Switched Brushless Doubly-Fed Reluctance

Machine Drive for Wide-Speed Range Applications

Shivang Agrawal, Arijit Banerjee
Department of Electrical and Computer Engineering
University of Illinois at Urbana-Champaign
Urbana, USA
[email protected], [email protected]

Abstract—Brushless doubly-fed reluctance machines (BDFRM) two sets of stator windings — commonly known as primary
utilize a partially rated power converter but are restricted winding and secondary winding. The rotor can be induction
to applications which have limited operating speed range. A or reluctance resulting in two types of machines — brushless
switched BDFRM architecture has been described in this paper
that re-configures one of the stator excitations on-the-fly to enable doubly-fed induction machine (BDFIM) and brushless doubly-
a wide-speed range operation. This approach can extend the fed reluctance machine (BDFRM). BDFRM is more efficient
use of BDFRM to many wide-speed range applications, such as due to absence of copper losses in the rotor winding; active
propulsion systems, while preserving the benefit of fractionally- and reactive power is easier to control; and since all the losses
rated power electronics. Design, analysis, and simulations have occur in the stator, more efficient cooling methods can be
been performed and are discussed here.
Index Terms—Doubly fed, brushless, rotor modulation, utilized [7].
switched drive, power converter, current rating, voltage rating. The drive architecture is as shown in Fig. 1 where a
controlled back-to-back converter is always connected to the
I. I NTRODUCTION secondary stator, while the primary stator electrical port is
A standard approach for constructing megawatt-scale elec- connected to different sources on-the-fly depending on the
tromechanical drivetrains for many industrial applications operation speed. In the low-speed, low-power mode, the
has been to use single-electrical-port machines. To enable primary stator is either connected to a dc source or shorted
complete torque-speed control of the drivetrain, an identical together while the power converter connected to the secondary
megawatt-scale power-electronic converter is connected be- stator processes all of the mechanical power. In this paper, we
tween the electrical machine and the ac source to process consider the case where a dc source is used in low-power
the total input power. The need to handle megawatts of mode. In the high-speed, high-power mode, the bulk of the
power makes these converters massive and complex. Often, mechanical power is delivered directly through the primary
multiple power-electronic devices are stacked in series/parallel stator winding while the power converter processes only the
configurations to match the high power requirement with differential power through secondary stator. The advantage of
the available device ratings. This complexity results in the using a switched drive over a conventional drive architecture is
limited adoption of variable-speed drives (VSDs) in land-based shown in Fig. 2 where it can be seen that the power converter
high-power applications. Infact, only 13% of global loads in voltage rating is reduced to half by using the switched drive
megawatt class high-power applications are driven by VSDs architecture for an ideal BDFRM (zero resistance, leakage,
[1]. and magnetizing current).
An alternative approach uses a switched-doubly-fed ma- A minimum power converter rating, for a required drive
chine (DFM) with a parallel architecture for power processing torque-speed capability, is obtained by choosing an appropriate
to reduce the size of the power-electronic converter by two-
thirds while operating over a wide speed range [2]. In this
architecture, the power converter is connected to the DFM
rotor port. Using a thyristor-based transfer switch, the stator
port is reconfigured to an ac or a dc source on the fly, based
on the operating speed [3]. Advantages include reduction
in size, weight, and cost along with increased efficiency.
However, the need to access the rotor port through brushes and
slip rings leads to poor reliability and frequent maintenance.
These drawbacks limit the drive's relevance in many emerging
applications, such as megawatt-class propulsion systems [4]
[5] [6].
This paper extends the parallel architecture of power pro- Fig. 1. Architecture of switched brushless doubly-fed reluctance machine
cessing to a brushless-doubly-fed machine. The machine has drive.

