CERAMIC STRUCTURES
473 | ASCAAD 2024
Figure 7. The same pattern printed on a horizontal print bed (a) and on an inclined (α=21.5°)
print bed (b). The inclined printed pattern shows a more coherent appearance, confirming the
positive impact of the modified print bed.
Adjustin g the angle of the print bed successfully en sured a continuous material
extrusion with fewer visible crack formations and material compression.
Notably, the technique of mid-air printing was of particular interest, as it
utilizes the clay's own weight to create arcs over existing extrusions. To
further explore the potential of this technique, the following tests used a
corrugated print bed. While the inclined print bed utilized gravity and the
material's natural settling to form interlinked patterns, the corrugated print bed
builds on this principle to explore additional ways of achieving woven-like
structures without intricate sewing movements. Furthermore, the bridging
capabilities of the material are tested through larger (15-20 mm) gaps between
the weft threads.
4.3. BRIDGING CORRUGATED PRINT BED
Various parameters were tested to determine the optimal settings for achieving
successful gap bridging without extrusion tearing. For this purpose, a
substructure made of PLA was printed. The substructure had a hexagonal
basic shape with a side length of 10 cm. It consisted of 5 supporting crossbars
with a thickness of 8 mm and a spacing of 13.6 mm.
The initial tests were conducted using a nozzle size of 8 mm and an
extrusion amount multiplied by a factor of 1.4. This increase in the volume of
material extruded during printing caused the threads to sag between the
crossbars. At an extrusion factor of 1.4, it was observed that the weight of the
sagging threads caused crack formation or a complete breakage of the threads
(Figure 8, a).
K. KLEIBEL-MCGEE, L. GOSCH, J. JAUK, K. RISTOSKI, M. STAVRIC
474 | ASCAAD 2024
Figure 8. a) With a higher extrusion factor the threads sag deeper and cause crack formation;
b) with each additional layer the sagging is reduced; c) top view of triaxially woven clay mass
printed on a substructure; d) bottom view of the printed elements, clearly showing the imprint
of the corrugated substructure.
Upon reducing the extrusion factor to 1.2, cracks continued to form, although
they were significantly smaller in size. This decrease in the extrusion factor
resulted in reduced sagging of the threads when bridging gaps, while still
preserving the aesthetic characteristics of a woven structure.
In further tests (Figure 8, b), a 6 mm diameter was utilized and the
extrusion factor was kept at 1.2, resulting in less material being transported
and consequently reducing the overall weight. The cracking behavior
remained consistent with the smaller nozzle diameter. In these tests, the
pattern was printed a second time onto the existing structure, offset by half. It
was observed that the sagging of the threads was reduced with each additional
layer. This can be attributed to the fact that the threads could sag freely in the
lowest layer. At each subsequent level, however, the sagging was limited by
the strands of the levels below. The reduced sagging of the threads also
reduced the formation of cracks at higher layers.
In the subsequent series of tests, the same weave pattern was utilized, with the
modification that the pattern was divided into three directions, resulting in a
triaxial woven structure. In textile weaving, increasing the number of
directions in the plane leads to a substantial improvement in slippage
resistance. This is because each interlacing unit then functions as a fully
defined three-hinged arch rather than a four-hinged arch (Cherif 2016). The
nozzle chosen for these tests had a diameter of 8 mm.
By changing the angle of the strands lying on top of each other, the
bridging distance increased (Figure 8, c, d). It was observed that bridging
distances up to 20 mm were achievable without failing. A test distance was
labeled as successful once it was possible to bridge two times out of three,
without a significant number of cracks forming on the surface to ensure
reliable production results.
INTEGRATION OF WEAVING AND SEWING MOTIONS INTO THE 3D
PRINTING OF CLAY FOR LIGHTWEIGHT CERAMIC STRUCTURES
475 | ASCAAD 2024
5. Pattern Analysis
After testing various weaving patterns and different printing methods, the
following criteria were prioritized to guide the selection of the optimal pattern;
a) Densities: Emphasis was placed on choosing a pattern that could be printed
reliably with minimal risk of failure with a wide range of pattern-densities. b)
Scalability: The chosen pattern should be adaptable in terms of size to offer
flexibility for future applications. Considering these criteria, a triaxial plain-
woven pattern (TPW) was selected. This pattern utilizes one warp yarn system
and two weft yarn systems, each of the yarn systems enclosing an angle of
60°. The warp yarn system is also used as a substructure and is the first layer
to be printed. It is printed planar and provides a bonding base for the following
layers. The following two levels are referred to as weft 1 and weft 2. As with
the PDW, the print paths of the weft yarns run in arcs from one warp yarn to
the other, also referred to as stitches.
