Lichtbogenverlauf Usa i

2028 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 40, NO. 8, AUGUST 2012

Arc Movement Inside an AC/DC Circuit BreakerWorking With a Novel Method of Arc Guiding:Part I—Experiments, Examination, and Analysis

Harald Hofmann, Christian Weindl, Malik I. Al-Amayreh, and Ove Nilsson

Abstract—Within this project, the mode of operation of thecircuit breaker and its new innovative method of arc guiding wereanalyzed and verified. The solution refers to a switching devicethat is able to deal with both ac switching loads and bidirectionaldc switching loads. It is primarily intended for use in UIC-capableswitching units for voltages up to 3 kV and currents up to 800 A.The concept uses a combination of permanent and electromagneticblowout fields. This completely new and innovative approachis intended to permit activation of the blast coils by the arcsthemselves to generate the electromagnetic blowout fields withoutneed for additional electrical switching contacts. The combiningof a newly developed optical data acquisition system togetherwith the conventional recording of electrical parameters made theverification of the working hypothesis of the switching processpossible. A numerical model is used for the simulation of thelater stage of the extinguishing process, where the arc is driventoward the arcing chamber by the superposed magnetic fields ofpermanent magnets and blast coils, until it is extinguished dueto elongation and cooling by the arc splitter stack. The resultsof the measurement data analysis and theoretical modeling andsimulation of the extinguishing process led to the identification ofcritical operational areas and resulted in a successful optimizationof the contactor.

Index Terms—Arc guiding, circuit breaker, electrical contactor,ionized gas guiding, thermal plasma.

I. INTRODUCTION

FOR a long time, requests for the development of circuitbreakers that are able to break ac currents as well as dc

currents have been made. The difficulty herein is to achieve acomplete extinction of the arc at shutoff cycle for both types ofload. Within dc circuit breakers, permanent magnets are usuallyused to generate a specific magnetic field that forces the arc tomove toward the arcing chamber. In the arcing chamber, theelongation or splitting of the arc and its extinction by coolingare achieved by the use of fan-shaped plates made of ceramic

Manuscript received December 22, 2011; accepted February 20, 2012. Dateof publication June 18, 2012; date of current version August 7, 2012. Thiswork was supported by the Bavarian Research Foundation under the referencenumber AZ 746-07.

H. Hofmann and C. Weindl are with Lehrstuhl für Elektrische En-ergieversorgung, Universität Erlangen, 91058 Erlangen, Germany (e-mail:[email protected]; [email protected]).

M. I. Al-Amayreh is with Lehrstuhl für Strömungsmechanik, Univer-sität Erlangen, 91058 Erlangen, Germany (e-mail: [email protected]).

O. Nilsson is with Schaltbau GmbH, 81829 München, Germany (e-mail:[email protected]).

Color versions of one or more of the figures in this paper are available onlineat http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/TPS.2012.2200697

or metal [1]. Due to the use of permanent magnets, the functionof the circuit breaker depends on the polarity of the current.If a reverse current were to be applied to the switch, the arcwould move into the opposite direction, away from the arcingchamber. AC circuit breakers, on the other hand, generally usean electromagnetically generated blowout field because, at accurrents, the polarity of the current changes with every halfsine wave. The coils used to produce this blowout field areconnected to the same electrical circuit as the contactor itselfso that the magnetic field and the arc oscillate in phase becauseof the resulting electromagnetic force

−→F =

−→B ×−→

I . The resultof this approach is that the arc is driven into the arcing chamber,regardless of the polarity of the current. A disadvantage of theelectromagnetic blowout in normal operation, however, is thepermanent current conduction and the resulting heating ofthe blast coils. Another problem arises for the breaking of smallcurrents because the strength of the magnetic field generated isnot strong enough to drive the arc into the arcing chamber [2].The problem of the permanent activation of the coils can besolved by using leading contacts [3]. This design can achievethe activation of the blast coils shortly before the release of themain contacts. Unfortunately, this technical approach has notperformed very reliably in practice. Here, the development of acontactor with a novel method of arc guiding is managed, whichis able to deal with both ac switching loads and bidirectionaldc switching loads [4]. It is primarily intended for use in UIC-capable switching units for voltages up to 3 kV and currentsup to 800 A. The concept uses a combination of permanentmagnetic and electromagnetic blowout fields. This completelynew and innovative approach is intended to permit the acti-vation of the blast coils by the arcs themselves to generatethe electromagnetic blowout fields without need for additionalelectrical switching contacts.

