Electric Power Systems Research 62 (2002) 67 /80

www.elsevier.com/locate/epsr

Impact of distributed generation on protection of single wire earth

return lines
M.A. Kashem *, G. Ledwich
School of Electrical & Electronic Systems Engineering, Queensland University of Technology, Brisbane, Australia

Received 16 October 2001; received in revised form 23 January 2002; accepted 25 January 2002

Abstract

Distributed generation (DG) inclusion within the grid system potentially introduces problems related to control, protection,
harmonics, and network transients. This paper analyses one of the key issues: protection of the network, by ascertaining the impact
of rotary DG inclusion on existing protection system of SWER (single wire earth return) lines and the DG sensitivity during faults.
The analysis is carried out by estimating fault-sensitivity for the worst-case situation, determining the DG impact on the existing
protection scheme, and comparing the network situation with and without DG during the fault. A model of arc voltage is used to
represent a fault on a SWER scheme. The size of DG is selected based on the SWER capacity and SWER load. The study is
conducted on an example SWER system by considering the SWER lines with and without DG and faults on the SWER backbone
and laterals, and simulation results are reported. In every case studied, the fault current from the DG significantly exceeded the DG
rating and the DG would have tripped. Thus the system reverts to the case with no DG. Even if DG did not trip, the fault current
from the source would be largely independent of the DG, and thus the original feeder protection would continue to provide the same
quality of performance. Hence, net sensitivity and existing protection system will not be adversely affected by DG inclusion in
SWER lines. # 2002 Elsevier Science B.V. All rights reserved.

Keywords: Distributed generation; Protection; Fault sensitivity and SWER System

1. Inroduction Ackermann et al. [1] have discussed the relevant DG

issues, and aimed at providing a general definition for
Power utilities invest large amounts of money in distributed power generation in competitive electricity
expanding transmission and distribution facilities. This markets. Environmentally benign generation technolo-
is necessary although such existing facilities are gener- gies are seen by regulators as environmental ‘adders’
ally under-utilised as they are built for infrequent but and able to attract higher returns or equity for clean or
large peak loads. To meet an expansion of peak load the renewable energy [2]. Barram [3] has focused on the
normal methods include for line re-conducting, more development of a model that attempts to optimise the
transformer installed capacity and new feeders or lines. design of an integrated remote area power supply to
With DG, peak load can be shaved by operating the provide a least cost of the life cycle and minimise
local generating units. With low penetration of DGs, the environmental impact. Woods [4] has identified issues
system impact might be negligible. But as penetration that a Network Service Provider (NSP) should be aware
levels increases the impact will become significant. DG of relative to the design and grid connection of
inclusion in distribution networks is an area for research embedded generation into a distribution network. In
as it raises the difficulties in control, protection, [5], a study has been conducted for interconnecting DG
harmonics, network transients, etc. into utility grids. The focus of this study was the
potential application of DG within the traditional
vertically integrated utility industry structure. Alderfer
* Corresponding author. Tel.:
/61-7-3864-9123; fax
/61-7-3864-
etl al. [6] have reviewed the barriers that DGs are
1516. encountering when attempting to interconnect to the
E-mail address: [email protected] (M.A. Kashem). electrical grid and proposed a ten-point action plan for
0378-7796/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved.
PII: S 0 3 7 8 - 7 7 9 6 ( 0 2 ) 0 0 0 3 1 - 7
68 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80

