Downloaded from http://iranpaper.
ir
Paper accepted for presentation at 2009 IEEE Bucharest Power Tech Conference, June 28th - July 2nd, Bucharest, Romania
Impact of Distributed Generation in Steady
State, Voltage and Transient Stability–Real Case
R. R. Londero, C. M. Affonso, M. V. A. Nunes
Abstract— This paper investigates the influence of a
synchronous distributed generation on a Brazilian real network.
The distributed generation represents a small hydro power plant
with total capacity of 30MW. The model of the network is
developed in a professional computer software package.
Simulations are carried out considering voltage dependency of
static loads and different DG penetration levels. The technical
aspects analyzed are steady-state voltage profile, electrical power
losses, voltage and transient stability. Also, some adequacy
aspects are analyzed, related to national and international
standards. It was found that the distributed generation enhanced
the overall system performance.
Index Terms— Distributed generation, synchronous machine,
steady-state voltage profile, transient stability, voltage stability.
I. INTRODUCTION
I N recent years, many distributed generators (DG) have been
installed in power systems worldwide, and many more units
are expected to be installed in the near future because of their
advantages over conventional generation. The European
Union goal is a 22% production from renewable energy
sources in 2010 [1]. This scenario is motivated by the
increasing concern over environment impact and the need for
eliminating the unnecessary transmission and distribution
costs.
Examples of renewable energy sources are wind turbines,
photovoltaic systems, biomass, fuel cells and small hydro
power plants. In particular, small hydro power plants have
obtained increasing interest due to their acceptable prices for
generating electrical power without producing harmful
pollution and green-house gases and their relatively low
environmental impact compared to large hydro power plants.
However, the connection of the DG has both benefits and
drawbacks of the technical, economical and environmental
aspects [2]. A high DG penetration level may influence the
whole system operation and control, leading to technical
impacts that must be identified [3, 4]. Then, such aspects must
be analyzed to prevent instability problems and assure
adequate system voltage and frequency levels, regarding
customers and utility interests.
In this context, this paper presents a complete investigation
Manuscript received October 31, 2008. This work was supported by
CNPQ, Brazil.
R.R.Londero, C.M.Affonso and M.V.A.Nunes are with the Faculty of
Electrical Engineering, Federal University of Para, Belem, PA, 66075-110
Brazil (e-mail: rafaelrorato@gmail.com, carolina@ufpa.br, mvan@ufpa.br).
978-1-4244-2235-7/09/$25.00 ©2009 IEEE
http://www.itrans24.com/landing1.html
of DG penetration level influence on system technical aspects
such as: steady-state voltage profile and losses, voltage and
transient stability. Also, some adequacy aspects are analyzed,
related to national and international standards. These aspects
are analyzed in an existing Brazilian system with real data by
using load flow program, continuation method and non-linear
time-domain dynamic simulations.
This paper is organized as follows. Section II describes the
system and models used to represent the main networks
components. The methodology used to carry out the
simulations is also explained in this Section. Section III
presents and discusses the simulation results. Finally, the
conclusions are summarized in Section IV.
II. NETWORK TOPOLOGY AND MODELING
A. System Description
The test system used in this paper is part of a real power
system sited at the North region of Brazil, and is
interconnected with the main transmission Brazilian system.
Its one-line diagram in presented in Fig. 1.
The main system is composed by 230kV and 138kV
transmission lines and has two hydroelectric: TUC, the main
generation with total capacity of 8.1 GW, and CUNA, a
distributed generation with total capacity of 30MW. The main
load points, which are indicated in the figure and represent
small cities, are: TUC, ALT, RUR, ITA, TAP and STA. The
system was divided in three different areas for clarity
purposes, as shows Fig. 1.
Table 1 presents system active power demand to each area
for different load pattern. Since the more pessimist situation is
the high load, the studies presented in this paper were
conducted to the high load pattern.
TABLE I
SYSTEM DEMAND (MW)
Area Area 1 Area 2 Area 3
Light load 19.0 28.2 89.0
Medium load 27.1 40.7 92.4
High Load 29.0 43.6 100.4
 Downloaded from http://iranpaper.ir
Fig.1. Single-line diagram representing part of the northern region of the Brazilian system.
B. Distributed Generator Modeling
The small hydro power plant (CUNA) is modeled as
synchronous salient pole generators with 2 machines of
12.5MVA and 1 machine of 10.84MVA, represented by a
sixth-order model in the electromagnetic transient simulations
[5]. The generator nominal voltage is 6.9kV with nominal
frequency of 60Hz.
There are two different modes of controlling the excitation
system of distributed synchronous generators. One aims to
maintain constant the terminal voltage (voltage control mode),
and the other one aims to maintain constant the power factor
(power factor control mode) [2]. Power factor control mode is
usually adopted by independent producers to maximize the
active power production. In this paper, the voltage control
mode is adopted following the utility practices. Voltage
regulators are included and modeled with an IEEE type 1
simplified model [5] and speed regulators according to [6].
C. Simulations
The DG penetration level (PL) can be calculated as a
function of the total DG power generation over the total load
demand:
PDG
PL (%) = × 100
∑ PL
http://www.itrans24.com/landing1.html
To avaliate the impact of the DG penetration level, the
following cases are considered:
• Case 1: PL = 15.6% (DG = 27 MW);
• Case 2: PL = 12.1% (DG = 21 MW);
• Case 3: PL = 8.6% (DG = 15 MW);
• Case 4: PL = 0% (DG = 0 MW).
In all cases, both active and reactive power demand of the
loads are kept constant. Thus, the power required from the
main generator TUC decreases with the increase of the DG
penetration level.
The simulations are performed by using the softwares
ANAREDE for power flow and PV curves analysis and
ANATEM for electromechanical transients analysis, both
developed by CEPEL [7,8].
The PV curves are obtained by varying the active and
reactive loads (with constant power factor) and keeping the
active power injected by the DG at the nominal level [9]. The
active and reactive power supplied by the generator was kept
constant because, usually, such generators are not rescheduled
by the system operator. Different load models were
considering during the simulations.
For transients analysis, voltage dependency of static loads is
(1)
considered. System active power loads at 13.8kV buses are
considered as 60% constant impedance loads, and reactive
power loads are considered as 100% constant impedance loads
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

