CIRED 20th International Conference on Electricity Distribution
Paper 0185
CALCULATION OF VOLTAGE SAGS ORIGINATED IN TRANSMISSION SYSTEMS
USING SYMMETRIC COMPONENTS
Carlos CASAROTTO
Comahue National University, Argentina
cfcasarotto@yahoo.com
ABSTRACT
The objective of this work is the study of voltage sags
presented at a particular point of an electric network, when
a short circuit fault takes place in some point of the
transmission system. Different types of short circuits were
analyzed, so much balanced as not balanced, applied to a
bar of a typical power system, and the resulting voltages
were obtained in the terminals of the sensitive equipment
that are connected at the point of common coupling (PCC)
on another bar. The habitual techniques of symmetrical
components for short-circuit analysis were used. The
sensitive equipment were considered threephase, star
connected with neutral. To contrast the results the power
system was implemented in the program DigSilent, and the
same events were simulated. Starting from the obtained
results, and carrying out the appropriate simplifications,
the voltage sags were classified according to the
approaches mentioned by Bollen.
INTRODUCTION
The voltage sag (Sag) can be defined as short duration
reduction of effective value (rms) of voltage, caused by
short circuits, overloads and motor starts. The interest in its
study has been lately extended to be approached as a Power
Quality Problem of global reach, and whose consequences
can affect to different types of sensitive equipment as
Adjustable Speed Drives (ASD), process control devices
(for instance PLCs), and computers.
The equipment of “end use” that is affected by the existence
of voltage sags, and in general for events of power quality,
is denominated sensitive equipment (SE). They are affected,
132 A
PCC
External Grid
LAT 300/50 mm2- 10 Km
r = 0.094 Ohm/ Km ro = 0.245 Ohm/ Km
x = 0.388 Ohm/ Km xo = 1.164 Ohm/ Km
b = 2.951 uS/ Km bo = 1.968 uS/ Km
Skk 3fhi = 1500 MVA
Skk 1fhi = 1400 MVA
T 132 / 33 / 13.2 kV
S= 30 / 30 / 30 MVA
Z12 = 11.2 % Z012 = 9.35 %
Z13 = 18.1 % Z013 = 14.45 %
Z23 = 5.52 % Z023 = 5.40 %
starting from the impossibility of fulfilling their design
CIRED2009 Session 2
Prague, 8-11 June 2009
Juan Carlos GÓMEZ
Rio Cuarto National University, Argentina
jcgomez@ing.unrc.edu.ar
specifications, up to the upsetting of their operation. [2].
A very important statistical aspect it is constituted by the
fact that the faults that take place in transmission systems
have an occurrence percentage of approximately 45% for
single phase, 21% for two-phase and 31% for three phase
faults [2, 3].
The presence of transformers between the high voltage
networks and the sensitive equipment introduces changes in
the way in that the voltages are presented in the terminals of
the low voltage equipment. These changes imply that the
mentioned percentages will change according to the number
and group connection of the involved transformers [4, 5].
Besides the magnitude and duration, most of the
bibliography indicates that the equipment behavior depends
on other variables, as the unbalance and the phase jump. For
that reason, the work was centered in the characterization of
voltage sags taken place by symmetrical and asymmetric
faults, using the symmetrical components methodology,
since it facilitates to extend the analysis per-phase to
unbalanced systems. This method has demonstrated to be
superior to the direct one to solve unbalanced problems [6,
7].
TYPES OF BASIC VOLTAGE SAGS
Were denominated basics the voltage sags that take place by
the four elementary types of short circuit that are usually
required in fault studies that are the three-phase short
circuit, single-phase short circuit, two-phase short circuit
and two-phase to ground short circuit.
As it is habitual in the studies of power systems, load
132 B 132 C
LAT 300/50 mm2- 35 km
Transmission
T 132 / 33 / 13.2 kV T 13.2 / 0.4 kV
Distribution S = 1 MVA
Zcc = 5 % Load A
50 mm2 - 4.5 km
RN
33 A 13.2 B 0.38 A
13.2 A
Subtransmission
Figure 1. Power System Network
current were considered negligible, when the fault takes
Paper No 0185
CIRED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0185

