Content uploaded by Nikolas Frueh
Author content
All content in this area was uploaded by Nikolas Frueh on Jul 25, 2018
Content may be subject to copyright.
Proceedings of the IASS Symposium 2018
Creativity in Structural Design
July 16-20, 2018, MIT, Boston, USA
Caitlin Mueller, Sigrid Adriaenssens (eds.)
Copyright © 2018 by ITKE and ICD, University of Stuttgart
Published by the International Association for Shell and Spatial Structures (IASS) with permission.
ICD/ITKE Research Pavilion 2016/2017: Integrative Design of a
Composite Lattice Cantilever
James SOLLY*, Nikolas FRUEH, Saman SAFFARIAN, Marshall PRADOa, Lauren VASEYa, Benjamin
FELBRICHa,, Daniel REISTa, Jan KNIPPERS, Achim MENGESa
*Institute of Building Structures and Structural Design, University of Stuttgart
Keplerstrasse 11, 70174 Stuttgart
j.solly@itke.uni-stuttgart.de
a Institute for Computational Design and Construction, University of Stuttgart
Abstract
This paper describes the design and fabrication of the ICD/ITKE Research Pavilion 2016/2017, the most
recent in the series of experimental installations developed as an outcome of the design-and-build studio
of the ITECH Masters Programme. The completed structure is a 12m long cantilevering lattice-
composite shell that was wound in one piece by a multi-machine fabrication system using coreless
filament winding. To realise such a structure through this fabrication process involved a negotiation
between architectural design, structural requirements and fabrication constraints, details of which are
found in this paper. Technical details of the multi-machine fabrication system were previously described
in Felbrich et al. [1].
Keywords: Fibre-Reinforced Composite, CFRP, GFRP, Shell Structure, Robotic Fabrication, Additive Fabrication
Figure 1: ICD/ITKE research pavilion 2016/2017 at the opening event in March 2017 in Stuttgart
1. Introduction
The ICD/ITKE Research Pavilion 2016-17 is the latest in an ongoing series of collaborative studies at
the named institutes on the generation of lattice composite structures through the process of robotic
coreless filament winding. While the previous pavilions have investigated and then demonstrated
fabrication strategies possible within the limited-reach of an industrial robot, the presented work
considers a major expansion of the fabrication space for winding. The extension of this volume within
Proceedings of the IASS Symposium 2018
Creativity in Structural Design
2
which a monocoque structure may be created offers both new structural possibilities and challenges,
defined by limits of the selected fabrication strategy, that are outlined within this paper.
The project was realised in a collaboration between researchers of both institutes and students on the
ITECH International Masters Programme and resulted in the creation of a 12m long monocoque
cantilever formed from a fibre-reinforced composite lattice and fabricated through the coreless filament
winding process by a novel multi-machine system that incorporated both UAVs and Industrial Robots.
The structure was installed on the Stadtmitte Campus of the University of Stuttgart in April 2017 and
was later moved to the Zentrum für Kunst und Medien in Karlsruhe, where it currently resides.
2. Background and context
2.1 Lightweight structures and composite materials
J.Schlaich and M.Schlaich (of Schlaich Bergermann and Partners) wrote in 2000 that any structure
designed intelligently and responsibly aspires to be “as light as possible” [2] and they note that a
structure can be considered as “lighter” if the ratio between dead load and live load is reduced. Thus the
use of modern high-tech materials with superior strength to weight ratios can immediately be considered
as a step towards achieving this definition.
Fibre Reinforced Polymers (FRP) are one such type of modern material. CFRP has a specific strength
of ~780kNm/kgm, over 10 times higher than structural steel at ~45kNm/kg and they have already been
deployed on several projects where this benefit has used to great advantage (for example see Part F
“Case Studies” of Knippers et al. [3]). In addition, the fabrication of FRP from fibres and resin (highly
flexible before curing) offers the opportunity for designed material placement at a local level and the
creation of freeform geometries.
Shell structures are known to reduce the quantity of material required to support a load if the correct
form can be found to enable loads to be primarily carried by in-plane forces, as described by
Adriaenssens et al. [4] for example. Discrete lattices can offer savings compared to continuous shells.
By utilising carefully-oriented bar-type elements, material can be directly aligned with the carried load
and in situations where little material is required for the stress constraint, the amount needed may be
gathered into bundles with greater local section heights, improving buckling resistance compared with
a thin sheet. This is well researched in aerospace where an example is the carbon fibre lattice tube by
IsoTruss Industries that requires only 56% of the material of an equivalent continuous CFRP cylinder
[5] to carry a load.
