Durham Research Online

Deposited in DRO:
19 May 2016

Version of attached le:

Accepted Version

Peer-review status of attached le:

Peer-reviewed

Citation for published item:

Widmer, J.D. and Mecrow, B.C. and Spargo, C.M. and Martin, R. and Celik, T. (2012) 'Use of a 3 phase full

bridge converter to drive a 6 phase switched reluctance machine.', in 6th IET International Conference on

Power Electronics, Machines and Drives (PEMD 2012). , B4.2.

Further information on publisher's website:

https://doi.org/10.1049/cp.2012.0260

Publisher's copyright statement:

This paper is a postprint of a paper submitted to and accepted for publication in the 6th IET International Conference
on Power Electronics, Machines and Drives (PEMD 2012) and is subject to Institution of Engineering and Technology
Copyright. The copy of record is available at IET Digital Library.

Additional information:

Use policy
The full-text may be used and/or reproduced, and given to third parties in any format or medium, without prior permission or charge, for
personal research or study, educational, or not-for-prot purposes provided that:
• a full bibliographic reference is made to the original source
• a link is made to the metadata record in DRO
• the full-text is not changed in any way
The full-text must not be sold in any format or medium without the formal permission of the copyright holders.
Please consult the full DRO policy for further details.

Durham University Library, Stockton Road, Durham DH1 3LY, United Kingdom
Tel : +44 (0)191 334 3042 | Fax : +44 (0)191 334 2971
https://dro.dur.ac.uk
 Use of a 3 Phase Full Bridge Converter to drive a 6 Phase
Switched Reluctance Machine
J. D. Widmer*, B. C. Mecrow*, C. M. Spargo*, R. Martin* and T. Celik†

* Newcastle University, UK, +44(0)191 2223016, james.widmer@ncl.ac.uk

†Dyson, UK, tuncay.celik@dyson.com

Keywords: SRM, Full Bridge Converter, Torque Ripple unipolar waveforms, allowing the emulation of a half bridge
Reduction, Low Cost Drive. converter. The paper concludes that this drive configuration
does not significantly impact the torque / speed or efficiency
Abstract performance of a 3 phase machine and provides comparable
performance to an induction machine.
This paper explores how the use of a modified full bridge
converter, coupled to a 6 phase switched reluctance motor, This paper takes this research to its logical extension,
can produce a torque dense drive with low torque ripple, exploring how such a methodology would allow a 6 Phase
combined with standard drive electronics and a good SRM to be driven from a conventional AC Full Bridge
converter VA rating. A description is provided of the new Converter.
drive configuration: a prototype motor and drive are
constructed and test results provided. Results are compared 2 Proposed Drive Configuration
with those to be expected from the same switched reluctance
Fig. 1 shows how diodes, arranged alternately between
machine driven from a conventional, asymmetric half bridge
phases, can be used to convert the bipolar current waveform
converter.
output from each phase of the converter into two unipolar half
wave forms, relating to the positive and negative regions of
1 Introduction the waveform respectively. Consequently the three phase
Two reasons are often cited as contributing to the under- converter is able to supply a six phase SRM, whilst having
utilisation of Switched Reluctance Machine (SRM) only three power connections between inverter and motor.
technology for commercial applications. These are: Configurations based on both star or delta configurations are
possible [8], and this paper will show examples of both.
(a) high torque ripple;
(b) non-standard asymmetric half bridge converters [1].
However, there are undoubted benefits associated with the
use of SRM technology. These include extremely robust
construction, high torque density and low cost, even
compared to the induction machine [2]. Therefore, a drive
topology which can overcome the SRM’s limitations is
welcome.
In [3] a basis for estimating the kVA requirements for an
SRM drive is presented and results compared with the
requirements of induction machine drives. In [4] and [5] (a) Star connection
schemes are proposed for operating star/delta connected
SRMs from conventional AC full bridge converters. A later
publication [6] reports that these configurations reduce torque
ripple, but suffer in terms of reduced average motor torque
capability. However, these configurations benefit through a
reduction in the number of connections between drive and
motor (from 6 to 3 for a three phase machine) and the
reduction in the number of current and voltage measurement
devices needed from three to two.
In [7] the authors present a drive configuration which would
allow a three phase SRM, with windings connected in a delta
configuration, to be driven from a 3 phase full bridge (b) Delta connection
converter. This utilises diodes in line with each phase in order Figure 1: Six-phase SRM driven by a 3-phase Full Bridge
to convert the bipolar output of the full bridge converter into a converter
With the above configuration the inverter is behaving exactly
as though it is supplying a three phase ac machine. It could
be configured to supply sinusoidal or quasi-square
quasi wave
outputs. In this work the converter is a standard commercial
drive, with a sinusoidal output current.
Fig. 2 shows idealised current waveforms for the six phase
SRM in the star configuration.

