1058 IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 17, NO.

6, NOVEMBER 2002

Mitigation of Saturation in Dynamic Voltage Restorer

Connection Transformers
Chris Fitzer, Atputharajah Arulampalam, Mike Barnes, and Rainer Zurowski

Abstract—During the transient period at the start of a voltage

sag, a DVR injection transformer can experience a flux-linkage
that is up to twice its nominal steady-state value. In order to
prevent the transformers from saturating it is normal to choose
a rating flux that is double that of the steady-state limit. An
alternative method is to limit the flux-linkage during the transient
switch-on period, thus preventing saturation. It is shown through
both simulation and experimental results that an adaptive form
factor can be applied to the DVR injected voltage, which mini-
mizes the disturbance seen by a sensitive load, while at the same Fig. 1. Equivalent electrical circuit diagram of the DVR.
time preventing saturation.
The proposed method removes the need for rating the series in-
jection transformers for the DVR transient switch-on period, and
therefore removes the redundancy normally associated with their
steady state operation. In economic terms, this may reduce the total
cost of a DVR system, thus making it a more attractive solution for
voltage sag mitigation.
Index Terms—Active power injection, DVR, sensitive loads,
space vector control, transformer saturation, voltage sags.

I. INTRODUCTION

V OLTAGE sags can be defined as momentary reductions in

supply voltage, lasting from a few milliseconds to a few
cycles. Due to the increased use of sensitive automated manu-
facturing and IT equipment, so-called sags or dips are becoming
an increasing problem.
When a sag causes an uncontrolled shutdown in an IT system
or industrial process the cost of the interruption may be dis-
proportionately large compared with the apparent severity of
the cause. This has been the fundamental reason for developing
mitigation devices, which effectively buffer (completely or par-
tially) the supply and a potentially sensitive load. A well-es-
tablished device in the information technology industry is the
“un-interruptible power supply” or “UPS” for short. This typi-
cally is a device connected between the load and supply, de-cou-
pling both, thus offering protection for all types of supply fluc-
tuations, whether they be short duration sags, or longer term out-
ages. For industry, where loads are normally rated in the MW
range, the UPS device is no-longer the most economic solu-
tion, as such a device would be required to have a power rating
greater or equal to that of the load. The majority of voltage re-
ductions are typically to about 0.7 pu of a nominal 1.0 pu supply
Manuscript received October 26, 2001; revised May 28, 2002. Recommended
by Associate Editor J. H. R. Enslin.
C. Fitzer, A. Arulampalam, and M. Barnes are with the Manchester Centre
for Electrical Energy (MCEE), Department of Electrical Engineering and
Electronics, University of Manchester Institute of Science and Technology
(UMIST), Manchester M60 1QD, U.K. (e-mail: mike.barnes@umist.ac.uk).
R. Zurowski is with the System Planning Department, SIEMENS AG, EV SE
NC4, Erlangen D91050, Germany.
Digital Object Identifier 10.1109/TPEL.2002.805584
0885-8993/02$17.00 © 2002 IEEE
Fig. 2. Typical installation location of a dynamic voltage restorer.
voltage and last for anything from a few milliseconds to a few
seconds [1]. It has been shown [1] that mitigation devices that
support rather than completely buffer the supply from a sen-
sitive load need only be a fraction of the power rating of the
load and have much lower energy storage requirement com-
pared with the traditional UPS. It has also been shown [1] that
series connected voltage support devices can have a lower rating
than the shunt connected variants. A series device or “Dynamic
Voltage Restorer” (DVR) [1]–[6], [8], [9] as it is more com-
monly known, injects a voltage in series with the supply to elim-
inate any voltage sags seen by a load (Fig. 1).
Fig. 2 highlights the problem a DVR is used to solve. Faults
on adjacent feeders cause the receiving end voltage seen by a
load to reduce. A single-phase to ground short is a common
fault. The reduction seen by the DVR in such a case is a two-
phase dip with phase angle shift due to the “delta–wye” trans-
former interconnection. Depending on the power factor of the
load, the DVR is required to supply both real and reactive power
so as to provide complete compensation for the dip.
A fast response feed-forward control system has already been
presented in [3]–[5] where a digital controller can quickly de-
FITZER et al.: MITIGATION OF SATURATION 1059

problems associated with this type of system, which still require

further investigation [8].
To correct for the effects of voltage sags, DVR’s are required
to inject both positive and negative sequence voltage vectors.
This unfortunately presents problems with unbalanced fluxes
if conventional three-limb injection transformers are used. To
avoid this, which mainly would be a result of a nonzero flux
in the center or common limb, three separate injection trans-
formers are utilized. This allows the flux-linkage in each to be
dealt with separately and remains true regardless of the external
configuration, star, or delta. The injection voltage across the pri-
mary coil (DVR side) of one of the three injection transformers
can be represented by

