POWER SYSTEM DESIGN FOR HIGH-POWER ELECTRIC SMELTING AND MELTING FURNACES T. Ma, G.J. Bendzsak and M. Perkins Hatch Associates Ltd., 2800 Speakman Drive, Mississauga, Ontano, Canada L5K 2R 7
ABSTRACT High power electric smelting furnaces operate typically at power levels in the 30 to 60 MW range, and frequently, the furnace dynamic load swings can have a significant impact on the generation equipment, transient stability of the power system and the power quality to other interconnected loads. Power systems for these furnaces are designed with the objectives of increasing the average furnace power levels while meeting utility load restrictions, disturbance limits and equipment performance limitations. System design considerations include generation frequency swings, bus voltage fluctuations, harmonic filtering, furnace and power system controls. A systematic design approach consists of estimation of furnace load fluctuations, dynamic numerical simulations of furnace and power system equipment , followed by simulation and analysis of process controls. Reprinted from: The Proceedings of the International Symposium on Non-Ferrous Pyrometallurgy: Trace Metals, Furnace Practices and Energy Efficiency Edmonton, Alberta, Canada August 23-27, 23-27, 1992 31st Conference of Metallurgists of the Metallurgical Society of CIM
INTRODUCTION High power electric smelting furnaces have to satisfy operating restrictions dictated by the utility. Specialized power system compensation equipment such as tuned harmonic filters, primary reactors or controlled reactive power compensation may be required. Electrical separation of the furnace bus from other loads may be necessary. These increasingly stringent measures arise when the furnace load becomes a large portion of the generation capacity on the line. In the situation where furnace load is supplied from captive (dedicated) generation, the furnace load can approach 80 % of the generation capacity. These circumstances underline the importance of designing and operating the power system and furnace load as a Combined System. Furnace load instabilities originating from both arc behaviour and from furnace operating upsets are discussed from process and electrical perspectives. The interactions of these disturbances with furnace power supply equipment and generation are presented. This paper presents a method of analyzing the operation of large electric furnace loads on a power system. The paper characterizes process specific furnace loads according to power conversion mechanisms and heat transfer. Simulation techniques for evaluating performance of the furnace
arc, power generation, control systems, as well as corrective measures are described. Application of simulation results to power system design is discussed.
CHARACTERIZATION OF SMELTING FURNACE LOADS Most electric smelting furnaces contain a molten bath of conductive metal or matte on the hearth, underlying a relatively resistive slag layer onto which unmelted charge mix is added. There are four distinctive types of electric smelting operations, characterized primarily by the mechanisms of power conversion to heat and transfer of the liberated heat to the furnace charge: ➤
Immersed Electrode
➤
Open Arc
➤
Shielded Arc
➤
Submerged Arc
In practice, the four methods are distinguished by the operational positions of the electrode tips relative to the molten bath and the presence and depth of unmelted charge cover surrounding the electrodes. These features, along with the associated secondary circuit electrical parameters typical for a 40-50 MW furnace, are summarized for each furnace type in Table I.
