Prototyping and testing of novel flood

protection systems

Kristian Vemund Haaland

Øystein Blix Walderhaug

Master of Science in Mechanical Engineering

Submission date: June 2016
Supervisor: Torgeir Welo, IPM

Norwegian University of Science and Technology

Department of Engineering Design and Materials
ii
Abstract

This thesis describes the early development stages of novel consumer flood protection systems.
The development process has emphasized prototyping and experimentation as a strategy to
front-load the development, which is outlined by a literature study.

A framework for the development and a set of quantitative and qualitative measures for the
systems being designed were established. The development has been based on rigid-flexible
hybrid systems, combining desired aspects from each.

As consumer flood protection systems represent an immature market with little knowledge to
base the development on, generating and capturing knowledge has been a priority.
Furthermore, the thesis describes the development and application of a program for capturing
and reusing the knowledge obtained during this project. Knowledge was primarily gained
through prototyping and by designing and conducting experiments.

The feasibility of using PVC canvas in hybrid systems in terms of stability was investigated.
The test setup used to investigate this also showed great potential for testing overall stability
of systems. In addition, a new innovative way for testing gaskets, independent of system
design, was designed and performed.

Lastly, the thesis describes two promising concepts for consumer flood protection systems and
points out further work that needs to be completed in order to move the products towards
industrialization.

iii
iv
Sammendrag

Denne avhandlingen beskriver de tidlige utviklingsstadiene av nye flomvernsystemer beregnet

for forbrukermarkedet. I utviklingsprosessen har prototyping og eksperimentering blitt brukt
som strategier for “front-loading” av utviklingen, disse strategiene er introdusert i
litteraturstudiet.

Et rammeverk for utvikling, og et sett med kvantitative og kvalitative mål for systemene ble
etablert. Arbeidet har tatt for seg hybrider av rigide og fleksible systemer som kombinerer
ønskede egenskaper.

Flomvernsystemer rettet mot forbrukerere representerer et umodent marked med lite kunnskap
tilgjengelig å basere utviklingen på. Derfor har det å generere og fange kunnskap vært en
prioritet.

Videre beskriver oppgaven utviklingen og anvendelsen av et program for å fange og gjenbruke

kunnskap tilegnet i løpet av dette prosjektet. Kunnskap har primært blitt opparbeidet gjennom
prototyping, og ved å designe og gjennomføre eksperimenter.

Muligheten for å bruke PVC-duk i hybridsystemer med tanke på stabilitet ble undersøkt.
Testoppsettet som ble brukt til å undersøke dette viste seg også å ha et stort potensial til å teste
stabiliteten til systemer. I tillegg ble en ny innovativ måte for å teste pakninger, uavhengig av
systemdesign, designet og utført.

Til slutt beskriver oppgaven to lovende konsepter for forbruker-flomvernsystem, og påpeker

videre arbeid som må utføres for å bringe disse nærmere industrialisering.

v
vi
Acknowledgements

We would like to express our gratitude to our supervisor professor Torgeir Welo and
AquaFence CEO Fred Dahl for giving us the opportunity to work with this project.

We are particularly grateful for the assistance given by Postdoc. Christer W. Elverum and PhD-
candidate Sigmund A. Tronvoll.

Thank you!

vii
Table of Contents
1 INTRODUCTION ............................................................................................................ 1

1.1 Background ................................................................................................................. 1

1.2 Objectives .................................................................................................................... 1

1.3 Project Structure .......................................................................................................... 2

2 LITERATURE STUDY: PROTOTYPING AND EXPERIMENTATION IN

PRODUCT DEVELOPMENT ............................................................................................... 3

2.1 Purpose ........................................................................................................................ 3

2.2 What is a prototype ..................................................................................................... 3

2.3 Classifying prototypes ................................................................................................. 4

2.4 The purpose of prototyping ......................................................................................... 8

2.5 Experimentation ........................................................................................................ 10

2.6 Decisions about prototyping and fidelity of prototypes. ........................................... 13

3 UNDERSTANDING THE CONTEXT OF FLOODS................................................. 15

3.1 What causes floods .................................................................................................... 15

3.2 The global impact of floods ...................................................................................... 15

3.3 The forms of flood protection ................................................................................... 16

3.4 Finding an Alternative to Sandbags .......................................................................... 17

3.5 Categorization of temporary flood protection systems ............................................. 19

3.6 AquaFence ................................................................................................................. 22

3.7 The Market for a New Type of Flood Barrier ........................................................... 24

4 ESTABLISHING A FRAMEWORK FOR THE DEVELOPMENT PROCESS..... 25

4.1 Development strategy................................................................................................ 25

4.2 Choosing a direction for the development ................................................................ 28

4.3 Failure mechanisms as design guidelines.................................................................. 30

viii
 4.4 Outlining the Key mechanisms ................................................................................. 32

4.5 Other considerations for development ...................................................................... 37

4.6 Summary of design restrictions ................................................................................. 39

5 GAINING AND CAPTURING KNOWLEDGE ......................................................... 41

5.1 The value of capturing knowledge ............................................................................ 41

5.2 Developing a Knowledge Capturing System ............................................................ 42

5.3 Investigating the contribution of canvas towards stability ........................................ 46

5.4 Investigating the seepage rates for different gaskets ................................................. 50

6 CONCEPT DEVELOPMENT ...................................................................................... 65

6.1 Resources .................................................................................................................. 65

6.2 Early phase concept development ............................................................................. 68

6.3 Tetris.......................................................................................................................... 69

6.4 Caterpillar .................................................................................................................. 80

6.5 FlexiFront .................................................................................................................. 87

6.6 Evaluation of concepts .............................................................................................. 91

6.7 Further work and industrialization of concepts ......................................................... 94

7 CONCLUSION ............................................................................................................... 97

8 REFERENCES ............................................................................................................... 99

APPENDICES ...................................................................................................................... 103

Appendix A - Code for the Knowledge Capturing System ................................................ 105

Appendix B - Equations used to establish relationships for Tetris .................................... 107

Appendix C - Arduino code for automated gasket experiment .......................................... 109

Appendix D - Documentation from initial development ................................................... 111

Appendix E - Risk assessment form .................................................................................. 137

ix
1 Introduction

1.1 Background
The market for flood protection systems has seen considerable growth in recent years. In the
wake of global natural disasters with increasing frequency and severity, the demand for flood
protection solutions is increasing.

AquaFence is a Norwegian company specializing in development and production of flood

protection systems that are tested and certified according to international standards. The
company is currently offering temporary medium- and large-scale systems. Since AquaFence’s
current products are mainly intended for large-scale applications and extreme flood levels, it is
desirable to develop a new, inexpensive system targeted at the consumer market.

Consumer flood protection systems represent an immature, and partially unexplored market,
demanding products that differ greatly from the products AquaFence currently offers. The
company will therefore benefit from a new product designed from the ground up.

1.2 Objectives
As AquaFence is a relatively small company with limited resources, their development efforts
mainly target incremental improvements of existing products. The circumstances of this project
allow a more scientific approach to development, that includes experimentation to generate
reusable knowledge that can be of great value to the company. Therefore, in addition to
generating promising concepts for consumer flood protection barriers, a system generating
reusable knowledge will be developed.

This project aims to attain the following objectives:

I. Develop novel concepts for consumer flood protection barriers

II. Develop new innovative ways for testing desired aspects of new concepts
III. Design and build a functional test set-up for comprehensive prototypes
IV. Establish a system for capturing knowledge gained during development

1
1.3 Project Structure
Chapter 2 presents a literature study on prototyping and experimentation. This is to gain an
understanding of the different methods that exist for prototyping and experimentation so that
they can be used in the best way possible to aid the development of a new flood protection
system. Chapter 3 outlines facts about floods and establishes the types of flood protection that
currently exists. Based on this, chapter 4 establishes a framework and a design space for the
development of new flood barrier concepts. Chapter 5 outlines the tools for capturing the
knowledge that is created in the project and finally, the concepts that have been developed based
on this knowledge are presented in chapter 6.

2
2 Literature study: Prototyping and
Experimentation in Product Development

2.1 Purpose
Creating prototypes and performing experiments are essential activities when developing new
products, and they can be useful tools for the duration of the development process. Because
different types of prototypes and experiments are required for different stages of development,
it is important to gain a thorough understanding of the various types, and how to best take
advantage of them. This is the main incentive for performing a literature study on prototyping
and experimentation for this master’s thesis.

In an experiment investigating the benefits of various techniques in product development, Tidd

& Bodley (2002) found that prototyping was useful in all projects, especially in high novelty
projects, which is the category a new flood barrier concept belongs to.

2.2 What is a prototype

As a starting point, it is useful to define what a prototype is. It is common to think of a prototype
as a rough version of the final product, used as a tool for validation and verification, usually
late in the development. However, prototypes have proven useful throughout the whole product
development process.

Houde & Hill (1997) define a prototype as “Any representation of a design idea, regardless of
medium” and further argue that anything can be considered a prototype, depending only on how
it is used and interacted with. For example, a brick could be considered a prototype if it
represents the weight and scale of a product. This concept is similar to the ideas presented by
Ulrich & Eppinger (2012) who define prototypes as “an approximation of the product along
one or more dimensions of interest”; meaning that anything that exhibits properties that are of
interest to the product developers can be considered a prototype, regardless of it being a physical
or a non-physical object.

3
2.3 Classifying prototypes
The definition of a prototype established so far is extremely broad, and as different types of
prototypes are useful in different scenarios and stages in the product development cycle, it is
convenient to classify and categorize the prototypes. There is no universally agreed upon way
to classify prototypes, but several frameworks have been proposed.

Ulrich and Eppinger

Ulrich & Eppinger (2012) propose a categorization in two dimensions: analytical as opposed
to physical, and focused as opposed to comprehensive (see Figure 2.1).

Figure 2.1: The two dimensions of Prototyping (Ulrich & Eppinger, 2012)

In physical prototypes, the aspects that are of interest to the product developer are built into
tangible objects that approximate the product. An example of this notion would be the brick
mentioned earlier, simulating the weight and scale of a product. Being able to interact with an
object in this way, allows a product developer to get a clearer idea about the final product. A
benefit with physical prototypes is that they can present the visual aspects more clearly than,
for example, a 3D model on a computer screen. Physical prototypes are also useful as they often
reveal aspects unrelated to the original intention of the prototype and can provide useful insight
into the final product.

4
Analytical prototypes, on the other hand, are mathematical approximations of a product and
consequently allow the parameters to be changed easily. For example, changing a specific
dimension on a CAD-model could be done in seconds, while changing a dimension of a physical
prototype usually requires building a new one. Therefore, for complex designs, analytical
prototypes are useful to establish a range of feasible parameters, which can then later be fine-
tuned by building high-fidelity physical prototypes. However, for simpler design parameters it
is often advantageous to build physical prototypes directly.

Whether a prototype is considered comprehensive or focused depends on how many attributes

are implemented in the prototype. A fully comprehensive prototype would therefore implement
all the attributes, while a focused prototype explores only a selection. As a result, focused
prototypes can be built earlier on in development, as they do not rely on every single attribute
being established before building. A common practice that Ulrich & Eppinger (2012) identify
is building two focused prototypes where one of the prototypes is a “looks-like” and the other
a “works-like” prototype.

Houde and Hill

Houde & Hill (1997) categorize prototypes into three dimensions, where each dimension is an
important part of the development process, and that each requires a different approach to
prototyping. The three dimensions are listed below, and visualized in Figure 2.2:

 Role: the function an artifact serves in a user’s life

 Look and feel: the sensory experience of using an artifact
 Implementation: the techniques and components through which an artifact performs
its function

Figure 2.2 Three dimensions describing what prototypes prototype (Houde & Hill, 1997)

5
The purpose of this classification is to divide prototypes into categories that “visualize the focus
of exploration” as each area requires a different approach to prototyping. “Role” refers to the
usefulness of a product in a user’s life and requires that the context of use be established. “Look
and feel” refers to the sensory aspects of using a product and requires that user experiences are
simulated. “Implementation” refers to the mechanisms that allow a product to perform its
functions, and it requires that a working system is built.

Houde & Hill (1997) identify a fourth type of prototype called Integration prototypes that are a
combination of the three dimensions: role, look and feel, and implementation. Integration
prototypes are built as a representation of the “complete user experience of an artifact” and are
useful to balance and resolve constraints across the design dimensions, and to verify that the
design is complete and coherent. Integration prototypes resemble a finished product the closest,
but as a result often requires them to be as complex as the final product, meaning they are
difficult and time-consuming to build.

Bryan-Kinns and Hamilton

Bryan-Kinns and Hamilton (2002) identify three dimensions that are useful for determining
what kind of prototype should be used. The three dimensions are illustrated in Figure 2.3 and
are:
 Fidelity: “The degree of exactness” of the prototype, ranging from low to high
 Audience: Who the prototype is intended for, ranging in terms of organizational
position
 Development stage: Where in the development process the prototype is being used.
This is not a linear process as the development process is iterative and different
aspects may be relevant at different stages of development

Figure 2.3 Three dimensions for determining what prototype to use (Bryan-Kinns & Hamilton, 2002)

6
Bryan-Kinns and Hamilton (2002) stress that although these three dimensions are the most
important, they are not the only factors that should be considered when deciding what type of
prototype should be made. Available tools and skills, time-constraints, and costs are examples
of other dimensions that should affect the form and content of the prototypes.

Schrage
Schrage (1993) writes about prototyping culture and proposes a diagnostic matrix to determine
how a company uses prototyping. The matrix (Figure 2.4) is composed of three axes: External
vs Internal (who is going to interact with the prototype), formal vs informal (roughness of the
prototype), and risk-management vs opportunity (the purpose of the prototype). Whether a
prototype is classified as an opportunity or as risk management is closely related to whether or
not it is designed for validation and verification, or primarily to learn and gain knowledge.

Figure 2.4 Prototype Culture Diagnostic Matrix (Schrage, 1993)

Schrage (1993) urges the use of creative prototypes, meaning informal and internal prototypes
as they have the ability to create new opportunities. He argues that a company's prototyping
culture can indicate its ability to innovate, and that there is a connection between the number
and quality of their prototypes, and the quality of their products. The prototypes can become a
tool to drive the innovation process forward, rather than simply being a product of the process
itself.

7
 Elverum and Welo
Elverum & Welo (2015) investigated the use of prototypes in the development of two novel
innovations in the automotive industry. Their findings led to a proposed explanatory model that
consists of two distinct types of prototypes: directional and incremental. Directional prototypes
are described as a tool for evaluating the direction in which a development is heading. In this
case, the prototypes are used to investigate the feasibility of a concept, and determine whether
development should continue in the direction by providing “good enough” or “not good
enough” as an answer. The main purpose of an incremental prototype is to optimize aspects of
a product while increasing the understanding.

2.4 The purpose of prototyping

Ulrich and Eppinger’s four purposes of prototyping
Ulrich & Eppinger (2012) argue that within a product development project, prototypes serve
four purposes: learning, communication, integration and milestones.

Learning: Is when prototypes are used to answer questions like “will it work?” and “how well
does it meet customer needs?”. Prototypes in this context can reveal new aspects, opportunities
or limitations that would not have been discovered otherwise.

Communication: Prototypes can aid communication in a product development process

between engineers, but also with customers and management. Not only do prototypes facilitate
team and firm communication, but they also offer an effective mechanism to provoke
discussions early in a development process.

Integration: Prototypes can be used to encourage the different members of a team to coordinate
their efforts by connecting all the parts and subassemblies that make up a product. Physical
comprehensive prototypes can therefore be a highly effective tool to ensure that components
work together.

Milestones: Milestone prototypes are used to show that the development of a product has
progressed and that it has reached a certain level of functionality. This includes formal tests that
a product must pass to achieve approval.

8
 The generative role of prototypes
Lim, Stolterman, & Tenenberg (2008) point out the lack of knowledge about the fundamental
nature of prototypes. The article reveals two main dimensions of prototyping: prototypes as
filters, and prototypes as manifestations.

They convey that a prototype filters qualities, and consequently the designer become aware of
factors of the design that generate certain qualities. This paper draws attention to the importance
of the designer’s ability to design prototypes that filters the qualities of interest (Figure 2.5).
Further, they argue that “the most efficient prototype is the most incomplete one that still filter
the qualities of interest to the designer”.

Design Idea Filter Manifestation

Refines

Figure 2.5 Prototyping as a filter (Lim, Stolterman, & Tenenberg, 2008)

The idea behind prototypes as manifestations is that the process of making a prototype in and
of itself contributes to further development of an idea in some way.

“When a designer creates and envisions an idea, she necessarily develops

the idea by moving it out in the world.” (Lim, Stolterman, & Tenenberg,
2008)

9
2.5 Experimentation
The term experimentation is comprehensive and hence difficult to define. Smith (2007)
suggests three definitions:

 Learning by trying things out

 Something one does deliberately to see what happens
 The process of an action followed by an observation

He emphasizes that any of these three definitions points in the direction of “You provoke a
situation and see how it responds”. Figure 2.6 illustrates that the learning outcome is greatest
when the outcome is unexpected, and Smith states that most developers use experimentation
mostly for validation and verification, meaning the learning outcome of the experiments are
low.

