1

Generic 12-Bus Test System for Wind Power

Integration Studies
Andrzej Adamczyk, Student Member, IEEE, Mufit Altin, Student Member, IEEE, Ömer Göksu,
Student Member, IEEE, Remus Teodorescu, Fellow, IEEE, and Florin Iov, Senior Member, IEEE

offerings for ancillary services will expand in the future with

Abstract--High wind power penetration levels into power new functionalities such as inertial response, power system
systems requires an appropriate power system model when damping, secondary voltage control, etc. These services will
assessing impact on the overall system stability. The model play an important role especially in networks with high wind
should capture the wide range of dynamics related to the wind
power penetration levels. In order to study the impact of these
integration studies, such as voltage control, synchronizing power
control, inertial response, frequency control, damping of new ancillary services on the power system, simulation studies
electromechanical oscillations, balanced and unbalanced fault must be performed on a test system.
management, etc. Hence, the power system components: Numerous test systems variants were proposed in the
conventional power plants with controls, transmission lines, literature. A simple test system represented by a voltage
transformers and loads should be represented accurately to source behind an impedance is proposed in Danish grid code
achieve realistic power system characteristics. Additionally, the
[8] to assess the stability of a wind turbine for symmetric
power system model should be simple and computationally
manageable in order to simulate multiple scenarios with different three-phase faults. However, this test system does not include
control parameters in a reasonable time. In this paper, a generic the dynamics regarding the frequency response, voltage
power system model is presented in order to comprehend the control, and electromechanical oscillations. Furthermore, for
wind integration studies with different penetration scenarios. inertial response and frequency control studies the single-mass
model is implemented in the literature [9]. The rotating masses
Index Terms-- wind power, integration studies, power system of all synchronous generators in CPPs are lumped as a single-
stability and dynamics, power system modeling
mass; and governor controls are acting on this single-mass.
The single-mass model represents the frequency control
I. INTRODUCTION
dynamics however neglecting the voltage variations and

W ind energy generation has increased by 27.4%

worldwide over five recent years [1]. Moreover, targets
for total electricity generation, ranging from 20% to 50%,
electromechanical oscillations. Another simplified model
comprising these two models’ behavior is implemented as a
single-bus model in the literature [10], where all the
have been set for wind energy installations by many countries, generating units are connected to a single bus with their
considering CO2 emissions, fossil fuel costs and energy voltage and frequency controllers. In the single-bus model, the
efficiency [2]-[4]. Transmission system operators (TSOs) of voltage dynamics are limited due to the small electrical
these countries have revised their grid codes for wind power distances between the synchronous generators. Nevertheless,
integration in order to provide the stable operation of the this model is sufficient enough to analyze power system
power systems while reaching these targets [5], and power oscillations and balanced/unbalanced faults. However, in order
systems are planned for more wind power integration. Wind to cover the wind integration studies thoroughly, a multi-
power penetration scenarios have been defined for varying machine multi-bus system is required with appropriate voltage
levels of penetration and consumption [6]. When wind power and frequency controls.
penetration is reaching high levels, it is planned that wind Different multi-machine multi-bus power system models
power plants (WPPs) replace old thermal conventional power that can exhibit particular phenomena exist in the available
plants (CPPs) [7]. literature, e.g. 9-bus system [11], 2-area 4-machine system
Beyond a certain level wind power penetration into a power [12], 12-bus system [13], [14], and 68-bus 16-generator
system poses challenges regarding wind power control and system [15].
stable power system operation. Nowadays, the wind power The 9-bus system is not adequate to create various
can offer capabilities for ancillary services that can contribute generation mixes, voltage profiles and electromechanical
to the power system stability. It is expected that these oscillation modes. The 2-area 4-machine system has been
developed for studying the theory of the small signal stability;
This work was supported in part by Vestas Wind Systems A/S under the however it is not representing a realistic power system layout.
Vestas Power Programme.
A. Adamczyk, M. Altin, Ö. Göksu and R. Teodorescu are with the
The voltage profile of the 2-area 4-machine system is
Department of Energy Technology, Aalborg University, Aalborg, 9220, relatively stiff due to small electrical distances between the
Denmark. (e-mail: [email protected], [email protected], [email protected] and system buses and the radial and symmetrical system structure
[email protected]). does not allow flexibility in mode creation. The original 12-
F. Iov is with Vestas Wind Systems A/S, Århus, 8200, Denmark (e-mail:
[email protected]) bus system was developed to test FACTS devices and it is
 2