978-1-4799-7312-5/18/$31.00 ©2018 IEEE 4473

Authorized licensed use limited to: Indian Institute of Technology Patna. Downloaded on March 19,2025 at 10:12:33 UTC from IEEE Xplore. Restrictions apply.
 III. S IZING FOR MINIMUM POWER ELECTRONICS
REQUIREMENTS FOR BDFRM

This section discusses sizing for a minimum power rated

power-converter such that the machine can operate at rated
torque for a wide speed range. A 42 KW, 380/380 V, 44.2/44.2
A, 750 rpm (synchronous speed) BDFRM is used to illustrate
the design process, which involves determination of current
(a) (b)
and voltage rating [9]. The current rating is selected such
that the candidate BDFRM is utilized to its maximum torque
Fig. 2. Required converter voltage using a (a) conventional drive and (b) capability in the high-speed mode. The primary stator flux
switched drive architecture for an ideal BDFRM.
and applied dc voltage are computed in low-speed mode, such
that the slope of the converter voltage versus rotor speed is
transition speed between the operating modes, maximum rotor minimized (Fig. 2(b)), while achieving the the desired torque.
speed, dc voltage and primary stator flux in the low-speed Finally the transition speed (speed at which mode switching
mode. For example, a proposed drive would require a power takes place) and maximum rotor speed are chosen to minimize
converter whose rating is about 48% of the maximum shaft the converter voltage rating. The limits on d and q-axis primary
power while operating over a wide speed range of 0 to 1.52 and secondary stator currents are computed to ensure that the
p.u., where 1 p.u. represents the natural synchronous speed. machine remains within its rated condition both in low and
The remainder of the paper is organized as follows. Section high-speed modes.
II presents the BDFRM steady-state model. In Section III, the
A. High-speed mode: Determination of converter current rat-
power-converter sizing design is explained with an example of
ing
a real BDFRM. Section IV shows the simulation results and
conclusions are presented in Section V. The first step is to decide the current rating of the power
converter connected to the secondary stator, such that both
II. BDFRM STEADY STATE MODEL primary and secondary stator currents remain within bounds,
while yielding maximum torque. Since the primary stator is
Power-converter sizing requires the steady state machine connected to a stiff ac voltage source,
model. The steady-state BDFRM per-unit flux equations in
2 2
the primary stator flux orientation are summarised as [8]: vpd + vpq =1 (11)

ϕp = xp ipd + xps isd (1) Substituting vpd and vpq from Eqs. (5) and (6), and using Eq.
0 = xp ipq − xps isq (2) (2),
ϕsd = xs isd + xps ipd (3) xps
ϕ2p + 2ϕp rp isq + rp2 i2p = 1 (12)
ϕsq = xs isq − xps ipq (4) xp
Solving the above quadratic equation and assuming rp << 1,
The corresponding voltage equations are [8]: the primary stator flux is estimated as
vpd = rp ipd (5) rp xps
ϕp = 1 − isq (13)
vpq = rp ipq + ωp ϕp (6) xp
vsd = rs isd + (ωp − ωr )ϕsq (7) To remain within primary and secondary stator currents limits,
vsq = rs isq − (ωp − ωr )ϕsd (8)
i2pd + i2pq ≤ 1 (14)
where p: primary stator winding, s: secondary stator winding i2sd + i2sq ≤ Is2rated (15)
and the torque is expressed as
The primary stator currents in Eq. (14) are substituted with
3 xps the secondary currents using Eqs. (1), (2) and (13):
τe = ϕp isq (9)
2 xp
 1 − rp xps isq − xps isd 2  2
where the base values are taken as xp xps
+ isq ≤1 (16)
vB xp xp
vB = vp ; ωB = ωp ; iB = iprated ; ϕB = ;
ωB
vB PB (10) Both current constraints, given by Eqs. (15) and (16), are
zB = ; PB = v B i B ; τ B = p r ; shown in Fig. 3. The common region occupied by these curves
iB wB
ABCD, is the area for safe operation in the high-speed mode,
Thus, in steady state, the primary stator-flux frequency wp is in which the q-axis converter current can vary in between -0.88
zero in low-speed mode and 1 p.u. in high-speed mode. and 0.88 p.u.

4474

Authorized licensed use limited to: Indian Institute of Technology Patna. Downloaded on March 19,2025 at 10:12:33 UTC from IEEE Xplore. Restrictions apply.
 power to and from the ac mains in the high-speed mode. The
primary stator reactive power is given by [8]:
3 3 ϕp − xps isd
Qp = (vpq ipd − vpd ipq ) = wp ϕp ( ) (18)
2 2 xp

This can be made zero by having

rp xps
1− isq
ϕp xp
isd = = (19)
xps xps

which implies a unity power factor operation of the primary

stator. This calculated d-axis current must also be less than
the maximum allowable value given by

