APPEEC2019 1570580655




EPON Communication Network Topology Optimization Based on Electrical Nodes’ Importance in

Active Distribution Networks



Yingqiao He1,2, Li Ma2, Lingfeng Wang2, Junjie Tang1, and Rao Rao1,2

1The State Key Laboratory of Power Transmission Equipment & System Security and New Technology, College of Electrical Engineering᧨


Chongqing University, 400044, China

2
Department of Electrical Engineering and Computer Science, University of Wisconsin-Milwaukee, Milwaukee, WI 53211, USA

Email: l.f.wang@ieee.org



Abstract— A large amount of DGs’ (distributed generations) hand, ICT (Information and Communications Technology)

accessing to the distribution network is a characteristic of the system’s random failures may lead to adverse consequences

present and future power distribution systems. According to the such as operational condition deterioration and widespread

actual operation state of power system, distribution network blackouts. The active distribution network can be regarded as

actively manages the massive distributed generations and a typical Cyber Physical System (CPS).

adaptively adjusts the network to satisfy the requirements of After the access of a large number of distributed energy
economical and safe power supply both in the physical network

and cyber network. This paper aims to propose a feasible
and intelligent user terminals, the distribution network takes


the responsibility of not only reliable electricity supply to
planning optimization model for the communication network

loads but also effective consumption of new energy. Thus,
based on the evaluation of all electrical nodes’ importance in the

physical distribution network. This solution not only meets the
this paper proposes a new importance index for electrical

current technical requirements, but also considers the possible nodes on the physical side. Focusing on the topology of cyber
side based on the node importance of physical side, this paper

expansion of communication network in the future. Based on
provides a communication network optimization model

the existing single-link network scheme for all the
considering reliability and economy simultaneously.
communication nodes, a certain portion of the important nodes


in the distribution network is chosen to be upgraded to dual-

link connection. The optimization result can provide a more DISTRIBUTION NETWORK IN POWER SYSTEM

economical solution for distribution network operators to build A. Distribution automation system structure

the corresponding communication network.


Distribution network locates at the end of power system.
Due to its direct connection to users, the quality and ability

Index Terms—cyber-physical system; active distribution
of power supply are greatly influenced by it. This paper pays

network; EPON; node importance; topology optimization;
complex network. more attention on 10kV distribution network. For the

convenience of management and maintenance, distribution
INTRODUCTION

automation system usually adopts hierarchical structure.
The emergence of new elements in power grids such as

distributed generators, energy storage devices, and flexible The communication channel is a communication

loads that have the ability to control itself according to the network that connects the main distribution station and the

users’ demands poses new challenges to the operation and distribution substation, and transmits automation

planning of distribution networks. The distribution system information between them. In particular, the communication

will transfer from the traditional power distribution network channel acts as the nervous system of the distribution


to current power exchange system, which takes the automation system, and is responsible for the uploading and

responsibility of power production, storage, transmission and dispatching of various monitoring information and control

distribution. These new characteristics bring profound commands. It is an important basic part for realizing

influences on the pattern of power supply in the distribution distribution automation, which leads to the discussion of the

network. Under the circumstances, the communication key issues in this paper—the communication network
network planning is more complex and important, and topology in the distribution automation system will be


requires planners’ systematic considerations. Planners should elaborated in the next chapter.

Obviously, in this heterogeneous network, if the

ensure the reliability of the distribution network and
corresponding distribution communication network on the importance of the node with a high degree of importance is

attacked firstly, and this will cause the connectivity
premise of guaranteeing a secure and stable power system

considering the load growth and power planning scheme performance and reliability of network to drop sharply, and


during the planning period [1-4]. even cause the corresponding cyber network cause serious

Structurally, CPS (Cyber Physical System) can be divided problem. Therefore, it is extremely important and necessary

into two major parts: cyber subsystem and physical to optimize the network for the CPS of the distribution

subsystem, which interact with and interdepend on each other network at the planning process to identify the nodes or

[5]. In CPS, the cyber subsystem represents the information regions with high importance [10].

