Académique Documents
Professionnel Documents
Culture Documents
I HC BCH KHOA H NI
B MN H THNG IN
PHN II
CUNG CP IN
82
347
I. INTRODUCTION
MONG power-quality (PQ) phenomena, the voltage sag
(dip) is dened in IEEE1159, 1995 as a decrease in rms
voltage to between 0.1 and 0.9 of the nominal voltage at the
power frequency for the duration of 0.5 cycle to 1 min. There has
been a greater interest in voltage sags recently due to problems
caused by the performance of sensitive electronic equipment
that is widely used.
Research about the voltage sag is usually related to a basic
process known as a compatibility assessment [1], [2] which
includes three steps.
Step 1) Obtain the voltage sag performance of the system of
interest.
Step 2) Obtain equipment voltage tolerance.
Step 3) Compare equipment voltage tolerance with the
voltage sag performance and estimate the expected
impacts of the voltage sag on the equipment.
Current research has shown evidence that obtaining the
voltage sag performance still needs more improvement. The
83
Bach Quoc Khanh, Dong-Jun Won, Member, IEEE, and Seung-Il Moon, Member, IEEE
This paper proposes the idea of using 2-D stochastic models instead (e.g., the bivariate normal distribution model as illustrated
in Fig. 1).
For large power systems, it is hard to obtain a converged 2-D
fault distribution model for various causes in a large area. However, for small-to-medium-size networks, such as the section of
distribution network fed from a bulk-point distribution substation, of which the monitored historical data of fault performance
shows that faults due to external causes occur concentratively on
one location (e.g., some lines pass through a small area which
is at high risk for faults due to industrial pollution or trunk fall),
it is the favorite condition to obtain a converged 2-D fault distribution model.
IV. PROBLEM DEFINITION AND SOLUTION
A. Case Study Denition
To illustrate the new method of fault distribution modeling
in the stochastic prediction of voltage sag in the distribution
system, this paper uses the IEEE 123-bus radial distribution
feeder [7] as the test system. It can be seen as the distribution
system is fed from a bulk point. It does not narrow the scope of
application of the study with the following assumptions.
Since line segments in the test system come in one, two,
and three phases, distribution transformers at load nodes
are the single phase type for separate single-phase loads.
For three-phase loads, the connection of the distribution transformer is 4.16-kV grounded wyelow-voltage
grounded wye.
Voltage sags are only caused by faults in the test system.
If the test system is supposedly a section of a large distribution system, only faults occurring in it are considered.
The faults in sections fed from other distribution substations can be skipped as the transformer impedance in distribution substations, in reality, is rather high. Similarly, the
faults in low-voltage networks are also ignored because of
the large impedance of distribution transformers. This assumption only neglects voltage sags caused by faults in the
transmission system. It will be considered if the stochastic
prediction of voltage sag in large transmission systems [4]
is included.
In terms of reliability, the test system is modeled on two
main components: lines and distribution transformers. The
reliability of any other distribution equipment is supposedly included in the reliability of these two components.
The fault positions are selected as mentioned in Part II. For
transformers, one fault position at each load node (i.e., the
nodes connected with distribution transformers) is applied.
84
348
B. Fault-Rate Modeling
Faults are random events and as previously indicated, they
can be simulated by stochastic distribution models. According
to the analysis in Section III, the fault rate of each fault type at
each fault position is equal to the sum of equipment failure rate
and fault rate due to external causes. The equipment failure rate
is supposed to follow the uniform distribution model. Therefore,
for the fault position of the transformer , the failure rate is calculated as follows:
The joint probability density function of bivariate normal distribution is expressed as follows:
where
(5)
(1)
where
,
number of transformer faults of the test system;
total distribution transformers;
contributory percentage of equipment failure.
where
number of line faults of the test system;
total line segments;
length of the line segment (in feet).
The distribution of the fault rate due to external causes depending on fault positions is supposedly in compliance with the
2-D stochastic
model. This paper uses bivariate normal distribuB mn H thng in - i hc Bch Khoa H Ni
For lines, one fault position is also applied for each line
segment. Due to the short line segments, this paper selects
the fault position at the end of each line segment (For the
test system, there are 122 line segments and 87 load nodes.
Therefore, 209 fault positions in total are selected).
Fault types (single phase to ground, phase to phase, two
phases to ground, and three phases to ground) are applied
to fault positions depending on the number of available
phases. The fault impedance is assumed to be negligible.
The fault rate of a distribution transformer is a random
variable depending on the position of the load node it is
connected to. The fault rate of a line segment is also a
random variable depending on the fault position and the
length of this line segment.
Based on the previous denitions and assumptions, the computation of voltage sags at all load nodes on the primary side
of distribution transformers throughout the test system is performed on MATLAB [8]. The voltage sag frequency at each load
node is obtained when applying the fault rate to each fault position. The fault rates at the fault positions are calculated based on
the new fault distribution modeling presented in Part B. Finally,
related voltage sag indices are calculated.
349
TABLE I
SYSTEM FAULT-RATE BREAKDOWN
(8)
PQ indices are used to estimate the quality of supplied electric energy for the power system. To date, many PQ indices
have been proposed for various PQ events. A well-known index
of voltage sag is the system average rms voltage variation frequency index for voltage sag down to under X% of the nominal
voltage value
. It is often used for evaluating the PQ
of a three-phase power system based on monitored limited segmentation [3]. The assessed system is segmented so that every
point in the system is contained within a section monitored by
an actual PQ measuring instrument.
In distribution systems, because various phase loads (phase
to neutral, phase to phase, three-phase loads) are available,
asymmetrical faults, which account for most faults, never result
in voltage sags to all single-phase loads (e.g, phase A-to-ground
faults may not cause voltage sags to the loads connected between
phase B and neutral or phase C and neutral or loads connected
between phase B and phase C). Therefore, using
regardless of the number of phases involved, may not exactly
reect the voltage sag performance of the distribution system.
From the demand sides, the indices are more interesting because
they can estimate the voltage sag performance for phase loads.
In order to take the availability of various phase loads in the
distribution system into account, this paper newly develops
in regard to phase loads as follows:
(10)
(11)
(12)
where
,
Fig. 3. Sag frequency spectrum and SARFI of different phase loads for the case the mean value is at node 13 and deviation = = 0:51
C. Results Analysis
Based on aforementioned procedures of stochastic prediction,
the following are remarkable results.
In Fig. 3, the indices of voltage sag for different phase loads,
including voltage sag frequency spectrums, corresponding
,
, and
for X ranging
from 10% to 90% of the nominal voltage are shown. In this
case study,
,
.
Besides that,
for the whole test system for different mean values (at nodes 13, 51, 67, and 85) of the fault
distribution models regardless of the number of involved
phases are also depicted in Fig. 4. Obviously, there are big
differences between
of different phase loads or
between
of phase loads and
of the whole
system.
of phases A, B, and C are different
because the number of single-phase loads on each phase are
different.
are rather low as single-phase loads
just experience sags due to single-phase-to-ground faults on
the same phase. Generally,
are greater because
phase-to-phase loads are impacted by more faults (faults on
two phases) than phase-to-neutral loads (faults on one phase).
For phase-to-phase loads, there is a little deep sag frequency;
meanwhile, the shallow sag frequency rises greatly because almost phase-to-ground faults (80% system fault rate) just cause
shallow sags to phase-to-phase loads.
for three
phases is the greatest and
for
is equal
to 500 sags per load because three-phase loads will experience
voltage sag for any fault type. The aforementioned remarks
also explain why
, dened for phase loads, is for more
useful indices for estimating the voltage sag performance in the
distribution system where many single-phase loads exist.
Fig. 4 also shows that different positions of the mean value of
fault distribution models result in different spectrums of voltage
sag frequency. It is notable that if the position of mean value gets
closer to the bulk point of supply, the deep sag frequency will
increase, that is, mainly because of the radial network topology
of the distribution system.
87
351
Fig. 4. Sag frequency spectrum and SARFI of the whole system for different mean positions for the case that the deviation is = = 0:51
Fig. 5. Voltage sag frequency spectrum of the load-bus 63 on phase A for different deviations. The mean value is at node 67 (upper) and node 13 (lower).
Fig. 7. Voltage sag frequency spectrum and SARFI for the whole system for
different deviations for the mean value at node 67.
Fig. 8. Voltage sag frequency spectrum and SARFI for the whole system for
different deviations for a mean value at node 13.
Fig. 6. Voltage sag frequency spectrum for loads on phase A for different deviations. The mean value is at node 67 (upper) and node 13 (lower).
Figs. 5 and 6 plot the voltage sag frequency for load node
63 (see
Fig. 2) on phase A and for all loads on phase A for
B mn H thng in - i hc Bch Khoa H Ni
352
Fig. 9. Voltage sag frequency distribution for sags lower than 10%, 40% to 50%, 60% to 70%, and 70% to 80%,
node 67.
for the whole test system also for different deviation values
and for the mean values at
node 13 and node 67. Increasing the deviation values
and
will turn the normal distribution into the uniform distribution. It causes shape variations to the voltage sag frequency
spectrum. The clear increase of the frequency of deep sags is
shown in all cases of the sag performance demonstration. If
the mean position of the distribution model is located at node
13, which is very near the bulk point, the frequency of sags
below 10% is even raised by about 50% for the small deviation
. That is also explained as the result of
the radial network topology of the distribution system.
The spectrum of the voltage sag frequency for different case
studies (from Figs. 38) is quite similar in which deep sags account for a large number mainly due to short feeders in the distribution system. The frequency of 40% to 60% sags is also high
as the network topology consists of one trunk line with many lateral taps in the middle. That means the point of common coupling of many load nodes is on the middle of the trunk line. Few
load nodes connected to the trunk line near the bulk point of
supply (the distribution substation) explain why the shallow sag
frequency is very low. Fig. 9 gives us a closer look at the voltage
sag frequency distribution for different sag magnitudes. It is,
without doubt, that deep sag frequencies appear at the nodes
on branches connected close to the far end of the trunk line.
Voltage sags 40% to 50% are distributed rather uniformly excepting nodes near the bulk point. The shallow sag frequencies
mainlyBoccur
at several nodes near the bulk point of supply.
mn H thng in - i hc Bch Khoa H Ni
= = 0:51
= 50%, mean at
VI. CONCLUSION
This paper presented a new method of fault distribution modeling in the stochastic prediction of voltage sag for the distribution system using 2-D distribution models. When using 2-D
distribution models for modeling fault distribution, parameters
of the distribution model should be selected properly to match
the monitored historical data of fault performance of the system
of interest. By using the bivariate normal distribution for modeling fault distribution, this paper also analyzed the inuences
of its parameters on voltage sag performance. It is notable that
the alteration of the deviation value of the distribution has a
much stronger impact on sag performance, especially for the
deep sag frequencies pattern than switching the position of the
mean value. The more concentrated occurrence of faults on one
location in the distribution system of interest will increase the
number of deep sags. The results are also evidence that the typical radial network topology of the distribution system is also
another important reason for the high frequency of deep sags.
2-D stochastic models, such as the bivariate normal distribution used for modeling fault distribution, can provide a good
overview of fault performance of the whole system of interest.
Thus, it is possible not only to analyze the relation between
faults and voltage sags at individual locations of the system,
such as a specic load node or a segment of line, but also to
compute system indices of voltage sags, such as
.
The application of 2-D stochastic models has some limits to
the size of the system of interest. For the sections of the
dis89
353
tribution system, of which the size is so large as the one supplied from a bulk distribution substation, it is practical to use
this fault distribution modeling. The accuracy will be further
improved for the distribution systems, of which the topology
features the uniform arrangement of components. In addition,
the stochastic prediction of the transmission system should be
included if the inuence of fault occurring in the transmission
system on voltage sag performance in the distribution system of
interest is considered.
The presence of different phase loads in the distribution
for the whole system without
system indicated that
considering the number of phase of the loads cannot reect
voltage sag performance properly. To have a better assessment
of the voltage sag, this paper develops modied
regarding phase loads. The results proved that there are
,
, and
big differences between
for different phase loads and
for the
whole system. This modication of
is more practical
from the customers point of view when power-supply contracts
are set up.
Dong-Jun Won (M05) was born in Korea on January 1, 1975. He received the B.S., M.S., and Ph.D.
degrees in electrical engineering from Seoul National
University, Seoul, Korea, in 1998, 2000, and 2004, respectively.
Currently, he is a Full-Time Lecturer with the
School of Electrical Engineering with INHA University, Incheon, Korea. He was a Postdoctoral Fellow
with the Advanced Power Technologies Center,
Department of Electrical Engineering, University of
Washington, Seattle. His research interests include
power quality, dispersed generation, renewable energy, and hydrogen economy.
90
REFERENCES
[1] R. C. Dugan, M. F. McGranaghan, and H. W. Beaty, Electric Power
System Quality. New York: McGraw-Hill, 1996.
[2] M. H. J. Bollen, Understanding Power Quality ProblemsVoltage
Sags and Interruptions. New York: IEEE Press, 2000.
[3] D. L. Brooks, R. C. Dugan, M. Waclawiak, and A. Sundaram, Indices
for assessing utility distribution system RMS variation performance,
IEEE Trans. Power Del., vol. 13, no. 1, pp. 254259, Jan. 1998.
[4] M. R. Qader, M. H. J. Bollen, and R. N. Allan, Stochastic prediction
of voltage sags in a large transmission system, IEEE Trans. Ind. Appl.,
vol. 35, no. 1, pp. 152162, Jan./Feb. 1999.
[5] J. V. Milanovic, M. T. Aung, and C. P. Gupta, The inuence of fault
distribution on stochastic prediction of voltage sags, IEEE Trans.
Power Del., vol. 20, no. 1, pp. 278285, Jan. 2005.
[6] R. E. Brown, Electric Power Distribution Reliability. New York:
Marcel-Dekker, 2002.
[7] IEEE Distribution Planning Working Group Report, Radial distribution test feeder, IEEE Trans. Power Syst., vol. 6, no. 3, pp. 975985,
Aug. 1991.
[8] W. H. Kersting, Distribution System Modeling and Analysis. Boca
Raton, FL: CRC, 2002.
[9] T. A. Short, Electric Power Distribution Handbook. Boca Raton, FL:
CRC, 2004.
[10] G. Olguin, Voltage dip (sag) estimation in power system based on stochastic assessment and optimal monitoring, Ph.D. dissertation, Dept.
Energy Environ., Div. Elect. Power Eng., Chalmers Univ. Technol.,
Gotteborg, Sweden, 2005.
Bach Quoc Khanh received the B.S. and Ph.D. degrees in power network and systems from Hanoi University of Technology, Hanoi, Vietnam, in 1994 and
2001, respectively. He received the M.S. degree in
system engineering from the Royal Melbourne Institute of Technology (RMIT), Melbourne, Australia, in
1997.
He is currently a Lecturer with the Faculty of
Electrical Engineering, Electric Power System
Department, Hanoi University of Technology. He
was a Postdoctoral Fellow with the Power System
Laboratory, School of Electrical Engineering and Computer Science, Seoul
National University, Seoul, Korea. His special elds of interest include power
distribution system analysis, DSM, and power quality.
The 2009 ASEAN Symposium on Power and Energy Systems - EEE.RC.ASPES 2009
I. INTRODUCTION
SARFI X
N
i
X (i )
(1)
where
X rms voltage threshold; possible values 10-90%
nominal voltage
NX(i) Number of customers experiencing voltage sag with
magnitudes below X% due to measurement event i.
N number of customers served from the section of the
system to be assessed
Despite being widely used, SARFIX only considers the
magnitude of voltage sag and, of course, its value is maybe
much greater than the actual number of tripping electrical
appliances, especially when the duration of sags is small
enough (less than a half second). To take the sag duration
into account, SARFIX is developed into SARFIX-CURVE [2],
[4], [6] which is defined below
B mnNo.
H thng
Paper
PS-1in - i hc Bch Khoa H Ni
91
'
X (i )
i 1
SARFI X CURVE
(2)
where
N X' (i ) : Number of customers tripped when experiencing
voltage sag with magnitudes below X% due to measurement
event i.
SARFIX-CURVE corresponds to voltage sags below an
equipment compatibility curve. So far, frequently used
curves are CBEMA, ITIC and SEMI [1]. Obviously,
SARFIX-CURVE can provide a better understanding of the
influence of voltage sag on the operation of electric
equipment in electric networks. This paper presents the
method of calculating SARFIX-CURVE using ITI curve (Figure
1) as a case study.
III.
A. Problem definition
The problem of stochastic prediction of voltage sag can
obtain the voltage sag performance of a specific electric
system by using data of events leading to sags. In fact, more
than 90% sag events are resulted from short-circuits and it is
possible to use fault modeling and short-circuit calculation
tools to predict voltage sags in the power system (Figure 2).
tPD
E
~
Vt
ZF
ZS
Vt
Short circuit
tPD
Load at PCC
E
VSag
circuit faults per year) for all fault types at all possible fault
positions throughout the system of interest [3]. It consists of
the selection of fault position and fault type and the
distribution of fault rate for selected fault positions and fault
types.
Fault positions are generally chosen in the way that a
fault position should represent short-circuit faults leading to
sags with the similar characteristics [1]. For the distribution
system with typically radial network topology, small line
segments and distribution transformers along the trunk
feeders, it is possible to apply only one fault position for
each distribution transformer and also one fault position for
each line segment.
Different fault types should be applied to each fault
position mainly depending on number of phases available at
the selected fault positions. The fault rate of each fault type
is normally referred from the observed historical data.
Fault rate mainly depends on fault position, fault type
and fault cause. For a segment of distribution system that is
geographically seen as small area, it possible to assume that
fault rate of each fault type follows uniform distribution for
all fault positions. [3]. In this sense, the fault rate at each
position is identical to component failure rate that is based
on component reliability. In reality, uniform fault
distribution is a practical assumption for distribution
systems because the service area of a certain distribution
line outgoing from a distribution substation is normally
small.
C. Assumptions
Besides fault distribution modeling, for the distribution
system, following assumptions are possibly considered [3].
- Voltage sags are only caused by faults in the
distribution system.
- If the distribution system is supposedly a section of a
large distribution system, only faults occurred within it are
considered. The faults in sections fed from other distribution
substations can be skipped as the transformer impedance in
distribution substations, in reality, is rather high. Similarly,
the faults in low voltage networks are also ignored because
of the large impedance of distribution transformers. This
assumption only neglects voltage sags caused by faults in
the transmission system. It will be considered if the
stochastic prediction of voltage sag in large transmission
systems [7] is included.
- In terms of reliability, the distribution system is
modeled on two main components: lines and distribution
transformers. The reliability of any other distribution
equipment is supposedly included in the reliability of these
two components.
- The fault positions are selected as mentioned in the Part
III.B. For transformers, one fault position each load node
(i.e. the nodes connected with distribution transformers) is
applied. For lines, one fault position is also applied for each
line segment. Because of short line segments, the paper
selects the fault position at the end of each line segment.
- Fault types (single phase to ground, phase to phase, two
phases to ground and three phases to ground) are applied to
fault positions depending on the number of available phases.
92
The 2009 ASEAN Symposium on Power and Energy Systems - EEE.RC.ASPES 2009
The 2009 ASEAN Symposium on Power and Energy Systems - EEE.RC.ASPES 2009
ON FAULT-LINE CALCULATION
Calculate fault current IN and
sags VS at nodes on fault line
Circuit
breaker
Fuse
Distribution
transformer
SARFIX CALCULATION
Distribution
transformer
t PD
TRIPPING TIME
tPD = f(IN)
Fuse
a
(I * )b 1
(3)
where
I*: Ratio of fault current IN and pickup current IP.
a, b: Constants that are selectable.
calculation
SARFIX-CURVE CALCULATION
Sag quantification by duration
STOP
Figure 3. Block-diagram of voltage sag prediction
and SARFIX-CURVE in distribution systems
IV. A CASE STUDY
A. Case study definition
This work illustrates the method by predicting voltage
sag performance and resulting SARFIX-CURVE for a 24kV
feeder network in Hanoi, Vietnam. Preliminary data is as
follows
The network segment under consideration: Feeder 482E14, 24kV, underground cable, outgoing from 110/35/22kV
Giam substation. Its a radial network with 99 nodes and 98
branches. Fault positions can be selected at load nodes for
distribution transformer fault and at all nodes for line fault.
