1. Introduction
Wind energy has the potential of playing an important role in the future electricity generation. Within the few past decades, wind turbine technology has become quite mature and is moving forward with a fast-growing rate. According to global wind energy report [
1], installed wind energy capacity in the world doubles every three years. However, as the suitable places for onshore wind turbines are filled up, wind turbine installations shift offshore. Higher wind speeds and higher use rate also increase the attraction for offshore installations. In [
2], it is predicted that a majority of wind turbines will be installed offshore in the next 20 years. However, lower availability due to unexpected failure is a major drawback for offshore wind turbines. It is more difficult to access offshore wind turbines compared to onshore turbines, so their maintenance and repair costs are higher. To overcome this problem, reliable concepts and components should be used in offshore wind turbine systems to reduce the need for regular maintenance and repair [
3].
Direct Drive (DD) and geared drive are two main power take-off concepts for wind turbines. The direct drive system offers higher reliability as the mechanical gearbox, which requires regular lubrication and maintenance, is eliminated. Furthermore, mechanical losses in the gearbox is also eliminated, which increases the potential energy yield [
4]. However, direct drive wind turbine system is challenged by the large and heavy generator, which introduces serious difficulties in both construction and installation stages, especially for multi-MW wind turbines [
5]. One solution to cope with this issue is the use of high torque density generators to reduce the generator size; however, reliability should not be sacrificed to achieve higher torque density. Hence, both torque density and reliability should be taken into account as critical factors when choosing a generator topology for offshore wind turbine applications.
To improve torque density, various topologies of electrical machines are proposed for DD wind turbine generators in the literature, which can be classified into two main categories, Permanent Magnet Synchronous Machines (PMSM) and Electrically Excited Synchronous Machines (EESM). There is great deal of flexibility in the geometry of PMSMs so that different configurations of PMSMs, including Radial Flux Permanent Magnet (RFPM), Axial Flux Permanent Magnet (AFPM), Transverse Flux Permanent Magnet (TFPM) and Permanent Magnet Vernier (PMV) or magnetically geared machines are discussed and evaluated for DD wind turbine applications [
6,
7,
8,
9,
10,
11,
12,
13]. The PMSMs promise higher efficiency and energy yield, higher torque density than the EESM owing to the absence of field winding. However, EESMs are more cost-effective than PMSMs due to the absence of rare-earth PMs. Economic, environmental and geopolitical issues like magnet price, depletion of magnet resources and concentration of rare-earth magnet resources in china are the main concerns regarding usage of rare-earth magnets in PMSM [
14]. In EESMs, maximum allowable temperature is limited by insulation class, while in PMSMs both demagnetization of PM and insulation put restriction on operating temperature, hence, EESMs are potential of operating at higher temperatures. Furthermore, output voltage can be fully controlled once the wind turbine is equipped with an EESM. In addition, field flux can be controlled at different power to minimize generator loss and maximize annual energy yield [
9]. PMSMs may be considered more reliable than EESM thanks to the absence of brushes and slip rings, while the risk of PMs demagnetization at high temperature and harsh atmospheric conditions decreases PM generators reliability.
Magnetically geared or Vernier machines have recently gained lots of attention because of their potential high torque density. Field excitation of Vernier machines can be either PM-based or electrically excited. Owing to magnetic gearing effect and flux modulation poles, mechanical speed is multiplied by the magnetic gear ratio so that a higher frequency MMF (magneto-motive force) is produced [
15]. This phenomenon makes Vernier machines a suitable option for low-speed high-torque applications. Various topologies are proposed for permanent-magnet Vernier (PMV) machine [
16,
17,
18,
19]. In [
16], a novel dual PM Vernier machine is proposed in which PMs are placed on both stator and rotor sides to increase fundamental value of the air gap flux density. In [
17], a high-power factor dual stator single rotor Vernier machine is proposed. Although both the power factor and the torque density are improved in that topology with respect to single stator option, complex mechanical structure makes it less reliable. In [
18], a magnetically geared generator called Pseudo Direct Drive (PDD) is proposed for direct drive wind turbine applications. This generator promises a high torque density and reasonable power factor; however, it is composed of two rotating parts, which make the structure complex and less robust to be used as an offshore wind turbine generator. Magnetically geared PMV machines require several times more PM compared to conventional DDPM generators, which makes them less attractive. For instance, the topology presented in [
18] requires excessive mass of PM which is about 25% of the generator active materials mass. Using field winding instead of PMs may be considered as a solution to reduce the capital cost and make Vernier machines economically feasible for large scale wind turbines. For that purpose, the proposed generator in [
18] is revised in [
20] and PMs in high speed rotor are replaced with field windings to reduce PM usage.
