A Low Cost, Fast-Rising, High-Voltage Pulsed Power ... - CAU

IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 48, NO. 12, DECEMBER 2020 4387

A Low Cost, Fast-Rising, High-Voltage PulsedPower Modulator Based on a Discontinuous

Conduction Mode Flyback ConverterChang-Yu Liu , Chan-Gi Cho , Graduate Student Member, IEEE, Seung-Ho Song , Member, IEEE,

and Hong-Je Ryoo , Senior Member, IEEE

Abstract— Treatment of gas and water for environmentalapplications requires a high-voltage, fast-rising pulse to act onthe plasma gas or water treating reactor, but the existing stackedhigh-voltage pulse generator is arduous to use owing to its largesize, high cost, and relatively slow pulse rise time. In this study, wepropose a low-cost, small-volume pulsed power modulator thatcan significantly reduce the cost and volume of the device andshorten the pulse rise time to meet the requirements of gas andwater treatment. The proposed pulsed power modulator is basedon a discontinuous conduction mode (DCM) flyback converter,and it generates a high-voltage pulse through the resonance of theflyback transformer secondary side inductance and the parallelcapacitor. A straightforward spark gap sharpens the generatedhigh-voltage pulse. Finally, a fast-rising, high-voltage pulse with anarrow pulsewidth run on the load is generated. Through PSIMsimulation and actual experiments, we validated the feasibility ofthe proposed pulse generator and obtained 23 kV, 500-Hz pulseswith a width of 0.5-µs width and rise time of 5 ns in actualexperiments.

Index Terms— Converters, high-voltage techniques, pulsepower systems.

I. INTRODUCTION

PULSED power technology is widely used in modernnational defense, industry, and medical treatment [1],

[2]. A high-voltage pulse can treat waste gas and sewage bygenerating radicals to convert harmful substances into harm-less ones, which is more efficient than traditional chemicaltreatment and is also environmentally-friendly [3]–[8].

For exhaust gas or water treatment, a pulsed power generatorneeds to achieve fast pulse-rising time, minimum pulsewidth,

Manuscript received December 30, 2019; revised February 28, 2020,July 28, 2020, and September 14, 2020; accepted October 20, 2020. Date ofpublication November 10, 2020; date of current version December 11, 2020.This work was supported in part by the Korea Institute of Energy TechnologyEvaluation and Planning (KETEP), Ministry of Trade, Industry and Energy,Korea through the Project Development of DC Arc Interruption Technologyand Performance Evaluation Facility for Medium and Large PV SystemDevelopment under Grant 20192910100090 and in part by the Ministry ofTrade, Industry & Energy, Republic of Korea through the Human ResourcesProgram in Energy Technology of the Korea Institute of Energy TechnologyEvaluation and Planning (KETEP), under Grant 20184030202270. The reviewof this article was arranged by Senior Editor R. P. Joshi. (Correspondingauthor: Hong-Je Ryoo.)

The authors are with the School of Energy Systems Engineering,Chung-Ang University, Seoul 06974, South Korea (e-mail: [email protected]).

Color versions of one or more of the figures in this article are availableonline at https://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/TPS.2020.3034724

and high repetition rate. To meet these conditions, most ofthe previous studies have used stacking structures [9]–[12].However, stacking involves the use of multiple switches andenergy-storage elements, which results in high expenditure andrequires large space. Additionally, because of the high cost ofthe existing pulsed power supply devices, commercializationis difficult, and extensive effort is needed to reduce the costsinvolved.

In this study, the proposed low-cost simplified structurecombines a flyback converter circuit and a simple spark-gap.The converter comprised an insulated gate bipolar transis-tor (IGBT), a transformer, and a straightforward spark gap,and several high-voltage diodes that are optional to use.A discontinuous conduction mode (DCM) flyback converterconverts grid power electricity into high-voltage pulse output,as discussed in [13]. Furthermore, according to other studies,plasmas produced by extremely short rise time (nanosecondsor less, the shorter the better) and high-voltage pulses areeffective in gas treatment [14], [15]. To achieve a rapid risetime, a simplified spark gap is used to sharpen the pulse, whichyields a fast-rising time of approximately 5 ns. In Section II,the working principle of the DCM flyback converter and ablock diagram of the overall system is introduced. Section IIIintroduces the component selection and overall structure ofthe device. Section IV proves the feasibility of the proposeddevice through simulation and experimental results.

