IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 21, NO.

4, JULY 2006

1013

Dynamic Control and Performance of a Unied Power Flow Controller for Stabilizing an AC Transmission System
Hideaki Fujita, Member, IEEE, Hirofumi Akagi, Fellow, IEEE, and Yasuhiro Watanabe
AbstractThis paper presents dynamic control and performance of a unied power ow controller (UPFC) intended for installation on a transmission system consisting of two sets of three-phase transmission lines in parallel. When no UPFC is installed, interruption of either three-phase line due to a fault reduces an active power ow to half, because the line impedance becomes double before the interruption. Installing the UPFC makes it possible to control an amount of active power owing through the transmission system. The validity of the theoretical analysis developed in this paper is veried by experiments using a 10-kVA laboratory setup, as well as a computer simulation. Index TermsLine interruption, power swings, transmission systems, unied power ow controller (UPFC).
Fig. 1. Circuit conguration of a UPFC used for experiment.

I. INTRODUCTION HE unied power ow controller (UPFC) [1][3], which is one of the most promising devices in the FACTS concept, has been researched and put into practical use. The American Electric Power (AEP) company has built and installed a 160-MVA UPFC at the Inez substation in eastern Kentucky for the rst time in the world [4], [5]. The UPFC consists of combined series and shunt devices, and the dc terminals which are connected to a common dc-link capacitor. The series device controls active power ow from the sending to the receiving end by means of adjusting the phase angle of the output voltage. On the other hand, the shunt device performs regulation of the dc-link voltage as well as control of reactive power. The UPFC realizes power ow control, stability improvement, and so on. Damping performance against the so-called power swings is essential to improving the sending capacity of a power transmission system. The power swings are low-frequency oscillations of active and reactive powers, which are caused by resonance between a line inductance and a moment of inertia in synchronous generators. The oscillating frequency of the power swings is usually in a range from 0.3 to 2 Hz. Thus, a conventional UPFC for the purpose of damping out the power swings is designed to have a response time as slow as about 100 ms in power ow control. Line-to-line and line-to-ground faults may cause a considerable variation in power ow as fast as 10 ms.

However, the conventional UPFC could not mitigate such a variation. A fast response is required for a UPFC to eliminate the variation from power ow. Dynamic control methods based on feedback control of instantaneous active and reactive power has been proposed to realize fast power ow control [6], [12]. The authors have already proposed the advanced control method, which is characterized by providing a response as fast as 3 ms in power ow control without any oscillation or overshoot [13], [14]. This paper presents dynamic control and behavior of a UPFC under a fault condition in a transmission system consisting of two parallel three-phase lines.

II. EXPERIMENTAL SYSTEM CONFIGURATION Fig. 1 shows the laboratory setup of the UPFC used in the following experiments and simulations. The circuit parameters of the UPFC are shown in Table I. The main circuit of the series device consists of three single-phase H-bridge voltage-source inverters rated at 1 kVA. The inverters perform ac voltage control by means of pulsewidth modulation (PWM) with a switching frequency of 1 kHz. The ac terminals of each H-bridge inverter are connected in series to the transmission line through a singlephase transformer with a turns ratio of 1:12. It is possible to replace the three single-phase inverters with a three-phase inverter. The circuit conguration using three single-phase inverters seems to be suitable for implementation of a practical UPFC rated at 100 MVA or higher. Each leg in such a large-rating converter generally consists of series and/or parallel connection of high-voltage large-current switching devices. The volt-ampere rating of the three single-phase inverters

Manuscript received June 17, 2004; revised March 24, 2005. Recommended by Associate Editor V. Staudt. H. Fujita and H. Akagi are with the Tokyo Institute of Technology, Tokyo 152-8552, Japan (e-mail: fujita@ee.titech.ac.jp; akagi@ee.titech.ac.jp). Y. Watanabe is with the Japan Atomic Energy Agency, Ibaraki 319-1195, Japan (e-mail: yasuhiro@linac.tokai.jaeri.go.jp). Digital Object Identier 10.1109/TPEL.2006.876845

0885-8993/$20.00 2006 IEEE

1014

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 21, NO. 4, JULY 2006

TABLE I SYSTEM PARAMETERS OF THE EXPERIMENTAL SYSTEM

Fig. 3. Control circuit of the series device.

