2004 Intematlonal Conference on

-
Power System Technology POWERCON 2004
Slngapore, 21-24 November 2004

Voltage Flicker Analysis and Mitigation-

Case Study in AC Electric Arc Furnace Using
PSCADEMTDC
C.Sharmeela, G . Uma, M.R. M o b and K.Karthikeyan

the power system quality. The power quality is mainly

Abstmct-Voltage flicker, a phenomenon of light intensity affected by the voltage and current harmonics, low power
fluctuation is caused by large and rapid industrial load factor and voltage fluctuations. Harmonics problems arise due
changes. The AC Electric Arc Furnace is one of the major to increasing use of power electronic systems and voltage
sources generating voltage fluctuations, deteriorating the fluctuation occurs because of large industrial loads such as
power quality. In the present work, the voltage EAF,rolling mills and pumps operating periodically in a weak
fluctuations produced by a 4 tonne AC EAF and its effect power distribution system. It causes voltage fluctuation at the
on the llkV/0.566kV distribution system is analyzed. The Point of Common Coupling and annoys residential consumers
EAF load is modeled based on field measurement data by causing visible lighting flicker on incandescent or
obtained from the AC EAF at a modern steel fluorescent lamps. The severity of the voltage flicker depends
manufacturing plant, Chennai, Tamil Nadu, India. on the fiequency and amplitude of the voltage variation and
Voltage flicker mitigation studies with and without static the short circuit rating at the PCC. It is reported that a small
var compensator are carried out, The low frequency voltage fluctuation of less than 0.5% in the frequency range of
voltage oscillations due to EAF load currents io the 5-1OHz can cause visible and uncomfortable incandescent
distribution system are eliminated by controllingthe f i g flicker [1]-[3]. Additionally they also cause reduction in
angle of the SVC. The PSCADmMTDC software is used electrical equipment efficiency, interference in protection
for modeling and simulation and the results are presented. systems, torque and power oscillations [43.
Voltage flicker has therefore been an important power
Index Terms-Electric Arc Furnace, Harmonics, Nonlinear quality concern for both power companies and customers. To
Load modeling, Static Vnr Compensator, Voltage flicker analysis quantify the degree of voltage flicker and to facilitate the
solution of flicker related problems, various definitions and
I.NOMENCLATURE standards have been proposed [5]-[6]. The IEEE standard 519-
AC E M - Altemathg Current Electric Arc Fumace 1992 [6] defines maximum permissible voltage flicker levels
AF bus - Arc Fumace bus with respect to frequency and allowable hannonic levels for
HV bus - High Voltage bus distribution systems. The voltage fluctuation problem can be
IF bus - Induction Fumace bus overcome by using voltage stabilisation devices like
PCC - Point of Common Coupling synchronous condenser with buffer reactor [7]-[SI.
p.f - Power factor The synchronous condenser during installations needs a
SVC - Static Var Compensator buffer reactor in series with the furnace busbar at the upstream
TCR-TSC- Thyristor Controlled Reactor-Thyristor Switched of the condenser. As a result, the flicker improvement is
Capacitor achieved only at the PCC and at the expense of an increased
THD - Total Harmonic Distortion flicker at the furnace busbar. The conventional compensator
requires regular maintenance and has limited flicker
II. INTRODUCTlON suppression capability even with buffer reactor. The onIy type
An AC EAF being used in the iron and steel industry is an of compensator capable of realizing the benefits of voltage
unbalanced, time-varying, non-linear load causing problems to stabilization are active compensators which has rapid
response. The SVC is one of the active compensator proving
C.Sharmeela, Research Scholar is with Department of Electrical and to be economic and technically advantageous. Hence, the
Electronics Engineering, CEG, Anna University, Chennai-6OOO25, India (e- present work investigates the application of SVC for the
mail: s h & p 2 0 0 2 ~ o o . c o . i n ) mitigation of voltage flicker due to EAF.An user-defined
G.Uma, Assistant Professor is with Department of ElectricaI and three-phase EAF load model is developed based on field
Electronics Engineering, CEG, Anna University, Chennai-600025, India (e-
mail: [email protected]) measurement data and the voltage flicker phenomenon due to
M.R.Mohan, Professor is with Department of EIectrical and Electronics EAF in llkV/O.566kV distribution system is modeled and
Engineering, CEO, Anna University, Chmai-600025, India simulated. The PSCADEMTDC tool is used for simulation.
KKarthikeyaoJdaster’s student is w i h Department of Electrical and The effective susceptance of SVC is controlled by
Electronics Engineeting, CEG,Anna University, Chennai-60(1025, India (e-
mail: [email protected]) continuously varying the firing angle of the thyristor and the