The pulsed power modulator proposed in this study imple-ments a simple structure device with a fast pulse-rising ratethrough various simulations, and actual load tests verify thepossibility of practical application.

II. ANALYSIS OF THE PROPOSED PULSED

POWER MODULATOR

As mentioned, the function of the proposed pulsed powermodulator is to generate a fast-rise, high-voltage pulse usingthe 110-V grid as input. To introduce the principles of theproposed pulse power modulator, operation mode diagrams,and voltage and current waveforms are shown in Figs. 1 and 2.Different from conventional DCM flyback converter, whichoutputs dc voltage, the proposed pulsed power modulator pro-duces a high-voltage pulse because of the resonance betweenthe parallel capacitor (Cp) and the transformer secondaryinductance (Lsec) in an extremely short duty cycle. The

0093-3813 © 2020 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See https://www.ieee.org/publications/rights/index.html for more information.

Authorized licensed use limited to: Chung-ang Univ. Downloaded on December 14,2020 at 00:57:24 UTC from IEEE Xplore. Restrictions apply.

4388 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 48, NO. 12, DECEMBER 2020

Fig. 1. Operation mode diagrams of the proposed pulsed power modulator.

high-voltage pulse can also be approximated as a half-sinewave. When the diode is used in the transformer secondaryside, the operation mode is analyzed as follows.

Mode 0 is the period before the main switch is closed.Due to the extremely short duty cycle, all the energy storageelements are discharged before the switch closes again. Hence,at the initial point, energy stored in the energy storage elementsis zero.

Mode 1 starts when the main switch is turned on. Duringthe time the switch turns on, magnetizing inductance (Lm)of the transformer charges and stores energy. The secondaryside of the transformer does not conduct because of thediode cutoff. When the main switch is turned off, mode 2starts.

In mode 2, similar to the operation of a conventional flybackconverter, the energy stored in the magnetizing inductance

Fig. 2. Operation waveforms of the proposed pulsed power modulator.Characteristics of spark gap input voltage (VS−In), spark gap output sidevoltage, i.e., output load voltage (VS−O), the parallel capacitor current (ICp ),capacitive load current (ICload ), magnetizing inductance current (ILm ), andmain switch gate signal (VGate) for various modes.

begins discharging to the transformer’s secondary side. Thetransformer secondary side current charges the parallel capac-itor (Cp) through the diode. As mentioned above, becauseof the resonance between the parallel capacitor (Cp) and thetransformer secondary inductance (Lsec), an approximate half-sine high-voltage pulse is generated. The spark gap input pulseenergy, which is equal to the energy stored in the parallelcapacitor (Cp), can be expressed as

EP = 1

2× CP × V 2

S−In (1)

where VS−In is the spark gap input pulse voltage, marked inFig. 1. Hence, we can figure out VS−In as below

VS−In =√

2 × EP

CP. (2)

The next mode begins when the pulse voltage reaches thecritical point of spark gap break down. In mode 3, the airbetween the spark gap breaks down due to which the outputside of the spark gap reaches the same potential as the inputside in an instant, which results in a fast-rising pulse on theload side. As a result, the approximate half-sine high-voltagepulse is sharpened to an approximate triangle pulse with afast-rising time. The breaking down of the spark gap can alsobe seen as a short circuit. Next, the sharpened high-voltagepulse is applied across the load.

Identical to the operation of a DCM flyback converter, theenergy reserved in the magnetizing inductance (Lm) will befully discharged, subsequently starting mode 4. In mode 4,because Lm is fully discharged, the Cp and capacitive load(Cload) begin to discharge through the parallel resistor (Rp).The discharging time depends on the values of Cp, Cload, andRp, which will be mentioned in Section III. The Cp, Cload, andRp get entirely discharged before the next switch-on signal,finally returning to mode 0, because the gate duty is extremelyshort.

