Access-2020-20869 Proof Hi

IEEE Access
Electric Power Grid Resilience to Cyber Adversaries: State of

the Art
Journal: IEEE Access
Manuscript ID Access-2020-20869
Manuscript Type: Regular Manuscript
Date Submitted by the

20-Apr-2020
Author:
Complete List of Authors: Nguyen, Tien; The George Washington University, Electrical and
Computer Engineering
Wang, Shiyuan; The George Washington University, Electrical and
Computer Eng
Alhazmi, Mohannad; The George Washington University, Electrical and
Nazemi, Mostafa; The George Washington University, Electrical and
Estebsari, Abouzar; Politecnico di Torino, Department of Energy
Dehghanian, Payman; The George Washington University, Electrical and
Keywords: <b>Please choose

keywords carefully as they
Cybernetics, Cyberspace, Smart grids, Resilience, Detection algorithms
help us find the most suitable
Editor to review</b>:
Subject Category<br>Please
select at least two subject
Power engineering and energy, Systems, man, and cybernetics
categories that best reflect
the scope of your manuscript:
Cyber Physical Systems (CPS), Cyber Attacks, Intrusion Detection

Additional Manuscript
System (IDS), False Data Injection Attack (FDIA), Energy Management
Keywords:
System (EMS)
For Review Only

Page 1 of 22 IEEE Access
1
2 Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000.
3 Digital Object Identifier 10.1109/ACCESS.2017.DOI
4
5
6
7 Electric Power Grid Resilience to Cyber
8
9 Adversaries: State of the Art
10
11 TIEN NGUYEN1 , SHIYUAN WANG1 , (Student Member, IEEE), MOHANNAD ALHAZMI1 ,
12 (Student Member, IEEE), MOSTAFA NAZEMI1 , (Student Member, IEEE), ABOUZAR
13 ESTEBSARI1,2 , (Member, IEEE), and PAYMAN DEHGHANIAN1 , (Member, IEEE)
14 1
Department of Electrical and Computer Engineering, The George Washington University, 800 22nd St NW, Washington, Suite 5900, DC 20052, USA.
2
15 Department of Energy, Politecnico di Torino, Corso Duca degli Abruzzi, 24, 10129 Turin, Italy.
16 Corresponding author: Payman Dehghanian (e-mail: payman@gwu.edu).
17
18
19
ABSTRACT The smart electricity grids have been evolving to a more complex cyber-physical
20
ecosystem of infrastructures with integrated communication networks, new carbon-free sources of power
21
generation, advanced monitoring and control systems, and a myriad of emerging modern physical hardware
22
technologies. With the unprecedented complexity and heterogeneity in dynamic smart grid networks comes
23 additional vulnerability to emerging threats such as cyber attacks. Rapid development and deployment of
24 advanced network monitoring and communication systems on one hand, and the growing interdependence
25 of the electric power grids to a multitude of lifeline critical infrastructures on the other, calls for holistic
26 defense strategies to safeguard the power grids against cyber adversaries. In order to improve the resilience
27 of the power grid against adversarial attacks and cyber intrusions, advancements should be sought on
28 detection techniques, protection plans, and mitigation practices in all electricity generation, transmission,
29 and distribution sectors. This survey discusses such major directions and recent advancements from a lens
30 of different detection techniques, equipment protection plans, and mitigation strategies to enhance the
31 energy delivery infrastructure resilience and operational endurance against cyber attacks. This undertaking
32 is essential since even modest improvements in resilience of the power grid against cyber threats could lead
33 to sizeable monetary savings and an enriched overall social welfare.
34
35
36 INDEX TERMS Cyber physical systems (CPS), cyber attack, intrusion detection system (IDS), false data
37 injection attack (FDIA), energy management system (EMS), power grid resilience.
