IEEE COMMUN.

SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID 1

Combining power and communication network

simulation for cost-effective smart grid analysis
Kevin Mets, Member, IEEE, Juan Aparicio Ojea, Chris Develder, Senior Member, IEEE

Abstract—Today’s electricity grid is transitioning to a so-called typically amount to 8% [1]). Further, DSM/DR approaches can
smart grid. The associated challenges and funding initiatives have help to reduce required generation capacity to deal with peak
spurred great efforts from the research community to propose demand only (for which around 20% of current generation
innovative smart grid solutions. To assess the performance of
possible solutions, simulation tools offer a cost effective and capacity is deployed [1]).
safe approach. In this paper we will provide a comprehensive While the smart grid transition happens at the various grid
overview of various tools and their characteristics, applicable in levels (i.e., generation, transmission and distribution), much
smart grid research: we will cover both the communication and research attention is going to the distribution grid, where today
associated ICT infrastructure, on top of the power grid. First, limited control is available. Also, typically the roots of power
we discuss the motivation for the development of smart grid
simulators, as well as their associated research questions and system issues trace back to this distribution level [1].
design challenges. Next, we discuss three types of simulators in Central to the smart grid concept, is the convergence of in-
the smart grid area: power system simulators, communication formation and communication technology with power system
network simulators, and combined power and communication engineering. Modern monitoring, analysis, control, and com-
simulators. To summarize the findings from this survey, we munication capabilities are being added to the aging infrastruc-
classify the different simulators according to targeted use cases,
simulation model level of detail, and architecture. To conclude, ture of the electricity grid, to more accurately get insight in the
we discuss the use of standards and multi-agent based modeling current grid state and use that knowledge to operate it more
in smart grid simulation. efficiently. The latter also implies environmental constraints,
which are an important underlying motivation for the smart
grid evolution, as exemplified by e.g., the European Union’s
I. I NTRODUCTION “Climate and Energy Package” definition of the famous 20-20-

T ODAY’S electricity grid is transitioning to a so-called

smart grid. This is driven by the objective of making
electricity delivery more reliable, economical and sustainable.
20 targets, to be met by 2020: (i) 20% of energy supply should
stem from renewable energy sources, (ii) reduce greenhouse
gasses with 20%, (iii) 20% increase in energy efficiency.
Given the reliance of critical services (e.g., transportation, Undeniably, aforementioned challenges and associated
communication, finance) on the power grid, demand for a funding initiatives have spurred great efforts from the research
resilient and self-healing grid is high. The challenge to re- community to propose innovative smart grid solutions. Smart
alize it is complicated by the ever increasing penetration of grid technology typically results in an increased complexity
renewable and distributed energy, adding an extra uncertainty of the power grid, and implies uncertainty (to be dealt with
dimension and thus the need for efficient responses to not by, e.g., stochastic control models). To assess the performance
only varying customer demand, but also to varying (and of possible solutions, simulation tools offer a cost effective
less controllable) production levels: demand-side management approach. In this paper, we will provide a comprehensive
(DSM), in particular demand response (DR) is increasingly overview of the various tools and their characteristics, appli-
important to keep the grid operation economically viable (i.e., cable in smart grid research: we will cover both the communi-
feasible without excessive infrastructure investments). Indeed, cation and associated ICT infrastructure, on top of the power
the power grid since its inception was designed to deliver grid.
power from large centralized generation units unidirectionally The aim of our work is to assist (i) smart grid researchers
over transmission networks towards the consumers connected looking for tools that target a certain use case, as well as
to distribution nets. To make it more economical, distributed (ii) smart grid simulator developers that wish to gain insights
sources could help reduce the distance between production and learn more about simulator paradigms, architectures, stan-
and consumption (thus limiting transmission losses, which dards, etc. However, it is not our intention to provide a detailed
implementation guide for smart grid simulators.
Manuscript received April 15, 2013. Revised version received November 3,
2013. Camera-ready version received March 3, 2014. Work described in this The remainder of this introduction outlines the main power
paper was partly funded by the European Commission through the 7th ICT- grid challenges and indicates how they call for communication
Framework Programme (FP7-ICT-2011-8) project C-DAX (grant agreement infrastructure to be added. In Section II in general, and more
no. 318708). K. Mets was funded through a Ph.D. grant from the Agency for
Innovation by Science and Technology (IWT). specifically in Section II-A, we motivate the choice for a
K. Mets, J. Aparicio Ojea, and C. Develder are with with Dept. of simulation approach in the domain of smart grids. Section II-B
Information Technology – IBCN, Ghent University – iMinds, Ghent, Belgium, points out possible pitfalls to aspiring developers of a smart
e-mail: {firstname.lastname}@intec.ugent.be.
J. Aparicio Ojea is also with Siemens Corporate Research, Princeton, NJ, grid simulator, through an overview of the related design
USA. challenges. From a researcher’s perspective, the same overview
2 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

of design challenges can serve as a guide whether to develop sections we present some high level examples showcasing the
custom simulation tools, or rather aim to reuse existing tools need for communication for both measurement/monitoring and
where possible. A general overview of smart grid simulation control. The latter calls for combining accurate models of in-
paradigms is given in Section III. Specifically, Section III-A formation and communications technology (ICT) components
provides insights into the two main approaches used to achieve as well as power networks, e.g., allowing the impact of such
combined simulation of communication networks and power control on power system transients [5]. In the context of such
grids, and Section III-B goes into more detail regarding the smart grid applications, some examples of communication
differences in modeling time in both domains. Although this requirements and performance metrics are [2], [4]:
survey is focused on software based simulation, we briefly • Latency requirements are concerned with the time re-
discuss the related concepts such as emulation, real-time quired to send data from a source to a destination. Certain
simulation, and hardware-in-the-loop in Section III-C. Next, applications, such as real-time state estimation using
we will discuss the three types of simulators in the smart PMU data requires very low latency (few tens of ms).
grid area: power system simulators in Section IV, commu- For applications such as smart meters data collection or
nication network simulation tools in Section V, and combined demand response the latency requirements are less critical
power and communication simulation in Section VI. From a (up to seconds).
researcher’s perspective, these respective overviews can help • Data rate requirements are concerned with the speed at
to assist in the tools to select for a particular task, while for which data can be sent, i.e., the data volume that can be
a developer it might be worthwhile to select one (or more) sent within a certain period of time. For example, video
as a starting point (resp. building block(s) in a co-simulation data used in wide area monitoring and control requires
approach, see further). We will finally provide a summarizing high data rates, whereas data rates for AMI can be low.
discussion in Section VII and conclude in Section VIII. • Reliability requirements deal with ensuring the communi-
cation system remains available and is able to send data.
A. The role of communication networks in smart grids Remote protection applications require a very reliable
Communication networks already play an important role communication network to ensure the safe operation of
in the power system. However, from a communication per- the grid.
spective, existing power grid networks suffer from several • Security requirements aim to protect the system from
drawbacks [2], such as: (i) fragemented architectures, (ii) a a wide range of attacks. Concepts related to security
lack of adequate bandwidth for two-way communications, are confidentiality (i.e., prevent the disclosure of infor-
(iii) a lack of inter-operability between system components, mation to unauthorized parties), integrity (i.e., maintain
and (iv) the inability to handle increasing amount of data and assure the accuracy and consistency of data over its
from smart devices. As we will show in the next sections, entire life-cycle), availability (i.e., the information must
communication networks will play an even more crucial role in be available when needed), authenticity (i.e., validate that
the development of smart grids, and hence are subject of many parties are who they claim to be), and non-repudiation.
research efforts, studying the most efficient topology of the Power line communication (PLC) reuses existing power
communication network, physical media, protocols, etc. [3]. wires for data communication. i.e., the power grid itself
To gain a better understanding of the type of communication becomes the communication network. Different types of PLC
networks present in smart grids, the overall smart grid com- technology exist [6]: (i) ultra narrowband PLC technology
munications layer is often considered to consist of three types operating in 300 to 3000 Hz range with very low bit rate (100
of networks, each having a distinct scale and range: bps), (ii) low data rate (few kilobits per seconds) narrowband
• Wide Area Networks (WAN) provide communication be- PLC operating in the 3-500 kHz range, (iii) high data rate
tween the electric utility and substations, and as such narrowband PLC (500 kbps), (iv) broadband PLC operating
operate at the scale of the medium voltage network and in 1.5–30 MHz range and data rates up to 200 Mbps.
beyond. WAN are typically high-bandwidth backbone Narrowband PLC technologies that operate over the medium
communication networks that handle long-distance data voltage or low voltage power grids have been proposed by
transmission. e.g., PRIME [7], PLC G3 [8], and IEEE 1901.2 initiatives.
• Field Area Networks (FAN), Neighborhood Area Net- Targeted applications include monitoring (e.g., AMI), grid
works (NAN), and Advanced Metering Infrastructure control, etc. Broadband PLC is being used for e.g., home
(AMI) provide communication for power distribution ar- multimedia services. However, PLC is challenging because the
eas (low voltage network). FAN/NAN/AMI interconnect communication channel, i.e., the power grid, was not designed
WAN and the Home/Building/Industrial Area Networks for that purpose.
(HAN/BAN/IAN) of the end-users.
• Home Area Networks (HAN), Building Area Networks
(BAN), and Industrial Area Networks (IAN) provide B. Advanced metering and demand response
communication between electrical appliances and smart Distribution grids have limited monitoring and control ca-
meters within the home, building or industrial complex. pabilities and today in practice still depend largely on manual
Various smart grid applications have specific (challenging) actions. As part of the efforts to transition to more automated
communication requirements (see [4]), and in the next sub- solutions, advanced metering infrastructure (AMI) has been
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 3

the focus on the distribution system level. It provides distribu- such as coal or hydroelectric. As such, large amounts of
tion system operators not only with system state information, distributed energy sources have to be monitored and man-
but also provides remote control capabilities. AMI systems aged [11] to ensure optimal integration. Demand and supply
originate from automated meter reading (AMR) systems capa- must be in balance in the power grid. As a result, large shares
ble of remotely reading consumption and production records, of renewable energy require stand-by controllable generation
alarms and status information from the customer. However, or the presence of storage to cope with sudden changes
AMR is limited by one-way communication capabilities and in power output. Small controllable energy sources can be
does not enable control actions based on received information. aggregated in so called virtual power plants. Distributed algo-
AMI on the other hand provides two-way communication, rithms must be developed to make decisions on power system
and therefore supports control over the demand: AMI is state and control actions [3]. In this context, communication
considered as a possible basis for distributed command and protocols, standards and data formats will be essential to make
control strategies [1]. Note that AMI will need to scale to very these components inter operable. Therefore, it is essential that
large number of participants (e.g., every electricity meter). these are evaluated in detail before deployment [3], [11]. Also,
Indeed, energy demand levels and their patterns over time DRES may be located in regions where no communication
are undergoing changes as a result of emerging technologies infrastructure is currently available and possibly difficult to
such as electric vehicles, heat pumps, µ-CHP, etc. Demand deploy. For example, DRES located in mountainous terrain or
response (DR) technologies aim to adapt the energy demanded offshore may require wireless or power line communication
over time. A classic example of DR is a dual tariff scheme based solutions due to the complexity and cost of deploying
for energy consumption, i.e., an expensive peak hour tariff, alternative wired solutions (e.g., fiber).
and a cheap off-peak hour tariff. In such a scheme, consumers
are provided an incentive to modify their energy consumption
patterns. Communication technologies such as AMI will en- D. Wide-Area Monitoring, Protection & Control (WAMPAC)
able much more fine-grained levels of control using variable To prevent instability and collapse of the system (e.g.,
pricing or even real-time pricing. Electric appliances that are because of DG behavior), control and protection schemes
equipped with a smart grid interface could react automatically are essential. Traditional protection schemes depend on local
to these price signals (thus relieving the consumer from having measurements sent to a central control system that is part of
to take manual actions based on the changing prices). the supervisory control and data acquisition (SCADA) sys-
One particular area of specific interest in the DR sphere is tem [12], and which sends adjusting (low bandwidth) control
the charging of plug-in (hybrid) electric vehicles (P(H)EV), signals over dedicated communication networks. However,
which show great promise for the transport sector in reducing modern protection and control schemes measure and send
the associated emissions and costs (esp. if the energy is information at a much higher rate: e.g., measurement and
supplied by renewable sources). However, such vehicles rep- communication of coherent real-time data is considered an
resent a significant new load to the power grid, especially for enabling technology for improving monitoring and control of
distribution grids that are already operating near their limits. the power grid [13]. Synchronized phasor measurements (syn-
The load stemming from uncontrolled EV charging (which for chrophasors), representing both magnitude and phase angle of
full-electric EVs amounts to the same order of magnitude of voltage or current waveform at particular points in the grid,
a complete household!) thus may require substantial (distribu- are obtained by phasor measurement units (PMU) devices
tion) grid infrastructure investments. Hence the importance of and further collected by phasor data concentrators (PDC).
applying DR-like techniques to avoid overloading the grid. On This offers real-time state information with microsecond time
the other hand, electric vehicles also present new opportunities accuracy, thanks to synchronization using Global Positioning
for utilities. For example, the vehicle batteries could be used System (GPS) clocks. Such PMU data supports detailed and
for so-called vehicle-to-grid (V2G) applications [9], [10]: accurate state estimation, and enables multiple applications
provide peak power, or cope with the intermittent behavior of including distributed wide area control, protection, wide-area
renewable energy sources by storing excess energy and feeding situational awareness, post-event analysis, etc. While such
it back into the grid when needed. Intelligent management PMU networks initially were considered in the context of
(based on ICT technology in the power grid) of these vehicles transmission networks, today PMU applications are considered
will be essential to deal with these challenges and to benefit to also improve the observability of the distribution grid. These
from the opportunities. safety- and time-critical applications clearly need fast com-
munication networks, with requirements beyond best-effort
internet technologies. Therefore, there is a need for modeling
C. Distributed renewable energy sources (DRES) the communication network and evaluating its impact on
Another major cause of the smart grid challenges stems modern protection and control schemes [14], [15].
from distributed renewable energy sources (DRES): their large
scale deployment has a significant impact on the power system,
II. M OTIVATION
since the output of solar and wind power is difficult to control
given its dependence on variable local weather conditions. To study aforementioned smart grid innovations, simulation
Therefore, the effect of such distributed generation (DG) units is considered an important tool. However, writing a new
on system stability is less predictable than on-demand sources simulation engine from scratch is complex, costly and time
4 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

