Content uploaded by Simon Schleicher
Author content
All content in this area was uploaded by Simon Schleicher on Nov 15, 2016
Content may be subject to copyright.
260
Bending-Acve Plates
1 Berkeley Weave installed at the
courtyard of UC Berkeley’s College
of Environmental Design (CED).
University of California, Berkeley
Department of Architecture
University of Stugart, Instute
of Building Structures and
Structural Design (ITKE)
Form-Finding and Form-Conversion
1
With this paper, the authors aim to contribute to the discourse on bending-acve structures by
highlighng two dierent design methods, form-nding and form-conversion. The authors compare
the two methods through close analysis of bending-acve plate structures, discussing their
advantages and disadvantages based on three built case studies. This paper introduces the core
ideas behind bending-acve structures, a rather new structural system that makes targeted use of
large elasc deformaons to generate and stabilize complex geometrical forms based on inially
planar elements. Previous research has focused mainly on form-nding. As a boom-up approach,
it begins with at plates and recreates the bending and coupling process digitally to gradually
determine the nal shape. Form-conversion, conversely, begins with a predened shape that is then
discrezed by strategic surface ling and informed mesh subdivision, and which in turn considers
the geometrical and structural constraints given by the plates. The three built case studies exem-
plify how these methods integrate into the design process. The rst case study applies physical and
digital form-nding techniques to build a chaise lounge. The laer two convert a desired shape into
wide-spanning construcons that either weave mulple strips together or connect distant layers
with each other, providing addional rigidity. The presented case studies successfully prove the
eecveness of form-nding and form-conversion methods and render a newly emerging design
space for the planning, fabricaon, and construcon of bending-acve structures.
261
PROGRAMMABLE MATTER
In recent years, the architecture community has witnessed an
increased availability and constant improvement of computa-
onal tools that enable not only advanced geometrical modeling,
but also the integraon of real-me physics-based simulaons
into the design process in common CAD environments. Programs
like Kangaroo Physics or SOFiSTiK are used, for example, to
rapidly form-nd and interact with parcle systems or accurately
analyze structures on the bases of Finite Element Methods (Piker
2013; Lienhard et al. 2011). With the help of these programs,
one can describe and evaluate the mechanical behavior and
structural capacity of a model under simultaneous consideraon
of external forces and internal material stresses.
With the rise of these tools, architects and engineers are
becoming more and more interested in structural systems whose
forms and load states cannot easily be predicted, but instead
result from a delicate balance between geometry, interacng
forces, and material properes. It is here, in parcular, where
physics-based simulaons that provide real-me feedback can
demonstrate their strength. Bending-acve structures illustrate
these interrelaonships and as such are chosen in this paper as a
detailed example (Figure 1).
This newly established structural system is characterized by the
use of large elasc deformaons of inially planar building mate-
rials to generate geometrically complex construcons (Knippers
et al. 2011). While the tradional maxim in engineering is to
limit the amount of bending in structures, this typology actually
harnesses bending for the creaon of complex and extremely
lightweight designs. The underlying idea of exploing a struc-
ture’s exibility in a controlled way is rather simple and extremely
versale. It can be used, for example, as a form-giving and
self-stabilizing strategy in stac structures or as compliant mech-
anism in kinec structures (Lienhard 2014, Schleicher 2015).
Bending-acve structures can be generally divided into two
main categories, which relate to the geometrical dimensions of
their fundamental components. One-dimensional (1D) systems
can be built, for instance, by bending slender rods, while
two-dimensional (2D) systems use thin plates as basic building
blocks. While extensive knowledge and experience exists for 1D
systems, with elasc gridshells as most prominent applicaon,
plate-dominant structures have not received much aenon
yet and are considered more dicult to design. One reason is
certainly that plates have a limited formability, since they bend
mainly along the axis of weakest inera and thus cannot easily be
forced into complicated geometries. However, what makes this
subset of bending-acve structures parcularly interesng from
a mathemacal point of view is the fact that plates have a clear
scale separaon. They are typically very large in one dimension
and progressively smaller in the other two. Their length is spec-
ied in meters, their width in cenmeters, and their height only
in millimeters. Having hierarchical geometrical features facilitates
the further design process of bending-acve plate structures and
makes it easier to assess the structural behavior and accurately
ancipate their deformed geometry with digital simulaons.