place, and the possible contribution from the motors 1 1

Z eq 1 = = = 0 .0667 pu
connected to the industrial circuits was underrated. The S cc 3φ pu 15
representative loads of the sensitive equipment are
considered, as first approach, three-phase and connected in For transmission lines, the impedance is given by:
star with neutral. The rest of the power system was replaced Z L1= j 3 .880 Ω ⇒ Z L1 = j 0 .02226 pu
by an equivalent one static with three-phase and single-
Z L 2 = j 13 .58 Ω ⇒ Z L 2 = j 0 .07793 pu
phase short circuit power of 1500 MVA and 1400 MVA
respectively. The modeling of the high voltage lines for the The direct sequence short circuit impedance value results:
manual calculations was based on the equations of short
line, rejecting the resistive losses and the influence of Z cc1 = j 0.1669 pu
parallel admitances. The short circuit voltages of the three
winding transformer were considered only inductive, their The sequence currents are calculated as:
losses neither the presence of tap changers were taken into E
I 0 = 0; I1 = ; I2 = 0
account. It was considered the contribution of the zero Z S1 +Z f1
sequence components for the faults that involve earth. In all
the cases the fault impedance was considered null and the
As the load currents are neglected, the EMF value of the
presence of generators in the low voltage systems were not
Thevenin equivalent on the faulted point is:
taken into account. All the values were referred to the high
E = 1∠ 0º (phase A reference).
voltage system, with a common power base of 100 MVA
and a voltage base of 132 kV. 1 ∠0º
I1 = = 6 ∠ − 90 º pu
0 . 1669 ∠ 90 º
1- Three-phase short circuit: Type A The sequence voltages at the PCC are obtained by:
The simplest case of being analyzed presents when a three- V1 = I 1 Z L 2 = 0.4676 ∠ 0º pu
phase fault takes place in the bar 132 C of the transmission
system. To quantify the voltage sag magnitude in radial V 2 = V0 = 0
networks, the simple circuit that is illustrated in Figure 2
can be used. In the circuit, the equivalent of the system or The phase values can be obtained from the Fortescue
source it is represented by means of an impedance Zs transformation matrix.
connected to the point of common coupling (PCC), and by
means of Zf the impedance between the PCC and the fault
V abc = T V012[ ]
location is represented. ⎡V A ⎤ ⎡1 1 1 ⎤ ⎡V 0 ⎤ ⎡ 0 . 467 ∠ 0 º ⎤
⎢ ⎥ ⎢1 ⎥ ⎢ ⎥ ⎢
= 2
⎥ = ⎢ 0 . 467 ∠ − 120 º ⎥⎥ pu
In the present case Z S = Z eq + Z L1 and Z F = Z L 2 . ⎢V B ⎥ ⎢ a a ⎥ ⎢V 1
⎢V C ⎥ ⎢⎣1 a a 2
⎥⎦ ⎢V 2 ⎥ ⎢⎣ 0 . 467 ∠ 120 º ⎥⎦
⎣ ⎦ ⎣ ⎦
Figure 3 shows the connection of the sequence nets for this
With a = 1∠ 120 º
fault type. For being a symmetrical type of fault only direct
sequence values are involved. These results were compared with those obtained by
simulations. The obtained voltages at bar 132B are given.
⎡V A ⎤ ⎡ 0 . 475 ∠ − 5 . 37 º ⎤
⎢ ⎥ ⎢ 0 . 475 ∠ − 125 . 37 º ⎥ pu
⎢V B ⎥ = ⎢ ⎥
⎢V C ⎥ ⎢⎣ 0 . 475 ∠ 114 . 63 º ⎥⎦
⎣ ⎦

Figure 2. Simple circuit model. It is observed that besides a change in the magnitude, a
phase jump appears. In the simplified method this change
I1 doesn't appears because all the resistive losses were
Z S1 ZF1
rejected. For short circuits in sub-transmission or
+ distribution systems, this simplification can give important
E V1 errors, due to the high resistance values for small cross
− section conductors. Also, the difference of impedance of the
high and low voltage conductors can introduce phase jumps
Figure 3. Sequence nets connection for three-phase faults. higher than those studied up to now. The cited errors are
The direct sequence impedance of the external equivalent acceptable, for what the simplified method can provide very
circuit can be obtained by: good results with minimum calculations. This is important
in cases where the study is done without specialized
programs. This sag was classified by Bollen as Type A [1].

