1

Analysis, Design and Construction of Origami

Inspired Structures.


Rahul Vaish


School of Computing, Science and Engineering
University of Salford


This dissertation is submitted in partial fulfillment of the
requirements for the Msc degree in structural engineering.


2

DECLARATION

I, Rahul Vaish declare that this dissertation is my own work. Any section, part or
phrasing of more than twenty consecutive words that is copied from any other work or
publication has been clearly referenced at the point of use and also fully described in the
reference section of this dissertation.

















3



Acknowledgements

I am grateful to the faculty of structural engineering at the school of computing, science
and engineering at the University of Salford for their support and guidance during the
course of my dissertation work.
I am particularly thankful to my advisor, Mr Neil Currie for his time, ideas and valuable
discussions that were a constant source of motivation and encouragement during the
period of my work.













4

Abstract
The study aims to look at solutions for spatial designs in the realm of plated shell
structures. A systematic review into the Japanese art of paper folding reveals suitable
patterns that readily adapt to spatial requirement of a dome shaped structure.
An approach to model the pattern geometry over various forms has been devised and
three standard spatial forms the Paraboloid, Catenoid and the Hemisphere are compared.
Experimental investigation is thereby conducted to quantify the aspects of interest such as
stress and efficiency through a parametric study of the most suitable form with a chosen
pattern namely the Miura Ori.
Further research aims to look closely at the effects of the parameters, or the number of
corrugations on the weight and stresses that govern their forms via a Finite Element
Formulation.
Intuitive notions of rigidity are preserved although certain behaviours such as stress
distribution with the parameters are not so straightforward. Recommendations towards
design and analysis have been made based on the findings of the experiments.

Keywords: Origami, Plated Shell, Miura Ori.












5


List of Figures
Figure no description Page no
2.1 Cantilever barrel vaults by Diesto
2.2 The hyperbolic Paraboloid
2.3 Concrete Shell and formwork by Candela employing the Hypar geom-
etry

2.4 Tensile membrane structure
2.5 Tensile membrane millennium dome.
2.6 Various configurations for a lattice grid
2.7 Folded art form with its crease pattern
2.3.1 Mountain and valley fold
2.3.2 Origami wave
2.3.3 Origami earwig
2.4.1 Uniaxial base
2.4.2 Flat foldability
2.4.3 Non flat foldable as it self-intersects
2.5.1 Spherical representation and Gaussian curvature
2.5.2 Polyhedral vertex and Gaussian curvature
2.5.3 Four fold origami vertex
2.5.4 Developable and non -developable four fold vertexes
2.6.1 The miura ori pattern
2.6.2 Freeform variations in the miura ori pattern
2.6.3 Fold mechanism and the parameters
2.6.4 The Yoshimura pattern
2.6.5 Foldable tube
2.6.6 The hexagonal pattern
2.6.7 Axial crushing of PVC tube
2.6.8 The pattern
2.6.9 Vault with hexagonal pattern
2.6.10 Vault with diamond pattern
2.7.1 Reverse fold
2.7.2 Reverse fold applied to create a frame
2.7.3 Origami frame for the chapel of St Loup.
2.7.4 Changing the pattern
2.7.5 Changing the corrugation amplitude
2.7.6 S shaped curves
2.7.7 Barrel vault with diamond pattern
2.7.8 Vaults with varying sectional profile and height
2.7.9 Miura ori vault
2.7.10 Radially folded yoshimura pattern and crease pattern
2.7.11 Radial miura ori pattern
2.7.12 Radial diamond pattern
2.7.13 Twist foldable diamond pattern
2.7.14 Radially folded hexagonal cone
6

2.7.15 Groin vault approximation
2.7.16 Combination of radial and barrel diamond pattern
2.7.17 Double curvature shapes
2.7.18 Radial diagonal pattern (orientation of quadrilaterals)
2.7.19 Radial diagonal pattern(crease alignment)
2.7.20 Diagonal pattern(crease orientation)
2.7.21 Radial hexagonal(orientation of quads)
2.7.22 Hexagonal (size orientations)
2.7.23 Hexagonal(shape of quads)
2.7.24 Radial hexagonal(orientation of crease)
2.7.26 Origami frame
2.7.27 Parametric variation A
2.7.28 Parametric variation B
2.7.29 Yoshimura and miura ori vault with parametric variation
2.7.30 Concrete folded plate structure
2.7.31 IBM pavilion, Renzo Piano
2.7.32 Deployable structure based on the yoshimura pattern
2.7.33 The Yokohoma cruise terminal, Japan
2.7.34 The chapel of St Loup, Buri
2.7.35 Assembly hall, university of Illinoi
2.7.36 Artistic rendition of origami inspired market hall
2.8.1 The Resch pattern
2.8.2 Four fold mechanism and a detailed view of the hinge
2.8.3 Triangle based designs allow more flexibility
2.8.4 Proposals for rigid origami
2.8.5 Fabrication of hinge using vacuumatics
2.8.6 Cross laminated timber panels
2.9.1 Modular origami(rigid and kinematic)
2.10.1 Curved crease domes
2.11.1 Forces on folded plates
2.11.2 Deformation of vertex
2.11.3 coupling
2.11.4 Forces and displacement at plate edges
2.11.5 Two different prototypes
2.11.6 Failure of hinges
2.12.1 Membrane forces in a hemispherical dome
2.12.2 Boundary conditions
2.12.3 Variable thickness for an arch and a shell
2.12.5 Hanging weight model and force polygon for an arch
2.12.6 Dome of the mausoleum of Farag Ibn Barqak,Cairo
2.12.6.1 Construction methodology
2.12.7 Dome failure
2.12.8 Vault at kings college London
2.12.9 Mideaval timber vault
2.12.10 Ribbed masonry dome
2.12.11 Geodesic hemisphere
2.12.12 Geodesic dome with plated vertex
2.3.1 Structural optimisation
7

3.1 Transfer of forces in a shear connection and a miura ori dome
3.2.1 Radially folded hybrid dome
3.2.2 Radial crease pattern
3.3.1 Experimental procedure
3.4.1 trapezium
3.4.2 Spatial position of the plates
3.4.3 Arc length discretization of the parabola
3.4.4 Plate angles
3.4.5 Plate areas
3.5.1 Plate element capable of bending and membrane action

3.5.2 Displacement of a sector
3.5.3 Forces at base
3.5.5 Stress field for rotation free(P,H,C)
3.5.6 Stress field for fixed rotations(P,H,C)
3.5.7 Displacement field
3.6.1 Boundary condition
3.6.2 Range of plate boundaries
3.6.3 Variations in maximum stress and deflections(case 1)
3.6.4 Variations in maximum stress and deflections(case 2)
3.6.5 Variation in weight and thickness with meridian segments
Variation in stresses and displacement with meridian segments
Variation in stresses(Nv=5,10,12,15,20)
Displacement field
Variation in thickness with P
Variation in stresses and displacemant
Variation in thickness and weight
Stress distribution(Nh=12,16,20,24)
Displacement field
Typical deflection profile and the use of tie
Uneven discretization of vertical segments
Hybrid stiffenend dome
Connection details








8



List of tables:


















9






















10


Contents
List of Figures ............................................................................................................................. 5
CHAPTER 1 .............................................................................................................................. 12
INTRODUCTION ..................................................................................................................... 12
1.1 Structure of the Dissertation .......................................................................................... 13
1.2 Aims and Objectives............................................................................................................ 14
CHAPTER 2 .............................................................................................................................. 15
LITERATURE REVIEW ......................................................................................................... 15
2.2 ORIGAMI: Historical Development ................................................................................... 21
2.3 Origami folding: Basic Considerations .............................................................................. 25
2.3.1 Crease, Crease Pattern, Folding Pattern, Mountains and Valleys ................................ 25
2.4 Origami Design: .................................................................................................................. 28
2.4.1 Flat Foldability: ............................................................................................................ 30
2.5 Origami and Curvature: ....................................................................................................... 32
2.6 Few Common Patterns ........................................................................................................ 35
2.6.1 The Miura- Ori pattern ................................................................................................. 35
2.6.2 The Yoshimura Pattern: ................................................................................................ 38
2.6.3 The Diamond Pattern: ................................................................................................... 40
2.7 Origami for architectural considerations: ............................................................................ 42
2.7.1 Radial origami: ............................................................................................................. 48
2.7.2 Parametric variations in origami: ................................................................................. 54
2.7.3 Some Notable Examples: ............................................................................................. 61
2.8 Rigid Origami: ..................................................................................................................... 66
2.8.1 Kinematics: ................................................................................................................... 66
2.9 Modular origami: ................................................................................................................. 71
2.10 Curved crease: ................................................................................................................... 72
2.11 Behavior under Load: ........................................................................................................ 73
2.12 Axisymmetric Domes: ....................................................................................................... 79
2.13 Optimization of Shell Structures: ...................................................................................... 90
2.14 Conclusion: ........................................................................................................................ 93
11

CHAPTER 3 .............................................................................................................................. 94
DESIGN OF EXPERIMENT .................................................................................................... 94
3.1 Choice of pattern: ................................................................................................................ 94
3.2 The Miura - Ori radial pattern: Relationship between 2D and 3D form............................. 96
3.2.1 Range of parameters: .................................................................................................. 100
3.3 Design of Experiment: ......................................................................................................... 99
3.4 Geometric modeling of the Miura-Ori pattern .................................................................. 103
3.4.1 Results ........................................................................................................................ 109
3.5 Finite Element Modeling ................................................................................................... 111
3.5.1Restraints ..................................................................................................................... 112
3.5.2 Loading ....................................................................................................................... 112
3.5.3 Results ........................................................................................................................ 113
3.5.4 Discussion .................................................................................................................. 116
3.6 Boundary conditions:......................................................................................................... 117
CHAPTER 4 ............................................................................................................................ 121
RESULTS AND DISCUSSION.............................................................................................. 121
4.1 Experiment Stage 1: .......................................................................................................... 121
4.2 Experiment Stage 2: .......................................................................................................... 122
4.2.1 Discussion: ................................................................................................................. 124
4.3 Experiment stage 3: ........................................................................................................... 125
4.3.1 Discussion: ................................................................................................................. 125
4.4 Experiment stage 4: ....................................................................................................... 126
4.4.1 Discussion: ................................................................................................................. 128
CHAPTER 5 ............................................................................................................................ 130
CONCLUSION AND FUTURE WORK ................................................................................ 130
5.1Conclusion: ......................................................................................................................... 130
5.2 Future Work: ..................................................................................................................... 133
List of References: ................................................................................................................... 146

12

CHAPTER 1
INTRODUCTION
Modern engineering is advancing at a rapid pace today. New paradigms emerge as we
push our limits across boundaries which are constantly being challenged as our
understanding of the world around us increases.
With the advent of computers, the modern engineer is empowered just like a biologist is
with a microscope. Today we see a rather interdisciplinary approach to problem solving
and new discoveries in materials and computation has seen a rapid modernization in the
field of structural engineering.
The modern engineer today has variety of tools at his disposal and structures are
becoming more slender, economic and cheap. We look towards a solution to explore the
use of plated structural forms to create efficient, lightweight, modern yet aesthetically
pleasing structures as it is an important responsibility as designers and engineers of the
modern fabric of society to explore interesting forms which will have a pleasant visual
impact as well as an efficient structural mechanism.
Naturally we look towards nature as it never fails to awe and inspire. Creating efficient
utility based structure which also appeal to us aesthetically can be a topping on the cake
and Origami can help us reassert this lacking aesthetic aspect in modern designs as the
structural benefits of folding are well known.
It is with this aim in mind that we proceed to undertake a study to reveal as to what
extent are we justified to use this ancient art as an inspiration to construct modern
structures.


13

1.1 Structure of the Dissertation
The study begins with a brief overview of the class of structures considered, spatial
structures, their coming of age, the problems associated and some possible alternatives
which are being successfully implemented today. Then some attention is paid to the
history and development of the art of paperfolding and how it can help us realize efficient
solutions for architecture and the problems associated with transferring the concept from
paper to a buildable rigid material with regard to the difficulties associated with their
analysis and design. Some successfully implemented designs have been mentioned and
discussed briefly.
Consideration then is given to axisymmetric domes and possible solutions in origami are
considered to arrive at a suitable pattern. Next it is of interest to take a closer look at the
structural properties after deciding upon a suitable finite element model which is obtained
after deciding upon the nature of constraints and boundary conditions to be used.
Experiment is designed to quantify the aspects of interest for structural engineering such
as stresses and deflections and define an efficient structure with regard to behavior and
weight. Origami designs, being repetitive and controlled by a few parameters are suitable
for a parametric investigation and hence the experiment compares three different forms,
namely the hemisphere, paraboloid, catenoid. The best form is forwarded to be studied
for its parameters in four stages.
The dissertation concludes with a discussion and recommendations based on the findings
of the experiment suggesting possible future work in the field.





14

1.2 Aims and Objectives
Aims: To arrive at a suitable pattern for a dome like spatial structure and investigate its
structural properties via a suitable experiment.
Objectives:
- Undertake a comprehensive literature review of various solutions for spatial
structures to understand their behavior.
- Undertake a literature review into existing patterns of folded forms to rationalize a
suitable design with consideration to the behavior of the pattern.
- Undertake a numerical experiment to understand the nature of stresses to compare
and shortlist the best form for the design via a comparative study.
- Experimentally investigate into the parameters of the pattern to quantify the
structural aspects of interest to better understand the behavior of these structures.
- Suggest recommendations as per findings.