reducing barriers to DG. In [7], potential market issues, vi) SWER line capacitance is neglected in the fault
merits and demits of DG implementation for Utilities study.
and Customers have been discussed. The authors in [8] vii) The size of DG is generally selected based on the
have presented a numerical algorithm for arcing faults SWER capacity and SWER load. For this study,
detection and fault distance estimation on overhead DG size is chosen from the range of 25 /300% of
lines. maximum amount load lumped at SWER end that
Single Wire Earth Return (SWER) distribution lines causes 6% voltage drop at its connection point.
are typically constructed in rural areas to supply
electricity to remotely located customers. The revenue
from SWER customers is a small gross return and not
well enough for capital investment for network aug- 2. Modelling of SWER line and DG sensitivity
mentation or reconductoring. SWER is a single wire
system using the ground as a return path. Therefore, an The main supply at the connection of SWER recloser
isolating transformer is generally installed to make sure is modelled as a Thevnin equivalent to ascertain the
that no earth leakage current can flow in three-phase sensitivity of the recloser during a fault. The Thevnin
and single-phase lines which feed a SWER scheme. voltage is assumed the same as the substation voltage,
In this paper, a study is conducted on a SWER system which is 19.1 kV and Thevnin impedance is obtained
to evaluate the impact on existing protection scheme from the short-circuit MVA level at that point. A L-G
and DG sensitivity during faults. Rotary DG with a fault is modelled on the SWER scheme as an arc voltage
synchronous generator is considered for this study. The limited to 8 kV at the fault. Arc voltage is in phase with
following assumptions are made during the analysis. fault current [8]. The study is carried out on an example
remote SWER system by considering the cases: (1)
i) The maximum permitted voltage rise or drop in SWER system without DG, (2) DG connected at
SWER lines is 6%. SWER end, (3) DG connected at SWER line but not
ii) The existing protection system is capable of protect- at the end, and (4) DG connected at the end of SWER
ing the SWER system (without having DG) in all line but fault in SWER lateral.
abnormal situations (including high impedance
fault). This means that, in the absence of DG,
SWER breakers and reclosers are assumed capable 2.1. Case 1: SWER line without DG
of detecting of a fault in the SWER system and
isolating of faulty part. Fig. 1 shows a SWER line containing an automatic
iii) Fault-voltage at the point of fault in the SWER circuit recloser (ACR) and L-G fault at the end. A
feeder is always lower than or equal to 8 kV. The simplified diagram with modelling of source and L-G
fault is modelled with arc resistance and treated as fault is shown in Fig. 2. The formula for fault current
constant voltage source [8]. contributed by main source has been derived below. L-G
iv) Loads are disconnected to estimate worst-case fault is created at various points of SWER line to
situation during fault. It is noted that load inclusion determine worst-case situation. Phasor diagram for this
will increase the measured current during faults. situation is shown in Fig. 3.
v) The phase difference between DG and source Total impedance for source and whole line is
voltages is very small during a fault and hence it is
Z  (Rs
jXs )
(RL
jXL ) (1)
neglected. Therefore, DG and source voltages are
assumed identical. Justification has been discussed Applying Sine rule (a/sin A /b/sin B /c/sin C ) in
in Case 2. the triangle of Fig. 3, we get

Fig. 1. SWER line and L-G fault.

 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80 69

Fig. 2. Simplified diagram of Fig. 1.

Vs V IZ assumed to be Vs /19.1
/j0 kV and source impedance

 f  f (2)
sin 8 sin b sin a is Zs /88.16
/j 72.16 V during fault. N /20 locations
have been identified in the SWER line to analyse L-G
Line resistance and reactance are known and hence u fault and determine sensitivity of existing SWER
(and 8 ) can be obtained. Vs and Vf are given. b , a and If protection system.
can be calculated from Eq. (2) as follows, Fig. 4 shows that fault current decreases when

 distance from the source increases and it is very low
1 Vf sin 8
b sin (3) for faults at the end point of SWER line. Therefore, the
Vs
worst-case situation for protection will be when a fault
a pb 8 (4) occurs on the end of the SWER line. It may be difficult
Vs sin a to protect the network in this situation if load current
If (kA) (5) magnitude is very close to fault current. In this study, we
Z sin 8
have assumed that existing protection system can
N points, including load buses, are identified in the protect SWER network at any situation. Our main
above SWER line in Fig. 1. At each point, a L-G fault is objective is to investigate whether the connection of DG
modelled with the suitable fault impedance to obtain 8 makes the situation of SWER network worse, especially
kV fault-voltage and all the points are investigated during a fault.
individually. The sensitivity of ACR during the fault is observed. It
is found that the critical point of worst-case situation is
2.1.1. Analysis results for case 1 the end point of the SWER line, at which the amount of
A 60 km SWER network has been modelled with line fault current is low but may be sufficient to cause ACR
impedance Zl /1.828
/j0.876 V/km. Source voltage is to respond in the existing protection scheme.

2.2. Case 2: DG connected at SWER end

Fig. 5 shows a SWER line with DG connected at the

end. A L-G fault is modelled in SWER line to estimate
the main source and DG sensitivity. A simplified
diagram to reflect the sensitivity during L-G fault is
shown in Fig. 6. Impedance for main source and the
portion of line from connection point to the fault is
Zsf (Rs
jXs )
(R1
jX1 ) (5?)

Impedance for DG and the portion of line from DG

to the fault is
Zdgf (Rdg
jXdg )
(R2
jX2 ) (6)
Fig. 3. Phasor diagram during fault. Total impedance is
70 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80

Fig. 4. Sensitivity of existing protection system.

