Comparison of Permanent Magnet Generators for a Very Low Speed Renewable Energy Application

David G. Dorrell, Sze Song Ngu, Calum Cossar

Abstract -- This paper describes the design of a very low speed direct-drive brushless permanent magnet generator for use in a Bristol cylinder device which is used for generation of electrical energy from sea waves. It studies both air-gap wound and slotted stator types of machine. It studies the design of a scaled down machine using air-gap windings then scales the machine up to obtain a full-scale generator. A comparison is made to a slotted alternative. The performance is studied1in steady state terms. Index TermsBrushless permanent magnet generator, renewable energy, Bristol cylinder, low speed, direct drive.

HE functions of a sea wave energy device are to harness the energy of sea waves and convert the energy into useful forms of energy for domestic or industrial use. It is therefore known as wave energy converter. There are several significant reviews of wave energy devices which describe and discuss the various forms of device developed which attempt to harness sea wave energy. Basically, wave devices can be categorized into several types of device; these categories are: 1. The oscillating water column (OWC). 2. The point absorber. 3. The flap or surge device. 4. The attenuator or contouring device 5. Overtopping devices. 6. Other types that are unique and do not fall into any category above. In this paper we will study a generator suitable for a direct-drive generator in a Bristol cylinder. In this section we will put forward a brief review of some specific devices. We will then describe the Bristol cylinder. A. Wave Generator Systems In this section we will review some basic systems and their power take-off arrangement. There are no established systems and most of the systems, even those undergoing sea trials, are of a prototype nature. The OWC operates much like a wind turbine, applying the principle of wave-induced air pressurization in an enclosed chamber. The semi- air chamber has a top outlet though which reciprocating airflow and this drives the bidirectional turbine. Some examples of OWCs and component bidirectional turbines are the Limpet and the Breakwater Turbine developed by Wavegen [1][2] (which uses a Wells turbine with variable-speed induction generators). Other examples use the Denniss-Auld Turbine
D. G. Dorrell is with the University of Technology Sydney, NSW 2007, Australia (email: david.dorrell@uts.edu.au). S. S. Ngu is with the Universiti Malaysia Sarawak, 94300 Kota Samarahan, Sarawak, Malaysia (email: ssngu@feng.unimas.my) C. Cossar are with the University of Glasgow, Glasgow, G12 8LT, UK (email: c.cossar@gla.ac.uk)
1

I.

INTRODUCTION

as developed by Oceanlinx [3]. Point absorbers, or a buoys, that is small relative to the wavelength of the waves. Wave energy from all directions can be absorbed by the vertical movement of the buoy as the waves pass. These often use linear generators. Examples of point absorber wave devices are the PowerBuoy by the Ocean Power Technologies [4], the CETO by Carnegie Wave Energy [5] and the Linear Generator by Seabased [6]. Surge wave devices harness energy from the horizontal movement of the water particles in waves. Examples for this form of device are the Oyster by Aquamarine Power [7] and the WaveRoller by AW-Energy [8]. The reciprocating nature of these devices can complicate the systems. Hydraulic systems are often needed. Attenuator/Contouring devices are elongated floating devices that are parallel to the wave direction. When incident wave propagates along the device, movement within the device is generated which produces energy. Examples are the Pelamis by Pelamis Wave Power [9], the Wave Star by Wave Star Energy [10], the Anaconda by Checkmate Seaenergy [11] and others. The Pelamis used hydraulic pumps that drive constant-speed synchronous machines. In this respect, the electrical system is more conventional. Overtopping devices rely on using a funnel or barrage arrangement on the device to elevate part of the incident waves above the mean sea level to fill a raised reservoir. The seawater returns to the sea via a low-head turbine. Examples of this device are the Wave Dragon by Wave Energy AS [12] and the Multiple Stage Overtopping Device by Wave Energy [13]. The use of low head turbines means that constant speed generators can be used, again, helping to simplify the system. Some of the wave energy devices need gears and hydraulic systems to generate electricity, as already mentioned; while some are direct drive wave energy devices [14]. The wave devices each have their own advantages and disadvantages when compared to the others. In the next section the application conditions are discussed then the Bristol Cylinder is described in more detail. B. Bristol Cylinder The Bristol Cylinder wave device that is studied here does not come under any of the categories discussed above. It is a semi-submerged device that is marginally buoyant which rotates synchronously with the incident wave, given an appropriate speed control. The wave power absorption characteristic of a submerged cylinder was studied in [15] and [16], and these proved that power generation by such a device is feasible and worthy. Fig. 1 shows a basic arrangement which uses hydraulic pumps as described in [17]. As described later, water particles rotate in orbits and

