Hydrofoil Design and Construction

Author: Pol Bernad Serra

Director: Aleix Báez

UPC ESEIAAT

Grau en Enginyeria en Vehicles Aeroespacials

Delivery Date: 10th of June, 2019

Pol Bernad Serra
Hydrofoil Design and Construction

Abstract

This TFG project reports the design and prototyping of a surf hydrofoil. The hydrodynamic design
has been done with a NASA’s open source software Open VSP together with a simulation
platform on Simulink. Some geometries have been designed, 3D printed and tested in a scale
bench pulled from a boat. The best geometry has been constructed with glass fibre and steel
and tested in real scale. All the work has been done at home with own elaborated machines like
a 3D printer or a CNC.

2
Pol Bernad Serra
Hydrofoil Design and Construction

Table of Contents

1. Introduction .......................................................................................................................... 9
1.1 Aim ................................................................................................................................ 9
1.2 Scope ............................................................................................................................. 9
1.3 Justification ................................................................................................................... 9
1.4 Background.................................................................................................................. 10
1.5 Acknowledgment ........................................................................................................ 11
2. State of the art .................................................................................................................... 12
3. Design .................................................................................................................................. 15
3.1 Objectives and constrains ........................................................................................... 15
3.2 Methodology ............................................................................................................... 16
3.3 Geometry design ......................................................................................................... 17
3.3.1 Alpha ................................................................................................................... 18
3.3.2 Beta ..................................................................................................................... 22
3.4 Platform design ........................................................................................................... 26
3.4.1 Solid Block ........................................................................................................... 27
3.4.2 Hydrodynamics Block .......................................................................................... 28
3.4.3 Control Block ....................................................................................................... 31
3.4.4 The smaller blocks ............................................................................................... 33
3.5 Geometry comparison................................................................................................. 34
3.5.1 Alpha ................................................................................................................... 34
3.5.2 Beta ..................................................................................................................... 37
3.5.3 Comparison ......................................................................................................... 40
4. Model .................................................................................................................................. 41
4.1 Construction ................................................................................................................ 41
4.1.1 Geometry ............................................................................................................ 42
4.1.2 Surf Board ............................................................................................................ 44
4.1.3 Sliding Mass ......................................................................................................... 44
4.1.4 Articulated quadrilateral ..................................................................................... 51
4.2 Tests ............................................................................................................................ 53
4.2.1 First test............................................................................................................... 53
4.2.2 Second test .......................................................................................................... 53
4.2.3 Third test ............................................................................................................. 54

3
Pol Bernad Serra
Hydrofoil Design and Construction

4.2.4 Fourth test ........................................................................................................... 55

4.3 Decision making .......................................................................................................... 56
5. Real Scale Model ................................................................................................................. 57
5.1 Structural design ......................................................................................................... 57
5.1.1 Internal structure ................................................................................................ 57
5.1.2 Lifting surfaces .................................................................................................... 59
5.2 Construction ................................................................................................................ 61
5.2.1 Internal Structure ................................................................................................ 61
5.2.2 Molds ................................................................................................................... 63
5.2.3 Lifting surfaces .................................................................................................... 68
5.2.4 Board ................................................................................................................... 75
5.3 Tests ............................................................................................................................ 76
5.3.1 First test............................................................................................................... 76
5.3.2 Second test .......................................................................................................... 77
5.3.3 Third test ............................................................................................................. 78
6. Future work ......................................................................................................................... 79
7. Timing .................................................................................................................................. 79
8. Conclusions ......................................................................................................................... 81
9. Bibliography ........................................................................................................................ 82

4
Pol Bernad Serra
Hydrofoil Design and Construction

List of Figures
Figure 1: Side force and angles diagram ..................................................................................... 10
Figure 2: Bottom force and angles diagram ................................................................................ 11
Figure 3: Hydrofoil 1 .................................................................................................................... 12
Figure 4: Hydrofoil 2 .................................................................................................................... 12
Figure 5: Hydrofoil 3 .................................................................................................................... 13
Figure 6: Hydrofoil 4 .................................................................................................................... 13
Figure 7: Hydrofoil 5 .................................................................................................................... 14
Figure 8: Design methodology .................................................................................................... 16
Figure 9: Alpha geometry ............................................................................................................ 18
Figure 10: Alpha Cm .................................................................................................................... 19
Figure 11: Alpha CL ...................................................................................................................... 19
Figure 12: Alpha Cn ..................................................................................................................... 20
Figure 13: Naca 23012 (Airfoil Tools, 2019) ................................................................................ 20
Figure 14: Sweep effect............................................................................................................... 21
Figure 15: Beta ............................................................................................................................ 22
Figure 16: Beta Cm ...................................................................................................................... 23
Figure 17: Beta CL........................................................................................................................ 23
Figure 18: Beta Cn ....................................................................................................................... 24
Figure 19: ESA40 (Airfoil Tools, 2019) ......................................................................................... 24
Figure 20: Platform overview ...................................................................................................... 26
Figure 21: Solid block .................................................................................................................. 27
Figure 22: Solid block structure................................................................................................... 27
Figure 23: Hydrodynamics block ................................................................................................. 28
Figure 24: Hydrodynamics structure ........................................................................................... 28
Figure 25: Coefficients calculator structure ................................................................................ 29
Figure 26: Depth effect ............................................................................................................... 30
Figure 27: Control block .............................................................................................................. 31
Figure 28: Control block structure .............................................................................................. 32
Figure 29: Alpha body angles control off .................................................................................... 34
Figure 30: Alpha velocity control off ........................................................................................... 35
Figure 31: Alpha submergence control off.................................................................................. 35
Figure 32: Alpha body angles control on..................................................................................... 36
Figure 33: Alpha Velocity control on ........................................................................................... 36
Figure 34: Alpha submergence control on .................................................................................. 37
Figure 35: Beta body angles control off ...................................................................................... 37
Figure 36: Beta velocity control off ............................................................................................. 38
Figure 37: Beta submergence control off.................................................................................... 38
Figure 38: Beta body angles control on ...................................................................................... 39
Figure 39: Beta velocity control on ............................................................................................. 39
Figure 40: Beta submergence control on .................................................................................... 40
Figure 41: Scale model test ......................................................................................................... 41
Figure 42: Open VSP geometry ................................................................................................... 42
Figure 43: Ready to print geometry ............................................................................................ 42
Figure 44: 3d printed Alpha geometry ........................................................................................ 42
Figure 45: Surface treatment for gluing ...................................................................................... 43
Figure 46: Gluing process ............................................................................................................ 43

5
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 47: 3d printed Beta geometry .......................................................................................... 43

Figure 48: Final Alpha geometry ................................................................................................. 43
Figure 49: Final Beta geometry ................................................................................................... 43
Figure 50: Components of the sliding mass system .................................................................... 45
Figure 51: Central piece assembled ............................................................................................ 45
Figure 52: Central piece printed.................................................................................................. 45
Figure 53: The final design .......................................................................................................... 46
Figure 54: The first design approach ........................................................................................... 46
Figure 55: Custom Arduino NANO pcb ........................................................................................ 47
Figure 56: Connections photo ..................................................................................................... 47
Figure 57: Custom driver pcb ...................................................................................................... 47
Figure 58: Tin channels................................................................................................................ 47
Figure 59: Failed electronic components .................................................................................... 48
Figure 60: Final sliding mass system ........................................................................................... 49
Figure 61: Final scale model constructed.................................................................................... 49
Figure 62: Real model for comparison ........................................................................................ 50
Figure 63: Render of the model .................................................................................................. 50
Figure 64: Articulated quadrilateral on a test ............................................................................. 51
Figure 65: Articulated quadrilateral build ................................................................................... 51
Figure 67: Plastic tub ends joints ................................................................................................ 51
Figure 67: Aluminium hinge inserts ............................................................................................ 51
Figure 68: Quadrilateral tip ......................................................................................................... 52
Figure 69: Quadrilateral root ...................................................................................................... 52
Figure 70: Definitive bearings position ....................................................................................... 52
Figure 71: First bearings position ................................................................................................ 52
Figure 72: Second test. No tension on the ropes ........................................................................ 54
Figure 73: Final internal structure render ................................................................................... 57
Figure 74: First central tub wing load diagram ........................................................................... 58
Figure 75: First central tub tail load diagram .............................................................................. 58
Figure 76: Final central tub wing load diagram ........................................................................... 58
Figure 77: Final central tub tail load diagram ............................................................................. 58
Figure 78: Tail load diagram ........................................................................................................ 59
Figure 79: Wing load diagram ..................................................................................................... 59
Figure 80: Naca 23015 bending inertia (Online Wing Bending Inertia Calculator, 2019) ........... 60
Figure 81: Ensuring perpendicularity with a machined to square aluminium plate ................... 61
Figure 82: Welded joint ............................................................................................................... 61
Figure 83: The results after the first welding day ....................................................................... 62
Figure 84: Finished internal structure ......................................................................................... 62
Figure 85: Finished wing mold .................................................................................................... 63
Figure 86: Wing mold render ...................................................................................................... 63
Figure 87: Face operation toolpath............................................................................................. 64
Figure 88: Adaptative clearing machining ................................................................................... 64
Figure 89: Adaptative clearing toolpath...................................................................................... 64
Figure 90: Parallel machining ...................................................................................................... 65
Figure 91: Parallel toolpath ......................................................................................................... 65
Figure 92: Parallel 2 toolpath ...................................................................................................... 65
Figure 93: Contour toolpath ........................................................................................................ 65
Figure 94: Surface finish after treatment .................................................................................... 66
Figure 95: Surface finish before treatment ................................................................................. 66
Figure 96: Wing mold after varnishing ........................................................................................ 67
Figure 97: Finished tail mold ....................................................................................................... 67

6
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 98: Perforated polyethylene film ..................................................................................... 68

Figure 99: Vacuum fridge pump .................................................................................................. 68
Figure 100: Template transferred to the fibre mat..................................................................... 69
Figure 101: Paper template......................................................................................................... 69
Figure 102: Plastic adapter for the vacuum tub .......................................................................... 69
Figure 103: Manual-laying the composite material .................................................................... 70
Figure 104: Resin excess being absorbed by the paper .............................................................. 70
Figure 105: Pressing the vacuum bag to avoid bridging ............................................................. 70
Figure 106: First composite part ................................................................................................. 71
Figure 107: Checking the wall thickness ..................................................................................... 71
Figure 108: Leveling the tail ........................................................................................................ 72
Figure 109: Wing joint procedure ............................................................................................... 72
Figure 110: Tail joint procedure .................................................................................................. 72
Figure 111: Joint surface problem............................................................................................... 72
Figure 112: Joint solution ............................................................................................................ 73
Figure 113: Leading edge fibre strip............................................................................................ 73
Figure 114: Clamping the halves ................................................................................................. 73
Figure 115: Final result after sanding .......................................................................................... 74
Figure 116: Polyurethane foam fill .............................................................................................. 74
Figure 117: Finished prototype ................................................................................................... 75
Figure 118: First test.................................................................................................................... 76
Figure 119: Second test ............................................................................................................... 77
Figure 120: Third test .................................................................................................................. 78
Figure 121: Project charter’s Gantt diagram ............................................................................... 80