(2)

Fig. 3. Compensation strategy of the DVR for voltage sags with a phase jump.
If the voltage is injected ( ) at an angle , then the flux-
 : Angular position of synchronous frame.
: Angle of supply vector. V : linkage in the transformer at a later angle is given by
Actual supply vector. V : “d” coordinate of the supply. V : “q ” coordinate of
the supply. V : “” coordinate of the supply. V : “ ” coordinate of the supply.
V : Compensation vector required to correct for fault. V : Reference
vector.
termine the injected voltage that is required to ride-though a
voltage sag. The control topology is based around a rotating ref-
erence frame [ ] locked on to the fundamental positive se-
quence component of the supply voltage, which is represented
in a fixed [ ] frame [3]–[5], [7], [10]–[12]. and
are generated by
(1)
Fig. 3 illustrates a typical instantaneous snapshot of the con-
trol system vectors when a single-phase outage on a nearby
feeder causes both positive and negative sequence error infor-
mation to be generated. In the plane this is seen as a magni-
tude dip (positive sequence error) of the phasor , along with a
100 Hz superimposed oscillation (negative sequence error). At
any instant a compensation vector (or phasor) can be calculated
by subtracting the actual vector from a reference vector, gen-
erated by a phase-locked-loop (PLL). The error (or compensa-
tion) vector can then be used in various PWM schemes. In [3]
space-vector-pulse-width-modulation (SVPWM) was chosen as
the DVR inverter control pattern. If a wye–wye interconnection
transformer is used in place of a delta–wye (Fig. 3), then the
control and PWM scheme can be modified to so as to deal
with any zero-sequence information generated [10].
II. CONVENTIONAL METHODS TO LIMIT FLUX
The aim of this paper is to highlight and propose a possible
solution to one of the transient switch-on issues surrounding
the use of series injection transformers as part of a dynamic
voltage restoration system. Transformer-less operation has been
suggested as a possible remedy, however there are a number of
(3)
If and , then
(4)
As illustrated by (4) and Fig. 4, if , or the start of a voltage sag,
occurs at (zero crossing), the flux-linkage during the
first few cycles or more, depending on the transformer loading,
can be as much as double its steady-state value. This is due to
an inherent dc offset that is created when the voltage applied, or
injected, does not commence at a voltage maxima or minima. If
the injected voltage amplitude is close to the maximum voltage
rating of the injection transformer, then in the first few cy-
cles can be as much as double the maximum rating and
this could result in the transformer saturating. The above re-
mains true provided that the interval between successive sags
is long enough to allow transformer fluxes to decay to zero.
It is assumed in this paper that apart from the flux-linkage is-
sues there are no other transient delays in bringing the DVR on-
line. In practice this can be achieved by injecting a zero voltage
during “normal” or “no-fault” conditions, rather than short-cir-
cuiting the DVR with an external crowbar. In the investigation
into transformer saturation a small voltage vector is injected
during “no-fault” operation to correct for the slight fundamental
volt drop across the filter impedance.
A starting point in eliminating saturation would be to rate the
injection transformers for the worst-case scenario, which would
be double the steady-state flux-linkage rating. For large loads
(in the MW range), doubling the transformer rating involves a
significant increase in the cost, which may result in the DVR not
