IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO.

4, JULY/AUGUST 2002 1123

Considerations of a Shunt Active Filter Based

on Voltage Detection for Installation
on a Long Distribution Feeder
Keiji Wada, Associate Member, IEEE, Hideaki Fujita, Member, IEEE, and Hirofumi Akagi, Fellow, IEEE

Abstract—This paper deals with a curious phenomenon referred Installation of an active or passive filter on a long-distance
to as the “whack-a-mole” that may occur in a long-distance distri- power distribution feeder may result in a strange phenomenon:
bution feeder having many capacitors for power-factor correction. voltage harmonics are mitigated at the point of installa-
The idea of whack-a-mole is that installation of an active or pas-
sive filter on the feeder makes voltage harmonics increase on some tion, whereas they are magnified on other busses where
buses, whereas it makes voltage harmonics decrease on other buses, no filter is connected. This phenomenon, referred to as the
especially at the point of installation. The distributed-parameter “whack-a-mole”1 in this paper, has been known among some
representation is applied to a simplified feeder, thus making it pos- researchers and engineers in the fields of power systems and
sible to perform analysis of the whack-a-mole. As a result, this anal- power electronics. However, nothing has been published or
ysis yields such a basic way as to avoid the whack-a-mole. More-
over, both theory and experiment clarify that installation of the presented in the literature on the analysis of this phenomenon.
active filter acting as a harmonic terminator on the end bus of the The authors have presented a shunt active filter based on
feeder can damp out harmonic propagation throughout the feeder voltage detection for the purpose of harmonic termination of
without causing any whack-a-mole. a radial feeder [6]–[8]. Installing the active filter on the end
Index Terms—Distributed-parameter representation, harmonic bus of the feeder makes it possible to damp out harmonic prop-
damping, harmonic propagation, power distribution systems, agation. The purpose of this paper is not only to analyze the
shunt active filters. whack-a-mole in a long-distance distribution feeder, but also to
discuss the installation effect of a voltage-detection-based shunt
I. INTRODUCTION active filter characterized by acting as a harmonic terminator.
It is assumed that line inductors and capacitors are uniformly

R ECENTLY, harmonic resonance or harmonic propagation

between capacitors for power-factor correction and
inductors present in cables, lines, transformers, and so on
dispersed in a radial trunk feeder having no branch feeders, al-
though these inductances and capacitance in a real system do not
have to be equal in size. Invoking such a daring assumption al-
has multiplied voltage and current harmonics in industrial lows us to use the “distributed-parameter representation” or the
power plants and utility power distribution systems [1]–[5]. so-called “transmission line theory,” thus making it possible to
Reference [1] presents measured and calculated values for achieve theoretical analysis. The reason why the whack-a-mole
voltage and current multiplication due to harmonic resonance occurs in the simplified radial feeder is clarified by both theo-
at a medium-sized industrial plant. References [3] and [4] retical analysis based on the transmission line theory and com-
describe how harmonic propagation actually occurs in the puter simulation using a PSpice software package. As a result,
downtown area of 6.6-kV utility power distribution systems this analysis leads to the following essentials.
in Japan under light-load conditions at night. Reference [5]
• Installation of the active filter on the end bus does not
points out that the fifth-harmonic voltage at the end bus is
cause any whack-a-mole at a specified harmonic fre-
actually magnified 3.5 times as large as that at the beginning
quency when the feeder length is shorter than a quarter of
bus in a 6.6-kV 17-km-long utility power distribution feeder
wavelength at the frequency.
having power-factor-correction capacitors with a total capacity
• As long as the active filter has an adequate gain re-
of 245 kVA.
lated to the characteristic impedance of the feeder, no
whack-a-mole occurs, irrespective of the feeder length.
Paper IPCSD 02–012, presented at the 2001 Industry Applications Society These general understandings of the whack-a-mole in the
Annual Meeting, Chicago, IL, September 30–October 5, and approved for pub- simplified radial feeder give broad hints on how to avoid the
lication in the IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS by the Indus- whack-a-mole in a real feeder. In addition, these findings are
trial Power Converter Committee of the IEEE Industry Applications Society.
Manuscript submitted for review November 1, 2001 and released for publica- verified by experimental results obtained from a three-phase
tion April 15, 2002. 100-V 5-kVA laboratory system.
K. Wada is with the Department of Electrical Engineering, Tokyo
Metropolitan University, Tokyo 192-0397, Japan (e-mail: [email protected] 1The term “whack-a-mole” originally meant an indoor amusement for chil-
u.ac.jp). dren. The activity can be described as follows. As soon as a child whacks a mole
H. Fujita and H. Akagi are with the Department of Electrical and Electronic emerging from a hole with a wooden hammer, the mole goes back into the hole.
Engineering, Tokyo Institute of Technology, Tokyo 152-8522, Japan (e-mail: At the same time, another mole appears from another hole. Immediately, as the
[email protected]; [email protected]). child whacks it, another mole emerges from another hole. He repeats the play
Publisher Item Identifier 10.1109/TIA.2002.800584. until he gives it up.