List of tables
Table 1: Alpha characteristics ..................................................................................................... 18
Table 2: Beta Characteristics ....................................................................................................... 22
Table 3: Solid block inputs and outputs ...................................................................................... 27
Table 4: Hydrodynamics block inputs and outputs ..................................................................... 28
Table 5: CL as function of submergence ..................................................................................... 30
Table 6: Control block inputs and outputs .................................................................................. 31
Table 7: Model characteristics .................................................................................................... 50
Table 8: OWA decision making.................................................................................................... 56

7
Pol Bernad Serra
Hydrofoil Design and Construction

List of Abbreviations

CD Drag coefficient
CY Sideslip coefficient
CL Lift coefficient
Cl Rolling moment coefficient
Cm Pitching moment coefficient
Cn Yawing moment coefficient
𝛼 Angle of attack
𝛽 Sideslip angle
𝜇 Velocity roll angle
𝛾 Velocity pitch angle
𝜒 Velocity yaw angle
𝜙 Roll angle
𝜃 Pitch angle
ψ Yaw angle
𝛿 Wave slope

8
Pol Bernad Serra
Hydrofoil Design and Construction

1. Introduction

1.1 Aim

The aim of the TFG is to design, build and test a hydrofoil for foil surf. The design will focus on
finding a hydrofoil geometry capable to ensure stable and exciting surfing. Some geometries will
be built at scale with a 3D printer and tested. Finally, a winning geometry will be built in full scale
with composite materials and tested in real conditions. At the end of the project, a hydrofoil
prototype will be ready for industrialization.

1.2 Scope

• Understand the differences between a foil working in air and working in water.

• Analyse the state of use of the hydrofoils.

• Understand and model the requirements of surfing hydrofoils.

• Create a simulation platform with Simulink.

• Design several hydrofoil geometries, test on the simulation platform and understand
the relations between geometrical parameters and response.

• Build some scale models of the representative geometries with 3D printing.

• Testing of the scale models.

• Define a final hydrofoil geometry.

• Construction of the final prototype with composite materials.

• Test the final prototype in full scale.

• Report the results.

• The project should be environmentally friendly.

1.3 Justification

Hydrofoils are causing a revolution in water sports. In foil surf they allow surfing weaker waves
and riding common waves for longer periods. This allows surfing in the Mediterranean coast of
Catalonia, and Southern Europe more days of the year. This opens a market opportunity for a
design platform to help predicting the behaviour of these devices prior to construction and for
the technologies leading to hydrofoil manufacturing at reasonable costs.

I, as an aerospace student with a background in parts construction and with attraction to the
sea and sailing would like to take this opportunity to apply my aerospace engineering knowledge
to water sports.

9
Pol Bernad Serra
Hydrofoil Design and Construction

1.4 Background

Cavitation

Cavitation is the phenomenon that describes the phase change from liquid to gas of a fluid due
to a reduction of the pressure at constant temperature (Eisenberg, 1968). The forming of
bubbles affects the flux changing the hydrodynamic forces. When the bubbles collapse,
structural erosion and noise are produced. This is one of the reasons why hydrofoil boats have
not been taken place over the traditional floating technology. However, in this study cavitation
has not been considered because it is a phenomenon which occurs at much higher speed, and
thus pressure difference, than foil surfing does (Acosta, 1973).

Depth effect

The depth effect considers the lift reduction due to the proximity of the wing to the water
surface. The transition between air and water must have a pressure of 1atm and thus, there is a
contour condition at a variable distance, not like in unrestricted airflow (Acosta, 1973). This lift
reduction increases when the submergence of the foil reduces and at the surface the lift must
be 0. This phenomenon will be considered in the design section as will be later explained.

Foil surf diagrams

In this section, two diagrams are presented to show the hydrodynamic forces and the angles
that will be used to compute and project them. The wave has been considered to be a constant
slope sea level as it will be explained in detail in the platform design section.

Figure 1: Side force and angles diagram

10
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 2: Bottom force and angles diagram

1.5 Acknowledgment

This project would not be possible without the support of a lot of people which have helped me
throw the whole process.
First, I would like to thank my family because they have given me their support and ideas to
constantly upgrade the work done. They have also taken care of the entire cost of the
elaboration of this project.
Secondly, I would like to acknowledge my uncle who has also helped me with the welding
process. I would also like to thank the local metal worker Josep Vidal and the local carpenter
Fustes Martínez who have supplied material without any cost.
Finally, and not less important, I would like to thank the director of this project, Aleix Báez, who
has always guided me since the first day and has shown interest in every step of the project.
I would like to thank you all because without you this project would not have been possible.

11
Pol Bernad Serra
Hydrofoil Design and Construction

2. State of the art

This section describes the actual technology used in the industry of foil surf.
Slingshot 2019 Hover Glide Fsurf

This hydrofoil mast and fuselage are

made of aluminium and its wing and
tail are made of composite material.
The mast is 61cm long, the aspect
ratio 3.7, the surface area is 1534 cm2
and the wingspan is 77cm. Its price is
1170€.

Figure 3: Hydrofoil 1

Naish 2019 Foil Surf Thrust XL

The fuselage and the vertical support

are also made of aluminium. The wing
and the tail are made of monocoque
carbon and glass fibre with a foam core.
The mast is 55cm long, the aspect ratio
is 4.75, the wingspan is 86cm and the
surface is 1572cm2. The price for this
hydrofoil is 1470€.

Figure 4: Hydrofoil 2

12
Pol Bernad Serra
Hydrofoil Design and Construction

Ride Engine 2019 Futura Sup Foil

This foil construction consists again of

aluminium fuselage and mast and
carbon fibre reinforcement plastic
wings. Its mast is 61cm high, the aspect
ratio is 3.3, the surface area is 2066cm2
and its wingspan is 84cm. the price of
this foil is 1040€.

Figure 5: Hydrofoil 3

Naish 2019 Foil Wind Thrust

This foil is not for wave surf but for

windsurf. Its construction is like the
others seen before, aluminium fuselage
and mast and composite material wings.
Its mast is 70cm long, its aspect ratio is
3.46, the wingspan is 65cm and its
surface area is 1220cm2. Its price is
1360€.

Figure 6: Hydrofoil 4

13
Pol Bernad Serra
Hydrofoil Design and Construction

Fanatic Aero Sup Foil 2000 Long Fuselage 2019

This is a wave surf hydrofoil and its

construction is like the others’. Its
fuselage and mast are made of
aluminium and the wings are made out
of carbon composite material. The
mast is 75cm long, the wingspan is
99cm and the surface area is 2000 cm2.
The selling price is 1350€.

Figure 7: Hydrofoil 5

From the state of the art it can be seen that the current technology for the foil surf uses the
same construction methods, aluminium fuselage and mast and composite material wings. The
surf hydrofoils have a surface area between 1500 and 2000 cm2, a wingspan between 60 and
100cm, a variable aspect ratio between 3 and 5 and a mast length between 55 and 75cm. The
average price is above 1000€. All the information for the elaboration of the state of the art has
been obtained from (B3 Proshop, 2019).

14
Pol Bernad Serra
Hydrofoil Design and Construction

3. Design

This section describes the design process and final geometry of the hydrofoil.

3.1 Objectives and constrains

The objectives that need to be accomplished are:

• Develop a hydrofoil which is stable enough to be ridden by a beginner.

• The hydrofoil should lift 80Kg at 6knts.

• The design should allow the construction to be done at home with the tools and
machines available.
• The wing structure will be made of composite material and the internal structure of
stainless-steal.

• The whole surf should weight under 20kg because it has to be carried by only one
person.
• It should allow surfing in low waves condition

The constrains that limit the design are:

• The 3D printer maximum print dimensions which restrict the size of the scale model.

• The CNC maximum dimensions which restricts the molds dimensions.

• The budget.

• The deadline.

15
Pol Bernad Serra
Hydrofoil Design and Construction

3.2 Methodology

The hydrodynamic design started with generating and optimizing different geometries with an
open source software Open VSP. To do so, different parameters of the geometry were changed
and then the coefficients which describes the flight behaviour were computed again. The
software uses a Vortex Lattice Method (VLM) to compute the coefficients. The computation time
for each iteration took about 2min. Once the static stability was accomplished, the dynamic
stability was studied with an own elaboration platform on Simulink. To do so, all the coefficients
for all the possible angles were computed by the Open VSP. This process took about 1:30h. Then,
the results were imported into the Simulink Platform which took about 10 seconds to compute
the dynamic behaviour. Finally, this platform was used to compare the different designed
geometries. In Addition, a scale test is also used to choose the best geometry. The testing
requires 3D printing the geometry and all the test operation. This process is much slower. The
following image shows a block diagram of the design structure:

Figure 8: Design methodology

16
Pol Bernad Serra
Hydrofoil Design and Construction

3.3 Geometry design

This section describes the design of the geometry using the NASA’s open source software Open
VSP. The design should be done efficiently since the optimization of the geometries needs a lot
of different iterations. So, the computation time required for each calculation should be short.
Every time the geometry is changed, the software needs to recalculate all of the coefficients
which takes about 2min.

For this design process there have been three curves considered for each iteration:

• Cm-alpha: this curve represents the pitching moment coefficient as function of the angle
of attack. For the plane to be stable the slope of this curve must be negative. This means
that when a perturbation, for example, increases the angle of attack, the moment
resultant should be more negative than before because this means the plane is trying
to minimize this increase of alpha. The other important design aspect is the 0-moment
angle. This angle is where the geometry will try to equilibrate itself. The position of the
centre of mass has a major effect over this curve. This is the principal problem of this
design, the centre of mass is located at 1.5m over the plane’s wing, which is more than
its wingspan. The tail volume is another key factor for design because it changes the
slope of the curve and also changes the cut-off point. The angle of attack of the wing
and the tail can be also used in addition with the wing’s plant form. The profile used also
affects to the moment distribution.
• CL-alpha: this curve represents the lift coefficient as function of the angle of attack. The
before commented 0-moment alpha should have enough lift to maintain a flight at a
reasonably low speed. The profile and the wing’s angle of attack have a huge effect over
this curve.
• Cn-beta: this curve represents the yawing moment coefficient as function of the angle
side-slip angle. This curve is used to size the vertical stabilizer and its slope should also
be negative to dissipate perturbations.