15

CHAPTER 2
LITERATURE REVIEW
2.1 Spatial Structures: An Overview
Spatial structures are designed to cover large spaces. Over the years they have evolved
considerably and most modern structures today are fabricated as single or double layered
lattice grids. Over the years the art of creating space structures has seen considerable
experimentation and advancement and today we see have a number of varieties such as
apace frames, cable and strut models, air inflated, tension membranes, cable net, geodesic
domes, plated shells, folded plates etc.
The earlier designs were made from reinforced concrete while today mainly tubular
members are preferred for the lattices for they do not contribute to planar problems.
Shells can fail catastrophically so a good judgment of the forces is required for their
design. Moreover it is mostly form that contributes to strength.
Earlier designers relied on basic intuition of shell behavior and many designed the form
such that the compression field tends to neutralize the tension field in the structure. Most
notable of such designs are works of Felix Candela, Diesto and contemporaries.
Candela exploited the geometry of the hyperbolic paraboloid to create very efficient
forms (figure no 2.2). A lot many designs have been based on using a section of standard
space geometries although the configurations are infinite. Figure 2.1 shows a cantilever
barrel vaults by Diesto which has been prestressed at the crown to avoid tension.
Figure shows the millennium dome which works under tension. The principle is not new
as the use of tents made by animal skin was as early as prehistoric times. The membrane
of these structures is in a state of tension (prestress) supported by masts which go in
compression due to the pull. These are efficient, lightweight, economical structures,
highly form dependent. It is because the geometric stiffness of the membrane arising
from change in geometry and membrane prestress is more significant than the extensional
16

stiffness of the material.(R.Bradshaw) These are usually optimised for dynamic
amplification. As the tensile strength of fabrics is greater for uniaxial than in biaxial
loading, the failure is mainly due to tear propagation rather than tensile rupture.
Early form finding techniques were mainly empirical in nature with limitations when
nonlinear behavior is incorporated.
Today with the advent of computers we have sophisticated techniques which can be
accommodated to run various kinds of optimization to arrive at a suitable form.
Advancement in the field of computational geometry lets us analyse the geometrical
considerations like discretizations, segmenting etc. with ease.
However shells do not lend themselves to accurate analysis very easily .When the
thickness becomes small, the shell behaviour falls into one of two dramatically different
categories; namely, the membrane-dominated and bending-dominated cases. The shell
geometry and boundary conditions decide into which category the shell structure
falls, and a seemingly small change in these conditions can result into a change of
category and hence into a dramatically different shell behaviour.(Bathe 1997)
Numerous attempts have been made over the decades to model shell behavior and one of
the most reliable model today seems to be where the bending and membrane surfaces are
treated as two different surfaces.(CR 1983). See reference for a complete discussion.
However we also have the powerful numerical technique, FEM that most modern
designers use to analyse these structures.
There are problems of buckling associated with spatial structures. For positive Gaussian
curvature models, the surface is in compression and they tend to buckle while anticlastic
surfaces have problems with material failure. Modern lattice grids have problems such as
snap through buckling which is the major concern for these types.


17

One major disadvantage of all dome shaped space structures can be poor acoustic and
light capabilities.
Origami domes will be an efficient solution as they can be designed to work under
membrane action and can have certain advantages such as better acoustic performance
and light modulation capabilities.
However as far as the structural properties are concerned we would be interested to know
the nature of forces arising in the plates.


Figure no 2.1: Cantilever barrel vaults by Diesto



18


Figure no 2.2: The Hyperbolic Paraboloid



Figure no 2.3: Concrete Shell and formwork by Candela employing the Hypar geometry.


19


Figure no 2.4: tensile membrane structure




Figure no 2.5: Tensile membrane millennium dome



20





Figure 2.6: Various configurations for a lattice grid







21


2.2 ORIGAMI: Historical Development: From an ancient art form to a
modern scientific tool.
Etymologically, the word Origami originates from the Japanese words ori or oru meaning
to fold and kami meaning paper.
It's generally accepted among scholars and artists that origami began approximately in
the first century AD in China, where people discovered the simple thrill of folding paper
into various shapes and forms. Many of these primitive methods survived to this day as
basic origami shapes and moves. As the skill of making paper moved across the Eastern
world the art became quite popular in Japan where it melded with the culture and religion
at the time. Soon the Shinto religion integrated various origami shapes and creations in
their ceremonies where they remain until this day.
In 1797, "How to Fold 1000 Cranes" was published. This book contained the first written
set of origami instructions which told how to fold a crane which was considered a sacred
bird in Japan. It was a Japanese custom that if a person folded 1000 cranes, they would be
granted one wish. Origami became a very popular form of art as shown by the well-
known Japanese woodblock print that was made in 1819 entitled "A Magician Turns
Sheets of Birds". This print shows birds being created from pieces of paper.

In 1845 another book, "Window on Midwinter", was published which included a
collection of approximately 150 origami models. For centuries there were no written
directions for folding origami models. The directions were taught to each generation and
then handed down to the next. This form of art became part of the cultural heritage of the
Japanese people. One major problem of ancient origami was that the techniques and
designs were all oral; passed on through the mother to the daughter much like the oral
traditions in other parts of the world.
22

Origami requires one to make a succession of folds, creating a complex pattern of creases
that turns the piece of paper into a form made up of polygonal facet, usually in three
dimensional space.
Origami has begun to find relevance beyond merely an art form and is continuing to
expand in intricacy. In the last two decades, significant technical and artistic
developments have been made in the field, attributable to a growing interest and
development of a systematic aproach to its study.
Margherita Piazzolla Beloch gave the first set of axioms in 1936 to analyse the geometric
construction of origami. This being possibly the first contribution to origami
mathematics. Later contributions were by Huzitas when he proposed hissix axioms.
Fundamental theorems on local crease patterns around a single flat-folded vertex were
established by Jun Maekawa, Toshikazu Kawasaki, and Jacques Justin. Over the years,
several studies have explored the use of origami folding to prove geometrical
constructions and theorems. Moreover it has been demonstrated that folding technique is
more efficient in explaining geometrical proofs than traditional methods and this has
given origami a successful space in teaching and education. (Boakes.N 2010)
Mathematical Origami or technical folding as it began to be referred to, was christened
Sekkei in Japanese. Thomas Hull extended this work into the area of flat-foldability.
Robert Lang one of the pioneers in origami design, developed an algorithm around 1993
and a software program thereafter for designing origami, which he called the Tree Maker,
because it is based on graph theory that uses trees. He goes on to describe how the theory
can convert the problem of finding efficient crease pattern into one of many nonlinear
constrained optimization problems which is well researched by the computer science
community.
Lang also describes techniques such as circle packing to design efficient art forms.
He has recently published Origami Design Secrets ((Lang.R (2003)) unfolding a
computational approach to origami design.
23

Mathematical origami research generally revolves roughly around two foci -foldability
and design (Demaine .E.D 2007).The first focus origami foldability, generally asks
which crease patterns can be folded into origami that uses exactly the given creases.



Figure no 2.7: A folded art form with its crease patterns.

The simplest forms can be creases emanating from a single vertex. In these cases, we can
completely characterize which folding sequence will lead to a successful flat folded state.
The problem arises with crease patterns which have many vertices. This is where origami
sekkei or technical folding begins to play a part.
The second focus origami design is, generally, the problem of folding a given piece of
paper into an object with certain desired properties, for example, a particular shape, and
specifically in the case of this paper, into architectural forms which can be dealt with
using the tree maker or other algorithmic procedures.


24

Mathematicians today are involved in the study of the various contributions that Origami
can make to a better understanding of phenomena in the physical world. For instance a
mechanism in nature, from which we can learn much, is how leaves of some plants are
folded or rolled when un-blossomed inside the bud and how they unfurl thereafter during
blossoming. (Kobayashi H 1998)
Cedar or Beech tree leaves have simple and regular corrugated folding patterns. These
patterns can recommend ideas for the design of deployable forms and structures such as
solar panels and light-weight antennae of satellites, or for the folding of membranes such
as tents, clothes or other coverings such as large scale parasol umbrellas, which need to
be tightly packed and reduced to a small size during transportation / pre-deployment and
then, to expand to their full size at the site.
These concepts have successfully adapted by NASA to unfold satellites, car
manufacturers for airbags, cardiologists for heart stents, researchers for optic systems,
and genome scientists for DNA exploration etc. It has also been demonstrated that
origami shaped box, has better absorption characteristics and a higher buckling load than
a conventional tubular or box structure. This may have potential application in design for
impact. (Ma J and You Z (2011))
Robert Lang has used the principles of folding to create a 5m wide aperture, foldable
Fresnel transmissive telescope lens which is a prototype for an eventual 100m wide
aperture lens for a telescope. The lens is designed to be spin stabilized.
The possibilities are only increasing with each passing year as it is a very efficient way to
manipulate spatial forms, the utility of which being limited only to human imagination.




25

2.3 Origami folding: Basic Considerations
Certain basics of folding are presented here which are the fundamental building blocks of
the art form.
2.3.1 Crease, Crease Pattern, Folding Pattern, Mountains and Valleys
A crease is a line segment or a curve on a piece of paper. Creases may be folded either as
a mountain or valley fold forming a ridge or a trough. (figure3)



Figure no 2.3.1: Mountain and valley fold
A Crease Pattern (CP) is a means of conveying a folding instruction.(R). It may be seen
as a collection of lines drawn on a square of paper, meeting only at common endpoints,
which is usually a graph. A crease on a crease pattern may or may not be folded.
Moreover there may not be a unique step by step folding sequence to reach the desired
form.
Many modern origami designs, particularly if they were designed using tree theory,
circle packing, box pleating, or any of the other tools of modern design, are designed in
an all-or-nothing way. The creases all work together when they are fully folded, but it
26

is often the case that there are no intermediate states no subsets of the creases that
can be folded together, which would form the individual steps. For such a model, the only
way to assemble the model is to precrease all of the creases, and then gently bring them
altogether at once. That method of assembly, as it turns out, is most efficient and is the
approach used for folding a model from a Crease Pattern in most cases.
And thus, a crease pattern provides the folder insight into the thought processes of the
orgami composer in a manner that a step by step sequence cannot. The figure below
shows two kinds patterns and their final folded form. In figure 2.3.2, all the creases are
precreased and folded together in cohesion to obtain the desired form.



Figure no 2.3.2: origami wave (sourcehttp://www.langorigami.com/art/gallery/wave24)
27


Figure no 2.3.3: origami earwig (source:
http://web.mit.edu/chosetec/www/origami/earwig/)

Figures 2.3.2 and 2.3.3 show paper models and their respective crease pattern. However
they are different in certain aspects. All creases in figure 4 have been folded while certain
lines in figure 5 are merely for understanding purposes.

A Folding Pattern (FP) is an identification of which creases should be folded as
mountains and which as valleys. Together, a Crease Pattern (CP) and a Folding Pattern
(FP) describe a MountainValley Assignment (MVA).





28

2.4 Origami Design:
To design an origami model, one mainly needs a crease pattern and folding assignment
that will allow us to reach the final model. For various different types of models there
may be a particular approach that may be most suitable. To understand the process one
needs to be familiar with the fine interplay between the 2D representation of the fold
lines and the corresponding enclosed facet that is sent in the 3D representation upon
folding. Although difficult to a first reader, the subject is quite friendly and interesting.
The subject has received methodological treatment by pioneers in the field and one of the
most sophisticated technique may be the tree maker developed by Lang to fold insects
and figures. The main idea is to decompose the folding process into two steps: first fold a
base that roughly distributes the paper in the right places, and then shape the base
into the actual Origami model.

The input for Tree Maker is a two-dimensional stick figure that captures the essential
features of the target object. Theoretically, the stick figure is a weighted tree, the weights
being the lengths of the various edges. The computer works out a way to fold a square of
paper so that the result, which is called the base, sits exactly over the stick figure. In the
final (nonalgorithmic) step, standard origami techniques are used to flesh out the base,
making the model aesthetically satisfying and realistic.
Consider, for example, the design of a starfish where the task for Tree Maker, in effect,is
to pack five circles of equal radius inside a square.


29


Figure no 2.4.1: Uniaxial base (source (Shenk.M 2008))

It does so by formulating the design challenge as a problem in nonlinear constrained
optimization, turning the weighted tree into a set of algebraic equations.
It finds a local maximum, Lang says. And with a little bit of human intelligence, you
can convince yourself that youve found the global optimum.(Lang.R (2003))
Two important conditions that need to be satisfied for a 2D surface to be able to go in a
3D folded state when referring to a CP or an FP are, that the conditions of Isometry and
Non-Crossing.(Demaine .E.D 2007)
Isometry here implies that the distances between two points, measured by the shortest
path on the surface of the paper, are preserved by the mapping, i.e., the mapping does not
shrink or stretch the paper. The Non-Crossing condition specifies that the paper does not
cross through itself when mapped by the folded state. Portions of paper are allowed to
come into geometric contact as multiple overlapping layers, but the layers must not
penetrate each other, i.e., the mapping must not tear or cut through the paper.
In folding an assignment one may have to twist one or more facet or tuck one facet
beneath other. This may not be possible if one is working with a rigid material like
cardboard.
30

Also one may wonder, given any arbitrary crease pattern, is it possible to find a mountain
valley assignment that sends it to a flat folded state?
Origami foldability is governed by certain interesting rules and axioms which will be
presented below.

2.4.1 Flat Foldability:
One can speak of local and global flat foldability. For a structure to be deployable it
needs to be flat foldable and therefore this condition is of prime interest. Two important
theorems which provide a necessary but not sufficient condition for flat foldability are
(Hull.T 2010):


Theorem 1: Kawasakis theorem: A single-vertex crease pattern defined by angles
u
1 +
2 + n = 2
t
is flat foldable iff n is even and the sum of the odd angles (2i+1) is
equal to the sum of the even angles (2i), or equivalently either sum is equal to .


Figure no 2.4.2: Flat foldability

u u u
u t
31

Theorem 2. Maekawa-Justin theorem: In a flat-foldable single-vertex mountain-valley
pattern defined by angles 1 + 2 +. ..+ n = 2 , the number of mountains and the
number of valleys differ by 2. (M V=

2)

These two theorems although necessary are not sufficient a crease pattern can be found
that satisfies the above two theorems but is not flat foldable.