0093-9994/02$17.00 © 2002 IEEE

1124 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO. 4, JULY/AUGUST 2002

TABLE II
RELATIONSHIP AMONG FREQUENCY, WAVELENGTH, AND FEEDER LENGTH

Fig. 1. Power distribution feeder (lumped-parameter model).

TABLE I
The characteristic impedance of a general feeder is given
CIRCUIT PARAMETERS by

(1)

where is a conductance per kilometer. Note that the conduc-

tance is excluded from Fig. 1 and Table I. The wavelength and
the propagation constant are defined as

(2)

(3)

Such a relationship as and exists in a real

feeder. Hence, and in (2) approximate

(4)
Fig. 2. Distributed-parameter model when a harmonic voltage source exists
K
on bus 1, and the active filter with a gain of is installed on bus 10.
(5)
A lossless feeder allows us to substitute a relationship of
II. A POWER DISTRIBUTION FEEDER and into (1)–(3), thus resulting in the following simpli-
A. Distributed-Parameter Model fied equations:
Fig. 1 shows a three-phase power distribution feeder. Table I
(6)
shows circuit parameters of the feeder for the following anal-
ysis. The circuit parameters are designed on the basis of a real
distribution feeder rated at 6.6 kV and 3 MW in Japan. However, (7)
in order to clearly demonstrate the occurrence of the “whack-a-
(8)
mole,” the feeder length and the total capacity of capacitors for
power-factor correction are modified to 9 km and 3 MVA, which
are three times as large as those in the real feeder. It is assumed Table II summarizes a relationship among harmonic fre-
that each line inductor between a bus and its adjacent bus has quency, wavelength, and feeder length. The feeder length of 9
the same inductance as , and each capacitor has the same ca- km corresponds to half of the fifth-harmonic wavelength and
pacitance as for the sake of simplicity. Attention is paid to a three fourths of the seventh-harmonic wavelength.
radial trunk feeder, so that it is assumed that no branch exists on
B. Shunt Active Filter Based on Voltage Detection
any bus. Moreover, no-load conditions are taken to discuss the
most serious harmonic propagation. The so-called “skin effect” The purpose of the active filter is to damp out harmonic prop-
in conductors at the most dominant fifth- and seventh-harmonic agation throughout the feeder. The active filter detects voltage
frequencies is not so serious in a real system, and therefore it harmonics at the point of installation, and then injects a com-
is not considered in this paper. Moreover, no iron loss exists pensating current as follows:
in overhead lines and power cables having no magnetic com- (9)
ponent. Note that a small amount of iron loss at the harmonic
frequencies may appear in distribution transformers that are ex- where is a control gain of the active filter. The above
cluded from Fig. 1. equation implies that the active filter behaves like a resistor
Fig. 2 depicts a distributed-parameter model of the power dis- of to the external circuit for harmonic frequencies,
tribution feeder, where is the feeder length, and is the dis- whereas the active filter makes no contribution to the external
tance from bus 1. A harmonic voltage source exists on bus 1, and circuit for the fundamental frequency. For the following anal-
a resister of is connected to the end bus. Note that the ysis and experiment, the active filter is connected to the end
resistor implies a shunt active filter based on voltage detection bus (bus 10) in order to achieve the most effective harmonic
when attention is paid to voltage and current harmonics. damping [7], [8].
WADA et al.: SHUNT ACTIVE FILTER BASED ON VOLTAGE DETECTION 1125