world including discrete computing processes, logical B. The indicator of node importance in distribution
network

communication processes and feedback control processes; the
The distribution network is a complex structure

physical subsystem represents processes, objects or
components in the physical world [6]. On the one hand, timely consisting of lines, equipment and multiple loads. In order to

extract key information of distribution network, it needs to

and accurate transmission and decision-making are the basis
for enabling a variety of active control functions; on the other be simplified, which means that a simple topology composed












1
of nodes and lines is used to describe a complex distribution
network. The distribution network could be modelled as a
graph ‫ ܩ‬ൌ ሺܸǡ ‫ܧ‬ሻ, where ܸ is a set of nodes and ‫ ܧ‬is a set of
m edges.
Each node of the distribution network may be connected
to distributed generations, energy storage devices, and loads
with different types and importance [7]. This paper not only
considers these factors in the new indicators, but also
considers the contribution of this node to the entire
distribution network. The node importance indicator is
expressed as follows:
I node (i ) = α ∗ Finput (i ) + β ∗ Fourtput (i ) + δ ∗ Δ E (1)
where, ߙ, ,  is the correction factor, and α + β + δ = 1 .
Finput (i ) = Gi / G max ; ‫ܩ‬୧ is the amount of distributed
generation connected to node i; ‫ܩ‬୫ୟ୶ is the largest
distributed generation among all the nodes;
Foutput (i ) = Li / Lmax ; ‫ܨ‬௢௨௧௣௨௧ ሺ݅ሻ represents the total amount
of load, which connected to node i; ‫ܨ‬௢௨௧௣௨௧ ሺ݅ሻ reflects the
load of node i; ‫ܮ‬୧ is the load of node i; ‫ܮ‬୫ୟ୶ is the largest load
E (Y ) − E (Y − 1)
among all the node; ΔE (i ) = , E (Y ) is
E (Y )
original global efficiency, E (Y − 1) is the global efficiency
when one node is removed, ΔE (i ) is the efficiency change
after removing node i.
2 ˄dij - d˅ vij -1
E (Y ) = ∗ ¦ i∈{G}, j∈{ D} exp˄ ∗ ˅
n ∗ (n − 1) d vij ,max
where, node i belongs to generator set {G} and node j
belongs to load set {D}, and other nodes belong to set {C}.
n is the number of nodes in power grid, ݀ҧ is the average
COMMUNICATION NETWORK OPTIMIZATION
C. Introduction of distribution communication network
and EPON technology
Due to the wide variety of distribution network points,
complex topological structure and poor operating conditions,
planning of topology has become the key and difficult
problem in the construction of distribution network
communication system [8]. Passive optical network (PON) is
an optical communication network composed of many
passive optical components using pure optical transmission
medium, and has excellent transmission characteristics of
optical fiber communication. From the perspectives of
transmission speed and transmission distance, PON has
incomparable virtues, and it has gradually become the
mainstream technology in several countries. Now many
operators have started to increase the investment in the
construction of passive optical network. EPON/GPON have
become mature [11]. EPON (Ethernet passive optical
network) adopts PON technology in physical layer, using
Ethernet protocols in link layer and PON topology structure
to realize Ethernet access. This paper uses EPON, which
reduces the cost of network construction and operation,
facilitates network construction and expansion, and provides
more users with flexible access options. EPON is a single-
fiber and bidirectional Optical access network from point to
multiple points. It consists of Optical Line Terminal (OLT)
installed in Distribution station, Optical Network Unit
(ONU) installed in Distribution Terminal and Optical
Distribution Network (ODN) in Distribution Network. ODN
can connect one OLT and multiple ONUs, which consist of
main optical cables, splitters and branch optical cables, thus
(2)
providing bidirectional transmission of optical signals. The
structure of EPON is shown in figure 2. In this paper, the
location of OLT is fixed. The planning is focusing on the
topology of ODN.
ODN
ONU1

