Content uploaded by Steffi Silbermann
Author content
All content in this area was uploaded by Steffi Silbermann on May 12, 2020
Content may be subject to copyright.
Textile Architecture for Wood Construction
Name, Affiliation
Steffi Silbermann, University of Kassel, Bau Kunst Erfinden, Germany
Jannis Heise, Daniel Kohl, Stefan Böhm, University of Kassel, Department for Cutting and Joining
Manufacturing Processes, Germany
Zuardin Akbar, Philipp Eversmann, University of Kassel, Department for Experimental and
Digital Design and Construction, Germany
and Heike Klussmann, University of Kassel, Bau Kunst Erfinden, Germany
Short Summary
Our project is an investigation of the design, shaping, simulation, manufacture, and
construction of lightweight load-bearing structural components made of wood-based
continuous-fiber textiles. Our aim is to innovatively adapt established concepts in wood
construction, such as panelized construction and wood framing, to textile construction. We are
developing a continuous filament out of solid wood that can be made into wood-textile
structures. Textiles have many advantages: excellent suitability for light construction, versatility
of form and function, refined and tested manufacturing and processing technologies, and a
characteristic, ever-changing, deeply familiar aesthetic of parallel and crossing threads. Our
ultimate goal is to develop a material-efficient, functional, and aesthetically appealing
architecture based on solid-wood textiles.
Introduction
Textile structures have been known for millennia. Over time, a variety of textile construction
principles have been developed for building cohesive structures out of long, thin, flexible
elements. For most of that time, textiles were made primarily by hand, but with the advent of
industrial textile production, highly complex automated processes have become available,
which are capable of joining different types of fibers to produce a range of textile structures
(Cherif et al. 2011).
2
Such structures are of interest for construction, where they are used as semifinished products,
primarily as the reinforcing element in fiber-reinforced plastics. In these materials, textiles
(typically flat fabrics) are impregnated with plastic, providing stiffness once the plastic has
hardened. Before it hardens completely, the composite can be formed into two- or three-
dimensional shapes, taking advantage of the flexibility of textile structures. The reinforcing
fabrics can be prepared for shaping during the manufacturing process, or even preshaped. The
combination of a rigid, pressure-resistant plastic matrix with the tensile strength of fiber
creates composites with an excellent ratio of strength to weight.
Continuous-fiber textiles are especially interesting since they can be used to produce high-
performance textile fabrics that can be designed for specific applications and formed into
complex shapes (Cherif et al. 2011). Such fabrics are made from fine, flexible continuous fibers,
typically made of plastic, glass, or carbon, as well as flax and sisal. Thanks to their lightweight
design, such textile-based structural elements are seeing increasing use – for example, in
vehicle manufacturing. By reducing weight, they produce greater fuel efficiency while still
providing the strength required in such materials (Branchenbericht 2015).
Between 1956 and 1970, architects made a number of unsuccessful attempts to establish new
construction methods using fiber-based materials (Knippers 2007). Now, 40 years later, those
attempts are being revived and adapted to robot-aided manufacturing by the Institute of
Building Structures and Structural Design (ITKE) and the Institute for Computational Design and
Construction (ICD), both in Stuttgart. Two experimental pavilions made of wound carbon and
glass fibers impregnated with epoxy resin, built in 2014–15 and 2016–17, displayed the
effectiveness of fiber-composite structures, as well as a textile aesthetic unusual in the building
context (Knippers and Menges 2015, 2017).
Lightweight materials can help reduce energy consumption, and therefore CO₂ emissions, not
only in use, but also during transport before or after use . The global reduction of CO₂ in all
Textile Architecture for Wood Construction
3
areas of human culture is an extremely high-priority goal in the 21st century (UNEP 2018). The
energy consumed by a textile structure depends not only on its weight, but also on the
environmental footprint of the fiber type used. Since the fibers currently used in light
construction are predominantly mineral or petrochemical in origin and require a great deal of
energy to produce, there is a strong interest in investigating the capabilities of textile structures
made from previously unused fiber materials that are more environmentally sustainable.
