Investigation of a New Topology of Hybrid-Excited

Flux-Switching Machine with Static Global Winding:

Experiments and Modeling
Agathe Dupas, Sami Hlioui, Emmanuel Hoang, Mohamed Gabsi, Michel
Lecrivain

To cite this version:

Agathe Dupas, Sami Hlioui, Emmanuel Hoang, Mohamed Gabsi, Michel Lecrivain. Investigation of a
New Topology of Hybrid-Excited Flux-Switching Machine with Static Global Winding: Experiments
and Modeling. IEEE Transactions on Industry Applications, Institute of Electrical and Electronics
Engineers, 2015, �10.1109/TIA.2015.2497305�. �hal-01233372�

HAL Id: hal-01233372

https://hal.archives-ouvertes.fr/hal-01233372
Submitted on 25 Nov 2015

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Investigation of a New Topology of Hybrid-Excited
Flux-Switching Machine with Static Global
Winding: Experiments and Modeling
Agathe Dupas, Sami Hlioui, Emmanuel Hoang, Mohamed Gabsi and Michel Lecrivain
SATIE, CNRS, ENS Cachan
CNAM, Universud
61, av President Wilson, F-94230 Cachan, France
[email protected]

Abstract—This paper presents a new structure for a hybrid topology”). The second category includes machines whose flux
excitation, flux-switching synchronous machine. The particularity lies both in the plane and perpendicular to the plane (hence a
of this machine is its global excitation winding component that ”3D topology”). This paper will work with a ”3D topology”.
creates a 3D excitation flux path to control the global air
gap flux. A 3D finite element analysis and a comparison with The hybrid flux-switching (HEFSM, Tab.I) machine pre-
experimental results are both provided. This model makes it sented in [3], as derived from the flux-switching machine
possible to: show the excitation flux path, explore the flux- introduced by Rauch in [4], has a ”2D topology” and features
regulation capability of a new structure, and identify the influence many optimizations since the original structure. Z.Q. Zhu
of material characteristics. The experimental performance of this [5] provided a complete review of various flux-switching
machine in a DC generator is presented and compared with
simulation output. structures and optimizations proposed in the literature.
Index Terms—BH-curve, Diode Bridge Rectifier, Flux- Due to a salient passive rotor, the hybrid excited, flux-
Switching, Hybrid excitation machine, Measurements, Syn- switching permanent magnet machines exhibit high robustness
chronous alternator, 3D-FEA and high torque density thanks to permanent magnets located
on the stator. A review in [6] presents some hybrid excited,
TABLE I
NOMENCLATURE flux-switching machines with the topology shown in [7] ; these
incorporate iron flux bridges in order to improve the effective-
FEA Finite-Element Analysis. ness of field coil excitation. The new topology discussed in this
PM Permanent Magnet. article is a hybrid, flux-switching machine derived from [9],
HEM Hybrid Excited Machine. and its ”3D-topology” differs from the initial ”2D-topology”.
HEFSM Hybrid Excited Flux-Switching Machine. Some hybrid-excitation machine topologies give rise to
DBR Diode Bridge Rectifier. magnetic fluxes that cannot be represented in a plane. It
δexc Excitation current density. becomes more complicated to classify existing topologies
δcc Short-circuit current density. than the ”2D-topologies” since adding a dimension multiplies
Icc Short-circuit current. the intentions of designers and multiple structures wind up
EMF Electromotive force. being proposed. A number of structures however can still
be distinguished. First of all, some machines combine two
synchronous designs, one a permanent magnet synchronous
I. I NTRODUCTION machine the other containing windings. For example, the
Hybrid Excited Machines combine the advantages of both machines presented in [8]-[16] are indeed combinations of two
wound-field excitation machines and permanent-magnet ex- synchronous configurations. Flux-concentrating machines, like
cited machines, with on the one hand electric excitation offer- in [17], are also included herein. The permanent magnets are
ing control of the magnetic field by controlling the excitation located inside the rotor. Due to the flux-convergence effect,
current and on the other a permanent magnet benefiting from the air gap flux density of the machine is high.
high-energy density. Adding an excitation source however can The literature contains relatively few studies of hybrid
complicate the structure. excited, flux-switching permanent magnet machines with a 3D
Various hybrid excited topologies have been proposed in flux path. One such device is an axial Flux-Switching, Hybrid
the literature and some authors have presented an overview Excitation Synchronous Machine and is described in [18].
of the different structures, e.g. [1][2]. This introduction will In this study, a new hybrid excitation, flux-switching ma-
focus on the path taken by the magnetic flux, which allows chine is being presented. The main feature of this device is its
us to distinguish two categories. The first comprises machines global winding hybrid excitation with claw poles. This solution
whose flux can be represented in a plane (a so-called ”2D has been explored in order to reduce the copper mass and
 TABLE II
PROTOTYPE MACHINE SPECIFICATIONS