38
39
40
I. INTRODUCTION operational functions such as automatic generation control
41
(AGC), state estimation (SE), and energy management
42 MART GRIDS have transformed the monitoring,
43
44
S control, and operation of bulk power grids via modern
communication, signal processing and control technologies.
systems (EMS) which have been globally observed to
be on the rise in recent years [5]. Cyber attacks are
malicious intrusions triggered by disrupting the cyber layers
45 While the smart grids allow for power networks to
of the communication systems in the power grid. There are
46 be effortlessly and wide-area monitored, the widespread
generally four types of attacks that the power grid may
47 deployment of modern information and communication
be vulnerable to: physical-only, cyber-only, cyber-enabled
48 technologies (ICTs) engenders a significant security concern
physical and physical-enabled cyber attacks [6]. Disruptions
49 and vulnerability to malicious cyber attacks: adversaries
appear when either the system operator makes a detrimental
50 which may alter the underlying physical systems and
error based on compromised sensor measurements or the
51 processes, thereby potentially compromising the national
power grid is remotely or directly controlled by a malicious
52 security [1–3]. With the extensive integration of cyber
intruder [7]. An intruder may be motivated to initiate a
53 infrastructure in smart grids is formed an expanded
cyber attack for many reasons including financial benefits,
54 attack surface characterized by intensified complexity,
large blackouts, or a combination of both [8]. The gravity
55 heterogeneity and number of resources [4] (see Figure
of the attack is dependent on the resources the attacker
56 1). This is evidenced by the frequency, complexity, and
has access to and the knowledge he/she possesses on the
57 severity of cyber attacks targeting several key power system
58
59 VOLUME xx, 20xx For Review Only 1
60
IEEE Access Page 2 of 22
Author et al.: Electric Power Grid Resilience to Cyber Adversaries: State of the Art
1
2
3
4
5
6
7
8
9
10 PM Us PM Us PM Us Sm ar t M eter s RTUs IEDs Sm ar t M eter s
11
12 Com m u n act i on Net w or k
13
14
15 SCADA Syst em
16 Net w or k Ser ver SCADA Ser ver Sm ar t M et er Ser ver PDC Appl i cat i on Ser ver
17 M easu r em en t : St at u s: Con t r ol Com m an d:
18 Bus Data (load/gener ation): P, Q, V, F Cir cuit Br eaker Data Tr ip/Reset Relay
19 Substation Data: P, Q, V Gr id Topology
Pr otection Data
Open/Close Br eaker
... ...
20 ...
21
22
Con t r ol Cen t er
23
24
Autom atic Decision System Dynam ic Recor d/Analysis. System -w ide Fr equency/Voltage/Flow Contr ol
System Stablity M onitor ing and Contr ol. Low -fr equency oscillation Analysis and Suppr ession
25 Hum an Decision Load Balance and Congestion M onitor ing ...
26
27 Figure 1: General view of the power grid cyber vulnerabilities.
28
29
30 system topology. The more accurate model the attacker penetrated a computer network at the Davis-Besse nuclear
31 has access to, the larger the deception attack that can be power plant located in the US [13]. There have also been
32 executed undetected [9]. Attackers may take advantage of reports that an experimental cyber attack was launched by
33 their knowledge of the grid and launch coordinated attacks researchers which caused a generator malfunction and self-
34 to critical substations in the network which may eventually destruct [14]. Energy theft is another common cyber attack
35 cause brownouts/blackouts with significant techno-economic practice in which the electric power is misused or "stolen"
36 consequences [10]. From a realization perspective, a cyber by a malicious intruder. Reports reveal that the United
37 attack can be considered measure-based or control-based. States loses ∼ $6 billion due to energy theft alone while it
38 A measure-based cyber attack targets the tie-line flows and accounts for ∼ $25 billion loss by the electric companies
39 frequency measurements, i.e., the measurements PL45, PL69 globally [15]. Even advanced metering infrastructure (AMI)
40 and the system frequency being sent to the control center. platforms, which are used to moderate the power flows in
41 A control-based cyber attack targets the area control error the grid, have been compromised and abused for energy
42 (ACE) values once they are sent from the AGC algorithm theft. In 2009, the FBI reported a wide and organized energy
43 and before they arrive the designated substations. An attack theft scheme which may cost a utility company up to $400
44 can send the opposite-direction ramping commands to the million annually following the deployment of AMIs [16, 17].