Zsf Zdgf Zsf

Z (7) Idg  If (9)
Zsf
Zdgf Zsf
Zdgf
In the absence of DG, Thevnin equivalent voltage at Is  If Idg (10)
the fault point should lie in the range of 0.94 /1 p.u. One N points (including load buses) are identified in the
of the aims of DG inclusion is to improve the voltage SWER line shown in Fig. 5. At each point, a L-G fault is
profile of the network, and therefore the equivalent modelled with the suitable fault impedance to obtain 8
voltage at the fault will be improved by the presence of kV fault-voltage and all points are examined individu-
DG and it will be closer to 1 p.u. It is also noted that if ally. The location of the DG remains fixed to determine
phase difference between DG and source voltages the level of sensitivity. The sensitivity of ACR and DG
increases, the magnitude of Thevnin equivalent voltage during fault is observed. Both the main source and the
decreases. As the maximum voltage drop is 6% in the DG will contribute fault current toward fault, and
SWER systems, the Thevnin equivalent voltage at the therefore the magnitude of overall fault current will be
fault point cannot be less than 0.94 p.u. Thus, modelling increased.
of the Thevnin source at the fault point as 1 p.u. will be
more true in the presence of the DG and will cause
negligible difference in the detectability of the fault. As 2.2.1. Analysis results for case 2
main source and DG voltage are assumed identical, we The same test system as described in Case 1 is used to
get analyse L-G fault in the presence of DG at SWER end
Vs Vdg (8) in Fig. 5. The same locations as in Case 1 have been
identified and for each location the size of DG has been
The same phasor diagram (shown in Fig. 3) can be changed, from 25 to 300 kVA with a step of 25 kVA, to
used to analyse fault for SWER line with DG and Eqs. observe the level of sensitivity during fault. These sizes
(2) /(5) are used to determine fault current, If. Fault have been selected from the range of 25/300% of
currents contributed by DG and main source are maximum load lumped at SWER end which causes 6%
calculated as, voltage drop at the connection point. The maximum

Fig. 5. SWER line with DG and L-G fault.

 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80 71

Fig. 6. (a) Simplified diagram with L-G fault and DG. (b) Simplified diagram for identical voltage source with L-G fault and DG.

tolerable load is calculated in Appendix A and found to Fig. 7(a) shows the fault currents for different fault
be 100 kVA for this sample SWER system. Thus a 25% locations with 25 kVA DG installed at SWER end. The
load corresponds to 25 kVA. The calculations in similar results for different DG sizes are reported in
Appendix A show that the DG rating is always lower Appendix B (Figs. B1 and B2) to visualise the impact of
than 20% of the short-circuit MVA at the connection DG size. DG is located at point 20 and hence DG rated
point (Fig. A1). The DG voltage is assumed the same as current is shown at point 20 in these figures. Fig. 7(b)
source voltage and DG impedance is Zdg /0
/j1.76 p.u. summarises all the above figures and indicates the
on its own base. reflection of fault currents for different DG sizes due

Fig. A1. Modelling of SWER system with load.

72 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80

Fig. 7. Fault currents in Case 2: (a) 25 kVA case (b) different DG sizes.

to a L-G fault at SWER end (point 20). From Fig. 7 and situations, total fault current decreases when the dis-
Figs. B1 and B2 (in Appendix B), it is seen that DG tance of fault location from source increases. However,
rated current is very much low compared with DG fault for DG size 100 kVA or higher DG fault contribution
current. Also, source fault current most likely remains increases and it is appreciably high for large size of DG.
the same as the fault current without having DG in Source fault current always decreases when distance of
SWER line, which is also the source fault current in the fault location from source increases. It is observed that
absence of DG. This means that the same setting of for DG size above 200 kVA the fault current
ACR or overcurrent relay will work for both cases, with contributed by the source increases compared with the
DG and without DG. In all the above cases DG fault current without DG. Therefore, it can be con-
sensitivity is high during fault and DG will be discon- cluded that the existing protection scheme will not be
nected based on its own overcurrent relay. For the size affected and no modification in the protection system
of DG less than 100 kVA, DG fault contribution is will be required even though DG is installed in SWER
almost constant, irrespective of fault location. In these network.
 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80 73

Fig. B1.