Number of phases 3
Number of stator slots 12
Number of rotor teeth 10
External stator radius 100 mm
External rotor radius 45.5 mm
Air gap 0.25 mm
Active axial length 30 mm
Number of turns per phase 184

TABLE III
MATERIALS

Rotor FeSi (laminated sheets)

Stator FeSi (laminated sheets)
Stator Yokes XC18 (NF: A35-551: 1986)
Permanent magnets NdFeB

increase excitation winding efficiency. One of the most widely

used alternators in automotive applications is a claw pole
alternator whose claws are located on the rotor. The prototype
introduced in this article is based on the same principle yet
with claws located on the stator, which allows its rotor to
be passive in rotating at higher speeds without slip-rings or
brushes. Furthermore, the advantages of the double excitation Fig. 1. 3D model and photograph of the Hybrid Excited, Flux-Switching
are cumulative. The article will first describe the structure Machine
and operating principles of this new hybrid excitation, flux-
switching machine. The second part will explore the no-load
flux linkage and the back-electromotive force on a no-load aligned with one of the stator teeth. Some flux paths created by
using a 3D finite element analysis; the results of this analysis permanent magnets are actually short-circuited in the yoke, as
will then be compared with measurement data. The no-load shown in Fig.3. Without a claw pole structure, the maximum
flux linkage investigation serves to highlight: the influence no-load flux linkage for a single coil equals 350 µWb, whereas
of the stacking factor, the B-H curve definition, and the with a claw pole structure at an excitation current density
permanent magnet residual induction value (Br ). Moreover, (δexc ) equal to 0 A/mm2 , the maximum no-load flux linkage
the load testing of this machine will be displayed. Short-circuit for one coil equals 204 µ Wb (some 40% of all flux paths are
currents will be calculated and measured in order to determine short-circuited in the yoke).
the output power capability while operating in generator mode. The ”2D-structure” has been is derived from the hybrid
excited, flux-switching machine presented in [9][20]. The main
II. H YBRID EXCITED FLUX - SWITCHING PERMANENT difference between these two machines lies in the hybrid
MAGNET PROTOTYPE excitation circuit. In [9], the excitation windings are composed
Fig.1 shows the proposed 12-slot and 10-pole machine com- of 12 windings and positioned on the permanent magnets
posed of NdFeB permanent magnets, an excitation winding The second part of the stator is composed of a global wound
and armature windings; the pertinent specifications are listed inductor separated from the first part, which serves to avoid
in Tab.II and Tab.III. This machine is flux-switching, hence end-winding. An end-winding introduces copper losses and
the rotor is composed of ten salient poles without a permanent produces no flux linkage, whereas with a global winding all
magnet (PM) or windings. the ”copper” is used for producing flux. Each stator yoke,
The stator can be divided into two parts. The first is which maintains the winding, is composed of six claws and
composed of armature coils, permanent magnets and laminated made from solid iron materials. Fig.4 and Fig.5 show the
sheets. The armature windings themselves consist of twelve excitation circuit. The DC current in the excitation winding
windings located around teeth and the PM. The magnetic is independently controllable, and the pertinent control circuit
structure displays axial symmetry; this structure can be mod- is exhibited in Fig.4.
eled in two dimensions and the flux path can be utilized The flux path created by the DC-excitation winding and
without claws, as depicted in Fig.2. This figure indicates the without a PM is a 3D path and has been represented in Fig.6.
flux-switching principle, with both the negative and positive As this figure reveals, the flux paths generated by the winding
maximum flux linkages occurring when a rotor pole is roughly pass by one claw and return by the next claw. Fig.7 shows
flux paths with permanent magnets and excitation windings.
According to this configuration, the permanent magnet flux
path and excitation winding flux path are opposite, thus the
no-load flux linkage decreases.

