Emerging Tech in Maritime Industry

INTRODUCTION

As technology advances, the trend of introducing new technologies in the maritime industries????. Among the new technologies, ththe one which attracts attention is the unmanned ship which operates without people. This is a very attractive technology not only to save man's labour but also to reduce accidents for human beings. This unmanned ship technology has been applied to marine robots, which are smaller in size than the technology applied to large vessels so far. There are two major types of robots used in this field: AUV (Autonomous Underwater Vehicle) and ASV (Autonomous Surface Vehicle). They are ideally free to move in any direction towards which an operator could want them to, so in the 3-D situation of a marine environment, they will be essential as one of the research tools (Blair Thornton, 2007). It can also be used for ship hull plate inspection of naval vessels and marine pollutants; which humans have difficulty to do due to danger. It will also be utilized for marine mine detection and treatment and coastal monitoring for military purposes (Townsend, 2017). Moreover, the Unmanned vehicle has a diverse field of applications and the various applications demand enormous numbers of different types of robots. For using marine robots, the robots must take care of crew's work and navigate the ship, so a naval control system, a camera system, a location tracking system and a communication system should be installed in the body of marine robots. Sufficient power must be supplied to operate these systems unattended. Also, such a system is expensive, and it is essential to supply an ample power supply for a complete return of robots after tasks. Therefore, the energy is a crucial issue for long endurance and operating them for performing desired tasks safely. ASVs and AUVs have been conducted only for a few hours or days at a time because the main energy sources are the batteries and they to need to be charged. However, the technology to charge the battery in real-time at sea is practically tricky and requires docking for charging. To answer these problems, energy saving technology of marine robots using renewable energy such as solar energy, wind and waves has been recently studied. Among them, oceans waves have great potential as renewable energy on this planet because there is no effect of the difference of day and night like solar energy, and the sea water is about 850 times heavier than air, which means when water is in motion then so is a lot of energy (Higgins, 1982).Besides, since the wave energy has a more significant effect as it approaches the sea surface and shallow water, it has an incentive to obtain a substantial energy harvesting effect when applied to an ASV running on the water surface.

Therefore, the motivation comes, which uses huge wave energy potential to generate power to help the endurance of ASVs. Following this, the next step is to choose which methods of wave energy harvesting technology to apply the ASVs. Among various methods, in this paper, the method, which uses the Control Moment Gyro system (CMGs) in the autonomous surface robots, was considered. The first reason to choose the gyroscope system is the operating property of gyroscope. By using wave induced motion and gyroscopic effect, a large torque is generated. Following this principle, using generated torque by gyroscopic effect of CMG to operate a generator, is considered to verify wave energy harvesting. Futhermore, because the gyroscope has been used as a gyrostabilizer in the huge ferries and luxury yachts which are developed as autonomous ships and similar structure to ASVs, this method will be easily adapted to an actual marine robotic industry by merely adding the power which is necessary to adapt the device to the existing gyrostabilizer systems.

In conclusion, Marine robots are attracting attention as the first technology of the next generation marine industry. The types of marine robots are largely divided into ASVs on water as vessels and AUVs submerged in under water. However, one of the drawbacks in the marine robots is the low power capacity of power sources represented by batteries. To solve this problem, A method of finding other energy sources has been considered, and among them, a method using a wave energy in which a marine robot is exposed has been noted. Wave energy will focus more on ASVs because ASVs that are active on the surface rather than AUVs can be used more efficiently. In addition to the effective wave, energy harvesting method is motivated by using CMGs which is used as gyrostabilizer in the actual ships similar to the ASV structure currently. To verify this mentioned method, in this paper, It will be investigated the effect of energy harvesting and the effect of ASV motion stabilization obtained by installing a CMG used as a gyrostabilizer in ASV body during wave conditions.

The aim and objectives

The aim of this project is an investigation into the potential of using control moment gyro (CMG) on autonomous surface vehicles (ASVs) for wave energy harvesting and roll-motion stabilization of ship motion.

On sea surface the wave condition; pitching and rolling, gives rotation motion to the body of ASVs. In this paper, the ASV motion of roll is focused. Then, the ASV motion which is induced by a wave generated torque for rotating gimbal of CMG, is called precession. The purpose of this thesis is to verify how much energy could be generated by this precession of CMG in different wave conditions and flywheel speed of CMG. Besides, when operating CMG on ASV, the stabilization effect of ASV motion by the gyroscopic effect of CMG will be expected. Therefore, it also will be investigated with energy harvesting.

The he numerical calculations for roll motion of ASV considering wave motion and CMG systems.

The final roll motions of the ASV. These will have to be combined to get the final motions of the ASV including wave induced force, a torque generated by CMG.

The he numerical calculations for wave energy harvesting system

The generated energy from CMG precession. These will be investigated by installing the device at gimbal axis of CMG which is rotated by precession motion.

The verification through experimental methods.

The experiment of the ASV model with the CMG system in the towing tank. The experiment will be set-up in regular wave conditions considering numerical calculations above to verify the aim of this project.

The analysis and comparison

Analysis of experimental results. The results of each test of the experiment will be compared. Discussions would involve the determination of the possibility of using the experimental results and the future work.

LITERATURE STUDY

The purpose of the literature study is to provide a theoretical background to the different effects and forces that need to be taken into consideration when looking at this specific case. It contains substantive findings relevant to the topic, including theoretical and methodological contributions.

Autonomous Surface Vehicles (ASVs) and operating hours

Autonomous Surface Vehicles (ASVs) are robotic vehicles which are self-propelled and unmanned platforms which float on the sea surface with by itself through autonomous capability, which researches marine environment across a range of variables on the sea surfaces. Also, ASVs can be utilized broadly in the marine sector from research and environmental monitoring programs to naval and defense applications. ASV has a higher battery capacity than AUVs (N.O.C, 2018). Moreover,ASVs are operated by a lot of methods of propulsion, principally wave-powered or propeller driven. As per the development of ASVs, In the literature, the first effort to the autonomous vessel by wind propelled is a project named SKAMP (Station Keeping Autonomous Mobile Platform). It was developed in 1968 by E. W. Schieben. The second published autonomous sailing attempt was the Relation Ship project of the University of Applied Science in Furtwangen, Germany in 1995. But the project was failed by the operating plan which was exchanged by using satellite, not autonomous (Jafarmadar, 2011). From the program from the MIT Sea Grant College, Autonomous Surface Vessels (ASVs) including ARTEMIS which is the first one was developed in 1993. This vessel was then used to collect simple bathymetry data in the Charles River in Boston, MA. It is quite meaningful due to the first prototype; however, this is restricted by endurance and seakeeping (Vaneck, 1997).