Figure no 2.4.3: non flat foldable as it self-intersects.






u u u t
32

2.5 Origami and Curvature:
For a curved surface, Gaussian curvature is useful to represent the curvature as an
intrinsic property of the surface. (Cohn-Vossen.S 1952))
The Gaussian curvature can be estimated by using the spherical representation as
illustrated in figure no 9. A closed loop is drawn on the curve around the point under
consideration and its spherical image is mapped on a unit sphere by mapping the normals
at selected points. It the sense of traversal of the points is same it is said to have a positive
Gaussian curvature, otherwise negative.

Figure no 2.5.1: spherical representation and Gaussian curvature.
Mathematically it can be expressed as the limit


0 F
G
k Lim
F

= or
1 2
1 1
k
R R
=
Where
1
R and
2
R are the principle curvatures for the point considered.

33

The idea can be extended to creased surfaces as well. Miura in his paper, demonstrates
the use of the principle to validate the existence of a set of creases with a given mountain
valley assignment. Since a folded surface is developed from a sheet, the Gaussian
curvature of the folded form should exhibit zero Gaussian curvature as the curvature of
the flat sheet is zero. For a single and two fold lines at a vertex, this is easily shown in
figure to be zero as the spherical representation is an arc enclosing nil area.
For three vertices, if all the normals have different orientations, then the enclosed area
cannot be zero and this case is not admissible for origami.
For four fold vertex, first consider a roof vertex with all mountain folds. The spherical
representation has a positive area and hence it cannot be developed. Furthermore
Calladine demonstrates that the solid angle subtended by the roof is merely the spherical
excess which is given by
1
2
n
i
i
t u
=
A =

, and the Gaussian curvature is given by

A
A
, where
A is the area associated with the vertex. (CR 1983). This does not change by addition of
creases and the curvature is entirely contained in the vertex. The Gaussian curvature is
invariant under in extensional deformation.


Figure no 2.5.2: polyhedral vertex and Gaussian curvature
34


Figure no 2.5.3: Four fold origami vertex
Miura demonstrates that the only developable configuration with four vertices is three
mountains and one valley. It is the simplest origami. Furthermore they cannot be
orthogonal otherwise the folding cannot be simultaneous. A fourfold inclined vertex is
also the generic case when a piece of paper is crushed. It possess only a single degree of
freedom and this remarkable property makes the miura ori pattern, which is a repetition
of identical four fold units(see later) simultaneously deployable in two directions.
The figure below shows two paper works make by the four fold vertex. The first uses the
foldable vertex and is developable while the second work cannot be developed.


Figure no 2.5.4: Developable and non-developable four fold vertex designs.

35

2.6 Few Common Patterns
Literature review has suggested that few patterns are more useful than others. These
patterns arise from a systematic study of buckling behavior of plates and cylindrical
shells and are therefore flat foldable and can be readily adapted to suit architectural
needs. Three common patterns are described below namely the herringbone or Miura-
Ori, diamond and hexagonal pattern.

2.6.1 The Miura- Ori pattern
Take a piece of paper and crumple it. Unfold it and observe the pattern of the folds that
appear. Normally these will be a superimposition of various modes of buckling of the
sheet but the problem, devised as an experiment to study the buckling modes of an
infinite elastic plate uniformly compressed from all sides has been studied thoroughly and
various modes have been isolated and studied for their energy index. i:e the amount of
energy taken to deform in a particular mode. It was found that the miura ori pattern has
the lowest energy index and is the mode that requires the least energy for distortion
(Koryo.M 2009).
One of the most unique properties of this pattern is its deployable property. (Koryo.M
1989)
1) It can be deployed simultaneously in two orthogonal directions.
2) It possesses only a single degree of freedom.
3) The deployment and retraction follow the same path.
It follows that when we deploy a sheet from a flat folded state to become a planar surface,
the miura pattern will strain the material least. Combined with the 1DOF mechanism, this
pattern has been successfully used in space solar panels. Also given the fact that the
resulting planar quadrilaterals form a good criterion for rigid foldability (Naohiko and
Ken-ichi 2009), its variations can be explored to design rigid architectural origami. The
36

flexibility of the pattern can be made clear from the following figure where a parametric
variation allows a flexible adaptation to an arbitrary surface. The figure below shows the
miura ori sheet and its crease pattern with variations of the angles (also referred to as
miuras). The flexibility of the pattern to adapt to various surfaces can be indicated by
figure no 21. The miura ori is the simplest and perhaps the most useful pattern.



Figure no 2.6.1: The miura ori pattern

37




Figure no 2.6.2: Freeform variations in miura ori pattern (source (Sternberg.S)









38

2.6.2 The Yoshimura Pattern:
This pattern was observed to appear on the surface of a cylinder while buckling under a
twisting moment. The transformation happens via bending only at the folds and not
stretching i: e the process is purely an origami. The pattern parameters, the fold angle and
the number of segments were sensitive to the length of the cylinder made to buckle and
the pattern provided post buckling strength. The pattern is also called as the Yoshimura
pattern as it was first discovered by Yoshimura.



Figure no 2.6.3: Fold mechanism and the parameters.

39

Figure below shows the crease pattern and its application to create a folded tube
structure.


Figure no 2.6.4: The Yoshimura pattern


Figure no 2.6.5 : Foldable Tube


40

2.6.3 The Diamond Pattern:
This pattern arises out of the buckling of a cylinder under axial compression as shown in
the figure below.


Figure no 2.6.6: The hexagonal pattern. Figure no 2.6.7: axial crushing of PVC tube

There are two possible representations of the pattern as shown below. The hexagonal and
the diamond case. They differ in the valency of the vertex. Under repetitive and
symmetrical conditions, both the patterns are flat foldable. These patterns are very
flexible and a large amount of variations is possible to produce varied shapes and designs
in space. The diamond case is very similar to the Yoshimura pattern but the difference
lies in the orientation of the folds.

41


Figure no 2.6.8: The crease patterns


Figure no 2.6.9: Vault with hexagonal pattern

Figure no 2.6.10: Vault with diamond pattern

42

2.7 Origami for architectural considerations:
For architectural applications we want the folded configuration to enclose a certain
desired space and have some structural rigidity to withstand loads.
It should be noted that all folds are derived from origami. Although folded plate models
for roof spans are well implemented, complex variations in architecture and structural
engineering have not been very popular, probably due to demanding geometries and
complex requirements of the design layout.
Before proceeding further it is desirable to introduce a basic kind of fold which has been
exploited by spatial designers. A figure 9 describes the reverse fold which can be made in
two ways and the spatial position is controlled via the angle shown.


Figure no 2.7.1: Reverse Fold
43

This fold technique can be used in several ways to map a given spatial configuration.
In a study by Buri (Buri 2010), he illustrates how various surfaces can be mapped by
using a cross sectional profile and a corrugation profile.
The figure below illustrates the application of this fold to a corrugated profile to create
frame geometry where the cross sectional profile is shown in red and blue zig-zag
represents the corrugation profile. The two ways of applying the fold creates two
different configurations.


Figure no 2.7.2: reverse fold applied to create a frame
44


Figure no 2.7.3: Origami frame for the chapel of S.t loup.
The cross sectional profile can be modelled over various curved profiles to create barrel
shaped vaulted geometries, example figure no 2.6.9.
In the patterns thus produced, each vertex is a four valent vertex i: e there are four creases
meeting at a given vertex. If the distance between the apex and the next base is collapsed
to zero we get a changed pattern with a six valent vertex as demonstrated in figure no
2.7.4.


Figure no 2.7.4: changing the pattern

45

This also has certain implications on the depth of the plates. Figure no 2.7.5 shows the
side view of the pattern, from the gable direction, where one can define a term called the
amplitude of the corrugations denoted by A.
When this amplitude acquires the maximum available value, we change the pattern and
the valence of the vertex with an increase depth of the plates.

Figure no 2.7.5: Changing the Corrugation Amplitude.

The pattern changes from the hexagonal to the diamond pattern. Figures no 2.7.6 show
certain S shaped curves modelled with a hybrid pattern formed with the hexagonal and
diamond pattern. As the number of divisions increase, the curve is modelled more
closely. If we change the sectional profile to vary, we can map curvatures of double
curvature and hence a lot of flexibility can be achieved by manipulating the sectional and
46

the corrugation profile. Figure 2.7.8 show two different, maximum amplitude patterns
modelled over varying section and varying height.


Figure no 2.7.6: S shaped curves.


Figure no 2.7.7: barrel vault with diamond pattern.
47

Figure no 2.7.8: Profiles with varying section and heights with diamond pattern.

. Figure below shows a barrel vault made from the herringbone pattern.

Figure no 2.7.9: Miura Ori vault
The crease pattern can be interpolated from the space configurations by taking
appropriately scaled projections on the plane below. This gives an easy way to fabricate
the sizes of the panels required for the implementation of the design.


48

2.7.1 Radial origami:
Although until now we have produced geometries of double curvatures, slightly different
configurations are achieved by using the radial form of the above stated patterns.
Axisymmetric domes of revolution can be obtained by folding the radial patterns. Figures
below show the radially folded diagonal pattern and miura ori pattern.



Figure no 2.7.10: Radially folded yoshimura pattern and crease pattern.(dashed lines
are valleys)
49



Figure no 2.7.11: radial miura ori pattern and crease pattern.

By changing the inclination of the diagonals in figure 2.7.10, we produce geometries with
varying curvatures or the climb angles of the pattern. It suggests that by varying the
angles different geometrical configurations can be approximated.
We can only approximate the curves since these surfaces are not developable.
Figures no 2.7.12 and 2.7.13 show the radial diamond pattern and its variation. The first
dome can be folded along its circumference, like a curtain while the second dome can be
50

folded by applying a twisting moment at the top. Notice the difference in the seam
arrangements of the variation.


Figure no 2.7.12: radial diamond pattern.


Figure 2.7.13: twist foldable diagonal pattern
In an architectural study (Mitra.A 2008-09), the author has tabulated a list of various
architectural geometries that can be mapped by various arrangements of crease patterns.
51

Several interesting shapes have been made possible and the figures below are few
important illustrations taken from the study.



Figure 2.7.14: Radially folded hexagonal cone.



Figure no 2.7.15: Groin vault approximation with diamond pattern


52

Various interesting shapes can be achieved by combining the radial pattern with others.
Figure no 2.7.18 shows a combination of radial and barrel diagonal pattern.

O.Tonon in his paper geometry of spatial folded form describes the formation of different
shapes using the combination of the above patterns. His approach is somewhat different
as he begins with the crease pattern in 2D state. Figure no 2.7.19 shows a picture taken
from his study.(O.Tonon 1993)
He remarks about the rigidity of the pattern with different sectional profiles. As the
profile closely models the curve, the rigidity begins to decrease and vice versa. There is
an intermediate form where the rigidity is proper.


Figure no 2.7.16: combination of radial and cylindrical pattern.


53


Figure no 2.7.17: Double curvature shapes (source Tonon)











54

2.7.2 Parametric variations in origami:
Origami designs as manifested in 3D space are quite sensitive to variation in parameters
in the 2D crease pattern. As a result various shapes are produced by changing the
parameters of the crease pattern. These parameters can be anything like the shape, size,
orientation of the quadrilaterals, the crease orientation etc.
Figures below illustrate certain examples as applied to folded tubes and conical shapes
produced from the radial pattern. Figures 2.7.20, 21, 22 show variations in the diamond
pattern and the rest show the same for the hexagonal pattern. The parameter varied is in
parentheses.

Figure no 2.7.18: Radial Diagonal Pattern (orientation of quadrilaterals)

Figure no 2.7.19: Radial Diagonal Pattern (crease alignment)
55


Figure no 2.7.20: diagonal pattern (crease orientation)

At times it becomes quite puzzling to classify the pattern, especially while considering
parametric variations as Yoshimura or diagonal pattern as they are very similar.


Figure no 2.7.21: Radial hexagonal (orientation of quads)
56


Figure no 2.7.22: Hexagonal (size, orientation)

In the above figure the quadrilaterals have to be twisted to accommodate the shape. They
will not be planar.



Figure no 2.7.23: Hexagonal (shape of quads)



57


Figure no 2.7.24: Hexagonal (shape of quads)


Figure no 2.7.25: Radial Hexagonal (orientation of crease)
Various other shapes can be realised and a complete list can be found in the study. These
tubes have larger applications in robotic systems.
Origami structures, as they suggest lend themselves to parametric analysis and design.
For any given architectural functionality requirement a number of forms can be derived
by choosing to vary the parameters that may govern the design. For example in a frame
design we can choose to vary the corrugation amplitude, corrugation profile or the cross
sectional profile to produce a variety of different forms.
58

Figures below illustrate an example of a frame, showing the original sectional and
corrugation profile and few of the variations.


Figure no 2.7.26: origami frame



Figure no 2.7.27: parametric variation A


59



Figure no 2.7.28: parametric variation B

Naturally we are faced with a problem of selecting the best possible configuration that
serves our purpose most efficiently. Little literature is available regarding the design and
analysis of these kinds of structures and a general framework is still lacking.
Efficiency is usually described as the strength to weight ratio of a structure. Origami
structures are an arrangement of plates that are stiff due to their orientations and folded
plates are considered a lightweight solution with a high span to depth ratio.
But the structure has various hinges at odd angles and as it may be quite intuitive that the
fabrication of hinges will play an important role in the design of these structures.
Change in parameters most often changes the plate configuration as well as the hinge
characteristics (the angle and their number). For example in the above example if we
choose to use the maximum amplitude of corrugations, we will deepen the plates,
increase the total length of creases and increase the valency of any joint. Additionally, if
we flatten the corrugation amplitude, we will flatten the plates and change the angle of
hinges. Which one of these will produce a more efficient design is not very easy to judge
by intuition alone.
However certain obvious parametric changes can be readily seen and implemented to
strengthen our design. Figure below shows a barrel vault with a sharp variation in
60

corrugation profile as compared to the inner part. This produces plates with larger
inclination, hence stiffer. This helps minimise the deflection observed at the edges.
Also compare the total no of hinge lines and their load bearing characteristics. The miura
vault seems to be heavier as it has finer corrugations. But then the load carrying
mechanism is different and hence the load to be borne at hinges also differs.