Fig. 3. Analytical and computer-simulated results of a voltage-magnifying Fig. 4. Analytical and computer-simulated results of a voltage-magnifying
factor on each bus when a fifth-harmonic voltage source is connected to bus factor on each bus when a seventh-harmonic voltage source is connected to bus
1. 1.

III. ANALYSIS OF HARMONIC-VOLTAGE STANDING WAVES

A. Harmonic Voltage Source Existing on Bus 1
Fig. 2 gives a harmonic-voltage standing wave and a
harmonic-current standing wave as a function of distance
[9]
(10) Fig. 5. Distributed-parameter model on harmonic current source on the feeder.

(11) bus 1. The vertical axis shows a voltage-magnifying factor, that

is, a ratio of a harmonic voltage on a bus to that on bus 1. The
Hence, constants and can be determined by two terminal
marks “ ,” “ ” and “ ” indicate simulation results using the
conditions at the sending and receiving ends. A harmonic
PSpice software package, and the lines “—,” “– –,” and “- - -”
voltage source is connected to the sending end, and a
are analytical results obtained from (14). In both figures, “
resistor of is connected to the receiving end, instead
” corresponds to no installation of active filters, “ ”
of the actual active filter. Therefore, and are given by
means that the active filter acts as the characteristic impedance,
and “ ” forms a short circuit across the end bus for
(12)
harmonic frequencies. Note that the line resistance in Table I
is considered in achieving both analysis and simulation.
(13) Fig. 3 shows the fifth-harmonic-voltage standing waves. In
the case of , a voltage-magnifying factor on each bus is
Substituting and into (10) induces the harmonic-voltage less than unity. When the active filter with a gain of
standing wave is installed on the end bus, no fifth-harmonic voltage appears on
the end bus. However, the harmonic voltage on bus 5 is magni-
(14) fied by five times, compared with that on bus 1, thus resulting in
the occurrence of the whack-a-mole. The harmonic voltage on
The distribution feeder in Fig. 1 consists of nine inductors each bus is equal to that on bus 1 when the active filter acts as
and nine capacitors. A small difference in harmonic voltage may a resistor of . This implies that installation of the
occur between the lumped-element and distributed-parameter active filter does not cause any whack-a-mole in the feeder.
models because the division number is not large enough. For As for the seventh-harmonic voltage, harmonic propagation
example, the seventh-harmonic voltage on the end bus is 4.0 can be damped out throughout the feeder, as shown in Fig. 4,
times as large as that on bus 1 in the lumped-element model in when is set to .
Fig. 1, while it is 2.4 times in the distributed-parameter model
with the feeder length of 9 km. Thus, the distributed-param- B. Harmonic Current Source Existing Inside the Feeder
eter model needs an adjustment to the feeder length. When the Fig. 5 shows a distributed-parameter model when a harmonic
feeder length is modified to 9.5 km in the distributed-parameter current source exists inside the feeder. Here, no harmonic
model, the harmonic voltage on the end bus agrees with that voltage source exists on bus 1, and is the distance from bus 1
in the lumped-element model. Thus, the distributed-parameter ( ) to the position of the harmonic current source. The cur-
model with a feeder length of 9.5 km is used in the following rent injected by the harmonic current source is divided into
analysis. and , that is, two upstream and downstream currents. In
Figs. 3 and 4 show voltage standing waves caused by the fifth this case, the feeder should be dealt with as two distributed-pa-
and seventh harmonic voltage sources on bus 1, respectively. rameter models for and . Thus, two constants and in
The horizontal axis shows the bus number and the distance from (10) have to be calculated for each distributed-parameter model.
1126 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO. 4, JULY/AUGUST 2002

