PROGRAMMABLE FOLDING:

COMPUTATIONAL GEOMETRY WITH EMBEDDED ASSEMBLY LOGIC

Annie Locke Scherer

Integrative Technologies M.Sc. Candidate
M. Arch University of Michigan
B.S. Arch University of Virginia
 02

ITECH M.Sc. Programme

ICD
Institute for Computational Design

ITKE
Institute for Building Structures and Structural Design

Thesis Advisors
Prof. AA Dipl.(Hons.) Arch. Achim Menges
Prof. Dr.-Ing. Jan Knippers
Dipl.-Ing. Oliver David Krieg
M.Arch, B.Arch.Sci David Correa
Dipl.-Ing. Benjamin Felbrich

2014-2015
 03

To mom, Danny, Grandad and Pat:

Thank you for your unending love and

support, that made this all possible.
04
 05

CONTEXT
INTRODUCTION.........................................08

AIM............................................................15

CONTEXT...................................................19

STATE OF THE ART.....................................25

METHODS..................................................33

DESIGN RESEARCH DEVELOPMENT........41

RESEARCH PROPOSAL.............................85

DISCUSSION..............................................91

OUTLOOK..................................................95
06
 07

INTRODUCTION
CHAPTER 01
 08

INTRODUCTION

One does not normally design leftover waste material. More

commonly, one designs and what is left over is discarded.
Programmable Folding is centered around strategic cuts.
Careful design of removed material has the ability to imbue a
flat sheet with intrinsic curvature.

The aim of this thesis is to create parametrically-derived folding

patterns that approximate irregular, doubly-curved surfaces
when assembled. Looking at fundamentals of cut patterns, this
thesis investigates the relatively unexplored world of origami’s
lesser-known cousin: kirigami. This, like origami, is the art of
folding sheets but also utilizes simple cuts to program inherent
curvature into a flat material.

Programmable Folding analyzes the basic geometrical rules of

kirigami folding, identifying which parameters are flexible and
which ones must be followed. The resulting geometrical pattern
encodes a sheet material with an inherent assembly logic and
provides the necessary information to fold the material into a
complex, 3-dimensional surface.

This project builds upon research from much of the origami

world, particularly referencing the work of Tomohiro Tachi
(University of Tokyo), Daniel Piker (Foster + Partners), and Toen
Castle (University of Pennsylvania). Grasshopper and Kangaroo
for Rhino are the primary design tools to realize the geometrical
complexities within double-curvature folding. These tools allow
for easy manipulation of areas and degrees of curvature, quick
generation of a cut and tabbing pattern, and simulation of the
final folded geometry.

Programmable Folding explores design possibilities of geometry

on multiple levels and delivers a final product whose design is
embedded in the generated folding pattern. This thesis simplifies
complex geometrical problems of approximating an irregular
surface, while simultaneously investigating typology, assembly
logic and structural performance.
“All designers fold.

That is, all designers crease, pleat, bend, curve or wrap two-
dimensional sheets of material, and by these processes of
folding, create three-dimensional objects... Since almost all
objects are made from sheet materials (such as fabric, plastic,
sheet metal or cardboard), or are fabricated from components
used to make sheet forms (such as bricks - a brick wall is a sheet
form), folding can be considered one of the most common of
all design techniques.”

-Paul Jackson, origamist

ORIGAMI

When most people think of origami, the first image that comes
to mind is that of a folded paper crane or a child’s “cootie-
catcher.” While these are the most simple and common uses for
folding paper, the same fundamental geometrical concepts
can be applied in a much more complex manner. With careful
programming, it is possible to achieve irregular surfaces and
forms that could be applied in an architectural context.

While a great deal of intricacy can be achieved in origami

surfaces, Programmable Folding examines the inherent
geometric complexities of folding and demonstrates that
origami does not allow for easy asymmetrical manipulation
of patterns. While some patterns such as Figure 01.01 have
the freedom to bend and form around sub-structures, it is not
possible to create a rigid surface without additional support.
Figure 01.01 Work by Thomas Diewald, applying Ron Resch’s classic
origami pattern to create architectural surfaces.
 12

KIRIGAMI

Deriving inspiration from the University of Pennsylvania

Astronomy & Physics Department, this research builds on the
question of how to create three-dimensional forms from flat
sheet materials. Try to wrap a sphere, and soon the geometrical
problems become increasingly complicated. No amount
of folding can change a flat sheet of paper to have more
“intrinsic” curvature.

Kirigami, like origami, is the art of folding paper, but with

strategic cuts. This small addition has allowed artists, scientists,
and designers to create complex curvature and shapes that
have never before been possible with a single sheet. Kirigami
is an elegant solution to imbue surfaces with more or less
curvature based on simple geometric principles.

As seen in Figure 01.02, by removing material from a sheet and

folding it, it is possible to fold a flat sheet and achieve zero,
positive, or negative Gaussian curvature. These fundamentals
will later be extrapolated to create more complex curvature
and surface approximations.
 13

Figure 01.02 Simple kirigami

切り紙
Kirigami:
“kiru” = to cut,
“kami” = paper
14
 15

AIM
CHAPTER 02
 16

AIM

Architecture mainly consists of planar elements, with sheet

materials being the most common base of building design.
In the past, architecture has kept a clear division between
architectural skin and structure. The substructure is independent
from the cladding, and the two do not necessarily have a
relationship.

The aim of this thesis is to bridge the gap between skin and
structure. It investigates the potentials of a light-weight,
flat-pack, rigid panel system that has structural and spatial
capabilities. This type of system would have a differentiated
pattern with an embedded assembly technology.