consuming [3], [14], especially if we consider the interdisci- Simulation is also considered an important tool for educational
plinary nature of the smart grid comprising both power system and research support [17].
engineering and ICT as key components. The alternative, i.e.,
reuse existing (off-the shelf, commercial) simulation environ-
B. Smart grid simulator design challenges
ments as is, or combine them into a (distributed) simulation
environment, may have the benefit of better reliability and In this section, we further motivate the need for smart grid
scalability [3]. However, the interdisciplinary nature of the simulators, and also discuss the challenges associated with
smart grid complicates the assurance of the model validity for the design and development of smart grid simulators. The
both power and communication networks, requiring extensive provided information not only assists developers in the devel-
expertise of the most appropriate tools (and their settings) for opment process, but also enables users to evaluate the different
both domains. solutions. We discuss (i) the need for combined simulation
As such, the primary objective of this survey is to provide of the power system and ICT infrastructure, (ii) selection
a comprehensive overview of existing simulation tools in the of the appropriate abstraction level for simulation models,
individual fields of power systems and communication net- (iii) requirements for simulation scenario’s, (iv) differences in
works, and the interdisciplinary field of smart grids combining modeling time, and (v) practical considerations such as user
power and ICT simulation. To assist in selection of the right friendliness, flexibility, etc.
tool for the job, this survey provides a detailed overview and The underlying challenge associated with smart grid simula-
classification of existing tools and their capabilities, illustrated tion is that it requires combined simulation of both the power
by example use cases. . system and the ICT infrastructure, as well as the applications
Although reusing existing simulation tools offers many (e.g., control algorithms) running on top of them, especially
benefits, it is sometimes necessary to design custom tools, e.g., considering the large scale those systems [17], [18]. As pointed
due to missing features. Therefore, the secondary objective of out previously, the operation of the power grid increasingly
this survey is to give insights in the design and implementation depends on ICT [21] and it is therefore crucial to under-
of smart grid simulators, indicating common pitfalls, lessons stand the impact of the performance of the communication
learned from earlier experiences, and methods to integrate network on the operation of the power grid [17], [23]. The
different simulators. smart grid, comprising many heterogeneous communicating
Next we first motivate the use of simulation tools for smart devices, thus needs to deal with issues such as safety, security
grid research, and continue by pointing out the most apparent (including protection against potential cyber attacks [17]),
challenges in designing such tools. interoperability, and performance [24]. Yet, current power grid
simulators typically do not model the network communication
A. Why simulation? protocols, or even traffic patterns involved in such a smart
Historically, simulation is an important tool for the design grid [14], [24]. On the other hand, the operating mode of the
of power systems [16]–[18] as well as communication net- smart grid has an impact on the traffic in the communication
works [19]. Communication network simulation environments network [23]. Thus, integration of power and ICT components
are used to develop and evaluate new ICT architectures and in the operational power grid also requires similarly integrated
network protocols, while similarly power system engineers simulation frameworks [17].
use simulation environments for power system planning and A first main challenge that thus arises is to decide on the
operations. In a smart grid context, simulators allow to study appropriate abstraction level for smart grid simulator models,
complex interactions between these interconnected systems that should cover the power grid, and ICT components ranging
and the monitoring and control elements on top of them [20]. from the communication network, middleware (e.g., [13],
Motivations for resorting to simulation has both economical [25]), control strategies (which constitute the key smart grid
and practical origins. Simulation is used to reduce the costs innovations, see Section I), etc. One of the key challenges is
associated with upgrades to the power system and communi- the different time resolution (see below) and fidelity of the sim-
cation network infrastructures: costs related to performing the ulation [20]. Furthermore, the simulator should allow flexible
upgrades (installation, testing, etc.), but also to the potential specification of varying scenarios [20], and possibly definition
loss of service that can occur as a consequence. Indeed, of the level of detail (e.g., time resolution). In this respect,
upgrades can have severe economic and social impacts, even scalability is an important concern: simulators should scale
for a short period of time [21]. Simulation reduces these risks, to support the complexity of modern large scale smart grid
enabling the design and evaluation of different solutions before scenario’s, e.g., when considering nation wide smart grids. As
actually deploying them the in the field, and moreover in a such, deciding on the level of modeling detail has to account
fully controlled environment. The latter implies that future for computational efficiency [17]. Furthermore, simulations
power systems or communication networks can be studied should not only aim to achieve technical objectives, but also
under varying conditions and for different scenario’s [20]. consider financial and business criteria as dictated in industry
Another benefit is that simulation can happen faster than standards [26].
real-time, depending on the complexity of the simulation On the modeling part, it should be noted that traditional
model [22]. This can reduce the time required to develop new simulation tools will need to be extended with models specific
technologies. Therefore, simulation offers much more flexibil- of the advanced smart grid scenarios. On the power side, this
ity compared to studies that depend on real-life deployments. includes appropriate characterization of renewable sources:
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 5

dealing with their intermittent and stochastic behavior is a

crucial research topic [17]. In view of the DR approaches,
correct modeling of the user behavior [26], and especially the
flexibility of his load (e.g., time shifting of appliance usage,
state-of-charge and charging deadlines for EVs), is crucial.
Such models should be accompanied by explanatory meta-
data to allow correct application of the models, respecting the
assumptions under which they were constructed.
Another complexity stems from different models of time by
various simulators: continuous simulation is common in power
systems, whereas communication network simulators typically
are discrete event simulators [3], [15], [20], [27]. Thus, when (a) Co-simulation
combining such tools in so-called co-simulation approaches
(see Section III), synchronizing the time of different co-
simulation components is a recurring topic [3], [14], [22],
[28]. Clearly, the synchronisation of various simulation model
constituents has to be carefully managed, as we will explain
in Section III-B.
Beyond aforementioned technical aspects, the design of a
smart grid simulator should also take into account more practi-
cal aspects, including user friendliness. Not only is simulation
is an important tool to support education and research [17],
[29], [30], consumer involvement in smart grid simulation
is also considered [17], [30]. As such, a smart simulator
should be an open and flexible environment, that supports user- (b) Integrated simulation
defined models [17], and easy reuse of already established and
validated models. The latter suggests that possible integration Fig. 1. Conceptual approaches to combining power and communication
network simulation: (a) Co-simulation: Multiple simulators with specialized
with different programming languages could give such support tasks, each having their own simulation interface for data input/output, control,
to a broad audience [17], [20]. To achieve this, the use of etc. The arrows indicate that interaction between the simulators is required.
a common simulation interface and existing communication (b) Integrated or comprehensive simulation: One combined simulator provides
an integrated environment for combined simulation of power system and ICT.
methods (e.g., web services) is suggested, as to enable inte-
gration of existing models, independent of the programming
language or simulation tools used [20]. Related to this is
the use of data formats for input/output: simulators should extensively in both domains, it is the combined simulation
limit the dependency on proprietary input formats, operating of the power system and communication network that has
systems or third party libraries. Ideally, a smart grid simulator recently attracted more attention due to rising interest in smart
should be able to incorporate actual power system components, grid from governments, industry, and academia. It can be
i.e., hardware in the loop simulations [17], [18], [23]: thus, achieved using a variety of approaches, of which two will be
existing components can be tested in a controlled environment, discussed in more detail: (i) co-simulation, (ii) comprehensive
or used as building blocks to speed up development. However, or integrated simulation.
this requires real-time operation of the simulator and hence Constructing a new combined simulation environment is
appropriate modeling of time. potentially time-consuming and expensive. Therefore, a co-
simulation approach combines existing specialized simulators.
III. S MART G RID S IMULATION PARADIGMS In the context of smart grid co-simulation, a co-simulator
In the following sections we will present simulation envi- would consist of a specialized communication network simula-
ronments that are used for simulating power systems, commu- tor (e.g., OMNeT++) and a specialized power system simulator
nication networks, as well as their combination in the context (e.g., OpenDSS). Figure 1(a) illustrates the co-simulation
of smart grids. First however, we will discuss the overall approach. Multiple simulators are used, each having their
simulation paradigms they are built on. After sketching how own distinct simulation interface for data input, configuration,
to combine power and ICT simulation constituents, we will result output, control, etc. Therefore, the main challenge is to
outline specific time modeling approaches and the complexity connect, handle and synchronize data and interactions between
of combining them. both simulators using their respective simulator interfaces.
Especially time management between both simulators is chal-
lenging, because each simulator manages their simulation time
A. Combined simulation of power and communication systems individually. Nonetheless, the main advantage is that existing
We briefly discuss the combined simulation of the power simulation models, algorithms, etc. that have already been
system and communication network. Although power sys- implemented and validated can be reused. Indeed, the majority
tem or communication network simulators are being used of the development effort is put into modeling of additional,
6 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

smart grid specific components: systems such as photovoltaics,

wind turbines, etc. and composite sub-systems such as low
or medium voltage power grids [20]. Hence, a co-simulation
approach reduces development time and the risk of errors.
Notwithstanding the development advantages, running the (a) Continuous time simulation
simulators separately and the necessary synchronization likely
will imply performance penalties. E.g., in [31] the authors
present an example in the context of video streaming where
synchronization overhead accounted for 90% of the total
simulation time. To further illustrate potential performance (b) Discrete event simulation
examples, we consider a co-simulation approach in which each
simulator is run in sequence. For each simulation run, the
simulation environment must be loaded (i.e., start-up time is
the performance penalty), configured and input data must be
provided (i.e., reading and processing configuration and input
data is the performance penalty). Next, the simulation model is
executed and results are gathered and output. Data input/output
often requires intermediaries to store the data, e.g., files on a
file system, a database, web services, etc., in which case the
access time and the time required to read the data will incur
a performance penalty. Also, input/output data must be pre-
processed before using it in a next step (e.g., due to different
file formats used), introducing pre-processing delay. (c) Synchronisation issues
An alternative for co-simulation is an integrated or com-
prehensive approach to simulation, in which both the power Fig. 2. Continous time vs discrete event simulation: (a) Time stepped sim-
ulation of a continuous time simulation model. (b) Discrete event simulation
system and communication network are simulated in one (DES). (c) Example of simulation errors in an approach based on predefined
environment. Figure 1(b) illustrates the concept. A single synchronization points.
simulation interface is provided, instead of having distinct
interfaces for each simulator. Another advantage of this tightly
coupled approach is that the management of time, data, modeled as discrete event systems characterized by events
and power/communication system interactions can be shared such as sending and receiving of packets, expiration of timers,
among the simulator constituents. Hence, no performance etc. Such events occur unevenly distributed in time. This is
penalty due to synchronization is expected. However, the main clearly different from the time stepped approach commonly
challenge is the combination of both models in one environ- used for power system dynamic simulation, where a fixed
ment. The main challenge is to provide a simulation interface interval between events is selected. An event scheduler is
that provides sufficient level of detail for the different aspects responsible for maintaining a time-ordered list of all scheduled
of the smart grid simulation model. A possible implementation events, and simulation time progresses from event to event as
approach to integrated simulation is to select a communication sketched in Fig. 2(b).
network, power system or other platform as the basis for the One approach to combine both approaches is the use of
smart grid simulator, and implement the other components predefined synchronization points, indicated by the dashed
from scratch or link existing libraries or tools. lines in Fig. 2(c). Each simulator pauses when their simulation
clock reaches a synchronization point. After each simulator is
paused, information is exchanged. This however can lead to
B. Continuous time and discrete event simulation models simulation inaccuracies: messages that need to be exchanged
As stated earlier, power system and communication net- between both simulators are delayed if they occur between
work simulators tend to adopt different modeling approaches. synchronization points. A solution to this problem is to
Dynamic power system simulation commonly uses continu- reduce the time step between synchronization points (and
ous time modeling, where state variables are described as possibly refining the timescale used for the continuous time
continuous functions of time. Thus, power system element simulator), yet this clearly degrades performance. Thus, co-
dynamics are expressed by differential equations defining the simulation needs to strike the right balance between accuracy
relations between continuous state variables. However, some and simulation speed. Also, not all time instants at which
discrete dynamics are introduced by circuit breakers, relays, communication between the different simulators must occur
etc. Hence, a time stepped approach is used: since exactly are known a priori.
solving the equations analytically is only possible for trivial
cases, numerical algorithms using discrete time slots are used. C. Emulation, Real-Time Simulation and Hardware-in-the-
This leads to the time model illustrated in Fig. 2(a). Loop Simulation
Communication networks typically are packet switching So far we only considered pure software-based simulation
networks (cf. IP based technologies), which are adequately approaches, i.e., both power grid and ICT infrastructure are
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 7

time model not only depends upon the numerical computation,

but also on the timeliness with which the simulation model
interacts with external components (hardware or software).
Applications of real-time simulation include testing of physical
control and protection equipment.
(a) Offline or non real-time simulation: fast Hardware-in-the loop (HIL) simulation is a technique used
to develop complex real-time embedded systems (e.g., in the
domain of power electronics) in which some components
are real hardware, whereas others are simulated. Components
may be simulated because they are unavailable, or because
experiments with the real components are too costly, time
consuming, or are too hazardous. Typically, a mathematical
(b) Offline or non real-time simulation: slow
model of the simulated system is used to provide electrical
emulation of sensors and actuators that are connected to real
hardware.

IV. P OWER S YSTEM S IMULATION

(c) Real-time simulation In this section we discuss power simulation, mainly target-
ing readers with an ICT background: we introduce different
Fig. 3. Non real-time (offline) simulation and real-time simulation: (a) Non power simulation types, and an overview of existing power
real-time simulation in which computation takes less time than the simulated
event: simulation clock progresses faster than the real-time clock. (b) Non simulators, in terms of their main features, example studies,
real-time simulation in which computation takes more time than the simulated and options for integration of external tools.
event: simulation clock progresses slower than the real-time clock. (c) Real- Simulators for power system analysis have been extensively
time simulation: simulation clock and real-time clock are synchronized.
used by professionals for network planning, operations and
price forecasting. Over-voltages, harmonics, short circuits,
transient stability, power flow, and optimal dispatch of generat-
simulated: the physical world components are abstracted as ing units are examples of important power system phenomena
software models. However, some approaches aim for a more that need to be captured and parameterized in the simulations.
realism and therefore provide support for emulation, real-time Power system simulations are usually classified into one of
simulation, and/or hardware-in-the loop experiments. In this these two categories:
section, we provide an introduction to these concepts. 1) Steady state simulations form the basis for power grid
In an emulation approach (integrated or co-simulation), the network planning studies. Researchers and engineers perform
emulated component more closely mimics the real world in “what-if” studies to measure the impact of modifications in the
hardware. For example, a network emulator such as Emu- power system. The system is analyzed in a stable equilibrium
lab [32] can be used instead of simulators such as ns-2/ns-3 state, and focus lies on checking whether the power system
or OMNeT++, resulting in a more realistic but still control- variables are within proper boundaries (e.g., validation of
lable environment: i.e., Emulab allows specifying an arbitrary voltage limits). The different simulators specialized in steady
network topology, resulting in a controllable, predictable, and state studies offer a full range of analysis methods, from
repeatable environment. To provide an even higher level of power flow studies, load estimation and load balancing, to
detail, it is possible to use actual smart grid components, fault analysis or optimal capacitor placement. Steady state
e.g., GridSim [18] uses the GridStat [33] communication simulations also cover optimal power flow studies. In these
middleware platform. studies, the system conditions that minimize the cost per kW/h
Next, we discuss real-time simulation. The difference with delivered are analyzed using linear optimization. Other optimal
non real-time or offline simulation is illustrated in Fig. 3. power flow methods that incorporate Artificial Intelligence
Figure 3(a) and Fig. 3(b) illustrate two possible scenarios for (AI) techniques are described in [35].
non real-time simulation: the simulation clock can progress 2) Transient dynamics simulations study transitions between
either faster than the real-time clock (i.e., time in the real equilibrium points due to a major changes in the power grid
world) or slower. However, in a real-time simulation approach, configuration, e.g., disturbances. A major goal of such studies
the simulation clock and real-time clock are synchronized as is to determine if the load angle reaches a new optimal
illustrated in Fig. 3(c). For these examples, we have assumed steady state. Simulations performed include electromagnetic
a simulation model with discrete time and constant time step transient studies with finer time granularity (in the order of
(see also Section III-B). Note that techniques exist for sup- microseconds to milliseconds) than the steady state ones. In
porting variable time steps, but they are less suitable for real- these simulations, time varying and short term signals are
time simulation [34]. Put more formally, a real-time simulator studied. If the equilibrium is lost due to continuous small dis-
must accurately produce the internal variables and outputs of turbances, dynamic stability simulations, also known as small-
the simulation model within the same length of time as its signal stability simulations, are needed. Simulators specialized
real-world counterpart would. I.e., the correctness of a real- in transient dynamic power characteristics enable to model the
8 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