978-1-4673-0141-1/12/$26.00 2012 IEEE

113

the phase difference to produce sea waves and the cylinder rotates with the water the resistance of the cylinder to the rotation should extract energy and damp the waves. The system described in [17] is really only suitable for shallow water areas where the water orbit is more elliptical. In deeper water the cylinder could be mounted on submerged towers with an armature structure at each end as shown in Fig. 2. This system was outlined in [18]. The cylinder does not rotate on its own centre but it does make an orbital motion with the waves and the armature structure will drive a generator of some description. Two are required one at each end.
Cylinder marginally buoyant and just under surface Wave direction Cylinder rotation

factor. However it may be that one linear direction or the wave periods are not scaled according to the Froude number [20].
TABLE I COMPARISON OF SMALL PROTOTYPE AND FULL SCALE DEVICE KEY
OUTLINE PARAMETERS

Parameter Target wave period Target wave height Cylinder diameter (using diameters) Cylinder length Wave power per m Device input power

Full scale device 10 s 3m 15 m --50 m 90 kW/m 4.5 MW 2.25 MW (50 % efficiency) 1790.5 kNm

Small prototype (Figs. 2 and 3) 3s 0.16 m (using ratio of diameters) 0.8 m 0.5333 1.5 m (short compared to ) 75.4 W/m 113 W 28.3 W (25 % efficiency) 13.5 Nm

Sea

Output power Rated torque (per generator)

Hydraulic pumps Hydraulic link

Fig. 1. Bristol cylinder as described in [17].
Sea wave with cylinder marginally buoyant just below water surface

W ave propagation Cylinder rotates around this orbit (but stays upright in water)

Generator rotor turns around this axis as cylinder moves in a circular orbit in water Generator maximum stator diameter Armature structure

The scale model was an early stage project to test a simple small Bristol cylinder. Fig. 3 shows this and some simulations of the operation of the direct drive machine were given in [18]. This prototype has a power-take-off that was not very successful. There was a 10:1 gearing of a small M machine and this was subject to failure because it was not designed for the application. However, this small cylinder was used for modeling and simulation. The full scale model was designed for conditions in the north Atlantic. An assessment of the projected location (or locations) needs to be used. In Fig. 4, an occurrence spectrum is shown for the north Atlantic. This seems to indicate that the target design should be at about 10 s in period and a wave height of 3 m. However, the higher period waves should not be ignored so possibly the generator should be generating up to about 17.5 s. However, as shown later the power in the waves increases as the period increases. This is an important point because as the generator gets slower the power increases. This is a very demanding application. Therefore a compromise design is very much required for this application.

Fig. 2. Basic Bristol Cylinder arrangement for deeper waters side view.

In this paper we will investigate the design of the generators in Fig. 2. The target wave conditions are given in Table I. The scaling factor is given by the ratio of the wave heights where Hl = (1) H FS Where Hl is the small scale model wave height and HFS is the full scale model. Scaling was discussed in [19]. The time can also be scaled where, from [19], the scaling factor for the wave periods is
Tl Dl = = TFS DFS

Fig. 3. Small prototype (10:1 gearing in cylinder and DC machine sizes given in Table I).

