techcommentarytechcommentaryPower Quality for Induction Melting in Metals Production

IntroductionFast, efficient batch meltingusing the modern inductionfurnace can improve operatingflexibility and productionyield, as well as reduce the costof environmental protection.The compelling advantages ofinduction batch melting haveencouraged foundries tochange the way they handletheir melting operations.Among the chief advantages iscost. The operational cost of atypical induction-melt furnaceis in the neighborhood of$130 per ton for steel, whichcompares favorably with theoperational cost of a typicalelectric arc furnace.

Since the 1970s, induction hasbeen the number one methodof melting in non-ferrousmetal foundries and animportant tool in ironfoundries. New technology is improvinginduction power supplies, furnacerefractory linings, heat recovery, andoverall system control. In the last tenyears, the use of induction melting hasincreased by as much as 20% per year,making it the fastest growing electrictechnology in metals production. Overtime, induction may even surpassconventional use of electric arc furnacesin both tons of production and kilowatt-hours of energy use.

Only a fraction the size of an electric arcfurnace, the induction-melt furnace maystill cause power quality problems in theelectric utility system. Power qualityproblems are more likely when aninduction-melt furnace is connected atdistribution-level voltages, where thefurnace current is relatively largecompared to the utility supply. In a fewcases, new installations of medium-frequency induction-melt furnaces havelead to difficult-to-resolve power interfaceproblems. This TechCommentary is

intended to help both furnace users andtheir energy providers better anticipateand resolve power quality problems relatedto induction melting.

Electrical Characteristics ofInduction FurnacesThe growth in use of induction meltinghas come about primarily because ofsignificant technologyadvances in the furnacepower supply and itsresonant circuit. This growthis primarily in medium-frequency systems, sized from0.2 to 16 MW and operatingat frequencies from 150 to3000 Hz. These medium-frequency furnaces haveproven to be versatile andefficient at a relatively largescale. The older low-frequency models, whichconnect directly to the60-Hz utility source, cannot

compete because of control andefficiency limitations. High-frequency systems, whichoperate at greater than 3 kHz,are relatively small and limitedto special applications.

Despite the appeal of themedium-frequency inductionfurnace, the same advances thatmake it so effective alsoengender problems with thepower interface. For example,consider harmonic distortion.Today, the most efficientfurnaces run at full power andvary the frequency to optimizethe melt. The furnace generatesfixed- and variable-frequencyharmonics that may lead toadverse interactions between thefurnace and the utility system.This is particularly true of thepopular high-power-densitycoreless-induction furnaces,which typically operate as a

relatively large load at distribution-levelvoltage.

The Furnace Circuit

Electrically, an induction-melt furnace issimply a loosely coupled transformer. Asshown in Figure 2, current in the powercoil surrounding a ceramic cruciblegenerates a magnetic field. Laminated iron

Figure 1. Modern variable-frequency induction enablesefficient batch-melting processes. The furnace is pouredempty after each melt, and successive melts are started usingunheated or preheated metal. Because there is no need tomaintain a molten heel, smaller furnaces can be used,electrical energy goes further, alloy changes are easier, andloading safety is enhanced. Induction batch melting hasbrought about meaningful increases in both the efficiencyand productivity of the modern melt shop.

Figure 2. Components of a Large Induction-Melt Furnace

2forms a magnetic yoke that also surroundsthe crucible. The crucible helps todistribute and contain the field, whichinduces current in the conductive metalto be melted. The current that penetratesthe metal is controlled to ensure properstirring of the metal and prevent over-stirring. A concentrated current on theouter layer of metal to be melted generatesthe melting power as it quickly heats tothe melting point. The refractory liningand cooling jacket separate the hot metalfrom the furnace power coil.

The power supply of an induction-meltfurnace provides both the power andcontrol required to properly melt metal.Early induction melting was carried outat line frequency, with power provided bya special transformer and tuning circuits.Switching capacitors provided power-factor adjustment, and changingtransformer taps regulated the powerlevel. For maximum melting power, theresonant frequency of a tuned LC circuithad to be matched to the line frequency.This condition limited the coil currentand therefore the efficiency of the furnacebecause the 60-Hz line frequency resultsin relatively high penetration into themelt and excessive stirring. Also, earlyinduction-melt furnaces were single-phaseloads, which draw heavily from only onephase of a three-phase system and limitthe power available from the utilityservice.