In addition to it, in 1997, the Remote Mine-hunting Operational Prototype (RMOP) was operated from USS Cushing in the Persian Gulf to conduct mine hunting(R Yan, 2010). In the 2010, in accordance with National Oceanography Centre of the UK, the government-backed Small Business Research Initiative (SBRI) launched a competition to improve reliable long-endurance ASVs. Moreover, in 2014, the two vehicles, AutoNaut (Fig.1) from the MOST Ltd and C-ENDURO (Fig.2) from the ASV Ltd were showcased to the marine community at the NOC in Southampton. These two vehicles firstly embarked on their official purpose named as ‘Exploring Ocean Fronts’ (N.O.C, 2018).

Among the development of ASVs, the most significant feature of ASVs considered was the ability to use power capacity. This is very important because it is directly related to the loss of driving ability in a remote situation. Besides, this capacity includes not only the main propulsion power but also other measuring sensors, cameras, accessories and various communication devices. Until now, batteries are the major power supply of the autonomous vehicles. Batteries have a limited life, sometimes need to recharge period for re-operating from an operator or other support vessels (Townsend, 2016).

In table 1, there is the summary of the power requirements of various kinds of battery powered ASVs which currently used to. For covering these drawbacks, the improvement of energy storage method or more efficient energy devices is required. In summary, ASVs are performing well in a variety of fields, from initial models to models currently being developed, including ecosystem surveys, military purposes and environmental protection. Among various properties of ASV, overcoming endurance hours of ASVs in the marine environment is one of the major projects for researchers and developer.

Wave energy harvesting

The marine robots such as ASVs is one of the remarkable technology. However, there is still a significant problem of ensuring a long operating time for the operation of marine robots including ASVs for the desired purpose. Currently, most of the robots use the batteries as their main power source but the batteries have problems such as limited time use and time for recharging (Townsend, 2016). Referring to the research of G.Griffiths, Long endurance propeller-driven AUVs are likely to require at least 100 kWh of energy. The study found that using Li-Ion cells would produce 1000 kWh of energy, with the AUV expected to be more than 30 tones. It is costly and unrealistic (G.Griffiths, 2004). Therefore, for practical application, it was suggested to increase propulsion efficiency or to develop another efficient energy source. For developing new technologies to alter the batteries, much research has been carried out in the field of AUVs whose operating time is shorter than that of ASVs until now and these technologies apply to ASVs because they can be used on the sea surface. Following this, in the case of marine robots, the importance of getting renewable energy from the environment has been recognized for a long time because they are continuously exposed to the natural environment.

Many kinds of energy harvesting methods have been developed and applied in marine robot industry by developers and researchers (H.Kanki, 2009). Among the renewable energies available to ASVs, the most effective power is investigated as the harvesting of wave energy. According to the researches, the density of the wave energy contained in the world oceans is more than 2 kW/m, which is more than 90% of the year (Zheng, 2014). On the other hand, the average wind energy density is 50 W/m (Hermann, 2006), and the average solar energy density at sea level is 168 W/m (Wei Zheng, 2014). Also, wave energy harvesting is cost-effective because the energy sources of the robot can be charged even during missions and there is no need for additional devices to help in recharging or propulsion. Although tides have the potential to generate huge amounts of energy (approximately 3,000 GW worldwide) less than 3% of areas are suitable to actually harness tidal power (Esteban, 2012). However, the wave energy has no disadvantage such as tidal power because they are found at many more locations throughout the world, as can be seen in Figure 3.

There are numerous devices that have been established in the wave energy industry which attempt to most efficiently convert the potential energy of waves into power. The first suggestion to harvesting ocean wave energy device was proposed more than 30 years ago (S.H.Salter, 1974). A wide variety of wave energy harvesting devices which have been named Wave Energy Converters(WECs), have been proposed to extract power from the sea (J.Cruz, 2007). One of the most fundamental types of WECs is to use the point absorber developed by Oregon State University which is shown in figure 4. These devices float on the surface of the water and generate energy through the periodic passing of waves, causing the device to bob up and down. Point absorbers depend on an internal hydraulic system which could pump air to power a generator as the cylinder is compressed and released by wave energy. The system resonates at the dominant period of the waves and this resonant motion generates electricity (Ted K.A. Brekken, 2009).The drawback of point absorbers is that they can’t adapt to the varying height and frequency of waves and this prevents optimization through consistency.

The other type of WECs is ‘Oscillating water columns (Figure 5)’. It is operated based on a pressure differentials which occur between air and the ocean water within a structure that is partially submerged. These devices are located along the shoreline to catch wave crashing on the coast, generating the wave pressure to push air through a hollow cavity that is attached to a turbine.