45 generating units. Modifications to the ACE signals (e.g., a One major known attack was the cyber attack that occurred
46 sign change of the ACE value) can lead to the generation on the Ukrainian power grid, happened on December 23rd,
47 ramping-up for load reduction and vice-versa [11]. 2015, where a third party from Russian security services
48 Historically, there were reported incidents in which power illegally entered the SCADA systems and computers, and
49 systems and industrial control systems (ICS) had their ultimately caused a blackout with massive consequences: a
50 systems cyber compromised. In the United States, the power service outage that left 225,000 customers without electricity
51 grid was penetrated in 2009 by cyber spies and a key for 2-6 hours [18]. Such blackouts are detrimental in that
52 infrastructure was compromised by an undetected intrusion: they cause financial losses and disruptions in all aspects of
53 Siemens supervisory control and data acquisition (SCADA) our everyday life [19]. Hence, characterization, modeling,
54 systems were attacked by computer worm Stuxnet. In 2010, and assessment of the power grid cyber vulnerability and
55 Stuxnet was able to infiltrate Iran’s Natanz nuclear fuel- designing solutions to protect the grid and enhance its
56 enrichment facility which was a part of Iran’s nuclear resilience against cyber adversaries is essential. This is
57 development project [12, 13]. In 2003, a cyber-attack because even modest improvements in resilience of the power
58
2 VOLUME xx, 20xx
59 For Review Only
60
1
2
3
4 Con t r ol Layer Net w or k Layer Sen sor /Act u at or Layer Physi cal Layer
5 Sensor s
6 PM U
7
8
9
10 Actuator s
11
12
13
Figure 2: General architecture of a CPS with multiple layers.
14
15
16 grid against cyber threats (through advanced monitoring, of dynamic swing equations for generator rotor dynamics
17 efficient threat detection, and recovery algorithms) could lead and algebraic load-flow equations for power flow through
18 to sizeable monetary savings and an enriched overall social network buses [22].
19 welfare. More critically, it could help reduce undesirable Aiming at reporting the existing state of knowledge on the
20 social, psychological, and physical outcomes associated with topic, this paper is structured as follows: a background on
21 the prolonged power outages resulting from cyber intrusions, cyber attacks including definitions, potential attack surfaces,
22 e.g., premature death, injury, social unrest, etc. and the impacts on bulk power grids are presented in
23
Section II. Section III discusses the power grid resilience to
24 Various studies have investigated the impact of cyber cyber attacks and how the smart grid cyber layer should be
25 attacks against different day-to-day operation and control characterized to resist cyber threats, ensuring the operational
26 mechanisms in power grids, including but not limited to endurance and resilience. Section IV reviews some protection
27 state estimation (SE), electricity markets, power system mechanisms in power systems against cyber adversaries to
28 protection, renewable forecasts, and power system dynamics prevent failures, followed by Section V where mitigation
29 and control [1, 2, 20–31], among many others. A cyber- solutions are reviewed. The paper will be concluded in
30 resilient power grid entails fault tolerance, fast response, Section VI with several concluding remarks.
31 recovery and reliability. Ensured resilience of the power
32 grids against extremes does not only reduce the volume of
II. CYBER ATTACKS: MODELING AND CLASSIFICATION
33 outages, but also ensures that the grid timely responds to
34 a variety of cyber catastrophes and man-made faults [32]. In this section, the root causes of cyber attacks and the attack
35 In the case of power transmission systems, difficulty in surfaces are firstly reviewed to overview where in power grid
36 maintaining system security arises in that intelligence is such threats would emerge. The impacts of cyber attacks on
37 only applied and available locally by protection systems power systems are next discussed, considering the technical
38 and by central control through SCADA systems. In some failures and the consequent effects of triggering events.