The advent of large-scale solid-state powersupplies has greatly improved inductionmelting. Three-phase converters, can beoperated at a high power factor, therebyincreasing the practical power ratings in atypical application. These power suppliesalso precisely control frequency and thedepth of penetration to efficiently melt thematerial without over-stirring. The processis more efficient because the variable-frequency power supplies are able tomatch the varying electrical characteristicsof different metals during melting. Theelectronic power supply that is used onthe modern furnace has opened the wayfor batch operation by eliminating theneed to maintain a molten heel. Figure 3shows a typical furnace and its electronicpower supply.

To achieve an efficient and fast melt, morepower per unit weight of the metal beingmelted is desired. Typical power densitiesin the medium-frequency furnace range

from 600 to 1000 kW per ton, comparedto 200 to 400 kW per ton in the line-frequency furnace. Electric arc furnacesalso operate at high power densities.However, the arc melting temperaturesare significantly higher than inductionmelting-in the neighborhood of 6000F(3300C), compared to 2600F (1400)for induction-melt furnaces. The highertemperature in an electric arc furnaceenables the melting of somewhat dirtiermetal but also leads to a few percent higherlosses of metal to ionization and creationof ash.

Figure 4 illustrates recorded electricalparameters of an actual medium-frequencyfurnace during a typical batch cycle.

1. Metal is loaded into the furnace usinga heavy duty conveyor. The meltingprocess forms a molten bath in thebottom of the ceramic crucible. In thiscase loading continues until the crucibleis filled with 10 tons (9 Mg) of moltenmetal, at about 2600F (1400C).

2. After all of the metal has melted, thebatch is superheated to the desired taptemperature of about 2800F (1500C).

3. Then, the power is removed so that themolten metal can be poured into a ladle.

Depending on the relative size andconfiguration, furnace operation can affectpower quality at the point of commoncoupling (PCC) or at the interface withthe public power supply. Many differentelectronic power supply configurations areavailable. The choice of configuration willhave a big impact on both furnaceperformance and electrical compatibility.The following is a summary of practicalpower supply options that may help toavoid costly corrective actions or operatingcurtailments when they are consideredbefore furnace installation.

The Electronic Power Supply Circuit

Electronic power supplies control currentand frequency for efficient inductionmelting. Most electronic power supplies

Figure 3. Typical Electronic Power Supply and Medium-Frequency Induction-Melt Furnace

Figure 4. Electrical Parameters During a Typical Batch Cycle for a 6-MW, 2400-VoltFurnace (scrap iron loaded at 10 tons/hour (9 Mg/h); power supply running at 200 to275 Hz, 36,000 Amps)

3rectify the AC line current to provide aDC source of energy. This DC is theninverted at a frequency to obtain thedesired induction from the furnaceresonant circuit. The two main types ofsolid-state power supplies are the current-fed power supply (parallel furnaceresonant circuit) and the voltage-fed powersupply (series furnace resonant circuit),both of which are used in medium-frequency induction melting systems.These different power supply designs havedifferent impacts on the furnaceperformance, as indicated in Table 1.

Voltage-Fed Power SuppliesThe voltage-fed power supply, which is thenewer of the two designs, takes advantageof switching technology that is capable ofhandling high currents. As shown inFigure 5, it employs a simple input dioderectifier to produce DC, and a parallel-connected DC capacitor for energy storageand filtering. The output inverter controlsmelting power by its commutationfrequency and can fully regulate thecurrent to the series tuning capacitor andthe furnace. Consequently, the inverter isexposed to the full current and partialvoltage of the furnace. The DC capacitorprovides or absorbs excess energy forstarting and stopping the inverter.

Current-Fed Power SuppliesAs shown in Figure 6, the basic current-fedconverter uses a phase-controlled rectifierto convert AC to DC and to regulate thevoltage on the DC link. When current isflowing in the DC link, the two series-connected inductors provide energystorage and filtering. Consequently, astarter circuit is needed to energize theseinductors, and a crowbar circuit is used todischarge them when the melt is complete.The inverter commutates or reverses thecurrent to obtain the desired outputfrequency, which, along with varying therectifier output voltage, controls themelting power.