However, if the diode is not used, in mode 4, part ofthe energy will be exchanged between Lm and Cp until itis exhausted, which causes negative component of spark gapoutput pulse voltage (VS−In) and spark gap output pulse voltage

Authorized licensed use limited to: Chung-ang Univ. Downloaded on December 14,2020 at 00:57:24 UTC from IEEE Xplore. Restrictions apply.

LIU et al.: LOW COST, FAST-RISING, HIGH-VOLTAGE PULSED POWER MODULATOR BASED ON A DCM 4389

TABLE I

SPECIFICATIONS OF THE PROPOSED PULSED POWER MODULATOR

Fig. 3. Total schematic of the proposed pulsed power modulator.

(VS−O), the tradeoff between whether use the diode or not willbe discussed in Section III.

III. DESIGN OF THE PROPOSED PULSED

POWER MODULATOR

A. Design of Components of the Proposed PulsedPower Modulator

The design of the proposed pulsed power modulator is basedon the design specifications listed in Table I. Additionally, asshown in Fig. 3, the overall system can be divided into threeparts, namely the input part, DCM flyback converter part, andthe output part. The component design is described from leftto right.

1) Design of the Input Part: As mentioned at the beginningof Section II, as a 110-V electricity grid power is used as aninput to this system, a rectifier and input filter capacitors areneeded to feed dc input to the flyback converter.

A full-bridge rectifier and the dc link, which contains teninput filter capacitors, are shown in Fig. 3. As the peak valueof the 110-V sine ac voltage is

Vin = VSupply × √2 ≈ 156 V (3)

capacitors with a maximum voltage of 200 V were selected.Furthermore, to reduce the volume of the device and improvethe integration, ten parallel 220-μF capacitors instead of one2200-μF capacitor were selected. Section III-B addresses thisissue in more detail.

Fig. 4. Operation waveforms of the proposed pulsed power modulator withoutdiode. Characteristics of spark gap input pulse voltage (VS−In), spark gapoutput side voltage, i.e., output load voltage (VS−O), and main switch gatesignal (VGate) for various modes.

2) Design of the DCM Flyback Converter Portion: Themain switch used is an Infineon IGBT FZ600R12KE4, rated1200 V/600 A, in practice. The spike voltage across the mainswitch can be high, hence, a compact design is implemented tominimize the inductance component responsible for the spikevoltage.

A UU 125 mm × 235 mm × 17 mm core is selected for thetransformer. Also to prevent voltage across the IGBT exceed1200 V, the following equation should be satisfied:

VCES = VS−In

N+ Vin < 1200 V (4)

where IGBT rating voltage VCES = 1200 V, flyback converteroutput pulse voltage, which is the same as the spark gapinput pulse voltage VS−In = 23 kV, and input voltage afterrectification and filtering Vin = 156 V. Hence, the transformerturns ratio can be calculated as N > 22.

Furthermore, the parallel capacitor (Cp) determines theoutput pulse energy, according to Table I and (2), and it canbe calculated as

CP = 2 × EP

V 2S−In

= 1.13 nF. (5)

Ideally, the peak value of the spark gap output side voltageis equal to the input side. However, in practice, loss occurs inthe spark gap, so lowering the value of the parallel capacitor(Cp) slightly increases the spark gap input pulse voltage.As the voltage across the parallel capacitor (Cp) is high, itis designed to be implemented in series with several high-voltage capacitors, which results in the final design value of0.94 nF.

Using the transformer secondary side diode is optional.If the diode is not utilized, as shown in Fig. 4, whichis different from Fig. 2, the output pulse has a negativecomponent because of the discharging loop of the parallelcapacitor (Cp), and the capacitive load (Cload). Also, the sparkgap input pulse voltage (VS−In) ringing appears while the mainIGBT is conducting. However, it does not affect the outputvoltage (Vs−o) because the spark gap is not conducting at thatmoment.

In contrast, if the diode is used, the negative componentof the output pulse will almost disappear. However, it needsplenty of diodes in series because of the resultant high voltage

Authorized licensed use limited to: Chung-ang Univ. Downloaded on December 14,2020 at 00:57:24 UTC from IEEE Xplore. Restrictions apply.

