This material is posted here with permission of the IEEE.

Such permission of
the IEEE does not in any way imply IEEE endorsement of any of Helsinki
University of Technology's products or services. Internal or personal use of
this material is permitted. However, permission to reprint/republish this
material for advertising or promotional purposes or for creating new collective
works for resale or redistribution must be obtained from the IEEE by writing to
[email protected].

By choosing to view this document, you agree to all provisions of the copyright
laws protecting it.
 PUBLICATION V

©2003 IEEE. Reprinted, with permission, from the publisher

Heine, P., Lehtonen, M., Voltage Sag Distributions Caused by Power System Faults,
IEEE Transactions on Power Systems, Vol. 18, No. 4, November 2003, pp. 1367-1373.
IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 18, NO. 4, NOVEMBER 2003 1367

Voltage Sag Distributions Caused

by Power System Faults
Pirjo Heine, Member, IEEE, and Matti Lehtonen

Abstract—Voltage sag distributions caused by faults at different

voltage levels and experienced by low-voltage customers were es-
tablished for four different power system areas. The shares of dif-
ferent fault types at each voltage level and the sag propagation
throughout the power system were taken into account. The results
show that the origin of sags in urban and rural areas tends to be
different. These data are needed when, for example, planning mea-
sures for sag mitigation in different parts of the power system.
Index Terms—Power distribution, power quality, power system
modeling, power systems, power transmission, voltage sag.

I. INTRODUCTION

A CCORDING to IEEE standard 1159-1995, a voltage sag

is defined as a decrease to between 0.1 and 0.9 p.u. in
root mean square (rms) voltage at the power frequency for du-
rations of 0.5 cycle to 1 min [1]. Voltage sags have always been
present in power systems, but only during the past decades have
customers become more aware of the inconvenience caused by Fig. 1. Origin of fault positions that cause sags experienced by an LV
them. customer.
A power system fault is a typical cause of a voltage sag [2].
Faults occur in transmission (EHV), subtransmission (HV), A. Sag Distribution
medium-voltage (MV), and low-voltage (LV) systems, and A sag distribution can be determined for each LV customer,
the sags propagate throughout the power system. The sag categorized in terms of the part of the network in which the fault
distribution experienced by a low-voltage customer includes all occurs (Fig. 1):
these sags of different origin.
1) transmission and subtransmission systems;
It is not essential that all power system areas are modeled
2) local MV distribution systems;
and included in voltage sag distribution calculations. This issue
3) adjacent MV distribution systems;
is studied in this paper. In addition, voltage sag distributions are
4) local LV distribution systems;
calculated for two urban and two rural power system areas. The
5) adjacent LV distribution systems.
sag propagation throughout the power system and the probabil-
ities of different fault types at each voltage level are taken into
The sags experienced by an LV customer can be described by
account in the calculations.
a cumulative distribution function
II. VOLTAGE SAGS CAUSED BY FAULTS
(1)
Voltage sags can generally be characterized by sag mag-
nitude, duration, and frequency [3]. Network impedances where is the fault frequency of fault type at fault position
determine the sag magnitude. When considering sags caused , the sagged voltage experienced by an LV customer at node
by faults, the protection practices specify the sag duration, and , and the sag duration for the fault place and fault type
the fault frequencies determine the number of voltage sags. in question.
Because different fault types cause sags with different char-
acteristics, the fault frequencies of each fault type should be de-
termined. Thus, the fault frequency at a certain fault position
includes the fault frequencies of all fault types: single-phase,
two-phase, and three-phase faults, with and without earth con-
Manuscript received January 23, 2003. This work was supported in part by
TEKES and in part by the Helsinki University of Technology. nections (2)
The authors are with Helsinki University of Technology, FIN-02015 HUT,
Finland (e-mail: [email protected]; [email protected]). (2)
Digital Object Identifier 10.1109/TPWRS.2003.818606