as suggests the ONS (Brazilian System National Operator)
[10]. Two protections schemes were considered in the
simulations, as shows Table 2. The undervoltage protection
acts when voltage becomes lower than 0.9 pu. This circuit
breaker is indicated in Fig. 1 as CB.
TABLE II
PROTECTION SCHEMES
protection type
distance protection
undervoltage protection
III. SIMULATION RESULTS
A. Steady-State Analysis
In this section, both voltage drop and active power loss are
analyzed. Fig. 4 shows system voltage profile to high load for
the main buses in the corridor between the main generator
(bus 6420) and the DG (bus 5170). The results show that as
the DG penetration level increases, system voltage profile also
B. Voltage Stability Analysis
The voltage stability margin measures the distance from the
initial operation point to the maximum loading point in the PV
curve, and can be expressed in MW or in percentage [5].
In order to study the impact of DG on the voltage stability,
three scenarios with different DG penetration levels are
clearing time (s) investigated: case 1, case 2 and case 3. Fig. 6 shows the
80 ms voltage stability PV curve at bus 5210 (138kV) for different
60 ms DG penetration levels to high demand, considering constant
power loads. Although the active and reactive power supplied
by the DG was kept constant during the simulations, what
means that only the main generation assumes the increase in
system demand, the results show that the presence of the DG
improves the system voltage stability margin. It can be
explained because the DG provides active and reactive power
to local loads, decreasing system losses and increasing system
voltage stability limit.
0,984

increases. This fact is reflected in system losses. Since the DG

0,944
is connected close to a huge consumer center (STA), at low
DG penetration level system losses increases for all load 0,905

profiles, as shows Fig. 5.