Fig. 3. BDFRM stator current ratings limit the maximum permissible
converter d- and q-axis currents, which set the converter current rating.
isdmax = Is2rated − i2sq (20)

The d-axis secondary stator current is hence given by

1) Maximum torque capability: Using Eqs. (9) and (13), ϕp  2
torque is expressed as a function of isq as isd = min( , Israted − i2sq ) (21)
xps
3 xps rp xps Figure 5 shows the d-axis converter current choices for various
τe = (1 − isq )isq (17)
2 xp xp q-axis currents satisfying the Eq. (21). For −0.47 ≤ isq ≤ 0.5,
a unity power factor on the primary stator can be obtained,
Torque is plotted against the converter q-axis current in Fig.
by choosing isd given by Eq. (19). Beyond this range of
4. The maximum possible positive torque is 1.34 p.u. (for
isq , Eq. (20) gives the value of isd for operating at the
isq = 0.88 p.u.) and maximum negative torque = -1.42 p.u.
maximum possible primary power factor while keeping both
(for isq = −0.88 p.u.). Considering equal magnitudes of
stator currents within their respective limits.
maximum torque for both acceleration and deceleration, the
Having obtained isd and isq , the other machine currents
torque capability range is regarded as -1.34 p.u. to 1.34 p.u.,
and voltages can be computed using Eqs. (1)-(8) and (13).
as denoted by points E and C, respectively, in Fig. 4. The
The converter orsecondary stator voltage in the high-speed
operating point C, which defines the machine torque capability,
mode (vshigh = vsd 2 + v 2 ), when operating at the maximum
is used to calculate the converter current rating. Because this sq
point lies on the ellipse given by Eq. (15), the converter current positive torque (acceleration), is
rating is equal to Israted . 
2) Maximum grid power factor: While the torque is con- vshigh |acc = 1.383wr2 − 2.699wr + 1.317 (22)
trolled by the converter q-axis current, its d-axis current is
commanded to control the flow of primary stator reactive

Fig. 5. The choice of d-axis converter current is found on the path AXYZC
Fig. 4. Electromagnetic torque for varying converter q-axis current. Maximum as the q-axis current varies to obtain the best possible power factor on the
positive and negative torque capability of the machine is also shown. primary stator.

4475

Authorized licensed use limited to: Indian Institute of Technology Patna. Downloaded on March 19,2025 at 10:12:33 UTC from IEEE Xplore. Restrictions apply.
B. Low-speed mode: Determination of dc voltage and primary Replacing ipd , ipq and ϕp using Eqs. (26)-(28):
stator flux xp xs rs xp rs 2τe
vsd = wr (xps − )ip sinδ − ip cosδ +
In the low-speed mode, the primary stator is connected to a xps xps xps 3ip sinδ
dc source. The goal is to find the dc source voltage (vdc ) and xp xs r s xp xs 2τe
vsq = wr (xps − )ip cosδ + ip sinδ + wr
low-speed mode primary stator flux (ϕp ), which minimizes the xps xps xps 3ip sinδ
converter voltage rating for a required torque. The required 
2 + v 2 becomes
The converter voltage vs = vsd
low-speed torque is chosen as 75% of the maximum high- sq