Prominent examples for bending-acve plate structures are
Buckminster Fuller’s plydomes and the ICD/ITKE Research
Pavilion 2010 (Figures 2 and 3). While the rst example follows a
raonal approach in which the shape of a sphere is approximated
with a regular ling of idencal plates (Fuller 1959), the second
example takes advantage of computaonal mass customizaon
and joins 500 individual parts together (Fleischmann et al. 2012).
A previous study idened three main design strategies for
bending-acve structures: a behavior-based, geometry-based,
and integrave approach (Lienhard et al. 2013). According to
this study, the rst category refers to the tradional approach
by skilled crasmen who bend building materials intuively
on the construcon site. The other two categories describe a
more scienc approach, in which hands-on experiments and
analycal tests were conducted beforehand and informed the
further design and construcon process. While the geome-
try-based approach relates to the idea of forcing an object to
match a specic target geometry without further consideraon
of material properes, the integrave approach takes exactly
these liming factors into account when exploring a reachable
design space. In order to best contribute to the above-men-
oned classicaon of bending-acve structures, it is the aim
of this paper to further elaborate emerging design trends in the
integrave approach by having a closer look at the techniques of
form-nding and form-conversion.
2 3
2 Buckminster Fuller’s geodesic plydome in Des Moines, Iowa, 1957. The hemi-
sphere spans 7.3 m and is made out of marine plywood sheets with a thickness
of 6.4 mm.
3 ICD/ITKE Research Pavilion 2010 by the University of Stugart spans 10 m and
consists of 80 birch plywood strips with a thickness of 6.4 mm.
262
The term form-nding is best known for its role in the design
of membranes and shell structures and refers to the concept
of using physical models and numerical simulaons to nd
an opmal geometry of a structure in stac equilibrium with
a design loading (Adriaenssens et al. 2014). From the 1950s
onwards, architects and engineers focused on form-nding stra-
tegies that both incorporated materials and forces while enabling
a systemac exploraon of lightweight construcons. They
became an essenal part in the work of people like Buckminster
Fuller, Félix Candela, Heinz Isler, and Frei Oo (Chilton 2000;
Oo 2005). While these early pioneers implemented form-
nding strategies to design shells and membranes determined by
the shape of hanging chains and cloth, these techniques can also
be applied for bending structures.
In the context of bending-acve plate structures and digital
simulaon, form-nding is oen used for a boom-up design
approach. It starts with planar sheets or strips to create the
bending and coupling process in the nal shape (Lienhard,
Schleicher, and Knipper 2011; Fleischmann et al. 2012). By using
spring-based simulaons in Kangaroo Physics or nite element
methods in SOFiSTiK, one can not only determine the resulng
geometry of the deformed structure but also visualize the evolu-
on of stresses within the material while the system is deforming
(Figure 4) (Schleicher et al. 2015). Based on this informaon,
one can cauously bend a component, for instance by following
an ultra-elasc cable approach unl a permissible stress state
is reached (Lienhard et al. 2014). The nal shape and caused
stresses are oen unknown at the beginning, especially when
mulple parts are bent and fastened together, which is a consid-
erable drawback. A designer with a certain aim in mind would
therefore have to conduct mulple simulaons with gradually
changing parameters to move closer to the design objecve.
In comparison to the previous method, form-conversion pursues
a dierent approach when integrang geometrical and material
consideraons into the design of bending-acve plate struc-
tures. Here, the process begins with a predened target surface
or mesh, which is then discrezed and further subdivided into
smaller bent les based on the exibility of the used plates. The
main restricon in this regard is the knowledge concerning the
plates’ material formability. Here, it is parcularly important to
know that for strips and plate-like elements, the basic shapes
that can be achieved by pure bending without stretching are
conical and cylindrical surfaces. These shapes are also referred to
as single-curved or developable surfaces. Aempng to bend a
sheet of material in two direcons simultaneously either results
in irreversible, plasc deformaons or ulmate material failure.
Thus, to expand the range of achievable shapes, it is necessary to
develop other methods for the inducon of Gaussian curvature
into the system.