Architecture – a field with little experience of textile-based structural components thus far –
also presents an opportunity to discover and utilize other, architecture-specific advantages of
textile structures. Under the heading of “Textile Architecture for Wood Construction,” our
project explores the possibilities of solid-wood-based textile construction, combining the
outstanding environmental performance of wood with the exceptional structural potential of
textile structures.
Research
Solid Wood Monofilament
Wood-weaving and the use of short wood fibers and drawn wood are recorded from the 19th
and early 20th centuries (Purfürst 1880; Klausegger et al. 2016). Willow withes have been used
by basketmakers for centuries, providing durable strips for weaving. In the 1980s, the Forestry
Institute of the East German Academy of Agronomic Sciences in Eberswalde attempted to
produce continuous fibers from whole willow withes for use in looms (Gutwasser 1990), but
German reunification put an end to those efforts. Today, continuous veneer ribbons are
produced to conceal the edges of wood-composite panels, and can also be used in artisanal
weaving (Janson 2001).
Building on the knowledge of basketmakers, we are using split willow withes as the raw
material for a solid-wood monofilament that can be used in the large-scale automated
manufacturing of high-performance structural elements. A monofilament is a continuous fiber
4
consisting of a single filament, typically more than 0.1 mm in diameter and effectively unlimited
in length. Split willow withes are thin strips about 1.5 m long that are irregular in cross-section,
measuring up to 8 mm wide and 1.5 mm thick. They are usually obtained from one-year-old
withes using a cutting device that splits them off from the outer surface of the branch. Split
willow withes are extremely flexible and tolerant of the bending required of continuous fibers
(Fig. 1).
Fig. 1: Split willow withe, knotted while wet ©Bau Kunst Erfinden
The morphology of willow withes is naturally variable. Because growing withes are wider at the
bottom than at the top, when they are sliced longitudinally the narrow ends of the resulting
split withes are about 25 percent narrower than the wide ends. The straightness of the whole
withes, which can vary dramatically in their curvature, carries over to the split withes (Figs. 2
Textile Architecture for Wood Construction
5
and 3). After splitting and stripping, the sides, top, and bottom remain wavy. Average thickness
is about 1 mm, with a standard deviation of just under 19 percent.
All this means that the cross-section varies over the entire length of the split withe, with
tolerances of less than a millimeter. Standardizing the cross-sectional shape would be beneficial
for all subsequent steps in the automated manufacturing process. Preliminary studies have also
shown that standardization can increase the tensile strength of the withe’s cross-section in
comparison to split but otherwise unmodified withes. In addition, standardizing the thickness
or width of the split withes has been observed to produce more uniform tensile-strength and
maximum tensile-force ratings. Furthermore, one of our aims is to produce different
monofilaments in a range of standard cross-sectional shapes, which can then be manufactured
into textiles with varying characteristics.
Fig. 2: Unmodified split willow withes, approx. 120 cm long ©Bau Kunst Erfinden
Fig. 3: The withes from Figure 2, reduced to 3.1 mm wide (length unchanged) ©Bau Kunst
Erfinden
6
Our first step has been to develop and compare several methods of standardizing the width and
thickness of the split withes. Since these are quite small in cross-section, generally measuring
less than 8 x 1.5 mm, and tend to twist and bend during processing, familiar techniques for
modifying the width and thickness of wood have had to be modified to account for these
characteristics. This has primarily involved the development of semi-automated tools of our
own construction.
To modify the width, we are experimenting with saws and shapers. Using a double-sided
shaping tool on unstandardized withes measuring 120 cm long and 4–7.5 mm wide, we have
succeeded in obtaining a uniform width of 3.1 mm, with variations of less than 0.1 mm, in over
80 percent of our test pieces. We are also conducting comparative analyses to assess the
quality of the material removed. The object here is to prevent frayed edges and avoid defects
arising from common flaws in wood, such as knots. With careful handling of the workpiece, it is
also possible to make maximum use of the material by following the curve of the withe with the
shaper.
To combine individual withes into a continuous fiber, we have analyzed classic woodworking
joints and adapted them to the characteristics and demands of willow withes to develop a
suitable mechanical process for joining split withes end to end. The focus here is on developing
the technology and geometry of the joint, taking into account influences such as residual
moisture from the gluing process and aging caused by warmth, moisture, and other
environmental factors. This process takes place in tandem with the sectional processing.