In this paper, a novel Electrically Excited Claw Pole Vernier (EECPV) generator is proposed for Direct-Drive wind turbine applications. This topology adapts claw pole rotor with a single loop field winding and a conventional laminated stator carrying three phase winding. Due to the magnet-free structure, unlike Permanent Magnet Synchronous Machines (PMSM), it does not suffer from high and varying permanent magnet price [
14]. Thanks to the rotor claw pole structure, the generator has a single loop field winding. As results of such a simple field winding, the proposed generator has lower field copper loss and higher efficiency compared to conventional EESM machines. Furthermore, a new structure is devised for the rotor which makes it possible to construct the rotor from laminated steel and get rid of the problems like, difficult manufacturing process and poor magnetic characteristic, associated with the soft magnetic composite (SMC) materials [
21,
22]. To achieve higher torque densities, the number of rotor claw poles and stator teeth are designed so that magnetic gear effect is created and the generator benefits from frequency multiplying effect [
23]. In addition to these advantages, the proposed generator has a reliable and robust structure for offshore wind turbines, because of its single rotating part. The paper is organized as follows. In
Section 2, the proposed generator topology is introduced and its operating principle is discussed. An analytic-numeric design procedure is developed and explained in
Section 3. In
Section 4, a simplified parametric sweep is combined with the developed design procedure and design parameters maximizing the generator torque density is obtained. The geometrical and performance parameters of the final design are presented and the mechanical structure design is included. Finally, in
Section 5, a conclusion is made based on the provided results. Enercon 7.5 MW, 12 rpm wind turbine generator (Enercon GmbH, Bremen, Germany) is chosen as the reference design and the proposed EECPV generator is designed for the same specifications.
3. Design Procedure of EECPV Generator
In this section, the developed design methodology for the proposed generator is discussed. The applied design methodology is an analytic-numeric method which is a combination of analytic equations and FE method. Since the main aim of this paper is to introduce topology of the proposed generator and show its potential sutibility for Direct-Drive wind turbine applications, the applied design procedure is kept as simple as possible. The detailed design procedure will be disscused in a separate paper. Torque expression of the EECPV generator can be expressed in terms of its main dimensions as,
where
Dg and
Lstk are bore diameter and axial length of the generator,
Bg and
q are magnetic loading and electric loading and
kw is the winding factor. It is well-known that choice of electric loading (
q) depends on the cooling type and thermal behavior of the generator.
Bg is defined as the peak pole flux divided by pole area. In conventional machines such as EESM, Bg can be analytically calculated using equivalent magnetic circuit; however, the magnetic circuit of the EECPV machine is too complex and it is difficult to obtain an accurate analytical model for calculating Bg. Therefore, 3D Finite Element (FE) analysis is used to have an accurate estimation of Bg.
According to (5),
Bg is required for calculating generator main dimensions; however, the main dimensions and geometrical parameters of the generator are required for building an FE model and calculating
Bg. Therefore, a set of design parameters is specified as a design vector which has adequate information to establish a primary FE model. Parameters of the design vector are introduced in
Table 1.
The generator geometrical parameters, such as tooth width,
tw, tooth height,
ht and claw-pole width,
wcp are calculated using (6), (7) and (8) using design vector parameters given in
Table 1.
Using the given design vector and the calculated geometrical parameters, a primary FE model is established. Magnetostatic FE simulations are used to obtain
Bg. Adaptive mesh which is controlled by the FEM software is used to for mesh generation. The stator back core height,
hsbc and the rotor back core,
hrbc are calculated using (9).
where,
Bsat is the saturation flux density of the core material.
Number of turns per phase is calculated using (10),
where,
Ea is induced EMF and
Ap is the area under a pole.
For a given design vector the design procedure is summarized in the following steps:
Choose an initial value for axial length, Lstk.
Calculate
ht,
tw,
wcp (shown in
Figure 5) analytically using (6), (7) and (8) in the terms of design vector and design constants (see
Table 2).
Establish the primary 3D FE model and obtain Bg
Calculate values of
hsbc and
hrbc (shown in
Figure 5) analytically using (9) in the terms of estimated
Bg and the design vector
Calculate number of turns per phase using (10).
Calculate Lstk using (5) for a desired torque.
Repeat this procedure until Lstk is converged.