Figure 2.6 Learning outcomes (Smith, 2007)

Incorporating prototyping in the early stages of a new product development process can be a
useful tool for communicating ideas and gaining knowledge. By doing this, the use of
experimentation with prototypes fundamentally change, from a context where unexpected
outcomes of experiments are considered non-compliance and would call for corrective actions,
to a context where unexpected outcomes are not deemed as failures, but rather important means
for learning and communication (Elverum & Welo, 2015).

10
 Experimentation and Set-based design
Since the cost of rework increases in the later stages of development, efforts should be made
early in development to avoid rework later on. However, this strategy is often difficult to
enforce as development is often driven by short-term incentives, which make savings early on
seem attractive. Set-based design is an attempt to prevent rework in the later stages by
representing initial requirements as a range of acceptable values that are refined during
development, rather than the traditional point-based method where specifications are pinned
down to a single value. Development teams might often want to pin down specifications as it
gives the impression that the project is moving forward, while in reality it may delay the
process.

The sets established as a result of a set based approach are refined by adding constraints as
additional knowledge is gained during development. This gradually reduces the design space
until the final specifications are decided. Restraints can be added for many reasons, such as the
laws of physics, the available technologies and laws governing production methods and
materials. These restraints can be understood from the beginning of development, while other
restraints become apparent as more knowledge is gained. Experimentation is a valuable method
to gain this kind of knowledge to limit the design space. While the traditional method is to
design and then test, the set-based approach is to test and then design based on the results.

Kennedy, Sobeck II, & Kennedy (2013) demonstrate the values of set based experimentation
by exploring why the Wright brothers were successful in designing the first functioning airplane
despite having considerable fewer resources in terms of aeronautical knowledge and budget
than the competition. The Wright brothers identified and isolated critical knowledge gaps and
then performed experiments to fill these gaps. For example, they constructed a wind tunnel to
test different isolated parts of a wing design, like aspect ratio and airfoil shape, to learn how
each of these parts affected lift. They would not have to take into consideration customer
requirements and other airplane specifications when performing these experiments, and as a
result were able to perform many tests over a short period of time to gain a lot of knowledge.
They then used this acquired knowledge to design a functioning and reliable wing on their first
try. Kennedy, Sobeck II, & Kennedy (2013) argue that another reason the process of the Wright
brothers was so successful was because they front-loaded the development process and thereby
avoiding time-consuming late-process rework.

11
 Limit curves and set-based knowledge
The Wright brothers needed a way to capture the knowledge
gained from the experiments in a way that enabled them to
quickly and easily reuse the knowledge. To accomplish this,
they created limit curves (Figure 2.7) that visualized the sets of
acceptable values for given parameters in a way that was easy
to understand and extract information from. These types of
curves can be made for multidimensional, as well as simple 2-
dimensional relationships, and they are an important tool in
shifting development from a “guess and test and then rework”
Figure 2.7 A limit curve
to a development process based on sets that are known to work. created by the Wright Brothers

Experience prototyping
Buchenau & Suri (2000) discuss what they call “experience prototyping” and relate it to what
Houde & Hill (1997) refer to as the “look and feel” of a prototype, but describe it as beyond
“concrete sensory”. By engaging in more of an “exploring by doing” way of developing, it is
easier to understand the subtle differences between design decisions and allows us to engage
with new problems in novel ways. Research has been carried out studying the benefits of
“learning by doing” that suggest it can lead to lower costs and increased production (Macher &
Mowery, 2003).

Buchenau & Suri (2000) outline three situations where experience prototyping is especially
useful:

Understanding and evaluating design ideas: Creating a high fidelity situation to simulate a
user situation as a way to identify the needs and possibilities

Exploring and evaluating design ideas: Taking the ideas and creating situations that allow
them to be tested by interacting with them

Communicating ideas to an audience: Allowing someone external to the prototyping process

actively interact with a design idea, usually to persuade them of something

12
 Collaborative prototyping
Bogers & Horst (2013) argue that by including stakeholders in the prototyping process, it
changes from a process of discrete steps to a continuous process of iterative problem solving.
It allows all involved with development to see their changes being implemented, while being
exposed to and aware of the design constraints. As Buchenau & Suri (2000) also discussed, the
prototypes act as a communication device when involving stakeholders, acting as a tool for
“cross-fertilization” of knowledge, creating a balance between functionality and usability.

By creating prototypes you can involve the user and receive feedback during development. It
is a lot easier for potential users to interact with a physical prototype than specifications. By
introducing the users at an early stage of development the prototypes can be used as a platform
for discussion to improve the product (Schrage, 1993). If the stakeholders are only introduced
towards the end of development, the prototype instead becomes a tool of persuasion since it is
at this point too expensive and time-consuming to go back and change the design.

It is noted, however, that there are windows of opportunity where collaborative prototyping has
the optimal effect, and that efforts should be made to involve the appropriate stakeholder at the
right time.

2.6 Decisions about prototyping and fidelity of prototypes.

Elverum & Welo (2015) argue that using prototypes has several limitations and shortcomings.
Prototyping can be a time-consuming and expensive activity. Therefore, making the right
choices about when to prototype, and choosing no more than adequate fidelity of a prototype
should be a priority. High-fidelity prototypes at the wrong stages of development does not only
require many resources but tends to lead to superficial feedback regarding details and
appearances.

13
McCurdy, Connors, Pyrzak, Kanefsky, & Vera (2006) highlight that in many cases, a prototype
cannot be categorized as either low or high fidelity; prototypes with different fidelity in each
area should be emphasized, and is especially useful in the early parts of a development project.
Focused prototypes should be designed so that the area that is to be targeted by the prototype is
of higher fidelity than other parts.

“Low-fidelity methods have received a great deal of recognition in the field

for their ability to validate designs and predict large problems at an
extremely low cost” (McCurdy, Connors, Pyrzak, Kanefsky, & Vera, 2006)

Ulrich & Eppinger (2012) argue that the anticipated benefits in reducing risk by making a
prototype, must be assessed on the basis of time and resources used to make and evaluate the
prototype, especially when dealing with comprehensive prototypes. The novelty of the product
should also be considered, as products are often related to new technologies, and as a result,
uncertainty.

14
3 Understanding the Context of Floods

3.1 What causes floods

A flood is defined as “an overflow of a large amount of water beyond its normal limits,
especially over what is normally dry land” (Oxford University Press, 2016), and it can be the
result of many events. The most common cause is rivers or streams that overflow due to natural
events like heavy rainfalls or rapid snow melting, or due to unnatural events like a dam failing.
Another type of flood is coastal flooding, which occurs when the sea surges inland, usually due
to a large storm or a tsunami.

Some floods take hours or even days to develop, while others are flash floods that generate
quickly and with little warning. Efforts are made to predict when and where flooding will occur
so that flood warnings can be issued and appropriate measures can be taken. It is the accuracy
and speed of these warnings that determine the available lead-time to deploy flood protection
measures.

3.2 The global impact of floods

Flooding has become the most common natural disaster in the world, and is the third most
damaging (The World Bank, 2010). Furthermore, with rapid increases in urbanization, as well
as aging populations in developed countries, the overall vulnerability is increasing (UNISDR,
2011). As the risk of flooding is expected to rise with increasing global temperatures
(Hirabayashi, et al., 2013), the effects of climate change will further increase the risks of flood
among the global population.

Even small floods can have a large impact as it is estimated that nearly 40% of small businesses
in USA are unable to reopen after a small flood (Federal Emergency Management Agency,
2015). As a result, the market for flood protection systems has experienced considerable
growth, and is expected to continue to increase. This is being aided by legislation and insurance
discounts in some countries, and by businesses that are trying to minimize their losses.

15
3.3 The forms of flood protection
As a starting point, it is useful to get an overview of the types of flood protection. Ogunyoye,
Stevens, & Underwood (2011) outline several ways to deal with the risk and associated
problems of flooding, which include:

 Development planning: Avoiding development in areas prone to floods, as well as

implementing SuDS (Sustainable Drainage Systems), which are management practices
and facilities designed to drain surface water
 Flood storage: Controlling flood water by retaining and releasing it at a rate that avoids
flooding
 Channel improvements: Increasing the flow of water so floods are unable to build up
 Diversion channels: Leading the water away from a certain area
 Floodwalls/embankments: Defense structures that prevent floodwater from getting
into a protected area

The focus of this project will be on floodwalls as other types of flood protection consist and
depend on large infrastructures, which is beyond the scope of this project. However, it should
be noted that floodwalls can be used as diversions channels, but this will not be addressed
directly.

The floodwalls can be categorized into the following three groups:

 Temporary: Flood protection systems consisting of fully removable parts, meaning it

is erected before a flood occurs and completely removed afterwards
 Demountable: Flood protection systems that are at least partly pre-installed and that
requires further installation to be fully operational
 Permanent: A completely pre-installed system that is passive, meaning it requires no
further action before a flood event

The most effective type of flood protection systems is permanent flood barriers because they
can incorporate cut-off barriers below ground to control underground seepage, something
temporary barriers are not able to do. Furthermore, they do not rely on any operational systems,
which reduces the risk of failure. However, with high costs and its obtrusive nature, it is often
not a viable solution, especially for areas that are not usually affected by floods. In addition, if
the requirements for a flood protection system changes, for example if protection against higher
flood levels is required, the cost of changing the setup is high.

16
The main benefit of temporary barriers is that their flexibility allows them to be erected
potentially anywhere in a short period of time. They are also relatively inexpensive, and can be
stowed away when not needed. However, the main weakness of the temporary flood barriers
currently on the market, is that they suffer from high seepage rates compared to the permanent
systems. This is mainly due to the challenges in sealing the flood barrier against the ground as
well as water seeping through permeable ground. Demountable systems can address this issue,
but this comes with higher costs and less flexibility than a fully mobile flood barrier.

The focus of this project will be on temporary flood barriers, with the possibility of exploring
some demountable systems. Permanent flood barriers will not be addressed any further as they
are dependent on their specific location, and require research that is outside the scope of this
project.

3.4 Finding an Alternative to Sandbags

Bags filled with sand and stacked on top of each other is a flood protection method that has
been around for a long time, and it remains one of the most popular ways to stop floods. The
popularity of sandbags stems from it being inexpensive and low-tech, requiring few tools and
little experience. Empty bags can easily be sent in large quantities to areas they are needed, and
with enough people and shovels (Figure 3.1), they can be filled with sand found on site to create
a functioning flood barrier.

Figure 3.1 Members of the Georgia National Guard filling sandbags in preparation for floods

17
However, there are huge problems with the use of sandbags as flood protection. These problems
include:

 Filling and stacking sandbags high enough to stop a flood is a labor intensive process
 Sand is not always readily available
 Dismantling the barrier after a flood is a labor intensive process
 Flood water is often contaminated, rendering the sandbags as hazardous waste
 UV-light degrades the bags

Despite these issues and a multitude of alternatives available on the market, sandbags are still
a popular method for flood protection, meaning the high flexibility and low cost of sandbags
are highly valued. Therefore, to develop a new flood protection system that can compete with
sandbags, it will not only need to improve on the unfavorable aspects of sandbags, but also
compete in terms of price and flexibility.

A parallel can be drawn between sandbags and mousetraps. More than 4400 patents for
mousetraps exist, but despite this, the traditional spring-loaded mousetrap (Figure 3.2) invented
in 1894, is still one of the most popular mousetraps in use. In the same way, there are many
alternatives to sandbags on the market today without having any drastic impact on the
popularity of sandbags. Consequently a new flood protection system will need to be a
drastically improved system compared to sandbags in order to be able to capture a large portion
of the consumer market.

Figure 3.2 The traditional spring loaded mousetrap made by Victor

18
3.5 Categorization of temporary flood protection systems
Table 1 identifies and describes the different types of temporary flood protection systems.
Advantages and disadvantages are outlined for each.

Table 1 Categorization of temporary flood protection systems

Temporary Flood Protection Systems
Air filled Tubes Description:
Consists of impermeable tubes filled with air. Achieves
stability and a seal against the underlying surface by using
a weighted skirt, or by anchoring it to the ground with pins.
Advantages:
 Low storage volume
 Quick installation
Example: NOAQ Tubwall
 Easy to clean and reuse
 Highly versatile
Disadvantages:
 Requires a flat foundation
 High width to height ratio
 Vulnerable to vandalism and sharp objects
 Small punctures can lead to total failure

Water filled tubes Description:

Consists of pre-fabricated impermeable membrane tubes
filled with water. Stability is achieved by the weight of the
water. Typical applications are situations where longer
flood protections are required, such as along roads or
riverbanks.
Advantages:
 Low storage volume
Example: Aqua Dam  Quick installation
 Very versatile
 Less dependent on flat foundation
Disadvantages:
 High width to height ratio
 Vulnerable to vandalism and sharp objects
 Small punctures can lead to total failure
 Not suitable at low temperatures

19
Impermeable filled Description:
containers The barriers are made out of separated rigid containers
filled with a fill mass like water or gravel. Because they
use the weight of the fill mass for stability they are
classified as gravity dams. They are most suited for urban
applications.
Advantages:
 Water proof, independent of fill material
 Can be filled with any available material
Example: Floodstop  Reusable and easy to wash
 Easy to repair
Disadvantages:
 Dependent on even terrain
 Large storage volume
 Long installation time

Permeable filled containers Description:

Permeable containers filled with aggregates and often
strengthened by a rigid structure. Typically used in rural
areas and for the protection of roads or damming along
riverbanks. Sandbags fall into this category.
Advantages:
 Low storage volume
 Adapts well to uneven foundations
 Does not require skilled labor to deploy
Example: Defense Cell Disadvantages:
 Labor intensive to install
 High setup time
 Contaminated material needs to be disposed of
 Limited re-use
 High seepage rates

20
Flexible freestanding barrier Description:
Consists of self-supporting, freestanding sections made
out of flexible and impermeable materials. Utilizes either
the weight of the floodwater or anchoring to achieve
stability and avoid seepage. The systems can be used in a
wide variety of situations, but some systems are dependent
on anchoring and therefore limit use to foundations such
as grass or gravel.
Advantages:
 Quick installation
Example: Rapid dam  Requires minimal labor to install
 Low storage volume
 Mobile
 Easy to clean and reuse
Disadvantages:
 High seepage at low water depths
 Vulnerable to wind prior to flood peak
 Vulnerable to vandalism and sharp objects

Rigid freestanding barrier Description:

Self-stabilizing rigid barriers that utilize the weight of the
flood water to achieve stability. Typically used in urban
areas that call for smaller or mid-sized closed barriers.
Advantages:
 Fast installation
 Ease to clean and reuse
 Durable
Example: NOAQ Boxwall  Less dependent on anchoring
Disadvantages:
 High seepage rates on uneven terrain
 Require high storage volume

21
 Rigid Frame Barriers Description:
Impermeable surfaces supported by rigid frame systems.
These systems are very solid and are most useful in
situations that call for usage over long periods of time.
Advantages:
 Very solid
 Easy to clean and reuse
Example: Hydro Response Steel  Minor repairs can be performed under while in use
Barrier
Disadvantages:
 High seepage at low flood depths
 Vulnerable to wind prior to flood peak
 High storage volume
 Low mobility

3.6 AquaFence
AquaFence’s current system
The company’s current list of products are rigid
freestanding barriers that are L-shaped modules (Figure
3.3) connected together with canvas. The system takes
advantage of flood water to stabilize the module and to
create a seal against the underlying surface. Each module
consists of two separate marine grade plywood sheets that
are strengthened by aluminum frames and connected to
each other with PVC canvas. Underneath each module is a
thin foam gasket used as a seal between the module and the
Figure 3.3 AquaFence Model V2100
foundation. The modules can be anchored to the foundation
for improved stability and reduction of seepage. The system
is constructed in such a way that each module is collapsible,
meaning it can fold flat and be stacked for storage.

The modular structure accommodates a system that can be arranged into any length and with
the addition of separate corner modules, the system can be angled and create a closed system.
Currently, the available system heights are: 1.2m, 1.8m, 2.1m, and 2.4m.

22
 Performance test of AquaFence system
Ward D. (2012) summarizes a laboratory test carried out in 2012. The goal of the test was to
gather data on the performance of the 1.2-meter-high AqF system, and comparing it to the
performance of sandbags. This test is valuable to examine, as the data can be used to see which
aspects the current system is better than sandbags.

The test was performed on a 1.2-meter-high and 23-meter-long AqF system which was
anchored to a concrete foundation (Figure 3.4). The results were compared to a sandbag barrier
that was 0.9 meters high. The results are shown in Table 2.