Split willow withes are available in Germany from a number of different countries. We have
conducted a comparative analysis of material originating from Poland and from Spain. Not only
were the Spanish withes a lighter, more characteristic willow color, they had fewer flaws and
broke much less frequently during initial manual processing tests. They were also about 1 mm
wider on average (Silbermann n.d.). They have thus become our choice for further testing.
Textile Architecture for Wood Construction
7
Besides disparities in width, the split withes also display varying textures and visual qualities
that could potentially come into play when the withes are processed into monofilaments or
textiles. These differences involve the integrity, straightness, width, flatness, and color of the
material. We are developing a grading system for split willow withes based on these
characteristics. By constantly shifting perspectives between classifying the raw material and
modifying its width and thickness via the processes we are developing, outlined above, we can
refine our grading criteria to better understand and exploit the relationships between raw
material, processing, and monofilament performance.
Making Textiles out of Solid-Wood Monofilament
Our novel wood monofilament is next formed into textile fabrics (Figs. 4 and 5). The most
important types of fabrics for us are woven, laid, wound, and braided constructions. These
possess the greatest strength, because the fibers in them lie flat. Variations in the cross-
sectional shape of the monofilament have a direct effect on the qualities of the textile. For
example, monofilament that is wider and thicker produces a textile that is coarser in texture,
stiffer, and faster to manufacture, whereas a fabric woven of thin and narrow monofilament is
much more delicate and laborious to produce but much stronger in composites.
Using textiles as load-bearing elements requires a fundamental understanding of their
structural connections and resulting properties. Each of the four fabrication methods
mentioned has its own unique parameters that affect the textile’s strength. These parameters,
such as fiber spacing, weave pattern, fiber direction, and fiber density, interact both with each
other and with the monofilament cross-section, which can be adjusted as needed. To discover,
define, and consistently achieve promising combinations, we are investigating the basic
characteristics of textile wood-monofilament constructions, including both biaxial (anisotropic
and orthotropic) and unidirectional structures. The parameters can be varied to obtain stiffness
as well as pliability in different component geometries. These technical-functional properties
can also be studied and specified in combination with the aesthetic qualities of the wood-textile
components to create structures that are both functional and attractive.
8
Fig. 4: Woven fabric made from solid-wood monofilament ©Bau Kunst Erfinden
Fig. 5: Braided tube made from solid-wood monofilament ©Bau Kunst Erfinden
Textile Architecture for Wood Construction
9
To make textile fabrics out of solid-wood monofilament, we plan to use established textile
manufacturing processes and robots. The goal is an industrial, fully automated process.
Although willow monofilament is unusually flexible in comparison to other solid-wood
products, it is still relatively stiff compared to more common fibers. Therefore, our processing
methods must be adapted to this new material. One of the first steps is to develop appropriate
spools on which the monofilament can be wound and transported (Fig. 6). Other adaptations
will be necessary to be able to feed the material into weaving machines or robots.
A spool we have developed for weaving is 13 cm in diameter and 18 mm wide (Fig. 7). It can
hold approximately 15 m of 1-mm-thick willow monofilament. To hold longer lengths, the
diameter can be increased. This results in large spools, meaning that sufficient space must be
provided on and around the machines used in processing. Normally, the warp beam of a loom
holds all the fibers running lengthwise; the loom we are using has duplicate warp beams
positioned at various heights. The fibers from these beams converge on one level to form the
warp, into which the weft fibers are woven.
Fig. 6: Solid-wood monofilament wound on a spool for transport ©Bau Kunst Erfinden
10
Fig. 7: Solid-wood monofilament on a spool adapted for weaving ©Bau Kunst Erfinden
Simulation
A special challenge for digital simulations is to analyze and optimize the monofilament in terms
not only of its material composition and joinery, but also of its conceptual applications and
workability on the object or the architecture level. These discrete criteria place thoroughly
different demands on the simulation technology. The goal here is to investigate the
parameterizability of the joint geometry and implement an appropriate simulation method,
using the finite element method and dynamic relaxation simulation, and ultimately developing
a general material simulation model for strength-optimized textiles made from solid wood
monofilament.
