Journal of Civil Engineering and Architecture 14 (2020) 10-19 doi: 10.17265/1934-7359/2020.01.002

UHPFRC Folded Pavilion

Raphael Fabbri1,2, Catalina Francu3, Beatrice Gheno3, Mattia Federico Leone4 and Jenine Principe4

1. Maitre de Conference des Ecoles d’Architecture, ENSA Paris-Belleville, 60 boulevard de la Villette, Paris 75019, France

2. ATELIER MASSE, 14 rue des Jeuneurs, Paris 75002, France

3. ENSA Paris-Belleville, 60 boulevard de la Villette, Paris 75019, France

4. Department of Architecture, University of Naples Federico II, Via Toledo 402, Naples, NA 80132, Italy

Abstract: The aim of the Student Workshop “Material Optimization and Geometric Exploration” (ENSA Paris-Belleville and University of Naples Federico II) is to discover the possibilities offered by new materials, starting from their characteristics. The final goal is to build a synthetic pavilion, which—in the last session—demonstrates ultra-high performances fibre reinforced concrete (UHPFRC) capacities. Designing with UHPFRC requires thinking simultaneously about the geometry, the static, the casting (mainly precast) and the implementation process. The design of the pavilion starts with a widespread geometric exploration using a phylogenetic tree. This approach has the advantage of exploring different designs at the same time without enclosing the creative process in one path. The geometry of the final pavilion is based on a folded surface, called “Yoshimura”, made out of rows of triangles. The profile of the pavilion is bent in order to create a double curvature and so, more stability. The modules are multiplied asymmetrically to minimize the number of the moulds, having at the end just one mould for each row of triangles. The moulds are made with polyethylene terephthalate glycol (PETG) laser-cut sheets which have been folded afterwards. This process has been chosen for both the smooth finishing it delivers and the simplicity of the fabrication process.

Key words: UHPFRC, folded mould, phylogenetic tree, parametric design, design pedagogy, structuralism.

1. Introduction

1.1 “Techno-Evolution” and “Bio-Evolution”

How does technical evolution happen? Although it

uses different ways, “techno-evolution” has been often

compared to “bio-evolution” by historians and

prehistorians as André Leroi-Gourhan, or by

philosophers of techniques as Gilbert Simondon:

“If we propose to juxtapose Invention and Mutation,

Tradition and Transmission of acquired traits, it is not

to take sides, by extension from technological values

to biological values; the complexity of biological

problems is familiar enough so that we can observe

the utmost caution. Biology is going through its

puberty crisis and Technology is hardly vague, but it

is to be expected that in the future the proximity of the

two disciplines will become increasingly clear and

Corresponding author: Raphael Fabbri, architect, maitre de

conference, research fields: UHPFRC, advanced geometry, parametric design, pedagogy.

that, through the confrontation of the two series of

creations of Nature and the creations of Human

Industry, we will reach a deeper understanding of the

general phenomena of Evolution.” [1].

But both have noted the necessity to “objectify”

(“Objectiver” in French) the process of technical

evolution and not to reduce it to a purely biological

one:

“The classification is made according to specific

modalities that distinguish the genesis of the technical

object from those of other types of objects: aesthetic

object, living being. These specific modalities of

genesis must be distinguished from a static specificity

that could be established after genesis by considering

the characteristics of the various types of objects: the

purpose of using the genetic method is precisely to

avoid the use of a classifying thought occurring after

genesis to divide all objects into genera and species

appropriate to speech.” [2].

It is important to note that “techno-evolution” does

D DAVID PUBLISHING

UHPFRC Folded Pavilion

11

not take place through DNA transmission but through

“structural re-organisations” [2] created by humans.

These “designers” are all “located” in the social world

of the human production. Then the selection process is

dependent on social factors.

In both cases, the evolution can be separated

conceptually in two phases: the “structural

re-organisation” and the “selection process”. In

biology, the “structural re-organisation” takes place

through the combination of alleles and through

mutations, while the “selection process” is based on

the organism’s capacity to survive and to transmit its

own genetic heritage. In the field of technical science,

the “structural re-organisation” (in its broad meaning)

is conceived by designers as a strategy to deal with

problems, while the “selection process” is a

consequence of the convergence process, which is a

progressive assimilation of all the project

requirements. It is called “Concrétisation” by

Simondon:

“Like in a phylogenetic lineage, a defined stage of

evolution contains within it dynamic structures and

patterns which are at the heart of the evolution of

forms. The technical being (l’être technique) evolves

through convergence and adaptation to oneself; it

unifies inwards according to a principle of internal

resonance.” [2].