The first idea was to optimize 4 different type of geometries: conventional, biplane, flying wing
and canard. The optimization was made by proposing a geometry and then changing parameters
like its wingspan, sweep or chord distribution and computing again each time until obtaining the
desired results; all this process was made with the software Open VSP. However, after the first
iterations of the design it was clear that the cm-alpha was the principal problem because of the
before commented height of the centre of mass. It was very difficult to obtain a negative slope
and also a positive lift angle of equilibrium. For this reason, the canard geometry which are the
more difficult geometries to stabilise was discarded. As the one wing geometries were already
giving enough lift, the biplane was also discarded because of its constructive complexity. So, the
remaining type of geometries were the conventional and the flying wing.

17
Pol Bernad Serra
Hydrofoil Design and Construction

3.3.1 Alpha

The alpha geometry is the conventional design. One major advantage over the flying wing is that
the major font of lift, the wing, is detached from the font of stability, the tail. This made possible
to change the equilibrium angle or the slope of the Cm-alpha graph without compromising the
lift too much. For this case, there were some construction limitations. In order to fit in two pieces
inside the CNC machine, the wingspan could not be bigger than 1m and the tail wingspan should
be 0.5m to fit in one piece. The resultant geometry after all the iterations was the following:

Figure 9: Alpha geometry

Next, a table with the geometry properties:

Property Value
Wingspan 1m
Tailspan 0.5m
Wing surface 0.19 m2
Aspect ratio 5.26
Wing Foil Naca23015-Naca23012
Tail foil Naca0012
Tail volume coefficient 1.67
Angle of equilibrium 1º
CL at equilibrium 0.27
Sweep at 0.25 chord 35º
Twist 0º
Wing’s angle of attack 3º
Tail’s angle of attack -2º
Cn-alpha slope -0.005
CoM X position from leading edge -0.05 m
Table 1: Alpha characteristics

18
Pol Bernad Serra
Hydrofoil Design and Construction

Next, the three before commented graphs are presented:

Figure 10: Alpha Cm

As it can be seen, the equilibrium angleFigure

is 1º and the whole
11: Alpha Cm slope of the graph is negative.

Figure 11: Alpha CL

In this graph it can be observed that at 1º the CL is equal to 0.27.

19
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 12: Alpha Cn

It can be extracted from this graph that its slope is -0.005.

As it can be seen in the table, the root foil has a 15% and a 12% thickens at the tip. This is done
for structural proposes. It can also be seen that the tail volume coefficient is 1.67 which is much
bigger than a regular plane which normally have between 0.3 and 0.6 (Basic Aircraft Design
Rules, 2019). This is because the height of the centre of mass need a big tail to counter its
unstable moment when pitching. The different angle of attack between wing and tail also made
possible to obtain the 1º angle of equilibrium. The foil Naca23012 has given good results because
despite of not being the most lift producer, its moment is also low and in combination is the one
that offered more lift at its equilibrium angle, 0.27. This means that the required speed is
(Franchini & López, 2011):

𝐿𝑖𝑓𝑡 800𝑁
𝑣=√ =√ = 5.6𝑚/𝑠
1 1 1000𝑘𝑔 2 · 0.27
· 𝜌 · 𝑆𝑟𝑒𝑓 · 𝐶𝐿 · · 0.19𝑚
2 2 𝑚3

Figure 13: Naca 23012 (Airfoil Tools, 2019)

20
Pol Bernad Serra
Hydrofoil Design and Construction

It is important to point that the sweep helped on the Cm graph, without it the equilibrium point
was at the largest angles of attack and the slope it was already positive, making an unstable
point of equilibrium. The next figure shows the comparison between Cm of an early design, the
left one had sweep and the right one did not have:

Figure 14: Sweep effect

As it will be explained later, for constructive reasons the 3d printed design and the real scale
model are not exactly the same, but the only change is in the vertical support and stabilizer.

21
Pol Bernad Serra
Hydrofoil Design and Construction

3.3.2 Beta

The Beta geometry is the flying wing design. The major advantage of this design is the simplicity.
There are no tail and fuselage, so the efficiency is normally better than the conventional designs.
However, the lack of tail proposes a difficult challenge to develop a stable geometry with the
centre of mass at that high. The size restrictions for this design was that the wing should fit in
three parts on the CNC machine, so the maximum wingspan or longitudinal length should be
1.5m. The following image shows the result of the different design iterations:

Figure 15: Beta

Next, a table with the geometry properties:

Property Value
Wingspan 1.3m
Wing surface 0.23 m2
Aspect ratio 7.35
Foil ESA40
Angle of equilibrium 0.2º
CL at equilibrium 0.06
Sweep at 0.25 chord 45º
Twist -5º
Wing’s angle of attack 3º
Cn-alpha slope -0.011
CoM X position from leading edge -0.05 m

Table 2: Beta Characteristics

22
Pol Bernad Serra
Hydrofoil Design and Construction

Next, the three before commented graphs are presented:

Figure 16: Beta Cm

As it can be seen, the equilibrium angle is 0.2º and the whole slope of the graph is negative.

Figure 17: Beta CL

23
Pol Bernad Serra
Hydrofoil Design and Construction

In this graph it can be observed that at 0.2º the CL is equal to 0.06.

Figure 18: Beta Cn

It can be extracted from this graph that its slope is -0.011.

In the flying wing geometry design there is an essential component, the profile. A regular airfoil
has negative moment which means that the tendency is to pitch down. In a regular geometry
this effect is counter by the tail. In a flying wing there is no tail so the profile needs to have a
moment equal to zero or even positive. These type of airfoils are called reflex. For this study
case a regular reflex airfoil was not enough again because of the centre of mass position. The
profile selected is the ESA40 which has an average Cm of +0.04. The problem with this type of
profiles is that the more Cm they have the less lift is produced.

Figure 19: ESA40 (Airfoil Tools, 2019)

24
Pol Bernad Serra
Hydrofoil Design and Construction

However, the airfoil is not the only solution needed to develop a stable tail-less plane. The wing
needs to have sweep. In this case, 45º. In addition, to counter the negative moment the wing
also has negative twist or wash-out. This lowers the lift produced by the most delayed part of
the wing or even it makes it negative. This helps to pitch up. In this case -5º were given to the
tip of the wing. However, this technique also reduces the global lift so an equilibrium needs to
be found. Lastly, the winglets are necessary not for aerodynamic efficiency but for yaw stability
acting as vertical stabilizers.
The final geometry has optimized till obtaining a CL of 0.06 at its equilibrium angle. This means
that the required speed is:

𝐿𝑖𝑓𝑡 800𝑁
𝑣=√ =√ = 10.8𝑚/𝑠
1 1 1000𝑘𝑔 2 · 0.06
2 · 𝜌 · 𝑆𝑟𝑒𝑓 · 𝐶𝐿 · · 0.23𝑚
2 𝑚3

25
Pol Bernad Serra
Hydrofoil Design and Construction

3.4 Platform design

The Simulink platform is a comparative tool that can compute the dynamic behaviour of a surf
hydrofoil with different geometries. It considers the depth effect and a human controlled
response.

The Simulink model is divided in 3 main blocks which are connected together:

• Solid

• Hydrodynamics

• Control

There are 4 smaller blocks that feed the main blocks:

• Angles transformation

• Geometry

• Summation

• Initial Conditions

Figure 20: Platform overview

26
Pol Bernad Serra
Hydrofoil Design and Construction

3.4.1 Solid Block

Figure 21: Solid block

Figure 22: Solid block structure

The solid block task is to compute the dynamic behaviour assuming that the ship is a rigid solid.
All the parameters are computed using an inertial reference frame.

There are 2 equations that need to be solved (Gómez & Pérez, 2012):

𝐹 =𝑚·𝑎

𝑀 =𝐼·𝛺

Once the linear and angular accelerations are computed, they are integrated to obtain velocities.
Then, the initial velocities from the initial conditions block are added. After that, the velocities
are integrated to obtain position and angles. The initial position and angles are also added to
the computed values.

The inputs and outputs of the block are the following:

Inputs Outputs
Mass Linear Acceleration
Forces Linear Velocity
Inertia Position
Moments Angular Acceleration
Initial conditions Angular Velocity
Angles
Table 3: Solid block inputs and outputs

27
Pol Bernad Serra
Hydrofoil Design and Construction

3.4.2 Hydrodynamics Block

The hydrodynamics block task is to compute the static forces and moments produced by the
hydrofoil inside the water.

Inputs Outputs
Fluid Properties Aero Forces
Wind Velocity (Vw) Aero Moments
Aero Angles
Depth
Surface&Chord
Table 4: Hydrodynamics block inputs and outputs

Figure 23: Hydrodynamics block

The main equations that need to be solved in this block are:

1
𝐹= 𝜌 · 𝑣𝑤 2 · 𝑆𝑟𝑒𝑓 · 𝐶𝑜𝑒𝑓𝑓
2
1
𝑀= 𝜌 · 𝑣𝑤 2 · 𝑆𝑟𝑒𝑓 · 𝐶𝑟𝑒𝑓 · 𝐶𝑜𝑒𝑓𝑓
2
The density 𝜌, the wind velocity 𝑣𝑤 , the reference surface 𝑆𝑟𝑒𝑓 and the reference chord 𝐶𝑟𝑒𝑓 are
inputs of the block. So, the only variables that need to be computed in this block are the 3 force
coefficients 𝐶𝐷, 𝐶𝑌, 𝐶𝐿 and the 3 moment coefficients 𝐶𝑙, 𝐶𝑚, 𝐶𝑛. This can be seen in the
following image:

Figure 24: Hydrodynamics structure

28
Pol Bernad Serra
Hydrofoil Design and Construction

In order to compute the coefficients, this block imports data from a NASA’s aerodynamics open
source software: OpenVSP. This data is imported to Matlab where is selected and organized in
a matrix which will be used by Simulink. Then the hydrodynamic block can compute the
coefficients as function of 𝛼 and 𝛽 and interpolate when it’s needed.

Figure 25: Coefficients calculator structure

In addition, in this block is where the depth effect is considered. This phenomenon reduces the
hydrodynamic forces the closer the hydrofoil gets to the water surface. To model this situation
2 identical wings with a symmetrical foil were positioned at a variable vertical distance. The
upper wing had 𝛼 = −3º and the lower one 𝛼 = +3º. The non-variable pressure line is located
in between the wings. Having the lift and half the distance between the two wings is possible to
adjust a function that describes the forces as function of the submergence. It is important to
impose 0 hydrodynamic forces value when the hydrofoil is out of the water. This effect was
studied using another open source aerodynamic software called XFLR-5.