0885-8950/03$17.00 © 2003 IEEE

1368 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 18, NO. 4, NOVEMBER 2003

and the share of the various fault types will satisfy two TABLE I
properties (3) and (4) FAULT FREQUENCIES AND SHARES OF DIFFERENT FAULT TYPES FOR
FINNISH TRANSMISSION SYSTEMS [8]

(3)
(4)

B. Sag Magnitude
The calculation of voltages in meshed transmission systems
is based on Thevenin’s theorem and the network impedance ma-
trix [4]. To calculate the sagged voltage at bus caused by a fault
at node , (5) or (6) can be applied • 20 kV medium voltage, radially operated, neutral un-
earthed, or compensated;
(5) • 0.4 kV low voltage, radially operated, solidly earthed.
Typical transformer connections
(6) • 400/110 kV, 400/220 kV, 220/110 kV: YNyn0d11 (d11,
typically 21 kV, is used for compensation);
where and are the sagged voltages during the fault • 110/20 kV: YNd11;
at the nodes and , respectively. and are the prefault • 20/0.4 kV: Dyn11.
voltages. is the driving element that corresponds to the di- Y refers to wye connected, d to delta connected, and n to
agonal element of the node impedance matrix, the transfer neutral earthed systems. Capitals refer to the primary side and
element of the node impedance matrix that corresponds to nodes small letters to the secondary side of the transformer.
and , and the fault impedance. Sags caused by symmetrical three-phase faults propagate
For radially operated distribution systems, the calculation can without changes through transformers. In the case of unsym-
be simplified and a voltage divider model can be used [3]. For metrical faults, however, the transformer connections have a
example, in the case of a symmetrical three-phase short circuit strong effect [4]. The phase voltages on the secondary side
fault in one radially supplied MV distribution feeder, the sagged are derived from the phase voltages on the primary side
voltage on the substation busbar can be calculated using (7) as follows:
(9)
(7)

(10)
where is the impedance between the substation and the fault
location, the impedance of the primary transformer, and
the source impedance of the transmission system. (11)
When considering faults behind a neighboring HV/MV trans-
former, the PCC (= point of common coupling) is on the HV side
of the transformer and (8) should be applied (12)

(8) In (9), matrix transforms the phase voltages to symmet-

rical components, while matrix does the opposite. Matrix
The index 2 refers to the neighboring transformer and feeders determines the transformer type. The element depends
connected to this transformer. on how the zero sequence component propagates through the
transformer. If the zero sequence current cannot penetrate both
C. Sag Propagation the windings, then is set to zero. In a YNyn transformer
Although most customers are connected to LV networks, with both neutrals earthed, . Angle is determined
faults occur at all voltage levels. Hence, the sag propagation by the change in the positive sequence voltage.
throughout the entire power system should be modeled. The
fault type, earthing practices, and transformer connections D. EHV and HV Faults Experienced by an LV Customer
determine which voltages are of interest when considering sags 1) Fault Frequencies of Transmission Systems: In trans-
at the LV customer location. In this paper, the Finnish power mission systems, fault frequencies are typically small and the
system is used as an example to determine the voltage sag share of single-phase to earth faults is about 80% [5]–[8]. In
distributions experienced by LV customers, that is: Table I, the long-term fault frequencies and shares of different
• 400 and 220 kV transmission systems, looped, neutrals fault types for Finnish transmission systems are presented [8].
impedance earthed, or solidly earthed; 2) Sag Propagation From Transmission to LV Sys-
• 110 kV subtransmission system, looped, neutrals tems: When an earth fault occurs in a transmission system,
impedance earthed, or unearthed; one phase voltage is sagged, the neutral point voltage will rise,
HEINE AND LEHTONEN: VOLTAGE SAG DISTRIBUTIONS CAUSED BY POWER SYSTEM FAULTS 1369

and the voltages of the sound phases will increase. In Finland,

400 and 220 kV systems are earthed either solidly or through
an inductance. The earth fault factor is , which
means that during earth faults the maximum phase to ground
voltage in sound phases is 1.39 times the rated phase to ground
voltage [8]. The subtransmission system is either earthed via
an impedance or left unearthed so that in 110 kV systems the
earth fault factor is typically [8]. In a resistance or
high-impedance grounded system, the zero sequence source
impedance differs significantly from the positive and negative
sequence voltages. When a single-phase earth fault occurs in
phase R (assuming and ), the phase
voltages (p.u.) are [3]

Fig. 2. Shares of different faults for one power distribution company: faults
cleared by high-speed (H-S) reclosers, time-delayed (T-D) reclosers, and
permanent faults.