Besides, contributory percentages of different fault type
B mnNo.
H thng
Paper
PS-1in - i hc Bch Khoa H Ni
The 2009 ASEAN Symposium on Power and Energy Systems - EEE.RC.ASPES 2009
45
Sag
Sag leading load failure
40
35
30
25
VII. REFERENCES
20
[1]
15
[2]
10
5
[3]
0
0-10
10-20
20-30
30-40
40-50
50-60
60-70
70-80
80-90
[4]
[5]
120
SARFIX
SARFIX-CURVE
100
[6]
80
60
[7]
40
[8]
20
0
<10
<20
<30
<40
<50
<60
<70
<80
<90
VIII. BIOGRAPHIES
Bach Quoc Khanh received B.S. and Ph.D. degrees in power network
and systems from Hanoi University of Technology, Hanoi, Vietnam in 1994
and 2001 respectively. He received M.S. in system
engineering from RMIT, Melbourne, Australia in
1997. He is a teaching staff of Electric Power
System dept., Electrical Engineering Faculty,
Hanoi Univeristy of Technology. His special
fields of interest include power distribution
system analysis, DSM and power quality.
VI. CONCLUSIONS
This paper presented a method of assessing voltage sags
in distribution systems with regard to tripping time of
protective devices. The assessment bases on SARFIX-CURVE
that combines SARFIX and equipment compatibility curves.
Therefore, the results of assessment provide a better
understanding of the influence of voltage sag on loads.
This method is also found useful for power quality
assessment and power supply contracting principles for
power distribution utilities in Vietnam in the process of
electricity market establishment because the management of
distribution system is becoming financially separated from
the power system.
The application of the method has some limits that can be
B mnNo.
H thng
Paper
PS-1in - i hc Bch Khoa H Ni
94
NH GI ST GIM IN P NGN HN
TRN LI TRUYN TI IN 220KV VIT NAM
PREDICTION OF VOLTAGE SAGS IN THE 220KV TRANSMISSION SYSTEM OF VIETNAM
Phng Th Anh
Cng ty CP T vn Xy dng in I
TM TT
Bi bo trnh by phng php nh gi mt hin tng cht lng in nng (CLN) trn li
truyn ti in (LTT) l st p ngn hn (SANH - voltage sag) [1]. M hnh nh gi SANH da trn
phng php d bo ngu nhin SANH [2] trong h thng in (HT). Vic nh gi ny da trn ch
tiu tn sut SANH trung bnh ca HT vi c tnh X (SARFIX) v SARFIX-CURVE [3] cho php xt n
khng ch c trng bin ca SANH m cn c c trng thi gian tn ti SANH. i tng tnh
ton l h thng truyn ti in 220kV ca Vit Nam theo tng s 6 vi t l sut s c ngn mch
thc t ca nm 2008. Vic nh gi ny l mt c gng u tin nh lng ha tnh hnh mt hin
tng cht lng in nng ph bin trn mt li in din rng thc t gip cho vic nh gi cht
lng in nng ni chung ca h thng in Vit Nam hin nay.
This paper presents a method of predicting a power quality phenomena in distribution systems,
voltage sag [1]. The calculation of voltage sag performance follows the model of stochastic prediction
of voltage sag in power systems [2]. The voltage sag performance is predicted basing on the System
Average RMS variation Frequency Index (SARFIX) and SARFIX-CURVE [3] that considers not only the
characteristics - magnitude, but also the characteristics duration of voltage sag. The objective of
research is the whole 220kV transmisson systems in Vietnam as per the 6th master-plan with actual
data of fault rate of the year 2008. This prediction is the first effort of quantifying the voltage sag
performance for such a large transmission system that helps assess the power quality of the electric
power system in Vietnam now.
CLN c cung cp v nh gi tc ng ca
CLN i vi ph ti. Vic xc nh yu cu
CLN ca cc ph ti thuc v cc nh sn
xut thit b dng in m in hnh l c tnh
chu in p ca ph ti CBEMA, ITIC hoc
SEMI [1] (Hnh 1).
I. T VN
Theo IEEE-1159, 1995, SANH (voltage
sag) l hin tng CLN trong gi tr in
p hiu dng ca li in st gim cn t 0,1
n 0,9 in p nh mc trong thi gian t 0,5
chu k n 1 pht [1]. SANH c th lm cho
cc thit b in nhy cm nh in t cng
sut, cc b iu tc hay my tnh c nhn
ngng hoc lm vic khng mong mun. Hin
tng ny li rt hay xy ra, trong khi nng
cao hiu sut qu trnh hay vic ng dng cc
cng ngh mi, cc thit b in ng dng in
t cng sut ngy cng c s dng nhiu, do
SANH ngy cng c quan tm nghin
cu. Trc khi xem xt nhng gii php khc
phc tc ng ca SANH, yu cu nh gi
SANH trong HT lun c t ra. Qu trnh
nh gi CLN ni chung v SANH ni ring
thng tri qua ba khu ch yu [1] l i. Nhn
dng tnh hnh CLN c cung cp, ii. Xc
nh yu cu CLN ca cc ph ti, iii. So
snh yu cu CLN ca ph ti vi tnh hnh
95
ABSTRACT
Vng
an ton
Vng mt an ton
Vng mt
an ton
73
96
Start
M phng phn b s c trn
li truyn ti 220kV
N(1)
N(1,1)
N(2)
N(3)
Sut s c
ng dy
Trm bin p
92,95
29,25
28,60
9,00
14,30
4,50
7,15
2,25
- Chn v tr nt ph ti cn xc nh SANH :
Trn li truyn ti, nt ph ti l cc nt c
trm 220kV cp in xung cc li c in p
thp hn. Li 220kV li c dng mch vng
nn nhn chung trn mi nhnh ng dy
220kV, bo v c t ti c hai u v khi
xy ra s c ngn mch trn nhnh ng dy
no th nhnh s b c lp ring. Do , tt
c cc nt (66 trm 220kV) trn li in u
b SANH khi s c, khng c nt no b mt
in duy tr v ta phi tnh SANH cho 66 nt
ny.
Loi
97
SARFIX
N(1,1)
N(2)
N(1)
SARFIX-CURVE
X
X
98
li truyn ti in cn c xt thm cc
SANH do ngn mch phn ngun v li
truyn ti in 500kV. Hn na, nghin cu
cng cn c th pht trin vic xem xt cc yu
t nh hng n phn b s c khi li truyn
ti in ca Vit Nam tri trn mt phm vi
rng ln vi tnh hnh s c khc nhau. Cc m
hnh ngu nhin vi cc lut phn b xc sut
ph hp vi tnh hnh xy ra s c thc t c
th c xem xt [2, 6, 8].
V. KT LUN
Bi bo trnh by phng php nh
gi SANH trn li truyn ti in 220kV ca
Vit Nam thng qua ch tiu SARFIX v
SARFIX-CURVE. y l c gng u tin nh
lng ha vic nh gi tnh hnh SANH ni
ring v CLN ni chung trong HT Vit Nam
trn mt din rng.
Nghin cu trong bi bo cng cn c
pht trin thm. Kt qu nh gi SANH trong
1.
M. H. J. Bollen; Understanding power quality problems - voltage sags and interruptions; IEEE
Press, 2000.
2.
Bach Quoc Khanh, Dong Jun Won, Seung Il Moon; Fault Distribution Modeling Using Stochastic
Bivariate Models For Prediction of Voltage Sag in Distribution Systems; IEEE Trans. Power
Delivery, Vol.23, No.1, pp.347-354, Jan. 2008.
3.
Bach Quoc Khanh; Prediction of Voltage Sags in Distribution Systems With Regard to Tripping
Time of Protective Devices; Proceeding, EEE.CR.ASPES2009, Tech. Section 2.1., Hua Hin,
Thailand, Sep. 28-29, 2009.
4.
D. L. Brooks, R. C. Dugan, Marek Waclawiak, Ashok Sundaram; Indices for Assessing Utility
Distribution System RMS Variation Performance; IEEE Trans. Power Delivery, Vol.13, No.1,
pp.254-259, Jan. 1998.
5.
6.
7.
8.
Recommended practice for the establishment of voltage sag indices, Draft 6, IEEE P1564, Jan
2004.
9.
Tng s pht trin H thng in Vit Nam, Bn IV, Vin Nng lng, 2006.
76
99
I. INTRODUCTION
978-1-61284-788-7/11/$26.00
2011
B mn H thng in - i hc Bch
KhoaIEEE
H Ni
SARFI X
X (i )
(1)
where
X rms voltage threshold; possible values 10-90% nominal
voltage
NX(i) Number of customers experiencing voltage sag with
magnitudes below X% due to measurement event i.
N number of customers served from the section of the system
to be assessed
Despite being widely used, SARFIX only considers the
magnitude of voltage sag. Unfortunately, the magnitude value
maybe much greater than the actual number of tripped
electrical appliances, especially when the duration of sags is
small enough (less than a half second), such as for
transmission system in Vietnam where the total fault clearing
time of protection system is typically less than 5 to 7 cycles of
the mains frequency. To take the voltage sag duration into
account, SARFIX is developed into SARFICURVE-X [5], [6]
which is defined below
m
SARFI CURVE X
'
X (i )
i 1
(2)
where
event i.
If we plot voltage sag as a point with co-ordinates being its
magnitude and duration on the graph of the equipment
compatibility curve, SARFICURVE-X corresponding to voltage
sags falling out of the equipment voltage tolerant area (Fig. 1)
will be obtained. So far, well known curves are CBEMA, ITIC
and SEMI [1]. Obviously, SARFICURVE-X can provide a better
understanding of the influence of voltage sag on the operation
of electric equipment in electric networks. This paper presents
the method of calculating SARFIX-CURVE using ITIC and SEMI
curve (SARFIITIC-X and SARFISEMI-X) as case studies.
SARFIX-CURVE
disqualified
SARFIX-CURVE
qualified
101
220/110/35kV
Mai Dong substation,
2x250MVA, Hanoi
102
Regional Power
Transmission
Company
220kV
Substation
fault rate
(per year)
PTC1 (North)
0.00093
0.02504
0.0397
0.00562
0.00536
0.0408
0.00173
0.01279
0.0161
PTC4 (South)
0.0077
0.00808
0.0229
NPT
0.00407
0.01478
0.0306
Short-circuit
calculation and
determine voltage sag
magnitude at selected
load node by PSS/E
Fault distribution
modeling, determine
fault rate of the fault
under calculation
Sag frequency
spectrum by
the fault under
calculation
(event index)
Sag frequency
spectrum at
selected load
node by all
faults
Sag frequency
spectrum at all
load nodes by
all faults
(site index)
Fig 3.
Block diagram
of the problem
of prediction of
voltage sag in
the transmission
system of
Vietnam.
Are
all fault type
selected ?
START
False
True
Are
all fault position
selected ?
False
True
Are
all load nodes
selected ?
False
True
SARFIX
calculation
Check ITIC
curve ?
True
SARFIX-CURVE
(system index)
STOP
103
Sag Magnitude
(p.u)
SP-G
PP-G
P-P
Sag magnitude
(p.u)
Fig 4. Voltage sag frequency spectrum (per year)
by fault types at load node 220kV Mai Dong substation
Sag Magnitude
(p.u)
3P-G
Fig 5. Voltage sag frequency spectrum (per year) for all fault
events at 220kV Mai Dong Substation, Hanoi, Vietnam
B mn H thng in - i hc Bch Khoa H Ni
SARFISEMI-X
Sag Magnitude
(p.u)
Sag Magnitude
(p.u)
SARFIITIC-X
D. Result Analysis
From the results, therere some following remarks:
- The SARFIX and SARFICURVE-X values obtained from this
calculation are useful for utilities as a benchmark for reducing
the frequency of fault for solving the problem of voltage sag.
This result also helps customers know the voltage sag
performance and choose suitable location for less voltage sag
frequency.
- The frequency of voltage sag as the result of an individual
fault type is proportional to fault rate of that fault type for
shallow sags (Fig. 4).
- Shallow sags (0.7-0.9 p.u) feature a rather high frequency
while the frequency of deep sags is very small. Furthermore,
the frequency of voltage sag with X lower than 0.9 for either
the 220kV Mai Dong substation (about 33 times, Fig. 5) and
the system average load node (about 22 times, Fig. 7) is also
khanh
Sag magnitude
(p.u)
Sag magnitude
(p.u)
Sag magnitude
(p.u)
IV. CONCLUSIONS
This paper presented the first effort of predicting voltage
sag performance for a large transmission system as a case
study of Vietnam. In this work, fault distribution modeling is
proposed basing on actual fault performance for different
regions in Vietnam. Using SARFIITIC-X gives a better
assessment of voltage sag influence on loads operation. The
results of this work will be a useful reference for utilities in
power system quality assessment toward electricity market
operation. This research still needs to develop as faults in the
generation part has yet to take into consideration. Besides, if a
better fault data is achieved (by monitoring), a more detailed
fault distribution can be made and finally a better voltage sag
performance can be obtained.
[3]
Bach Quoc Khanh, Dong Jun Won, Seung Il Moon, Fault Distribution
Modeling Using Stochastic Bivariate Models For Prediction of Voltage
Sag in Distribution Systems, IEEE Trans. Power Delivery, pp. 347354, Vol.23, No.1, January 2008.
[4] D. L. Brooks, R. C. Dugan, Marek Waclawiak, Ashok Sundaram,
Indices for Assessing Utility Distribution System RMS Variation
Performance, IEEE Trans. Power Delivery, vol.13, no.1, pp.254-259,
Jan. 1998.
[5] Juan A. Martinez, Jacinto Martin-Arnedo, Voltage Sag Studies in
Distribution Networks - Part II: Voltage Sag Assessment, Part III Voltage Sag Index Calculation, IEEE Trans. Power Delivery, pp. 16791697, Vol. 21, No. 3, July 2006.
[6] Bach Quoc Khanh, Prediction of Voltage Sags in Distribution Systems
With Regard to Tripping Time of Protective Devices, Proceeding,
EEE.CR.ASPES2009, Tech. Section 2.1., Hua Hin, Thailand, September
28-29, 2009.
[7] J.V.Milanovic, M.T.Aung and C.P.Gupta, The Influence of Fault
Distribution on Stochastic Prediction of Voltage Sags, IEEE Trans.
Power Delivery, vol.20, no.1, pp.278-285, Jan. 2005.
[8] P. Saninta, S. Premrudeepreechacharn Assessment and prediction of
voltage sag in transmission system in northern area of Thailand,
Proceeding, 13th International Conference Harmonics and Quality of
Power, ICHQP, Sept.28-Oct.1 2008, Wollongong, NSW, Australia.
[9] E Inan, B. Alboyaci, C. Leth Bak, A Case Study Of Turkish
Transmission System For Voltage Dips, The Journal on Power and
Energy Engineering, Vol 1, No. 2, April 2010
[10] National Institute of Vietnam, Master Plan VI, 2006.
VII. BIOGRAPHIES
Bach Quoc Khanh received B.S., M.S.
and Ph.D. degrees in power network and
systems from Hanoi University of Science
and Technology, Hanoi, Vietnam in 1994,
1997 and 2001 respectively. He has been a
faulty member of Electric Power System
dept., Electricity Faculty, Hanoi University
of Science and Technology since 2002. He
is currently a visiting scholar in the Center
for Advanced Power System, IESES,
Florida State University. His research
interests include power system analysis,
DSM, power system quality, distributed
generation.
Nguyen Hong Phuc received BS in Electrical Engineering Faculty,
University of Thai Nguyen in 2006, Vietnam. He is currently a master student
in the Electric Power Systems Department, Electricity Faculty, Hanoi
University of Science and Technology, Hanoi, Vietnam.
V. ACKNOWLEDGEMENT
It is acknowlegded that this work received helps from
Phung The Anh (Msc), Nguyen Anh Tu (Msc) with Power
Engineering Consulting Joint-Stock Company No. 1, Hanoi,
Vietnam for data collection and power system simulation as
well as technical consultancy from Professor David A. Cartes,
Senior Member, IEEE, and Dr. Bhuvaneswari Ramachandran
with the Center for Advanced Power Systems, Institute for
Energy Systems, Economics and Sustainability, Florida State
University, USA.