0-7803-8610-8/04/$20.00 Q 2004 IEEE 707

 melting process in a 4 tonne AC EAF at a modem steel
voltage oscillations due to EAF are eliminated. Also, it is
proposed to connect the SVC at the llkV/OS66kV manufacturing plant, Chennai, Tamil Nadu, India.
A simplified one-line diagram of the modern steel plant in
distribution power system that comprises o f a constant voltage
source connected to arc furnace load through transmissionChennai, India is presented in Fig. 2.The PCC is 1lkV HV
system and step down transformer. bus. The steel plant is having an arc furnace of 4 tonne
capacity connected to the EAF bus through 2MVA,
m. ELECTRIC ARC FURNACE 1 lkV10.566kV dedicated transformer. The AC EAF i s a major
load producing harmonics and voltage fluctuation in the
The typical EAF load cycle varies from 4 to 7 hours
distribution system. It is observed that when EAF is in
depending upon the size of the furnace and metallurgical
operation the voltage fluctuates at the PCC and other
sequin
residential consumers supplied fiom the PCC experience
visible lighting flicker. To carry out flicker and harmonic
studies in an EAF system, an exact load model o f EAF is
necessary. Hence in the present work, an attempt has been
made to model the AC E M based on the field measurement
data.
A. Current controlled non-linear resistance model of EAF
load
The arc melting process is a very complicated process. It
converts the electrical energy into thermal energy. The electric
arc is used to melt the raw materials held by the furnace. The
random movement of the melting material results in heavy
current fluctuations during the arc melting process. During the
r e f i i g period, the scrap metal is at a molten form and hence
fluctuations are small.
The furnace operation depends upon arc voltage, arc
current and arc length, which is determined by the position of
the electrodes. By examining the actual V-I characteristic of
the arc furnace, the arc melting process is divided into three
periods and explained as follows.
1. In the first period, the arc begins to reignite from
Fig. 1 A 4-Tonne AC FAF during melting
extinction. As the arc voltage increases to the reignition
The. operation of the arc fumace is divided into melting and
voltage Vi,, the E M acts as an open circuit. However, a
refining stages [9]-[11]. The random property of arc melting small leakage current exists, which flows through the foamy
process contributes to lowering of power factor resulting in
slag parallel with the arc. The foamy slag is assumed to be a
additional voltage drop through the power system yielding a
constant resistance Rg and the reignition voltage is assumed to
lower system voltage on the plant buses. Fig. 1 shows the
be proportional to the arc length.

11 kV

M
IXlSflNG P.F
CQRLtCTlON

605kVA 7
Fig 2 Simplified one-line diagram of the modem steel plant, Chennai, Tamil Nadu, India.

708
 2. In the second period, the arc is established. A voltage the voItage drop per unit arc length. Thus, the arc length
transient appears at the beginning of arc melting process. Then controls the nonlinear resistance.
the arc voItage drops suddenly flom V, to a constant value
B. Hannonic$eld measurements
vd. This process is expressed as an exponential function with
time constant 91. Field measurements are made at the PCC [12], in the AF
3. During the third period, the arc begins to extinguish. The bus and at the machining shop to study the levels of current
arc voltage continues to drop smoothIy, except a sharp change and voltage harmonics due to EAF load. The current and
after the arc extinction. This period represented by an voltage transformers are used as signal sources at all
exponential function with a time constant ~ 2 . measurement points. An adequate measurement is
Based on the above approximation, the EAF is modeiled as a accomplished by several days on-site. In addition to harmonic
current controlled non-linear resistance and the V-I data, power flow and power factor data are also captured to
characteristics is shown in Fig.3. The mathematical aid in analyzing the system behavior. The field measurement
representation of EAF is given by (1) and the arc model setup for recording the harmonics at the secondary side of arc
parameters are calculated using (2) to (4). h a c e transfonner is shown in Fig.4