4390 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 48, NO. 12, DECEMBER 2020

across the diodes, which implies higher cost and volume.Hence, there is a tradeoff between higher cost and volumeand the presence of negative components.

3) Design of the Output Side: A straightforward spark gap,which is exposed in the air and simply made by iron screws,is used to sharpen the pulse in the proposed pulsed powermodulator. In theory, air begins to break down when stressedby a sufficiently high voltage such as about 3 kV/mm indry air. However, the breakdown voltage varies with the airhumidity conditions, hence, the spark gap spacing is designedto be adjustable. The spark gap can also be adapted forsharpening the high-voltage pulse to account for the swiftrate of breakdown of the air, which can significantly speedup the pulse rising time. In addition, as the spark gap spacingis adjusted for breakdowns at the highest pulse voltage on theinput side, the pulsewidth on the output side is reduced to halfof the input side. Thus, the spark gap can shorten the pulserising time and narrow the pulsewidth, making the output pulsemore in line with the requirements of gas or water treatment.However, the spark gap will be unstable after a long timeoperation, which means that periodic exchanges of the sparkgap are needed. Because of its simple construction, periodicexchanges of the spark gap do not incur much cost.

As mentioned in Section II, a parallel resistor (Rp) is usedfor discharging the parallel capacitor (Cp) and the capacitiveload (Cload). As the value of Cload is much smaller than Cp, itcan be approximately ignored. The value of the output pulseis determined from the value of parallel resistor using the RCdischarging equation, given as

= EPulse × e− tRP CP . (6)

From (6), after t = 2 × RpCp, the residual voltage of theparallel capacitor (Cp) is calculated 0.135Epulse, which impliesthat it is nearly discharged. Thus, we can assume that aftertwo RpCp times, the discharge is complete. Because the risetime of the output pulse is extremely short, output pulsewidth(WOPW) can be designed considering only Cp dischargingtime. According to Table I, output pulsewidth (WOPW) is0.5 μs, so the parallel resistor (Rp) can be calculated as

RP = 0.5μs

2CP= 266 �. (7)

We use an approximate value of 250 � for the resistor, Rp.Specifications of the proposed pulsed power modulator arelisted in Table I.

B. Structure Design of the Proposed PulsedPower Modulator

As mentioned in Section III-A, the primary side of theflyback converter needs to be designed to reduce the leakageinductance component to decrease the spike voltage acrossthe main switch. Therefore, 18 high-voltage wires connectedin parallel are used as the primary winding of the transformerto reduce the leakage inductance significantly. The measuredleakage inductance is 0.092 μH. Also to deal with the over-heating problem caused by enormous peak current throughthe primary winding, 18 parallel windings cause the current

Fig. 5. 3-D structure diagram of the flyback converter part of the proposedpulsed power modulator.

Fig. 6. Simulation circuit of the proposed pulsed power modulator.

flowing through each winding to become one eighteenth ofthe total.

Fig. 5 depicts the 3-D structure diagram of the flybackconverter of the proposed pulsed power modulator. Afterrectification, the input power is directly connected to the twoends of the ten parallel input filter capacitor, whose structure isshown in the figure. The rectified dc energy passes through theinput filter capacitor and flows into the 18 parallel transformerprimary winding. It also passes through the 22 turns of thesecondary winding wound around the primary winding, whichis omitted in this figure. Along with this, the main IGBT isalso shown. In the actual experiment, a heat sink can be addedabove the main IGBT.

IV. SIMULATION AND EXPERIMENT RESULTS

A. Simulation Results

The simulation of the proposed pulsed power modulator isimplemented using PSIM.

As there is no spark gap element in the PSIM, an idealswitch is used to model the spark gap, as shown in Fig. 6.The conduction interval of this ideal switch is set to startfrom the time when the spark gap input pulse voltage (VS−In)reaches the maximum value, and to end when the discharge iscompleted. The simulation results of the voltage waveforms ofspark gap input pulse voltage (VS−In) and spark gap output sidevoltage (VS−O) are presented in Fig. 7. We note that the pulserising time on the output side of the spark gap is significantlyshorter than that on the input side. The function of the sparkgap in sharpening the pulse is also confirmed in the simulation.Furthermore, an output pulse of 23 kV was obtained.

