Novel Technique for Power Generation in HAWP Systems

Novel Technique for Power Generation in HAWP Systems

A. Sheryl Arulini, Dept. of EEE, BIE, Anna University

Sardar Patel Road, Guindy, Chennai, Tamil Nadu 600025

Abstract This paper presents a power generation technique using high altitude wind power generating system buoyed by a aerostat filled with light gas by which electrical energy is extracted with the help of high-altitude streamlined wind. The best generation and transmission mechanisms that provide the right power-to-weight (P/W) ratio and efficiency of the overall system are examined. The differences in weight and total losses with deviations in generation voltage and pole-pair number (frequency) of the permanent magnet synchronous generator have been examined. Also, the design of the tether that transports electrical power to the ground-based station is presented. To find the optimal weight of the tether, AC and DC transmission mechanisms using conductors that use aluminum/copper are studied and compared. It is found that aluminum conductor offers better P/W ratio than using the copper conductor. By means of the detailed analysis of generation and transmission mechanisms, it is determined that the optimal electrical power architecture is medium voltage (MV) AC generation and also transmission. It reveals better P/W ratio and efficiency in contrast with low-voltage AC generation and MV DC transmission. The designated electrical design simplifies the electric system by transporting the power electronic converter from the aerial unit to the ground base station and thus the overall P/W ratio is increased by a factor of 7% approximately.

Index Terms AC Transmission, Dc Transmission, High- Altitude Wind Power (HAWP), Low Voltage (LV), Medium Voltage (MV), Permanent Magnet Synchronous Generator (PMSG), Power-To- Weight (P/W) Ratio, Tether.

1. INTRODUCTION

   Solar and wind have emerged as two major sources of renewable energy in the last two decades [1]. Solar power generating system has a lower power density (150250 W/m2

   ) as compared with the power density of conventional ther- mal power generating system (10001200 W/m2 ). Whereas a conventional wind power generating system requires huge civil constructions and suffers from low capacity factor (30%35%) [2] (capacity factor is defined as the ratio of actual output energy over a period of time to potential output energy, if it were possible for it to operate at the rated power indefinitely) [3]. Due to these reasons, the penetration of renewable energy sources have not significantly increased in present power market [1]. However, true potential of wind power could be extracted using high-altitude wind. The speed of wind increases with the increase in the altitude from the ground surface [4], as expressed in (1). In addition, at higher altitudes, the wind flow is streamlined and consistent in

   nature. Since the wind power is proportional to the cube of the wind speed and directly proportional to the turbine area, AT , as mentioned in (2), a large amount of electrical power can be extracted with reduced turbine size.

   (1)

   (2)

   where a is density of air, Prated is rated power of high-altitude wind power (HAWP) generating system, AT is swept area of rotor blade in m2 , 0 is the known velocity of wind in meter per second at earth surface, (h) is the speed of wind in meter per second at an altitude h in meter above sea level, 0 is the known wind speed in meter per

   second at a known altitude h0 above sea level in meter, CP () is the coefficient of power extraction by the turbine, and is the Hellmans coefficient of the surface that depends on the terrain.

   Various concepts of harnessing HAWP have been explained in [5][10]. HAWP generation system based on the motor generator concept is discussed in [5] and [6]. The turbine rotor acts as a propeller and electrical machine acts as a motor to lift the complete airborne system. When the airborne system attains the desired altitude, where high- speed wind flows, and electrical machine operates as a generator and extracts the cross wind power. Consequently, the extracted wind power is sent back to the ground station using tether cables. The concept of an air- borne wind turbine (AWT)-cum-generator supported by buoyancy provided by light gas filled blimp/aerostat is framed in

   [7] and [8]. At high altitude above the earths surface, a stationary AWT extracts wind power and sends it to the ground using suitable power conversion and transmission mechanisms. The airborne unit can be actuated to move up down and sideways to orient itself in the direction of wind for maximum power extraction. This concept of generating wind power without using cross wind can generate a power of two to three times higher than the power generated by the conventional wind turbine [9]. The electric system for generating HAWP consists of generation, conversion, and transmission system in the airborne unit. This paper focuses on determining the optimal solution for power generation,

   conversion, and transmission methods for a HAWP generating system sup- ported by light filled blimp/aerostat.

   Fig.1.Conceptualized and animated figure of blimp supported HAWP generating system developed by Altaeros Energy [11].

   HAWP generating system supported by blimp, as shown in Fig. 1, can be deployed easily without any grid connection. Hence, this system can be used for remote power supply, emergency power requirements, as well as for grid- connected power supply [8], [9].

   The conceptualization and optimization of the designed electric power system for AWT based on generatormotor concept is carried out in [5] and [12]. In [5], the airborne wind power generating system consists of eight generators that work in both generator and motor modes and the complete system is rated for 100 kW. The generation-side and the transmission-side voltage levels are determined for rated 100-kW application. In [13], multi objective optimization of brushless dc motor has been accomplished for solar airplane application. Optimization of permanent magnet synchronous generator (PMSG) for hydraulic lifting system is presented in [14]. The complete design and optimization of the generator for direct-driven wind turbine has been discussed in [15]. The size and weight of direct- driven wind power system has been assessed in [16] for scaling up the power level. Design and optimization of tether for dc transmission have been conducted in [5], but ac transmission mechanism is not included in its study. However, the study of various mechanisms of HAWP generation and transmission system to find the optimal solution for designing a complete electric system has not been investigated yet. This paper finds the optimal solution for generation and transmission of power and decides the requirement of power electronic converters for the airborne unit.