Table 2: Seepage and setup time for 1.2 m AquaFence barrier and Sandbags barrier

AquaFence Sandbags
Install/remove Man-hours [hours]
Construction 15.0 205.1
Repair 1 0.2 2.0
Repair 2 2.0
Repair 3 2.0
Disassembly 2.0 9.0
Depth [m] Seepage [l/min/m]
0.30 1.73 0.62
0.60 1.99 2.86
0.87 6.58
0.90 2.48
*units have been converted from USC to metric

The results clearly show that AqF’s system is superior to sandbags when it comes to setup time
as the man-hours required to set up the sandbags is over ten times that of the AqF barrier. The
disassembly time is also much shorter for AqF’s system. However, it should be noted that no
information about the procedure used for setting up the sandbags is given, and that this could
have an impact on the results.

In addition, AqF’s system provided a superior performance compared to sandbags with regard
to the seepage rates for high water depths, with the seepage rates for the sandbags being more
than twice as high. However, for lower water depths (lower than 0.3 meters) the sandbags
performed better. This has to do with the fact that the AqF system relies on the water pressure

23
on the base of the system to create a good seal against the underlying surface. When developing
a new system with a focus on smaller floods, it is evident that it would need to be designed
differently than the current system to improve the seepage rates at flood levels below 0.3 meters.

Overall, the AqF system performs better than the sandbag barrier despite high seepage rates at
low water depths as it drastically improves as the water depth increases. However, it should be
noted that the test was performed on a flat concrete foundation, with AqF’s system anchored to
the foundation, which is an ideal situation and may not necessarily reflect a real flooding
scenario.

Figure 3.4 AquaFence installed in laboratory from (Ward D. , 2012)

3.7 The Market for a New Type of Flood Barrier

The systems that AqF currently deliver are not designed for protecting against minor floods as
their smallest system is 1.2 meters high. As “a few inches of water from a flood can cause tens
of thousands of dollars in damage” (Federal Emergency Management Agency, 2016) there is
reason to believe there would be a market for small-scale systems. Using a system that is 1.2-
meter-high is not a viable solution for protection against smaller floods due to size and cost.
Each module of the system weighs about 70kg, and therefore requires a team of several people
to assemble. AqF recommends a team of 10 people to assemble 100 meters of the system
(Ogunyoye, Stevens, & Underwood, 2011). The cost of installing the system (in 2008) was
£475/m which adds up to 100 meters of the systems costing roughly half a million GBP.
Consequently, expenditure alone would exclude a significant segment of a potential market.

24
4 Establishing a Framework for the Development
Process
4.1 Development strategy
The development strategy used in this project can be divided into four main processes:

 Initial development
 Narrowing of the design space
 Development of concepts
 Refinement of concepts

An overview of the design strategy, including key activities for each process can be viewed in
Figure 4.1. Except for “the narrowing of the design space”, neither of these processes were
linear and ideas and concepts were developed in parallel.

25
 Overview of development strategy
Initial development

Brainstorm

Build Documentation
Learning cycles

Test

Context for flood protection system

Narrowing of the design space

Defining the user

Using failure mechanism Documentation

as guidelines

Outlining key mechanisms

Design Space
Development concepts

Brainstorming Documentation

Prototyping Developing knowledge

Learning cycles
capturing system
Testing
Targeted experiments

Early concepts
Refinement of concepts

Applying knowledge Documentation

Brainstorming Universal experiments

Prototyping Targeted

Testing

Refined concepts

Figure 4.1: Development strategy

26
 Initial development

The initial development process consisted of development before any restrictions of the design
space were made. The design space was explored by quickly prototyping and testing ideas
(Figure 4.3) that surfaced during brainstorming (Figure 4.2). Concepts and ideas that showed
promising performance were redesigned, built and tested again. Using learning cycles in this
manner is a quick way to discover the design space. Although most concepts were rejected, this
approach accelerated learning by improving intuition and understanding of associated
problems. Documentation from the initial development phase can be examined in Appendix D
- Documentation from initial development.

Figure 4.2: Brainstorming

Figure 4.3: Prototyping and testing of ideas

27
4.2 Choosing a direction for the development
Defining the user
Before developing concepts for a new type of flood barrier, it is important to establish whom
the new concept is being developed for. The current products developed by AqF are sold to
customers who are protecting large and extremely valuable properties, such as Nestlé’s mineral
bottling factory in Thailand, a skyscraper in Lower Manhattan and Teterboro Airport in New
Jersey. These customers are willing and able to purchase flood protection systems at a high
price to be able to safeguard significant values.

This project is focused on developing a system to stop small floods for the consumer market or
small businesses, where the values being protected are substantially lower compared to the
customers of AqF’s current products. The user that the development will be focused on is
someone who owns a property in an area, which is at risk of experiencing floods. The user
would purchase the system and store it somewhere on the property (or at least nearby), and
erect the barrier swiftly in the event of a flood alert.

The following considerations for the user have been established:

 Does not have access to expensive tools or machinery

 Limited number of people to erect the system (maximum two)
 Highly sensitive to the price of the system
 Lives in an urban or suburban area
 Has limited storage space

The type of flood protection system

An understanding that a consumer flood protection system needs to be different from AqF’s
current products has been established. The development will therefore not focus on improving
the design of the current system, but rather develop a new system from the ground up.

The different types of flood barriers (see section 3.5) were evaluated and it was decided to
explore freestanding barriers further. Filled tubes were rejected because the filling material is
not always available, and the high base to height ratio is an issue for most properties in an urban
environment. Even though air is always available, air-filled tubes require pumps, which in
return require electricity, which is not always available before a flood. Filled containers were
disregarded for the same reasons, as well as for being a lot more labor-intensive to install.

28
There are both advantages and disadvantages for rigid and for flexible freestanding barriers.
Instead of deciding which type of barrier would be most suitable for a new consumer model,
the idea that a system could combine the desired features from each by creating a hybrid
emerged. The desired characteristics of typical flexible and rigid systems are outlined in Table
3.

Table 3: Desired system characteristics, rigid and flexible systems

Rigid Flexible

Stability Flexibility
Concentration of gasket pressure Low Weight
User perception Mobility
Strength Storage volume
Set-up time
System cost

Development will be focused on a combination of the desired characteristics from each type of
system.

The main advantages of implementing canvas in a flood barrier for the consumer market is that
it enables light weight concepts to be developed. The canvas increases the flexibility of the
flood barrier when in use, as it can easily be bent without affecting the properties of the canvas.
The flexibility also applies to storing the system as it can be rolled or folded to minimize the
space required when not in use. However, the canvas needs to be supported to maintain a desired
shape which needs to be taken into consideration when designing for storage.

Wear from repeated use is an issue of greater importance for canvas structures than for rigid
structures due to the thickness. As floodwater often contains debris, it is susceptible to tears or
punctures if debris hits the flood barrier. Even though the low weight is a desirable feature when
erecting the barrier, it is less desirable when in operation. If the barrier as a whole is less dense
than water, there is a risk that it will float and have no effect on stopping the flood. An increased
weight leads to a higher pressure on the gasket against the ground, which affects the seepage
rates positively.

29
Furthermore, it was decided that the system should protect against flood depths of up to 0.6
meters and that the focus would be on closed barrier systems, meaning that end connections to
other structures will not be explored. End connections can be developed independently of the
rest of the system and are not needed in every flood situation, and can therefore be developed
at a later stage.

4.3 Failure mechanisms as design guidelines

Outlining the failure mechanisms and their significance

For temporary and demountable flood protection systems there are three main types of failure
that can occur. These are functional failure, structural failure and operational failure (Koppe &
Brinkmann, 2010). It is important to have a good understanding of each failure mechanism
when designing a flood protection system so that necessary precautions can be taken to avoid
them, either by knowing the limits of use, or by including certain features into the system that
prevent certain types of failure. It does not mean that the failure mechanisms should have a
large impact on early phase concepts, as they can often be addressed in later stages of
development, but the knowledge may however prevent expensive rework in the later stages of
development.

Table 4: failure modes

Failures of temporary and demountable flood barriers

Functional failure: Overtopping of flood barrier or excessive seepage through

protection system

Structural failure: Insufficient strength or stability

Operational failure: Failure of closure of flood barrier

The different failure modes are explained in the next sections, and the mechanisms that affect
the failure modes are explored in section 4.4.

Functional failure
This type of failure is based on pre-determined performances of the system that are not met. If,
for example, a system is designed and specified to a given maximum amount of seepage and
this is exceeded, it is considered a functional failure.

30
This failure mechanism is avoided by gaining a proper understanding of the environment the
flood protection system is operating in which demands appropriate hydrological and hydraulic
analyses, topographical surveys and an understanding of the seepage characteristics of the
surface the system is deployed. It is not possible to know exactly what environment the barrier
will be deployed in, but it is important to understand which parameters affect the seepage rates.

Structural failure
This type of failure occurs as a result of insufficient strength or stability. This can lead to
breaching, collapse, overturning, or sliding, and failure is defined as the point at which the
system is no longer able to meet its predefined performance.

Two main factors make structural failure more likely to occur in temporary flood barriers
compared to other forms of flood protection. The first is that mobile systems usually consist of
many parts and subcomponents, and a failure in one of these parts can often lead to a failure of
the system as a whole. The second factor is that mobile systems are usually not designed for
one particular place and usage, making it difficult to predict performance. Conditions could
further aggravate if the system is needed on short notice with insufficient time for detailed study
of the location and environment.

Operational failure
Operational failure is defined as a failure to fully erect the system before the flood reaches a
critical height, which is the height at which it is too late to set up a functional system. This can
either mean that the system was not closed off in time or not erected correctly. The risk of
operational failure will often be increased in temporary systems as they often need to be
mounted in dark or stormy conditions. This can be the result of very short lead-times before
expected arrival of a flood.

The four processes that are involved when installing a temporary flood barrier are forecast,
alert, mobilization and closure. If either one of these fail, the whole system fails. By improving
each of these stages when installing the flood barrier, the risk of operational failure is reduced.
Forecasting and alerting are independent of the system, but it is however possible to affect the
mobilization time and closure time through design choices.

31
4.4 Outlining the Key mechanisms
Structural strength
There is a considerable number of forces that are involved in a flood. In addition to the
hydrostatic pressure, there are forces from waves, currents, wind and impact from debris
floating in the water. The flood barrier as a whole needs to withstand all of these forces, meaning
that every component and sub-component also needs to take these forces into consideration. It
is not necessary to do detailed calculations and simulations early on in the development process,
but decisions about design should account for the possibility of needing to change dimensions,
material selection or other properties in order to strengthen the component if necessary.

Stability
In addition to having the structural integrity to withstand the forces exerted by the water, the
system is required to have sufficient stability to avoid the two structural failure mechanisms:
sliding and overturning (Figure 4.4). Sliding takes place when the friction at the tipping point
is too low, and overturning occurs when the friction at the tipping point is too high.

Sliding Overturning

Figure 4.4 Sliding and overturning

The stability of a module can be deconstructed into the forces acting on the base of the module,
and the forces acting on the wall of the module. The higher the forces are on the base compared
to the wall, the more stable the system is. Mechanisms for achieving this are outlined in Table
5.

32
Table 5 Mechanisms to increase the stability of a system

Mechanism Comments

Increasing the base area A simple solution that can be integrated into most systems.
However, this increases the width to height ratio, which
means it requires more space when in operation, limiting the
places where it can be erected.

Increasing the weight Increases the stability, but makes the system harder to
assemble for a small group of people. Can also increase the
cost of the system as it requires more material.

Adding external weights Useful when extra weight can be found on site, which is not
always the case.

Anchoring to the ground A lightweight solution that can increase stability

considerably. However, there are a limited amount of surfaces
that allow quick and simple anchoring. For example,
anchoring to asphalt is possible, but requires expensive tools.

Creating a seal against the underlying surface

One of the major challenges involved with temporary flood barriers is creating a proper seal
against the ground. To create the seal, rigid systems tend to use a gasket mounted underneath
the base, towards the front, while flexible systems tend to use a long skirt constructed of the
system material.

It was established that the sealing properties of a temporary flood protection system depends
on a series of variables that can be divided into two groups: circumstantial and system design.
The variables are outlined in Table 6.

33
Table 6 Seepage variables

Circumstantial System design

Foundation material Gasket material

Water pressure (flood depth) Gasket dimensions (width and height)

Seepage through foundation Pressure in gasket

The gasket material is best evaluated through testing as there are many options, each with
different properties. The dimensions of the gasket can affect the seepage rate as the narrower
the width is, the higher the resultant pressure on the gasket becomes, which in turn reduces the
seepage rates. The pressure on the seal can also be increased by increasing the force on the base
of the module, and can be affected by the same mechanisms as discussed in Table 5.

In freestanding barriers, the weight of the water is usually the most significant contributor to
the pressure on the gasket, which in return improves the sealing against the underlying surface.
This means that seepage rates are often very high at low water depths, which was made clear
during testing of AquaFence’s system outlined in section 3.6.1. This means that the force on
the base affects the seepage rates, but the seepage rates also influence the force on the base.
This is due to the fact that the stability that comes from the weight of the water, is dependent
on a pressure difference over and under the base of the module. If the seepage rate is high, there
is a lot of water underneath the base which can counteract out the hydrostatic forces, and
consequently reduce the stability of the system. Creating a good seal is therefore extremely
important, especially at low water depths.

34
 Flexibility
Since consumer flood protection systems should accommodate installation at various locations
with little or no modification, the system should be as flexible as possible. In this context
flexibility means that the system can adapt to uneven foundations as well as having the ability
to follow the terrain through bends, or go around corners. In Figure 4.5, the different modes of
flexibility between two modules are illustrated. The different modes of flexibility are
translations along each axis, and rotation around each axis. Increasing flexibility of each mode
corresponds to an increase in complexity for the system, so only the most important ones should
be addressed. Each mode of flexibility is therefore explained in detail in Table 7.

Figure 4.5 Different types of flexibility

35
Table 7 Evaluating the importance of each degree of freedom

Mode Importance Extent Achieved characteristics

Translati high low Affects “wiggle room” in the space between the
on x-axis modules. Allows the barrier to be closed without
planning out the position of each individual module
beforehand.

Rotation moderate moderate Allows adjacent modules to be placed at different

x-axis angled foundations, while maintaining sealing
capabilities.

Translati low moderate Allows adjacent modules to be placed with a distance

on y-axis between them, which can be an advantage in case of
obstacles.

Rotation high high Allows adjacent modules to be placed at different

y-axis angled foundations, while maintaining sealing
abilities.

Translati moderate moderate Allows adjacent module to be placed at foundations

on z-axis with different heights

Rotation high high Allows a module to be placed at an angle to the

z-axis adjacent module

36
Rotation around the z-axis and the y-axis have been identified as the most critical modes of
flexibility and needs to be addressed in the development of new concepts. If the maximum
rotation around the z-axis is large enough it is possible to create an enclosed system using only
one type of barrier. AquaFence’s current lineup allows for some rotation, but is dependent on
additional corner modules like the one seen in Figure 4.6. Being able to avoid the use of extra
modules can help drive the cost of the system down, as well as allowing for more flexibility
when erecting the system.

Figure 4.6 AquaFence V1200 outside Corner

4.5 Other considerations for development

User perception
When developing products for the consumer market, it is not enough to create a great product.
It is equally important to convince the customer that the product is good, and it is not necessarily
sufficient to back up a concept with numbers. The appearance of the product is a useful tool in
convincing the customer that the flood barrier is stable and secure, and that it can protect their
property. This can for example be done through material selection, as solid beams would
portray higher strength than a cable would, even though tensile strength could be equal.

Canvas has the advantage of being extremely lightweight and flexible compared to a rigid plate.
These advantages must however be weighed up against the perceived effectiveness of the
system, as canvas may be perceived as “flimsy”. Using canvas therefore demands extra
attention to the need of portraying security to the user.

37
 Ease of use
AquaFence’s current systems are designed for large applications, and require trained personnel
to ensure correct mounting. Since the new system is being targeted at consumers it means that
the system will be mounted without training. This means that the system will need to be
designed in a way that helps reduce the risk of operational failure. Steps that are likely to be
performed incorrectly for a concept should be identified and addressed accordingly.

By providing feedback to the user when mounting the system, the user can be assured that the
system is correctly mounted. Examples of feedback can be “torque-control” when tightening
bolts or the use of symbols to indicate which parts interconnect.

Improving the user-friendliness by making the system as simple as possible to set up can, assure
the user that the system is mounted correctly and will thereby increase the trust in the system.
However, this should not affect the effectiveness of the system in use. Making the system as
simple as possible would be beneficial in reducing the time needed to fully erect the system,
and limiting the likelihood of operational failure.

Production
As the flood barrier being developed is aimed at the consumer market, reducing the cost of the
system is critical. Since production cost is a large part of the total cost, efforts should be made
early on to come up with design decisions that allow for cheaper production methods. This can
be done, by among other things, reducing the number of components needed for each module,
simplifying the geometries, using inexpensive materials and focusing on effective use of
materials. Some aspects, like material selection, can only be made later on in development,
while other aspects, like reducing the number of components, can be done early on.