Fig. 2. First part of the stator: 2D flux paths utilized

Fig. 5. Prototype: One half of claw pole stator yoke

Fig. 6. Machine model without permanent magnets: 3D flux paths

Fig. 3. First part of the stator: 3D flux paths utilized with claws

Fig. 4. Second part of the stator: Excitation circuit composed of two

armatures and one winding, plus the control circuit Fig. 7. Machine model with permanent magnets and excitation windings:
3D flux paths

III. 3D F INITE -E LEMENT A NALYSIS AND EXPERIMENTAL

VALIDATION : N O - LOAD TESTS
model precision, and defining a realistic constitutive relation
To validate the operating principle of this new structure for the material. To avoid this second difficulty, our model
and evaluate the potential of this concept, a prototype and results were compared to experimental measurements. The
simulations have been run. In this part, the operating principle meshing offered a compromise between accuracy and com-
on a no-load will be explored and measurements compared putation time. Fig.8 depicts the 3D finite element mesh of
with 3D finite element analysis (FEA). The 3D finite element the studied machine. Different mesh sizes were compared in
model was chosen since it takes the saturation effect into order to identify the mesh configuration that yields consistent
account and observes the influence of material definitions or results with respect to experimental measurements, yet without
the lamination effect. Two major difficulties of FE modeling excessive computation times. The final model was composed
consist of: reducing computation time without deteriorating of 365,888 tetrahedron elements. All 3D FEA were performed
with the ANSYS Maxwell Software. The solver formulation is
based on a current vector potential in conductors and a scalar
potential over the entire field domain. The models employed
were all supplied by current. A finite-element model of the half
machine has been made by considering symmetry in order to
reduce computation time. The half model was composed of
163,737 tetrahedron elements and provides the same results
as presented in this paper with computation time divided by
around two
In this part, for no-load test, the experimental test bench
used is shown in Fig.9. As the prime mover of HEFSM, a
5.5-kW induction motor was used.