Fig. 1 illustrates the basic working hypothesis of the contac-tor. The moving contact separates from the two fixed contacts,and an arc arises between the moving contact and each of thefixed contacts. The two fixed contacts are mounted in a defineddistance to the arc-guiding rails. The two arcs are deflected hor-izontally in the same direction by the permanent magnetic fields[5]. The arc moving inward commutes from the fixed contact tothe arc-guiding rail and activates the first electromagnetic blastcoil, which generates an electromagnetic field that deflects thearc in vertical direction. The arc commutates to the oppositearc-guiding rail, whereby the second blast coil, which generatesan electromagnetic field with the same direction, is activated.The arc moving outward extinguishes due to the contact link

0093-3813/$31.00 © 2012 IEEE

HOFMANN et al.: ARC MOVEMENT INSIDE AN AC/DC CIRCUIT BREAKER—PART I 2029

Fig. 1. Stages of the switching operation. Phase 1: Ignition. Phase 2: Commutation. Phase 3: Migration.

Fig. 2. Regions and associated positions of the optical fibers.

going dead. The remaining arc is accelerated in the verticaldirection and driven into the arcing chamber [6], where it isextinguished due to elongation and cooling. The aim of thisproject was to verify the movements of the arcs and to optimizethe design of the contactor.

II. EXPERIMENTALS

A. Test Setup

The task includes the detection and modeling of the arcmovement inside the circuit breaker resulting from the switch-ing process and the verification of the working hypothesis ofthe contactor. The voltage and current waveforms as well asthe positions of the arcs have to be recorded at any time of theswitching process to be able to prove the functional principle byexperiment. Light detectors in conjunction with optical fibers

have shown very good performance for the investigation of arcposition and arc movement [7]. A new optical measurementsystem, based on photodetectors with transimpedance ampli-fiers, which are connected to the device via optical fibers, wasdeveloped and successfully tested. For the first time, the record-ing of the position, intensity, and propagation characteristicsof the arcs within the switch with high temporal and spatialresolution could be achieved with this device. Holes had to bedrilled into the case of the circuit breaker at defined positionswhere the optical fibers are inserted. The positions of the holeswere selected to achieve maximum technical and scientificknowledge from the readings of the photodetectors as well asminimum possible attenuation of the magnetic field. The devicehad to be divided into defined regions, each of them assigned toa specific stage of the switching process, to simplify the evalua-tion of the different stages of the switching process (see Fig. 2).

2030 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 40, NO. 8, AUGUST 2012

Fig. 3. Tapping of the electrical and optical signals.

For preliminary investigations, the power for the test circuitwas supplied by the battery system of the Institute for ElectricalEnergy Supply of the University of Erlangen, which has a no-load voltage of 436 V and a short-circuit current of 5800 A. Dueto the limited voltage of the battery system, the existing high-power ac system at the department, which provides a three-phase adjustable output voltage in the range of 52–250 V with amaximum output current of 40 kA, has been extended to supplythe voltages and currents needed for the inspections and tests.An additional transformer is used to convert the output voltagerange of the high-current system to the operating voltage rangeof the contactor. Depending on the wiring, voltages up to 8 kVat currents up to 1.4 kA or 4 kV at currents up to 2.8 kAcan be provided. The current limitation is done via externalimpedances on the primary side of the transformer. To be able touse the large current and voltage range of this configuration fordc tests as well, a diode bridge rectifier was developed, whichis composed of 24 single high-power diodes and matching heatsinks. Each diode is rated for a continuous current of 830 A ata reverse voltage of 5 kV. The three-phase construction of therectifier with four diodes per string allows the interconnectionof the diodes either for maximum reverse voltage (5 kV ×4 = 20 kV), maximum current (830 A × 4 = 3.32 kA), orcombined (10 kV, 1.66 kA). The output voltage can be adjustedin the range of 1–11 kV and has a ripple of less than 13.4%without smoothing. The entire system is protected by surge andovervoltage arresters at the input and output.

B. Data Acquisition

Measurement systems, which are able to acquire the voltageand current waveforms as well as the light emissions at specificpoints of the device, are used to determine a correlation betweenthe electrical parameters and the migration of the arc within

the specimen. The electrical parameters are recorded with anIMC Polares universal meter for power measurement, andlight emissions are recorded with photodetectors and a high-speed camera. All measurement systems are synchronized by acommon trigger signal to provide synchronized time stamps fordata acquisition. A LabView application is used for recording,visualization, and analysis of the measured data. The measure-ment setup is shown in Fig. 3.