speed mode torque. Let the A phase of primary stator be 

vs = awr2 + bwr + c (31)
connected to the positive terminal of dc source, and B and
C phases are connected to the negative terminal. Therefore where
2 1 1 xp xs xs 2τe 2
vAN = vdc ; vBN = vdc ; vCN = vdc (23) a =((xps − )ip )2 + ( ) +
3 3 3 xps xps 3ip sinδ
Using the abc − αβ transformation in Eq. (23), we get 4τe xs xp xs cosδ
(xps − )
2 3 xps xps sinδ
vpα = vdc ; vpβ = 0 (24)
3 b =2τe rs
Since the primray stator is connected to a dc source, the rs x p 2 rs 2τe 2 4τe rs 2 cosδ
primary stator current is c =( ip ) + 2( ) − xp ( )
xps xps 3ip sinδ 3 xps sinδ
2 (32)
ip = vdc rp (25)
3 Using the chosen BDFRM parameters, it is observed that the
Figure 6 shows the primary stator flux and voltage steady- first term awr2 in Eq. (31) dominates. This is particularly true
state space vectors. The d and q-axis primary stator winding at high speeds (wr > 0.3 p.u.), where the voltage rating will
currents are expressed as be decided. Thus, in order to minimize the converter voltage
ipd = ip cosδ (26) rating, the a function must be minimized, while the primary
stator current ip and flux angle δ are the constrained variables.
ipq = ip sinδ (27)
The upper bound on the angle δ for steady operation is 90◦ .
where δ is the angle between the primary stator voltage and The primary stator current rating must not be exceeded, and the
primary stator flux space vectors. Using Eqs. (4), (9) and (27), secondary stator current must not surpass the converter current
torque in this mode is calculated as rating (derived in the high-speed mode), both for steady state
3 and during transients.
τe = ip ϕp sinδ (28) To keep the dc primary stator current within its r.m.s rating,
2
The objective is to express the converter voltage in terms of 1
ip ≤ √ (33)
primary stator current ip , flux angle δ, and minimize it while 2
keeping both stator currents in bounds. Using Eqs. (1)-(4), the
Using Eqs. (1), (2) and (28), the secondary stator currents are
converter voltage as described by Eqs. (7) and (8) is expressed
expressed as
as
xp xs rs x p rs 2τe xp
vsd = wr (xps − )ipq − ipd + ϕp (29) isd = − ip cosδ (34)
xps xps xps 3xps ip sinδ xps
xp xs rs x p xs xp
vsq = wr (xps − )ipd + ipq + wr ϕp (30) isq = ip sinδ (35)
xps xps xps xps
Therefore, for the secondary stator current is to remain within
bounds in steady state,
2τe xp xp
( − ip cosδ)2 + ( ip sinδ)2 ≤ Is2rated (36)
3xps ip sinδ xps xps
The converter current must also be within its limit during the
transients. A step change in torque demand of τe from zero
can be considered as the worst-case scenario in this regard.
The transient per-unit d and q-axis primary voltage equations
in the w = wp reference frame are given by
1 dψp
vpd = rp ipd + (37)
wB dt
Fig. 6. Low-speed mode: Steady-state orientation of the primary stator 1 dδ
flux, voltage, and current space vectors in the primary stator ABC winding vpq = rp ipq + ψp (38)
reference frame. wB dt

4476

Authorized licensed use limited to: Indian Institute of Technology Patna. Downloaded on March 19,2025 at 10:12:33 UTC from IEEE Xplore. Restrictions apply.
With an initial zero torque demand, the angle δ is zero (Eq.
(28)), implying the primary stator flux and dc source voltage
vectors are aligned along the primary stator A-phase axis. So
2vdc
ipd = = ip (39)
3rs
Using Eqs. (1), (39) and (28), the d-axis converter current is
2τe xp
isd = − ip (40)
3xps ip sinδ xps
Equation (38) shows that the angle δ cannot change instantly.
Using Eqs. (2) and (27), a step change in torque will require
a q-axis converter current
xp
isq = ip sinδ (41)
xps
Using Eqs. (40) and (41), the limit on the d and q-axis
Fig. 7. Converter voltage rating of 0.65 p.u. can enable an operating speed
converter current in the dynamic state, is written as till 1.52 p.u. for the chosen BDFRM.
2τe xp xp
( − ip ) 2 + ( ip sinδ)2 ≤ Is2rated (42)
3xps ip sinδ xps xps
1 This equation yields a transition speed (wT ) = 0.5 p.u. and
Since ip ≤ √ , ϕp ≤ 1, sinδ ≤ 1, the maximum torque the power converter voltage requirement (vP E ) = 0.65 p.u. at
2
possible in low-speed mode is 1.06 p.u. (Eq. (28)), whereas it this transition speed. The maximum rotor speed (wrmax ) is
was calculated as 1.34 p.u. in high-speed mode. For this study, computed by equating the converter voltage equation in high-
the low-speed mode torque is taken as 75% of the maximum speed mode for positive torque (Eq. (22)) to vP E , i.e.,
high-speed mode torque, i.e., equal to 1 p.u. The function a