38
 Storage and Transportation
As one of the main advantages of temporary flood barriers is that
they can be stowed away when not needed, it is important to
design a system that allows it to be stored an effective way, and
be transported in an effective way to where it is going to be set
up. AquaFence solves this by having modules with hinges, so that
they can be folded into a flat shape and be stacked. Another way
to make storage and transportation easier with L-shaped modules
is by making them stackable (Figure 4.7) without having to fold
the modules first. Stacking like this does not require hinges or Figure 4.7 A stack of NOAQ-
Boxwall modules
other extra parts and is therefore a good option when making
simple modules.

Another concern that needs to be addressed when storing and transporting a system, is the
ability to clean it properly subsequent to use. Floods are often contaminated by sewage and
other waste, and thus the modules need to be designed in a way that allows proper cleaning
after use. Complex or hidden geometry should therefore be avoided in order to allow for easy
cleaning.

4.6 Summary of design restrictions

Table 8: summary of design restrictions

Quantitatively Qualitative

Protect against 0 m to 0.6 m flood depths Low cost system

Freestanding barrier Compact when not in use

Maximum two people required for installation High flexibility

No advanced tools or machinery required Fast installation

Closed barrier Build user trust/confidence

Hybrid of rigid and flexible system Avoid operational failure

Avoid structural failure

Avoid functional failure

39
40
5 Gaining and Capturing Knowledge
5.1 The value of capturing knowledge
There are two main reasons why capturing the knowledge is useful. Firstly, the development
process is often long, and if the knowledge acquired early on is not captured, it is likely
forgotten later when it may be useful. This means that the time spent gaining the knowledge
was wasted and that value was later lost. The second reason for capturing knowledge is to avoid
expensive rework. Processes are often repeated several times during development, and if
relevant knowledge is lost between each time, unnecessary rework is needed.

For flood barriers, there are several aspects that apply to any system. For example, as almost
all systems on the market today implement a gasket, it is reasonable to expect that the new
system also will implement a gasket. Therefore, instead of redoing work about gaskets for each
concept that is being developed, knowledge about gaskets should be captured in a general way
that allows it to be applied to any system. This was the reason for performing the gasket
experiments, which is discussed further in section 5.4.

Another situation where it is valuable to be able to reuse knowledge is when choosing the
dimensions for certain concepts, as there are multiple variables that make up a virtually
unlimited number of combinations. It is not feasible to evaluate every combination individually,
but by establishing the general relationship between each property of the flood barrier, it is
possible to quickly evaluate the parameters without having to repeat the same process over and
over.

Several methods for capturing knowledge were used for this project. After brainstorming
sessions, a document would be created which quickly outlined what had been discussed and
which conclusions had been drawn, and photos would be taken off the whiteboard if this had
been used during the brainstorm. Building prototypes and performing experiments is another
useful tool used during this project as a way to both create and to capture knowledge (discussed
in Section 2). Also, a program for capturing knowledge was developed based on Knowledge
Based Engineering (KBE), as the main objective with KBE is to be able to capture engineering
expertise and to be able to present this to the engineer in an effective way when it is needed
(discussed in Section 5.2).

41
5.2 Developing a Knowledge Capturing System
The requirements

For a knowledge capturing system to be a useful tool in the development process, several
qualities the system needs to fulfill were outlined. These were as follows:

 The ability to apply the system to new flood barrier concepts

 Easy to customize the inputs and outputs of the system
 Present knowledge in a clear manner that is easily interpreted by a person
 Provide information that can be used for evaluating new concepts

As creating a program is a time-consuming process, it is important to make sure that the net
gain is positive. To ensure this, the program needs to be easily applied to any new flood barrier
concept to avoid creating a new program each time. Also, by being able to apply the program
to different concepts, it can be used to easily compare different concepts against each other.

As knowledge is gained and relationships between parameters are established, there needs to
be a simple way to implement this into the system. Also, new aspects that are of interest are
likely to arise during the development process so new parameters need to be simple to add or
change.

Being able to present the knowledge in a way that is useful means that it needs to be presented
clearly, not only for the person who created the program. For the program to be useful to the
development team, anyone who uses the program should be able to understand how to use it,
and what is being presented.

Lastly, it is important that the program provides information that can be used in the development
of new concepts. This can be, for example, to evaluate and optimize parameters, identify gaps
in knowledge that needs to be filled or comparing concepts to each other.

Developing the program

The program was created using Microsoft Excel using ‘Visual Basic Editor’. Microsoft Excel
was chosen because it is a powerful application that is familiar to almost everyone, allowing
the Knowledge Capturing System to implement a user-interface people already know how to
use. Also, because most people have access to Microsoft Excel, it means that the system can be

42
used without requiring users to purchase and install new software. Another advantage is that it
is easy to export data from Microsoft Excel to other applications.

The code for the program is included in Appendix A – Code for the Knowledge Capturing
System.

The main function of the program is to calculate properties of new flood protection concepts
based on inputs from the user, and then present it in a useful way. It was therefore decided to
divide the program into the following six areas:

1. Controls
2. Constants (Input)
3. Variables (Input).
4. Output (Output)
5. Graphs (Output)
6. Data points (Output)

A Controls area (Figure 5.1) was established to let the user select the information that they wish
to display. The user can select which parameters will be plotted using drop down menus. The
program can be updated and reset using the “Update” and “Clear” buttons.

Figure 5.1: The Controls area

Input was divided into the two areas: Constants and Variables. This was done in order to clearly
separate the values that can be affected by design (like the dimensions) and those that cannot
(like gravity and the density of water). The user can input a range for the parameters in the
Variables area by inserting the lower limit into the “Value (0)” column and the upper limit into
the “Value (n)” column.

Output was divided into the three areas: Output, Graph and Data Points, where each area
presents the information in a different way. The Output area shows the values that have been
calculated for each property based on the values from the input areas. This is useful when values
of properties of specific configurations are of interest. The Graph area acts as a visual
representation, which is useful to see how changing one parameter affects and for identifying
critical points. Also, because it allows multiple graphs to be drawn on top of each other, it is a
43
useful tool for quickly comparing two or more situations against each other. The Data point
area is useful for looking closer at specific points of the graph, or for exporting the data to
another program.

The program was created in a such a way that customization is made easy by allowing the user
to insert new rows to any of the three tables without needing to worry about the programming
(Figure 5.2). This was done by the program “reading” where each table starts and where each
ends so that each table is treated correctly when performing the calculations. The relationships
can be added in column B of the output area in the same way that functions are added to cells
in Microsoft excel. To use values from any of the other areas, the cells in column B are used.
The different formulas in column B are included in Appendix B - Equations used to establish
relationships for Tetris.

Figure 5.2 Screenshot illustrating how relationships between parameters are added

44
 How to use the system

Figure 5.3: User interface

The program (Figure 5.3) is used in the following way:

1. Choose the relationship that is going to be explored by selecting the x-axis and the y-
axis parameters in the drop down menu
2. Fill in the values in the “Constants” table
3. Fill in the values in the “Value (0)” column in the “Variables” table
4. Fill in the values in the “Value (n)” column in the “Variables” table. This is only for the
parameters that span a range, for example see how the stability of a system changes as
the flood depth increases from one level to another, and is otherwise left blank
5. Press the “update” button to calculate the values in the “Output” table and to plot the
graph. The values in the “Value (0)” and the “Value (n)” column of the “Output” table
corresponds to the values in the columns of the same names in the “Input” table
6. Steps 2-5 can be repeated to compare different plots on the same graph
7. Press the “clear” button to reset the program. This clears the graph and the values in the
“Variables” and “Output” table.

Establishing the relationships

Changing the parameters of a design will often have an effect on multiple aspects of a design.
For example, increasing the thickness of a structure will increase the weight and cost while also
affecting the structural properties like stiffness and strength. Some of the relationships are

45
clearly defined, like the effect the flood depth will have on water pressure where the two are
linked by the formula for hydrostatic pressure (P=ρgh). Other relationships are more
complicated, like how seepage rates are affected by the gasket material. It these cases it is easier
to establish the relationships through experimental testing.

When performing experiments to establish relationships it is important that the knowledge

created is reusable. By breaking down the product into subsystems and isolating each subsystem
for testing, the results can be applied to many concepts. Two subsystems were identified that
apply to the type of barriers that are going to be developed that require experiments to establish:

 Contribution of canvas towards stability (Explored further in section 5.3)

 Seepage rates for different gaskets (Explored further in section 5.4)

Limitations
The program is meant as a quick way to evaluate the different concepts during the development
process and is not meant as a way to certify different systems as the relationships established
are only approximations. There is also the possibility that some relationships are defined
incorrectly, so further testing should be conducted to verify the relationships.

5.3 Investigating the contribution of canvas towards stability

Loss of stability and gasket pressure when using canvas
Using canvas does introduce some complications regarding
the stability and sealing of a system. When the canvas rests
against the foundation the hydrostatic pressure transfers to
the ground rather than supporting the structure. The area of
the canvas in contact with the ground will not contribute to Figure 5.4 Rigid plate vs canvas

stability or gasket pressure, (see Figure 5.4).

Therefore, the stability of the system increases as the canvas is elevated from the foundation.
This can done by tightening the canvas, raising the attachment points or distributing the rigid
area (see Figure 5.5).

Figure 5.5 Ways to increase contribution from canvas towards stability of a module

46
 Designing the experiment
Water supply
Generalized module Scale Ratchet strap

Flood Bucket

Flood water

Raisers

Figure 5.6 Diagram of the test setup

Defining the area of canvas resting on the ground is dependent on multiple variables which are
difficult to define as the canvas is not a rigid material. Initially, a conservative assumption was
made that there was no contribution from the canvas at all, meaning that stability is only
achieved through the rigid components of the barrier. To gain a better understanding of how
the canvas contributes to stability, a test setup was designed (Figure 5.6).

The setup consists of a generalized freestanding module (Figure 5.7) that is mounted in a flood
bucket. A ratchet strap is attached to the top of the module that can be tightened to increase the
force until the front of the module is lifted. The force applied is measured with a digital scale
that is attached between the module and the ratchet, and based on the magnitude of this force it
is possible to estimate the contribution of the canvas towards the stability by performing a
moment calculation around the tipping point.

Raisers were added to the module to test how lifting the canvas affects stability.

47
 Figure 5.7 The generalized Figure 5.8 Backside of base, Figure 5.9 Module rigged in test
Module attachment of raisers in progress tank

The following test procedure was performed:

1. Add raisers to the base of the module (see Figure 5.8)

2. Mount the module in the test tank (see Figure 5.9)
3. Attach the ratchet strap to the module
4. Fill the test tank with water up to the desired water level
5. Zero the scale
6. Tighten the ratchet strap until the front of the module starts to lift
7. Record the force displayed on the scale
8. Repeat steps 4-6 three times for more accurate measurement
9. Repeat steps 1-8 for each canvas attachment height

Results

All of the canvas was above the foundation when the attachment height was 24 mm (only one
layer of raisers), so the experiment was only performed for two canvas attachment heights.
Because the hydrostatic forces on the rigid parts of the modules are simple to calculate, the
contribution was calculated by finding the sum of moments around the tipping point. The results
are presented in Table 9.

48
Table 9 Results

Results
Canvas
Attachment height [m] Water depth [m] Ratchet Force [N]
Contribution [Nm]
0.05 11.8 23.8
0.10 109.3 45.1
0,012
0.15 200.9 71.8
0.20 299.0 99.3
0.05 394.4 33.1
0,024 0.10 11.8 60.6
0.15 122.6 94.4

Since the canvas was completely lifted at an attachment height of 24 mm, the canvas
contribution is 100% of a rigid plate of the same area, and at 12 mm, the contribution is roughly
75%. From this, the first conclusion that can be drawn is that canvas contributes considerably
to stability, even when the attachment point is relatively low. This was unexpected, especially
since the canvas was not attached particularly tight. This indicates that the use of canvas in
freestanding consumer flood protection systems may be a viable solution.

The outcome of the experiment shows a noticeable increase in the moment contribution as a
result of an increase in canvas height. This supports the idea of lifting the attachment point as
a way to increase stability on modules consisting of canvas.

Limitations
This was a quick experiment set up to learn more about whether freestanding modules that
incorporate canvas can depend on stability by more than just the rigid sections. The measured
values and calculations are indicative as there are factors that may affect the results that were
not controlled for. The results are not meant to be used to determine properties of a system
accurately. However, later on in development when promising concepts have been created and
need to be refined, a modified version of this experiment would be useful. It could be used to
determine the exact contribution from the canvas and the overall stability of various
configurations.

49
5.4 Investigating the seepage rates for different gaskets
Establishing the variables

Most systems incorporate a gasket in their design to prevent water from seeping underneath the
barrier. It is therefore useful to gain an understanding of how a gasket performs in a given flood
situation, and how to obtain the best seal. Instead of performing tests on each new flood barrier
concept that developed, the gasket can be isolated as a subsystem. The results from the
experiments that are performed can then be applied to new concepts without further testing.

Seepage Pressure

Gasket Pressure

Figure 5.10 Gasket pressure and seepage pressure

As the seepage rate is the most relevant aspect of a gasket it is important to understand which
variables affect the seepage rate, so that they can be controlled accordingly. An hypothesis has
been established maintaining that the main variables that affect the seepage rate is the “Gasket
pressure” and the “Seepage pressure”. The Gasket pressure is defined as the combination of
forces acting over the area of the gasket against the underlying surface. The main forces
contributing to the gasket pressure are the water pressure (determined by the depth of the flood
water), the weight of the module, and any additional weight placed on top of the barrier. The
seepage pressure is defined as the hydrostatic pressure at the gasket that causes seepage as
opposed to the pressure of the water that has seeped through. This is illustrated in Figure 5.10.

These are therefore the two independent variables that can be changed (gasket pressure and
seepage pressure) to affect the dependent variable: seepage rate. A gasket experiment will
therefore need to allow these two independent variables to be changed, while at the same time
measuring the seepage rate.

50
 Designing the experiment
This experiment was designed as a quick way to quantify the relationship between the seepage
rate, the gasket pressure, and the seepage pressure for gaskets on different ground surfaces.
Furthermore, it was designed in such a way that the information learned could be used for any
type of flood barrier that uses a gasket.

Water supply Measuring tape

Flood bucket

Flood water

Added weight

Gasket plate

Collection bucket

Surface plate

Figure 5.11 Diagram of the test-setup

Figure 5.12 Top and bottom of the surface plate Figure 5.13 Top and bottom of the gasket plate

The setup (Figure 5.11) consists of a container, called the flood bucket, which is filled up with
water to simulate a flood of a certain depth. In the bottom of the flood bucket is a hole that is
covered by the gasket plate. The gasket plate is a solid plate that has the gasket being tested
attached around the perimeter of its base (Figure 5.13) so that the water that comes out of the
hole in the bottom of the flood bucket is the seepage rate for that gasket. The seepage pressure
is controlled by varying the depth of the water in the flood bucket, using the measuring tape
attached to the side of the flood bucket as a guide. The gasket pressure is controlled by a
combination of the flood depth and added weight placed on top of the gasket plate.

Because flood barriers are deployed on different types of surfaces, it is useful to be able to
perform the experiment for different surfaces without having to build a new test setup for each
51
surface. A plate called the surface plate (Figure 5.12) that can easily be mounted and demounted
is therefore added between the base of the flood bucket and the gasket plate (see Figure 5.14).
It has a hole in the middle for seepage and is sealed around the edges to make sure that the only
water that comes out the bottom of the flood bucket comes from the seepage between the gasket
and the surface plate.

Gasket plate

Surface plate

Flood bucket (base)

Figure 5.14 Configuration of gasket plate and surface plate

The seepage water is collected in a container, called the collection bucket. The collection bucket
is placed underneath the outlet for a set amount of time, after which the amount of water in the
collection bucket is measured to calculate the seepage rate (see Figure 5.15).

Flood bucket

Collection bucket

Figure 5.15 Experiment setup

52
 Test Method
1) The test setup was prepared by attaching the surface plate to the flood bucket, assuring
that seepage only occurs through the center hole, and not between the seal plate gasket
and the flood bucket. This was done by taping the hole with a strong water-proof tape
and filling the flood bucket with water. As no seepage occurred, the surface plate was
attached correctly.
2) The gasket plate was placed over the hole in the surface plate and the desired amount of
weight was placed on top of the gasket plate.
3) The flood bucket was filled with water from the water source until the water depth was
at the desired level.
4) The collection bucket was placed under the flood bucket to begin the experiment. The
flood bucket was refilled during the experiment to compensate for the seepage, using
the measuring tape as help to maintain a constant flood depth.
5) After 2 minutes, the collection bucket was removed from underneath the flood bucket
to stop recording seepage.
6) The amount of water in the collection bucket was measured by weighing the bucket.
7) Steps 2-6 was repeated for the different flood depths.
8) The gasket plate was changed and steps 2-7 were repeated for each gasket plate.

Data collection
The four gaskets tested were all made from the same foam material, but with different
thicknesses (0.8 cm and 1.8 cm) and widths (3 cm and 6 cm) and an outer diameter of 30 cm.
The gaskets were:

 Narrow (3 cm) and Thin (0.8 mm)

 Narrow (3 cm) and Thick (1.8 mm)
 Wide (6 cm) and Thin (0.8 mm)
 Wide (6 cm) and Thick (1.8 mm)
The gaskets were tested at flood depths of 0.1 m increments up to 0.6 m, with additional
measurements at 0.05 m and 0.15 m due to large variations at low flood depths.