(2)

The power is proportional to LH2T therefore if we scale all three linear distances and the time then we find that the power scales by 3.5. Essentially is the linear scaling
114

7
Percentage Time of Occurance [%]

6 5 4 3 2 1 0

0-1 m 1-2m 2-3m 3-4m 4-6m

Input power [MW]

5 10 15 20 Spectral Modal Period [s] Fig. 4. Percent time of occurrence of spectra in the North Atlantic as a function of spectral modal period for different significant wave height bands [19].

total input power to the cylinder generators and the torque per generator. The rationale for these profiles is that there is constant power increase up to 10 s (since the power in a wave is proportional to the period) and then constant torque above 10 s. This is not the usual representation for a generator. A torque characteristic with respect to the frequency is required and this is illustrated in Fig. 6. This illustrates that this machine has a very narrow constant torque region before going into an extended reduced torque region. This looks similar to a field weakening region although the power in not constant. This is a demanding profile. Many modern applications such as electric and hybrid vehicles [22] have this sort of wide speed range.
3 2.5 2 1.5 1 0.5 0 0 5 10 15 Wave Period [s] 20 Power Torque Target torque 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0

Torque [M-Nm]

II. GENERATOR DESIGN The design of the small prototype generator was reported in [18]. Here we will address the design of the large generators for the full scale device. We will first address some basic sea wave theory and the load demands for the generator. The paper will then outline some simple design procedures that can be implemented to formulate a design for an air-gap wound brushless permanent magnet machine. This procedure can also be used for a slotted machine. A design is also put forward for a slotted machine. The same procedure can be used for the slotted machine with some minor differences.

Fig. 5. Generator power (total) and torque (per generator) characteristics as functions of the wave period.
3 2.5 2 1.8 Power Torque Target torque 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 0 0.2 0.4 Cylinder Frequency [Hz] 0.6

Sea Wave Energy and Loading Generally, high amplitude waves are more powerful. The wave energy can be determined by wave height, wave speed, wavelength, and water density. Using the standard equation for deep water conditions, the mean power per meter of wave front, Pwave is [19]: g 2 H 2 981.2 H 2 Pwave = = = 981.2 H 2T W/m (3) 32 f f where is the sea water density of 1025kg/m3, H is the wave height, f is the wave frequency, T is the wave period, g is gravity and 2 = = 2 f rad/s (4) T where is the wave rotational velocity. These can be used to assess the power in the waves. For the target waves of 10 s and 3 m then there is 88.3 kW/m. If the target swell waves are 10 s 3 m waves and a device that is in excess of 2 MW is required then it was assumed that the conversion rate was 50 % of the available power. If the cylinder is 50m long then this is 4.4 MW of wave energy. With a 50 % conversion rate then 2.2 MW is the rating of the electrical generation system. This is a total torque of 3581kNm or 1790.5 kNm per generator. This is at a speed of 0.1 Hz which is a very low speed. This would mean that a very high pole number is needed and this suggests that a surface magnet machine is required. However, it is not only the target speed and load that need to be considered. A load requirement over the full operating speed range (which covers a spectrum of dominant waves periods and heights) needs to be considered. The specification leads to a possible power and torque characteristics as shown in Fig. 5. This shows the

Input power [MW]

A.

Torque [M-Nm]

2 1.5 1 0.5 0

Fig. 6. Generator power (total) and torque (per generator) characteristics as functions of generator rotational frequency.

Generator Sizing and Design Slotless and Slotted The first issue to be decided is the pole number. At the target speed the electrical frequency can be set to 12.5 Hz as a low frequency choice. The generator will start to generate with 2 s waves which corresponds to 62.5 Hz. The generator speed is 0.1 Hz. Therefore: 2f 2 12.5 (5) = 250 248 P= s = 0.1 fr The reason for reducing the pole number slightly is to make the pole number divisible by 8 which adjusts the frequency to 12.4 Hz. Obviously more poles would be advantageous to raise the frequency and obtain better rate of change of flux linkage. However, there is also the dichotomy of considerations for the torque constant and considerations for the back-EMF constant. While these are essentially the same thing, as can be easily proved, the torque constant needs to be maximized to limit the size. Meantime the voltage constant may be too high because of the speed range requires as illustrated in Fig. 6.

B.