The rapid decrease in water pressure then pulls air back through the bi-directional turbine, generating electricity through air flow in both directions. The advantage of this device is low maintenance costs due to the lack of complex internal machines. However, even it, has economic effect which could be understood by the factor the constructing this structure near the coast causes larger environmental implication than other devices. At last, by using Control Moment Gyroscopes(CMGs) wave energy generating method has been developed in these days. As interest in the use of wave energy as renewable energy has expanded, methods for generating energy using gyroscopic effects and torques generated from CMGs installed in marine robots have been proposed. In 2009, Spain developed a new offshore wave energy converter named OCEANTEC WEC. The feature of this product is that the electric generator is operated by force generated in the pitching direction by the wave on the gyroscopic device installed inside the floating device. Although it was limited to the pitching direction, it is meaningful to test the mathematical model that was designed theoretically in actual sea conditions and to confirm the development potential of the wave energy converter design model (F. Salcedo, 2009). From 2000 to 2010, H. Kanki in Japan developed an improved high-efficiency wave energy converter using gyroscopic moment through a device consisting of float and gyroscope (H.Kanki, 2009). Till recently, research on wave energy harvester in underwater robots through CMGs has been continuing (Townsend, 2016) and researches on the endless propulsion of underwater robots have been carried out using them (Townsend, 2016).This method by using CMGs to harvest wave energy is appropriate for application of marine robotsbecause the size of the device is smaller than the others, which makes it suitable to be installed easily on robotic platforms. Also, by using only power for operating flywheel of gyroscope, it can easily generate energy through various wave conditions. In conclusion, to solve the energy problem of marine robots, wave energy harvesting technology is considered as effective solution. there are various methods of wave energy harvesting methods such as point absorbers, Oscillating water column and Control moment gyros. Among them, because it is easy to install and small, Control moment gyros are seen to be a suitable method for use in marine robots. Also, since the ship already uses Gyroscopes, it is possible to use it. This method was examined in more detail in the next section.

Ship stabilization including gyro stabilization

Since ships operate at sea, any vessel exposed to wind and waves experience motion in the six degrees of freedom such as surge, sway, heave, roll, pitch and yaw. Among them, the rolling motion is a phenomenon in which the left and right sides of a ship are alternately vibrated around the centre of the ship by the waves. Such rolling movement interferes with the comfort of the occupants of the ship, causes accidents during maritime operations, and could produce severe accidents in which the ship capsizes. The equipment used to reduce the rolling of a ship is called a ship roll motion stabilization system.

Ship stabilization systems are separated into passive and active systems. Some of the most commonly used systems are listed below.

Passive stabilization systems:

- Bilge keels.

- Trim tabs.

- Passive anti-roll tanks.

Active stabilization systems:

- Rudder stabilization.

- Fins.

- Active anti-roll tanks.

- Gyroscopic stabilization systems.

The bilge keels are the most generally used passive ship stabilization system. The bilge keels are fitted in pairs, towards the bottom of the side of the hull and often run along much of the length of the ship. How the bilge keel could influence the roll motion of a vessel depends on the bilge radius, length, width, position along the ship, angle and the size of the vessel. Bilge keels are widely used on different kinds of vessel types. This is because the bilge keels are easily applied to any kind of vessel, at a low cost. Moreover, it does not occupy any internal area and does not affect the original hull shape. On the other hand, bilge keels do not contribute as much to roll damping as active stabilization systems. Therefore, they are often used in combination with other systems. The bilge keels also affect the hydrodynamic resistance of the vessel negatively, hindering forward motion. However, it is not a concern for the model used in this project. The other method is passive anti-roll tanks. Anti-roll tanks were invented by William Froude. Froude installed water chambers in the upper part of a ship to decrease the movement in roll motion. These anti-roll tanks are free of surface tanks that are supposed to counteract the roll motion of the ship. As the water flows from each side to the other, inside the chamber, the roll motion of the vessel can be diminished. By changing the level in the chambers, one can control the natural frequency of the tank. Froude’s free surface tanks did not receive much attention until the middle of the 20th century when it became widely used in naval vessels. Another type of anti-roll tank is the U-tube tank, introduced by Frahm in 1910. The U-tube tank works one the same principles as those of Froude’s anti-roll tank. By moving water from one side of the U-shaped tank to another, a counteracting roll moment is produced. The movement of water is controlled by baffles. However, these above mentioned methods without gyroscopic stabilization have a restriction that the ship does not generate lifting force in the straight line. Since the early 2000s, research has been carried out to actively and efficiently control the rolling motion even when the ship is running or in the right direction. For knowing about gyroscopic stabilization, first, the principle of the control moment gyro that produces the stabilization torque would be described.

The Control Moment Gyro(CMG) is a torque generating actuator using the principle of physics gyroscope and is composed of the flywheel, spin motor and gimbal motor functionally. The flywheel is mounted on the spin motor rotary shaft and the gimbal motor rotary shaft is disposed perpendicular to the spin motor rotary shaft. Now the angular momentum (h) is generated when the spin motor drives the flywheel. When the spindle motor rotates about the centre axis (∅) of the spindle motor by the driving of the gimbal motor, the gyro torque (c) is generated on the axis perpendicular to the two spindles (See the figure. 6). As a way of doing this, researchers have developed the ship roll-motion stabilization system applying gyroscoped and applied it to leisure vessels, special purpose vessels and military vessels. Founded in 2003, SeaKeeper of the United States developed a prototype through four years of research and started selling the product in 2007. SeaKeeper's rolling damping system consists of a flywheel, a spin motor and a hydraulic device for gimbal speed control. The flywheel is characterized by its rotation in a vacuum housing to reduce air resistance (Seakeeper, 2018). Furthermore, Gyro marine’s rolling damping system consists of active drive with rotor that offers gyroscopes with stabilizing capacities from 300 [KNm] to 1850 [KNm] and offsets the roll disturbance using the gyroscopic torque generated at that time (gyromarine, 2018).

In recent years, researchers have been carried out to apply CMGs that generate torque for using motion stabilization in underwater robots and surface vehicles. In 2007, Blair Thornton first applied the CMGs technology, which had not been used in marine robots, while building the robot IKURA. According to his research, he proposed a method to control the motion of a robot in zero gravity using robot motion control technology through CMGs in motion with underwater robots requiring accurate 3-D mapping technology (Blair Thornton, 2007). In this process, we found that CMGs can be used for propulsion, motion control, etc. and in marine robots including ASVs. However, the use of CMGs in marine robots has been limited to underwater robots. In addition, if we utilize wave energy as the energy source of these CMGs, we also know that it is a great innovation that the robot can produce energy by itself without any other energy source. In this paper, the method of stabilization using torque generated after driving CMGs using wave energy will be investigated for ASVs without external energy source.