The aerodynamic block is now able to compute the forces and moments produced by the
hydrofoil in wind reference frame so an Euler angles transformation will be applied in the
summation block.

29
Pol Bernad Serra
Hydrofoil Design and Construction

Height CL CL Norm
0,010 0,01 0,015
0,250 0,444 0,663
0,050 0,488 0,728
0,100 0,548 0,818
0,250 0,623 0,930
0,500 0,657 0,981
Table 5: CL as function of submergence

Figure 26: Depth effect

30
Pol Bernad Serra
Hydrofoil Design and Construction

3.4.3 Control Block

Inputs Outputs
Toggle (On/Off) Control Moments
Depth
Angles
Table 6: Control block inputs and outputs

Figure 27: Control block

The control block task is to mimic the response of a person who is over the surf board and tries
to stabilize it.
In order to recreate the response of a person, a negative feedback is implemented. The desired
angular position is imposed, and the actual angle of the foil is subtracted. The result is the error
between the premise and the signal. Then, a PID can be applied to that error in order to create
a response against fluctuations. However, the human response is not immediate, there is a
delay; the human response is about 0.25 seconds (Human Benchmark, 2019). So, all PID have a
0.25s delay. Furthermore, the person over the surf board is able to produce limited torque.
These torque limitations have considered different conditions for each direction:
• Pitch (Y axis): the feet are 0,5m separated and the 620N of weight can be located at the
right or left foot completely, so:
0.5
𝑀 = 𝐹 · 𝑑 = 620 · 2
= ±155𝑁𝑚.

• Roll (X axis): the feet are 0,25m long and the weight can be located at the tip or at the
0.25
heel, so: 𝑀 = 𝐹 · 𝑑 = 620 · 2 = ±77.5𝑁𝑚.

• Yaw (Z axis): the yaw moment comes from inertial rotation forces of the whole body.
This means the torque is not sustained on time. This makes it difficult to model so, a
quarter of the maximum Roll moment has been assumed to be the maximum value 𝑀 =
20𝑁𝑚.
There is another PID for recreating the will of maintaining a constant height above the water
surface which is the same as a constant depth. It follows the same negative feedback structure
before explained. This moment is added to the pitching torque.
The constants KP, KI and KD, which multiplies the proportional, integral and derivative parts of
the PID respectively, is what an experienced hydrofoil surfer would be adjusting in order to
respond accurately to perturbations. In this study, the constants have been tuned iteratively
comparing the different responds obtained.

31
Pol Bernad Serra
Hydrofoil Design and Construction

Finally, in this image it can be seen the internal structure of the control block:

Figure 28: Control block structure

32
Pol Bernad Serra
Hydrofoil Design and Construction

3.4.4 The smaller blocks

There are 4 smaller blocks that feed the main 3 which are:

• Angles transformation: in this block is where all the different angles are computed in the
different reference frame. There are three types: the angles of the wind with respect to
ground, the angles of the wind with respect to the body and the angles of the body with
respect to ground. In this block is where the depth is also computed assuming that the
wave has a certain slope.

• Geometry: this block is where properties needed for other blocks such as mass, inertia
or the position of the centre of mass are introduced.

• Summation: in this block is where the forces and moments coming from the
hydrodynamic and control blocks are added. Then, an Euler transformation is needed to
pass from wind reference, where the hydrodynamic forces and moments are referenced
to, to ground reference (Gómez & Pérez, 2012).

• Initial Conditions: the initial velocity and position are introduced in this block. This can
be used to simulate perturbations setting the initial pitch angle position to, for example,
5º.

33
Pol Bernad Serra
Hydrofoil Design and Construction

3.5 Geometry comparison

In this section is where the geometries are introduced in the Simulink platform and their results
are compared. Three graphs are presented for each study case: body angles, the velocity and
the submergence of the wing. All these graphs are function of the time in seconds. To simulate
perturbations, the initial angle 𝜃 = −0.2 𝑟𝑎𝑑 and the initial angle 𝜓 = 0.2 𝑟𝑎𝑑.

3.5.1 Alpha

Alpha is able to ride a wave of at least 𝛿 = 10º of inclination. The following results were
obtained with the control turned off, which means the rider do not react to the perturbations:

Figure 29: Alpha body angles control off

34
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 30: Alpha velocity control off

Figure 31: Alpha submergence control off

35
Pol Bernad Serra
Hydrofoil Design and Construction

And next, the results of the platform with the control turned on, so, the rider tries to minimize
the oscillations:

Figure 32: Alpha body angles control on

Figure 33: Alpha Velocity control on

36
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 34: Alpha submergence control on

3.5.2 Beta

Beta needed at least 𝛿 = 19º of wave inclination to reach the take-off speed. The following
results are with the control turned off:

Figure 35: Beta body angles control off

37
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 36: Beta velocity control off

Figure 37: Beta submergence control off

38
Pol Bernad Serra
Hydrofoil Design and Construction

The next graphs show the results of Beta with the control turned on:

Figure 38: Beta body angles control on

Figure 39: Beta velocity control on

39
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 40: Beta submergence control on

3.5.3 Comparison

From Alpha, it can be seen in Figure 29 and in Figure 32 that over angle 2, which is 𝜃, the control
does not have an appreciable effect. However, the angle 3, which is ψ, needs the control to
dissipate the oscillation. In both Figure 30 and Figure 33 can be seen that the velocity needed to
maintain the flight is 6.5m/s. It can be seen in Figure 31 and Figure 34 that when the control is
turned on the height of the board becomes steady faster than when it is turned off. Also, the
submergence of the foil without control is higher.
From Beta, it can be observed the same phenomenon reported in Alpha, the control is needed
for stabilizing the angle ψ, as it can be seen in Figure 35 and Figure 38. However, in these figures
and also in Figure 37 and Figure 40, when the control is turned on it becomes more oscillating.
Finally, in both Figure 36 and Figure 39 the speed needed to maintain the flight is 11m/s.
These results show that Beta should have worst behaviour than Alpha, is more oscillating, it does
not beneficiate from the active control and its take-off speed is almost twice the value of Alpha’s.
For this reason, the wave needed for Beta is also almost twice as sloped. Considering that low
wave conditions was and objective, is clear that Alpha should be better option. However, both
geometries will be constructed at scale and tested. After the tests, a decision making will choose
the final geometry to build at real scale.

40
Pol Bernad Serra
Hydrofoil Design and Construction

4. Model

The objective of this part is to build at 1:5 scale model that is able to accept different plane
configurations printed with a 3D printer. The centre of mass position needs to be variable at will.
The tests developed in this section will be used to decide the final geometry, but they can also
be compared with the Simulink platform results. If those results coincide, in the future, the
Simulink platform could be used to assess the dynamic stability without needing the scale tests.

4.1 Construction

In the following pages, the construction of this model will be explained part by part. The
following figure is an image of the whole model in a test.

Figure 41: Scale model test

41
Pol Bernad Serra
Hydrofoil Design and Construction

4.1.1 Geometry

The first step was to export the Open VSP geometry to a Cad software, in this case, Autodesk
Fusion 360. This software has a student license which is open source. The Open VSP geometry
was exported as and STL, which describes the geometry with triangles. Also, the analysis
geometry has a shorter vertical stabilizer because not all of it is inside the water when the
hydrofoil is lifting. This was then imported in fusion 360 which converts the hole shape into a
body. The wing, the tail and the fuselage had to be combined because they were interpreted as
separated bodies. Then, a perpendicular support plate to the vertical stabilizer was designed
with 4 holes for M3 screws. The model needed to be scaled to the correct size, 1:5. This is the
maximum scale that fit on the 3D printer. Before printing, the whole model was cut by its X-Z
symmetry plane. This made possible to print the wing and tail in the best orientation for avoiding
layer manufacture inaccuracy and also avoids the use of support material. The central fuselage
was also enlarged in order to resist the bending loads.

Figure 42: Open VSP geometry Figure 43: Ready to print geometry

From fusion 360 the model is exported in STL

to the slicer program Simplify3D.
The printing settings were 100% infill density
in order to obtain a solid piece of plastic which
is more resistant and also has no lick in water
problems. The layer high was set to 0.25mm
with 3 outside perimeters. The nozzle
temperature was 235 ºC and the heated bed
was set to 60ºC. The plastic used was PLA from
Sunlu brand. Each half took 5h to print and 82g
of plastic were used.

Figure 44: 3d printed Alpha geometry

42
Pol Bernad Serra
Hydrofoil Design and Construction

After the 2 halves were printed, the inner flat surface was sanded, preparing it to be glued with
epoxy glue of Supertite brand. This glue is optimal for this job because it is water resistant and
sticks very well the plastic used to print. After assembling the two parts together, the excess
glue of the joint was sanded to obtain a smooth transition between halves.

Figure 45: Surface treatment for gluing Figure 46: Gluing process

This explanation was for the Alpha configuration.

However, Beta was done exactly like described
before but adding a plastic arm to place the wing
in the desired position. In addition, for Beta
design was not possible to avoid support material
because the winglet that acts like a vertical
stabilizer was perpendicular to the wing. Each
half took 7h to print and 115g of plastic were
used.
Figure 47: 3d printed Beta geometry

Next, a photo of the final result for each design is presented.

Figure 48: Final Alpha geometry Figure 49: Final Beta geometry

43
Pol Bernad Serra
Hydrofoil Design and Construction

4.1.2 Surf Board

A wood surf board have been constructed to contribute floatability and central attachment for
all the components. The explanation of its construction will be found in the Annex section.

4.1.3 Sliding Mass

The objective of the sliding mass is to control the centre of mass position at will.
The design chosen is a 3d printer like motor with a pulley on its axis and a static belt. The battery
was mounted over the motor and the whole weight can slide on two rails and can precisely
control its position. The first idea was to implement an accelerometer and a PID to active
stabilize the model but this idea was declined as it will be explained later on. Instead of the
accelerometer, the position was finally controlled by radio control receiver. Furthermore, all the
electronics need to be in a waterproof box.
The first step was to design everything in fusion 360. One major problem was the need to put
everything inside a hermetic box. So, there is no space for oversizing the parts. In fact, the first
idea was to build an acrylic box and all the dimensions were adjusted to 100mm wide but when
all the pieces were printed and was successfully working a better solution was found; which was
an acrylic food case found on the local store. Unluckily, the maximum wide accepted by this case
was 80mm. Also, the height was limited to 90mm. However, the food case was overall better
option because of its industrial manufacture which ensures hermeticity. So, a redesign and
reprint of the pieces was done. There are 8 3d printed pieces in this construction. All of them
were printed with the same settings as the wings, 100% infill and pla plastic. The new wood base
was CNC cut. The next image shows all the components bought for the construction of the sliding
mass.
• A nema 17hs4401 stepper motor with its A4988 stepper driver
• 1m of gt2 open belt
• A 32 teeth gt2 pulley
• An Arduino Uno
• A 3 cells Lithium Polymer battery with 1.3 Ah of capacity
• The accelerometer MPU6050 and the radio transmitter Turnigy 6ch are not in this figure
because it was not necessary to buy them.