supporting small signal stability analysis. However, it does not

include a wide range of parameters and settings for CPPs. The
68-bus 16-generator system gives great flexibility in
formulating system dynamics and enables approaching a
realistic system behavior. However, due to the system size it is
too complex to handle it analytically in a small-signal stability
analysis or other wind integration studies.
The existing power system models described above are not
able to accommodate wind integration studies when they are
used in their current form. For wind integration studies, the
wind power locations in the power system and different wind
penetration scenarios should be specified in detail. Thus, a test
system that is able to support wind power studies should be
relatively weak – system should be stressed close to a stability
limit. Still, the test system should be stable under steady-state
conditions. In the base case, voltage profile of the network
Fig. 1. Structure of generic 12-bus system for wind power integration studies.
should not be uniform - i.e. deviations around 1pu should
exist, while staying in the limits of ±5% from the nominal
There is a large power flow, over long lines, from Area 1 to
values, as commonly required by the grid codes. Moreover,
Area 3. The infinite bus in the original 12-bus system (bus-9)
some of the system load buses should be relatively weak -i.e.
is replaced with a CPP, G1. Traditionally, it is assumed that in
be electrically distant from the conventional generating units.
power system studies the reference machine is the largest plant
The primary frequency control requires a network with
in the system. Therefore, size of G1 is set to 768 MVA to
different conventional generation mixes, including settings for
make its capacity largest among the generation mix. Most of
governors as well as load variations. Inertial response studies
the loads in Area 2 are shifted to Area 1 in order to balance the
have the similar requirements with different wind penetration
power flow.
scenarios. On the other hand, studies regarding
Active power control for CPPs is included by implementing
electromechanical oscillation damping capabilities from wind
the prime mover and governor models. Active power dispatch
power are requiring presence of different oscillation modes in
between the areas is adjusted to improve voltage profile and
the power system that are mainly coming from CPPs in
achieve N-1 contingency criterion.
combination with the network parameters and layout.
Instead of simplified 1st order exciters, more detailed
Taking into account the abovementioned requirements, a
models with realistic parameters are used for each CPP. The
number of test cases should be defined (e.g. line and generator
main buses exhibit a poor voltage profile and different
tripping, step load change) to impose abnormal conditions on
electromechanical oscillations occur when triggered.
the system and excite different kinds of instabilities. Since
Furthermore, both the residential and industrial loads in the
wind integration studies are targeted, different realistic wind
test system are modeled as voltage and frequency dependent.
penetration scenarios should be established based on existing
The system can be considered to be similar to UK power
trends seen around the world. Overall, it is possible to relate a
system in terms of direction of power flows as shown in
generic test system used for wind integration studies with real Fig. 2, where Area 4 is representing offshore WPPs, Area 2 is
power grids including their trends in accommodating wind
hydro dominant Scotland, finally Area 1 and 3 are the
power.
congested central-south generation regions of the country.
The present paper is proposing a generic test system model
based on the 12-bus test system [13], [14] that is able to
address all the above mentioned characteristics.

II. GENERIC 12-BUS TEST SYSTEM

A. Model structure
The layout of the generic 12-bus system for wind
integration studies is given in
Fig. 1. The system represents a small islanded power
system with four areas dominated by thermal power plants like
power systems in UK, USA and Germany. Area 1 is the
biggest thermal power generation area with industrial and
residential loads. Area 2 has dominant hydro power generation Fig. 2. Generic 12-bus system in context of UK grid [16].
with small amount of rural loads. Area 3 is a heavily industrial B. Load Flow
load center with thermal power generation, while Area 4 is The load flow for the base case without any wind power
rich in wind resources. Hence, it is considered to have wind contribution is presented in Fig. 3.
power connected mainly at bus-5.
 3

50% as shown in Fig. 5

Fig. 5. Wind penetration scenarios for generic 12-bus system.