Authorized licensed use limited to: Chung-ang Univ. Downloaded on December 14,2020 at 00:57:24 UTC from IEEE Xplore. Restrictions apply.

LIU et al.: LOW COST, FAST-RISING, HIGH-VOLTAGE PULSED POWER MODULATOR BASED ON A DCM 4391

Fig. 7. Simulation waveforms of the proposed pulsed power modulator.

Fig. 8. Photograph of the gas treating reactor.

B. Experiment Results

In this experiment, a gas treating reactor, shown in Fig. 8,was used as the load to prove the experimental results. Thephotograph of the flyback converter portion of the proposedpulsed power modulator is shown in Fig. 9. The pulse outputfrom it is sharpened by the spark gap and connected to the gastreating reactor. Furthermore, we measured the output pulseacross the gas treating reactor and compared the advantagesand disadvantages of the presence or absence of the diodes.The pulse output was measured with a Tektronix P6015A high-voltage probe.

The waveforms of the spark gap input and the spark gapoutput pulse voltage when the diodes are employed are shownin Fig. 10. The spark gap sharpens the pulse and dramaticallyshortens the pulse rising time as shown in Fig. 10(a). More-over, as shown in Fig. 10(b), the pulse rising time is measuredas 5 ns and it indeed meets the requirement of exhaust gas orsewage treatment.

Fig. 9. Photograph of the flyback converter portion of the proposed pulsedpower modulator.

Fig. 10. Experimental results when diodes are present. (a) Flyback converteroutput voltage VS−In and spark gap output pulse voltage VS−O (200 ns/div).(b) Spark gap output pulse voltage VS (20 ns/div).

The comparison between the waveforms with and with-out the diodes is shown in Fig. 11. When diodes are absent,the output pulse has a larger amount of negative componentcompared to the case when the diodes are present in which

Authorized licensed use limited to: Chung-ang Univ. Downloaded on December 14,2020 at 00:57:24 UTC from IEEE Xplore. Restrictions apply.

4392 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 48, NO. 12, DECEMBER 2020

Fig. 11. Spark gap input pulse voltage VS−In and spark gap outputpulse voltage VS−O waveforms with and without diodes. (a) Without diodes.(b) With diodes.

Fig. 12. Output pulse voltage operated at 500 Hz.

the negative component of the output pulse is almost entirelyremoved. However, both cases have advantages and disadvan-tages. When the diodes are present, a more stable nonnegativepulse can be obtained, which can be used on a load withhigh-quality pulse requirements. However, it also leads to a13.3% of voltage drop in Fig. 11 and cost increase. On theother hand, when diodes are not used, the issue of cost andvolume can be mitigated but only by compromising on thequality of the output pulse. The proposed device can be easilyswitched between the two modes, with or without diodes, tosuit different output requirements. In the case where the loadis a gas treating reactor, the characteristic of the output pulseshould only be considered only the voltage value and the risetime. Since the usage of diodes is only relating to the polarityof output voltage, we removed the diodes to minimize the cost.The pulse waveform operation at 500 Hz is shown in Fig. 12,for which an output pulse of about 24 kV is verified.

V. CONCLUSION

This article presented a new type of pulsed power modulatorbased on a DCM flyback converter. Through simulation andexperiment, the feasibility of the design parameters was con-firmed. In addition, the two topologies of the proposed pulsedpower modulator were analyzed. In the case where the load is a

gas treating reactor, the topology without diodes can minimizeboth cost and volume. We confirmed 24 kV, 0.5-μs pulsewidth,5-ns rising time, and 500-Hz output pulse in the experiment.While the stacking structured pulsed power modulators aremainly used in industrial applications because of their highcost, the proposed pulsed power modulator fulfills the aims ofboth small volume and low cost, which are aligned with theneeds of small household appliances. Further, the topology,with or without diodes, is selectively considered for differentapplications.