Fig. 10. Measurements: Evolution of no-load flux linkage vs. excitation

current density per turn per slot

Fig. 8. FEA: View of the meshing configuration

A. No-load flux linkage

The first intrinsic machine characteristic observed was the
no-load flux linkage, along with the influence of hybrid
excitation on flux. The main advantage of hybrid excitation is,
Fig. 11. FEA: Magnetic flux density distribution (a)δexc = 0A/mm2
like with wound excitation for a synchronous machine, the flux (b)δexc = 6A/mm2
regulation capability. In Fig.10, the no-load flux linkage can
be viewed as a function of excitation current density. Thanks
to the excitation current, the air gap magnetic field may be a) Material definitions: The 3D finite element model
adjusted, as the no-load flux linkage per turn per slot ranges allows determining the influence of material definition on the
from 180µW b to 380µW b. Fig.10 reveals that the maximum no-load flux. A comparison between two material definitions
flux is reached for δexc = 9A/mm2 . The flux decreases for will be drawn first. The laminated parts of the stator and
δexc ≥ 9A/mm2 , with this decrease being due to magnetic rotor are produced using FeSi sheets, which are defined by
saturation phenomena in the stator. The saturation is located means of a B-H curve. Two curves are used, and the no-load
on stator teeth, as shown in Fig. 11. flux linkage is calculated for both curves. The first curve,
The no-load flux linkage is explored through measurements, BH(1), is acquired with a permeameter in the laboratory.
and this knowledge can then be used to readjust the finite The permeameter yields a B-field for higher H-field values
element model. The influence of three parameters on no-load than with an Epstein frame. The B-field is measured for an
flux is investigated, namely: the BH curve of FeSi sheets, H-field between 200A/m and 80, 000A/m.. The values for
permanent magnet remanence, and the stacking factor. H ≤ 200A/m have been removed in order to avoid potential
convergence problems during the FEA. An extrapolation gives
a B-field to 600, 000A/m. The second curve stems from the
manufacturer’s catalogues and has been acquired using an
Epstein frame. The B-field is measured for an H-field between
0A/m and 10, 000A/m. Both curves are seen in Fig.12.
Fig.13 presents the results obtained from the FEA with both
BH curves. BH(1) offers a better agreement with measurement
output. The existence of high saturation zones makes it es-
sential to have a BH characteristic measured for high H-field
values in order for the model to be realistic. These zones are
Fig. 9. Experimental test bench: No-load flux linkage and EMF measure- mainly located in the stator. This study has therefore revealed
ments the importance of material knowledge in an optimization
process. A rough approximation of the BH curve definition has
in fact introduced a 15% error on the no-load flux estimation.
The second material definition explored is the permanent
magnet definition. In order to avoid remanence value uncer-
tainty, a measurement has been conducted to determine the
permanent magnet remanence: Br equals 1.2T . The magnetic
field is measured in the air gap of a ring-shaped core using
a Hall effect probe. As indicated in Fig.14, the finite element
analysis provides good agreement with measurements when-
ever the magnetic remanence is well known. As a matter of
fact, an uncertainty of approx. Br equal to 0.1T prevents easily
obtaining a realistic model (Fig.14). The difference between
measurements and no-load flux calculations with FEA and
Br = 1.1T amounts to 10% at δexc = −6A/mm2 .

Fig. 14. FEA and measurements: No-load flux linkage vs. hybrid excitation
current density, influence of Br value

the stator (2D). Two configurations have been explored with

FEA, one without lamination the other with a stacking factor
(klam ) equal to 0.97 (i.e. a realistic value). The stacking factor
reflects the proportion of steel with respect to insulation. The
major influence of lamination occurs for δexc ≤ 5A/mm2 , as
shown in Fig.15. Due to insulation, the no-load flux linkage
decreases by 7%.

Fig. 12. Part of BH curves used in FEA

Fig. 15. FEA: No-load flux linkage vs. hybrid excitation current density,
influence of lamination

B. Electromotive force
The electromagnetic performance of the prototype machine
is predicted by a 3D finite element analysis and then com-
pared with measurement results. Measurements and transient
Fig. 13. FEA and measurements: No-load flux linkage vs. hybrid excitation
current density, influence of the BH curve simulations are both conducted at a rotational speed of 1,500
rpm. The back-EMF is obtained by numerically differentiating
b) Lamination: Three-dimensional finite element models the flux linkage, as the following equation illustrates:
allow taking lamination into account; in this machine, two dΦ
elements are laminated, namely the rotor and the first part of e = −Nr Ω (1)
dθ
where Ω is the angular speed of the rotor θ the angular been adopted: 1) permeability is assumed to be infinite without
position, Nr the rotor pole number and Φ the flux-linkage of any saturation effect; and 2) the prototype is considered to be
the phase windings. Fig.16 compares the EMF obtained from a smooth pole machine.
measurements and simulations at a single excitation current
density: the EMF simulated by FEA is indeed similar to A. Short-circuit current
measurement results. The short-circuit current density is determined in order to
assess the potential of such a structure when in generator mode
connected to a diode bridge rectifier (DBR). The maximum
output power when the machine is connected to a DBR can
actually be defined by the following formula [19]:
2
Pmax = 3Icc √ UDC (2)
π 2
where the DC-bus voltage UDC is assumed to be constant, and
Icc is the mean short-circuit current. For measurement data, the
experimental test bench used is the same as that displayed in
Fig.9 with short-circuited armature windings. The short-circuit
current is also determined by means of FEA and corresponds
to the current that cancels the flux; in this case, only the
fundamental part of the flux is being considered. To calculate
this current, the d-axis position must first be determined. Next,
three-phase currents, in phase opposition with the no-load flux,
are injected. Lastly, the total excitation flux in the machine
Fig. 16. FEA and measurements: Back EMF vs. Time is computed and compared to zero. If the flux equals zero,
then the short-circuit current density is obtained; otherwise,
The back-EMF can be adjusted by the DC excitation cur- the current amplitude is increased until the fundamental flux
rent; moreover, as the flux linkage increases, a peak value component is cancelled.
appears at a high excitation current density (Fig.17). Fig.18 presents the evolution in short-circuit current density
with respect to hybrid excitation current.