Table I – Types of Smelting Furnaces
Electrode Tip Position Relative to Slag Bath Charge Cover at Electrodes
Furnace Resistance
Immersed Electrode
Open Arc
Shielded Arc
Figure 2
Figure 3
Figure 4
Immersed in bath
Above Bath
Above Bath
Above Bath
Bath Covered or Open Around Electrodes
Open Around Electrode Tips
Charge Cover Around/Over Electrode Tips
Deep Charge Cover Over Electrode Tips
Slag Bath Resistance Only
Arc + Slag resistance
Arc + Slag Resistance
Mainly Arc
Resistance Distribution
R Arc R Bath
Power Liberation Heat Transfer to Charge
≅
R Arc
0
R Bath
2
>3
2
I R Bath
I ( R Bath + R Arc )
Convection by Hot Circulating Slag Currents
R Arc R Bath
Submerged Arc
R Arc
= 3 to 5
2
R Bath
>5
2
I ( R Bath + R Arc )
I ( R Bath + R Arc )
Arc Radiation & Convection via Hot Furnace Freeboard
Arc Radiation & Convection Direct to Charge Banks
Arc Radiation & Convection Direct to Charge / Coke Bed
Typical Electrical Parameters (sec’dry) Resistance per Electrode (mohms)
5-10
5-15
10-40
1-2
Reactance per Electrode (mohms)
2-2.5
3-5
3-5
~1-2
Sec. Volts/Electrode
100-300V
100-500V
500-700V
~75V
Sec. Current kAmps
20-50
20-50
10-30
~100
0.97-0.98
0.8-0.9
0.95
~0.7
Small Swings
Large Swings
Large Swings
Moderate Swings
About Avg. Levels
Infrequent
Frequent
Frequent
Moderate
Electrode Speed
Slow
Very Fast
Fast
Very Slow
<50 cm/min
>150cm/min
100-150 cm/min
< 50 cm/min
Power Factor Current /Power Swings
Regulation Required
Immersed Electrode -Operation To assist an understanding of the specific electrical features of each furnace type, it is useful to examine the behaviour of electrode resistance with electrode tip position, as illustrated in Figure 1. In the IMMERSED ELECTRODE mode, the electrode tips are immersed into the slag bath, which forms the only significant resistance in the circuit, and power is liberated solely by Joule heating: 2
P E = I R Bath
The liberated energy superheats the slag locally establishing circulating flows that distribute the heat to the charge banks (Figure 2). Electrical conversion to heat energy is very stable. Bath
resistance fluctuations and the associated power swings are very small. Low speed electrode regulation is sufficient for power set point regulation. The power factor is high, typically above 0.95. The merits of immersed electrode smelting are found in these important features. This smelting mode suffers from limited hearth power densities. Large furnace sizes are required to control erosion of the sidewall refractories by the hot circulating slag currents.
Figure 1 – Load Resistance versus Electrode Tip Position
Figure 2 – Immersed Electrode Operation and Figure 3 – Open Arc Operation
The slag bath resistance is dependent primarily on the slag resistivity, which is dependent on slag composition and temperature, electrode size and the immersed depth of the tips into the slag. For a given operation, the resistance increases as the electrode tip is moved upward from the slagmetal interface to the slag surface, as shown in Figure 1. This simple resistance to immersion depth relationship provides the basis for regulating the furnace power. The transformer secondary voltage tap is set at the desired value and the electrodes are raised or lowered to maintain the set point resistance or impedance. Control of load resistance essentially controls furnace power through the relationship: P E =
(
V
x
PF
R Bath
where P E V PF R Bath
2
)
= Electrode Power = Electrode Voltage to Bath = Power Factor = Bath resistance per Electrode
Arcing Operation Referring to Figure 1, it is noted that the resistance begins to rise more rapidly as the electrode tips approach the slag surface, mainly due to the reduced contact area of the pointed tips with the slag bath. Micro-arcing between the electrode and the slag commences before the electrode tip is free from the slag surface. This condition is commonly referred to as "Brush Arcing". Steady operation in this zone is generally undesirable since it loses the inherent stability of the IMMERSED mode, does not provide the desired resistance boost needed for high voltage levels and suffers from the instabilities derived from variable electrode tip geometry and the wave motion of the slag surface. As the electrodes are raised above the brush arc zone, the rate of resistance increase with electrode position assumes a much steeper slope, as shown in Figure 1, reflecting the onset of stable arcing conditions. Energy liberated in the high voltage arcs is transferred either directly to the charge in the case of Shielded-Arc operation (Figure 4) or a combination of direct transfer and freeboard re-radiation in the Open Arc mode (Figure 3). Since transfer of the arc energy to the charge does not rely on superheated slag recirculation, the main cause of sidewall refractory attack, much higher power densities can be used than are possible in Immersed Electrode smelting. Since Shielded-Arc smelting results in more moderate freeboard and off-gas conditions than are achieved in the Open-Arc mode, it should be adopted, wherever conditions such as gas evolution rate, charge mix size, porosity and sintering temperature of the banks permit. The arc resistance increases with arc length, but not in an invariant manner. In practice, the voltage drop across an arc increases virtually proportionally to the arc length over a fairly wide range of current levels. Typically, the arc voltage gradient is about 15 volts per cm of arc length, in both the open and shielded-arc modes. Since the rate of resistance change with electrode position is much steeper in the arcing mode than with immersed electrode - operation, the furnace electrical parameters are much more sensitive to arcing mode upsets in electrode position relative to the bath surface. This paper excludes discussion of low-resistance submerged arc smelting operations (ferroalloy, pig iron, phosphorous) where the electrodes are deeply buried in a conductive charge mix, with micro-arcing from the tips to a floating coke bed.