When the harmonic current source exits at on the feeder, the

harmonic-voltage standing wave at position , can be ex-
pressed as follows:

(15)

Fig. 6. Analytical and computer-simulated results of standing voltage waves

Assuming that no loss occurs in the feeder, the propagation when a fifth-harmonic current source of 3.4% is connected to bus 6.
constant in (15) is imaginary from (7), and the harmonic-
voltage standing wave can be simply explained by sinusoidal
functions.
In the case of the gain of , the amplitude of the
standing wave is given by

(16)

The largest harmonic voltage appears on the end bus, irrespec-

tive of the feeder length and the position of the harmonic cur-
rent source. When the current source is connected to the end
bus ( ), a large harmonic voltage appears throughout the
feeder. In particular, the feeder length of or may
cause quite a large harmonic voltage throughout the feeder, be- Fig. 7. Analytical and computer-simulated results of standing voltage waves
cause the denominator in (16) is equal to zero. A nonnegligible when a seventh-harmonic current source of 3.4% is connected to bus 4.
resistor in practical feeder could make a contribution to reducing
the harmonic voltage. as shown in Fig. 6. When is set to , the active filter can
In the case of the gain of , the standing wave is suppress the harmonic voltage at bus 10 completely. However,
given by the whack-a-mole appears; the harmonic voltage reaches 10%
on bus 6 where the harmonic current source is connected. On
the other hand, when is set to , the harmonic voltage is
kept under 2% on each bus.
(17) When no active filter is connected, the seventh-harmonic
. voltage reaches 10% at buses 4 and 10, as shown in Fig. 7.
Installing the active filter with the gain of or on
bus 10, the harmonic propagation can be well damped.
When the harmonic current source exists at a distance of
from the end bus ( ), the largest harmonic voltage
appears at the position of from bus 1. If the feeder length C. Rated Power Capacity of the Active Filter
equals , an excessive harmonic voltage may occur.
Figs. 6 and 7 show standing waves when a fifth- or seventh- Although it would be difficult to estimate the rated power ca-
harmonic current source of 3.4% is connected to bus 6 or bus 4 pacity of the active filter in a real feeder, it is easy in the simpli-
in the feeder, respectively. The vertical axis means a ratio of the fied feeder shown in Fig. 2. It is assumed that when a fifth-har-
harmonic voltage with respect to the fundamental line-to-line monic voltage of 2% of the rated phase voltage of 6.6
voltage. exists on the bus 1, the fifth-harmonic voltage on the end bus is
In the case of disconnecting the active filter, that is, , damped out to be 2% as a result of having installed the active
no fifth-harmonic voltage exists between bus 1 and bus 6, but filter on the end bus. Note that the active filter is operated with
the harmonic voltage tends to increase downstream of bus 6, an optimal gain of , where is 8.9 from (6).
WADA et al.: SHUNT ACTIVE FILTER BASED ON VOLTAGE DETECTION 1127

Fig. 8. Experimental system.