2.3. Case 3: DG connected at SWER line but not at the Impedance of the portion of line from DG point to
end the SWER end is
Z2 R2
jX2 (13)
In this case, DG is connected at a point of SWER line
far from SWER recloser but not at the SWER end. Fig. Total impedance is
8 shows a typical diagram for such a case. L-G fault at a
point between ACR and DG can be analysed in similar Zsl Zdg
Z
Z2 (14)
way as described in Case 2 and the sensitivity in this part Zsl
Zdg
of SWER will be high to respond during fault. Fault at a
point in SWER line from DG connection to the SWER Same phasor diagram (shown in Fig. 3) can be used to
end will be analysed here intensively. Fig. 9 shows the analyse fault at SWER line with DG and Eqs. (2) /(5)
impedance diagram with the model of 8 kV fault. are used to determine fault current, If. Fault currents
Impedance for main source and the portion of line contributed by DG and main source are calculated as,
from source to the fault is Zsl
Idg  If (15)
Zsl (Rs
jXs )
(R1
jX1 ) (11) Zsl
Zdg
Impedance of DG is Is  If Idg (16)
Zdg Rdg
jXdg (12)

Fig. B2.
74 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80

Fig. 8. SWER line and DG with L-G fault.

2.3.1. Analysis results for case 3 DG and source, fault current without DG, all are
The test system of Case 2 is used but DG location has decreasing as impedance increase with the distance,
been moved toward source, 20 km far from the SWER from DG location toward SWER end. Therefore, the
end and 40 km far from the source. The procedure point of worst-case situation is the SWER end at which
discussed in Case 2 has been employed to test the all of them are low. At this point, DG fault current is
sensitivity during a fault on the portion of SWER line very high compared with its normal current and it will
(20 km from SWER end), between DG connection and be tripped off by overcurrent relay during fault. A L-G
SWER end. Here also 20 points have been selected to fault is modelled at SWER end (point 20) and then DG
analyse L-G fault and at each point L-G fault is size is changed from 25 to 300 kVA to observe the trends
modelled using arc resistance in such way that 8 kV of fault currents for various DG sizes, and the results are
fault voltage is attained at the point of fault. The size of summarised in Fig. 10(b). In Fig. 10(b) DG fault current
increases with the increase of DG size and causes total
generator is changed from 25 to 300 kVA to observe the
fault current to increase gradually with respect to DG
fault current contributed by DG and source and
size. Fault current contributed by source and fault
ascertain the sensitivity.
current without DG, both are almost same in magnitude
A L-G fault at various points on the portion of
(especially for small DG) in all the above figures and
SWER line (from DG connection to SWER end) in the
therefore existing protection system will not be affected
presence of 25 kVA DG is examined and the results are
due to DG inclusion.
shown graphically in Fig. 10(a). Fig. C1 and Fig. C2 in
Appendix C show the trends of fault currents for
different DG sizes. These figures show total fault 2.4. Case 4: DG connected at the end of SWER line but
current, DG fault current, source fault current, fault fault at SWER lateral
current without DG at various fault locations. DG is
connected at 0 location and therefore DG normal Fig. 11 shows a model of SWER line with its lateral.
current is shown at zero location in all figures. In this DG is connected at main SWER end. L-G fault is
case, total fault currents, fault currents contributed by modelled at various points of SWER lateral. SWER line

Fig. 9. Simplified diagram of network given in Fig. 8.

 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80 75

Fig. 10. Fault current sin Case 3: (a) 25 kVA case (b) different DG sizes.

and its lateral (without having DG in SWER network) Zdgl (Rdg
jXdg )
(R2
jX2 ) (18)
are protected by the existing protection system. It is
Impedance of the lateral from the connection point to
assumed that no current will flow in SWER lateral if
the fault is
there is a fault at SWER backbone. Therefore, at the
presence of DG, L-G fault at SWER backbone can be Z3 R3
jX3 (19)
analysed in similar manner as discussed in Case 2. L-G
fault at lateral will only be investigated in this case. L-G Total impedance is
fault in SWER lateral will be modelled with suitable arc Zsl Zdgl
resistance to achieve 8 kV fault voltage at the point of Z
Z3 (20)
Zsl
Zdgl
fault. Fig. 12 shows a simplified diagram of SWER
impedance, DG and L-G fault. The same phasor diagram (shown in Fig. 3) can be
Impedance for main source and the portion of line used to analyse fault at lateral with DG at the end of
from source to the connection of lateral is SWER line and Eqs. (2) /(5) are used to determine fault
current, If. Fault currents contributed by DG and main
Zsl (Rs
jXs )
(R1
jX1 ) (17) source are calculated as,
Zsl
Impedance of DG and the portion of line from DG to Idg  If (21)
the connection of lateral is Zsl
Zdgl
76 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80

Fig. C1.