shortest distance, ݀௜௝ is the electrical distance between node

i and node j, this paper calculates the electrical distance using
the resistance between two nodes and the resistance per unit
length. ܸ௜௝ is the amount of electricity exchange between the
power generation node i and the node j. This process of
calculating ܸ௜௝ is described in the below flow chart.
start
Find the farthest node
from the node 1
is generation node
yes
Sort the nearest load node to node i
yes
Traversed all nodes
no
yes
OLT ONU2
ONUn
Fig.2. The structure of EPON
D. Network topology of access layer in distribution
communication network
The Hand-in-hand network of whole-chain network
topology is shown in figure 3. OLT is placed at electronic
station in this network structure, and two optical directions
are extended by 1:2 splitters in two optical directions. The
link of each ONU implements "1+1" redundancy protection
of the link through the dual PON port. Topology is the most
NO
intuitive and essential attribute of communication network.
In conclusion, in order to ensure the reliability of power grid
and communication network, the dual-link network structure
is more reliable than the single-link network structure. In the
´=i case that one of the OLTs devices or one of the OLTs device
Sort the nearest
is attacked or damaged, the dual-link network structure could
no generation node ´ to still enable ONUs to connect to OLT and realize information
node j
transmission.
ODN
ONU1

End
OLT ONU2

Fig. 1 The process of calculating electricity exchange

ONUn
When the power transmission distance is short and more
power is delivered between node ‹ and node Œ, these two
Fig.3. Hand-in-hand network of whole chain protection
nodes take more responsibility of keeping global efficiency.

2
 ேೀಿೆ
E. Model of communication network σ௜ୀଵ ‫ܤ‬௜ ‫ͻ ا‬͹Ͳ‫„ܯ‬Ȁ• (9)
1᧥ Objective function The calculation formulation of total flow for all users
Min C=F᧤ܰ௉ைௌ ,†, ‫᧥ܦ‬ (3) connected to the PON is as below:
௧
The total cost C of an EPON communication network σே ೀಿೆ ேೀಿೆ
௜ୀଵ ‫ܤ‬௜ ൌ σ௜ୀଵ σ୫ୀଵ ߮௜௠
೔
(10)
can be expressed as:
CASE STUDY
C =ܰை௅் *‫ܥ‬ை௅் +ܰைே௎ *‫ܥ‬ைே௎ +ܰ௉ைௌ *‫ܥ‬௉ைௌ +᧤‫ܥ כ ܦ‬ଵ ൅ † ‫ܥ כ‬ଶ ᧥
(4) F. Case study of distribution network
where, ܰை௅் ǡ ܰைே௎ ,ܰ௉ைௌ is the number of OLT, ONU and 1᧥The test system
POS, respectively. ‫ܥ‬ை௅் is the cost of one OLT; ‫ܥ‬ைே௎ is the According to ‫ܫ‬௡௢ௗ௘ ሺ݅ሻ , the important nodes in
cost of one ONU,‫ܥ‬௉ைௌ is the cost of one POS;ܰை௅் is the distribution network with various distributed generations
amount of OLT. ‫ ܦ‬is the total length of optical fiber setting will be found. This section selects the IEEE 33-bus system
through the existing cable channel; † is the total length of as the experimental object. The structure of this system is
optical fiber setting through the newly building cable trench; shown in figure 4. In addition, a time-span data is used in the
‫ܥ‬ଵ is the cost of the unit length optical fiber, ‫ܥ‬ଶ is the unit evaluation process, which is the average of typical scenes.
cost of optical fiber and the installation cost of unit length These nodes with blue lines represent the node is connect to
optical fiber on the new cable trench. the distributed generations in the distribution network.
2᧥Constraints
Constraint 1: reliability
The reliability of the communication network is that the
communication path could connect from ONU to OLT
without interruption. The path of the communication network
is a collection of nodes and links, and the availability of one
path represents that all nodes and links on this path are
available. If the two nodes have higher redundancy, it Fig.4. The topology of distribution network
indicates that the connectivity of these two nodes is stronger
and the reliability of their communication is higher. If an 2᧥Simulation results and analysis
ONU has more than one path to OLT, the ONU is more
reliable from the perspective of topology. We should make
some important ONUs have higher reliability. The reliability
of path from ݅௧௛ ONU to OLT:
Ri = rONU ∗ (1 − ∏ t =1 (1 − ∏ s =1 rt , s ))
T S
(5)