This thesis investigates the fundamentals of kirigami patterns,

seeing what parameters are flexible and which rules must not
be broken. Connections between folds are programmed within
the plastic sheets, so no added materials are needed for the
joints. By programming the underlying geometry and intrinsic
curvature, the resulting kirigami patterns are used to investigate
potential design systems that apply this logic to architectural
design.
 17

Figure 2.01 Kirigami fold pattern for

monkey saddle geometry
18
 19

CONTEXT
CHAPTER 03
 20

ORIGAMI FOLDING
In order to create a bespoke folding system, it is first necessary
to understand classic origami folds and their architectural
capabilities. Ron Resch’s patterns give a high degree of
variability as well as potential structural capacity, while folds like
those of Yoshimura and Miura Ori have the ability to flat pack in
multiple configurations.

Although many of these folding patterns have interesting

architectural implications formally, most of the patterns are
impossible to manipulate asymmetrically for they must retain
symmetry in order for the global system to function.

Miura Ori Paul Jackson

Ben Parker Hans Buri

Yoshimura Ron Resch

 21

John Mckeever Paul Jackson

Ron Resch Ben Parker

Folds Flat
Variability
Deployability
Rigidity

Miura Ori

Figure 03.01 Common origami folding patterns

 22

KIRIGAMI
With the simple addition of a few strategic cuts, kirigami allows
the generation of intricate fold patterns with substructure bases
of triangles, rectangles, or even hexagons. While it is possible to
create objects or flat patterns, Programmable Folding focuses
on the possibilities of approximating complex curvature. The
process of manipulating kirigami geometry patterns will be
explored further in design research development.

objects

triangle base

quad base

hex base
 23

3-D surface approximation

Figure 03.02 A wide array of kirigami variations, as explored by Mike Tanis

24
 25

STATE OF THE ART

CHAPTER 04
 26

FOLDING IN ARCHITECTURE
Many architectural works have been inspired by origami,
whether in shape or in function. In looking at precedents, The
operability, asymmetry, and potential reconfigurability of the
systems are related to one another in Figure 04.01. Note that
most are symmetrical, becasue of the difficulty of generating
variable patterns.

Manual

Canary Warf Kiosk | Make Architecture Evolution Door | Klemens Torggler

Operable
Actuated

Hoberman Arch | Salt Lake City Actuated Rigid Plates | Erik Hull
Static

Bowoos Pavilion Shenzhen International Airport

Figure 04.01 Origami Inspired Precedents

 27

Reconfigurable

Novi Sad | Origami Forum Woodskin | MammaFotogramma

Resonant Chamber | RVTR Ron Resch Origami Appended Space | Mohamad Al Khayer

Enoc Armegnol Folded Chair Cruise Terminal Yokohama Fold Plate Hut | Osaka

Asymmetrical Panels
 28

ASYMMETRIC FOLDING PATTERNS &

CORRUGATION

PARAMETRIC MIURA ORI

Because of the geometrical complexities inherent in origami, it
is extremely difficult to manipulate a symmetrical pattern locally
without making an unfoldable pattern. Here, Tachi looks at
asymmetrically adjusting patterns to create complex curvature.
Figure 04.02 shows a simple manipulation of the classic Miura Ori
fold pattern to create asymmetrical double-curvature. Because θ2
of the Kawasaki-Justin Theorem, the sum of the mountain and θ1
valley folds around a vertex must equal zero to make a flat θ3
folding pattern. This severely limits the geometrical possibilities θ4
for asymmetrically manipulating classic origami folds.

FREE FORM ORIGAMI TESSELLATION

Tomohiro Tach is one of the only origamists to research
complex, asymmetrical forms from single sheets. As illustrated in
Figure 04.02, based on generalizations of Ron Resch’s patterns,
Tachi demonstrates how to fold a sheet of stainless steel into a
freeform origami tessellation. Kawasaki-Justin Theorem

ORIGAMI CORRUGATION θ1 + θ3 = θ2 + θ4
Tessellated Group develops mechanically folded sheet
material and utilizes them as structure for double-layer systems.
This corrugation improves the typical performance while
reducing manufacturing cost and waste. While none of their
designs accommodate asymmetrical patterns, Tessellated
Group’s products can be applied in a multitude of industries,
ranging from architecture to aerospace to construction and
packaging.

Figure 04.02 Pattern for variable curvature

 29

Figure 04.03 Free-form origami tessellation

Figure 04.04 Industrial origami

 30

KIRIGAMI RESEARCH IN PHYSICS

Physicists Randall Kamien and Toen Castle at the University of
Pennsylvania are investigating how to approximate a curved
surface with kirigami. Their work is based on French researchers’
study of how sunflower seeds cover the dome of the sunflower.
The seeds have a precise pattern of 5, 6, or 7 sided; this natural
variation allows for the seeds to pack evenly around the dome.

While origami uses the technique of “tucking” to hide excess

material in the final shape, kirigami addresses these limits and
removes the material completely, allowing folding without
any excess material. Kamien and Castle have simplified these
principles into unique patterns, each with their own spatial and
structural potentials. By introducing a step in the modules, they
can approximate curves much like that of a voxel.

Figure 04.05 Sunflower Lattice of Seeds. Blue, red, and

green cells are pentagons, hexagons and heptagons,
respectively

Figure 04.06 Variation within single 5/7 kirigami module

 31

Figure 04.07 Projected stepping pattern of a ziggurat, along with cut

pattern

Figure 04.08 Left: A duopotent lattice of sixons in the flat state, where
grey hexagons denote excised regions of paper. Right: the two folded
state configurations of the duopotent sheet
32
 33

METHODS
CHAPTER 05
 34

METHODS
COMPUTATIONAL KIRIGAMI
Programmable Folding develops a robust computational model
that uses the principles of Gaussian curvature as a substructure
to kirigami cut patterns in order to approximate curvature.
Strategic, computationally determined cuts imbue the pattern
with intrinsic curvature that will allow the flat pattern to take its
pre-determined shape once assembled.