Current-fed designs have been aroundlonger and take advantage of rugged andeconomical switching-device technology.The inverter is exposed to the full furnacevoltage. However, it only sees about 10%of the furnace resonant current because thereactive component of the furnace currentbypasses the inverter via the parallel tuningcapacitor. Consequently, the current-fedpower supply has less control over the

furnace current than the voltage-fed powersupply.

One side effect of using a phase-controlledrectifier in the current-fed power supply isvoltage notching. The line voltage isnotched because a momentary line-to-linefault occurs as each phase rectifier device isturned on before the other phase devicehas commutated off. Depth of the notchdepends on the circuit impedance between

furnace transformer and the rectifier.Width depends on the timing betweenturn-on and turn-off. Notches are moresevere near the converter, as illustrated inthe voltage waveform shown in Figure 6.Notching can cause equipment operatingproblems when propagated in a plantelectrical system. The most commonproblem caused by notching is tripping ofother power supplies and DC drives.

Table 1. Performance and Interface Comparisons for Medium-Frequency Melt Furnaces

Characteristics Current-Fed Inverter Voltage-Fed Inverter

Controllability of Melt Poor Excellent

Efficiency of Melt 70-80% 75-85%

Power-Line Interface Phase-Controlled Rectifier Diode Rectifier

DC-Energy Storage Inductive, Dynamic Capacitive, Static

Figure 5. Voltage-Fed Power Supply Driving a Series-Resonant Furnace Circuit

Figure 6. Current-Fed Power Supply Driving a Parallel-Resonant Furnace Circuit

Common Power QualityConcerns About InductionFurnacesSometimes thought to be a panacea for theinduction melting process, modern solid-state power supplies have actually been amixed blessing for achieving compatibilitywith the electric utility system. Largethree-phase power supplies for furnaces,while providing economies of scale withhigher productions levels, also bring largefluctuating load currents withvarying levels of harmonicdistortion. These varyingfurnace currents can affectdistribution line-voltageregulation and quality.Table 2 provides a summaryand comparison of powerquality concerns for both thecurrent- and the voltage-fedpower supply.

Generation of CurrentHarmonicsBoth current- and voltage-fedinverters generate harmonicsback into power lines in theprocess of rectifying AC toDC. In the larger furnaces, itis popular to provide more than onerectifier bridge, along with phase-shiftingtransformers. This reduces the amountof current per bridge and the level ofharmonics in the combined current drawnfrom the utility. Each three-phase bridgerequires six devices, and one positive andone negative pole for eachphase. A single bridge, such asshown in Figures 5 and 6, iscalled a six-pulse rectifier, twobridges a 12-pulse, as shown inFigure 7, and so on.

Increasing the number ofrectifier bridges adds moresteps in the waveform of the linecurrent, making it moresinusoidal. The harmonicsproduced by 6-, 12-, and24-pulse rectifiers are shownin Table 3. Assuming and idealsquare wave, a rectifier shouldonly have harmonics that are aninteger multiple of the numberof pulses 1. For example, in a12-pulse rectifier, the harmoniccomponents should be 11, 13,23, 25, and so on. However,

Rectifier Individual Harmonic Order and Levels (% of Fundamental)

Harmonic 5th 7th 11th 13th 17th 19th 23rd 25th Total

6-Pulse (I) 20 14.3 9.1 7.7 5.9 0.5 4.3 4.0 29%

6-Pulse (P) 17.5 11.0 4.5 2.9 1.5 1 0.9 0.8 21.5%

12-Pulse (I) 0 0 9.1 7.7 0 0 4.3 4.0 15.5%

12-Pulse (P) 2.6 1.6 7.9 5.5 0.2 0.1 2.3 0.8 10.4%

24-Pulse (I) 0 0 0 0 0 0 4.3 4.0 5.9%

24-Pulse (P) 2.6 1.6 0.7 0.4 0.2 0.1 1.9 0.8 3.8%

I = Ideal Square WaveP = Practical Case

Table 3. Ideal Square Wave and Practical Harmonic Spectrum for Furnace Rectifiers

due to unbalances, other harmonics arepresent in practical applications, as shownin Table 3. Even so, the total harmonics inmost practical applications are moderatelyless than theory predicts for ideal square-stepped waveforms where each individualharmonic (N) is 1/N of the fundamental.

Power Factor of Furnace andPower Supply

The term power factor is well definedfor 60-Hz systems as the phase differencebetween the fundamental current andvoltage. In the presence of harmonics,

power factor is best definedas the ratio of the wattsover the total kVA for allfrequencies. Distortedsystems have limited powerfactors even when thefundamental voltage andcurrent are in phase. Forexample, when the currentis 21.5% distorted, as withthe practical case of asix-pulse rectifier, themaximum power factor is0.98 instead of 1.0. At 60%distortion, the maximumpower factor is about 0.86.The expected power factorsfor 6-, 12-, and 24-pulsebridge rectifiers in full-waverectification mode arerelatively high, as shown in

4

Table 2. Power Quality Comparisons (Single-Rectifier-Bridge Configuration)

Current-Fed Voltage-FedCharacteristics Inverter Inverter

Line-Voltage Notching Yes (Caused by Phase Control) No

Harmonic Generation High Moderate

System Power Factor 0.7-0.95 (Depends on Phase Control) 0.95

Generates Inter-Harmonics Yes (Depends on Furnace Frequency) No

Figure 7. Typical 12-Pulse Bridge Rectifier Configuration with Phase-Shifting Transformers

Table 4. The greater the number of pulsesthe greater the expected power factor.Compared to the full-wave rectifier involtage-fed power supplies, the phase-controlled rectifier in current-fed powersupplies uses a delay in turning onswitches to control power levels and toregulate the DC bus voltage. It shouldbe noted that the power level and powerfactor in a phase-controlled rectifier dropsrapidly with the increase of delay angle.At a delay angle of 30, the maximum is0.95, at 90 it is less than 0.7, and at 120it drops to 0.46. Therefore, if the powersupply is current-fed, low power factorscan be expected when the power supplyreduces power to the furnace.

Voltage Fluctuations Caused by theFurnace CircuitIn addition to the harmonics that arenormally expected from different pulserectifiers large furnaces operating at a fewhundred hertz can generate significantnon-characteristic harmonics. Theseharmonics, which fluctuate with thefrequency of the furnace resonant circuit,are usually not multiples of the supplyfrequency, making them difficult to filter.This phenomenon, known as inter-harmonics, can overload power systemcapacitors, introduce noise intotransformers, cause lights to flicker,instigate UPS alarms, and trip adjustable-speed drives (see Inter-Harmonics inPower Systems).

The typical scenario for the generationof inter-harmonics is a relatively largefurnace with a current-fed power supplyoperating between 100 and 500 Hz ona distribution feeder. When the powersupply inverter is operating at frequencyf0, the frequency reflected back to the

rectifier is two times f0. This frequency

combines with the line frequency (60 Hz),resulting in line currents containingharmonics of two times f

0 60, 4f

0 60,

and so on. For example, a 12-pulserectifier feeding a furnace operating at123 Hz may have inter-harmonics at 186and 306 Hz, 432 and 552 Hz, and so on.These frequencies are not characteristic ofthe 12-pulse rectifier and are fed back intothe power system from the current of thefurnaces resonant circuit.

When inter-harmonics combine with thefundamental voltage, modulations of thepower system voltage may interact withother equipment. Light flicker is probablythe most common interaction problem.Many utilities have dealt with electric arcfurnaces as a large-scale cause of flickeringlights. Induction melting can also causeannoying lamp flicker. However, themechanism is related to voltagefluctuations resulting from inter-harmoniccurrents rather than from arcing currents.Figure 8 shows the fundamental 60-Hzvoltage with amplitude modulation ofapproximately 6 Hz, resulting from a186 Hz inter-harmonic. This voltagecauses a strong light flicker in most lamps.

The interaction level depends on therelative size of the furnace, its operatingfrequency, and loading. Also, the effectmight be aggravated by resonance in thesystem, which causes amplification of theinter-harmonic frequencies at certainpoints in the power system. New IEEE/

Inter-Harmonics in Power Systems

Inter-harmonic is a relatively new classification of power system distortion. Its effecton the power system is unique, as are the methods for measuring inter-harmonics andmitigating its effects. Inter-harmonics can be thought of as voltage or currentcomponents that are not related to fundamental frequency or to integer-harmoniccomponents of the system. As furnace power supplies become more sophisticated, thefrequencies of the current they draw are less likely to be limited to harmonics of thefundamental.

The equipment that causes inter-harmonics includes induction furnaces, static-frequencyconverters, cycloconverters, induction motors that drive shakers, and DC arc furnaces.Generally, any equipment that draws a load current that pulsates asynchronously withthe fundamental power system frequency generates inter-harmonics. In the case ofinduction melting, the variable frequency of the furnace is likely to cause inter-harmonics in the power system. The typical impacts on other equipment are flickeringlights or computer screens, tripping of certain power electronic equipment, and heatingin the power system similar to the heating caused by harmonic currents.

Standards are still emerging on this subject. IEEE 519-1992 indirectly addresses inter-harmonics in the discussion on cycloconverters. A future IEEE standard is expected toprovide general technical descriptions of the phenomenon, methods of measurement,and guidelines for limits. The IEC 61000-2-1 currently defines the inter-harmonicenvironment, and IEC 61000-4-7 describes a measurement technique. Even with thesestandards, agreement among popular harmonic monitors does not exist. If a monitorcan detect inter-harmonics, the most common result is an under-registration of theinter-harmonic levels.

IEC standards in flicker prediction,measurement, and assessment can be a bighelp in dealing with light flicker caused bythe operation of induction furnaces (seeStandards for Assessment of VoltageFluctuations and Lamp Flicker).

Solutions to InductionFurnace Power QualityProblemsWhen an induction furnace is causingpower quality problems, other customersare often involved. Both end user andpower provider want to consider allpractical solutions. Tools for avoidingand resolving typical problems includemeasurement and assessment methods,application of standards, changes in thefurnace or utility power supply, specialoperating procedures, and powerconditioning. Pre-installation planningand post-installation problem-solving for a

Figure 8. System Voltage Modulated by186-Hz Components

5

Number Power of Pulses Factor

6 0.955

12 0.988

24 0.997

Table 4. Expected Power Factor forFull-Wave Rectifiers

6specific foundry case willdemonstrate options for preventingand resolving power qualityproblems.

Pre-installation planning usuallystarts with an assessment of therelative size of the end usercompared to the utility powersource. Consider the installation ofa 2-MW furnace in a foundry on a12-kV distribution feeder. Concernsare harmonic generation, lightflicker, and interaction with otherequipment at the foundry. The firststep is to establish a point ofcommon coupling and calculate ashort-circuit ratio (SCR) of availableshort-circuit power (SSC) dividedby average maximum demandpower. Figure 9 illustrates thiscalculation for a foundry that has an SCRof 25.8 at the point of common couplingwith the distribution system.

Options to Control FurnaceHarmonics and Power Factor

IEEE Standard 519 providesrecommended harmonic distortion limits

for both end-user current and power-supplier voltage. The current limitsdepend on the relative size of the plantor its SCR. Current limits in 519 arecalibrated for the harmonic spectrum ofa six-pulse-rectifier. These current limitswill relax if the rectifier is a higher pulsenumber. Limits are given for both total

demand distortion (TDD)and individual harmonicdistortion. In most practicalcases, these individualharmonic limits are the mostrestrictive.

For example, assuming thereare no power-factor-correction capacitor banks atthe foundry, all the harmoniccurrents from the furnace arelikely to flow into the utilitydistribution system. At thepoint of common coupling(PCC), given a short circuitratio of 25.8, the totaldemand distortion limit inIEEE 519 is 8%, and theindividual single harmoniclimit will depend on the

rectifier type. With this information,the following procedure can be used todetermine the maximum furnace size atthe location:

1. Identify the harmonic spectrum of theparticular furnace to be connected.This will depend primarily on thefurnace rectifier type and the supply

Annoying lamp flicker can occur when rapid changes in loadcurrent cause the power system voltage to fluctuate. Bothincandescent and fluorescent lamps can flicker during voltagefluctuations. The standards for measuring and limiting lampflicker are based on the 60-Wincandescent lamp.

Assessing whether or not voltagefluctuations might result inobservable flicker can be done usinga flickermeter calibrated for a typicallamp and human eye-brain response.The best methods for thismeasurement were developed inEurope for 230-V incandescentlamps and are contained in standardspublished by the InternationalElectrotechnical Commission (IEC).These standards were recentlyadapted for 120-V lamps used inNorth America. In 1999, the Instituteof Electrical and ElectronicsEngineers (IEEE) accepted the IECmethod and will publish it as IEEEStandard 1354. As shown in thefigure, the flicker thresholds in this new standard are similarto flicker curves based on the early General Electric studiesconducted in the US in the 1920s. The difference between thenew standard and the IEEE flicker curves is that a measurementmethod is also specified in the new standard.

In the IEC method, the threshold of irritation is defined asPst=1, based on 60-W incandescent lamps and a short-term(10-minute) measurement period. For a long-term (two-hour)measurement period, a Plt is defined as the cube root of 12

successive Pst measurementsaveraged. The allowed percent ofvoltage change for Pst = 1 varieswith the frequency of voltagefluctuations. For example, at 120changes per minute (two changesper second), the IEC curveindicates that irritating flicker willresult from voltage fluctuationsthat are about 0.8 percent ormore of the nominal voltage. Atthat same frequency, the originalIEEE curves give a similar resultof about 0.7 percent. BecauseIEEE had no standard way tomeasure flicker, the IEC flickerstandards have been adopted.

Three IEC standards may help inresolving a flicker dispute. Limitson voltage fluctuation for

equipment greater than 16 amps are provided in IEC 61000-3-5.Methods for assessing fluctuating loads at medium and highvoltage are covered in IEC 61000-3-7. Flicker measurement isspecified in IEC 61000-4-15. The same measurement method isnow also included in IEEE Standard 1354.

Standards for Assessing Voltage Fluctuations and Lamp Flicker

Figure 9. Induction Furnace Point of Common CouplingShowing Relative Size with Power System(Max. DemandAVE =3.1 MVA, SCR = 80/3.1 = 25.8)

7This method demonstrates that forpractical cases, the allowed furnace sizeincreases as the pulse number of the powersupply increases. Figure 10 illustrates theIEEE 519 recommended limit for therelative size of the furnace based onindividual harmonic levels from Table 3.Another rule of thumb sometimes used isthat furnaces greater than 2 MW shouldbe 12-pulse, and furnaces greater than10 MW should be 24-pulse.

Sometimes the actual furnace harmonicsare higher than predicted because ofinsufficient series reactance, unbalanceloading, or resonance with other powersystem filters and equipment. When IEEElimits are violated, some form of series-balancing reactor or parallel-tuned filtermay be required. Consult the furnacemanufacturer to determine if changes tothe furnace are practical. If harmonicfrequencies are changing during furnaceoperation, an on-site filter may be verydifficult to apply. In nearly every case, astudy is needed to select the best solutionfor the application.

Options to Control Furnace-Related Light Flicker

Predicting Flicker ComplaintsThe first step in assessing furnace-relatedlight flicker is to monitor furnace operationand voltage fluctuations simultaneously.Flickermeters calibrated for 120-V lampsare available. Measurements can be quiteeffective in predicting flicker complaints,

even when voltage fluctuations are causedby inter-harmonics. By comparing theflicker levels with the furnace operatingmode, useful correlations may beobtained. When attempting to reduceflicker levels, the meter will provide quickfeedback following changes in the powersystem configuration, the furnaceoperation, or the PCC. Measurementsbefore and after furnace installation areusually helpful in diagnosing andcorrecting flicker problems.

Several standard assessment methods areavailable for use prior to furnaceinstallation. Use a short-circuit ratio testfor initial screening where flicker problemsare not expected. This assessment is simplybased on the ratio of the power change,S, divided by the available short circuitpower, S

SC, at the PCC. The limits to be

applied for automatic acceptance of thefluctuating furnace load also depend onfrequency of load changes, Cf, as shown inTable 5. Note that this simple method isnot effective in cases where either S or C

f

are not predictable. Inter-harmonicsintroduced into the power system from thefurnace resonant circuit may be one ofthese unpredictable cases.