Figure 3: 6 Phase SRM with a single phase excited.

Figure 2: Current waveforms for a SRM being driven under
current control by the new unipolar full bridge configuration
in a star configuration. The colour of each phase in the
diagram relates to the output phase of the power converter.

3 6 Phase SRM Magnetic Design

The design of the SRM forr this application is based on
standard best practice: machine dimensions are summarised
in Table 1. Too maximise the torque capability the tooth width
to rotor pole pitch (tw/λ) ratio has been chosen to be 0.4.

Table 1: 6 Phase SRM design parameters

Number of Stator Teeth 12
Number of Rotor Teeth 10
Axial Length (Lamination Stack) 150.0mm
Stator Outer Diameter 150.0mm
Turns per Phase 100
Figure 4: Constructed 6 Phase SRM.
SRM
The core backs of the machine are relatively deep compared
to the tooth width. This is necessary because three phases
will be conducting at all times and so the core backs have to
simultaneously carry the flux of several phases.
The six phase machine has twelve stator teeth and so the
natural manner of excitation is to wind around single teeth,
with each phase comprising two coils. Fig. 3 illustrates such
an arrangement: the two coils of one phase are geometrically
opposite and wound in such a manner that the two coils share
the same flux.
Figs. 4 and 5 show photographs of the constructed
nstructed prototype
motor. Figure 5: 10 pole conventional rotor.
For this machine the measured unaligned
u flux-linkage results
4 Static Test Results are higher than predicted by finite element modelling,
because the 2D model neglects end winding effects upon
Experimentally derived Flux-Linkage
Linkage MMF and static torque motor performance. In the aligned position,
position for low currents,
characteristics are shown in Figs. 6 and 7, compared with 2D the measured flux-linkage
linkage indicates that the air-gap
air length is
FE results. slightly larger than the design value. These effects can also be
seen in the torque curves, with average measured torque Each motor phase is connected in parallel with another one
slightly less than predicted. via two diodes. The two motor phases are one half electrical
cycle apart so that, ideally, when one conducts the other does
not. The voltage applied to the two phases is equal in
magnitude and of opposite sign. Consequently the voltage
applied to a phase switching on is identical to that of one
switching off.

In the case of a star connection each inverter line output

current is directly fed into two back-to-back phases. There is
no access to the star point, so there is no direct control of the
voltage applied to each pair of phases.

In contrast in the case of a delta connection there is direct

control over each phase voltage, with the condition that the
phase voltages V1+V3+V5=0 and V2+V4+V6=0. The
inverter line currents are controlled, but there is no direct
control over motor phase currents.

6 Dynamic Test Results

Figure 6: Experimental static magnetization curves (solid In order to demonstrate that this 6 Phase SRM can be driven
lines) shown with modelled data (dashed lines). from a commercial, three phase AC drive it was tested using a
Control Techniques SP3410 3-phase drive cabinet under
It is generally assumed that mutual coupling effects in SRMs
speed control with a DC link voltage of 560V and switching
are negligible; this assumption is also made in the simulated
frequency of 16 kHz. The output of the motor drive was
results presented. Whilst transient FE modelling has
connected directly to the diode conversion unit, previously
demonstrated how for this motor mutual coupling may have a
described in Fig. 1. This diode unit houses three 60A diode
more significant, the purpose of this paper is to concentrate
pairs mounted on a heat sink, which converts the 3-ph bipolar
upon the converter and so this effect will not be addressed
output of the drive to the required 6 unipolar signals which
further.
are fed into the prototype machine.
It was possible to ‘Autotune’ the drive parameters so that PI
controller values were correctly and automatically identified.
The motor was configured as a 20 pole, synchronous PM
servo motor; the expected 10 bipolar poles being emulated by
the 10 unipolar rotor teeth.
The machine was tested from standstill to up to 4000rpm
maximum speed. The load at each operating speed was
steadily increased until torque could not be sustained. To
achieve the best torque per unit current it was necessary to
manually advance the phase offset angle to up to 90 electrical
degrees at 4000rpm.
Fig. 8 shows measured currents with the star connection,
generating 10Nm at 3000 rpm. The line current output from
the converter is controlled to be sinusoidal, but the phase
currents associated from this line current deviate significantly
from half sinusoids. Current in a phase ramps up quickly, but
Figure 7: Experimental static torque (solid lines) curves continues to flow after the incoming line current has changed
shown with modelled data (dashed lines). polarity. This is because there is a significant current flowing
round the loop created by the two back to back phases.
5 The Converter Configuration Operation with the delta connected configuration is shown in
Fig. 9 generating a mean torque of 20Nm at a speed of 3000
When driving an SRM from an asymmetric half bridge there
rpm. In general, operation is similar to that of the star
is complete control over the magnitude and polarity of the
connection, though of course the phase voltage can now equal
voltage applied to each phase. With the three phase bridge
the dc link voltage. The controlled current continues to be the
configuration this is no longer the case.
inverter line current, which comprises the sum of four phase
currents. There is therefore less direct control over phase
current, but this does not seem to have a significant effect The same machine, driven from six asymmetric half bridges,
bridges
upon performance. will be used for comparison in the sections below for
operation at 8000 rpm.