44
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 50: Components of the sliding mass system

The central piece is the one that holds the motor. It also holds the two 608 bearings. These are
used to guide the belt throw the pulley attached to the motor axis. This is because the motor is
mobile and the belt is static. This central piece also acts as the linear bearings for the 8mm
aluminium tubs that act as linear guides. This was possible by making the hole slightly bigger
than the tub diameter, 8.2mm, together with some grease to reduce the friction. This piece
required support material and took 4h to print. A few images are now attached to show this
mechanism.

Figure 52: Central piece printed Figure 51: Central piece assembled

45
Pol Bernad Serra
Hydrofoil Design and Construction

There are 4 pieces that act as a tub holder to join the two aluminium tubs to the piece of wood.
There are 2 more plastic pieces which fix the belt to the wood. The last piece is a battery and
Arduino holder which is mounted over the motor. These other pieces took 3h to print.

Figure 54: The first design approach Figure 53: The final design

The next point is the electric circuit. The brain of this circuit is the Arduino Uno. The Arduino
Uno can be programmed using the Arduino IDE, an open source software based on C++. The
microcontroller is powered by the 12v LiPo battery. It has a dc to dc converter which can reduce
the voltage to 3.3 and 5V. These voltages are suitable for powering the other devices: the
accelerometer, the radio receiver and the logic part of the stepper driver. The Arduino receives
information from the accelerometer or the radio controller, it processes the inputs and
elaborates a suitable output which goes to the stepper motor driver. The driver transforms the
serial data into the input for the motor which is the high-power part of the whole electric circuit
and it is powered directly from the battery. Then, the motor spins precisely and maintains its
position until the next move.

46
Pol Bernad Serra
Hydrofoil Design and Construction

A connections photo is attached but with the Arduino Nano instead of the Uno as will be later
explained. As it can be seen, the Arduino Nano and the stepper motor driver needed a pcb to be
connected. These pcbs were made out of a virgin breadboard and the inner connections were
done using tin channels.

Figure 56: Connections photo

Figure 57: Custom driver pcb Figure 58: Tin channels Figure 55: Custom Arduino NANO pcb

As it can be seen in Figure 54, which was the first design before the food case, the height was
not a problem and the Arduino was mounted over the battery and the battery over the motor.
However, the food case introduced height limitation and the Arduino Uno was not an option so
an Arduino Nano was bought. The Nano has almost the same features than the Uno in a much
smaller form factor. The C++ code is the same for both. The code uses interrupts to send
information to the motor driver and once every 10ms the accelerometer is read and the desired
position changes with the angle of the accelerometer. If the surf board pitch down the mass
move backwards and it moves to the front if the surf pitch up.

47
Pol Bernad Serra
Hydrofoil Design and Construction

The design of the Figure 54 was fully functional and had no problems. Then, the dimensions
reduction made the Arduino Nano better option than the Uno. But there was a problem with
the Arduino Nano. After elaborating the pcb and reconnecting all the cables; the same code
uploaded to the Uno did not work properly on the Nano. The mass slide properly during a few
seconds but after that it always moved to one end of the rail and crashed. After revising all the
connections, the problem was not resolved. Suddenly after a few tries, the accelerometer blew
up. After buying another accelerometer and installing it again, the randomly crashes did not
resolve. Then, the Arduino Uno which has worked well was connected again. Unluckily, the
connections were wrong and the microcontroller blew up because of a short circuit. After buying
another Arduino Uno and soldering all the connections again, the same problem of the Arduino
Nano started to occur. The system just worked during about 20 seconds. It was not a code
problem because disconnecting the motor and observing the plot signal of the Arduino IDE there
were no cracks or suddenly stops, everything was smooth and running right during more than 1
hour. After an online search, it was found in (Arduino Forum, 2019) this problem can come from
irregularities in input voltage produced by the a4988 driver and the solution found was to
connect a ceramic capacitor of 100nF in parallel with the 5 or 3.3V and the ground. This method
blew up the accelerometer again and it was at this point where the idea of active stability
controlled by the accelerometer was discarded. Next, an image of all the components bought
that ended burnt or without use:

Figure 59: Failed electronic components

The final solution was to connect a radio-controlled receiver which can be read by the Arduino
and transform the signal to be proportional to the position of the mass. In this manner, the stick
position of the transmitter controls the position of the centre of mass of the surf board. Next, a
photo of the final sliding mass system.

48
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 60: Final sliding mass system

Finally, to ensure a proper water resistance, a plastic like window seal was glued to the
perimeter of the wood piece in the location where the food case made contact. Some plastic
flanges ensure a strong union between the acrylic case and the wood. Two 300mm aluminium
tubs with a diameter of 12mm where located between the surf board and the sliding mass in
order to have the centre of mass located at 300mm above the wing tip, just one fifth of the
distance between the wing tip and the centre of mass of the real model.
At this point, the model was ready for the tests. A test bench was needed to develop the test
and will be explained in the next section. A few images of the final model constructed are
attached.

Figure 61: Final scale model constructed

49
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 63: Render of the model Figure 62: Real model for comparison

Characteristics Value
Mass 1910g
Floatability 15N
Centre of mass displacement 46mm
Speed range 0-3m/s

Table 7: Model characteristics

50
Pol Bernad Serra
Hydrofoil Design and Construction

4.1.4 Articulated quadrilateral

The bench test function is to create a controlled environment where the desired characteristics
can be studied. In this case was very important to restrict some degrees of freedom because
there are six in total. However, only two of them have been studied, the vertical position and
the pitching angle.
In order to restrict the other 4 degrees of freedom an articulated quadrilateral has been
developed. This quadrilateral needs to have a tub at the end where two bearings are located
and attached to the surf board. In this way, the quadrilateral allows vertical movement and the
bearings allow pitching freedom. This idea was inspired by the trapezoidal suspension of a car
and its wheel bearing. Some images are now attached for the better understanding of the
system:

Figure 65: Articulated quadrilateral build Figure 64: Articulated quadrilateral on a test

The big piece of white wood where the quadrilateral is attached is collocated on the boarding
of a boat which can vary its speed. So, at the end, the speed of the board can be controlled, it
cannot move sideways, the roll and yaw rotations are restricted and the only free movements
are the vertical position and the pitching angle.
In order to control the pitching moment and the vertical position, two ropes were fixed to the
surfboard. One of them was located at the tip tub of the quadrilateral so it only affects the
vertical position. The other was fixed at the front aluminium tub and allowed to rotate the model
at will. These ropes can be seen in Figure 64.
The construction of the articulated quadrilateral consisted in 4 aluminium tubs, two of them
were 1m long and the other two were 30cm long; all of them were 12mm in diameter. The short
ones were located at the root and form a 45º angle. Their propose was to avoid bending

Figure 67: Plastic tub ends joints Figure 67: Aluminium hinge inserts

51
Pol Bernad Serra
Hydrofoil Design and Construction

moments on the root joints. All the white pieces that can be seen on the Figure 65 were 3d
printed. There was a total of eleven printed parts which took 12 hours to print. Six of them were
tub ends with a 6mm hole. Short aluminium tubs were located in those holes to act as a hinge
in conjunction with M4 inox screws.
There were two more pieces which act as the other half of the hinge between the white wood
and tub ends. In these parts was where the screws were tightened. The bearing axel was also
joint with a plastic piece as it is shown in the next image:

Figure 69: Quadrilateral root Figure 68: Quadrilateral tip

Finally, two kp008 bearings were attached to the surfboard. The position of them was modified
after the first test, as it would be explained on the next section.

Figure 71: First bearings position Figure 70: Definitive bearings position

52
Pol Bernad Serra
Hydrofoil Design and Construction

4.2 Tests

All the tests developed in this study were qualitative. In these tests, there were 3 aspects to
evaluate.

• Stability: is the most important aspect. To evaluate it, only the central rope was pulled.
This means that the whole model was free to pitch. In this situation, the surf board
needed to be able to maintain an angle. If it did, the other rope (the one that can change
the angle) could be pulled to simulate perturbations. The board needed to dissipate the
perturbations and return to its equilibrium position.
• Lift at its equilibrium point: this aspect was studied by knowing how much force was
required in the central rope to maintain the vertical position constant. If a lot of force
was needed, the hydrofoil was lifting poorly but if the force required was lower or even
zero, the hydrodynamic forces were compensating the weight.
• Centre of mass position: In this test, the sliding mass is positioned most delayed position
possible. The more delayed the centre of mass, the higher the angle of attack and
greater the lift but it becomes more unstable till the point where is not possible to
maintain the equilibrated position.
Four tests have been done to evaluate the geometries.

4.2.1 First test

The first test was on 11th of April and only the geometry Alpha was been printed for this day. In
this test day, the bearings were located in the centre of mass height. This did not produce any
torque in the centre of mass reference but does not represent the mechanics of surfing. The
force that moves forward the surf board is the component of lift projected forward, like a sailing
plane, which is applied in the aerodynamic centre. Pulling the board from the centre of mass
caused the drag force, multiplied by the 30cm distance in between, to produce a moment that
made the board pitch down. This effect was incremented the faster the test was done. This
leaded to a major problem which was that there was or not enough speed or not enough angle
of attack and the hydrofoil was not able to sustain itself.
To make the board pitch up the centre of mass can be delayed but this also destabilize the
model. With the sliding mass almost at the most delayed position, the foil was able to maintain
an equilibrated angle with only a slight tension on the central rope. However, the smallest
perturbation introduced was not able to be dissipated.

4.2.2 Second test

To solve the problem encountered the first test day, the pull position was lowered on the base
of the surfboard as it can be seen in Figure 70. Ideally, the pulling position should be the
aerodynamic centre where the lift is produced and therefore the pulling force. However, the tip
tub of the articulated quadrilateral would have a major effect on the fluid, so, the lowest possible
position was just over the water. These leads in a noticeable improvement of the tests.

53
Pol Bernad Serra
Hydrofoil Design and Construction

In this second test day, Alpha was also the only geometry available. The new pulling position
made possible that the foil had a much better performance and, in addition with the not so
delayed centre of mass, the model was way more stable and less force was needed in the central
rope; even reaching zero force as it can be seen in the following image. The two ropes are
hanging and not transmitting any force to the model:

Figure 72: Second test. No tension on the ropes

This was a successful test and established a reference to compare with the other geometry. The
tension in the rope, the level of perturbation dissipation and the optimum and most delayed
position of the sliding mass.