The weaving technique we have developed for this research is based on matrix multiplication
Textile Architecture for Wood Construction
11
and defines a fundamentally binary structure: When the weft interlaces with the warp, every
warp thread is either up or down. We have developed an interface based on Grasshopper (Fig.
8), a visual scripting plugin for Rhinoceros 3D, to build a digital twin to help with modeling and
organizing a material data space based on different matrix configurations. This data space will
support the early design process by enabling us to develop different variations of patterns on a
single surface or structure. A single pattern is developed using an interactive matrix editor as an
instance of the object, with particular attributes related to pattern parameters such as matrix
domain, pattern binary, and willow size, as well as parameters for later physics simulation, to
be used in pattern form-finding. This model will work as a micro-scale material-system model
that will then be calibrated with experimental test results. A multiscalar modeling approach will
be investigated to develop a robust design tool with the flexibility to adjust the parameters of
various details. The use of a GPU-accelerated application will also provide opportunities for
further high-resolution modeling during the macro-scale design phase.
Fig. 8: Digital representation of a weaving binary pattern in Rhinoceros 3D and the interactive
matrix editor ©Department for Experimental and Digital Design and Construction
12
Conclusion
Our project is making possible the automated manufacturing of solid-wood textiles and
harnessing the potential of their light weight, shapability, and aesthetic appeal. The key
element is the continuous solid-wood fiber we are using: a novel monofilament made from
willow withes, a rapidly renewable resource in Europe. By adapting existing textile
manufacturing processes to the new material once development of the monofilament is
complete, we can produce semifinished wood-based textile products for further processing into
high-performance composites for construction. Because textile construction is a form of
joinery, other fiber types can be added to the wood as well – for example, stabilizing or
functional fibers – thus laying the foundation for a wide variety of possible wood-textile-based
structural elements. These can be refined in an iterative process involving concepts from
architecture, vehicle design, and product design, and connections between these areas can be
developed as well. We are merging the aesthetics and structural advantages of textiles with
those of wood, giving rise to a new formal language for wood construction – the language of
textile tectonics.
References
Commerzbank. “Branchenbericht Technische Textilien”. Industrieverband Veredelung - Garne –
Gewebe – Technische Textilien e.V. http://www.ivgt.de/de/home/details/article/-
0c83fa017a.html. (10.12.2018).
Cherif, Chokri. Textile Werkstoffe für den Leichtbau. Berlin-Heidelberg: Springer, 2011
Gutwasser, Frank. Chemische Vergütung und Verklebung von Weidenflechtmaterialien mit
technologischen Konsequenzen für den Verarbeitungsprozess. Thesis. Eberswalde: Akademie
der Landwirtschaftswissenschaften der DDR, Institut für Forstwissenschaften, 1990.
Textile Architecture for Wood Construction
13
Janson, Manfred. Flechtmaterialstreifen zur Herstellung eines Flächenflechtwerkes oder eines
Formflechtwerkes. Filed by Janson, Manfred on 12.12.2001. Publication No.: DE 201 20 158 U1,
2001.
Klausegger, Anton. “Holzdraht”. Atterwiki. www.atterwiki.at/index.php?title=Holzdraht
(12.12.2018).
Knippers, Jan. “Faserverbundwerkstoffe in Architektur und Bauwesen.” Themenheft Forschung
Leichtbau 3 (2007): 58–68.
Knippers, Jan; Menges, Achim. “ICD/ITKE Forschungspavillon 2014-15”. itke. www.itke.uni-
stuttgart.de/de/archives/portfolio-type/icditke-research-pavilion-2014-15 (12.12.2018).
Knippers, Jan; Menges, Achim. “ICD/ITKE Forschungspavillon 2016-17”. Itke. www.itke.uni-
stuttgart.de/de/archives/portfolio-type/icditke-research-pavilion-2016-17 (12.12.2018).
Purfürst, Otto. “Die Sparteriewaarenerzeugung”. Die Gartenlaube 9 (1880): 148–151
Silbermann, Steffi. Gewebe aus Weidenholz-Monofil zum Einsatz in textilen Bauteilen.
Unpublished thesis. Kassel: University of Kassel, n.d.
UNEP. “Emissions Gap Report 2018”. United Nations Environment Programme.
www.unenvironment.org/resources/emissions-gap-report-2018 (10.2.2019)