First, the wind power is inserted at bus-5 for penetration

Fig. 3. Load flow results for generic 12-bus system.
levels up to 30%. Then, the distribution of wind power to bus-
In the base case, voltage profiles and load flows are 3 and bus-4 is assumed to be more realistic for higher
balanced in order to represent steady-state operating penetration levels (>30%). The total active power capacity of
conditions of the test system with proper reactive power the generic 12-bus test system is 2248 MW, including
compensation. Bus voltages are within the specified limits as frequency responsive reserve. For the base case (0%), total
±5% of the nominal value, and accordingly the generator active power of the load is established as 1450 MW and the
loadings (around 70%) and line loadings (around 50%) are total active power from CPPs is 1480 MW including the active
within the acceptable limits. N-1 contingency criterion is power losses.
achieved except the tripping of the longest line between bus-7
TABLE I
and bus-8. In addition to the steady-state analysis, the stiffness WIND POWER PENETRATION SCENARIOS IN GENERIC 12-BUS TEST SYSTEM
of the system buses is evaluated by a short circuit power 0% 5% 10% 20% 30% 40% 50%
calculation, and illustrated in Fig.4. CPP (MW) 1480 1470 1470 1470 1260 1120 1020
WPP (MW) 0 100 200 400 600 750 850
Load (MW) 1450 1550 1650 1850 1850 1850 1850

The sharing of production and consumption for the

considered wind penetration levels is given in Table I. It
should be noticed that the active power losses are included in
the CPP generation.

Fig. 4. Short-circuit power levels for each buses in generic 12-bus system. III. CONVENTIONAL POWER PLANT MODEL

C. Operational Scenarios The structure of the CPP model implemented in the test
system is shown in Fig. 6.
Various operational scenarios are considered for the generic
power system model. These scenarios reflect different wind
penetrations into the power system. Two trends in
development of the power system have been considered. The
first trend assumes that the increased loads are covered by the
increase of wind power while the installed capacity of the
CPPs is kept constant. This is a typical situation in countries
with incentives for wind power in the early stages of wind
power development e.g. Denmark, Germany, UK, etc. For Fig. 6. Structure of conventional power plant model.
these cases some levels of wind power penetration may
require reinforcement of lines; however no major changes in As given in the literature [17] this model comprises of a
the network layout are expected. The second trend assumes synchronous generator, prime mover, governor, excitation
that the demand will not change significantly, but new wind system, and automatic voltage control (AVR). In the following
farm installations would replace some for the existing CPPs. subsections, brief descriptions of those components are given.
This is typically in the countries where wind power is reaching A. Generator
a relatively high penetration level e.g. more than 20%. A
The generator model is described by a set of six differential
typical example for this trend is Denmark where 50% of
equations in a synchronous reference frame that describe rotor
electricity demand is planned to be covered from wind energy
circuit dynamics as defined in [17]. Two additional steady-
by 2025. Thus, the generic power system model can
state equations are describing the stator circuit.
accommodate wind penetrations from 0% (base case) up to
 4

B. Prime mover and governor while CPPs and WPPs are described by differential equations.
The CPPs considered for the model are of fossil, steam and The implemented approach in RMS toolbox to solve these
hydro types. Various types of governors in the power system equations is defined as partitioned-explicit method using a
for CPPs are available on literature. However, it is difficult to current-balance form [17]. The Matlab/Simulink RMS toolbox
find parameters for a wide range of sizes. Thus, a governor currently includes the analyses such as load flow, short-circuit
and turbine model as defined in [11] is selected and shown in calculation for symmetrical and asymmetrical faults, small-
Fig. 7. The different parameters for the model are tested and signal stability analysis, and time domain simulations for
selected with respect to the response rates. Three sets of power system stability studies.
parameters are benchmarked for different power system An alternative implementation has been accomplished in
characteristics. RSCAD, running in real time on RTDS. This model is suitable
for Hardware-In-the-Loop (HIL) electro-magnetic transient
(EMT) studies with detailed models of network components.