Figure no 2.7.29: Yoshimura and Miura Ori vault with parametric variation.






61

2.7.3 Some Notable Examples:
Figures below depict some successfully implemented folded plate designs. The list is not
exhaustive but several different types of ideas are captured.
Figure 2.7.32 shows a concrete folded plate building The construction on site of the
structure was complicated and needed time consuming guidance of the designing
engineer. The sheet thickness was just 7cm. Yet, this approach enabled a material saving
construction.



Figure no 2.7.30: Concrete folded plate structure in Neuss-Weckhofen, Polny/Schaller, 1969

In contrast, Figure 2.7.30 shows the IBM Pavilion which was assembled and
disassembled various times for a touring exhibition in twenty European cities. In this
example folding represents a practicable light-weight construction principle. (Trautz. M
2011)
62

Figure no 2.7.31 shows yet another deployable structure based on the yoshimura pattern used for
a travelling fair. The crease lines have been replaced by linkages to produce a truss like
mechanism.

Figure no 2.7.31: IBM-Pavilion, Renzo Piano


Figure no 2.7.32: Deployable structure based on yoshimura pattern

63


Figure no 2.7.33: The Yokohoma cruise terminal, Japan.

Figure above shows the modern Yokohoma cruise terminal, Japan which used a hybrid
structural system of steel trussed folded plate and concrete girders especially adequate in
coping with the lateral forces generated by seismic considerations.
Figure no 2.7.33 shows an origami structure made of cross laminated timber panels. The
joints are fabricated from nail using steel plates and several prototypes are studied in the
study (Buri H 2009). The roof plates are 60mm thick and span 9 meters (slenderness of
1/150) while the wall plates are 40mm thick. The enormous efficiency of folded plate
structures is evident by this very high slenderness. Figure 2.7.34 shows a concrete folded
roof which uses the principles of origami. Figure 2.7.36(a, b) shows a Miura Ori shell
made from timber plates which have been stiffened at the folds from inside.


64


Figure no 2.7.34: the chapel of St Loup. (Buri)



Figure no 2.7.35: Assembly hall, university of Illinoi.
65



Figure no 2.7.36(a,b): Miura Ori shell structure with stiffened plates, Germany.



66


2.8 Rigid Origami:
Rigid origami is when the design is realized through rigid, thick panels and not
necessarily developed from a sheet. It is essentially a plates and hinge model for origami
where plates do not stretch or bend and there is synchronized motion between all the
plates. A classic example is the everyday shopping bag which can only be folded by
bending the facets and not otherwise. Rigid foldability deals with the question of
existence of a route from an unfolded state to a final folded state.(Naohiko and Ken-ichi
2009)
For architectural application the main problems encountered are to accommodate the
finite thickness of plates. It is especially nice if we develop a 1DOF model so that the
deployment can be semi-automatic.

2.8.1 Kinematics:
In order to apply kinetic rigid origami to various architectural and other engineering
purposes, one must consider the geometry of the plates in motion and provide
generalized methods that produce controlled variations of shapes that suit the given
design conditions.
Currently most deployed constructions are based on the use of textile materials as the
deployable element or completely rigid building elements, which can be removed
entirely. Deployable structures using folded plate constructions are rarely realized despite
the fact that it is possible to create wide variety of high performances structures with
enclosing and formative character. The articulated design of the folds allows the structure
to provide kinematical properties. Thus a structure can be designed which combines the
advantages of folded plate structures with the possibility of a reversible building element
through folding and unfolding [(Trautz.M 2009)]
67


Several designs of rigid-origami structures have been proposed from around 1970s. For
example, the developable double corrugation surface (DDC) Miura-ori pattern. Resch and
Christiansen have proposed a kinetic plate mechanism, folded from a planar sheet that
forms a three-dimensional existing in two different configurations, one with and other
without curvature.(figure 2.8.1).(Resch.R (1971))
This is particularly useful if we are looking towards morphing structural forms to allow
modulation of light, acoustics etc. A number of such patterns exist and are studied under
the context of origami tessellations which can exist in two as well as three dimensional
states. An engineering approach based study can be found in (Shenk.M 2008).

Figure no 2.8.1: The Resch pattern
.
Rigid origami transforms in a synchronized motion based on multiple non-linear
constraints and the design of rigid origami is not a trivial problem given by an arbitrary
design approach without geometric considerations remarks Tachi who has developed the
68

rigid origami simulator which can allows us to simulate the motion of an origami and
detect collisions. (Tachi.T (2009))
As for every deployable structure, the loads bearing capacity is reduced due to the
influence of the characteristics resulting from the associated structural design. The
principle load bearing characteristics of folded plate structures normally include two
aspects which cannot be generated.
One is the bending resistance between the plates in the area of the fold. Since the hinges
are articulated and we cannot allow any rigidity, this characteristic is disabled. The
second aspect results from allowing translational hinge deformations depending on the
chosen folding pattern. Most often, hinges will not allow adequate transfer of shear forces
in the folds or unbalanced live loads on the structure, which will reduce the load bearing
capacity.
The deformation of the folded plate structure under dead load results in a self-
deployment and therefore in further hinge translation. The figure below illustrates the
difficulties encountered at a typical hinge in a quadrilateral based design of rigid
origami.Rigid origami can be realized through triangle based or quadrilateral based
design approach where the former are more flexible.(Tachi.T 2010). Figure no 2.8.3
illustrates this fact via hypar designs which cannot exist in these states if they are not
triangulated.


69

Figure no 2.8.2: Four fold mechanism and a detailed view of the hinge

Figure no 2.8.3: Triangle based design allow more flexibility

The problem of hinges has been dealt by Tachi and he proposes several articulation
techniques to allow for realizing thick origami with finite thickness. (Tachi.T 2011).
Figure 2.7.42 shows few proposals to accommodate finite thickness rigid origami where
the hinges are put on the valley side and the edges are trimmed to accommodate folding
motion.



70

Figure 2.8.4: Proposals for Rigid Origami
One of the latest studies involves exploring the use of vacuumatics to allow for the
complex and demanding stiffness characteristics of the hinges to allow for deployment
and strength. (Tomohiro Tachi 2011).This way a number of flexible configurations may
be possible by varying the boundary geometry and support conditions. The basic idea is
that of a double membrane structure where the compression forces on the valley side can
be manipulated via a vacuum. This varies the stiffness of the structure. This compressive
force is balanced by the tension force on the hinges due to external load.
Figure 2.8.5 illustrates such a hinge. The valley side of the hinge is filled with inflatable
material. When vacuumed, moment is generated this aids folding and can be controlled to
stiffen the hinge.


Figure no 2.8.5: fabrication of hinge using vacuumatics.

However an interesting development is the fabrication of origami inspired design of
folded plate geometries with using cross laminated timber panels. However only static
properties of these forms have been used where the main concern will be to design
efficient connection methodologies based on numerical studies.(Haasis. M 2008)
The problem is again not trivial as it requires calculating the offset based on the angle and
thickness into the cutting machine.
71

The process is cumbersome if the angles and thickness vary.

Figure no 2.8.6: cross laminated timber panels



2.9 Modular origami:
Modular origami is about building shapes using a basic origami unit as a starting point.
These manufactured folded modules are built into spatial forms. The form variety reaches
over simple platonic bases such as cubes and tetrahedron up to complex polyhedrons.
The advantage is that the total form is always based on basic geometrical form and is
mathematically recordable. The number of different modules and detail connections are
limited here and are therefore organized for planners and engineers.
Rigid as well kinetic structures are possible in modular origami and several domes have
been realized.


72



Figure no 2.9.1: modular origami (rigid and kinematic)
2.10 Curved crease:
Curve creases on paper also show promising possibilities in the domain of designs of
structures from paper models. Curved folds are characterized by larger DOF than regular,
straight line origami. Folding flat sheets of paper at a small scale along curved creases
results in shapes with remarkable strength and stiffness properties yet their deployment
trajectory is not understood. Curved folding is an advantageous product of folding and
bending: the surface consisting of developable surface patches and a relatively small
number of separate curved creases can have a single degree of freedom only.

Figure no 2.10.1: curved creased domes
73

2.11 Behavior under Load:
Figure below illustrates forces occurring in a folded plate system. The structural
condition of the hinge, bearings and the angle of inclination of the plates play an
important role in determining the force response.
The downward force on the plate can be resolved into a force perpendicular (W1) and
parallal (W2) to the plane of the plate. These are resisted through slab action and in plane
plate action which are generally coupled. The behavior changes from a pure bending
action to a pure membrane action as the angle changes from 0 to 90.


Figure no 2.11.1: Forces on folded plates

Figure no 2.11.2: Deformation of vertex

74

Let the total vertical deflection be
A
with components
1
A
and
2
A
. Since the plate is very
rigid for in plane actions,
2
A
is small and correspondingly
1
A
is also small.
Consider the pure bending action where a reaction force is obtained at the hinge aligned
with W1 but in the opposite direction. This reaction has a component R2 which is resisted
by in plane action so the coupling is evident.

Figure no 2.11.3: Coupling

The hinges encounter shearing forces and transmit bending moments from one plate to
another. Ideal analysis will incorporate a rotational stiffness and a translational stiffness
at the hinge.
If the joints are rigid and deflections are small, a linear elastic solution under the
assumption of superimposition to hold can be used to arrive at an exact solution for
simply supported plate system using an elasticity formulation.
The method assumes four displacements and four forces at each edge of the plate which
are analyzed individually. The figure below illustrates the displacements (three
translations and a rotation) and the corresponding forces for the system shown in the
local element coordinates.

75


Figure no 2.11.4: Forces and displacement at plate edges


The displacements for the longitudinal direction(x) can be represented by a half range
Fourier series as shown below. These confirm to the boundary conditions and the analysis
essentially becomes one dimensional.

0
( )
( , ) ( ) cos
m
m
m x
u x y u y
L
t

=
=

0
( )
( , ) ( ) sin
m
m
m x
v x y v y
L
t

=
=

0
( )
( , ) ( ) sin
m
m
m x
w x y w y
L
t

=
=

These are related by the equilibrium equations for plate bending and plane stress as:
76


1)
4 4 4
4 2 4 4
2 0
w w y
x x y y
c c c
+ + =
c c c c

2) 0
xy
x
x y
t
o
c
c
+ =
c c

3) 0
y xy
y x
o t c c
+ =
c c

Where

2
( )
1
y
E v u
y x
o u
u
c c
= +
c c

2
( )
1
x
E u v
x y
o u
u
c c
= +
c c

( )
2(1 )
xy
E u v
y x
t
u
c c
= +
+ c c

E= Youngs modulus
u = poisons ratio.

The analysis assumes that the forces are applied at the joints also expressed as a Fourier
series. The equations are solved by equating the in plane translations normal and
tangential to plate edges, the normal translations of the edges and the rotations for
symmetric and anti-symmetric cases in the global coordinate system.
The author demonstrates that the solution thus produced matches well with a finite
element solution obtained by using elements capable of in plane and bending capabilities.
The idea can be extended to folds in both edges as well, although no attempt as such was
found available for the elasticity method.(Hassan.A 1971)


77

Figure no 2.11.5 illustrates two different types of prototypes, the first folded from metal
sheet where the connections have a finite rigidity along their length. This is because the
behaviour of material at a fold is like a cylinder of very small radius in compression. It is
therefore difficult to crumple a sheet of paper beyond a certain limit due to the buckling
force required to break a fold increases with diminishing length of fold.
In the second type, timber plates are joined via nails and longitudinal and transverse shear
forces are resisted by the nails at the connection. Their behaviour and failure mechanism
are also different. The overall behaviour of a folded plate structure and the choice of a
failure mechanism depend on the distribution of stiffness of the folds and the plates and
the type of connections, if any.


Figure no 2.11.5: Two different prototypes(source Buri)
The results for the metal prototypes are not available but the timber prototype was studied
via the finite element method.
A test model of a single hinge was loaded to find the stiffness against rotation which was
fed into the FEM procedure. It was found that the prototype was much more flexible than
the computer model.
78

Moreover three different softwares used showed different results for hinges with
rotational freedom but two of them aligned for rigid hinges suggesting that analysis for
rigid hinges is more reliable.
Figure below shows the prototype rupture. The folds have opened up indicating that the
joints needed strengthening.



Figure no 2.11.6: failure of hinges

79


2.12 Axisymmetric Domes:
Axisymmetric domes are a class of spatial structures which are generated by revolving a
curve about an axis and they are also called shells of revolution.
Dome structures have a long history and the earliest occurrences can be found in parts of
Greece, Egypt, China and India. These structures have served as roof spaces for places of
worship and were an inherent part of medieval architecture. Early designs utilized cut out
stone, sun dried mud, for their construction.
In modern days these shapes are usually utilized for storage, funerary, and utilitarian
purposes such as defense, storage, kilns etc. these are robust shapes but modern
architectural forms have seen several variation such as lattice formed, geodesic shapes
etc.
Elementary load carrying mechanism may be understood by visualizing the shape as a
number of arches in different planes. There is a horizontal thrust at the base which is
taken care of by providing a ring which works in tension. The structure then only
transmits vertical reaction forces. (Figure2.12.2 b). Figure 2.12.1 below shows typical
forces under equilibrium analysis in a spherical dome under symmetric loading. These
are membrane forces, assumed to act in the middle plane of the shell. The meridian forces
are compressive all along and increase as we move down and the hoop forces are
compressive at top and tensile at the bottom. For a hemisphere the angle of change is 51.5
degrees. The forces on a differential element and expressions for the meridian and hoop
forces are shown in the figure
Full shell analysis is rather complex and the reader is referred to (CR 1983)but under
large curvature the surface of the shell exhibits rigidity towards in plane stretching and
usually these forces are dominant if the thickness to radius ratio(h/R) is < <1. (Eduard
Ventsel and Krauthammer)and under suitable boundary conditions as shown in figure
below (a) where the shell acquires inextensible rigidity as it mimics the condition for
80

closed surface. In addition some localized moments do occur near the edges(c) which are
taken care of by providing additional thickness (figure2.12.4). However, the response of a
shell to localized forces is by bending which can cause significant bending stresses.

