Advanced PMSG Designs for Efficient Wind Energy Conversion

Motors and Generators
By Henda Zorgani Agrebi | 25/11/2022

Introduction

With the rapid growth in global energy needs, wind turbines have emerged as a reliable solution to face the problem of climate and ecological changes. Direct-drive permanent magnet synchronous machines (PMSMs) used in wind turbines have received significant attention thanks to the elimination of gearbox noise and their low maintenance cost. PMSMs provide high energy efficiency and a better permanent magnet yield. In addition, the absence of a circuit exciter reduces the maintenance cost of the generator’s rotor and minimizes its losses, which helps achieve better thermal effectiveness and higher performance.


Machines Used for Energy Conversion of Wind Systems

The most popular wind generators currently used in wind conversion chains are asynchronous generators (such as squirrel cage generators, and double-fed asynchronous generators) and permanent magnet synchronous generators. How to choose the right machine for your wind turbine conversion chain? 


Squirrel Cage Induction Generators (SCIGs)

SCIGs operate at a fixed speed. They are standardized and have a low maintenance cost thanks to the use of a simple electrical interface and the elimination of power electronics. They are used for high power wind turbines (above 100KW). High speed wind turbines use gearbox which affects their mass. Besides, the energy captured is not optimal because the speed is not controllable, and its range of variation is limited. The concept of fixed speed inspired wind turbine manufacturers to use alternative types of machines adequate for the variable speed such as doubly-fed induction generators (DFIGs) and permanent magnet generators (PMSGs).


Doubly-Fed Induction Generators (DFIGs)

The DFIGs have the same operating principle of the SCIGs with a wound rotor instead of squirrel cage rotor. They are innovative compared to SCIGs working with variable speed. However, the design of the DFIGs is a complex task due to the presence of brush-ring assembly and the mechanical speed multiplier. They demand regular maintenance. In addition, their connection with the network requires power electronic interface and converters which are expensive, sensitive to over-current, and present complex control strategies during the network disturbances. To overcome the drawbacks of DFIGs, wind turbine designers used PMSGs.


Permanent Magnet Synchronous Generators (PMSGs)

The PMSGs operate at a variable speed with reduced cost and volume. They can be used directly or via mechanical gearbox. The absence of the gearbox eliminates noise problems and reduce maintenance costs. The use of the permanent magnets (PMs) provides a strong magnetic field in the air gap. In fact, the absence of the excitation circuit at the level of the rotor minimizes the losses by Joule effect which improves the energy efficiency of the generator. The brush-ring assembly is eliminated, and the design of the machine becomes simpler especially for the rotor thanks to the PMs. PMSGs with a large number of poles provide considerable mechanical torques. They use power electronics interfaces, which make them economically viable and real competitors to DFIGs. The major downside of this type of machine is the high cost of PMs.


Popular Topologies of PMSGs

    The topologies of PMSGs are determined by the types of flux, rotor, and permanent magnet.

    Flux Types:

    There are two types of PMSGs which are axial flux and radial flux generators.

      Radial-flux PMSGs are preferred for a wide range of velocity variations. They are slightly more efficient and require less active materials than axial flux generators.

      Rotor Types:
      PMSGs have 2 types of rotors which are internal and external rotors .
      External rotor generators have less iron volume in the stator. Hence, they are lighter compared to internal rotor generators. Moreover, for the same number of slots and poles machines, the external rotor generator offers a better efficiency than its internal rotor counterpart. This gain is explained by the reduced magnetic losses resulting mainly from minimizing the stator iron volume. The presence of magnets leads to a construction complexity and mechanical and magnetic embrittlement at the rotor level.

      Permanent Magnet Types: 
      There are 4 types of PMs which are namely, the surface mounted magnet, the inserted permanent magnet, the flux concentration magnet, and the V-shape magnet.

      For radial flux generators, the magnets can be surface-mounted, inserted, buried, or flux concentrator. These topologies present both advantages and limitations. The surface magnets, attached directly to the periphery of the rotor, present the simplest construction. Although they offer better performances in terms of efficiency, mass power and power factor, surface permanent magnets risk coming off due to the centrifugal force applied directly to them.
      The buried magnets are difficult to insert, which causes greater torque ripples than those present in surface magnet generators. This topology is more sensitive to rotor eccentricities compared to its surface mounted counterpart. Nevertheless, such topology allows to add a salience torque to the interaction torque between the stator windings and the magnets. For this configuration, the magnets are not susceptible of being detached or demagnetized. Generators with flux-concentrating magnets represent a more complex design than that of inserted magnets but allow the use of less expensive and less temperature-sensitive ferrite magnets.

      Analytical Modelling of a Surface PMSG with Internal Rotor

      The principle of geometry determination is based on fixing the generator's electromagnetic torque on its nominal value. Once the torque is fixed, the bore radius is deduced. Then all machine dimensions are obtained by referring to an analytical model. Some requirements and hypotheses should be considered. The table below represents the generator requirements.

      Generator Specifications Value
        Rated power (W) 1100
        Rated angular speed (rpm) 382
        Pole pairs number 6
        Stator yoke induction (T) 1.4
        Slot number per pole per phase 1
        Current surface density (A.mm2) 2.7

      Specification of Surface PMSG

      According to the following equations, the geometry and the electromagnetic parameters of the machine are calculated as given in the table below.