Fig. 10. Control circuit of the active filter.
TABLE III
CIRCUIT PARAMETERS IN EXPERIMENT
Fig. 10 shows a control circuit of the active filter. Three-
phase voltages, which are detected at the point of installation,
are transformed to and on the coordinates. Two first-
order high-pass filters (HPFs) with the same cutoff frequency as
5 Hz extract ac components and from and . The ac
components are applied to the inverse transformation circuit,
so that the control circuit yields three-phase harmonic voltages
at the point of installation. Amplifying each harmonic voltage
by a gain of produces each current reference as follows:

(18)

In this experiment, is set to instead of

because it is impossible to set because of causing
instability. Note that is half of the characteristic
impedance of the feeder.
Fig. 9. Main circuit of the active filter.
B. Experimental Results of a Harmonic Voltage Source
Existing on Bus 1
These conditions enable us to derive the rated power capacity as
follows: Figs. 11 and 12 show experimental waveforms when a fifth-
or seventh-harmonic voltage source exists on bus 1, in which
kVA each bus voltage is a line-to-line voltage. Note that the voltage
source can be considered as a background harmonic voltage ex-
Since the power capacity per feeder in Japanese 6.6-kV utility isting upstream of a primary distribution transformer in an ac-
distribution power systems is around 3 MW, the active filter ca- tual power system. Here, no harmonic-producing load exists in
pacity corresponds to 3.3% of the feeder capacity. the power distribution line simulator.
Figs. 13 and 14 show harmonic standing waves obtained from
IV. EXPERIMENTAL RESULTS Figs. 11 and 12. In the case of disconnecting the active filter, a
A. Experimental System magnifying factor of the fifth-harmonic voltage is less than unity
on each bus. When the active filter with a gain of
Fig. 8 shows an experimental system using a power distribu-
is installed, the harmonic voltage on bus 6 is twice as high as
tion line simulator, which is rated at 100 V and 5 kVA. Table III
that on bus 1, while the harmonic voltage on bus 10 is smaller
summarizes circuit parameters of the simulator which is a down-
than that on bus 1. This means that installation of the active filter
scaled model of the distribution feeder shown in Fig. 1. Under
causes the whack-a-mole. When the active filter is disconnected,
no-load conditions, the fundamental voltage on bus 10 is 1.4
a magnifying factor of the seventh-harmonic voltage on buses 4
times as high as that on bus 1, because of the Ferranti phe-
and 10 is 2.5 times as large as that on bus 1, as shown in Fig. 14.
nomenon. Although this simulator is originally designed for a
200-V system, a step-down transformer with turn ratio of 2 : 1 Connecting the active filter with a gain of can miti-
is introduced to withstand an overvoltage across the capacitor gate the seventh-harmonic voltage throughout. It is concluded
on bus 10. that neither fifth- nor seventh-harmonic voltage is magnified
Fig. 9 shows the main circuit of the active filter for the exper- throughout when the active filter with a gain of
iment. It consists of a three-phase voltage-fed pulsewidth mod- is connected.
ulation (PWM) inverter using six insulated gate bipolar transis-
tors (IGBTs). The active filter is installed on the end bus in the C. Experimental Results of a Harmonic Current Source
feeder via a three-phase transformer with turn ratio of 2 : 1. An Existing Inside the Feeder
electrolytic capacitor of 3300 F is connected to the dc side, the Figs. 15 and 16 show experimental results of the standing
dc voltage is 200 V, and an inductor of mH (7.5% on waves produced by a fifth- or seventh-harmonic current source
a 3 50-V 60-Hz 500-VA base) is connected in series to an ac of 1.0 A (3.4%). Note that a small background harmonic voltage
line each phase. exists on bus 1 in this experimental system.
1128 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO. 4, JULY/AUGUST 2002

Fig. 11. Experimental waveforms when a fifth-harmonic voltage source is connected to bus 1.

Fig. 12. Experimental waveforms when a seventh-harmonic voltage source is connected to bus 1.

Fig. 13. Experimental voltage standing waves when a fifth-harmonic voltage Fig. 14. Experimental voltage standing waves when a seventh-harmonic
source is connected to bus 1. voltage source is connected to bus 1.