Fig. 7. Block diagram for prime mover and governor model. V. SMALL-SIGNAL STABILITY ANALYSIS
In order to select suitable types of excitation systems that
C. Excitation system meet the requirements for the base case of the test system a
Excitation systems can be categorized into three main comprehensive small-signal stability analysis is performed.
groups [11], [18]: Final selection of parameters regarding the excitation
• DC excitation systems, system and generator models is done to obtain a desired mode
• AC excitation systems, profile for electromechanical oscillations in the generic 12-bus
• Static excitation systems. test system. Based on the analysis, two sets of parameters are
Such categorization is also adopted in this paper. selected which correspond to two mode shape profiles.
Nevertheless, for each category many models exist in the Mode profile 1 is characterized by low damping of all
literature. Additionally, a wide scope of realistic parameters electromechanical modes. All damping ratios are in the critical
for excitation systems is hardly available in the literature. The range of 0 to 10%. Moreover, for all modes of interest, a WPP
parameters for abovementioned categories are selected from at bus-5 has a relatively high potential (controllability index)
[11]. These models can be viewed as equivalent or simplified to contribute to damping of this mode profile. For some modes
versions of models recommended by the current IEEE active power controllability is higher, while for others reactive
standard [18]. Hence adopting notation used in [11], for the power controllability is higher. Mode profile 2 is characterized
generic 12-bus test system: by highest average controllability index for reactive power
• DC exciters are modeled as Type A (equivalent to DC1A modulation at bus-5, where most of the wind power is
in [18]), connected.
• AC brushless exciters are modeled as Type C (simplified Frequency and damping ratio for mode shapes
version of AC1A in [18]), corresponding to the mode profile 1 are given in Table II. The
• Static exciters are modeled as Type G (simplified version generic 12-bus test system exhibits swing modes as well as
of ST1A in [18]). control modes in the typical range for a power system. The
participation factors for all swing modes in the mode profile 1
are shown in Fig. 9 for all CPPs.
TABLE II
OSCILLATORY MODES FOR MODE PROFILE 1

Fig. 8. Block diagram of an AC brushless exciter [18].

The different parameters for the model are tested according

to the related IEEE standard [18]. Three sets of parameters are
benchmarked for different power system characteristics.

IV. SYSTEM IMPLEMENTATION The response of the generators’ speed in time domain for a
The proposed generic 12-bus test system model for wind small step in the voltage reference is shown in Fig. 10. Poorly
integration studies is implemented in Matlab/Simulink and damped mode in which G4 is highly participating is clearly
tested against an implementation in Power Factory DigSilent. noticed in Fig. 13. Two groups are swinging with respect to
Currently, the model is available in the RMS toolbox each other; the first group is G1 with G2, and the second
developed by the authors in Matlab/Simulink. The power group is G3 with G4.
system is represented algebraically in a complex current form
 5

Fig. 9. Normalized participation factors for swing modes related to Mode

Profile 1; Blue; Mode 1, Green: Mode 2, Red: Mode 3.
Fig. 11. Generic 12-bus test system frequency response for different governor
and prime mover settings.

Fig. 10. Time domain responses of generators’ speed to a step in voltage

reference.