Figure 4 – Shielded Arc Operation
Figure 5 – Ideal Arc Voltage and Current Waveforms
Figure 7 – Furnace Power Supply Equipment 6 Electrode Furnace
Figure 9 – Frequency Versus time Captive Generation 44MW Furnace Tip
Arc Instabilities A major electrical instability introduced by arcing arises from the requirement to re-ignite the arc each half cycle, e.g. 120 times per second for 60 Hz power. If the arc fails to ignite, current flow is interrupted in the electrode, causing a 33 % to 50 % reduction in furnace power and excessive levels of power unbalance and negative sequence current. Re-ignition difficulties and the corresponding electrical instabilities are more severe in the open arc mode since hot plasma is swept from under the tips into the freeboard, thereby leaving a much colder, less conductive environment for arc re-ignition. Arc re-ignition produces reactive power (MVAR) swings on the power system which cause supply voltage fluctuations. These disturbances, referred to as voltage flicker [2], range in frequency from 1 to 20 hz, resulting in objectionable light flicker. By contrast, in the shielded arc mode, the environment in the arc crater under the electrode is contained and protected from the furnace atmosphere. Re-ignition is much easier and hence power instabilities are considerably reduced. This benefit of shielded arcing is evident from the reduced incidence of resistance and power fluctuations seen in Figure 4 compared to the open arcing data shown in Figure 3. The improved stability of shielded arcs allows the use of higher voltage levels.
The ignition voltage requirement of the arc introduces a phase delay angle between the arc current and supply voltage, denoted by e in Figure 5. This phase shift is referred to as an equivalent flare reactance" and is manifested as a reduction in furnace power factor. The arc reactance follows the relationship: X arc
= K 1 - Rarc
where K 1 is empirically measured and is process specific. The non-sinusoidal arc current of Figure 5 contains higher order frequencies, known as harmonic components. Both utility grids and captive generation are sensitive to current harmonics for the following reasons: ➤ ➤
Distortion of the supply voltage waveform to other customers. Amplification of supply voltage distortions due to resonance at the power factor correction capacitors of other customers [1].
➤
Interference with communication circuits and sensitive electronic equipment.
➤
Derating of power system equipment due to increased harmonic current losses.
Operation Load Disturbances In addition to the power system disturbances resulting from arcing, instabilities also result from normal operating upsets themselves such as power ramping and unbalances. Start up and shut down power ramp rates are normally governed by the rate at which the generation can be regulated to meet the load. Furnace unbalanced operation due to electrode breakage and rebaking, or tapping temperature adjustments is governed by the power supply and motive load negative sequence capability. Lastly, furnace full load trips result in an immediate step mismatch between load and generation. This step mismatch results in frequency oscillations as the generator controls adjust to the loss of load.