A fifth-harmonic current source is connected to bus 6 in on bus 6 reaches 3.2%. A seventh-harmonic current source is
Fig. 15. In the case of disconnecting the active filter, the largest connected to bus 4 in Fig. 16. The seventh-harmonic voltages
fifth-harmonic voltage appears on bus 10, which is only 2% of on buses 4 and 10 reach about 6% in case of disconnection the
the fundamental voltage. When the active filter with a gain of active filter. When the active filter with a gain of or
is installed on the end bus, the harmonic voltage is installed on the end bus, the seventh-harmonic voltage
WADA et al.: SHUNT ACTIVE FILTER BASED ON VOLTAGE DETECTION 1129

• Installation of the active filter on the end bus does not

cause any whack-a-mole at a specified harmonic fre-
quency as long as the feeder length is shorter than a
quarter of the wavelength at the frequency.
• No whack-a-mole occurs irrespective of the feeder length
and the total capacity of capacitors for power-factor cor-
rection as long as the active filter acts as the characteristic
impedance of the feeder.
These conclusions make a significant contribution to practical
installation of the active filter in terms of having given broad
hints on how to avoid the whack-a-mole in a real feeder.
Fig. 15. Experimental voltage standing waves of fifth-harmonic voltage when
a harmonic current source is connected to bus 6. REFERENCES
[1] E. J. Currence, J. E. Plizga, and H. N. Nelson, “Harmonic resonance at a
medium-sized industrial plant,” IEEE Trans. Ind. Applicat., vol. 31, pp.
682–690, July/Aug. 1995.
[2] D. Andrews, M. T. Bishop, and J. F. Witte, “Harmonic measurements,
analysis, and power factor correction in a modern steel manufacturing
facility,” IEEE Trans. Ind. Applicat., vol. 32, pp. 617–624, May/June
1996.
[3] K. Oku, O. Nakamura, and K. Uemura, “Measurement and analysis of
harmonics in power distribution systems, and development of a har-
monic suppression method” (in Japanese), Trans. Inst. Elect. Eng. Jpn.,
vol. 114-B, no. 3, pp. 234–241, 1994.
[4] K. Oku, O. Nakamura, J. Inoue, and M. Kohata, “Suppression effects
of active filter on harmonics in a power distribution system including
capacitors” (in Japanese), Trans. Inst. Elect. Eng. Jpn., vol. 115-B, no.
9, pp. 1023–1028, 1995.
[5] H. Sumiya and M. Kaneko, “Development of a measuring instrument
for harmonic frequency” (in Japanese), Denki Genba Gijyutu, vol. 34,
Fig. 16. Experimental voltage standing waves of seventh-harmonic voltage no. 402, pp. 23–28, 1995.
when a harmonic current source is connected to bus 4. [6] H. Akagi, “New trends in active filters for power conditioning,” IEEE
Trans. Ind., Applicat., vol. 32, pp. 1312–1322, Nov./Dec. 1996.
[7] , “Control strategy and site selection of a shunt active filter for
is mitigated throughout. Thus, the active filter with a gain of damping of harmonic propagation in power distribution systems,” IEEE
can damp out harmonic propagation caused by Trans. Power Delivery, vol. 12, pp. 354–363, Jan. 1997.
both fifth- and seventh-harmonic current sources. [8] H. Akagi, H. Fujita, and K. Wada, “A shunt active filter based on voltage
detection for harmonic termination of a radial power distribution line,”
IEEE Trans. Ind. Applicat., vol. 35, pp. 638–645, May/June 1999.
V. DISCUSSION OF THE WHACK-A-MOLE [9] F. E. Terman and W. W. Harman, Analysis, Transmission, and Filtering
of Signals. New York: McGraw-Hill, 1963, pp. 354–376.
When attention is paid to a harmonic frequency specified by
a relation between the feeder length and the wavelength , de-
pendency of the gain on the occurrence of the whack-a-mole Keiji Wada (S’98–A’01) was born in Hokkaido,
in a feeder is derived as follows. Japan, in 1973. He received the B.S. and M.S.
degrees from the Polytechnic University, Sagami-
• —When no active filter is connected, a harmonic hara, Japan, and the Ph.D. degree from Okayama
voltage on a bus goes up gradually as the bus number in- University, Okayama, Japan, in 1995, 1997, and
crease. This is similar to the Ferranti effect. Installation of 2000, respectively, all in electrical engineering.
a shunt active filter based on voltage detection on the end Since 2000, he has been a Research Associate in
the Department of Electrical Engineering, Tokyo
bus of the feeder is most effective in achieving harmonic Metropolitan University, Tokyo, Japan. His research
damping. Moreover, this conclusion is valid, irrespective interests are active filters for power conditioning and
of the position of harmonic voltage and/or current sources. photovoltaic power generation.
• —Installation of the active filter on the end bus
may cause the whack-a-mole at a specified harmonic fre-
quency. However, if the active filter can be controlled to Hideaki Fujita (M’91) was born in Toyama Prefec-
ture, Japan, in 1965. He received the B.S. and M.S.
present the characteristic impedance, the whack-a-mole degrees from Nagaoka University of Technology, Na-
disappears throughout the feeder. This corresponds to the gaoka, Japan, and the Ph.D. degree from Tokyo Insti-
so-called “termination” in the transmission line theory. tute of Technology, Tokyo, Japan, in 1988, 1990, and
2000, respectively, all in electrical engineering.
From 1991 to 2002, he was a Research Associate in
VI. CONCLUSION the Department of Electrical Engineering, Okayama
University. Since 2002, he has been an Associate Pro-
This paper has presented considerations of a shunt active filter fessor in the Department of Electrical and Electronic
based on voltage detection for installation on a long distribution Engineering, Tokyo Institute of Technology. His re-
feeder. An effort has been made to analyze a curious phenom- search interests are static var compensators, active filters, and resonant con-
enon referred to as the “whack-a-mole.” verters.
Dr. Fujita received First Prize Paper Awards from the Industrial Power Con-
Analytical and experimental results obtained from a simpli- verter Committee of the IEEE Industry Applications Society in 1990, 1995, and
fied feeder have led to the following conclusions. 1998.
1130 IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 38, NO. 4, JULY/AUGUST 2002