FABRICATION
Programmable folding utilizes 0.5mm x 800mm x 1200mm
plastic sheets to express the full geometrical possibilities of
computational kirigami. These plastic sheets allow enough
flexibility for incremental folding for assembly. Additionally, the
plastic is structural enough to withstand the plastic deformation
along the dashed fold lines. A flexible material was specifically
chosen as rigid kirigami would require simultaneous actuation
of all folds. Furthermore, a rigid structure would be comprised
of multiple materials within the joints, which requires more
assembly instruction and materials.

The Zünd cut plotter allows Programmable Folding to be realized

at a real architectural scale. Because of the sheet limitations
of the laser cutter, the cut plotter is the best option for taking
advantage of full polypropylene sheets with as little waste as
possible.

FOLDING
Instead of adding additional material to join the kirigami edges,
Programmable Folding takes advantage of the kirigami cut outs
and programs the tabs into the material that would normally be
removed as waste. One set of tabs closes the holes, while the
other reciprocally locks each module into the other to create a
more rigid structure.

Figure 05.01 Zero, positive, and negative Gaussian curvature

 35

0.5mm Polypropylene
1m² Sheets: 800x1200mm

Bed size: 1500mmx3000mm

4.5m²

Zünd Knife Cutter

Universal Cutting Tool: cuts material
up to 5mm

Figure 05.02 Fabrication tools and flexible material

Figure 05.03 Tabbing and reciprocally interlocking modules

 36

GAUSSIAN CURVATURE
Gaussian curvature is the basis for programming material with
intrinsic curvature. It is the product of both principle curvatures
at any given point, which is equal to the intrinsic curvature. By
looking into typology of curvature with tessellation and adding
or subtracting the number of vertices around a node, it is
very simple to manipulate and program a surface’s resulting
Gaussian curvature. If the vertex has 5 instead of 6 vertices, the
corresponding surface has positive Gaussian curvature while
adding a 7th results in a saddle-like negative curvature shape.

This basic theorem is the underlying logic for parametrically

manipulating generic kirigami folding patterns. After
understanding the different typologies that can be created with
simple manipulation of node vertices and developing a tool to
easily do so, it is possible to use this logic as a sub-structure and
generate more controlled, geometrically complex kirigami.

{3;5}

(+) Gaussian curvature

Σ Θ > 360

{3;7}
no Gaussian curvature
Σ Θ = 360

(-) Gaussian curvature

Σ Θ < 360

Figure 05.04 Positive and negative Gaussian curvature by adding and removing vertices
 37

(+) gaussian curvature

no gaussian curvature

(-) gaussian curvature

Figure 05.05 Positive and negative Gaussian curvature applied to kirigami patterns
 38

DIGITAL METHODS
Programmable Folding utilizes a robust computational model
to analyze global formations and irregular surfaces. This model
takes an irregular input surface, approximates its underlying
Gaussian curvature, and creates a flat sub-structure for the
kirigami based on the required intrinsic curvature. After running
the kirigami pattern through a kangaroo script, the necessary
cut outs are determined and a flat-pack folding pattern can
be generated. All folding and cut lines, along with tabbing, are
pre-programmed and the final pattern is ready for production.
 39

COMPUTATIONAL MODEL

base geometry for global curvature program local curvature

OUTPUT

triangulated surface approximation sheet specs required for construction

flat kirigami fold pattern
& percent deviation
40
 41

DESIGN RESEARCH
DEVELOPMENT
CHAPTER 06
 42

SIMPLE KIRIGAMI MODULES

I explored a few simple modules of kirigami and parametrized
the variables. It is important to note that not all variables can
be changed without inserting additional folds. These are some
of the study models I fabricated in an effort to have a better
understanding of the folds, cuts, and relationship between
panels, and how to produce more or less curvature.

w2
w2

w3
w3

h1
h1 w1
w1

h2
h2

Module 1
43
 44

SIMPLE KIRIGAMI MODULES

This series of experiments investigates triangular cuts and “glide”
modules. Here I discovered that not all parameters can be
modified without folding implications. Some of the tests show
gaps in the folded module and begin to reveal the parameters
of kirigami that cannot as easily be manipulated without
consequences.

h3

h3

h2
d1
h2
d1

d2
h1

d2
h1
w1 w2

Module 2 w1 w2
45
 46

SIMPLE KIRIGAMI MODULES

These hexagonal cuts are a variation of Module 1, with two
additional “relief” folds on each side to allow synclastic or
anticlastic bending between modules.

d1 d1
h3 h3

h2 h2

h1 h1

Module 3
47
 48

KIRIGAMI FUNDAMENTALS

After numerous studies with the basic kirigami cuts and folds,
some key principles of kirigami can be abstracted. For example,
a right angle between a cut will produce a vertical panel, while
making it larger or smaller will change the relationship to the
lower panel.

Changing width between mountain and valley folds will

influence a module’s height, while manipulating the width of
the cut will simply change the dimensions between modules.
Finally, changing angle between the mountain and valley folds
to be smaller or larger will result in an downward or upward
curvature (respectively) between modules.

acute slit

right angle slit

obtuse slit
 49

wide slit

tapering edges

widening edges
 50

MANIPULATING KIRIGAMI MODULES

SYMMETRICAL QUAD PATTERN

These modules are arranged in a simple quad pattern with
slight variations in order to understand their relationship to one
another. These studies look at the resolution of panelization
and experiment with changing single folds from mountain to
valley. A single fold can change the global pattern from flat,
to cylindrical, to undulating.