- - - Piccc-wise lincar
approximate model

C -Actual model
Fig. 3 Actual aad piece-wise linear approximation ofV-I characteristic of arc

Fig. 4 Field measurement setup for recording harmonics

at Qe secondary side of arc finnace (ransfonna

The individual harmonic content of arc m a c e and %

THD of current and voltage are given in Table I. The
intensive field measurement study signifies the presence of
2nd,3', 5", 7" and 9" order harmonics. They are predominant
when compared to other harmonic orders.

Table I Individual harmonic content and %THD of arc furnace

dt

Where,

Vi, F: 1.15 *Vd (2)

6 I"HD(%) 38.8 41.9 5.0

v,= - [
lmc;']
*vd (4)
7 VTHD(%) 28.5 15.3 5.0
The average arc voltage V, has a linear relationship with the C.Mode1ing and Simulation of AC EAF using
average arc length I. This is given in (5). PSCAD/EMTDC simulation fool
The EAF load is modeled assuming rectangular arc voltage
V, = A + B.1 (51 to generate the V-I characteristics of E M . This is carried out
where A and B are constants. A is calculated based on using the practical system data given in APPENDIX. The
anode and cathode voltage drops and the value of B is equal to PSCADEMTDC simulation tool is used for modeling the AC

709
E M . The variation of arc resistance, source voltage, arc
voltage and arc current with respect to time is obtained by
500 7
simulation are shown in Figs.5 and 6.The variation of arc
resistance, arc voltage with arc current are shown in Figs.7
and 8 (a).The actual characteristics obtained by field
measurements are shown in Figs.8 @) & 8 (c). These results
are in close agreement with the theoretical waveforms
reported in [133. The results reveal that the V-I characteristics
obtained by the modeling of EAF using field measurement
data is in close agreementwith the theoretical result.

:-==I==
v
8
>

-20-8 -6 4
-80

-2 0 2 4 6 8 10

Fig. 5 Arc resistance curve

-12-10 -8 -6 4 -2 0 2 4 6 8 10 12
AnMIKpJ
(c)
Fig. 8 (a) Simulated V-Icharacteristic of EAF using typical values of EAF (b)
Observed V-I characteristic using field measurement values of EM during
Fig. 6 Source voltage, arc voltage and arc current waveform melting cycle (c) Observed V-I characteristic ushg field measurement values
of EAF drtring refining cycle
IV.VOLTAGE FLICKER SIMULATION AND
MITIGATION STUDY IN AC FAF
CURRENT
M E PHRSE REDER POlNl OF ARC FURNACE CONlROUED
FOURCE lYPEDRNCE COMMON COUPUNC TFk4NSFORMER RESISTWCE

-1% -100 -50 0 50 100 150

Arc Current [kA)

Fig. 7 Arc resistance Vs arc Current curve

Fig9 The PSCADEMTDC simulation model of AC EAF

In the present work voltage flicker mitigation studies are

carried out by modeling the EAF as a non-linear resistance.

710
Tbe PSCADEMTDC simulation model o f EAF is shown in
Fig. 9.
During the stable arcing period in the arc fumace
operation, the voltage fluctuation is noticed. Tbis is due to low
:r I -QC/QL