THE FURNACE POWER SUPPLY Smelting Furnace Electrode Configurations The majority of large electric furnaces are either three electrode round furnaces or six-electrode-inline rectangular furnaces. Furnace operation and dynamic control is influenced by the number of electrodes. A six electrode furnace has an additional degree of flexibility by virtue of individual furnace transformer tap changer control as well as individual regulation on each of the six electrode columns. This provides for a wide range of power control flexibility in furnace operations. The three electrode furnace differs in that the three furnace transformer secondary windings are electrically connected to each other by secondary bus connections made either in the transformer vault or near the electrode clamps. This close electrical phase coupling of the three electrode furnace causes a current fluctuation in one electrode to be reflected into the other two electrodes. Following the loss of arc conditions in a six electrode furnace, the total power is reduced by 33%. For the same conditions, the total power reduction of a three electrode furnace is 50%. A six electrode furnace therefore presents reduced MW fluctuations, and power unbalance to the power supply system during loss of arc.
The Furnace Transformer and Power Supply The operating ranges of furnace electrical parameters are established by the transformer specifications. The transformer voltage tap range, currents and furnace load range are graphically displayed using a Power/Voltage/Current (PVI) diagram. A typical PVI diagram for a 3 electrode furnace is shown in Figure 6. An operating point of 50 MW, 32 kA, 1042 volts, 18 mohm is indicated for a 72 MVA furnace transformer. This diagram also displays both the immersed electrode operating region as well as high voltage arcing operation. A typical power supply arrangement for a 6 electrode furnace powered by 3 single phase furnace transformers is shown in Figure 7. The key equipment items with power control capability are the furnace transformer tap changer, and the electrode regulation. The transformer voltage tap changer enables a range of discrete secondary voltage levels as shown in the PVI diagram. The electrode regulator controls the load resistance, and hence the power level, by raising or lowering the electrodes.
GENERATION, TRANSMISSION AND DISTRIBUTION OF FURNACE LOAD Generation The electrical transmission system connecting generation to the furnace load is shown in Figure 8. Furnace load can be supplied from either Captive Generation or a Utility Grid. A utility grid consists of multiple generation sites, each site consisting of two or more synchronous generators. A summary of high power electric furnace loads supplied from both Utility grids and captive generation is shown in Table II. Table II – Power Supply to Furnace Load Load Type
Project
Fce
Generation
MW
Type
Captive Open Arc
Power System Iscott
2x45
Gas Steam
Open Arc
Steel Plant
1x17
Diesel
3x88 +100 MW Grid 4x12
Shielded Arc Shielded Arc UTILITY
Falcondo PT Inco GRID Cerro Matoso Ipsco
1x80 3x55
Steam Hydro
3x70 3x60
1x48 2x40
Grid Grid
Sysco Falconbridge
1x60 2x21
Cyprus Impala
Immersed electrode Immersed electrode Immersed electrode
MW
PF Correction
Electrode Resistance
MVAR
mohm
Fault Level Fur. Ty. Pri. MVA
Fault Level Utility Supply MVA
SVC 2x65 SWC30
5
600
1200
SVC+45/10 None None
5
157
240
660 700
800 950
600 334
1340
Grid Grid
None Synchrono us Condenser s 120 MVAR SVC 1x40 None
135/2 15 45 5 5
5 35/2
900 300
2100 1500
1x40
Grid
None
12/2
370
1180
1x30
Grid
None
15/2
700
2000
Figure 6 – Power/Voltage/Current (PVI) Curves 3 Electrode Furnace
Figure 8 – Power System Block Diagram The traditional approach has been to obtain smelter power from the closest utility grid provided it has the required capacity and acceptable reliability. In new greenfield projects, an adequate utility supply is often not available. Consequently, installation of a local generation plant is required. The choice of generation type depends mainly on the relative supply costs of competing energy sources. The potential for using waste heat from the smelter to augment the energy supply as well as waste heat from the generation plant in the smelter should also be examined.
A brief discussion of the main types of captive generation follows.
Hydroelectric Generation The stored water energy is directed through penstocks to turbines which supply mechanical energy to the synchronous generators. The turbine wicket gates control the mechanical energy to the turbine according to the speed control set by the governor. Fast furnace load sheds result in water flow diversion from the turbine to a bypass valve. Frequency swings of up to 15% are possible. Start Up/Shut Down power ramps are controlled by the turbine wicket gates. The wicket gate and bypass valve also provide protection from the furnace load dynamics. Hydroelectric generation operates at low speeds and high inertia, and is well suited for furnace load dynamics.