Fig. 4. Time scales of different power phenomena and power control: depending on the time scale, different model representations are adopted. The time
scale considered depends on the use case, which typically is related to a particular part of the grid (generation, transmission, distribution, etc.).

network at circuit level, reproducing the time domain wave A. PSCAD/EMTDC

forms of state variables at any point in the system. PSCAD/EMTDC is a commercial simulation tool for the
Power System Computer Aided Design and Electromagnetic
In addition to the “steady state” vs “transient dynamic” transients for DC. An example of PSCAD/EMTDC simula-
classification, power system simulations usually focus to one tions of power system control in a smart grid context is [36],
of the hierarchical power grid domains: Power Generation, where Fazeli et al. present a novel integration of wind farm
Transmission, Distribution or Utilization (residential, commer- energy storage systems within microgrids. PSCAD/EMTDC
cial and industrial loads). Depending on the domain of interest can be coupled with external tools like Matlab, as exemplified
and the power phenomena, the time steps of the simulation in [37], where Luo et al. combine PSCAD/EMTDS’s elec-
would vary. Figure 4 gives an overview of the timescale for tromagnetic transient simulation capability and with advanced
different phenomena and control strategies in power systems. matrix calculations in Matlab for testing a new network based
Phenomena that require higher frequency studies (transients) protection scheme for the power distribution grid. Similarly,
would require a smaller duration of calculation time steps. Mahmood et al. have designed a three-phase Voltage Source
Note that such smaller time steps would deliver more accurate Converter (VSC) for distributed generation, developed their
results, but come at the price of increasing the total simulation linear model in Matlab and validated it using a detailed
runtime [11]. Figure 4 also captures the different power system switching model in PSCAD/EMTDC [38].
domains, example studies and the mathematical representation
of the various power phenomena. The top part of the diagram
focuses on steady-state analysis, while the bottom groups the B. DigSilent - PowerFactory
transient dynamics. DigSilent Power Factory allows the modelling of generation,
transmission, distribution and industrial grids, and the analysis
As pointed out in Section II-B, smart grids pose specific of their mutual interactions. Load flow, electromechanical
challenges, such as high penetration of renewable DG units RMS fluctuations and electromagnetic transient events can
and microgrid operation, implying importance of energy stor- be simulated. Thus, both transient grid fault and longer-
age and decentralized energy management. In energy transmis- term power quality and control issues can be studied. As an
sion and distribution, the increment in sensing and communi- example of power flow studies using DigSilent, Coroiu et al.
cation capabilities enables new automation and control strate- evaluate the continuity of power supply using the comparative
gies for remote condition monitoring or blackout prevention. methods of the probabilistic load flow and the stochastic load
Moreover, new intelligent consumption strategies are possible flow [39]. Transient studies is performed by e.g., Chen et al. ,
thanks to more frequent meter readings, demand response who studied the transient stability of a micro-grid supplied
plans and smart appliances with different load management by multiple distributed generators [40]. Models of voltage
features. These all need to be appropriately modeled. In the controllers, generators, motors, dynamic and passive loads,
following subsections, we present an overview of the main transformers, etc. are part of DigSilent’s built-in electrical
simulators found in research literature and illustrative applica- components library, but the algorithms inside these models
tions thereof in smart grid studies. We also indicate interfaces are not accessible. However, users can create models using
offered by the simulation tools to expand its functionality, and the DigSilent Simulation Language (DSL). An example of
e.g., link with other components to realize co-simulation. such a study on dynamic wind models can be found in
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 9

Table I
C LASSIFICATION OF POWER SIMULATORS

Simulation Type Power Subsystem - Domain License

Simulator Steady State Transient Dynamics Generation Transmission Distribution + Commercial Open
(min, hours, days) (s, ms, µs) RCI loads Source
Cymdist x x x
DigSilent x x x x x x
EMTP-RV x x x x
ETAP PSMS x x x x x x
EuroStag x x x x x
homer x x x (v.2.68Beta)
ObjectStab x x x x
OpenDSS x x x x x
PowerWorld x AO x
PSCAD/EMTDC x x x x
PSS R
E x x x x x
PSS R
Sincal x AO x x x

RCI: Residential, Commercial and Industrial loads AO: Add-on

[41]. In addition, DigSilent supports the exchange of power Napolitano et al. use transient modeling using EMTP-RV
data with external tools. For example, in [42] Andren et al. software to model the MV feeder response to indirect lightning
combine DigSilent with Matlab andt present a framework for strokes [47]. Other potential uses of EMTP-RV include studies
the simulation of power networks and their components, using in insulation coordination, switching surges, capacitor bank
an Open Process Control (OPC) interface for exchanging data switching, motor starting, etc. Users can develop customized
between simulators. modules and interface them to EMTP-RV via dynamic-link
library (DLL) functionality.
C. Siemens PSS
R

E. PowerWorld
The Power Systems Simulator (PSS R
) product suite in-
cludes several software solutions targeting different domains PowerWorld Simulator is an interactive, visual-approach,
and time scales. Among others, PSS includes PSS SINCAL power system simulation package designed to simulate high
and PSS E. PSS SINCAL targets utility distribution system voltage power system operation on a time frame ranging
analysis: it is a commercial (with special licenses for research from several minutes to several days. PowerWorld’s add-
and education) network planning and analysis tool with capa- on SimAuto allows to control the simulator from external
bility to perform, among others, power flow, load balancing, applications. SimAuto acts as a Component Object Model
load flow optimization and optimal branching simulations. PSS (COM) object for interfacing with external tools, such as
SINCAL’s COM-server interface facilitates the integration into Matlab or Visual Basic. Such combination is illustrated by
existing IT architectures. The COM interfaces can be exploited Roche et al. , who combine PowerWorld with external artificial
in Smart Grid simulations, where PSS SINCAL can be used intelligence (AI) decision making tools to realize smart grid
in the analysis of distributed generation and smart meter data. simulations studying feeder reconfiguration and large-scale
As an example of such studies, Chant et al. investigate the demand response [48].
impacts on integrating photo voltaic panels on the utility grid
in terms of harmonic distortion, voltage fluctuation and load F. ETAP PSMS
rejection issues [43]. PSS SINCAL allows users to link each ETAP PSMS is a real time power management system.
Smart Grid equipment model (e.g., e-cars, micro-turbine, smart ETAP software has more than 40 modules for load flow
meter, etc. ) with their correspondent generation and load analysis, short-circuit analysis, device coordination analysis,
profiles [44]. For transmission system planning, the PSS E motor starting analysis, transient stability analysis, harmonic
tool allows users to perform load flow analysis and transient analysis, etc. In [49], Mehra et al. applied principal component
analysis. For example, Mohamad et al. use PSS E for transient analysis (PCA) to simulated phasor data, generated by ETAP
stability analysis [45]. software.
PSS E can interact with user scripts using the Python
scripting language. Such integration is used by Hernandez
G. Cymdist
et al. : modeling Synchronous Series Compensators (SSSC)
in Python, they simulate the control of power flow through Cymdist is designed for planning studies and simulating
transmission lines [46]. the behavior of electrical distribution networks under different
operating conditions and scenarios. It offers a full network
editor and it is suitable for unbalanced load flow and load bal-
D. EMTP-RV ancing studies. The software workspace is fully customizable.
EMTP-RV is a commercial software for simulations of The graphical representation of network components, results
electromagnetic, electromechanical and control systems tran- and reports can be built and modified to supply the level of
sients in multiphase electric power systems. For instance, detail needed. Furthermore, the CYME COM module allows
10 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

different environments to communicate with the CYMDIST Table II

software for accessing different pre-defined functions and C LASSIFICATION OF M ATLAB - BASED POWER SIMULATORS
calculations. An illustrative distribution system modeling study Package PF CPF OPF TD EMT SSA
using Cymdist can be found in [50]. DCOPFJ x
EST x x x
INTERPSSS x x x x
H. EuroStag MatEMTP x x
EuroStag is a power systems dynamics simulator developed MATPOWER x x x
PAT x x x
by Tractebel Engineering GDF SUEZ and RTE (electricity PSAT x x x x x
system operator of France). It allows a range of transient PST x x x x
and stability studies. Supplementary tools, such as Smart PYLON x x
SIMPOWER x x
FLow, enable load flow calculations. An example of such SPS x x x x
studies can be found in [51], where Asimakopoulou et al. TEFTS x x
compared various load control scenarios for the power system VST x x x x
in the island of Crete, using EuroStag as the basis for their
PF: Power Flow CPF: Continuation Power Flows
simulations. OPF: Optimal Power Flow TD: Time Domain
EMT: Electromagnetic transients SSA: Small-signal Stability Analysis

I. Homer
HOMER is a power generation simulator. It can be used for industrial power systems, etc. It supports simulation of very
designing hybrid power systems containing a mix of energy large power grids with a time step as low as 20 microseconds.
sources: conventional generators, combined heat and power, It can also be used for simulation of power electronics found
wind turbines, photo voltaics, batteries, etc. Both grid tied in distributed generation (e.g., wind farms, photo voltaic cells)
or standalone systems can be simulated. In addition, green and Plug-in Hybrid Electric Vehicles (PHEV). RT-LAB [57]
house calculations are also possible. An illustrative micro grid is the core technology behind eMEGAsim and enables dis-
sizing and dynamic analysis study using Homer and EuroStag tributed real-time simulation and hardware-in-the-loop test-
is presented in [52]. ing of electrical, mechanical, and power electronic systems,
and related controllers. ARTEMIS is a suite of fixed-step
J. OpenDSS solvers and algorithms that optimize real-time simulation of
SimPowerSystems [58] models of electrical, power electronic,
OpenDSS is an open-source distribution system simula-
and electromechanical systems. Opal-RT products are fully
tor developed and maintained by EPRI. It is designed to
integrated with MATLAB/SimuLink.
support power distribution planning analysis associated with
The Real-Time Digital Simulator (RTDS) [16] from RTDS
the interconnection of distributed generation to the utility
Technologies [59] is a power system simulator that solves
system. Other targeted applications include harmonic studies,
electromagnetic transient simulations in real-time. It supports
neutral-earth voltage studies, volt-var control studies, etc. Co-
high-speed simulations, closed-loop testing of protection and
simulation interfaces (e.g., COM and scripting interfaces)
control equipment, and hardware-in-the-loop applications. Par-
are provided and users can define their own models [53].
allel processing techniques enable the simulation of large scale
OpenDSS is considered a suitable platform for smart grid re-
power systems: power system equations are solved fast enough
search as it supports the analysis of intermittent and stochastic
to continuously produce output conditions that realistically
processes associated with renewable energy sources [17].
represent conditions in the real network. RTDS supports IEC
61850 device testing. As a result, the simulator can be con-
K. ObjectStab nected directly to power system control and protective relay
ObjectStab [54] is an open source power system library with equipment.
capabilities to perform power system transient simulations.
It is based on Modelica, a general purpose object oriented M. Classification
modeling language. An example of high voltage DC (HVDC)
power transmission studies can be found in [55], where Meere A characterization of the previously mentioned simulators
et al. designed optimised power system models for variable can be found in Table I, which presents a classification of
speed wind turbine machines with a HVDC link for grid popular power simulators according to the time-scale of the
interconnection. The electrical performance of the system is simulations (steady-state vs transient), the domain (power
verified using ObjectStab. generation, transmission, distribution, consumption) and their
licensing (open-source vs commercial).
In addition, simulation platforms based on Matlab/Simulink
L. Real-time hardware-based simulation environments are also widely used. Examples of power system
Opal-RT [56] develops real-time digital simulators and simulators based on MATLAB include Power System Analysis
hardware-in-the-loop testing equipment. eMEGAsim from Toolbox (PSAT) [60], Power System Toolbox (PST) [61], Ed-
Opal-RT is a real-time hardware-based simulator that has ucational Simulation Tool (EST) [62], SimPowerSystem [58],
been developed to study, test, and simulate large power grids, Power Analysis Toolbox (PAT) [63], Voltage Stability Toolbox
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 11

(VST) [64] and MATPOWER [65]. Note that although several (instead of OTcl), improved scalability, more attention to
of these tools are open source, MATLAB is a commercial realism, better software integration, etc. [69]. However, when
and closed product. Yet, PSAT can also run on GNU/Octave, selecting a specific version of ns, it is important to consider
which is a free Matlab clone, therefore resulting in a complete that ns-3 is not backwards compatible with ns-2: i.e., existing
open source solution that is freely available. In addition, ns-2 simulation models must implemented again for ns-3.
PYPOWER is a translation of MATPOWER to the Python Both are widely used for networking research in general, and
programming language. Table II summarizes the different unsurprisingly also in a smart grid context both ns-2 and ns-
MATLAB modules and their capabilities, based on [17], [60], 3 are adopted in e.g., a co-simulation approach [11], [22],
[64] [24], [27], [70], [71]. In [72] a suite of software modules
Note that in addition to the major tools discussed above, for simulation of PLC networks using ns-3 is presented and
additional open source tools are described by Milano et al. source code is made available at [73].The simulation model is
in [66]: UWPFLOW (power flow, implemented in the C pro- based on transmission line theory (TLT), which relies on the
gramming language), TEFTS (transient stability, C), InterPSS knowledge of the topology, wires, and the load characteristics
(load flow and transient studies, in Java), AMES (whole sale of the power grid underlying the PLC system. This approach
power market, Java), DCOPFJ (DC optimal power flow, Java) supports networks with multiple node-to- node links. An
and PYPOWER (DC and AC power flow and DC and AC interface to the ns-3 framework is provided, which allows the
optimal power flow). integration of higher level protocols such as TCP/IP. A GUI is
provided that enables users to draw the topology and specify
V. C OMMUNICATION N ETWORK S IMULATION node and line properties, and also noise present in the network.
In this section, we present an overview of communication
network simulators, which are widely used for the develop- B. OMNeT++
ment and evaluation of communication architectures and pro-
The open-source OMNeT++ discrete event simulation envi-
tocols. We present a short overview of the different simulators
ronment [74] has been designed for the simulation of commu-
that have been successfully used in a smart grid context:
nication networks (wired and wireless) and distributed systems
ns-2/ns-3, OMNeT++, NeSSi and OPNET Modeler R
. This
in general. The simulation environment has a general design
section will primarily serve readers with a power systems
(i.e., it is not limited to simulating communication networks)
background, since ICT experts will be presumably be familiar
and therefore has been used in various domains, such as
with some of these tools. Yet of particular interest for ICT
wireless network simulations, business process simulation
researchers will be the highlighted sample smart grid use cases
and peer-to-peer networking. However, OMNeT++ is mostly
for which they have been used. We limit our selection of
applied in the domain of communication network simulation.
examples to those that focus on the communication aspects in
A comprehensive set of internet based protocols is provided
the smart grid, and as such do not require (detailed) modeling
by means of the INET framework extension which includes
or simulation of the electric behavior of the power grid.
support for IPv4, IPv6, TCP, UDP, Ethernet, and many other
Simulators and use cases that focus on the combined sim-
protocols. Other extensions provide simulation support for
ulation of the power system and communication network are
mobility scenarios (e.g., VNS), ad-hoc wireless networks
considered in Section VI. Note that general purpose tools such
(e.g., INET-MANET), wireless sensor networks (e.g., MiXiM,
as MATLAB have also been applied to study communication
Castalia), etc. Distributed parallel simulation is supported to
networks in a smart grid context [67], [68], but we will not
enable simulation of large scale networks. Additionally, feder-
further elaborate on those studies here.
ation support based on the High-Level Architecture (HLA)
standard is provided in OMNEST, the commercial version
A. Network Simulator (ns-2 and ns-3) of OMNeT++. An OMNeT++ simulation model consists of
The Network Simulator version 2 (ns-2) is a widely used simple modules implemented in C++. Compound modules
open source discrete event network simulator created for consist of other simple or compound modules, and are defined
research and educational purposes. It is targeted at networking using the OMNeT++ Network Description Language (NED).
research, with a strong focus on internet systems. Therefore, it Modules communicate by passing messages via gates, which
includes a rich library of network models to support simulation are the input and output interfaces of the modules that are
of e.g., IP-based applications (including TCP, UDP, etc.), rout- linked to each other by so-called connections forming com-
ing, multicast protocols, over wired and/or wireless networks. munication links between modules. Apart from the networking
The ns-2 core is written in the C++ programming language. community, OMNeT++ has also received substantial attention
Users can create new network models or protocols using the from the smart grid community for developing smart grid
C++ language. Simulation scripts to control the simulation and simulators [5], [29], [75]–[80].
configure aspects such as the network topology are created Example use cases that focus on the communication aspect
using the OTcl language interface. As a result, users can of the smart grid include the design and evaluation of different
create and modify simulations without having to resort to C++ smart grid communication architectures, performance of smart
programming and recompiling ns-2. Development of ns-3, the grid protocols, etc. For example, a demand side management
successor to ns-2, is ongoing: new features include support communication architecture based on orthogonal frequency-
for the Python programming language as a scripting interface division multiplexing (OFDM) power line communication
12 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