CIRED2009 Session 2 Paper No 0185

CIRED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0185

2- Single Phase short circuit: Type B 4- Two phase to ground short circuit: Type E
For this fault type the sequence nets are connected in series. In this case, the direct, inverse and zero sequence nets are
The three windings transformer will be studied for the connected in parallel. The results of the manual
connection group more commonly used in power systems. calculations and simulations are the following ones:
The presence of the delta winding allows the circulation of
Manual Simulation
the necessary harmonic component for the magnetic circuit,
excitation without important voltage distortions. The high ⎡V A ⎤ ⎡1 . 028 ∠ 0 º ⎤ ⎡V A ⎤ ⎡1 . 028 ∠ 0 .08 º ⎤
⎢ ⎥ ⎢ 0 . 547 ∠ − 125 . 4 º ⎥ ⎢ ⎥ ⎢ 0 . 532 ∠ − 132 . 30 º ⎥
voltage windings are connected in star to ground, which ⎢V B ⎥ = ⎢ ⎥ ⎢V B ⎥ = ⎢ ⎥
allows lower insulation level. It was decided to include in ⎢V C ⎥ ⎢⎣ 0 . 547 ∠ 125 . 4 º ⎥⎦ ⎢V C ⎥ ⎢⎣ 0 . 510 ∠ 123 . 38 º ⎥⎦
⎣ ⎦ ⎣ ⎦
the circuit the existence of neutral reactor, although it
doesn't participate in the calculations. For the analysis of the This voltage sag type has been proposed as sag Type E. In
circulation of the zero sequence components, it should be this case the not failed phase it is not affected and a
kept in mind that the delta coil allows the current decrease of the remaining two phases magnitude is verified
contribution coming from earth toward the fault, through the but the phase change can be disregarded with little
transformer primary neutral. In consequence this path calculation error, which notably simplifies the mathematical
should be included in the net of zero sequence, adding a expressions. The following table is a summary of the types
branch in parallel with the sum of the external impedance of of PCC voltage sags, according to the faults that originated
the equivalent one and the line L1. The impedance between them.
primary and third is taken in the common base of 100 Table 1. PCC voltage sags types
MVA. In general equals impedances of direct and inverse
sequence are adopted. This is acceptable in the case of Type of fault Star with neutral
faults electrically far from rotating machines. The results of Three phase A
the manual calculations and the simulations are:
Single phase B
Manual Simulation
Two phase C
⎡V A ⎤ ⎡ 0 . 572 ∠ 0 º ⎤ ⎡V A ⎤ ⎡ 0 . 579 ∠ − 3 .74 º ⎤
⎢ ⎥ ⎢1 . 018 ∠ − 121 . 7 º ⎥ ⎢ ⎥ ⎢1 . 019 ∠ − 121 .69 º ⎥ Grounded two phase E
⎢V B ⎥ = ⎢ ⎥ ⎢V B ⎥ = ⎢ ⎥
⎢V C ⎥ ⎢⎣1 . 018 ∠ 121 . 7 º ⎥⎦ ⎢V C ⎥ ⎢⎣1 . 017 ∠ 121 .78 º ⎥⎦
⎣ ⎦ ⎣ ⎦ TRANSFER AMONG DIFFERENT VOLTAGE
This event type has been defined as voltage sag Type B. In LEVELS
fact in the classification proposed by Bollen the presence of To connect the transmission net with the rest of the system,
the voltage sag only affects the failed phase, rejecting for the existence of a 132/33/13.2 kV transformer station has
the healthy phases so much the voltage elevation like the been supposed in the bar 132 B that provides the electric
phase change (although small) that is evidenced in the energy to the subtransmission and distribution nets. These
calculated values and also in the results of the simulation nets extend the feeding to the low voltage systems where it
that are given next. is supposed that the sensitive equipment is connected. The
transformers can have diverse connection forms, but in
3- Two phase short circuit: Type C order to explain the voltage sags transfers is enough to
classify them in three types.
To maintain the phase A as symmetry reference, it is
supposed that the fault takes place in the phases B and C. Type I: Transformers that don't change the voltages
For this fault type the direct and inverse sequence nets are between primary and secondary. For this type the voltages
connected in parallel, while the zero sequence net is opened in pu in both sides are the same, this means that the
by not being included the contact to ground. The results of transformer type I doesn't produce any change in the voltage
the manual calculations and simulations, applying the same sag type when transferring the voltages between levels. The
methodology of the previous examples, are: transformation matrix is the identity. Example YNyn