in Eq. (32) is minimized while satisfying the constraints in vshigh |acc = 1.383wr2 − 2.7wr + 1.317 = 0.65 (47)
Eqs. (33), (36), and (42). For the given BDFRM, the values
of primary stator current ip and flux angle δ achieved are 0.707 The maximum rotor speed obtained by solving the above
p.u. and 80.9◦ , respectively. The primary stator dc voltage quadratic function is 1.52 p.u. Figure 7 shows the converter
and flux are then calculated using Eqs. (25) and (28). The voltage requirement for the entire speed range for both positive
secondary stator currents in the low-speed mode are then and negative torque, along with the transition and maximum
computed using Eqs. (1) and (2) while the voltage is calculated rotor speed as calculated above.
using Eqs. (3), (4), (7) and (8), and is given by
 If the converter voltage rating is computed using a conven-
vslow |acc = 1.507wr2 + 0.051wr + 0.0014 (43) tional drive architecture (primary stator always connected to
the ac grid), the rating obtained is 1.15 p.u. for the entire-speed
A similar analysis is performed for the deceleration or the
range, as shown in Fig. 8. The switched-drive architecture thus
maximum negative torque. The converter voltage equations for
reduces the converter voltage rating by about 44%.
the example BDFRM, computed for maximum negative torque
in the high-speed mode and 75% of the maximum negative
torque in the low-speed mode, are given by

vslow |dec = 1.507wr2 − 0.051wr + 0.0014 (44)

2
vshigh |dec = 1.585wr − 3.237wr + 1.654 (45)
C. Determination of transition, maximum rotor speed and
converter voltage rating
The final step to calculate the converter voltage rating is to
find the transition speed. To minimize the required converter
voltage rating for operating the drive under all conditions, a
mode transition speed is chosen such that required converter
voltage in low-speed mode for maximum positive torque (ac-
celeration) equals that in high-speed mode for the maximum
negative torque (deceleration). From Eqs. (43) and (45),

vslow,acc |wT = 1.507wr2 + 0.051wr + 0.0014

= 1.585wr2 − 3.237wr + 1.654 = vshigh,dec |wT Fig. 8. Converter voltage rating of 1.15 p.u. is required for the conventional
(46) drive architecture for the chosen BDFRM.

4477

Authorized licensed use limited to: Indian Institute of Technology Patna. Downloaded on March 19,2025 at 10:12:33 UTC from IEEE Xplore. Restrictions apply.
D. Power-electronic converter power rating
The converter VA rating (including both the active and
reactive power) is calculated using the computed current and
voltage rating, considering both acceleration and deceleration.
Pconverter = 3vP Erms iP Erms = 1.5vP E iP E = 0.975p.u.
(48)
Maximum shaft power is found by using the maximum rotor
speed and rated torque as
Pshaf tmax = τerated wrmax = 1.34 ∗ 1.52 = 2.04p.u. (49)
The ratio of the power-converter VA rating to the rated shaft
power is 0.975/2.04 = 0.48. This number is around 0.33
for a DFIM [2]. The higher leakage flux, which demands
Fig. 10. Designed drive torque-speed capability of the chosen BDFRM using
much more reactive power flowing through the converter, switched-drive architecture with a propulsion-load torque.
increases the converter rating for a BDFRM. Figure 9 shows
the sharing of active and reactive power between the primary TABLE I
and secondary stators, while the machine operates in full-speed C OMPUTED P OWER C ONVERTER S PECIFICATIONS FOR THE EXAMPLE
range. The maximum active and reactive power handled by the BDFRM
converter is 0.76 p.u. and 0.6 p.u. respectively. The primary Drive Specifications Switched architecture Conventional architecture
stator winding power factor is 0.92 in high-speed mode. Converter current rating 62.5 A 62.5 A
Figure 10 shows the steady-state torque-speed capability of Converter voltage rating 201 V 357 V
the proposed BDFRM drive. A typical propulsion load torque, Converter power rating 18.9 KVA 33.4 KVA
Maximum torque 330.5 N-m 330.5 N-m
which is proportional to the square of the rotor speed is also Operating speed range 0 - 1140 rpm 0 - 1460 rpm
shown. In this case study, the low-speed torque was chosen as Maximum Shaft power 39.6 KW 50.7 KW

75% of the high-speed mode. This looks sufficiently larger

than the load-torque. Therefore, the same study can be carried
for a lower value of the low-speed mode torque, reducing even
further the converter voltage and the power rating. Table 1
summarizes the obtained results and outlines the benefit of a
switched-drive architecture over a conventional one.