0,866

1.12 PL = 15.6%
0,826 PL = 12.1%
1.1 PL=15.6%
PL = 8.6%
PL=12.1%
0,787
1.08 PL=8.6%
PL=0%
1.06 0,748
169,9 189, 208,1 227,3 246,4
System Loading (MW)
Voltage (pu)

1.04
Fig.6. PV curves for different DG penetration levels considering constant
1.02
power load.
1
Simulations were also performed considering the others ZIP
0.98 load models: constant impedance (Z), constant current (I) and
0.96 constant power (P) [5]. The constant power load model is the
most pessimist situation for static loads, with lower voltage
0.94
stability margin as shows Table III. Another simulation was
6420 6510 6530 6540 6550 5150 5165 5170
Buses conducted replacing 60% of the static loads modeled as constant
Fig.4. Steady-state voltage profile for different DG penetration levels. power with dynamic loads (induction motors). This brings the
system to a voltage stability margin even worst, of 39.6%.
12 However, for all cases, simulations results considering
High Load
11 Medium Load
different DG penetration levels show that the presence of the
Light Load DG always increased the voltage stability margin.
10
TABLE III
9 LOAD MODEL AND SYSTEM VOLTAGE STABILITY MARGIN
System Losses (MW)

Load Model Voltage Stability Margin

8
Static - costant impedance 64.80%
7
Static - constant current 56.94%
6 Static - constant power 45.02%
5 60% of dynamic loads 39.60%
4

3 C. Transient Analysis
2
Although many simulations were analyzed, this paper
0 2 4 6 8 10
DG penetration level (%)
12 14 16
presents only the results analyzed for two faults, which are
indicated in Figure 1.
Fig.5. Active power losses for different DG penetration levels.

• Fault 1: three phase fault applied at 1 second at 40% of

 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

line 5200 - 5205 (area 1), which is eliminated by (a)Simulation

60.3
tripping this branch. PL = 8.6%

Fig. 7 presents the voltage sag at bus 751 close to the DG, 60.2 PL = 12.1%

Frequency (Hz)
PL = 15.6%
occasioned by this fault for different DG penetration levels.
60.1
The results show that the increase in the DG penetration level
results only in a slightly increase in voltage sag magnitude 60
during the transient, which is not significant. The DG slightly
increases the current in the branch 5170–751, as shows Fig. 8. 59.9
0 2 4 6 8 10 12 14 16 18
Time (s)
1.08
(b)Transient Details
PL = 8.6% 60.2
1.05 PL = 8.6%
PL = 12.1%
PL = 15.6% 60.15 PL = 12.1%

Frequency (Hz)
PL = 15.6%
Nodal voltage at bus 751 (pu)

60.1
1

0.85 60.05
0.845
0.95 60

0.835 59.95
0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
0.9 Time (s)
0.825
Fig.9. Frequency behavior during fault 1 for different DG penetration levels.
34
0.85 0.815
31 PL = 8.6%

DG Active Power (MW)

PL = 12.1%
28
0.805 PL = 15.6%
0.99 1 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09
0.8 25
0.5 1 1.5 2 2.5 3 22
Time (s)
19
16
Fig.7. Nodal voltage at bus 751 for different DG penetration levels.
(a)Simulation 13
0.14 10
0 1 2 3 4 5 6 7 8
0.12 PL = 8.6% Time (s)
PL = 12.1% 170
Current (pu)

0.1 165
PL = 15.6%
PL = 8.6%
Active Power (MW)

0.08 155 PL = 12.1%

Main Generator

PL = 15.6%
0.06 145

0.04 135
0 1 2 3 4 5 6 7
125
Time (s)
(b)Transient Details 115
0.14 0 1 2 3 4 5 6 7 8
PL = 8.6% Time (s)
0.12 PL = 12.1%
Fig.10. Active power during fault 1 for different DG penetration levels.
PL = 15.6%
Current (pu)

0.1

0.08

0.06 • Fault 2: three phase fault at 40% of line 5150 - 5155

(area 2), which is eliminated by tripping this branch;
0.04

0.98 1 1.02 1.04 1.06 1.08 1.1 1.12 1.14 1.16

Time (s) Fig. 11 presents the voltage sag at bus 751 close to the DG,
Fig.8. Current in the branch 5170-751 during fault 1 for different DG occasioned by this fault for different DG penetration levels.
penetration levels. Since this fault is much closer to the monitored bus, this
voltage sag is more severe than the one occasioned by fault 1.
Since the DG plants do not participate in system frequency Also, the increase in the DG penetration level does not affect
regulation, one can see from Fig. 9 that frequency excursions significantly the voltage sag magnitude.
are larger when DG penetration level is bigger. Differently from fault 1, to this fault the increase in the DG
Fig. 10 shows the active power generation in CUNA (DG) penetration level results in a slightly decrease in voltage sag
and TUC (main generation). The demand of 29MW from area magnitude. This can be explained by the current in branch
1 is not supplied due to the short circuit. Then, the TUC 5170-751 during the fault, which also decreases, as shows Fig.
decreases its generation to meet the demand after the fault, 12.
since CUNA output generation remains constant after the
transient.
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