Fig. 18. FEA and experiments: Evolution of short-circuit current density vs.
excitation current density

Fig. 17. Measurements: Maximum back-EMF vs. hybrid excitation current The FEA results are similar to measurement output even
density
when harmonics are not taken into account. From Equation 2,
for this structure in generator mode with a DBR connected to
IV. 3D FE A NALYSIS AND EXPERIMENTAL VALIDATION : a 300V DC-bus, the potential maximum power lies between
L OAD TESTS 2,850 W and 5,800 W.
In this part, the operating principle relative to load will be
explored. The load calculation with FEM is time-consuming, B. Output power measurements and comparison with calcu-
so an analytical model has been implemented herein and then lation
verified via a comparison between measurements and analyt- For the configuration depicted in Fig.19 and Fig.20, the
ical results. To make use of this model, two hypotheses have DBR is connected on the same DC-bus voltage supplying
 Fig. 21. Single-phase equivalent circuit

Fig. 19. Diagram of the experimental set-up (generator mode) where Φ is the total flux-linkage, and δ the phase difference
between e and v. The circuit equation of the machine when
operating as a generator is given by
di(t)
e(t) − L − v(t) = 0 (8)
dt
In the sinusoidal steady state for each harmonic, this equa-
tion can be written using complex vectors, leading to a
solution.[21] The DC current is determined by
3
IDC = I (9)
π
And the DC power is:
s  2
3 UDC 2UDC
PDC = 1− (10)
π L πΦNr Ω
This equation 10 is only valid for
Fig. 20. Experimental test bench: The alternator is connected to both a DBR 2UDC
and a DC-bus
≤1 (11)
πΦNr Ω
with Nr being the rotor pole number. The definition of a speed
the 3-phase inverter that controls the induction motor; the Ωb (below this speed the output power equals to zero) thus
alternator output power can be defined by (3): appears as:
2UDC
Ωb = (12)
πΦNr
PDC = UDC IDC (3)
Φ
Moreover, the short-circuit current Icc is defined by L. This
Some assumptions have been adopted in order to calculate equation can thus be rewritten as follows:
the alternator output power as a function of no-load flux s  2
linkage and short-circuit current. The armature coil resistances 3 Ωb
PDC = UDC Icc 1 − (13)
have been neglected (R = 0.7Ω). The DC-bus voltage remains π Ω
equal to 300 V. The phase-to-ground voltage can be decom-
To verify the model, measurements have been recorded and
posed into a Fourier series
compared to the calculation results. The no-load flux and
+∞
X short-circuit current values have been calculated with a 3D-
vAN (t) = V2k−1 sin ((2k − 1)ωt) (4) FEA. Output power was measured at various speeds for a
k=1 single hybrid excitation current density (7A/mm2 ). Fig.22
The current is assumed to be AC and controls the opening and presents a comparison between measured and model-based
closing of switches. In addition, the current curve shape and output power. These good results reveal that the model enables
magnitude is mainly given by the fundamental current (6): estimating the potential of a structure in generator mode
more quickly than with a 3D-FEA combined with a circuit
+∞
X model. Measurements conducted in generator mode with the
ia (t) = In sin(ωt) + In sin(nωt) (5) experimental test bench shown in Fig.19 and Fig.20 have
n=1
made it possible to determine the copper, iron and mechanical
ia (t) ' I sin(ωt) (6) losses. Copper losses in the armature and excitation circuits
are known thanks to resistance and current values. Total
Based on these assumptions, the machine can be described by copper losses vary between 85 W and 150 W. Torque is
Fig.21, with: measured using a torque sensor, while mechanical power can
e(t) = Φωsin(ωt + δ) (7) be deduced from both torque and speed values. The DC
 Fig. 24. FEM and prototype: Single-turn sense coil on a claw pole