Figure no 2.12.1: membrane forces in a hemispherical dome


Figure no 2.12.2: boundary conditions
81

1 cos
aq
N
|

=
+

1
( cos )
1 cos
N aq
u

|
=
+

Figure no 2.12.3: element equilibrium

Figure no 2.12.4: variable thickness for an arch and a shell

Historical methods of analysis are not available but the ancient builders had a command
over the forces and geometry. Studies available regarding the investigation of
construction techniques of the domes of Cairo suggest the ancient designs are indeed
quite elegant for example the figure 2.12.6 depicts the dome of the Mausoleum of Farag
Ibn Barquq which has stood for over hundreds of years. The dome is less than 15 inches
thick and spans 47 feet. (h/R=0.02) and sits on a 25 feet cylindrical wall. The dome has
no reinforcement despite the fact that modern analysis predicts it to be unstable without
tensile reinforcement.(.W.Lau)
82


Earliest reported techniques have been mainly empirical in nature such as graphic statics,
utilizing the force polygon to find an equilibrium solution.
The method works on finding a thrust line in accordance with the geometry and the
weight distribution of the dome via a force polygon construction. The line of thrust is a
hypothetical path over which the internal forces transport the external loads to the
supports. A geometrical construction is demonstrated in figure 2.12.5 where an inverted
cable with distributed weights assumes a form which will stand in compression as
demonstrated by Poleni in 1748. He concluded that if this line lies within the effective
thickness of the structure, it will be safe.
This lower bound approach is quite safe as in neglects the hoop forces which have a
stabilizing effect on the dome.

Figure no 2.12.5: hanging weight model and force polygon for an arch.
83


Figure no 2.12.6: Dome of the Mausoleum of Farag Ibn Barquq, Cairo.

This basic idea has been refined by researchers like Eddy, Heynmenn Wolfe to
incorporate hoop forces obtain upper and lower bound theorems for analysis of masonry
domes.
The methods of construction are illustrated in the figure below, from a study devoted to
the understanding of these domes. Masonry needs to be supported over a formwork
before it can acquire strength. Modern bricks are made from clay and dried at a furnace
before they acquire strength. Masonry can fail in shear by slipping and in tension by
rupture. However it is extremely efficient for compression. Ancient masons had a good
knowledge of forces for they aligned the joints in a staggered fashion to avoid lining the
joints with the flow of force. Modern masonry is usually reinforced by steel by using
especially manufactured bricks which provide grooves and holes in them.

84


Figure no 2.12.6.1: construction methodology(source(Cipriani.B 2005))

Ancient and medieval architecture has withstood the test of time. Figure 2.12.8 depicts a
vault at kings college London where the thickness to span ratio is less than that of an
eggshell speaking of a very daring design.
Figure 2.12.7 shows a typical collapse mechanism for a dome where first cracks occur
along meridians, separating them into lunes (pie shaped arches). Then cracks along the
hoop develop forming a mechanism. The top part bends inwards while the bottom part
rotates outwards. This mechanism suggests that reinforcement elements along the
meridians will be important for a strong design.

85


Figure no 2.12.7: Dome failure

Later designs saw domes to be reinforced along the meridian for example there are
several ribbed domes where the stiffest element is along the meridian. Figure 2.12.9
depicts a medieval timber vault. Figure 2.12.10 depicts a ribbed masonry dome. A large
number of designs have been successfully implemented by meridian reinforced elements.


Figure no 2.12.8: Vaults at the Kings College London

86

Modern versions of domes are articulated lightweight lattice structures which have
elements along meridian and hoop directions and are analysed through computer
programs.




Figure no 2.12.9: medieval timber vault.

87


Figure no 2.12.10: Ribbed masonry dome

In 1951 fuller introduced the geodesic dome which was a paradigm shift in dome
construction. The dome is formed by triangulating the facets of an icosahedron and
mapping them to the surface of a sphere. The points thus lie on the geodesics and this
proved to be a very efficient load carrying mechanism. The triangulation can be done in
two ways as shown and are referred to as class 1 or class 2 subdivisions.
Fuller claimed that geodesic domes built upon principles embodying force distributions,
similar to those of atoms, molecules, and crystals, would form the lightest, most efficient
forms of construction.

88




Figure no 2.12.11: Geodesic hemisphere

While radial domes exhibit greater stiffness for uniform loads, geodesic domes exhibit
larger stiffness for non-uniform loads. The forces in a geodesic network are a
combination of tension and compression, tension forces being global and continuous,
while compression forces are local and discontinuous.(Kubik 2009)
89

Figure below shows a geodesic surface formed with plated vertex and stiffened by
tubular members.


Figure no 2.12.12: Geodesic dome with plated vertex











90

2.13 Optimization of Shell Structures:
Shell structures are known to be extremely parameter sensitive. even small changes of
the initial design, e.g., to the shape of the shell, may drastically change the internal
stress state or an initial imperfection may significantly affect the buckling load.(Ramm
1993). For a concrete shell the ideal case is a state of pure membrane state of
compression. In many situations where the ideal form is not obvious, form finding
methods are employed to arrive at a form to suit a required optimal condition.
A typical problem of structural optimization is characterized by an objective function
f(x) and constraints g(x) and h(x) which are non-linear functions of the
optimization variable x.
It can be stated as:
Minimize: f(x)
Subject to: h(x) = 0; g(x) = 0
Because of their general formulation, methods of structural optimization can tackle
problems with many load conditions, arbitrary design objectives and loads such as
changing boundary conditions and forces.
For example strain energy nay be chosen as the minimization function for shapes that act
in membrane state of stress i:e compression and tension but no bending.
Where the total strain energy given by:

1
( ) . .
2
v
F x dV o c =
}

Where G(x) may represent a reliability constraint such as stress or displacement limit.
If a leveled state of stress is desired, the function may be chosen as

2
( ) ( ) .
a
v
F x dV o o =
}

91

If we want to maximize stiffness for a given mass,
( ) h x
may be used to represent a
constant mass otherwise unrealistic heavy solutions may be obtained. This is usually the
case when external loads are dominant.
The minimization function generally can be anything from a cost function, weight or
natural frequency or any parameter of interest.
Depending on the objective (strain energy, weight, etc.), the constraints
(equality, non-equality) and their combinations, the optimization problems can vary
from totally unconstrained (stress leveling) to semi constrained (strain energy
minimization with fixed mass, displacement limit) to highly constrained problems
like weight minimization, which tend to reduce mass until the limit of material
resistance is reached.
Sophisticated computer methods are employed as they involve extensive search
algorithms, structural analysis FEM, and design modeling (CADG) to work together.

Figures below illustrate an example taken from the study. The initial configuration is a
parabolic shape shell( uniform snow load of
2
5 / KN m
and hinged supports) and the
various forms it achieves under optimality of different conditions.
2
s
and
1
s
are the
heights at the middle and edge cross sections respectively.
Various conditions that were simulated are:
b) Strain Energy Optimization with fixed thickness and no stress constraint.
c) Stress leveling (
2
100 /
a
KN m o =
)
d) Weight minimization with constraint on maximum von- mises stress(400 KN/m2)

92

Only linear material properties and geometry formulations were considered but Nonlinear
relations can also be implied and the subject is under much research.



Figure no 2.13.1: Structural Optimization.



93

2.14 Conclusion:
Origami inspired structures have been few despite the fact that a number of spatial forms
are possible using the patterns overviewed in the review. The major shortcoming may be
the demanding geometry and fabrication of elements and the problems associated with a
realistic analysis. Except for the pleated corrugation, which has been used for concrete
roofs, more elaborate designs with double curvatures have not been successfully
implemented although research and work towards the same is ongoing.
In this dissertation I will aim to investigate the rationale behind producing a design to
cover a three dimensional dome structure for commercial purposes.
The major considerations for any design would be the choice of pattern, and the spatial
form most suitable for it. Next it would be to research into the effect of the parameters
which control the rigidity of the structure and which configurations would be the most
suitable. This requires a numerical analysis and immediately we face the problem for a
realistic hinge design. If we are to keep the thickness of the plates different for different
levels, as the forces in the dome suggest, it requires us to test several connections.
Although it is not possible to undertake a realistic analysis with a prototype in this study,
it definitely is worthwhile to study the sensitivity and trends for a particular design
variable of interest under the assumption that the reflection will provide an initial
judgment of the nature of forces irrespective of the hinge characteristics.

Question:
What crease pattern and spatial form would be best to realize a dome shaped structure?
Further how would the parameters influence the rigidity?


94

CHAPTER 3
DESIGN OF EXPERIMENT
Two main issues addressed will be the choice of pattern with its geometry and design of a
suitable experiment to study the effect of parameters on the structural properties that
control the pattern such as deflections, state of stress, internal forces etc.

3.1 Choice of pattern:
As understood, under conventional arrangement of material, the states of internal forces
comprise a meridian and hoop component. By conventional arrangement, it means that
we are not looking towards unconventional seam arrangements, such as geodesic
patterns, of twisted layout of seam. Although these may be efficient but as a start if was
decided to stick to conventional pattern.
Figure below shows an efficient shear connection which transmits the shear via an axial
force in the steel triangles. The miura- ori pattern as understood by the layout of plates
will form an efficient design as it will mimic such an arrangement. The Yoshimura
pattern on the other hand will align the hinges along the meridian which unless folded
from a metal sheet, does not seem very efficient as such.
All the patterns form a closed polyhedron which are infinitesimally rigid on account of
their configuration and only deform due to material stretching. Miura ori has a four valent
vertex and is a quadrilateral based design which seems stronger with regard to the loading
and the detailing required at a six valent vertex.
95


Figure 3.1: Transfer of forces in a shear connection and the miura ori dome.

Lightweight plated structures can be realized if the load path is mainly through in plane
action and miura-ori pattern, due to its flexibility was given preference over other
patterns and was chosen to be studied further.








96

3.2 The Miura - Ori radial pattern: Relationship between 2D and 3D
form.
The pattern was folded from paper to identify the parameters and model the layout of the
plates. Figures below show the crease pattern and the folded model.
Three parameters that create the geometry were identified in the folded forms which
completely define the model. These are:
a) No of Vertical discretizations (meridian).(Nv)
b) No of Hoop discretizations in circular plan.(Nh)
c) The position of the apex (as a percent rise of the arc length from the hinge
lines).(p)
It was observed through paper models that the ends and apex of each successive triangle
lie on a plane which is formed by rotating the x-z plane around the z axis by a certain
constant angle, like the slices of an orange. (Fig 3.2.1). The planes ZOA, ZOB contain the
base and the apex of the triangulations respectively. Lines Ha, Hb, Hc show different
configurations of parameter p. Green dots lie on hinge lines and red ones lie above, on or below
indicating the value of p>0, p=0, p<0 respectively. The blue shaded plates arise from the pattern.
Brown shade represents boundary conditions.
The typical crease pattern for such a dome is shown in figure 3.2.2. It consists of a
polygon of sides 2Nh where Nh is the number of hoop discretizations. Mountain and
valley fold instructions are indicated in figure 2.7.13. In the figure Nh=12 and Nv= 4,
and P is constant. The parameter P dictates the angle A marked in the crease pattern. The
angle marked u is determined by mathematical underpinnings of origami. The points A,
B, E, F lie on the surface while C, D lie inwards. The angle marked | depends on the
number of hoop discretizations. At the boundary the last plates will have to be extended
to meet the ground.

97


Fig 3.2.1: Radially folded hybrid dome showing meridian and hoop
discretizations.(Nv=8,Nh=12)

98

Fig 3.2.2: Crease pattern for miura ori dome.
To model the pattern over a spatial geometry one must start from the 3D state and work
backwards towards the crease pattern. Detailed procedure is explained later.
Apart from these the thickness of plates can also be an important parameter, but to assess
the variation of plate thickness along the height of the dome is not trivial and requires a
parametric approach to modeling which complicates the process and hence it was decided
to stick to constant thickness model.
This greatly reduces the number of models to be considered. It seems quite fair since
there is little evidence to consider the contrary.













99

3.3 Design of Experiment:
Our basic intuition tells us that if an arch is replaced by a triangulated framework, it will
be more rigid i:e deflect less as the load carrying mechanism becomes more efficient. The
nature of stresses will also change as now we will see stresses corresponding to axial
deformations. As we increase the no of points the rigidity will degrade till we replicate
the arch itself.
Does it hold true in the realm of plated shell structures?
The forces in a shell are sensitive to the form and hence the stress in an origami dome
will be sensitive to the parameters which dictate the form. Sensibility tells us that we
should get larger rigidity with lesser discretization and vice versa. What about weight and
thickness which totally lie at the disposition of origami mathematics? To realize such a
structure one must have an idea of the variation of geometrical and structural
characteristics both and arbitrarily applying just any configuration may not be suitable.
However quantifying these aspects can only be done with experimentation. The
arrangement of plates is complex for hand calculations so a numerical experiment using
ANSYS is devised to test:
- The influence of spatial form on design variables.
- The influence of parameters on design variables.
and whether the existing notions of rigidity still remain sensible.
Two main design variables to be studied were identified to be the maximum
displacement, the von -mises stress which is an indication failure criterion and also
Efficiency which is the measure that takes care of the strength as well as the weight. This
seems necessary as we cannot disregard weight that is also a variable and changes with
other parameters.

100

Study variables were the form and the pattern parameters which are explained below.