controls power ow from the sending to the receiving end. Inand represent line inductance in each transmisductors sion line. A three-phase solid-state relay (SSR) installed in series with line 1 is used as a line breaker to simulate the interruption of line 1. The SSR remains closed in a normal condition. While it is opened for a few cycles, the total line impedance increases, so that the power ow decreases. In an actual power system, a periodic change in the rotor speed and angle of a synchronous generator induces power swings. If the power swings exceeded a stable region, the generator would fall into out-of-phase conditions. III. CONTROL METHODS Fig. 3 shows a block diagram of the control circuit for the series device. The threetwo-phase transformation obtains and from the three-phase currents and . Then the transformation is applied with the help of sinusoidal signals of and . These blocks are implemented by using analog multipliers and operational ampliers. The phase is generated by a phase-lock-loop (PLL) circuit. information and signals are obtained by connecting digThe ital-to-analog converters to a read only memory (ROM). Thus, the control circuit shown in Fig. 3 has almost no delay. The PWM operation of the inverter produces the most dominant delay. This delay in the experimental setup is about 0.5 ms, because the switching frequency is 1 kHz. The following sections discuss transient responses of the cross-coupling control method in [7] and the advanced control method in [13]. A main difference between the two methods exists in whether they can damp out power swings. The cross-coupling control method is based on voltage and current phasors, while the advanced control method starts from differential equations in the transient states. The advanced control method makes it possible to improve its transient stability even when its feedback gains are higher than that in cross-coupling control method. As a result, the advanced control method brings a faster response to active and reactive power than the cross-coupling control method [13]. A. Cross-Coupling Control Method The cross-coupling control method in [7] has two control and , for active- and reactive-power feedbacks. gains, This makes it possible to control both active and reactive powers independently. Integral gains are widely applied for a practical

Fig. 2. Experimental system conguration.

is twice as large as that of the three-phase inverter, when each leg of the inverters consists of the same number of the same switching devices. Moreover, the three single-phase inverters produce less switching ripples than the three-phase inverter. The shunt device consists of a three-phase PWM inverter, the ac terminals of which are connected in parallel with the transmission line via a three-phase transformer with a turns ratio of 2:1. The shunt device regulates the dc-link voltage 200 V. A practical shunt device may consist of a as multilevel inverter. However, the simple three-phase inverter is introduced to the experimental setup, because this paper pays attention to the control and performance of the series device. For the same reason, no reactive-power control is achieved in the shunt device. Fig. 2 shows the 10-kVA laboratory model including a transmission system consisting of two sets of three-phase lines; lines and is assumed to be sending- and re1 and 2. Here, ceiving-end voltages, respectively. The model assumes that the sending end corresponds to a power plant while the receiving end to an electric power network. The receiving-end voltage may not cause any phase-angle change, because can be considered as an innite-bus voltage. The phase angle of is adjusted according to a power demand for the power plant. A phase-shifting transformer is employed to produce a difference in phase angle between the sending-end and receiving-end voltages [14]. The UPFC installed near the sending end effectively

FUJITA et al.: DYNAMIC CONTROL AND PERFORMANCE OF A UNIFIED POWER FLOW CONTROLLER

1015

feedback control in order to reduce steady-state error. The crosscoupling control method with integral gains gives the following and to the series device: voltage references (1) and are proportional gains for active and reactive where powers, respectively, and is a integral gain [8][11]. B. Advanced Control Method The advanced control method in [13] has an additional control gain for the purpose of improving the stability of the active- and reactive-power control. The voltage references and are given by the following equation: (2) where swings. is a control gain capable of damping out power
Fig. 4. Bode diagrams of the transmission line, 0.02
, and = 0.5 mH.

G (s) and G (s) when R =

IV. TRANSFER FUNCTIONS OF THE TRANSMISSION SYSTEM Fig. 2 gives the following three-phase voltage and current equation: . The resonant angular frequency appears very close to , in a transmission line. because Fig. 4 shows the bode diagrams of and . In a is higher than frequency range lower than 60 Hz, in gain, and the phase angle is almost zero. Thus, adjusting can control effectively, while can control . This means that the cross-coupling control may operate in the lowshows frequency range successfully. On the other hand, a high gain in a frequency range higher than 60 Hz, compared . Therefore, in (2) is more effective than and with for controlling and in the high-frequency range. Therefore, the advanced control method may show dynamic performance better than the cross-coupling control and phase control methods.