where, ‫ݎ‬ைே௎ is the reliability of ݅௧௛ path, which is from ONU

to OLT;‫ݎ‬௧ǡୱ is the reliability of the ‫ݏ‬௧௛ series link of the ‫ݐ‬௧௛
parallel branch; ܵ represents the number of series
components on one parallel branches, here we didn’t
distinguish links from device nodes, just considering all links
and devices as components. ܶ represents the number of
different parallel branches of an ONU arriving at the OLT.
ƒ ‫ܴ ا‬௜ (6)
where a is a constant, the value of a will be obtained after Fig.5. The significance of nodes in power grid
evaluating the original communication network that adopts In figure 5, the red area and grey area show the value of
the single-link fiber network structure. According to the ‫ܫ‬௡௢ௗ௘ ሺ݅ሻ of 33 nodes, which are different from the weight of
evaluation result of original communication network, a is set each part when calculating ‫ܫ‬௡௢ௗ௘ ሺ݅ሻ. This figure indicates the
as 0.9623. importance of all nodes. In figure 5, these blue lines, yellow
Constraint 2: optical power loss limitation of one path
The transmission index of one path calculate according lines and green lines are the values of every part of ‫ܫ‬௡௢ௗ௘ ሺ݅ሻ.
to the worst value method. The transmission loss between the The most important node is node 1, because node 1 connect
OLT and the ONU should meet the following formulas. to the large grid, most of the electricity power of the
distribution network is provided from node 1. The second
Ltotal = ¦ i =1 Li + ¦ i =1 K i + ¦ i =1 M i + ¦ i =1 Fi + Lredun (7)
n m p h
and third important nodes are 24, 25, respectively, whose
loads are very large. Nodes 12, 30 are ranked as the top rank
(8) of importance since more distributed generation devices are
Ltotal ≤ Ptotal
connected to these nodes.
where, σ௡୧ୀଵ ‫ܮ‬୧ is the sum of optical fiber attenuation of all G. Case study of distribution communication network
optical channel, σ௡୧ୀଵ ‫ܮ‬୧ ൌ σ௡୧ୀଵ ܿ ‫ܦ כ‬௜ , c is a constant,‫ܦ‬௜ is 1᧥Original communication network
the length of optical fiber; σ௠୧ୀଵ ‫ܭ‬୧ is the sum of the insertion
௣
attenuation of m optical active connectors; σ୧ୀଵ ‫ܯ‬୧ is the
sum of attenuation of p fiber fusion splices; σ୦୧ୀଵ ‫ܨ‬୧ is the
sum of the insertion attenuations of the h optical splitters;
‫ܮ‬୰ୣୢ୳୬ is the redundancy of optical power, generally ‫ܮ‬୰ୣୢ୳୬ =
3dB; and ܲ୲୭୲ୟ୪ is the total optical power budget of EPON. In
most literature, ܲ୲୭୲ୟ୪ ൌ 26dB;
Constraint 3: bandwidth capacity of network
Fig.6. Original communication network

3
 As shown in figure 6, the original communication 23 24 25 26 27 28 29 30 31 32 33

network includes three ODNs. The blue lines in figure 6 are

original communication lines, which means cable trench in 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 34

those locations has been existed. The orange lines present the 19 20 21 22

possible cable trench, which is designed to form the dual-link

protection. Though the construction unit price of these two
new cable trenches is higher than other cable trenches, 23 24 25 26 27 28 29 30 31 32 33

adopting new cable trench will reduce the distance in some

schemes. This paper selects the important nodes of the 33- 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

node system according to different proportions. Then, the 19 20 21 22

communication network topology optimization could

upgrade according to the corresponding communication
Fig.9.The communication planning scheme in ratio=0.4
nodes.
As shown from Figures 7 to 12, the topologies of 23 24 25 26 27 28 29 30 31 32 33

distribution network have changed with the increase of the

ratio of selecting important nodes. The blue line is original 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 34

communication network as shown in figure 6, which 19 20 21 22

connects to OLT1 located on node 1. Other nodes connect to

another OLT2 located on node 34. The upper blue area
presents the distribution network, which is cyber network of 23 24 25 26 27 28 29 30 31 32 33

the active distribution network; the grey area presents the

distribution communication network, which is physical 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

network of the active distribution network. When the ratio is 22

over 0.4, to satisfy the technical requirements, topologies

19 20 21

have gradually evolved into two subsystem that called ODN

Fig.10.The communication planning scheme in ratio=0.5
in communication system. In figure 9 to 12, the topology of
communication network includes one yellow line (from node 23 24 25 26 27 28 29 30 31 32 33