5.1. DESNITIES
To determine the printable limitations of the chosen pattern, elements were
printed with a variation of pattern-densities. The printing process remained
the same as with the horizontally printed elements. The experiments were
printed on a smaller scale, with a side length of 50 mm and a nozzle diameter
of 4 mm. This scaling offered the advantages of reduced printing time and
shorter drying time. A range of pattern-densities were tested from 73% to
100%. The density was changed by adding warp threads. The density refers
to the percentage of closed area when orthogonally projected onto a flat
surface. It was observed that the addition of warp threads resulted in a
logarithmically increasing density (Figure 9).
Figure 9. TPW range of densities from a) 73.6%, b) 84.1%, c) 90.8%, d) 95.1%.
At a density below 73%, misprints occurred more frequently as the
distances between the warp threads became too large for the weft threads and
they tore off. From a density of 95%, the visual differences become
insignificant and the material is increasingly squeezed as the distances
K. KLEIBEL-MCGEE, L. GOSCH, J. JAUK, K. RISTOSKI, M. STAVRIC
476 | ASCAAD 2024
between the wefts are smaller than the nozzle diameters. The optimal range
for the TPW is between 73% and 95%.
5.2. SCALABILITY
For this test series, the same pattern-density was printed in different
dimensions, with both the size of the element and the diameter of the nozzle
being adjusted to ensure a proportional scaling of the elements. Starting with
nozzle diameters of 2, 4, 6 and 8 mm, the radius of the hexagon was scaled by
the same ratio to 25, 50, 75 and 100 mm (Figure 10).
The pattern-density of the 3D-printed objects is strongly influenced by the
viscosity of the clay mixture. If the nozzle is to be increased in size, the feed
through the CFS must also be increased. As a result, the pressure in the
connecting hose increases and more material can be extruded.
All mentioned nozzle sizes were successfully used for printing; however,
the 2 mm nozzle presented the greatest challenge. Due to its small radius, the
extrusion factor required precise adjustment. With this nozzle size, the ideal
range between excessive material (causing compression) and insufficient
material (leading to weft tearing) was critical. Notably, printing time increased
linearly with nozzle diameter.
Figure 10. Scalability of the elements was achieved
5.3. GEOMETRY COMPARISON
5.3.1. Area to Volume Ratio
The woven-like elements, due to their intricate surface texture, possess a
greater surface area compared to a flat element of equivalent size and volume.
An element with a hexagonal radius of 50 mm was chosen for the
measurement. The woven-like element has a density of 95%, a volume of
V1=29.5 cm³ and an area of A1=227.8 cm². This element was compared with
a flat element with the same volume V1=V2=29.5 cm³, A2=155.2 cm³, what
resulted in V1=V2, A1:A2=146.7:100. The higher surface-to-volume ratio of
the woven-like structures indicates potential areas of application where such
INTEGRATION OF WEAVING AND SEWING MOTIONS INTO THE 3D
PRINTING OF CLAY FOR LIGHTWEIGHT CERAMIC STRUCTURES
477 | ASCAAD 2024
properties could be beneficial, such as in acoustic protection, cooling systems
or air filtration systems. However, further research is required to confirm the
effectiveness of the elements in these specific applications.
5.3.2. Drying Duration
Woven-like and flat elements of the same size and weight (volume), and
surface ratio of 1:1.467, were printed to test the drying duration. The aim of
this test was to determine whether the larger surface area of the woven-like
elements enables faster water evaporation than the flat printed counterparts.
Three different sizes with a hexagonal radius of 50, 75 and 100 mm, nozzle
diameter of 4, 6 and 8 mm and a pattern density of 95.1% were analyzed. The
elements were weighed after printing and then at 6-hour intervals. The water
content in the clay mass was calculated from these measurements. The tests
were carried out at a room temperature of 22° and a humidity of 57.2%.
The woven-like elements reached complete dryness after 86 hours,
exhibiting a 20% faster average drying time compared to the flat elements.
Drying time was significantly influenced by size, with the largest elements
(100 mm radius) taking 58% longer to dry than the smallest ones (50 mm
radius). These findings clearly demonstrate that a higher surface area to
volume ratio accelerates drying due to increased evaporation.
5.3.3. Non-Uniform Shrinkage
Extruding clay aligns its particles, leading to more shrinkage along the
extrusion direction during drying, which causes slight asymmetries in
woven-like elements after drying and firing. To address this, an experiment
was conducted where elements were measured after firing in the three
weaving directions, and their respective shrinkage was determined (Table 1).