(3) where and are line inductance and resistance in parallel connection of lines 1 and 2. Applying three-to-two-phase conversion and transformation to (3) can be represented by (4) is a current component corresponding to an active Here, corresponds to a reactive power. This means power, while that is in phase with the -phase supply voltage, while is 0, because the receiving perpendicular to it. (4) assumes end is assumed to be connected to an innite bus. Equation (4) gives the following transfer functions from the output voltage of the series device to the line current: (5) where (6) (7) (8) The transfer functions and exhibit a second-order response with a resonant angular frequency of . The existence of results in cross couplings of with and with

V. TRANSIENT RESPONSE WHEN NO UPFC IS INSTALLED This section discusses transient response of the active power, to considering the line inductance change from due to a line fault. For the sake of simplicity, it is assumed that the -axis voltage on the sending end is equal to that on the . Substituting 0 and receiving end, that is, 0 to (5) gives a step response of the -axis current as follows:

(9) From (9), the initial current line fault is given by just before occurrence of the

(10)

1016

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 21, NO. 4, JULY 2006

and the nal value

is

Invoking the approximation in (18)) results in the following steady-state error in active power: (11)

and Taking represented as

into account, a transient response of

is

(12) The time response of can be obtained as follows:

(19) is set as (17), a large amount of steady-state When the gain error may appear in active power. 2) In Case of Having the Integral Gain: In particular, the integral gains are more dominant than the proportional gains in a . Applying the approximalow frequency range of tion in (18), the transfer function of the whole system is given by

(13) The dc component of continues to decrease during the line fault, while an ac component with the line frequency appears in . In order to eliminate the ac component from , the UPFC requires capability of power ow control as fast as 23 ms. VI. TRANSIENT RESPONSE WHEN THE UPFC IS INSTALLED A. Cross-Coupling Control Method Let us consider the transient response of the transmission system with an UPFC, to which the cross-coupling control is applied. 1) In Case of No Integral Gain: For the sake of simplicity, the integral gain in (1) is neglected as follows:

(20) The time constant of the transfer function is

(21) The approximation in (18) is applicable for a steady state or a slow transient state. Therefore, the time constant has to be larger , in order to avoid the coupling between active and than 1 reactive power: 1

(22) is summarized as

(14) As shown in Fig. 4, the cross-coupling control are applicable to a frequency range lower than . In this range, almost no coupling and because . appears between , Paying attention to the resonant angular frequency the following approximations are derived: (15) (16) Although both transfer functions have the same amplitude, the lags by 90 . If 2 and 2 , the phase angle of open-loop gain is 0 dB and the phase margin is 90 . However, and tends induce overshoots and oscillations, increasing is almost 180 in the frequency because the phase angle of range over . In order to avoid the overshoot and oscillaand should be set as tions, the control gains (17) The gain of at 0 is given by

The requirement for the integral gain

(23) From (19) to (23), the time response of the active power control with the proportional and integral gain can be represented as follows: (24) Note that (24) is valid only for a slow transient response without any overshoot nor oscillation in active power. The error equal to (19) appears in at the instant of occurrence of the line fault. However, the active power gradually recovers with the time constant of . The nal value of (24) is equal to its ini. Thus, no steady-state error exists. tial value of B. Advanced Control Method 1) In Case of No Integral Gain: The voltage reference for the advanced control method in (2) is represented by the following equation, when the integral gain is neglected:

(18)

(25)

FUJITA et al.: DYNAMIC CONTROL AND PERFORMANCE OF A UNIFIED POWER FLOW CONTROLLER

1017

Assuming in (5) makes almost no change in even during the line fault. Substituting (25) into (5) obtains the following simplied equation related with :

TABLE II GAIN SETTINGS FOR THE EXPERIMENTS AND SIMULATIONS

(26) is controlled by the difference between The -axis current itself and its reference . Therefore, (26) is assumed as a feedback control of . The open-loop transfer function of the assumed feedback control, which is a transfer function from to , is given by

TABLE III VALUES OF  , !; I =I , AND 

(27) Neglecting the line resistance quency response of derives the approximated fre-