VI. SYSTEM FREQUENCY RESPONSE DURING NETWORK

EVENTS
For the grid frequency performance profile, there are
Fig. 12. Voltage and power variations at the connection point of WPP (bus-5).
indices available in the grid codes such as the minimum
frequency point (nadir or dynamic frequency deviation),
VII. FUTURE WORK
quasi-steady-state frequency deviation, and rate of change of
frequency (ROCOF) [19]. In the load-frequency control As described in the previous sections, the developed
regulations of ENTSO-E, the maximum permissible frequency generic 12-bus test system is appropriate for wind integration
deviation is defined as ±800 mHz (0.016 pu) including a studies and can be used for analysis purposes such as small-
safety margin for the automatic load shedding frequency (1 signal stability, electromechanical oscillation damping,
Hz). Moreover, the quasi steady-state frequency deviation is frequency response, etc. Additionally, the generic 12-bus test
specified between ±180 mHz (0.0036 pu) range. system is open for modifications and gives the opportunity to
In islanded power systems like UK and Ireland, TSOs be improved by enhancements.
defined ROCOF relay settings for the distributed generation In order to analyze the impact of wind integration on the
connected to distribution system. The ROCOF relay trip value power systems for cases where HVDC connections and
for EirGrid is 0.6Hz/s (0.012 pu/s). Taking into account the FACTS devices are installed, the generic 12-bus test system
above regulations and practical implementations, a set of can be expanded with these units as shown in Fig. 16. For
operational metrics is defined for the generic 12-bus test instance, the electromechanical oscillations in the power
system. Based on a comprehensive analysis involving system can be analyzed where the HVDC and STATCOM
different sets of parameters for governors and prime movers, units are also contributing in damping of oscillations.
three sets of parameters are selected for each CPP. The system
frequency response for the N-1 contingency which is the
largest infeed loss is given in Fig. 14 considering the defined
operational metrics. These parameter sets correspond to slow,
medium and fast response of the CPPs respectively.
The voltage variations and electromechanical oscillations
should be taken into account for the frequency control and
inertial response studies of the WPPs. It is possible to observe
these variations and oscillations in the generic 12-bus test
system. An example of these variations at the connection bus
of the WPP (bus-5) for the base case following the largest
infeed loss is shown in Fig. 15.
Fig. 13. Future expansion of generic 12-bus system with FACTS devices.
 6