To overcome the limitaons related to Gaussian curvature,
muldireconal bending can be induced by strategically removing
material and freeing the plates from the sening constraint of
their surroundings. This principle is illustrated in Figure 5 and
a similar approach was presented by Xing et al. (2011). Here, a
connuous rectangular plate is reduced to two orthogonal strips.
Once again, the strips are bent using the ultra-elasc cable
approach of Lienhard et al. (2014). The bending sness of the
plate, depending proporonally on its width, results in a radical
increase of sness in the connecng area between the strips.
Schleicher, La Magna
4 The form-nding approach
starts from a at sheet and uses
contracng elasc cables elements
to generate the nal bent shape.
Simulaons in Kangaroo Physics
allow for quick and interac-
ve models while the soware
SOFiSTiK enables precise shape
and stress analysis based on Finite
Element Methods. (Schleicher et al.
2015).
4
263
PROGRAMMABLE MATTER
As a result, the connecng area remains almost planar and the
perpendicular bending axis remains unaected by the induced
curvature. In this way, it becomes possible to bend the strips
around mulple axes, spanning dierent direcons but sll main-
taining the material connuity of a single element. The center
image of the cross-like strips in Figure 5 depicts the resulng von
Mises stresses and clearly displays an area of unstressed material
at the intersecon between the two strips, which supports the
previous arguments. A local stress concentraon appears at the
juncon of the strips due to the sharp connecng angle as well
as the inevitable geometric sening in that area.
The result of muldireconal bending can be compared with an
analysis of Gaussian curvature on the right. From the plot, it is
clear that the discrete Gaussian curvature of the deformed mesh
is zero everywhere apart from a small, localized area at the inter-
secon of the two branches. This conrms the assumpon that,
for inextensible materials, most developable surfaces (or slight
deviaon thereof) are achievable. Based on this approach, other
arbitrary freeform surfaces can be converted, following the logic
of strategic material removing and dened zones of local bending
and planarity, as demonstrated in the lower images of Figure 5.
To further illustrate the design potenals of a form-nding and
form-conversion approach, the following secon will have a
closer look at three built case studies. While the rst case study
takes advantage of a boom-up, form-nding technique to
design a load-bearing chaise lounge, the laer two demonstrate
the form-conversion of pre-dened shapes into wide-spanning
construcons that gain rigidity by either weaving mulple strips
together or connecng distant layers with each other.
The rst case study represents the previously menoned form-
nding process and demonstrates its design possibilies in the
context of a furniture-scale object. The goal of this bending-ac-
ve plate structure is to meet the highest structural demands yet
only use a minimum amount of material. The applicaon chosen
to address this challenge in built form was a chaise lounge for
one person.
In this project, bending is primarily induced by strategically
removing material from the center of a thin sheet and then
pinching its naked edges together and fastening the deformed
shape with rivets. This technique has mulple benets: gener-
ang intricate forms out of a single planar surface and achieving
three-dimensional shapes that perform structurally in the bent
state. The general design of the chaise lounge followed a gradual
form-nding approach that comprised a series of physical models
and digital simulaons. In a rst step, the geometry of dierent
reference chairs and the relaonship with the human body was
studied and a group of target angles were idened that allow
for a comfortable seang posion (Figure 6). Based on this
informaon, quick sketch models were built out of paper to gain
a beer understanding of the interdependencies between the
cung paern and the angles of the deformed structure aer
the pinching.
The second step was to turn these cung paerns into digital
models. This was done in Grasshopper and the pinching simu-
lated in Kangaroo Physics. At this point, the main advantage of
using this type of spring-based simulaon was the possibility for
real-me feedback and user-interacvity to further modify the
cung paern and improve the design (Figure 7). Furthermore,
5 The form-conversion approach
is informed by the mechanical
characteriscs of a bent material
and applies these principles to
the subdivision process of a
given target geometry. The upper
row shows the muldireconal
bending of a cross-like strip based
on contracng elasc cables.
The center image indicates the
distribuon of von Mises stress.
The image on the right shows the
Gaussian curvature. The boom
row uses these liming factors
for a form-conversion of a target
mesh into an assembly of bent
components.
5
264
6 Dierent reference chairs were analyzed for their seang posion and the angles
therein were recreated by pinching at sheets into a deformed shape.