(PLC) is proposed in [76], [77]: the authors test business cases user interface offering access to a large library of C and
and benchmark overall network performance in a controlled C++ source code blocks, representing the different models
environment, and use OMNeT++ results to iteratively improve and functions. It comes with an open interface for integrating
the network design. As part of that research, a full simulation external object files, libraries, other simulators (co-simulation)
model of PRIME protocols has been developed that enables and even hardware-in-the-loop.
simulation of IP communication over a PLC network. Another The Smart Grid Communications Assessment Tool (SG-
PLC simulation model for OMNeT++ is presented in [81]. It CAT), introduced in [84], is a simulation, modeling and
is a generic model that does not implement a specific variant analysis platform, targeted to utilities that want to develop
of PLC, but provides a toolkit that should enable the user to a holistic smart grid communications strategy. It has been
model the desired PLC variant. Simulation of broadband PLC developed to assess the performance of different smart grid
in a home environment is demonstrated. applications under various terrains, asset topologies, technolo-
Another example is a simulation environment to study gies and application configurations. SG-CAT has been built
geographical routing in multi-hop wireless networks in the on top of OPNET Modeler, taking advantage of OPNET’s
context of smart grid energy applications [78]. There, the modular design, which allows the exchange and customization
authors purely focus on communication, i.e., without power of applications, communication technologies, terrain profiles
system modeling and simulation. That work is extended and and path-loss models. The same authors also discuss the
a modular and distributed simulation environment is proposed scale-up concerns when approaching large scale simulations
in [79], focusing on scalability analysis of smart grid ICT in OPNET, and offer a solutions to these challenges based on
infrastructures. It allows distributed simulation and provides the unique characteristics of smart grid scenarios [85].
additional simulation management features (scenario genera- Furthermore, OPNET is used in multiple co-simulation
tion, model repository, dependency management, management approaches (see further in Section VI) that consider both the
GUI, etc.). Main research questions include topology-specific communication network and power system in detail [15], [28],
influences on the scalability of different technologies and [86]–[88]. Smart grid use cases that focus on the communi-
various traffic patterns for smart grid applications. cation network without detailed modeling of the power grid
A last example is related to the evaluation of smart grid are described in [89]–[91]. The authors of [89] consider a
standards and protocols. An important standard in smart grids wide area monitoring and control scenario system that uses
is the IEC 61850 standard, targeted at substation automation. a WiMAX/IEEE 802.16 network to transport delay-sensitive
A IEC 61850 simulation platform is described in [29] based PMU data: several IEEE 802.16 scheduling services (UGS,
on OMNeT++. The platform is designed to support commu- rtPS, BE) are evaluated in terms of delay, uplink use and
nication network performance analysis, hardware-in-the-loop signaling overhead, using a simulation model developed in
simulations, and algorithm development and evaluation. An OPNET. The same authors also propose a heterogeneous
overview of other IEC 61850 simulation platforms that are WiMAX-WLAN network architecture for advanced metering
limited to communication network performance analysis is infrastructure (AMI) communications [90], and compare the
also presented in [29]. performance of the WiMAX-WLAN network architecture to
that of a pure WLAN network architecture. In [91], the authors
C. NeSSi study the performance of a Long Term Evolution (LTE) based
networks (frequency- vs time-division multiplexing mode) for
NeSSi (Network Security Simulator) is an open source
up-link biased smart grid communication in terms of latency
discrete event network simulator developed at DAI-Labor
and channel utilization.
(Distributed Artificial Intelligence Laboratory) and sponsored
by Deutsche Telekom Laboratories. We include NeSSI because E. Discussion
the primary focus of the tool is on network security related
The communication network simulators discussed in this
scenarios in IP networks [82]. Features described to support
section have been used successfully in context of smart
security related scenarios are attack modeling, attack detection,
grid research. OMNeT++ and ns-2/ns-3 are used extensively
security metrics, etc. Distributed simulation is supported to
in academia due to their open-source nature. In terms of
enable simulation of large scale networks. Example uses in
supported simulation models, we believe that a wide range
the smart grid domain include a security analysis of a smart
of models is available for each simulator, and the choice
measuring scenario through federated simulation [83] and to
mainly depends on prior knowledge and preferences of the
use an integrated approach for evaluating and optimizing an
user regarding modeling language and tools, extensibility and
agent-based smart grid management system [82].
supported programming languages, presence of extensive GUI
tools, etc. For example, OMNeT++ and NeSSi provide an
D. OPNET Modeler
R
integrated development environment (IDE) that includes GUI’s
OPNET Modeler R
is a powerful commercial discrete event for building and configuring simulation models, visualization
network simulator with built-in, validated models including of topologies, result processing, etc. However, ns-2/ns-3 lacks
LTE, WIMAX, UMTS, ZigBee, Wi-Fi, etc. It enables mod- an extensive set of GUI tools as found in OMNeT++, making
eling of various kinds of communication networks, incorpo- it more complex in its usage. OPNET Modeler R
on the
rating terrain, mobility, and path-loss characteristics in the other hand is a commercial simulator that has a visual high-
simulation models. OPNET Modeler has a visual high-level level interface. Another aspect that may influence the choice
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 13

of simulator is commercial support, which is available for

OMNeT++ (i.e., OMNEST) and OPNET. NeSSi, also an open-
source simulator, distinguishes itself from the other tools due
to its primary focus being network security.

VI. S MART G RID S IMULATION

In this section, we present an extensive overview of smart
grid simulators, i.e., those that support the combined sim-
ulation of the power system and the communication net-
work, and/or model and study higher layers such as market
mechanisms (e.g., for the development of demand response
algorithms). We will categorize such smart grid simulators
in two types, which we dub tools, resp. environments. A Fig. 5. Example of a co-simulation approach [24].
smart grid simulation tool is defined as providing a combined
simulation of the power grid and communication network for
a specific use case, i.e., the simulation tool is designed for that distribution networks.We include GridLAB-D in the smart grid
specific use case and others are not supported. As such, these simulator overview instead of the power system simulator
tools are used to provide answers to very specific research overview because it focuses on smart grid technologies and
questions, and are not extensible. On the other hand, smart aims to incorporate simulation of the communication network.
grid simulation environments do not target a specific use case,
1) GridLAB-D: GridLAB-D can be considered as a power
but their design supports a wide range of use cases. As such,
distribution system simulation and analysis tool [92] targeted
these environments are used to provide answers to a broad
at the smart grid. It allows the simultaneous simulation of
range of research questions, and are much more extensible.
power flow, end use loads, and market functions and in-
teractions. The software consists of a system core that can
A. Specialized smart grid simulation tools determine the simultaneous state of millions of independent
A smart grid co-simulation tool to study the impact of devices (each can be described by multiple differential or
delays in the communication network on the performance of difference equations) resulting in a detailed and accurate
the power grid is presented in [24]. A wireless communication system model. GridLAB-D is designed as a modular system:
network is simulated. A control strategy uses the wireless the system core can load additional modules that add specific
network to activate distributed storage units to compensate for functions and models to the simulation environment. Modules
temporary loss of power from a photo voltaic (PV) array, a can be developed and distributed independently. Basic features
phenomenon called “cloud transient” or “solar ramping”). The provided by these modules include power flow calculations
tool is used to determine if the distributed storage units can and device control, end use loads and controls, data collection,
be dispatched quickly enough in case such a cloud transient etc. Additional, more advanced features, such as consumer
occurs. A model of an actual distribution feeder is used to behavior models (e.g., different types of demand profiles, price
which small-scale storage batteries and a large scale PV array response, contract choice), energy operations (e.g., distribution
are connected. The wireless communication system is based on automation, load-shedding programs, emergency operations),
IEEE 802.11 (Wi-Fi). OpenDSS is employed to simulate the and business operations (e.g., retail rate, billing, market-
distribution system and the ns-2 network simulator is used to based incentive programs) are also provided or under develop-
simulate the wireless communication network. Figure 5 illus- ment [93]. Although the original focus of GridLAB-D was on
trates the sequential co-simulation approach that is employed. the distribution system, research into the transmission system
OpenDSS outputs data regarding the time of the PV ramp is also supported (e.g., the power flow module consists of
event, the geographical coordinates of the storage nodes, and both a distribution module and a transmission module [93]) as
the power output of the storage nodes. Scripts parse this output illustrated by [94] in which the influence of distributed energy
and configure ns-2 with the storage node topology. Ns-2 then sources on the transmission grid is evaluated. Although the
simulates the arrival of the dispatching messages at the storage current version (2.3.1) of GridLAB-D does not support explicit
units. Next, the arrival times of these messages are used to modeling of the communication network, a communication
create OpenDSS scripts that are fed back to the OpenDSS network module and a co-simulation approach are mentioned
environment, which then performs a sequence of power flow in the context of the next version (3.0): i.e., a communications
solutions. Note that this implies careful synchronization, as module will allow users to simulate latency and dropped
discussed in Section III-B. messages [95], [96]. The addition of such a module will enable
users to determine the impact communications systems have
on the operations of smart grid technologies. GridLAB-D is
B. Smart grid distribution system is also reported to be used as a basis for other smart grid sim-
In this section we discuss (i) the power distribution system ulation frameworks [97], [98] (although some raise concerns
simulation and analysis tool GridLAB-D, and (ii) a hardware- on the limited flexibility of composing GridLAB-D with other
in-the-loop test platform for real-time state estimation in modules [20]). An electricity market simulator and GridLAB-
14 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

D distribution system simulator are combined to simulate Modeling thereof started with tools such as SEPIA [12] to
integrated retail and wholesale power system operation in [97]. which additional control strategies would be added. Hence,
In [98] the authors show that demand response resources our reasoning for including SEPIA in this discussion of smart
can be used to maintain a flat and stable voltage profile grid simulators.
over the feeder. For this, the authors extended GridLAB-D 1) SEPIA: Simulator for Electric Power Industry Agents
with a demand response controller, and adapted the existing (SEPIA) [26] is an agent-based simulation approach to mod-
volt/var controller is adapted to make use of the added demand eling and simulation of physical and business operations in
response controller. Note that no communication network is an electric power system. SEPIA is aimed to be a proof-of-
simulated in [97], [98]. concept to illustrate an agent-based simulation approach for
2) Hardware-in-the-loop test platform: A hardware-in-the- the power industry. Possible applications targeted by SEPIA
loop [99] test platform for real-time state estimation of relate to the integration of physical and business operations
active distribution networks using phasor measurement units in a power system. A power system structure can be defined
is presented. Active distribution networks refer to electrical by components that represent generators, loads, and business
grids of which the resources are controlled by an energy entities. These components are interconnected by links, repre-
management systems (EMS) to perform optimal voltage con- senting power grid links, ownership, or money flows. Basic AC
trol, fault detection and management, etc. Such functions are and DC power flow simulations are supported. SEPIA consists
deployed in time frames that vary between a few hundreds of three main components: (i) a graphical user interface to
of milliseconds (fault management) to few tens of seconds. design, monitor and steer simulations, (ii) domain specific
As such, they require real-time information about the network agents, and (iii) a simulation engine. Domain specific agents
state. For this purpose, real-time state estimators (RTSE) that consist of traditional power system agents (e.g., generators,
use PMU measurements are being developed. However, it loads, transmission lines) and ancillary agents (e.g., markets,
is difficult to assess the accuracy of such RTSE in a real weather and speculators). Agents can transmit messages to
operational grid, as the true network state is unknown. Real- each other. Each message is sent with an associated deliv-
time simulators overcome this problem by enabling researchers ery time, which enables modeling of communication delay.
to reproduce realistic power network conditions in a controlled The simulation engine has three major functions: (i) keeping
environment. track of simulated time, (ii) managing all communication
The authors use the eMEGASim PowerGrid Real-Time Dig- between agents, and (iii) enforcing constraints set by the model
ital Simulator from Opal-RT to generate three-phase voltage topology. SEPIA supports studying agent learning in a power
and current analog signals of the monitored network buses, system by including a learning module that is based on the Q-
which are captured by a number of PMUs, which encapsulate learning algorithm (for agents to learn actions to take based
the processed signals according to IEEE Std. C37.118.2- on their observations of the system state). An example use
2011 [100] and send them over a real communication network case considers generator agents that learn how to take price
to a workstation running a Phasor Data Concentrator (PDC) decisions in electricity markets.
that processes and stores the information. The RTSE, also 2) MASGriP: Similarly to Sepia, the authors of [101] pro-
running on the workstation, uses the information to estimate pose a multi-agent based smart grid environment, but explicitly
the network state in real-time. focuses on smart grid use cases e.g., in the context of residen-
The real-time digital simulator accurately simulates the tial demand response. The simulation environment consists
electromagnetic transients required by power grid and fast of two parts that are integrated in one environment: (i) the
power electronic and converters systems. The true network multi-agent smart grid simulation platform (MASGriP), and
state is known because it is recorded by the real-time simulator. (ii) the multi-agent system for competitive electricity markets
Therefore, the performance of the RTSE algorithm can be (MASCEM) [102]. Thus, MASGriP considers the technical
assessed. Also, because a real communication network is used, aspects, whereas MASCEM considers the economical aspects
the impact thereof (e.g., latency and/or data errors and loss) of the smart grid, as discussed in more detail below.
can be evaluated. MASGriP models the distribution network and the involved
players. Power system entities such as consumers (residential,
commercial, industrial) and (distributed) generators are mod-
C. Electricity Markets eled as agents. Each agent represents a physical entity in the
In this section we discuss smart grid simulators that focus smart grid and includes information regarding the electrical
on simulation of electricity markets in smart grids. Although properties, location, etc. Additionally, demand response (DR)
these simulators do not explicitly model the communication functions, micro-generation units, and/or electric vehicles can
network, we include them because of they incorporate specific be assigned to these consumer types. These consumer agents
smart grid technologies (e.g., VPP). Also, agent based simula- establish contracts with aggregator agents: Virtual Power
tors for electricity markets such as SEPIA could be seen as the Players (VPP) or Curtailment Service Providers (CSP). Since
predecessors of the smart grid simulators of today. Agent based individual consumers have insufficient flexibility required by
approaches were gaining attention as a concept for self-healing for example DR programs, a CSP aggregates the demand
distributed control of the power grid. Clearly, concepts such response participation from small and medium consumers.
as self-healing, distributed control, and agent based system CSP tasks include: identifying curtailable loads, enrolling
are currently still active research domains in the smart grid. customers, manage curtailment events, and calculate payments
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 15