The resulting configuration is called “technical

object”, conceptually defined by its “structural

organisation”, or using the words of the American

architect Louis I. Kahn, by its “Form” (in contrast to

its Design. “All the spoons have the same Form, but

each spoon has a different Design” [3]).

1.2 Exploration in Building and Architecture Field

In the field of building and architecture, the phrase

“technical object”, widely used in the industrial sector,

is usually replaced with “typology” or “construction

system”. When a designer (an architect, an engineer or

a craftsman) chooses a typology or a construction

system for a project, they usually work by induction

and analogy: they study something similar that has

already been done. Making a choice through analysis

and deduction, which means starting with all the

requirements of the project and ending with the

technical solution, is very unusual. This fact is due to

the time-efficiency of the “Concrétisation” process.

We can observe it in the industrial sector too: for

example, when a technician designs a motorized

vehicle, he starts from the “structural organisation” of

the existing cars or trucks, instead of defining all the

requirements and then trying to find a way to fulfil

them.

Although the induction method saves time, it has

two limitations:

Are the existing typologies the best technical

solutions available?

How to proceed with a new material or with a

new problem?

We already talk about the first limitation in a

precedent research [4]. The second one is the topic of

this research, focussing on the ultra-high performances

fibre reinforced concrete (UHPFRC). As academics

and designers, we can hardly act on the “selection

process” of “Concrétisation” (resulting from complex

social forces), but we can work on the “structural

re-organization” by questioning old configurations and

proposing new ones.

1.3 The Workshop “Material Optimisation and

Geometric Exploration”

The UHPFRC is a new material whose

characteristics allow new geometries [4]: smooth

shapes due to the absence of passive reinforcements,

precise finishing due to the ultra-thin aggregates, and

flat cross-sections due to the high strength. Designing

with UHPFRC requires the designer to think

simultaneously about the geometry, the static, the

casting process (mainly precast) and the

implementation. The aim of the workshop “Material

Optimisation and Geometric Exploration” (ENSA

Paris-Belleville & University of Naples Federico II) is

UHPFRC Folded Pavilion

12

to discover the possibilities offered by new materials,

through exploring their potential [5]. During the last

session of the workshop we focussed on the UHPFRC.

After presenting the material, the work starts with a

widespread exploration, using phylogenetic trees. The

work continued by specifying the design according to

the “rules” of the material. The final goal is to build a

synthetic pavilion, which demonstrates UHPFRC

capabilities. This approach has the advantage of

exploring at the same time different designs,

questioning the “structural organisation” without

narrowing the creative process in one path.

2. Design Phase: Widespread Exploration

2.1 Organisation

The workshop is divided in two phases: the design

phase, hosted in Naples, and the building phase, in

Paris.

The design phase starts with a presentation of the

material UHPFRC: its history, its mechanical

behaviour, the casting process and the fabrication of

mould. The students are then split in 5 bi-national

teams, each of whom focused on one kind of surface:

ruled surface;

revolution surface;

reciprocal frame;

folded surface;

sweeping surface.

To promote a widespread exploration and to avoid

the connatural fascination for a specific design, each

group has to propose 24 small designs, according to a

phylogenetic diagram.

2.2 The Phylogenetic Diagram

The phylogenetic tree is firstly a tool of

classification, used in biology to show the

evolutionary relationships among various species or

other entities. Each branch split-up is based upon

similarities and differences in the physical or genetic

characteristics of the species analysed—the closer the

split between two organisms, the more related they are;

the further the split is, the less related they are. All life

on the earth is part of a single rooted phylogenetic tree,

indicating common ancestry (the tree of life). The

linguistic field defines rooted phylogenetic trees to

show the common origins of the languages (as the

well-known phylogenetic tree of Indo-European

language). Although the scientific validity of this last

diagram is largely discussed, the linguists consider it a

tool for understanding the evolution of languages.