   HAWP generating system, which floats at an altitude of 1 km from ground surface, of 10-, 50-, and 100-kW power ratings are considered in this paper. The turbine rotor radii and rotational speeds are determined so as to obtain optimal tip speed ratio (TSR) for maximum power extraction. Furthermore, this paper evaluates the weight, efficiency, and power-to-weight (P/W) ratio of the generator with respect to generation voltage and frequency. At different power levels, optimal generation voltage and frequency are calculated to attain a better compromise between P/W ratio and efficiency of the generator. A tether cable, which serves as a power cable and provides mechanical strength, has been designed for both dc and ac transmissions. In addition, a comparison between ac and dc transmisson systems is performed with respect to P/W ratio and overall weight of the tether. Optimal transmission voltage at the

   desired transmission efficiency and weight of tether, for various power levels, are calculated and presented in this paper. Depending on the power generation and transmission mechanisms, two different electric power architectures for harvesting HAWP have been proposed and compared. In [5], an electrical system is designed for flying

   Fig. 2. Variation of wind power extraction coefficient with respect to

   TSR.

   Electric generator to harvest HAWP and it had an overall efficiency >90% at rated power. In this paper, an electrical architecture for blimp supported HAWP generating system is designed for minimum weight of the overall airborne system with an overall system efficiency >90% at rated power.

   Section II calculates optimal rotational speed of rotor to extract maximum power for three different power levels. The optimal rotational speed is used to calculate the optimal number of poles of the generator in Section III. The section also elaborates on the selection and optimization of generation mechanism. The design of electromechanical tether is presented in Section IV. In addition, the determination of optimal transmission voltage is carried out in Section IV. Comparison between two proposed electric systems is explained in Section V.

2. OPTIMAL TSR FOR MAXIMUM POWER EXTRACTION

   The wind speed at an altitude h, measured in meter, above earth surface is given by (1). Considering wind speed of 10 m/s at the ground level, the wind speed (h) calculated at an altitude of 1 km is 24 m/s (surface roughness , 0.3). The power extracted by the AWT as a function of high-altitude wind speed, turbine area, and power extraction coefficient is conveyed by (2). Power extraction coefficient, C P, is not a constant parameter and varies with TSR, (maximum limit of 59%). The power extraction coefficient as a function of TSR, , is given in [17].

   (3)

   Fig. 2 shows the variation of power extraction coefficient with respect to variation in TSR. For the proposed HAWP systems, optimum power extraction coefficient is found to be

   0.442 at a TSR of 7.1. High rotational speed of the rotor may lead to lessen the power extraction if optimal TSR is not matched. For the rated power, the optimal turbine rotational speed can be obtained using

   TABLE I

   (4)

   cooling setup have been excluded from the calculation). The weight of the machine can be expressed as

   Mgen = Mcu + Mins + Mpm + Mrot + Msta (5)

   OPT IM A L ROTAT I O NA L SP EED OF TURBINE ROTOR FOR MAXIMUM POWER EXTRACTION

   Power Level (kW)	Rotor Radius (m)	Optimal Turbine Rotational Speed (rpm)
   10	0.98	1690
   50	2.2	755
   100	3.1	535

   Where R is rotor diameter in meter and wopt is optimal turbine rotational speed in radian per second.

   For three different power levels, the desired optimal turbine rotational speed is listed in Table I. Three different optimal rotational speeds of the turbine are used to calculate the optimal generation frequencies in Section

3. The use of gearbox in the drive train is eliminated due to sufficient rotor speed to drive the generator.

III. POWER GENERATION SYSTEM FOR HAWP

Existing wind power generation systems use induction genorators (IGs), doubly fed IGs, SGs, and PMSGs depending on the system requirements. As explained in [16][18], PMSG exhibits better efficiency and P/W ratio than other types of machines. IG has higher reliability (due to brushless con- figuration and rugged rotor design) than PMSG, but IG is not preferred for direct- driven wind power generating system with variable speed operation. In addition, PMSG (a brushless machine) requires less maintenance and, therefore, provides higher reliability as compared with other brushed synchronous machines. Hence, PMSG is preferred choice as an airborne electric generator for HAWP generating system.

Comparative studies of different PMSGs are presented

In [19] and [20]. Radial flux machine (magnets made up of NdFeB) exhibits better efficiency and low cost than the axial flux machine [20]. In addition, radial flux machine allows better cooling, as stated in [13]. Hence, three-phase radial flux machines with concentrated winding are used as an electric generator in HAWP generating system.

1. Weight Modelling of PMSG Generator

   Optimization of various parameters of PMSG has been carried out in [13] and [14] for several industrial applications. For a HAWP generating system, an efficient and low weight machine is required in order to reduce the weight of the airborne system. Hence, the overall weight and losses in the machine are expressed mathematically to evaluate the efficiency and P/W ratio of the machine.

   The overall weight of the machine depends on the generation voltage and operating frequency (number of poles in the machine) of the generated power. Total approximate weight of the PMSG is the sum of weight of copper, copper insulation, rotor iron, stator iron, and permanent magnets (weights of external housing and

   Where Mgen , Mcu , Mins , Mpm , Mrot , and Msta are estimated weights (in kilogram) of the generator, copper winding, insulating material, permanent magnets, rotor iron, and stator iron of the machine, respectively.