VI. REFERENCES
[1]
[2]
106
&&&&"#'- $-$!'- - # &'% )' $#- *&'"- % #- - #%( #'*$- #- &- &'% )'- #%' $#-
*32+328 +*5-8 +-*38 68 5$)8 08 08 $08 8 3)2$*)8 * 8 2#8
5$)8 0+8 2#28 $08 02*#02$8 $)8 )23-8 68 30$)"8 2#8 +-*$'$02$8
(*'$)"8 * 8 5$)8 0+8 28 2#8 8 $)02''2$*)8 0$28 2#8
3)-2$)268 * 8 5$)8 -"$(8 $08 3$'28 $)2*8 2#8 +-*'(8 * 8 02*#02$8
+-$2$*)8 * 8 4*'2"8 0"8 $)8$02-$32$*)80602(8 5$2#8 5$)8 )-"68
-0*3-08*))28 #8 ++-8 000008 4*'2"8 0"8 +- *-()8
$)8
8 EH *08 028 -8 8 *))28 )25*-&8 (*'$)"8
)8 3'28 )'60$08 -8 *)8 30$)"8
8 M&3& *'2"8 0"8
#-2-$02$08 )8 -,3)68 -8 '0*8 2-($)8 $)8 -'2$*)8 2*8
5$)8 0+8 *-8 $!-)28 0)-$*08 * 8 8 0$7$)"8 )8 +'()28
26+08 * 8 5$)8 +*5-8 ")-2$*)8 )8 2#8 $(+28 * 8 2#8 0602(8
0*3-8 +$268* 82#8 $02-$32$*)80602(8* 8$)2-028
. .4=;6&?#=.64
xlFqNvqqRNRPv`R{FqmddvqqRv{j vF\{Rmv{R
PRNRq{FldRP Rm d` mFllR{ ^RqR{Fdq^ qd qFmRl
Pd{dLRP ^RqR{Fdvq &* NvqqRNRP v Pd{dLdvq
qRv{jv{RRqPd{RNlv`RxvR{Nvqmxdvq "RdPRF
F{dR vT RN`qdNFl FqP RNvqvmdN LRqRX v dlddR RqP
R{ FqP `R Rqd{vqmRq `R dqFllFdvq vT &* dq xvR{
RmFlv{FdRN`FllRq^RvdlddR dqF{dvFxRNvq
qRv{jxlFqqdq^PRd^qFqPvxR{Fdvq=`R&*N`FllRq^Rv
qRv{j vxR{Fdvq F{R PR v `R TFN `F `R Pd{dLdvq
Rmd{FPddvqFllPRd^qRPd`dq^lRRqR{^v{NRFqP
`Rx{RRqNRvT&* N`Fq^RRmT{RzRqNxvR{Yvdq
Rd`R{ RFP FR v{ {FqdRq FR !mvq^ qmLR{ vT
vxR{FdvqFl dR NFRP L &* dqvlRmRq dmxFN vq
xvR{zFld`FRLRRqvT^{RFNvqNR}vdlddR
:GA$?/>BN @'$'/J'%N '"?G ?MN
0'B: 7N %H$!B0<:N <H:& B0<:N (<?N B,'N 0A/B0:*N $,<4!?N ?<*@ 7N @4:4@4dd$
$,N G<$N , :,N /AN L/B-N 5'$D?/$N <L'?N MAE'8AN '> ?B9':BN 5'$D?/$ 5N
:*/:''@1:*N $G6BMN :</N :1J'@A1FMN <)N$0':$'N :%N '$-:<5<*MN
0'BN % N !:</N /'B:!7N '7!04N
XSU;IOSS5SSQQ-;;-AQ(55A5;;
B
mn H thng in - i hc Bch Khoah___
H Ni
e /N =N
107
r
"
8nj<Tld Ftcdz=444Tdz|8ddcdxn|lZ1:zrtcdqdctsKRR*14KTN'4
B89d`{RR x`FRRlRmRq ddvqlxlFNRPvq`RmFdq
TRRPR{ FqP `R `{RR x`FR lFR{Fl &9 lvNFdvq d qvuFll
PRRudqRP LFdq^ vq RR{Fl vxdmFl xvR{ Yv xlFqqdq^
TvulFdvq DE ,vRR{ Tv{ `R x{xvR vT `d v{j
B89 d RlRNdRl lvNFRP F qvPR v{ F
NFR PdR 4vPR TRF{R `R &9 qRF{ Rm v{NR
qvPR FqP F{Rdq`RmdPPlR vT mFdqTRRPR{FqPqvPR
ddqFa{RR x`FRlFR{FlRqP &9dNvqqRNRPvmRPdm
vlF^R TRRPR{ `{v^` Pd{dLdvq {FqTvuR{ (v{ qvPR
`d xFxR{ mvPdXR `R dR vT C3( {FqTvuR{ F
j@!Tv{NvqdPR{ds^`R&9d`xRqR{Fdvq
4vFPF`{RRxRvTB89F{RRlljqvq DE DEF
XRP xRRP dqPNdvq mFN`dqR F{dFLlR xRRP PvLl TRP
dqPNdvq mFN`dqR FqP F{dFLlR xRRP qNb{vqv ^RqR{Fv{
&RFdlRP mvPRl vT `RR `{RR xR vT dqP xvR{ {Rv{NR
`FR LRRq x{RRqRP dq F{dv xLldNFdvq ,vRR{ d`
{R^F{P v &9 {RxvqR v TFl dq Pd{dLdvq Rm `R
`v{ Nd{Nd xvR{ dqhRNRP L &9 dq `R qRv{j `Rq F
TFlvNN{mFdqlPRxRqPvq`RxRvTmFN`dqRPRd^qRP
Tv{ B89 .mxFN vT xvR{ RlRN{vqdN TFNdlddR dq `R
dqR{NvqqRNdvqLRRRqB79FqP`RqRv{jF{RqR^ld^dLlR
=`R{RTv{R `dxFxR{R`RmvPRlvTqN`{vqvmFN`dqR
FqPdqPNdvqmFN`dqR FFdlFLlRdq:<<!&':=FdmxldXRP
mvPRlvTB89xR
B mn H thng in - i hc Bch Khoa H Ni
... EG14C?F:9*UCT6W4FTB/>-&9VF-4OT*=FT\
J 56>9>7J (35>1C9?DJ?CJ(IDE5=J(8?CE9C3G9EJ%?H5CJ!>J
9DEC92GE9?>J(IDE5=J -9E8J -9>4J1D54J
108
8njOdq_|ntsad|ddsnsczvddcY _scZHIzt|v|vtdxL
$% ,7 ,
57,
5
37
*%7 &7
"7
&7
`R{R
#7 ;FRPvxxvR{vT B&*
# dqdqPxRRP4v{mFll d o
BdqP xRRP F `dN` `R B&* vx xvR{ {RFN`R
{FRPFlR=`dFlRv[RqTFlldq`R{Fq^RvT o
x~Y # vdqPxRRP=`dFlRNFqLR o
1@J 1J FqPIF{RNvqFq`FF{RNFlNlFRPFTvllv DE
,17
,7
,7
7
7
7
7
7
)
0-7
`R{R
/77 <`FxRdqPR J7 <NFlRfqPR
=`R {RldFLdld vT BRdLll xPT dq dqP {R^dmR FqFld
109
-
;R^F{Pdq^xR vTB&* v xR
dqPNdvqmFN`dqR
FqPqN`{vqvmFN`dqRd``RFmR{Fdq^ jB Tll
lvFPRPF{RRRPTv{NvqqRNdvqv`RqvPR
"Fdq^ vq `R FLvR x{vxvRP NRqF{dv vlF^R F^
N`F{FNR{ddN F qvPR d NvmxRP d` `R `v{ Nd{Nd
dFxxldRPFqvPR !l`v^`qvPRd qvFlvFPqvPR
d d RRq F F N{ddNFl qvPR `Rq NvtsRNRP d` B&* FqP
vlF^RF^NFqLRNvqdPR{RPF`dlvNFdvq !llxvdLdlddR
vT TFl xR <dq^lR ldqR v ^{vqP lfqR v ldqR v ldqR v
^{vqPFqP `{RR x`FR v ^{vqP F{RNvqdPR{RP . d Flv
qvdNRFLlR `F Tv{ `R x{xvR vT `d v{j vlF^R F^
xR{Tv{mFqNR d vql PRR{mdqRP LFdq^ vq `R RRq dqPR
=`R Rm dqPR NFq LR vLFdqRP F[R{ fmdlF{ RRq dqPR
NFlNlFdvq F{R mFPR Tv{ Fll TFl RRq dq `R Rm vT
dqR{RFqPdRdqPdNRvTFlllvFPqvPRF{RdqR^{FRP
J
a
cw
` aw
y
b
^_
m
o{
??T2
82
7W22
8L>2
8 2
RP2
DT2
2
cd+
z
hi
v
ef+x
8nj Tt|_qf_q|x_|d_sc[1:zt|v|vtdx_j_nsz|nsczvddc
)7
HP
8nk C! [enaqqvcf
`
4 `xnec@ O`qenklcn{|xna|nus
a
4 #4`xnec
B`S{S
5A TJl ZSzSqO wT `S TJl SSq 9J `J wOO{ `Sq
dqPxSSPJB&*lwOJdwqSzFl z
5 =wFlTFl{FSFFTFlxwddvqF#J
J
'7 :S{NSqF^S wT `S TFl xS Nw{{SxwqPdq^ w `S TFl
SSq 9 J ;STS{{dq_w /0J `dw{jFmS`SxS{NSqF^SwT
`STFlxSdq`SSSmdJTwllw
<dq_lSx`FSw_{wqP:4
J
=wx`JSw_{wqP::4
J
=wx`FSw^S`S{::
J
=`{SSx`FSv^{wqP:4
J
i@ BSdMllxPTF`SdqPxSSP y
)d^ J dll{FSFNFSPdq`dN``SPd{dMdwqwTTFl
{JSwT`STJlSSqJqwPS J J_JdqdqPxSSPJqPd
{Sldq_ wx xwS{ wT B&* lwOJSP J qwPS
J dq_
BSdMll Pd{dMdwq d` ;J JqP $ =`S xwS{ O{S wT
`SdqP{MdqS J dOwqdPS{SPd` %&
:.J ,=DJ
,&=DJ JqP ,A=DJ "Jdq^wq`S{SlJdwqMSSSqTFl
Pd|dMdwq JqP dqP xSSP JqP `S xwS{ O|S wT dqP
{MdqS SNJqd`P{J`S {SlJdwqMSSSqB&* wx
xwS{JqP`SNw{{SxwqPdq_TJl{FS)d_
6T Ow{S `S JMwS wJl TJl {JS Pd|dMdwq J_Jdq
B&* wx xwS{ dll MS M{wjSq Pwq \|`S{ Tw{
dqPddPFl TJl xS FNNw{Pdq_ w (7 Tw{ wMJdqdq^ `S
wOO{|SqOS{FSwTdqPddPFlTFlSSq
B mn H thng in - i hc Bch Khoa H Ni
[
XYZ
%
%
n.+7
"Jdq^wqdqPxSSPxPT SNJqmwPSl`SPd{dMdwqwT
wNN{{SqNS {FS wT M SSq wT TJl dq Pd|dMdwq Sm
d` B&* .qSJP wT `S JmS Sm `w{Nd{Od xwS{
JxxldSPwJTJlSSqd`J_dSqTJl{JSJJ{dJMlSdqP
xSSP PSxSqPSP Sm `w|Nd{Nd xwS{ d JxxldSP JqP `S
TJl{FSwT`STFlSSq`wlPMSPd{dMSPT{`S{dqldqS
d` dqP xSSP x{wMFMdld Pd|dMdwq mwPSl BSdMll
mwPSl =`S{STw{S {S_F{Pdq_`S TFl SSq FqwPS J Tw{
dqFqNS dT`SwFlTFl{JSFd_qSPw `FTFlxwddwqd
FmSPwMS
J +Tw{FqFld `STFl{JSPd{dMSP
wFM SSqwTTFl`FwOO{`SqB&*wxxwS{
d x{wPNSP F F xSNdXN dqP xSSP `wlP MS NFlNlFSP M
FTwllw
C3J
,r
j 6V ;
<
>
-EM
8nj"Wtq|_jdz_j_|stcd%!ftx
cngdxds|f_q||vdz_|stcd% "
GO
8nj4 Wtq|_jdz_j_|stcd%!ftx
cngdxds|ZH:zndz_|stcd%!
(d^$`v`RvlF^RF^FqvPRPRv`{RRx`FR
v ^{vqP TFl F qvPR Tv{ NFR vT PdTTR{Rq B89
qvmdqFl dR PdWR{Rq /0 xRqR{Fdvq NvtqRNRP Flv F
qvPR (d^ $^^R{vq^R{dmxFNvTB89dR`Fq
`R F{dFdvq vT B89 vx xvR{ =`F d LRNFR B89
dRdmmRPdFRldmxFNvqB89{FqdRqFqPL {FqdRq
dmxRPFqNR`F`d^`ldqYRqNRvq`v{ Nd{NdN{{RqFqP
R{mdqFlvlF^RQvx`Rq`v{ Nf{NdvNN{
B3K
67
[
DL
7
Z
?
8
@
i 6K
+-
J
qQ
: -FN
8nj%Wtq|_jdz_j_|stcd%!
ftxcnffdxds|ZH:qtb_|ntsz
8nj&Wtq|_jdz_j_|stcd%!ftx
cnffdxds|zz|drztxbdb_v_bn|
u} q}k}a .tl
8nj8xdwdsbzvdb|xrtftq|_jdz_j_|stcd%!
n|l|ldnscxdjnrdrtcdqdcaO_qdnjlvcf
4
(d^ FqP (d^ `v `R dqR^{Fdvq vT vlF^R F^
xR{Tv{mFqNR PR v Fll TFl xR FqP B89 vx xvR{
FlRNv{{RxvqPdq^ v `R`vlR{Fq^R vTB89vxR{Fdq^
dqPxRRPTv{`RBRdLllxPTd`PdTTR{RqdqPdNR /7FqP J
.q Lv` (d^{R `R {Fq^R vT mF^qdPR lvR{ `Fq x
FNNvq Tv{ vlF^RF^PRv`RTFl vNN{`RqdqP
{LdqRdNvLRNFRvTRd`R{R{lvv{R{`d^`dqP
xRRPvdPR`RN dqvN vdqPxRRP{Fq^R@vlF^R
F^ZvmxvxF{RPRvFllTFlxRRNRxdq^
dq^lRx`FRv^{vqPTFl @vlF^RF^`d^`R{`Fqx
F{R `R {Rl vT dq^lR x`FR v ^{vqP TFl =`R ZRzRqN
vT vlF^R F^ dq `d {Fq^R d zdR `d^` LRNFR `R {FR vT
dq^lR x`FRv^{vqP TFl FNNvq Tv{vTvFlTFl{FR
.q (d^ `R BRdLll xPT d NvqdPR{RP Tv{ k ;FlRd^`
111
J
'
'
(
#
(
8nj<<Wtq|_jdz_jfxdwdsb_|stcd%!n|lZ1:ns|ldnscxdjnrdz
rtcdqdcaO_qdnjlvcfn|lcngdxds|_qdztf|ldzb_qdnscd
%+ +
$"'*+
"
"
&
o
%+ +
(+ +
+
8njWtq|_jdz_jidwdsb_|stcd%!n|lZ1:ns|ldnscxdjnrdz
rtcdqdcaZdnaqqvcf.(# n|lcngdxds|_qdztf |ldzl_vdnscd
X/GF-CUR=GF
=`dxFxR{x{RRqFqRFRmRqvTvlF^RF^PRv
TFl dq Pd{dLdvq Rm d` B89NvqqRNRP ;FqPvm
F{dFdvq vT dqP xRRP mFjR N`Fq^R dq B89 vx
xvR{ dqhRNRP v `R qRv{j `Rq TFl vNN{ FqP `R
vlF^R F^ xR{Tv{mFqNR d N`Fq^RP FNNv{Pdq^l =`R xFxR{
LdlP `R vN`FdN mvPRldq^ vT `R qNR{Fdq vT dqP
{R^dmR dq PRXqdq^ TFl RRq FqP TFl Pd{dLdvq
mvPRldq^ vT `R x{vLlRm vT vN`FdN x{RPdNdvq vT vlF^R
F^ =`R{RTv{R `R vlF^R F^ xR{Tv{mFqNR d FqFlRP dq
{RlFdvqd`dqP{R^dmRF&9lvNFdvq&dTTR{RqNRqF{dv
vT B89 FxxldNFdvq F{R NvqdPR{RP =`R {Rl ^^R `F
N`Fq^RdqvlF^RF^N`F{FNR{ddNPRvdN`dq^vqFqP
vTT B89 F{R `FxR{ `Fq PR v N`Fq^R dq dqP xRRP
PRxRqPRP vx xvR{ vT B89 BdqP {R^dmR d` `d^`R{
FR{F^R dqP xRRP dll ^dR mv{R dmx{vRmRq Tv{ vlF^R
F^ =`R{RRF{N`{RlF{RLhRNRPv`R{RlFdRlldmdRP
NFxFLdld Tv{ B&9 mvPRldq^ dq :<<!8':= FqP `R
B mn H thng in - i hc Bch Khoa H Ni
112
P584O4F-4R
V. V. Thang
Keyword:
Distributed Generator (DG), Planning of
Distribution System (DS), Competitive Electricity Markets (CEM)
I. INTRODUCTION
Recently, the issue of planning of DS has gained
remarkable achievements. Because of the construction of
competitive electricity market (CEM), technological
development and environmental pollutions, the development of
DG and renewable energy resources, in particular, is fostered
[1].
DGs connected directly to DSs or supplied straight to
customers [2], normally use new electric generating
technologies such as gas turbines, Combined Heat and Power
(CHP), Fuel Cells, solar energy, geothermal and wind energies.
The power of DGs can reach to 300MW depending on
particular technologies, but the power in use of DGs is
normally less than 5MW. The DG is installed close to loads so
that it gets some main advantages including the elimination of
transmission and distribution cost, enhancement of flexibility
and reliability of distribution systems, reduction of power loss,
and improvement of differential voltage at nodes as well as
reduction of environmental pollution because of using
renewable energy resources [2]. However, DGs require high
B mn H thng in - i hc Bch
Khoa IEEE
H Ni
978-1-4577-0365-2/11/$26.00
2011
631
113
J =
t
NS
+ ( C SF + C SC .SiS,t ) +
i =1
NS
+ Tmax .
NDG KDG
CkDG .SiDG
,k ,t
i =1 k =1
(1)
( PS .Pi S,t
QS .QiS,t
i =1
NDG KDG
+ Tmax .
i =1 k =1
( kPDG .Pi DG
,k ,t
kQDG .QiDG
,k ,t
) Min
1
calculated total cost at base year;
( 1 + r )t
Components in 1 are upgrading cost of feeders for year t with
fixed capital cost (CFF) and variable capital cost (CFC).
substation transformers upgrading costs in year t with fixed
capital cost (CSF) and variable capital cost (CSC) in 2 . 3 are
new investment costs of DGs at node i, year t with DG
technologies k. Electrical energy purchased cost from CEM in
4 and 5 are O&M and fuel costs of DG depending per
technology.
TABLE I. SETS, INDICES, VARIABLES AND PARAMETERS
No
Symbol
Definition
I. Sets and Indices
1
N
Set of buses in distribution system
2
i, j
Bus (i, j N)
3
NL
Set of load buses in distribution system
4
NS
Set of substation buses in distribution system
5
NDG
Set of DG buses in distribution system
6
T
Overall planning period, year
7
t
Planning period (t T)
8
KDG
Total number of DG's technology
9
k
Technology of DG (k KDG)
II. Variables
Active power purchased from CEM at node i, for year t
10
PSi,t
(MW)
Reactive power purchased from CEM at node i, for year t
11
QSi,t
(MW)
F
12
S i,j,t
Upgrading capacity of Feeder ij for year t (MVA)
13
SSi,t
Upgrading capacity for Substation i, at year t (MVA)
New investment capacity of DG node i for technology k,
14
SDGi,k,t
at year t (MVA)
Active power of DG node i, for technology k, at year t
15
PDGi,k,t
(MW)
Reactive power of DG node i, for technology k, at year t
DG
16
Q i,k,t
(MVAr)
17
Ui,t
Voltage for node i, at year t (pu)
Voltage angle at bus i, for year t (radian)
18
i,t
III. Parameters
19
r
discount rate (%)
20
CFF
Fixed capital cost of Feeder ($/km)
21
CFC
Variable capital cost of Feeder ($/MVA.km)
22
Li,j
Length of Feeder ij (km)
23
Yi,j
Magnitude of admittance matrix element (1/)
II.
632
114
Where:
1
N N
1
FF
FC
F
.
(
C
+
C
.
S
)
L
i , j ,t
i, j
( 1 + r )t i =1 j =1
TABLE I. Continued
Symbol
i,j
CSF
CSC
CDGk
SP
SQ
30
kPDG
31
kQDG
32
33
PDi,t
QDi,t
34
PDGmax,k
35
QDGmax,k
36
37
Umax
Umin
38
39
40
41
fSL
U min U i ,t U max
U i ,t = cons tan
DG
DG
DG
Pi DG
,k ,t = Pi ,k ,t 1 + P ,Qi ,k ,t = Qi ,k ,t 1 + Q
(6)
(7)
(2)
SiS,t f SL .SiS,1
j =1
i , j N
Constraint of nodal power balance in (2) is only used for
nodal loads, when DSs planning considers the use of DG with
different time and technologies, the previous formula is
rewritten as (3 and 4).
t 1, i NS
(8)
SiS,t SiS,t 1 + S t 1, i NS
(9)
S
,k ,t + Pi ,t PDi ,t =
PiDG
Yij .Ui,t .U j,t .cos(ij j,t i,t )
(3)
PiDG
,k ,t 0,i N,t T ,k KDG
(5)
i NS
DG
DG
DG
Pi DG
,k ,t Pmax,k , Qi ,k ,t Qmax,k i NDG , k KDG
i, j N,t T ,k KDG
i NL
j =1
j =1
(4)
k =1
j =1
k =1
QiDG
,k ,t 0,i N ,t T ,k KDG
KDG
i, j N,t T , k KDG
KDG
S
,k ,t + Qi ,t QDi ,t = Yij .Ui ,t .U j ,t . sin(ij j ,t i ,t )
QiDG
Definition
Angles of admittance matrix elements (radian)
Fixed capital cost of Substation ($)
Variable capital cost of Substation ($/MVA)
New investment cost for DG technology k ($/MW)
Active power purchased cost from CEM ($/MWh)
Reactive power purchased cost from CEM ($/MVArh)
O&M cost and Fuel cost of DG for active energy
($/MWh)
O&M cost and Fuel cost of DG for reactive energy
($/MVArh)
Active power demand at bus i, for year t (MW)
Reactive power demand at bus i, for year t (MVAr)
Maximum DG capacity limit for active power with
technology k (MW)
Maximum DG capacity limit for reactive power with
technology k (MVAr)
Maximum voltage limit at bus (pu)
Minimum voltage limit at bus (pu)
Active power ramp-up limit for DG in planning year
(MW)
Reactive power ramp-up limit for DG in planning year
(MVAr)
Capacity ramp-up limit for Substation transformer in
planning year (MVA)
Load factor of Substation transformer base year
633
115
No
24
25
26
27
28
29
No
1
2
22
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
23
24
25
26
27
28
29
30
31
32
DG technology
Gas turbine
Solar PV
Capital cost
(M$/MW)
0.5
5
O&M cost
($/MWh)
10
1
Fuel cost
($/MWh)
80
0
33
1) Assumptions in analyis
This research utilizes some economic and technical
assumptions for the ease of computation:
Planning period is 5 years. Annual developing rate of load
demand is constant at 5% during planning time. Hence, the
total increased load demand is 25%.
a) Case A
This case aims to estimate the working capacity of current
feeders and substations as load demands in future.