The gasket pressure was varied by adding the following weights:

 0 kg added weight
 4 kg added weight
 8 kg added weight

53
 Results
The raw data is presented below (Figure 5.16) in four charts, one for each gasket tested.

Narrow and Thick Narrow and Thin

8 8
7 0 kg 7
4 kg

Seepage Rate [l/min]

Seepage Rate [l/min]

6 6
5 8 kg 5
4 4
3 3 0 kg

2 2 4 kg

1 1 8 kg

0 0
0 0,2 0,4 0,6 0 0,2 0,4 0,6
Flood Depth [m] Flood Depth [m]

Wide and Thick Wide and Thin

8 8
7 0 kg 7
4 kg
Seepage Rate [l/min]

Seepage Rate [l/min]

6 6
5 8 kg 5
4 4
3 3 0 kg
2 2 4 kg
1 1 8 kg
0 0
0 0,2 0,4 0,6 0 0,2 0,4 0,6
Flood Depth [m] Flood Depth [m]

Figure 5.16 Results

What is immediately clear when interpreting the data is that adding weight reduces the seepage
rate, especially at very low flood depths (0.05 meters). This suggests that adding weight is most
effective at reducing seepage rates at low water depths.

It is also apparent that a thick gasket yields the best results in terms of seepage rate as it
converges at roughly 2-4 liters per minute for the “narrow and thick” and the “wide and thick”
gasket, while ending up at roughly 4-5 liters per minute for the thin gaskets. This can be
explained by the thicker gaskets allowing for larger deformations to “fill in” the ground surface
to create a seal.

54
For the thick gaskets, the narrow gasket yields lower seepage than the wide gasket. This is as
predicted, since a narrow gasket leads to a higher gasket pressure, while the seepage pressure
remains constant.

The main reason for performing this experiment was to investigate and establish a relationship
between the seepage rate, the gasket pressure, and the seepage pressure. To get more data points
to work with, the data from the ‘narrow and thick’ and ‘wide and thick’ gaskets were combined
and plotted against the ratio: gasket pressure/ seepage pressure (Figure 5.17) to see if there is a
relationship.

Thick Gasket
6

5 0.20 m

0.30 m
Seepage rate [l/min]

4
0.40 m

0.50 m
3
0.60 m
2

0
1 1,5 2 2,5 3 3,5 4 4,5 5
Gasket Pressure/ Seeepage Pressure Ratio

Figure 5.17 Relationship between seepage, gasket pressure and seepage pressure

The trend lines in Figure 5.17 seem to be parallel to each other, which suggests that there is a
direct relationship between the seepage rate and the gasket pressure/seepage pressure ratio. The
plan was to redo the test to eliminate error in the data. This would be done by, among other
things, doing multiple recordings at each flood level, adding more weight to get a bigger span
of results, and testing all water depths for each loading case before adding more weight. Adding
weight and removing weight during the experiment might have a significant impact as it
compresses and decompresses the gasket and might not give it sufficient time to return to its
new equilibrium. However, it turned out that time had a greater impact on the results than
anticipated (see section 5.4.6) and the experiment therefore needed to be repeated.

55
 Developing an improved test set up
It was discovered that the seepage rates were time-dependent, so the initial experiment was
conducted again but with 15 minutes between each recording to allow the seepage rates to
stabilize. When it was discovered that the seepage rate continued to decline after 15 minutes, it
was deemed necessary to redesign the test setup. Since the time needed for stabilization was
potentially long, it was deemed beneficial to create an automated set up that conduct the
experiment and collect data automatically.

The new test set-up would need improvements on two aspects of the initial test setup:

 Maintaining a constant flood depth

 Periodically recording the seepage rate.

In the initial test setup, the flood depth was kept constant by monitoring the depth visually, and
adjusting the rate of water supply into the bucket accordingly. This method required constant
monitoring and was therefore not suitable for maintaining a constant flood depth in an
automated setup. The seepage rate was measured by periodically weighing the seepage water
over a given amount of time. This is a labor-intensive process that requires multiple steps, and
it is therefore not suitable for the new setup.

The new setup (see Figure 5.20) was built around an Arduino controlling the rate of water
entering the flood bucket and recording the seepage rates (see Figure 5.21). The code used is
included in Appendix C – Arduino code for automated gasket experiment.

Figure 5.18 Float switch Figure 5.19 Solenoid valve and flowmeter attached
to water supply

A float switch was built (Figure 5.18) by attaching a foam material on a lever to a toggle switch
that when triggered sends a signal to an input switch on the Arduino. This lets the Arduino
know if the flood level is higher or lower than the desired depth, allowing it to adjust the water

56
level by turning on or off a solenoid valve (an electromechanically operated valve that switches
on and off by an electric current) (see Figure 5.19) connected on the hose supplying water to
the flood bucket. The solenoid valve operates at 12V so it is controlled using a relay switch.
The hose supplying water to the flood bucket is connected to a faucet so the Arduino can control
the amount of water supplied.

As the flood level stabilizes on the desired depth, the frequency that the solenoid valve is turned
on and off increases. Initial testing of the system had the solenoid valve turning itself on and
off more than once every second. To reduce the wear on the solenoid valve, a delay was
incorporated to limit the frequency. This is the reason the float switch was not connected
directly to the solenoid valve as was originally the plan, but rather through the Arduino.

There was no quick and easy way to measure the amount of seepage. However, since the flood
level is being kept constant by adding water to the flood bucket, the amount of water going into
the system is equal to the amount of water going out of the system. A flowmeter was therefore
added to the water supply hose and connected to an input pin on the Arduino.

Flowmeter

Solenoid valve Water supply

Measuring tape
Float switch
Flood bucket

Flood water

Added weight

Gasket plate

Surface plate

Figure 5.20 Diagram of test setup

57
 Solenoid valve Flowmeter

Toggle switch
(Float switch)

Relay switch

12V Power supply

Arduino Uno (Microcontroller)

Figure 5.21 Layout of Arduino components

58
 Results
The raw data from the experiment is presented on the left in Figure 5.22 on the next page, and
is divided into the graphs, one for each of the three different depths tested. For each depth, three
plots are shown, one for each of the three load cases (0kg, 10kg and 20kg added weight). As
expected, an increase in weight correlates to a decrease in seepage rates. This is most clearly
seen at a flood depth of 0.2 m, while not as clear for the two other depths. This is because the
pressure on the gasket comes from the water pressure and the added weight, and as the flood
depth increases, the amount of contribution from the added weight decreases relative to the
water pressure. At 0.2 m the gasket pressure increases by 53% when using 20kg added weight
compared to 10kg, while at 0.6m the gasket pressure increases by only 23%. In addition to this,
the seepage pressure is three times higher at 0.6m than at 0.2m which is why the effect is more
significant. This can be seen in the combined chart, where the highest seepage rate and the
lowest seepage rate are both at a flood depth of 0.2m.

The graphs on the right side of Figure 5.22 present the same data as on the left, but in
standardized form by multiplying the seepage rate with the gasket/seepage pressure ratio. This
means the data shows what the seepage rate would be expected to be given that the
gasket/seepage pressure ratio was 1. For this assumption to be correct, it would mean that the
curves for the seepage rates would all be the same in the standardized form, which seems to be
correct. This can be seen when comparing the combined graph of the standardized data to the
raw data in Figure 5.23.

The seepage rate in the standardized graphs are also presented as “per meter”. The seepage rates
were divided by the length of gasket used, so the seepage rate is per 1 minute, per 1 meter. The
expected seepage rates for a system can then be estimated by multiplying the value with the
length of barrier being used, and by the time.

59
 7 10
0,2m 0,2m (standardized)
9

Standardized Seepage Rate [l/min/m]

6
8
0kg 0kg
5 7
Seepage Rate [l/min]

10kg 10kg
6
4
20kg 20kg
5
3
4

2 3
2
1
1
0 0
0 20 40 60 80 0 20 40 60 80
Time [min] Time [min]

7 8
0,4m 0,4m (standardized)
6 7
Standardized Seepage Rate [l/min/m]

0kg 0kg
6
5 10kg 10kg
Seepage Rate [l/min]

5
4 20kg 20kg
4
3
3
2
2

1 1

0 0
0 20 40 60 80 100 120 0 20 40 60 80 100 120
Time [min] Time [min]

7 8
0,6m 0,6m (standardized)
0kg
Standardized Seepage Rate [l/min/m]

6 7 0kg

10kg 6 10kg
5
Seepage Rate [l/min]

20kg 5 20kg
4
4
3
3
2
2
1 1

0 0
0 20 40 Time60
[min] 80 100 120 0 20 40 Time60[min] 80 100 120

Figure 5.22 Results

60
 Combined Combined (standardized)
7 10
0kg (0,2m) 0kg (0,2m)
9 10kg (0,2m)

Standardized Seepage Rate [l/min/m]

6 10kg (0,2m)
20kg (0,2m) 8 20kg (0,2m)
0kg (0,4m) 0kg (0,4m)
5 10kg (0,4m)
Seepage Rate [l/min]

10kg (0,4m) 7
20kg (0,4m) 20kg (0,4m)
4 0kg (0,6m) 6 0kg (0,6m)
10kg (0,6m) 10kg (0,6m)
20kg (0,6m) 5 20kg (0,6m)
3
4
2 3
2
1
1
0 0
0 20 40 60 80 0 50 100
Time [min] Time [min]

Figure 5.23 Combined

10

0kg (0,2m)
9
10kg (0,2m)
20kg (0,2m)
8
0kg (0,4m)
Standardized Seepage Rate [l/min/m]

7 10kg (0,4m)
20kg (0,4m)
6 0kg (0,6m)
10kg (0,6m)
5 20kg (0,6m)

0
0 20 40 60 80 100
Time [min]

Figure 5.24: Combined (Standardized)

61
The data points in the standardized graphs do not line up perfectly which could be because it is
not a perfect assumption to make, or from errors in the measurements, or a combination of the
two. However, they line up close enough so that it can be used as an estimate, and by creating
an upper and lower limit (represented by the black lines in the graph in Figure 5.24), it is
possible to estimate a range that the seepage rates will fall into.

The expected seepage rate is calculated by finding the standardized seepage rate from the graph
and dividing it by the gasket/seepage pressure ratio.

For example, the seepage rate at 60 minutes is calculated by finding the upper and lower limit
at 60 min in the graph above (shown by the dotted line). The standardized seepage rate’s upper
value is roughly 4.3 and the lower limit is roughly 3.2. If the Gasket/seepage pressure ratio is
2, then the predicted seepage rate is between 2.15 l/min/m and 1.60 l/min/m.

It is often more useful to see what the seepage rates will be after they have stabilized. Figure
5.25 shows the stabilized seepage rate for the gasket/seepage ratio. The graph is constructed by
averaging the ten last data points of each data set and plotting them against the gasket/seepage
pressure ratio. Again, the points do not match up perfectly, so an upper and lower limit
(represented by the black lines) is added to define a range the seepage rate is expected to lie
within. Since this graph shows the seepage rates not in standardized form it means the values
can be read directly from the graph.

2,5
Seepage Rate [l/min/m]

1,5

0,5

0
0 0,5 1 1,5 2 2,5 3 3,5 4
Gasket/Seepage Pressure Ratio

Figure 5.25 Time stabalized and standardized

62
 Limitations
Only one gasket and one surface has been tested for this project. When using this experiment
to determine which gasket would be the best, it is necessary to test multiple gaskets on different
surfaces.

There are a lot of factors that affect the seepage rate, and perfectly replicating each scenario in
a lab is not feasible. The results are also only valid within the range tested. Even though it looks
like the data can be extrapolated, there might be unknown effects as the gasket/seepage pressure
ratio goes towards zero or as the value gets very high. If the area of interest lays outside of the
data collected, it is recommended to do more testing.

The experiment has only been performed once per data point, meaning there might be sources
of error not discovered. For more reliable data, the experiment should be performed multiple
times.

The experiment also does not take ground-water seepage into consideration, which on some
surfaces can be a considerable factor. In this case, geological analysis is required to get an
estimate for seepage rates, as it is not possible to test with the current test set up.

The seepage rates calculated in this experiment have been for gaskets that have not been
exposed to water or pressure over an extended period of time, which would be expected for a
gasket used in an actual flood protection system. It is therefore necessary to gain an
understanding of how the gasket behaves over longer time intervals than done with this
experiment. Gaining information about the life expectancy of the gasket is crucial if it is to be
used in a finished product.

63
64
6 Concept Development

6.1 Resources
Prototyping equipment
Most of the physical prototypes were built using simple materials like paper, cardboard and
plywood. To enable testing with water, a PVC canvas was obtained. In addition to having access
to basic tools, a CNC wood router was available to use. CAD models were created with Siemens
NX.

Size template
Physical prototypes are easier to relate to, compared to drawings and CAD models. To get a
better understanding, two generalized prototypes (Figure 6.1) were built in full scale (0.6 m x
0.6 m). The prototypes were used as a tool for communication, and were used in brainstorms
for explaining new designs ideas. The understanding of several aspects of the system was
improved:

 The size of the system

 The magnitude of forces acting on a system
 Interaction between a system and the user
 Interaction between modules
 Feasible module length
 Indication of module weight

Figure 6.1: Cardboard template prototypes

65
 Universal test container
Conducting tests involving water for flood protection systems is a comprehensive task, and it
is therefore valuable to develop test rigs that can be reused for different concepts. As most
experiments consists of a container that is filled with water, a universal test container was
developed. The container was required to allow prototypes and test equipment to be mounted
in it, and needs to withstand the forces that occur at high water depths. It is critical that the
container is secure, as it will potentially hold hundreds of liters, which means the consequences
of a failure are high.

The first iteration of the test container (Figure 6.2) was built out of plywood sheets, reinforced
with wooden beams, and sealed with silicone. Measuring 0.4 m in width, the tank
accommodated testing of half scale modules (Figure 6.3). The tank was filled using a bucket to
move water from a buffer tank to the test tank (Figure 6.4). Seepage water would exit the tank
through a hole at the end, and fall into the buffer bucket that was placed underneath.

Figure 6.2 First iteration of a universal test Figure 6.3 Test of a prototype in the first
container iteration container

66
 Figure 6.4: First iteration of test rig in use

The test tank served as a tool to investigate sealing characteristics of early prototypes, as well
as the stability of single modules. However, several shortcomings were identified, and
suggested improvements for the next iteration of a universal test rig were identified:

 Accommodate testing of full scale modules

 Higher durability and not degrading when exposed to water over long periods of time
 Automated filling of water
 Accommodate implementation of tools for measuring water flow in and out of the tank

The second iteration of a universal test container (Figure 6.5) was built out of stainless steel
sheets that were welded together. The container was 1.5 m long, 0.8 m wide and 0.7 m high.
The bottom of the tank was rectangular to be able to test systems of two lengths by either
mounting the system across the length or the width.

A crib was built for the container that enabled the rig to be moved using a pallet lift. Also, by
raising the container off the ground it was easier to empty it through a drain pipe attached to
the base.

67
 Figure 6.5 Second iteration of a Universal test container

6.2 Early phase concept development

Early phase concept development was done through iterative brainstorms and prototyping
within the design space established in Section 4.6. Numerous ideas surfaced and several
concepts were explored. The most promising concepts were selected:

 Tetris (Section 6.3)

 Caterpillar (Section 6.4)
 FlexiFront (Section 6.5)

68
6.3 Tetris
First Idea and explanation of concept
The name “Tetris” developed from the structures resemblance to the blocks of the video game
with the same name.

The Tetris concept is a modular system based on

rigid L-modules where the amount of rigid structure
is minimized and canvas is used as the main material
for the barrier. This increases the maximum angle
that two adjacent modules can have to each other, as
they can be rotated without colliding. It was also
meant as a measure to reduce the weight of the
system, as the previously heavy rigid structure
would be replaced by canvas. Figure 6.6 Prototype of the first idea

The Tetris system is made out of two main components: the frame and the canvas. The frame
is the rigid part that supports the canvas and gives the modules its strength, while the canvas
acts as the barrier stopping water from passing through.

Figure 6.7 Different configurations for the structure of the Tetris concept

Simple prototypes were created from paper (Figure 6.6) to quickly visualize which
configurations seemed to work best. The concept evolved into consisting only of straight beams,
and several configurations were explored (Figure 6.7). The connection between the structures
could either be on the front (towards the flood) or on the back (towards the area being
protected). Connecting the structures on the back means the length of the back is constant,
which makes it is easier to strengthen the structure while still allowing the modules to rotate.
However, the front lengths would not be constant and this is where the gasket would need to
be. Because creating a seal that can vary in length is more of a challenge than strengthening the
structure, a front connected structure was chosen (Figure 6.8). Additionally, having the

69
connection at the front leads to an increase of base area when adjacent modules are placed at
an angle, meaning that the hydrostatic pressure on the rigid parts gives a greater contribution to
stability.