€tequency sinusoidal variation of arc resistance. It is modeled

as a sinusoidal variation with frequency 3 H z as shown in Fig.
.Qc/QL' I '
10. The Fig. 11 shows the simulation result of RMS voltage at
the PCC when the SVC is not connected to the PCC. It is seen Fig. 12 An SVC control system
that the voltage flicker of 2% for a 3 Hz modulated
fluctuation, which exceeds the permissible limits. The The Fig. 13 shows the RMS voltage at PCC with SVC
m a x i " permissible voltage flicker is 0.5% for a 5 Hz which shows a voltage variation of 0.005 %. The result
modulated fluctuation. reveals that there is significant decrease in the voltage
fluctuation by connecting SVC. Since the SVC compensates
the reactive power absorbed by the AF, the reactive power
drawn into AF bus from the distribution system is now very
small. It is also observed that the voltage distortion is well
below the E E E -519 standard limits. Table I1 shows the
comparison of % current and voitage THD, p.f without and
with SVC. These static compensators not only reduce the
flicker but also bring an increase io the plant's production
cost. Hence the SVC can be used for improving the load p.f to
unity.

Fig.10 Low frequency sinusoidal variation of an:resistance

Fig. 13 Three-phase R M S voltage with SVC

Table I1 Comparisonof % cumnt and voltage THD,p.f for Without and witb
svc
Fig. 1IVoltage fluchition at PCC due to AC EAF
I I I Without SVC I With SVC I
3.VoltageFlicker Mitigarion using SVC Current THD
I. 4.975 4.954
The digital simulation is performed to study the voltage (%)
flicker mitigation without and with SVC connected at the 2. VoltageTHD
(a.1
4.784 2.21
PCC.The SVC consists of a thyristorcontrolled reactor and \ 'U, I I
I
thyristor switched capacitor. The main objective of the SVC is 1
I
3. I
I
P.f I
1
65.69 1
I
85.72 I
I
to maintain the R M S voltage at the PCC within the limit. The
block diagram of SVC control system is shown in Fig. 12. V.CONCLUSION
Here, the reference voltage and actual voltage are compared.
The error voltage is amplified and given to the lead-lag Voltage flicker has emerged as a major concem in the area of
network. The output of the lead-lag network is given to the power quality. The voltage fluctuation problem in an
limiter. This gives the value of susceptance to be included to 11kV/O.566kV distribution system due to AC EAF is
the system to regulate load voitage. The effective susceptance investigated in this paper. The analysis and simulation results
of SVC is controlled by varying the f ~ of gthe thyristor of SVC application for the voltage flicker mitigation are
angle. The voltage oscillations occurring due to EAF presented and discussed. The validity and effectiveness of the
operation are eliminated by varying the firing angle of proposed SVC and EAF Ioad modeling has been demonstrated
SVC.The following simulation parameters are chosen for SVC through PSCAD/ EMTDC simulation tool. The dynamic
control system.Vref = 1.0, K 4.0,G =1.0, TI = 0.05, T2 = response and flexible control of SVC allows for effkient
0.05, Qc 50 and QL c 100. voltage flicker mitigation in distribution systems. Field
measurements recorded in modem steel plant during melting
and refining stages are quite useful in the analysis, providing