Furthermore, upgrading time and capacity of them will be
decided to ensure requirements within planning period.
634
116
19
20
21
Substation
01
Resource
Substation
Feeder
Substation Transformer
upgrading capacity in each
year (MVA)
1
2
3
4
5
16
-
Capacity
in
base year
(MVA)
Total capacity of
Substation
Transformer
(MVA)
16
32
1
2
3
4
5
6
7
Capacity in
base year
(MVA)
No
Feeder
1-2
2-3
3-4
4-5
5-6
8-9
9-10
16.96
16.96
12.57
12.57
12.57
4.95
4.95
- 240 - 185
150 - 150 - 120 35
-
* Where: Cca - Capital cost of DG; Cpo, Cqo - O&M and Fuel cost of DG
1
2
3
Cost
Substation Transformer
upgrading
Feeder upgrading
O&M and Electrical
energy
Total
2
-
3
-
4
-
5
-
0.25
0.51
0.48
0.5
0.24
8.57
9.06
9.55
10.06 10.58
Total
cost
(M$)
1.00
1.2
2.0
0.6
3.8
0.0
0.4
0.2
1.2
1.8
0.0 0.0
Capacity in
Total capacity
base year
of DG (MVA)
(MVA)
0
0
0
0
0
0
0.4
0.2
2.4
2.0
0.6
5.6
1.98
47.82
50.80
18
33
18
22
33
DG capacity invested in
each year (MW)
No
Cca
Cpo
Cqo
DG
Pmin Pmax
Qmin
Qmax
(M$/
($/
($/
technology (MW) (MW) (MVAr) (MVAr)
MW) MWh) MVArh)
Solar PV
0
2
0
0
5
1
0
Gas turbine
0
2
0
1.5
0.5
90
10
635
117
No
Substation
Transformer
1
2
Loi thit b
Feeder, i-j
Substation
Transformer, 1
Cost
Total
cost
(M$)
2
3
Substation
Transformer
upgrading
Feeder upgrading
Investment DG
1.9
0.0
3.6
0.0
0.0
0
5.5
Total
80.00
70.00
Electrical Energy (10e6kWh)
TT
40.74
46.24
60.00
50.00
40.00
30.00
20.00
10.00
0.00
1
-10.00
-20.00
Case A
2
Case B
5
Planning Period (years)
3
2
Comparison cost
between Case B
Case B
and Case A
1
2
3
4
Cost
Substation
Transformer upgrading
Feeder upgrading
O&M and Electrical
energy
Investment DG
Total
Case A
1.00
0.00
-1
1.98
0.00
-1.98
47.82
40.74
-7.08
0.00
50.8
5.5
46.24
5.5
-4.56
Note
Total cots is
reduced -8.99%
TT
1
0
1
-1
-2
-3
Case A
Case B
636
118
No
20
15
10
1.1
1.05
Voltage (p.u)
5
0
-5
11
13
15
17
19
21
23
25
27
29
31
33
-10
-15
0.95
-20
0.9
-25
Case A
Case B
Bus
0.85
0.8
0.75
0.7
1
5
7
Year 1
11 13
Year 2
15 17
Year 3
19 21
Year 4
23
25 27
Year 5
29
31
33
Bus
0.9
0.85
0.8
0.75
0.7
1
Year 1
11
13
Year 2
15
17
19
Year 3
21
23
Year 4
25
27
Year 5
29
31
33
Bus
Voltage (p.u)
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
1
2
Case A
Case B
3
Comparison Case A and Case B
5
Planning Period (years)
637
IV. CONCLUSION
Planning of distribution systems have been change
significantly for recent years because of reconstruction process
of CEM, technological development and environmental
pollutions, promoted the development of DGs and renewable
energy resources, in particular. The new proposed model for
planning of distribution systems has allowed selecting DGs as
an optimum. The objective function is minimum quantity of
total costs of investing and upgrading costs of feeders and
substation transformers, building cost of new DGs, operating
fees as well as fuel cost of DGs and electricity energy expenses
purchase from CEM via connected substation transformers. All
these costs are converted to the first year of planning period.
Nodal power balance between supplies and load demands, DG
power limitation, efficiency of existing equipments, required
nodal voltage drop are constraints of this research model so
that they can guarantee technical requirements. An optimal
process can be estimated (time and power) to upgrade and
reconstruct feeders and substation transformers, and choose
optimal locations, power and investing process of DGs.
Particularly, DG technology can be selected in this model via
economic and technical indices. It can be seen from the results
that planning together with using DG usually provides better
economic and technical outcomes. Total investing and
operating costs of planning of distribution system, converted to
the base year, have been reduced, active power and electricity
energy losses have been decreased, voltage profiles have been
supported and upgrading time of existing feeders and
substation transformer has been delayed. Furthermore,
reduction of electricity energy purchased CEM has limited the
use of traditional energy resources, which contributes to the
decrease of environmental pollution.
119
Voltage (p.u)
1
0.95
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
QD0
(kVAr)
160
240
250
130
310
300
300
120
160
110
135
435
180
460
200
320
No
Bus
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Total
PD0
(kW)
490
190
290
190
390
390
420
220
660
560
360
420
300
550
310
460
10,675
27
28
29
30
31
32
QD0
(kVAr)
440
140
220
140
340
350
350
200
525
525
310
370
200
470
260
390
9,040
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
1.2
2.3
3.4
4.5
5.6
6.7
7.8
8.9
9.10
10.11
11.12
12.13
13.14
14.15
15.16
16.17
17.18
2.19
19.20
20.21
21.22
3.23
23.24
24.25
6.26
26.27
Fij
(mm2)
AC-150
AC-150
AC-95
AC-95
AC-95
AC-35
AC-35
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
AC-35
AC-35
Smax.ij
(MVA)
16.96
16.96
12.57
12.57
12.57
6.67
6.67
4.95
4.95
4.95
4.95
4.95
4.95
4.95
4.95
4.95
4.95
4.95
4.95
4.95
4.95
4.95
4.95
4.95
6.67
6.67
Lij
(km)
1.5
0.6
1.6
1.5
2.1
2.5
1.1
1.6
1.5
0.2
0.4
1.2
1.4
1.3
1.6
1.2
0.8
0.2
1.3
0.3
0.7
1.4
0.8
0.8
1.5
2.2
Rfij
()
0.2910
0.1164
0.5024
0.4710
0.6594
1.9325
0.8503
1.8336
1.7190
0.2292
0.4584
1.3752
1.6044
1.4898
1.8336
1.3752
0.9168
0.2292
1.4898
0.3438
0.8022
1.6044
0.9168
0.9168
1.1595
1.7006
4.95
4.95
4.95
4.95
4.95
4.95
1.7
1.8
1.3
2
2.2
1.4
1.9482
2.0628
1.4898
2.2920
2.5212
1.6044
0.7310
0.7740
0.5590
0.8600
0.9460
0.6020
REFERENCES
[1]
[2]
[3]
[4]
[5]
AC-25
AC-25
AC-25
AC-25
AC-25
AC-25
* Where: PD0, QD0 - active and reactive power demand at bus in base year of planning period
No
27.28
28.29
29.30
30.31
31.32
32.33
[6]
Xfij
()
0.5760
0.2304
0.6352
0.5955
0.8337
1.0725
0.4719
0.6880
0.6450
0.0860
0.1720
0.5160
0.6020
0.5590
0.6880
0.5160
0.3440
0.0860
0.5590
0.1290
0.3010
0.6020
0.3440
0.3440
0.6435
0.9438
[7]
[8]
[9]
[10]
[11]
[12]
[13]
638
120
No
PD0
(kW)
200
290
320
160
360
300
300
160
220
145
160
460
220
560
260
360
I. INTRODUCTION
Peak
Clipping
Valley
Filling
Strategic
Load Growth
DEMAND-SIDE
MANAGEMENT
Load
Shifting
Flexible
Load Shape
978-1-4577-1002-5/11/$26.00
2011
IEEE
B
mn H thng in - i hc Bch Khoa
H Ni
121
B. Network modeling
For analyzing DSMs impacts on energy losses in
distribution system, this research chooses the IEEE 34
Nodes Test Feeder as the test distribution system. Among
radial distribution test feeders proposed in IEEE Distribution
System Analysis Subcommittee Report [3], IEEE 34 Nodes
Test Feeder is characterized by long feeder with the highest
rate of power loss (13,37%) meanwhile it also displays a
wide variety of components and topological features
(multiple single-phase and three phase laterals, transformers,
voltage regulators, capacitor banks). In this work, the test
system is modeled by Siemens Power System Simulator
Advanced Distribution Engineering Productivity Tool
(PSS/ADEPT 5.0) as plotted in Fig. 2. This is an effective
tool for distribution system modeling. PSS/ADEPT 5.0 has
several useful functions for analyzing power system and
reporting results. In load flow analysis, reports on branch
power loss and total system power loss are included. That
helps fasten analyzing process concerning energy loss as
used for this research. Some remarks for this network
modeling are as follows
- PSS/ADEPT only accepts sequence values for threephase line modeling, so the inherent unequal mutual
coupling in the phase impedance matrices for distribution
lines [2] is not considered. Normally, its only accepted for
the transposed lines. Since distribution lines are rarely if
ever transposed, mutual coupling between phases is not
equal, so it may cause errors. However, because coupling
inductances are much smaller than self inductances, the
error should be negligible.
- Distributed load lines are modeled using lumping load
modeling in which one-third of the load should be placed at
the end of the line branch and two-thirds of load placed at
one forth of the way from the source end [2]. Therefore, a
number of fictional load nodes labeling A are additionally
generated in the test system. For example, the line branch
802-806 is load distributed, and it is modeled by two
lumping loads. One-third of the load is allocated to node 806
and other two-thirds of load is placed at an additional load
node namely 806A at one forth of branch length from source
end (Node 802).
- The source transformer labeled XMF-1 is added in the
test system for power loss calculation. In order to have the
122
P(MW)
P(MW)
800
800
600
600
400
400
200
200
6
12
18
24 t(h)
a) Load Curve
12
18
24 t(h)
t(h)
Pr
24
18
0.75
12
0.5
P(MW)
200
400
600
c) Inverted LDC
800
0.25
P(MW)
200
400
600
800
d) Normalized ILDC
Pr
1
ILDC
0.75
ELDC
0.5
0.25
800
600
200
400
Fig. 4. ILDC and ELDC
P(MW)
X Pmin
(1)
exp C .
f ( X ) 1
P
P
P
P
max
min
min
max
= 1
= 0
for X > Pmax
where
Pmax: Peak load, Pmin : Base load, X: Load
C is a constant depending on Pmax, Pmin and E
n
E T . Pmin
0.5
max Pmin )
n 1
(2)
where
E: Supplied energy, T: Time period of LDC,
An: Coefficients of the inverse function of the following
function
m
ai . x i
i 1
(i 2)!.x
(3)
i 1
with m = 7 we have
A1 a11
A2 a13 a 2
A3 a15 (2a22 a1a3 )
A4
(4a)
(4b)
(4c)
(4d)
A5
A6
2
1
2 2
1 3
4
2
3
1 5
2
1 2 3
A7
13
1
(4f)
(8a a2 a6 8a a a 4a a 120a a a
4
1
4
1 3 5
4
1
2
4
2
1
3
2 4
(4g)
1
for i=(17) (5)
(1 i )!
Comparative tests on using VPI model and similar models
shown that this model give more exact results in calculating
power system indices using ILDC and ELDC. This paper
proposes the application of VPI model for representing
DSM impacts in system load curve for analytical assessment
of DSM impacts of energy loss in distribution system.
and ai
Qi (t ) Qi .P * (t )
where
Pi, Qi: Load values at
load node i given by
IEEE test system
P*(t): The per-unit
peak demand at the
instant t in LDC (Fig. 6)
A T .( P
'E
48
'P (t ).t
(7)
t 0
In calculating system
power loss, we also build
up the discrete relation
between system power
124
m 1
24.Pmin t ( X i ).G P
EB E X
(10)
i 0
Pmax Pmin
and X i
m
m is an integer.
where G P
B i.G P (i=0,m)
t (h)
24
a) ILDC
EB
EX
P*
Pmax = 1
Pmin
t (h)
24
b) IPLDC
EB
EX
P(kW)
Pmin
Pmax
'EB
'E X
24.'Pmin
(12)
m 1
t ( X )>'P ( X
i
i 1
) 'P ( X i ) @
(13)
i 0
TABLE II
VPI MODELED LDC TESTING FOR ENERGY LOSS CALCULATION
Residential
Industrial
Load
E
E
'E (kWh)
'E (kWh)
Original
14.548
3203.087
17.780
4249.698
VPI modeled
14.628
3166.703
17.840
4194.661
Error (%)
-0.547
1.136
-0.337
1.295
125
loss and system load, P(P*), for the IEEE test system with
unit diversity factor as Fig. 7 and Appendix B. This curve
will be a useful tool for analysis of DSMs impact on
distribution system energy loss in Part B of this section.
In order to test VPI model for loss calculation, it needs to
build VPI modeled ILDC. From data of CLC, VPI modeled
ILDC for industrial and residential loads are given in the
equations (8) and (9).
- For residential loads
X 0.405
X 0.405
t 1
.24 (8)
exp 1.29253.
1 0.405
1 0.405
X 0.485
X 0.485
t 1
.24 (9)
exp 0.01954.
1
0
.
485
1 0.485
load curve shape becomes, the lower the system energy loss
gets. However this effect of energy loss reduction depends
on the shape of LDC without DSM. For the industrial load,
because the original LDC is quite flat, the effect of loss
reduction is lower in comparison with residential load. In
Fig. 10, big errors appear either in the area where deeply
clipped peak goes with small base increase and the opposite
area where slightly clipped peak is paired to high raised base
load. If we also consider cases with GE < 0.5%, the best
possible strategy of DSM load shift is summarized in the
Table IV.
TABLE IV
OPTIMAL STRATEGY OF DSM LOAD SHIFT FOR THE BEST ENERGY LOSS
REDUCTION FOR TYPICAL INDUSTRIAL LOAD
Fig. 10. System energy loss variation and supplied energy error
of VPI model ILDC in load shift modeling for industrial load
0
5
4151
1.035
0.338
5
5
4121
1.749
0.337
10
10
4064
3.119
0.338
15
15
4022
4.114
0.338
20
20
3996
4.732
0.338
25
25
3988
4.936
0.363
G'E
'E 'EDSM
.100[%]
'E
(14)
where 'E and 'EDSM are system energy losses without and
with DSM application respectively.
TABLE III
OPTIMAL STRATEGY OF DSM LOAD SHIFT FOR THE BEST ENERGY LOSS
REDUCTION FOR TYPICAL RESIDENTIAL LOAD
Peak-clip (%)
Base-fill (%)
'E (kWh)
G'E (%)
Error GE(%)
5
0
3142
0.782
0.452
10
2.5
3091
2.41
0.439
15
7.5
3025
4.461
0.489
20
10
2993
5.497
0.471
25
12.5
2967
6.308
0.453
30
15
2948
6.892
0.436
Fig. 12. System energy loss reduction and supplied energy error
of VPI model ILDC in peak-clip modeling for residential load
Peak-clip (%)
Base-fill (%)
'E (kWh)
G'E (%)
Error GE(%)
TABLE V
OPTIMAL STRATEGY OF DSM PEAK-CLIPPING FOR THE BEST ENERGY LOSS
REDUCTION FOR TYPICAL RESIDENTIAL LOAD
5
0
3142
0.782
0.452
10
2.5
3012
4.895
0.44
15
7.5
2792
11.82
0.481
20
10
2682
15.3
0.461
25
12.5
2580
18.54
0.445
30
15
2484
21.55
0.431
Fig. 14. System energy loss reduction and supplied energy error
of VPI model ILDC in peak-clip modeling for industrial load
Peak-clip (%)
Energy saved (%)
'E (kWh)
G'E (%)
GE(%)
5
0
4153
1.002
0.337
10
5
3798
9.468
0.337
15
10
3470
17.27
0.338
20
15
3171
24.41
0.338
25
20
2900
30.86
0.341
Fig. 15. System energy loss reduction and supplied energy error
of VPI model ILDC in valley-fill modeling for residential load
Fig. 16. System energy loss reduction and supplied energy error
of VPI model ILDC in valley-fill modeling for industrial load
TABLE VII
SYSTEM ENERGY LOSS INCREASE IN DSM VALLEY FILLING OBJECTIVE
Residential
Industrial
Max 'E
Min 'E
Max 'E
Min 'E
B/E
B/E
B/E
B/E
G'E
G'E
G'E
G'E
5/30
46.7
5/10
12.99
5/10
16
5/0
-1.035
10/35 55.17 15/20 26.51 10/15 24.67
15/5
5.642
15/40 64.52 25/30 42.73 20/20
33.7
20/10
14.05
30/45 73.17 30/35 51.84
25/15
23.14
Peak-clip (%)
Energy saved (%)
'E (kWh)
G'E (%)
GE(%)
V. CONCLUSIONS
This paper presented a new proposal for analytical
assessment of DSMs impacts on technical energy loss in
distribution system using VPI model. The results show the
suitability of VPI model application for energy loss analysis
because the error in supplied energy always helps find the
set of practicable DSM strategies. The results also show the
quantitative effect of energy loss reduction for DSM peakclipping and load-shifting. That will be an additional
benefice for estimating the performance of DSM programs
in distribution system. The calculation and modeling in this
proposal will give more detailed results if the value of m in
(10) gets greater and more levels of DSM parameter
variation (peak, base, energy) are considered.
[7]
APPENDIX A
[8]
[2]
[3]
[4]
[5]
[6]
Residential Industrial
0.435
0.53
0.43
0.52
0.415
0.505
0.405
0.5
0.405
0.5
0.405
0.5
0.41
0.5
0.42
0.485
0.46
0.495
0.51
0.51
0.555
0.55
0.585
0.6
0.575
0.67
0.55
0.74
0.52
0.83
0.495
0.92
h
8.5
9
9.5
10
10.5
11
11.5
12
12.5
13
13.5
14
14.5
15
15.5
16
Residential Industrial
h
0.5
0.98
16.5
0.54
0.995
17
0.595
1
17.5
0.65
1
18
0.655
0.945
18.5
0.625
0.85
19
0.585
0.815
19.5
0.55
0.8
20
0.52
0.84
20.5
0.485
0.9
21
0.465
0.93
21.5
0.475
0.96
22
0.49
0.975
22.5
0.52
0.985
23
0.58
0.96
23.5
0.69
0.925
24
Residential Industrial
0.8
0.88
0.88
0.825
0.95
0.8
0.98
0.795
1
0.79
0.95
0.78
0.91
0.775
0.88
0.72
0.85
0.685
0.815
0.655
0.755
0.645
0.69
0.645
0.625
0.63
0.56
0.61
0.5
0.58
0.465
0.55
[9]
[10]
[11]
[12]
[13]
APPENDIX B
Discrete System Power Loss vs Demand Relation P(P*) of IEEE 34
Nodes Test Feeder with Unit Diversity Factor modeled by PSS/ADEPT 5.0
P*
1
0.995
0.99
0.985
0.98
0.975
0.97
0.965
0.96
0.955
0.95
0.945
0.94
0.935
0.93
0.925
0.92
0.915
0.91
0.905
0.9
0.895
0.89
0.885
0.88
0.875
0.87
0.865
0.86
0.855
P(kWh)
288.969
286.09
283.23
280.389
277.567
274.764
271.979
269.217
266.471
263.746
261.039
258.352
255.683
253.034
250.405
247.794
245.203
242.631
240.079
237.544
235.031
232.536
230.06
227.604
225.167
222.749
220.351
217.972
215.612
213.272
P*
0.85
0.845
0.84
0.835
0.83
0.825
0.82
0.815
0.81
0.805
0.8
0.795
0.79
0.785
0.78
0.775
0.77
0.765
0.76
0.755
0.75
0.745
0.74
0.735
0.73
0.725
0.72
0.715
0.71
0.705
P(kWh)
210.951
208.649
206.367
204.104
201.861
199.637
197.432
195.247
193.082
190.936
188.809
186.702
184.614
182.546
180.497
178.468
176.459
174.469
172.499
170.548
168.617
166.706
164.814
162.942
161.09
159.257
157.444
155.651
153.878
152.124
P*
0.7
0.695
0.69
0.685
0.68
0.675
0.67
0.665
0.66
0.655
0.65
0.645
0.64
0.635
0.63
0.625
0.62
0.615
0.61
0.605
0.6
0.595
0.59
0.585
0.58
0.575
0.57
0.565
0.56
0.555
P(kWh)
150.39
148.676
146.982
145.308
143.653
142.019
140.404
138.809
137.234
135.68
134.145
132.63
131.136
129.661
128.206
126.771
125.357
123.963
122.588
121.235
119.901
118.587
117.294
116.021
114.768
113.536
112.324
111.133
109.961
108.81
P*
0.55
0.545
0.54
0.535
0.53
0.525
0.52
0.515
0.51
0.505
0.5
0.495
0.49
0.485
0.48
0.475
0.47
0.465
0.46
0.455
0.45
0.445
0.44
0.435
0.43
0.425
0.42
0.415
0.41
0.405
P(kWh)
107.68
106.57
105.48
104.411
103.363
102.335
101.328
100.341
99.375
98.43
97.506
96.602
95.719
94.856
94.015
93.194
92.394
91.616
90.858
90.121
89.405
88.71
88.036
87.383
86.751
86.14
85.551
84.983
84.435
83.91
ACKNOWLEDGMENT
It is acknowledged that this work received technical
supports from Prof. D.A. Cartes, Senior Member, IEEE, and
researchers with the Center for Advanced Power Systems,
Institute for Energy Systems, Economics and Sustainability,
Florida State University (FSU), USA as well as Dinh Cong
Bang (Ph.D candidate, College of Communication &
Information, FSU) for IT supports.