being cost effective. A more desirable option is to avoid the dou-
bling of the flux-linkage during the first few cycles altogether.
1060
Fig. 4. Idealized representation of injected voltage and transformer flux for a
sag of 100 V (0.7 pu of a 230 Vrms nominal supply).
Apart from rating the transformer for the transient switch on
period, another qualitative method would be to only switch the
transformers on at a voltage peak. Unfortunately this could lead
to no voltage being injected by the DVR for up to half a cycle,
which is undesirable.
III. NOVEL METHODS TO LIMIT FLUX
An effective method of preventing saturation can be achieved
by applying a form factor to the injected voltage, which
would have the effect of manipulating the injected voltage in
a controlled manner, until the first zero crossing is detected.
This method can be compared to the gradual increase used
in some conventional transformer systems, but here a much
faster ramp-up time is used, i.e., in milliseconds rather than
cycles. The rest of this paper focuses on the development and
verification of such algorithms.
A. Constant Form Factor
Consider a constant factor applied for a fixed time . If a
voltage sag occurs at an angle , the form function which
multiplies the injected voltage for a further angle , can be rep-
resented by the series
(5)
This leads to a residual flux-linkage in the injection transformer
after the angle of , with the general flux-linkage at any
IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 17, NO. 6, NOVEMBER 2002
subsequent angle radians later given by
(6)
where can be evaluated [Appendix (A16)–(A19)] to be
(7)
Referring back to (4), one method to prevent saturation is to
entirely remove or prevent a DC flux-linkage component .
This can be achieved by the solution of (7). Analytical methods
give one solution as
for
This results in only half the required voltage being injected
for the first half-cycle. The solution correctly positions the
flux-linkage of the injection transformer so that subsequent full
voltage may be applied after the first half cycle, as has
been removed, Fig. 5.
This approach is a better solution than the previous quan-
titative methods, as the dc offset and the need for rating the
transformers for the transient switch-on period is removed. This
method is, however, still far from ideal, as the load side volt-
ages are not fully compensated when the flux-linkage limiter
(form factor) is operational. Certain types of industrial drives
and process equipment may be intolerant of this effect [1], [13],
[14].
B. Adaptive Form Factor
Depending on the amplitude of a voltage applied to a trans-
former, in order to prevent it from saturating, it is not neces-
sary to completely remove the D.C offset. This means for sags
less severe than the nominal rating of the DVR, a proportion-
ately larger voltage can be initially injected. From the above dis-
cussion, a DVR may inject up to half its nominal rated voltage
in the first half cycle without exceeding the peak transformer
flux-linkage rating. For lesser sags, the higher voltage may be
used. This results in a residual flux-linkage offset but the peak
flux is not exceeded. This result leads to the development of
more sophisticated form factors, which aim only to prevent the
maximum flux-linkage excursion from being exceeded. This
helps to reduce the voltage disturbance seen by a sensitive load
during the initial stage of the voltage sag, while still maintaining
correct transformer operation.
The adaptive form factor developed in this paper predicts the
amount the flux-linkage maximum will be exceeded during the
FITZER et al.: MITIGATION OF SATURATION 1061