   1. Weight of Copper Used in Stator Windings: For the calculation of copper weight of the machine, total volume of copper used in the stator windings is required. Thus, the radius of the copper strand needs to be estimated to calculate the volume of copper used. The radius of the copper wire is a function of generated phase voltage, copper current density, and rated power. The radius of copper strand, rc , total volume, Vcu , and the weight, Mcu , of the copper used in the windings of the machine are given by [21].

      (6)

      (7)

      (8)

      where N is the total number of turns in each phase, r is the radius of the copper winding in meter, Vph is generated phase voltage (in volts), cu is the density of copper in kg/m3, Jcu, Vcu, and cu are current density of copper in A/m2, volume of copper in m3, and electrical rc

      =conductivity of copper in seimen/meter, respectively.

   2. Weight of Insulations Used in Stator Windings: : The winding insulation also contributes to the overall weight of the machine. The weight of insulating material depends upon the thickness of insulating material and total length of copper strand. The thickness of the winding insulation depends on the generated phase voltage of the machine. The thickness of insulating material required and the estimated weight of insulation are calculated using [9], [22]

      (9)

      (10)

      where tins is the thickness of winding insulation in meter, ins is the density in kg/m3, and S dielectric constant in volt per meter of the insulation material.

   3. Weight of Permanent Magnets: The weight of permanent magnets is another component in overall weight of the machine. It depends on the shaft speed, pole-pair numbers, and properties of magnet. The mathematical expression for estimation of the weight of permanent magnets used in the generator is expressed as [23], [24]

      The output coefficient, inner diameter, and axial length of the generator are required to calculate weight of the stator iron. The output coefficient, inner diameter, and axial length to estimate the weight of stator are given by [23], [24]

      (16)

      (11)

      Where, pm is the density of the magnet used in kg/m3, Cv is the coefficient of utilization of permanent magnet, Ns is rotational speed of generator shaft in revolution per minute, Hc is maximum coercive force of the magnet in ampere per meter, and P is total number of pole pair in the geerator.

   4. Weight of Rotor Iron: The rotor iron adds up to the overall weight of the machine. The volume of rotor iron depends mainly on the rotational speed, tangential stress, and density of iron. The volume, Vrot, and mass of the rotor iron used, Mrot, are given by [23], [24]

      (12)

      (13)

      where iron and F tan are density in kg/m3 and tangential stress in pascal of rotor, respectively.

   5. Weight of Stator Iron: Stator iron is the heaviest component in the generator. The volume, Vsta, and the weight, Msta , of the stator iron pivot mainly on the rated power of the machine are stated in [23] and [24]. In addition, maximum flux, output coefficient, winding coefficient, current density, and turns of copper windings also affect the volume of stator iron used in the machine. The expression for stator iron volume and weight can be expressed as [23], [24]

      (14)

      (15)

      where ts is the estimated thickness of stator iron in meter, Lin is the axial length, and Din is the inner diameter of the machine in meter. The ratio of inner diameter to the length of the machine is assumed to be 4:5 [23], [24].

      (17)

      where p is output coefficient, e is armature reaction factor of the generator, kw is winding coefficient (0.91), and Am is linear current density in ampere per meter of PMSG.

      The required number of turns of the copper winding deter- mines the weight of copper inside the machine. Since air-gap flux density of the machine is related to the turns of the copper required, air-gap flux density needs to be estimated to calculate the turns ratio of the copper winding [23], [24]. The air-gap flux density depends on the magnetic properties of permanent magnets and dimension of the machine. The air-gap flux density can be expressed as [23], [24]

      (18)

      where g is air-gap flux density of the machine in Wb/m2.

      The number of turns required for the machine depends on

      the flux, g , phase voltage, Vph, and operating speed, Ns , in revolution per minute and can be expressed by [23], [24]

      (19)

      Finally, from the required turns of the winding using (19), the weight of copper and insulation can be estimated by (6)(10). Similarly, the weight of permanent magnets, rotor iron, and stator iron are calculated using (11), (13), and (15). Hence, the overall weight of the machine as a function of generation voltage and number of pole pair (operating frequency) for various power ratings, of the machine can be estimated using (5).

2. Loss Modelling of the generator

   For evaluation of total losses of PMSG, losses are divided into copper loss and iron loss. Copper loss depends on the volume of copper used and iron loss depends on the maximum flux, operating frequency, and volume of the stator iron [25]. Losses of the machine are calculated as [5], [25]

   Pcu = 3 I 2 Rph = J 2 Vcu /cu (20)

   Fig. 3. Variation of the weight of the generator with respect to generation voltage for 100-kW power level.

   Fig. 4. Variation of the weight of the generator with respect to number of pole pair for 100-kW power level.

   (21)

   Ploss = Pcu + Pir (22)

   where Pcu, Pir, and Ploss are copper, iron, and total losses (in watts) of PMSG, respectively, Iph and Rph are the phase current in ampere and phase resistance in ohm of PMSG machine, respectively, Kh and Ke are hysteresis and eddy current constants of the permanent magnet, respectively, and Bmax, , s, and Vsta are maximum flux density in W/m2, Steinmetz constant, generated frequency in radian per second, and the total volume of stator iron in m3, respectively. The efficiency g (which depends on the overall losses of the machine) and P/W ratio (which depends on the overall losses and weight of the machine) for PMSG are expressed as

   (23)

   (24)

   Using (23) and (24), the variation of generators efficiency and P/W ratio with respect to the variations of generated phase voltage and number of pole pair (operating frequency) at different power levels can be assessed.