Figure 6.8 A CAD-model of the first iteration of the Tetris concept with front connection

Proof of concept
As rigid freestanding barriers rely on the rigid base plate for stability, and in the Tetris concept
most of this area is removed, the largest uncertainty was whether the system would be stable
enough.

To investigate the feasibility of the concept in terms of stability a conservative calculation was
made. It was assumed that the hydrostatic pressure exerted on the base canvas gives no
contribution to the stability (see section 5.3.1). The stability calculation was done for a
maximum water depth, by calculating the moment around the tipping point of the module.

The calculated moment was positive, meaning that the simplified module would be self-
stabilizing. The width of the structure was chosen arbitrarily and can be varied for a more stable
structure, but the calculations indicated that it is possible to construct a flood barrier based on
the Tetris concept that would be stable enough.

70
 Physical prototype
Two identical physical prototypes were built (Figure 6.10) as a tool to visualize the concept and
to explore the interaction between two modules. The prototypes were built in a 1:2 scale to
make them easier to work with, while still being large enough for useful testing later on. A
connection between the frames was developed (Figure 6.11), based on the ‘tabs and blanks’ of
a jigsaw puzzle piece, that allows the modules to pivot while still being connected to each other.

Figure 6.9 Routing plywood parts Figure 6.10 The assembled Figure 6.11 Close up of connector
Tetris prototypes

The prototype was designed as a CAD model, and using a CNC woodworking router the pieces
were cut out of a 12mm plywood sheet (Figure 6.9) and glued together. A woodworking router
was chosen to build the prototypes because they could be built quickly and with high accuracy,
at a low cost.

Due to the asymmetric nature of the Tetris module there was a concern that it would be
diagonally unstable, but the physical prototype indicated that this was not a problem. Another
concern was how the canvas would behave with regards to buckling. The prototype was
modified by attaching a fabric to the modules (Figure 6.12), and testing showed that buckling
was not an issue. The fabric was later replaced with a PVC canvas (Figure 6.14) to more closely
resemble a material that would be used in a final product.

Figure 6.12 Module with Figure 6.13 Interaction between two Figure 6.14 A later prototype
fabric modules with strengtheners with PVC canvas

71
 Canvas fastening mechanism
The canvas from two adjacent modules needs to be connected in a way that creates a continuous
seal along the flood barrier. In AqF’s current systems, the modules are connected to each other
with a section of canvas that is clamped down to the adjacent module. This securely fastens two
modules together as well as adding flexibility to the positioning as the canvas can be bent.
However, for the Tetris module the canvas needs to be as stretched out as for the canvas to
contribute to the stability of the module. The clamping mechanism in AqF’s current system is
also relatively complicated as it consists of many additional parts, which increases the cost and
risk of operational failure.

Different ways to create a flexible connection between the modules were brainstormed. It was
concluded that the best mechanism would allow the canvas to attach to any point to the
following module, thereby allowing the canvas to remain fully stretched, regardless of the angle
between modules.

Three general fastening mechanisms were suggested:

 Adhesive: “Sticking” the two canvases together with some type of adhesive. For
example, static forces, hook and loop (Velcro) or a non-permanent adhesive
 Cam Locking: Cam mechanisms that easily allow the canvas to be tightened, but
prevents the canvas from moving in the opposite direction
 Clamping: Attaching the canvas by pressing the canvas directly onto the next module

Several ideas involving cam locking were explored but no solutions that could accommodate
the angle created between the base plate and the wall plate was discovered. The cam locking
concepts were therefore rejected.

Adhesive based solutions would not be greatly affected by the corner, and would allow quick
set-up times, but issues finding an appropriate material that would be strong enough while also
being reusable, meant these ideas also were rejected. Precautions would also need to be made
to prevent dirt from coming in contact with the adhesive during storage and set-up. A hook-
and-loop mechanism is less affected by dirt, but it is difficult to create a watertight seal using
this mechanism.

72
A clamping mechanism was chosen as the most promising mechanism for connecting the
canvas of one module to the next, mainly due to its simplicity. A triangular piece was created
that could be placed over the canvas and tightened using straps (see Figure 6.15). The straps
were extensions of the connected canvas, making the design simple, yet effective.

Figure 6.15 Triangle canvas fastening mechanism

This fastening mechanism also contributes to the strength of the system, potentially removing
the need for extra support. Integrating functions in this way reduces the amount of parts needed
for the system, which reduces the cost of the system and makes the system easier to use and
store.

The design showed promising performance during testing (section 6.3.7), but further
improvements should be made to:

 Increase the pressure between the connecting canvases to reduce seepage

 Simplify user experience
 Improve the reliability and durability

73
To increase the pressure between the connecting canvases the triangle was redesigned into an
arch. (Figure 6.16). Tightening the strap will then push the triangle towards the corner and
distribute the forces more evenly over the canvas. A ratchet strap can be used to gain sufficient
pressure between the connecting canvases.

Figure 6.16 Suggested design with ratchet strap and a reshaped triangle

Frame joint
The original mechanism, which was based on a jigsaw puzzle, was meant as a temporary
solution to evaluate other aspects of the concept. However, this mechanism proved to be a
suitable solution for a connection between modules. This is because it is simple and does not
require extra components, which reduces the cost of the system. The mechanism is also easy to
understand, since it is familiar reference to most people.

As the joint is prone to seepage, a watertight sealing material will have to be applied to either
the male or the female part of the joint. In addition, the joint has to be tight enough to ensure
there is pressure on the sealing material. Figure 6.17 illustrates how a gasket could be applied
to the female part of the joint. In addition to sealing the joint, it will make the connection
between the modules firmer.

74
 Figure 6.17 Suggested seal in the joint

User feedback on the frame joint and fastening mechanism

To get initial feedback on how intuitive the combination of the proposed joint and fastening
mechanism was, a quick test was set up to see how two users, unfamiliar with the system, would
interact with the design (Figure 6.18). Two people were each given a pair of modules and were
instructed to connect the two modules together, without instruction on how to do it.

Figure 6.18 A person connecting two modules for the first time

Both test participants were able to set the system up correctly. However, confusion about the
orientation of the modules resulted in high set up time.

This test was used as a quick way to get feedback and only gave an indication for how intuitive
the configuration was. For more detailed feedback, more test participants should be recruited
and efforts made to ensure a diverse group is selected.

75
 Transport and storage
To decrease storage volume and increase mobility, the wall could be hinged, making the module
collapsible (Figure 6.19). Since the canvas fastening mechanism adds strength and rigidity to
system, the hinge mechanism can be made by separating the wall from the base, letting the
canvas act as a hinge. This way, adding a hinge mechanism makes the design less complicated,
by avoiding the need for joining of the wall and the base during production. Collapsible modules
allow for compact stacking (Figure 6.20), which is useful for storage and transportation.

Figure 6.19 Collapsible Tetris module

Figure 6.20 AquaFence modules stacked for storage and transport

76
 Directional Prototype
The feasibility of the concept was tested by building a comprehensive prototype and performing
a water test (Figure 6.21). The purpose of the prototype and the test was to investigate whether
or not to continue the development in the direction of a Tetris system, and to look for
unexpected outcomes.

Figure 6.21 Test of directional prototype

The most crucial property to investigate at this point was stability. There was some uncertainty
linked to the way the system was mounted, and to what extent this would provide extra stability
to the system. However, the stability was deemed “good enough” to invest further work in the
concept.

The test revealed the following:

 The flexibility in the joint enabled each of modules to be placed at uneven foundations
(rotation in y-axis).
 The gasket was too thin to properly seal against the uneven foundation in the test tank
 All of the canvas was elevated from the foundation
 There was no leak detected through the connection between the modules.

77
 Analysis of the frame dimensions
The structure of the system can be divided into three sections (Figure 6.22):

 Wall
 Middle
 Front
The length of the wall is determined by the maximum
flood protection height (0.6m). The width of the wall is

0.6
not believed to have any major effects on stability but
wall
does have an impact on the strength of the system.
Therefore, the width of the wall will need to be
determined based on a structural analysis of the module.
middle front

The length of the middle section has a great impact on

the overall stability. Making the middle section longer
increases the base area, and consequently the Figure 6.22 Tetris frame and its three sub-
components, with initial dimension [m]
contribution from the hydrostatic downforce increases.
This is verified by the knowledge capturing system in Figure 6.24. However, effort should be
made to keep this length as short as possible to minimize the base to height ratio, and to reduce
storage volume. The width does not affect stability and needs to be based on a structural
analysis. The thickness needs to be high enough to raise the canvas off the ground (see section
5.3.1)

Increasing the length of the front section (Figure 6.23)

reduces the number of modules, which means the
number of connections are also reduced. This improves
the set up time, and reduces the number of complex parts
per meter of barrier, which in return reduces the cost.
However, the longer the front section is, the fewer wall
sections exist, meaning each wall section will be subject
to higher forces. It also affects the flexibility of the Figure 6.23 Tetris module with 1.2 m front
system as there are fewer pivot points.

78
 9
VARIABLES
8 Description Value (0) Value (n) Units
Safety factor (stability)

7 Wall length 0,6 [m]

Wall width 0,12 [m]
6
Middle length 0,1 0,8 [m]
5 Middle width 0,12 [m]
4 Front length 0,6 [m]
3 Front width 0,12 [m]
Frame thickness 0,02 [m]
2
Sealant width 0,04 [m]
1 Sealant thickness 0,02 [m]
0 Overlapping canvas 0,3 [kg]
0 0,5 1 Added weight 0 [kg]
Middle length [m] Flood depth 0,6 [m]

Figure 6.24 Graph showing the effects of middle length on stability (left) and variables used for the calculation (right)

Estimating seepage rates

The seepage rates of a Tetris system can be estimated using the knowledge capturing program
(Section 5.2.3), which has incorporated the results from the automated gasket experiment
(Section 5.4.7). The estimate is based on time-stabilized seepage and the seepage rate on the
gasket pressure.

The estimated seepage rates do not account for seepage through/beneath the surface nor leaking
through the system (connection between modules). The rates presented in Figure 6.25 will
therefore be a low estimate.

2,5

2
Seepage [l/min/m]

1,5

0,5

0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7
Flood depth [m]

Figure 6.25: Estimated seepage rates (blue graph: no external weight, green graph: one sandbag per module)

79
The blue plot represents the Tetris system without any additional weight. The seepage rates is
approximately 1.5 l/min/m, which means that a small gasoline-powered water pump, which is
low cost solution, would be able to pump seepage water out of an enclosed barrier stretching
over several 100 meters.

Sandbags can be used to increase the gasket pressure. By assuming that a sandbag (17kg)
increases gasket pressure evenly, the expected seepage rate is seen as the green plot in Figure
6.25. The improvement is considerable, especially for low flood depths. In addition, the stability
of the system will improve because of the added weight. Using sandbags to improve overall
performance should therefore be considered. The negative aspects of using sandbags have been
discussed in section 3.4, but the amount needed for the extra stability is negligible compared to
the amount needed for a sandbag barrier.

6.4 Caterpillar
First idea and explanation of concept

The Caterpillar concept evolved from the desire to reduce the number of connections in a
system, which has been identified as one of the main causes of system complexity. This is
achieved by having long modules with permanent connections. The starting point for designing
the shape of the system was a long canvas L module (Figure 6.26). Two rapid prototypes were
made out of fabric that were used as a tool when investigating possible design solutions.

Figure 6.26 rapid prototyping

80
 Designing the rigid structure

Two concepts for the rigid structure were explored. Either, the canvas could to be lifted above
the foundation to contribute to the stability, or rigid plates would at the base of the system would
give the system its stability. Early 3D models illustrating both concepts can be viewed in Figure
6.27.

Figure 6.27 CAD model of Caterpillar with lifted canvas (left) and rigid plate elements (right)

The concepts that lifts the canvas could be designed to be lighter than the concepts with plates.
However, it would rely on complicated solutions to allow canvas to be tightly attached and be
collapsible.

The rigid plate concept naturally divides the structure into elements, which enables the system
to be made collapsible in a simple way. In addition, the plate concept will perform well in terms
of stability. Generally, the plate concept represented great performance in many key areas of
the initial design stage.

First iteration and prototype

The plate concept was considered the best solution and was therefore developed further. By
dividing the rigid elements into a base and a wall, it allows each element to be collapsible. By
making a slot in the base and a corresponding tab in the wall, they could easily be mounted
together. A prototype made out of PVC canvas and plywood elements was made (Figure 6.28).

81
 Figure 6.28 Caterpillar plywood and PVC canvas prototype

Connection between modules

To form a continuous waterproof barrier a solution to connect the modules together needed to
be developed. The first idea was to clamp the end of the canvas from one module onto the frame
of the adjacent module, using a rigid section.

Solution 1: Solution 2:

Figure 6.29 Clamping of canvas, solution 1 and 2

82
Two solutions for fastening the canvas is presented in Figure 6.29. Solution 1 fastens the canvas
on top of the element, while solution 2 fastens the canvas underneath the element.

When fastening the canvas on top of the element, the clamping part needs to be a L-shaped
cross section to seal between the side of the element and the canvas. In addition, the clamping
part is required to seal against both the base and the wall. Fastening the canvas underneath the
base, requires the base to have a gasket, and may be a source of seepage if the foundation is
uneven, especially in the corner between the base and the wall.

Another possibility that was explored was using waterproof zippers to connect the modules. By
equipping each module with a male zipper at one end, and two female zipper on the other, two
different attachment points for the adjacent module is available:

1. Directly to the end of the L-cross section to form a straight continuation of the barrier
2. To the backside of the wall to from a 90 degree corner

1 2

Figure 6.30: Connection option 1 and 2

Figure 6.30 illustrates the two connection options. This solution offers highly improved rotation
around z-axis flexibility (see section 4.4.4), and makes it possible to form an enclosed barrier
using only one type of module. This reduces the system complexity by eliminating the need for
special modules to form corners.

83
 Storage and transportation
Due to each module being (potentially) several meters long, they need to be collapsible to be
simple to handle in transportation, and to reduce storage volume. By allowing the wall and base
to be separated, with each element’s wall and base only connected by the flexible canvas, the
system can be folded. Investigating the prototype gave an answer to how the modules could be
efficiently packed (see Figure 6.31). In addition to being compact, the system needs to be as
light as possible to make it easier to handle in transportation and when mounting the barrier.

Figure 6.31 Packing of module

Directional prototype
A directional prototype was tested (Figure 6.32) in the universal test container to investigate
the feasibility of the concept further. It performed well in terms of stability but did not seal
satisfyingly. However, sealing capabilities are believed to improve noticeably in the case of a
larger scale prototype. Sealing between the elements proved to be the greatest issue, which
points out the need for a design that increased gasket pressure between the elements.

Figure 6.32 Directional prototype of Caterpillar

84
 Sealing between the elements
Creating a good seal between the elements is the biggest challenge for the Caterpillar design,
especially at low flood depths. The seepage rate needs to be decreased by increasing the gasket
pressure. As increasing the weight of the modules is not a desired solution, sandbags can be
used as external weight (Figure 6.33). This is a low-cost solution which allows the modules to
be light-weight for transportation.

Figure 6.33: Using sandbags to improve sealing between elements

Improving element strength

The current design relies on the connection between the base and the wall to take up all the
forces acting on the wall of the module. The strength of the joint depends on both the material
of the elements, and the design of the joint. A structural analysis will need to be performed to
determine if the joint is strong enough, and the addition of strengtheners (Figure 6.34) should
be considered. For the modules to continue to be collapsible, the strengtheners need to be
flexible.

Figure 6.34: Strengtheners added to Caterpillar element

85
 Analysis of the dimensions
The properties of the system can be affected by changing the dimensions (Figure 6.35): module
width, number of elements, element distance, element depth, and module length.

Element width

Element distance
Module length

Figure 6.35: Important dimension of Caterpillar modules

One of the most important design choices will be the ratio between elements and canvas. Firstly,
this will affect the stability of the system, as the greater the share of elements, the more stable
the system will the. Secondly, a module with a low share of elements will lead to a lighter,
cheaper, and more flexible system.

The stability of the system is greatly affected

3
by the element distance and was therefore
explored further using the Knowledge 2,5
Safety factor (Stability)

Capturing System (see Figure 6.36). In the 2

case of element distance being 0.6m, the
1,5
stability factor of safety is about 1.1, which is
not acceptable. By reducing, the element 1

distance to 0.4 the safety factor is improved to

0,5
about 2. By reducing the element distance
0
further, some of the key advantages of the 0 0,2 0,4 0,6 0,8 1
system, like system flexibility and the weight Element distance[m]
Number of elements: 4
of the modules will be affected negatively.
Element width: 0.6m
With more comprehensive testing at a later
Element depth: 0,6m
stage, adjustments can be made to optimize the Figure 6.36 Stability factor of safety vs element
element distance. distance

In order to fulfill the idea of reducing connections, the length of the modules is desired to be as
long as possible. The module length is mainly limited by the module weight. The requirement

86
in terms of weight, is that two people are able to move the system. Considering that, the material
of the elements has not been determined, choosing a final number of elements is not optimal at
this stage. However, calculations performed with the Knowledge Capturing System indicate
that the modules should consist of between 3-5 elements.