[14]
[15]
[16]
[17]
[18]
128
REFERENCES
[1]
I.
INTRODUCTION
129
322
A. Wind generators
The wind energy is converted to electric energy by wind
turbines and generators. There are two basic configurations of
wind turbines, the horizontal axis and the vertical axis. The
vertical axis turbines are isotropic, receiving wind from any
direction. The gearbox and generator can be mounted at or
near ground so they have higher durability and lower
investment cost than horizontal axis ones. However, some
main disadvantages of this type consisting of more expensive
produce cost and smaller power ratio compared to horizontal
axis counterparts. Therefore, the only horizontal axis wind
turbines has been researched and applied widely on the
industry [18]. The change of wind speed depends strongly on
climatic conditions. These high variations are dependent
geographical location, both spatially and temporally. Hence,
wind generators power output are also naturally intermittent
and highly variable [19]. Wind generators cannot operate
when wind speed is lower than the Vin speed or higher than
the Vout speed. Power output of wind generators is
represented as follows [20]:
0
,Vi ,h < Viin
1
h .PS h
BP =
.
365
.
( 1 + r )t
(5)
t = 1
h =1
t T ,h H
B. Increased cost of wind generators
All costs on model also are calculated by the first time of
planning using discount rate r.
1. Investment cost of BESS
Electric energy can be charged, discharged and stored by
BESS. Hence, investment cost of BESS is related to
130
323
II.
CB =
t =1 ( 1 + r )
. PB max .C PC + EB max .C EC
t T
(6)
COM =
t
t =1 ( 1 + r )
(7)
t T
where, CEOM is increased at operation time and it is
assumed to go up by 20% per year [17].
3. Electricity energy loss cost on BESS
The efficiency factor of BESS is about 65% to 85% [5, 6,
7] so electricity energy loss cost on itself is determined as
follows:
T
H
1
CF =
.
365
.
( 1 k F ).PBh . h (8)
( 1 + r )t
t = 1
h =1
t T ,h H
4. Transmission access fee
The transmission access fees in the competitive electricity
markets rely on power value and accessing times. During the
periods of high electric price, the accessing fees are raised
and vice-versa. Therefore, this cost is increased because
BESS delivers power at high electric price times.
T
H
1
CT =
.
365
.
PS h . F ,h
( 1 + r )t
(9)
t =1
h =1
t T ,h H
5. Upgrading cost of connected substation transformers
The transmission power through substation transformers
can be increased when BESS discharges at high electric price
times. At the same, generator power of wind generators is
also high. Therefore, upgrading cost of connected substation
transformers is calculated as follows.
T
CS =
t =1 ( 1 + r )
PS max S
.C S t T
Cos
(10)
J = BP CB COM CF CT CS
=
h =1
( 1 k F ).PBh . h
(11)
h =1
H
PS h . F ,h
365.
h =1
PS
.C S Max
Cos
max
t T ,h H
131
324
WGs
1
Sub
2
i =1
BESS
(12)
i N ,h H
PBh + k F .PBh = 0
h =1
h =1
h H , H 1 + H 2 = H
3. Constraint of BESS power and capacity limits
With the assumption that operation cycle of BESS is one
day. The BESS power and capacity limits depend on
generator powers characteristics of wind generators.
PGh PBh PGh
EB EB max =
PBh
(14)
h =1
h H
4.
h H
5. Constraints of BESS power limit on planning period
During
planning
duration,
generator
powers
characteristics of wind generators in an operation cycle are
assumed constant. This constraint is given as follows.
PBh ,t = PBh ,t 1 h H ,t T
(16)
The proposed model including equations from (1) to (16)
is a nonlinear programming model. In order to investigate this
model, Nonlinear Programming with Discontinuous
Derivatives (DNLP) solver in GAMS program language [1,
22] is used to find out an optimal solution.
IV.
4.0
3.0
2.0
1.0
0.0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (h)
2000
Power
(kW)
g
( )
H1
1500
1000
500
0
1
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (h)
132
325
Distribution
systems
140
12000
120
10000
100
8000
80
6000
60
4000
40
2000
20
14000
0
1
Demand (MW)
1
2
Power ($/kW)
Capacity ($/kWh)
Investment costs
CP C, CEC
500
100
Operation and
Maintenance costs
CP OM, CEOM
10
0.5
Periods (h)
1-6
7-9
10-22
23-24
Time (h)
Parameters
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
No
Fees ($/kW.month)
2.0
2.17
3.55
2.0
Note
Analyzing results
a) Case A
In this case, power of wind generators is sold out to
market as the characteristics in figure 3. Total electric energy
value that is provided to market on calculation period is
7,700,059$ and transmission access fees are 265,731$.
Therefore, total profits are 7,434,327$ as in table IV.
TABLE IV. TOTAL PROFITS, COSTS FOR CASE A AND CASE B
5.
No
1
2
3
4
Case A
($)
Comparison
cost between
Case B and
Case A ($)
Case B
($)
7,700,059.0
8,508,204.0
808,145.0
265,731.0
273,879.0
-8,148.0
0.0
634,082.0
-634,082.0
0.0
15,261.0
-15,261.0
0.0
0.0
0.0
7,434,327.0
7,584,981.0
150,654.0
133
326
1. Assumptions in analysis
This research utilizes some economic and technical
assumptions for the ease of computation:
2000
120
1500
90
1000
60
500
30
0
1
Power for W Gs
0
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Time (h)
Power to sell on market )
Electric energy price ($/MWh)
6
500
150
400
120
300
90
200
60
100
30
Power (MW)
-100
2 3
4 5
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
-200
-300
-400
-500
V.
REFERENCES
[2]
[3]
-600
-700
Power for BESS )
Time (h)
CONCLUSION
[1]
Electric energy Price ($/MWh)
600
[4]
134
327
150
Power (MW)
g
( )
2500
b) Case B
The optimal decision of the proposed model shows the
optimal selling power of wind generators through substation
transformers in a calculation cycle depicted in figure 5. In this
case, BESS is considered as an optimal selection to improve
efficiency of wind generators. Although, high investment
cost, electric energy sale profits are high. Thus, operation
power of BESS in a calculation cycle is selected as in figure
6. The optimal power and capacity of BESS are decided to
invest by 632.43kW and 3178.66kWh, respectively.
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[13]
135
328
I. INTRODUCTION
DVR
Supply
VDVR
Filter
Control
DVR
Voltage
Controlled
Bus
VSC
Energy
Storage
VDVR
IM
Sensitive
Load
VPCC
VIM
Fig. 3. DVR application to protect individual sensitive loads
as (2). After the motor speed reaches near the rated speed,
terminal voltage of the induction motor VIM increases close to
1 p.u, DVR is switched to stand-by mode (VDVR = 0).
Starting
Motor
ZS
DVR
IM
VIM
VDVR
VPCC
a)
Sensitive
Load
I.ZS
VS
VS
I.Zs
VPCC
VPCC
VDVR
VIM
VPCC = VIM
b)
c)
d)
Fig. 4. DVR application for mitigating voltage sag due to starting of induction
motor. a) System scheme, b) Phasor diagram before starting motor IM, c)
Phasor diagram at the point of starting motor IM without DVR, d) Similar to
c) with DVR. I supply current during starting motor.
B. Topologies of DVR
There are different topologies for DVR application as
discussed in [6]. For the problem under our consideration, the
paper only considers two typical system configurations for
DVR application as shown in Fig. 5.
Supply
IM
Starting
Motor
+
_
Series
Converte
Supply
Starting
Motor
Shunt
Rectifier
+
_
Series
Converte
138
VS
Va.PCC
--0
Vb.PCC
a-b-c
V0
Block 2
d-q-0
Vd
Vc.PCC
Vq
--0
Block 6
_
+
--0
_
+
Three
phase
PLL
Block 7
a-b-c
V
V0
d-q-0
Va.Con
Vb.Con
Vc.Con
--0
Block 8
PWM
Generator
Block 5
V
--0
Vb.Ref
Vc.Ref
a-b-c
V
V0
Block 3
PWM1-6
switch
t
d-q-0
--0
Vd.Ref
Voltage Sag
Detection
Vq.Ref
DVR
Breaker
Block 4
Fig. 6. Control Scheme for mitigating voltage sag due to starting of induction motor
V cos
V =
sin
Va
1 2 1 2
(3)
Vb
3 2 3 2
Vc
Block 2 transforms measured voltages from - frame to
synchronously rotating d-q frame using Parks transform (4).
Vd cos sin V
(4)
V =
q sin cos V
Similarly, Block 3 and Block 4 converts the reference
voltage to two phase d-q frame. The reference voltage VRef is
three-phase symmetrical fundamental frequency, sinusoidal
component of the source voltage.
Block 5 - phase locked loop (PLL) tracks continuously the
fundamental frequency of the measured supply voltages and
generates a unit sinusoidal wave in phase with system voltage.
Measured supply voltages and reference voltage are compared
to each other
Vdq.Con = Vdq.PCC Vdq. Re f
(5)
Va
V =
b
Vc
V
V =
2 1
3 0
sin Vd
cos Vq
(6)
0
1
V
2
(7)
3 2
1 2
V
3
1 2 3 2
This straight-forward comparison allows to obtain the
desired extract voltage (compensating voltage) without any
further control techniques (PI control or feedback loops). So,
the control strategy is simplified and consistent.
Block 8 is the PWM generator that generates firing signals
to control three phase voltage source converter for injecting
the compensating voltage in series with motor voltage. Block
voltage sag detection is used to control the operation of DVR
between active and inactive modes. The voltage sag is
detected when Root Mean Square (rms) value of motor
terminal voltage drops under 0.9 p.u. During voltage sag
duration, DVR is activated by opening the DVR breaker. In
steady state operation, DVR is short-circuited by closing the
DVR breaker.
IV. SIMULATION AND RESULT ANALYSIS
A. Test System
The test system simulation in the Matlab/Simulink is
printed out as Fig.7. The parameters of the main components
are listed in the Table I.
139
Va.Ref
oltages
Ratings
400 V, 50 Hz
1
7.5 kW, 400V, 1445rpm, 2P
0.7384
3.045mH
0.7402
3.045mH
124.1mH
0.0343kg/m2
10kVA, 400/400V
8mH
10, 20F
IGBT based
6 pulses, 1080Hz
300V
2000F
Diode based
1kVA, 400/230V
Vpcc
(p.u)
VApcc
(p.u)
Vpcc
(p.urms)
VApcc
(p.urms)
Im(A)
VAdvr
(p.u)
Im
(Arms)
VAdvr
(p.urms)
VAm
(p.u)
nm
(rpm)
VAm
(p.urms)
time(s)
Fig. 8. Test system parameters when motor starts without DVR
Vpcc
(p.urms)
Im(A)
DVR
deactivated
Im
(Arms)
nm
(rpm)
Te
(N.m)
time(s)
Fig. 9. Test system parameters when motor starts with battery supplied DVR
time(s)
Fig. 10. Bus voltages when motor starts with battery supplied DVR
Te
(N.m)
QS
(VAr)
Vdvr
(p.u)
PDVR
(W)
Vm
(p.u)
time(s)
Fig. 11. DVRs output voltage and motor terminal voltage when voltage sag
detection threshold is set to 0.9.
QDVR
(VAr)
PM
(W)
QM
(VAr)
time(s)
Fig. 13. Power flow when motor starts with battery supplied DVR
Idc (A)
Vpcc
(p.urms)
Vdc
(V)
c)
PS (W)
Im
(Arms)
time(s)
Fig. 12. DC bus voltage and current when DVR is supplied by battery
QS = QIM + QDVR
(9)
nm
(rpm)
Te
(N.m)
Idc
(A)
Vdc
(V)
time(s)
Fig. 14. Test system parameters when DVR is energized through rectifier
Vdc
(V)
time(s)
Fig. 15. DC bus voltage for different capacitances
142
V. CONCLUSION
This paper presents a new solution for mitigating voltage
sag due to the starting of three phase induction motors in
industrial electric networks. For the problem of voltage
compensation like the case under papers consideration, using
the series compensator like DVR proves it a suitable solution.
The feed-forward control strategy also proves its reliability
and suitability for DVR application. The simulation results
shows that this compensation solution gives a better
performance in comparison with previous researches using
DSTATCOM because the voltage sag is totally compensated
(not partly compensated). Apart from the above said power
quality improvement, a better motor starting condition
(reducing starting current and electro-magnetic torque) is
obtained by this proposed solution as well. The
Matlab/Simulink simulation tested two DVR configurations conventional DC supply using battery and DVR using rectifier
DC supply and both prove the same performance. This work
can be further developed with regards to further different
optional energy storage applications and the combination with
DSTATCOM for including other compensation objectives in
industrial distribution systems.
VI. REFERENCES
M.H.J. Bollen, Understanding power quality problems - voltage sags
and interruptions, IEEE Press, 2000.
[2] R.C. Dugan, M.F. Granaghan, S. Santoso, H.W. Beaty, Electrical Power
Systems Quality, 2nd Edition, McGraw Hill, 2004.
[3] A. Ghosh, G. Ledwich, Power Quality Enhancement Using Custom
Power Devices, Kluwer Academic Publishers, 2002.
[4] H. Akagi, E. H. Watanabe, M. Aredes, Instantaneous Power Theory and
Applications to Power Conditioning, IEEE Press, 2007.
[5] A. Ghosh, G. Ledwich, Compensation of Distribution System Voltage
Using DVR, IEEE Trans. Power Delivery, vol. 17, no. 4, pp. 10301036, Oct. 2002.
[6] J.G. Nielsen, F. Blaabjerg, A Detailed Comparison of System
Topologies for Dynamic Voltage Restorers, IEEE Trans. Industry
Application, vol. 41, no. 5, pp 1272-1280, Sep/Oct 2005.
[7] T. Devaraju, V.C. Veera Reddy, V. Kumar, Modeling and Simulation
of Custom Power Devices to Mitigate Power Quality Problems,
International Journal of Engineering Science and Technology, vol. 2(6),
pp. 1880-1885, 2010,
[8] B. Singh, P. Jayaprakash, D. Kothari, A. Chandra, Kamal-Al-Haddad,
Indirect control of capacitor supported DVR for power quality
improvement in distribution system, Presented in Power & Energy
Society General Meeting, 2008.
[9] Mahmoud A. El-Gammal, Amr Y. Abou-Ghazala, and Tarek I. ElShennawy, Dynamic Voltage Restorer (DVR) for Voltage Sag
Mitigation, International Journal on Electrical Engineering and
Informatics, vol. 3, no. 1, 2011.
[10] R. Omar, N. A. Rahim, M. Sulaiman, Modeling and Simulation for
Voltage Sags/Swell Mitigation Using Dynamic Voltage Restorer (DVR),
Journal of Theoretical and Applied Information Technology, pp. 464470, 2005-2009.
[11] D. Masand, S. Jain, G. Agnihotri, Control Strategies for Distribution
Static Compensator for Power Quality Improvement, IETE Journal of
Research, Vol. 54, Issue 6, pp. 421-28, Nov/Dec. 2008.
[1]
143
144
145
146
147
I.
INTRODUCTION
148
Department of Electrical Engineering, Faculty of Engineering, Chulalongkorn University, Bangkok, 10330, Thailand
Email: tien_lv@yahoo.com and tayjasanant@yahoo.ca
A. Fault at Buses
When faults take place at a bus of a power system, the sag
magnitude at bus can be calculated by the Z-bus matrix.
Therefore, the voltage at bus when a fault occur at bus is
given by
(1)
(2)
(3)
TABLE I
THE ZERO, POSITIVE AND NEGATIVE SEQUENCE FAULT CURRENT PHASE AT
BUS
Type of
fault
SLGF
LLF
3PF
(7)
DLGF
(6)
(8)
(9)
C. Voltage Threshold
Information Technology Industry Council (ITIC) provides
the ITIC curve which defines the withstand capability of
sensitive loads and devices for protection from power quality
variations. Each type of sensitive equipment has different
curve. The equipment sensitive limit can be evaluated from the
curve.
An example of ITIC curve represents magnitude and
duration of the event is shown in Fig.2 [3]. Points outside the
envelope are presumed to cause damage to the equipment. The
lower envelope is used to determine the acceptable sag
magnitude and duration level. In this paper, ITIC curve is
defined as the voltage threshold of a sensitive load.
(4)
line .
where
and
, respectively.
(5)
149
Probability
0.80
0.05
0.10
0.05
Durations (ms)
100
200
TABLE IV
FAILURE RATE FOR BUSES AND LINES
Type of
Fault
SLGF
LLF
DLGF
3PF
Fault type
Single Line-to-Ground fault (SLGF)
Line-to-line fault (LLF)
Double Line-to-Ground fault (DLGF)
Three-Phase fault (3PF)
TABLE II
FAULT TYPE DISTRIBUTION
150
D. Area of Vulnerability
The area of vulnerability (AOV) of a sensitive load is
essential to estimate the expected number of voltage sags from
the utility systems. The sag magnitudes at a sensitive load bus
can be approximated to quadratic curve for .
Therefore, there are four cases to determine the AOV by using
voltage threshold:
Case 1: , line is outside the AOV.
Case 2: , line is inside the AOV.
Case 3: If the voltage threshold is a value between the
remaining voltage at and , there is one
part of line inside the AOV.
Case 4: If the voltage threshold is greater than the remaining
voltage at and , and lower than the
maximum remaining voltage, there are two parts
of line inside the AOV.
TABLE V
VALUES OF ESF AT VARIOUS BUSES
Bus
3
4
6
7
9
10
12
14
ESF
2.10
2.61
3.80
1.32
5.47
6.15
3.90
7.09
Bus
15
16
17
18
19
20
21
22
ESF
9.27
7.44
7.22
9.59
9.10
8.75
6.75
6.45
Bus
23
24
25
26
27
28
29
30
ESF
9.69
8.49
6.51
7.55
3.73
1.32
4.91
5.49
ESF values at buses near the generator buses are lower than
buses further away. The area which covers buses with ESF
value above 8.3 sags/year is drawn with the dotted line in
Fig.3. Buses in this area need more attention if sensitive loads
are about to connected to.