Considering all of the above, FRP placed in a lattice arrangement over a global shell-type form can be
considered as a highly suitable system for the creation of lightweight structures. The wider benefits of
lightness in structures have been covered in many papers, i.e. Schlaich and Schlaich [2]. In this research
pavilion, one highly beneficial property of lightness was in the ability to transport the completed pavilion
by road without significant cost or complication.
2.2 Coreless filament winding
2.2.1 Coreless filament winding method
Coreless Filament Winding (CFW) is a fabrication method for the creation of fibre-reinforced polymer
(FRP) parts that has been actively developed at the ITKE and ICD since 2011 for the creation of
composite building systems. Details of the process were first described in La. Magna et al. [6]. CFW is
based on the FRP fabrication process of Filament Winding that involves the wrapping of fibre bundles
around a rotating mandril, a process well-used in industry due to the speed of material placement.
As outlined in [6], the use of a core-based fabricated method has significant limitations for use in
architecture (and the wider AEC industry), primarily because each required geometry must have its own
core and therefore a unique part must have a unique core.
CFW involves the winding of resin-impregnated fibre-bundles in a programmed sequence around
winding points that are (typically) located at the boundary of the element being produced. Each
Proceedings of the IASS Symposium 2018
Creativity in Structural Design
3
arrangement of winding points enables multiple geometries through variation in the placement sequence
of fibres. In previous projects these points have been mounted to minimal skeletal frames that can be
demoulded once the composite has cured. The first fibres form a pseudo-mould for all later placements
and final fibre locations depend on fibre-fibre interactions in space. Both of these points highlight the
criticality of using the correct winding sequence.
The material deposition concept of CFW is therefore extremely simple, with fibre bundles pulled from
a spool and wrapped around winding points and one other. As fibres must be under tension during
winding, the formed geometries are examples of tensile anticlastic surfaces. Thus the fabrication process
is a type of form finding similar to that described in Adriaenssens et al. [4] and wound geometries are
often efficient shell forms by default.
The described limitations on fibre placement and form creation and were a critical input to the
engineering design of the Research Pavilion 2016/2017.
2.2.2 Coreless filament winding previous projects
The development of CFW by the ITKE and ICD has been widely disseminated to the public, AEC
industry and the academic community through the presentation of a series of architectural installations.
The ICD/ITKE Research Pavilions of 2012 and 2013-2014 developed as studio projects with students
initiated investigations that were then progressed through research projects within the institutes.
Figure 2: ICD/ITKE research pavilions made from coreless-wound fibre-reinforced polymers
The Research Pavilion 2012 [8] was a monocoque structure fabricated by mounting the skeletal winding
frame onto a rotary axis then placing fibres using an industrial robot. This pavilion effectively
demonstrates the largest scale easily achievable with a single industrial robot mounted in a static
location.
The Research Pavilion 2013-2014 [9], the University of Stuttgart Fair Stand [10] and the Elytra Filament
Pavilion [7] all demonstrate the use of CFW for the creation of components that are then assembled into
a larger structure. These modular elements enabled the creation of larger structures without increasing
the size of the fabrication setup but added complexity in the necessity for joints.
The Research Pavilion 2016-2017 studio took its starting point from the intention to create a jointless
structure, considered in the 2012 demonstrator, but with the greatly improved CFW system knowledge
developed through the, more recently completed, modular projects.
3. Research Pavilion 2016/2017
3.1 Integrative design
The research aim of the ICD/ITKE Research Pavilion 2016/2017 was to build a fibre-reinforced polymer
structure that extended beyond the reach of a single industrial robot.
The underlying ITECH approach is to identify the coupled implications of design constraints on form,
function and fabrication of the structure. Due to a fully model-based process, the impact of various
options could be iteratively simulated and visualised in order to progress towards an optimal solution
within the range of available possibilities. This process included physical prototypes as well as virtual
Proceedings of the IASS Symposium 2018
Creativity in Structural Design
4
ones in order to understand the influence of winding sequence on the geometry of the pavilion. The final
sequence was derived from virtual models using combinatorial algorithms then tested on physical
models that provided information to be fed back into the algorithms for later iterations.
An unmanned aerial vehicle (UAV) was chosen to pass material between two independent fabrication
environments that each contained a fixed-position industrial robot. The use of a UAV showcases the
potential for this fabrication system to be scaled. In combination with precise control over the pretension
of fibres during fabrication, the completed setup would allow an easy extension of the length of the
structure along the axis between the stations. The details of the multi-machine fabrication system are
explained in [1].
3.2 Structural design methodology
The thin structural surface of the CFW cantilever was designed to support external loads (wind, snow
etc.) and the structural form-finding and design to achieve this was divided into two phases. In the first
phase, a parametric model of the pavilion surface was meshed into shell elements and evaluated as a
normal structural shell for stress concentrations and buckling performance. The highly anisotropic
constituent materials were characterised for this phase by the creation of representative sample lattices
that were tested in the lab to generate pseudo-isotropic material properties for the shell elements.