ASYMMETRICAL QUAD PATTERN

Taking some of the knowledge from module placement
manipulation, the next step is to move around and rotate
the modules. While moving a module up and down is simple
and successful, module rotation is less so as it creates many
unforeseen issues that are much more mathematically
complicated to solve. The diagrams on the next page explain
these limits.
51
 52

MODULE MANIPULATION

Further investigation in global configurations requires a more in-

depth look at translation of modules. While moving a module
up, down, and side to side, the kirigami cut remains unchanged
because all edge cut sides remain equal.

However, if a rotation is made, the cut edges are no longer

equal and must be adjusted to compensate for the additional
length. This also involves the addition of another crease to
alleviate any tension between plates and subsequently
complicates the kirigami pattern.

module translation

module rotation
 53

unlimited translation

unequal edge lengths need to

be compensated for
 54

SIXTON PATTERN

Hexagonal patterns allow much more freedom when designing

a global pattern as the geometry of triangles is embedded in
the global pattern as well. These studies attempt to examine
at synclastic and anticlastic surfaces, and investigate how one
could potentially approximate curvature and sustain variation
within the patterns.
55
 56

SURFACE DECONSTRUCTION

These surfaces have the opposite approach of previous tests, as

they begin with the surface design, decompose the curvature
and slope, and apply the learned fundamentals of kirigami
modules. After the surface is decomposed, a kirigami cut
pattern is arranged with rigid plates sandwiching a thin sheet
of flexible plastic to act as a hinge. The mitred joints inform the
angle between two adjacent plates, and rubber bands act as
the actuation to pull the structure into its final configuration.
 57

mitre joint for

flexibility with stops

held in place wood sandwich

with tension panels

polypropelyene for
live hinge

Figure 06.01 Joint possibility with multi-material kirigami

Global Form Deconstruction

Figure 06.02 Sandwich assembly with rigid kirigami

 58

RIGID KIRIGAMI
This prototypical model is constructed of two layers of wood
panels, sandwiching a sheet of polypropylene plastic between
them. Mitred joints, similar to the Woodskin project, allow folding
between panels with a specific programmed angle. The final
configuration is held in place with rubber bands to simulate
tension cables and small bolts.
59
 60

Figure 06.03 Folding and unfolding of rigid kirigami

61
 62

KIRIGAMI TRIPLET COMBINATION

Kirigami cuts can also be combined into double-layered

structures in many base forms: triangles, squares, and hexagons.
These “triplets” result in more complex forms that create an
interesting sandwiching structure. The structural capabilities
of these modules have enormous amount of potential for
thick plate systems, and they can be varied to approximate
curvature.

After extensive research into the kirigami modules and

potentials, this Programmable Folding focuses on the triangle
triplets, as they allow easy surface approximation and eliminate
planarity issues of quads and hexagons.

section and axon of ki

formulating kirigami triplet

section and axon of kirigami triplet double-layer structure

Figure 06.02 Kirigami double-layer structural potentials

 63

Figure 06.01 Potential kirigami triplet combinations

64

MANIPULATING TRIANGLE RATIOS

One of the two methods for creating curvature in kirigami is

to take advantage of the triangle relationship of each side
of the sandwich structure. By subtly adjusting these ratios of
upper and lower triangles, one can control the intensity and
location of curvature. Because the triangles are being adjusted
symmetrically around their center, there is very little additional
computation that needs to be applied to create a working
kirigami pattern.

Figure 06.02 Changing top and bottom triangle

ratios to achieve varying levels of curvature
 65

:
66

APPROXIMATING OVERALL CURVATURE

Based on the relationship between the top triangle, bottom
triangle, and depth of the pattern, one can calculate the
curvature at any specified point. This subdivision is important to
the resolution and aesthetic of the final design. These simple
angle calculations are embedded within Programmable
Folding’s computational design tool so that geometrical
constraints predict unfoldable kirigami patterns.

W1

H
θ

W2 W1 -W 2

W1

H θ θ

W
2
W
1 -W
2

θmax= tan
-1
( ) ΔH
ΔW
θ
total
= 180 - 2 θ θ = 180 - 2θ
max

Figure 06.03 Calculating curvature at a specific point

CALCULATING THE ANGLE BETWEEN SEGMENTS:

Based on the relationship between the top triangle,
bottom triangle, and depth of the pattern, one can
calculate the curvature at any specified point.
 67

100 100 100

50 25 10

θ max = 233° θ max = 208° θ max = 191°

68

1D CURVATURE

All of the previous kirigami studies have investigated double

curvature (dome) and zero curvature (flat surfaces). In
manipulating kirigami to curve in one direction, it is impossible
to do so without scaling the triangles unevenly. Figure 06.05
and Figure 06.06 demonstrate the only ways to create a 1D
curvature with equilateral triangle kirigami.

Figure 06.04 explains the geometrical problems with 1D kirigami

curvature. The desired inside and outside layers are represented
by the cylinders nested inside each other. However, the inside
cylinder must be scaled in three dimensions instead of only
two. While the kirigami works for the first few layers, one moves
up the cylinder, the connection between inside and outside
triangles become more and more skewed. Furthermore,
unrolling the geometry (Figure 06.05) creates a pattern that
requires additional joining. Figure 06.06 utilizes truncation as an
alternative solution, although the truncation forms additional
holes within the folded pattern that cannot be closed.