Fig. 6. DVR injection voltage (one-of-three phases).

If then , setting and suitably rear-

ranging equation (10) to make “ ” the subject, will yield the
appropriate scaling factor by which the injected voltage must
be multiplied

(12)

Referring to Fig. 7, as it is the positive portion of the injected

voltage that may cause the start-up flux-linkage to exceed the
maximum value, only the positive portion of the waveform re-
quires scaling. Equation (10) is now changed to represent this
Fig. 5. Flux-linkage with the dc offset removed and corresponding load
voltages for a 0.7 pu balanced dip.

(13)
initial stage of the voltage sag, if at all. The basis of the adaptive
In order to determine the “ ” value, needs to be predicted
form factor is
(13). If then as previously “ ” requires
scaling
(8)

(14)
where can be evaluated as

(9) (15)

“ ” now replaces . Applied to Fig. 6 the formula be-
comes
(10)
As it is the negative portion of the injected voltage that may
cause the start-up flux-linkage to exceed the maximum value
in Fig. 6, only the negative portion of the waveform requires
scaling (10). The limit in the second integral (10) indicates
that after this angle ( ) in Fig. 6 the flux-linkage DC offset
will have reached its steady state value. In order to determine
“ ” value required, needs to be first predicted, or evaluated,
using
(11)
In order to be able to use equations (8)–(15), one must determine
the injected voltage amplitude , and the point on wave that
the voltage sag commences with reference to a cosine wave “ ,”
as above. Also if the magnitude of a voltage sag may change
within the first cycle or so compared with its initial value, “ ”
must be adjusted if the correct transformer operation is to be
ensured.
IV. SIMULATION RESULTS
A model of the proposed DVR system has been implemented
in the simulation package PSCAD/EMTDC. The overall circuit
layout and control topology, are illustrated in Fig. 8.
The simulation is based around a nominal load line voltage of
415 Vrms, which allows for direct comparisons with the experi-
mental results, Figs. 17–23. The system voltage and load power
rating, although significantly lower than in a practical DVR
1062
Fig. 7. DVR injection voltage (one-of-three phases).
Fig. 8. PSCAD circuit and system control topology.
Fig. 9. Supply voltage during a balanced dip (phase A, simulation).
system, still allow for feasible results to be obtained, which are
not overly damped.
A. Summary of the Results
The adaptive form factor operates on each of the three-phases
separately. Any type of sag, balanced or unbalanced, can be pre-
sented to the DVR system in order to verify the flux-limiting
process. The fault generator, Fig. 8, is set up to give a balanced
three-phase dip to 70% of a nominal 230 V(rms) supply, Fig. 9.
The fault occurs, appears to clear, then re-occurs again. The
flux-linkage limiter is programmed in such a way that when a
sag ends, it is assumed that the flux-linkage of the transformer
drops close to zero before another fault occurs.
IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 17, NO. 6, NOVEMBER 2002
Fig. 10. Voltage required to be injected to fully compensate for the dip (phase
A, simulation).
Fig. 11. The level of scaling required in order to prevent the injection
=
transformer (phase A, simulation) from saturating (1 no scaling required).
To prevent a load from exposure to the sag, the DVR in Fig. 8
must inject a voltage of correct magnitude and phase, Fig. 10.
The voltage appears free of unwanted harmonics, as it is an
internal analog signal developed by the control system block
set, as opposed to the actual injected voltage, which is a fil-
tered PWM waveform. The small voltage that is injected during
the no-fault condition is required in order to compensate for
the slight volt-drop across the filter impedance, during normal
operating conditions. This action ensures that the transformer
flux-linkage at the sag onset and the control assumed value both
match, which is important, as the controller is purely feed-for-
ward.
The form factor to be applied to the injected voltage is cal-
culated by the flux-linkage-limiter algorithm. The form factor
calculated is effectively a continuously changing variable by
which the injected voltage is multiplied, Fig. 11. As the magni-
tude of the injected voltage varies slightly with time, the scaling
required to prevent transformer saturation must also change,
Fig. 11. Scaling is only required when the DVR is going from an
off to on state. In the simulation this occurs twice. The second
fault occurs close to a zero crossing, hence the severe scaling
as expected. The step appearance of Fig. 11 is due to the flux-
linkage-limiter algorithm only operating on the analog signal
Fig. 10 at a sampling rate of 1 ms.
The modified injected voltage is shown in Fig. 12. This
produces a maximum flux-linkage excursion of not more than
0.32 Wb-turns, Fig. 13. The peak flux-linkage limit varies
depending on the magnitude of voltage dip a DVR is required
to compensate for. For this simulation a flux-maximum of 0.32
FITZER et al.: MITIGATION OF SATURATION 1063

Fig. 12. Injected voltage after scaling (phase A, simulation).

Fig. 13. Limited flux-linkage envelope magnitude (phase A, simulation).

Fig. 15. (a) Ideal and average 3-phase rms voltage. (b) Percentage deviation
from the ideal (simulation).

Fig. 14. Load voltage (phase A, simulation).