REFERENCES

[1] H. Akiyama, T. Sakugawa, T. Namihira, K. Takaki, Y. Minamitani, andN. Shimomura, “Industrial applications of pulsed power technology,”IEEE Trans. Dielectr. Electr. Insul., vol. 14, no. 5, pp. 1051–1064,Oct. 2007, doi: 10.1109/TDEI.2007.4339465.

[2] H.-S. Kim, C.-H. Yu, S.-R. Jang, G.-H. Kim, and H.-J. Ryoo, “Highvoltage pulsed power modulator with high reliability and fast switchingspeed for medical lasers,” in Proc. IECON-42nd Annu. Conf. IEEEInd. Electron. Soc., Florence, Italy, Oct. 2016, pp. 3415–3418, doi: 10.1109/IECON.2016.7793694.

[3] H.-J. Ju, J.-H. Park, K.-C. Ko, and E. Hotta, “Discharge processes ofNO gas using bidirectional pulsed voltage,” in Pulsed Power PlasmaSci. 28th IEEE Int. Conf. Plasma Sci. 13th IEEE Int. Pulsed PowerConf. Dig. Papers, Las Vegas, NV, USA, Jun. 2001, pp. 1106–1109,doi: 10.1109/PPPS.2001.1001738.

[4] M. G. Grothaus, M. K. Khair, P. Paul, E. R. Fanick, and D. R. Bannon,“A synergistic approach for the removal of NOx and PM from dieselengine exhaust,” in Dig. Tech. Papers. 12th IEEE Int. Pulsed PowerConf., Monterey, CA, USA, vol. 1, Jun. 1999, pp. 506–510, doi: 10.1109/PPC.1999.825521.

[5] K. Takaki, S. Kudo, S. Kato, S. Mukaigawa, and T. Fujiwara, “Multipointbarrier discharge for removal of NOx from diesel engine exhaust,” inProc. Pulsed Power Plasma Sci. 28th IEEE Int. Conf. Plasma Sci. 13thIEEE Int. Pulsed Power Conf., Las Vegas, NV, USA, vol. 2, Jun. 2001,pp. 1126–1129, doi: 10.1109/PPPS.2001.1001743.

[6] S. Boev and N. Yavorovsky, “Electropulse water treatment,” in Dig.Tech. Papers. 12th IEEE Int. Pulsed Power Conf., Monterey, CA, USA,Jun. 1999, pp. 181–184, doi: 10.1109/PPC.1999.825442.

[7] H.-S. Jin, S.-H. Song, C.-G. Cho, S.-M. Park, and H.-J. Ryoo, “Study ofexhaust air treatment from a ship building factory painting facility usingpulse plasma technology,” IEEE Trans. Plasma Sci., vol. 46, no. 10,pp. 3552–3556, Oct. 2018, doi: 10.1109/TPS.2018.2829494.

[8] S. R. Jang, H. J. Ryoo, Y. S. Jin, S. H. Ahn, and G. H. Rim,“Application of pulsed power system for water treatment of the leachate,”in Proc. IEEE Pulsed Power Conf., Washington, DC, USA, Jun. 2009,pp. 980–983, doi: 10.1109/PPC.2009.5386150.

[9] S.-R. Jang, C.-H. Yu, and H.-J. Ryoo, “Simplified design of a solidstate pulsed power modulator based on power cell structure,” IEEETrans. Ind. Electron., vol. 65, no. 3, pp. 2112–2121, Mar. 2018,doi: 10.1109/TIE.2017.2745472.

[10] D. Malviya and M. Veerachary, “A novel boost converter basedhigh-voltage pulsed-power supply,” in Proc. IEEE Int. Conf. Sustain.Energy Technol. (ICSET), Bhubaneswar, India, Feb. 2019, pp. 353–358,doi: 10.1109/ICSETS.2019.8745027.

[11] M. Zhang et al., “Design of high voltage pulsed power supply forHFRC,” IEEE Trans. Plasma Sci., vol. 48, no. 6, pp. 1688–1692,Jun. 2020, doi: 10.1109/TPS.2020.2987038.