(28) In a frequency range higher than represented by , the phase angle of is

(29) The gain contributes to increasing the phase angle. The is set to be two. In phase margin is more than 60 when the case of the cross-coupling control method, the phase angle is equal to zero. Thus, in (29) is 180 , because the gain almost no phase margin exists in the cross-coupling control. The steady-state error of the advanced control method is and can be set to a also obtained by (19). The gains large value, compared with the cross-coupling control method, because the cross-coupling control method has the limitation makes it possible to reduce shown in (17). Setting the steady-state error in (19). 2) In Case of Having the Integral Gain: Equation (24) is derived for the cross-coupling control method under a assumption of a slow transient response. This equation is also applicable to the advanced control method when the time constant in (20) is adequate to the range shown in (22). The advanced control method has capability of reduction of the error in at the instant of occurrence of the line fault, because the advanced control than the cross-coupling method allows setting a higher gain control method does. As a result, the advanced control method makes it possible to maintain almost constant even under the line fault condition. VII. EXPERIMENTAL RESULTS The control parameters used in experiment and simulation are summarized in Table II. The proportional gains in the cross-

0.05 V/A. Since coupling control method are set to these gains are slightly greater than the values in (17), a small amount of overshoot appears in power ow. The integral gains 2.5 V/A/s. Then, the time constant shown in are set as 75 ms during the SSR is opened, while it is (21) is 38 ms during the SSR is closed. The proportional gains and in the advanced control 0.5 V/A, which are ten times as large as are set to is set to those in the cross-coupling control. The gain 0.9 V/A, so that the damping factor is 0.8. The integral gain is also set to 0.5 V/A/s, and thus, the time constant is one tenth of that in cross-coupling control method. Table III shows damping factor , resonant angular frequency , error in active , and time constant . The time constant of the power advanced control method is 4 ms during the SSR is turned on, as shown in Table III. The time constant dominates the transient performance of active and reactive powers because it is eight times as long as a delay time of 0.5 ms that is caused by PWM with a switching frequency of 1 kHz. Therefore, the delay time has almost no effect, although the delay time is not considered in the analysis in the previous section. However, the delay time may cause a stability problem in the active- and reactive-power control if the time constant were close to or less than the delay time. Figs. 5 10 show simulated and experimental waveforms. The SSR installed in series with line 1 had been turned on before the marked point (SSR off), it remained turned off for ten cycles (200 ms), and then, it was turned on again at the other marked point (SSR on). An actual amount of active power ow was set to 10 kW by adjusting the phase-shift transformer, when the SSR was turned on and the UPFC was not installed. In this situation, a reactive power of 3 kvar was owing through the transmission lines. The computer simulation was carried out by using the EMTDC software package. An ideal switch model was used as the SSR, and it was programmed to be turned off at a zero-current point [15].

1018

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 21, NO. 4, JULY 2006

Fig. 5. Simulated waveforms when no UPFC controls active and reactive powers.

Fig. 7. Simulated waveforms when the cross-coupling control method is applied.

Fig. 6. Experimental waveforms when no UPFC controls active and reactive powers.

Fig. 8. Experimental waveforms when the cross-coupling control method is applied.

A. No UPFC Figs. 5 and 6 show simulated and experimental waveforms when no UPFC controls active and reactive powers. In this experiments, the lower-arm switching devices of the H-bridge inverter are turned on, while the upper-arm devices remain off. Thus, the ac terminal voltage of the inverter should be zero. However, a small amount of voltage appears in , as shown in

Fig. 6. It is caused by the leakage inductance of the single-phase transformer. The active power decreased from 10 to 5 kW within 5 ms after the SSR was turned off. After the SSR was turned on, the active power regained to 10 kW with an overshoot of 3 kW. Fig. 6 has a smaller overshoot in than Fig. 5, because an on-state resistance exists in the SSR used in this experiment.