It is essential to record electrical measurement data frominside the switch to be able to do a detailed analysis of theswitching process. Thus, the internal connections from the blastcoils to the arc-guiding rails are cut and redirected to theoutside. High-voltage dividers are used to adapt to the input-voltage range of the data acquisition system. Because of thepotential-free current measurement via Hall effect transducersand the common ground of the high-voltage dividers, there isno need to use isolating amplifiers. The waveforms of the coilcurrents and voltage potentials of the coils and arc-guiding railsas well as the overall current and voltage can be determinedwith this test setup. Additional data such as arc voltage and arcpower as well as the strength and shape of the magnetic fieldcan be derived from the measured values [8].

The measurement of light emissions is performed via Si PINphotodetectors, which are connected to the device under test viaoptical fibers. A total of 96 holes are drilled into the case of thetest sample at defined positions to hold the optical fibers. Dueto the very low output voltages and the large internal resistanceof the photodetectors, transimpedance amplifiers are used forcoupling with the data acquisition system. Two 16-channel dataacquisition systems with a sampling rate of 50 kHz are usedfor capturing of the signal shapes, so up to 32 photodetectorsignals can be recorded simultaneously. Both the intensity ofthe arc and its propagation speed can be determined from theacquired data. The advantages of the optical measurement by

HOFMANN et al.: ARC MOVEMENT INSIDE AN AC/DC CIRCUIT BREAKER—PART I 2031

Fig. 4. Current and voltage characteristics of the switching process.

photodiodes are short response time (< 50 ns), which allowsgood temporal resolution of the signals, and a wide dynamicrange, so that very weak signals still can be resolved well,whereas very bright signals do not lead to saturation. However,the number of input points is limited due to the fact thateach photodetector requires a separate amplifier and a separatechannel on the data acquisition system. In order to overcome thedisadvantage of limited spatial resolution due to the small num-ber of photodetectors, parallel recording of the arc migration isperformed with a high-speed camera. The camera used is capa-ble of recording 10 000 images per second with a resolution of512 × 192 pixels. Thereby, the arc migration can be observed,but a detailed analysis of the intensity and propagation speedof the arc is only possible to a limited extent due to the limiteddynamic range and smearing effect of the camera.

III. RESULTS

Several sets of experiments with different types of voltageand current loads were performed. The load was varied to beable to determine the switching characteristics over the entireoperating range. For detailed analysis of each stage of theswitching operation (ignition, commutation, migration, and ex-tinction), it is necessary to assign a defined area of the contactorto each of the stages. The contactor is divided into differentregions, in which a particular stage of the switching process canbe observed (see Fig. 2). Each area of the specimen is visualizedin a separate window within the LabView application. Thisallows the direct observation of the influence of load and pa-rameter changes to the different stages of the switching process.In Fig. 4, the curves of breaking a current of 280 A at 3.6 kV acare shown. The contacts of the switch separate at t = 80 ms.

Stage 1: Ignition of the Arcs and Deflection by the PermanentMagnetic Field

The ignition and vertical migration of the two arcs can beobserved within regions 1A and 1B (see Fig. 5). The deflectionof the two arcs is solely caused by the permanent magneticfield in the horizontal direction. The inward shifting arc (region1B) initially moves with a speed of 20 m/s, which raises upto 40 m/s at commutation to the arc guide rail. The outwardshifting arc (region 1A) moves at first with a speed of 20 m/s,

and then, the speed decreases continuously until the arc finallystays at the position of the permanent magnet (see Fig. 6).The different voltage levels in Fig. 6 result from the differentspatial distances between the arc and the optical fiber head ofthe corresponding hole.

Stage 2: Commutation of the Inner Arc Into the Coil Channel

The inner arc commutes onto the arc guide rail and isexpanding into the right coil channel (region 2AB; see Fig. 7).The propagation speed is initially at 90 m/s and decreases untilthe arc reaches the blast coil at a speed of 45 m/s (see Fig. 8).By reignition, the regions are also passed several times. Fig. 8displays the first reignition within this switching process.

Stage 3: Expansion of the Arc Into the Arc Splitter Stack

The arc in the coil channel has died out because the voltageon the blast coils has dropped below the required arc voltage.Because of the electrical potential between the arc-guidingrails, the arc ignites in area B. Now, the total current is flowingthrough both of the blast coils and the arc and produces amagnetic field that deflects the arc horizontally toward thesplitter plates. The propagation speed is about 120 m/s inthe range of the sensors. At this point, it becomes plainlyvisible that the propagation characteristics of the arc can onlybe determined with access to a high temporal and spatialresolution. Despite the very high frame rate of the high-speedcamera of 10 000 frames/s and the very short exposure time of50 μs, the position of the arc is no longer uniquely determinedwithin single frames (see Fig. 9). The arc migration, however,can still be resolved well by the photodetectors with their shortresponse and decay time and the high sampling rate of the dataacquisition system (see Fig. 10).