4.2.3 Third test

For this test day the geometry Beta was already printed and prepared to be used. Even though,
the first part of the test was done with Alpha to corroborate the results obtained in the second
test.
After developing the test, the geometry Beta did not obtain the same results as Alpha. The
equilibrium angle was possible but the lift produced at this position was small and a lot of force
on the central rope was needed. In addition, the perturbations were fast amplified and the
equilibrium was lost in a few seconds. Moreover, the mass needed to be past the half line, any
more delayed position was not able to be equilibrated. It also was very sensible to mass moves.

54
Pol Bernad Serra
Hydrofoil Design and Construction

4.2.4 Fourth test

This test was a replica of the third session, the two different geometries were put to the test.
However, Aleix Báez, the director of this degree final project was able to assist and help with the
development of the test. The results obtained where very similar in comparison with the third
session, overall, the geometry Alpha had a better performance than Beta. Nevertheless, to take
the adequate geometry an OWA decision making needs to be done because there are other
factors which affect the decision.

55
Pol Bernad Serra
Hydrofoil Design and Construction

4.3 Decision making

In order to take the appropriated decision an ordered weighted averaged method will be used,
an OWA. To use this methodology, all the parameters need to be explained and weighted
before. All the parameters will have a punctuation from 0, meaning very poor behaviour, to 5
which is perfect performance. Next, a list of the parameters considered.

• Stability: it represents the ability of dissipating the perturbations induced and remaining
on the equilibrated angle. It has a weight of 5.
• Lift: it represents the sustentation force made by the foil at its equilibrated position. It
has a weight of 3.
• CoM: centre of mass, it represents the range of possible positions of the CoM where the
board remains equilibrated. It has a weight of 2.
• Molds: the number and difficulty of the molds that need to be manufactured with the
CNC machine in order to create the outer shape of the real scale geometry. It has a
weight of 4.
• Structure: The internal stainless-steel structure needed to withstand the loads. It has a
weight of 3.
• Cost: The material cost. It has a weight of 2.
• Aesthetics: how much beautiful it looks, it is a subjective aspect. It has a weight of 1.
OWA Alpha Beta
Criteria P PxWeight P PxWeight Weight
Stability 4 20 2 10 5
Lift 4 12 2 6 3
CoM 3 6 2 4 2
Molds 3 12 2 8 4
Structure 1 3 4 12 3
Cost 3 6 3 6 2
Aesthetics 2 2 4 4 1
Sum of PxWeight - 61 - 50
Max of PxWeight 100
OWA 0.61 0.5

Table 8: OWA decision making

The OWA shows Alpha as the geometry to choose and therefore it will be the one to build in real
scale.

56
Pol Bernad Serra
Hydrofoil Design and Construction

5. Real Scale Model

In this section the construction and testing of the real scale model of the geometry Alpha will be
explained.

5.1 Structural design

The structural design will be separated in two parts, the internal stain-less structure and the
composite material lifting surfaces.

5.1.1 Internal structure

The internal structure is the responsible of withstand the loads produced by the wing and tail
and transport them into the board where the rider stands on. Next, a photo of the final model
on fusion 360:

Figure 73: Final internal structure render

Stainless steel has been the choice to build the internal structure because its oxidation
resistance, its strength and its ability to be soldered and accomplish strong joints. From the
design, the distance between the leading edge of the wing and tail is 1m. So, the central tub will
have the same length. In order to size the tub, the maximum load factor that will be assumed is
n=4. This means that if the total weight is 800N, during a n=4 maneuver the lift produced is
3200N. It can be assumed that the wing is doing all the lift, to be more conservative. Since the
tail has 3 times less surface than the wing, it can also be assumed that the maximum lift
produced by the tail it is one third of the produced by the wing, 1067N.
It was at this point when it was clear that a change in the design was needed. In the first
approach of geometry Alpha, there was not vertical stabilizer because the vertical support acted
as it. To do so, the vertical support needed to be delayed leaving a large distance between it and
the wing; the principal font of hydrodynamic forces. In the first approach, the vertical stabilizer
was located at 0.6m of the wing. This would produce a large torque as it can be seen in the next
image:

57
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 74: First central tub wing load diagram

The central tub needed to withstand 1920Nm. This is a very high torque and would need a large
inertia which would be excessive for resisting the tail loads. The next image describes the
situation:

Figure 75: First central tub tail load diagram

In the tail case, the tub would need to resist a moment 6 times smaller. To solve this problem,
the vertical support can be located at 0.25m from the front end and, because the central tub is
1 m long, at 0.75m from the back end. This leads in the same torque produced by the wing and
tail at maximum lift condition as it can be shown in the following images.

Figure 76: Final central tub wing load Figure 77: Final central tub tail load
diagram diagram

𝑀
This means that using the structures equation: 𝜎 = · 𝑦 (Wikipedia Bending, 2019), the
𝐼
maximum stress which is able to resist the stainless steel: 550 MPa (Azo Materials, 2019) and
𝜋
the inertia of a circular hollow tub: 𝐼 = 4 · (𝑟𝑜𝑢𝑡 4 − 𝑟𝑖𝑛 4 ) (Wikipedia Second Moments of
Inertia, 2019) the 30mm tub with 3mm wall give these results:
𝜋 𝜋
𝐼= 4
· (𝑟𝑜𝑢𝑡 4 − 𝑟𝑖𝑛 4 ) = 4 · (15𝑚𝑚4 − 12𝑚𝑚4 ) = 2.35 · 104 𝑚𝑚4

𝑀 8 · 105 𝑁𝑚𝑚
𝜎= ·𝑦 = · 15𝑚𝑚 = 510 𝑀𝑃𝑎
𝐼 2.35 · 104 𝑚𝑚4
𝜎𝑚𝑎𝑥 550
𝜂= = = 1.08
𝜎 510

Once the central tub was chosen, the vertical support tub will be the same because it implicates
a cost reduction. Both the central and the vertical tub are chosen to be circular in order to reduce
drag.

58
Pol Bernad Serra
Hydrofoil Design and Construction

5.1.2 Lifting surfaces

The lifting surfaces will be constructed in composite materials. More precisely, in chopped
glass fibre of 300g/m2 with unsaturated polyester resin as matrix.
The structure chosen to build the real scale model is monocoque. To avoid buckling of the fibre
parts polyurethane expansion foam will be added as a solid core. This technique will also prevent
water to enter inside. Since the foam expands and adopts any shape, the minor manufacture
problems such as irregular skin thickness will self-address. It will be used in the tail, in the vertical
stabilizer and also in the wing. In addition, in all of them there will be a central stainless-steel
beam which will also take part of the loads. However, for this study only the fibre resistance will
be considered.
As will be later explained in the section 5.2.3 Lifting surfaces, a first trial was done to know the
fibre to resin ratio and the thickness of the composite as function of the number of layers. These
values are 60% in weight and 0.375mm per layer respectively.
To be more conservative, a constant distributed load will be considered in this study. In the next
image, only half of the wing is showed in the load diagram. The wing distributed load is 3200N/m
at n=4 since the span is 1m and the load 3200N. The same applies for the tail but since its span
is 0.5m and the load at n=4 is 1067N, the distributed load is 2134N/m.

Figure 79: Wing load diagram Figure 78: Tail load diagram

For the wing, it can be seen that the total moment applied at the root is 400Nm. The maximum
tensile strength of the glass fibre with 60% of fibre to resin ratio cured at room temperature is
130MPa (Elahi, Hossain, Afrin, & Khan, 2014). However, the maximum compression strength of
the composite is 90MPa (CES Edupack, 2019). To avoid fatigue problems, the stress supported
should never exceed the 50% of the maximum tensile strength so, the value used to compute
the number of glass fibre layers will be 45MPa. From the design section, the geometry of the
root wing is known, its chord is 300mm and it is 45mm thick. The minimum inertia of the section
can be computed using the equation below:
𝑀 4 · 105 𝑁𝑚𝑚
𝐼= ·𝑦 = · 22.5𝑚𝑚 = 2.00 · 105 𝑚𝑚4
𝜎 45𝑀𝑃𝑎

The inertia of the root section is computed using an online free software (Online Wing Bending
Inertia Calculator, 2019). Next, the value of the Inertia for a 2mm wall thickness:

59
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 80: Naca 23015 bending inertia (Online Wing Bending Inertia Calculator, 2019)

The inertia is 2.6·105mm4, but since the thickness is discrete, six layers of glass fibre were applied
to both halves of the wing because either of them can be under compression. Six layers of fibre
means 2.25mm wall thickness and this represents and inertia of 2.88·105mm4. This represents a
𝐼𝑚𝑎𝑥 2.88·105
safety factor of: 𝜂 = 𝐼
= 2.00·105 = 1.44 .

For the tail, the moment at the root is 67Nm, the chord and thickness are 180 and 22mm, so the
inertia needed is:
𝑀 6.7 · 104 𝑁𝑚𝑚
𝐼= ·𝑦 = · 11𝑚𝑚 = 1.64 · 104 𝑚𝑚4
𝜎 45𝑀𝑃𝑎
The inertia of the tail with 4 layers each half would be 3.02·104mm4 which would give a safety
𝐼 3.02·104
factor of: 𝜂 = 𝑚𝑎𝑥𝐼
= 1.64·104 = 1.84 which is a bit too high. However, since the lower part was
made with 4 layers and 1.49mm skin thickness, the upper part was made with the same number
of layers. The vertical stabilizer was also made with 4 layers of glass fibre each half which is also
more than what is needed but it can resist better the possible hits it could have. One example
of hit will be explained in the annex section.

60
Pol Bernad Serra
Hydrofoil Design and Construction

5.2 Construction

The construction method used in this model has been an internal stainless-steel structure with
fibre glass plastic reinforcement wings and wood and foam sandwich structure for the board.
The construction will be separated in four different parts:
• Internal Structure: the internal stainless structure which have required arc soldering.

• Molds: the fusion 360 design and toolpath, the CNC machining and the post treatment
to achieve good surface finish.
• Composite material skin: the lamination process using vacuum technology. Also, all the
joints and final sanding of the parts.
• Board: the board construction which have to develop enough resistance and floatability.

5.2.1 Internal Structure

The joints of all the stainless-steel tubs have been done with arc soldering with 2mm electrodes
of Castolin brand. To obtain a good surface contact and ensure resistance, all the circular unions
needed a tip shape which matched the shape of the other tub. To avoid soldering deformation
as much as possible, the perpendicular joints where always ensured with a machined to square
plate of aluminium.