39 cases, the central control system is slow to respond to cyber
40 attacks and the protection systems are limited to a few local A. CYBER ATTACK ROOT CAUSES AND SURFACES
41 components [33]. There are many methods to model a cyber The smart grid is a hybrid of power and communication
42 attack: an Unmanned Aerial Vehicle (UAV) trajectory plot systems, the latter of which renders vulnerabilities which can
43 can model the different paths it can take based on the type be compromised during a cyber attack; these vulnerabilities
44 of the cyber attack and the impact the attack imposes on the are confidentiality, integrity and availability (CIA) [38]. In
45 power grid [34]. Attack trees can be devised to model many today’s standards, the power grid is characterized as a cyber-
46 types of cyber attack scenarios encompassing all possible physical system (CPS) shown in Figure 2, which contains
47 approaches an attacker may take [15, 35] or other methods physical, sensor/actuator, network, control, and information
48 such as Markov decision process (MDP) to enumerate all layers. Manipulation of each layer is possible but does
49 possible attack scenarios [36]. In order to model a control not necessarily mean an intrusion detection component or
50 system, however, a graph-based topological network model system needs to be applied in all layers. Information flows
51 or graph theory is proposed for a target control system. in between all layers as they operate only in tandem [39].
52 Integrated with logical connection information, it permits the Cyber attacks appear in many different forms, where its
53 implementation of a simple Prolog-based expert system to most basic definition is man-made manipulation of the power
54 represent a device visibility path and allows assessment of the grid and redirecting power flow to where it is unassigned
55 device vulnerability [37]. A classical mathematical model to by the network operator (see Table 1). Some of the main
56 describe the power transmission grid is commonly referred to common attacks are denial of service (DoS), false data
57 as the structure-preserving power network model. It consists injection attack (FDIA), energy theft [17], insertion of
58
VOLUME xx, 20xx 3
59 For Review Only
60
1
2 Table 1: Potential Attack Surfaces in Power Grids
3
4 Transmission Distribution
Device System Attack Type
System System
5
Data Concentrator (DC) X X X FDIA/Delay/Jamming
6 SCADA X X X FDIA/DOS
7 Control System X X X FDIA/DOS
8 State Estimator X X FDIA
Communication Channel X X X DOS/Jamming/Delay
9 Power Market X X FDIA/DOS/Delay
10 Remote Termial Unit (RTU) X X X FDIA/DOS/Delay/Jamming
11 Phasor Measurement Unit(PMU) X X X FDIA/Delay/Jamming
12 Programmable logic controller (PLC) X X X FDIA/Delay/Jamming
Advanced Meter Infrastructure (AMI) X X FDIA/Jamming
13 Intelligent Electronic Device (IED) X X FDIA/Jamming
14
15
16 malware or worms, as well as physical damage of the power infrastructure increases the system connectivity and
17 grid such as causing equipment to self destruct [14, 35, 40]. autonomous decision-making by employing standardized
18 • DoS attacks are often realized when the attacker information protocols that often have (or will have in the
19 jams the communication channels, compromises the future) publicly documented vulnerabilities. Motivations
20 electronic devices, and attacks the routing protocols for cyber attacks also abound. Market deregulation and
21 which ultimately lead to delays and congestion in privatization of the energy industry has increased the
22 the communication channels. Generally, a DoS attack competition among energy providers to enhance consumer-
23 restricts a legitimate user’s access to the services and centricity. Threats also exist in the form of dissatisfied utility
24 resources by flooding the communication network with insiders, electricity consumers, and cyber terrorists.
25 unnecessary traffic [12, 41].