Diesel The generators are driven by diesel engines coupled to large flywheels, which provide additional inertia for dynamic load changes. Fast diesel throttle control and a specially designed fuel turbocharger provide a proper response for full load rejections. Diesel generation, operating at low speeds of 400 rpm is a solution for the supply of furnace power.
Steam Generation Steam is produced in a coal, oil or gas fired boiler, and is admitted to the turbine through a main stop valve and downstream throttle valves, freely expanding through the turbine and exhausting as low pressure steam to a condenser. In the event of a furnace load shed, over speed protection is furnished by the main stop valve, limiting over speed to 6%. Load regulation is provided by the steam control valves and the speed regulator. Steam turbines operate at speeds of 3000 rpm and are mechanically more sensitive to furnace dynamic load changes.
Gas Turbines Gas powered turbines have been used for the supply of open arc furnace load. In this instance, the system frequency excursions were limited to 0.4% or 0.24 hertz, requiring a thyristor-switched ballast resistor (static WATT control-SWC) for power system stabilization [3].
Generation Controls The generating plant supplies two power components to the furnace; the MW or real power component that is used for smelting and the MVAR or inductive component, which is the stored energy in the magnetic fields of the power transformers and secondary bus. The MW and MVAR furnace power requirements are both met by the synchronous generators. For a steady furnace operation, balances for both the MW and MVAR loads exist between the generation and the furnaces. A furnace load disturbance, however, upsets the balances. A transient power system MW mismatch is first drawn from the rotating inertia of the generation machines, resulting in a speed change, and corresponding frequency swing. Generation speed governor controls then react to restore system frequency to the set point value by regulating the prime mover mechanical power. The speed control loop is typically slow, in the order of 5 to 30 seconds. Significant frequency excursions may result, as seen in the case of hydroelectric captive generation. Where tighter frequency limitations must be met, additional MW compensation is required. Reactive power MVAR mismatches resulting from arc instabilities cause a system voltage change. The generation voltage regulator senses the change and adjusts the internal field of the generator
to re-establish set point voltage. In doing so, the MVAR output of the generation is matched to the corresponding new furnace MVAR requirements. This voltage regulation control loop has a response of 0.5 to I second and is generally too slow for open arc flicker control. In this situation, an additional source of controlled MVAR compensation may be required. Static VAR compensation (SVC), consisting of fixed harmonic filters and a thyristor-controlled reactor (TCR) may be appropriate. Other solutions involve raising the utility fault level with additional generation and transmission. The fault level is a measure of the short circuit current at the point of interest and is expressed as the short circuit volt-amperes supplied by the generating sources. Table II provides typical fault levels for furnace installations. In general, increasing fault level represents a "stiffer" power supply which reduces the impact of arc-generated voltage flicker and harmonic voltage distortion to other customers. System fault levels normally increase with increasing supply voltage. Consequently, many smelter operations are supplied at voltages of 69 kV and higher. The next section is devoted to the analysis and correction of furnace load disturbances on the power system. ANALYSIS, CORRECTION AND CONTROL OF POWER SYSTEM DISTURBANCES Proper selection and specification of the power equipment for a furnace load requires the detailed analysis of furnace load profiles operating on selected generation and transmission system configurations. The analysis involves the examination of furnace steady state and dynamic loads, solution of the furnace supply circuit equations, simulation of appropriate generator and regulator control systems, and simulation of corrective compensation equipment. Presently available software packages, running on engineering work stations, are sufficiently powerful for power system simulations [4]. Application of furnace resistance and arc models, equipment control systems and reactive power compensator schemes [5] can be incorporated within the software packages. A summary of furnace load disturbances, power supply standards and corrective measures is presented in Table III. Technical selection of corrective measures is assisted by power system studies. The results of representative case studies for furnace power supply analyses covering voltage flicker, frequency fluctuations, arc harmonic filtering and furnace unbalance are presented below.