An example use case discussed is a communication-based

backup distance relay protection scheme. The present distance
relay protection framework is extended with an underlying
network infrastructure. Distance relays can communicate with
each other through their software agents thereby forming a
coordinated system protection scheme. The objective of the
scheme is to have faster backup relay protection and additional
Fig. 6. The GECO Architecture. Power system is simulated by PSLF and robustness to prevent tripping. Depending on the type of
state information and control commands are exchanged between PSLF and ns-
2 using a bidirectional interface (indicated by Sync). Control models (PMU, communication, two related protection schemes are discussed:
intelligent agents, etc.) are implemented in ns-2. supervisory (master-slave) and ad-hoc (peer-to-peer). Both
schemes achieve faster backup relay protection than traditional
non-communication based schemes, and also false-tripping
and penalties for participators. A VPP manages energy re- (i.e., due to faulty measurements) is avoided.
sources (DG, DR, SS, EV) and participates in the energy 2) ORNL Power System Simulator: Another example, based
negotiation process (DR contracts, markets, etc.). Hence, a on a co-simulation approach using the ns-2 and A Discrete
CSP is responsible for the technical management of energy EVent system Simulator (adevs) simulation tools, is presented
resources, whereas a VPP is responsible for the economical in [27], and in [5] the authors present a similar approach
activities associated with these resources. using OMNeT++ instead of ns-2. In [27], the authors discuss
MASCEM is a modeling and simulation tool to study com- in detail the problem of integrating the discrete event nature
plex and restructured electricity markets. Following agents are of communication systems and the continuous time models
defined: market operator, system operator, market facilitator, of power systems. An approach based on Discrete Event
buyer agents, seller agents, VPP agents, and VPP facilitators. System Specification (DEVS) is proposed to ensure formally
Although the focus of MASCEM is on the economical aspects that simulation correctness is preserved, enabling an integrated
(i.e., electricity markets), technical constraints influence the simulation of both domains. DEVS is a formalism to model
operation of electricity markets (e.g., supply and demand must and analyze general discrete event systems. The Toolkit for
be balanced). Therefore, the system operator agent ensures that HYbrid Modeling of Electric power systems (THYME) is built
all constraints are met in the system and is therefore connected on adevs and provides power system models (loads, transmis-
to a power system simulator to perform power flow analysis. sion lines, generators, etc.), a power flow model, and a limited
model for electro-mechanical transients [5]. A wide area load
control use case demonstrates the simulation environment.
D. Wide-Area Monitoring, Protection and Control
Example results link the performance of the communication
Now we discuss three approaches that target use cases re- network to the operation of the power system: e.g., network
lated to wide-area monitoring, protection and control: (i) Two flows affect load shed order and available bandwidth and
co-simulation approaches (GECO [3] and ORNL PSS [27]), network latency affects the control behavior.
(ii) a federated co-simulation approach (EPOCHS), and (iii) A 3) EPOCHS: The electric power and communication syn-
real-time co-simulation approach (GridSim). chronizing simulator (EPOCHS) [14], [104] is a platform for
1) GECO: A global event-driven co-simulation framework agent-based electric power and communication simulation.
for interconnected power systems and communication net- The main use cases supported by the EPOCHS simulation
works (GECO) is proposed [3], [70]. It is based on the PSLF framework are related to wide area monitoring, protection and
(steady state and dynamic power system simulations) and ns- control. Example use cases are: (i) evaluation of the benefits
2 (communication network) simulation environments. GECO and drawbacks of using communication in an agent-based
has been used to evaluate wide area monitoring, protection special protection system, (ii) a backup protection system,
and control schemes [3], [103]. (iii) monitoring of power system to prevent blackouts caused
The GECO architecture is illustrated in Fig. 6. A subcom- by voltage collapse. Instead of designing and building a
ponent in ns-2 is responsible for managing the co-simulation. new combined simulation environment, multiple specialized
It implements a global event scheduler designed as the global simulation environments (PSCAD/EMTDC, PSLF, ns-2) are
time reference and coordinator. A bidirectional interface be- linked into a distributed environment (federation).
tween ns-2 and PSLF is used to exchange information (e.g., EPOCHS is a combined simulation environment that links a
power system data, control commands), which is a tighter cou- power system simulator and communication network simulator
pling than the co-simulation approach of e.g., [24]. Network- (“federates”) in a distributed environment (a “federation”).
based power system control strategies are implemented in ns-2 Figure 7 gives an overview of the EPOCHS architecture. The
based on the Application class in ns-2: control models for digi- user of the simulation environment has the choice between
tal relays, phasor measurement units, and intelligent electronic two power system simulators, depending on the target use
devices. Agents make control decisions that are communicated case: the PSCAD/EMTDC electromagnetic transient simulator
using the simulated network and communication protocols (power system modeling), or the PSLF electromechanical
based on TCP and UDP. Synchronization of the simulators is transient simulator (transient timescales). Support for these
based on a global event driven mechanism, therefore it does different power system simulators is required due to the large
not exhibit the accuracy problems illustrated in Section III. differences in time scales between the electromagnetic and
16 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

is that it operates in real-time to ensure optimal interfacing

with actual power system elements, either hardware or soft-
ware, i.e., it enables hardware-in-the-loop (HiL) experiments.
GridSim provides a flexible simulation framework that sup-
ports power system simulation, data delivery, flexible sensor
deployments, and integration of actual power system compo-
nents, protocols, and algorithms. GridSim components can be
organized in four groups: power system simulation, substation
simulation, communication and data delivery, and control
center applications. TSTAT, a transient stability simulator, is
used for power system simulation. GridStat, is used to deliver
data between the different components in GridSim. GridStat
is a wide area data delivery framework based on a publish-
subscribe architecture. Examples of control center applications
that are included in GridSim are: (i) an oscillation monitor,
Fig. 7. The EPOCHS Architecture. Intelligent agents implement distributed and (ii) a state estimator, both built using the OpenPDC
wide area control and protection schemes. RTI routes all messages between applications set, which is an open-source software system for
simulation components and manages simulation time. collecting and processing PMU measurements.
Summarized, GridSim is a real-time simulator for the power
grid, the communication network and the control systems.
Real-time operation ensures that actual power system elements
can be integrated. Instead of using a communication network
simulator, a real communication middleware platform is used.

E. Demand-Response/Demand-Side Management
This section gives an overview of simulators that are used to
perform simulations related to demand-response or demand-
side management applications. The simulators have been se-
lected because they have distinct features. The IBCN smart
grid simulator provides an integrated environment that has
been used to evaluate DSM algorithms for electric vehicles.
The SGiC simulator for example aims to involve end-users in
their simulations, whereas GridSpice demonstrates how cloud
Fig. 8. The GridSim Architecture. The Power System component generates technology can be used to enhance smart grid simulation
PMU measurements that are encapsulated in C37.1.18 data format and
forwarded to simulated substations that use real communication middleware
scalability.
(GridStat) to transmit them to OM and SE applications. 1) IBCN Smart Grid Simulator: An integrated smart grid
simulator that considers the combined simulation of the power
system and ICT infrastructure is proposed in [75]. A case
electromechanical simulations. The communication network is study demonstrates the capabilities of the environment by
modeled in Network Simulator 2 (ns-2). The federation is man- investigating the impact of control algorithms for distributed
aged by a central component, the runtime infrastructure (RTI). generators (i.e., PV panels) has on a distribution grid, i.e., on
The RTI routes all messages between simulation components the voltage profile and load profile of a household. Another
and ensures that the simulation time is synchronized across area for which the simulator has been used extensively is
all components. The AgentHQ provides a unified view on the demand side management of electric vehicles, e.g., [105].
federation and provides a framework for implementation of The smart grid simulation environment is designed as lay-
intelligent agents, for example to implement distributed wide- ered architecture in which three layers are defined: application,
area control and protection schemes. EPOCHS uses a time middleware and support layers. The architecture is illustrated
stepped synchronization approach as discussed in section Sec- in Figure 9. The simulation environment is implemented in
tion III and as such may exhibit accuracy problems. OMNeT++ using the INET framework, and power system
Summarized, EPOCHS is a distributed simulation environ- simulator module implemented in Matlab is integrated into
ment that considers the combined simulation of the power grid the environment.
and communication network. Supported use cases are related The application layer consists of high-level applications or
to wide-area monitoring, protection and control. services, for example AMI, DSM/DR, or billing services. The
4) GridSim: simulates the power grid, the ICT infrastruc- services in the application layer make use of the middleware
ture that overlays the grid, and the control systems running on layer, which provides generic functionality that can be used
top of it, in real-time. It focuses on the design and testing of by any service. This includes a communication interface which
wide area control and protection applications using PMU and can be used to send messages between components indepen-
other high-rate time stamped data. Distinctive about GridSim dent of the underlying networking technology (e.g., ZigBee
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 17

Fig. 9. IBCN Smart Grid Simulator presented in [75]

or PLC; TCP or UDP) that is being simulated, discovery of

devices or services, etc. The goal of this middleware layer
Fig. 10. The layered SGiC architecture [17]. AS: ancillary services, VPP:
is to support a broad range of applications while reducing virtual power plant, DSM: demand side management, DR: demand response,
the effort required to develop these services to a minimum. MAS: multi-agent system.
The support layer, composed of the network and electrical
components, provides support functions for the layers above.
Communication between services is simulated by the network and send decisions back to the users. Data from network
component that provides simulation models for multiple types operators, markets, DR participation, etc. is recorded in a
of physical media and communication protocols. The simu- common database in the data access layer.
lation environment must be able to model and interact with 3) GridSpice: is cloud-based simulation package developed
(virtual) electrical devices. This is supported by the power to provide a framework to model all interactions of a smart
system component of the simulator which provides power grid, i.e., power flows, communication and market operations,
flow simulations. Basic electrical models are provided (e.g. in distribution and transmission networks. Built on top of
PV panel, battery, electric vehicle), but the user can add his GridLab-D and MATPOWER, the initial applications it targets
own models. are: renewable energy integration, home area control and smart
2) SGiC: The Smart Grids Information & Communication algorithms, electric vehicle infrastructure, distributed energy
(SGiC) [17] is web-based software for distributed decision resources, micro-grids, demand response and distribution op-
support and performance analysis. Target use cases for the eration, and utility scale storage. In [98] the authors use the
SGiC framework are power routing, power balancing, virtual GridSpice simulation platform to simulate volt/var control,
power plants, or price based control. The software enables demand response, and distribution automation in order to
active participation of researchers, engineers and customers maintain a flat and stable voltage profile over the feeder.
(residential or commercial). The latter is the unique aspect
of this simulation tools. SGiC provides a end-user interface F. Generic smart grid simulation environment
that supports social network interactions, which are considered In this section we discuss generic smart grid simulation
appropriate incentives for consumers participating in DR, environments. Such environments do not target a specific use
DSM, and virtual power plant (VPP) programs. case, but aim to be general enough to support a wide range of
The SGiC software has a three-layer architecture, illustrated use cases. The coupled simulator presented in [11] uses IEC
in Fig. 10: presentation, service, and data access layers. The 61850 to provide standards based distributed simulations. Mo-
presentation layer provides web-based services to the end user saik [106] is an automatic simulation composition framework
that assist in participating in VPP, DSM, DR and local balanc- for the smart grid.
ing programs. Customers are encouraged to share information, 1) The Coupled Simulator: A coupled power system and
in order to obtain information on interesting programs in which communication network simulator is presented in [11]. Ex-
to participate. An agent framework is used in the service layer ample use cases include the monitoring and control of large
to share information between users, network operators, and amounts of distributed energy resources in the context virtual
markets. Based on input from the users, an analysis agent power plants. Nevertheless, the simulator is described as not
(based on OpenDSS) will perform power system simulations being limited to specific use cases (e.g., time step can be
18 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

Fig. 13. The Mosaik architecture for the selection and combination of
simulation components into valid smart grid simulation systems

the data that can be exchanged using the interfaces defined

at the syntactic level. The scenario level deals with scenario
Fig. 11. The GridSpice Architecture definition and depends on a scenario meta model which is a
formal scenario description. A prototype scenario meta model
has been implemented using domain specific language (DSL).
The control level layer provides a standardized API for the
control strategies to analyze and manipulate the system at run-
time. The Mosaik prototype consists of two components (both
implemented in Python) [20]: (i) Master Control Program
Fig. 12. Architecture of the Coupled Simulator described in [11]
(MCP) (ii) simulation interface (SimAPI). The MCP manages
. the composition of the simulation scenarios and controls the
execution of the scenarios. SimAPI must be implemented by
the simulation models to integrate with Mosaik. An example
chosen in function of the phonomena under study). A message use case is presented that composes a variety of simula-
format and communication protocol based on the IEC 61850 tion models: (i) electric vehicles (Python/SimPy and JADE),
specification is used to communicate over sockets, enabling a (ii) photovoltaics (MATLAB/Simulink), (iii) residential loads
standards based distributed approach. (CSV timeseries), (iv) distribution grids (single-phase power
An overview of the architecture is given in Fig. 12. The flow analyses with Python/Pylon). Although not described,
authors define the concept of interaction points, which are a the SimAPI should allow a communication network simulator
subset of the IEC 61850 logical nodes. These nodes are the to be part of the framework. A future resource management
elements of the data model used for communication. Access component will enable simulations to be distributed over
(reading/writing) to those interaction points is provided via multiple machines thereby enhancing scalability.
JNI (Java Native Interface) interfaces. A network simulator
is placed between the smart grid applications and the power
simulator. All messages are routed through this network. A G. Summary
GUI enables the user to view the topology and simulation Figure 14 displays a classification of smart grid simulators
results of the simulated smart grid. Real-time simulation is according to their modeling capabilities in terms of com-
supported enabling real-time testing of hardware. munication network and power system. The communication
2) Mosaik: is a modular smart grid simulation framework network model level of detail is divided in three parts: (i) no
supporting automatic composition of existing, heterogeneous model, (ii) black box communication network model, (iii) de-
simulation models for the evaluation of control strategies for tailed communication network model.
heterogeneous DER and loads [106]. As such, the framework Figure 15(a) illustrates the cases where no communication
aims to provide support for scenario specification, simulation network simulation model is implemented by the smart grid
composition and scenario result analysis. simulator. Information is exchanged without modeling mes-
Mosaik adopts a layered approach to the simulation com- sage sizes, bandwidth, delay, errors, congestion, protocols, etc.
position problem, which deals with the selection and combi- In other words, an ideal network with infinite bandwidth, zero
nation of simulation components into valid simulation sys- delay, no errors, etc. is modeled.
tems, according to specific user requirements. The layered Figure 15(b) illustrates a black-box communication network
architecture is illustrated in Figure 13. The syntactic level model which provides a simplified and abstract model of the
defines the interactions between the simulation models: i.e., simulated communication network. The example black-box
to integrate a simulation model in Mosaik, the modeler has to communication network is modeled using two parameters: the
provide an implementation of a predefined interface (SimAPI delay and errors. For simplicity, we assume a fixed delay,
– XML/RPC API) that enables Mosaik to progress time of independent of the source, destination, message size, etc. In
the simulation model and to get and set model data in a such a scenario, a source that wants to transmit a message
uniform way. The semantic level uses a reference data flow to a destination, forwards the message to the black-box
model to add a semantic description (i.e., data type, units) of communication network model, which delivers the message
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 19