(13) E. MV Faults Experienced by an LV Customer

where index refers to the source impedance and to the 1) Fault Frequency of MV Systems: While transmission net-
impedance from the PCC to the fault. Further, index 1 refers works typically consist of overhead lines, MV networks con-
to the positive, 2 to the negative, and 0 to the zero sequence sist of underground cables in urban areas and overhead lines in
impedances. rural areas. The fault frequency of an MV overhead line network
When considering a single-phase earth fault in a 110 kV can be remarkably high. This is primarily due to construction
system in terms of the voltages seen by an LV customer, the and reclosing practices. Traditionally, the MV faults reported
zero sequence voltages do not propagate through the YNd and by power distribution companies only include data from perma-
Dyn transformers from the 110 kV to the 0.4 kV network. nent faults. A typical frequency of permanent MV faults is four
Hence, when applying (9) to the LV the side, (13) reduces to faults per year per 100 km [9]. However, these data are not ad-
the form equate for voltage sag calculations, where the faults cleared by
automatic circuit reclosers and the shares of different fault types
should also be known.
A detailed study of MV fault frequencies based on the fault
statistics of a Finnish power distribution company has been per-
formed. The material includes 720 faults over a period of 2.5
(14) years. The company has an MV network of bare conductor over-
When , the sagged voltage remains high, even in the head lines (50%), covered conductor lines (15%), and under-
case of a terminal fault. Thus, the lowest phase voltage at an LV ground cables (35%) with an unearthed neutral. Reclosers are
customer location caused by earth faults in a 110 kV system is in use, but not in purely underground cable feeders. The main
quite high (in Finland, typically ). Equation (14) is results of the study were (Fig. 2)
also valid for earth faults in the EHV system. Because the zero 1) Fifty-two percent of MV faults were cleared by high-
sequence impedances in the EHV system are smaller than in the speed (H-S) reclosers, 21% by time-delayed (T-D) re-
HV system, an LV customer experiences lower sagged voltages closers, and 27% remained as permanent faults.
during earth faults in the EHV system than during earth faults 2) The shares between earth faults and short circuits de-
in the HV system. pend on the network type. In cable feeders, the respec-
Further, because the transformer connections in the HV/MV tive shares were 80% for earth faults and 20% for short
transformer are YNd11 and in the MV/LV transformer are circuits. In feeders where reclosers are in use, the shares
Dyn11, sagged phase voltages caused by three- and two-phase of fault types depend on the fault clearing type. In faults
short circuits in the HV system have a similar appearance on the cleared by H-S reclosers, the share of earth faults was
LV side. In the case of a three-phase short circuit of 75%, and for T-D reclosers and permanent faults was
in an EHV or HV system, the sagged voltages are 0% in all 50%.
phases at an LV customer location. In the case of a two-phase 3) The probability of two-phase short circuits was four times
short circuit of , the sagged voltages of the faulted the probability of three-phase short circuits.
phases are 50% of nominal at an LV location while the phase
voltage of the sound phase remains unchanged. In two-phase to In comparison with the national statistics, values like 76% for
ground faults in transmission systems, the minimum voltages H-S, 17% for T-D, and 7% for permanent faults have been re-
at the LV location are slightly lower than in two-phase short ported [9]. Otherwise, the results are quite similar to the findings
circuits [3]. in [9] and [10].
1370 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 18, NO. 4, NOVEMBER 2003