7.1 Asymmetric Half Bridge

Fig. 10 shows the flux-linkage
linkage current loci for one phase of
the machine, driven from an asymmetric half bridge under
full voltage control. 180° conduction is employed to
maximise the torque capability. The phase advance angle is
varied to help choose the optimum angle for torque
production. Fig. 11 shows the resultant current waveforms.
waveforms
There is little change in the mean torque produced for
advance
ce angles in the range 80 to 120 degrees, but increased
advance angle gives much larger phase currents. For this
reason an advance angle of 80°° was chosen, combining high
torque with low current.
Fig. 12 shows the instantaneous shaft torque produced by the
Figure 8:: 3000rpm at 10Nm load with the star connection, machine: this has a mean value of 17.4 Nm, with much less
showing measured inverter line current for phase A and motor ripple than most SRMs under voltage control because of the
phase currents in phases 1 and 4. high phase number.

0.2
0.18
0.16
0.14
Flux linkage (Wb-t)

0.12
0.1
0.08 60 deg
0.06 80 deg
100 deg
0.04
120 deg
0.02
140 deg
0
0 10 20 30 40
Current (A)

Figure 10. Flux-linkage

linkage loci for one phase with an
Figure 9:: 3000rpm at 20Nm load with the delta connection, 000 rpm Voltage control (180o
asymmetric half bridge. 8000
showing inverter line current for phase A and motor phase conduction.) Range of advance angles shown.
currents in phases 1 and 4.

Subjectively noise levels were perceived to be lower for this

machine, probably
ably as a result of its low torque ripple, as
compared to a conventional 3 phase SRM though still higher
than for an equivalent PMBL machine.

7 Operation under Voltage Control

As the speed rises the drive needs to move from current to
voltage control. Ultimately the inverter reaches a position
where the output voltage from each line of the inverter
becomes simply a square wave. The commercial drive was
not designed to operate inn this mode and so simulation results
will be used to examine performance under these conditions.
Figure 11. Current waveforms for one phase with an
000 rpm Voltage control (180o
asymmetric half bridge. 8000
conduction.) Range of advance angles shown.
7.2 Three phase bridge, delta connection. currents for this condition are shown in Fig. 16, showing how
they remain relatively close to sinusoidal in nature.
With the delta connection the voltage applied to any one
phase becomes 120° of positive voltage, then 60 degrees of
0.14
freewheeling, before negative voltage is applied to bring the
flux-linkage back down to zero, as shown in Fig. 13 below. 0.12
Compared to the asymmetric half bridge the peak flux-linkage
is reduced by one third and there is consequently a significant 0.1

Flux linkage (Wb-t)

reduction in current and torque. In order to produce the same
0.08
torque under current control it is therefore necessary to reduce
the number of turns in the machine. The following results 0.06
correspond to the same machine, but with the number of turns 60 deg
80 deg
per phase reduced from 50 to 40. The same slot fill factor is 0.04
100 deg
used, with the conductor cross-section adjusted accordingly.
0.02 120 deg
In reducing the number of turns by 20% the overall effect
140 deg
upon phase resistance is therefore a reduction to 64% of the 0
original value. 0 10 20 30 40
Current (A)
25
T per phase (Nm) T sum (Nm)
Figure 14. Flux-linkage loci for one phase with the delta
20 connection. 8000 rpm Voltage control. Range of advance
angles shown.
15
45
Torque (Nm)

10 60 deg
40 80 deg
5 35 100 deg
120 deg
0 30
140 deg
Current (A)

0 60 120 180 240 300 360

25
-5
20
-10 15
Electrical Angle (deg)

Figure 12. Total torque produced with an asymmetric half 10

bridge. 8000 rpm Voltage control (180o conduction.) 5

0
800 0.16 0 60 120 180 240 300 360
Electrical Angle (deg)
600 0.14

0.12
Figure 15. Motor phase current waveforms for one phase with
400 the delta connection. 8000 rpm Voltage control. Range of
Flux Linkage (Wb-t)

200
0.1 advance angles shown.
DC Voltage

0.08
0 50
0 60 120 180 240 300 360 0.06
40
-200
0.04
30
-400
0.02 20
Inverter Current (A)