Table III - Correction of Furnace Load Disturbances
Arc Based Distrubances Furnace MVAR swings cause excessive voltage flicker which disrupts other sensitive loads on the system.
Standard
Spec
Corrective Measures
UIE 2
0.28 % at 7 hz (sinusoidal)
➤ ➤ ➤ ➤ ➤
Furnace MW swings cause excessive frequency swings which damage motive load drives
Excessive Arc Harmonics Operations Based Disturbances Start up and shut down load can be delayed by scheduling of water or steam
Individual Utility Requirements
➤
Separate Motive Load Bus Electrode Regulation Control Generation Controls Dynamic MW Compensator
➤
Harmonic Filters
2-10 MW/Minute
➤
Computer Assisted Control
2% Voltage Unbalance. Derate Motors to 95% of Nameplate
➤
➤ ➤ ➤
IEEE Std. 519
Utility
Sustained unbalanced furnace load occurring due to electrode breakage or other sustained furnace operation with unbalanced furnace power.
Separate the furnace and motive bus Increase Power Supply Fault Level MVAR Compensation Equipment Series Buffer Reactors Restrict, Change Operations
➤
Increase unbalance ratings of rotating machines Automated control of furnace phase power set points.
CASE 1 VOLTAGE FLICKER Open arc furnace operation on a Utility grid was measured and calculated with and without operation of MVAR compensation by a thyristor-controlled reactor (TCR) and fixed capacitor scheme [6]. The furnace, power system, and arc parameters, denoted in Figure 7 are: Generation Fault Level at Incoming Bus Furnace Operating Power Power Factor Total Fixed Capacitors TCR Rating Equation (1) kl MEASURED DATA Flicker Voltage Magnitude at 7 hertz Flicker Voltage Magnitude at 20 hertz
GRID 2100 MVA 60 MW 0.9 110 MVAR 140 MVAR 0.20 WITH TCR 0.3 % - 0.4 % 0.4 % - 0.8 %
Initial GRID MW and MVAR power flows were calculated using a Gauss Seidel load flow pro-ram [7]. An empirical arc model was developed which includes the described arc instabilities,
measured arc reactance, and measured harmonic current source generation for open arcs. Numerical calculations were carried out using a commercial software package [4]. Simulation calculations verifying SVC correction of low frequency flicker and amplification of high frequency flicker are in progress.
CASE 2 FREQUENCY DYNAMICS Furnace load trips on a captive power system were causing excessive frequency fluctuations of + 15 %. Calculations were required to determine the magnitude and ramp rate of the MW compensation needed to reduce the frequency fluctuations to 6% for a 50 MW load shed. The furnace, power system, and arc parameters, denoted in Figure 7 are: Generation Fault Level at Incoming Bus Furnace Operating Power Power Factor Equation (1) kl
190 MW Captive 850 MVA 52 MW Shielded Arc 0.92 0.217
The calculated frequency swings for both uncompensated and compensated furnace load trip are shown in Figure 9. A calculated ramp rate of 10 MW/sec, immediately following a f urnace trip reduces the frequency swing to 6%. Process control and equipment modifications are presently in progress to implement the above strategy. This design parameter was obtained from numerical simulation of the generator electromechanical and control loop differential equations representing the machine equations, voltage regulator and speed governor control, a shielded arc model and electrode regulator equations. The simulation software was written using commercial software [8].
CASE 3 REVIEW OF HARMONIC FILTER DESIGN A performance review of harmonic filters for open arc load and thyristor-controlled reactor operation was carried out. The furnace, power system, and arc parameters, denoted in Figure 7 are: Generation Fault Level at Incoming Bus Furnace Operating Power Power Factor 2nd Harmonic Filter 3rd Harmonic Filter 5th Harmonic Filter TCR Rating Equation (1) k1
45 MW CAPTIVE 240 MVA 27 MW Open Arc 0.7 18 MVAR 12 MVAR 15 MVAR 55 MVAR 0.20
Harmonic filters are designed to provide a current sink for arc and TCR generated harmonics, thereby reducing the magnitude of the net harmonic injection into the utility grid. The filters also minimize parallel resonance of the power factor correction capacitors, and limit the harmonic voltage distortion of the utility bus [9].
