Adaptability of a frequency changing

harmonically excited pendulum for wave

energy harvesting

EG4013 Individual Project in Engineering

Author: Manuel Beltrá Soler

Student Number: 51877053

Supervisor: Dr Vahid Vaziri

Word Count: 8837

A dissertation submitted in partial fulfilment for the requirements of the award of Masters
degree of Mechanical and Electrical Engineering
at the University of Aberdeen

May 2022
Abstract

As technology advances, so does the demand for electrical power, and the development and
investigation of new renewable sources have been of great interest in the past few years.
This thesis has been motivated by the idea of obtaining renewable energy from wave excita-
tion via a parametric pendulum. This innovative system uses waves kinetic energy to generate
mechanical energy in the pendulum system, where the energy will be extracted from. Different
pendulum configurations have been considered, including pendulum excited along a tilted axis
and a pendulum whose pivot point is inclined by an angle, therefore having less gravity element
to overcome and making rotation easier.
Initially, a preliminary study on the dynamics of both pendulum systems has been conducted
via Matlab simulations for wave values on a chosen sea location, obtaining different pendulum
responses for different initial conditions. Since rotational response of the pendulum systems
is characterised with significantly higher kinetic energy and higher energy harvesting potential
than systems with an oscillatory response, a control method has been developed for initiating
and maintaining the desired period-1 rotational response.
Finally, the controlled pendulum systems have been simulated in order to find the pendulum
system least influenced by wave changes and more optimal for energy harvesting considera-
tions.

i
Acknowledgements

I would like to express my gratitude to Dr Vahid Vaziri for his help throughout this project. He
took personal interest in the project and helped me, not only with technical tasks but also with
his guidance. It help me go all through the project in a straight forward path without making
me loose any time in irrelevant matters. I would also like to thank him for the time he took to
answer my emails and for all the meetings, I couldn’t have asked for a better advisor.

ii
Contents

Nomenclature vi

1 Introduction 1
1.1 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2 Aims and Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 Literature Review 5
2.1 Renewable Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
2.1.1 Wave energy harvesting . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.1.2 Wave energy harvesting limitations . . . . . . . . . . . . . . . . . . . 10
2.2 Dynamics of pendulum systems . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2.1 Vertically and horizontally excited pendulum system . . . . . . . . . . 11
2.2.2 Low gravity force pendulum system . . . . . . . . . . . . . . . . . . . 12
2.2.3 Pendulum’s physical properties . . . . . . . . . . . . . . . . . . . . . 14
2.2.4 Types of wave excitation . . . . . . . . . . . . . . . . . . . . . . . . . 14

3 Dynamics of two pendulum systems 18

3.1 Systems modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
3.1.1 Different pendulum responses . . . . . . . . . . . . . . . . . . . . . . 19
3.2 Preliminary study for pendulum’s behaviour . . . . . . . . . . . . . . . . . . . 21
3.2.1 Case study: Saltire platform . . . . . . . . . . . . . . . . . . . . . . . 21

iii
Contents

4 Rotational control of pendulum systems 29

4.1 Time-delay feedback controller implementation . . . . . . . . . . . . . . . . . 29
4.1.1 Simulation results . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
4.2 Parametric study with time-delay feedback controller . . . . . . . . . . . . . . 31
4.2.1 Parametric study on the effect on wave amplitude . . . . . . . . . . . . 31
4.2.2 Parametric study on the effect on wave frequency . . . . . . . . . . . . 33
4.2.3 Parametric study on the effect on damping . . . . . . . . . . . . . . . . 34

5 Discussion 36

6 Conclusions and recommendations for future work 38

6.1 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

References 39

Appendices 42

A X-Y pendulum Matlab code 43

B Low gravity pendulum Matlab code 45

C X-Y controlled pendulum Matlab code 47

D Low gravity controlled pendulum Matlab code 50

E Risk assessment 53

iv
List of Figures

1.1 Two pendulum systems schematics . . . . . . . . . . . . . . . . . . . . . . . . 4

2.1 ”OT” schematics [3] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.2 ”OWC” schematic [3] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3 Oscillating bodies mechanical systems . . . . . . . . . . . . . . . . . . . . . . 10
2.4 ”X-Y” pendulum structure and displacement schematic [14] . . . . . . . . . . 12
2.5 Low Gravity pendulum schematic . . . . . . . . . . . . . . . . . . . . . . . . 13
2.6 Two shaker systems schematics . . . . . . . . . . . . . . . . . . . . . . . . . . 17

3.1 Two pendulum systems schematics . . . . . . . . . . . . . . . . . . . . . . . 18

3.2 Oscillatory behaviour graphical representation . . . . . . . . . . . . . . . . . 20
3.3 Rotational behaviour graphical representation . . . . . . . . . . . . . . . . . . 21
3.4 Set 1 two pendulum’s simulation for initial conditions θ̇ = 0 and θ = 0 . . . . . 23
3.5 Set 2 two pendulum’s simulation for initial conditions θ̇ = 0 and θ = 0 . . . . . 24
3.6 Set 1 two pendulum’s simulation for initial conditions θ̇ = 0 and θ = 1 . . . . . 25
3.7 Set 2 two pendulum’s simulation for initial conditions θ̇ = 0 and θ = 1 . . . . . 25
3.8 Set 1 two pendulum’s simulation for initial conditions θ̇ = 1 and θ = 0 . . . . . 26
3.9 Set 2 two pendulum’s simulation for initial conditions θ̇ = 1 and θ = 0 . . . . . 26
3.10 Set 1 two pendulum’s simulation for initial conditions θ̇ = 1 and θ = 1 . . . . . 27
3.11 Set 2 two pendulum’s simulation for initial conditions θ̇ = 1 and θ = 1 . . . . . 27

4.1 Set 2 two controlled pendulums simulation . . . . . . . . . . . . . . . . . . . 30

v
 4.2 Minimum wave amplitude values . . . . . . . . . . . . . . . . . . . . . . . . 32
4.3 Maximum wave amplitude values . . . . . . . . . . . . . . . . . . . . . . . . 32
4.4 Minimum wave frequency values . . . . . . . . . . . . . . . . . . . . . . . . 33
4.5 Maximum wave frequency values . . . . . . . . . . . . . . . . . . . . . . . . 34
4.6 Maximum damping values . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

E.1 Risk assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

List of Tables

3.1 Saltire platform amplitude and frequency values [8] . . . . . . . . . . . . . . . 22

3.2 Low gravity pendulum’s physical properties for different wave parameters . . . 22
3.3 X-Y pendulum’s physical properties for different wave parameters . . . . . . . 23

4.1 Pendulum and controller parameters for rotational response . . . . . . . . . . . 31

4.2 Range of wave amplitudes for pendulum’s rotational behaviour . . . . . . . . . 32
4.3 Range of wave frequencies for pendulum’s rotational behaviour . . . . . . . . 33
4.4 Maximum damping values for pendulum’s rotational behaviour . . . . . . . . . 34

vi
Nomenclature

Constants

g Gravitational force constant, 9.81 m/s2

Variables

β Inclination angle, degree

θ̇ Angular velocity, rad/s

Ω Pendulum’s natural frequency, Hz

ω Excitation frequency, Hz

ω Wave excitation frequency, Hz

θ Rotation angle, rad

ε Energy extraction coefficient

A Wave amplitude, m

c Damping coefficient, %

f (t) Wave forcing element, N

l Pendulum’s arm length, m

m Pendulum’s bob mass, Kg

vii
Chapter 1

Introduction

Energy demand is exponentially increasing and is needed more than ever, due to the advance in
technology and the dependence that modern society has on it. Throughout history, energy has
been obtained by the burning of fossil fuels. Although this energy model cannot be raised any-
more due to the devastating environmental effect that it has. When fossil fuels are burned, they
release large quantities of carbon dioxide, a green house gas, that traps heat into the atmosphere,
having a vast contribution to climate change. For this reason alternative ways of producing en-
ergy in a eco-friendly way have been taken place and the research on these renewable energy
methods will not stop in a near future.
The law of conservation of energy states that ”energy can neither be created nor destroyed,
only converted from one form of energy to another”. This is the principle applied for renew-
able energies, which are obtained by converting energy from a wide range of natural factors to
electrical energy.
The ocean covers more than 70% of the earth surface, with approximately 361 million square
kilometers. Ocean is never still, this vibrations are commonly caused either from wind or tide,
commonly called waves. Ocean waves have an average height of 3 meters, this will give an
approximate mass displaced by waves of 541.5 ∗ 105 tonnes, carrying an incredible amount of
kinetic energy due to all the displaced mass. This kinetic energy can be converted into electrical
energy by different wave harvesting methods. Wave energy harvesting is a source of renewable

1
Chapter 1. Introduction

energy that has not been reviewed in depth and researched to find an optimal way to extract all
the energy that waves accumulate. This is why the natural energy source that has been chosen
to be reviewed for this project are waves.