Fig. 22. Measurements and calculation: Output power vs. speed

current measurement allows proceeding with an output power

calculation. The mechanical and iron losses are found by
taking the differences between the mechanical power, copper
losses and output power. Fig.23 displays output power and
losses at various speeds for a single hybrid excitation density
current (7 A/mm2 ). Iron and mechanical losses are separated Fig. 25. Measurements: single-turn sense coil flux vs. Time
using a mechanical model [22]. As expected, iron losses
increase with speed, and the mechanical losses vary between
0.68 W and 17.6 W. V. C ONCLUSION
This paper has presented one topology of a hybrid ex-
cited, flux-switching permanent magnet machine with claws
located on the stator. Flux paths have been explored using a
permanent magnet or else with hybrid excitation current in
order to highlight axial flux paths. The investigation of no-
load flux linkage has shown remanence, lamination and BH-
curve influences on the 3D-FE model. This 3D model offers a
good level of agreement for both no-load flux linkage and
electromotive force. The final part has presented prototype
characteristics relative to the load. First, short-circuit currents
have been computed to obtain an output power estimation.
Second, output power has been measured and compared with
calculation values. The model allows estimating output power
of the structure when running in generator mode (with a DBR).
The operating principle of this new structure has been
validated. The 3D FEA model developed in this paper has laid
Fig. 23. Output power measurement and separation of losses by calculation a foundation for the design optimization and further research
and measurements on prototype performance.