In the case of a two-phase short circuit of fault resistance

, two of the phase voltages sag to a voltage of
while the other remains unchanged,
[Fig. 3(b)]. Further, only one phase-to-phase
voltage collapses to zero while the other two phase-to-phase
voltages decrease to 87% of the nominal voltage [Fig. 3(c)].
Thus, during two-phase short circuits on the MV side, only one
phase voltage on the LV side is substantially disturbed.
When an earth fault occurs in a high impedance earthed
medium voltage distribution network, the neutral point voltage
rises and the phase voltages of the sound phases increase. If
the fault resistance is 0 , the phase voltages of the sound
phases reach the value of the phase-to-phase voltages. Because
of the shift in the neutral point, there is no change in the
phase-to-phase voltages on the MV side. Further, there will be
no collapsed phase voltages on the LV side [Fig. 3(d)].

III. MODELED POWER SYSTEM AREAS

A. Sags Caused by Faults in EHV and HV Systems

Because of the meshed structure of transmission systems, the
sagged area caused by a transmission fault is typically large.
The affected line length may be hundreds of kilometers. The
sag distribution caused by transmission faults is highly depen-
dent on the location of the supply point in the transmission
system. In addition, if local generation exists, this will also have
a strong effect on the sag distribution [11], [12]. To evaluate the
sagged area, the network impedance matrix of the meshed net-
work should be calculated (5), (6).

B. Sags Caused by Faults in MV Networks

Most of the faults affecting LV customers occur in MV
networks. An LV customer experiences sags caused by faults
in the neighboring MV feeders and also via the HV systems,
from faults in the MV networks located behind the neighboring
substations.
1) Faults in Neighboring MV Feeders: MV networks are op-
erated radially. The most serious sags caused by MV faults are
those in the neighboring feeders of the sag sensitive customer. In
faults near an HV/MV substation, the voltage collapses to nearly
zero (7). The further away from the substation the fault occurs,
Fig. 3. Sagged p.u. voltages of (a) a three-phase short circuit, (b) a two-phase the higher the remaining voltage on the substation busbar will
short circuit (phase voltages), (c) a two-phase short circuit (phase-to-phase be. A strong subtransmission system and a large transformer
voltages), and (d) an earth fault (phase voltages) in an MV network with an
unearthed neutral. contribute to a higher remaining voltage (Fig. 4). These proper-
ties have approximately the same impact on the sagged voltage.

2) Sag Propagation From MV to LV Systems: The typical 2) Faults Behind Neighboring HV/MV Transformers: In the
connection of an MV/LV transformer is Dyn11. This means that case of a fault behind a neighboring HV/MV transformer, the
in (12) the element is zero and the angle . sagged voltage will not collapse to a high degree because the
Further, the phase-to-phase voltage on the MV voltage side is transformer impedance is now on the load side of the PCC (8).
seen as a phase voltage on the LV side. The further away the neighboring substation is located, the
In the case of a symmetrical three-phase fault on the MV side, less severe the sag will be. However, the effect of the HV line
all three phase-to-phase and phase voltages will collapse to the length is small (Fig. 5). The sagged voltage is lowest in the case
same degree and propagate without changes to the LV side. In of a weak transmission system and a large neighboring HV/MV
the example presented in Fig. 3(a), a three-phase short circuit transformer (Fig. 6). The short circuit level of the transmis-
causes a remaining voltage of 50%. sion system is the most critical factor in this analysis. In typical
HEINE AND LEHTONEN: VOLTAGE SAG DISTRIBUTIONS CAUSED BY POWER SYSTEM FAULTS 1371