In summary, through a review of the literature, at first the definition and evolution of ASVs are identified. Second, the transformation of wave energy into renewable forms of energy is described. Finally, the principles of control moment gyro and stabilization using gyro in the marine field are examined.

METHODOLOGIES

Through the literature review, this paper summarized the utilization of marine robot industry and CMGs technology. Within this content, we have broadened the idea of ASVs roll motion control and energy harvesting using CMGs and embodied it in the aim and objectives part. To achieve these goals and objectives, the methodology section will describe more specific goals and planning processes. This section is divided into two major parts; the numerical calculation and experiments set-up. The first part is to understand the roll motion of ASVs with wave induced force and the torque which is generated by CMG system. Also, the harvested energy through generator installed in CMG system will be investigated. Then, the investigated values will be formulated into equations. The second part is to set-up the experiment procedures. The ASV in combination with CMG system was planned to verify the numerical calculation and to investigate actual results of wave energy harvesting and motion stabilization result. This section will introduce the configuration of the system required for the experiment, the types of sensors required to measure the data, the method for obtaining the required data and the towing tank as well as the procedures for the experiment.

Numerical calculation

This chapter will provide the necessary theory to understand the following chapters. Most important is to inform the concept of the ASV motion with gyroscope torque and wave induced force. Moreover, how gyroscope can generate a stabilizing torque and harvest energy.

Gyroscopes motion calculation

For the representing the gyroscope motion and vessel motion, the frame was set-up. (XE,YE,ZE) represent hydrodynamic fixed frame. It is not moved with ASV’s motion. (XA,YA,ZA) represents ASV’s body fixed frame. The frame (XA,YAZA) is able to rotate or move with respect to the frame (XE,YE,ZE). When the frame (XA,YA,ZA) rotates with respect to the frame (XE,YE,ZE), the denoted angles at each axis are , ∅, respectively. Then, (XF,YF,ZF) represents the flywheel of CMG unit fixed frame.

In this paper, the flywheel rotation and precession are, respectively. In this paper, the centers of all frames coincide. These frames which were mentioned are described at figure 7. According to Newton’s second law, when the flywheel is operated, the spinning flywheel will create an angular momentumHf as follows,

Hf=Ifspin

where Hf is the moment of inertia in direction of the spin axis (Y3) and spinis angular velocity of flywheel.

These angular momentum is referred to flywheel unit fixed frame. Then this can be expressed in the ASV’s body fixed frame (XA,YAZA) by using rotation matrix AfA. AfAis the rotation matrix which transformation of angular moment factors from (XF,YF,ZF) to (XA,YAZA).

Following this, this equation is expressed,

HA*=AfA.Hf

where HA*is the angular momentum of flywheel in the ASV body fixed frame (XA,YAZA). Then, including the ASV motion, the moments acting around each axis in the ASV body fixed coordinated frame (XA,YAZA) can be expressed as;

HA=HA*+HA*

Where HA represents the moments acting around each axis in the ASV body fixed frame (XA,YAZA), represents the skew-symmetric form (equivalent to the cross product operation) of the ASV body motion experienced by the flywheel. HA*is the angular momentum of flywheel in the ASV body fixed frame (XA,YAZA). When the flywheel and precession axis is assumed to be restricted at Yf and Zf respectively, the flywheel’s angular velocity is in Yf axis and precession angular velocity is in Zf axis. In addition, assuming that both of the centre of mass between the flywheel and ASV lie at the origin of the ASV body fixed frame and the body-frames ((XA,YAZA) of reference coincide with the principal axes(XE,YE,ZE) of inertia of the bodies, then the rotation matrix has been expressed using Euler angles.

Equation of the Gyroscopic power

Expanding Eq.(3) the gyroscopic moment can then be fully expressed as;

HA= Iyysin +Iyycos -Iyycos +Izz Iyycos -Iyysin +Iyysin -Izz Izz-Iyysin +Iyycos

Where the roll, pitch and yaw motion of ASV body represent , , respectively. Neglecting gimbal effect of CMG, the following equation is expressed as follows;

HA= Iyysin +Iyycos Iyysin -Iyysin

In this paper, the motion of roll is focused to induced ASV motion and energy harvesting motion. Therefore, for getting the required data, the following assumptions are used to simplify the equations. The first thing is ship structure is port/starboard symmetry structure. Secondly, the surge force (ship propulsion) is zero. Thirdly, it could be assumed that the pitch and yaw motion of ASV body is very small, so the motion would not affect ASV’s roll motion. Then, the following equation is obtained.

g= Iyycos

In this equation, g is the moment from gyro system in roll motion

Equation of the ASV motion with Gyroscope.

The equation of ASV roll motion on the wave condition can be expressed as follows;

I44+A44+B44+ C44=Fwave

Where I44 is the roll inertia term, A44is the added mass term,B44is the damping term and C44 is the restoring term respectively. Then, Fwave is the roll moment induced by encounter wave motion. By using the obtained equation, the whole component (Both ASV and CMG) roll motion equation also can be expressed as follows;

I44+A44+B44+ C44=Fwave- g

I44+A44+B44+ C44=Fwave- Iyycos

From this equation, by increasing and decreasing precession angle speed,the roll motion of ASV, the effect of stabilization, is changeable. So, with controlling precession angle actively against roll motion of ASV, more safe motion of ASV will be acquired about the roll motion. But in this paper, the precession in the gimbal axis is passive which is used to generate energy so that the precession motion cannot be controlled. So, the effect of roll motion stabilization without control in various waves will be discussed.

Wave energy harvesting calculation

To calculate the generated power by precession motion of CMG in ASV induced by wave, It is considered that the small generator is installed in the precession axis of CMG (Zf) as shown in the figure 9. The operation torque of the generator is gyroscopic power of the system caused by flywheel angular momentum and wave induced ASV roll motion.