Manual Type II: Transformers that remove the zero sequence. The
Simulation
voltages of both sides are the same except for the
⎡V A ⎤ ⎡1 ∠ 0 º ⎤ ⎡V A ⎤ ⎡1 ∠ 0 . 006 º ⎤
⎢ ⎥ ⎢ 0 . 643 ∠ − 141 º ⎥ ⎢ ⎥
component of zero sequence. Example Yy Dd
⎢ 0 . 677 ∠ − 142 . 74 º ⎥
⎢V B ⎥ = ⎢ ⎥ ⎢V B ⎥ = ⎢ ⎥
⎢V C ⎥ ⎢V C ⎥ Type III: Transformers that eliminate the zero sequence
⎣ ⎦ ⎣⎢ 0 . 643 ∠ 141 º ⎦⎥ ⎣ ⎦ ⎣⎢ 0 . 617 ∠ 138 . 4 º ⎦⎥ and that exchange line by phase voltages. For these
This event type has been classified by Bollen as voltage transformers each secondary voltage is the difference
sag Type C. It can be observed that the not failed phase between two primary voltages. YNd Yd Dyn Dy
doesn't modify, while the two phases that participate in the In matrix form: V abc = [TR 2 ] V ABC . Where TR 2 indicate the
short circuit suffer a magnitude decrease and a phase
change. transformation matrix.

CIRED2009 Session 2 Paper No 0185

CIRED 20th International Conference on Electricity Distribution Prague, 8-11 June 2009

Paper 0185

TipoI TR 1 TipoII TR 2 TipoIII TR 3 In table 2 the types of sags that can be obtained by transfer
⎡1 0 0⎤ ⎡2 −1 − 1⎤ ⎡0 1 − 1⎤ from high to medium voltage and from medium to low
⎢0 0 ⎥⎥
1⎢
− 1⎥⎥
j ⎢
⎢ 1 −1 2 −1 0 1 ⎥⎥ voltage are summarized, with the consideration of the group
3⎢ 3⎢ obtained by means of the multiplication of transformation
⎣⎢ 0 0 1 ⎦⎥ ⎣⎢ − 1 −1 2 ⎦⎥ ⎣⎢ 1 −1 0 ⎥⎦
matrixes. As it is not very probable in the real world the
To maintain the phase “A” as reference with angle 0º, had presence of two type I transformers in cascade I, only
to rotate the values 90º multiplying the matrix given by j. equivalents of the group TR2 or TR3 are possible. In
3 is included so that the secondary values in pu are consequence the sags type B and E do not take place in low
similar to those of the primary one. If it is wanted to voltage nets, since zero sequence component is blocked.
analyze the transfer toward the low voltage sensitive loads, Table 2. Summary of different voltage sags types
it can be considered the connection of the transformers in
cascade by means of the multiplication of the Primary Secondary
transformation matrix, in way of obtaining an equivalent Sag Group TR1 Group TR2 GroupTR3
group. A A A A
TR 1 . TR 1 = TR 1 TR 2 .TR 2 = TR 2 B B D* C*
TR 1 . TR 2 = TR 2 TR 2 .TR 3 = TR 3 C C C D
E E G F
TR 1 .TR 3 = TR 3 TR 3 .TR 3 = TR 2

