See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/311534824

Bio-inspired Kinetic GFRP-façade for the Thematic Pavilion of the EXPO 2012
in Yeosu

Conference Paper · May 2012

CITATIONS READS

41 8,856

4 authors:

Jan Knippers Florian Scheible

Universität Stuttgart Knippers Helbig Advanced Engineering
340 PUBLICATIONS 4,323 CITATIONS 8 PUBLICATIONS 80 CITATIONS

SEE PROFILE SEE PROFILE

Matthias Oppe Hauke Jungjohann

Knippers Helbig Gmbh 14 PUBLICATIONS 124 CITATIONS
44 PUBLICATIONS 140 CITATIONS
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Jan Knippers on 09 December 2016.

The user has requested enhancement of the downloaded file.

Bio-inspired Kinetic GFRP-façade for the Thematic Pavilion
of the EXPO 2012 in Yeosu
Jan KNIPPERS1,2, Florian SCHEIBLE1, Matthias OPPE1, Hauke
JUNGJOHANN1

1
Knippers Helbig Advanced Engineering, Stuttgart, New York, [email protected]
2
Institute of Building Structures and Structural Design (ITKE), University of Stuttgart, Germany

Summary
A new type of kinetic façade system is presented which was inspired by flexible deformation
principles found in plant movements. The project is a role model for a novel application of glass
fiber reinforced polymers (GFRP) for deployable structures as well as for advanced biomimetic
research and design.

Keywords: kinetic façade, FRP, biomimetic research, plant movements, deployable structures
Introduction

1. Introduction
The kinetic media façade is an integral part of the Thematic Pavilion, a major and permanent
building for the Expo 2012 in Yeosu, South-Korea which was designed by SOMA Architecture,
Vienna.
Convertible systems in architecture are usually realized through the combination of stiff elements or
soft textiles with particular movable elements like hinges, rolls, etc.. The use of GFRP allows for
large and reversible elastic deformations and enables us to think about a completely new
interpretation of convertible structures. In nature, plants have evolved a multitude of kinetics based
on elastic deformation without any hinges to serve the opening and closing of flowers, leaf
orientations etc. [1], [2], [3], [4]. The analysis of these principles inspired the kinetic façade
system of the Yeosu pavilion.

Figure 1: Thematic Pavilion (SOMA Architecture), night view (rendering @isochrom)

The facade is 140 m long and between 3 and 13m high. It consists of 108 kinetic GFRP louvers,
which are supported at the top and bottom edge by fixed supports on one corner and extendable
actuators on the other corner. These actuators push the upper and lower edges together and lead to
an elastic bending and a side rotation of the GFRP element. The 13-metre-long louvers are only 9
mm thick and are stiffened at both longitudinal edges with a 200 mm rib and a 30 mm rib including
a hard rubber buffer bar. The rubber bar protects the GFRP material in the closed state, when one
louver rests on the other. In very strong wind conditions which occasionally occur at the South
Korean coast, the façade will be closed and locked automatically. For the different open positions
pressure values were derived from wind tunnel tests.
Within the operational mode the louvers are individually actuated and create animated patterns
along the façade. The potential choreography ranges from subtle localized movement to waves
patterns moving over the whole elevation of the façade.

2. Operational Modes
2.1 General architectural intention
The operable louvers fulfil a climatic function and allow different modes of operation depending on
the user’s needs. Within the operation mode the louvers are individually actuated and create
animated patterns along the façade. The potential choreography ranges from subtle local motion to
overall waves affecting the whole length of the façade.

Figure 2. Thematic Pavilion (SOMA Architecture), closed / opened louvers (rendering @isochrom)

After sunset the analogue visual effect of the moving louvers is intensified by linear LED bars,
which are located at the inner side of the front edge of the louver. When open the LED can
illuminate the next louver depending on the angle at which it is opened. Continuous GFRP flanges
minimize the amount of light (artificial and natural) that enters when in the closed position.

2.2 Movement
The louvers are moved by actuators located on both the upper and lower edge, inducing
compression forces to create the complex elastic deformation. They reduce the distance between the
two bearings and in this way induce a bending which results in a side rotation of the louver.

Figure 3. Movement concept by KnippersHelbig

The actuator of the louvers is a screw spindle driven by a servomotor. A computer controlled bus-
system allows the synchronisation of the actuators. Each louver can be addressed individually
within a specific logic of movement to show different choreographies and operation modes. The
operation system requires feedback information from the electric servo motors regarding the actual
position of the louvers and can be linked to internet. This allows uploading movement scenarios and
information about façade condition to be sent to a maintenance company. Upper and lower motors
often work with opposite power requirements (driving – braking). Therefore generated energy can
be fed back into the local system to save energy.