3. Optimal generation voltage and Operating frequency of PMSG

The constraints taken into account to find the optimal operating point for the generator are as follows:

Fig. 5. Weight variation of 100-kW generator with respect to generating voltage and number of pole pair.

1. generator efficiency should be >95% that allows sufficient forced air cooling of the generator without thermal runaway.

2. Total number of pole pairs should not exceed 30 [5]. Higher number of pole pair increases complexity in manufacturing and cost of PMSG. In this section, generators weight, losses, and P/W ratio are evaluated against the variation of pole-pair number and generation voltage using [5]-[24]. The major equations involved for the generation of plots are as follows:

1) equations (5), (8), (10), (11), (13), and (15) are

involved

for evaluating weight;

1. equations (20)(22) are involved for evaluating losses;

2. equations (5) and (22)(24) are involved for evaluating P/W ratio.

The total weight of the generator is mainly decided by the generating voltage and the operating frequency (pole pairs). The additional requirements of the insulations with an increase in the generation voltage increase the weight of the generator. However, increment in the weight of the generator due to increase in the generation voltage level is not significant

<5000 V, as shown in Fig. 3. The percentage increase in the generator weight is 20% when generation voltage increases from 500 to 5000 V.

Fig. 4 shows the variation of generator weight with respect to the number of pole pair. Increase in the pole-pair number increases the operating frequency of the machine, which decreases the weight of the required stator iron, rotor iron, and permanent magnets of the machines. However, the rate of reduction in the weight of the generator due to increase in pole pairs is small at pole-pair number >30, as shown in Fig. 4. When pole-pair number increases from 5 to 30, the weight of generator decreases by 80%. However, the increase in pole pairs from 30 to 40 gives reduction on the generator weight by 4% only. The deviation in the generator weight with respect to pole pair and generation voltage is shown in Fig. 5. The generator exhibits the least weight at low generation voltage and high pole-pair number.

Fig. 6. Variation of the generator losses with respect to pole pairs for 100- kW power level.

Fig. 7. Variation of generator losses with respect to generation voltage for 100-kW power level.

The core loss of the generator depends on the volume of stator iron and operating frequency. Increase in the pole number of the machine increases the operating frequency, which leads to substantial increase in core loss of the machine. However, copper loss of the machine is independent of pole number of the machine. Equations (20)(22) are employed to study the loss variation of generator with respect to pole-pair number and generation voltage. The effects of pole number on core loss, copper loss, and total losses of the machine are shown in Fig. 6.

The copper loss of the machine depends on the volume of copper used in the windings, as expressed in (20). At low generation voltage, thicker copper wire is used, which has a lower number of turns. In contrast, at high generation voltage, thinner copper wire is used, but it requires a large number of turns. Hence, at any generation voltage, volume of copper used is constant, resulting in constant copper loss of the machine. The core loss of the machine has no relation with the generation voltage. Fig. 7 shows the relation of various losses with respect to generation voltage. Fig. 8 shows the total losses of the machine with respect to generation voltage and pole-pair number.

Efficiency and P/W ratio of the machine are calculated using (23) and (24). Fig. 9 shows the decreasing trend of the P/W ratio of the machine wth the increase in generation voltage. However, the rate of decrease in the P/W ratio with the increase in generating voltage is minimal up to 5000

1. The effect of pole-pair number on the P/W ratio of the machine is shown in Fig. 10. At very low pole number, P/W ratio of

   Fig. 8. Total losses of the generator as the function of generating voltage and number of pole pair for 100-kW power level.

   Fig. 9. Variation of P/W of the generator with respect to generation voltage for 100-kW power level.

   Fig. 10. Variation of P/W of the generator with respect to pole pairs for 100 kW power level.

   the machine is very low; but it increases sharply and saturates, maintaining nearly constant P/W ratio in the order of 800 W/kg at pole-pair number >30. Higher pole-pair number leads to higher operating frequency that reduces the size of rotor, stator, and permanent magnets in the machine. However, higher pole pair number increases the core loss of the machine that limits the increment in P/W ratio.

   The change in P/W ratio of the machine for three different power levels of HAWP system is shown in Figs. 1113. It i s clear t h a t l o w -voltage ( LV) m a c h i n e g i v e s b e t t e r P/W ratio than medium voltage (MV) machine. Nonetheless, the operating voltage of the machine can be noticeably increased with slight compromise in P/W ratio of the machine, as shown in Figs. 1113.

   Fig. 11. P/W ratio variation of 100-kW generator with respect to generating voltage and pole pair.

   Fig. 12. P/W ratio variation of 50-kW generator with respect to generating voltage and pole pairs.