FUJITA et al.: DYNAMIC CONTROL AND PERFORMANCE OF A UNIFIED POWER FLOW CONTROLLER

1019

and were a small gain to the case of no UPFC, because of 0.05 V/A. After that, the active power gradually increased to 10 kW by the integral gains. A large amount of overshoot and oscillations appeared in the active power and reactive power when turning the SSR on, because the integral gains could not respond to the rapid change of the line impedance. While active power in Fig. 8 had almost no oscillation, a large amount of oscillations appear in Fig. 7. In the simulation, any resistor is considered in SSR, so that large oscillations appear in Fig. 7. C. UPFC With the Advanced Control Method Figs. 9 and 10 show simulated and experimental waveforms when applying the advanced control. The active power was kept and were as 10 kW during the fault, because the gains set to 0.5 V/A, which is ten times as large as that in Figs. 7 and has the capability of damping the active 8. Since the gain and reactive power oscillations, no overshoot nor oscillations appeared in the active and reactive power. These results reveal that the advanced control makes it possible to signicantly improve stability of transmission systems. VIII. CONCLUSION
Fig. 9. Simulated waveforms when the advanced control method is applied.

This paper has presented dynamic control and performance of a UPFC under a fault condition in a two-parallel three-phase transmission system. The transient analysis claries that a line interruption generally causes a transition in the active power as fast as 23 ms. Thus, a conventional UPFC with a response as slow as 100 ms has difculty in suppressing the power variations caused by the faults. Moreover, a conventional UPFC may cause an overcurrent after nishing the fault, due to the slow response of the integral gains in the control loop for the active and reactive power. The advanced control in this paper shows good transient performance without any overshoot or oscillation. The advanced control may contribute not only to achieving fast power ow control but also to improvement of stabilizing the transmission systems. REFERENCES
[1] L. Gyugyi, Unied power-ow control concept for exible ac transmission systems, Proc. Inst. Elect. Eng. C, vol. 139, pp. 323331, Jul. 1992. [2] L. Gyugyi, C. D. Schauder, S. L. Williams, T. R. Rietman, D. R. Torgerson, and A. Edris, The unied power ow controller: a new approach to power transmission control, IEEE Trans. Power Del., vol. 10, no. 2, pp. 10851097, Oct. 1995. [3] N. G. Hingorani and L. Gyugyi, UnderStanding FACTS: Concept and Technology of FlexibleAC Transmission Systems. Piscataway, NJ: IEEE Press, 2000. [4] C. Schauder, E. Stacey, M. Lund, L. Gyugyi, L. Kovalsky, A. Keri, A. Mehraban, and A. Edris, AEP UPFC project: Installation, commissioning and operation of the /spl plusmn/160 MVA STATCOM (phase I), IEEE Trans. Power Del., vol. 13, no. 4, pp. 15301535, Nov. 1998. [5] B. A. Renz, A. Keri, A. S. Mehraban, C. Schauder, E. Stacey, L. Kovalsky, L. Gyugyi, and A. Edris, AEP unied power ow controller performance, IEEE Trans. Power Del., vol. 14, no. 4, pp. 13741381, Nov. 1999. [6] B. S. Rigby and R. G. Harley, An improved control scheme for a series capacitive reactance compensator based on a voltage source inverter, in Proc. IEEE/IAS Annu. Meeting, 1996, pp. 870877. [7] Q. Yu, S. D. Round, L. E. Norum, and T. M. Undeland, Dynamic control of a unied power ow controller, in Proc. IEEE/PELS PES96 Conf., 1996, pp. 508514. [8] Y. Jiang and A. Ekstrom, Optimal controller for the combination system of a UPFC and conventional series capacitors, in Proc. EPE97, 1997, vol. 1, pp. 372337.

Fig. 10. Experimental waveforms when the advanced control method is applied.

A practical system may show a good agreement with the simulation results, because a practical line breaker has a very small resistance. B. UPFC With the Cross-Coupling Control Method Experimental and simulation results using the cross-coupling control are shown in Figs. 7 and 8, respectively. The active power decreases to 5 kW after the SSR was turned off similarly