Figure 3: Calibration process for defining pseudo-isotropic material properties of a carbon and glass fibre
composite lattice structure
The geometry of the surface was generated through a membrane-relaxation between the two boundaries
to give an approximation of the tensile wound form which could be converted into a structural model
and the relative performance measured. With this as an output metric, the boundary geometry and
membrane form-finding parameters were handed to an optimisation algorithm in order to determine the
most performative possible form. The controllable parameter ranges were limited to ensure that any
output geometry would be fabricatable using CFW.
Early in the design process it was noted that the edges of the shell were in compression and susceptible
to lateral buckling. The “folded” edges where therefore added to generate a semi-closed profile formed
through wrapping carbon fibres from outside to inside around the initially-laid sheet geometry. Both
arrangement of anchor points and fibre winding syntax were tailored towards enabling this essential and
final step.
The results of this first shell-analysis phase was a shell model of the most optimal structural geometry
within the boundary conditions selected. The results from this model were used to determine critical
buckling regions and areas of high and low stresses, generating a “criticality map” of the surface.
Proceedings of the IASS Symposium 2018
Creativity in Structural Design
5
In the second phase of analysis, this criticality map was combined with a map of windable fibre paths
(determined, as mentioned earlier through a combination of algorithmic design and physical
prototyping) to select candidate fibres for inclusion in the final design and discard those that would cover
non-critical regions along their length. These candidate fibres were used as inputs to a beam element
model of the structure, which was then run through a size-optimisation algorithm that could select from
a series of possible fibre-beam cross-sections. This iterative process was repeated for each of the layers
of carbon fibre strands to achieve a stable, optimal fibre arrangement that could then be queried to
determine the number and sequence of fibre-roving placements needed.
Figure 4: FEA models (left to right): initial model with shell elements for geometry evaluation, refined model
using beam elements for final analysis
3.3 Fabrication
Following several design iterations the pavilion consisted of four key elements: (i) initial bent flat sheet
of the portal, (ii) volumetric portal, (iii) cantilever shell, (iv) folded edge.
The initial GFRP sheet of the portal was produced flat on a timber frame with embedded sleeves for
winding later fibre layers and to allow connection to the steel foundations (feet).
Once cured this was elastically bent into shape over the foundations then reinforced using carbon fibre
bundles. The fibre rovings were pre-impregnated through a resin infusion process (tow-preg) developed
by the Institute of Aircraft Design, University of Stuttgart. Glass fibres were wound within this bent
sheet to generate the inner surface of an arch with a “D”-shape cross-section. This was later reinforced,
again using carbon fibre, to complete the volumetric portal, the first complete part of the final structure.
This was loaded to confirm its stiffness and strength and scanned to compare as-built and design
geometry. Three configurations of load test (1-5 kN) were performed: (i) centric horizontal load, (ii)
eccentric horizontal load left, (iii) eccentric horizontal load right. The maximum deformation recorded
was just 10 mm and geometry deviation was within 5-70mm of the digital design. These results were
used to update the central model, something crucial for the form finding and structural analysis of the
cantilever.
The cantilever shell started with a glass fibre layer serving as a supportive mould for the primary load-
carrying CFRP ribs. During winding the as-built geometry of the strands was surveyed to ensure the
digital model reflected reality so the structural model could be updated if required and the correct
amounts of reinforcement placed. While deviations were observed during fabrication, the model did not
vary so significantly during fabrication that extra material was required.
The folded edge was constructed in a similar manner, with a glass-fibre guiding layer placed before
carbon fibre reinforcement was added.
Proceedings of the IASS Symposium 2018
Creativity in Structural Design
6
Figure 5: Glass fibre body of the cantilever with initial carbon fiber strands
Figure 6: Pavilion fibre trajectories with different fibre path-types shown in different colours
3.4 Support systems
Aside from the high-tech fabrication system, the pavilion required temporary supports that could bear
the relatively high tension forces and resulting bending moments during winding.
The portal rests on project-specific steel foundations (feet), to which the 184 km of resin impregnated
fibres were anchored. The shape of the feet defined the possible winding geometry for both the portal
and the cantilever.
Figure 7: Shape defining, project specific foundation components anchored to the formwork system
The beams below the portal are a permanent part of the structure, resisting winding forces during
fabrication and also preventing overturning when installed. The workshop had rail systems embedded
in the floor allowing the formwork to be connected to the foundation of the building. On site the space
in between the beams was filled with gravel (20 kN) to prevent the cantilever from lifting and tipping.
The tip of the cantilever was held up by push-pull props and support beams, designed to withstand
winding forces. The necessary arrangement of anchor points for the tip was provided by a custom steel
extension that could be disassembled so only the anchor points remained embedded in the composite.
