Transverse Flux Machines What for

Document Sample
Transverse Flux Machines What for Powered By Docstoc
					IEEE MULTIDISCIPLINARY ENGINEERING EDUCATION MAGAZINE, VOL. 2, NO. 1, MARCH 2007                                                                  4

                    Transverse Flux Machines: What for?
                    Jo˜ o S. D. Garcia, Graduate Student Member, IEEE, Mauricio V. Ferreira da Luz,
                   Jo˜ o Pedro A. Bastos, Senior Member, IEEE, and Nelson Sadowski, Member, IEEE

  Abstract— Transverse flux machines are becoming extremely
popular in the design of critical systems. This paper presents
some of its advantages and disadvantages and illustrates its
functioning using finite element analysis.
  Index Terms— Electrical machines, Transverse Flux Machines.

                        I. I NTRODUCTION

      have just recently had a great increase in popularity. Its
high energy density values have taken the TFM into high
regards in aerospace and other critical applications. The advent
of permanent magnet technologies, which made possible the
commercial use of materials such as neodymium-iron-boron
                                                                           Fig. 1.   TFM topology with soft magnetic composite [3].
(NdFeB) or samarium-cobalt (SmCo5 ) has contributed to the
popularization of this type of machine.
   The TFM was first introduced and named by Weh in 1986
[1]. Since the TFM allows the pole number to be increased
without reducing the magnetomotive force per pole, it is
capable of producing power densities much higher than a
conventional machine. In a TFM the electromagnetic force
vector is perpendicular to the magnetic flux lines. In all
standard or longitudinal flux motors the electromagnetic force
vector is parallel to the magnetic flux lines.
   The achievable power to total machine weight ratio for
active rotor TFM ranges between 0.5–2.0 kW/kg compared
to 0.24–0.8 kW/kg for conventional machines [2].                           Fig. 2.   Double-rotor, double-sided TFM geometry with flux concentration
   Figure 1 illustrates the structure of a TFM developed at the            [1].
University of Technology Aachen, Germany [3]. The external
rotor consists of permanent magnets and flux concentrating
iron parts, which are made by soft magnetic composite. The                 distribution in the air-gap which interacts with the heteropolar
U-yoke laminations of the inner stator are bent so that the two            magnetomotive force produced by the permanent magnets’
limbs are shifted by 180 degree electrical.                                rotor excitation. The TFM behaves like a synchronous machine
   There are three concepts of TFM. The first is TFM with                   and the homopolar features allows it to avoid the limitation
active rotor where the exciting permanent magnets are placed               of the “BiL” principle of force production (force caused by a
on the rotor as shown in Figs. 2 and 3. The second concept has             magnetic field on a current carrying conductor) [4]. The TFM
permanent magnets on stator part and it is called TFM with                 has a true three-dimensional flux distribution as one can see
passive rotor. The third concept of TFM is the electrically                in Fig. 3(a), where for a better view it is shown only a part
excited reluctance motor, which has no permanent magnets as                of one phase in a linear arrangement.
shown in Fig. 4. In these three cases the TFM can be single                   The advantages of TFM topology against the classical
sided (Figs. 3 and 4) or double sided (Fig. 2). Single-sided               longitudinal concept are [6]:
machines are easier to manufacture and have better prospects                  (i) An increase of pole number does not reduce the mag-
in practical applications.                                                        netomotive force per pole;
   The TFM is basically a homopolar machine since the stator                 (ii) The magnetic flux geometry and the coil section can be
winding carries current on the circumferential direction and                      varied without compromising the dimensions of either,
produces a homopolar magnetomotive force that is modulated                        giving design freedom;
by the stator poles. It results in a heteropolar flux density                (iii) Very simple armature coils are employed and the total
                                                                                  conductor length is relatively short;
     a                                               a
  Jo˜ o S. D. Garcia, Mauricio V. Ferreira da Luz, Jo˜ o Pedro A. Bastos    (iv) The phases in a TFM are magnetically independent
and Nelson Sadowski are with the Universidade Federal de Santa Catarina,
Florian´ polis 88040–970 BRAZIL +55 48 3331–9649 (
       o                                                                          and this decoupling in structure simplifies the control
  Publisher Identification Number is 1559-7908-012007 01.                          method.