The from the equation (8), the torque (e)of precession axis (Zf) can be obtained as below;

e=Izz-Iyysin +Iyycos

Where Iyy, Izz are the mass moment of inertia about the flywheel fixed frame (XF,YF,ZF). The torque (e) represents the torque about the precession axis generated by CMG unit. In addition, ,, represent the flywheel angular velocity, pitch motion angular velocity of ASV and roll motion angular velocity of ASV. From this equation(14), if the ASV experiences have roll and pitch motion by induced wave force and has rotating flywheel inside of it, both the precession angular velocity and the precession torque (e) would be obtained. By using this, the power is possible to be generated in precession axis of flywheel (Zf).

Measuring the moment of inertial

As discussed in the background theory, Inertial moment and radius of gyration are important properties for deciding roll motion of a ship. In this paper, by using simple pendulum technique, these values are obtained. The method is mainly divided into two axes, Pitch and Roll. Practically, the correct estimation of the centre position of ASV is so tricky that it is assumed that the centre position of longitudinal, transversal.

Figure 10 represents the method to measure the pitch and roll moment of gyration. The model is firstly suspended in a light frame so that the centre of gravity is h meters below a pivot point. After installing, the whole rig is rolled through a small angle by hand, and the natural frequency T was measured. For the accuracy of the test, the natural frequency T was measured ten times in complete oscillations and the mean value was calculated.

The total roll moment of inertia of the complete rig is,

I= mmk42+ mmh2+IF

And the stiffness of the compound pendulum is

c =mmh+ mFhFg

Where mF, IF and hF represent the mass of the supporting frame, the mass moment of inertia of the supporting frame and the distance from the centre of gravity of the supporting frame to the pivot point.

By pendulum test, the natural oscillation frequency is approximately:

*=2T= cI

And the model’s radius of gyration is given by

k4=mmh+ mFhFgT242mm-h2-IFmm

And the roll moment of inertia can be written as follows

Ixx= I44= k42*mm

Following the above equation, the roll moment of inertia was obtained. Like roll-moment of inertia, yaw-moment of inertia can be determined as follows; Figure 11 represents the set-up for measuring of the inertia by utilizing an adoption of the compound pendulum, the "bifilar suspension rig". The model was installed with two wires under the carriage. Then, the model was yawed by hand as a small angle and the natural period of oscillation T was measured. In order to increase accuracy, ten complete oscillations were conducted and the mean value was calculated as well. The moment of inertia of the wires supporting the model is negligible and the natural oscillation frequency is as follows,

*=2T= cmmk62=gh*xrk6

The model’s radius of gyration is given by

k6=Txr2gh

and the moment of inertia can be written as follows

Izz= I66= k62*mm

The Yaw and Roll direction test was investigated as above methods and the pitch direction was assumed to be the same as the Yaw direction (Lloyd, 1989).

Experimental Set-up

In this chapter the set-up of experiment will be explained. For verifying the aim of this thesis and numerical calculation of above mentioned method, the experiments were planned and tested as follows. The main feature of the whole system is to mount the CMG system in the ASV model. As the first step, ASV model is selected.

ASV model set-up

The scaled model is determined according to the representative of a proto-type of Lifeboat; Arun-class boat from the keel to the main deck. Since this paper studies the hydrodynamic properties of an autonomous surface vehicle, the scaled model is chosen by following reasons. First, it is preferable to find the scaled model as large as possible, considering the limitations imposed by the wave tank. This is due to effects, which are difficult to scale (i.e. viscous effects). Secondly, the shape of the vessel would be the symmetric structure with port and starboard. Finally, the scale ratio (λ) used is 10.2941 and is defined by this equation (23), by Lloyd’s seakeeping (Lloyd, 1989).

λ= LOAF/LOAm

The full-size boat has a length overall of 14 meters, is 4.5 meters wide and has a height of 10 meters. The displacement of Model is determined after ASV ballasting weight and pendulum test was determined. The model which has been used in this project was made by F.S.I group of the University of Southampton as Model No.56. The model was constructed of Glass Fiber Reinforced Polyester (GRP). The picture of ASV model is shown in figure 12. The final value parameter in both model and full-scale vessel are as follows table.

ASV ballasting and pendulum test

The mass of ballasting and the position of ballast weight is ballasted to obtain the correct weight, mass moment inertia, longitudinal and vertical location of the centre of gravity according to the specifications obtained in the scaled process. The position of the gyro system was located on the longitudinal axis of the boat. After that, by changing the position of two Lead-Acid batteries inside of the model, the degree of the heel is adjusted as 0 degrees. Finally, the final trims of the model boats were 0.135 / 0.134 [m] for the fore and aft.

After all ballasting weight is arranged, the pendulum tests were performed in the robotics lab, University of Southampton. Figure 13 represents the picture of the pendulum test about the Yaw axis. The pendulum test methods were followed and calculated through the guidance which was described in the section 4.1.5. The result was also shown in the table#.

During measuring the ASV ballasting test, the sailed displacement value is more significant than estimation because the towed post and connecting device’s weight are added when the experiments were performed. So, the displacement value which is shown in the Table # is the final value of experiment which means that it includes all loading material, such as gyroscopic systems, batteries, electric cables, towed post and connecting device.

Gyroscope system set-up

The required gyroscope system was constructed according to the purpose of the experiment. The control moment gyro was fixed at the position assumed the centre of gravity of the ASV with the wooden plate which was installed in the bottom of ASV. The flywheel of the gyro is free to rotate at Yf direction and the gimbal of gyro is free to rotate at Zf direction. Then, Arduino Uno and breadboard including cable wires of the system were well fixed at the wall of ASV body and were well watertight. Two 12 [V] lead-acid batteries were placed in the forward inner side of ASV. They were strongly fixed in the position which was selected in ballasting and pendulum test process with Velcro tape. Finally, there were three external USB cables which were connected with the Arduino program, Escon motor control studio program and xSENS MT manager program. All cables were long enough not to affect the motion of the ASV. After preparing the above process, the towed post and connecting device were installed at the position which is assumed as the centre of the longitudinal axis in the wooden plate of the bottom of the ASV body. The above description is the overall installation process and the completed system is shown in figure 14 below.

The detailed components of the system were presented at following sections.