Proceedings of the IASS Symposium 2018
Creativity in Structural Design
7
A component based steel system, typically used for concrete formwork, was used for the major beams
and props. All of the parts belonging to this system were returned following project completion.
Figure 8: Formwork system of the cantilever (left to right): portal frame and permanent part of the structure,
temporary tip frame with push-pull props and steel extension
4. Conclusion and Outlook
The project demonstrates that CFW lattice composite systems can be used to create performative,
lightweight, monocoque shell structures. While using industrial robots as fabrication tools, the
dimensional limitations of a single unit were expanded. The final structure currently retains its integrity
after being moved twice, experiencing significant weather conditions through winter and being climbed
upon by many adventurous interlopers (not a recommended activity).
The project furthered the two institute’s research on design and simulation of windable geometries,
structural modelling of CFW structures, the technical process of robotic winding and recorded some
additional constraints for the process. For example, to achieve the specified fibre directions and quantity
at the root of the cantilever resulted in more material than required at the tip due to many fibre paths
needing to travel to this winding frame.
Future CFW structures could utilise multiple winding frames to avoid this problem, the winding process
could be automated or manual (depending on project requirement) and components could be
prefabricated and later assembled or fully wound on site. The combination of high strength-to-weight
ratio, longevity of the material and the compact format of fibres and resin before winding and curing
makes these structures suitable for use in remote or inaccessible areas.
Below we present a concept for a CFW bridge that stems from the completed Research Pavilion. The
two sides could be considered as separate components but there is also the possibility for the central
form to be a integrated winding frame, supported in space by a few initial fibres to provide an
intermediate deviation point for all remaining fibres. This helps to overcome the issues noted above in
expanding a fibre-wound tensile surface over longer distances.
Figure 9: CFW Future Bridge Concept
Proceedings of the IASS Symposium 2018
Creativity in Structural Design
8
Figure 10: CFW Future Bridge Concept
References
[1] B. Felbrich, N. Früh, M. Prado, S. Saffarian, J. Solly, L. Vasey, J. Knippers, and A. Menges, “Multi-
Machine Fabrication: An Integrative Design Process Utilising an Autonomous UAV and Industrial
Robots for the Fabrication of Long-Span Composite Structures,” in ACADIA 2017: Disciplines +
Disruption, 2017, pp. 248–259.
[2] J. Schlaich and M. Schlaich, “Lightweight structures,” in Widespan Roof Structures, M. R. Barnes,
Ed. 2000.
[3] J. Knippers, J. Cremers, M. Gabler, and J. Lienhard, Construction manual for polymers membranes:
materials, semi-finished products, form-finding, design. Basel: Birkhauser, 2011.
[4] S. Adriaenssens, P. Block, D. Veenendaal, and C. Williams, Shell structures for architecture: form
finding and optimization. Abingdon, Oxon: Routledge, 2014.
[5] “Structural Capability,” IsoTruss Industries. [Online]. Available:
http://www.isotruss.com/structuralcapability/. [Accessed: 20-Apr-2018].
[6] R. La Magna, S. Reichert, T. Schwinn, F. Waimer, J. Knippers, and A. Menges, “Prototyping
Biomimetic Structures for Architecture,” in Prototyping Architecture, The Conference Papers, 2013,
pp. 224–244.
[7] M. Prado, M. Dörstelmann, J. Solly, A. Menges, and J. Knippers, “Elytra Filament Pavilion: Robotic
Filament Winding for Structural Composite Building Systems,” in Fabricate 2017: Rethinking
Design and Construction, A. Menges, B. Sheil, R. Glynn, and M. Skavara, Eds. Stuttgart: UCL Press,
2017, pp. 224–231.
[8] J. Knippers, R. La Magna, A. Menges, S. Reichert, T. Schwinn, and F. Waimer, “ICD/ITKE Research
Pavilion 2012: Coreless Filament Winding Based on the Morphological Principles of an Arthropod
Skeleton,” Archit. Des., vol. 85, no. 05, pp. 48–53, 2015.
[9] M. Doerstelmann, J. Knippers, A. Menges, S. Parascho, M. Prado, and T. Schwinn, “ICD/ITKE
Research Pavilion 2013-14: Modular coreless filament winding based on Beetle Elytra,” Archit. Des.,
vol. 85, no. 5, pp. 54–59, 2015.
[10] “University of Stuttgart Fair Stand, Hannover Fair 2015,” Institute Of Building Structures And
Structural Design. [Online]. Available: https://www.itke.uni-stuttgart.de/archives/portfolio-
type/university-of-stuttgart-fair-stand-hannover-fair-2015. [Accessed: 20-Apr-2018].