Hirofumi Akagi (M’87–SM’94–F’96) was born in

1951. He received the Ph.D. degree in electrical engi-
neering from Tokyo Institute of Technology, Tokyo,
Japan, in 1979.
In 1979, he joined the Department of Electrical
Engineering, Nagaoka University of Technology, as
an Assistant Professor. He later became an Associate
Professor. In 1987, he was a Visiting Scientist
at Massachusetts Institute of Technology for ten
months. From 1991 to 1999, he was a Professor in
the Department of Electrical Engineering, Okayama
University. From March to August 1996, he was a Visiting Professor at
the University of Wisconsin, Madison, and then Massachusetts Institute
of Technology. Since 2000, he has been a Professor in the Department of
Electrical and Electronic Engineering, Tokyo Institute of Technology. His
research interests include ac motor drives, high-frequency resonant inverters
for induction heating and corona discharge treatment processes, active/passive
EMI filters, and utility applications of power electronics such as active filters,
self-commutated BTB systems, and FACTS devices. He has authored approxi-
mately 130 peer-reviewed journal papers, including 50 IEEE TRANSACTIONS
papers and an invited paper in the PROCEEDINGS OF THE IEEE. He has made
numerous presentations as a keynote or invited speaker internationally.
Prof. Akagi has received nine IEEE Prize Paper Awards, including the IEEE
TRANSACTIONS ON INDUSTRY APPLICATIONS Prize Paper Award for 1991 and
the IEEE TRANSACTIONS ON POWER ELECTRONICS Prize Paper Award for 1998.
He was elected a Distinguished Lecturer of the IEEE Industry Applications and
IEEE Power Electronics Societies for 1998–1999. He was the recipient of the
IEEE William E. Newell Power Electronics Award in 2001.