The voltage sag type A is symmetrical and in consequence CONCLUSIONS

it is transferred without losing their identity anyone it is the
group of selected transformer, for what only the asymmetric
cases will be analyzed. For the group TR1, the
transformation matrix is the identity, in consequence the sag
that is applied to the primary winding is transferred without
modifying to the secondary one. Next the results are given
obtained for voltages in the secondary of the group TR2,
when on the primary one (HV) voltage sags type B, type C
and type E are applied.
HV Sag B HV Sag C HV Sag E
0,72 ∠ 0º 1 ∠ 0º 0,85 ∠ 0º
0,94 ∠ -112º 0,64 ∠ -141º 0,64 ∠ -132º
0,94 ∠ 112º 0,64 ∠ 141º 0,64 ∠ 132º
For the type B, a sag is transferred for which modifies the
voltage magnitude of the failed phase and also of the real
component of the not failed phases, it was classified as
Type D *. The sag type C doesn't modify for this group, as
long as for the type E the sag that appears it classified as
Type G and is characterized by a modification in the
voltage real and imaginary components in the failed phases
and a change in the voltage module of the not failed phase.
As in the previous case, the results are given obtained for
the voltages in the secondary of the group TR3, when in the
primary one sags type B, type C and type E are applied.
HV Sag B HV Sag C HV Sag E
1 ∠ 0º 0,47 ∠ 0º 0,55 ∠ 0º
0,80 ∠ -129º 0,90 ∠ -105º 0,78 ∠ -110º
0,80 ∠ 129º 0,90 ∠ 105º 0,78 ∠ 110º
For the type B, the sag that appears in the secondary coil
possesses characteristic similar to the type C, although it is
less severe in magnitude and phase change. It was
classified as Type C *. The voltage sag type C generate
similar sag to the one seen in the previous case for sag B
and for Bollen is Type D, while for the type E the sag that
appears modifies all the components, and it was as Type F.
The method of the symmetrical components has
demonstrated to be a powerful tool for the analysis of
unbalanced systems. The application to the case of voltage
sags originated by faults in transmission system, with the
habitual simplifications, allows the classification of the
events in a simple way and with a minimum calculation
effort. Keeping in mind the habitual transformer connection
groups, only will be presented in the low voltage loads
symmetrical voltage sags type A, due to balanced faults,
and the asymmetric forms type C and D due to single and
two phase fault, and type F and G due to grounded two
phase faults. In reference to the statistical indexes shown at
the beginning, it can be affirmed that the most probable sag
types for industrial systems are C and D. The types F and G
are usually consider as distorted forms of the previous ones.
REFERENCES
[1] M. Bollen “Understanding Power Quality Problems:
Voltage Sags and interruptions”. NJ. IEEEPress, 2000.
[2] J. C. Gómez: “Calidad de potencia: para usuarios y
empresas eléctricas” Edigar, 2005.
[3] G. Yalcinkaya, M. Bollen, P. Crossley,
“Characterization of Voltage Sags in Industrial
Distribution Systems”. IEEE Transactions on Industry
Applications. Vol. 34, N° 4. pp. 682-688. 1998.
[4] M. McGranaghan, D. Mueller, M.. Samotyj, “Voltage
Sags in Industrial Power Systems”. IEEE Trans. on
Industry Applications. Vol.29, pp. 379-403, 1993.
[5] L. Zhang and M.H.J. Bollen, “Characteristic of
Voltage Dips in Power Systems”. IEEE Transactions
on Power Delivery. Vol. 15, N° 2. pp. 827- 832. 2000.
[6] P. M. Anderson, “Analysis of Faulted Power
Systems”. Iowa State Univ. Press. 1976.
[7] J. J. Grainger-W. D. Stevenson Jr. “Análisis de
Sistemas de Potencia”. Mc Graw Hill, México.

CIRED2009 Session 2 Paper No 0185