1020

IEEE TRANSACTIONS ON POWER ELECTRONICS, VOL. 21, NO. 4, JULY 2006

[9] Y. Chen, B. Mwinyiwiwa, Z. Wolanski, and B. T. Ooi, Unied power ow controller (UPFC) based on chopper stabilized multilevel converter, in Proc. IEEE/PELS PESC97 Conf., 1997, pp. 331337. [10] L. Gyugyi, C. D. Schauder, and K. K. Sen, Static synchronous series compensator: A solid-state approach to the series compensation of transmission lines, IEEE Trans. Power Del., vol. 12, no. 1, pp. 406413, Feb. 1997. [11] L. Gyugyi, C. D. Schauder, and K. K. Sen, Improving power system dynamics by series-connected FACTS device, IEEE Trans. Power Del., vol. 12, no. 4, pp. 16351641, Nov. 1997. [12] B. T. Ooi, M. Kazerani, R. Marceau, and Z. Wolanski, Mid-point siting of FACTS device in transmission lines, IEEE Trans. Power Del., vol. 12, no. 4, pp. 17171722, Nov. 1997. [13] H. Fujita, Y. Watanabe, and H. Akagi, Control and analysis of a unied power ow controller, IEEE Trans. Power Electron., vol. 14, no. 6, pp. 10211027, Nov. 1999. [14] H. Fujita, Y. Watanabe, and H. Akagi, Transient analysis of a unied power ow controller and its application to design of the DC-link capacitor, IEEE Trans. Power Electron., vol. 16, no. 5, pp. 735740, Sep. 2001. [15] H. Fujita, S. Tominaga, and H. Akagi, Analysis and design of a dc voltage-controlled static var compensator using quad-series voltagesource inverters, IEEE Trans. Ind. Appl., vol. 32, no. 4, pp. 970978, Jul./Aug. 1996.

Hirofumi Akagi (M87SM94F96) was born in Okayama, Japan, in 1951. He received the B.S. degree from the Nagoya Institute of Technology, Nagoya, Japan, in 1974, and the M.S. and Ph.D. degrees from the Tokyo Institute of Technology, Tokyo, Japan, in 1976 and 1979, respectively, all in electrical engineering. In 1979, he joined the Nagaoka University of Technology, Nagaoka, Japan, as an Assistant and then Associate Professor in the Department of Electrical Engineering. In 1987, he was a Visiting Scientist at the Massachusetts Institute of Technology (MIT), Cambridge, for ten months. From 1991 to 1999, he was a Professor in the Department of Electrical Engineering, Okayama University, Okayama. From March to August of 1996, he was a Visiting Professor at the University of Wisconsin, Madison, and then MIT. Since January 2000, he has been a Professor in the Department of Electrical and Electronic Engineering, Tokyo Institute of Technology. He has published about 170 peer-reviewed journal papers, including about 70 IEEE TRANSACTIONS papers and an invited Proceedings of the IEEE paper. His research interests include power conversion systems, ac motor drives, active and passive EMI lters, high-frequency resonant-inverters for induction heating and corona discharge treatment processes, and utility applications of power electronics such as active lters, self-commutated BTB systems, and FACTS devices. Dr. Akagi received two IEEE IAS TRANSACTIONS Prize Paper Awards in 1991 and 2004, two IEEE PELS TRANSACTIONS Prize Paper Awards in 1999 and in 2003, nine IEEE IAS Committee Prize Paper Awards, the IEEE William E. Newell Power Electronics Award in 2001, and the IEEE IAS Outstanding Achievement Award in 2004. He has made presentations many times as a keynote or invited speaker internationally. He was elected as a Distinguished Lecturer of the IEEE IAS and PELS for 19981999.

Hideaki Fujita (M91) received the B.S. and M.S. degrees in electrical engineering from the Nagaoka University of Technology, Nagaoka, Japan, in 1988 and 1990, respectively. In 1991, he became a Research Associate with the Okayama University, Okayama, Japan. Since 2002, he has been an Associate Professor in the Department of Electrical Engineering, Tokyo Institute of Technology, Tokyo, Japan. His research interests are static var compensators, active power lters, and resonant converters. Dr. Fujita received Prize Paper Awards from the Industrial Power Converter Committee, IEEE Industry Applications Society, in 1990, 1995, 1998, and 2003, respectively.

Yasuhiro Watanabe was born in Gifu, Japan, on May 16, 1972. He received the B.S. degree in electrical engineering from Shizuoka University, Hamamatsu City, Japan, in 1995, and the M.S. and Ph.D degrees from Okayama University, Okayama, Japan, in 1997 and 2000, respectively. In 2000, he joined the National Laboratory for High Energy Accelerator Research. Since 2001, he has been a Research Scientist in the Japan Atomic Energy Agency (formerly Japan Atomic Energy Research Institute), Ibaraki, Japan. His research interests are magnet and magnet power supply for synchrotron accelerator.