   Table II shows operating points of HAWP generating systems without constraints. The P/W ratio depends on the total losses and the weight of the machine. Total losses are subjected to the volume of iron, volume of copper and operating frequency. The gross weight of the machine relies on the weight of iron, permanent magnets, copper, and insulating material. The optimal operating points for different power level HAWP generating systems are listed in Table II. The P/W ratio of the machine increases with the increase in the rated power up to 50 kW. However, in the range of 50 100 kW, increase in rated power does not have significant effect on P/W ratio. The operating points indicated in Table II with bold letters violate the constraints, i.e., efficiency <95% and pole-pair number >30.

   Table III shows the optimal operating voltage and frequency for three different output power levels, when generator is subjected to specified constraints. As a consequence, the P/W ratio is reduced and it does not have a linear relation with the rated power. The P/W ratio increases with increase in power rating up to 50 kW and then saturates, thus maintaining

   Fig. 13. P/W ratio variation of 10-kW generator with respect to generating voltage and pole pairs.

   TABLE II

   OPTIMAL GENERATION FOR VARIOUS POWER LEVELS HAWP SYSTEMS WITHOUT CONSTRAINTS

   Power Level	Generator Weight	P/W ratio	Overall loss	Generator
   10	22.5	414	680	92.8
   50	43	1053	4700	90.6
   100	86	1062	8700	91.3

   TABLE III

   OPTIMAL GENERATION PARAMETERS FOR VARIOUS POWER LEVELS HAWP SYSTEMS WITH CONSTRAINTS (EFFICIENCY OF 95%)

   Power Level	Generator Weight	P/W ratio	Generated frequency
   10	23	413	86
   50	45	1006	106
   100	119	800	85

   constant P/W ratio irrespective to power rating. The variation of P/W ratio with power levels, with and without constraints, is shown in Fig. 14. The detailed study of PMSG in this section leads to the following conclusions:

   1. three-phase LV generation using PMSG gives better P/W ratio for the machine;

   2. higher number of pole pair reduces the weight but increases the losses of the machine;

   3. Generation voltage can be significantly increased with slight compromise in P/W ratio of the machine.

Fig. 14. P/W ratio variation of the generator with respect to power rating

of the machine.

Fig. 15. Designed tether cable for (a) dc transmission and (b) ac transmission.

1. TRANSMISSION IN HAWP GENERATING SYSTEM The tether serves as power transmission cable as

   well as it provides mechanical strength that holds the

   airborne unit. In addition, the tether is used to deploy and move the blimp to the required location. Hence, it should have enough mechanical strength to support the blimp, better electrical efficiency for efficient power transmission, and should be able to withstand adverse environmental conditions.

   A tether designed for the blimp supported HAWP generating system consists of four major components:

   1. Core to provide tensile strength;

   2. Conductors to transmit power to ground base station

   3. Insulator to insulate the power cables;

   4. Outer jacket layer for physical protection.

   The use of coaxial cable forms a kink due to continuous stress and strain [5]. Therefore, tether geometry, as shown in Fig. 15(a) and (b), is used for dc and ac transmission.

   1. Tensile Strength of the Core

      Aramid (Kevlar-49) and high modulus poly- ethylene (HMPE) are two options that can be used for high tensile strength at low specific weight. Kevlar-49 has a

      Fig. 16. Decrease in conductor radius with respect to increase in the transmission voltage for Al and Cu conductors for 100-kW power level.

      Fig. 17. Increase in insulation thickness with respect to increase in the transmission voltage for 100-kW power level.

      tensile strength of 3620 Mpas at a specific weight of 1.44 and HMPE has a tensile strength of 2400 Mpas at a specific weight of 0.97. Kevlar core exhibits better strength-to-weight ratio as compared with HMPE, henceforth, Kevlar core is used. The required diameter of the core and the weight of the Kevlar core are expressed as

      (25)

      (26)

      (27)

      (28)

      (29)

      where Fx and Fy are the horizontal force and vertical force in N acting on the blimp due to wind gusts, Ab is the cross- sectional area of blimp in m2 , T is the tension acting on the tether cable, a is the intended upward acceleration of the blimp in m/s2 , me is the total mass of nongaseous items in the blimp in kilogram, m g is the mass of gas inside blimp in kilogram, g is the density of filled gas in kg/m3, and T is the tension on the tether in N . Ts , Kev , dKev, and WKev

      are the tensile strength in Pa , density in kg/m3, diameter in meter, and the overall weight in kilogram of Kevlar core used.

   2. Design of Transmission Cable

      Conductor cables are either made up of aluminum or copper material, which are insulated using cross linked

      Fig. 18. Variation in P/W ratio of tether with respect to transmission efficiency and transmission voltage for dc transmission with Al/Cu conductor for 100-kW power level.

      Polyethene. Extra protection for the cables is provided with outer jacket layer made up of elastomer and synthetic fibers [5]. Transmission voltage determines the weight of the conductor. Higher transmission voltage reduces the conductor volume, as shown in Fig. 16. However, higher transmission voltage requires thicker insulation, as shown in Fig. 17. Therefore, optimal transmission voltage needs to be determined that gives a better P/W ratio and desirable transmission efficiency for both ac and dc transmission system.