IV. IDENTIFICATION OF CRITICAL

OPERATING CONDITIONS

Due to the fact that the electromagnetic blowout force isdirectly dependent on the current flowing through the blastcoils and because of the current dependence of the arcs’ ownmagnetic field [9], it becomes evident that the force on the arcis very limited at low currents. Only a small arc was observedin the test when breaking low currents (< 10 A) at a voltage of2.5 kV dc. After raising the voltage to 3.6 kV, slowly movingarcs were detected within the region of the contacts, sometimescommuting to the coil channels. The arcs moved along theedge of the contact bridge, as described in [10], and led to theoutgassing of the plastics at the inner walls of the contactorhousing. As a result, the switching times and, thus, the timethat the arcs resided at the predamaged walls increased fromtime to time until it finally led to a fire exit from the vent of thecircuit breaker due to burning plastic (see Fig. 11).

V. OPTIMIZATION OF THE CONTACTOR

The results of the investigation of the critical opera-tional conditions required a modification of the test object,

2032 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 40, NO. 8, AUGUST 2012

Fig. 5. Deflection of the arcs caused by the permanent magnetic field.

Fig. 6. Photodetector signals of regions 1A and 2A.

Fig. 7. Migration of the inner arc within the coil channel.

particularly to ensure that the switch is not damaged while cut-ting off low currents. As a solution, a ceramic plate was placedat those areas of the switch where outgassing of the plastichad occurred. To ensure that the modification of the contactorhas no adverse effects on the ultimate breaking capacity, themeasurements taken with the unmodified version were repeatedwith the modified version. It was observed that, within themeasurement accuracy, the modification has no effect on theswitching process when breaking high currents. The analysisof the performance at low currents showed that the switching

times increased even after the modification from time to time.The switching operations were performed at 5-min intervalswith a voltage of 3.6 kV dc and a current of 10 A. The switchingtime increased from 50 ms initially to over 1 s at the eighthswitching operation. After an idle time of 3 h, the experimentwas repeated. The first breaking also took 50 ms, and furtheroperations again showed a significant increase in switchingtimes. Further experiments showed that low switching timescan be achieved again by blowing out the test item. After eachswitching operation, compressed air was blown through the

HOFMANN et al.: ARC MOVEMENT INSIDE AN AC/DC CIRCUIT BREAKER—PART I 2033

Fig. 8. Photodetector signals of region 2AB.

Fig. 9. Migration of the arc within the expansion channel.

Fig. 10. Photodetector signals of regions B and C.

Fig. 11. Fire resulting from a nonmoving arc at a switching operation withcritical current.

sample for 30 s, and the switching times were back at 50 ms.Apparently, the blowing will remove conductive substances thataccumulate inside the device with every breaking operation, sofurther improvement of the contactor may be achieved by usingan improved arc gas venting system [11].

VI. MODELING AND SIMULATION

The mathematical modeling of flow and heat transfer withinthe contactor is based on the basic equations of fluid mechanics(Navier–Stokes equations) and the energy balance equations.These equations contain different mass, momentum, and

2034 IEEE TRANSACTIONS ON PLASMA SCIENCE, VOL. 40, NO. 8, AUGUST 2012

energy transport processes, including the convective transportof mass, momentum, and energy, the diffusive transport ofmass, momentum, and energy, and the thermal radiation. More-over, the Maxwell equations in conjunction with the covering ofchemical effects are taken into account. The numerical solutionof these equations is calculated by means of the finite volumemethod. The results of the modeling and the simulation areshown in part 2.

ACKNOWLEDGMENT

The authors would like to thank R. Kralik who inventedthe contactor studied and his colleague A. Ignatov for theprototypes and advice regarding the experiments.

REFERENCES

[1] K. Nakano and K. Takemura, “Circuit breaker with an arc suppressor,”Eur. Patent FR2 465 308, May 13, 1981.

[2] J. G. J. Sloot and G. M. V. Bosch, “Some conditions for arc movementunder the influence of a transverse magnetic field,” Holectechniek, vol. 3,pp. 98–106, 1972.

[3] F. Hollmann, “Mittelspannungs-Lasttrennschalter,” Eur. PatentDE3 332 684, Oct. 9, 1983.

[4] R. Kralik, “Schütz für Gleichstrom- und Wechselstrombetrieb,” Eur.Patent DE 102 006 035 844, Feb. 6, 2008.

[5] T. E. Browne, Circuit Interruption—Theory and Technique. New York:Marcel Dekker, 1984, pp. 641–679.