115

This pole number allows a fractional slot winding with 8 poles and 9 slot or coil-side bundles in a similar manner to [19]. This gives the number of coils or slots for 248 poles are 279. For a slotted machine this will be important to reduce cogging and torque ripple. 1) Air-gap Wound Machine Sizing Air-gap windings should be radially-narrow to enable reasonable linkage with the magnet flux. It was decided to make the magnets 30 mm thick and limit the winding layer thickness Lw to 20 mm with a 1 mm air-gap. If we allow a fill factor Sfill of 0.4 and a rms current density J in the conductor of 5 A/m2 then the surface current density J is
(6) = 5 106 20 103 0.4 = 40000 A/m Assuming that the air-gap flux density from the magnets is half the remanent flux density then the sheer stress can be calculated for the air-gap wound machine as shown in [24]: 4B J mean air-gap = ( Bn H t ) mean pk ( trap ) surface ( rms ) winding 2 (7) 1.05 4 2 40000 = 18907 N/m 2 = 2 which is half of the Torque-per-rotor-volume (TRV) [23]. This gives the approximate mean value. A TRV of 37.8 kNm/m3 is high. Typical TRVs are given in [23] and this form of machine has TRVs up to about 30 kN/m3.The torque per rotor volume is obviously related to the torque T PT TRV = = (8) 2 R L 4 2 R 3 L = 4 R J surface ( rms )
air-gap winding

= H t = J Lw S fill

1 2 R = Ht = J S fill 2 2P R = 5 106 0.5 0.4 (10) 248 3 = 12.67 10 R A/m This maintains a slot fill of 0.4. Obviously this is now a function of the rotor radius. The sheer stress will likewise be a function of the rotor radius. This time we will assume the air-gap flux density is higher and a crude approximation that it is 0.75 of Br so that 4B J mean Slotted = ( Bn H t ) mean pk ( trap ) surface ( rms ) stator 2 3 1.05 4 4 12.67 103 R = (11) 2 = 8.98 103 R N/m 2 TRV = 2 mean 248 1789.3 103 (12) = 17.97 103 R = 4 2 R 3 L = 4 R J surface ( rms )
Slotted Stator

where R is the rotor radius and L is the axial length. This assumes that the axial length is about double the pole pitch so that L = 4R/P. Therefore
PT winding 4 2TRV (9) 248 1789.5 103 3 = = 297.3 = 6.67 m 4 2 37814 and L = 0.338 m. These are very crude sizing exercise. 2) Slotted Stator Sizing We can carry out a similar sizing exercise for the slotted machine but equation (6) can be modified. This time there is more area for the winding. There are two options here. An increased area can be used with a lower current density. In [24] it was shown that when the operating frequency is low, the current density in the windings should be low to maintain good efficiency. Alternatively, the machine size can be reduced by increasing the loading since the slot area will be much larger than the equivalent area in an air-gap wound machine. A good general approximation is to assume that the tooth and slot pitch are equal. The slot depth is very variable. In low pole number machine the slot depth may be very high. However, this does produce a long path for the flux to flow round and there is added slot leakage. As a rough sizing value let us assume that the slot depth for this fractional slot winding is about half a pole pitch. This is somewhat arbitrary but inspection of many machine designs suggests this is not unreasonable. This means that the slots are then approximately square for this slot number since there is approximately one slot per pole. We can then modify (6) so that R Air-gap =
3

248 1789.3 10 = 4 625.5 = 5 m (13) 4 2 17.97 103 and L = 0.253 m. The sheer stress is 44.9 kN/m3 which is much higher since the slot depth is 63.3 mm with a slot width of 56.3 mm. This is significantly smaller than the airgap wound machine although it will exhibit cogging torque. If we maintained the same rotor diameter then the current density in the winding can be calculated by taking ratios of the winding areas. For the air-gap winding the area is 150 mm 20 mm = 3000 mm2; for the slotted winding, the slot depth is 84.5 mm and the slot pitch is 75 mm so that the area is 6562 mm2. This gives an equivalent current density of 5 3000 / 6562.5 = 2.28 A/mm2 if the current is maintained although this is likely to produce higher torque. The machine will have to exhibit certain characteristics. If the synchronous reactance is too high then there may be issues with insufficient short-circuit current in a similar manner to a synchronous machine [25]. Alternative designs to this straightforward design may be to use a permanentmagnet induction generator [26]. This may improve dynamic performance. It is also possible to consider the cage more as a damper cage (again, in a similar manner to the synchronous machine) [27] although the literature on these alternatives is still sparse The next stage is to test this using a motor design package. It is necessary to define the winding and carry out simulations using finite element analysis to adjust the basic geometrical sizes. In this case SPEED software from The University of Glasgow (now CD-Adapco) is used and this is assessed in the next section. R=
4