-600 0 10

-800 -0.02 0
Electrical Angle (deg) 0 60 120 180 240 300 360
-10
Figure 13. Voltage and flux-linkage in one phase with a delta
-20
connection.
-30
Fig. 14 shows the flux-linkage current loci for one phase of
the machine. The period of zero voltage applied to the phase, -40
IA (A) IB (A) IC (A)
whilst conducting, results in a period when the flux-linkage is -50
Electrical Angle (deg)
constant, resulting in flat topped loci. Fig. 15 shows the
resultant motor phase currents. In this case an advance angle Figure 16. Inverter line current for the delta connection. 8000
of 90o was judged to form the best compromise between rpm.
torque capability and current requirement. The inverter line
Fig. 17 shows the instantaneous shaft torque which, at 17.4 Table 2. Comparison between converter arrangements
Nm, is virtually identical to the torque produced with the Asymmetric half 3 phase bridge
asymmetric half bridge arrangement. Once more, the torque bridge (delta connected)
ripple is very small, despite being under full voltage control. Number of IGBTs 12 6
Additional diodes none 6, rectifier grade
25 No. of power 12 3
T per phase (Nm) T sum (Nm)
cables between
20 converter & motor
No of current 6 2
15 sensors
Torque (Nm)

No. Of motor 100 80

10 turns per phase
Phase resistance 0.627 Ohms 0.401 Ohms
5 DC link voltage 560V 560V
Speed 8000 rpm 8000 rpm
0 Mean torque 17.4 Nm 17.5 Nm
0 60 120 180 240 300 360 Motor phase rms 10.5 A 13.1 A
-5
current
Electrical Angle (deg) Motor winding 418 418
Figure 17. Total torque produced with a delta connection. loss
8000 rpm Voltage control. Peak IGBT 20.7 A 44.3 A
current
7.3 Comparison between Converter Arrangements. Inverter peak VA 69.5 kVA 74.5kVA
rating
A fair comparison between arrangements must take into
kVA/kW 9.5 10.2
account a range of factors, relating both to the machine and
the converter. Table 2 summarises the findings, based upon
the above results. References
The results show that both configurations produce the same [1] T. J. E. Miller, “Electronic Control of Switched
torque, with virtually identical winding loss. It seems Reluctance Machines”, Newnes, 2001
therefore that the machine size is unaffected by the choice of [2] D. Dorrel, A. Knight et al, “Comparison of different motor
converter. The three phase bridge configuration has only half design drives for hybrid electric vehicles”, IEEE ECCE, 2010
the number of IGBTs, but each IGBT has to take around [3] T. J. E. Miller, “Converter Volt-Ampere Requirements of
double the peak current, so the overall converter volt-ampere the Switched Reluctance Motor Drive”, Industry
rating is little changed. Applications, IEEE Transactions on, vol. 21, pp. 1136-1144,
1985.
The converter kVA/kW ratings are similar to those described [4] A. Jin-Woo, O. Seok-Gyu, M. Jae-Won, et al., “A three-
in [3] where a rating of 10.5 was determined experimentally phase switched reluctance motor with two-phase excitation”,
for an 8/6 four-phase SRM and contrasted with a rating of 9.2 Industry Applications, IEEE Transactions on, vol. 35, pp.
for a comparable induction machine. 1067-1075, 1999.
[5] P. Somsiri, K. Tungpimonrut and P. Aree, “Three-phase
8 Conclusions full-bridge converters applied to switched reluctance motor
drives with a modified switching strategy”, Proc. of IEEE
This paper has demonstrated the feasibility of operating a 6
ICEMS, Conf., pp. 1563-1567, 2007.
Phase SRM from a three phase, AC full bridge converter
[6] X. Liu, Z. Q. Zhu, M. Hasegawa, et al., “Performance
through the simple addition of 6 rectifier grade diodes.
comparison between unipolar and bipolar excitations in
switched reluctance machine with sinusoidal and rectangular
The arrangement requires the addition of six rectifier diodes,
waveforms”, in Energy Conversion Congress and Exposition
but it offers the following features:
(ECCE), 2011 IEEE, 2011, pp. 1590-1595.
[7] A. Clothier, B. C. Mecrow, “Inverter Topologies and
• A standard three phase inverter; Current Sensing Methods for Short Pitched and Fully Pitched
• Only thee connections between motor and inverter; Winding SR Motors”, APEC, 1999
• Only two current sensors; [8] T Celik, “Segmental Rotor Switched Reluctance Drives”,
• Low torque ripple throughout the operating range; Newcastle University, August 2011
• No increase in motor loss, compared to standard
SRM drives;
• Very similar converter VA rating.