Wb-turns was specified which corresponds to a steady-state

Fig. 16. 10 KVA medium power hardware rig.
injected voltage of 100 V per phase.
The voltage (phase A) as seen by a sensitive load is shown in
It is important for a sensitive load that the supply be relatively
Fig. 14. Even with the flux-linkage-limiter operational, the DVR
free from any voltage fluctuations. It has been shown [1] that
still has sufficient compensation ability. When the second fault
some loads may trip or fail if the 3-phase rms supply reduces
occurs on phase A, the injected voltage required to compen-
excessively from a nominal 100% supply. Fig. 15 indicates that
sate has a magnitude that is equal to the maximum steady-state
the DVR can boost the receiving end voltage while and still
voltage rating of the transformer (100 v), but also commences
maintain correct transformer operation.
at a zero crossing. As explained in Section II this type of fault
produces the greatest flux dc offset. The only way in which to
V. EXPERIMENTAL RESULTS
prevent the transformers from saturating in this case is to halve
the injected voltage for a few milliseconds, Fig. 11. Due to this An experimental rig with potential rating of 10 kVA has
the load experiences a modest disturbance on phase “A” while been developed [3] to verify the control of the DVR in real
the flux-linkage-limiter is operational. The simulation results time. The control is realized via a pair of TMS320C4X DSP
however indicate this level of scaling is not required in all three processors.
phases at the same time, which results in the average 3-phase To introduce a fault, Fig. 16, a variable impedance is placed
rms voltage as seen by the load varying only slightly, Fig. 15. in series in-between the source and the delta–wye transformer.
1064
Fig. 17. Supply voltage during a balanced dip (phase A, experiment).
Fig. 18. Magnitude of the required injected voltage (phase A, experiment).
Fig. 19. Level of scaling required in order to prevent the injection transformer
=
(phase A, experiment) from saturating (1 no scaling required).
This configuration gives fault types that are comparable with
those created by the fault generator in Fig. 8. As with the simu-
lation arrangement the load is connected to a nominal 415 V rms
3-phase supply. The load is rated at 3 kW, which is sufficient to
test the DVR dynamic response. As with the simulation control,
the DSP/Control coding for the adaptive form factor operates
on each phase separately. The fault generator for the practical
rig is set up only to give one fault, as opposed to the simulation
where a fault occurs, appears to clear, then re-occurs again. The
following results illustrate the DVR’s ability to compensate for
a balanced three-phase fault.
A balanced three-phase fault occurs as illustrated in Fig. 17.
The magnitude of the ideal injected voltage required to com-
IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 17, NO. 6, NOVEMBER 2002
Fig. 20. Injected voltage after scaling (phase A, experiment).
Fig. 21. Limited flux-linkage envelope magnitude (phase A, experiment).
Fig. 22. Load voltage (phase A, experiment).
pensate for the dip is shown in Fig. 18, where it is apparent
that there is a gradual rise in the voltage magnitude level.
This is due to the response of the digital control system,
where the phase shifters and filters that are used to obtain
correct steady-state operation, degrade the dynamic response.
It is also noticeable that at no-fault conditions there is a
small internal injected DVR voltage when the actual injected
voltage across the transformers is approximately zero. This
is for the same reason as in the simulation where the DVR is
set to inject a small voltage, in order to compensate for the
volt drop across the transformer filter, which is created by
the load curren This action is necessary to ensure that at a
FITZER et al.: MITIGATION OF SATURATION
Fig. 23. (a) Ideal and average 3-phase rms voltage. (b) Percentage deviation
from the ideal (experiment).
sag on-set the actual transformer flux-linkage is sufficiently
small that is can be considered approximately zero.
The other results Figs. 19–23, illustrate a control response
that matches that seen in the simulation results. The apparent
noise on some of the signals is due to harmonics on the supply
voltage, which are caused by a nearby computer cluster. The
harmonics introduce small errors within the magnitude detec-
tion algorithm, as the equations used have been developed only
to deal with fundamental frequency information.
VI. CONCLUSION
This paper has proposed and verified a method that removes
the need for rating the series injection transformers of the DVR
for transient switch-on period, and therefore removes the redun-
dancy normally associated with their steady-state operation. In
economic terms, this may reduce the total cost of a DVR system,
thus making it a more attractive solution for voltage sag mitiga-
tion.
The simulation and experimental results illustrate that the av-