7 Digital simulaons in Kangaroo Physics were used to quickly test dierent
cung paerns and form-nd the desired geometrical form.
8 Structural validaon of the form-found shape based on FEM simulaons and a
thorough analysis of the appearing minimal bending radii.
9 A series of paper and plasc sketch models were used to gradually approach the
nal shape and ascertain the required cung paern.
10 Full-scale prototype of the bending chaise lounge is built out of 1.6 mm thin
HDPE plasc and is able to carry the weight of a person.
11 Berkeley Weave installaon spans over 4 m and is built out of 480 individual
plywood strips with a thickness of only 3 mm.
tracking the curvature of the mesh and idenfying minimal
bending radii allowed us to draw rst conclusions if this form
could be built out of a specic target material.
The following third step provided much more accurate results
regarding the materializaon of the chaise lounge. Here,
SOFiSTiK was used to generate nite element models with
dened material characteriscs (Figure 8). All surfaces were given
the material properes of high-density polyethylene (HDPE) with
a Young’s modulus of 1200 N/mm2 and a thickness of 1.6 mm
or 3.2 mm for a single or double layer. The model was then
deformed using the ultra-elasc cable approach. Consulng this
slightly more me-consuming method at this stage of the design
process had mulple benets. It allowed us to calculate the
exact geometry of the highly deformed structure and assess the
stresses within. Thus, this simulaon is a much more complete
descripon of the mechanical behavior and structural capacity
of the bending-acve system. Furthermore, the feedback on
the structure’s complex equilibrium state also allowed localizing
potenally dangerous stress concentraons. And last but not
least, having done the simulaon in typical engineering soware
also made it possible to further analyze the chair’s structural
performance once the weight of a human body was added.
As a proof of concept, the chaise lounge was built both as a
series of small-scale models as well as a full-scale chaise lounge
with the dimensions of 2.44 m x 1.22 m x 1.6 mm (Figures
9–10). The construcon material was HDPE, the paerns for
the dierent sheets were cut on a Zünd blade cuer, and the
pieces were connected together with steel rivets. Riveng, and in
parcular blind riveng, was used in this project both to perma-
nently pinch each surface as well as to connect mulple plates
with each other. Since rivets can only transmit tension and shear
forces, their exact posion needed to be determined carefully. In
this regard, the iterave from-nding over mulple simulaons
9
8
6
7
Schleicher, La Magna
265
PROGRAMMABLE MATTER
played an important role. It provided crucial informaon about
the precise geometry of the deformed plates as well as the exact
posion of the rivet holes, which was needed to guarantee the
alignment of all layers. As far as the actual assembly of the chaise
lounge is concerned, the bending of the plasc was rather easy
and could be done manually. In fact, it was surprising how rigid
the structure became once all pieces were fastened together
and the stored elasc energy began to pre-stress the structure.
The nal chaise lounge was capable of carrying the weight of a
person, of course under some deecons but within permied
tolerances (Figure 10).
In contrast to the previous example, which uses only a small
number of parts, designing bending-acve plate structures out
of mulple components is very challenging and thus requires a
dierent approach. For this reason, the second case study applies
a dierent method and aims to demonstrate the design poten-
al of form-conversion (Figure 11). It invesgates an integrave
approach that considers not only bending but also torsion of
slender strips. The saddle-shaped design of the Berkeley Weave
is based on a modied Enneper surface (Figure 12a). This par-
cular form was chosen because it has a challenging anclasc
geometry with locally high curvature. The subsequent conversion
process into a bending-acve plate structure followed several
steps. The rst was to approximate and discreze the surface
with a quad mesh (Figure 12b). A curvature analysis of the
resulng mesh reveals that its individual quads are not planar but
spaally curved (Figure 12c). The planarity of the quads, however,
will be an important precondion in the later assembly process.
In a second step, the mesh was transformed into a four-layered
weave paern with composed strips that feature pre-drilled
holes. Here, each quad was turned into a crossing of two strips
in one direcon with two other strips at a 90-degree angle. The
resulng interwoven mesh was then opmized for planarizaon.