Flywheel motor and speed controller

The EC motor to drive flywheel was a Maxon EC 22 brushless motor which is shown in figure 15. This motor was connected directly to a variable control voltage so that the motor speed is changeable following control voltages. In each of these experiments, the voltage fed to the motor was constant and the rpm of the flywheel motor was constant which was required by each experiment condition. The speed controller was an Escon 50/5 model which is shown in figure 16. It gave control signal to flywheel motor through the Escon studio program in PC which is shown in figure 17. So, before each experiment starts, the speed of flywheel was pre-set as required speed through the Escon studio program. Besides, the hall sensor which was installed with EC motor gave the speed information of flywheel to speed controller. By using this signal as output from the speed controller, the Arduino Uno could get the information of average flywheel speed and record it at SD card in each experiment.

Gimbal motor generator and speed controller

The DC motor was chosen to be installed at the gimbal axis of CMG flywheel. However, in this paper, the gimbal axis rotation is operated by induced flywheel momentum and roll-motion of ASV, so the gimbal DC motor was just passively used so worked as generator. The generator related to gimbal axis of CMG by using Gear which is the ratio rate 50.8: 1 as shown in the figure 18. Also, the encoder which was installed with DC generator was connected with the speed controller unit to check the average rpm of gimbal axis. The used speed controller was of the same type of the flywheel speed controller (Escon 50/5) and the output signal was connected with Arduino Uno to check the average rpm of gimbal axis and record it in SD card at each experiment. Moreover, The DC gimbal generator had power taking off system which gave harvested energy. For measuring the generated voltage, the wire from power taking off system was connected with Arduino Uno’s voltage measuring circuit which is detailed in section 4.2.4

Arduino Uno

The Arduino Uno microcontroller which has been shown in figure 19 was used for all computations, measurements via sensors and data exports regarding the CMG’s control system. All code for the Arduino Uno units was written in the Arduino IDE. All Arduino software and hardware are open sources which greatly simplified the developing process since example scripts were readily available. In this experiment, the Arduino Uno board was mainly used to communicate between all other connected sensors and the PC and to record the data at each experiment at the same time. The data which was recorded in the SD card through Arduino Uno is as follows;

- Sampling time, angular velocity and acceleration of ASV motion, Harvested instant power from gimbal DC generator, average rpm of Gyro’s flywheel and gimbal, Wave height from wave probe.

Measurement sensors

During each experiment, the wave height, vessel motions, induced voltage were recorded to collect the data in the Arduino Uno. The required sensor and data acquisition methods have been decided not only in consideration of the types of data, but also in consideration of acquiring data in an ASV’s unmanned environment.

Firstly, the two IMU, which can measure the motion of the 6 DOF, was installed in the expected centre of gravity of the model. Two types of IMU were selected to increase the accuracy of measuring and useful comparison of each sensor data. One of them is MPU 6050 which is shown in figure 20. The MPU 6050 was an embedded 3-axis MEMS gyroscope, a 3-axis MEMS accelerometer. The code to calculated 6 DOF motion to get angular velocity and acceleration was programmed through the Arduino program. The other one is xSENS MTi-100 IMU which is shown in figure 21, which was directly connected to PC. It can get and record directly angular velocity and acceleration from the roll pitch and yaw each experiment through xSENS Manager program in PC.

By using collected angular velocity from these two sensors, the amplitude of roll, which is required to be obtained in this paper, was calculated through MATLAB program. The programming code can be confirmed at Appendix B. When it comes to measuring the voltage produced from the gimbal DC generator, the voltage distribution was used to construct a circuit in which two resistors; 100K [ohm] and 10K [ohm] were connected in series to measure the voltage across the resistor. Following the same method, Wave height is also measured as a voltage from the wave probe. After recording, the wave height was calibrated from voltage to meter to know the actual wave height. In conclusion, 6 DOF motions through the IMUs mentioned above, the voltage generated by the gimbal generator, flywheel rpm, gimbal rpm measured through the Escon speed controller and the wave height were all recorded through the SD card installed in the Arduino at the same time during each experiment. All relevant Arduino codes used in the above procedure are attached in Appendix B.

Power sources

To power the one flywheel EC motor and two Escon speed controller two 12 [v] Lead-Acid batteries were used as shown in the figure 22. The Arduino Uno microcontrollers and MTi-100 IMU sensor were supplied power via the PC’s USB ports.

In conclusion, the gyroscope system consists of flywheel motor, gimbal motor generator, motor speed controller, various sensors and power sources. The combined CMG system is presented in figure 23. The final wiring diagram of this system will be found in Appendix. A.

To measure the voltage which is the output of the wave probe and harvested power, a voltage divider was implemented in order to scale down the input voltage to the Arduino analog pin since the range of the voltage sensor in the Arduino is 5 [V]. There are two same divider circuits which are constructed and these recorded the measured voltage in SD card of Arduino same time with other measured values. The voltage divider circuit consisted of above figure 24 and the voltage across the R2 resistor can be calculated;

Since the Arduino Uno analog pins can only handle 0 ~ 5 [V]. The voltage divider was then put in series with that pin set. The resistors in the implementation was set to R1 = 100k and R2 = 10k [ohm], resulting in a reduction of Udivided= 111UInput.The magnitude of resistance was chosen to be large in order to prevent damage to the resistors and Arduino Uno from large currents.

Tank testing

Wave Tank

All experiments were carried out in the towing tank at the Southampton University. The towing tank is 138 meters long, 6 meters wide and 3.5 meters deep. There was a flap-based wave generation system in one end. Controlling the wave generation had been done by a control station which is installed near the wave flap. The wave generating process is shown in the below figure 25. The wave probe was installed at next to the model to correct the same direction of encounter frequency with the model and the corrected wave height data was stored in the SD card of Arduino at the same time with other sensor values. The model was fixed to the tow post at the carriage as has been shown in figure 26, which remained at a constant distance from the wave Wflap.

Experiment plan

The ASV was attached to the tow post of the carriage and constrained the motion in yaw, pitch, surge and sway. The installed position of the tow post connection device was secured approximately at the longitudinal centre of gravity(LCG) of the ASV model. The ASV model was constrained in beams in the seas during the whole experiment period.