(a) No communication network simulation model

(b) Black-box communication network simulation model

Fig. 14. Classification of smart grid simulators according to supported power

grid simulation type and communication network model.

to the destination after the specified delay. An example of (c) Detailed communication network simulation model
a possible “error” model could be a given probability that a Fig. 15. Level of detail of the communication network simulation model:
message is lost in the network, and as a result is not received (a) No communication network simulation model. (b) Black-box: high level
at the destination. Note that other parameters (e.g., bandwidth, abstract simulation model. (c) Detailed communication network simulation
model.
congestion, message sizes) could also be included in the black-
box model.
Figure 15(c) illustrates a detailed communication network grid simulators are built on. Next, we will discuss the use
model which provides a realistic model of the simulated of standards, communication protocols, data formats, etc. in
communication network. The example network consists of smart grid simulators. Finally, we briefly indicate the role of
a source and destination host connected to a switch, which on multi-agent based systems in smart grid simulators.
are connected to the core network that consists of multiple
interconnected routers. Each communication link may be con-
A. Smart grid simulator architectures
figured with specific bandwidth, delay, error, etc. parameters.
Source and destination hosts contain models for the applica- In this section we give an overview of the different smart
tion, transport, network, link and the physical layers of the grid simulator architectures, for which we will indicate the re-
network. Switches contain models for the link and physical lationship between the four high level functional components:
layers of the network. Routers contain models for the network, Power system models the power grid.
link and physical layers of the network. Network models the communication network.
Summarized, a black-box model does not explicitly model Control models the smart grid applications (WAMS,
the network topology, links, protocols, background traffic, etc., DSM/DR, AMI, etc.).
whereas a detailed communication network model provides Sync synchronizes time, data and interactions be-
support for this tween the different simulator constituents.
The power system model level of detail is divided in As discussed in Section III, in an integrated simulation
two levels: (i) steady state, (ii) transient dynamics. Summa- architecture (see Fig. 1(b)) a single simulation environment
rized, steady state simulations analyze the system in a stable combines simulation of the power system, communication
equilibrium state, and focus lies on checking whether the network and control. Synchronization between the various
power system variables are within proper boundaries (e.g., components in this approach is straightforward, since there is
validation of voltage limits). Transient dynamics simulations only one core simulation engine keeping track of (simulated)
study transitions between equilibrium points due to a major time. This is the approach taken in, e.g., [75], [107].
changes in the power grid configuration, e.g., disturbances. We In a co-simulation approach (recall Fig. 1(a)), multiple
refer to Section IV for more information about these power specialized simulators are used, thus requiring synchronization
system simulation types. between them. Therefore, in practice typically one simulator
is selected as a master simulator for the synchronization logic,
VII. D ISCUSSION which usually (although not strictly required) is also the one
Above, we presented a survey of power system, commu- where control logic is implemented: this amounts to a master-
nication network and smart grid simulators. In this section, slave configuration, as illustrated in Fig. 16(a). Control and
we first synthesize the architectural schemes these smart synchronization thus are possibly limited by the capabilities
20 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

of the master simulator. An example of this approach is [22],

combining ns-2 (master) and Modelica, which identified pos-
sible drawbacks: (i) the master controls the slave model,
therefore, sending data from slave to master is not possible
(i.e., slave cannot push messages in response to internal events
to master) (ii) no parallelism is exploited, both run each in
turn. Also, such an architecture does not naturally lend itself
to distributed or federated simulation. Other examples that
use this approach are GECO [3], the ORNL Power System
Simulator [5], [27], VPNET [28].

Fig. 17. Layered simulation architecture. Sync A is responsible for synchro-

nization between Control and Network components whereas Sync B provides
synchronization between the Power System and Network components. The dif-
ference in naming indicates that completely different approaches/technologies
may be used.

is put on simulators that consider the combined simulation

(a) Master-slave of the power grid and ICT infrastructure. However, certain
examples do not consider both components, but are still
included due to the specific smart grid applications they target.
For each simulator, we indicate the main use case that is being
targeted, the power and network communication components,
and lastly if the simulator can be used in a distributed setting.
The support for distributed simulation can be beneficial the
increase the scalability of the simulator, and enables easier
integration with other simulators (e.g., based on HLA, see
below).

B. Standards and smart grid simulation

(b) Dedicated synchronization and control
Federation is identified as a common mechanism to co-
Fig. 16. Two fundamentally different approaches to co-simulation. simulation in this survey and in [82]. The High Level Architec-
ture (HLA) is an open standard developed by the Simulation
Figure 16(b) illustrates a co-simulation in architecture in Interoperability Standards Organization and published in IEEE
which a single dedicated component is responsible for syn- Standard 1516. It is a technology for developing distributed
chronizing and connecting all the different components. Not simulation and describes the components of HLA, their inter-
only does it provide synchronization between the multiple faces and properties. Several smart grid simulators use this
simulators, it also offers a unified interface for the control technology or a similar approach to perform a combined
logic. This approach also lends itself to distributed or federated simulation of the power system and ICT infrastructure [14],
simulation: dedicated hosts could be used for each individual [15], [83]. A federation consists of a number of simulators
host. Examples that use this approach are EPOCHS [14] (federates) that are connected to a service bus called the
and GridSpice [18], Mosaik, and the HLA-based simulator Runtime Infrastructure (RTI). Figure 18 gives an overview of
proposed in [15]. a topology of a HLA federation. The RTI provides informa-
Figure 17 illustrates a layered approach using different tion, synchronization, and coordination services. Information
synchronization layers between each simulator. An example exchange occurs according to a publish/subscribe paradigm.
that uses this approach is [11]. Also, we could consider the Synchronization services handle time, synchronization points,
co-simulation approach presented in [24] to be of this type. snapshots, etc. Coordination services are used to manage
Although the SGiC does not explicitly model the communi- the execution of the federation and the different federates.
cation network, it also uses a layered approach. Mapping the A Federation Agreement is a document that describes how
SGiC architecture [17] to the functional blocks defined in this federates are exchanging services. It consists of Federation
section (power system, network, control, sync), we could say Object Models (FOM) that contain a description of the data
that network and synchronization layers have been merged in exchange in the federation (e.g., objects, interactions). Main
the service layer. advantages of the HLA are standardized interface specifica-
Table III provides an overview of smart grid simulators tions and documentation. However, concerns are also raised
using an integrated or co-simulation approach. It can be used regarding: (i) added complexity of developing a federated
to identify simulators based on the targeted use case, illustrate simulation, (ii) the requirement to modify existing simulators
different co-simulators that have been used, etc. Emphasis to make them HLA conform.
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 21

Table III
OVERVIEW OF SMART GRID SIMULATORS .

Reference Use Case Power Communication Distributed

GridLab-D [92] Distribution system - - No
GridSim [18] Wide Area Monitoring and Control TSTAT GridStat Yes
ORNL [5], [27] Wide Area Monitoring and Control adevs ns-2, OMNeT++ No
EPOCHS [14] Wide Area Monitoring and Control PSCAD/EMTDC, PSLF ns-2 Yes
GECO [3] Wide Area Monitoring and Control PSLF ns-2 No
Zhu et al. [87] Wide Area Monitoring and Control Matlab/Simulink OPNET No
Georg et al. [15] Wide Area Monitoring and Control DigSilent OPNET Yes
Lugaric et al. [21] Wide Area Monitoring and Control PowerWorld Anylogic No
Mets et al. [75] DSM/DR Matlab OMNeT++ No
GridSpice [98] DSM/DR GridLab-D, MATPOWER - Yes
Godfrey et al. [24] DSM/DR OpenDSS ns-2 No
SEPIA [26] Electricity Markets MP 2 - Yes
MASGrip [101] Electricity Markets - - No
Bergmann et al. [11] Virtual Power Plants PSS NETOMAC ns-2 Yes
Davis et al. [108] SCADA Security PowerWorld RINSE ?
Mallouhi et al. [109] SCADA Security PowerWorld OPNET ?
Liberatore et al. [22] Networked control Modelica ns-2 No
VPNET [28] Networked control VTB OPNET No
Xiaoyang et al. [88] Networked control - OPNET No

are acquired or developed over time, as well as the exchange of

simulation models. It also adds an extra level of realism to the
simulation models. Another advantage is that users can easily
select and combine components according to their specific
requirements, reducing cost, time and risk.

C. Multi-agent based systems

Agents are a natural way to extend the power system
Fig. 18. Topology of a HLA Federation without drastic changes in the architecture of the power
system [14]. Main benefits that are associated with agent based
approaches are their (i) autonomous nature, (ii) ability to share
The smart grid simulators presented in [11], [15] make information, (iii) ability to coordinate actions. Hence, multi-
extensive use of smart grid standards such as IEC 61970 agent based systems are being used in a variety of ways
Common Information Model (CIM)/Energy Management, IEC for smart grid applications [111]. For example, protection
61968 Common Information Model (CIM)/Distribution Man- schemes, demand side algorithms, etc. are being implemented
agement and IEC 61850 Power Utility Automation. In [15], the using (market based) multi-agent systems, in which the agents
power system topology is provided as input to the simulator contain the intelligence required to take appropriate actions.
using CIM, which defines a description language used for As such, the (simulated) multi-agent architecture and the
the power system topology. Ontology matching is used to intelligence implemented in the agents could eventually be
convert the CIM topology to the IEC 61850 based model implemented in the field and thus is not only for simulation
description used internally. Support for both technologies purposes. Another example of the application of multi-agent
using ontology matching approaches is considered beneficial systems is the increased use of agents that is observed in
considering the ongoing CIM and IEC 61850 harmonization. devices deployed in the field such as Intelligent Electronic
Similarly, Mosaik [20] supports CIM to define the topology Devices (IED) [14].
of the power system. A message format and communication This survey on smart grid simulation has pointed out that
protocol based on the IEC 61850 specification is used to also in the context of simulation, agent based approaches
communicate over sockets between the different simulator play a significant role: i.e., agents are used as a model for
components [11]. GridSim [18] uses a communication protocol simulator components, which would not necessarily corre-
defined by IEEE C37.118 when simulating the exchange of spond to an actual components in the real world. Agent
PMU data. For static power flow analysis, a CIM compliant based approaches are typically used in simulators that consider
tool chain for Python has been identified in [110], comprising: electricity markets such as SEPIA and MASGriP. Examples
(i) PyCIM to import grid topology as CIM XML/RDF file of other smart grid simulators that use agent models include
(ii) CIM2BusBranch to convert CIM node breaker topology GECO, EPOCHS and SGiC. The ILIas framework presented
to the bus branch topology, and (iii) PyPOWER to perform in [82] focuses on integration of simulation and multi-agent
load flow analysis. based management systems. For the requirements analysis
Use of standards based approaches (HLA, IEC 61850, CIM, for Mosaik, additional emphasis was put on supporting agent
etc.) facilitates the interoperability of different simulators that based control strategies [106]. In [112] the authors describe a
22 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

simulator based on software agents that simulates the dynamic applications. However, support for simulation of PLC
behavior of a smart city: heterogeneous devices that consume networks in popular network simulators is only limited
and/or produce energy, and that are able to act autonomously and not available by default.
and collaborate. Agents are also considered to model the • Smart grid simulators that offer a combined simulation
human factor within simulations [21]. or focus on applications that characterize smart grids are
found for use cases related to active distribution systems,
VIII. C ONCLUSION electricity markets, wide-area monitoring, protection and
control (WAMPAC), and demand-response/demand side
Smart grid technology typically results in an increased management.
complexity of the power grid, and implies uncertainty (to • Generic smart grid simulation tools are being developed
be dealt with by, e.g., stochastic control models). To assess that support a wide range of use cases instead of focusing
the performance of possible solutions, simulation tools offer on one specific area. However, most simulators focus on
a cost effective approach. A comprehensive overview of the one specific area.
various tools applicable in smart grid research, as well as • When power network (resp. communication) details can
their main characteristics, shows they fall into three groups: be highly abstracted, an integrated simulator taking a
(i) power system, (ii) communication network, and (iii) smart detailed power (resp. communication) simulator as a base
grid simulators. Power simulation tools broadly are either seems appropriate.
targeted at steady state analysis (typically power flow studies), • When a detailed simulation of both domains can be most
or at transient dynamics simulations (typically upon distur- efficiently (esp. in terms of development effort) realized
bances or sudden system changes). They typically adopt a using a co-simulation approach that reuses existing tools.
continuous time model, studying the system state at fixed, • However, supporting combined simulation remains chal-
equidistant points in time. Communication network simulators lenging because of the need to manage and synchronize
on the other hand typically adopt a discrete event simulation actions and state (especially time) of the components.
approach, where time intervals between successive events (i.e., • Federated smart grid simulators are a promising to
system changes) can greatly vary. Thus, combining them both achieve large-scale and detailed smart grid simulations:
into real smart grid simulators requires careful synchronization distributed simulation is supported and other co-simulator
when a so-called co-simulation approach is followed, where components could be added more easily (e.g., transporta-
models from both domains in different tools are combined. tion, weather). Use of standards (e.g., HLA, IEC 61850,
More integrated solutions have a tighter coupling between CIM) may play an important role in this.
the two domain models, avoiding more tedious model syn-
chronization interactions. In terms of use cases, we found
two major types of studies: either on wide-area monitoring,
R EFERENCES
protection and control (WAMPAC), or on demand-response
(DR). The latter also imply extensive models studying market-
[1] H. Farhangi, “The path of the smart grid,” IEEE Power and Energy
based control, where typically multi-agent system (MAS) Magazine, vol. 8, no. 1, pp. 18–28, 2010.
approaches are adopted. [2] V. Gungor, D. Sahin, T. Kocak, S. Ergut, C. Buccella, C. Cecati, and
Our survey details current state-of-the-art grid simulation G. Hancke, “A survey on smart grid potential applications and commu-
nication requirements,” Industrial Informatics, IEEE Transactions on,
approaches, in terms of their use cases, architecture and vol. 9, no. 1, pp. 28–42, 2013.
example studies. We believe this synthesis thus will assist [3] H. Lin, S. Veda, S. Shukla, L. Mili, and J. Thorp, “GECO: Global
(i) smart grid researchers looking for tools that target a certain event-driven co-simulation framework for interconnected power system
and communication networks,” IEEE Trans. on Smart Grid, vol. 3,
use case, as well as (ii) smart grid simulator developers that no. 3, pp. 1444–1456, Sep. 2012.
wish to learn more about simulator paradigms, architectures, [4] K. C. Budka, J. G. Deshpande, T. L. Doumi, M. Madden, and T. Mew,
standards, etc. To conclude, lessons learned from the current “Communication network architecture and design principles for smart
grids,” Bell Labs Tech. J., vol. 15, no. 2, pp. 205–227, Sep. 2010.
state of the art seem to be:
[5] J. Nutaro, “Designing power system simulators for the smart grid:
• Power system simulation is supported by a wide variety Combining controls, communications, and electro-mechanical dynam-
of tools that can be classified in steady state and transient ics,” in Proc. IEEE Power and Energy Society General Meeting 2011
(PES ’11), 2011, pp. 1–5.
dynamic simulators according to the phenomenon under [6] T. Papadopoulos, C. Kaloudas, A. Chrysochos, and G. Papagiannis,
investigation. “Application of narrowband power-line communication in medium-
• For well-defined, specific use cases, dedicated simulation voltage smart distribution grids,” Power Delivery, IEEE Transactions
on, vol. 28, no. 2, pp. 981–988, 2013.
tools exist in both power and communications domains, [7] “Prime (powerline intelligent metering evolution) alliance,”
but for cross-domain issues, combined simulations are http://www.prime-alliance.org/.
required. [8] “G3-plc alliance,” http://www.g3-plc.com/.
• Combined simulation of power system and ICT infras-
[9] W. Kempton and J. Tomic, “Vehicle-to-grid power fundamentals:
Calculating capacity and net revenue,” Journal of Power Sources,
tructure can be achieved using a co-simulation or inte- vol. 144, no. 1, pp. 268 – 279, 2005. [Online]. Available:
grated approach. http://www.sciencedirect.com/science/article/pii/S0378775305000352
• Power line communication (PLC) technologies transform [10] ——, “Vehicle-to-grid power implementation: From stabilizing the
grid to supporting large-scale renewable energy,” Journal of Power
the power grid into a data communication network, and Sources, vol. 144, no. 1, pp. 280 – 294, 2005. [Online]. Available:
are being considered for a wide range of smart grid http://www.sciencedirect.com/science/article/pii/S0378775305000212
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 23