However, only the regions where strips overlapped were made
planar, while the quads between the intersecons remained
curved (Figure 12d). A second curvature analysis illustrates the
procedure well and shows zero curvature at the intersecons of
the strips (blue areas) while the connecng arms are both bent
and twisted (Figure 12e). Specic rounes in the form-conver-
sion process guaranteed that the bent zones stayed within the
permissible bending radii. In the last step, this converted shape
was used to generate a fabricaon model that featured all the
connecon details and strip subdivisions (Figure 12f).
A closer look at the most extremely curved region illustrates the
complexity related to this last step (Figure 13). To allow for a
proper connecon, bolts were only placed in the planar regions
between intersecng strips. Since the strips are composed out of
smaller segments, it was also important to control their posion
in the four-layered weave and the sequence of layers. A paern
was created which guaranteed that strip segments only ended in
layer two and three and are clamped by connuous strips in layer
one and four. A posive side eect of this weaving strategy is
that the gaps between segments are never visible and the strips
appear to be made out of one piece. The drawback, however, is
that each segment has a unique length and requires individual
posions for the screw holes (Figure 14).
To demonstrate proof of concept for this design approach, this
case study was built in the dimensions of 4 m x 3.5 m x 1.8 m
(Figure 11). The structure is assembled out of 480 geometri-
cally dierent plywood strips that were fastened together with
400 bolts. The material used is 3.0 mm thick birch plywood
with a Young’s modulus of EmII = 16471 N/mm2 and Em
1029 N/mm2. Dimensions and material specicaons were
employed for a nite element analysis using the soware
SOFiSTiK. Under consideraon of self-weight and stored elasc
energy, the minimal bending radii are no smaller than 0.25 m
10 11
266
and the resulng stress peaks are sll below 60% of permissible
material ulizaon.
The third case study showcases another take on form-conversion
for bending-acve plate structures that consists of many compo-
nents. This project is a mul-layered arch that spans over 5.2
m and has a height of 3.5 m. It was built to prove the technical
feasibility of using bending-acve plates for larger load-bearing
structures. In comparison to the previous case study, this project
implements a dierent ling paern and explores the possibility
of signicantly increasing a shape’s rigidity by cross-connecng
distant layers with each other. To fully exploit the large deforma-
ons that plywood allows for, the thickness of the sheets had
to be reduced to the minimum, leading once again to the radical
choice of employing 3.0 mm birch plywood. Since the resulng
sheets are very exible, addional sness needed to be gained
by giving the global shell a peculiar geometry, which seamlessly
transions from an area of posive curvature (sphere-like) to one
of negave curvature (saddle-like) (Figure 15a). This pronounced
double-curvature provides addional sness and helps avoid
undesirable deformaon of the structure. Despite the consid-
erable strength achieved by the shape alone, the choice of
using extremely thin sheets of plywood at that scale asked for
addional reinforcement to provide further load resistance.
These needs were met by a double-layered structure with two
cross-connected shells.
As in the previous example, the rst step of the process was to
convert the base geometry into a mesh paern (Figure 15b).
In the next step, a preliminary analysis of the structure was
conducted and informed the oseng of the mesh to create
a second layer. As the distance between the two layers varies
to reect the bending moment calculated from the preliminary
analysis, the oset of the surfaces changes along the span of
the arch (Figure 15c). The oset reects the stress state in the
individual layers, and the distance between them increases in the
crical areas to increment the global resistance of the system.
The following form-conversion process was once again driven
by material constraints and previously determined permissible
stress limits with respect to bending and torsion. The resulng
ling logic that was used for both layers aected the size of
the members and guaranteed that each component could be
bent into the specic shape required to construct the whole
surface. More precisely, this is achieved by strategically placing
voids into target posions of the master geometry, ensuring that
the bending process can take place without prejudice for the
individual components (Figure 15d). Although inially at, each
element undergoes mul-direconal bending and gets locked
into posion once it is fastened to its neighbors. The exible 3.0
12 Form-conversion process and analysis of the Berkeley Weave.
13 Analysis of Gaussian curvature in the area with the highest deformaon.
14 Schemac drawing of the technical details in the four-layered weave.
12
13
14
Schleicher, La Magna
267
PROGRAMMABLE MATTER
mm plywood elements achieve consistent sness when joint
together, as the pavilion, although a discrete version of the inial
shape, sll retains substanal shell sness. This was validated in
another nite element analysis that considered both self-weight
as well as undesirable loading scenarios (Figure 15e).