Moreover, WPPs can be installed in the generic 12-bus test TABLE VI

BUS DATA FOR 12-BUS SYSTEM BASE CASE SCENARIO (100MVA BASE)
system through HVDC connections, when analysis of wind
Bus V Type Pgen Pload Qload Bshunt Vmag
integration with DC connection is under interest. [kV] [pu] [pu] [pu] [pu] [pu]
1 230 PQ - 3.00 1.86 - -
VIII. SUMMARY 2 230 PQ - 2.50 1.21 - -
3 230 PQ - 3.50 1.15 - -
This paper presents a generic 12-bus test system with wind 4 230 PQ - 3.00 1.86 2 -
power penetration scenarios for wind integration studies. The 5 230 PQ - 1.00 0.48 0.4 -
developed test system is based on [13], but the original 12-bus 6 230 PQ - 1.50 0.49 - -
test system has been modified in order to investigate and 7 345 PQ - - - -1 -
8 345 PQ - - - - -
demonstrate the requirements of wind power on the power 9 15.5 SL - - - - 1.00
system. The scope of the analysis includes voltage control, 10 15 PV 4.0 - - - 1.01
synchronizing power control, inertial response, frequency 11 18 PV 2.7 - - - 1.01
control, damping of electromechanical oscillations, balanced 12 13.8 PV 3.3 - - - 1.01
and unbalanced fault management.
The realistic power system characteristics are demonstrated X. ACKNOWLEDGEMENTS
briefly using small-signal and frequency stability analysis. This work was supported by Aalborg University-Vestas
Accordingly, the generic 12-bus test system is simple and Wind Systems A/S partnership under Vestas Power Program.
computationally manageable to simulate multiple scenarios Any opinions, findings, and conclusions or recommendations
with different control parameters. The test system is open for expressed in this material are those of the authors and do not
modifications regarding different power system components necessarily reflect those of Vestas Wind Systems A/S.
with control concepts such as HVDC, FACTs, and WAMs.
XI. REFERENCES
IX. APPENDIX – SYSTEM DATA [1] BTM Consult ApS, "International Wind Energy Development, World
Market Update 2010," Mar. 2010.
Here is summarized data for the generic 12-bus system. [2] Ecofys Final Report for Work Package 3 prepared for EirGrid, "All
Layout of the network is the same as show in Fig. 1. Line Island TSO Facilitation of Renewables Studies," June 2010.
lengths and impedances are shown in Table III. Transformer [3] EcoGrid Phase 1 Work Package 3, "Step toward a Danish Power System
data is given in Table IV. Bus configuration is shown in Table with 50% Wind Energy", Jan. 2007.
[4] IEA Wind Task 25 Final Report Phase 1, "Design and Operation of
VI. Generator data, including controllers, is given in Table V. Power Systems with Large Amounts of Wind Power,” 2009.
Due to space limitation only unit names, as used in the source [5] M. Altın, O. Goሷksu, R. Teodorescu, P. Rodriguez, B.-Bak Jensen, L.
references, are given. Hence, the detailed list of parameters for Helle, "Overview of recent grid codes for wind power integration," in
CPPs can be found in, according to [11] and [18]. Tables III- Optimization of Electrical and Electronic Equipment (OPTIM), 2010
12th International Conf., pp.1152-1160.
VI describe the base case cofiguration of the system. [6] National Grid UK, "Frequency Response Technical Sub Group Meeting
Summary," Mar. 2011.
TABLE III [7] National Grid UK, "National Electricity Transmission System Seven
LINE DATA FOR 12-BUS SYSTEM BASE CASE SCENARIO (100MVA BASE) Year Statement," May 2011.
Line Length R X B [8] Grid connection of wind turbines to networks with voltages above 100
[km] [pu] [pu] [pu] kV, Regulation TF 3.2.5, Eltra and Elkraft System.
1-2 100 0.01131 0.08998 0.18377 [9] G. Lalor, A. Mullane, and M. O'Malley, "Frequency Control and Wind
1-6 300 0.03394 0.26995 0.55130 Turbine Technologies," IEEE Transactions on Power Systems, vol. 20,
2-5 400 0.0453 0.3599 0.7351 no. 4, pp. 1905-1913, Nov. 2005.
3-4 100 (x2) 0.0057 0.0450 0.3675 [10] J. F. Conroy and R. Watson, "Frequency response capability of full
4-5 150 0.0170 0.1350 0.2757 converter Wind turbine generators in comparison to conventional
4-6 300 0.03394 0.26995 0.55130 generation," IEEE Transactions on Power Systems, vol. 23, no. 2, pp.
7-8 600 0.0159 0.1721 3.2853 649-656, May 2008.
[11] P.M. Anderson and A.A Fouad, Power System Control and Stability,
TABLE IV IEEE Press, 2003, ISBN 0-471-23862-7.
TRANSFORMER DATA FOR 12-BUS SYSTEM BASE CASE SCENARIO [12] P. Kundur, Power System Stability and Control, McGraw Hill, 1994,
From-To Type MVA Uk Vector ISBN 0-07-035958-x.
Bus Capacity (%) Group [13] Shan Jiang, U. D. Annakkage, and A. M. Gole, “A platform for
1-7 Autotransformer 500 13 YNyN validation of facts models”, Power Delivery, IEEE Transactions on,
1-9 Step-up 800 12 YNd11 21(1):484{491, 2006.
2-10 Step-up 700 12 YNd11 [14] CIGRE Working Group B4.39, "Integration of Large Scale Wind
3-8 Autotransformer 500 13 YNyN Generation using HVDC and Power Electronics," Feb. 2009.
[15] B. Pal, B. Chaudhuri, Robust Control in Power Systems, Springer, 2010,
3-11 Step-up 400 10 YNd11
ISBN 978-1-4419-3853-4.
6-12 Step-up 500 11 YNd11
[16] www.nationalgrid.com/uk/
[17] Sauer, Pai, Power System Dynamic and Stability, Prentice Hall, 1997,
TABLE V ISBN-13: 978-0136788300.
CPP DATA FOR 12-BUS SYSTEM BASE CASE SCENARIO [18] IEEE recommended practice for excitation system models for power
# of Total MVA Generator Governor system stability studies, IEEE Sandard 421.5-1992, 1992
CPP Exciter type
units Capacity type type [19] A. Adamczyk, R. Teodorescu and P. Rodriguez “Adaptation of 12-bus
1 6 750 F6 H13 F10 System for Wind Power Integration Studies”.
2 4 640 F8 H13 F10 [20] ENTSO-E, "Load-Frequency Control and Performance," Jun. 2004.
3 2 384 F9 F8 F10
4 3 474 H15 ST1A H16