[6] N. Behrens, “Arc motion between opening and diverging electrodes,” inProc. Conf. Elect. Contact Phenom., 1978, pp. 243–247.

[7] J. W. McBride and P. M. Weaver, “High speed and medium resolution arcimaging,” in Proc. 17th ICEC, 1994, pp. 113–119.

[8] P. M. Weaver and J. W. McBride, “Magnetic and gas dynamic effects onarc motion in miniature circuit breakers,” in Proc. 39th IEEE Holm Conf.Elect. Contacts, 1993, pp. 77–85.

[9] R. Michal, “Theoretical and experimental determination of theself-field of an arc,” in Proc. 26th Holm Conf. Elect. Contacts, 1980,pp. 265–270.

[10] P. E. Secker and A. E. Guile, “Arc movement in a transverse magnetic fieldat atmospheric pressure,” Proc. Inst. Elect. Eng. A—Power Eng., vol. 106,no. 28, pp. 311–320, Aug. 1959.

[11] J. W. McBride and P. A. Jeffery, “The design optimisation of currentlimiting circuit breakers,” in Proc. IC-ECAAA Conf., 1997, pp. 354–360.

Harald Hofmann was born in Nürnberg, Germany,in 1968. He received the Dipl.Ing. degree in electricalengineering from the Friedrich-Alexander UniversityErlangen-Nürnberg, Erlangen, Germany, in 2002.

In the same year, he was recruited by the ModernDrive Technology GmbH as a Designing Engineerand became the Head of development in 2005. Hejoined the Institute of Electrical Power Systems atthe University of Erlangen-Nürnberg in 2008. Hisprimary research interests are electrical measure-ment engineering, switching behavior of ac/dc circuit

breakers, and novel measurement methods for the estimation of the remaininglifetime of electrical distribution systems.

Christian Weindl was born in Nürnberg, Germany,in 1965. He received the Dipl.Ing. degree in electricalengineering from the Friedrich-Alexander UniversityErlangen-Nürnberg, Erlangen, Germany, in 1993 andthe Dr. Ing. degree (cum laude) from the Instituteof Electrical Power Systems at the University ofErlangen-Nürnberg, Erlangen, in 1999/2000.

He worked in the High-Voltage Transmission andDistribution Department (Group System Planning)of Siemens AG, Erlangen, from 1993 to 1995 andjoined the Institute of Electrical Power Systems at

the University of Erlangen-Nürnberg in 1994. His primary research interestsare harmonic stability, control of converters and FACTS equipment, and theinteractions of these devices with the surrounding network. Since 2005, he hasheaded an international project in the field of the artificial aging of power cablesand the estimation of the remaining lifetime of electrical distribution systems.

In 1999, one of his papers won the literature award of ETG/VDE, and in2002, his Ph.D. work was awarded a research price by a major German utility(E-ON Bayern AG).

Malik I. Al-Amayreh was born in Erlangen,Germany, on October 09, 1981. He received the B.S.degree and the M.S. degree (with honor) from theMechanical Engineering Department, University ofJordan, Amman, Jordan, in 2004 and 2007, respec-tively. He is currently working toward the Ph.D.degree in the Institute of Fluid Mechanics, Univer-sity of Erlangen-Nürnberg, Erlangen, and is awardedwith an Alexander Mayer scholarship.

In 2007 to 2008, he was a Lecturer in the En-gineering Technology College at AlBalqa Applied

University, Salt, Jordan. In 2008 to 2010, he worked as a Researcher in theInstitute of Fluid Mechanics (LSTM) at the University of Erlangen-Nürnberg.His research interests include the applications of the flow field ionized gasesand gasification of oil shale using plasma.

Mr. Al-Amayreh is a member of the European Mechanics Society.

Ove Nilsson was born in Tavelsjö, Sweden, in 1956. He received the B.Sc.degree in material physics from the University of Umeå, Umeå, Sweden, in1980 and the Ph.D. degree from the Department of Experimental Physics,University of Umeå, in 1986.

At the Department of Experimental Physics at the same university, he didhis doctorate developing a new hot-wire method for the determination ofthermal conductivity and heat capacity under high pressure. From 1987 to1991, he was a Post-doc at the University of Würzburg, Würzburg, Germany,continuing in the field of thermal physics. In 1992, he joined the newly foundedBavarian Center of Applied Energy Research, Würzburg, where he worked asan Administration Manager and Scientist until 1998. After a period as SalesManager for Vitec GmbH, Würzburg, he joined Schaltbau GmbH, Munich,Germany, in 2001, where he currently works as a Research Engineer. Thecompany produces contactors, snap-action switches, connectors, and mastercontrollers.