Finally, aer fabricaon, the structure was assembled on site.
The built structure employs 196 elements unique in shape and
geometry (Figure 16). 76 square wood proles of 4 cm x 4 cm
were used to connect the two plywood skins (Figure 17). Due
to the varying distance between the layers, the connectors had
a total amount of 156 exclusive compound miters. The whole
structure weighs only 160 kg, a characterisc that also highlights
the eciency of the system and its potenal for lightweight
construcon. The smooth curvature transion and the overall
complexity of the shape clearly emphasize the potenal of the
construcon logic. Furthermore, the implemented form-con-
version process can be applied to any kind of double-curved
freeform surface, not only the one built at UC Berkeley’s campus
(Figure 18).
In summary, it can be concluded that the three case studies
clearly illustrate the feasibility of form-nding and form-conver-
sion techniques for the design of bending-acve plate structures.
All three examples showcase an integrave approach that is
directly informed by the mechanical properes of the thin plasc
and plywood sheets, which were employed in the dierent proj-
ects. Their overall geometry is therefore the result of an accurate
negoaon between the mechanical limits of the materials and
16
17
15
15 The form-conversion process of
Bend9 pavilion started from a base
geometry (A) and approximated
this shape with a mesh (B). Based
on a rst structural analysis, the
mesh was oset and turned into
a double layer. This structure was
then converted into an assembly of
bent plates (D). Aer another nite
element analysis (E), a fabricaon
model was generated (F).
16 Detail of the assembled structure
shows the layering of dierent
components and the strategically
placed voids to prevent conicts
between the bent parts.
17 Custom wood proles were used
to cross-connect the two layers
together and thus increase the
structural capacity of the pavilion
signicantly.
268
their deformaon capabilies. The very nature of all three case
studies required a ght integraon of design, simulaon, and
assessment of fabricaon and assembly constraints.
Due to its small number of parts, the bending chaise lounge
was a good case study to demonstrate the potenal of design
processes based on iterave form-nding. Depending on the
simulaon soware used, this method can be very quick and
interacve or parcularly accurate and reliable regarding its
results. This precision, however, comes at the expense of simula-
on speed. Therefore, form-nding meets its natural boundaries
when the number of parts exceeds a certain limit.
The second and third case studies aimed to tackle this challenge
by presenng form-conversion as an alternave design approach
for bending-acve plate structures that consist of many parts.
Furthermore, the Berkeley Weave and the Bend9 pavilion
exemplify the capacity of bending-acve plate structures to be
employed as larger scale, space-framing architectural interven-
ons. For future research, the presented case studies and the
underlying design rounes of form-nding and form-conversion
will serve as rst prototypes for the exploraon of more complex
surface-like shell structures that derive their shape through
elasc bending.
The authors would like to thank the following student team for their
amazing work on the bending chaise lounge and their indirect contribu-
on to this paper: Cindy Hartono, Fei Du, Shima Sahebnassagh, Eleanna
Panagoulia as well as their addional supervisors Prof. Kyle Steinfeld, Prof.
Jonathan Bachrach, and Luis Jaggy. For the Berkeley Weave installaon, the
authors would parcularly like to thank Sean Ostro, Andrei Nejur, and Rex
Crabb for their support. Finally, the Bend9 pavilion would not have been
possible without the kind support of Autodesk’s Pier 9 and its enre sta.
Adriaenssens, Sigrid, Philippe Block, Diederick Veenendaal, and Chris
Williams, eds. 2014. Shell Structure for Architecture: Form Finding and
Opmizaon. London: Routledge.
18 View of the Bend9 structure assembled out of 3 mm thin birch plywood in the courtyard at UC Berkeley’s College of Environmental Design (CED).
Schleicher, La Magna
269
PROGRAMMABLE MATTER
Chilton, John. 2000. The Engineer’s Contribuon to Contemporary
Architecture: Heinz Isler. London: Thomas Telford.
Fleischmann, Moritz, Jan Knippers, Julian Lienhard, Achim Menges,
and Simon Schleicher. 2012. “Material Behaviour: Embedding Physical
Properes in Computaonal Design Processes.” Architectural Design 82
(2): 44–51.