The source harmonic currents and the calculated net harmonic injection into the generation are provided below. Harmonic Number
Arc & TCR Harmonic Generation I3 % of Fundamental 1.7 3.3 1.3 3.5
2 3 4 5
Residual Utility Harmonic Current Iu % of Fundamental 0.22 0.2 0.2 0.2
Attenuation Factor I3 /Iu 7 17 6 17
The filters reduced the total harmonic distortion factor [10] to 4.6% on the interconnected rolling mill bus, thereby meeting the maximum distortion limit of 5.0%. The calculations involve representing the power system with equivalent harmonic impedances and solving for harmonic filter currents and residual utility currents at each harmonic frequency.
CASE 4 UNBALANCE COORDINATION OF FURNACE AND MOTIVE LOAD Both immersed six electrode load and large synchronous motors are supplied from a common bus connected to a Utility grid. The maximum sustained furnace unbalance acceptable to the motive load operation was calculated. The calculations were used to set the negative sequence relay settings. The furnace, power system, and arc parameters, denoted in Figure 6 are: Generation Fault Level at Incoming Bus Furnace Operating Power Power Factor Motive Load
GRID 11 80 MVA 18 MW Immersed Electrode 0.97 3x5OOO HP
The furnace transformer connection was grounded WYE with an insulated transformer neutral cable. Maximum furnace power with 4 electrode operation (two out of service) is 12 MW, resulting in a bus voltage unbalance of 1.2%. Furnace unbalance limits are normally given' by either the negative sequence rating of the generation, or the NEMA bus voltage unbalance of 2% [10]. Analysis of these limits involve calculation of furnace load negative sequence current, by multiplying the furnace three phase currents, in vector format by a transformation matrix.
CONCLUSIONS The design of electrical power systems for smelters incorporating one or more high power electric furnaces must take into account: ➤
Furnace operating characteristics, whether immersed electrode, open or shielded arc.
➤
Furnace power load relative to generation capacity.
➤
The method and source of power generation.
Arc-based disturbances can be corrected enabling high power furnace operation on weak power supplies. Appropriate power system calculations are required for the design of the power supply system. Each furnace power supply application requires individual evaluation to establish the parameters at the interface between generation and furnace load.
REFERENCES 1.
J. Arrillaga, D. A. Bradley, P. S. Bodger, "Power System Harmonics", Wiley, New York, 1985, Chapter 4.
2.
R. Seebald, J. Buch, D.J. Ward, "Flicker Limitations of Electric Utilities", IEEE Transactions on Power Apparatus and Systems, Vol. PAS 104, No. 9, Sept. 1985, 2627-2631.
3.
D.J. Chee-Hing, F.M. Wheeler, "The New ISCOTT Meltshop, In Trinidad", Electric Furnace Conf. Proc., 1981, 39, 48-56.
4.
V. Brandwaj n, " Electromagnetic Transients Program (EMTP) Revised Rule Book, Version 2.0", EPRI EL-6421-L, Vol. 1, Research Project 2149-4.
5.
Hatch Associates, "Static VAR Compensation of Voltage Flicker from Electric Arc Furnaces", CEA Project 042-T-818, in progress.
6.
R. M. Mathur, Ed., "Static Compensators for Reactive Power Control", CEA-Cantext, Winnipeg, Canada, 1984.
7.
Roettger Eng. Co., "LOADFLOW 2.2", Naples, Florida, May 1989.
8.
The Math Works, Inc., "MATLAB", MA, USA, 1985 - 1990.
9.
M.A. Pesonen, "Harmonics, Characteristic Parameters, Methods of Study, Estimates of Existing Values in the Network", Electra, No. 77, Pg- 35-53.
10.
NEMA Standard MG1-20.55.