Figure 3. The area of vulnerability for sensitive
load at bus 18 by using MCS
[3]
[4]
[5]
[6]
[7]
[8]
[9]
151
IV. CONCLUSION
I. I NTRODUCTION
Voltage sag is a short-duration reduction in RMS voltage
between 0.1 and 0.9 p.u. with duration from 0.5 cycles to 1
min [1-3]. Voltage sags that affect sensitive load are usually
caused by faults somewhere on transmission and distribution
systems.
Voltage magnitude and duration are essential characteristics of voltage sag. The magnitude of voltage sags mainly
depends on the fault location, fault type and some other
factors such as the pre-fault voltage, transformer connection,
and fault impedance [2-3]. The voltage sag magnitude, which
is expressed in percent or per unit, is calculated by shortcircuit analysis. The voltage sag duration is dened as the
ow duration of the fault current in a network. Therefore, the
duration is determined by the characteristics of the system
protection devices such as overcurrent relays, circuit breakers
and fuses. Generally, the duration is calculated by adding the
* The Department of Electrical Engineering, Chulalongkorn University,
Bangkok 10330, Thailand.
** The University of Tokyo, School of Frontier Science, Department of
Advanced Energy, Hongo 7-3-1, Bunkyo-ku, Tokyo 113-8656, Japan.
152
Fig. 2.
Fig. 3.
153
Fig. 1.
B. Recloser-Fuse Operations
k=1
Lij LF Ri
N SF =
(2)
i=1 j=1
N SFF T,P
(3)
N SFSP =
F T =1P =1
where F T is type of fault (i.e. single-line-ground fault; lineto-line fault, double-line-ground fault and three-phase fault);
P is the number of phases.
Fig. 4.
Protection coordination
(4)
(5)
(6)
154
10
100K Fuse
Recloser S
10
Breaker
Time(s)
10
Recloser F
10
Fuse TC
10
Fuse MM
2
10
10
10
Current(A)
10
10
80K Fuse
10
Breaker
10
Time(s)
Fig. 5.
Recloser S
Recloser F
10
10
Fuse MM
Fuse TC
10
10
10
Current(A)
10
Fig. 6.
155
Fig. 7. 100K fuse and 80K fuse coordination with the recloser to clear
threephase fault at bus 18
(b)
(c)
Fig. 8. The areas of vulnerability for sensitive load at bus 2 without protection
coordination. (a), (b), and (c) are areas of vulnerability due to SLGF and 3PF,
LLF, DLGF respectively.
(a)
(b)
(a)
(c)
Fig. 9. The areas of vulnerability for sensitive load at bus 2 with protection
coordination. (a), (b), and (c) are areas of vulnerability due to SLGF and 3PF,
LLF, DLGF respectively.
phases for all type of faults in the test system. The tables
clearly show that the number of sag frequency decreases
with protection coordination consideration. In this case, the
numbers of sag frequency for a single phase with uniform and
normal distribution along lines are 115.85 and 117.73 sags,
respectively.
VI. C ONCLUSION
This paper presents a new method to detect fuse and
recloser operations with consideration of impacts on sensitive
156
TABLE I
T HE NUMBER OF LINE SEGMENTS INSIDE THE AREA OF VULNERABILITY
FOR SENSITIVE EQUIPMENT AT BUS 2 WITHOUT PROTECTION
COORDINATION CONSIDERATION
Type of fault
SLGF (A)
LLF (B,C)
DLGF (B,C)
3PF
Phase A
124
0
0
131
Phase B
0
34
109
131
Phase C
0
145
143
131
TABLE II
T HE NUMBER OF SAG FREQUENCY FOR SENSITIVE EQUIPMENT AT BUS 2
WITHOUT PROTECTION COORDINATION CONSIDERATION
Type
of fault
SLGF
LLF
DLGF
3PF
Uniform distribution
PhaseA PhaseB PhaseC
292.52
0
0
0
7.62
32.26
0
15.16
20.72
7.27
7.27
7.27
Normal distribution
PhaseA PhaseB PhaseC
293.42
0
0
0
7.29
32.75
0
14.68
20.46
7.42
7.42
7.42
TABLE III
T HE NUMBER OF LINE SEGMENTS INSIDE THE AREA OF VULNERABILITY
FOR SENSITIVE EQUIPMENT AT BUS 2 WITH PROTECTION COORDINATION
R EFERENCES
[1] IEEE Std 1159-1995: IEEE Recommended Practice for Monitoring Electric Power Quality.
[2] M.H.J. Bollen, Understanding Power Quality Problems: Voltage Sags and
Interruption, New York: IEEE Press (2000)
[3] R.C. Dungan, M.F. McGranaghan, and H.W. Beaty, Electrical Power
Systems Quality, McGraw Hill (1996)
[4] J.M. Gers and E.J. Holmes, Protection of Electricity Distribution Networks. London, U.K.: Inst. Elect. Eng. (2004)
[5] L. Comassetto, D.P. Bernardon, L.N. Canha and A.R. Abaide: Automated
Coordination and Optimization Tool of Protection Devices for Distribution System , POWERENG2007, pp.388-393 (2007)
[6] S. Chaitusaney and A. Yokoyama: Prevention of Reliability Degradation
from Recloser-Fuse Miscoodination Due To Distribution Generation,
IEEE Transactions on Power Delivery, Vol. 23, No. 4, pp.2545-2554
(2008)
[7] C.H. Park and G. Jang: Stochastic Estimation of Voltage Sags in a Large
Mesh Network, IEEE Transaction on Power Delivery, Vol. 22, No. 3,
pp. 1655-1664 (2007)
[8] Le Viet Tien and Thavatchai Tayjasanant: An Analytical Approach for
Stochastic Estimation of Voltage Sags in Power Systems, ECTI-CON
2008, Vol. 2, pp. 1029-1032 (2008)
[9] R.N. Allan, R. Billiton, I. Sjarief, L. Goel, and K.S. So, A Reliability
Test System for Education Purposes Basic Distribution System Data and
Results, IEEE Transactions on Power System, Vol. 6, No. 2, pp. 813-820
(1991)
[10] J.V. Milanovic, Myo Thu Aung, C.P. Gupta, The Inuence of Fault
Distribution on Stochastic Prediction of Voltage Sags. IEEE Transactions
on Power System, Vol. 20, No. 1, pp. 278-285 (2005).
CONSIDERATION
Phase A
117
0
0
123
Phase B
0
20
99
123
Phase C
0
62
143
123
Type of fault
SLGF (A)
LLF (B,C)
DLGF (B,C)
3PF
TABLE IV
T HE NUMBER OF SAG FREQUENCY FOR SENSITIVE EQUIPMENT AT BUS 2
WITH PROTECTION COORDINATION CONSIDERATION
Type
of fault
SLGF
LLF
DLGF
3PF
Uniform distribution
PhaseA PhaseB PhaseC
274.90
0
0
0
4.60
13.86
0
13.74
19.89
6.84
6.84
6.84
Normal distribution
PhaseA PhaseB PhaseC
279.87
0
0
0
4.60
14.19
0
13.56
19.95
7.01
7.01
7.01
equipment in a distribution system. The method requires sensitive equipment data, and time-current characteristic curves
of the protective devices. Two types of fuse were used to
evaluate performance of coordination of protective devices
with sensitive equipment.
Further, two kinds of fault distributions along lines are
also simulated to show the numbers and characteristics of sag
frequency. The fault position method is used to determine the
number of line segments inside the area of vulnerability of the
sensitive equipment load. The effectiveness of the technique is
tested for impacts of voltage sags and protection coordination
on sensitive equipment. It has been make clear that sag
frequency taking the protection coordination into account is
lower than sag frequency without the protection coordination
consideration.
157
Paper
Nowadays, a lot of sensitive electronic equipment is widely used in modern power systems such as power converters
and adjustable speed drivers. Voltage sags have gained more interest due to their consequences on the performance of
sensitive equipment (SE). Malfunction or failure of the equipment that leads to work or production losses can be caused
by voltage sags. As a result, it is essential to have information on equipment sensitivity. If the magnitude and duration
of voltage sag exceed the equipment sensitivity threshold, the equipment can malfunction, and such a consequence can
aect an entire automatic process, resulting in high economical losses. Reclosers and fuses are the main overcurrent
protection devices in distribution systems. Poor coordination may adversely impact on the sensitive equipment. This
paper presents a method to analyze impacts of voltage sags and protection coordination on sensitive equipment. A
fault position method and mathematical equations describing characteristics of protection devices are used to calculate
voltage sags and the protection settings. Voltage tolerance thresholds and protective device characteristics are used to
analyze protection and sensitive equipment coordination. Based on the results, new settings for protective devices can
be adjusted for sensitive equipment in distribution systems. The Roy Billinton Test System (RBTS) bus 2 is used to
analyze impacts of voltage sags and protection coordination.
Keywords: sensitive equipment, voltage sag, protection coordination.
1.
Introduction
1
158
Le Viet Tien
Thavatchai Tayjasanant
Akihiko Yokoyama
Bundhit Eua-Arporn
This paper presents a novel approach to detect which protective device will operate to clear a short-circuit and protect
sensitive equipment. Proper protection coordination can provide correct relay settings and increase power quality.
Protection Coordination
where t and I are the associated time and current; a and b can
be known from the curve tting.
Circuit breakers (CB) and reclosers usually located at the
beginning and middle of main feeders. The general characteristics of these devices can be shown as the following equation:
A
t(I) = p
+ B (2)
M 1
where t is the operating time of inverse-time overcurrent device; I is fault current seen by the device; M is the ratio of
I/I pickup (I pickup is relay current set point); A, B, p are constants for selected curve characteristics.
For the protection settings, this paper uses the mathematical equations for overcurrent relays and the straight line I 2 t
log-log curve to formulate protection coordination (5) (6) . The
protection settings are done in the initial or existing condition. CBs and reclosers characteristics are assumed to be
equipped with the extremely inverse characteristic of overcurrent relays. A typical radial distribution system is shown
with protection devices in Fig. 1. In a typical distribution system, all demand loads are supplied from the main feeder.
2.2 Recloser-Fuse Coordination
Figure 2 shows traditional recloser-fuse coordination in distribution systems (6) .
A recloser has two characteristic curves which are fast and
slow (buck up) operation. In recloser function, there is an
interval between each operation when the recloser remains
open. If the fault is temporary, recloser will clear before the
fuse. If the fault persists after the recloser closes then the
fault has to be a permanent one and hence fuse must operate
to cut it o.
The general coordination is that the fuse should only operate for a permanent fault on the load feeder. However if
2.
li
(3)
lk
k=1
NS F =
Li j LFRi (4)
sensitive equipment (1) (7) . This curve denes the withstand capability of sensitive loads and devices for protection from
power quality variations. Each type of sensitive equipment
has dierent curve. An example of ITIC curve represents
magnitude and duration of the event as shown in Fig. 3.
SEMI F47 is a standard that denes the voltage threshold that
semiconductor processing must be controlled through without interruption during conditions identied in the area above
the standard line (Fig. 3). Points above the threshold are presumed to cause damage to the equipment. The lower region
will use to determine the acceptable sag magnitude and duration level.
In this paper, ITIC curve is dened as the voltage threshold
of a sensitive load. Voltage threshold can be determined by a
given sag duration.
3.2 Area of Vulnerability and Fault Distribution
Along Line
The region of the network that includes a
part of line or whole line where the occurrence of the faults
will lead to voltage sag at a sensitive load is called the area
of vulnerability (AOV). The area of vulnerability responding to the sensitive load is essential to estimate the expected
number of voltage sags from the utility systems as shown in
Fig. 4. The fault distribution along line will cause changes in
the expected number of sag frequency at the sensitive load.
3.3 Voltage Sag Estimation
Frequently, adverse
environmental and weather conditions can lead to sections of
lines exposed to higher fault rates than others. This can have
a considerable impact on the expected number of sags at the
B130 6 2010
B mn H thng in - i hc Bch Khoa H Ni
NS FS P =
1
NS F FT,P (5)
3 FT =1 P=1
where FT is type of fault (i.e. single-line-ground; line-toline, double-line-ground and three-phase faults); P is the
number of phases.
4.
i=1 j=1
fuse link (80 K and 100 K) are used as examples in this paper.
Parameters a and b of the straight line I 2 t of (1) are shown in
Table 1.
Let us make the following assumptions: (1) timecurrent
characteristic curves of fuses and reclosers used in the feeder
and (2) sensitive equipment characteristics are available (9)
(see Fig. 7). Each line is divided into ve equal line segments.
500 faults occurring in the test system are assumed. Besides
that, the details of system fault statistics are also assumed as
follows (11)
Single-Line-Ground fault (SLGF): 85%
Line-to-Line fault (LLF): 8%
Double-Line-Ground fault (DLGF): 5%
Three-Phase fault: 2%
Based on the list of percentages of dierent fault types,
number of SLGF, LLF, DLGF and 3PF are 425, 40, 25 and
10 faults, respectively.
5.1 Simulation of RecloserFuse Coordination with
Consideration of Sensitive Equipment
Assume that the
distribution network is protected by a 100 K fuse link and a
recloser on the main feeder upstream from the fuse. Threephase fault occurs at bus 18 in the test system. Voltage sag
Total clearing
aTC
bTC
-2.5132 -2.9199
-2.6951 8.4269
5.
Minimum melting
a MM
b MM
-2.9199 9.4467
-2.8378 8.5925
Type of
fuse
100 K
80 K
Fig. 8. 100 K fuse and 80 K fuse coordination with the recloser to clear threephase fault at bus 18
Fig. 9. Protection coordination with 100 K fuse and 80 K fuse when fault occurs along line from bus 5 to bus 18
Type of fault
SLGF (A)
LLF (B,C)
DLGF (B,C)
3PF
Phase A
124
0
0
131
Phase B
0
34
109
131
Phase C
0
145
143
131
Type
of fault
SLGF
LLF
DLGF
3PF
Uniform distribution
Phase A Phase B Phase C
292.52
0
0
0
7.62
32.26
0
15.16
20.72
7.27
7.27
7.27
Normal distribution
Phase A Phase B Phase C
293.42
0
0
0
7.29
32.75
0
14.68
20.46
7.42
7.42
7.42
(a)
(a)
(b)
(b)
(c)
(c)
Phase B
0
20
99
123
(1)
Phase C
0
62
143
123
(2)
(3)
(4)
(5)
Uniform distribution
Phase A Phase B Phase C
274.90
0
0
0
4.60
13.86
0
13.74
19.89
6.84
6.84
6.84
(6)
Normal distribution
Phase A Phase B Phase C
279.87
0
0
0
4.60
14.19
0
13.56
19.95
7.01
7.01
7.01
(7)
(8)
(9)
( 10 )
( 11 )
( 12 )
Appendix
When a fault concurs along line k j at a fault position
p which is dened as the ratio of length between bus k and
fault location (Lkp ) to the length of the line k j (Lk j ) or
p = Lkp /Lk j . The voltage at bus m, when a fault occurs at
fault position p between k and j, can be calculated from
012 012
Vm012 = Vm012p f Zmp
I p (A1)
012
is the transfer bus impedance beIn equation (A1), Zmp
tween bus m and a fault position p on the line k j and can
be shown as following
012
012
[Zmp
] = (1 p)[Zmk
] + p[Zm012j ] (A2)
Conclusion
This paper presents a new method to detect fuse and recloser operations with consideration of impacts on sensitive
equipment in a distribution system. The method requires sensitive equipment data, and time-current characteristic curves
of the protective devices. Two types of fuse were used to
evaluate performance of coordination of protective devices
with sensitive equipment.
Further, two kinds of fault distributions along lines are also
simulated to show the numbers and characteristics of sag
frequency. The fault position method is used to determine
the number of line segments inside the area of vulnerability of the sensitive equipment load. The eectiveness of the
technique is tested for impacts of voltage sags and protection coordination on sensitive equipment. It is clear that sag
frequency taking the protection coordination into account is
lower than sag frequency without the protection coordination
consideration.
(Manuscript received Sep. 14, 2009,
revised Jan. 13, 2010)
B130 6 2010
B mn H thng in - i hc Bch Khoa H Ni
V p1p f
Z 0pp + Z 1pp + Z 2pp + 3z f
(A4)
I 1p = I 2p =
V p1p f
(A5)
Z 1pp + Z 2pp + 3z f
and I 0p = 0.
7
164
Type of fault
SLGF (A)
LLF (B,C)
DLGF (B,C)
3PF
References
I 1p =
V p1p f
Z 1pp +
I 2p = I 1p
I 0p = I 1p
(A6)
Z 0pp + 3z f
Z 0pp + Z 2pp + 3z f
Z 2pp
Z 0pp
Z 2pp
+ 3z f
(A7)
Akihiko Yokoyama (Member) was born in Osaka, Japan, on October
9, 1956. He received the B.S., M.S. and D.Eng. degrees from The University of Tokyo, Tokyo, Japan, in
1979, 1981, and 1984, respectively. He has been with
the Department of Electrical Engineering, The University of Tokyo, since 1984 and is currently a Professor in charge of Power System Engineering. He is
a member of IEEJ and CIGRE.
(A8)
I 1p =
V p1p f
(A9)
Z 1pp + z f
and I 0p = I 2p = 0.
The value of matrix [Z 012
pp ] to be used in (A4) to (A9) can
be calculated from
2 012
2 012
[Z 012
pp ] = (1 p) [Zkk ] + p [Z j j ] +
012
+ 2p(1 p)[Zk012
j ] + p(1 p)[zk j ]
(A10)
where [z012
k j ] is a diagonal 3x3 matrix of impedance of distribution line k j.
Le Viet Tien (Student Member) was born in Vietnam, in 1980. He
received the B.Eng. in electrical engineering from
Hanoi University of Technology, Hanoi, Vietnam in
2003. He is currently pursuing the Ph.D. degree under the AUN/SEED-Net Scholarship from JICA at the
Power System Research Lab., Department of Electrical Engineering, Faculty of Engineering, Chulalongkorn University. He is now staying at the University of Tokyo for short-study program of JICA scholarship. He is a student member of IEEJ.
8
B mn H thng in - i hc Bch Khoa H Ni
I. INTRODUCTION
Recloser
Fuse
Fuse
Main feeder
Fuse
Fuse
Lateral feeders
A typical radial distribution system is shown with protection devices in Fig. 1. In typical distribution system, all demand loads are supplied from a bulk.
B. Protection Coordination
Fig. 2 shows traditional fuse-recloser coordination in distribution systems [6]. In recloser function, there is an interval
between each operation when the recloser remains open. If the
fault is temporary, recloser will clear before the fuse. If the
fault persists after the recloser closes then the fault has to be a
permanent one and hence fuse must operate to cut it off.
The general coordination is that the fuse should only operate for a permanent fault on the load feeder. However if the
fault is a temporary fault or the fault occurs behind the recloser, the recloser should disconnect the circuit with fast operation and give the fault a chance to clear. Recloser also provides back up function when a fuse fails to blow up. In order
to have a correct operation, the fuse must be coordinated with
upstream recloser on the main feeder.
In Fig. 2, the TC curve of the fuse is below slow curve of
recloser in coordination range. Therefore, for a permanent
B mn H thng in - i hc Bch Khoa H Ni
fault, fuse will open before recloser will back it up by operating in slow mode and finally locking out. The coordination
curves of recloser and fuse have to be modified. The fuserecloser coordination range graph between I fault max
and I fault min . Therefore, as long as the fault current values for
faults on lateral feeder are within coordination range, the fuserecloser coordination is accepted. We can see that the fast
characteristic of the recloser lies below the MM characteristic
of fuse between I fault max and I fault min . So, in coordination
range the recloser operates in less time than the time sufficient
to damage the fuse.
167
012
[ Z mp
]
012
012
(1 p)[ Z mk
] p[ Z mj
]
(4)
I 012
is calculated based on the type of faults. The value of
p
matrix Z 012
pp can be calculated from
Z 012
pp
(1 p) 2 [ Z kk012 ] p 2 [ Z 012
jj ]
012
kj
(5)
line k j .