Figure 4. Screw spindle used as actuator

The opening procedure of a 13m louver requires a movement of 450mm for an opening angle of
60°. The maximum movement speed of one actuator is about 3,80m/minute. Short louvers open
less. A detailed geometrical analysis of the movement has been carried out in order to determine
potential collisions. Those have been solved by minor adjustments of the movement paths and
bearing locations [5].

3. Structural Design of GFRP Louvers

3.1 Design Verification Concept
Design rules for glassfibre-reinforced polymers, which exist in international and national guidelines
are far from being complete and consistent. Therefore the safety concept applied consists of an ´
action part´ where all relevant load combinations and accompanying safety factors have been
considered according to Eurocode 0 (EN 1990) [6].
Reduction factors applied on the ´resistance part´ have been taken into account according to
different German guidelines, e.g. BÜV-recommendation [7]. The short term material properties
determined for laminate build-ups as specified within DIN 18820 [8] are reduced in order to
consider long-term behaviour, influence of media and temperature as well as a dynamic load
history.

f
f 5
γ A t A A A A

where: fk = characteristic resistance t0

M = partial safety factor
A1 (t) = reduction factor – loading duration = 1 + η [A(t=20a) -1]
A2 = reduction factor – influence of media
A3 = reduction factor – temperature influences
Adyn = reduction factor – opening procedure (dynamic amplification)
Afat = reduction factor – opening procedure (fatigue)
In terms of long-term behaviour, it has been assumed, that the louvers are subjected to max. stresses
caused by opening process 1/3 of their lifetime, which is 20 years. Reduction factor A2 for media
influences has been determined under consideration of outside conditions (water combined with
temperature influences). Max. design temperature is 70°C (45°C ambient + 25°C ext. surface).
Based on the duration of one opening/ closing process (about 30 seconds) and the natural
frequencies of the louvers the dynamic amplification factor Adyn has been considered to be equal to
1,0 (cp. Figure 5-left). The reduction factor Afat covers fatigue effects caused by cyclic long-term
loading. Since the allowable load for 106 load cycles and a frequency of f1 = 0,03Hz for each load
cycle is about 45% of the max. load, AFAT during opening process is similar to 2,0 (cp. Figure 5-
right).

Figure 5. Determination of reduction factors (dynamic amplification – left [5] / fatigue – right [9])

Overall safety factor consisting of partial safety factor and reduction factors is between 1,9 for short
term loading (wind loads) and 9,6 for stress in transverse fibre direction during opening process.

3.2 Load Assumptions / Wind Tunnel Tests

The facade had to be designed for very high wind speeds of up to 35m/s occurring on the South
Korean coast under Typhoon conditions. The wind loads were derived from wind tunnel tests in
order to achieve a safe and economic design of the louvers. Besides this realistic assessment of the
local peak wind pressures acting on the louvers, the aerodynamic stability had to be determined,
since slender or large spanned structures with a low stiffness tend to wind induced instabilities [10].
The operational concept based on the aforementioned studies basically identifies two different
situations: As long as the wind is less than 12m/s (6 Beaufort) the façade can operate. If the wind is
stronger than 12m/s the façade will automatically close and pretension will be applied by pre-
stressing the weak edge accordingly to increasing wind loads.

3.3 Numerical Models / Calculation Results

The structural analysis is based on the geometry of the longest (13,6m) and shortest (6,0m) louver.
Assuming that the values of displacements, natural frequencies, reaction forces and stresses are
critical values, these louvers have been chosen for exemplary investigations.
In order to obtain reliable results numerical calculations have been carried out using two different
Finite Element Method (FEM) programmes SOFISTIK 23 by Sofistik AG, Germany and MSC
Nastran by MSC Software, USA (cp. Figure 6).
Through use of composite layered shell elements strains and stresses of the laminate and each ply
can be calculated with MSC Nastran. Laminate failure index (cp. Figure 6) is representing results of
a max. stress criterion. This stress based criterion is a decoupled failure criterion, assuming a
material failure occurs if one of the stress components within the local direction exceeds the
relevant design stress resistance.
 Type 1
(tapered)

Type 2

Type 1

Type 2 1.0 0.3

closed opened Laminate types Laminate Failure Index (max. opening)

SOFISTIK MSC Nastran
Figure 6. Numerical models and results for louver with max. length

, , ,
max ; ; 5
, , ,

where: x,Ed = design value of normal stress in local x-direction

y,Ed = design value of normal stress in local y-direction
12,Ed = design value of in-plane shear-stress

For the analysis three major critical positions have been investigated:

Position 1 “Closed”: louvers are completely closed and pre-tensioned.

Position 2 “almost closed”: louvers are still in closed position but not pre-tensioned
Position 3 “max. opening”: louvers have been opened to their maximum opening angle, actuator is
max. extracted

In addition exceptional load cases due to malfunction of the technical elements within the façade
system have been carried out.
3.4 Dimensions / Connection Details
The 13-metre-long louvers consist of an orthotropic laminate (Type 1) with a thickness of about
9mm, they are stiffened at both longitudinal edges with a 200mm rib and a 30mm rib (Type 2)
respectively. A hard rubber buffer bar protects the GFRP laminate, when the façade is closed and
one louver rests on the other.