TABLE 1. After firing, measurements were taken in each of the thread directions to assess the
non-uniform shrinkage and inform the parametric model. Measurements of samples with
densities ranging from a) 91.0%, b) 94.4% to c) 96.6% to assess the relative shrinkage.
Densities Warp Weft 1 Weft 1
a) 91.0% -12% -12.5% -13%
b) 94.4% -12.17% -12.67% -13.17%
c) 96.6% -12.28% -12.78% -13.28%
The results showed less shrinkage in the warp thread direction, likely due to
shorter print paths and less material compared to weft threads. Shrinkage was
also tested in relation to density, with elements of varying densities (91.0%,
K. KLEIBEL-MCGEE, L. GOSCH, J. JAUK, K. RISTOSKI, M. STAVRIC
478 | ASCAAD 2024
94.4%, and 96.6%) measured in the three weaving directions. Higher-density
elements exhibited greater shrinkage due to the increased material volume,
whereas lower-density elements, with wider spacing between warp and weft
threads, underwent less shrinkage. These findings enabled the prediction in
shrinkage of the elements and could therefore be integrated into the script to
modify print paths, applying both appropriate scaling and deformation to
ensure that, after firing, all elements formed regular hexagons of the same
size.
5.4. LIGHT TRANSMISSION
This study used light transmission tests to compare the actual pattern-density
of fired elements to their pre-calculated densities. Six elements, each with a
50 mm radius and 4 mm nozzle size, were chosen for measurement (Figure
11). These elements had calculated densities ranging from 73.6% to 99.4%.
They were secured with a PCB holder and placed 10 cm in front of an LED
panel for the light transmission tests.
Figure 11. Light transmission a) 91.0 %, b) 94.4%, c) 96.6 %.
To document the light transmission, images were taken with a Canon EOS
6D SLR camera and a 50 mm f/1.8 lens orthogonal to the positioned elements
from a distance of 50 cm. The shooting parameters were: an aperture of f/10,
a shutter speed of 1/40 second and an ISO value of 320 to ensure optimal
image quality under the given lighting conditions. The captured images were
processed to contain only two colors: black, representing solid areas, and
white, indicating voids. This binarization allowed for a clear distinction
between the solid and void regions. Subsequently, the percentage of the total
area occupied by the solid regions was calculated to provide a precise
quantification of the actual density of the elements. This data was used to
compare the measured densities with the previously calculated values. The
aim of this analysis was to determine any discrepancies between the expected
and actual pattern-densities and to identify potential causes for these
differences. Upon analysis of the results, an average deviation of 1.5% was
observed. This discrepancy can be attributed to several potential factors.
Firstly, the viscosity of the clay during the printing process influences the
actual density of the elements. At a constant extrusion factor, clay with a
INTEGRATION OF WEAVING AND SEWING MOTIONS INTO THE 3D
PRINTING OF CLAY FOR LIGHTWEIGHT CERAMIC STRUCTURES
479 | ASCAAD 2024
slightly higher water content leads to more material being extruded through
the nozzle than with drier clay resulting in a larger cross section of the
extrusion. This influences the distances between the warp and weft threads
and therefore the density of the manufactured elements. Secondly, the
discrepancies observed are due to the deformation of the elements during the
drying and firing process. This deformation made it difficult to photograph
the elements completely orthogonally, which was particularly noticeable in
the case of elements with a higher density, as these exhibited the greatest
deformations.
6. Conclusion
This study has successfully demonstrated the feasibility of integrating
traditional weaving patterns and sewing motions into the 3D printing process
to fabricate lightweight ceramic structures. By leveraging the flexibility of
clay extrusion and precise control of printing paths, a range of complex
woven-like structures with varying densities and sizes were produced. The
research highlighted the potential of this method for applications in
architecture, interior design, and industrial manufacturing, including
enhanced cooling systems, acoustic optimization, and light or air filtration.
The developed software script enabled efficient design and fabrication of
woven-like elements, allowing for customization of parameters such as
pattern-density, element size, and nozzle diameter. The investigation into
material properties and behavior during drying and firing provided valuable
insights into shrinkage and deformation patterns, which were subsequently
integrated into the script for improved accuracy. Light transmission tests
revealed discrepancies between pre-calculated and actual densities,
emphasizing the influence of material viscosity and deformation during
drying and firing. The primary limitation of this study is the viscosity of the
material during printing, which restricts the possibilities for non-supported
printing paths. The use of filament-reinforced extrusions presents an
opportunity to overcome these limitations (Jauk et al., 2023).