Unrooted phylogenetic trees (without common

ancestry) are today used in the fields of logic,

computer science and taxonomy. Each branch split-up

represents a level of classification. In Architecture,

Foreign Office Architects (FOA, founded by Farshid

Moussavi and Alejandro Zaera Polo), frequently uses

phylogenetic trees to classify their own projects [6]

and to present the different structural organisation of

historical buildings [7]. In their opinion, the

phylogenetic tree is not only an ex post representation

graph, but also an ex ante generative tool. A good use

of phylogenetic trees as a generative tool was made by

the architect George L. Legendre. In his book [8],

Legendre uses a phylogenetic tree of different kinds of

pasta to determine their parametric equations. The

phylogenetic tree of pasta has four levels of

embranchments: (1) the longitudinal profile (twisted,

helicoidal, pinched, bunched, straight, sheared or

bent); (2) the cross-section (solid, hollow or

semi-open); (3) the surface (smooth or striated) and (4)

the edges (smooth or crenelated). The first level gives

the general typology of the parametric equation, and

the other three levels add more complexity to the

equation. The parametric equations might seem too

complex to be quickly understood at a first glance but

on a closer look, we understand that they are built by

the different levels of the phylogenetic tree.

From a tool of classification, the unrooted

phylogenetic trees are becoming a useful exploration

tool, able to show all the combinatorial possibilities.

In this research, the phylogenetic tree is used as a tool

for the exploration of structural re-organisations.

UHPFRC Folded Pavilion

13

Using the phylogenetic tree (see Figs. 1 and 2), each

student group (see above) briefly designs 24 pavilion

proposals, one for each final branch of the tree. The

bifurcations of the phylogenetic tree are organized in

three levels: at first the “composition” (through the

“shape generation” process the pavilion can be

designed as a unique surface or as an addition of

single elements), then the “static scheme” (arch, side

wall, dome and funicular) and finally the “membrane”

(continuous, openwork or gridshell/wireframe

membrane).

Afterwards, each group decides to develop

specifically one of the final branches. As we do not

have the space to comment all 120 (24 × 5 groups),

only the five pavilions that were developed are placed

at the end of the phylogenetic tree (see Fig. 2). For

each of them, the morphology, the static scheme, the

assemblies and the implementation have been defined.

3. Building Phase: Synthetic Pavilion

3.1 Morphogenesis

The final pavilion incorporates principles from

every group and puts the UHPFRC’s capacities to the

test. The first group, “Rositalian Spiral”, offered the

form of two spirals entwined in plane—a kind of

“rolling in itself” effect—cited in the synthetic

pavilion’s tendency to turn inwards. The “FRA NA”

group’s design is responsible for the folding pattern as

well as for the casting process. “Pentathing” inspired

the polygonal division of the initial surface as well as

the manner of component assembly. The

function—seating with sunshade incorporated—was

added into the final design from the concept of the

group “Micro Cosmos” while the group “Spritz”

offered the idea of using adaptive moulds, as well as

the courtyard implementation. Fig. 2 showcases each

of the five designs and their relationship on the

branches of the phylogenetic tree for a better

understanding of the final pavilion as well as out of a

need for transparency in terms of the design process.

The geometry of the final pavilion is based on a

folded surface, called “Yoshimura folded” or

“Whirlpool”, made with rows of triangles. The profile

of the pavilion is bent in order to give it a double

curvature and so, more stability. The modules are

multiplied in axisymmetric rotation to minimize the

number of the moulds, having at the end just one

mould for each row of triangles (see Fig. 3).

The vertices of the triangles are cut to avoid sharp

angles. The centre of the element is open, for both

light and weight reasons.

Figs. 4 and 5 present the main steps of the

morphogenesis:

(1) The 10 basic triangles generated by the bending

of the Yoshimura folding pattern;

(2) The cut of vertices of the triangles;

(3) The contact surfaces between blocks;

(4) The internal void of the blocks;

Fig. 1 The three levels of bifurcations in the phylogenetic tree.

UHPFRC Folded Pavilion

14

Fig. 2 The whole phylogenetic tree with the five projects explored by each team.

UHPFRC Folded Pavilion

15

Fig. 3 Repetitive triangles in the “basic” Yoshimura folded (left) and in the “bent” one (right).

Fig. 4 Morphogenesis: (1) basic triangles, (2) cut of vertices, (3) contact surfaces between blocks, (4) internal voids.