E. Synchronizer: Transition between low and high-speed

modes
The primary stator flux ϕp is set by the grid in high-speed
mode, whereas in low-speed mode, it is set to a value necessary
for minimizing the converter voltage requirement. Switching
(a)
between these two modes can cause large perturbations in the
primary stator flux. An appropriate choice of transition instant
is needed to minimize the primary stator flux perturbations
during the mode transition [10].
1) Low to high-speed mode transition: During the low to
high-speed mode transition, Vpd and Vpq are the d and q-
axis components, respectively, of the dc source voltage in the
primary stator flux reference frame, as shown in Fig. 11(a).
The ac voltage vector Vac represents the to-be-connected ac
grid supply. For minimal perturbation of the primary stator
flux magnitude, dϕp /dt must be equal to zero during mode
transition. Equation (37) implies that the d-axis dc source
voltage must be equal to the d-axis component of the incoming
ac voltage (ignoring the effect of the primary stator resistance).
(b)
The two time instances where this occurs are t = t2 and t3
Fig. 9. Sharing of (a) active power and (b) reactive power between primary in Fig. 11(a). Also, the primary stator flux frequency must
and secondary stators. transition from zero to the ac grid supply frequency, which is

4478

Authorized licensed use limited to: Indian Institute of Technology Patna. Downloaded on March 19,2025 at 10:12:33 UTC from IEEE Xplore. Restrictions apply.
 the machine inertia. The sharing of active and reactive power
among the primary and secondary stator is shown in Fig. 14.
These graphs are consistent with the analytical ones in Fig. 9.

(a) (b)

Fig. 11. Determination of the correct instant of transition based on the existing
primary stator voltage, incoming source voltage, and primary stator flux space
vector (a) low to high-speed mode transition (b) high to low-speed mode
transition.

governed by Eq. (2). A positive q-axis component of the ac

grid voltage drives the primary stator flux frequency towards
the ac grid supply frequency. Thus, t = t3 represents the best
time to switch to high-speed mode.
2) High to low-speed mode transition: Assuming that the
machine is operating at steady state in high-speed mode, Vpd Fig. 13. Rotor speed as a function of time when a step command of 1140
and Vpq are the d and q-axis components, respectively, of the rpm is given at t = 0.5 s.
ac voltage in the primary stator flux reference frame, as shown
in Fig. 11(b). Using the same argument as in the low to high-
speed mode transition – matching the d-axis component of
the incoming dc voltage with Vpd and ensuring that the q-axis
component of the dc voltage is negative or zero – makes t3
the optimal switching instant from high to low-speed mode.
Figure 12 shows the effect of the synchronizer on the primary
stator flux when the mode transition occurs.

IV. S IMULATION R ESULTS

The simulations are performed using the chosen BDFRM
dynamic model in MATLAB/Simulink. A maximum rotor
speed command of 1.52 p.u. (=1140 rpm) is given at t = 0.5
s, and the rotor speed is observed as in Fig. 13. While the
calculated transition speed is 0.5 p.u. (=375 rpm), the actual
transition speed is seen as 410 rpm due to synchronizer action.
This difference depends upon the ac source frequency and (a)

(b)

Fig. 12. Synchronizer reduces the perturbations in the primary stator flux Fig. 14. Simulation results: Sharing of (a) active power and (b) reactive power
during the transition between modes. between primary and secondary stators.