33 to node 13), which means the optical fiber cross this new
cable trench. This result is based on the proposed 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 34

optimization model that considers the different cost between 19 20 21 22

building new cable channels and utilizing existing cable

channels. According to relevant technical requirements, the
number of nodes, the practical results, when ratio is over 0.5, 23 24 25 26 27 28 29 30 31 32 33

the optimal scheme should include more than one OLT. In

figures 10, 11, 12, the optimal schemes include more than 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

one OLTs. 19 20 21 22

23 24 25 26 27 28 29 30 31 32 33
Fig.11.The communication planning scheme in ratio=0.6

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 34 23 24 25 26 27 28 29 30 31 32 33

19 20 21 22

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 34

19 20 21 22

23 24 25 26 27 28 29 30 31 32 33

23 24 25 26 27 28 29 30 31 32 33

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

19 20 21 22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

19 20 21 22

Fig.7.The communication planning scheme in ratio=0.2

Fig.12.The communication planning scheme in ratio=0.7
23 24 25 26 27 28 29 30 31 32 33 Table 1 reflects the impact of ratios of important nodes
in distribution power network on the constructing cost of a
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 34
corresponding distribution communication network.
Table 1 Cost of different ratio
19 20 21 22

The Ratio of
0.2 0.3 0.4 0.5 0.6 0.7
important node

23 24 25 26 27 28 29 30 31 32 33

Investment᧤$᧥ 16,715 17,670 19,222 20,472 22,855 25,458

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

19 20 21 22
In the communication network, the most important
factor is the connectivity between the ONU and OLT. In this
Fig.8.The communication planning scheme in ratio=0.3
figure, the value of the reliability is based on all edges in
communication network. In figure 13, it is clear that the
reliability of whole communication system has been
significantly improved after optimizing the topology of
communication network using the proposed model in this
paper. At the very beginning, the reliability of original