III. SPEED SIMULATION The machine was simulated using SPEED PC-BDC and PC-FEA. This is an analytical calculation tool with a finite element analysis (FEA) bolt-on package. FEA is necessary because the field solution for a large air-gap (due to the winding layer for the slotless design). Air-gap Winding Machine The air-gap diameter was set to 13.34 m and an axial stack length of 338 mm. The winding layout for 8 poles is shown in Fig. 7 and this has to coil sides of 10 mm A.

116

Torque [MNm] or Power [MW]

thickness with a slot fill of 0.4. This is quite high so careful fabrication, possibly with rectangular conductors, could be used. There are 4 turns per coil but there may be many parallel conductors to avoid additional eddy current loss in the wire. The magnet thickness was set to 30 mm and the winding thickness to 20 mm with a 1 mm air-gap. The magnet material was set to NeIGT 30 H with a remanent flux density Br of 1.12 at 25C and they had a pole pitch of 170 elec. PC-BDC was used to obtain a torque of 1.87 MNm and 94 % efficiency. However, the FEA simulation produced a torque that was lower than this at 1.32 MNm. Surface windings require FEA for accurate simulation.

instantaneous torque for each of the 45 steps in the I-Psi loop. The disadvantage is that there is limited winding area leading to higher copper losses than an equivalent slotted machine. It also uses a lot of magnet material. For comparison two comparable slotted designs are investigated as sized and discussed above.
0 2.5 2 1.5 1 0.5 0 0 Mechanical torque per generator Total electrical output power Target Point Efficiency 0.1 Wave frequency [Hz] 0.2 0.3 0.4 0.5 100 98 96

Efficiency [%]

94 92 90 88 86 84

Fig. 7. Winding layout for 9 poles. There are 4 turns per coil and two coilside layers.

10 20 Cylinder speed [rpm]

30

Flux linkage [mVs]

This torque was confirmed using both current fluxlinkage loops (I-Psi loop - [24]) and Maxwell stress integrals. A flux plot under loading is shown in Fig. 8. The I-Psi loop used 45 steps as the rotor was moved through two pole pitches and at each point the Maxwell stress was calculated and this showed a very low torque ripple.

Fig. 9. Machine performance with lengthened axial stack.

10000 5000 0 -1200 -800 -400 -5000 -10000 Current [A] 0 400 800 1200

To obtain the correct torque then the axial length can be increased. Therefore Ttarget 1.79 = 0.338 = 0.458 m (11) L/ = L 1.32 TFEA loaded With this new axial length the simulations were run over the cylinder frequencies given in Fig. 6. The torque per generator, total output power from both generators and efficiency are shown in Fig. 9. This shows good efficiency despite the low frequency of the back-EMF. The I-Psi loop at 6 rpm and full load is shown in Fig. 10. This is the same for all the phases. This is a very large machine with 4015 Kg of magnet. This is a major consideration when costing the machine manufacture. Bulk material is about US$40 per Kg (20/12/2011 price in news item on http://www.asianmetal.com) which could double with processing. This gives a price of US$160K to US$320K just for magnet material. The slotted machine is smaller and used less material. The advantage with this machine is that there is no cogging torque (or drag torque) and the power factor is very high (near unity). The load torque is shown in Fig. 11. This is at 6 rpm and full load. It was obtained from the FEA using two Maxwell stress integrals in the air-gap to get the

Fig. 8. Flux plot under target loading (5 A.mm in the conductors and current on q axis). (on open-circuit magnet flux density reduces to 0.63 T).

Fig. 10. Current flux linkage loop at 6 rpm and full load for air-gap winding machine.