1.1 Motivation

This study has been motivated by the idea of wave energy harvesting via an harmonically excited
pendulum.
In the past years, several ideas for extracting energy from sea waves have been developed.
All of them based in built structures that interact in different ways with the waves, structures on
which, the energy will be extracted from. These built components could be generally classified
in three types, oscillating water columns (columns that move when waves enter in a chamber),
over-topping (waves entering in a reservoir accumulating potential energy) and oscillating bod-
ies (like the harmonically excited pendulum that will be further reviewed in this thesis). The
most promising and innovative method out of this three is the oscillating bodies. Hence The
built structure that will be reviewed in this dissertation will be harmonically excited pendulum
systems.
Waves have set initial conditions and parameters that vary through time, such as frequency,
amplitude, wavelength and speed. Depending on this initial conditions and parameters, the pen-
dulum can respond with movement of the pendulum’s rod, from where energy can be extracted
from, via a collector placed in the pivot point of this rod. There are several possible scenarios
for the pendulum’s motion. The pendulum system can respond either with oscillations, with a
chaotic response or a rotational motion depending on the pendulum’s physical characteristics
and wave parameters. Previous studies have emphasized in the oscillation response of the pen-
dulum, however the rotational motion will provide higher kinetic energy and hence will have
the potential to deliver higher electrical energy than an oscillatory behaviour. This is why this
thesis will be focused on the rotational response of different pendulum systems.

2
Chapter 1. Introduction

1.2 Aims and Objectives

This dissertation project is aimed to obtain green energy at a constant rate for an indefinite
period of time via wave harvesting by an harmonically excited pendulum, finding an optimal
way of making this energy extraction without needing human intervention.
This study will help transforming the kinetic energy of sea waves into mechanical energy
via the rotational motion of the pendulum system, and then convert the mechanical energy into
electrical in the most optimal way possible. Contribute in the narrowing of the gap between
supply and demand of energy, and help with the transition from the burning of fossil fuels to a
sustainable green energy consumption.
For the pendulum system to be autonomous, the pendulum will need to be able to readjust
its physical characteristics in order to adapt for different wave parameters.
To achieve this objectives, two pendulum systems will be reviewed. The first pendulum
that will be introduced will be tilted in its axis, getting pendulum’s motion in the x-y plane as
shown in Figure 1.1a. The second system to be reviewed is more complex in its dynamics,
this pendulum system is tilted in its base, getting a motion in the x-y-z plane. It is also called
pendulum with low gravity force since its inclination makes lower the gravity force to overcome
for a full rotation. The side view is shown in Figure 1.1b. This two pendulum system dynamics
will be simulated and compared to finally obtain the pendulum system that will harvest energy
in the most optimal manner.
The pendulum’s behaviour will be study form simulations. For all the simulations, the
MatLab software, with the ODE 45 numerical function will be used. With the aim of obtaining
as results the range of wave parameters and pendulum’s physical parameters that will give a
rotational response of the pendulum systems and to find which pendulum system will be more
optimal for energy extraction purposes.
The results of the simulations will be critically analysed to choose the pendulum system
that will have a better balance between the width of range of wave parameters that will give a
rotational response and the amount of energy that could be extracted from them.

3
Chapter 1. Introduction

(a) Schematic of rotational (b) Schematic of low gravity force pendulum

x-y motion pendulum system system [20]
[14]

Figure 1.1: Two pendulum systems schematics

4
Chapter 2

Literature Review

In this chapter, different methods of wave energy harvesting will be reviewed, focusing on an
innovative method based on a parametric pendulum. Whose equations of motion, excitation and
physical characteristics will be studied for two different pendulum systems.

2.1 Renewable Energy

Worlds energy consumption was of 13864.9 Mtoe (5.805 ∗ 1014 MJ) in 2018, an increase of
18.45% from 2008. Following with an increase of the 4.6% from 2019 to 2021. The sources
of energy that supply this energy demand are oil (33.3%), coal (27.2%), natural gas (23.9%),
renewable energies (10.8%) and nuclear (4.4%) [16]. Renewable electricity generation jumped
to (35%) in 2021, up from a (27%) in 2019, the fastest year-on-year growth since the 1970’s
(8% increase from 2020) [9]. The main producers of renewable energy in 2017 are hydro-
power (65.87%), which mainly corresponds to dams, wind (17.97%), solar (6.93%), bio-energy
(7.85%), geothermal (1.36%) and marine (0.02%). Even tho the biggest contribution for renew-
able energy is hydro-power with a total energy in 2017 of 4.1 billion TWh, wind is the one with
the most significant increase in renewable energy production, growing by a rate of (17%/year)
due to the continuous installation of new plants [17].
Marine energy harvesting is a renewable and carbon-free way of extracting energy, it has the

5
Chapter 2. Literature Review

smallest contribution in the production of renewable energy, the (0.02%) mainly corresponds to
tidal energy generation [17]. Because of the energy accumulated by the mass of water moving
due to tides and ocean waves, it has the potential to contribute significantly to future energy
supply. It could end up being the most significant contribution source of renewable energy
in the UK, with a predicted potential of 80 TWh/year, most of it coming from tidal power
barrages and wave energy [10]. Tidal energy harvesting is the most developed method from
marine energy harvesting up to date, having the biggest energy contribution. Although, this
doesn’t mean that tidal energy has more potential than wave energy, moreover it is expected
that from the 80 TWh/year predicted from marine energy in the UK, 30 TWh/year will be
tidal and 50 TWh/year wave based [10]. If we have a look into the predicted global marine
energy production, the results don’t differ on which source has the biggest energy harvesting
potential. According to the International Energy Agency, the resource with largest potential
is wave power, which may offer up to 80,000 TWh of energy per year. The second greatest
source is ocean thermal energy technology (OTEC), which has a potential of 10,000 TWh/year,
followed by salinity gradient, which has a potential of 2000 TWh/year, and finally tidal current,
which has a potential of 800 TWh/year [3].

2.1.1 Wave energy harvesting

Wave energy harvesting is an upcoming way of energy extraction with a vast potential as previ-
ously mentioned. The idea of harvesting energy from waves has delivered a mixture of mechan-
ical devices that transform the wave motion into electricity. This mechanical devices have one
thing in common. They transform wave energy into mechanical energy and then it’s converted
to electrical energy. In spite of the fact that they have a common thing, their working principles
of this devices change widely between them, making difficult the classification of this devices.
Although we can generally classify their operations into three groups.

6
Chapter 2. Literature Review

Over-topping (OT)

Over-topping devices are the only out of the three reviewed methods that is based on the extrac-
tion of the potential energy of waves. The working principle of this harvesting method is based
on pressure head of sea water, created by an arise of the sea level. This head can be achieved
with several methods, but their working principle is very similar. Building an off-shore reser-
voir above the sea water level on which only wave crest water will enter and captured, creating a
pressure head. The water in the reservoir will finally flow back to the sea through turbine, from
where energy will be extracted [3]. The over-topping schematic is shown in Fig. 2.1 below.

Figure 2.1: ”OT” schematics [3]

Oscillating water columns (OWC)

Oscillating water columns schematic is shown in Fig. 2.2. It’s the most common mechanical
device for wave energy harvesting out of the three mentioned in this section. This method is
based on a chamber connected to the sea with an aperture below the sea level, from where waves
will enter and exit. This chamber is connected to an outlet to the atmosphere with a turbine. As
waves enter and exit the chamber, the water level on it will rise and fall, creating a pressure force
that will make the air inside the chamber to enter and exit. The air will flow through the turbine
placed in the outlet, converting the turbine rotation into electrical energy. Since air flows in and
out the turbine, a special type of turbine is needed. A turbine that rotates in just one direction
without mattering the flow direction of the air. This type of turbine is called Wells turbine [3].

7
Chapter 2. Literature Review

Figure 2.2: ”OWC” schematic [3]

Oscillating bodies (OB)

There are numerous oscillating bodies methods for energy harvesting. Most of them have a
communal potential, mainly based on the wave height and period. Three of the most common
methods will be reviewed in this section with the addition of one method that is still under
development but has a promising future projection.
The first method introduced is the heave wave converter, shown in Fig. 2.3a. This method
has a very simple working principle, just needing vertical excitation, hence, the only wave
parameter that will affect the energy extraction of this method will be the wave height. The
structure of this system is based on a buoy fixed to the seabed. Between the buoy and the
seabed, an hydraulic system formed by a piston, an accumulator and an hydraulic motor will
be attached. With the vertical wave excitation pressure will be created inside the piston passing
then to the accumulator, to the hydraulic motor and then to a rotatory generator from where the
electrical energy will be extracted from [7].
The next method is an oscillatory wave surge converter, shown in Fig. 2.3b. It has similar
working principle than the previous mentioned, but instead of vertical motion, it will work with
horizontal excitation, hence, the wave parameter that will affect the energy extraction will be
the wave period. This system can have several models for its construction. The moving part,
because of its interaction with waves it can either be a buoy as show in Figure 2.3b or a paddle.