[11] J. Bergmann, C. Glomb, J. Gö andtz, J. Heuer, R. Kuntschke, and [31] D. Kim, Y. Yi, and S. Ha, “Trace-driven hw/sw cosimulation using
M. Winter, “Scalability of smart grid protocols: Protocols and their sim- virtual synchronization technique,” in Design Automation Conference,
ulative evaluation for massively distributed DERs,” in Proc. 1st IEEE 2005. Proceedings. 42nd, 2005, pp. 345–348.
Int. Conf. on Smart Grid Communications 2010 (SmartGridComm ’10), [32] C. Siaterlis, A. Garcia, and B. Genge, “On the use of Emulab testbeds
oct. 2010, pp. 131–136. for scientifically rigorous experiments,” IEEE Communications Surveys
[12] S. Amin and B. Wollenberg, “Toward a smart grid: power delivery for Tutorials, vol. PP, pp. 1–14, 2012.
the 21st century,” IEEE Power and Energy Magazine, vol. 3, no. 5, pp. [33] H. Gjermundrød, D. Bakken, C. Hauser, and A. Bose, “GridStat: A
34–41, 2005. flexible QoS-managed data dissemination framework for the power
[13] D. Bakken, A. Bose, C. Hauser, D. Whitehead, and G. Zweigle, “Smart grids,” IEEE Trans. Power Deliv., vol. 24, no. 1, pp. 136–143, Jan.
generation and transmission with coherent, real-time data,” Proceedings 2009.
of the IEEE, vol. 99, no. 6, pp. 928–951, 2011. [34] J. Blanger, P. Venne, and J.-N. Paquin, “The what, where and why
[14] K. Hopkinson, X. Wang, R. Giovanini, J. Thorp, K. Birman, and of real-time simulation,” in IEEE PES General Meeting, Minneapolis,
D. Coury, “EPOCHS: a platform for agent-based electric power and MN, USA, 25–29 July 2010.
communication simulation built from commercial off-the-shelf compo- [35] S. K.S.Pandya, “A survey of optimal power flow methods,” Journal of
nents,” IEEE Trans. on Power Systems, vol. 21, no. 2, pp. 548–558, Theoretical and Applied Information Technology, vol. 4, pp. 450–458,
May 2006. 2008.
[15] H. Georg, C. Wietfeld, S. C. Müller, and C. Rehtanz, “A HLA [36] M. Fazeli, G. Asher, C. Klumpner, L. Yao, and M. Bazargan, “Novel in-
based simulator architecture for co-simulating ICT based power system tegration of wind generator-energy storage systems within microgrids,”
control and protection systems,” in Proc. IEEE SmartGridComm 2012 IEEE Trans. on Smart Grid, vol. 3, no. 2, pp. 728–737, 2012.
Symposium - Performance Analysis and Simulation, 2012. [37] C. Luo, J. Yang, Y. Sun, M. Cai, J. Wu, and L. Xu, “A network based
[16] R. Kuffel, J. Giesbrecht, T. Maguire, R. Wierckx, and P. McLaren, protection scheme of distribution system,” in IEEE Int. Conf. on Power
“Rtds-a fully digital power system simulator operating in real time,” in System Technology 2012 (POWERCON ’12), 2012, pp. 1–6.
WESCANEX 95. Communications, Power, and Computing. Conference [38] H. Mahmood and J. Jiang, “Modeling and control system design of a
Proceedings., IEEE, vol. 2, 1995, pp. 300–305 vol.2. grid connected VSC considering the effect of the interface transformer
[17] J. de Haan, P. Nguyen, W. Kling, and P. Ribeiro, “Social interaction type,” IEEE Trans. on Smart Grid, vol. 3, no. 1, pp. 122–134, 2012.
interface for performance analysis of smart grids,” in Proc. 1st Int. [39] F. Coroiu, C. Velicescu, and C. Barbulescu, “Probabilistic and deter-
Workshop Smart Grid Modeling and Simulation (SGMS 2011) at IEEE ministic load flows methods in power systems reliability estimation,”
SmartGridComm 2011, oct. 2011, pp. 79–83. in Proc. IEEE Int. Conf. on Computer as a Tool 2011 (EUROCON
[18] D. Anderson, C. Zhao, C. Hauser, V. Venkatasubramanian, D. Bakken, ’11), 2011, pp. 1–4.
and A. Bose, “A virtual smart grid - real-time simulation for smart [40] X. Chen, W. Pei, and X. Tang, “Transient stability analyses of micro-
grid control and communications design,” IEEE Power and Energy grids with multiple distributed generations,” in Proc. Int. Conf. on
Magazine, vol. 10, no. 1, pp. 49–57, jan.-feb. 2012. Power System Technology 2010 (POWERCON ’10), 2010, pp. 1–8.
[19] E. Weingärtner, H. Vom Lehn, and K. Wehrle, “A performance [41] A. Hansen, C. Jauch, P. Sørensen, F. Iov, and F. Blaabjerg, “Dynamic
comparison of recent network simulators,” in Proc. IEEE international wind turbine models in power system simulation tool DIgSILENT,”
Conference on Communications 2009, ser. ICC’09. Piscataway, Risø National Laboratory, Roskilde, Tech. Rep., 2003.
NJ, USA: IEEE Press, 2009, pp. 1287–1291. [Online]. Available: [42] F. Andren, M. Stifter, T. Strasser, and D. Burnier de Castro, “Frame-
http://dl.acm.org/citation.cfm?id=1817271.1817510 work for co-ordinated simulation of power networks and components
[20] S. Schutte, S. Scherfke, and M. Troschel, “Mosaik: A framework for in smart grids using common communication protocols,” in Proc. 37th
modular simulation of active components in smart grids,” in Proc. 1st Annual Conference on IEEE Industrial Electronics Society (IECON
Int. Workshop Smart Grid Modeling and Simulation (SGMS 2011) at ’11), 2011, pp. 2700–2705.
IEEE SmartGridComm 2011, 17 Oct. 2011, pp. 55–60. [43] T. Chant, G. M. Shafiullah, A. Oo, and B. Harvey, “Impacts of
[21] L. Lugaric, S. Krajcar, and Z. Simic, “Smart city - platform for emer- increased photovoltaic panel utilisation on utility grid operations - a
gent phenomena power system testbed simulator,” in Proc. IEEE/PES case study,” in Proc. IEEE/PES Innovative Smart Grid Technologies
Innovative Smart Grid Technologies Conference Europe 2010 (ISGT Asia 2011 (ISGT Asia ’11), 2011, pp. 1–7.
Europe ’10), oct. 2010, pp. 1–7. [44] U. Sachs, “Smart grid offering within PSS SINCAL,”
R Siemens, Tech.
[22] V. Liberatore and A. Al-Hammouri, “Smart grid communication and Rep., 2012.
co-simulation,” in Proc. IEEE Energytech 2011, may 2011, pp. 1–5. [45] A. Mohamad, N. Hashim, N. Hamzah, N. F. N. Ismail, and M. Latip,
[23] F. Ponci, A. Monti, and A. Benigni, “Simulation for the design of smart “Transient stability analysis on sarawak’s grid using power system
grid controls,” in Proc. 1st Int. Workshop Smart Grid Modeling and simulator for engineering (PSS/E),” in Proc. IEEE Symposium on
Simulation (SGMS 2011) at IEEE SmartGridComm 2011, oct. 2011, Industrial Electronics and Applications 2011 (ISIEA ’11), 2011, pp.
pp. 73–78. 521–526.
[24] T. Godfrey, S. Mullen, R. C. Dugan, C. Rodine, D. W. Griffith, and [46] A. Hernandez, P. Eguia, E. Torres, and M. Rodriguez, “Dynamic
N. Golmie, “Modeling smart grid applications with co-simulation,” simulation of a SSSC for power flow control during transmission
in Proc. 1st IEEE Int. Conf. on Smart Grid Communications 2010 network contingencies,” in Proc. IEEE Trondheim PowerTech 2011,
SmartGridComm ’10), 2010. 2011, pp. 1–6.
[25] Y.-J. Kim, M. Thottan, V. Kolesnikov, and W. Lee, “A secure decen- [47] F. Napolitano, A. Borghettia, M. Paolonea, and M. Bernardib, “Voltage
tralized data-centric information infrastructure for smart grid,” IEEE transient measurements in a distribution network correlated with data
Commun. Mag., vol. 48, no. 11, pp. 58–65, Nov. 2010. from lightning location system and from sequence of events recorders,”
[26] S. Harp, S. Brignone, B. Wollenberg, and T. Samad, “SEPIA. a Electric Power Systems Research, vol. 81, pp. 237–253, 2011.
simulator for electric power industry agents,” IEEE Control Systems [48] R. Roche, S. Natarajan, A. Bhattacharyya, and S. Suryanarayanan, “A
Magazine, vol. 20, no. 4, pp. 53–69, aug 2000. framework for co-simulation of AI tools with power systems analysis
[27] J. Nutaro, P. Kuruganti, L. Miller, S. Mullen, and M. Shankar, software,” in Proc. 23rd Int.Workshop on Database and Expert Systems
“Integrated hybrid-simulation of electric power and communications Applications 2012 (DEXA ’12), 2012, pp. 350–354.
systems,” in Proc. IEEE Power Engineering Society General Meeting [49] R. Mehra, N. Bhatt, F. Kazi, and N. Singh, “Analysis of PCA based
2007 (PES ’07), june 2007, pp. 1–8. compression and denoising of smart grid data under normal and fault
[28] W. Li, A. Monti, M. Luo, and R. Dougal, “VPNET: A co-simulation conditions,” in Proc. IEEE Int. Conference on Electronics, Computing
framework for analyzing communication channel effects on power and Communication Technologies 2013 (CONECCT ’13), 2013, pp.
systems,” in Proc. IEEE Electric Ship Technologies Symposium 2011 1–6.
(ESTS ’11), april 2011, pp. 143–149. [50] Y. Tang, X. Mao, and R. Ayyanar, “Distribution system modeling using
[29] J. Juárez, C. Rodrı́guez-Morcillo, and J. A. Rodrı́guez-Mondéjar, CYMDIST for study of high penetration of distributed solar photo-
“Simulation of IEC 61850-based substations under OMNeT++,” in voltaics,” in Proc. North American Power Symposium 2012 (NAPS ’12),
Proc. of the 5th Int. ICST Conference on Simulation Tools and 2012, pp. 1–6.
Techniques. ICST (Institute for Computer Sciences, Social-Informatics [51] G. Asimakopoulou, E. Voumvoulakis, A. Dimeas, and N. Hatziar-
and Telecommunications Engineering), 2012, pp. 319–326. gyriou, “Impact of large-scale integration of intelligent meters to the
[30] R. Podmore and M. Robinson, “The role of simulators for smart grid operation of the power system of crete,” in Proc. 16th Int. Conf. on
development,” IEEE Trans. on Smart Grid, vol. 1, no. 2, pp. 205–212, Intelligent System Application to Power Systems 2011 (ISAP ’11), 2011,
Sept. pp. 1–6.
24 IEEE COMMUN. SURVEYS & TUTORIALS – SPECIAL ISSUE ON ENERGY AND SMART GRID