• fuses limit the fault current and thus also the voltage drop.
As for sags caused by MV faults, the most serious sags
caused by LV faults are those in the neighboring feeders
near the MV/LV transformer. These sags are rare, however.
Further, when considering sags caused by LV faults behind
neighboring MV/LV transformers, the impedance between
the MV busbar (PCC) and the fault location consists of the
MV/LV transformer and LV lines. The sagged voltages remain
high because the MV/LV transformers have a rather large
impedance and the impedance of the LV line referred to the
MV side also represents a high value. For example, in Finland
one typical MV/LV transformer rating is 315 kVA, .
This represents a transformer impedance of 63.5 on the MV
Fig. 4. Effect of the short circuit level of the subtransmission system and the
HV/MV transformer rating on voltage sag magnitude, when a three-phase fault side. In addition, when the LV line impedance is referred to the
occurs in the neighboring MV feeder, z = 0:4 + j0:4
=km, Z = 0
. MV side, the LV line impedance is multiplied by a factor of
. In a strong city network [40 MVA 110/20
kV transformer and kA (110 kV)], this means a sagged
voltage of on the MV busbar and a corresponding
value of in a weak rural area (16 MVA, 2 kA).
Thus, in practice, sags caused by faults behind MV/LV power
transformers can be neglected in sag analysis.

IV. STUDY CASES: SAGS IN RURAL AND URBAN SYSTEMS

As study cases, sag distributions are calculated for four dif-
ferent LV points in Finland: 1) Urban1, east; 2) Rural2, east;
3) Urban3, south; 4) Rural4, south. In this study, the sag dis-
tribution includes sags caused by faults in 110 kV and 20 kV
Fig. 5. Sag magnitude as a function of the HV line length for typical rural
systems. The 220 kV and 400 kV transmission networks are ne-
[2-kA (110-kV), 16-MVA transformer] and urban [20-kA (110-kV), 40-MVA glected because of the small quantity of 220 kV line length and
transformer] systems, z = 0:1 + j0:4
=km (110 kV), Z = 0
. the low fault frequency in the 400 kV system. It was shown ear-
lier that the sags caused by LV faults only have a very marginal
effect on sag distribution and are not taken into account in this
paper. Different fault types are taken into account according to
(1) and (2) [13].
Two models of the 110 kV systems have been determined,
one for the eastern and the other for the southern area [14]. The
eastern area of 200 200 km has 1600 km of 110 kV trans-
mission lines. The model has two supply points to the 400 kV
network (the distance between the 400 kV nodes is 150 km).
The study points 1 and 2 are situated at a distance of about 150
km from the 400 kV supply points. Study point 1 is situated at a
point where four loops branch off, and point 2 is located at the
far end of a 30 km branch of a single loop.
Fig. 6. Sagged voltage caused by a three-phase short circuit behind a The modeled southern part includes the 110 kV networks of
neighboring HV/MV transformer Z = 0
. two cities having 400 km of 110 kV lines. In southern Fin-
land, the number of 400/110 kV substations is higher and, thus,
urban areas that have strong transmission systems, the sag im- the distance between substations is substantially smaller (30–50
pact of networks behind neighboring substations may even be km) than in the eastern area. In addition, the cities in southern
neglected. Finland have a considerable amount of local generation. Study
point 3 is situated in a looped city area at a distance of 15 km
C. Sags Caused by Faults in LV Networks
and 25 km from a 400 kV node. Point 4 is located at a distance
Typically, sags caused by LV faults are not taken into account of 10 km and 30 km between two 400 kV nodes.
in a sag distribution [3] because: The share of earth faults is 80% (Table I). In this paper, it is
• LV faults are rare [9]; assumed that the share of three-phase short circuits is 3%, and
• one distribution transformer supplies only a small number two-phase and two-phase-to-ground faults 17%. As was shown
of customers and, thus, these sags have only a minor and earlier, sags caused by earth faults in an HV system are barely
very local impact on the overall sag distribution; discernible at an LV customer location and can be neglected in
1372 IEEE TRANSACTIONS ON POWER SYSTEMS, VOL. 18, NO. 4, NOVEMBER 2003