3.3.1 Range of parameters:
The range of the parameters has been kept limited via practical considerations.
1) Form: Three popular forms were chose namely, Hemispherical, Parabolic and
Catenoid of equal height and span. This was because the hemisphere has equal
height as well as span and in order for a fair comparison, the other geometries
must also have the same dimensions. A large radius of 25m was chosen.
2) Nv :( number of vertical discretization): These are studied between 5 and 20 as
this range would be sufficient to capture the trend.
3) Nh: (number of hoop discretization): These were studied as 12,16,20,24
discretizations in circular plan.
4) P: This is a continuous parameter and is discretized before studying. Only one
value on either side of the middle (p=0) is studied i:e P=20%,P=-20%
Since a large number of configurations are possible by choosing different permutations of
parameters and spatial configurations it has been assumed that the trend in variation of
parameters is same for different ranges i:e the trend for stresses against vertical
discretizations either,10,15,20 will be same for all range of horizontal discretizations.
.
Design variables: Von- mises Stresses, displacement, Efficiency.
Study variables: Geometrical Configuration, Pattern parameters (N (h), N (v), p)



101

The experiment is conducted in four stages, each one being independent of the other to
establish relationships between the design and study variables.
Efficiency is defined as the ratio of strength to weight. Strength is inversely proportional
to the maximum stress developed as failure is defined when the first plate reaches its
limit stress. Therefore, Efficiency is accounted for as

max
*
k
Efficiency
W o
= where W is the weight and
max
o is the maximum stress. K is
proportionality constant.

The experiment takes the form:

Figure 3.3.1: Experimental Procedure





INPUT: study
variable
ANSYS(numerical
experiment)
OUTPUT:Design
variable
102

The various stages can be summarized in the table below
Experiment(objective) Study variable(Input) Design
variable(Output)
Suitable form Form(geometry)[Nv=Nh=20,P=0]
(radius=25m)
Efficiency
Relationship b/w stress,
disp and Nv
. Nv(5,10,12,15,20)
[Nh=20,P=0]
Von-mises stress
field, displacement
field, Efficiency
Relationship b/w stress,
disp and P.
P (-20%, 20%)
[Nh=20,Nv=10]
Von-mises stress
field, displacement
field, Efficiency
Relationship b/w stress,
disp and Nh
Nh(12,16,20,24)
[Nv=10,P=0]
Von-mises stress
field, displacement
field, , Efficiency
Table 3.1: Experiment schedule

In testing the relationship between one variable, all other variables will be kept constant
(shown in square brackets) so that the observed differences are purely attributable to the
parameter studied.
Each part of the experiments is accompanied by geometrical results and a FEM result.
Before proceeding with the experiment, it is required to solve the geometry of the pattern
which is explained below.


103

3.4 Geometric modeling of the Miura-Ori pattern
The shape to be modeled is discretized in the number of required divisions in the
meridian and hoop directions and the pattern is fitted into the segments by well-defined
equations which are derived below.
Consider a typical portion (ABCDEF) (fig 3.4.1) of the crease pattern which forms two
planes in the final configurations. As a start we will focus on the problem of finding the
position of these planes in terms of their coordinates in 3D space.


Figure no 3.4.1: Trapezium





104

Due to symmetry of the pattern, CEDF form an isosceles trapezium where:
1) CE=BD
2)
EDC DCB Z =Z
(parallel angles)
3)
90 CBD EDF o u Z =Z = =


Now consider the position of this unit as it is mapped onto the 3D configuration.(fig
3.4.2). Since the points A, E, B, F form the ends of a triangle, they lie on two separate
plane with A, E and B, F being coplanar.
The point C lies on the plane through A and E, and the point D lies on the plane through
B and F. Since CD is the fold line and EFCD and ACDB form two rigid planes the
essential conditions we get are:
1) CE=BD. This implies that ED and CB are two parallel lines in space hence the
parallel angle condition is duly satisfied.
2)
90 CBD EDF o u Z =Z = =
. This condition is satisfied as the fold does not
alter this in any way.
3) The points A,C,D,B lie on a plane.
4) The points C,E,D,F lie on a plane.

The above conditions are all the conditions we have about the pattern. This implies that
we can solve a four parameter model at max. i:e we can have a maximum of four
independent constraints that govern the spatial configuration of these two planes.
Now since the geometry that is being modeled requires the hinge lines to lie along the
discretizations, we can fix the position of the points A,E.
105

If we fix the parameter p (position of the apex) we can fix the position of the points B and
F. So now the problem essentially is to find the position of two points in space C,D which
satisfy the four conditions stated above.
The parametric position of a point in a plane is characterized by two unknowns and hence
we have four unknown and four equations and the problem is determinate.
Consider the parametric position of a point P in a plane formed by the plane through 3
points A,B,C..

{ , , } (1 ){(1 )( , , ) ( , , )} {( , , )}
P P P A A A B B B C C C
X Y Z s t X Y Z t X Y Z s X Y Z = + +

Where s, t are the two parameters for the point P.
Similarly, placing the origin of coordinates as shown (fig 3.4.2) for convenience we can
assume the position of C as:
{ , , } (1 ){(1 )(0, 0, 0) (0, 0,1)} {(1, tan , 0)} { , tan , (1 ) }
c c c c c c c
X Y Z s t t s s s s t u u = + + =

Similarly, for D:
{ , , } (1 ){(1 )(0, 0, 0) (0, 0,1)} {(1, tan , 0)} { , tan , (1 ) }
d d d d d d d
X Y Z s t t s s s s t u u = + + =

Points, A,E B,F are known but it helps for the formulation of equations to write them as:

{ , , } { , tan , (1 ) }
a a a a a a a
X Y Z s s s t u =

{ , , } { , tan , (1 ) }
b b b b b b b
X Y Z s s s t u =

{ , , } { , tan , (1 ) }
e e e e e e e
X Y Z s s s t u =

{ , , } { , tan , (1 ) }
f f f f f f f
X Y Z s s s t u =


106



Figure 3.4.2: spatial position of the plates. The angle marked is on the axes is and on the
plates is alpha.

Formulation of equations:
1) Planarity on ABCD.
{ ( )} { ( )} { ( )} { ( )}
a b b a b a
u s w s v s s x A w x s B x s w + = +
(1)
2) Planarity on CDEF.
{ ( )} { ( )} { ( )} { ( )}
e f f e f e
u s w s v s s x E w x s F x s w + = +
(2)

u
107

3) CE=BD
2 2 2 2 2
( ) ( ) sec {( ) ( ) } 0
e b
u E v B s w x s (u) + =
.(3)
4)
90 EDF o u Z = =

2 2
( ){ (1 tan ( )) sec ( )} ( )( )
f e
s x s x E v F v u u +
=
2 2 2 2 2 2
sin( ) {( ) ( ) tan ( ) sec F 2( )( )
e e f
s x s x s x E x F x u u u + (E ) + + +
(4)

Where:
(1 )
a a
A s t =

(1 )
b b
B s t =

(1 )
e e
E s t =

(1 )
f f
F s t =

(1 )
(1 )
c
d
c c
d d
w s
x s
u s t
v s t
=
=
=
=

These are four nonlinear equations in four variables (u, v, w, x) and were solved with the
help of Mathematica.
The coordinates of points C and D can be easily computed from the roots
{ , , } { , tan u}
c c c
X Y Z w w u, =



The crease pattern can be interpolated as it is merely the representation of successive
polygons formed whose lengths can be computed since all points are now known.

{ , , } { , tan v}
d d d
X Y Z x x u, =
108

For the experiment, the chosen parameters were Nh = Nv = 20 and p = 0 and the form
equations were:
1) Hemisphere:
2
25 z x =

2) Parabola:
2
25
25
x
z =

3) Catenary:
15.47cosh 40.47
15.47
x
z
| |
= +
|
\ .



Figure 3.4.5: Arc length discretization of a Parabola on the x (horizontal) and z axis.






109


3.4.1 Results
The results were successfully obtained and it reveals that the distribution of weight is
linear in all the curves. The heaviest configuration is hemispherical and the lightest is the
parabola. It can be seen that the plates ABCD and CDEF are almost identical with respect
to their areas and hence weights. However the exact shape can only be ascertained with
the help of crease pattern. The inner points C, D were seen to lie on a similar curve as
modeled with a constant thickness, 23.6mwhich gives a thickness of 1.4m.this result was
the same for the three geometries.

Dome type Meridian length(m) Total plate area(m2)
Hemispherical
39.25 160.781
Paraboloid
36.973 136.507
Catenary
37.394 141.881
Table 3.2: comparison of the weights

Figure no 3.4.6 and 7 below show the variation in plate areas and angles for the
Paraboloid. The trend is typical for all the geometries. In figure no 3.4.6, Series 2
represent angles between plates ABCD and CDEF (outer angles) as they appear to a
viewer outside the dome. Series 1 represent inner angles. i:e between CDEF and
EFC1D1. In fig no 3.4.7 series 1 and 2 represent plate areas ABCD and CDEF
respectively. Series 3 represents the combined areas. Complete results are available
in appendix 1.

110



Figure 3.4.6: plate angles


Figure 3.4.7: plate areas

After establishing the geometry it was fed into ANSYS to determine the nature of stresses
and deflections.

1 3 5 7 9 11 13 15 17
0
10
20
30
40
50
60
70
80
Series1
Series2
1 3 5 7 9 11 13 15 17
0
2
4
6
8
10
12
14
Series1
Series2
Series3
111

3.5 Finite Element Modeling
A finite element model in ANSYS was made to ascertain the nature of these forces. The
geometry of the plates was solved for as described above and the model was created. The
radial symmetry of the dome allowed only to model a sector.
The plates were modeled as a single four noded element, with six DOF, three translations
and three rotations at each node and the joints were modeled with two scenarios,
completely rigid and free pinned. In reality the joints will have a finite stiffness against
rotation and shear but that needs a joint test which was not feasible in the study. The best
configurations was picked to be used in subsequent experiments.


Figure3.5.1: Plate element capable of bending and membrane action

Element properties: The elements were modeled as concrete plates with the following
values:
10 2
2*10 / E N m =

0.3 u = Poissons Ratio.

112

3.5.1Restraints
Since the plates form a rigid closed system, and the loads are only symmetrical, the
displacements can only be along the radial direction as shown below. Hence the nodes
were restrained to move only in the plane. The DOF corresponding to the radial direction
in cylindrical coordinates was ceased. At the base two cases were tried and are presented
later.

Figure 3.5.2: displacement of a sector

3.5.2 Loading
Only self-weight of the plates was applied. The density of concrete was taken as
2400Kg/m2 and plates were modeled as 80mm thick. Linear behavior was assumed upto
a stress value of 3.0KN/m2.
The plate thickness was adjusted in subsequent experiments to stay within linear
behavior.



113

3.5.3 Results
The following results are for the catenoid. These are generic to all cases.
The base experiences forces as illustrated. A net moment is transferred to the support
below.



Figure 3.5.3 forces at the base







114

The table illustrates the values obtained from FEM models. The stresses and deflections
show considerable increase when the joints are pinned.

Joints free to rotate Joints fixed against rotation

Form Max
stress(10^7N/m2)
Max disp
(cm)
Max stress Max disp
(cm)
Parabola(P) 2.38 10.75 2.79 26.38
Hemis(H) 2.185 11.438 2.38 23.83
Catenoid (C.) 2.37 10.93 2.62 25.54
Table 3.3: Comparison of results for pinned and fixed rotations

The figures below show the contour diagrams for the von mises stress values for the
pinned and fixed cases respectively.


115


Figure3.5.5: Stress fields for rotations free case.(P,H,C)


Figure3.5.6: Stress field for fixed rotation case.(P,H,C)
116

3.5.4 Discussion
The predominant load carrying mechanism is through in plane action in the plates as the
moments observed are maximum upto the range of 30KNm while the forces are higher.
The stresses were all in the positive range indicating the absence of tension in the
structure.
The state of stress is quite different for both cases. Numerical values and ranges are
higher for pinned joints as illustrated in the figures below.
For the pinned case plates A, B, C, D and C, D, E, F are equally stressed and a larger
region is under a higher state of stress while for the fixed case the plates C, D, E, F are
relieved of stresses and the region of high stresses is lower and is towards the base
suggesting that fixed joints allow for greater economy.
There is a region of stress concentration where the last plate C, D, E, F meets the cap
observed in all models.








117

3.6 Boundary conditions:
The figures below show possible boundary conditions. At the top only the one shown
makes sense while at the bottom, there are a number of possible solutions. Figure no 37
shows the range of positions the boundary plates (denoted by point P) can take. Figure 36
shows one such configuration where the end points of plates lie on a circle of radius less
than the maximum i: e 25m. In a sensitivity study via a FEM formulation, various
positions were tested against the maximum stress (von mises) and deflections observed
in the structure. Figures 3.6.3 and 4 show the results for the paraboloid normalized over
the first value. (For R=22.5m)


Figure 3.6.1: boundary condition 1. Dome with p=0. The lower plates bend inwards.

118



Figure no 3.6.2: Range of plate boundaries (P)

The inner radius was fixed to be 22.5 meters and the outer radius was chosen to be 26
meters. The increment chosen was 0.5 or 0.25 m. It would suffice to consider this range
as the behavior was captured with enough accuracy. In figures below, Series 1 represents
the maximum stress and series 2 represents maximum deflection observed.

Case 1 represents the condition when only the translational DOF are fixed and rotations
are allowed while Case 2 represents the condition where all DOF s are ceased. The values
for the first stress (at R=22.5m) and deflections are 2.34*10^7N/m2 and 10.31cm
respectively for both cases.


119

The difference in stresses and deflections is negligible in both cases till around inner
radius of 24 meters after which the rise is larger in the first case. It seems feasible to keep
the boundary at the interpolated inner radius as it suits the geometry and the stresses and
deflections are small.



Figure no 3.6.3: variations in max stress and deflections. (Case1)


Figure no 3.6.4: variations in maximum stress and deflections. (Case2)
0.000
0.200
0.400
0.600
0.800
1.000
1.200
1.400
1.600
1.800
22 23 24 25 26 27
Series1
Series2
0.000
0.200
0.400
0.600
0.800
1.000
1.200
1.400
1.600
1.800
22 23 24 25 26 27
Series1
Series2
120



For all subsequent models, the inner radius was chosen as the position and rotations were
kept free at the boundary.