[12] S. Mohsenzade, M. Zarghani, and S. Kaboli, “A high-voltage series-stacked IGBT switch with active energy recovery feature for pulsedpower applications,” IEEE Trans. Ind. Electron., vol. 67, no. 5,pp. 3650–3661, May 2020, doi: 10.1109/TIE.2019.2921297.

[13] S.-H. Song, C.-G. Cho, S.-M. Park, H.-I. Park, and H.-J. Ryoo, “Designand implementation of novel series trigger circuit for xenon flash lampdriver,” IEEE Trans. Plasma Sci., vol. 46, no. 10, pp. 3584–3590,Oct. 2018, doi: 10.1109/TPS.2018.2842200.

[14] T. Matsumoto, D. Wang, T. Namihira, and H. Akiyama, “Energy effi-ciency improvement of nitric oxide treatment using nanosecond pulseddischarge,” IEEE Trans. Plasma Sci., vol. 38, no. 10, pp. 2639–2643,Oct. 2010, doi: 10.1109/TPS.2010.2045903.

[15] E. J. M. van Heesch, G. J. J. Winands, and A. J. M. Pemen, “Evaluationof pulsed streamer corona experiments to determine the O∗radical yield,”J. Phys. D: Appl. Phys., vol. 41, no. 23, Dec. 2008, Art. no. 234015.

Authorized licensed use limited to: Chung-ang Univ. Downloaded on December 14,2020 at 00:57:24 UTC from IEEE Xplore. Restrictions apply.

LIU et al.: LOW COST, FAST-RISING, HIGH-VOLTAGE PULSED POWER MODULATOR BASED ON A DCM 4393

Chang-Yu Liu received the B.S. degree inautomation from the Central South Universityfor Nationalities, Wuhan, China, in 2016, andthe M.S. degree in electrical energy engineeringfrom Chung-Ang University, Seoul, South Korea,in 2020.

He is currently with the Critical Power SolutionDivision (CPSD), APAC, Eaton.

Chan-Gi Cho (Graduate Student Member, IEEE)received the B.S. degree in information display engi-neering from Kyung-Hee University, Seoul, SouthKorea, in 2016, and the M.S. degree in energysystem from Chung-Ang University, Seoul, in 2018,where he is currently pursuing the Ph.D. degree withthe Department of Energy System Engineering.

His current research interests include resonantconverters and high-voltage pulse power systems.

Seung-Ho Song (Member, IEEE) received the B.S.degree in electrical engineering from Kwang-WoonUniversity, Seoul, South Korea, in 2016. He iscurrently pursuing the M.S. and Ph.D. degrees withthe Department of Energy Engineering, Chung-AngUniversity, Seoul.

His current research interests include soft-switchedresonant converter applications and high-voltagepulse power systems.

Hong-Je Ryoo (Senior Member, IEEE) received theB.S., M.S., and Ph.D. degrees in electrical engineer-ing from Sungkyunkwan University, Seoul, SouthKorea, in 1991, 1995, and 2001, respectively.

From 1996 to 2015, he joined the Electric Propul-sion Research Division as a Principal ResearchEngineer and the Korea Electrotechnology ResearchInstitute, Changwon, South Korea, a Leader withthe Pulsed Power World Class Laboratory and theDirector of the Electric Propulsion Research Center.From 2004 to 2005, he was a Visiting Scholar

with the Wisconsin Electric Machines and Power Electronics Consortium(WEMPEC), University of Wisconsin–Madison, Madison, WI, USA. From2005 to 2015, he was a Professor with the Department of Energy ConversionTechnology, University of Science and Technology, Deajeon, South Korea.In 2015, he joined the School of Energy Systems Engineering, Chung-AngUniversity, Seoul, South Korea, where he is currently a Professor. His currentresearch interests include pulsed-power systems and their applications, as wellas high-power and high-voltage conversions.

Dr. Ryoo is the Academic Director of the Korean Institute of PowerElectronics, a Senior Member of the Korean Institute of Electrical Engineers,and the Vice President of the Korean Institute of Illuminations and ElectricalInstallation Engineers.

Authorized licensed use limited to: Chung-ang Univ. Downloaded on December 14,2020 at 00:57:24 UTC from IEEE Xplore. Restrictions apply.