The pre-fault voltage at the fault position p is
A. Sensitive Equipment
Normally, sensitive equipment is analyzed by using voltage threshold [7-8]. The voltage threshold is determined by
fixed duration of fault current. Therefore, if the duration of
fault current is greater than fixed duration, it will impacts on
sensitive equipment. This paper will present an approach to
determine vulnerability time t fault SE based on remaining voltage at sensitive equipment bus.
Sensitive equipment will be impacted by voltage sags
V fault when a fault occurs in system. Then t fault SE will be de-
(6)
and j , respectively.
When a three-phase fault occurs, the voltage sag at bus m
can be expressed as
1
Z mp
(7)
Vmfault Vmpref 1 V ppref
Z pp
V ppref
B. Recloser-Fuse Operations
When a permanent or temporary fault occur, a recloser or
fuse should blow to isolate the short-circuit condition. Times
corresponding to I fault on the fuses time-current curves
(TCC) and the reclosers TCC curves are determined as
t fault fuse MM and t fault fuse TC , t fault rec fast and t fault rec slow respectively as shown in Fig. 4.
For considering sensitive equipment, the time t fault SE when
the fault occurs must be greater than t fault fuse MM or
t fault fuse TC ,
t fault fuse MM d t fault SE or t fault fuse TC d t fault SE
(8)
(10)
168
012
The value of matrix Z mp
is
I fault
results indicate that the fuse coordinations well with the recloser considering sensitive equipment are 100 K-type and 80
K-type, respectively. Figures 7 and 8 show the operation point
on the recloser and fuse coordination curve.
3
10
Step 2)
Step 3)
Step 4)
Step 5)
10
10
Time(s)
Step 1)
10
-1
10
10
10
Current(A)
10
Fig. 7. Fuse 100 K-type coordination with the recloser to clear three-phase
fault at bus 18.
3
10
10
Time(s)
10
10
-1
10
100
-2
10
90
80
10
10
Current(A)
10
Fig. 8. Fuse 80 K-type coordination with the recloser to clear three-phase fault
at bus 18.
70
60
50
Fig. 9 shows the coordination of protective devices (recloser and fuses: 80 K-type and 100 K-type) in four main
feeders with the sensitive equipment located at bus 2 in feeder
1. In Figure 9b, 80 K-type fuse may not protect the sensitive
equipment when fault occurs at bus 23 in the feeder 2.
40
30
20
10
0.2
0.5
1.0
10
100
169
-2
10
170
Fig. 10. Protection coordination with 100 K-type and 80 K-type fuse when fault occurs along line 11 (from bus 5 to bus 18).
[6] S. Chaitusaney and A. Yokoyama, Prevention of Reliability Degradation from Recloser-Fuse Miscoodination Due To Distribution GeneraB. Case 2: Simulation with Fault Along Lines
tion, IEEE Transactions on Power Delivery, Vol. 23, No. 4, October
Based on the fault location method, a three-phase fault will
2008.
be occur along line 11 (from bus 5 to bus 18) on the feeder 1 [7] C.H. Park, G. Jang, Stochastic Estimation of Voltage Sags in a Large
Mesh Network, IEEE Transaction on Power Delivery, Vol. 22, No. 3,
as shown in Fig. 10. It is shown that 100 K-type fuse can coJuly 2007.
ordinate with recloser to protect sensitive equipment if the [8] Le Viet Tien and Thavatchai Tayjasanant, An Analytical Approach for
fault occurs. The estimated fault current flowing in the fuse
Stochastic Estimation of Voltage Sags in Power Systems, ECTI-CON
2008, Vol. 2, May 14-17, 2008.
were 2.33 kA, 2.62KA and 2.45kA corresponding with fault
location 0.1, 0.5 and 1, respectively. Vulnerable duration is [9] Y.S. Lim and G. Strbac, Analytical Approach to Probabilistic
Prediction of Voltage Sags on Transmission Networks, IEEE Proc.0.5s based on the sensitive equipment characteristic. Figure 10
Gener. Transmi. Distrib, Vol. 149, No. 1, January 2002.
shows that if a fault occur at midline, 80 K-type fuse cannot [10] R.N. Allan, R. Billiton, I. Sjarief, L. Goel, and K.S. So, A Reliability
protect the sensitive equipment, but 100 K-type fuse can proTest System for Education Purposes Basic Distribution System Data
and Results, IEEE Transactions on Power System, Vol. 6, No. 2, May
tect with the protection coordination.
1991.
VI. CONCLUSION
VIII. BIOGRAPHIES
The proposed method is to detect fuse and recloser opera- Le Viet Tien was born in Vietnam, in 1980. He received the B. Eng. in elections with consideration of impacts on sensitive equipment in trical engineering from Hanoi University of Technology, Hanoi, Vietnam in
2003. He is currently pursuing the Ph.D. degree at the Power System Research
distribution system. The method requires sensitive equipment Lab., Department of Electrical Engineering, Faculty of Engineering, Chudata, and time-current characteristic curves of the protective lalongkorn University.
Journal of International Council on Electrical Engineering Vol. 2, No. 1, pp. 45~52, 2012
45
1. Introduction
The currently generalized industrial standard requires
disconnecting all the DG sources from the islanded entity as
fast as possible. For example, under the IEEE 1547
standard, that period is within two seconds [1]. For a
smooth transition from a grid-connected mode to an
islanded mode of operation, fast islanding detection (ID) is
essential to allow DGs to adapt their control strategy to the
new operational one.
The ID methods can be separated into two basic
categories: communication and local one. Local detection
methods can further subdivided into active and passive
detection schemes [2]. In 2008, C.Wrinch developed a
simple control scheme which is easy to accomplish in
software as well as hardware [3]. It presents passive ID by
using
negative-sequence
impedance
measurement.
Information about the step change of actual grid negativesequence impedance could be used for detecting an
islanding situation. That method helps detect an islanding in
as fast as a few cycles for the specific settings described
afterwards in the simulation in Section 4. Thus, it facilitates
an islanding's subsequence actions which fall into two
options. The first one is to connect immediately to a backup
system. The second option is to regulate control topology of
DG which is still in operation inside the islanded system.
*
**
172
Subsequence Action to Eliminate Blackout after Detecting Islanding using Solid State Transfer Switch Implemented in ~
Z DG
Z utility
Z load
E utility
V DG
Z 0sys
Z sys
Z load
V
sys
+
Z +sys
I V
m
I+
Z mload
I0 + I + I+
3. Transfer Switch
Z +load
V+
Z 0load
V0
(1)
173
46
Nguyen Duc Tuyen, Goro Fujita, Toshihisa Funabashi and Masakatsu Nomura
47
time delay, TD. That is a time delay (in ms) before the poles
actually commence to close after the Timed Breaker Logic
component (or other signal) has initiated closing. By
randomly changing TD with the Multiple Run component
over a period of one cycle (0 - 16.667 ms) during successive runs, line energizing studies can be accomplished.
3.2 Solid state transfer switch
Bka
BKA_backup
Bk
Trip
TripBK_backup
SWIT...
BREAKER D...
16.6667
Bkb
Statistical
Breaker
Close Bkc
TD
BKB_backup
BKC_backup
msec
BREAKER DELAY
10
(
) are transformed into positive-, negative-
174
Subsequence Action to Eliminate Blackout after Detecting Islanding using Solid State Transfer Switch Implemented in ~
u (PoM
a ,b ,c )
i (PoM
a ,b , c )
vc-Backup Terminal
u1(PoM
a ,b , c )
+0
vb-Backup Terminal
vb-CP
vc-CP
i1(PoM
a ,b , c )
48
Z line
Z line
SSTSa
Utility Bus
SSTSb
SSTSc
650
646
645
632
633
CP
u (PoM
)
i (PoM
)
634
PoM
tolerence
611
684
Z sys (t )
671
692
675
Z sys (t + delay )
652
680
Wind Generators
175
Nguyen Duc Tuyen, Goro Fujita, Toshihisa Funabashi and Masakatsu Nomura
Im p e d a n ce M a g n itu d e (O h m )
Z2
Z2filter
0.0
2.925
2.950
2.975
3.000
3.025
3.050
3.075
3.100
3.125
1.0
2.0
3.0
4.0
5.0
176
49
Subsequence Action to Eliminate Blackout after Detecting Islanding using Solid State Transfer Switch Implemented in ~
5.0
Voltage(kV)
Vbackup-c
Vbackup-a
-5.0
2.90
3.00
3.10
4.0
Voltage(kV)
50
Vload-b
Vload-c
-4.0
2.90
3.00
3.10
LoadVa
BackupVa
-4 0
4.0
Trip
-5 0
3.070
SSTSb
3.080
3.090
LoadVb
3.100
2.950
BackupVb
3.110
3.120
3.130
Z2
0.80
0.70
0.60
0.50
0.40
0.30
0.20
0.10
0.00
Z2filter
2.975
3.000
SSTSc
LoadVc
3.090
3.100
3.110
3.120
3.100
3.125
3.150
3.175
BackupVc
1.25
3.080
3.075
3.130
Ia
Ib
Ic
i [kA]
Magnitude
-4 0
3.070
Trip
3.050
3.025
-1.00
1.0
2.0
3.0
4.0
5.0
Voltage(kV)
5.0
Vload-a
time
Vload-b
Vload-c
-5.0
0.0
0.0
1.0
2.0
3.0
4.0
5.0
4.0
5.0
Z2filter
4.0
2.0
0.0
0.0
1.0
2.0
3.0
177
Magnitude
4.0
SSTSa
Nguyen Duc Tuyen, Goro Fujita, Toshihisa Funabashi and Masakatsu Nomura
0.0
1.0
2.0
3.0
4.0
5.0
3.0
4.0
5.0
References
-0.50
0.0
1.0
2.0
Qbackup
Power
4.00
-0.50
0.0
1.0
2.0
3.0
4.0
5.0
1.20
Qwind
Pwind
-0.60
0.0
1.0
2.0
3.0
4.0
5.0
3.0
4.0
5.0
PUti
QUti
-0.50
0.0
QUti
Pow er
3.50
7. Conclusion
1.0
2.0
178
Pow er
Pwind
2.00
1.50
1.00
0.50
0.00
-0.50
51
52
Subsequence Action to Eliminate Blackout after Detecting Islanding using Solid State Transfer Switch Implemented in ~
[10] http://ewh.ieee.org/soc/pes/dsacom/testfeeders.html
179
Paper
Adaptive Notch Filter for Synchronization and Islanding Detection using
Negative-sequence Impedance Measurement
Nguyen Duc Tuyen , Student Member
Goro Fujita , Member
Toshihisa Funabashi , Member
Masakatsu Nomura , Member
1. Introduction
The growing interest in environmental issues, the progress in
technologies to couple renewable energy sources to the grid, and
the liberalization of the energy market have led to a growing share
of grid-connected dispersed generation (DG). In the context of
DG units, an island is formed when one or more DG units are
disconnected from the utility grid due to accidental or preplanned
events while they are still in operation. The current generalized
industry standard requires disconnecting all the DG units from the
island as fast as possible, for example, the IEEE 1547 standard is
in 2 s [1]. For a smooth transition from a grid-connected mode
to an islanded mode of operation, fast islanding detection (ID) is
essential in order to allow DG units to adapt their control strategies
to new operational strategies. In fact, the detection time interval,
i.e. 2 s, is not fast enough for controller reaction in autonomous
operation.
There exist many ID methods that can be fundamentally
classied into two basic categories: the communication method
and the local one, and the latter can then be subdivided into active
and passive detection schemes [2]. In [3], a simple control scheme
has been developed that is easy to accomplish in software and
hardware as well. With the low cost used for ID, this can support
the negative impedance measurement method. It initially presents
a passive ID by using negative-sequence impedance measurement
a
2012 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.
B mn H thng in - i hc Bch Khoa H Ni
180
Dispersed generation (DG) has been found promising for satisfying the requirements of high power quality in distributed
systems. One operation situation, namely, an island, is formed when one or more DG systems and an aggregate of local loads are
disconnected from the main grid and remain operational as an islanded entity. Islanding is either due to intentional events, e.g.
maintenance outage, or due to unintentional events, such as faults, and their subsequent switching actions. Islanding is usually
undesirable because of the potential harm it causes to the existing equipment, human safety, power reliability and quality, etc.
Accordingly, anti-islanding schemes are used to immediately detach a DG system from the feeder after islanding. This paper
presents an active method of detecting islanding for DG systems by injecting a small negative voltage into the point of common
coupling, and then measuring the negative system impedance. In this study, an adaptive notch lter is introduced as a synchronous
part instead of a phase-locked loop and as a signal processing unit as well. The proposed control strategy allows DG systems
to detect properly the occurrence of islanding in a balanced distribution system. Simulation results show the overall system
performance including synchronization, power control, and islanding detection capability of the simulated DG system. 2012
Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.
Tng hp cc bi bo khoa hc
giai on NOTCH
2007-2012
ADAPTIVE
FILTER FOR SYNCHRONIZATION AND ISLANDING DETECTION
I0
Z0sys
Z0load
V0
Zsys
Zload
I V
m
E+sys
Zmload
I0 + I + I+
I+
Z+sys
Z+load
V+
sys
V
I
(1)
sys
E utility
CB
sys
sys
sys
E DG
sys
I
Z DG
Z load
Z utility
241
sys
v
2 1 1/2 1/2 aPCC
v
vbPCC
=
v
3/2
3 0 3/2
vcPCC
Vdcref
Vdc(t)
PID
iDGa(t)
iDGb(t)
IDGd
IDGdref
PID
IDGq
T(q)
V DGqref
V DGdref
PID
iDGc(t)
2
V 2DGqref + V DGdref
tan1
IDGqref
vaPCC(t)
vbPCC(t)
IDGqref = 0
(Qref = 0)
and
V +DGqref
cos
vd
=
vq
sin
V +DGdref
ref
vcPCC(t)
Inverter
V DGaref
VDGaref
Switches
control
PWM
i
iq
vq 0
c
id
q
a
C d
v aPC
C
PC
va
(3)
3.3. Three-phase adaptive notch lter in synchronization role One important part of the grid-connected DG units
vd
e(t) = u(t) x
where u(t) is the input signal, e(t) is the error signal, is the
estimated frequency, and and are adjustable real positive
parameters that respectively determine the estimation accuracy and
the capability of the algorithm in tracking the signal characteristics
variations. A trade-off between the (steady-state) accuracy and
(transient) convergence speed can be obtained by adjusting those
design parameters. By increasing , one can achieve faster
convergence speed; however, at the same time, should be
increased to avoid oscillatory behavior. The ANF structure has
three integrators. The initial condition for the one that outputs the
frequency is set to the nominal power system frequency of 2.60
rad/s. The initial conditions for all other integrators are set to zero.
For a single sinusoidal input signal u(t) = A sin (t + ) where A
v
v
sin
cos
V +DGaref
ANF
(2)
242
Pref
N. D. TUYEN ET AL.
Tng hp cc bi bo khoa hc
giai on NOTCH
2007-2012
ADAPTIVE
FILTER FOR SYNCHRONIZATION AND ISLANDING DETECTION
x..
1
s
theta_xa
x.
ea
1
Fundamental (x.)
1
s
2*zeta
ea
xa.
xa_
x
theta_xa
theta theta*xa
u2
Input
signal
u(t) 1
theta_xb
theta
2
Frequency (w)
theta
ANF-a
eb
u_a(t)
eb
xb.
e(t)
e(t)
1
s
-gamma
theta
x.
u_b(t)
abc
u(t)
5
Fundamental_90 (-theta*x)
sqrt
ec
ec
xc.
xc_
xc.
theta_xc
theta theta*xc
xb_
3
Amplitude (A)
theta_xb
ANF-b
x*theta
u2
xb_
xb.
theta theta*xb
Frequency
estimator
u_c(t)
4
Harmonic(h)
x.
xa_
xa.
ANF-c
u2
theta_xc
x.
theta
xc_
ua
4
x e (t)
-gamma
eb
1
s
4
theta
3
ec
7
xb.
8
theta*xc
uc
6
ec
9
xc.
x
cos (t + )
0
p (t) = x = A sin (t + )
(9)
(6)
and = . For the three-phase ANF in the steady state, the dened
outputs x and x are
x = A sin (t + ) u (t)
x = A cos (t + ) S90 u (t)
(10)
(7)
=a,b,c
Aa sin (t + a )
ua (t)
2
eb
3
theta*xb
ub
5
ea
2
xa.
1
ea
1
theta*xa
A
x
cos (t + )
.
(5)
o = x = A sin (t + )
3.4. Negative-sequence injection topology The proposed ID method is based on injecting a negative-sequence and
fundamental frequency current component through the power electronic interface. Afterward, the ANF detects the corresponding
(8)
243
N. D. TUYEN ET AL.
e_alpha
x_alpha.
1
1
s
2*zeta
1
s
2
theta*x_alpha
negative-sequence voltages at the PCC and DG units current owing into the PCC, which results in a negative system impedance.
To inject the negative-sequence currents, the conventional positivesequence current controller of the inverter is slightly augmented
with a negative-sequence voltage reference. In this study, the controller uses the ANF to measure a three-phase signal and derives
a single-phase reference signal. The distinct phase references for
individual phases are obtained by adding or subtracting 2/3 radians to the measured phase angle. Such a three-phase design has the
advantage of triple-harmonic cancelation. As shown in Fig. 4, for
the positive-sequence component, three reference phases a, b, c are
composed from the modulation index m and angle ref , formed by
Fast and accurate detection of the negative-sequence components of the utility voltages and currents is a prerequisite for the
proposed ID method. The ANF might be extended to meet this
function. This section proposes an ANF-based signal processor
that has the three-phase synchronization scheme proposed in the
previous section as the main cell. It is developed to extract key
power system information that is required for ID. The input signals
u(t) can be decomposed to positive-, negative- and zero-sequence
components u(t) = u+ (t) + u (t) + u0 (t) and are related to the
input signal u(t) by the following equations [8]:
+
= m sin m +
uaref
+
= m sin m + 2/3
ubref
+
= m sin m + + 2/3
ucref
(11)
0
1
1
0
1
T1 = 21 3 1
1
1
0
(14)
1
0.5 0.5
1
0.5
T2 = 13 0.5
0.5 0.5
1
uaref
= m sin m +
= m sin m + + 2/3
ubref
= m sin m + 2/3
ucref
(13)
(12)
1/3
1
xa.
1/2
4
ua
va
Va+
1/(2*sqrt(3))
1
ua+
6
theta_xa
vb+
3
xb.
m
1
[0 2*pi/3 2*pi/3]
vb
4
theta_xb
2
xc.
2
phi_m
sin
1/(2*sqrt(3))
3
theta
Uref P1
0.005
[0 2*pi/3 2*pi/3]
1
Pulse
vc
1/2
Discrete 3-phase
PWM generator
% Negative
sequence
injection
vc+
5
theta_xc
2
ub+
5
vb
3
uc+
6
uc
1/3
sin
244
1
2
theta
Tng hp cc bi bo khoa hc
giai on NOTCH
2007-2012
ADAPTIVE
FILTER FOR SYNCHRONIZATION AND ISLANDING DETECTION
PQ_DG
100
0
Active power
Reactive power
100
15 000 Hz
900 V
0.8 mH
4m
30 F
1
480 V
200
0.1
0.2
0.3
0.4
0.5
0.6
0.4
0.5
0.6
0.4
0.5
0.6
PQ_load
100
50
0
50
0.1
0.2
0.3
PQ_utility
48830 W
45 891.68 Var
45 891.68 Var
200
0
200
400
60 Hz
600 V
40 MV
7
0.1
0.2
0.3
Time
Ki = 4, Kp = 0.04
Ki = 2, Kp = 0.05
Ki = 2, Kp = 0.05
= 18 000, = 0.6
(a)
1.5
1
0.5
In order to show the performance of the proposed control strategy based on the ANF under grid-connected and islanding operating conditions, the simulation model of the proposed DG system as
in Fig. 1 has been built and evaluated on Matlab/SimpowerSystem
environment. The system parameters of simulation are given in
Table I.