Fig. 5 Morphogenesis: (5) holes for the bolts, (6) axisymmetric rotation, (7) mirror and trimming.

(5) Holes for the bolts;

(6) Axisymmetric rotation;

(7) Mirror and trimming.

3.2 Casting Process

The moulds are made out of laser-cut polyethylene

terephthalate glycol (PETG) sheets which have been

folded into the necessary shape (Figs. 6 and 7). This

process has been chosen for both the finishing and the

simplicity of fabrication. The PETG sheets are easy to

fold and clean, but they have weak bending stiffness,

so they had to be reinforced by wood boards cut to

size. The internal void in the overall shape is made

with the help of a polystyrene block. The structural

UHPFRC Folded Pavilion

16

strength of the pavilion lies in the contact surface

between elements, necessary to sustain load forces. To

ensure the stability of the aforementioned connection,

we used (8Φ × 130 mm) bolts that traversed both

pieces and were fixed into place by nuts (Fig. 8). The

mould required us to protect the mobility of the bolt

while still casting its void in the piece; therefore, we

chose to use a Vinyl/PVC flexible tube (as long as the

thickness of the element) to protect it.

The UHPFRC used for the casting is DUCTAL

NAW3 from LafargeHolcim. The mould was

externally watertight, but sometimes concrete flowed

between polystyrene and PETG. To solve this issue,

we added a compressed foam-joint in-between. The

mould has been slightly improved during the 5 days of

casting, but the general concept has been maintained for

Fig. 6 Pictures of the moulds.

Fig. 7 First cast elements.

Fig. 8 Pictures of the cast elements and the assembly.

UHPFRC Folded Pavilion

17

its efficiency.

3.3 Implementation

The pavilion was implemented in April 2019 during

the second phase of the workshop with the help of the

students. After the realization of the wood base, the

construction started from the pre-assembly of the

triangles which would have been anchored to the

ground. They were fixed together using two bolts. A

cork sheet was placed in between the connecting

surfaces, which were folded to block the shear and to

guarantee good contact. The actual implementation

had to proceed symmetrically, in order to conclude the

building process with the keystone (also to exploit the

scenic effect). After setting the first pieces of the

principal arch and the double triangles immediately

near them, the construction started to become more

complex. Despite it, it was possible to build it without

any temporary support elements, besides some

movable wood trusses. Immediately before putting the

keystone, the smaller pieces without any structural

function were rapidly installed. After removing the

temporary supports, the pavilion had a small

adjustment, followed by an uneasy but harmless

creaking. Without considering the wooden base

(whose designing and building process took two days),

the construction of the pavilion, pre-assembly

included, lasted for 12 hours with a maximum of 8

people who worked on it.

4. Final Pavilion and Future Works

The pavilion was built in the interior garden of

Ecole Nationale Supérieure d’Architecture de

Paris-Belleville (Figs. 9 and 10). The experience made

us think further, to what future research might

uncover, and a few ideas that seemed important are

building methods that do not involve

shoring—through a design in which the gravity centre

is always kept inside the cupola-shaped structure’s

projection on the ground. Nowadays, repetitive

geometries are used for reducing the moulds’ cost

in precast building [9] so, an important issue to further

Fig. 9 Render of the final pavilion (left) and picture in May 2019 (right).

Fig. 10 Pictures of the final pavilion in May 2019.

UHPFRC Folded Pavilion

18

explore could be the optimisation of adaptive PETG

folded moulds to meet the requirements of both

geometry and cost optimisation. In other words, the

aim is to achieve the optimal balance between variety

and practicality in the building process, because the

first one makes way for innovation while the latter

makes it possible.

UHPFRC has a huge shaping capacity, as its liquid

form and the absence of passive reinforcements permit

a large degree of imagination. One of its biggest

advantages is its high load-bearing capacity, even

when cast in thin dimensions. This, together with its

relatively light weight, makes it much easier to shape

than stone, both in its liquid and precast form; even if

we must remember that the issue of contact surface has

to be further developed. This need emerges from the

necessity to fix the degrees of liberty in order to

better control the global shape. The precision and the

strength of the connections are crucial for this kind of

pavilion.