TABLE II
INPUT MV DATA FOR THE CASE STUDIES

TABLE III
EXPERIENCED SAGS PER YEAR BY AN LV CUSTOMER

this analysis. The effect of two-phase short circuits is taken into

account by a factor of 2/3. This assumption slightly exaggerates
the frequency of the most serious sags, however.
Sag distributions caused by MV faults are calculated using
MV data in Table II [15]. The fault frequency in Table II has
been determined from data for permanent faults. In the calcula-
tions, it is assumed that the fault frequency is constant along the
entire feeder length. It is assumed that in urban areas there are
two transformers at a substation while substations in rural areas
have only one.
In point Urban3, south, the reclosers are not in use and the
share of short circuits is 20%. In points 1, 2, and 4, it is assumed
that 7% of the faults are permanent (of which 50% are short
circuits), 17% are cleared by time-delayed (30% short circuits),
and 76% by high-speed reclosers (30% short circuits). Further,
it is assumed that a third of the short circuits are three-phase
short circuits.
Faults were applied to all of the lines in the modeled areas and
the voltage profiles for the LV study points were computed by
applying (1). Different fault types were taken into account by (2)
and the sag propagation by applying (9). Sag frequencies caused
by HV faults, by MV faults in the neighboring feeders, and by
faults behind neighboring HV/MV substations are presented in
Table III and Fig. 7.
Results show that sag distributions are highly dependent on Fig. 7. Cumulative number of sags experienced by an LV customer caused by
faults in HV, adjacent MV and local MV systems.
the power system characteristics. In addition, the following
characteristics also hold.
• The LV customer Urban1, east, experiences the highest due to the longer critical line lengths of 110 kV networks
number of sags caused by transmission faults. In the in the eastern area.
eastern area, the study points 1 and 2 are situated far away • In southern urban city areas, sags caused by MV faults are
from the 400 kV supply points and the amount of local rare, and transmission faults are the most prominent cause
generation in the area is small. These characteristics are of sags.
HEINE AND LEHTONEN: VOLTAGE SAG DISTRIBUTIONS CAUSED BY POWER SYSTEM FAULTS 1373