In this case study, the output power of the DG unit is limited to
50 kW. Moreover, the inductance and the capacitance of the load
should be selected so that their resonant frequency equals the fun1
= 2. (60 0.1Hz).
damental frequency of the system, i.e. LC
Satisfying this condition, the overall load circuit has unity power
factor and appears to be purely resistive. Therefore, the load
draws a reactive power which is equal to the DG capacity minus
the loss in the lter and almost no reactive power, i.e. Q =
1
V 2 L
C 0. This scenario represents the worst case for
ID, since no noticeable change in either the voltage or the frequency at the PCC occurs when utility circuit breaker (CB) is
opened. Voltages and currents in this simulation are measured in
p.u.
1
1.5
0.1
0.2
0.3
Time
0.4
0.5
0.6
0.4
0.5
0.6
(b)
2
1.5
1
0.5
0
0.5
1
1.5
The
test system initially operates in a matched power grid-connected
mode. On opening the switch CB , the system is islanded during t =
0.4 and 0.5 s. The interval from 0.0 to 0.1 s is the system transient
initiated time of the simulated system. Figure 12 shows the reactive
and active power of the DG unit, load, and utility. The active power
generated by the DG unit is matched with the local loads power
demand, while the reactive power of load draws almost zero power
because of the condition of the fundamental frequency resonance.
The DG unit controls its output reactive power at zero to meet
unity power factor.
Figure 13 shows the instantaneous point of the common
couplings voltages and the inverter output currents. Mainly due to
the injection of the negative-sequence currents, the PCCs voltages
and the DG units output currents (before and after the islanding
event) are not balanced and cause almost no change. However,
after reclosing the CB, these voltages and currents still deviate
from the normally operated range even in case of nearly matching
power because of the ripple power generated from the DG unit. In
Fig. 13 also, since a tiny amount of unbalanced voltages is injected,
B mn H thng in - i hc Bch Khoa H Ni
0.1
0.2
0.3
Time
0.5
N. D. TUYEN ET AL.
ANF-phase
U+ref
6
4
0.1
10
0.2
0.3
0.4
0.5
0.6
0.1
0.2
0.3
Uref
0.4
0.5
0.6
0.4
0.5
0.6
0.5
0.6
PLL-phase
6
0
5
4
2
0.1
0.2
0.3
0.4
0.5
0.6
0.1
0.2
0.3
Uref
1
0
1
0.1
0.2
0.3
Time
0.4
0.5
0.6
1
0
0.1
0.2
0.3
Time
0.4
380
0.15
370
0.05
360
0.05
0.1
0.1
350
0.2
340
330
0.1
0.2
0.3
Time
0.4
0.5
0.1
0.2
0.3
Time
0.4
0.5
0.15
0.6
0.6
0.02
0.015
0.01
0.005
0
0.005
0.01
0.015
0.02
0.1
0.2
0.3
Time
0.4
0.5
0.6
and DG-
5.2.2. Islanding detection time and tolerant level Figure 18 indicates that before islanding (t < 0.4s), the system
negative-sequence impedance is dictated by the low grid impedance
and remains small, i.e. about 0.06 p.u.
Based on the mechanism of impedance ID, a simple method
is proposed to set tolerance levels so that the protection system
functions properly. As mentioned in Refs 9 and 10, the impedance
ID has a small NDZ, which means that choosing the tolerance
levels can be implemented simply in practical systems. The
sensitivity concept is then referred to. The relationship between
the impedance of grid-connected system and that of islanded entity
should be considered. On one hand, the grid-connected impedances
246
over the tolerance level many times before the islanding truly
occurs. Thus, in those cases, the ID scheme should consider
that no islanding appears. The one-time checking design could
make the ID relay work incorrectly. Therefore, multichecking
is necessary for a proper operation. By contrast, the more the
number of checking times, the longer the detection time is
pushed to increase.
Regarding the tolerance levels, determining the xed value
that is suitable for almost all operation scenarios is never an
easy task. The tolerance level is then determined on the basis
of negative-sequence impedances of the grid-connected mode.
As in Fig. 18, the grid-connected impedance is roughly 0.05
p.u. Therefore, the higher value of the tolerance level should
be chosen. However, how much higher this value should be
above 0.05 p.u is left to the operating engineers. This chosen
tolerance value is a compromise between two things. The rst
one is the disturbance during the grid-connected mode such
as opening induction motors (which also makes the negativesequence impedance to increase [3]) or other perturbation, as
seen in Fig. 18. For instance, if the tolerance level is set at a
small value as much as 0.08 p.u, then at the 0.13 s instance the
increasing impedance phenomenon is detected as an islanding.
And the second one is of high values that even the occurred
island could not bring negative-sequence impedance to reach.
The tolerance level is set at 0.9 p.u is such an example.
In Fig. 19, an example of theoretical analysis shows how to
set up the tolerance levels. This grid is an IEEE 13 bus
system [10]. This system involves multiphase distribution line
segments including three-phase, two-phase, and single-phase
lines. Instead of the innite bus in the previous simulation,
the multiload grid with existing unbalance components is
considered. Modication of this grid was taken into account
in order to be suitable for this analysis. One unit DG is
inserted at node 675. It should be understood that there are
power-consuming loads at every node. Therefore, under this
conguration, once line interruption happens at one of the four
sections (I, II, III, IV as shown in Fig. 19), the islanding takes
place. If islanding happens at section I, the DG unit will supply
for local loads at node 675. The change symptom of negativesequence impedance manifests, which has been disconnected as
shown in Fig. 20. The grid-connected impedance is retained at
the smallest value during the normal operation of the system.
The order of tolerant levels, i.e. TLi (i = 1 to 4), depends on
the respective interruption parts. The highest impedance of the
islanded part occurs when opening section I, while the smallest
islanded impedance is when opening section IV. The tolerance
level of each section is chosen as clearly shown Fig. 20.
2.5
p.u
2
System
impedance
reaches
tolerant
level
1.5
Tolerant level 2
Islanding detected
Tolerant level 1
Checking-time 2
Checking-time 1
0.5
0.1
0.2
0.3
Time
0.4
0.5
0.6
Detection-time2
Detection-time1
Ztolerant level
gridconnected
(15)
Fig. 19. Modied IEEE 13 node using for tolerance level analysis
247
Tng hp cc bi bo khoa hc
giai on NOTCH
2007-2012
ADAPTIVE
FILTER FOR SYNCHRONIZATION AND ISLANDING DETECTION
4.5
4
3.5
Tolerant level
for 0.4 MW
p.u
3
2.5
Tolerant level
for 4MW
2
1.5
7. Conclusion
0.5
0
0.1
0.2
0.3
Time
0.4
0.5
0.6
248
N. D. TUYEN ET AL.
Tng hp cc bi bo khoa hc
giai on NOTCH
2007-2012
ADAPTIVE
FILTER FOR SYNCHRONIZATION AND ISLANDING DETECTION
Speed
Main characteristics
Fast
Fast
Very fast
Positive-sequence impedance
Zpositive =
Harmonic impedance
Zharmonic =
Negative-sequence Impedance
Znegative =
dVpositive
dIpositive
Fast
dVharmonic
dIharmonic
Slow
Vnegative
Inegative
Very fast
References
(1) IEEE-SA-Standards-Board. IEEE STANDARD for Interconnecting
Distributed Resources with Electric Power Systems, IEEEStd 1547
Standard, 2003.
(2) Xu Wilsun S, Mauch K. An assessment of distributed generation
islanding detection methods and issues for Canada. Natural Resources
Canada, Technical Report, 2004.
(3) Wrinch MC. Negative sequence impedance measurement for distributed generator islanding detection. Ph.D. dissertation, University
of British Columbia, 2008.
Complicated implementation
Difcult implementation
Neutralize by interferences from grids
Large NDZ with high Q
Malfunction by other transients
Large NDZ with high Q
Depend on harmonic available levels
Facilitates an autonomous operation
Small NDZ, simple implementation
Sensitive threshold setting
249
Active method
N. D. TUYEN ET AL.
Masakatsu Nomura (Member) received the B.E. degree in electronics engineering and the M.S. degree in electrical engineering, both from Nagoya University,
Nagoya, Japan, in 1975 and 1977,respectively,
and the Ph.D. degree from Meiji University,
Tokyo, Japan, in 2009. Since 1977, he has been
with Meidensha Corporation, where he is currently working in the New Product Development
Group, in Tokyo. His research interests include power electronics,
control, and factory automation. Dr Nomura is a member of the
IEEE and the Japan Society for Simulation Technology.
250
Extended Summary
pp.501509
2
RT PCH4 PO2 T
ln 2
Ecell = E0 kE (T 298) +
E (1)
ne F
PH2 O PCO2 0
4
B mn H thng in - i hc Bch Khoa H Ni
191
Paper
That ever increasing electricity consumption, progress in power deregulation, and rising public awareness for environment have created more interest in fuel cell distributed generation. Among dierent types of fuel cells, solid oxide
fuel cells (SOFCs) manifest themselves as great potential applications due to many advantages such as low emission,
high eciency, and high power rating. On the other hand, SOFC systems are benecial because they can convert fuel
such as natural gas (almost CH4 ) which is supplied by widespread systems in many countries into electricity eciently
using internal reforming. In facts, the load demand changes exibly and fuel cell life time decreases by rapid thermal
change. Its lifetime may be extended by maintaining in appropriate temperature. Therefore, it is important to acquire
the load following performance as well as control of operation temperature. This paper addresses components of the
simple SOFC power unit model with heat exchanger (HX) included. Typical dynamical submodels are used to follow
the variation of load demand at a local location that considers temperature characteristics using the Matlab-SIMULINK
program.
Keywords: SOFC, dynamic simulation
1.
the fuel and air directly, the system eciency therefore can
be improved (1) . And overall heat balance inside SOFC power
unit eecting on operating temperature afterward on output
voltage will be pointed out.
Introduction
2.
U min
f
2Kr
iref cf qin
CH4
U max
f
2Kr
re f
max
imin
f c i f c i f c (2)
fc )
min
re f
min
ifc
(i f c < i f c ) (3)
ifc =
i , otherwise
fc
qin
CH4
Kr /U f
1 + f .s
.iref cf (4)
2
RT pCH4 pO2
ln
Ecell = E0 kE (T 298) +
ne F p2H O pCO2
2
(6)
pj =
methodology (8) is used to model the pre-heaters in CH4 SOFC system, which makes use of two non-dimensional
groups: and NTU.
By comparing two heat capacity of hot and cold streams,
the lower and higher values are assigned as Cmin and Cmax , respectively. The ratio of heat capacity rates is then available.
Thus, Cr = Cmin /Cmax and NT U = UA/Cmin ; where, U is
overall heat transfer coecient, which is dened largely by
the system and in many cases it proves to be insensitive to the
operating conditions of the system. With our simulation, we
take U to be a constant value and U = 0.5 [W/(cm2 .K)] with
high pressure gas and A = 20000 [cm2 ] is total heat transfer
area.
The value is dened as the ratio of actual heat transfer
rate (q), and the maximum possible heat transfer rate between
the 2 streams (qmax ): = q/qmax , where the theoretical maximum heat transfer is:
VC = i C
dVC
(Ract + Rcon ) (13)
dt
where the values (T h,in ; T c,in ) are simply the temperature difference between the 2 inlet ports.
We have the NT U relation by these equations with
counter ow and parallel ow, respectively:
1 e[NT U(1Cr )]
ct =
1 Cr e[NT U(1Cr )]
(16)
[NT U(1+Cr )]
1e
pr =
1 + Cr
In this model,
both counter ow and parallel ow types of heat exchangers
are selected for comparison. There are 2 HXs. The rst one,
HX1 is used to increse temperature of air and the second one,
HX2 is used to increse temperature of fuel. To determine the
outlet temperatures of the heat exchangers, which vary with
the inlet conditions, the heat capacity rates of the cold and
hot gas streams are calculated:
( 1 ) HX1-air preheater
78 in
q CN
21 O2 2
Ch [CH4 +N2 +O2 +CO2 +H2 O]
1 U opt
f
qin + 78 qin CN
=
opt
CH4
21 O2 2
Uf
1 U opt
f
out
out
qin
+
O2 C O2 + qH2 O C H2 O + qCO2 C CO2
opt
Uf
Cc [Air] = qin
O2 C O2 +
( 2 ) HX2-fuel preheater
Cc [CH4 ] = qin
CH4 C CH4
Ch [HX2 ] = Ch [HX1 ]
out
Ch [CO2 + H2 O] = qout
CO2 C CO2 + qH2 O C H2 O
in
out
Cc [CH4 + N2 + O2 ] = qout
CH4 C CH4 + qN2 C N2 + qO2 C O2
503
194
Assuming the SOFC conguration as parallel one, the eciency can be calculated:
=
1 e[NT U(1+Cr )]
(18)
1 + Cr
and
q = qmax (19)
Hence,
T h,2 = T h,1
q
;
Ch
T c,2 = T c,1 +
q
(20)
Cc
where
T h,1 = T (operating temperature)
T c,1 =
HX2 out
in
qCH4 + T OHX1
(qout
T CH
O 2 + qN 2 )
4
2
out
in
qout
CH4 + qO2 + qN2
S OFC
T h,2 = T CO
= T HS 2OFC
O
2
GS HHX + H i f c V
dT
+ Kh (T T 0 ) (22)
= GS HS OFC + R f c
dt
The dynamic operating temperature can be derived from
Eq. (22) shown in the following Equation:
T = T ini
T S OFCexhaust
=
T HS 2OFC
O
out
out
out
in
qout
H2 O +qCO2 + T S OFCHX qCH4 +qO2 +qN2
out
out
out
in
qout
H2 O + qCO2 + qCH4 + qO2 + qN2
(21)
t
0
GS HHX + H i f c V
dt
Kh (T T 0 ) GS HS OFC
(23)
1
Rfc
Cr =
in
qin
O2 ex = qO2 1 + Kair
3.
GS HHX + H i f c V = GS HS OFC + Kh (T T 0 )
(24)
Kh =
T 1273
(26)
1273
505
196
Fig. 11. HX operating temperature consideration between Counter and parallel Conguration
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
( 10 )
( 11 )
( 12 )
( 13 )
4.
( 14 )
( 15 )
Conclusion
Appendix
S.H. Chan, C.F. Low, and O.L. Ding: Energy and exergy analysis of simple
solid oxide fuel cell power systems, J. Power Sources, Vol.103, pp.188200
(2002)
F. Jurado: Modeling SOFC plants on the distribution system using identication algorithms, J. Power Sources, Vol.129, pp.205215 (2004)
D. Sanchez, R. Chacartegui, A. Munoz, and T. Sanchez: On the eect of
methane internal reforming modelling in solid oxide fuel cells, J. Power
Sources, Vol.33, pp.18341844 (2008)
S.H. Chan, K.A. Khor, and Z.T. Xia: A complete polarization model of a
solid oxide fuel cell and its sensitivity to the change of cell component thickness, J. Power Sources, Vol.93, pp.130140 (2001)
P.R. Pathapati, X. Xue, and J. Tang: A new dynamic model for predicting
transient phenomena in a PEM fuel cell system, J. Power Sources, Vol.30,
pp.122 (2005)
S. Nagata, A. Momma, and T. Kato: Numerical analysis of output characteristics of tubular SOFC with internal reformer, J. Power Sources, Vol.101,
pp.6071 (2001)
J. Larminie and A. Dicks: Fuel Cell Systems Explained, 2nd ed., Wiley, England (2003)
T. Kuppan: Heat Exchanger Design Handbook, Illustrated ed., Marcel
Dekker, Inc., New York (2000)
M. Uzunoglu and M.S. Alam: Dynamic Modeling, Design, and Simulation
of a Combined PEM Fuel Cell and Ultracapacitor System for Stand-Alone,
IEEE Trans. Energy Conversion, Vol.21, No.3 (2006)
C. Wang and M.H. Nehrir: A Physically Based Dynamic Model for Solid
Oxide Fuel Cells, IEEE Trans. Energy Conversion, Vol.22, No.4, pp.887
897 (2007)
P. Piroonlerkgul, W. Kiatkittipong, A. Arpornwichanop, A. Soottitantawat,
W. Wiyaratn, N. Laosiripojana, A.A. Adesina, and S. Assabumrungrat:
Integration of solid oxide fuel cell and palladium membranereactor:
Technical and economic analysis, Int. J. Hydrogen Engergy, Vol.34, No.9,
pp.38943907 (2009)
EG and G Technical Services, Inc.: Fuel Cell Handbook, 7th ed., Morgantown, West Virginia, USA: U.S. Department of Energy (2004)
C. Wang, M.H. Nehrir, and S.R. Shaw: Dynamic Models and Model Validation for PEM Fuel Cells Using Electrical Circuits, IEEE Tran. Energy
Conversion, Vol.20, No.2, pp.442451 (2005)
http://www.engineeringtoolbox.com/specic-heat-metals-d 152.html
P. Aguiar, D. Chadwick, and L. Kershenbaum: Modelling of an indirect
internal reforming solid oxide fuel cell, Chemical Engineering Science,
Vol.57, pp.16651677 (2002)
1.
Nomenclature
v or V: Voltage, V
E: Voltage potential, V
P: Power, kW
Pdemand : Required power of load, kW
R: Resistance,
R = 8314: Universal gas constant, J/(kmol.K)
R f c : Heat capacity of SOFC power unit, J/K
C: Capacitor or Heat capacity rate
i: Output current density, A/cm2
T : Operating temperature, K
T 0 = 293: Ambient temperature, K
T ini = 1273: Initial temperature, K
p: Pressure, atm
q: Flow rate, kmol/s or heat transfer rate
F = 96487: Faraday constant, C/mol
N0 = 384: Number of cells connected in series
U f : Fuel Utilization, %
Kr : Constant, mol/C
ne = 8: Number of electrons are transfer between
layers respect to 1 mol CH4
K j = q j /p j : Valve molar constant, kmol/(s.atm)
507
198
References
T : Temperature
2.
T OHX1
: Temp. of O2 after HX1, K
2
3
3H2 + O2 3H2 O (A2)
2
1
CO+ O2 CO2 (A3)
2
from
SOFC, K
Combining these three Eqs. (A1), (A2) and (A3) match the
combustion reaction of methane:
Greek letters
j = Van(ca) /(K j RT ): Time constant in reaction, s
Van(ca) : Volume of anode or cathode, m3
1 , 2 : Empirical coecients of activation loss
: Eectiveness, %
Subscripts
j: CH4 , O2 , H2 O, CO2
f: fuel
fc: fuel cell
e: electron
act: activation
con: concentration
c: cool
h: heat
ct: counter
pr: parallel
SOFC: Solid Oxide Fuel Cell
HX: Heat Exchanger
GSH: Gas Sensitive Heat
ini: initial
Nguyen Duc Tuyen (Non-member) received the B.S. in electrical engineering from Hanoi University of Technology, in
2006 and M.S. from Shibaura Institute of Technology,
in 2009. He is now the Ph.D. Candidate at Shibaura
Institute of Technology, Japan. Currently, he is a Faculty Member in the Department of Electrical Power
System, Hanoi University of Technology, Vietnam.
His current research interests include modeling and
control of distribution power systems and dispersed
power generation systems.
Superscripts
r: react
min: minimum
max: maximum
opt: optimal
re f : reference
in: input
out: output
508
B mn H thng in - i hc Bch Khoa H Ni
S OFC
T CO
: Temp. of CO2 exhaust from SOFC, K
2
T HS 2OFC
O : Temp. of H2 O exhaust from SOFC, K
T S OFCHX : Temp. of non-reactants exhaust
Time delay
Masakatsu Nomura (Member) received the B.E. degree in electronics engineering and the M.S. degree in electrical engineering from Nagoya University, Nagoya, Japan, and
the Ph.D. degree from Meiji University, Tokyo, Japan
in 1975, 1977 and 2009 respectively. Since 1977, he
has been with Meidensha Corporation, where he is
currently working in the New Product Development
Group, Tokyo, Japan. His interests are power electronics, control and factory automation. Dr. Nomura
is a member of IEEE, IEEJ and Japan Society for
Simulation Technology.
B130 5 2010
B mn H thng in - i hc Bch Khoa H Ni
509
200