Figure 82: Welded joint Figure 81: Ensuring perpendicularity with a

machined to square aluminium plate

61
Pol Bernad Serra
Hydrofoil Design and Construction

Also, shorter tubs were soldered in the points where the fiberglass parts need to be joined with
the internal structure. This increments the surface of bounding and also allows glass fibre
reinforced joints as later will be explained.

Figure 83: The results after the first welding day

The upper horizontal tub does not go underwater and can be square shaped. Its function is to
rigidize the whole board. Two lateral tubs were also soldered to ensure that the surf board was
properly attached to it. It is also important to notice that the two perpendicular joints were
reinforced with diagonal tubs. The next image shows the finished internal structure with the tail
already joined.

Figure 84: Finished internal structure

62
Pol Bernad Serra
Hydrofoil Design and Construction

5.2.2 Molds

The elaboration of the molds is the most important part of the construction. They are
responsible of the outer shape of the foil and must truly represent the geometry elaborated in
the design section. Their surface finish will also have huge effect on the performance of the
hydrofoil.

Figure 85: Finished wing mold

The first step is to design the molds in Fusion 360. To do so, there are some limitations that need
to be considered. The CNC dimensions and the stock material used.
The maximum machining size, 500x500x120mm, is not big enough to machine the whole wing
all at once because its span is 1m. to solve this problem, the wing mold needs to be divided in
two parts. The stock material used in the elaboration of these molds was mdf. The reason is
because its low price and its lack of wooden betas. It is also easy to machine and can be sanded
until very good surface finish. However, the deepest section of the upper wing mold was 28mm.
There were no thick enough mdf in the local store. So, two pieces of 16mm were glued together
to form a 32mm thick piece of mdf. The glue used for the joints was wood glue from the brand
Ceys. So, the stock material dimensions will be 500x500x32mm of mdf.
Once the stock is chosen, the Fusion 360
design can continue. The first operation
was to separate the wing in four parts,
the upper and lower section of each side.
Then, the stock material was drawn as a
500x500x32mm square box. Next, the
wing shape is cut from the box to obtain
the final mold geometry. In Figure 86 a
render of the shape which will be
machined is shown.

Figure 86: Wing mold render

63
Pol Bernad Serra
Hydrofoil Design and Construction

Now, all the 5 machining operations will be explained.

• Face: the face operation was used to obtain a flat surface and correct the possible
deviations introduced when the piece was screwed to the CNC table. The stepover of
each pass was 8mm and the cut was 1mm deep. Next, a photo of this operation is
presented:

Figure 87: Face operation toolpath

• Adaptative clearing: this was the roughing operation, the responsible of the major
material elimination. It consisted in making deep passes, 5mm, but with low tool load,
4mm. This allowed a bigger part of the tool flute to be in contact with the material and
has reduced its tip wear. The finest passes of this operation leaved 1mm stairs. Next, a
photo of the operation and another of the actual machining:

Figure 89: Adaptative clearing toolpath Figure 88: Adaptative clearing machining

64
Pol Bernad Serra
Hydrofoil Design and Construction

• Parallel: This was the finishing operation which leads the smoothest surface finish. It
consisted on parallel passes separated 2mm between them. The direction of cut was
parallel to the X axis. This has ensured a good finish on the leading and trailing edge.
The following images shows the toolpath and the real cutting operation:

Figure 90: Parallel machining

Figure 91: Parallel toolpath

• Parallel 2: This operation was like the one explained before but 90 degrees shifted to
finish the tub section where the fuselage is attached. This is shown in the following
image:

Figure 92: Parallel 2 toolpath

• Contour: this final operation is to clean the edges and ensure the perpendicularity of
the face which will be joined with the other half. The passes were 2mm deep.

Figure 93: Contour toolpath

65
Pol Bernad Serra
Hydrofoil Design and Construction

All the operations have been done at the maximum feedrate able for the machine,
3000mm/min. This is because the mdf needs a very high surface cutting speed, 195m/min and
a chip load of 0.5mm with these values, the optimal feedrate needed would be 7500mm/min at
6200rpm as it is computed in the online feedrate fablab calculator (Fablab Speed and Feeds
Calculator, 2019). If the surface cutting speed and the chip load are lowered to 160m/min and
0.3mm respectively, the obtained feedrate is 3000mm/min at 5000rpm. However, the rpm in
this CNC machine cannot be known so they are adjusted until good chips are obtained.
The next step was to glue the molds together. To do so, wood glue and screws were used. After
the glue had dried, the sanding began. The first attempt to sand the surface made clear that
small fibres of the wood would never disappear because they were not rigid enough to be
sanded. To solve this problem, all the surface was varnished with 4 layers of mtn brand spray
varnish and then sanded again. This time, all the small fibres were rigid enough to be eliminated
and a smooth glossy finish was accomplished. The sanding process started with 120 grid sand
paper for the roughing sanding and ended with 400 grid paper for the smoothest finish. The next
pictures show the difference finish quality before and after applying the varnish and sanding:

Figure 95: Surface finish before treatment Figure 94: Surface finish after treatment

The last step of the molds preparation was the application of 4 layers of wax from Liberon brand.
The wax ensures a good demolding and prevents the resin to stick to the mold. All the four layers
were applied using a cotton cloth.
For the rest of the molds the process was exactly the same. They were a total of 7 pieces of
wood machined. The time needed for machining the 4 wing parts was 1h 30min for each one,
then 1h for each one of the tail and also 1h for the vertical stabilizer one. These 3 last pieces
were able to be machined each at once, without gluing two finished molds. This was because
the span of the tail was exactly 500mm. This made a total machining time of 9h but the
preparation of each piece of wood took a lot of time in between.

66
Pol Bernad Serra
Hydrofoil Design and Construction

The molds section ends here and some images of the final results are showed next.

Figure 96: Wing mold after varnishing

Figure 97: Finished tail mold

67
Pol Bernad Serra
Hydrofoil Design and Construction

5.2.3 Lifting surfaces

It has been chosen to work with composite materials because is the actual technology and it
presents an interesting challenge because of its extended applications once the technology is
mastered. It can adopt almost any shape and, after curing the resin, the mechanical properties
are very good. It can also be used under water without any major considerations apart from
using the appropriated resin to do so. Despite the vanguard technology is carbon fibre
reinforcement plastic, in this case glass fibre has been used. The reason is because the high price
of the carbon fibre and the resin used with it. If the model had been made out of carbon fibre it
could had been lighter but the mass of the glass fibre in comparison with the mass of all the surf
board, rider included, represents a very low percentage and does not compensate the cost
increase.
The fibre bought to build this model was 10m2 of 300gr/m2 chopped mat glass fibre. This mat
has short fibres, about 100mm each, randomly distributed along the cloth. Its mechanical
properties are lower than the ones of fabric cloths where the fibres are woven but, again, the
price of woven fabrics is much higher and the weight reduction does not compensate the cost
increase, for this project. The resin used is Polyester isophthalic which is adequate for naval
usage (Nazza, 2019). For this job, 2kg of this type of resin were at disposal with its respective
hardener, methyl ethyl ketone peroxide; which is mixed in a 2.5% mass ratio with the polyester
resin.
The methodology used to elaborate all the fibre parts is vacuum forming. This consists in using
the atmospheric pressure to compress all the fibre layers against the mold, ensuring that the
cloth adopts the desired shape and also extracts all the excess resin. The order of the different
layers is: the mold, the demolding agent, the impregnated composite, a perforate film of
polyethylene which does not stick to the resin and allows the resin excess to pass throw, some
paper to absorb the resin excess and, lastly, the vacuum bag. The air inside the vacuum bag is
extracted using a vacuum pump which was extracted from an old fridge.

Figure 99: Vacuum fridge pump Figure 98: Perforated polyethylene film

The first part elaborated was the lower half of the tail. This part would be used as a first trial to
know which is the thickness obtained in relation with the number of layers and the fibre to resin
weight ratio.

68
Pol Bernad Serra
Hydrofoil Design and Construction

The first step was to elaborate a template to cut the fibre. To elaborate this template, a paper
sheet was pressed to suit the mold and then it was marked and cut. Then, this template was
transferred to the glass fiber mat and cut using a cutter. For this first attempt, 4 layers of glass
fiber were used, which had a total weight of 95g.

Figure 101: Paper template Figure 100: Template transferred to the fibre mat

The second step was to prepare the vacuum bag. A layer of plastic film was placed under the
mold and double-sided tape was added to the perimeter. To attach the vacuum tube to the bag,
a 3d printer adapter was designed. This piece had a flat surface where a big strap of double-
sided tape could be bounded and a hole to let the air pass throw. Next, a photo of the adapter
is presented.

Figure 102: Plastic adapter for the vacuum tub

The third step was to prepare and lay the composite material. For this case, 200 g of resin were
mixed. Then, each layer was collocated on the mold and impregned with resin using a roller,
which also was used to extract big air bubbles and ensure a good positioning of the cloth.

69
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 103: Manual-laying the composite material

Finally, the perforated film was laid over the last coat of resin and over it, the absorbing paper.
Then, the vacuum bag was fully closed by bounding another layer of plastic film with the double-
sided tape before located. Then, the vacuum pump started to suck all the air inside. A very
important step is to press the vacuum bag while the pump is on to avoid the phenomena called
bridging, which consists in the fibre not reaching small concavities. Some pictures of this process
are presented next.

Figure 105: Pressing the vacuum bag to avoid bridging Figure 104: Resin excess being absorbed by the paper

70
Pol Bernad Serra
Hydrofoil Design and Construction

The demoulding process required a spatula but it came out relatively easy. This was the piece
just after extracting it from the mold:

Figure 106: First composite part

The mass of the cured part was 164g and the mass of fibre
used was 95g. This means that the fibre to resin weight
ratio obtained throw this methodology is 58%. Also, the
thickness of the four layers composite was 1.49mm as it
can be seen in the Figure 107. These two values are very
important because they allow to search for information
about the properties and the required layers needed to
withstand the loads. These results mean that every layer
of glass fibre will increase the thickness in average
0.375mm.
Figure 107: Checking the wall thickness

All the other parts were developed following the same procedure. The number of layers for each
part is chosen in the before explained 5.1.2 Lifting surfaces.
Once the 6 halves of the 3 lifting surfaces were done, it was time to joint them with the internal
stainless-steel structure. A very important consideration was that all the pieces were aligned as
they should. To do so, the upper half was placed at its correct position, which was measured
with a level, and then it was glued with the same 5 min epoxy glue used to glue the plastic parts
in Figure 46. Then, strips of glass fibre were laid over the metal tubs, that were previously
soldered, ensuring a resistant bond between the composite parts and the stainless-steel
structure. The reason of choosing the upper half to be glued to the internal structure was
because is the one that is suctioned away from the structure, instead, the lower one is
compressed upwards. Some photos are now shown:

71
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 108: Leveling the tail

Figure 109: Wing joint procedure Figure 110: Tail joint procedure

After the upper halves were bounded to the metal structure, the other halves needed to be
joined. As it can be seen in the image below, the trailing edge joint has enough gluing surface
and can be joined without major problems. However, the leading edge has very small surface
and cannot be joined only with glue.