These experiments were a crucial turning point in the research

for Programmable Folding, as it quickly became evident that
a much more robust computational tool was necessary to
program the required kirigami cuts.

connection panels become

increasingly skewed

Figure 06.04 Diagram of internal scaling that must happen in

order to achieve one dimensional curvature
 69

Figure 06.05 Approximating 1 dimensional curvature

with without truncation

Figure 06.06 Approximating 1 dimensional curvature

with truncation
 70

MANIPULATING TRIANGLE RATIOS

As seen in the previous research development and in Figure
06.07, it is possible to create double curvature by having a
consistent ratio of top to bottom triangles in a kirigami sandwich
structure.

The next step is to program local curvature by asymmetrically

manipulating the top and bottom sides. In Figure 06.09, the top
side of the triangles are scaled from one end of the pattern to
the other. This creates zero curvature on one end, and gradually
transitions into double curvature.

Figure 06.08 uses the same approach, but scales the triangles
on each side inversely to each other, creating a simple double
curvature. While triangle scaling to achieve curvature works
when applied gradually, further computational tools are
required when working with more complex surfaces.

Figure 06.07 Consistent triangle ratios

 71

Figure 06.08 Change triangle ratios with indirect proportions

:
:

: :

Figure 06.09 Change triangles with single attractor on one side

 72

ASYMMETRICAL MANIPULATIONS

After thoroughly investigating symmetrical scaling of triangles to

create curvature in kirigami sandwich structures, Programmable
Folding looks to asymmetrical manipulation. Because adjacent,
connecting edges of kirigami have to be equal, it is necessary
to relax modified geometry to create a foldable pattern.

Remember the complex problems with one dimensional

curvature? By asymmetrically scaling one side of the sandwich
structure and relaxing the geometry with kangaroo, it is possible
to create a one dimensional curvature from one sheet without
truncation.
 73

Figure 06.10 Asymmetrical manipulation and edge relaxation

:
 74

KIRIGAMI RULES

After achieving a wide range of variability in kirigami, there are

two rules which cannot be broken. As shown in Figure 06.11,
the edge lengths must be equalized after asymmetrically
manipulating the triangles. This is to ensure that the kirigami
pattern can fold and have all the seams match. Figure 06.12
shows the geometric limits of manipulating the pattern, and
how the scaling of triangles past the limits of adjacent triangles
is not possible.

Figure 06.11 Edge length correction to ensure foldability

 75

Figure 06.12 Over-manipulation of kirigami pattern

 76

SADDLE SHAPE
After experimenting with manipulating kirigami in one
dimension, I applied the same logic to scale the pattern in two
dimensions. This scaling, after cut out lengths are equalized,
exhibits negative Gaussian curvature and folds into a saddle
shape.

Figure 06.13 Manipulating the pattern in two dimensions

 77

Figure 06.15 Saddle simulation. Highlighted: direction of triangle scaling

Figure 06.14 Simple saddle prototype

 78

MANIPULATING KIRIGAMI SUB-STRUCTURE

Apart from manipulating kirigami triangle ratios, it is possible
to determine general forms by modifying the hexagonal sub-
structrue. This catalogue shows the wide range of curvature
possible by simply changing a simple vertex. The research
proposal will use ths as one of the methods for approximating
curvature.

1x_octagon_lowres 1x_octagon_highres

1x_quad, 1xoctagon_highres
1x_quad, 1x_octagon_lowres

1x_heptagon, 1xpentagon_lowres 1x_heptagon_highres

Figure 06.16 Mesh substructure catalogue and resulting fold geometry

 79

1x_quad_lowres 1x_quad_highres

1x_decagon_lowres 1x_decagon_highres

1x_dodecagon_lowres 3x_octagon_highres
 80

MANIPULATING KIRIGAMI SUB-STRUCTURE

Building on the research of kirigami sub-structures, this monkey
saddle is formed by replacing the central hexagon of a
pattern with an octagon. The extra two vertices program the
flat pattern with negative Gaussian curvature. This rest of the
kirigami triangles are applied normally, relaxed into a foldable
pattern, and assembled.

Figure 06.17 Monkey saddle base pattern

Figure 06.18 Flat folding pattern folding into a monkey saddle

 81

Figure 06.19 Monkey saddle folding simulation

82

JOINT DETAIL TESTS

The joint details went through a number of iterations beginning
with the paper models. A system of tabs was developed out
of necessity to assemble intricate models more quickly. After
experimenting with different scales and secondary materials, I
settled on a 2-2-tabbing joint system. The first “2” refers to the
number of tabs per side, to add stability to the edges. The
second “2” refers to the number of adjacent modules that each
module is connected to. After the module interlocks its two
open sides together, it locks into the next module. The second
one, in turn, is locked into the third and the third is locked back
to the first. This reciprocal tabbing system balances structural
stability and assembly time.

Figure 06.20 2-2 tabbing system

 83

Figure 06.21 Single, large tabs Figure 06.22 Two tabs per side (final design)

Figure 06.23 Zip tie joining experiments

84

SURFACE RESOLUTION
One interesting discovery during the design research phase was
playing with the scale and resolution of kirigami fold patterns.
The monkey saddle pattern used slightly less than a full sheet of
polypropylene (because of its symmetric nature and asymmetry
of the plastic sheet). The next prototype takes 1/6th of the
monkey saddle pattern and scales the resolution up. Despite
the difference in triangle numbers, both prototypes take up
approximately the same area, only with differing thicknessess.

monkey saddle prototype

average triangle length: 40mm
m
400m 0m
m 40

large prototype
average triangle length: 120mm

m
0m
400m 60
m

original polypropylene sheet

m
m
00
12
800m
m

Figure 06.24 Size comparison of sample prototypes and their volume

 85

Figure 06.25 Monkey saddle pattern (400mm x 400mm x 40mm)

Figure 06.26 1/6th of monkey saddle pattern (400mm x 600mm x 120mm)

84
 85

RESEARCH PROPOSAL
CHAPTER 07
 86

DESIGN GENERATION
The overall design is generated from the two driving factors
that have been explored in the design research phase:
approximating curvature from bespoke mesh sub-structures
and manipulating local curvature with top and bottom triangle
ratios.