Fig. C2.

Is If Idg (22) at each point L-G fault is modelled using a suitable arc
resistance to obtain 8 kV fault voltage at the fault. The
size of generator has been changed from 25 to 300 kVA
to observe the sensitivity during fault.
2.4.1. Analysis results for case 4 DG is located at SWER end but rated DG current
In this study, the test system of Case 2 has been used (calculated at DG terminal) is shown at point 0, (which
but a 10 km lateral is connected at a point in SWER is 20 km far from DG terminal), with a symbol ‘o’ in
backbone which is 20 km far from the SWER end and Fig. 13(a) and Figs. D1 and D2 (in Appendix D). As the
40 km far from the source. The same line parameters point 0 is 20 km far from DG terminal (or SWER end),
have been used for the SWER line and its lateral. DG is the actual DG rated current at this point will be lower
connected at the end of SWER backbone. SWER lateral than the rated DG current at DG terminal. Figs. D1 and
will not be involved in calculations if there is a fault in D2 shows the same fault currents as Fig. 13(a) but for
SWER backbone and this situation has been well different DG sizes. Fault current contributed by DG is
presented in Case 2. Therefore, fault at SWER lateral relatively high; it does not matter where the fault is
has only been simulated in this study. 20 points have located in SWER lateral. Hence, overcurrent relay in
been selected on SWER lateral to analyse L-G fault and DG system will respond and DG will be disconnected
 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80 77

Fig. 11. SWER line with DG and L-G fault.

from the SWER system. These figures shows that fault has been carried out by considering a prototype SWER
currents decreases when the distance to the fault line with and without lateral. L-G fault has been
location increases. It is also seen in these figures that examined at various points in the line with and without
the fault current contributed by the source in the the presence of DG. Four phototype test systems have
presence of the DG and fault current in absence of been considered for this study and all of them are
DG, are both almost coincident. Fig. 13(b) highlights created by considering the worst situations of actual
the trends of the fault currents for different DG sizes SWER systems. From the analysis it is found that DG
with a fault at far end in SWER lateral. From Fig. 13(b) connection does not adversely affect the existing protec-
it is seen that DG size does not adversely affect the tion system, and the existing relay settings and coordi-
source fault current that remains almost same. Only nation can be retained. DG may be installed in a SWER
total fault current will increase or decrease if DG size system without any adverse affect on existing SWER
increases or decreases, respectively. Therefore, DG protection and SWER augmentation can be avoided.
inclusion will not affect existing protection scheme and DG may use overcurrent relay or ACR to detect L-G
setting and which may be retained after DG installation. fault with (almost) similar setting followed by the
nearest SWER ACR provided that existing overcurrent
relay and ACR protect SWER system in any abnormal
3. Conclusions situations. The results of this study highlight that fault
current contributed by DG is many times higher than
Impact of small size DG inclusion on SWER protec- the DG normal current and fault location causes a little
tion scheme have been explored in this study. The study change in fault currents. In all the four cases, DG

Fig. 12. Simplified diagram of network given in Fig. 11.

78 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80

Fig. 13. Fault currents in Case 4: (a) 25 kVA case (b) different DG sizes.

sensitivity is analysed during fault at various points on personnel for their cooperation in providing data and
SWER systems and is found high in all cases. It is seen advice on the operation of SWER lines.
that DG inclusion does not reduce the main source
sensitivity or ACR sensitivity during fault. If existing
protection system can detect high impedance fault, DG Appendix A: DG size relating with maximum load and
protection will also be able to detect. It is found that for short-circuit MVA
DG less than feeder rating, its size does not significantly
reduce the existing net sensitivity or adversely affect the Applying voltage divider rule in Fig. A1 we obtain
existing protection scheme.
Zload
Vs Vload (A:1)
Zsource
Zline
Zload
According to SWER rule maximum voltage drop in
Acknowledgements SWER system is 6%, which normally occurs at SWER
end. All the loads are lumped at SWER end to
The authors gratefully acknowledge the financial determine the maximum tolerable load at the worst-
support from Queensland Electricity Transmission and case situation. Therefore, load voltage in this situation is
Distribution (QETD). Special thanks to Ergon Energy 94% of source voltage Vs. Source voltage and source
 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80 79

Fig. D1.