                                 1558-7908 c 2007 IEEE Education Society Student Activities Committee (EdSocSAC)
IEEE MULTIDISCIPLINARY ENGINEERING EDUCATION MAGAZINE, VOL. 2, NO. 1, MARCH 2007                                                                      5

Fig. 3. Single side TFM with flat permanent magnets rotor (a) and with
concentrating flux rotor (b) [4].                                               Fig. 5.   TFM developed by WEG, Brazil [7].

                                                                                  The periodicity boundary conditions can be directly applied
                                                                               to the moving band connection. The connection between the
                                                                               moving and the stationary regions (both being separately
                                                                               meshed), through the moving band, is similar to a periodicity
                                                                               connection (direct identification of the degrees of freedom).
                                                                               Such connection conditions have to be updated for each
                                                                               position during the movement. When the calculation domain
                                                                               angle is exceeded, the moving part must be relocated in
                                                                               front of the stationary part. If anti-periodicity conditions are
                                                                               used, all the rotor field sources signs are inverted [7]. The a-
Fig. 4. Basic structure of a transverse flux reluctance motor with claw-poles
                                                                               formulation, with a magnetic vector potential a and an electric
[5].                                                                           scalar potential v, was obtained from the weak form of the
                                                                               Amp` re equation [7], i.e.,

  The disadvantages of TFMs include low power factor and                          (ν curl a, curl a′ )Ω − (ν br , curl a′ )Ω + n × hs , a′   Γh   +
complex construction with three-dimensional magnetic fields                                       ′                   ′           ′
                                                                                         (σ∂t , a )Ωc + (σ grad v, a )Ωc − (js , a )Ωs = 0
[6]. The use of lamination in TFMs is complicated forcing
                                                                                                                              ∀ a′ ∈ Fa (Ω), (1)
the use of soft magnetic composite materials which increase
producion costs considerably.                                                  where n × hs is a constraint on the magnetic field associated
                                                                               with boundary Γh of the domain Ω and ν = 1/µ is the
                                                                               magnetic reluctivity.
                 II. F INITE E LEMENT A PPROACH
                                                                                  Fa (Ω) denotes the function space defined on Ω which
   Since the TFM’s flux pattern is quite complex, a great                       contains the basis and test functions for both vector potentials
amount of simplification is required when finding an analytical                  a and a′ . Using edge finite elements for a, a gauge condition
model. Non-linearaties due to iron-core saturation and changes                 associated with a tree of edges is generally applied [7]. (·, ·)Ω
in the operation point of permanent magnets contribute to this                 and ·, · Γ denote a volume integral in Ω and a surface integral
complexity. The magnetic and electric parameters of TFMs,                      on Γ of products of scalar or vector fields.
such as the one shown in Fig. 5, are better computed by three-                    A three-phase TFM, 10 kW, 17.3 A, 220 V and 45
dimensional finite element analyses (FEA) [7].                                  poles was constructed by WEG Industries, Brazil, and it is
   In FEA, the calculation domain is discretized, and the partial              shown in Fig. 5. The permanent magnets used were Nd-Fe-B.
differential equations describing the field distribution, i.e.,                 Considering the electromagnetic symmetries and using anti-
Maxwell’s Equations, is substituted by a set of algebraic equa-                periodic boundary conditions, the smaller domain of study
tions which aproximate their solution. In three-dimensional                    consists of an 8 degree sector of the original structure. The
simulations, these representations of discrete units, or simply                three-dimensional mesh of this domain is shown in Fig. 6.
finite elements, are usually tetrahedra, bricks, and triangular                 Hexahedra in the moving band and prisms elsewhere were
prisms.                                                                        used. The mesh of the structure has 40 divisions along two
   The rotor displacement is modeled by means of a layer                       moving bands, one in each air gap.
of finite elements placed in the air gap. This method, named                       Figure 7 illustrates the electromotive force (emf) obtained
Moving Band Method, uses an automatic relocation of peri-                      with a magnetodynamic formulation. This result was presented
odicity or anti-periodicity boundary conditions allowing the                   for a speed of 200 rpm and when the generator operates at no-
simulation of any displacement between stationary and moving                   load condition, i.e., only the permanent magnet excitation is
parts of the electrical machine.                                               considered [7].