To verify the numerical calculation about wave energy harvesting and roll motion stabilization effect of ASV, totally 20 tests were developed, each of which was run with the flywheel having 4 different rotational speeds, 5 frequencies of the wave and 1 amplitude of the wave. Each test has the same condition about ASV’s loading and displacement. In addition, each test was repeated one more time to collect more accurate data and to prevent the missing of data. Table 3 shows the parameters of the wave used for each rotational speed of the flywheel.

Results

In this result section, the obtained results from regular wave tests in towing tank are presented as follows;

First, in the harvested energy section, the instantaneous voltage through different flywheel speed and different frequencies of the wave is presented.

In the second section, the roll-motion properties of ASV are presented in different frequencies of the wave with the zero-speed condition of the flywheel.

Third, the roll-motion stabilization effect is presented over a range of frequencies of waves and flywheel speed.

The experiments were performed following a conventional sea keeping methodology, in towing tank, the Boldrewood campus, University of Southampton 28th Sep.2018 ~ 30th Sep.2018.

The experiments were performed as the plan in table #. For clear comparison of data, all results were based that the ASV encountered ten waves after the flywheel speed was enough to reach the required values.

Wave motion data

First, the examples of wave motion were presented in figure 27. The measured wave height by wave probe had been recorded in Arduino Uno as a voltage form. Then, all the measured voltage values are found to be linearly proportional to the wave height, so they were all converted to wave height [m]. The initial setting was such that the wave probe showed 1 [V] and 6 [V] at 0.3 [m] intervals. Therefore, the wave height change of 0.06 [m] per 1 [V] can be measured in this section and the following first order linear equation (28) is obtained between voltage and wave height where the towing tank water was calm, the wave height was 0 [m] which showed 3.7465 [V] as average voltage value.

wave height m=0.06*measured voltage V-0.22479

The below figure 27 shows that the measured voltage and converted valueof wave height in the calm water condition. It seems that the measured voltage values (average: 3.7465 [V]) were well exchanged as average 0 [m].

There are examples of measured wave height with time histories during the experiments in figure 28. Also, there are examples of measured wave height with fast Fourier transforms (FFT) frequency ranges during the experiments in figure 28. In figure 27, it is shown that the wave motion was well generated in the required wave height and figure 28 shows that the desired wave amplitude can be obtained at the desired wave frequency. From this, the wave generator and wave probe used in this experiment worked well and that the required wave height 0.06 [m] and various frequencies [0.4, 0.5, 0.6, 0.8, 0.9 Hz] are well generated. Of course, there are some points which were a little bit lower or higher than the required value because of the sampling rate of Arduino Uno. However, there is no problem in the process of checking wave generator motion and wave motion.

Flywheel operation data

During each experiment, there are four different ranges of flywheel rpm (0, 1000, 4000, 8000). The average rotational speed was recorded in Arduino Uno using the output signal of the Escon speed controller as described above to ensure that the flywheel reached the required rpm. Above figure 29 represents examples of measured flywheel rotation speed during the experiments. The yellow line, the red line and the blue line were measured from 8000 rpm with 0.4 [Hz] frequency, 4000 rpm with 0.4 [Hz] and 1000 rpm with 0.4 [Hz] frequency respectively. Three graph shows that the flywheel rotation is controlled using the speed measured by the Hall sensor installed through the Escon controller and seems to maintain the required rotational speed well. Also, at 8000 rpm and 4000 rpm, feedback control can be used to control rpm at the constant speed during the experiment.

ASV roll motion data

For checking the ASV roll degree amplitude, measured angular velocity data were integrated through the MATLAB program. The angular velocity had been measured by the MTi-100 IMU sensor and the MPU-6050 sensor for each experiment. Because both sensors have different sampling frequencies (2.78 [Hz] – MPU-6050 with Arduino, 400 [Hz] – Mti-100 IMU), the measured values and calculated degree values are little different. Figure 30 represents roll-motion of ASV model which was measured through MTi-100 sensors in the range of frequency from 0.4 to 0.9 [Hz] with no operation of the CMG system. As shown in the above figure, at the measure of the 0.9 [Hz], the roll-degree amplitude is the highest among these ranges of frequency. The figure 31 shows that roll-motion of ASV model was measured through MPU-6050 sensors with Arduino Uno in the range of frequency from 0.4 to 0.9 [HZ] with no operation of the CMG system as well. This graph also has trends like those of the preceding graph, and the roll degree is also proportional to the frequency rising. However, when the data collected by the MPU 6050 is used, the error due to the low sampling rate is generated and the data passed through the low pass filter is used. As above, after recognizing the motion of the ASV when the CMG is not in operation, to provide roll-motion stabilization effect, the ASV roll motion was investigated by changing the flywheel rpm of the CMG system. For easy comparison of the stabilization effect, the response amplitude operator (RAO) was constructed over the frequency range and rotation speed of the flywheel. The results were also based on ten wave encounters and the values of roll degree used are the mean value of the ASV roll motion during ten wave encounters.

Figure 32 and 33 represents the mean value of roll motion response of ASV in regular wave condition measured by the MPU 6050 IMU sensor and xSENS Mti-100 IMU sensor. The measured values and calculated degree values are little differences, but the trend is similar. When the gyro system is turned off, the corresponding roll amplitude also increases with increasing frequency. When the flywheel rpm of the gyro system was increased to 1000, it was found that there was no significant change of the values when the gyro system was not actuated. On the other hand, when the flywheel rpm was increased to 4000 and 8000, remarkable roll-motion stabilization effect was confirmed. When it comes to the mean roll amplitude degree which was measured by the xSENS Mti-100 IMU sensor, especially, at low frequencies of 0.4 Hz, 0.5 Hz and 0.6 Hz, the roll motion was affected by the vibration of the rotating flywheel than wave frequency. So the roll motion on higher rpm (4000, 8000 rpm) is more significant than that of low range rpm (0,1000 rpm). However, at a frequency of 0.8 Hz, a stabilisation effect of 80 % when operating the gyro at 4000 rpm and 82 % when operating at 8000 rpm can be confirmed comparing with that of system off. Also, at 0.9 Hz, which is the highest frequency irradiated, the roll motion stabilization effect of 82 per cent at 4000 rpm and 93 per cent at 8000 rpm was confirmed.