4479

Authorized licensed use limited to: Indian Institute of Technology Patna. Downloaded on March 19,2025 at 10:12:33 UTC from IEEE Xplore. Restrictions apply.
 The converter A-phase current and voltage are plotted with V. C ONCLUSION
respect to time to illustrate the reduced power electronics This paper presented a switched-drive architecture for BD-
requirement, while operating in the complete speed range FRM, similar to that for a DFIM. While the primary sta-
for positive torque. The secondary stator/ converter current tor windings are always connected to the ac source in the
frequency ws is given by [11] traditional architecture, they are either shorted or connected
to a dc source at low-speed in the proposed architecture.
ws = pr wrm − wp = wr − wp (50)
This effectively reduces the converter voltage requirement. An
Since wp = 0 p.u. in low-speed mode, ws is proportional to example BDFRM was taken to show the drive-design process.
the rotor electrical speed wr . In high-speed mode, wp = 1 p.u., While operating at rated torque capability in a wide speed
and the magnitude of ws decreases until the synchronous speed range, the candidate BDFRM required the power-electronic
(wr = 1 p.u.). It increases beyond the synchronous speed, as converter rating to be 48% of the rated shaft power. The
seen from Figs. 15 and 16. To produce maximum torque in minimum power rating was computed using the current rating
high-speed mode, the phase-current magnitude is the same as derived in the high-speed mode and voltage rating in the
the secondary stator current rating of 62.5 A, as discussed low-speed mode. Although the drive was designed for a
in Section III. The converter phase-voltage magnitude, as propulsion type load, this method can be iterated for other
shown in Fig. 16 grows in the low-speed mode until the desirable load-torque profiles. The switched-drive architecture
transition speed is reached, reduces in high-speed mode until opens opportunity to use BDFRM in many wide-speed range
the synchronous speed is reached, and increases beyond that applications, such as propulsion, with the combined advantage
speed, as observed earlier in Fig. 7. of reduced power converter requirement and better reliability.
ACKNOWLEDGMENT
This research was supported by NASA through Universities
Space Research Association. The authors would also like to
thank Mr. Raymond F. Beach for his comments and sugges-
tions.
R EFERENCES
[1] O. of Energy Efficiency and renewable energy (EERE), “Next generation
electric machine: Megawatt class motors,” 2015.
[2] A. Banerjee, M. S. Tomovich, S. B. Leeb, and J. L. Kirtley, “Power
converter sizing for a switched doubly fed machine propulsion drive,”
IEEE Transactions on Industry Applications, vol. 51, no. 1, 2015.
[3] A. Banerjee, A. H. Chang, K. N. Surakitbovorn, S. B. Leeb, and
J. L. Kirtley, “Bumpless automatic transfer for a switched-doubly-fed-
machine propulsion drive,” IEEE Transactions on Industry Applications,
vol. 51, no. 4, pp. 3147–3158, 2015.
[4] E. National Academies of Sciences, Medicine, et al., Commercial
aircraft propulsion and energy systems research: reducing global carbon
emissions. National Academies Press, 2016.
Fig. 15. Simulation results: Converter A-phase current with time showing the [5] S. Agrawal, A. Banerjee, and R. Beach, “Brushless doubly-fed reluc-
phase current rating of 62.5 A. tance machine drive for turbo-electric distributed propulsion systems,”
presented at AIAA/IEEE Electric Aircraft Technologies Symposium
(EATS), Cincinnati, Ohio, 2018.
[6] J. L. Kirtley, A. Banerjee, and S. Englebretson, “Motors for ship
propulsion,” Proceedings of the IEEE, vol. 103, no. 12, pp. 2320–2332,
2015.
[7] M. G. Jovanovic, R. E. Betz, and J. Yu, “The use of doubly fed reluc-
tance machines for large pumps and wind turbines,” IEEE Transactions
on Industry Applications, vol. 38, no. 6, pp. 1508–1516, 2002.
[8] R. Betz and M. Jovanović, “Introduction to the space vector modeling
of the brushless doubly fed reluctance machine,” Electric Power Com-
ponents and Systems, vol. 31, no. 8, pp. 729–755, 2003.
[9] F. Zhang, L. Zhu, S. Jin, W. Cao, D. Wang, and J. L. Kirtley, “Developing
a new svpwm control strategy for open-winding brushless doubly fed
reluctance generators,” IEEE Transactions on Industry Applications,
vol. 51, no. 6, pp. 4567–4574, 2015.
[10] A. Banerjee, M. S. Tomovich, S. B. Leeb, and J. L. Kirtley, “Control
architecture for a switched doubly fed machine propulsion drive,” IEEE
Transactions on Industry Applications, vol. 51, no. 2, pp. 1538–1550,
2015.
[11] Y. Liao, L. Xu, and L. Zhen, “Design of a doubly fed reluctance
motor for adjustable-speed drives,” IEEE Transactions on Industry
Applications, vol. 32, no. 5, pp. 1195–1203, 1996.
Fig. 16. Simulation results: Converter A-phase voltage with time showing the
phase voltage rating of 200 V.

4480

Authorized licensed use limited to: Indian Institute of Technology Patna. Downloaded on March 19,2025 at 10:12:33 UTC from IEEE Xplore. Restrictions apply.