First, the desired surface is approximated and unrolled into a flat

pattern. Here we apply the first shape determining method: a
hexagonal mesh is generated with an octagon at the curvature
centroid to create negative curvature. Then the mesh is relaxed
and the user chooses the boundary conditions of the final form.

After the base mesh is finalized, a basic kirigami triangle

pattern is applied. From there, the second tool for manipulating
curvature is applied. Based on attractor points, one can
program in local positive or negative curvature. Next, the
pattern is run through kangaroo to equalize neighboring edge
lengths. Once complete, the lines are connected to make the
familiar kirigami pattern. After applying dashing and tabbing,
the pattern is ready for production.

Figure 07.01 Site planes generate design

ceiling-wall

stair-ceiling
central vertex

stair-wall

wall-wall

unrolled site conditions mesh pattern based on desired general form

user-determined boundary condition generate kirigami triangles from mesh base

 87

locally manipulate curvature generate kirigami pattern from base triangles

check line lengths equilaze lines

generate kirigami pattern cut sheet specs

 88

INSTALLATION PROPOSAL
The final architectural installation of Programmable Folding exhibits
the control that is achievable by asymmetrically manipulating
kirigami fold patterns. It will comprise of ~25 sheets of 0.5mm
polypropylene, which, when assembled, has a volume of
over 2 m². The design exhibits a response to pre-existing site
conditions, filleting the wall corner and making a nice transition
from the half-floor staircase into a wall element. Its specific
curvature is possible by both manipulation of kirigami geometry
substructure, as well as programmed local curvature based
on triangle ratios. This prototype showcases the full control that
Programmable Folding has achieved and provokes further
questions about design opportunities and architectural space
potentials of kirigami.
89
90
 91

DISCUSSION
CHAPTER 08
 92

DISCUSSION

GEOMETRICAL CONSTRAINTS
As seen in the mesh base catalogue, there are a few geometries
that are not compatible with kirigami triangle patterns. The
next step in this research would be to branch out, investigating
other geometries and understanding the full constraints. With
a wider range of geometries, a more detailed catalogue can
be developed and then the full understanding of architectural
implications of kirigami would be realized.

JOINTS
Although it is not a problem in larger sheets, smaller folding
patterns do not have the tabbing strength to combat the
internal forces. To the right is an example from the monkey
saddle prototype: even though the geometry works to keep
all folds planar, the tabs are simply not strong enough at the
center vertex where all the material forces are concentrated. It
is possible, here, to introduce apertures to release the stress. This
method acts as a design driver, and allows the general viewer
to understand the base mesh logic with a cursory glance.

1x_quad, 1x_octagon_lowres 1x_heptagon_highres 3x_octagon_highres

Figure 08.01 Problematic mesh bases, usually caused

by odd numbered vertices or the addition of too many
verticies next to each other
 93

Figure 08.02 Center vertex of monkey saddle prototype

94
 95

OUTLOOK
CHAPTER 09
 96

EMBEDDED ASSEMBLY INTELLIGENCE

Although Programmable Folding’s final installation is
incrementally folded, one possible avenue for further research
is different kinds of actuation. Figures 9.01-9.04 demonstrate
the possibilities of actuation explored in various engineering,
architecture, and biology research. These actuators range
from heat, robots, electricity or light.

Although Programmable Folding utilizes a flexible plastic

because of its’ ability to incrementally fold, it is extremely
relevant to investigate thick, rigid materials when applying
kirigami to larger, architectural scales. These would require
some kind of simultaneous actuation, as they are not flexible
enough for incremental folding. This design opportunity serves
as a driver to further embed the assembly process within the
programmed fold patterns. Integrating such intelligence would
challenge the traditional design process and existing tools, and
speculates on the potential of large scale deployable, folded
structures.

Figure 09.01 Self-deploying origami stent grafts

 97

Figure 09.02 Robotic folding

Figure 09.03 Electrical programmable matter Figure 09.04 Origami-enabled deformable silicon solar cells
 98

ARCHITECTURAL SYSTEM & FUNCTION

As an outlook, I would like to consider the design implications
of kirigami in a varying range of architectural systems.
Whether applied to industrial design, temporary shelters, or
full-scale permanent structures, kirigami’s spatial and structural
capabilities can be implemented across multiple scales. More
complex fold patterns, a higher degree of control, and varying
thicknesses are all topics to be further explored. Using kirigami
as a design driver for localized system differentiation, one
can discover new functional roles and material possibilities of
deployable sheet material in design and architecture.

industrial design temporary shelters

 99

permanent, structural design space

100
 101

ACKNOWLEDGMENTS:

Tutors:
David Correa
Oliver David Krieg
Benjamin Felbrich
Achim Menges
Jan Knippers

University of Pennsylvania
Physics & Astronomy Department:
Toen Castle, Randall Kamien

RoboFold : Ema & Gregory Epps

Design & Development:

Djordje Stanojevic
Boyan Milhalov
Kenryo Takahashi
Emily Scoones
Maria Yablonina
Yuliya Baranovskaya
Georgi Kaslachev

Fabrication:
Colin O’Keefe
Josh Few
Sasha Mballa
Bruno Knychalla
Leonard Balas
Becca Jaroszewski

My family:
Donna Stamps
Danny Noneaker
Don Stamps
Pat Stamps
 102

REFERENCES
N. M. Benbernou, E. D. Demaine, M. L. Demaine, and A. Ovadya (2011). “Universal Hinge Patterns for Folding Orthogonal Shapes
in Origami.” A K Peters Ltd., pp. 405-419.