121

CHAPTER 4
RESULTS AND DISCUSSION
4.1 Experiment Stage 1:
Out of the three geometries, parabola was found to be the most efficient, followed closely
by catenoid and then hemisphere. Although stresses are not greatly different, it is the
weight that makes the form efficient. The results are presented below.

form Max stress Max deflection weight efficiency
parabola 2.38 10.75 273.014 0.153k
catenoid 2.4 11.13 283.762 0.146k
hemisphere 2.185 11.438 321.56 0.142k
Table 4.1 Results stage 1

Further studies of parameters were undertaken for the Paraboloid.
1






1
The values of efficiency are not comparable within different stages of experiment as the best order and
round figure has been chosen for representation.
122

4.2 Experiment Stage 2:
The plates were modeled as 80mm thickness. The results are shown in the table below.
The graphs are plotted for values normalized against the first. Here Nh=20,P=0.

Nv Max
stress
(N/m2)
Max
dis(cm)
Plate
areas
M^2
Thickness
m
Efficiency
5 0.137 0.262 344.84 3.79 2.116k
10 0.628 2.36 295.018 2.21 0.539k
12 1.37 6.56 284.024 1.89 0.256k
15 2.16 7.756 275.964 1.623 0.167k
20 2.38 10.75 273.014 1.37 0.154k

Table 4.2: Results: Experiment stage 2

In fig no series 1 represents weight while series 2 represents thickness
In fig no series 1 represents stresses while series 2 represents deflections.

123


Fig no 4.2.1: Decrease in thickness and weight with increase in meridian segments


Fig no4.2.2: Variation in stress and displacement with meridian segments


Figures below show the state of stresses and their variation for Nv= 5,10,12,15 and 20.


0
0.2
0.4
0.6
0.8
1
1.2
0 5 10 15 20 25
Series1
Series2
0
5
10
15
20
25
30
35
40
45
0 5 10 15 20 25
Series1
Series2
124


Figure4.2.3: Variation in Stresses (Nv=5, 10, 12, 15, 20)

4.2.1 Discussion:
The thickness and the weight decreases as the discretizations better approximate the
curve and the rigidity begins to decrease. The decline is most rapid between 10 and 15
segments. We get rigid configurations with lower segments but at the cost of weight and
internal thickness which also increases the base moment.
The maximum value of stress and its range increases with the number of divisions. The
distribution of stress is also different and it seems to be moreover evenly distributed
between Nv=12-15.
From Nv=10 to 20 there is a reversal in the distribution of stress. There is a clear gap of
lower stress for Nv=10 whereas the same area experiences higher stresses for Nv=20.


125

4.3 Experiment stage 3:
The results are shown in the table for plate thickness of 100mm and Nv=10,Nh=20.
P Max stress deflections Plate areas efficiency
20% 0.792 2.9 322.844 0.391k
0 0.923 3.9 294.481 0.367k
-20% 1.08 4.3 254.441 0.363k
Table 4.3: results stage 3
4.3.1 Discussion:
The rigidity degrades as the position of the apex falls below the hoop discretization.
Although the weight decreases, it cannot compensate to increase efficiency and it does
not seem like a suitable configuration. It is a good option to increase this parameter to
achieve increased rigidity. The variation in stress field is not appreciable. The thickness
also increases with increasing P as shown in the graph below. The values are 2.57,2.21
and 1.83m respectively.



0
0.5
1
1.5
2
2.5
3
-30 -20 -10 0 10 20 30
126

4.4 Experiment stage 4:
The results are shown in the table below. Here Nv=10 and p=0. The plate areas are that of
a quarter dome as the enclosed volume is not constant for a particular configuration here.

Nh Stress
N/m2
Deflection
cm
Plate areas
(quarter
dome)
Thickness
m.
efficiency
12 2.8 13.39 618.636 2.58 0.57k
16 1.14 6.25 674.52 2.33 1.3k
20 0.628 2.36 737.55 2.21 2.15k
24 0.613 2.77 808.692 2.127 2.01k

In the graph below series 1 represents stress and series 2 represents deflections. In the
next graph, series 1 represents weight and series 2 represents thickness.


0
0.2
0.4
0.6
0.8
1
1.2
0 10 20 30
Series1
Series2
127

Fig no: variation in stresses and displacements






Figure no: variation in thickness and weight.


Figure below shows the stress distribution obtained via a FEM formulation.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 10 20 30
Series1
Series2
128


Figure: stress distribution (Nh=12,16,20,24)






4.4.1 Discussion:
As the curve is closely modeled, the thickness approaches a limiting value. The opposite
is true for weight as more of these lunes will be used up to make the dome.
The value of stress and deflections decrease with increase in the number of hoop
divisions and we get more rigid configurations. This is opposite to the result derived in
129

part B that larger vertical divisions decrease rigidity. The slight bent at Nv=20 suggests a
minima but further discretizations would be impractical and were not explored.

Although stresses decrease with the number of divisions, they are more or less evenly
distributed for Nh=12 whereas the clear gap can be observed with higher values.
The decreases in stresses easily overcome the increase in weight to produce more
efficient designs in the realm of larger segmentation.















130

CHAPTER 5
CONCLUSION AND FUTURE WORK
5.1Conclusion:
The experiments have helped quantify the aspects of rigidity for such structures modeled
with concrete. Since timber has a better strength to weight ratio, these designs should be
more appropriate with timber or a light material.
Certain non-intuitive aspects such as the variation of stress distribution have come to
light and should play an important role in proportioning the stiffness.
Plates of constant thickness have been used throughout but better designs may be realized
by reducing the thickness at the top parts. This may again change the distribution of
stress. Alternately the plates can be stiffened by providing stiffness elements along fold
lines.
Figure below shows typical deflection field in the structure. Ties can be used along the
meridians to distribute deflections and hence stresses to create an even stress profile.
This should be achieved with member capable of compression.
To hold deflections, hybrid models can be implied where the smaller Nh will act as
meridian reinforcement.(fig 5.3)
Alternatively, the geometry may be unevenly modeled as shown in figure to decrease the
weight on the top parts.(fig5.2)
Rigidity wherever required, may also be enhanced by controlling the parameter P.
Finally, weight will be an important consideration for these structures and optimisations
to minimize weight will be helpful in concluding the right distribution of parameters.

131


Fig5.1: Typical deflection profile and the use of tie

Fig 5.2: Uneven discretizations of vertical segment
132



Figure 5.3: Hybrid stiffened dome


Fig 5.4: Connection detail
133

5.2 Future Work:
Future work may revolve around using better methods which integrate the geometry and
FEM formulation more suitable for this kind of variational problem. Certainly the plate
thickness will have to be reduced at the top part which adds another parameter to the
problem. Studies can be undertaken to minimize weight which is an important to the
design of these structures.
Possibilities of hybrid geometries may be explored and the design and simulation of a
realistic hinge is crucial as the real models disagree with the FEM results indicated in the
study referenced (Buri).
Realistic loading conditions, suitable material properties if cross laminated timber panels
are used, which have a useful application here may be important.
Comparison studies of miura ori and other types of patterns may be useful and finally a
deployment mechanism may be crafted to realize a functioning rigid origami deployable
structure.









134

APPENDIX:
Results: Geometric modeling



135

Parabola Nv=Nh=20,P=0
Discretizations Cylindrical coordinates es (R,Z) plate areas
X(Radius) Z Roots C D ABCD CDEF
25.000 0.000 R Z R Z
24.162 1.647 23.573 0.31 22.720 1.78 5.873 5.828
23.300 3.283 22.742 1.943 21.863 1.715 5.651 5.605
22.413 4.905 21.890 3.561 20.984 1.646 5.428 5.38
21.499 6.512 21.012 5.165 20.076 1.575 5.197 5.152
20.554 8.101 20.109 6.74 19.139 1.501 4.981 4.915
19.576 9.670 19.169 8.31 18.166 1.425 4.736 4.69
18.563 11.216 18.207 9.86 17.164 1.346 4.484 4.436
17.510 12.736 17.203 11.378 16.120 1.264 4.242 4.196
16.413 14.224 16.170 12.866 15.033 1.18 3.988 3.936
15.268 15.675 15.092 14.317 13.900 1.09 3.734 3.673
14.069 17.082 13.963 15.73 12.709 0.997 3.487 3.422
12.811 18.436 12.789 17.09 11.471 0.9 3.221 3.156
11.485 19.724 11.571 18.393 10.166 0.8 2.956 2.878
10.085 20.931 10.289 19.625 8.792 0.69 2.692 2.602
8.605 22.038 8.947 20.772 7.346 0.576 2.436 2.326
7.038 23.018 7.539 21.822 5.832 0.457 2.178 2.038
5.384 23.840 6.093 22.75 4.233 0.332 1.987 1.769
3.646 24.468 4.614 23.59 2.543 0.12 1.769 1.465
1.842 24.864
0.000 25.000
Inner radius=23.63m
Thickness=1.37m
angles: 70.487 68.399 69.172 70.036 70.634 71.916 72.292 72.839
inner 72.975 73.376 73.186 72.544 70.879 68.335 64.371 58.752
50.518 39.365
outer: 64.011 64.615 65.49 66.162 67.583 67.886 68.717 68.965
69.441 69.646 69.388 68.377 66.504 63.586 59.161 52.81
43.966


136

catenary Nv=Nh=20P=0
discretizations ROOTS plate areas
X(Radius) Z C D
25.000 0.000 R Z R Z ABCD CDEF
24.270 1.723 23.56258 0.41 520.312 2.04 6.032 5.987
23.506 3.430 22.84104 2.106 486.113 3.72 5.831 5.791
22.707 5.120 22.08809 3.78 451.772 5.38 5.616 5.554
21.869 6.792 21.28561 5.45 417.090 7.03 5.378 5.354
20.988 8.441 20.49717 7.062 382.759 8.6 5.225 5.014
20.061 10.065 19.5699 8.695 346.605 10.25 4.919 4.93
19.084 11.658 18.67754 10.28 311.678 11.78 4.691 4.636
18.052 13.217 17.71761 11.84 276.798 13.294 4.42 4.366
16.960 14.735 16.70146 13.354 241.955 14.761 4.16 4.107
15.802 16.203 15.54594 14.831 207.544 16.175 3.888 3.949
14.574 17.612 14.49412 16.235 182.103 16.947 4.414 3.165
13.270 18.951 13.30072 17.58 141.769 18.78 3.356 3.283
11.884 20.206 12.03408 18.861 111.377 19.957 3.061 2.984
10.414 21.361 10.69903 20.05 83.103 21.022 2.783 2.69
8.857 22.395 9.28957 21.14 57.781 21.96 2.504 2.393
7.214 23.287 7.814081 22.116 36.175 22.795 2.187 2.077
5.493 24.015 6.281298 22.97 19.005 23.362 2.017 1.798
3.704 24.555 4.724511 23.72 6.907 23.79 1.835 1.486
1.865 24.887
0.000 25.000
inner radius: us: 23.6m
thickness=1.4m
angles 70.8 70.945 71.168 71.385 71.822 71.211 71.592 71.962
inner 71.769 70.26 64.121 69.406 68.203 65.769 62.147 57.846
48.866 38.213
outer 69.845 69.731 70.265 68.935 71.488 69.346 69.982 69.878
69.552 51.594 76.415 66.057 64.089 61.046 57.304 50.854
42.127





137

Hemisphere Nv=Nh=20,P=0
original discretizations ROOTS Plate angles
R Z C D
25 0 R Z R Z ABCD CDEF
24.92076 1.961 23.66589 0.895 23.50744 2.803 6.705 6.176
24.69007 3.911 23.54256 2.748 23.23056 4.639 6.594 6.577
24.30692 5.836 23.28678 4.586 22.81034 6.447 6.439 6.406
23.77432 7.725 22.88411 6.402 22.25958 8.22 6.236 6.187
23.09529 9.567 22.33865 8.18 21.56632 9.94 6.027 5.96
22.27282 11.35 21.66202 9.94 20.7439 11.62 5.706 5.678
21.31394 13.062 20.85081 11.601 19.79079 13.22 5.454 5.387
20.22367 14.695 19.90622 13.216 18.71367 14.747 5.148 5.087
19.00903 16.236 18.83744 14.75 17.51398 16.18 4.839 4.764
17.67603 17.678 17.65402 16.207 16.21397 17.525 4.481 4.41
16.2347 19.01 16.35739 17.563 14.80664 18.758 4.136 4.064
14.69308 20.225 14.95683 18.816 13.30395 19.887 3.772 3.703
13.06118 21.316 13.46248 19.958 11.7161 20.875 3.437 3.354
11.34905 22.275 11.8856 20.982 10.055 21.743 3.097 2.998
9.5667 23.097 10.23754 21.885 8.329635 22.476 2.771 2.649
7.725175 23.776 8.530353 22.667 6.551195 23.07 2.469 2.308
5.835506 24.309 6.783892 23.336 4.72758 23.523 2.198 1.967
3.910732 24.692 5.044489 23.927 2.862777 23.827 1.988 1.609
1.960883 24.923
inner radius us=23.6m
thickness 1.4m
angles: 70.487 68.399 69.172 70.036 70.634 71.916 72.292 72.839
inner 72.975 73.376 73.186 72.544 70.879 68.335 64.371 58.752
50.518 39.365
outer 64.011 64.615 65.49 66.162 67.583 67.886 68.717 68.965
69.441 69.646 69.388 68.377 66.504 63.586 59.161 52.81
43.966