71I
input data and information to validate system models. The
analysis reveals that %THD and % individual harmonic IX.BIOGRAPHIES
distortion for current and voltage are higher than the specified
KEE 519- 1992 standards. Hence installations of harmonic C.Shmeela was b o m on 30” August 1977
m Chennai She received B E degree in
filters in the steel plant are essentiaf and suggested to bring Electmal and Electromcs Engmeenng and
real payoffs in the form of improvements in plant M E degree m Power Systems from
productivity, Anoamalai UruverSity, Chidambaram. in
1999 and 2001 respecixvvely At present, she
is working as Teachmg Research Associate
VI.APPENDIX in SEEE,Anna Uruversity, Chemai and also
The various parameters of the llkV/O.566kV practical pursuing her PhD m Elecmcal Engmeeriug,
Anna University Her fields of interest
system are,,,E = 566 V, f = 50Hz, & = 0.05R,&-
mclude Power System Stability Stt&es, Voltage Stabiltty - Dynamics, Power
=0~0000528 + JO.000468 and ZM t Swm = 0.0003366 + electronics apphcatmns to power systems, ~armonicsin p w a systems
jO.OO322n. The load modef is carried out with the data V,, =
350.75 V, I- = 100 kA, tl= 0.01 Sec and r2= 0.02 Sec.
C,Um was b m in Mafladuthurai. She
received B.E. &gme in Electronics and
W.ACKNOWLEDGMENT Instrumentation from Annamalai Univmity,
Authors express their thanks to High Voltage Engineering Chidambaram, in 1989. She obtained M.E.
degree in Control and instrumentation and PhD,
Division, Anna University for helping us in giving the Three fiom Anna University, Chennai in 1991 and
Phase MAVOWAIT45s Harmonic Analyzer for the Field 2002 respectwely. Presently, she is working as
Measurements. an Asst. Prof in S E E , Anna University,
Chennai. Her fields of interest include resonant
converters, Microprocessor based Control of
W.REFERENCES Solid State Drives, Power quality and Power
electronicscontrol in power systems.
P.Ash mole and P.Amante, “System flicker disturbances from industrial
loads and their compensation,” Journal on Power Engineering, ~01.3,
pp.213-218,Oct.l997. MR Mohan was born in Madurai, TamilNadu, India,an July 1 I, 1950. He
W.N.Chang,C.J.Wu and S.S.Yen,”A flexible voltage flicker teaching graduated from Thiagarajar ColIege of Engineering, Madmi, India. He
facility for e l h c pwer quality education,”EEE Trans. Power obtained his Doctoral degree &om Anna University, Chennai,India.Presently,
Systems,vo1.13,pp.27-33, Feb.1998. he is working as a Professor io SEEE, Anna University, Chennai,India. His
[31 J.Suq D.Czarkowski and Z.Zabar,”Voltage flicker mitigation using field of interest is Power system optimization, operation and conhol and
PWM-based Distribution STATCOW, E E E Power Engineering power system modeling. He is student counselor, IEE of UK at Chennai
Society S-er Meeting, vol.1, pp.616-621,.tuly.2002. Chapter, Charted Engineer and Member ofIEE.
[41 G.R Peuoto and J.C.Oliveira,“A proposal for voltage flicker
measuremeut in eldcal pOWR systems”,
in Roc. of the 38th Midwest Symposium on Circuits and Systcm~ K Karthikeyan waa born in Sayalkudi, India
cod., vol.2,pp. 1309-13 I2,Aug.l995. (1981). He completed his undergraduation from
IEEE Recommemled Practices for Electric Power Disbibution for Syed Ammal Engg. College, Ramauathapuram
lndutrialPlants,EEE Standard 141-1993,1993. He is currently pursuing his masters in power
IEEE Rccommeadcd Practices and Requirements for HarmonicsControl system engineering fiom the college of
in Electrical Powa systems, IEEE Standard 519-1992, April, 1593. Engineering, Guindy, Chennai. His areas of
G.El-Saady,”Adaptive static var controller for simultaneous e l i t i o n interest include power quality, power electronics
of voltage flickm and phase current imbalances due to arc fumacts appliedto power systems d control system.
loads”, Electric power system research, vol.58, pp.133-140,2001.
T.J.E.Miller, Reactive power control in Electric Systems, USA: John
Wilcy & Sons, 1982, pp.299-327.
A . R . O l t r o g g e , ” F ~ t a lcriteria for large arc furrrave power supply
system,” Joumal ofMetals,pp.53-64,Jau.1971.
Y.Sundberg,’The nrc furnace as a load on the networr, MEA Journal,
~01.49,pp.75-87,1976.
Roger C.Dugau,“Simulation of Arc Furnace Power SystwS.’. IEEE
Transaftions on Industry Applications, v01.M-16, pp.813-
8 18,Nov./Dec.l98IL
C.Sharmeela,G.Uma, M.R.Mohan and D.Ravichanchn,“ Case Study on
Field Measurement of Harmonics in AC Electric Arc Furnace”, in P m .
of National d. tm Power Elechonics in Energy Conservation IEEE -
PEECON,Feb. 2004, India,pp.119-121.
T. Zheng and Elhm B. Ma!“, “An Adaptive Arc Furnace Model”,
IEEE Transactions on Power Delivery, vol. 15, pp. 931-939, July.2000

712