2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 0 90 180 270 Rotation [deg elec] 360

Fig. 11. Torque from FEA at 6 rpm and full load for air-gap winding machine using Maxwell stress method

B. Slotted Machine with Same Rotor Radius and Axial Length A slotted machine with similar rotor dimensions to the slotless machine is investigated in this section. The winding was similar although the conductor cross-section of the conductors is much higher due to the increased winding window. A flux plot is shown in Fig. 12. The magnet flux density was 1 down to 0.85 T. It should be stated that one disadvantage of this arrangement is that there will be significant flux ripple across the surface of the magnets producing eddy current loss (which is not calculated in the

117

Torque [MNm]

Torque [MNm]

FEA). The I-Psi loop for the machine at 6 rpm is shown in Fig. 13. This illustrates that there is much better linkage of the magnets with the windings and in fact only half the current is required to generator rated generating torque. This produces much higher efficiency. At 6 rpm and rated power, the efficiency is 97.8 % compared to just under 96 % for the slottless design. This may not seem very large but in absolute terms of loss, this is represents 26.8 kW of loss compared to 81.3 kW of loss in the slotless design.

than the slotless machine.

2 1.5 1 0.5 0 0 90 180 270 Rotation [deg elec] 360
Fig. 15. Torque from FEA at 6 rpm and full load for slotted machine with reduced rotor radius and axial length as slotless machine.

D.
Fig. 12. Flux plot under target loading for slotted machine with same rotor radius and axial length as slotless machine (1.23 A.mm2 in the conductors and current on q axis). (Open-circuit magnet flux density is about 1 T peak).

24000 18000 12000 6000 0 -600 -400 -200 -6000 0 -12000 -18000 -24000 Current [A]
Fig. 13. Current flux linkage loop at 6 rpm and full load for slotted machine with same rotor radius and axial length as slotless machine.

Inductance comparisons The three machines have phase inductances that can be compared. These were obtained from the FEA. It can be seen that the slotless machine has the highest inductance. The reduced size machine has the lowest inductance it has both lower diameter and axial length.
TABLE II MACHINE INDUCTANCES Machine Slotless machine Slotted machine Slotted machine-reduced size Self inductance [mH] 8.4 6 3.1 Mutual inductance [mH] -6.4 -0.3 -0.19

Flux linkage [mVs]

200

400

600

E.

Torque [MNm]

Magnet eddy current loss The argument in this paper assumes that there is no eddy current loss in the surface magnets for the slotted machine. This was studied in [28]. If the magnets are assumed solid then there is considerable eddy current loss. This can be significantly reduced by segmenting the magnets and the results in [28] show that this is very necessary. This is a common solution. Another solution is to bury the magnets to form an interior permanent magnet rotor. IV. CONCLUSIONS This paper explores the loading characteristics for a very low speed generator suitable for use in a Bristol cylinder application that is currently under investigation. The work highlights the demanding load requirements where the output power increases with decreasing speed. A sizing exercise is conducted for a slotless (air-gap winding) machine and a slotted machine using simple equations. This was aimed at a sea-going generator system with is 50 m in length and 15 m in diameter. These are found to give results that are not unreasonable which can be adjusted. A comparison between the slotless and slotted machine shows that the slotted machine is more compact and operates at a higher efficiency. However, there is torque ripple in the slotted machine and there is likely to be eddy current loss in the surface of the magnets. Therefore it can be concluded that both arrangements have advantages and disadvantages. The control of this machine is studied in [29]. V. REFERENCES

1.5 1 0.5 0 0 90 180 270 Rotation [deg elec] 360

Fig. 14. Torque from FEA at 6 rpm and full load for slotted machine with same rotor radius and axial length as slotless machine.

The torque is shown in Fig. 14 and this shows an increase in the torque ripple which may affect performance. Smooth torque is advantageous in this application. C. Slotted Machine Reduced Size The slotted machine is reduced in size as discussed in the sizing section (5m rotor radius and 0.253 m axial length). The slotless machine had to be increased in axial length compared to the calculation but here the dimensions are found to be sufficient such that they do not need adjustment. The current density in the slot is still only 3.8 A/mm2. The torque characteristic at 6 rpm and full load is shown in Fig. 15 and again this shows torque ripple. The efficiency at this point is 95.6 % which is slightly higher

[1]

T. V. Heath, The Development of a Turbo-Generation System for Application in OWC Breakwaters, Proc. of the 7th European Wave and Tidal Energy Conference, 2007.