                                   1558-7908 c 2007 IEEE Education Society Student Activities Committee (EdSocSAC)
IEEE MULTIDISCIPLINARY ENGINEERING EDUCATION MAGAZINE, VOL. 2, NO. 1, MARCH 2007                                                                        6

                                                                           Fig. 7.   Electromotive force versus time.

                                                                             One of the major impediments to use TFMs is in fact their
                                                                           complicated construction. Another significant aspect which
                                                                           must be taken into account when designing a TFM is that
                                                                           most of the prototypes already proposed used for their cores
                                                                           expensive soft magnetic composites, contributing its high
                                                                           cost. These are reasons which prevent the companies from
                                                                           having a mass production of this type of machine. For critical
Fig. 6.   The studied domain and three-dimensional mesh.                   applications demanding performance and compactness this
                                                                           type of machine fits perfectly.

  With the developed tools, the influence of the three-                                                   R EFERENCES
dimensional nature of the machine can be investigated. The
                                                                           [1] H. Weh and H. May, “Achievable force densities for permanent magnet
accurate computation of the eddy currents in some parts of                     excited machines in new configuration,” in International Conference on
the machine can be done as well. The tools are intended to be                                                    u
                                                                               Electrical Machines – ICEM, M¨ nchen, Sept. 1986, pp. 1107–1111.
used for the study of other TFM.                                           [2] G. Henneberger and I. A. Viorel, Variable Reluctance Electrical Ma-
                                                                               chines. Aachen: Shaker Verlag, 2001.
                                                                           [3] G. Henneberger and M. Bork, “Development of a transverse flux traction
                         III. C ONCLUSIONS                                     motion in a direct drive system,” in International Conference on Electrical
                                                                               Machines – ICEM, Helsinki, Aug. 2000, pp. 1457–1460.
   This paper is a report of design types and development                  [4] G. Henneberger, I. A. Viorel, R. Blissenbach, and A. D. Popan, “On
works of TFMs. To investigate the application of TFMs, sev-                    the parameters computation of a single sided transverse flux motor,” in
eral prototypes have been developed by different researchers.                  Workshop on Electrical Machines Parameters. Cluj-Napoca: Technical
                                                                               University of Cluj-Napoca, May 2001.
Recent topological investigations regarding TF geometries                           o
                                                                           [5] L. L¨ wenstein, “Kurbelwellen-Starter-Generatoren auf der Basis von
have lead to multiple arrangements with different character-                   Reluktanzmaschinen,” Master’s thesis, RWTH Aachen, Aachen, 2003.
istics. The TFM has a higher torque/volume ratio when com-                 [6] Y. G. Guo and J. G. Zhu, “Study of permanent magnet transverse flux
                                                                               motors with soft magnetic composite core,” in Australasian Universities
pared to conventional ones, but because of three dimensional                   Power Engineering Conference, Brisbane, Sept. 2004.
flux paths, the topology requires the use of isotropic materials            [7] M. V. Ferreira da Luz, P. Dular, N. Sadowski, and J. P. A. Bastos,
like soft magnetic composites, as well as 3D numerical design                  “Three-dimensional finite element analysis of a transverse flux permanent
                                                                               magnet motor,” in Conference on the Computation of Magnetic Field –
tools.                                                                         COMPUMAG, New York, 2003.

                                  1558-7908 c 2007 IEEE Education Society Student Activities Committee (EdSocSAC)

Shared By: