steel plant performance, power supply system design … · steel plant performance, power supply...

Draft Paper abstract - 54th ELECTRIC FURNACE CONFERENCE - Dec. 96. 96.11.08

1

STEEL PLANT PERFORMANCE, POWERSUPPLY SYSTEM DESIGN AND POWER

QUALITY ASPECTS

Power Quality leads to improved production andcosts savings

Antonio Silva

ABB Power T&D Company Inc.Raleigh NC

Lars Hultqvist and Aleksander Wilk-Wilczynski

ABB Power Systems ABVästerås Sweden

ABSTRACT

For a steel producer optimum performance,namely cost-efficient production, is of primaryinterest. Therefore, from his point of view, anyinvestment in the electric power supply is alwaysseen as a cost. This applies in particular toPower Quality equipment such as SVC, which isjustified only if it enables increased productionor energy savings.

Different aspects of this issue are highlightedin the paper.

1. INTRODUCTION

Views have been expressed on different occasionsthat all efforts to increase power quality in the steelplant supply systems only result in additionalinvestment costs for the steel producer, while retainingall advantages on the power utility side.

In the opinion of some people an SVC is a neces-sary evil that should be avoided whenever possible.Some “strategies” on how to “negotiate away” theSVC equipment have even been presented in papers.

Through a few simple examples we shall try toprove the opposite, namely that the improvement of

power quality in the supply system, and in particularthe installation of a properly designed SVC, is anadvantageous investment for steel manufacturers andmay bring substantial gains from the production pointof view.

We also wish to stress how important it is whendesigning steel plant power supply that the installationof an SVC should not be considered at the final stagewhen all other equipment has already been sized andordered but instead should be a part of anoptimization plan for a complete power supply systemfor a steel plant. This is valid both for brand-newgreen-field plants and for revamping and extensions ofexisting plants.

2. THE STEEL PLANT AS AN ELECTRIC LOAD

For a power distributor, a modern steel plantrepresents a somewhat dubious customer. On the onehand, the plant may be the biggest paying consumer inthe distribution system; on the other hand, the sameconsumer through the nature of his load disturbspower quality for the other consumers connected tothe network.

The short time varying load, as occurs in rollingmills, and largely represented by electric arc furnaces(EAFs) with their almost instantaneous fluctuations inboth active and reactive power, are the main sourcesof disturbance.

The interconnection point in the grid between thepower utility and the steel plant, generally designatedPCC (Point of Common Coupling), then becomes aterminal of high importance. This is particularly truenow that several countries have begun to deregulatetheir power transmission systems. Concepts such as“power quality” have been introduced in order to setrules and regulations for networks with an increasingamount of voltage pollution.

Fig. 1. ”Load” at a poor power factor of cos ϕ ≈ 0.6.

u

ui

i

X ohmR ohm

(R/X = 1.5)


2

For an inductive feeding network,which is the most common, it ismathematically simple to show that,setting the plant reactive power at a lowvalue with small fluctuation in time, thenetwork voltage quality will increaseconsiderably even if the fluctuation inactive power still exist. Fig. 1 shows anexample of an inductive load at a poorpower factor.

Fig. 2 illustrates typical features ofpoor voltage quality.

The specific short time undervoltageand overvoltage variations often de-scribed in literature as voltage sags andvoltage swells may have hazardous im-pacts on sensitive equipment connected tothe network, such as computers and X-ray units.

In addition, the varying loads also create networkdisturbances in the form of phase unbalance andvoltage flicker; nowadays a very frequently discussedsubject.

Fig. 3.

The heavy voltage fluctuation caused by thefurnace low frequency load current content varies withthe current as a typical amplitude-modulated curveshape as seen in Fig. 3.

The state-of-the-art device able to control and keepthe plant reactive power consumption within

Fig. 2.

acceptable limits is the Static Var Compensator(SVC). The typical SVC comprises a set of Fixedpower factor correction Capacitors (FCs), divided andtuned to form the harmonic filters. A ThyristorControlled Reactor (TCR) with a fast acting regulatorensures the SVC’s capability to change its reactivepower output, following the plant reactive loadclosely.

3. DESIGN OF POWER SUPPLY SYSTEM

The three basic cases can be identified here withdifferent preconditions.

• A. Revamping and installation of more powerful furnaces in an old plant

• B. Extending of an existing mill withelectric furnaces

• C. A brand-new green-field project

The examples relevant to these three cases aredescribed below.

The examples comprise application of both ACand DC types of EAFs. A point of interest is thevariation in the power quality stipulations caused bydiverse national or even local regional standards.

−V

0

V

UNDERVOLTAGE

TIME

−V

0

V

OVERVOLTAGE

TIME

−V

0

V

HARMONICS

TIME

−V

0

V

TRANSIENTS

TIME

0 1 2 3 4 510

15

20

25

30

TIME s

kVab

s

BUS VOLTAGE

0 1 2 3 4 5

−2

0

2

TIME s

kA

EAF CURRENT


3

Example A

Our first example illustrates the very common caseof upgrading of existing plants. This actual plant islocated in Singapore and the intention was to install,to start with, only one new DC EAF, but after somemonths another identical furnace was installed. Oldsites for AC furnaces had to be used. Two of theexisting AC EAFs would remain in operation. Thefour furnaces could be operated in all possiblecombinations, but with only two simultaneouslyrunning at full load.

The stringent power quality regulations set by theSingapore utility (PUB) limited the use of newfurnaces and made necessary the installation ofcompensation equipment. In other words, the impactof new, more powerful furnaces on the feedingnetwork had to be limited. The necessity of use ofexisting MV switchgear, however, limited thepossibility of installing a sufficient amount of reactivepower as the current handling capacity of existingcircuit breakers on the feeders had to be considered.The new SVC replaced the existing small sizecompensator and had to fit into its location. Thethyristor valve, control equipment and newsubdistribution switchgear were housed in a newcompact building. A simplified single line diagram ofthe plant is shown in Fig. 4 above.

The main factors to be considered were as follows:• Power factor limitation

(no over-compensation accepted)• THD in voltage

(THD = Total Harmonic Distortion)• Voltage unbalance• Voltage fluctuation

(flicker defined according to ERA)

ABB’s technical studies indicated that the demandscould be met by an SVC rated -2/56 Mvar and withthe capacitive power divided in a set of filters as perFig. 5. Field measurements performed confirmed thatthe plant meets the stipulated demands with theinstalled SVC.

EAFEAF

PUB 66 kV FEEDING LINES

PCC 66 kV S/S

ROLLING MILL ANDPLANT AUX. SUPPLY

TIE CB

NO

SVC

22 kV

Ssc 2059-2827 MVA

60 MVA12.5 %

36 MVAPF 0.77

36 MVAPF 0.77

12.5 %60 MVA12.5 %

(OPEN RING BUS)

EAF53 MVAPF 0.75

12-PULSE

60 MVA

EAF53 MVAPF 0.75

TCR

FC 3rd FC 4th FC 7th30 Mvar 16 Mvar 10 Mvar

nom. 58 Mvar 56 Mvar

SVC

22 kV

CB BUS

Fig. 5.

Fig. 4.


4

Example B

Our second example represents a steel plant withexisting two rolling mills, to be expanded on themetallurgical side with two new electrical furnaces: anAC EAF and a LF. The plant is located in a heavyindustrial district in southern Taiwan and the systemexpanded in this manner had to meet the severeregulations imposed by the Taiwan Power Company(TPC).

The simplified plant single line diagram is shownin Fig. 6 above.

The main TPC demands here covered:• Power factor

(no over-compensation accepted)• THD in voltage• Harmonic current distortion• Voltage fluctuation

(flicker measured according to V10)

The critical design criteria in this plant wereprimarily the limitation in the voltage fluctuation, butalso the fulfilment of the rather rigid limitation in theharmonic current content which necessitated extendedcomputer studies to define the best possible filtercombination. The EAF is required to remain inoperation under emergency conditions with one

transformer only and a closed tie-breaker, without anyrisk of resonances between the SVC and the mill’spower factor correction capacitors.

In this plant the SVC thyristor valve and thecontrol equipment are located in a small container-likebuilding within the compensator’s 22 kV switchyard.An SVC rated -5/70 Mvar with filter power divided asin Fig. 7 was finally implemented.

Fig. 7.

EAFLF

TPC 161 kV

PCC 161 kV S/S

ROLLING MILL ANDPLANT AUX. SUPPLY

TIE CB

NO

SVC

(INCLUDING POWER

22 kV

Ssc 5828-11098 MVA

70 MVA12.5 %

70 MVA12.5 %

62.4 MVAPF 0.75

9.2 MVAPF 0.8

FACTOR CORRECTION)

TCR

FC 2nd FC 3rd FC 4th

SVC

22 kV FURNACE BUS

20 Mvar 30 Mvar 20 Mvar

nom. 75 Mvar 70 Mvar

CB

Fig. 6.


5

Example C

Our third steel plant is a typical example of agreen- field project. This plant is located in the mid-west of the USA and the installation will represent adrastic change in the power transmission network in awide area surrounding the mill. Even if the strongestpossible connection to the interstate 230 kV system ismade, the fault level will still be very low.

To control and keep the “network pollution” downto acceptable limits the utility imposes quite strictdemands on plant operators, such as a flicker levelbelow IEEE 519 “border line of irritation”, also to beapplied at the lowest system fault level in use.

The main demands here again covered:• Power factor limitation

(no over-compensation accepted)• THD in voltage• Harmonic current distortion• Voltage fluctuation

(flicker according to flat top measurements)

To fulfil the demand of having a plant powerfactor close to unity it is also necessary to compensatethe reactive power consumption in the

rolling mills and the auxiliary systems. Therefore, afixed capacitor bank was added to the rolling mill’sbus. Step down transformers with high impedancewere used as well as smoothing reactors on the DCfurnace high current side as additional means togetherwith the SVC. The purpose was to prevent, as far aspossible, the disturbance from being fed into the utilitytransmission system. With limitations in both thevoltage and current distortion, the proper design andselection of the tuned capacitor banks are of utmostimportance and here again detailed computercalculations had to be performed.

The SVC was finally rated 95 Mvar, with the filterbanks divided as in Fig. 9.

In this plant the SVC outdoor switchyard isintegrated with the furnace substation and the indoorvalve and control equipment housed in the mainsubstation building.

LF

WAPA 230 kV

PCC 230 kV S/S

ROLLING MILLS ANDPLANT AUX. SUPPLY

SVC

34.5 kV

Ssc 1800 - 2300 MVA

75 MVA12.5 %/45 MVA

75 MVA12.5 %/45 MVA

20 MVAPF 0.75

EAF

75 MVAPF 0.73

12-PULSE

41 MVA

13.8 kV

41 MVA10.8 %/25 MVA 10.8 %/25 MVA

POWER FACTORCORRECTION(CAP. BANK)

Fig. 8.


6

Looking back to our three examples one canobserve details common to these plants and also to themajority of the steel plants around the world. Thepower from the utility is supplied by a transmissionnetwork at the highest possible regional fault andvoltage level. The furnaces, in turn, are located at amedium voltage level insulated from the transmissiongrid by one or more step down transformers. Themedium voltage is selected to suit the size of thefurnace and the market availability of equipment suchas the furnace breakers and tap changers. Usuallysome form of back-up power supply is arranged. Torely on only one step down transformer may be risky;a simple fault can result in a long plant standstill. Inplants with the same internal distribution voltage levelin the furnaces and mill sections, it can be a suitableprecaution to insert an emergency tie switch betweenthe buses. In an emergency parallel operation fed bythe mill supply, the mill suffers from the voltagefluctuations but steel production will still be possible.This type of rarely needed emergency operation isoften more cost-saving than the insertion of an extrastep down transformer, taking into consideration bothinstallation costs and the total active power losses.

Fig. 9.

The figure illustrates a typical SVC of today,built up with a TCR as the inductive elementand a set of capacitors forming differentfilters (three BP types and one C type).

However, one can also observe some distinctivedifferences between the examples. The feedingnetwork fault levels vary as well as the connectedloads. To comply with the differing power qualitystipulations all the SVCs must be tailor-made to fiteach plant. The size of installed reactive power willvary due to the load, and the division of the capacitivebanks into tuned filters will also be different. Eachplant will be a unique one of a kind if the very bestoperation behaviour is to be achieved, and extensivecomputer studies are needed.

As is obvious from our first example, the SVCdesign must involve several different considerations,such as future extension plans or more or lesstemporary changes in the network parameters. It isquite often the case that the system fault level inreality is outside (and normally below) the givenlimits. Here we must seize the opportunity to point outto the presumptive SVC buyer the importance of notonly estimating the SVC rating from the net powerinstalled but also taking into consideration theavailable capacitors’ gross power and the filterconfiguration. An SVC for industrial purposes mustbe built as a robust tool and with a good margin towithstand unforeseen stresses caused by ongoingchanges in plant performance requirements followingincreased production demands, etc.

SVC

TCR

FC 2nd FC 3rd FC 5th FC 11th

34.5 kV FURNACE BUS

20 Mvar 20 Mvar 25 Mvar 30 Mvar

nom. 95 Mvar

CB

95 Mvar

DISCONNECTOR


7

From the point of view of the electrical engineer,the ideal case is our third example as it gives thepossibility of optimum design of the system, with onlytwo limitations:

- The expected production requirements.- The actual PCC system fault level, together with

supplied voltage level and energy rates.

In the case of green-field projects even this can beoptimized as the geographical site for investment maybe subject to free choice. Once the production aimsare established and the size in tons of the furnacedefined, the electric supply system can be designed. Avery important element is here the step downtransformer whose rating and impedance will directlyaffect the production as well as the flicker disturbancelevel in the network. In the case of an AC furnace, ahigh impedance system in form of a series reactor,resulting in a long arc, is recommended. In the case ofa DC type of furnace, more powerful smoothingreactors should be considered.

In the second example, the electrical engineer hasless freedom of choice. But there is still the possibilityof optimizing the furnace circuit with the transformerand series reactor in a total approach together with thepower factor corrective equipment.

4. POWER QUALITY ASPECTS AND STEELPLANT PERFORMANCE

Section 3 above, with the three examples of steelplants and associated comments on electrical design,may have given the impression that the main purposeof power factor correction equipment is to fulfil rigiddemands forced upon the plant operator by the utility.In the following we should like to show the positiveinfluence and beneficial production features the plantoperator will obtain by increasing Power Quality.

4.1 General comments on utility demands

• • Power factor limitationThis demand may seem to be easy to fulfil, but

there are complications in compensation of fluctuatingloads, which means that careful calculations have tobe made.

The first important parameter is if over-compensation into the utility network is accepted ornot. According to our experience of dealing withseveral heavy steel mills, such a situation will seldombe accepted as a normal condition.

Power factor compensation demand withmechanically (CBs, etc.) switched capacitor bankswill force frequent switching operations, with bothheavy wear and network disturbance as a result.

Time s

Mvar

Typical EAF reactive power swing

QMEAN

QMAX

QDYN

Fig. 10.

Furnace mean active powerPmean Srated= × cosϕ

Furnace mean reactive powerQ Srated sinmean= × ϕ

Let us again use our second example. Tocompensate only this rather small EAF plant involvesa mean reactive power consumption up to a demandof power factor 0.95, a low demand today, seen over along time interval, say 24 hours. Simple calculationindicates that a range of 32 Mvar would be sufficientto compensate the furnaces. However, some moreMvar must be added to compensate the loss in the stepdown transformer up to the PCC as well, in all 36 to38 Mvar. With shorter integrating time intervals,typically 10 to 20 minutes, in some plants down to 3seconds, this fixed filter compensation will be toosmall to keep a mean power factor at the requiredvalue. In particular, one must take into account thedelays caused by the switching procedure (dischargetimes, etc.).


8

It is easy to understand that an insertion of thisrange of uncontrolled capacitor banks at the furnace'sbus would cause voltage variations far beyondacceptable standards. A voltage fluctuation in therange of 11 % would be seen at the furnace bus,starting with an overvoltage close to 8 % in no loadcondition and followed by a correspondingundervoltage of more than 3 % at full furnacesoperation. These figures are, of course, not acceptablefor the plant operator as they would result in muchless efficient furnace operation.

Installation of an SVC with continuous control ofthe reactive power will not only solve the problem ofvoltage variations originating from the fixed capacitoroperation but also give the possibility to compensatethe dynamically fluctuating power in the furnaces.

By compensating the “static” mean reactive powerand the power needed by the dynamic fluctuation, aswell as the losses in the step down transformer, thereis a need for an SVC rated 70 Mvar as in ourexample. This results in increasing the mean powerfactor at the PCC to a level above 0.98. In practice,the power factor will be close to unity most of thetime.

⇒ The nearby rolling mill, connected in parallel, willsuffer from the voltage fluctuations caused by thefurnace operation but will also, if not properlycompensated, have a negative influence on thenetwork. Together with the impact of the furnaces,the total network distortions may easily exceed thestipulated demands.

⇒ It may not be possible to operate the LF in theoptimum manner if the furnace bus voltagefluctuates rapidly and drastically due to EAFoperation. The bus voltage must be stabilized.

• • Reactive power consumptionThe above presented principle, i.e., calculating not

only the furnace mean power consumption but alsotaking into consideration the fluctuating dynamic partwhen designing compensating equipment is veryimportant. If the compensation power is not sufficient,the bus voltage will decrease drastically each time thislimitation in power output is reached. The voltagereduction will lower the melting capability and thusincrease the tap to tap time.

The fluctuation in the incoming feeding voltagemay lead to unstable arcing and most certainly to ahigh flicker level. Our examples include furnaces ofboth AC and DC type and without indicatingpreference for any of the two types, we as suppliers ofpower quality equipment must stress the similarity inreactive load fluctuation.

Fig. 11 above has been published in differentpapers as an impressive example proving that an ACtype furnace might need twice the value of dynamicreactive power compared with a DC furnace operatedwith same active power. However, this is generally anunfair comparison. Field measurements indicate, onthe contrary, that the fluctuation in reactive power(expressed as the value during 95 % of the time) isalmost in the same range. This is illustrated in thesimplified diagram in Fig. 12.

Fig. 12.

Some papers present various complicated theoriesas to how to keep the reactive power variation in DCfurnaces down, but so far practical experiences of

P

MW

QMvar

ACTIVE POWER

REACTIVE POWER

ACEAF

DCEAF

PHASEANGLE(pf)

SETPOINT

MW

MvarQMEAN

P MEAN

QDYN

MW

Mvar

QDYN

QMEAN

P MEAN

AC EAF

DC EAF

= FLUCTUATINGLOAD AREA

Fig. 11.


9

these are limited. In any case an action that may causelimitation in the control of the melting operation ormay generate additional network disturbance (forexample phase unbalance or increased harmonicgeneration) should be avoided.

• • Voltage fluctuation and flickerA further limitation common in our three examples

is a result of demands regarding voltage fluctuation.In the described cases only voltage fluctuation in theform of flicker is limited by the utility. Stipulationsimposed to control the slower voltage variations atPCC are also common. The latter are unfortunatelyoften mixed up with the more “aggressive” continuousflicker voltage modulation, as seen for instance in Fig.3.

Without going into the rare occurrence of furnaceshort-circuits, often even the normal furnace operationwith instant power swings from no-load to full-loadcondition will utilize most of the permitted voltagevariation range without the stabilizing effect of SVCon the furnace bus.

In general it is better to use terms such as voltagequality at a bus instead of talking about some shortcircuit voltage depression. The bus voltage variationshould not deviate by more than the stipulated level(say ±2 %) expressed in an integrated r.m.s. valueduring a given time interval (typically 3 seconds or 10minutes). Note also that the greatest impact on thenetwork from the furnace will come from thetransients caused by inrush current upon switching in.

Even if the voltage flicker, by definition, willmainly disturb the human perception of light sensitiveelectronic equipment will also be influenced by thedistortion. As an example, the steel plant’s internalcomputer system may temporarily be set out of orderby the influence of furnace flicker. This risk is not tobe disregarded.

A properly designed SVC will, to a great extent,attenuate the low frequency flicker down to a moreacceptable level, as per the example in Fig. 13.

Fig. 13.

Looking at Fig. 13 one may feel that the SVC isnot needed as no signals are above the borderline;however, it is important to understand that it is thetotal integrated amount of the modulated lowfrequency signals into the carrier fundamental (50 or60 Hz) which will create the flicker perception. Inpractice, it is very seldom that the flicker sensation isgenerated by only one single frequency and to achievea necessary margin an SVC will also be needed insuch a case as in Fig. 13. One must further have inmind that connecting a new melt shop to a “clean bus”is a rare case; the network is typically “polluted” closeto the limit set by the utility already before thefurnaces start to operate, for instance by the plant’sown rolling mill or by other heavy industries in theneighbourhood. Due to this fact, the last object to beswitched into the network must generate a much lowerdisturbance than the general limit in order to beacceptable by the utility.

As in our three examples, there are severaldifferent standards regarding flicker and flickermeasurements in use and valid in different countries.The different filament lamp types used (100 V or 240V, etc.) also make a comparison between standardscomplicated. As a rule of a thumb, in order to providea reference of typical flicker values in EAFinstallations, the following simple calculation may beused.

0 5 10 15 20 25 300

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1PCC VOLTAGE % dV/V

FREQUENCY Hz

IEC 555−3

WITH FC :

WITH SVC: IEC 555-3gives themaximumpermissiblepercentagevoltagechanges acc.to IEC


10

The standard procedure using a flicker meter ofUIE/IEC-type is to compare the furnace short circuitlevel (or short circuit impedance) seen from PCC withthe short circuit level at the PCC using the predictionformula (The Pst is a 10-minute integrated value):

Pst KstSsceafSscn

99% = ×

with

Kst = characteristic emission coefficient forPst

Ssceaf = short circuit level of the arc furnace atthe PCC

Sscn = short circuit level of the network at thePCC

The Kst value in the formula ranges from 45 to 85,depending on how “noisy” or unstable the furnaceoperation is expected to be.

Here it should be noted that, according to CIGRE,a normal accepted Pst95% reference value in aMV/HV power system is 1.0 and the correlationbetween Pst99% and Pst95% is 1.25 and as aconclusion, with furnaces used normally and a ratiobelow 50 between the two fault levels, unacceptabledisturbance can be expected without use of an SVC.

Fig. 14. Typical statistical registration of the flickerlevel and the use of the 95 % value.

4.2 Impact on steel production

If one follows the intention of decreasing thenetwork voltage fluctuation by insertion of an SVC,there will also be benefits to the steel plant operator.In the diagram in Fig. 15 the positive influence of analmost stiff network is clearly visible. In this example,the active power increase will be in the range of 15 %if an SVC is used. The voltage decrease from no-loadto the rated arc current will be approximately 5-6 %(which is less than in our second example), without

the SVC.Fig. 15.

An attempt to correct the furnace bus voltagequality by insertion of a tremendously oversized step-down transformer may help the furnace operation butwill in turn create unacceptably high flicker at PCCand thus this is not a practical solution.

• • Advantages of higher active power availabilityThe possibility of using more power in an existing

(or revamped) furnace with no need to make changesto the existing transformer and secondary system mayvary between steel mills for commercial reasons. Herewe wish to point out the use of power increase toshorten the tap-to tap time, as can be seen in Fig. 16.

0 20 40 60 80 100 1200

20

40

60

80

100

120

ELECTRODE CURRENT kA

MW

kV

POWER INCREASE WITH AN SVC

I nom

EAFPOWER

BUSVOLTAGE

OPERATION WITH SVC:

OPERATION WITHOUT SVC:

dP

0 0.5 1 1.5 2 2.5

10

20

30

40

50

60

70

80

90

100TIME %

FLICKER LEVEL Pst

WITH SVC WITHOUT SVC

=> Pst95 LEVEL


11

Fig. 16. A furnace load cycle without and with SVC.

The positive change in the melt down time cansimple be calculated based upon the old formula:

tmWliq

scr

ESrated ut

= × × × −

× ×η ϕ η60 10 3

cos

with

tm = Melt down time (min)

Wliq = Charge weight, liquid steel (tons)

ηscr = Yield factor, liquid/scrapE = Energy consumption (kWh/ton)

Srated =Furnace transformer rating power (MVA)

cos ϕ = Power factor

ηut = Power utilization factor (Pmean/Prated)

with typical figures using the example above one has

t m11350 9

465 60 10 3

59 0 8583 4= × × × −

×=

. .. min

t m21350 9

465 60 10 3

73 0 8567 4= × × × −

×=

. .. min

This represents a decrease of approximate 16 min!

If this steel plant with an expected tap-to-tap timeof 145 min prefers to use all the extra power in theproduction, at least one extra charge per day can bemade, which means an increased annual production ofaround 50 000 tons.

• • Electrode savingsDue to the decreased tap-to-tap time the electrode

consumption will also decrease. In this case thefurnace operates with the same electrode current,which means that a linear reduction of 12 to 15 % canbe expected, typically corresponding to 0.5 kg/ton.

• • Energy savingsWith shorter operating time and with a more stable

arc due to the stiffer network, the losses in thefurnace, as well as in the auxiliary systems (fans andpumps, etc.) will decrease and approx. 20 kWh/toncan be expected to be saved.

• • Refractory savingsWith the more efficient and stable arcing a

decrease in refractory wear will be expected. In thistype of installation a decrease of 0.8 to 1.0 kg/ton hasbeen reported.

4.3 Series reactor and long arc practice

Modern EAFs have a tendency to be more andmore powerful, with higher ratings in the furnacetransformers, and are of the so-called UHP type. Tooperate this type of furnace with low secondaryvoltages, according to previous practice would haveled to a drastic increase in the arc current, resulting inadditional electrode consumption and costs for a morerobust furnace secondary system e.g. involving armsand cables. A more suitable method of increasing thefurnace power is to increase the secondary voltage butkeep the current at the previous values. Due to thevery non-linear resistance characteristic in the arc,many of these UHP furnaces suffer from poor arcingdue to operation with too high a power factor, causingdifficulties in reigniting the arc after voltage zerocrossing. In addition to the losses in the furnaces, theunstable arcing also gives flicker and harmoniccurrent generation to the feeding network.

Installation of a furnace series reactor will nor-mally improve the situation for the furnace operatorby decreasing the cost per produced ton of steel andfor the utility by decreasing the voltage fluctuationthanks to the arc stabilizing effect of the reactor.

0 20 40 60 80 100 120 140 160 1800

10

20

30

40

50

60

70

80

90

100

TIME MINUTES

MW

EXPECTED TAP−TO−TAP TIME WITHOUT SVC

OP. WITH SVC


12

The Fig. 17 below illustrates a case with a furnacetransformer equipped with voltage taps hardlyoperable due to unstable arc. In this case, an increasein production was of minor interest due to otherlimitations in the network, but with the insertion ofseries reactor together with the stiffer network effectgiven by an SVC, a radical decrease in the electrodecurrent is possible.

Fig. 17.

The positive influence of the decreased electrodecurrent can be determined by dividing the main part ofthe electrode consumption into the two componentsbelow.

- side oxidation, mainly dependent on the furnacetap-to-tap time (as per section 4.2)

- tip consumption, mainly dependent on thesecond power of the electrode current

Typically, 30 to 70 % of the total electrodeconsumption can be referred to tip consumption andthus figures in the higher range represent the modernUHP furnace.

The tip consumption may be calculated as

Wtip ktip Irms2

0

T

dt= × ∫ktip= the fraction of the total consumption

In our case (Fig. 17), with almost identical powerand tap-to-tap time, the side oxidation should be equaland the difference in the total electrode consumptiondue to tip consumption only can be calculated by theformula:

Wtap WtapWtap

ktipItapItap

6 12

61

12

6

2−= × −

or with a ktip typical set to 0.5

Wtap WtapWtap

6 12

60 5 1

5365

20168

−= × −

=. .

This means a more than 16 % decrease in electrodeconsumption and in the actual example amounts toapproximately 3.5 kg/ton consumption, a decrease of0.5 kg/ton. The money savings resulting from thislower consumption are not insignificant. The lowerelectrode current will, in addition to the immediateadvantage of electrode saving, also decrease powerlosses in the secondary system, in this case byapproximately 2 MW or 15 kWh/ton).

4.4 DC EAFs and advantages of a stiffer network

Even if the example in section 4.1 shows an EAFof AC type similar results will occur with a furnace ofDC type.

Without an SVC, it is most likely that a higherrated furnace transformer must be installed, resultingin higher reactive power consumption. The lowerpower factor, in turn, imposes a need for more powerfactor correction equipment, followed by increasedover- voltage problems in no-load condition, etc.However, the capacitor power needed in the form offilters to correct the EAF harmonic current generation(in the furnace as well as in the rectifier), may alreadycreate such an overvoltage situation that theinstallation of an SVC becomes absolutely necessary.

It should further be noted, that in the case of DCfurnaces increasing the smoothing reactor to attenuateflicker is not as favourable as in the AC furnace casedue to the high increase in active power losses.

0 20 40 60 80 1000

20

40

60

80

100

Electrode current kA

MW

FUNDIA, Norsk Jernverk, Mo i Rana

cos fi * 100EAF xfmr rated power

Tap 12with reactor

Tap 6withoutreactor


13

As for the AC furnace, a decrease in the tap-to-taptime through a higher power input or a lower electrodecurrent made possible by a stiff network voltage willhave a positive economic effect.

5. ECONOMIC ASPECTS

In the simplified pay-off calculation in Table Iabove, we wish to demonstrate the typical economicbenefits with an SVC. The calculations are in prin-ciple, based upon one of the examples in section 4.2.Savings in refractory material are not included, andthese may be added as a further benefit.

6. SUMMARY

To summarize, in addition to lower power billswith more favourable rates, the installation of theSVC has a direct impact on several productionparameters such as:

- Increase in the available melting power whichleads to a shorter melt down time and thus higherproductivity.

- Decrease in specific electrode consumption due tothe shorter melt down time and further by themore stable arc.

- Decrease in specific energy consumption due tolower radiation losses. Further the lowerlosses in the involved auxiliary systems (fans andpumps) and in the power supply system (forinstance step down transformers) are significant.

REFERENCES

[1] L. Hultqvist and A. Wilk-Wilczynski“Voltage Criteria in Steel Mill Networks”Publ. ISS 53rd Electric Furnace Conference -Nov. 95

[2] S. Torseng "Shunt-connected reactors and capacitors controlled by thyristors"IEE PROC. Vol 128 Pt. C. No. 6. Nov. 1981

[3] SVC for voltage stabilization and harmonicsuppression in ladle furnace rolling millABB Pamphlet A02-0132 E

[4] Static Var Compensation of AC and DCfurnaces in joint operationABB Pamphlet A02-0143 E

Pay-off calculation without SVC with SVC

• • Production increase

Tap-to-tap time (min) 145 129Number of charges per day (at 135 ton)≈10 ≈11Production ton/year (325 days) 435 000 485 000Increase 50 000Profit with 40 USD/ton 2 000 000 USD

• • Electrode savings

Consumption savings 0.5 kg/tonΣ 242 500 kg

Profit at 3.5 USD/kg 849 000 USD• • Energy savings

Consumption savings 20 kWh/tonΣ 9 700 MWh

Profit at 0.03 USD/kWh 291 000 USDIn all 3 140 000 USDCompared with a typical SVC costturnkey, installation including civil works ≈4 000 000 USD

Pay-off time (4.0/3.1) ≈≈15 months

Table I

steel plant performance, power supply system design … · steel plant performance, power supply...

Documents