6.5 FlexiFront

Idea and explanation of concept

FlexiFront is a partial system that can be implemented as a part of other systems. The idea for
FlexiFront surfaced during the examination of the terrain in typical use cases. Most locations
have obstacles that would cause poor system performance if a system would be installed directly
over the obstacle. For example, the installation in Figure 6.37 would lead to critical seepage
rates. Most systems available today require either, improvement of the foundation, customized
modules or external components like sandbags to achieve satisfactory sealing capabilities.

Figure 6.37: Rigid system (Fluvial) placed directly on a curb

FlexiFront is a flexible canvas and gasket component that allows several add-ons to be attached
on top of the gasket to improve sealing in difficult areas. This allows the user to attach
components that are customized for a specific scenario.

FlexiFront is not a concept for a complete system but a feature that can be incorporated into
other concepts or systems to achieve better sealing in areas with uneven foundations. FlexiFront
can be a part of a systems design, or be added as a skirt in front of the system.

87
Three add-ons were developed:

 Weight belt: Fabric tube filled with a heavy fill material. It is designed to be used in
areas where the foundation is uneven.
 Weight box: A container filled with a heavy fill material that is designed to be used on
90-degree corners like on curbs. It functions much like a sandbag, but is able to seal
corners better due to its rigid structure.
 Flexible plate: Flexible plate of sheet metal. It can be formed to inorganic shapes and
used over curbs, but can also form natural shapes to be used on uneven foundations.

Prototyping and testing

A prototype of the FlexiFront system was made. Since the focus of the prototype was to see
how it performed without the rest of the system, the prototype consisted of a 60x60 cm canvas
square with an 8 cm wide seal at the front, and a strip of Velcro to attach add-ons, see Figure
6.38.

Figure 6.38: FlexiFront prototype

By making this simple prototype, quick but valuable feedback was achieved by testing on
different foundations. Although sealing capabilities weren't tested directly, an indication for
performance was gained by investigating the gap between the foundation and the gasket. The
feedback gained through testing of the prototype is outlined in Table 10.

88
Table 10 Evaluation of FlexiFront add-ons

Add-on: Findings and performance

None

The prototype showed that canvas and

gasket are too stiff to properly form to
the foundation. This supports the
assumption that the canvas itself does
not offer adequate sealing for low
flood levels.

Weight belt

The weight belt was not heavy enough

to sufficiently adapt to the foundation.
There was some improvement, so a
heavier weight belt is promising.

Flexible plate

Required some work to shape it to fit

the foundation. Showed promising
performance in most use cases.

89
Weight box

Performed great in closing the gaps

between the gasket and the foundation,
but is limited to 90-degree corners.

The weight box was made so that the

other two add-ons could be put on top
of it.

The weight belt had little impact on the

gasket pressure. Some improvement
on each side of the weight box.

The flexible plate improved gasket

pressure in the corner but lifted the
canvas slightly on each side of the
weight box.

90
 Implementing FlexiFront in Tetris and Caterpillar
FlexiFront can be implemented in most flood protection systems, including Tetris and
Caterpillar. This can be done in two ways:

 Attaching FlexiFront to the front of the system using a watertight fastening mechanism
 Integrating FlexiFront into the system directly, extending beyond the front of the system

FlexiFront could be especially useful in the Caterpillar system if it is placed between the
elements (Figure 6.39), as this has been identified as a critical point for seepage.

Figure 6.39: FlexiFront incorporated in Caterpillar

6.6 Evaluation of concepts

The extent to which extent the concepts fulfill the qualitative requirements established in
section 4.6, is evaluated in Table 11 for Tetris and in Table 12 for Caterpillar. The evaluation
is based on the tests performed, knowledge captured and intuition. A score from 0 to 1 is given
for each measure where 0 requires no further attention, and 1 requires a lot of attention in the
further development of the concepts.

91
Table 11Evaluation of qualitative measures for Tetris

Qualitative Score Explanation

Low cost system 0.75 The system consists of few mechanisms and few add-
ons, but effort needs to be made to keep production
costs low, especially due to the geometry of the rigid
parts.

Compact when not in 0.5 The folding mechanism needs to be optimized and
use stacking methods developed.

High flexibility 0 The flexibility of the current design is its strongest

feature.

Fast installation 0.25 Connecting modules is relatively fast, but due to the
high number of connections needed in a full system,
extra measures should be made to optimize the
connection mechanism.

Build user 0.5 The structure could be perceived as flimsy due to the
trust/confidence relative amount of canvas being high.

Avoid operational 0.75 The connection between modules is promising, but due
failure to the importance of correct use, extra attention should
be directed towards optimizing this.

Avoid structural failure 0.25 A structural analysis needs to be performed to select

satisfactory dimensions for the structure.

Avoid functional failure 0.5 Further testing needs to be performed to select a gasket
to minimize seepage rates. In addition, the connection
between modules is a potential source for function
failure and needs to be addressed.

92
Table 12 Evaluation of qualitative measures for Caterpillar

Qualitative Score Explanation

Low cost system 0.25 The system consists of few parts that each have a
simple geometry. The water-proof zipper is potentially
a high-cost part

Compact when not in 0 The folding mechanism renders the system compact,
use and does not require a lot of further development

High flexibility 0.25 The length of the modules restricts flexibility.

Fast installation 0.5 The main issue is mobility of modules due to their
weight. Efforts should be made to reduce the weight of
the modules.

Build user 0.25 The large rigid structures offers a solid impression, but
trust/confidence the canvas between could be perceived as “flimsy”.

Avoid operational 0.25 Insuring that the added weight is placed correctly.
failure

Avoid structural failure 0.25 A structural analysis needs to be performed to select

satisfactory dimensions for the structure.

Avoid functional failure 0.5 Further testing needs to be performed to select a gasket
to minimize seepage rates. In addition, creating a seal
between the elements must be addressed further.

93
6.7 Further work and industrialization of concepts
Materials and production

Although the design of the product could change during further work, investigating different
material and production options is valuable, as further development should be initiated with this
in mind. The components of each concept can be divided into four main categories: the canvas,
the rigid parts, gaskets and the add-ons.

6.7.1.1 Rigid parts

The rigid parts of the systems can be produced in a range of materials. The prototypes were
produced in plywood, because of easy and accurate manufacturing using a CNC woodworking
router. As AqF uses marine grade plywood for their current products, and considering the flat
and simple geometry of the rigid parts in the concepts (Table 13), this is a feasible production
material. For low production numbers marine grade plywood is believed to be a low cost option.
However, production time will be high, especially for the Tetris frame due to the connection
mechanism. In addition, the arch design of the canvas fastening part (see Table 13), requires
bending of the plywood, which is a complicated time consuming process. A different material
should therefore be considered for this part.

For larger production series, molded plastic is believed to be a low-cost alternative compared
to plywood. Although the initial investment of molds are high, they are believed to pay off if
the production series reaches a certain size. Plastic also offers advantages in terms of storage in
harsh humid conditions.

94
Table 13: Geometry of rigid parts

Tetris

Wall Canvas fastening part Base

Caterpillar

Wall Base

6.7.1.2 Canvas

The canvas used to build the prototypes was chosen mainly for accessibility in local hardware
stores. However, during testing, it proved strong and durable. The canvas was similar to the
canvas AqF uses, but it was slightly thinner. As the forces in the consumer model is lower, the
thinner canvas is a promising option.

Both Tetris and Caterpillar depend on the canvas being attached to the rigid parts (frame and
elements). The current solution that is used for the prototypes, is a contact adhesive. The
benefits of this is that it keeps the product complexity low as it avoids the use of additional parts
like bolts and brackets. If the frame is made out of plastic, welding the canvas to the frame is a
possible solution.

95
6.7.1.3 Gasket

Thorough testing should be conducted with several gasket types to select the best one. The
material used in the late prototypes of Tetris and caterpillar showed promising sealing
performance. However, testing of durability should be performed to ensure that the sealing
capabilities do not decrease after use and long-term storage. The gasket experiment that was
developed in section 5.4.6 is a useful tool for evaluating the gaskets.

6.7.1.4 Add-ons

Effort has been put into reducing the number of add-ons for the systems as a way to reduce the
costs, and to simplify the operational processes. Currently, the add-ons suggested are a
watertight zipper for Caterpillar, and ratchet straps for Tetris. Alternatives to these add-ons
should be investigated further.

Involving manufacturers
By introducing manufacturers to early concept designs, the development can advance with
feedback on how to design solutions can be produced at a low cost. AqF should include their
current manufacturer to get this feedback, and possibly other manufactures to investigate new
methods of manufacturing.

AqF possess unique knowledge about flood protection systems. Although they have provided
valuable insight about flood protection systems, it is believed that a more direct involvement of
AqF in the next phases of development can contribute to designing better systems. This includes
feedback on concepts and prototypes, as well as sharing general knowledge on flood protection
systems.

Certification

AquaFence’s success in the flood protection market is largely based on the certification
performed by FM Global. In addition to being a good selling point, FM Global informs potential
customers about AquaFence. Therefore, requirements for certification should be established as
design guidelines for further development.

96
7 Conclusion
This project is a resource that can be used for developing flood protection systems. A
framework and method for development has been established that outlines aspects of flood
protection systems that are especially important for the consumer market. Even if the two
concepts presented are not pursued, subsystems can be implemented into other systems. In
addition, useful tools and methods that can be used in development of new concepts have been
presented.

The conclusion will address each objective individually:

I. Develop novel concepts for consumer flood protection barriers

Based on the potential market for consumer flood protection systems, a design space for the
development was established. Qualitative and quantitative requirements were established for
the concepts being developed. Based on these requirements two promising complete concepts
(Tetris and Caterpillar) and one partial system (FlexiFront) were developed. Both complete
concepts met the quantitative requirements established. The qualitative requirements have been
addressed, and design solutions have been developed based on these requirements, but further
work (summarized in section 6.6) is needed to fully achieve this. Steps towards industrialization
of the concepts were outlined in section 6.7.1, but need to be examined further as development
moves forward.

II. Develop new innovative ways for testing desired aspects of new concepts

Two comprehensive experiments were executed: Seepage rates for gaskets (section 5.4) and
canvas contribution (section 5.3). The results from the seepage experiment verified the benefit
of adding external weights on the modules to reduce seepage rates at low water depths, and
gave an indication of the expected seepage rates for the concepts. The test setup that was
designed is also useful for development after this project for evaluating and selecting a gasket.
It can also be used for investigating the effects of gasket geometries and materials.

The canvas experiment proved that a rigid-flexible hybrid is a viable configuration for
consumer flood protection systems. The test setup that was designed is also useful to evaluate
the overall stability of concepts, including those not addressed in this project.

97
III. Design and build a functional test set-up for comprehensive prototypes

The test container was designed and built to be as universal as possible. This allowed quick
experiments to be conducted as a tool to explore the design space and to quickly gain knowledge
about specific design features. This was for example, useful during the testing of directional
prototypes in section 6.3.8 and 6.4.6. Furthermore, the test container allowed comprehensive
experiments to be rigged inside the tank.

IV. Establish a system for capturing knowledge gained during development

The knowledge capturing system (section 5.2) that was created was a useful tool to capture
knowledge from the experiments and the relationships established between the properties of the
concepts. The main potential for this system is reusing the knowledge that has been captured to
aid further development of the systems.

98
8 References
Bogers, M., & Horst, W. (2013). Collaborative Prototyping: Cross-Fertilization of Knowledge
in Prototype-Driven Problem Solving. Product Development & Management
Association.

Bryan-Kinns, N., & Hamilton, F. (2002). One for all and all for one? Case studies of using
prototypes in commercial projects. NordiCHI.

Buchenau, M., & Suri, J. F. (2000). Experience Prototyping. San Francisco: IDEO.

Elverum, C. W., & Welo, T. (2015). On the use of directional and incremental prototyping in
the development of high novelty products: Two case studies in the automotive industry.
Journal of Engineering and Technology Management.

Federal Emergency Management Agency. (2015, June 15). Protecting Your Businesses.
Retrieved December 13, 2015, from FEMA: http://www.fema.gov/protecting-your-
businesses

Federal Emergency Management Agency. (2016). Flood Facts. Retrieved from National Flood
Insurance Program: https://www.floodsmart.gov/

Ghosh, S., & Seering, W. (2014). Set-Based Thinking in the Engineering Design Community
and Beyond. ASME.

Hirabayashi, Y., Mahendran, R., Koirala, S., Konoshima, L., Yamazaki, D., Watanabe, S., . . .
Kanae, S. (2013). Global Flood Risk Under Climate Change. Nature Climate Change.

Houde, S., & Hill, C. (1997). What do Prototypes Prototype? In Handbook of Human-Computer
Interaction. Elsevier B.V.

Kennedy, B. M., Sobeck II, D. K., & Kennedy, M. N. (2013). Reducing Rework by Applying
Set-Based Practices Early in the Systems Engineering Process. Wiley Online Library.

Koppe, B., & Brinkmann, B. (2010). Opportunities and drawbacks of mobile flood protection
systems.

99
Lim, Y.-K., Stolterman, E., & Tenenberg, J. (2008, July). The Anatomy of Prototypes:
Prototypes as Filters, Prototypes as Manifestations of Design Ideas. ACM Transactions
on Computer-Human Interaction.

Macher, J. T., & Mowery, D. C. (2003, September). “Managing” Learning by Doing: An

Empirical Study in Semiconductor Manufacturing. Journal of Product Innovation
Management.

McCurdy, M., Connors, C., Pyrzak, G., Kanefsky, B., & Vera, A. (2006). Breaking the Fidelity
Barrier: An Examination of our Current Characterization of Prototypes and an Example
of a Mixed-Fidelity Success. Association for Computing Machinery.

Ogunyoye, F., Stevens, R., & Underwood, S. (2011). Delivering Benefits Through Evidence-
Temporary and Demountable Flood Protection Guide. Bristol: Environment Agency.

Oxford University Press. (2016). Retrieved from Oxford Dictionaries:

http://www.oxforddictionaries.com/

Schrage, M. (1993). The Culture(s) of Prototyping. Design Management Journal.

Smith, P. G. (2007). Flexible Product Development: Building Agility for Changing Markets.
Wiley.

The World Bank. (2010). Economics of Adaptation to Climate Change. Washington DC.

Tidd, J., & Bodley, K. (2002). The Influence of Project Novelty on the New Product
Development Process. R&D Management, p. 2002.

Ulrich, K. T., & Eppinger, S. D. (2012). Product Design and Development. McGraw-Hill.

UNISDR. (2011). Global Assessment Report on Disaster Risk Reduction.

Ward, A., Liker, J. K., Cristiano, J. J., & Sobek II, D. K. (1995). The Second Toyota Paradox:
How Delaying Decisions Can Make Better Cars Faster. Sloan Management Review.

Ward, D. (2012). AquaFence Summary of Laboratory Testing.