      Te radius of the conductor cable, thickness of insulation, weight of single cable, and overall weight of tether for dc transmission are expressed in [8], [9] and [22]

      (30)

      (31)

      (32)

      (33)

      Similarly, for ac transmission radius of the conductor cable, thickness of insulation, weight of single cable, and over- all weight of tether for ac transmission are expressed in [8], [9] and [22]

      (34)

      (35)

      (36)

      (37)

      where rc is conductor radius in meter, L is length of tether in meter, n is opted transmission efficiency, c is conductivity of the conductor in seimen per meter, V is transmission voltage level in volts, S is dielectric strength of insulator in volt per meter, tins is thickness of

      insulator in meter, con is density of conductor in kg/m3, in is density of insulator in kg/m3, and Wcab and Wjack are the weight of single cable and outer jacket layer.

      Using (25)(37), different transmission mechanisms using

      copper and aluminum conductors are thoroughly studied in Section IV-C.

   3. Tether Weight Optimization

      The weight of the transmission cable is determined by the required transmission efficiency and optimal transmission voltage. The variation of P/W ratio of aluminum and copper conductor tethers with respect to transmission efficiency and voltage is shown in Fig. 18. It shows that aluminum conductor exhibits a slightly better P/W ratio than that of the copper conductor for power transmission. Hence, aluminum conductor, which is cheaper than copper conductor, is a better choice for power transmission of HAWP generating system. The weight of single cable and the optimal transmission voltage is less

      in the case of ac transmission system than dc transmission system, as shown in Fig. 19. DC transmission system utilizes two p o w e r c a b l e s , w h e r e a s ac t r a n s m i s s i o n s y s t e m u s e s three power cables for transmission. Therefore, ac transmission system requires extra power cable with insulation and increased tether jacket volume. In consequence, dc transmission system exhibits slightly better P/W ratio and low weight than ac transmission system, as shown in Fig. 20.

      The conductors inside the tether cable have polyethene insulation and elastomer external jacket for physical protection. In addition, no internal cooling mechanism is present for the power cables. Therefore, transmission efficiency >97% is preferred in HAWP generating system [8], [9] that allows sufficient cooling of the power cable to prevent thermal meltdown. The optimal transmission efficiency at the best P/W ratio of tether is 88%; however, optimization of tether cable is carried out at an efficiency >97%. The P/W ratio of tether is compromised to allow for sufficient temperature rise in tether cable. Tables IV and V depict the comparison between aluminum and copper conductor in terms of P/W ratio, tether weight, and optimal transmission voltage at different transmission efficiencies, for ac and dc transmission systems. From these tables, it can be concluded that the use of aluminum conductor gives slightly. Better P/W ratio than copper and DC transmission shows slight edge over ac transmission in terms of P/W ratio but at higher transmission voltage. From the detailed study of tether optimization, it can be concluded that:

      1. Aluminum conductor exhibits better P/W ratio than copper conductor

      2. DC transmission system exhibits better P/W ratio than AC transmission system;

      3. optimal transmission voltage for dc transmission system is higher than ac transmission system.

   Fig. 19. Variation in tether weight with respect to transmission voltage for dc/ac transmission with aluminum conductor for 100-kW power level.

   Fig. 20. Variation in P/W ratio of tether with respect to transmission efficiency and transmission voltage for dc/ac transmission with aluminum conductor for 100-kW power level.

   Table Iv

   Op tImAL Gener atION Voltag e And TEther WEi Ght F OR Aluminum And CoPp e R Co nductor F Or Ac TRans M i S SION

   F oR 100-Kw Pow ER Level

   Transmission efficiency	Optimal Voltage	Tether weight		P/W ratio
   		Al	Cu	Al	Cu
   97	4400	148	160	655	605
   98	4800	160	175	608	560
   98.5	5200	172	189	572	520

   Fig. 21. Electric system of HAWP system with LV-ac generation and MV- dc transmission.

   Table V

   Op tImAL Ge ner at ION Vo ltage And TEt her WEi Ght F OR Alu mi Num And CoPp eR Co nductoR F Or Dc TRan s Mi SsiOn

   F oR 100-Kw Power Level

   Transmission efficiency	Optimal Voltage	Tether Weight Al Cu		P/W ratio Al Cu
   97	6800	140	151	689	639
   98	7600	154	167	638	586
   98.5	8100	162	178	605	552

   Table Vi

   Op tImAL Gener atION Voltag e And TEther WEi Ght F OR Alumi Num And CoPp e R Co nducto R F Or Ac TRans M i S SION

   F oR 100-Kw PowER LEvel

   DC (Transmission efficiency 98.5%)
   Power level	Optimal Transmission Voltage	Tether Weight	P/W ratio
   10	4600	83	117
   50	6800	129	380
   100	8100	162	572

2. SELECTION OF OPTIMAL ELECTRIC SYSTEM FOR HAWP GENERATING SYSTEM

   As calculated in earlier sections, three-phase LV ac (LV- ac) generation gives maximum P/W ratio of the generation system and MV dc (MV-dc) transmission gives maximum P/W ratio for transmission system. However, an airborne power electronic conversion (PEC) system is required to transform LV- ac power into MV-dc power. This additional PEC increases the weight of the airborne system and reduces the overall efficiency of the system. Based on the studies performed in the previous sections, the proposed electrical architectures are:

   1. Electrical syste m with LV-ac generation and MV- dc transmission;

   2. Electrical system with MV ac (MV-ac) generation and MV-ac transmission.

   1. Electrical System With LV-AC Generation and MV-DC Transmission

      The schematic of electric system of the HAWP system with LV-ac generation and MV-dc transmission is shown in Fig.