4
network is 0.9623, when setting ratio as 0.7, the average
reliability of whole network is over 0.98, which is a great
improvement.
Fig.13. The change of reliability after optimization
As shown in figure 13, in order to improve the reliability
of the distribution communication network, more investment
is necessary. If planners prefer to protect more nodes by
adopting dual-link structure, the reliability of distribution
communication network will increase with more investment.
Especially, when choosing more than half of nodes to adopt
double-link structure, the cost has risen sharply. At the same
time, the reliability of the system didn’t perform as well as
expected. The reliability of whole system is 0.9623 when
communication network uses original topology structure as
shown in figure 6, which has increased to 0.97924 after
adopting dual-link structure.
Fig.14. Sensitivity analysis of reliability and cost
According to the sensitivity analysis, it shows that the
reliability of communication system has been improved by
adding link to another OLT on some nodes. According to
figure 14, it is not necessary to choose many nodes. The ratio
of important nodes being under 0.5 is enough, which could
improve the reliability of system to large extent and keep the
cost affordable. As for some important node, such as node 2,
the value of reliability of this node improve obviously, which
indicates that the dual link strategy could improve the
reliability of important node. In addition, it is obvious that if
planner continue to increase investment, the effect of
reliability improvement is not obvious, especially when the
ratio of important nodes is over 0.6.
CONCLUSION
This paper studies the distribution communication
network topology planning based on node importance in the
power distribution network, and calculates the reliability
indices. Some new characteristics of distribution network are
considered in the importance index proposed in this paper.
The results of this new indicator show that some nodes in
distribution network take more responsibility with more
loads or more DGs. Then, an optimization model is
established to find the optimal network of communication
network adopting EPON technology. This model aims to
improve the reliability of communication network by
providing some important nodes a standby link to another
OLT, which will increase the investment to some extent. In
a word, planning optimization should take a balance between
the reliability of communication network depends on the
topology connection and the cost of constructing new fibers
and channels. The sensitivity analysis of reliability and cost
show that choosing a propriate ratio of important nodes to be
upgraded is reasonable and economical.
ACKNOWLEDGMENT
This work was supported in part by the National Science
Foundation Industry/University Cooperative Research Center
on Grid-connected Advanced Power Electronic Systems
(GRAPES) under Award GR-18-06.
REFERENCES
[1] X. Dong, Y.Zhang and J.Tong, “Planning-Operation Co-
Optimization Model of Active Distribution Network with Energy
Storage Considering the Lifetime of Batteries.” IEEE Access, vol.
6, 2018, pp. 59822–59832.
[2] J. Zhang, and S. Min, “Reliability Assessment of Distribution
Network System Considering Output Uncertainties of Distributed
Generators.” Journal of Physics: Conference Series, vol. 1087,
2018, p. 042017.
[3] C. Li, W. Han, L. Zhou and X. Li, “Coordinated Control Strategy
of Distributed Generation Based on Active Distribution Network
under Multi-Time Scales.” 2016 International Conference on
Smart City and Systems Engineering (ICSCSE), 2016, pp. 417–
420.
[4] W. Liu, Z. Li, B. Zeng, M. Yang, H. Qin, X. Zheng and D. Zhao,
“Collaborative Planning of DERs and Intentional Islands in
Distribution Network Considering Loss-of-Load Risk.” IEEE
Access, vol. 6, no. 99, 2018, pp. 45961–45973. 20
[5] W. Liu, Q. Gong, H. Han, Z. Wang, L. Wang, “Reliability
Modeling and Evaluation of Active Cyber Physical Distribution
System.” IEEE Transactions on Power Systems, vol. 33, no. 6,
2018, pp. 7096–7108.
[6] Z. Liu, Y. Wang, W. Li, S. Xu, W. He and K. Sun. “A planning
method of clustering ONUs based on reliability.” 2017 16th
International Conference on Optical Communications and
Networks (ICOCN), Wuzhen, 2017, pp. 1-3.
[7] K. Wang, B. Zang and Z. Zhang, “An Electrical Betweenness
Approach for Vulnerability Assessment of Power Grids
Considering the Capacity of Generators and Load.” Physica A:
Statistical Mechanics and Its Applications, vol. 390, no. 23-24,
2011, pp. 4692–4701.
[8] Y. Guo, M. Xu, “Research on Reliability Evaluation Model and
Path Optimization for Power Communication Network.” 2015
5th International Conference on Electric Utility Deregulation and
Restructuring and Power Technologies (DRPT), 2015, pp. 2495–
2500.
[9] Kaur, Hardeep, M. L Singh, “Performance Analysis of Mobile
WiMax for Frequency Selective Fading Channel Models.” 2016
International Conference on Communication and Signal
Processing (ICCSP), 2016, pp. 1889–1892.
[10] Emilio Ancillotti, Raffaele Bruno and Marco Conti, “The Role of
Communication Systems in Smart Grids: Architectures,
Technical Solutions and Research Challenges.” Computer
Communications, vol. 36, no. 17-18, 2013, pp. 1665–1697.
[11] X. Bao, “Analysis of Reliability and Cost Performance of EPON
Protection Network in Access Layer of Distribution
Communication Network.” Power System Automation
37.8(2013): 96-101.
[12] Shannon. C, “A Mathematical Theory of Communication.” ACM
SIGMOBILE Mobile Computing and Communications Review,
vol. 5, no. 1, 2001, pp. 3–55.
[13] Y. Qian, Y. Li, M. Zhang, G. Ma and F. Lu, “Quantifying edge
significance on maintaining global connectivity.” Scientific
Reports, 2017, 7:45380.
[14] Siddhartha K. Khaitan, James D. McCalley, “Cyber-Physical
System Security for the Electric Power Grid.” Proceedings of the
IEEE, vol. 100, no. 1, 2012, pp. 210–224.
[15] Y. Mo, Tiffany Hyun-Jin Kim and Kenneth Brancik, “Cyber-
Physical Security of a Smart Grid Infrastructure.” Proceedings of
the IEEE, vol. 100, no. 1, 2012, pp. 195–209.