H. Buri, Y. Weinand (2010). “Origami: Folded Plate Structuctures.”Laboratory for timber constructions IBOIS, - Ecole Polytechnique
Fédérale Lausanne, EPFL Lausanne, Switzerland
.
A. Byoungkwon, N. Benbernou, E. Demaine, D. Rus (2011). “Planning to Fold Multiple Objects from a Single Self-Folding Sheet.”

T. Castle et al. (2014). “Making the Cut: Lattice Kirigami Rules.” Phys. Rev. Lett. 113, 245502.

G. Epps et al (2015). “Foldable = Makable.” www.curvedfolding.com.

B. Felbrich. (2013). “Bionics in Architecture: Experiments with Multi-Agent Systems in Irregular Folded Structure.” Tutors: S.
Wiesenhütter, D. Lordick

J. Gattas, W. Wu, Z. You. “Miura-Base Rigid Origami Parameterizations of First-Level Derivative and Piecewise Geometries.”
Department of Engineering Science: University of Oxford.

F. Gioia, D. Dureisseix, R. Motro, B. Maurin (2014). “Design and Analysis of a Foldable/Unfoldable Corrugated Architectural
Curved Envelope.” Università Politecnica delle Marche

E. Hawkes, B. An, N. M. Benbernou, H. Tanaka, S. Kim, E. D. Demaine, D. Rus, and R. J. Wood (2010). “Programmable Matter by
Folding.” Proceedings of the National Academy of Sciences.

Hoberman, C. (1988). “Reversibly Expandable Three-Dimensional Structure.” United States Patent, No. 4,780,344.

K. Kuribayashi, K. Tsuchiya, Z. You, D. Tomus, M. Umemoto, T. Ito, M. Sasaki (2005). “Self-Deployable Origami Stent Grafts as a
Biomedical Application of Ni-rich TiNi Shape Memory Alloy Foil.”

Y. Klein, E. Efrati, and E. Sharon (2007). “Shaping of Elastic Sheets by Prescription of Non-Euclidean Metrics,” Science 315 1116-
1120.

P. Jackson(2011). “Folding Techniques for Designers: From Sheet to Form.” Laurence King Publishing Ltd. London, United Kingdom.

R.J. Lang (1996). A Computational Algorithm For Origami Design.” Proceedings of the Twelfth Annual Symposium on
Computational Geometry (ACM, New York), pp. 98–105.

L. Mahadevan and S. Rica (2005). “Self-Organized Origami.” Science 307, 1740.

S. Miyashita, D. Rus (2013). “Multi-Crease Self-Folding by Uniform Heating.” Computer Science and Artificial Intelligence
Laboratory: MIT.

B. Mihaylov (2014). “Encoded Surfaces: Topological Design of Woven Hybrid Textiles.” ICD Diploma

McNeel. “Grasshopper – Generative Modelling for Rhino.” http://grasshopper.rhino3d.com/

K. Miura (1985). “Method of Packaging and Deployment of Large Membranes in Space.” Proceedings of the 31st Congress of
the International Astronautical Federation, IAF-80-A 31, (American Institute for Aeronautics and Astronautics, New York, 1980), pp.
1-9.

J.H. Na et al. (2014). “Programming Reversibly Self-Folding Origami with Micro-patterned Photo Crosslinkable Polymer Trilayers.”
Adv. Materials 27, 79-85.

C. Onal, R. Wood, D. Rus (2011). “Towards Printable Robotics: Origami-Inspired Planar Fabrication of Three-Dimensional
Mechanisms.” 2011 IEEE International Conference on Robotics and Automation.

J. Paulose, B.G. Chen, and V. Vitelli (2015). “Topological Modes Bound to Dislocations in Mechanical Metamaterials.” Nature
Phys. 11, 153-156.

Piker, Daniel (2009). “Origami Electromagnetism.” https://spacesymmetrystructure.wordpress.com/2009/03/24/origami-

electromagnetism/

H. Pottmann, A. Asperl, M. Hofer, A. Killian, D. Bentley (2007). “Architectural Geometry.” Bentley Institute Press: Exton, PA.

J.F. Sadoc, J. Charvolin, and N. Rivier (2013). “Phyllotaxis on Surfaces of Constant Gaussian Curvature.” J. Phys. A: Math. Theor.
46, 295202.

M. Schenk (2011). “Folded Shell Structures.” University of Cambridge: Doctoral Thesis. Supervisor: S.D. Guest

M. Schenk, S. D. Guest. (2011). “Origami Folding: A Structural Engineering Approach.”

 103

J. L. Silverberg et al. (2014). “Using Origami Design Principles to Fold Reprogrammable Mechanical Metamaterials.” Science 345
647-649 (2014).

J.L. Silverberg, et al. (2015). “Origami Structures with a Critical Transition to Bi-stability Arising from Hidden Degrees of Freedom.”
Nature. Mat. 14, 389-393.

D. Sussman, Y. Cho, T. Castle, T. Gong, E. Jung, S. Yang, R. Kamien (2015). “Algorithmic Lattice Kirigami: A Route to Pluripotent
Materials.”

R. Tang, H. Huang, H. Tu, H. Liang, M. Liang, Z. Song, Y. Xu, H. Jiang, H. Yu (2014). Origami-enabled deformable silicon solar cells.
Applied Physics Letters 104.