Future research directions include exploring the mechanical properties of
the woven-like structures, expanding the range of weaving patterns, scaling
the printing path using mass customized 3D printers developed at the research
institution and investigating the effects of wove-like elements in areas such as
acoustic protection, cooling systems, and air filtration.
Acknowledgements
This work was funded by the Austrian Science Fund (FWF) project F77 (SFB
“Advanced Computational Design”). Open Access Funding by the Austrian Science
Fund (FWF).
K. KLEIBEL-MCGEE, L. GOSCH, J. JAUK, K. RISTOSKI, M. STAVRIC
480 | ASCAAD 2024
References
ALOTHMAN, S., IM, H.C., JUNG, F. and BECHTHOLD, M., 2019. Spatial print trajectory:
Controlling material behavior with print speed, feed rate, and complex print path. In: J.
Willmann, P. Block, M. Hutter, K. Byrne and T. Schork, eds. Robotic Fabrication in
Architecture, Art and Design 2018. Cham: Springer International Publishing, pp. 167–180.
Available at: https://doi.org/10.1007/978-3-319-92294-2_13 [Accessed 13 August 2024].
ANTON, A., BEDARF, P., YOO, A., DILLENBURGER, B., REITER, L., WANGLER, T. and
FLATT, R.J., 2020. Concrete choreography: Prefabrication of 3D-printed columns. In:
Fabricate 2020: Making resilient architecture. London: UCL Press, pp. 286–293.
Available at: https://doi.org/10.2307/j.ctv13xpsvw.41 [Accessed 13 August 2024].
BAGASI, ABDULLAH ABDULHAMEED, JOHN KAISER CALAUTIT, AND
ABDULLAH SAEED KARBAN. 2021. Evaluation of the Integration of the Traditional
Architectural Element Mashrabiya into the Ventilation Strategy for Buildings in Hot
Climates. Energies 14 (3): 530. https://doi.org/10.3390/en14030530. [Accessed 14 August
2024].
BAKER, I. and ASHBY, M.F., 2018. Fifty materials that make the world. Cham: Springer.
CHERIF, C., ed., 2016. Textile materials for lightweight constructions: Technologies - methods
- materials - properties. Berlin, Heidelberg: Springer Berlin Heidelberg. Available at:
https://doi.org/10.1007/978-3-662-46341-3 [Accessed 13 August 2024].
FRIEDMAN, J., KIM, H. and MESA, O., 2014. Experiments in additive clay depositions. In:
W. McGee and M. Ponce de Leon, eds. Robotic Fabrication in Architecture, Art and Design
2014. Cham: Springer, pp. 275–284. Available at: https://doi.org/10.1007/978-3-319-
04663-1_18 [Accessed 13 August 2024].
HÄNDLE, F., 2007. Extrusion in Ceramics, Engineering Materials and Processes. Springer.
JAUK, J., 2023. Termite [Software plugin]. Available at:
https://www.food4rhino.com/app/termite [Accessed 13 August 2024].
JAUK, J., GOSCH, L., VASATKO, H. KÖNIGSBERGER, M., SCHLUSCHE, J.,
STAVRIC, M., 2023.
JAUK, J., 2024. Advancing 3D Printing of Clay in Architecture. Doctoral Thesis at Graz
University of Technology. Filament-Reinforced 3D Printing of Clay. In: Materials 16 (18),
pp. 1–15, Available at: https://doi.org/10.3390/ma16186253 [Accessed 13 August 2024].
JENNY, D., MAYER, H., AEJMELAEUS-LINDSTRÖM, P., GRAMAZIO, F. and KOHLER,
M., 2022. A pedagogy of digital materiality: Integrated design and robotic fabrication
projects of the Master of Advanced Studies in Architecture and Digital Fabrication.
Architecture, Structures and Construction, 2 (4), pp. 649–660. Available at:
https://doi.org/10.1007/s44150-022-00040-1 [Accessed 13 August 2024].
RAHAMAN, M. N., 2017. Ceramic Processing and Sintering, Second. CRC press.
WAŚNIOWSKI, P., 2023. Modular ceramic designs through 3D printing. Available at:
https://www.instagram.com/piotrwasniowski [Accessed 13 August 2024].
WASP, 2023. 3D printers. Available at:
https://www.3dwasp.com/en/clay-3d-printer-delta-wasp-2040-clay/ [Accessed 7 August
2024].
WOLF, A., ROSENDAHL, P., and KNAACK, U., 2022. Additive Manufacturing of Clay and
Ceramic Building Components. In: Automation in Construction (133), pp. 6–13. Available at:
https://doi.org/10.1016/j.autcon.2021.103956 [Accessed 13 August 2024].