erage rms 3-phase voltage and the individual transformer flux-
linkages can be maintained within specified limits via a feed-
forward digital controller.
It is possible for particular loads to be sensitive to the varia-
tions in rms voltage caused by the adaptive form factor. The vari-
ations can be made less severe by using transformers of higher
flux ratings and adjusting the control such that it limits the in-
jected voltage proportionally less.
1065
APPENDIX
Derivation of
(A16)
(A17)
(A18)
(A19)
REFERENCES
[1] M. H. J. Bollen, Understanding Power Quality Problems; Voltage Sags
and Interruptions. New York: IEEE Press.
[2] Gyugyi, et al., “Apparatus and method for dynamic voltage restoration
of utility distribution networks,” U.S. Patent 5 329 222, July 12, 1994.
[3] C. Zhan, V. K. Ramachandaramurthy, A. Arulampalam, C. Fitzer, S.
Kromlidis, M. Barnes, and N. Jenkins, “Dynamic voltage restorer based
on voltage space vector PWM control,” IEEE Trans. Ind. Applicat., vol.
37, pp. 1855–1863, Nov./Dec. 2001.
[4] C. Zhan, C. Fitzer, V. K. Ramachandaramurthy, A. Arulampalam, M.
Barnes, and N. Jenkins, “Software phase-locked loop applied to Dy-
namic Voltage Restorer (DVR),” in Proc. IEEE PES Winter Meeting,
2001, pp. 1033–1038.
[5] K. Haddad, G. Joos, and S. Chen, “Control algorithms for series voltage
regulators in faulted distribution systems,” in Proc. IEEE PESC’99
Conf., vol. 1, 1999, pp. 418–423.
[6] C. S. Chang, S. W. Yang, and Y. S. Ho, “Simulation and analysis of
Series Restorers (SVR) for voltage sag relief,” in Proc. IEEE PES Winter
Meeting, 2000.
[7] V. Kaura and V. Blasko, “Operation of a phase locked loop system under
distorted utility conditions,” IEEE Trans. Ind. Applicat., vol. 33, pp.
58–63, Jan./Feb. 1997.
[8] B. H. Li, S. S. Choi, and D. M. Vilathgamuwa, “On some crucial tech-
nologies for dynamic voltage restoration,” in Proc. 5th Int. Power Eng.
Conf. (IPEC), 2001, pp. 426–431.
[9] R. S. Weissbach, G. G. Karady, and R. G. Farmer, “Dynamic voltage
compensation on distribution feeders using flywheel energy storage,”
IEEE Trans. Power Delivery, vol. 14, pp. 465–471, Apr. 1999.
[10] P. Verdelho and G. D. Marques, “Four wire current-regulated PWM
voltage converter,” IEEE Trans. Ind. Electron., vol. 45, pp. 761–770,
Oct. 1998.
[11] H. Fujita and H. Akagi, “The unified power quality conditioner: The
integration of series and shunt active power filters,” IEEE Trans. Power
Electron., vol. 13, pp. 315–322, Mar. 1998.
[12] B. Singh, K. Al-Haddad, and A. Chandra, “A review of active filters for
power quality improvement,” IEEE Trans. Ind. Electron., vol. 46, pp.
960–971, Oct. 1999.
[13] J. C. Smith, J. Lamoree, P. Vinett, T. Duffy, and M. Klein, “The impact of
voltage sags on industrial plant loads,” in Proc. Int. Conf. Power Qual/:
End-Use Applicat. Perspectives (PQA-91), 2002, pp. 171–178.
1066
[14] L. Conrad, K. Little, and C. Grigg, “Predicting and preventing problems
associated with remote fault-clearing voltage dips,” IEEE Trans. Ind.
Applicat., vol. 27, pp. 167–172, Jan. 1991.
Chris Fitzer was born in Preston, U.K., in 1977. He
received the B.Sc. (with honors) degree in electronics
from the University of Central Lancashire, Preston,
in 1999 and is currently pursuing the Ph.D. degree at
the University of Manchester Institute of Science and
Technology, Manchester, U.K.
His research interests are digital control of
power quality devices, dynamic voltage restorers,
transformer saturation prevention, digital signal
processing, and hybrid filters.
Atputharajah Arulampalam was born in Sri Lanka
in 1971. He received the B.Sc.Eng. (with honors) de-
gree from the University of Peradeniya, Sri Lanka,
in 1997 and is currently pursuing the Ph.D. degree at
the University of Manchester Institute of Science and
Technology, Manchester, U.K.
Since 1998, he has been a Lecturer with the De-
partment of Electrical and Electronic Engineering,
University of Peradeniya. His research interests are
the applications of FACTS devices, power quality,
and converter configurations.
IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 17, NO. 6, NOVEMBER 2002
Mike Barnes received the B.Eng. and Ph.D. degrees
from the University of Warwick, U.K., in 1993 and
1998, respectively.
He has worked as a Research Associate on a
DTI/LINK project on low cost switched reluctance
drives and as a Lecturer at Warwick. For the last
five years he has been a Lecturer at the University
of Manchester Institute of Science and Technology,
Manchester, U.K., where his research interests have
covered power electronics applied to power systems,
machine drives, and photovoltaics.
Rainer Zurowski (1965) received the Dipl.Ing.
degree in electrical engineering from Friedrich
Alexander University, Erlangen, Germany, in 1991
and the Dr.Ing. degree from Ruhr University,
Bochum, Germany, in 2001.
Since 1992, he has been with the System Planning
Department, SIEMENS AG, Erlangen. His main
activity is research and development of power
electronics based equipment for power quality
improvement in transmission and distribution
networks.