Tachi (2009). “Generalization of Rigid-Foldable Quadrilateral-Mesh Origami.” Journal of the International Association for Shell
and Spatial Structures.

T. Tachi (2013). “Freeform Origami Tessellations by Generalizing Resch’s Patterns.” Proceedings of the ASME 2013 International
Design Engineering Technical Conferences and Computers and Information in Engineering Conference. University of Tokyo.

T. Tachi (2013). “Interactive Form-Finding of Elastic Origami.” University of Tokyo.

T. Tachi (2009). “Simulation of Rigid Origami.” In Origami: The Fourth International Conference on Origami in Science,
Mathematics, and Education.

T. Tachi (2010). Freeform Variations of Origami , J. Geometry and Graphics 14, 203-215.

M. Tolley, S. Felton, S. Miyashita, L. Xu, B. Shin, M. Zhou, D. Rus, R. Wood (2013). “Self-Folding Shape Memory Laminates for
Automated Fabrication.”

Z. You (2014). “Folding Structures Out of Flat Materials: Reconfigurable machines and internally structured materials can be
created through folding.” Materials Design: Vol. 345, Issue 6197.
 104

LIST OF FIGURES
Figure 01.01 http://thomasdiewald.com/blog/?p=743

Figure 03.01
Folding Techniques for Designers: From Sheet to Form
https://spacesymmetrystructure.files.wordpress.com/2009/01/foldingpatterns.pdf
Folding Techniques for Designers: From Sheet to Form
Folding Techniques for Designers: From Sheet to Form
https://spacesymmetrystructure.files.wordpress.com/2009/01/foldingpatterns.pdf
Folding Techniques for Designers: From Sheet to Form
http://www.detail.de/architektur/themen/origami-faltkunst-fuer-tragwerke-000497.html
https://farm8.staticflickr.com/7028/6684347075_b272005316.jpg
https://spacesymmetrystructure.files.wordpress.com/2009/01/foldingpatterns.pdf
https://spacesymmetrystructure.files.wordpress.com/2009/01/foldingpatterns.pdf
http://blog.novedge.com/2014/05/the-edge-michele-calvano-and-the-architecture-of-folded-surfaces.html

Figure 03.02 https://instagram.com/hyperqbert/

Figure 04.01
http://www.evolo.us/architecture/bowoos-bionic-research-pavilion-is-inspired-by-marine-biodiversity/
http://laughingsquid.com/wp-content/uploads/2014/02/the-evolution-door-a-clever-door.jpg
http://www.a10.eu/news/headlines/sculptural_object_novi_sad.html
http://www.archdaily.com/362951/woodskin-the-flexible-timber-skin/
http://2.bp.blogspot.com/_Je8B4--Ky9E/S52NTUSBoII/AAAAAAAAADI/5bzAN2vZ_sE/s1600-h/olafur_eliasson2-1.jpg
https://vimeo.com/89379866
www.rvtr.com
http://www.eric-hull.com/appended-space/
http://www.evolo.us/architecture/bowoos-bionic-research-pavilion-is-inspired-by-marine-biodiversity/
http://thedesigninspiration.com/articles/shenzhen-baoan-international-airport-expansion-t3/
http://www.eikongraphia.com/?p=324
http://www.pleatfarm.com/2009/12/21/folded-plate-hut-in-osaka/

Figure 04.03 T. Tachi (2013). “Freeform Origami Tessellations by Generalizing Resch’s Patterns.” Proceedings of the ASME 2013
International Design Engineering Technical Conferences and Computers and Information in Engineering Conference. University
of Tokyo.

Figure 04.02 https://spacesymmetrystructure.wordpress.com/2009/03/24/origami-electromagnetism/

Figure 04.04 http://www.tessellated.com/

Figure 04.05 http://www.wired.com/2015/01/quanta-curves-from-flatness-kirigami/

Figure 04.06, - Figure 04.08 D. Sussman, Y. Cho, T. Castle, T. Gong, E. Jung, S. Yang, R. Kamien (2015). “Algorithmic Lattice
Kirigami: A Route to Pluripotent Materials.”

Figure 05.04 https://en.wikipedia.org/wiki/Gaussian_curvature

Figure 09.01 K. Kuribayashi, K. Tsuchiya, Z. You, D. Tomus, M. Umemoto, T. Ito, M. Sasaki (2005). “Self-Deployable Origami Stent
Grafts as a Biomedical Application of Ni-rich TiNi Shape Memory Alloy Foil.”

Figure 09.02 www.robofold.com

Figure 09.03 E. Hawkes, B. An, N. M. Benbernou, H. Tanaka, S. Kim, E. D. Demaine, D. Rus, and R. J. Wood (2010). “Programmable
Matter by Folding.” Proceedings of the National Academy of Sciences.

Figure 09.04 R. Tang, H. Huang, H. Tu, H. Liang, M. Liang, Z. Song, Y. Xu, H. Jiang, H. Yu (2014). Origami-enabled deformable silicon
solar cells. Applied Physics Letters 104.
105
106
107
108

ABSTRACT:
The aim of this thesis is to bridge the gap
between skin and structure. Programmable
Folding explores computationally
manipulating the geometry of kirigami
folding patterns in order to approximate
complex double curvature. These
differentiated patterns can be translated
to flat-pack sheets with an embedded
assembly logic. By programming the
underlying geometry and intrinsic curvature,
the resulting kirigami patterns investigate
potential design systems of how this logic
can be applied to architectural design.

KEY WORDS:
folding – pattern – flat-pack –
double-layer – kirigami - computational
geometry – plastic – corrugation