Fig. D2.

impedance are calculated from Thevnin equivalent Solving the above quadratic equation, we get R /
which are 19.1 kV and 88.16
/j 72.16 ohm, respectively. 3148 V
A typical SWER line’s impedance is 1.828
/j0.876 V/km Therefore, Zload /3148
/j0 V and Load kVA /107.8
and length is 60 km considered for this study. The two kVA.
following cases are considered to obtain the rough idea
of approximate maximum load. A.2. Case-A2: 0.8 p.f. (lag) Load

Let, Zload /R
/jX

A.1. Case-A1: pure resistive load For 0.8 p.f. load,
R
Let, Zload /R
/j0, Eq. (A.1) can be written as, pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi cos u 0:8 (A:4)
R
X2
2

j R
(Rsource
Rline
R)
j(Xsource
Xline ) j
Vs j0:94Vs j Rearranging Eq. (A.4), we get
R 1:33X (A:5)
(A:2)
Squaring both sides of the above equation, substitut- Eq. (A.1) can be written as,
ing the values for resistance and reactance of source, and
rearranging, we obtain j R
jX
(Rsource
Rline
R)
j(Xsource
Xline
X ) j
Vs j0:94Vs j

R 2 3015:52R416819 0 (A:3) (A:6)

80 M.A. Kashem, G. Ledwich / Electric Power Systems Research 62 (2002) 67 /80

Squaring both sides of the above equation and Appendix C: DG connected at SWER line but not at the
substituting R /1.33X and the values for resistance end (case 3)
and reactance of source and line, and rearranging, we
get Figs. C1 and C2.
2
X 2154:7X 151933:6 0 (A:7)
Solving the above quadratic equation, we get Appendix D: DG connected at the end of SWER line but
X /2223 V fault at SWER lateral (case 4)
From Eq. (A.5), R /2956.65 V
Therefore, Zload /2956.65
/j2223 V and Load Figs. D1 and D2.
kVA /91.76 kVA
Based on the above load calculation in Case-A1 and
Case-A2, the approximate maximum tolerable load is References
assumed to be 100 kVA.
[1] T. Ackermann, G. Andersson, L. Soder, ‘What is Distributed
A.3. Short-circuit level at the SWER end Generation?’, Royal College of Technology, Dept. of Electric
Power Engineering Electric Power Systems, Teknikringen 33,
VTH /Vs /19.1 kV 10044 Stockholm, Sweden, 1999.
[2] S.W. Chapel, L.R. Coles, J.J. Iannucci, ‘Distributed Utility
ZTH /Zsource
/Zline /197.84
/j124.72 V Valuation Project /Monograph’, Electric Power Research Institute
2 (EPRI) Report TR-102807, National Renewable Energy Labora-
jVTH j (19:1)2
Short-circuit MVA   1:56 MVA tory (NREL) Pacific Gas & Electric Company (PG&E) Report
jZTH j 233:87 005-93.12, 1993.
[3] F. Barram, ‘RAPS Design for Large Properties in Remote
Locations’, ESAA Renewable Energy Technologies and Remote
Power Supplies Conference, Cairns, Queensland, Australia, 1999.
A.4. DG size [4] G. Woods, ‘Embedded Generation */A Network Service Provider
Perspective’, Proceedings of Fifth International Transmission and
Distribution Conference & Exhibition (D2000), Brisbane, Queens-
Maximum DG size considered in this study /300 land, Australia, 9 /12 November 1999.
kVA /19.23% of Short-circuit MVA at DG con- [5] National Association of Regulatory Utility Commissioners,
nection /300% (approximate) of the maximum toler- ‘Review of Utility Interconnection, Tariff and Control Provisi-
able load. ons for Distributed Generation /A report to the NARUC
Minimum DG size considered in this study /25 Committee on Energy Resources and the Environment’, January
2000.
kVA /1.6% of Short-circuit MVA at DG con- [6] R.B. Alderfer, M.M. Eldridge, T.J. Starrs, ‘Making Connections /
nection /25% (approximate) of the maximum tolerable Case Studies of Interconnection Barriers and their Impact on
load. Distributed Power Projects’, National Renewable Energy Labora-
tory, Colorado, May 2000.
[7] Distributed Generation Report-2000, Energy Info Source, 1st
Edition, August 2000.
Appendix B: DG connected at SWER end (case 2) [8] M.B. Djuric, Z.M. Radojevic, V.V. Terzija, Time domain solution
of fault distance estimation and arcing faults detection on overhead
Figs. B1 and B2. lines, IEEE Trans. On Power Delivery 14 (1) (1999).