      3D Design of the Surface PMSG

      • Bore radius


      • Machine length


      • Air gap thickness


      • Slot width 


      • Magnet Angular Width per pole


      • Magnet thickness 


      • Stator yoke thickness 


      • Induced flux


      • Back EMF


      • Induced current


       

      Geometric Parameters Dimension (mm)
        Bore radius 83.16
        Air gap 1.18
        Magnet thickness 4.49
        Active length of the machine 41.07
        Slot width 9.68
        Stator yoke thickness 11.29
        Rotor yoke thickness 11.29
        Magnet width for one pole 36.28
        Tooth width 9.68
        Slot depth 30.44
      Electromagnetic parameters Value
        Maximum vacuum flux (Wb) 0.34
       Maximum vacuum induced EMF (V) 82.5
       Magnetic flux density in the air gap (T) 0.85

      Geometric Parameters and Electromagnetic Parameters of Surface PMSG


      Results & Discussion


      EMWorks2D Validation of the Reference PMSG

      Using the virtual prototyping software EMWorks2D, we designed the indicated machine. Then, we moved to the electromagnetic analysis to determine the following parameters:

      • Magnetic flux density in the air gap
      • Field lines and the magnetic flux density in the machine
      • Induced flux
      • Back EMF
      • Cogging torque

      Magnetic Flux Density in the Air Gap


      Magnetic Flux Density of the Reference PMSG


      Flux Linkage in the Reference PMSG


      Back EMF in the Reference PMSG


      Cogging Torque of the Reference PMSG

      EMWorks2D simulation results, the analytical values, as well as error percentage are displayed in the table below.

      Parameters Analytical value Numerical value using EMW2D Error (%)
        Magnetic flux density (T) 0.846   0.875   3.3
        Maximum vacuum-induced flux (Wb) 0.34   0.344   1.3
        Maximum vacuum-induced voltage (V) 82.5 63.47 30

      Analytical vs. Numerical Results 

      The percentage values are tolerable. The error for the flux and the maximum vacuum-induced flux is low. The back EMF error is caused by the effect of the winding technique on the PMSG. The windings are wound in a trapezoidal fashion and produce a trapezoidal back EMF. In fact, it is a BLDC machine used as a generator that has such back EMF waveform as shown in the corresponding figure. 


      Optimization of the PMSG

      We suggest keeping the reference PMSG’s characteristics while modifying only the pole pair number and the magnet width, i.e., we kept the same volume of magnets used in the reference PMSG. The following table shows the parameters for all the used models.


        Pole pairs number of the PMSGP = 6 P = 8 P = 10 P = 12
        Frequency (Hz) 38.5   38.5   38.5   38.5
        Angular velocity (rpm) 382   289   231   192.5
        Model length (mm) 41.07 41.07 41.07 41.07
        Magnet width (°) 27 20.25 16.2 13.5
        Other geometric parameters  Same geometry for the 4 simulations, only the poles number and magnet width changed

      Models Parameters

      The magnetic flux density of the optimized PMSGs



      Magnetic Flux Density for PMSG with p = 8



      Magnetic Flux Density for PMSG with p = 10



      Magnetic Flux Density for PMSG with p = 12

      Flux linkage of the optimized PMSGs


      Magnetic Flux Linkage for PMSG with p = 8



      Magnetic Flux Linkage for PMSG with p = 10



      Magnetic Flux Linkage for PMSG with p = 12

      Back EMF of the optimized PMSGs


      Back EMF for PMSG with p = 8



      Back EMF for PMSG with p = 10



      Back EMF for PMSG with p = 12



      FFT of the Back EMF of the Studied PMSGs 


      Cogging Torque of the optimized PMSGs


      Cogging Torque for PMSG with p = 8


      Cogging Torque for PMSG with p = 10


      Cogging Torque for PMSG with p = 12

      Summary

      The back EMF waveform of the PMSGs is optimal for the machine with p = 10 having the lowest THD. Also, the cogging torque has the lowest value for the PMSG with p = 10. So, we obtained a PMSG (BLAC machine) with optimal performances based on the initial design of the reference PMSG (BLDC machine) by keeping the same geometry and varying the number of pole pairs.


      Characteristics Reference Machine Optimal Machine
        P 6   10
        Slots number 36   36
        Slot number per pole per phase 1   0,6
        Winding type Distributed   Concentric
        Machine type BLDC   BLAC

      Reference and Optimized PMSG Characteristics


      In this blog article, an analytical model is considered to provide a feasible sketch for the surface PMSG. The EMWorks2D simulation was used to design, analyze, and validate the electromagnetic behavior of the used model. The percentage values of the error are tolerable and justified.  In the second part, the optimization of the reference generator, having high cogging torque, provided a PMSG with optimal performances. In fact, the slot and pole number combination affect the performance of the considered model such as the Back EMF wave and the cogging torque value.

      References 
      [1]: lal Bhukya, J. Modeling and analysis of double fed induction generator for Variable Speed Wind Turbine. in 2013 International Conference on Energy Efficient Technologies for Sustainability. 2013. IEEE.
      [2]: Goudarzi, N. and W. Zhu. A review of the development of wind turbine generators across the world. in ASME International Mechanical Engineering Congress and Exposition. 2012. American Society of Mechanical Engineers.
      [3]: Mirecki, A, Comparative study of energy conversion chains dedicated to a small wind turbine. 2005.
      [4]: Bazzo, T.d.P.M., et al., Multiphysics design optimization of a permanent magnet synchronous generator. 2017. 64(12): p. 9815-9823.
      [5]: Tran, D.H., Optimal Integrated Design of a "passive" wind turbine chain: robustness analysis, experimental validation. 2010, National Polytechnic Institute of Toulouse-INPT.
      [6]: Zhu, Z.Q. and Howe, D. (2000) Influence of design parameters on cogging torque in permanent magnet machines. IEEE Transactions on Energy Conversion, 15 (4). pp. 407-412. ISSN 0885-8969