Type 2

rubber
bar
Type 2

Type 1

Figure 7. Section - Dimensions of largest (13m long) louver in closed state [5]

Different build-ups for orthotropic und UD laminates as specified within DIN 18820 have been
used within schematic design phase (cp. Table 1). Further optimization of structural behaviour and
maximum opening angle of louvers was a possible option within construction document phase.

Table 1: Laminat layout for Type 1 (t=9mm – 45°/-45°) and Type 2 (t=30mm – 0°) (4)
TYPE 1 TYPE 2
(ORTHOTROPIC) (UD LAMINATE)
fibermass-content [%] 40 45
thickness t [mm] 8,96 29,84
number of plies [-] 11 55
orientation [°]1 45°/-45° 0°/90°
setup
chopped mat [g/m²]2 450 450
woven fabric [g/m²]2 580
proportion [x –y] 2,2 : 2,3 5 times
chopped mat [g/m²]2 450
UD-ply [g/m²]2 480 26
chopped mat [g/m²]2 300 times
UD-ply [g/m²]2 480
chopped mat [g/m²]2 450
1) related to local x-y coordinate-system 2) fiber-mass related to area

The connection detail between GFRP louver and actuator is realized by means of an inlay,
fabricated out of steel sheets and steel sleeves, laminated in tapered parts of the louver (cp. Figure
8).
 elevation section

Figure 8. Detail ´top connection of louver to actuator´ [5]

4. Performance Mock-Up Tests

Overall structural behaviour of the smallest and largest louvers respectively has been tested within a
full scale performance mock-up test.
In order to consider the stiffness and determine the influence on the overall load deflection
behaviour, the louvers have been attached to a steel sub-structure consisting of steel hollow sections
(cp. Figure 9).

Figure 9. Louvers during performance Mock-Up Tests

Within a time period of 35 consecutive days, the louvers have been operated successfully.
 5. Conclusions
During the planning phase different possible technical solutions for the kinetic media facade have
been investigated and their aesthetical and architectural implications have been considered. Based
on this process an optimal solution, supporting the initial idea of the design - a continuous building
skin that changes shape in smooth movements - has been found. Furthermore the biomimetic
approach meets the client’s wish to make the Thematic Pavilion a showcase of a future architecture
that learns from nature.
Structural design verification of GFRP louvers is based on a combined safety concept considering
various safety factors. For the analysis different critical positions, including exceptional load cases
due to malfunction of the technical elements have been investigated. Reliability of results has been
double-checked by means of a second FEM-software.
Overall load deflection behavior has been tested successfully within full scale performance mock-
up.

6. References

[1] Knippers, J., Speck Th.: “Design and Construction Principles in Nature and Architecture”,
Journal ‘Bioinspiration & Biomimetics’. in press.
[2] Lienhard, J. et al, “Form-finding of Nature Inspires Kinematics for Pliable Structures”,
Proceedings of the IASS Symposium: Spatial Structures Temporary and Permanent 2010,
Shanghai, CN.
[3] Schleicher, S. et al. “Bio-inspired kinematics of adaptive shading systems for free form
facades” Proceedings of the IABSE-IASS Symposium: Taller Longer Lighter, 2011,
London, UK.
[4] Schinegger, K., Rutzinger, S., Knippers, J., Scheible, F., “Biomimetic Media Façade”,
Journal or Conference Title, Vol. 12, No. 1, Month 2011, pp. 437-445.
[5] Scheible, F., Oppe, M., “Thematic Pavilion of EXPO 2012 YEOSU – Bionic Façade”,
Structural Narrative, June 2010, pp. 1-105
[6] EN 1990, “Basis of Structural Design”, CEN European Committee for Standardization,
Brussels, Belgium, 2001
[7] BÜV-Empfehlungen “Tragende Kunststoffbauteile im Bauwesen [TKB]“ – Entwurf,
Bemessung und Konstruktion, August 2010
[8] DIN 18820: Laminate aus textilfaserverstärkten ungesättigten Polyester und
Phenacrylatharzen für tragende Bauteile (GF-UP, GFPHA), Part 1-4, Beuth-Verlag, Berlin;
March 1991
[9] Meyer, H.-J., „Zur Bemessung von GFK-Bauteilen unter Zuhilfenahme der Linear-
Elastischen Bruchmechanik und probabilistischer Versagenskriterien“, Phd-thesis TU
Hamburg-Harburg, December 1992
[10] Wacker, J., “Investigation of aerodynamic stability”, Wind-tunnel test Report, June 2010

View publication stats