Flux density variations in the solid claw pole structure have R EFERENCES
been measured by a single-turn sense coil (Fig. 24). Eddy [1] Zhao, C. and Yan, Y., “A Review of Development of Hybrid Excitation
current losses in the yokes have been estimated. For example, Synchronous Machine”, Proceedings of the IEEE International Sympo-
sium on Industrial Electronics, pp. 857–862, 2005
at 7,200 rpm, iron losses due to induction variations equal 30 [2] Owen, R. and Zhu, Z.Q. and Wang, J.B. and Stone, D.A. and Urquhart,
W, whereas total iron losses amount to 500 W. Fig.25 shows I., “Review of variable-flux permanent magnet machines”, International
the flux measured in a claw pole. The portion of iron losses Conference on Electrical Machines and Systems, pp. 1–6, 2011
[3] Lipo, T.A., “A novel permanent magnet motor with doubly salient
occurring in the solid claw pole structure is about 6% of the structure”, IEEE Transactions on Industry Applications, pp. 1069–1078,
total iron losses. 1995
[4] Rauch, S.E. and Johnson, L.J., “Design Principles of Flux-Switch Alter- “ A Novel Dual-Stator Hybrid Excited Synchronous Wind Generator”,
nators ”, Transactions of the American Institute of Electrical Engineers. IEEE Transactions on Industry Applications, pp. 947–953, 2009
Part III : Power Apparatus and Systems, pp.1261–1268, 1955 [15] Capponi, F.G. and De Donato, G. and Borocci, G. and F. Caricchi,
[5] Zhu, Z.Q. and Chen, J.T., “Advanced Flux-Switching Permanent Magnet “Axial-Flux Hybrid-Excitation Synchronous Machine: Analysis, Design
Brushless Machines”, IEEE Transactions on Magnetics, pp. 1447–1453, and Experimental Evaluation”, IEEE Transactions on Industry Applica-
2010 tions, vol.50, no.5, pp.3173–3184, 2014
[6] Wang, Y. and Deng, Z., “Comparison of Hybrid Excitation Topologies for [16] Zhang Z. and Ma, S. and Dai, J. and Yan Y., “Investigation of Hybrid
Flux-Switching Machines”, IEEE Transactions on Magnetics, pp. 2518– Excitation Synchronous Machines With Axial Auxiliary Air-Gaps and
2527, 2012 Non-Uniform Air-Gaps”, IEEE Transactions on Industry Applications,
[7] Owen, R. L. Zhu, Z. Q. and Jewell, G. W., “Hybrid-Excited Flux- pp. 1729–1737, 2014
Switching Permanent-Magnet Machines With Iron Flux Bridges”, IEEE [17] Zhang, Z. and Yan, Y. and Yang, S. and Bo, Z., “Principle of Operation
Transactions on Magnetics, pp. 1726–1729, 2010 and Feature Investigation of a New Topology of Hybrid Excitation
[8] Fu, X. and Zou, J., “Numerical Analysis on the Magnetic Field of Hybrid Synchronous Machine”, IEEE Transactions on Magnetics, pp. 2174–
Exciting Synchronous Generator”, IEEE Transactions on Magnetics, 2180, 2008
4590–4593, 2009 [18] Yildiriz, E. and Onbilgin, G., “Design studies of Axial Flux Hybrid
[9] Hoang, E. Lécrivain, M.and Gabsi, M.,“ A new structure of a switching Excitation Synchronous Machine with magnetic bridge”, 8th International
flux synchronous polyphased machine with hybrid excitation”, EPE, pp. Conference on Electrical and Electronics Engineering (ELECO), pp. 234–
1–8, 2007 237, 2013
[10] Liu, X. and Zheng, A. and Wang, C., “A Stator-Separated Axial [19] Gaussens, B. Hoang, E. Manfé, P. Lécrivain, M. and Gabsi, M., “Im-
Flux-Switching Hybrid Excitation Synchronous Machine”, Journal of proved Output Power Capability of Hybrid Excited Flux-Switching DC-
international Conference on Electrical Machines and Systems, Vol. 1, Alternators : Analysis and Experiments”, Industrial Technology (ICIT),
No. 4, pp. 399-404, 2012 pp.960–965, 2012
[11] Laldin, O. and Sudhoff, S.D. and Pekarek, S., “Analysis and Design of [20] Hoang, E. Lécrivain, M.and Gabsi, M.,“ Flux Switching Dual Excitation
Axial Hybrid Synchronous Machines”, ICEM International Conference Electrical Machine”, US Patent, March 8, 2007
on Electrical Machines, pp. 2593–2599, 2014 [21] Caliskan, V., Perreault, D. J., Jahns, T. M., Kassakian, J. G. , “Analysis of
[12] Borocci, G. and Capponi, F.G. and De Donato, G. and Caricchi, F., three-phase rectifiers with constant-voltage loads”, Circuits and Systems
“Mixed-Pole Hybrid-Excitation Machine”, ICEM International Confer- I: Fundamental Theory and Applications IEEE Transactions on, 2003
ence on Electrical Machines, pp. 1–7, 2014 [22] Vrancik, J. , “Prediction of windage power loss in alternators”, R TN
[13] Syverson, C.D. and Mankato, N. and Cums, W. R., “Hybrid Alternator”, D-4849, NASA, 1968
US Patent, March 14, 1996
[14] Liu, X. and Lin, H. and Zhu, Z.Q. and Yang, C. Fang, S. and Guo J.,