8
Chapter 2. Literature Review

This moving part will be attached to the seabed with a frame with a pivot on top and bottom
allowing the horizontal displacement. The frame will have an hydraulic piston attached to it,
connected to an engine and then to a generator on which energy will be extracted from.
The third mechanism of oscillating bodies for wave energy harvesting is a hinged point
absorber shown in Fig. 2.3c. It consist in a floating device with several buoyancy parts hinged to
each other, each part of this device will move relative to the other. Energy can be extracted from
this junctions by using the motion to pump water, this water pumped will generate hydrostatic
pressure, that with a linear generator, energy can be extracted [3].
The less developed method but with a very promising future projection, is a parametric pen-
dulum excited by waves. The goal of this method is to convert the kinetic energy of waves into
mechanical energy by achieving a rotatory motion of the pendulum. The parametric pendulum
can exhibit a rotational motion in a bounded range of conditions. This range depend in the
pendulum set-up properties and the wave parameters. A lot of different pendulum models have
been developed in the past years contributing in the evolution of this energy extraction method,
in order find the most efficient way of harvesting energy. Figs 1.1a and 1.1b show two differ-
ent models of parametric pendulum for wave energy harvesting. These models have a more
complex dynamics than a simple parametric pendulum. For that same reason they offer a more
efficient way of energy extraction, as well as allowing more room to change in the pendulum’s
set-up parameters to give a rotational response for a wider range of wave frequencies and am-
plitudes. The working principle for a wave excited parametric pendulum is shown in Fig. 2.3d.
Fig. 2.3d show the schematics of the ”low gravity” pendulum.

9
Chapter 2. Literature Review

(a) Heaving wave converter [7] (b) Oscillatory wave surge converter [7]

(c) Hinged point absorber [7] (d) Working principle of the parametric pendulum
[20]

Figure 2.3: Oscillating bodies mechanical systems

2.1.2 Wave energy harvesting limitations

The overall amount of wave energy available on the planet is vast. Although only a tiny part
of total energy is possible to be retrieved because of several limitations that the ocean impose.
These limitations are the following: environmental considerations including shipping and other
limitations on the utilization of the sea, as well as technical factors like the capture device ef-
ficiency and mechanical and electrical losses that the system could have, and finally location,
many hundreds of kilometers of shoreline are efficient for energy harvesting purposes but iso-
lated from settlement, where the energy would need to be redirected. Many other locations have
a too hostile environment making the energy extraction more difficult and unsafe [10].

10
Chapter 2. Literature Review

2.2 Dynamics of pendulum systems

Parametric pendulum excitation has catched the attention of many researchers in this recent
years because of its working principle and wide range of engineering applications that can
offer. Parametric pendulum, can be thought as a mechanism that transform one form of motion
such as linear, into another form, like rotational motion [2].
Since one of the most common ways of generating electricity is by rotation, the parametric
pendulum systems were considered as a promising energy harvesting method.
The dynamics of parametric pendulum systems will differ depending on their functionality
and their set-up. The focus of this review is on the detection and investigation of pendulum’s
rotational behaviour, as this will be the what the future work will be based on. The first section
will review the working principle and equations of motion of a vertically and horizontally wave
excited pendulum. The second part will go over the operation of a low gravity force pendulum
excited by waves, giving its respective equations of motion. Since the pendulums reviewed will
be set for simulation, in the last part different excitation waves models will be as well reviewed.

2.2.1 Vertically and horizontally excited pendulum system

The vertically and horizontally excited parametric pendulum schematics is shown in Fig. 1.1a.
This system is based on a pendulum that achieves rotational motion by wave excitation. Pen-
dulum’s arm is attached to a pivot point shifted at an arbitrary angle (α). This pivot point will
be excited by waves, due to its angle of orientation, it will have a horizontal and vertical dis-
placement, this x and y motion will be transferred to the pendulum, resulting in higher values
of torque for the same amount of energy than a just vertically excited pendulum system. The
working principle of this system it’s shown in Fig. 2.4 [12].
The most straight forward way of obtaining the governing equation of this pendulum’s sys-
tem is by using the Lagrange’s or Hamilton’s equations [13]. The equation of motion describing
the dynamic response of this system is given by [14]:

11
Chapter 2. Literature Review

mlθ̈ + mẌcosθ + m(Ÿ + g)sinθ + clθ̇ = 0, (2.1)

Where m is the pendulum’s bob mass, l the arm length, θ the arm’s displacement angle, where
(”) and (’) denote the derivatives with respect to time, g is the gravitational force, c the damping
coefficient and X and Y arbitrary functions showing the horizontal and vertical displacement. X
and Y values will be given by the kinematic relationships imposed by the pendulums set-up ge-
ometry, in other words, they will be given by the room for vertical and horizontal displacement
that the pendulum’s structure has, depending on the wave excitation force (f(t)) and amplitude.
An example of the whole schematics of the ”X-Y” pendulum system is shown in Fig. 2.4.

Figure 2.4: ”X-Y” pendulum structure and displacement schematic [14]

2.2.2 Low gravity force pendulum system

A low gravity pendulum is a pendulum system on which the influence of gravity is reduced. The
reduction of gravity is achieved by shifting by an angle the common up and down rotations, ob-
taining a tilted rotation in the x-y-z plane that doesn’t need to cover the whole g for completing
a full rotation. An example of this low gravity rotation is shown in Fig. 2.5. The shifted angle

12
Chapter 2. Literature Review

is given by the angle between the water surface and the pendulum’s plane of rotation.
The working principle of the low gravity pendulum is shown in Fig. 2.3d, the buoy that acts
as the pendulum’s base moves up and down due to the excitation of waves. The buoy has a
sliding platform with a pendulum system, so the buoy’s motion will provide a parametric exci-
tation to the pivot point of the pendulum, where energy can be extracted from. As it’s shown in
Fig. 1.1b, the exciting force needed to overcome gravity, hence, the force for achieving rotation
will be mgsin(β). The mechanical energy of the pendulum will be dependant on the mass, the
square length of the pendulum’s arm and the inclination angle, having a large mechanical en-
ergy will help the pendulum system to overcome small fluctuations and maintain the rotational
behaviour desired.
The equation of motion describing the dynamic response of this system is given by [20]:

¨
ml2 θ̈ + cθ̇ + mglsinβsinθ = mlf (t)sinθ. (2.2)

Where β, the inclination angle, will be controlled depending on the moment of inertia of the
pendulum. Because the impact of viscous damping on the instability of a parametrically excited
system has been well studied in prior studies, it’s not a major issue here. This value will be held
constant throughout all simulations at c = 0.1.

Figure 2.5: Low Gravity pendulum schematic

13
Chapter 2. Literature Review

2.2.3 Pendulum’s physical properties

Previous researches have reached to the conclusion that the parametric resonance of the pen-
dulum’s system occurs when the wave’s excitation frequency is twice the pendulum’s natural
frequency [19]. Looking at the formulas given below 2.3, will give us an idea of what are the
physical properties that the pendulum should have in order to obtain a parametric resonance
response.

r
g
2Ω = ω; Ω = , (2.3)
l
Where Ω is the pendulum’s natural frequency and ω the wave’s excitation frequency.

2.2.4 Types of wave excitation

For simulation purposes, the wave excitation forcing function can be considered to be harmon-
ically (making all analysis and simulation process simpler), or to be a realistic wave profile,
giving a much more complex analytical and simulation process. In this section both wave ex-
citation forcing types will be introduced. Harmonic wave forcing will be considered for this
dissertation since it’s comparable to deep water.

Realistic wave excitation

There are numerous methods of obtaining a wave excitation function. The measurements statis-
tics of ocean waves have shown that waves have statistical properties that allow us to determine
a realistic wave excitation function with a reasonable accuracy [6].
The wave spectrum of Pierson-Moskowitz, is a probability density function of wave fre-
quencies and directions when the sea surface is considered as an infinite aggregate of waves
with infinitely tiny amplitude. Other observation is that the heights of ocean waves follow a
Rayleigh distribution, this distribution is shown below [5]:

14
Chapter 2. Literature Review

p
σ = Hs /2; Hs = 4 var(η), (2.4)

Where Hs is the significant height, σ is the distribution parameter and η is the surface elevation.
The linear (or small-amplitude) wave theory by [1] shown in equation 2.5, is the foundation
for ocean wave simulation.

Acos(κx0 − ωt + ϕ), (2.5)

Where κ is the wave number, x0 is the position at rest, ω the angular velocity and ϕ is the
uniformly distributed random phase term.
It’s shown in [18] that uniformly distributed random phase terms ϕ of equation 2.5 should
be changed with random amplitudes values. The expression 2.5 then becomes:

q
RAsin(κx0 − ωt + Φ), R = r12 + r22 , (2.6)

Where r1 and r2 are random numbers from the standard normal distribution.
Eq. 2.7 and R from Eq.2.6 follows a Rayleigh distribution.

arctan( r1 )

if r2 ≥ 0
r2
Φ= (2.7)
arctan( r1 ) + π

if r2 < 0.
r2

As we can see in this section, the modelling of real wave profiles can be a complex task to
achieve and even more for simulation purposes. This is why other type of wave modelling is
going to be introduced, making the simulation of this dissertation smoother.

Harmonic wave forcing

Waves can be assumed to be harmonic, this approximation is more suitable for ocean waves far
from the coast line, since there are less reflection perturbations and they maintain a more peri-
odic behaviour. This harmonic behaviour is commonly represented by a sinusoidal waveform

15
Chapter 2. Literature Review

in the form of:

F (t) = Asin(ωt), (2.8)

Where A is the excitation amplitude and Ω the waves frequency.

For experimental porpoises, this harmonic excitation response can be easily obtained by a
simple electrical shaker, with excitation in the y axis as shown in Fig. 2.6a. Although the shaker
system shown in Fig. 2.6b is a more appropriate model for the pendulum systems reviewed in
Sections 2.2.1 and 2.2.2, since they are coincident in its tilted orientation along an angle α/β,
that can be varied in order to adapt to different wave excitation parameters.
The pendulum system shown in Fig. 1.1a, it’s fully represented with this shaker model (fig-
ure 2.6b). The inclination angle (α) that Fig. 1.1a has in its pivot point, it’s now in the shaker’s
base. If shakers parameters are proportionally set, it will result in the same excitation direction,
and amplitude. If harmonic force is applied to the X-Y system, the X and Y parameters of Eq.
2.1 will then become:

Y = Asin(α), X = Acos(α). (2.9)

For the low gravity pendulum system reviewed in Section 2.2.2 to be characterized by the
shaker shown in Fig. 2.6b, few changes will need to be done. The simple pendulum system on
top of the shaker will need to be replaced by a low gravity pendulum system, as the one shown
in Fig. 1.1b. Here the inclination angle β will be set changing the shaker’s parameter α. The
forcing component ”f(t)” shown in Eq. 2.2 will then become:

f (t) = A∗ sin(ωt), A∗ = A/sinβ, (2.10)

Where A is the excitation amplitude, A∗ is the pendulum’s pivot excitation and β correspond to
shakers inclination angle α.

16
Chapter 2. Literature Review

(a) vertical simple shaker [11] (b) X-Y shaker [12]

Figure 2.6: Two shaker systems schematics

17
Chapter 3

Dynamics of two pendulum systems

In this chapter a preliminary study of the two pendulum systems will be carried. The study will
be carried in a certain location, which wave parameters from that location will be used in order
to analyse the pendulums behaviour.

3.1 Systems modelling

In this section the behaviour of the two pendulum systems presented in Sections 2.2.1 and 2.2.2
will be simulated.

(a) x-y pendulum schematic (b) Low gravity pendulum schematic from Figure 2.5

Figure 3.1: Two pendulum systems schematics

18
Chapter 3. Dynamics of two pendulum systems

Fig. 3.1 shows the two pendulum schematics. Both systems have been modelled with the
similar physical parameters, making the comparison between both systems easier.
As reviewed in Sections 2.2.1 and 2.2.2, each of this pendulum systems has an equation of
motion (Eqs 3.1 and 3.2) that will describe the behaviour of the system.
For the X-Y pendulum shown in Fig. 3.1a, the equation of motion is as follows:

mlθ̈ + mẌcosθ + m(Ÿ + g)sinθ + clθ̇ = 0, (3.1)

For the low gravity pendulum shown in Figure 3.1b, the equation of motion will be as follows:

¨
ml2 θ̈ + cθ̇ + mglsinβsinθ = mlf (t)sinθ. (3.2)

Some parameters of the models have been chosen to be changed over time due to their
practical ease to be modified without needing much intervention. The goal of this changing
parameters is to convert the physical composition of the systems in order to achieve the desired
pendulum behaviour. These parameters are L (pendulum’s rod length), parameter reviewed in
Section 2.2.3, it’s directly related with the wave frequency to obtain the rotational behaviour as
shown in 2.3. The parameter α / β are the pendulum’s pivot/base displacement angle respec-
tively, this parameter will affect the system’s excitation amplitude, as shown in Section 2.2.4
Eq. 2.10.
Other parameters like m (mass of the bob) and c (viscous damping) won’t be changed in
this thesis. This parameters will be a constant in time and will be chosen from previous studies
in order to optimize the system’s rotational response depending on different types of wave
parameters at any certain location.

3.1.1 Different pendulum responses

The behaviour of the pendulum system is going to be studied in the MatLab simulation. Phase-
plane and time-history graphs are used to analyse the type of behavior the system is facing. The
different behaviours can be classified in three groups. Pendulum system not moving, therefore,

19
Chapter 3. Dynamics of two pendulum systems

the system will be facing a stationary response. Oscillatory response of the system, meaning
that the pendulum will go back and forward in an angle range of −πrad < θ < πrad. Rota-
tional response, where the pendulum system will be achieving full rotations, this means that the
pendulum is rotating in the same direction in an angle range of −πrad ≤ θ ≤ πrad.

Oscillatory response example

In this example, the system behaviour is represented graphically in Fig. 3.2 for the x-y pen-
dulum under the initial conditions θ̇ = 0 and θ = 1. From the figure below we can see that the
pendulum is oscillating in the range of −0.5rad < θ < 0.5rad. An harmonic oscillation in the
time history graph will give an spherical graphical solution for the phase-plane graph since the
pendulum will be going back and forward in the same angle range.

(a) Phase-plane graph (b) Time history graph

Figure 3.2: Oscillatory behaviour graphical representation

Rotational response example

An example for rotational response of a pendulum system is shown in Figure 3.3 for the low
gravity pendulum under the initial conditions θ̇ = 1 and θ = 1. From this figures below it can be
seen that pendulum’s angle is in the range of −π ≤ θ ≤ π, facing a sudden change from π to
-π. Meaning that rotation is achieved since a full gyration of 2π is accomplished.

20
Chapter 3. Dynamics of two pendulum systems

(a) Phase-plane graph (b) Time history graph

Figure 3.3: Rotational behaviour graphical representation

3.2 Preliminary study for pendulum’s behaviour

A preliminary study of the behaviour for the two reviewed pendulum systems will be studied
in this section. As discussed in Section 2.2.4, deep water waves have an almost sinusoidal
behaviour. Since both pendulum systems will need a structure attached to the sea bed to work
successfully, offshore oil rigs about to be decommissioned are structures of interest. In order to
give them an other utility and make them useful again once their life span have come to an end.
An example of an offshore oil rig in deep waters and about to be decommissioned is the
Saltire platform. This is why the preliminary study will be done based on this location with its
corresponding wave scatter data.

3.2.1 Case study: Saltire platform

The Saltire platform was installed in 1992. Saltire is located on the northern margin of the Witch
Ground Graben in UK Central North Sea at the location point (1.900000, 57.20000). Offshore
oil rigs of this type have an average lifespan of 25 years, hence plans for decommissioning the
platform have already been made [15].
Wave parameters of this location are of interest for this project. Extreme values for this

21
Chapter 3. Dynamics of two pendulum systems

location have been taken into account, in order study of the pendulum’s behaviour for the widest
range of wave parameters in this location.
Stats for data-set: North Sea Wave, at point: (1.900000, 57.20000) between 1979-01-01 and
2015-12-31 gave the following values:

Wave set of parameters Amplitude (m) Frequency (Hz)

Set 1 6 0.05

Set 2 0.6 0.1667

Table 3.1: Saltire platform amplitude and frequency values [8]

Where set 1 of parameters corresponds to the maximum set of wave values and set 2 corre-
spond to the minimum set of wave parameters for Saltire platform location.

Simulation results

From the equations of motion for the two pendulum systems shown in Equations 2.1 and 2.2,
several simulations has been performed using a MatLab code. Wave parameters of Saltire plat-
form shown in Table 3.1 have been implemented in the code.
For this simulation, different parameters for the two pendulum systems have been chosen.
Table 3.2 shows the pendulum physical parameters for maximum wave values of the Saltire
location. Table 4.2 shows the pendulum’s physical properties for wave parameters in set 2.
Other parameters like the damping coefficient has been well studied in prior numerical studies,
therefore it has been kept constant throughout the simulation c=0.1 N/m [20].

Wave parameters L (m) β (rad) Mass (kg)

Set 1 20 0.5 20

Set 2 30 0.4 20

Table 3.2: Low gravity pendulum’s physical properties for different wave parameters

22
Chapter 3. Dynamics of two pendulum systems

Wave parameters L (m) β (rad) Mass (kg)

Set 1 45 0.4 20

Set 2 40 0.6 20

Table 3.3: X-Y pendulum’s physical properties for different wave parameters

For the simulations, the initial values for angular speed (θ̇) and displacement angle (θ) need
to be set. In this section a comparison of pendulum’s behaviour between different wave parame-
ters for different initial values will be performed. For set 1, the pendulum’s physical parameters
displayed in Table 3.2 will be used. For set 2, values in Table 4.2 will be used. This parameters
have been chosen in order to maximize pendulum’s performance.
The code was firstly ran with the initial values θ̇ = 0 and θ = 0 for the 2 different sets of
wave parameters. The obtained results are displayed in the following figures:

(a) Phase-plane graph x-y (b) Time history graph x-y

Figure 3.4: Set 1 two pendulum’s simulation for initial conditions θ̇ = 0 and θ = 0

23
Chapter 3. Dynamics of two pendulum systems

(a) Phase-plane graph x-y (b) Time history graph x-y

Figure 3.5: Set 2 two pendulum’s simulation for initial conditions θ̇ = 0 and θ = 0

For the set 1 and 2 of wave parameters, the pendulum systems will behave in a similar man-
ner. The low gravity pendulum will maintain a stationary behaviour in both set of parameters
because the forcing is in the pendulum direction as shown in Fig. 3.1b, so it won’t produce any
torque, while in the x-y pendulum system the forcing is on the pivot direction as shown in Fig.
3.1a so torque is produced and will show a non-harmonic oscillatory behaviour.
Secondly the simulation was ran for the next initial values θ̇ = 0 and θ = 1 for the 2 different
sets of wave parameters. The obtained results are displayed in the following figures:

24
Chapter 3. Dynamics of two pendulum systems

(a) Phase-plane graph x-y (b) Time history graph x-y

Figure 3.6: Set 1 two pendulum’s simulation for initial conditions θ̇ = 0 and θ = 1

(a) Phase-plane graph x-y (b) Time history graph x-y

Figure 3.7: Set 2 two pendulum’s simulation for initial conditions θ̇ = 0 and θ = 1

The second simulation with initial conditions θ̇ = 0 and θ = 1, proved that the pendulum
systems will behave in a similar manner for the different set of wave values as shown in Figures
3.6 and 3.7. The x-y pendulum and the low gravity systems will show an oscillatory behaviour
with different periods along time making it a non harmonic response.
At this point, the code was ran with initial values θ̇ = 1 and θ = 0 for the 2 different sets of

25
Chapter 3. Dynamics of two pendulum systems

wave parameters. The obtained results are displayed in the following Figures:

(a) Phase-plane graph x-y (b) Time history graph x-y

Figure 3.8: Set 1 two pendulum’s simulation for initial conditions θ̇ = 1 and θ = 0

(a) Phase-plane graph x-y (b) Time history graph x-y

Figure 3.9: Set 2 two pendulum’s simulation for initial conditions θ̇ = 1 and θ = 0

The third simulation results with initial conditions θ̇ = 1 and θ = 0, are displayed in Figs 3.8
and 3.9. For the x-y pendulum simulation, it will switch between oscillations and rotations. For
the set 1, the low gravity pendulum will show an oscillatory behaviour, while for the set 2, the
low gravity pendulum shows a rotational behaviour as desired. In Fig. 3.9a can be seen that

26
Chapter 3. Dynamics of two pendulum systems

the lines in the graph are in different high levels, this is because the angular speed is different
throughout time.
Lastly the simulation with initial values of θ̇ = 1 and θ = 1 was ran for the 2 different sets of
wave parameters. The obtained results are displayed in the following Figures:

(a) Phase-plane graph x-y (b) Time history graph x-y

Figure 3.10: Set 1 two pendulum’s simulation for initial conditions θ̇ = 1 and θ = 1

(a) Phase-plane graph x-y (b) Time history graph x-y

Figure 3.11: Set 2 two pendulum’s simulation for initial conditions θ̇ = 1 and θ = 1

The fourth simulation results with initial conditions θ̇ = 1 and θ = 1 are shown in Figs

27
Chapter 3. Dynamics of two pendulum systems

3.10 and 3.11 which describes the pendulum’s behaviour for the different set of wave values.
For both sets of wave parameters the x-y pendulum system switches between oscillations and
rotations. While for the low gravity pendulum will remain in a rotational behaviour.
Since a rotational behaviour will maximize energy extraction from the pendulum systems
due to an increase in potential and kinetic energy, it becomes of interest the specific values from
the previous simulation that will make the pendulum rotate. From the simulations results it can
be seen that the initial conditions θ̇ = 1 and θ = 1 are the values that gives a better rotational
solution.
Without an external intervention it’s impossible to set this desired initial conditions for the
pendulum. This is why the implementation of a controller that will help the system to maintain
a rotational behaviour will be reviewed in Section 4.

28
Chapter 4

Rotational control of pendulum systems

In this chapter, an implementation of a time-delay feedback controller for both pendulum sys-
tems will be reviewed, simulated and studied for energy extraction purposes [4].

4.1 Time-delay feedback controller implementation

The time-delay feedback controller is a controller that feeds back to the system an error, the error
is obtained by subtracting the current results to the one period before. This error is diminished
with the two controller variables k1 and k2 will multiply the velocity and displacement error by
a fixed value to feed again to the system an harmonically stable behaviour.
The time-delay feedback controller will then change the equations of motions for both sys-
tems reviewed in Eqs 3.1 and 3.2 to the ones shown in Eqs 4.1 and 4.2.
X-Y pendulum system with a time-delayed feedback controller:

mlθ̈ + mẌcosθ + m(Ÿ + g)sinθ + clθ̇ = uu. (4.1)

Low gravity pendulum system with a time-delay feedback controller:

¨
ml2 θ̈ + cθ̇ + mglsinβsinθ = mlf (t)sinθ + uu, (4.2)

29
Chapter 4. Rotational control of pendulum systems

where the controller parameter uu, has on it the two controller variables k1 and k2 that by
multiplying the error of θ and θ̇ by a fixed value will make the system behave as desired.

4.1.1 Simulation results

The time-delay feedback controller is implemented in the MatLab code and simulations were
ran for the different sets of wave parameters.
Simulations were ran for set 2 of wave parameters for the x-y and low gravity pendulums.
Obtaining the following results:

(a) X-Y controlled pendulum (b) Low gravity controlled pendulum

Figure 4.1: Set 2 two controlled pendulums simulation

Fig. 4.6 shows the behaviour of the two pendulum systems reviewed in this thesis with the
implementation of a controller.
The x-y pendulum behaviour it’s shown in Fig. 4.6a. It can be seen that the system is
initiated with non harmonic oscillations, then at 90s, non-harmonic rotations start to happen,
and finally, after 160s the system stabilizes with a period-1 harmonic rotational response as
desired, making the errors and the control signal go to zero.
The low gravity pendulum behaviour it’s shown in Fig. 4.6b. From this figure it can be seen

30
Chapter 4. Rotational control of pendulum systems

that the system initiates with non-harmonic rotations, this non-harmonic rotations remain until
100s has past. After 100s the system stabilizes and a period-1 harmonic rotational response is
shown. Since it’s the desired response errors and the control signal go to zero.
The pendulum and controller parameters that help the systems maintain a rotational response
are the following:

Pendulum parameters Controller variables

Pendulum system
L (m) m (kg) α(rad) K1 K2
X-Y 25 5 0.5 0.05 0.01
Low gravity 15 5 0.2 0.5 0.5

Table 4.1: Pendulum and controller parameters for rotational response

4.2 Parametric study with time-delay feedback controller

In this section, a parametric study on the effect of wave amplitude, wave frequency and damp-
ing will be made for both pendulum systems maintaining the same parameters as reviewed in
Section 4.1 above. The results discussed in the following sections will be on the extreme values
of wave amplitude, frequency and damping that will maintain a rotational behaviour of both
systems.

4.2.1 Parametric study on the effect on wave amplitude

A parametric study on wave amplitudes was made to obtain the range of wave parameters that
will maintain an harmonic rotational behaviour of the pendulum systems.
The range of wave amplitudes on which the pendulum systems will maintain a rotational
response are shown in the following table:

31
Chapter 4. Rotational control of pendulum systems

Pendulum system Min. amplitude values (m) Max. amplitude values (m)
X-Y 0.001 1.2
Low gravity 0.001 0.8

Table 4.2: Range of wave amplitudes for pendulum’s rotational behaviour

The graphical representation of the Table 4.2 above it’s shown below:

(a) X-Y controlled pendulum (b) Low gravity controlled pendulum

Figure 4.2: Minimum wave amplitude values

(a) X-Y controlled pendulum (b) Low gravity controlled pendulum

Figure 4.3: Maximum wave amplitude values

This parametric study made, based on wave amplitude, helping us to analyse which pendu-
lum system is more susceptible to wave amplitude changes. Therefore needing more pendulum

32
Chapter 4. Rotational control of pendulum systems

physical changes and interventions to adapt to those amplitude changes.

4.2.2 Parametric study on the effect on wave frequency

A parametric study on wave frequencies was made to obtain the range of frequencies that will
maintain an harmonic rotational behaviour of the pendulums systems.
The range of wave frequencies which will maintain a rotational response of the pendulum
systems are shown in the following table:

Pendulum system Min. frequency values (Hz) Max. frequency values (Hz)
X-Y 0.155 0.236
Low gravity 0.14 0.4

Table 4.3: Range of wave frequencies for pendulum’s rotational behaviour

The graphical representation of the Table 4.3 above it’s shown below:

(a) X-Y controlled pendulum (b) Low gravity controlled pendulum

Figure 4.4: Minimum wave frequency values

33
Chapter 4. Rotational control of pendulum systems

(a) X-Y controlled pendulum (b) Low gravity controlled pendulum

Figure 4.5: Maximum wave frequency values

This parametric study made, based on wave frequency, help us analyse which pendulum
system is more susceptible to wave frequency changes. Therefore needing more pendulum
physical changes and interventions to adapt the system to those frequency changes.

4.2.3 Parametric study on the effect on damping

A parametric study on damping coefficient was made to obtain the maximum damping that will
maintain an harmonic rotational behaviour of the pendulums systems.
The maximum damping which will maintain a rotational response of the pendulum systems
is shown in the following table:

Pendulum system Max. damping values

X-Y 0.175
Low gravity 150

Table 4.4: Maximum damping values for pendulum’s rotational behaviour

The graphical representation of the Table 4.4 above it’s shown below:

34
Chapter 4. Rotational control of pendulum systems

(a) X-Y controlled pendulum (b) Low gravity controlled pendulum

Figure 4.6: Maximum damping values

This parametric study made based on the damping, could be correlated to energy extrac-
tion. Energy extraction for the pendulum systems is angular velocity dependent, and the energy
extraction coefficient act as a damper. The harder the damper is the more energy that can be
extracted from the system.

35
Chapter 5

Discussion

In this section, results from Sections 3.2, 4.2 and 4.1 will be analysed and discussed. Finishing
the section with a clear idea of which pendulum system will be less susceptible to changes and
more optimal for for energy harvesting.
Section 3.2, shows that the desired rotational response of the pendulum systems can’t be
achieved without an external intervention. From this results the idea of the implementation of a
time-delay feedback controller was brought up.
In Section 4.1.1, Table 4.1 shows the pendulum physical parameters that help the controlled
systems to maintain a period-1 harmonic rotational behaviour. The x-y pendulum system, hav-
ing a length of 25m and the low gravity pendulum with a length of 15m, making the low gravity
pendulum easier to transport, build and decommission. From Fig. 4.1 it’s shown that the x-y
pendulum is stabilised faster than the low gravity pendulum. This characteristic of the system
is not of much relevance, but even though, it has been taken into account.
In Section 4.2, a parametric study on the effect of wave amplitude, wave frequency and
damping factor was made.
Table 4.2 shows the wave amplitude range on which the pendulum systems will maintain
a rotational behaviour. The pendulum systems were optimized for a wave amplitude of 0.6m.
Table 4.2 shows that the x-y pendulum has a range from 0.001m to 1.2m, therefore it has a tol-
erance percentage of −98.3% and +100%. The low gravity pendulum system has a tolerance of

36
Chapter 5. Discussion

−98.3% and +33.3%. Being the low gravity pendulum the more susceptible to wave amplitude
changes.
Table 4.3 shows the wave frequency range on which the pendulum systems will maintain a
rotational behaviour. The pendulum systems were optimized for a wave frequency of 0.1667Hz.
From Table 4.3, it can be seen that the x-y pendulum has a tolerance percentage of −7.01% and
+41.6%. The low gravity pendulum has a tolerance percentage of −16% and +139.9%. Being
the x-y pendulum the more susceptible system to wave frequency changes.
Table 4.4 shows the maximum damping values that will maintain the pendulum systems
with a rotational behaviour. The pendulum systems were optimized for a damping coefficient
value of 0.0828. From Table 4.4 it can be seen that the x-y pendulum can increase its damping
by a +111.3%. While the low gravity pendulum system can increase its damping coefficient by
a +181059%. Making the low gravity pendulum the less susceptible system for external forces
and as reviewed in Section 4.2.3 the best design for energy harvesting.

37
Chapter 6

Conclusions and recommendations for

future work

This project has reviewed two pendulum systems for energy harvesting purposes. It has inves-
tigated the influence that wave amplitudes, wave frequencies, damping coefficient and physical
pendulum parameters has on its behaviour.
For this investigation, equations of motion for both pendulum systems were obtained and
simulations with the MatLab software using the ODE 45 and ODE 5 numerical functions were
carried and analysed to carry out a detailed comparison between both systems.
The results of the simulations shown in Sections 3.2, 4.2 and 4.1 were discussed in Sec-
tion 5 above. This results show that for achieving the desired period-1 rotational behaviour
of both pendulum systems without external interventions, a time-delay feedback controller im-
plementation is needed. Simulations have been conducted to demonstrate that both controlled
pendulum systems have the ability to initiate and maintain a rotational motion for the studied
wave parameters. Further parametric studies were made in order to see which pendulum system
will be more tolerant to changes and which pendulum system will be more optimal for energy
harvesting considerations. These parametric studies done in Section 4.2, lead to the conclusion
that the x-y pendulum will have more tolerance for a wave amplitude changes by a 67%. But
the low gravity pendulum will have a greater tolerance to wave frequency changes by a 107%,

38
Chapter 6. Conclusions and recommendations for future work

being a significantly higher tolerance than the one the x-y pendulum has on wave amplitude
changes.
Finally a parametric study was made on the availability to increase the damping coefficient
without affecting the rotational behaviour of the system. Since the damping coefficient can be
directly related to an energy extraction coefficient, the harder the damping is the more energy
that can be extracted from the system. The results shown in Section 4.2 based on the damping
coefficient showed that the low gravity pendulum has the ability to have a damping 1800 times
higher than the x-y pendulum. Therefore the low gravity pendulum will be the most optimal
out of the two pendulum systems reviewed for energy harvesting considerations and the more
tolerant to changes.

6.1 Recommendations

• This study was made based on harmonic wave excitation, considering deep waver waves
act as a sinusoidal waveform. Any further work on this project should be on realistic
wave profiles exciting the pendulum, making it a more realistic system.

• The reviewed pendulum systems can be verified in the experimental rig with a wave tank.

• Other control methods can be designed in order to keep a rotational response of the pen-
dulum systems for a wider range of wave parameters.

• The pendulum systems reviewed in this thesis can be modified and studied with a four-
pendula system in order to further optimize them.

39
References

[1] Airy, G. B. [1845], ‘Tides and waves’, Encyclopaedia Metropolitana 5, 241–396.

[2] Andreeva, T., Alevras, P., Naess, A. and Yurchenko, D. [2015], ‘Dynamics of a parametric
rotating pendulum under a realistic wave profile’, International Journal of Dynamics and
Control 4, 233–238.

[3] Breeze, P. [2014], ‘Marine power generation technologies’, 14.

[4] De Paula, A. S., Savi, M. A., Vaziri, V., Pavlovskaia, E. and Wiercigroch, M. [2015],
‘Experimental bifurcation control of a parametric pendulum’, Journal of Vibration and
Control 23, 2256–2268.

[5] Fournier, A. and Reeves, W. T. [1986], ‘A simple model of ocean waves’, ACM SIG-
GRAPH Computer Graphics 20, 75–84.

[6] Fréchot, J. [2006], ‘Realistic simulation of ocean surface using wave spectra realistic sim-
ulation of ocean surface using wave spectra’.

[7] Gao, Q., Ertugrul, N. and Ding, B. [2021], ‘Mechanical energy storage impact on the
performance of wave energy converter’.

[8] Hindcast² [2021].

URL: https://app.metoceanview.com/hindcast-squared/?/stats-point/lng/-
1.85883/lat/57.16167

40
References

[9] IEA [2021], ‘Review 2021 assessing the effects of economic recoveries on global energy
demand and co 2 emissions in 2021 global energy’.

[10] Kerr, D. [2007], ‘Marine energy: getting power from tides and waves’, Civil Engineering
Innovation 1, 81–92.

[11] Liu, X. T. and Rao, S. S. [2018], ‘Vibration analysis in the presence of uncertainties using
universal grey system theory’, Journal of Vibration and Acoustics 140.

[12] Mann, B. P. and Koplow, M. A. [2006], ‘Symmetry breaking bifurcations of a parametri-

cally excited pendulum’, Nonlinear Dynamics 46, 427–437.

[13] Morin, D. [2007], ‘The lagrangian method’, Harvard p. Chapter 6.

[14] Najdecka, A., Narayanan, S. and Wiercigroch, M. [2015], ‘Rotary motion of the para-
metric and planar pendulum under stochastic wave excitation’, International Journal of
Non-Linear Mechanics 71, 30–38.

[15] Publicationslimited, O. [1992], The North Sea field development guide., Oilfield Publica-
tions.

[16] Sebestyén, V. [2021], ‘Renewable and sustainable energy reviews: Environmental impact
networks of renewable energy power plants’, Renewable and Sustainable Energy Reviews
151, 111626.

[17] Terrero González, A., Dunning, P., Howard, I., McKee, K. and Wiercigroch, M. [2021],
‘Is wave energy untapped potential?’, International Journal of Mechanical Sciences
205, 106544.

[18] Tucker, M., Challenor, P. and Carter, D. [1984], ‘Numerical simulation of a random sea: a
common error and its effect upon wave group statistics’, Applied Ocean Research 6, 118–
122.

41
References

[19] Yurchenko, D. and Alevras, P. [2013], ‘Dynamics of the n-pendulum and its application to
a wave energy converter concept’, International Journal of Dynamics and Control 1, 290–
299.

[20] Yurchenko, D. and Alevras, P. [2018], ‘Parametric pendulum based wave energy con-
verter’, Mechanical Systems and Signal Processing 99, 504–515.