Fuller, R. Buckminster, and Robert W. Marks. 1973. Dymaxion World of
Buckminster Fuller. New York: Anchor Books.
Fuller, R. Buckminster. 1959. Self-strued geodesic plydome. US Patent
2,905,113, led April 22, 1957, and issued September 22, 1959.
Knippers, Jan, Jan Cremers, Markus Gabler, and Julian Lienhard. 2011.
Construcon Manual for Polymers Membranes: Materials, Semi-Finished
Products, Form-Finding Design. Basel: Birkhauser Architecture.
Lienhard, Julian, Simon Schleicher, and Jan Knippers. 2011. “Bending-
Acve Structures—Research Pavilion ICD/ITKE.” In Proceedings of the
Internaonal Symposium of the IABSE-IASS Symposium. London, UK:
IABSE-IASS.
Lienhard, Julian, Holger Alpermann, Christoph Gengnagel, and Jan
Knippers. 2013. “Acve Bending, A Review on Structures Where Bending
Is Used as a Self-Formaon Process.” Internaonal Journal of Space
Structures 28 (3–4): 187–196. doi:10.1260/0266-3511.28.3-4.187.
Lienhard, Julian, Riccardo La Magna, and Jan Knippers. 2014. “Form-
nding Bending-Acve Structures with Temporary Ultra-Elasc
Contracon Elements.” In Proceedings of 4th Internaonal Conference
on Mobile, Adaptable and Rapidly Assembled Structures, edited by N. De
Temmerman and C. A. Brebbia. Ostend, Belgium: MARAS. 107–116.
doi:10.2495/mar140091.
Lienhard, Julian. 2014. “Bending-Acve Structures: Form-Finding
Strategies Using Elasc Deformaon in Stac and Kinec Systems
and the Structural Potenals Therein.” PhD Dissertaon, University of
Stugart.
Oo, Frei. 2005. Frei Oo: Complete Works: Lightweight Construcon,
Natural Design, edited by Winfried Nerdinger. Basel: Birkhäuser.
Piker, Daniel. 2013. “Kangaroo: Form Finding With Computaonal
Physics.” Architectural Design 83 (2): 136–137.
Schleicher, Simon. 2015. “Bio-Inspired Compliant Mechanisms for
Architectural Design: Transferring Bending and Folding Principles of Plant
Leaves to Flexible Kinec Structures.” PhD Dissertaon, University of
Stugart.
Schleicher, Simon, Andrew Rasteer, Riccardo La Magna, Andreas
Schönbrunner, Nicola Haberbosch, and Jan Knippers. 2015.
“Form-Finding and Design Potenals of Bending-Acve Plate Structures.”
In Modelling Behaviour, edited by M. Ramsgaard Thomsen, M. Tamke, C.
Gengnagel, B. Faircloth, F. Scheurer. Berlin: Springer. 53–64.
Xing, Qing, Gabriel Esquivel, Ergun Akleman, Jianer Chen, and Jonathan
Gross. 2011. “Band Decomposion of 2-Manifold Meshes for Physical
Construcon of Large Structures.” In Posters of the 38th Internaonal
Conference and Exhibion on Computer Graphics and Interacve Techniques.
Vancouver, BC: SIGGRAPH.
Figure 2: Marks, 1973
Figure 3: Schleicher, 2010
Figure 4: Schleicher et al. 2015
Figures 6–9: Hartono, Du, Sahebnassagh, Panagoulia, 2015
All other photography: Schleicher and La Magna, 2016
is an Assistant Professor in the Department of
Architecture at the University of California, Berkeley. Simon holds a
doctoral degree from the University of Stugart and worked for the
Instute of Building Structures and Structural Design (ITKE). His trans-
disciplinary work draws from architecture, engineering, and biology. By
cross-disciplinary pooling of knowledge he aims to transfer bending
and folding mechanisms found in nature to lightweight and responsive
systems in architecture.
is a structural engineer and PhD candidate at
the Instute of Building Structures and Structural Design (ITKE) at
the University of Stugart. In his research he focuses on simulaon
technology, innovave structural systems, and new materials for building
applicaons.