1.2 (a)Simulation
60.7
PL = 8.6%
PL = 8.6%
1.1 PL = 12.1%

Frequency (Hz)
60.5 PL = 12.1%
PL = 15.6%
PL = 15.6%
1 60.3
Nodal voltage at bus 751 (pu)

0.9 60.1
0.41

0.8 0.405
59.9
0 2 4 6 8 10 12 14 16 18
Time (s)
0.4
0.7 (b)Transient Details
0.395
60.7
0.6 PL = 8.6%

Frequency (Hz)
0.39 60.5 PL = 12.1%
0.5 PL = 15.6%
0.385 60.3

0.4 0.38
0.99 1 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09 60.1
0.5 1 1.5 2 2.5 3
Time (s) 59.9
0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8
Fig.11. Nodal voltage at bus 751 for different DG penetration levels. Time (s)

Fig.13. Frequency behavior during fault 2 for different DG penetration levels.

(a)Simulation

CUNA Active Power (MW)

40
0.3

0.25 30
PL = 8.6%
0.2 PL = 12.1%
Current (pu)

20 15MW
PL = 15.6%
0.15 21MW
10 27MW
0.1
0 1 2 3 4 5 6 7 8
0.05
Time (s)
0
0 1 2 3 4 5 6 7
TUC Active Power (MW)

Time (s)
160 PL = 8.6%
(b)Transient Details PL = 12.1%
0.275
PL = 15.6%
0.27 140

0.265
Current (pu)

120
0.26
0.255 0 1 2 3 4 5 6 7 8
PL = 8.6% Time (s)
0.25
PL = 12.1%
0.245 PL = 15.6% Fig.14. Active power during fault 2 for different DG penetration levels.

1 1.01 1.02 1.03 1.04 1.05 1.06 1.07 1.08 1.09

Time (s) D. Standards Adequacy Analysis
Fig.12. Current in branch 5170-751 during fault 2 for different DG penetration
levels.
During short circuits, synchronous generators usually
accelerate, and may become unstable due to the loss of
synchronism. The stability of synchronous generators can be
Fig. 13 shows frequency behavior when the system is determined by analyzing the dynamic response of the rotor
subjected to fault 2. The same performance is observed: angle. In order to evaluate this question, the critical clearing
frequency excursions are larger when DG penetration level is time was determined. The critical clearing time is the
bigger. maximum actuation time of the protection system such that the
Fig. 14 shows the active power generation in CUNA (DG) power system is transiently stable. The contingency simulated
and TUC (main generation). Again, the demand not supplied was fault 2 since it is more severe. The distributed generation
due to this fault is compensated with the decrease of the main is considered to supply 27MW (PL = 15.6%). Fig. 15 presents
generation. the rotor angle behavior to the critical clearing time, which is
0.24sec. This time is considerable superior than the clearing
time adopted by the utility (t = 80ms) and the maximum
clearing time of 150ms imposed by the Brazilian System
National Operator for 138kV transmission lines [11].
 Downloaded from http://iranpaper.ir
http://www.itrans24.com/landing1.html

system frequency regulation, frequency excursions are larger

400
T=0.24s
when DG penetration level is bigger. However, the DG
350
T=0.241s penetration levels do not impact voltage at DG terminals.
300 The critical clearing time evaluated is superior to the
250 clearing time currently adopted by the utility (t = 80ms) and
the maximum clearing time of 150ms imposed by the
Rotor Angle (degrees)