      21. It consists of an additional airborne PEC that converts LV-ac power into MV-dc power efficiently. This type of electrical

      Fig. 23. Increment in the weight of generator with respect to generating voltage for 10-, 50-, and 100-kW power level.

      Architecture is explained in [5] and [8]. In order to maintain the overall system efficiency >90%, transmission efficiency

      Table Ix

      Optimal Transmission Parameters For Ac Transmission With Aluminum Conductor When Transmission Voltage Matches Generation Voltage

      AC transmission efficiency of (97%)
      Power Level	Optimal Voltage	Tether Weight	Decrement in tether weight	P/W ratio
      10/p>	2500	77	6	126
      50	3700	118	11	410
      100	4400	148	14	655

      has to be maintained at 98.5% to compensate for the losses in PECs used in the airborne system. The optimal transmission parameters at a transmission efficiency of 98.5% are listed in Table VI. This electrical system uses LV machine with better P/W ratio than MV machine. However, the use of airborne PEC and requirement of higher transmission efficiency reduces the overall P/W ratio of electrical system. The estimated P/W ratio and efficiency for airborne PECs are 4 kW/kg and 96%, respectively, [5]. Table VII presents the overall efficiency, weight, and P/W ratio of complete airborne electrical system that includes power generation, power conversion, and power transmission for HAWP generating systems.

   2. Electrical System With MV-AC Generation and MV-AC Transmission

   The proposed electric system with MV-ac generation and MV-ac transmission is shown in Fig. 22. The proposed

   Fig. 22. Electric system of HAWP system with three-phase MV- ac transmission.

   System does not use any PEC in the airborne system. As stated in the previous sections, the generator losses remain constant with increase in the generation voltage.

   Generation voltage of the machine can be increased significantly with a slight increase in the weight of the machine without any increase in total losses. Hence, MV generation can be used in HAWP generation system with a slight compromise in P/W ratio of the machine. The degree of change in the weight of the generator with respect to the change in generation voltage at various power levels is shown in Fig. 23. The generation voltage of the machine is increased to match the optimal transmission voltage, which does not surge the weight of generator substantially, as presented in Table VIII. In addition, the generation-side voltages and corresponding increase in weight of the generators for various power levels have been demonstrated in the same table. For an ac transmission system, transmission efficiency can be reduced up to 97%. This increases the P/W ratio and decreases the weight of tether wire as compared with that required in dc transmission system (at 98.5% efficiency). In addition, it reduces the optimal transmission voltage to a lower value that matches the generation voltage. The increment in the machine weight due to MV-ac generation is compensated by the decrement in tether weight. The reduced optimal transmission voltage and weight of the tether at 97% transmission efficiency are shown in Table IX. The overall efficiency and the P/W ratio of the system at different power levels are presented in Table X. The efficiency and P/W ratio for both the proposed systems can be compared from Tables VII and X. The MV-ac generation with MV-ac transmission exhibits better P/W ratio and efficiency than LV-ac generation with MV-dc transmission. The P/W ratio and efficiency are increased by 7% and 2%, respectively, using MV-ac generation as well as transmission instead of using LV-dc generation and MV-dc transmission, as m e n t i o n e d in T a b l e s V I I a n d X . Hence, the use of MV-ac generation and MV-ac transmission in HAWP generating system provides an edge over previous system with simple electrical architecture without PEC in the airborne unit. Therefore, the need for complex design and control of airborne PEC is not required. For electrical isolation of HAWP system with the grid, isolated PECs can be used in the ground-based station.

3. CONCLUSION

HAWP generating system requires optimal electrical power architecture for power generation and transmission mechanism. The reduction in the weight of complete airborne system reduces the volume of light gas (H2 or He) required for buoyancy. Optimal turbine rotational speed is determined for maximum extraction of electrical power from high- altitude wind. The turbine shaft is directly connected to the generator shaft without using gearbox. Optimal rotational speed is used to calculate optimal operating frequency of the generator. Radial flux PMSG that exhibits better P/W ratio than for other machines is used as the airborne generator. The variations of weight and losses of PMSG at various

generating voltage and frequency (pole-pair numbers) are evaluated. Increase in generation voltage decreases the P/W ratio of the machine. An increase in pole-pair number reduces the size of generator but increases the core loss of the machine that limits the P/W ratio. Similarly, tether for ac and dc transmission are proposed and compared in terms of overall weight. For desired transmission efficiency, optimal transmission voltage is determined and corresponding tether weight is calculated. It is found that aluminum conductor yields better P/W ratio than copper; hence, it is used as power transmitting cable for HAWP. Two different electrical systems are proposed:

1. With MV-ac generation and MV-ac transmission and 2) with LV-ac generation and MV-dc transmission. The PMSG exhibits better P/W ratio at LV generation and optimal pole pair number (as calculated in this paper). Since the impact of generated voltage level on P/W ratio is not that substantial, the generation voltage is increased to match the transmission voltage; the system yields simple electrical power architecture without PEC in airborne unit. Thus, the optimal electrical power architecture for HAWP generating system consists of MV-ac generation and MV-ac transmission mechanism. This architecture gives the overall benefits on system efficiency by 2% and P/W ratio by 7% over LV-ac generation and MV-dc transmission mechanism.