Wilby, R. L., & Keenan, R. (2012). Adapting to flood risk under climate change. Sage.
100
101
102
Appendices

103
104
Appendix A - Code for the Knowledge Capturing System
'Define module variables
Dim emptyrow As Integer
Dim stepcounter As Integer
Dim incounter As Integer
Dim outcounter As Integer
Dim pos(1 To 6) As Integer

Sub Button2_Click()
'comments are posted above action

'Define sub variables

Dim i As Integer
Dim j As Integer
Dim xrow As Integer
Dim yrow As Integer
Dim stepsize As Double

i=1
j=0
Do While i <> 7
j=j+1
If IsEmpty(Cells(j, 1).Value) = False Then pos(i) = Cells(j, 1)
If IsEmpty(Cells(j, 1).Value) = False Then i = i + 1
Loop

'Select the number of datapoints to collect

stepcounter = 100

'Calculates input size

incounter = pos(4) - pos(3)

'Calculates output size

outcounter = pos(6) - pos(5)

'Find xrow and yrow

xrow = pos(3) + Cells(2, 2) - 1
yrow = pos(5) + Cells(3, 2) - 1

'Find empty slot

j=0
emptyrow = 0
Do While IsEmpty(Cells(34 + j, 10).Value) = False
j=j+1
emptyrow = j
Loop

'calculate step size

stepsize = (Cells(xrow, 6) - Cells(xrow, 5)) / stepcounter

For i = 0 To incounter
'copy input min to input buffer
Cells(pos(3) + i, 2).Value = Cells(pos(3) + i, 5)
Next i
For i = 0 To outcounter
'copy output buffer to output min
Cells(pos(5) + i, 5).Value = Cells(pos(5) + i, 2)
Next i

105
'Calculates datapoints
For i = 0 To stepcounter
'Copy min column + (stepsize*step) into Buffer
For j = 0 To incounter
Cells(xrow, 2).Value = Cells(xrow, 5) + stepsize * i
Next j
'Copy from buffer to data table
Cells(34 + emptyrow, 10 + i).Value = Cells(xrow, 2)
Cells(35 + emptyrow, 10 + i).Value = Cells(yrow, 2)

Next i

'Copy from buffer to output max

For i = 0 To outcounter
Cells(pos(5) + i, 6).Value = Cells(pos(5) + i, 2)
Next i
'print x/y title
Cells(34 + emptyrow, 9).Value = "x(" & (emptyrow + 2) / 2 & ")"
Cells(35 + emptyrow, 9).Value = "y(" & (emptyrow + 2) / 2 & ")"
'add series
ActiveSheet.ChartObjects("Chart 2").Activate
With ActiveChart.SeriesCollection.NewSeries
.Name = ActiveSheet.Cells(35 + emptyrow, 9)
.Values = ActiveSheet.Range(Cells(35 + emptyrow, 10), Cells(35 + emptyrow, 10 + stepcounter))
.XValues = ActiveSheet.Range(Cells(34 + emptyrow, 10), Cells(34 + emptyrow, 10 +
stepcounter))
End With

'label axes
With ActiveChart
'X axis name
.Axes(xlCategory, xlPrimary).HasTitle = True
.Axes(xlCategory, xlPrimary).AxisTitle.Characters.Text = Cells(xrow, 4) & " " & Cells(xrow, 7)
'y-axis name
.Axes(xlValue, xlPrimary).HasTitle = True
.Axes(xlValue, xlPrimary).AxisTitle.Characters.Text = Cells(yrow, 4) & " " & Cells(yrow, 7)
End With
End Sub

Sub Button82_Click()
'clear datapoints
Worksheets("Sheet1").Range(Cells(34, 9), Cells(35 + emptyrow, 10 + stepcounter)).Clear

'delete all series in chart

ActiveSheet.ChartObjects("Chart 2").Activate
For Each s In ActiveChart.SeriesCollection
s.Delete
Next s
'clear output
Worksheets("Sheet1").Range(Cells(pos(5), 5), Cells(pos(6), 6)).Clear
End Sub

106
Appendix B - Equations used to establish relationships for Tetris

=(B17*B18+B19*B20+B21*B22)*B23*B9 Frame weight

=(B17+B19+B22)*(B21+B26)*B12 Canvas weight
=B21*B24*B25*B10 Sealant weight
=B34+B33+B32 total weight

=((B41*((B19+B22)/2))+B40*(B21+(B22/2)))/((B42*B38)/3) Safety factor (stability)

=IF(B28>(B25+B23);1;0) Is flood over base
=(B28-B23-B25)*B37 Flood height over base

=(B19*B20*B23+B21*B22*B23)*B9+B21*B24*B25*B10+B21*B19*B12 Weight of base

=B27*B13 Force from sandbags

=B39*B13+B37*(B19+B22)*B21*B38*B11*B13 Force on base

=B38*B21*B38/2*B13*B11*B37 Force on wall
=(B41*((B19+B22)/2))+B40*(B21+(B22/2))-((B42*B38)/3) Moment at tipping point
=(B43*(B22+B19-(B24/2)))/(B21*B24)/1000 Gasket pressure
=B28*B13*B11/1000 Seepage pressure
=B44/B45 Gasket/seepage pressure
=3,4315*B46^-1,016 Seepage rate

107
108
Appendix C - Arduino code for automated gasket experiment
volatile int NbTopsFan; //measuring the rising edges of the signal
double Calc;
float CalcOld = 0 ;
int hallsensor = 2; //The pin location of the sensor
double Vol;
int delayLength = 200;
int realDelayLength = delayLength;
int floater = 6;
int relay = 4;
float timeNew = 0 ;
float timeOld = 0 ;
int floaterTest;
int val;
int timer = 0;
int timerTime = 60;
float seepage;

void rpm () //This is the function that the interupt calls

{
NbTopsFan++; //This function measures the rising and falling edge of the
hall effect sensors signal
}
// The setup() method runs once, when the sketch starts
void setup() //
{
pinMode(hallsensor, INPUT); //initializes digital pin 2 as an input
pinMode(floater, INPUT); //initializes digital pin 2 as an input
pinMode(relay, OUTPUT); //initializes digital pin 2 as an input
Serial.begin(9600); //This is the setup function where the serial port is
initialised,
attachInterrupt(0, rpm, RISING); //and the interrupt is attached
}
// the loop() method runs over and over again,
// as long as the Arduino has power
void loop ()
{
floaterTest = digitalRead(floater);

if (floaterTest == 0) {
digitalWrite(relay, 0);
timer = timerTime;
}
if ((floaterTest == 1) && (timer == 0)) {
digitalWrite(relay, 1);
}
if (timer > 0) {
timer = timer - 1;
}
NbTopsFan = 0; //Set NbTops to 0 ready for calculations
sei(); //Enables interrupts
delay (delayLength); //Wait
cli(); //Disable interrupts
Calc = (NbTopsFan / 7.5 * 1000 / realDelayLength); //(Pulse frequency) /
7.5Q, = flow rate in L/min
timeNew = millis();
realDelayLength = (timeNew - timeOld);
timeOld = timeNew;

109
 Vol = Vol + (Calc + CalcOld) / 2 * realDelayLength / (60e3);

if (timeNew > 10000) {

seepage = Vol / (timeNew / 1000);

Serial.print (seepage);
Serial.print ("\t");
}

Serial.print (Vol, DEC);

Serial.print("\t");
Serial.print (timeNew);
Serial.print("\n");
CalcOld = Calc;

110
Appendix D - Documentation from initial development
Brainstorm 1: Understanding the problem

Brainstorm to get a common understanding of the problem and lay out the main pain points and
concerns.

111
Summary of brainstorm 1:

Terrain Storage environment Obstacles Durability

 Grass  Compact  Bushy  UV
 Asphalt product  Ramp  Reuse
 Transition  Outside,  Curbs  Impacts
between variable  Walls  Fatigue
surfaces conditions  Fences  Chemical
 Brick  Rats/animals  Gardens/plants resistance
 Cobblestone  Access (full  Evacuation routes  Corrosion
 Clay garage)  Cars (parked)  Erosion
 Concrete  Costs  Hedge  Disposable
 Gravel  Sunlight/UV  Building geometry  Insects/animals
 Mud/dirt  Hot/cold (corners, curves)  Inspectability
 Sand temps  Not your  Minor repairs
 River rock  Humidity building/property  Replaceable
 Mulch  Time  Concrete cracks parts
 Slope  Shelves  Manholes  Active feedback
 Impacts (smart sensors
to predict
failure)
 Weather, wind
 Lightning
 Vandalism

Deployment/tear Use cases Clean

down  Flash floods  Accessible inside
 Possible for 1  Hurricanes  Able to dry
person  Tsunamis  Hydrophobic
 Time, 30  Overflowing  Disposable parts
min? rivers  Functional when
 Training  Lakes/dams dirty
 Feedback,
snap into
place etc.
 Tools,
included
 Cleaning
 Storing

Brainstorm 2: ways to protect your property/flood protection concepts

112
113
INITIAL CONCEPTS

Concept 1: Security Shutter

Inspiration is taken from security shutters, often found in front of shops after hours. It is flexible and
allows bending around one axis while offering high strength and stiffness in the other.

The flexibility is sought after to create modules that easily can accommodate corners and obstacles
the flood barriers have to tackle. Another issue this design tries to solve is varying terrain height. It
does this by allowing the vertical beams move up and down with respect to each other. The proposed
design for the links is shown on the left and a prototype proving the concept is beside it.

This wall can be implemented in a L-design by the vertical beams transitioning in an angle at contact
with the ground. There can however not be any links between the beams to allow bending in the
vertical part.

114
Another solution is by having support posts the modules can attach to. The support posts give stability
by one half of the base being submerged by water pushing down on the front and the other half acting
as support before the flooding occurs. These posts can also act as a drum that the wall can wrap around
during storage and transportation shown with a sketch below on the right.

Both of these solutions covers only the structural stability needed for a flood barrier but not the
waterproofing aspect of it. One solution to this is dressing the structure with a sheet of waterproof
material that is either permanently attached or that is attached after assembly by the user.

115
Concept 2: Accordion

Using the concept found in an accordion it is possible to create a sheet that is waterproof and that can
vary in length and angle. Triangular supports are connected with an accordion-folded material to
create each module of the flood barrier. To the left is the deployed module and to the right is the same
module compacted for storage and transportation.

Because this structure is not stable it needs support along x1, x2 and x3. By varying the lengths of these
three lengths it is possible to have custom lengths and angles of the module. A proposed system is to
attach bars that can be varied in length along each of the sides.
To create a good seal against the floor surface the accordion folded material has to lay in such a way
that it minimises seepage. One idea is to divide the sheet into two materials. One heavy and rigid
material (illustrated by two lines on the figure below) and a flexible material like rubber (one line). The
heavy material will fold over the flexible material creating a good seal. An extra option to secure a seal
would be having a device that attaches to the end and that can be fastened securely.

Concept 3: Line Dividers (Crowd Control)

Using a similar system AquaFence currently uses but combining it with a length of waterproof fabric
between each module to add flexibility of both length and angle, reducing the current system to only
one kind of module.

A retractable sheet is permanently connected to one side of the module on a spool. During deployment

116
the sheet is pulled out to the desired length and desired angle and attached to the next module and
the process is repeated.

Concept 4: Tetrahedral structure

“Out-there” concept based on an assembly of numerous of tetrahedral structures to accommodate
uneven ground.

117
Concept 5: Double spring concept

Use leaf-springs (or other material that can act as a spring) with canvas/PVC/flexible material
underneath. The essence of the concept is that the spring together with the flexible material will create
a good seal against the ground that can accommodate different types of surfaces, including surfaces
with variable height.

Sealing concepts:

Hinge/slide mechanism:

Rubber blocks on end of module. Hinged or with a sliding mechanism. Blocks will place against ground
in different heights

Sponge seal:

118
Seal in “waterproof sponge material” (noeprene?). Highly formable material will create a tight seal
caused by weight of module and pressure from water.

Power seal:

A hinged system with a spring that increases pressure against ground at the tip of the module for a
tighter seal.

Inflatable seal
Seal is made of a tube filled with water or air.
Double rubber seal:

119
Soft square foam:

Soft foam material covering all four edges of the module.

Squeegee:

TESTING

Testing of superabsorbent polymer:

120
Tested superabsorbent polymer to see how well it binds up water and if it will work with seawater
(about 35 g/l). Potential application is to absorb water to add ballast to the system and make a seal
against the surface.
Findings:
 Promising effect with tap water but does not work with water that has the salt content of
seawater
Next steps
 Test polymer embedded in fabric or similar
 Answer q: is it adequate/suitable for sealing off space between structure and ground?

Testing setup 1.0:

Tested two 25x25 cm models with 50 cm plastic canvas in a kiddie-pool.

Purpose:
 See how scale models behave in water
 See how a plastic canvas vs. solid structure behaves
Findings:
 Total failure
 Cardboard and canvas are too light and floats, need more solid and dense materials for testin
 Difficult to get create a working seal against the surface, even with PVA material
 A free-standing canvas is extremely exposed to wind
Next steps:
 Build and test with materials that are more durable and similar to those that could be used in
the actual products
 Re-use and modify the pool (if needed)

121
Test setup 2.0. :

A plywood test tank, to test stability and sealing for different concepts:

An acrylic glass L-module, to test different sealing concepts:

Different surface inserts plates to simulate foundation surface. Made out of “bedliner” and gravel.

122
Testing in setup 2.0.

Test 1: L-Module without surface plate and sealing

Worked surprisingly well. Only minor seepage was observed with the module, most of the seepage
occurred at the sides of the module (our temporary sealing with clay).
An effective seal as long as the rubberized wood was dry. Flipping and general stability was an issue
when the surface was wet

Test 2: Module with weather-stripping seal (double trouble) without surface plate

Findings:
 Our test setup and prototype seem to work well
 A mode of failure was detected if a layer of water was allowed to form underneath the
module: sliding and flipping
 Difficult to compare the two tests (no seal vs. seal) on flat, rubbery surface. Does not seem
like there are any major differences
 Major pain points to focus on

123
 o Surface. Important to test different surfaces that mimic those encountered in real-
life scenarios. Rubberized plywood is far from that
 Considerably better than test 1, but still a lot of seepage on both rubber surface and
medium bedliner(not thick enough). No tipping or other modes of failure
Test 3: Acrylic L-Module with weatherstripping seal (double trouble) on semi-rough bedliner
surface:

Findings:
 Surface seem to represent a real-life scenario
 Double trouble was not able to seal against the surface, most likely too low sealing.
However, the sealing was good enough to build up a significant difference in water level
 Pain points and next steps:

124
 o Test other surfaces (finer, coarser, grass etc.)
o Test custom-made types of sealing
o Test with more realistic weight. A module this size should probably weigh around
2 - 5 kg
o Investigate significance of weight distribution, could a completely different shape
(other than L) be advantageous?
Test 4: Foam sheet base, square and horseshoe
Attached a foam sheet to the base of the test module. Tested to configurations: Square with hole
in centre (left) and the horseshoe (right), the same square with a part cut out on the far side.

Findings
 Tested initially without extra weight on the module. The whole base of the module ended
up floating and therefore breaking the seal with the ground. Because of this the next tests
were performed with weights placed on the module.
 Square results: There was moderate leaking and low pressure holding the module down.
This could be because water that leaks under the module is built up and trapped in the
square that is cut out of the foam sheet and this creates a pressure underneath. The
horseshoe should remove this problem
 Horseshoe: As predicted the module stuck to the base as the water that leaked through
under the module is able to flow away and prevent pressure building up underneath.
There was leaking but at an acceptable rate.
 Based on the testing it is concluded that the horseshoe shape is superior to the enclosed
square
 Our assumptions that water could get trapped under the module and cause it to flip over
was confirmed
 The horseshoe solution allows excess water to flow out behind the module,
eliminating flipping
Test 5: Soft foam block - medium bedliner

125
As expected, no low-pressure zones were allowed to developed and the entire module ended
up flipping over.

Test 5: Initial superabsorbent polymer distributed in a plastic bag:

126
Findings:
 The superabsorbent polymer expanded expanded more than anticipated and we probably
used too much for the size of the sealing (about 15 grams powder)
 Very promising effect, the expansion elevated the module despite of being fixed with clay
along the edges and the water pressure forcing it down
 Pain points and next steps:
o Optimize powder to water ratio (gut-feeling: about 20 grams for 1 meter, 2 kg for
100 m, total cost 90 $ from amazon)
o Test different concepts of and designs (c-shaped, edge-balloon etc.)
o Investigate using a canvas in combination with superabsorbent polymer

Test 6: Superabsorbent polymer in fabric tube 1.0

127
128
A cloth tube (old t-shirt) was filled with superabsorbent polymer (17 g). The tube was 95 cm long
and made out of a 6 cm wide cut-out. Double-sided tape was used both to create the tube and fix
it to the module.

Findings
 The polymer expanded as planned and worked well as a seal
 Too much powder was used (again)
 Most of the seepage seemed to occur in the corners of the tube (this could be eliminated
with a more complex/higher resolution prototype)
 There was very little seepage with higher water levels
 Pain points and tings to test
o More rounded corners
o Less powder
o More durable tube to avoid collapse of tube (and possibly the entire module)
o Allow the superabsorbent polymer to absorb water before the water level rises
significantly. This should allow the material to swell and create a seal before being
flooded

Test 7: Superabsorbent polymer in fabric tube 2.0

Improved version of the previous test. 10 g of powder was used (vs. 17 in the last test) and the
following things were improved:
 increased radius of corners to avoid sharp, warped corners
 gorilla tape in addition to the double sided tape to make the tube stronger to avoid
rupture

129
Almost fully sealed at max water height.

Test 8: Superabsorbent polymer in fabric tube 3.0

130
Building and test of C-shape with superabsorbent polymer. This is a design that came up in a
brainstorm one of the first days. The C-shape is made from the PVC-coated polyester fabric (the C-
shape) and cloth (the front). The idea is that the module with lay flat before the deployment and
that it will absorb water and inflate once water reaches the front and gets absorbed by the
superabsorbent polymer.
Findings
 Superabsorbent polymer works extremely well, once again
 The seal managed to achieve minimal seepage despite being a very rough prototype:

131
 o Using a glue gun made the upper and lower parts a lot stiffer and rigid than
desired
o The PVC-coated polyester is stiffer and less flexible than desired, especially for the
scale that we are working on
Test 9: Without seal on medium bedliner

Findings:
 The module obviously leaks a lot, but the main mode of failure is that the module is prone
to tipping over when the pressure underneath the base is equalized

Test: without seal on medium bedliner - with extra ballast

Findings
 The added weight helps reduce the tipping
 The pressure underneath the module drops significantly when the water is allowed to
flow freely (Bernoulli) + avoiding accumulation of water
Seal against ground

132
Test 10: Double spring concept rapid prototyping and testing

Despite being tested only on rubberized plywood, it seems that the double spring solution
has potential to create a good seal. It does however seem to have worse load handling
properties than an L-structure.

133
Test 10: Ice-block and building foam

134
135
136
Appendix E - Risk assessment form
The risk assessment form starts on the next page.

137
138
139
140
141
142
143
144
145