• The annual number of sags is clearly higher in rural areas REFERENCES

than in urban areas. [1] IEEE Recommended Practice for Monitoring Electric Power Quality,
• In rural areas having long overhead line networks, the IEEE Std. 1159-1995, June 1995.
[2] IEEE Guide for Service to Equipment Sensitive to Momentary Voltage
share of shallow, less serious sags is noticeably higher than Disturbances, IEEE Std 1250-1995, Mar. 1995.
the share of the most serious sags. [3] M. H. J. Bollen, “Understanding power quality problems,” in Voltage
• It is only in the eastern rural area, where the transmission Sags and Interruptions. Piscataway, NJ: IEEE Press, 1999.
network is particularly weak, that faults behind a neigh- [4] I. J. Nagrath and D. P. Kothari, Modern Power System Analysis, 2nd
ed. New Delhi, India: TATA McGraw-Hill, 1989.
boring HV/MV transformer cause sags that affect the sag [5] D. Sikes, “Comparison between power quality monitoring results and
distribution. In the study cases, only sags having a higher predicted stochastic assessment of voltage sags—“Real” reliability
remaining voltage than 75% were included in this sag cat- for the customer,” IEEE Trans. Ind. Applicat., vol. 36, pp. 677–682,
Mar./Apr. 2000.
egory. In all of the other study cases, the faults behind [6] J. A. Martinez and J. Martin-Arnedo, “Voltage sag analysis using an
neighboring HV/MV transformers could be neglected. electromagnetic transients program,” in Proc. IEEE Power Eng. Soc.
Winter Meeting, vol. 2, 2002, pp. 1135–1140.
In principle, faults on all voltage levels contribute to the sag [7] Nordel Driftstörningsstatistik Fault Statistics 1999, 49 p.
distribution experienced by an LV customer but, because of net- [8] J. Elovaara and Y. Laiho, Sähkölaitostekniikan perusteet (in Finnish),
work characteristics, the sags caused by faults in some parts of 487 p., Otatieto Oy, Finland, vol. 499, 1988.
[9] Sähköenergialiitto ry, Keskeytystilasto 1998 (Interruption Statistics
the power system are shallow or even negligible. Thus, it is not 1998) (in Finnish), 18 p., Helsinki, Finland, 1999.
essential for all power system areas to be modeled and included [10] S. Hänninen and M. Lehtonen, “Characteristics of earth faults in
in voltage sag distribution calculations. electrical distribution networks with high impedance earthing,” Electric
The results shown have similarities to the findings in [16]. Power Syst. Res., vol. 44, pp. 155–161, 1998.
[11] J. V. Milanovic, R. Gnativ, and K. W. M. Chow, “The influence of
Knowledge of the network characteristics enables the sag dis- loading conditions and network topology on voltage sags,” in Proc.
tribution to be assessed by calculations, although long measure- Int. Conf. Harmon. Quality of Power, Orlando, FL, Oct. 1–4, 2000, pp.
ment periods would offer the most precise case- and site-sensi- 757–762.
[12] J. V. Milanovic and R. Gnativ, “Characteristics of voltage sags in radial
tive data. networks with dynamic loads and embedded generators,” presented at
the Proc. IEEE Porto Power Tech, Porto, Portugal, Sept. 10–13, 2001.
[13] P. Heine, P. Pohjanheimo, M. Lehtonen, and E. Lakervi, “A method for
V. CONCLUSION estimating the frequency and cost of voltage sags,” IEEE Trans. Power
Syst., vol. 17, pp. 290–296, May 2002.
Sag distributions are highly dependent on network character- [14] P. Heine and M. Lehtonen, “Influence of subtransmission system char-
istics. The sag distribution experienced by an LV customer in- acteristics on voltage sags,” presented at the Proc. Int. Conf. Harmon.
cludes sags caused by faults at all voltage levels. Because of the Quality of Power, Rio de Janeiro, Brazil, Oct. 6–9, 2002.
[15] M. Lehtonen, S. Kärkkäinen, and J. Partanen, Kokonaisvaltainen
construction of the meshed transmission system, sags caused by sähkölaitosautomaatiokonsepti Suomessa (Future distribution automa-
transmission faults propagate long distances, and affect urban tion system for Finnish utilities), (in Finnish), in Valtion teknillinen
and rural areas as well. tutkimuskeskus, VTT Tiedotteita—Meddelanden—Research Notes
1621, Espoo, 1995.
In urban areas, transmission faults are an important cause of [16] M. H. J. Bollen, “Fast assessment methods for voltage sags in distri-
sags, although the low fault frequency of underground cable MV bution systems,” in Proc. Ind. Applicat. Conf., Oct. 8–12, 1995, pp.
networks and short MV feeder lengths contribute to low sag 2282–2289.
frequencies. In addition, in urban areas, the strong transmission
system prohibits sags caused by faults behind the neighboring
substations from being experienced in other MV networks.
In contrast, in rural systems having typically long MV over- Pirjo Heine (M’99) was born in 1963. She received the Master of Science de-
gree from the Tampere University of Technology, Tampere, Finland, in 1987.
head line feeders and reclosers in use, MV faults represent the Currently, she is a Researcher with the Power Systems Laboratory of the
main cause of sags. The sag distribution mainly consists of sags Helsinki University of Technology, Helsinki, Finland. Her main interests are
caused by faults in the neighboring MV feeders. In the case of in power quality issues of distribution networks.
a weak transmission system, the faults behind the neighboring
substation may be of significance. The sag frequency of the shal-
lowest sags can be unpredictably high.
The kind of calculation and data concerning voltage sag char- Matti Lehtonen was born in 1959. He received the Master’s and Licenciate
acteristics and the origin of the majority of voltage sags pre- degrees from the Helsinki University of Technology, Helsinki, Finland, in 1984
and 1989, respectively, and the Doctor of Technology degree from the Tampere
sented in this paper is valuable when planning voltage sag miti- University of Technology, Tampere, Finland, in 1992.
gation measures in different parts of the power system. The re- Currently, he is with VTT Energy, Espoo, Finland, where he has been since
sults given were calculated using data from Finnish networks 1987. He has also been with the Helsinki University of Technology since 1999.
His research interests include earth fault problems, harmonic-related issues, and
and may, for example, not be entirely applicable to networks applications of information technology in distribution automation and energy
with different neutral earthing configurations. management.