[52] S. Kreckelbergh and I. Vechiu, “Sizing and dynamic analyses of a [76] R. Mora, A. Lopez, D. Roman, R. Sanz, F. Lobo, F. Carmona,
micro-grid supplying a harbor industrial area,” in Proc. 16th Int. Conf. D. Mora, A. Cabello, A. Sendin, and I. Berganza, “Demand man-
on System Theory, Control and Computing 2012 (ICSTCC ’12), 2012, agement communications architecture,” in Proc. 20th Int. Conf. on
pp. 1–5. Electricity Distribution (CIRED ’09), Prague, 2009, pp. 8–11.
[53] A. Anwar and H. Pota, “Loss reduction of power distribution network [77] R. Mora, A. Lopez, D. Roman, A. Sendin, and I. Berganza, “Commu-
using optimum size and location of distributed generation,” in Proc. nications architecture of smart grids to manage the electrical demand,”
21st Australasian Universities Power Engineering Conference (AUPEC in Proc. 3rd Workshop on Power Line Communications, Udine, Italy,
’11), 2011, pp. 1–6. October 1–2 2009.
[54] M. Larsson, “ObjectStab - an educational tool for power system [78] C. Müller, S. Šubik, A. Wolff, and C. Wietfeld, “A system design
stability studies,” IEEE Trans. on Power Systems, vol. 19, no. 1, pp. framework for scalability analysis of geographic routing algorithms in
56–63, feb. 2004. large-scale mesh networks,” in Proc. 3rd Int. ICST Conf. on Simulation
[55] R. Meere, M. O’Malley, and A. Keane, “VSC-HVDC link to support Tools and Techniques (SIMUTools ’10). ICST, Brussels, Belgium,
voltage and frequency fluctuations for variable speed wind turbines for Belgium: ICST (Institute for Computer Sciences, Social-Informatics
grid connection,” in Proc. 3rd IEEE/PES Int. Conf. and Exhibition on and Telecommunications Engineering), 2010, pp. 1–7.
Innovative Smart Grid Technologies 2012 (ISGT Europe ’12), 2012, [79] C. Müller, H. Georg, and C. Wietfeld, “A modularized and
pp. 1–5. distributed simulation environment for scalability analysis of smart
[56] “Opal-rt,” http://www.opal-rt.com/. grid ICT infrastructures,” in Proc. 5th Int. ICST Conf. on Simulation
[57] S. Abourida, C. Dufour, J. Belanger, G. Murere, N. Lechevin, and Tools and Techniques (SIMUTools ’12), ser. SIMUTOOLS ’12.
B. Yu, “Real-time pc-based simulator of electric systems and drives,” ICST, Brussels, Belgium, Belgium: ICST (Institute for Computer
in Applied Power Electronics Conference and Exposition, 2002. APEC Sciences, Social-Informatics and Telecommunications Engineering),
2002. Seventeenth Annual IEEE, vol. 1, 2002, pp. 433–438 vol.1. 2012, pp. 327–330. [Online]. Available: http://dl.acm.org/citation.cfm?
[58] Mathworks, “Simpowersystems,” available at id=2263019.2263070
http://www.mathworks.com/products/simpower/. [80] J. Tripathi, J. de Oliveira, and J. Vasseur, “Applicability study of RPL
[59] “Rtds technologies,” http://www.rtds.com/. with local repair in smart grid substation networks,” in Proc. First
[60] F. Milano, “An open source power system analysis toolbox,” IEEE IEEE International Conference on Smart Grid Communications 2010
Trans. on Power Systems, vol. 20, no. 3, pp. 1199–1206, aug. 2005. (SmartGridComm ’10), oct. 2010, pp. 262–267.
[61] D. Kwok, W. Cheung, and P. J. Chow, “Power system [81] H. Kellerbauer and H. Hisch, “Simulation of powerline communication
toolbox (PST),” Rensselaer Polytechnic Institute, available at with omnet++ in (static) smart grids,” in Proceedings of the
http://www.ecse.rpi.edu/pst/PST.html. 4th International ICST Conference on Simulation Tools and
[62] C. Vournas, E. Potamianakis, C. Moors, and T. Van Cutsem, “An Techniques, ser. SIMUTools ’11. ICST, Brussels, Belgium, Belgium:
educational simulation tool for power system control and stability,” ICST (Institute for Computer Sciences, Social-Informatics and
IEEE Trans. on Power Systems, vol. 19, no. 1, pp. 48–55, feb. 2004. Telecommunications Engineering), 2011, pp. 406–409. [Online].
[63] K. Schoder, A. Hasanovic, and A. Feliachi, “PAT: a power analysis Available: http://dl.acm.org/citation.cfm?id=2151054.2151126
toolbox for MATLAB/Simulink,” IEEE Trans. on Power Systems, [82] T. Konnerth, J. Chinnow, D. Grunewald, and S. Kaiser, “Integration of
vol. 18, no. 1, pp. 42–47, feb 2003. simulations and MAS for smart grid management systems,” in Proc.
[64] S. Ayasun, C. Nwankpa, and H. Kwatny, “Voltage stability toolbox 3rd Int. Workshop on Agent Technologies for Energy Systems (ATES
for power system education and research,” IEEE Trans. on Education, 2012) at 11th Int. Conf. on Autonomos Augents and Multiagent Systems
vol. 49, no. 4, pp. 432–442, nov. 2006. (AAMAS ’12), Valencia, Spain, June 5 2012.
[65] R. D. Zimmerman and D. Gan, “Matpower, documentation for version [83] J. Chinnow, K. Bsufka, A.-D. Schmidt, R. Bye, A. Camtepe, and
4.1,” Power System Engineering Research Center, Cornell University, S. Albayrak, “A simulation framework for smart meter security
2011, available at http://www.pserc.cornell.edu/matpower/. evaluation,” in Proc. IEEE Int. Conf. on Smart Measurements
[66] F. Milano and L. Vanfretti, “State of the art and future of OSS for power of Future Grids 2011, ser. SMFG ’11. Washington, DC, USA:
systems,” in Proc. IEEE Power Energy Society General Meeting 2009 IEEE Computer Society, 2011, pp. 1–9. [Online]. Available:
(PES ’09), july 2009, pp. 1–7. http://dx.doi.org/10.1109/SMFG.2011.6125758
[67] R. Amin, J. Martin, and X. Zhou, “Smart grid communication using
[84] A. Patel, J. Aparicio, N. Tas, M. Loiacono, and J. Rosca, “Assessing
next generation heterogeneous wireless networks,” in Proc. 3rd IEEE
communications technology options for smart grid applications,” in
Int. Conf. on Smart Grid Communications 2012 (SmartGridComm ’12),
Proc. IEEE Int. Conference on Smart Grid Communications 2011
2012, pp. 229–234.
(SmartGridComm ’11), 2011, pp. 126–131.
[68] C. Muller, M. Putzke, and C. Wietfeld, “Traffic engineering analysis of
smart grid services in cellular networks,” in Proc. 3rd IEEE Int. Conf. [85] J. Aparicio, A. Patel, N. Tas, M. Loiacono, and J. Rosca, “Scalable
on Smart Grid Communications 2012 (SmartGridComm ’12), 2012, smart grid simulations in OPNET Modeler,” Proc. OPNETWORK
pp. 252–257. 2011, Tech. Rep., 2011.
[69] T. R. Henderson, M. Lacage, and G. F. Riley, “Network simulations [86] T. Sidhu and Y. Yin, “Modelling and simulation for performance
with the ns-3 simulator,” in Proc. ACM Conf. of the Special Interest evaluation of IEC61850-based substation communication systems,”
Group on Data Communication 2008 (SIGCOMM ’08), August 17-22 IEEE Trans. on Power Delivery, vol. 22, no. 3, pp. 1482–1489, july
2008, p. 527. 2007.
[70] H. Lin, S. Sambamoorthy, S. Shukla, J. Thorp, and L. Mili, “Power [87] K. Zhu, M. Chenine, and L. Nordstrom, “ICT architecture impact on
system and communication network co-simulation for smart grid appli- wide area monitoring and control systems’ reliability,” IEEE Trans. on
cations,” in Proc. IEEE/PES Innovative Smart Grid Technologies 2011 Power Delivery, vol. 26, no. 4, pp. 2801–2808, Oct.
(ISGT ’11), jan. 2011, pp. 1–6. [88] X. Tong, “The co-simulation extending for wide-area communication
[71] J. Kim, D. Kim, K.-W. Lim, Y.-B. Ko, and S.-Y. Lee, “Improving the networks in power system,” in Proc. IEEE Asia-Pacific Power and
reliability of IEEE 802.11s based wireless mesh networks for smart grid Energy Engineering Conference 2010 (APPEEC ’10), March, pp. 1–4.
systems,” Journal of Communications and Networks, vol. 14, no. 6, pp. [89] R. H. Khan and J. Y. Khan, “Wide area PMU communication over a
629–639, Dec. 2012. WiMAX network in the smart grids,” in Proc. 3rd IEEE Int. Conf. on
[72] F. Aalamifar, A. Schlogl, D. Harris, and L. Lampe. Mod- Smart Grid Communications 2012 (SmartGridComm ’12), 2012, pp.
elling power line communication using network simulator-3. 187–192.
http://www.ece.ubc.ca/ faribaa/paper.pdf. [90] ——, “A heterogeneous WiMAX-WLAN network for AMI communi-
[73] ——, “Plc-ns3 module,” http://www.ece.ubc.ca/ faribaa/index files/Page417.htm. cations in the smart grids,” in Proc. 3rd IEEE Int. Conf. on Smart Grid
[74] A. Varga, “An overview of the OMNeT++ simulation environment,” Communications 2012 (SmartGridComm ’12), 2012, pp. 710–715.
in Proc. 1st Int. Conf. Simulation Tools and Techniques for Commun., [91] J. Brown and J. Y. Khan, “Performance comparison of LTE FDD and
Netw. and Systems (SIMUTools ’08), Marseille, France, 3–7 March TDD based smart grid communications networks for uplink biased
2008. traffic,” in Proc. 3rd IEEE Int. Conf. on Smart Grid Communications
[75] K. Mets, T. Verschueren, C. Develder, T. L. Vandoor, and L. Vande- 2012 (SmartGridComm ’12), 2012, pp. 276–281.
velde, “Integrated simulation of power and communication networks [92] D. Chassin, K. Schneider, and C. Gerkensmeyer, “GridLAB-D: An
for smart grid applications,” in Proc. IEEE 16th Int. Workshop on open-source power systems modeling and simulation environment,”
Computer Aided Modeling and Design of Communication Links and in Proc. IEEE/PES Transmission and Distribution Conference and
Networks 2011 (CAMAD ’11), Kyoto, Japan, 10–11 June 2011. Exposition 2008, april 2008, pp. 1–5.
METS et al.: COMBINING POWER AND COMMUNICATION NETWORK SIMULATION FOR COST-EFFECTIVE SMART GRID ANALYSIS 25

[93] K. Schneider, D. Chassin, Y. Chen, and J. Fuller, “Distribution power specification,” SG Simulations Working Group - OpenSG, Tech. Rep.,
flow for smart grid technologies,” in Proc. IEEE/PES Power Systems 2012.
Conference and Exposition 2009 (PSCE ’09), march 2009, pp. 1–7. [111] M. Pipattanasomporn, H. Feroze, and S. Rahman, “Multi-agent systems
[94] R. Guttromson, “Modeling distributed energy resource dynamics on the in a distributed smart grid: Design and implementation,” in Proc.
transmission system,” IEEE Trans. on Power Systems, vol. 17, no. 4, IEEE/PES Power Sys. Conf. and Expo. (PSCE 2009), Seattle, WA,
pp. 1148–1153, Nov. USA, Mar. 2009, pp. 1–8.
[95] GridLAB-D Wiki, “Gridlab-d communications [112] S. Karnouskos and T. de Holanda, “Simulation of a Smart Grid City
module,” http://sourceforge.net/apps/mediawiki/gridlab- with Software Agents,” in Proc. 3rd UKSim European Symposium on
d/index.php?title=Communications module, October 2013. [Online]. Computer Modeling and Simulation 2009. IEEE, 2009, pp. 424–429.
Available: http://sourceforge.net/apps/mediawiki/gridlab-d/index.php?
Kevin Mets received the M.Sc. degree in Computer
title=Communications\ module
Science from Ghent University, Ghent, Belgium, in
[96] ——, “Gridlab-d v3 applications concepts: Com- 2009. He is currently working at the research group
munications,” http://sourceforge.net/apps/mediawiki/gridlab- IBCN of the Dept. of Information Technology (IN-
d/index.php?title=V3 applications concepts, October 2013. [Online]. TEC) at Ghent University – iMinds, Ghent, Belgium,
Available: http://sourceforge.net/apps/mediawiki/gridlab-d/index.php? where he is working toward the Ph.D. degree in
title=V3\ applications\ concepts#Communications Computer Science. His research interests include
[97] D. Aliprantis, S. Penick, L. Tesfatsion, and H. Zhao, “Integrated retail smart grids, optimization, communication networks,
and wholesale power system operation with smart-grid functionality,” and demand side management algorithms for electric
in Proc. IEEE Power and Energy Society General Meeting 2010 (PES vehicles.
’10), july 2010, pp. 1–8.
[98] K. Anderson and A. Narayan, “Simulating integrated volt/var control
and distributed demand response using GridSpice,” in Proc. 1st Int.
Workshop Smart Grid Modeling and Simulation (SGMS 2011) at IEEE
SmartGridComm 2011, Brussels, Belgium, 17 Oct. 2011, pp. 84–89.
[99] M. Paolone, M. Pignati, P. Romano, S. Sarri, L. Zanni, and Juan Aparicio Ojea is currently working at Siemens
R. Cherkaoui, “A hardware-in-the-loop test platform for the real- Corporation, Corporate Technology (Princeton-NJ),
time state estimation of active distribution networks using phasor as an R&D engineer in the area of Intelligent Trans-
measurement units,” in Cigr SC6 Colloquium, Yokohama, Japan, 2013. portation Systems and Smart Grids. In parallel, he
[100] C37.118.2-2011 - IEEE Standard for Synchrophasor Data Transfer for is pursuing a Ph.D. degree at the research group
Power Systems, IEEE Standards Association Std. IBCN of the Dept. of Information Technology (IN-
[101] P. Oliveira, T. Pinto, H. Morais, and Z. Vale, “MASGriP - a multi- TEC), at Ghent University – iMinds, Belgium. In
agent smart grid simulation platform,” in Proc. IEEE Power and Energy 2010, he received a Master-level Double Degree in
Society General Meeting (PES ’12), july 2012, pp. 1–8. Information and Communication Engineering Tech-
[102] P. Oliveira, T. Pinto, H. Morais, Z. Vale, and I. Praca, “MASCEM - nologies (Lund Technical University - Sweden) and
an electricity market simulator providing coalition support for virtual Telecommunications Engineering (Madrid Polytech-
power players,” in Proc. 15th Int. Conf. on Intelligent System Applica- nic University-Spain), from the prestigious academic program T.I.M.E. (Top
tions to Power Systems 2009 (ISAP ’09), Nov. 2009, pp. 1–6. Industrial Managers for Europe). His research interests include smart grids,
[103] H. Lin, Y. Deng, S. Shukla, J. Thorp, and L. Mili, “Cyber security intelligent transportation systems, connected vehicles, computational intel-
impacts on all-PMU state estimator - a case study on co-simulation ligence, machine to machine communication, statistical signal processing,
platform GECO,” in Proc. IEEE SmartGridComm 2012 Symposium - complex networks simulations, distributed sensing and control architectures,
Wide Area Protection and Control (WAMPAC), 2012. wireless sensor networks and communication protocols and algorithms.
[104] K. Hopkinson, K. Birman, R. Giovanini, D. Coury, X. Wang, and
J. Thorp, “EPOCHS: integrated commercial off-the-shelf software for
agent-based electric power and communication simulations,” in Proc.
Winter Simulation Conference 2003, vol. 2, Dec., pp. 1158–1166.
Chris Develder is professor with the research group
[105] K. Mets, T. Verschueren, F. De Turck, and C. Develder, “Exploiting IBCN of the Dept. of Information Technology (IN-
v2g to optimize residential energy consumption with electrical vehicle TEC) at Ghent University – iMinds, Ghent, Belgium.
(dis)charging,” in Proc. 1st Int. Workshop Smart Grid Modeling and He received a M.Sc. degree in computer science
Simulation (SGMS 2011) at IEEE SmartGridComm 2011, oct. 2011, engineering (Jul. 1999) and a Ph.D in electrical
pp. 7 –12. engineering (Dec. 2003) from the same university.
[106] S. Schutte, “Composition of simulations for the analysis of smart grid He has been working in IBCN from 1999 to 2003 as
scenarios,” Energieinformatik, Tech. Rep., 2011. a research fellow of the Research Foundation – Flan-
[107] A. Awad, P. Bazan, and R. German, “Abstract-based methodology ders (FWO), on optical networks. From Jan. 2004
for modeling and simulation of smart grid components,” in Proc. to Aug. 2005 he worked at OPNET Technologies as
6th UKSim/AMSS European Symposium on Computer Modeling and senior engineer optical solutions. In Sep. 2005, he
Simulation 2012 (EMS ’12), nov. 2012, pp. 305–310. rejoined Ghent University – iMinds, as a post-doc (FWO scholarship 2006-
[108] C. M. Davis, J. Tate, H. Okhravi, C. Grier, T. Overbye, and D. Nicol, 2012). After a stay at UC Davis, CA, USA, in 2007 he became part-time (Oct.
“SCADA cyber security testbed development,” in Proc. 38th North 2007) and then full-time professor (2010) at Ghent University. His research
American Power Symposium 2006 (NAPS ’06), 2006, pp. 483–488. interests include dimensioning, modeling and optimizing optical (grid/cloud)
[109] M. Mallouhi, Y. Al-Nashif, D. Cox, T. Chadaga, and S. Hariri, “A networks and their control and management, smart grids, information retrieval,
testbed for analyzing security of SCADA control systems (TASSCS),” as well as multimedia and home network software and technologies. He
in Proc. IEEE/PES Innovative Smart Grid Technologies 2011 (ISGT regularly serves as reviewer/TPC member for international journals and con-
’11), 2011, pp. 1–7. ferences (IEEE/OSA JLT, IEEE/OSA JOCN, IEEE/ACM Trans. Networking,
[110] S. Schütte, M. Tröschel, J. Schoene, C. Develder, K. Mets, and Computer Networks, IEEE Network, IEEE JSAC; IEEE Globecom, IEEE
J. Taylor, “Smart grid simulation platform architecture & requirements ICC, IEEE SmartGridComm, ECOC, etc.).