   REFERENCES

   1. J. Sawin, Renewables global status report, Ren21 Steering Committee, Paris, France, Tech. Rep. REN21.2012, 2012.

   2. G. M. Masters, Renewable and Efficient Electric Power Systems. New York, NY, USA: Wiley, 2004.

   3. UMass. (Aug. 2014). UMass Amherst. [Online]. Available: http://www.umass.edu/windenergy/publications/published/community WindFact Sheets/RERL_Fact_Sheet_2a_Capacity_Factor.pdf

   4. M. L. Loyd, Crosswind kite power, J. Energy, vol. 4, no. 3, 1980, Art. ID 804075.

   5. J. W. Kolar et al., Conceptualization and multiobjective optimization of the electric system of an airborne wind turbine, IEEE J. Emerg. Sel. Topics Power Electron., vol. 1, no. 2, pp. 73103, Jun. 2013.

   6. B. W. Roberts et al., Harnessing high-altitude wind power,

      IEEE Trans. Energy Convers., vol. 22, no. 1, pp. 136144, Mar. 2007.

   7. M. Garcia-Sanz, N. White, and N. Tierno, A novel approach to airborne wind energy: Overview of the EAGLE system project, in Proc. IEEE Nat. Aerosp. Electron. Conf. (NAECON), Jul. 2011, pp. 167172.

   8. J. Adhikari, A. K. Rathore, and S. K. Panda, Harnessing high altitude wind power using light gas filled blimp, in Proc. 39th IEEE IECON, Nov. 2013, pp. 71637168.

   9. J. Adhikari and S. K. Panda, Overview of high altitude wind energy harvesting system, in Proc. 5th IEEE PESA, Dec. 2013, pp. 18.

   10. J. Adhikari, A. K. Rathore, and S. K. Panda, Modular interleaved soft-switching DC-DC converter for high-altitude wind energy application, IEEE J. Emerg. Sel. Topics Power Electron., vol. 2, no. 4, pp. 727738, Dec. 2014.

   11. (Aug. 2014). Altaeros Energies. [Online]. Available:http://www.altaerosenergies.com/wind.html.

   12. R. A. Friedemann, F. Krismer, and J. W. Kolar, Design of a minimum weight dual active bridge converter for an airborne wind turbine system, in Proc. 27th Annu. IEEE Appl. Power Electron. Conf. Expo. (APEC), Feb. 2012, pp. 509516.

   13. P. Ragot, M. Markovic, and Y. Perriard, Optimization of electric motor for a solar airplane application, IEEE Trans. Ind. Appl., vol. 42, no. 4, pp. 1053161, Jul./Aug. 2006.

   14. T. Wang and Q. Wang, Optimization design of a permanent magnet synchronous generator for a potential energy recovery system, IEEE Trans. Energy Convers., vol. 27, no. 4, pp. 856863, Dec. 2012.

   15. R. Vermaak and M. J. Kamper, Design aspects of a novel topology air- cored permanent magnet linear generator for direct drive wave energy converters, IEEE Trans. Ind. Electron., vol. 59, no. 5, pp. 21042115, May 2012.

   16. G. Shrestha, H. Polinder, and J. A. Ferreira, Scaling laws for direct drive generators in wind turbines, in Proc. IEEE Int. Electr. Mach. Drives Conf. (IEMDC), May 2009, pp. 797803.

   17. R. Melício, V. M. F. Mendes, and J. P. S. Catalão, Power con- verter topologies for wind energy conversion systems: Integrated modeling, control strategy and performance simulation, Renew. Energy, vol. 35, no. 10, pp. 21652174, 2010. [Online]. Available:http://www.sciencedirect.com/science/article/pii/S0960148 110001114

   18. H. Polinder, F. van der Pijl, G.-J. de Vilder, and P. Tavner, Comparison of direct-drive and geared generator concepts for wind turbines, in Proc. IEEE Int. Conf. Electr. Mach. Drives, May 2005, pp. 543550.

   19. J. Wang, Z. P. Xia, D. Howe, and S. A. Long, Comparative study of 3-phase permanent magnet brushless machines with concentrated,

       distributed and modular windings, in Proc. 3rd IET Int. Conf. PowerElectron., Mach. Drives, Apr. 2006, pp. 489493.

   20. A. Parviainen, M. Niemela, J. Pyrhonen, and J. Mantere, Performance comparison between low-speed axial-flux and radial-flux permanent- magnet machines including mechanical constraints, in Proc. IEEE Int. Conf. Electr. Mach. Drives, May 2005, pp. 16951702.

   21. A. Husain, Fundamentals of Electrical Engineering. New Delhi, India: Dhanpat Rai Pub., 2012.

   22. (Jan. 2014). Microwave Information Resource. [Online]. Available: http://www.microwaves101.com/encyclopedia/coax_power.cfm

   23. J. F. Gieras and M. Wing, Permanent Magnet Motor Technology: Design and Applications. New York, NY, USA: Marcel Dekker, 2002.

   24. J. Pyrhonen, T. Jokinen, and V. Hrabovcova, Main Dimensions of a Rotating Machine. New York, NY, USA: Wiley, 2008, pp. 281300. [Online]. Available: http://dx.doi.org/10.1002/9780470740095.cp

   25. C. Mi, G. R. Slemon, and R. Bonert, Modeling of iron losses of permanent-magnet synchronous motors, IEEE Trans. Ind. Appl., vol. 39, no. 3, pp. 734742, May/Jun. 2003.