If the voltage waveform at the PCC can bedescribed in a digital waveform, then astandard flickermeter simulation will givethe expected Pst. When the general shapeof the voltage fluctuation is known, IEC61000-3-3 provides shape-factor charts topredict flicker levels after installation. Inthe simple case of rectangular voltagevariations, with a known and a fixedfrequency, the traditional flicker curve canbe used to predict complaints. Table 6summarizes available assessment methods.

Existing Flicker ProblemsWhen the installed furnace is alreadycausing flicker complaints, the most likelysource is inter-harmonics. Experience hasshown that the current-fed converter,without sufficient filtering, promotes the

Figure 10. Limits on Furnace Size as a Percent of Plant Load forIdeal and Practical Conditions

Cf (Changes/Minute) S/SSC (%)

>200 0.1

10 to 200 0.2

EPRI Center for Materials Production 1251 Dublin Road Columbus, OH 43215614.225.2590 bmunk@tarateccorp.com

EPRI Corporate address 3412 Hillview Avenue, Palo Alto, CA 94304 PO Box 10412, Palo Alto, CA 94303 USA800.313.3774 650.855.2121 askepri@epri.com www.epri.com

TC-114625

8

1999 Electric Power Research Institute (EPRI), Inc.

All rights reserved. Electric Power Research Institute andEPRI are registered service marks of the Electric PowerResearch Institute, Inc. EPRI. POWERING PROGRESS is aservice mark of the Electric Power Research Institute, Inc.

Printed on recycled paper in the United States of America.

Photograph courtesy of Inductotherm Corp.

Applicable SIC Codes:332, 334, 336, and 339

To order additional copies of this publication call800.313.3774 or e-mail askepri@epri.com.

interaction of furnace and power-linefrequencies to cause voltage fluctuations.

Adding series inductance inside thefurnace power supply at the DC link oradding parallel capacitance on the powerline may reduce the flicker. The inductorswill also reduce the propagation ofnotching. However, side effects suchas reduced voltage at the furnace andovervoltage at the plant bus must also beconsidered. Flagging specific troublesomeoperating levels or frequencies andcontrolling the furnace to avoid theseoperating points can be an effective wayto mitigate inter-harmonic interactions.Less desirable measures include restrictingoperation to only certain hours duringthe day, increasing the service capacity,or requiring reconfiguration of thedistribution feeder to reduce interactions.

In some cases, a Pst level close to unitymay still result in isolated flickercomplaints. Most likely these flickeringlamps are more affected by voltagefluctuations than the standard 60-W

incandescent. Some fluorescents,particularly compacts, and low-wattageor dimmed incandescent lamps are veryprone to flicker. Also some people aremore sensitive to light changes thanaverage. In these cases, correction at pointof complaint, such as changing the lampsor adding a fast voltage regulator, may becost-effective.

Topics for FutureInvestigationActive control of the furnace frequencyand power to avoid adverse interactionswith the utility may be effective inmaintaining a compatible interfacebetween furnaces and utility service. Theimpact of utility voltage sags, momentaryinterruptions, and switching transients isgaining importance as furnaces becomemore sophisticated with sensitive processcontrols. Improved inter-harmonicmeasurement and elimination methodsare needed as more melting is carried outby induction.

AcknowledgmentsDr. Oleg Fishman of Inductotherm Corp.,and Nicolas Cignetti, private consultant,provided reference materials and valuabletechnical input.

Other Resources TechCommentary Induction Melting,

CMP-72, 11/91

CMP TechApplications

Induction Melting for aCompetitive Advantage, CMP-048,2/90

Induction Melting for HigherProductivity, CMP-1188-018

Induction Melting for BusinessBuilding, CMP-1289-020

Induction Melting for PollutionElimination, CMP-1289-010

Induction Melting for OperatingFlexibility, CMP-1289-021

Table 6. Methods for Predicting Load-Related Flicker Complaints (Pst or Plt)

Type of Voltage Change Data Assessment Method Standard Reference

Actual Load Operating Flickermeter IEEE 1354, IEC 61000-4-1

Predicted kVA Change Short-Circuit Ratio Test IEC 61000-3-7

Digitized Waveforms of Change Flickermeter Simulation IEC 61000-4-15

Typical Shape of Voltage Change Standard IEC Shape Factors IEC 61000-3-3

Rectangular Change, Fixed Rate Traditional Flicker Curve (Pst=1) IEEE 1354, IEC 61000-3-3