Figure 111: Joint surface problem

To solve this problem, it was necessary to increase the bounding surface. To do so, the best
solution was to add a fiberglass strip all around the leading edge to obtain this type of joint:

72
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 112: Joint solution

This solution does not only increment the bounding surface it also changes the pulling angle and
now the glue is working at shear stress which is better than tensile. In order to not influence the
design geometry, the joint must be in the inner part. This leads in the problem that once the
other half is put on top, there is no way to control if the strip is in the middle of the joint. To
solve this problem, the glass fibre strip was bounded only with the lower part, leaving the upper
section of the strip without resin. Once the half strip resin was dried, the reinforcement was
locked. At this point, it was possible to apply resin to the other half and accommodate the other
wing half over. To maintain its position, clamps were used in the leading edge and clothespins
in the trailing edge.

Figure 113: Leading edge fibre strip Figure 114: Clamping the halves

The next step was to fill the inner cavity with polyurethane insulation spray foam. To do so, the
spray cannula was inserted throw a small hole. Once there, the application begun. As this type
of spray expands and adopts any shape, only one application point was needed. Because of the
slightly translucence of the fibre glass, the progress of the foam could be visualised and the
complete fulfil confirmed.

73
Pol Bernad Serra
Hydrofoil Design and Construction

Finally, the excess of foam was sanded using sand paper 120 grid and the fibre excess of the
joints was eliminated using an angular grinder. Some photos of the final result are now shown:

Figure 116: Polyurethane foam fill Figure 115: Final result after sanding

74
Pol Bernad Serra
Hydrofoil Design and Construction

5.2.4 Board

The principal objectives of the board are to provide floatability and a rigid surface where the
rider can stand on. As it has been done with the scale board, the explanation will be found in
the Annex section.

Finally, the real scale model was complete.

Figure 117: Finished prototype

75
Pol Bernad Serra
Hydrofoil Design and Construction

5.3 Tests

The following tests were all done with a boat actin as a tug. This has allowed the tests to be
replayed many times a day without needing surf skills, without meteorological dependence and
having an accurate control over the speed.

5.3.1 First test

The first test was done on 19th of May and the director Aleix Báez was present. The way the
board was pulled was using a rope that the rider had in his hands. The first tries seemed that the
tugboat was too slow because the foil was not producing enough lift and the engine was at full
throttle, at 6.5knts, which is 3.2m/s. After a few tries, the lower wood of the board was
decoupled and also the lower foam. To correct this, a rope was rounded around the board to be
able to continue with the tests. The flat tip was also not helping because it was getting inside
the water and lifting up water to the rider’s face which was very uncomfortable. But suddenly,
when the rider moved its body backwards delaying the centre of mass position, the hydrofoil
begun lifting and taking off the water. The first time was a very short flight because the wing got
off the water but after a few tries the rider realised that for taking off the body needed to be
moved backwards but once it was off the water the centre of mass needed to be advanced again
to maintain a stable flight. In this occasion the only tried position was with the belly resting over
the board. In addition, the force needed to start the flight when the board was yet on the water
was very high and tiring for the rider, however, when the foil came out of the water the force
required to maintain the speed was way lower. The following photo was taken at the last flight
which was about 15s long and ended when the left wing got off the water and lost all of its lift.

Figure 118: First test

76
Pol Bernad Serra
Hydrofoil Design and Construction

5.3.2 Second test

The second test was made with a much faster boat. However, the speed did not seem to solve
the poor lift problem. Instead, much more pulling force was needed to maintain a 10knts take
off and the rider get tired much faster. To avoid the fatigue, the rope was passed a few times
throw the door handle attached to the board. This made possible to adopt a different position,
with the knees on the board. In this new position, the centre of mass can be moved more easily
and ended up with a much longer and controlled flight. Once again, the ability to change the
mass position is what made possible to accomplish a successful flight. This means that a stand
position should be better for controlling the board. Unfortunately, the tension needed in the
arm and the slippery wood made not possible for the rider to stand over the board. The next
photo shows the knee position in a successful flight.

Figure 119: Second test

77
Pol Bernad Serra
Hydrofoil Design and Construction

5.3.3 Third test

This test was again developed with the slower boat. The modifications for this test were the
addition of a non-slippery surface but the most important one was that the pulling rope was
attached directly to the board, which made possible for the rider to only worry about his
equilibrium with the help of a rope attached to the door’s handle. Not only the stand-up position
was possible with this configuration, it also was much easy to control and the flight was possible
even at 4knts, however, lower than 5knts was much difficult to sustain. The flights duration was
not a problem anymore. However, an important point is that the instabilities came from
sideways movements, roll instability, not from longitudinal. The position was difficult to recover
when the board tilted to one side. Contrary, the pitch angle was easy to change and maintain at
will. This was the most successful test and the one with longest and most controlled flights. The
next photo shows the stand-up position.

Figure 120: Third test

78
Pol Bernad Serra
Hydrofoil Design and Construction

6. Future work

There are three tasks to do as an extension for this project:

• Validate the Simulink platform: to do so, more cases should be studied and compared
with the tests results. However, the tests should also allow data recovery to not being
only qualitative.
• Use the hydrofoil to surf waves: for this future task, a professional surfer would be
needed. The tests have shown that the board is stable enough to stand on and it should
be possible to ride waves with the appropriated skills. From this point, a geometry
modification may be introduced in order to obtain the appropriated balance between
stability and manoeuvrability.

• Construction techniques: It must be noted that prototyping tools were used for the
manufacturing of the hydrofoil. A future improvement could be the mass production to
reduce the manufacturing costs and time.

7. Timing

The schedule for this project has been followed and the work did not stop from the first day.
The prove of it is that the proposed day for the first scale test in the Gantt diagram of the project
charter was on 1st of April and the first test was developed on 11th of the same month, just 10
days later. Something similar happened with the first real scale test, the assigned day was on
13th of May and the actual test was done on 19th of the same month, just one week late. The
following image is the Gantt diagram posted on the project charter:

79
Pol Bernad Serra
Hydrofoil Design and Construction

Figure 121: Project charter’s Gantt diagram

80
Pol Bernad Serra
Hydrofoil Design and Construction

8. Conclusions

This project has recreated the typical industry challenge. There was a list of objectives that need
to be accomplished within a deadline, a budget and the tools and machines available. The whole
process has passed throw a design stage, tests to validate the geometry designed and the
software implemented and a final real prototype.
The design section has shown that the principal problem of a hydrofoil surf is its centre of mass
height, which force the needing of a very stable plane. The static stability was achieved by
iterative design using the Open VSP and the dynamic stability was evaluated with an own
designed Simulink platform. The results from the platform were similar to the obtained on the
tests, however, the software predicted a higher flight speed than the actually needed.
In the scale model section, two geometries were 3d printed and tested in the test bench. Overall,
the conventional design was the one with better performance. This fits with the actual
technology used on hydrofoils since all of them have the same distribution of lifting surfaces.
The results obtained from the tests match with the ones obtained via Simulink platform which
means the platform could be used for future design iterations.
Finally, in the real scale model construction, the structure was designed and constructed using
composite materials and currently state of the art technology. The pulling tests have shown that
the geometry is stable and a beginner was able to ride the board without major problems; thus
accomplishing the most important of the objectives.
Overall, this project has presented a list of challenges which have all been overcome to end up
with a functional hydrofoil that could be used to ride waves and all of this was done in a regular
house, for a non-professional student and with own constructed machines.

81
Pol Bernad Serra
Hydrofoil Design and Construction

9. Bibliography

Acosta, A. J. (1973). Hydrofoils and Hydrofoil Craft. Pasadena, California: California Institude of
Technology. Retrieved from https://authors.library.caltech.edu/316/1/ACOarfm73.pdf
Airfoil Tools. (2019, March). Retrieved from http://airfoiltools.com/
Arduino Forum. (2019, April). Retrieved from
https://forum.arduino.cc/index.php?topic=579183.0
Azo Materials. (2019, Mayo). Retrieved from
https://www.azom.com/article.aspx?ArticleID=8211
B3 Proshop. (2019, Febrero). Retrieved from https://www.b3proshop.com/es/
Basic Aircraft Design Rules. (2019, March). Retrieved from
https://ocw.mit.edu/courses/aeronautics-and-astronautics/16-01-unified-engineering-
i-ii-iii-iv-fall-2005-spring-2006/systems-labs-06/spl8.pdf
CES Edupack. (2019, Mayo). Retrieved from
https://grantadesign.com/teachingresource/material-science/
Eisenberg, P. (1968). Cavitation. Massachusets: MIT. Retrieved from
http://web.mit.edu/hml/ncfmf/16CAV.pdf
Elahi, A. F., Hossain, M. M., Afrin, S., & Khan, M. A. (2014). Study on the Mechanical Properties
of Glass Fiber Reinforced Polyester Composites. International COnference on
Mechanical, Industrial and Energy Engineering, (p. 5). Khuina, Bangladesh.
Fablab Speed and Feeds Calculator. (2019, Abril). Retrieved from
https://pub.pages.cba.mit.edu/feed_speeds/
Franchini, S., & López, O. (2011). Introducción a la Ingeniería Aeroespacial. Madrid: Garceta
Grupo Editorial.
Gómez, M. A., & Pérez, M. (2012). Mecánica de Vuelo. Madrid: Garceta Grupo Editorial.
Human Benchmark. (2019, Marzo). Retrieved from
https://www.humanbenchmark.com/tests/reactiontime
Nazza. (2019, April). Retrieved from
https://www.nazza.es/img/cms/documentos%20PDF/Fichas%20T%C3%A9cnicas/FT-
resina-de-poliester-isoftalica_1.pdf
Online Wing Bending Inertia Calculator. (2019, Mayo). Retrieved from
http://www.wingbike.nl/Wingbike_Hydrofoil/Airfoil_bending_inertia.html
Strian. (2019, Mayo). Retrieved from http://structural-analyser.com/
Wikipedia Bending. (2019, Mayo). Retrieved from https://en.wikipedia.org/wiki/Bending
Wikipedia Second Moments of Inertia. (2019, Mayo). Retrieved from
https://es.wikipedia.org/wiki/Anexo:Momentos_de_inercia_de_%C3%A1reas

82
Pol Bernad Serra
Hydrofoil Design and Construction

83