200

150
Brazilian National Operator for 138kV transmission lines.
Also, only faults in lines next to the DG will cause the
100
generator disconnection by severe voltage sags. Based on the
50
results and discussion, it can be concluded that DG can
0 improve the steady-state and stability performance of power
-50 system.
-100
V. REFERENCES
-150
0 2 4 6 8 10 12 14 16 18 20
Time (s)
[1] EU Directive 2001/77/EC on the promotion of the electricity produced
from renewable energy source in the internal electricity market.
Fig.15. DG rotor angle for a three-phase short circuit. [Online]. Available:
http://europa.eu.int/comm/energy/res/legislation/electricity_en.htm
The ONS establishes a voltage sag tolerance curve for [2] N. Jenkins, R. Allan, P. Crossley, D. Kirschen, and G. Strbac,
generators (ride through capability) as shows Fig. 16 [12]. "Embedded. Generation", 1st ed. London, U.K.: Inst. Elect. Eng., 2000.
Generators which do not meet this tolerance curve must be [3] V. V. Thong, J. Driesen, R. Belmans, "Benefits and Impact of Using
disconnected and pay fine. The results show that these voltage Small Generators for Network Support", in Proc. 2007 IEEE Power
sags does not imply in generator disconnection. Only faults in Engineering Society General Meeting, pp 1-7.
lines 5150-5165 and 5165-5170 will cause the generator [4] A. M. Azmy, I. Erlich, "Impact of distributed generation on the stability
disconnection. of electrical power system", in Proc. 2005 IEEE Power Engineering
Society General Meeting, Vol. 2, pp. 1056-1063.
1.2 [5] P. Kundur, "Power System Stability and Control". New York: McGraw-
1.1
Hill, 1994.
[6] IEEE Working Group on Prime Mover and Energy Supply Models for
1 System Dynamic Performance Studies. IEEE Transactions on Power
0.9 Systems. Vol.7 No.1. Feb. 1992.
[7] CEPEL, Centro de Pesquisas de Energia Elétrica, "ANAREDE:
DG terminal voltage (pu)

0.8
Programa de Análise de Redes", Manual Guide, 07-08 Version. Manual
0.7
do Usuário, August 1999.
0.6 [8] CEPEL, Centro de Pesquisas de Energia Elétrica, "ANATEM: Análise
0.5 de Transitórios Eletromecânicos", Manual Guide, 09 Version, December
2002.
0.4
[9] T. Van Cutsem, "Voltage Instability: Phenomena, Countermeasures and
0.3 Analysis Methods", in Proc. 2000 IEEE, vol.88, No. 2, pp. 208 – 227.
Fault 2
0.2 Fault 1
[10] ONS Network Procedures 2008, “Guidelines and Criteria for Electrical
ONS curve Studies – item 23.3” [Online]. Available: www.ons.org.br
0.1
[11] ONS Network Procedures 2008, "Minimum Requirements for protection
0
0.8 1 1.2 1.4 1.6 1.8 2
systems – item 2.6". [Online]. Available: www.ons.org.br
Time (s) [12] ONS Network Procedures 2008, "Minimum Technical Requirements for
the connection to the basic network – item 3.6". [Online]. Available:
Fig.16. DG terminal voltage during fault 1 and fault 2. www.ons.org.br

IV. CONCLUSIONS VI. BIOGRAPHY

This paper addresses the impact of DG with different
penetration levels on steady-state voltage profile, losses, Rafael R. Londero received the B.S. degrees in electrical engineering from
Federal University of Para, Brazil, in 2005. Currently, he is a Master student
voltage and transient stability of power systems. A real at the Federal University of Para.
network is simulated with two hydro power plants: one
representing the main generation and the other the DG. The Carolina de Mattos Affonso received the B.S. degrees in electrical
engineering from Federal University of Para, Brazil, in 1998, the Master’s
analyses are conducted using load flow program, continuation degree from Federal University of Santa Catarina, Brazil, in 2000 and the
method and non-linear time-domain dynamic simulations. For Ph.D. degree in electrical engineering from State University of Campinas,
voltage stability analysis, the demand is considered as constant Brazil, in 2001. Currently, she is a Professor of Electrical Engineering with
power loads. However, for transients analysis, voltage the Federal University of Para. Her research interests are on power system
stability analysis, power quality and distributed generation.
dependency of static loads is considered.
Regarding the steady-state analysis, the utilization of DG Marcus V. A. Nunes received the B.S. and Master’s degrees in electrical
improves voltage profile and decreases active power losses. engineering from Federal University of Para, Brazil, in 1993 and 1996
respectively, and the Ph.D. degree in electrical engineering from Federal
With more power from the DG units, the voltage stability University of Santa Catarina, in 2003. Currently, he is a Professor of
margin is also improved, giving the system more active power Electrical Engineering with the Federal University of Para. His research
transfer capacity. Since the DG plants do not participate in interests are on power system stability and distributed generation.