The pedagogical methods and the research carried

out during the workshop “Material Optimization and

Geometric Exploration” are both experimental and

structuralist. Since the 2000s, experimental

pedagogical methods are spreading in architecture and

engineering universities, not only because of the

development of digital fabrication, but also because

they provide an important complement to other kinds

of teaching. The students are involved in collaborative

research leading to a final collective production.

These new teaching processes must be completed by

theoretical background and analytical feedbacks,

producing real experimentation and not only a simple

experience. Although structuralism is not as studied as

during the 60s and the 80s, it is an effective method of

exploration and creation. By focusing on the

structures, the student develops a creativity that

touches the essence of the technical objects and opens

up to possibilities for evolution.

Last but not least, we would like to touch upon the

new directions UHPFRC might open in the realm of

building and architecture. It is important to note that

with each new discovery we are bound to develop a

way of using it without falling into the mannerism of

rebuilding canonical shapes with materials that offer a

much wider spectrum of possibilities. We are firmly

advising against building with concrete following the

logic of the brick: we suggest that architects, as well

as engineers, would have a lot to gain from

approaching this new material through a critical and

curious lens that will most certainly unleash much

more richness than if they stick to the known. Today

we find ourselves in the same position of Louis I.

Kahn when he asked the brick what it wanted to be,

more than half a century ago: “You say to a brick,

‘What do you want, brick?’ And brick says to you, ‘I

like an arch’.” We have to understand this material

and know its answer to our question.

Acknowledgements

We would like to express our gratitude to Ductal

Team, to the ENSA Paris-Belleville, to the

Department of Architecture of the University of

Naples Federico II and to the French Ministry of

Culture for their support.

Special thanks to all the students and all the

teachers of this workshop:

Nunzia Ambrosino, Ata Gun Aksu, Iris Andreadis,

Maher Ben Hamed, Enrico Corvi, Thibaud de Zuttere,

Davide Ercolano, Olmo Galletti, Martin Gatto, Sarah

Husein, Elioth Jeantet, Daniele Lancia, Simona

Makoski, Simone Muscio, Aurora Naddeo, Giovanni

Nocerino, Bianca Pagano, Arianna Palumbo, Simone

Piccolo, Manuele Puopolo, Martina Pizzicato, Sergio

Pone, Alexandra Runcan, Edoardo Schettino,

Mariagrazia Serafino, Camille Vouin, William

Wattequant.

References

[1] Leroi-Gourhan, A. 1973. “Milieu et techniques.” In Évolution et techniques. Paris: Albin Michel.

[2] Simondon, G. 1989. Du mode d’existence des objets techniques. Paris: Aubier.

UHPFRC Folded Pavilion

19

[3] Kahn, L. I. 1996. “Réflexions and Vouloir être.” In Silence et Lumière (Silence and Light). Choix des conférences et d'entretiens. Paris: Editions du Linteau.

[4] Fabbri, R., and Corvez, D. 2013. “Rationalisation of Complex Shapes Facades.” In UHPFRC 2013—Proceedings of the RILEM-FIB-AFGC International Symposium on Ultra-High Performances Fibre-Reinforced Concrete.

[5] Fabbri, R., Principe, J., and Leone, M. 2017. “Exploring UHPFRC Possibilities.” In UHPFRC 2017—Proceedings of the RILEM-FIB-AFGC International Symposium on Ultra-High Performances Fibre-Reinforced Concrete.

[6] FOA, Kwinter, S., De Landa, M., Wigley, M., Moussavi, F., Cache, B., Mertins, D., Knapp, S., and Zaera-Polo, A.

2005. Phylogenesis Foa’s Ark: Foreign Office Architects. Barcelone: ACTAR.

[7] Moussavi, F. 2009. Function of Form. New-York: ACTAR and Harvard University Graduate School of Design.

[8] Legendre, G. L. 2011. Pasta by Design. London: Thames & Hudson.

[9] Aubry, S., Bompas, P., Vaudeville, B., Corvez, D., Lagrange, T., Mazzacane, P., and Brizou, A. 2013. “A UHPFRC cladding challenge: The Foundation Louis Vuitton Pour la creation, Iceberg.” In UHPFRC 2013—Proceedings of the RILEM-fib-AFGC International Symposium on Ultra-High Performances Fibre-Reinforced Concrete.