Of course, the values measured with the MPU 6050 IMU sensor show the same trend, but the roll motion stabilization effect is slightly less than that measured with xSENS Mti-100 IMU sensor. This is because that the sampling rate of the xSENS Mti-100 IMU sensor is higher than that of the MPU 6050 IMU sensor. In this paper, it is important to demonstrate the stabilization effect of the change in flywheel rpm, so it is notable that both graphs show the same trend. Therefore, it can be confirmed that the roll motion stabilization effect can be obtained by merely rotating the flywheel of the CMG in the ASV in the regular wave condition without installing a particular control system.

Harvested power data

As shown in figure 34, 35, 36 and 37, the generated instantaneous voltage from precession motion of gyroscope with ranges of frequencies in regular wave condition was presented according to different rpm 0, 1000,4000 and 8000 respectively. The rpm range was divided to collect the data from low ranges and middle range to high range (0,1000,4000 and 8000). At the 0 rpm, there were the gyroscopic motion of CMG and no power was generated in a whole range of frequencies. Also, when the flywheel rotated with the low range rpm (1000), the precession motion was not much in its occurrence, which is shown in figure #. The highest instantaneous voltage was generated at 4000 rpm, 0.9 Hz as is shown in figure #. in the graph, during the whole ten wave encounters, not only the precession motion occurred continuously but also it seems that the wave energy harvesting had happened continuously. One of the interesting points of these results is that when the rpm was in the high range (8000 rpm), the generated voltage was lower than that of middle range rpm (4000 rpm). This was confirmed during the test. The large torque generated by the 8000-rpm flywheel was in the same direction as the torque which was induced by the wave to the ASV during the precession at some points and no further precession was observed. This phenomenon was observed at both 4000 rpm and 8000 rpm, but at 4000 rpm, the torque generated by the flywheel was not too strong, so that when the wave-induced ASV motion was continuously applied, the precession motion was not completely fixed but continuously moving. This shows that 4000 rpm is more efficient than 8000 rpm regarding energy harvesting. In the next figures, they present how much power is stored by converting the voltage to power to get more energy.

As shown in Figure 38,39,40 and 41, the generated power by precession motion of CMG. The power was calculated by following ohm’s equation,

P=VI=V2/R, where R= 100K [ohm].

Therefore, the trend of graphs is the same as the voltage graphs in figure 34,35,36 and 37. The interesting point is that increasing the rpm on the flywheel does not increase the unconditionally produced power. Theoretically, at 8000 rpm, 0.9 [Hz], the largest and most energy should be produced, but not as shown in the graph above. Increasing the flywheel rpm results in a large momentum as has been shown in the previous section, resulting in a more significant roll motion stabilisation effect. However, in order to produce energy well, the precession motion should be well made, and the momentum which is too big sometimes interfered with the precession motion. For example, when compared at 0.9 [Hz], it can be seen that the precession is not performed at 8000 rpm, and the generated power is low. On the other hand, at 4000 rpm, the overall power is well produced, and the precession is continued. Therefore, it was confirmed that at 4000 rpm, more power is produced than 8000 rpm.

Then, in order to verify whether the wave-induced frequencies of the ASV will obtain the generated power values, the FFT (Fast Fourier Transform) was performed to obtain the power values according to the frequency domain. The following figures 42 and 43 are examples of FFTs of power values produced at 4000 rpm and 8000 rpm respectively.

Through the FFT transform, the amplitude value rises at the applied frequency as has been shown in the above figures 42 and 43. Finally, to comprehensively compare the harvested energy in each condition, the applied peak power and RMS (Root Mean Square) power over the frequency range were obtained according to the flywheel rotation speeds as follows in the figure 44.

The figure 44 shows that harvested power [mW] RMS and Peak respectively. When it comes to 1000 rpm of flywheel spin rate, the harvested power is very low overall ranges of frequencies. However, at flywheel spin rates of 4000 rpm, the CMG system could not harvest high energy from frequency ranges from 0.4 to 0.6 [Hz], but generated RMS power of 0.2077 [mW] and 0.2289 [mW] at 0.8 and 0.9 [Hz], respectively. In these ranges (0.8 and 0.9 [Hz]), the power harvested by the increase in frequency was also predicted to increase linearly. However, at 8000 rpm of flywheel spin rates, only remarkable power was harvested at 0.6 [Hz], but at 0.8 and 0.9 [Hz], it was observed that the harvested energy became smaller than expected. That is, the power which was produced as the frequency increased did not increase linearly. The CMG precession motion harvests energy by operating generator works by combining the torque produced by the flywheel spin and the roll motion transmitted by the wave to the ASV. The important point here is that these two values are vectors, so directionality must be considered. At 8000 rpm, the torque produced by the flywheel spin was so large that it was very difficult to change its position when placed in a direction that would not produce a precession motion. Therefore, when using the above device in other ASV types, it is thought that the rpm range of the flywheel should be selected according to the desired ASV size, which is a better way of energy harvesting. Also, if a device capable of actively manipulating the precession motion is added, it will be possible to produce a continuous precession motion, so that more efficient energy harvesting is expected to be obtained at higher rpm range. Of course, it is expected that not only the roll-motion but also the pitch-motion can be applied by the wave in the real ocean situation so that it will show different results. In summary, the result section shows wave data, CMG flywheel motion data, roll motion stabilization effect according to ASV motion, and energy harvesting data. All of the data were obtained well by the experimental plan and confirmed through figures. Notable results were that the precession motion of the CMG flywheel did not appear consecutively as expected and appeared to be a nonlinear phenomenon that was difficult to predict. However, it was confirmed that the wave energy harvesting process through the CMG installed in the ASV for the aim of this paper could be confirmed through the experimental results and the roll motion stabilization effect can be obtained at the same time.