138

parabola Nv=10,Nh=20,P=20% cylindrical coordinates
original discretization. ROOTS
base apex C D
R Z R Z R Z R Z
25.00003 0.000 22.94887 3.9339 22.17238 0.66 20.46503 4.17
23.30074 3.283 21.16612 7.079 20.52125 3.874 18.36155 7.233
21.49821 6.512 19.26532 10.153 18.79757 7.02 16.50129 10.21
19.57629 9.670 17.22219 13.135 16.96222 10.077 14.50409 13.07
17.50993 12.736 15.00675 15.9917 14.93498 13.327 12.324 16.127
15.26802 15.675 12.57974 18.669 12.92185 15.81 9.987818 18.234
12.81045 18.436 9.892966 21.085 10.72004 18.406 7.389718 20.368
10.08506 20.931 6.894236 23.098 8.453041 20.787 4.478775 22.27
7.038676 23.018 3.567227 24.49
3.646229 24.468 0 25
0 25
inner radius= us=22.427m
plate areas thickness=2.573m
ABCD CDEF
26.204 26.41
21.137 27.51
23.811 24.251
21.698 22.211
18.731 18.092
17.275 18.299
14.916 16.49
11.958 13.851



139

parabola Nv=10,Nh=20P=0 cylindrical coordinates
original discretizations ROOTS
R Z C D
25.00003 0 R Z R Z
23.30074 3.28347 22.47901 0.778 20.71484 3.778
21.49821 6.51197 20.80395 3.99 18.90801 6.91
19.57629 9.67011 19.02534 7.128 16.99213 9.961
17.50993 12.7356 17.14683 10.178 14.93099 12.887
15.26802 15.675 15.14666 13.11 12.69914 15.647
12.81045 18.4355 13.02539 15.88 10.26162 18.17
10.08506 20.9312 10.77319 18.46 7.583378 20.35
7.038676 23.0183 8.475736 20.8 4.594127 22.03
3.646229 24.4682
0 25
PLATE AREAS Inner radius= us=22.79m thickness=2.21m
ABCD BCDF
11.853 11.669
11.171 10.952
10.454 10.131
9.746 9.442
9.027 8.637
8.348 8.107
7.693 6.884
7.521 5.874








140

parabola Nv=10,Nh=20,P=-20%
original discretizations apex ROOTS
R Z R Z C D
25.00003 0 23.64803 2.63077 R Z R Z
23.30074 3.28347 21.82732 5.94264 23.04042 1.091 21.25492 3.624
21.49821 6.51197 19.88454 9.18426 21.27564 4.27 19.37272 6.81
19.57629 9.67011 17.7952 12.3333 19.28919 7.257 17.51566 9.73
17.50993 12.7356 15.52742 15.356 17.35437 10.295 15.3769 12.717
15.26802 15.675 13.03905 18.1993 15.29714 13.21 13.06986 15.53
12.81045 18.4355 10.27665 20.7756 13.12249 15.967 10.55253 18.114
10.08506 20.9312 7.181536 22.936 10.82343 18.51 7.783939 20.33
7.038676 23.0183 3.725072 24.4449 8.495675 20.83 4.704572 22.03
3.646229 24.4682 0 25
0 25
INNERR RAD= AD=23.172
PLATE AREAS THICKNESS= S=1.828
ABCD CDEF
19.302 18.593
18.429 18.016
17.414 17.943
16.359 16.97
15.232 16.012
14 15.075
13.299 12.711
11.125 13.961



141

parabola Nv=5 Nh=20,P=0
original discretizations ROOTS
R Z C D
24.99995 0 R Z R Z
21.50212 6.512 21.42368 -1.606 16.63475 1.302
17.50991 12.736 17.22399 7.81 11.57367 12.821
12.8104 18.435 13.9666 13.488 7.800599 17.66
7.038678 23.018
0 25
plate areas inner radius 21.21
thickness 3.79
ABCD CDEF
7.14 30.523
33.54 28.905
27.81 22.94


142

parabola Nv=15,Nh=20,P=0
original discretizations roots
R Z C D
24.99995 0 R Z R Z
23.8775 2.194 23.21445 0.471 22.06193 2.528
22.71194 4.366 22.10081 2.63 20.90144 4.66
21.49821 6.512 20.9453 4.774 19.6937 6.772
20.23126 8.626 19.74493 6.881 18.4328 8.841
18.90518 10.703 18.49393 8.95 17.11354 10.865
17.50991 12.735 17.18482 10.978 15.72618 12.836
16.03734 14.712 15.81369 12.952 14.26404 14.741
14.47557 16.618 14.3737 14.86 12.71614 16.56
12.8104 18.435 12.85695 16.688 11.06946 18.273
11.0269 20.136 11.25745 18.415 9.314728 19.846
9.108029 21.681 9.572426 20.018 7.439643 21.34
7.038678 23.018 7.804034 21.47 5.434722 22.381
4.812788 24.073 6.001 22.78 3.277171 23.22
2.448539 24.76
0 25
INNER RADIUS: 23.377
PLATE AREAS THICKNESS: 1.623m
ABCD CDEF
7.65 7.577
7.319 7.219
6.936 6.858
6.568 6.489
6.19 6.12
5.803 5.7
5.415 5.308
5.015 4.882
4.602 4.453
4.215 4.022
3.68 3.546
3.486 3.114
3.22 2.595



143

parabola Nv=12,Nh=20,P=0
original discretizations ROOTS
R Z C D
24.99995 0 R Z R Z
23.58958 2.739 22.85022 0.626 21.39029 3.163
22.11179 5.442 21.45089 3.317 19.91519 5.806
20.55423 8.101 19.98797 5.962 18.36155 8.396
18.90518 10.703 18.46324 8.546 16.7135 10.909
17.14974 13.235 16.76565 11.136 14.93291 13.481
15.26799 15.675 15.17174 13.553 13.14442 15.623
13.23676 17.99 13.31419 15.84 11.0842 17.83
11.0269 20.136 11.38408 18.019 8.912453 19.764
8.604438 22.037 9.358828 20.026 6.550198 21.41
5.944294 23.586 7.286329 21.84 3.952177 22.64
3.050377 24.627
0 25
INNER RADIUS: 23.111
THICKNESS: 1.89m
plate areas:
ABCD CDEF
9.619 9.514
9.127 9.002
8.606 8.459
8.096 7.896
7.299 7.492
6.896 6.546
6.414 6.379
5.833 5.497
5.389 4.816
5.054 4.078



144

parabola Nh=12,Nv=20,P=0
ROOTS
R Z C D
24.99986 0 R Z R Z
23.3003 3.283 22.2845 0.669 20.51648 3.627
21.4989 6.512 20.63647 3.889 18.74945 6.777
19.57645 9.67 18.89582 7.04 16.86511 9.837
17.50974 12.735 17.04881 10.103 14.83884 12.779
15.26764 15.675 15.07893 13.052 12.64115 15.563
12.81058 18.435 12.98346 15.848 10.23755 18.113
10.08622 20.931 10.90981 18.4 7.517285 20.235
7.038125 23.018 8.474482 20.808 4.595339 22.026
3.646204 24.468
0 25
inner radius: inner radius: 21.63
thickness: 3.37
plate areas
ABCD CDEF
18.474 18.148
16.985 16.679
15.468 15.125
13.908 13.553
12.307 11.908
10.703 10.225
9.569 8.534
7.872 6.754






145

parabola Nh=16 Nv=10 P=0
original discretizations ROOTS
R Z C D
24.99934 0 R Z R Z
23.30012 3.283 22.42221 0.745 20.63251 3.731
21.4985 6.512 20.75083 3.957 18.86175 6.872
19.57619 9.67 18.98633 7.101 16.95257 9.922
17.51029 12.735 16.85297 10.408 15.11264 13.042
15.26847 15.675 15.12478 13.091 12.68107 15.619
12.81062 18.435 13.00715 15.872 10.25636 18.151
10.08551 20.931 10.77179 18.45 7.582491 20.34
7.033827 23.018 8.467776 20.808 4.601793 22.029
3.645522 24.468
0 25
INNER RADIUS: 22.668
THICKNESS 2.33m
PLATE AREAS ABCD CDEF
14.164 14.485
13.168 13.417
12.179 12.294
10.048 11.4
10.152 9.961
9.135 8.728
8.195 7.611
7.538 6.153






146

parabola Nh=24 Nv=10 P=0
original discretizations Roots
R Z C D
24.99952 0 R Z R Z
23.29989 3.283 22.52122 0.798 20.74896 3.805
21.49902 6.511 20.83444 4 18.94263 6.939
19.57689 9.67 19.05083 7.14 17.01268 9.984
17.50948 12.735 17.16571 10.191 14.94796 12.905
15.26765 15.675 15.15743 13.12 12.71119 15.663
12.81044 18.435 13.02477 15.89 10.27187 18.182
10.08555 20.931 10.77476 18.436 7.586232 20.357
7.038049 23.018 8.475126 20.81 4.594804 22.031
3.645777 24.468
0 25
INNER RADIUS: 22.873 PLATE AREAS
THICKNESS: 2.127M ABCD CDEF
10.464 10.294
9.967 9.749
9.362 9.103
8.898 8.579
8.362 7.955
7.852 7.29
7.443 6.615
7.118 5.682







147

List of References:

Boakes.N (2010). "Origami-Mathematics Lessons: Researching its Impactand Influence
on
Mathematical Knowledge and Spatial Ability of Students."

Buri, H. (2010). origami-folded plate structures, Ecole polytechnique federal de lusanne.

Buri H , W. Y. (2009). "Origami - Geometry of Folded Plate Structures."

Cohn-Vossen.S, H. D. a. (1952). Geometry and the Imagination, AMS Chelsea
Publishing.

CR, C. (1983). Theory of Shell Structures, Cambridge University Press.

Demaine .E.D , O. R. J. (2007). Geometric folding algorithms: linkages, origami,
polyhedra, Cambridge University Press.

Haasis. M, W. Y. (2008). "ORIGAMI - Folded Plate Structures, Engineering." 10th
World Conference on Timber Engineering. Miyazaki, Japan, 2-5 June, 2008.

Heike Matcha, A. L. "parametric origami, adaptable temporary buildings."

Hull.T (2010) Maekawa and Kawasaki Revisited and Extended.

148

Kobayashi H, K. B. a. V. J. (1998). "The geometry of unfolding tree leaves." Proceedings
of the Royal Society B 265.

Koryo.M (1989). "A Note on intrinsic geometry of origami." first international meeting
of origami science and technology, ferrara, italy.

Koryo.M (2009). The Science of Miura-Ori. Origami 4, A K Peters/CRC Press: 87-99.

Lang.R ((2003)). Origami Design Secrets: Mathematical Methods for an Ancient Art, AK
Peters.

Ma J and You Z ( (2011)). "The Origami Crash Box." In Origami 5: Fifth International
Meeting of Origami Science, Mathematics, and Education (5OSME).

Mitra.A (2008-09). the grammar of developable double corrugations ( for formal
srchitectural applications). london, university college london. Msc adaptive architecture
and computation.

Naohiko, W. and K. Ken-ichi (2009). The Method for Judging Rigid Foldability. Origami
4: the fourth international conference of origami in science., A K Peters/CRC Press: 165-
174.

R, L.

Ramm, K.-U. B. a. E. (1993). "Form Finding of Shells by Structural Optimization "
Engineering with Computers (1993) 9:27-35.

149

Resch.R, C. H. ( (1971)). "Kinematic Folded Plate System." In Proceedings of IASS
Symposium on Folded Plates and Prismatic Structures, Vienna, Austria

Shenk.M (2008). Textured Shell Structures , first year Phd report., university of
cambridge.

Sternberg.S "symmetry issues in collapsible origami." symmetry: culture and science.

Tachi.T (2010). "Geometric Considerations for the Design of Rigid Origami Structures."
In Proceedings of the International Association for Shell and Spatial Structures (IASS)
Symposium 2010. Shanghai, China, November 8-12, 2010.

Tachi.T (2011). Rigid-Foldable Thick Origami. Origami 5, A K Peters/CRC Press: 253-
263.

Tachi.T ( (2009)). "Simulation of Rigid Origami." In Origami 4: Fourth International
Meeting of Origami Science, Mathematics, and Education (4OSME): pp. 175-188.

Trautz. M, C. S. (2011). "Folds and fold plate structures in architecture and
engoineering."

Trautz.M , K. A. (2009). "Deployable folded plate structures - folding patterns based on 4
fold mechanism using stiff plates." Proceedings of the international association of Shell
and Spatial structures(IASS) Sypnosium 2009, Valencia.

.W.Lau, W. Equilibrium Analysis of Masonry Domes. Department of Architecture,
M.I.T.

Bathe, D. C. a. K. J. (1997). "Fundamental Considerations For The Finite
150

Element Analysis of Shell Structures." Computers and Structures 66.

Cipriani.B (2005). Development of Construction techniques in the Mamluk Domes of
Cairo. Dept of Architecture, Massachusetts institute of technology. Masters of Science
in Architecture.

Cohn-Vossen.S, H. D. a. (1952). Geometry and the Imagination, AMS Chelsea
Publishing.

CR, C. (1983). Theory of Shell Structures, Cambridge University Press.

Eduard Ventsel and T. Krauthammer Thin Plates and Shells:Theory, Analysis and
Applications. Penselvania State University., Marcel Dekker, Inc.

Hassan.A (1971). A Refined Finite Element Analysis of Folded plate structures. Dept of
Civil Engineering, University of Ottawa,Canada. Master of Applied Science.

Kubik, M. (2009). Structural Analysis of Geodesic
Domes.
School of Engineering, Durham University.

Naohiko, W. and K. Ken-ichi (2009). The Method for Judging Rigid Foldability. Origami
4, A K Peters/CRC Press: 165-174.

O.Tonon (1993). "Geometry of the spatial folded form." Space Structures4.

R.Bradshaw, D. C., M.Gargari,A.mirmiran,P.Tripeni "special structures: past , present
and future." Journal of Structural Engineering 128.

Tomohiro Tachi, M. M., Masaaki Iwamoto (2011). "Rigid Origami Structures with
Vacuumatics: Geometric Considerations." Proceedings of the IASS-APCS 2012, Seoul,
Korea, May 21-24, 2011.


151