118

[2] [3]

[4]

[5] [6]

[7] [8]

[9] [10] [11] [12]

[13]

[14] [15] [16] [17] [18]

[19] [20]

Antonio F. de O. Falcao, Wave Energy Utilization: A Review of the Technologies, Renewable & Sustainable Energy Reviews, Vol. 14, No. 3, 2010. Antonio F. de O. Falcao, First Generation Wave Power Plants: Current Status and R&D Requirements, Journal of Offshore Mechanics and Arctic Engineering, Vol. 126., No. 4, pp. 384-388, 2004. K. Rhinefrank, E. B. Agamloh, A.von Jouanne, A. K. Wallace, J. Prudell, K. Kimble, J. Aills, E. Schmidt, P. Chan, B. Sweeny, and A. Schacher, Novel Ocean Energy Permanent Magnet Linear Generator Buoy, Renewable Energy, Vol. 31., No. 9, pp. 12791298, 2006. M. G. Hughes and A. D. Heap, National-scale Wave Energy Resource Assessment for Australia, Renewable Energy, Vol. 35., No. 8, pp. 1783-1791, 2010. C. Bostrom, O. Svensson, M. Rahm, E. Lejerskog, A. Savin, E. Stromstedt, J. Engstrom, H. Gravrakmo, K. Haikonen, R. Waters, D. Bjorklof, T. Johansson, J. Sundberg, and M. Leijon, Design Proposal of Electrical System for Linear Generator Wave Power Plants, Proc. of 35th Annual Conference of IEEE Industrial Electronics, pp. 4393-4398, 2009. M. S. Lagoun, A Benbouzid M. E. H. Benbouzid, Ocean Wave Converters: State of the Art and Current Status, Proc. of 2010 IEEE International Energy Conference and Exhibition, pp. 636-641, 2010. M. Folley, T. W. T. Whittaker and J. vant Hoff, The Design of Small Seabed-Mounted Bottom-Hinged Wave Energy Converters, Proc. of the 7th European Wave and Tidal Energy Conference, pp. 636-641, 2007. W. D. Jones, Update-Ocean Power Catches a Wave, IEEE Spectrum, Vol. 45, No. 7, pp. 14, 2008. Managing the Ups and Downs of Energy Generation for Wave Star, Rockwell Automation Inc., 2010. V. Heller, J. R. Chaplin, F. J. M. Farley, M. R. Hann and G. E. Hearn, Physical Model Tests of the Anaconda Wave Energy, Proc. of the European IAHR Congress 2010, 2010. Z. Zhou, W. Knapp, J. MacEnri, H. C. Sorensen, E. F, Madsen, I. Masters, and P. Igic, Permanent Magnet Generator Control and Electrical System Configuration for Wave Dragon MW Wave Energy Take-off System, Proc. of IEEE International Symposium on Industrial Electronics, pp. 1580-1585, 2008. J. P. Kofoed and E. Osaland, Crest Level Optimization of the Multi Level Overtopping Based Wave Energy Converter Seawave SlotCone Generator, Proc. of 6th European Wave and Tidal Energy Conference, pp. 243-250, 2005. N. J. Baker and M. A. Mueller, Direct Drive Wave Energy Converters, Proc. of Universities Power Engineering Conference , pp. 1-7, 2001. D. V. Evans, D. C. Jeffrey, S. H. Salter, and J. R. M. Taylor, Submerged Cylinder Wave Energy Device: Theory and Experiment Applied Ocean Research, Vol. 1, No. 1, pp. 2-12, 1979. P. Mclver and M. Mclver, Wave-Power Absorption by a Line of Submerged Horizontal Cylinders Applied Ocean Research, Vol. 17, No. 2, pp. 117-126, 1995. G. Boyle, Renewable Energy Power for a Sustainable Future, Oxford University Press, 2004. D. G. Dorrell, A. M. Knight, L. Evans and M. Popescu, Analysis and Design Techniques Applied to Hybrid Vehicle Drive Machines Assessment of Alternative IPM and Induction Motor Topologies, IEEE Trans. on Ind. Elect., 2012. S. S. Ngu, D. G. Dorrell and E. Acha, Control of a Bristol Cylinder for Wave Energy Generation, IEEE IPEC conference, Sapporo, Japan, June 2010. G. Payne, Guidance for the experimental tank testing of wave energy converters, SuperGen Marine, University of Edinburgh, 2008 (available at www.supergen-marine.org.uk).