42
 Appendix A

X-Y pendulum Matlab code

1 clc
2 close all
3 clear
4
5 global A w g m c beta A_r wn l
6 %A = 0.6; %wave amplitude set 1 [m]
7 %w = 0.1667*2*pi; %wave frequncy set 1 [rad/s]
8 %beta = 0.6; %rad set 1
9 %l = 40; %length set 1 %m
10 g = 9.81; %gravity
11 m = 20; %pendulum mass [kg]
12 c =0.1; %pendulum damping factor [Ns/m]
13 wn = 0.1667*2*pi; %(A_r*wˆ2)/g %rad/s
14
15 A = 6; %wave amplitude set 2[m]
16 w = 0.05*2*pi; %wave frequncy set 2 [rad/s]
17 l = 45; %length set 2 %m
18 beta = 0.4; %rad set 2
19
20 A_r = A/sin(beta); %pivot displacement
21
22 time_step=0.01;
23 time_vector=(0:time_step:100);
24
25 %initial condition
26 x0=[1;1];
27 options=odeset('RelTol',1e-3,'AbsTol',1e-3);
28 [time_vector2,solution]=ode45 (@dynamics_xypendulum,
29 time_vector,x0,options);
30
31
32 y_t = A*sin(beta)*sin(w*time_vector2); %y
33 h_t = A*cos(beta)*cos(w*time_vector2); %x

43
 Appendix A. X-Y pendulum Matlab code

34 A_w = A_r*sin(w*time_vector2);
35
36 figure
37 ax1=subplot(2,1,1); %graph angle and velocity
38 theta_cut = mod(solution(:,1)+pi,2*pi)-pi;
39 plot(time_vector2,theta_cut,'.r')
40 hold on
41 plot(time_vector2,solution(:,2),'.k') %time history
42 xlabel('Time (s)')
43 ylabel('Angle (rad), $\ddot{\theta}$ (rad/s)', 'interpreter', 'latex')
44 legend('Angle (rad)','$\ddot{\theta}$ (rad/s)', 'interpreter', 'latex');
45 title('Time history')
46
47 ax2=subplot(2,1,2); %excitation amplitude
48 plot(time_vector2,A_w)
49

50 xlabel('Time (s)')
51 ylabel('Excitation (m)')
52 A = get(gca,'position'); % gca points at the second one
53 A(1,4) = A(1,4) / 2; % reduce the height by half
54 A(1,2) = A(1,2) + A(1,4); % change the vertical position
55 set(gca,'position',A); % set the values you just changed
56 linkaxes([ax1,ax2],'x');
57
58 figure
59 plot(theta_cut, solution(:,2),'.r') %phase-plane graph
60 xlabel('\theta (rad)')
61 ylabel('$\stackrel{..}{\theta}$ (rad/s)', 'interpreter', 'latex');
62 title('Phase-plane')
63 axis square
64
65 function dx = dynamics_xypendulum(t,x)
66 global A w g m c beta A_r wn l
67
68 y = -1*A*sin(beta)*wˆ2*sin(w*t); %y_double_dot
69 h = -1*A*cos(beta)*wˆ2*cos(w*t); %x_double_dot
70
71 dx = [x(2,1); (-m*h*cos(x(1,1))-m*(y+g)*sin(x(1,1))
72 -c*l*(x(2,1)))/(m*l)];
73 end

44
 Appendix B

Low gravity pendulum Matlab code

1 clc
2 close all
3 clear
4
5 global A w g m c beta A_r wn l
6 g = 9.81;
7 m = 20; %pendulum mass [kg]
8 c =0.1; %pendulum damping factor [Ns/m]
9 A_r = A/sin(beta); %[m]
10 wn = 0.05*2*pi; %(A_r*wˆ2)/g %rad/s
11
12 A = 6; %wave amplitude [m]
13 w = 0.05*2*pi; %wave frequncy [rad/s]
14 l = 25; %g/((2*wn)ˆ2); %m
15 beta = 0.5; %rad
16

17 time_step=0.01;
18 time_vector=(0:time_step:100);
19
20 %initial condition
21 x0=[1;1];
22 options=odeset('RelTol',1e-3,'AbsTol',1e-3);
23 [time_vector3,solution1]=ode45 (@dynamics_lowgravitypendulum,
24 time_vector,x0,options);
25
26
27 y_t = A*sin(beta)*sin(w*time_vector3); %y
28 h_t = A*cos(beta)*cos(w*time_vector3); %x
29 A_w = A*sin(wn*time_vector3);
30 figure
31 ax1=subplot(2,1,1); %angle
32 theta_cut1=mod(solution1(:,1)+pi,2*pi)-pi;
33 plot(time_vector3,theta_cut1,'.r')

45
 Appendix B. Low gravity pendulum Matlab code

34 hold on
35 plot(time_vector3,solution1(:,2),'.k') %time history
36 xlabel('Time (s)')
37 ylabel('Angle (rad), $\ddot{\theta}$ (rad/s)', 'interpreter', 'latex')
38 legend('Angle (rad)','$\ddot{\theta}$ (rad/s)', 'interpreter', 'latex')
39 title('Time history')
40
41 ax2=subplot(2,1,2); %excitation
42
43 plot(time_vector3,A_w)
44
45 xlabel('Time (s)')
46 ylabel('Excitation (m)')
47 A = get(gca,'position'); % gca points at the second one
48 A(1,4) = A(1,4) / 2; % reduce the height by half
49 A(1,2) = A(1,2) + A(1,4); % change the vertical position
50 set(gca,'position',A); % set the values you just changed
51 linkaxes([ax1,ax2],'x');
52
53 linkaxes([ax1,ax2],'x');
54
55 figure
56 plot(theta_cut1, solution1(:,2),'.r') %phase-plane
57 xlabel('\theta (rad)')
58 ylabel('$\stackrel{..}{\theta}$ (rad/s)', 'interpreter', 'latex')
59 title('Phase-plane')
60 axis square
61
62
63 function dx = dynamics_lowgravitypendulum(t,x)
64 global A w g m c beta A_r wn l
65
66 f = -A_r*(wˆ2)*cos(w*t);
67
68 dx = [x(2,1); (-c*(x(2,1))-m*g*l*sin(beta)
69 *sin((x(1,1)))+m*l*f*sin((x(1,1))))/(m*(lˆ2))];
70 end

46
 Appendix C

X-Y controlled pendulum Matlab code

1 clc
2 close all
3 clear
4
5 global A w g m c alpha A_r wn l time_step
6 uu uu_saved x_saved x_dot_saved e_x e_x_dot
7
8
9 %Dimension from Aline's paper
10 g = 9.81; %gravity
11 w = 0.1667*2*pi; %1.0360 to 3.1250 %wave frequncy [rad/s]
12 l = 25; %g/((2*w)ˆ2); %length (m)
13 wn = sqrt(l/g); %(A_r*wˆ2)/g %rad/s
14 m = 5; %pendulum mass [kg]
15 c =0.175; %pendulum damping factor [Ns/m]
16 alpha = 0.5;%0.8; %rad
17 %A_r = A/sin(alpha); %[m]
18
19 A = 0.6; %wave amplitude [m]
20
21
22 time_step=0.1;
23 time_vector=(0:time_step:350);
24
25
26
27

28 uu=0;
29 uu_saved=zeros(length(time_vector),1);
30 x_saved =zeros(length(time_vector),1);
31 x_dot_saved=zeros(length(time_vector),1);
32 e_x=zeros(length(time_vector),1);
33 e_x_dot=zeros(length(time_vector),1);

47
 Appendix C. X-Y controlled pendulum Matlab code

34 %initial condition
35 x0=[0;0;0];
36 solution=ode5(@dynamics_xypendulum, time_vector,x0);
37 time_vector2=time_vector;
38
39 y_t = A*sin(alpha)*sin(w*time_vector2); %y
40 h_t = A*cos(alpha)*cos(w*time_vector2); %x
41 A_w = A*sin(wn*time_vector2);
42
43 figure
44 ax1=subplot(5,1,1); %angle graph
45
46
47 plot(time_vector2,mod(solution(:,1)+pi,2*pi)-pi,'.r--')
48
49 xlabel('Time [s]')
50 ylabel('\theta')
51
52
53 ax2=subplot(5,1,2); %velocity
54 plot(time_vector2,solution(:,2),'.k')
55 xlabel('Time [s]')
56 ylabel('$\ddot{\theta}$ (rad/s)', 'interpreter', 'latex')
57 b = get(gca,'position'); % gca points at the second one
58 b(1,4) = b(1,4) / 3; % reduce the height by half
59 b(1,2) = b(1,2) + b(1,4); % change the vertical position
60 set(gca,'position',b);
61
62 ax3=subplot(5,1,3); %excitation
63 plot(time_vector2,A_w)
64
65 xlabel('Time [s]')
66 ylabel('Excitation [m]')
67 p = get(gca,'position'); % gca points at the second one
68 p(1,4) = p(1,4) / 3; % reduce the height by half
69 p(1,2) = p(1,2) + p(1,4); % change the vertical position
70 set(gca,'position',p);
71
72 ax4=subplot(5,1,4); %feedback errors
73 plot(time_vector2,e_x,'b')
74 hold on
75 plot(time_vector2,e_x_dot,'r')
76 xlabel('Time [s]')
77 ylabel('Errors')
78 legend('e_x','e_{xdot}')
79 r = get(gca,'position'); % gca points at the second one
80 r(1,4) = r(1,4) / 4; % reduce the height by half
81 r(1,2) = r(1,2) + r(1,4); % change the vertical position
82 set(gca,'position',r);
83

48
 Appendix C. X-Y controlled pendulum Matlab code

84
85 ax5=subplot(5,1,5); %control signal
86 plot(time_vector2,uu_saved,'.k')
87 xlabel('Time [s]')
88 ylabel('Control signal')
89 z = get(gca,'position'); % gca points at the second one
90 z(1,4) = z(1,4) / 4; % reduce the height by half
91 z(1,2) = z(1,2) + z(1,4); % change the vertical position
92 set(gca,'position',z); % set the values you just changed
93