[21] P. Boccotti, Wave mechanics for ocean engineering, Elsevier Oceanography Series, Amsterdam, 2000. [22] K. K. McCreight, A Note on the Selection of Wave Spectra for Design Evaluation, Research and Development Report, Naval Surface Warfare Center, Bethesda, Maryland 20814, CRDKNSWCHD-974-02, January 1998. [23] J. R Hendershot and T. J. E. Miller, Design of Brushless Permanent Magnet Motors, Magna Physics Pub, Oxford , Clarendon Press, 1994. [24] D. G. Dorrell, Permanent Magnet Generators for Renewable Energy Devices with Wide Speed Range and Pulsating Power Delivery, International Journal of Computer Applications in Technology (IJCAT), Vol. 36, No. 2, 2009. [25] K. W. Klontz, T. J. E. Miller, M. I. McGilp, H. Karmaker, and P. Zhong, Short-circuit analysis of permanent-magnet generators, IEEE Trans. Ind. Appl. vol. 47, no.4, pp. 1670-1680, Jul./Aug. 2011. [26] R. Vermaak, J. H. J. Potgieter and M. J. Kamper, Grid-connected VSC-HVDC wind farm system and control using permanent magnet induction generators, International Conference on Power Electronics and Drive Systems, 2009 pp 897 902. [27] K. Kurihara, Effects of cage-bars for stability of circumferentially magnetized permanent-magnet synchronous generators, International Conference on Electrical Machines and Systems (ICEMS), 2010, pp 125 1130. [28] D. G. Dorrell, S. S. Ngu and C. Cossar, Comparison of High Pole Number Ultra-low Speed Generator Designs Using Slotted and Airgap Windings, IEEE Trans on Magn., Oct. 2012. [29] S. S. Ngu, D. G. Dorrell and C. Cossar, Design and operation of very slow-speed generators for a Bristol cylinder sea wave generating device, IEEE ECCE conference, Raleigh, USA, Sept. 2012.

VI. BIOGRAPHIES
David G. Dorrell is a native of St Helens, UK, and has a BEng (Hons) degree (1988), MSc degree (1989) and PhD degree (1993). He has held lecturing positions with The Robert Gordon University and The University of Reading. He was a Senior Lecturer with The University of Glasgow, UK, for several years. In 2008 he took up a post with The University of Technology Sydney, Australia, and he was promoted to Associate Professor in 2009. He is also an Adjunct Associate Professor with The National Cheng Kung University, Taiwan. His research interests cover the design and analysis of various electrical machines and also renewable energy systems with over 150 technical publications to his name. He is a Chartered Engineer in the UK and a Fellow of the IET. He is a Senior Member of the IEEE. Sze Song Ngu received the B.Eng. (Hons) degree in Electronics Engineering from Multimedia University, Cyberjaya (2003) and M.Eng. degree in Electrical Engineering from The University of Adelaide (2004). He is working as a lecturer in the Department of Electronic Engineering at University Malaysia Sarawak (UNIMAS), Malaysia. He is currently a Ph.D. student with the School of Engineering at the University of Glasgow. His research interests include electrical machines and drive, power electronics, control system and renewable energy. Calum Cossar was born in Hamilton, UK, in 1962. He received the B.Sc (Hons) degree in electronics and electrical engineering from the University of Glasgow, Glasgow, UK, in 1983. From 1983-1988, he was with Ferranti plc, Edinburgh, UK, where he worked on the design of high-speed digital signal processing for airborne radar applications. In 1988, he joined the SPEED Laboratory, Department of Electronics & Electrical Engineering, University of Glasgow, as a Research Assistant. He subsequently became a Research Technologist in 1990 and has been primarily involved in research and development of brushless motor control strategies.

119
Powered by TCPDF (www.tcpdf.org)