94 linkaxes([ax1,ax2,ax3,ax4,ax5],'x');
95
96 function dx = dynamics_xypendulum(t,x)
97 global A w g m c alpha A_r wn l time_step uu
98 uu_saved x_saved x_dot_saved e_x e_x_dot
99

100 t_index=round(x(3,1));
101 TT=2*pi/w;
102 k1=0.05; k2=0.01;
103 delay=round(TT/time_step);
104 %x_saved(t_index+1)=mod(x(1,1)+pi,2*pi)-pi;
105 x_saved(t_index+1)=x(1,1);
106 x_dot_saved(t_index+1)=x(2,1);
107 y = -1*A*sin(alpha)*wˆ2*sin(w*t); %y_double_dot
108 h = -1*A*cos(alpha)*wˆ2*cos(w*t); %x_double_dot
109 if x(3,1)*time_step>21
110 e_x(t_index+1)=x_saved(t_index-delay,1)-x_saved
111 (t_index,1)+2*pi*sign(x(2,1)); %error
112 e_x_dot(t_index+1)=x_dot_saved(t_index-delay,1)
113 -x_dot_saved(t_index,1); %error
114 uu=k1*e_x(t_index+1)*m*l*sign(x(2,1))+k2*e_x_dot
115 (t_index+1);
116 uu_saved(t_index+1)=uu;
117 end
118 dx(1,1) = x(2,1);
119 dx(2,1)= (-m*h*cos(x(1,1))-m*(y+g)*sin(x(1,1))
120 -c*l*(x(2,1))+uu)/(m*l);
121 dx(3,1)=round(1/time_step);
122 end

49
 Appendix D

Low gravity controlled pendulum Matlab

code

1 clc
2 close all
3 clear
4
5 global A w g m c alpha A_r wn l time_step uu
6 uu_saved x_saved x_dot_saved e_x e_x_dot
7
8 %Dimension from Aline's paper
9 g = 9.81;
10 w = 0.1667*2*pi; %1.0360 to 3.1250 %wave frequncy [rad/s]
11 l = 15; %g/((2*w)ˆ2); %length (m)
12 wn = sqrt(l/g); %(A_r*wˆ2)/g %rad/s
13 m = 5; %pendulum mass [kg]
14 c =150.0828; %pendulum damping factor [Ns/m]
15 alpha = 0.2;%0.8; %pivot angle (rad)
16 A_r = A/sin(alpha); %excitation [m]
17

18 A = 0.6; %wave amplitude [m]

19
20
21 time_step=0.1;
22 time_vector=(0:time_step:350);
23
24
25
26
27 uu=0;
28 uu_saved=zeros(length(time_vector),1);
29 x_saved =zeros(length(time_vector),1);

50
 Appendix D. Low gravity controlled pendulum Matlab code

30 x_dot_saved=zeros(length(time_vector),1);
31 e_x=zeros(length(time_vector),1);
32 e_x_dot=zeros(length(time_vector),1);
33

34 %initial condition
35 x0=[0;0.000001;0];
36 solution=ode5(@dynamics_xypendulum, time_vector,x0);
37 time_vector2=time_vector;
38
39 y_t = A*sin(alpha)*sin(w*time_vector2); %y
40 h_t = A*cos(alpha)*cos(w*time_vector2); %x
41 A_w = A*sin(wn*time_vector2);
42 figure
43 ax1=subplot(5,1,1); %pendulum time history angle
44 plot(time_vector2,mod(solution(:,1)+pi,2*pi)-pi,'.b--')
45

46 xlabel('Time [s]')
47 ylabel('\theta')
48
49 ax2=subplot(5,1,2); %velocity (rad/s)
50 plot(time_vector2,solution(:,2),'.k')
51 xlabel('Time [s]')
52 ylabel('$\ddot{\theta}$ (rad/s)', 'interpreter', 'latex')
53 b = get(gca,'position'); % gca points at the second one
54 b(1,4) = b(1,4) / 3; % reduce the height by half
55 b(1,2) = b(1,2) + b(1,4); % change the vertical position
56 set(gca,'position',b);
57
58 ax3=subplot(5,1,3); %excitation amplitude
59 plot(time_vector2,A_w)
60
61 xlabel('Time [s]')
62 ylabel('Excitation [m]')
63 p = get(gca,'position'); % gca points at the second one
64 p(1,4) = p(1,4) / 3; % reduce the height by half
65 p(1,2) = p(1,2) + p(1,4); % change the vertical position
66 set(gca,'position',p);
67
68 ax4=subplot(5,1,4); %error graph
69 plot(time_vector2,e_x,'b')
70 hold on
71 plot(time_vector2,e_x_dot,'r')
72 xlabel('Time [s]')
73 ylabel('Errors')
74 legend('e_x','e_{xdot}')
75 r = get(gca,'position'); % gca points at the second one
76 r(1,4) = r(1,4) / 4; % reduce the height by half
77 r(1,2) = r(1,2) + r(1,4); % change the vertical position
78 set(gca,'position',r);
79

51
 Appendix D. Low gravity controlled pendulum Matlab code

80 ax5=subplot(5,1,5); %control signal

81 plot(time_vector2,uu_saved,'.k')
82 xlabel('Time [s]')
83 ylabel('Control signal')
84 z = get(gca,'position'); % gca points at the second one
85 z(1,4) = z(1,4) / 4; % reduce the height by half
86 z(1,2) = z(1,2) + z(1,4); % change the vertical position
87 set(gca,'position',z);
88 linkaxes([ax1,ax2,ax3,ax4,ax5],'x');
89

90
91 function dx = dynamics_xypendulum(t,x)
92 global A w g m c alpha A_r wn l time_step uu
93 uu_saved x_saved x_dot_saved e_x e_x_dot
94 t_index=round(x(3,1));
95 TT=2*pi/w;
96 k1=0.5; k2=0.5;
97 delay=round(TT/time_step);
98 %x_saved(t_index+1)=mod(x(1,1)+pi,2*pi)-pi;
99 x_saved(t_index+1)=x(1,1);
100 x_dot_saved(t_index+1)=x(2,1);
101 %y = -1*A*sin(alpha)*wˆ2*sin(w*t); %y_double_dot
102 %h = -1*A*cos(alpha)*wˆ2*cos(w*t); %x_double_dot
103 f = -A_r*(wˆ2)*sin(w*t);
104 if x(3,1)*time_step>100
105 e_x(t_index+1)=x_saved(t_index-delay,1)
106 -x_saved(t_index,1)+2*pi*sign(x(2,1)); %feedback error
107 e_x_dot(t_index+1)=x_dot_saved(t_index-delay,1)
108 -x_dot_saved(t_index,1); %feedback error
109 %uu=k1*e_x(t_index+1)+k2*sign(x(2,1))*e_x_dot(t_index+1);
110 uu=k1*e_x(t_index+1)*m*l*sign(x(2,1))+k2*e_x_dot(t_index+1);
111 uu_saved(t_index+1)=uu;
112 end
113 dx(1,1) = x(2,1);
114 dx(2,1)= (-c*(x(2,1))-m*g*l*sin(alpha)*sin((x(1,1)))
115 +m*l*f*sin((x(1,1)))+uu)/(m*(lˆ2));
116 dx(3,1)=round(1/time_step);
117 end

52
Appendix E

Risk assessment

School of Engineering Risk Assessment Page 1 of 2

Risk Assessment
PROCEDURE:
 Complete risk assessment in consultation with PI/Supervisor and technical staff as appropriate.
 Risk assessment checked and signed by PI/Supervisor
 A copy or scan of the signed document to be given to the lab technician, School Safety Adviser and PI/Supervisor.
NOTES:
 No laboratory work is to commence without a risk assessment signed by the PI/Supervisor.
 The risk assessment must be reviewed when any changes are made to the equipment, materials, procedure or personnel.
 Technical staff can stop work if no risk assessment is in place or if, in their opinion, there is a risk to safety.
 Examples of how to complete this form are available at www.hse.gov.uk/risk/casestudies/

Project name: Adaptability of a frequency changing harmonically excited pendulum for wave energy harvesting
Location of work: University of Aberdeen
Principal Investigator/Supervisor: Dr Vahid Vaziri Date:

Signed:

Assessment Prepared by: Manuel Beltra Soler Signed:

Date: 24/04/2022

Outline description of the work: This project is about the adaptability of two pendulum systems excited by waves that aim to maintain a rotational response from
which energy will be extracted.

Names of persons carrying out the Dr Vahid Vaziri, Manuel Beltra Soler.
work:

16 April 2022

53
Appendix E. Risk assessment

School of Engineering Risk Assessment Page 2 of 2

What are the hazards? Who might be harmed and What are you already doing? Do you need to do anything else to Action by Action by Done
how? manage this risk? whom? when?
Screen brightness The student Try to keep the brightness as low as possible No X
and take some breaks

Maintaining same posture for The student Do some stretching and take some breaks No X
long periods of time

Covid propagation The student and Wear masks and make periodic tests No X
supervisor

16 April 2022

Figure E.1: Risk assessment

54