26 • FDIA scenarios are realized when an attacker injects B. IMPACTS OF CYBER ATTACKS ON POWER GRID
27 false data, usually on a communication line between the Control systems are becoming more vulnerable as they
28 field sensors and the control center, with the intent to get overwhelmingly coupled with modern information and
29 deceive the network operator and even disturb the SE communication technologies and the physical controllers
30 processes [8, 13]. FDIAs may result in a wide variety of in a CPS [57]. The critical equipment and systems which
31 outcomes depending on the intruder’s intention, some of can be mainly affected or exploited during an attack are
32 which include energy theft, miscalculation of locational in the energy management systems (EMS) in transmission
33 marginal prices (LMP) for illegal market profits, and networks or distribution management systems (DMS) in
34 physical damage upon the network. FDIAs can affect distribution networks. Such platforms collect data from
35 the LMPs by misleading the SE which then adversely remote and distributed meters and sensors across the network
36 affects the contingency analysis procedures [42]. and generate estimates of the system states at the intervals of
37 • Insertion of malware or worms can range in roughly 15 minutes [46–51, 53, 55]. When false meter data is
38 different types from malicious software which runs in injected through a cyber attack, the EMS or DMS functions at
39 backgrounds to slow down the operations of the electric the control center will be misled by the state estimators which
40 utility computers to insertion of Trojan software to steal may potentially make erroneous decisions on contingency
41 practical security certificates [40]. analysis, power dispatch, and even billing actions [14].
42 Cyber intrusion does not necessarily have to occur in The smart grid offers synchrophasor-based cyber security,
43 the power system itself since it can originate from separate which entails a CPS system that provides real-time data
44 systems that interact frequently with the grid such as electric to the EMS in order to manage (monitor and control) the
45 vehicle supply equipment (EVSE) [43]. In [40], a malware physical network [58–61]. However, the latest synchrophasor
46 attack model is able to attack the electric vehicle (EV) devices, e.g., phasor measurement units (PMUs), as well as
47 infrastructure and its communication systems when EVs are digital fault records (DFRs) and protective relays with PMU
48 plugged in for charging. In some instances, attacks can be functionality are susceptible to a wide range of errors [53,
49 undetectable such as malicious data injection attacks that 60] including cyber attacks; this is even further challenging
50 alter the values of measurements without being detected, considering the fact that such equipment are intertwined
51 which may result in serious consequences [44]. with a large number of legacy devices that have little or
52 From an engineering perspective, there is an opportunity no protection against cyber attacks [62]. In [34], the CPS
53 for cyber attacks in smart power grids due to the proliferation security is analyzed where a deception attack compromises
54 and reliance on distributed advanced metering infrastructure sensors, actuators and both sensors and actuators. The
55 (AMI) [45], intelligent electronic devices (IEDs) [46–56], probability of FDIA to be launched successfully usually
56 and wireless and/or off-the-shelf communications components depends on two assumptions: (i) the attacker has control
57 and systems across the power network. Such cyber over some sensor nodes and (ii) the attacker has complete
58
4 VOLUME xx, 20xx
59 For Review Only
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16 Figure 3: System view of advance metering infrastructures.
17
18
19 knowledge of the system or its exact topology at all moments they actually are not [42], injection of false data can affect
20 during the attack [63]. Generally, the highest impact of the stability and security of the system [66]. Spatiotemporal
21 an attack is realized when an intruder gains access to the cyber-state correlations can be used to detect the FDIA.
22 supervisory control access points of SCADA systems and Potential anomalies can be detected by monitoring the
23 launches control actions [64]. The attacker may compromise temporal consistencies of the spatial correlations between
24 raw data measurements which causes undetected errors to state estimations [67].
25 factor into estimates of state variables such as bus voltage Another way an intruder can affect the communication
26 angles and magnitudes. This can occur when the attacker network is by attempting to connect and dial up to a
27 takes advantage of small errors tolerated by SE algorithms. remote terminal unit (RTU) or an IED which can allow
28 Ultimately, this severely threatens power system security them to wiretap telecommunications, perform a local-area
29 [13]. Disturbances in SE can lead to increases in state network (LAN) or wide-area network (WAN) transmission
30 estimates mean square errors (MSE) and changes in the real- shown in Figure 3. They could also attack the corporate
31 time electricity market prices. The effect of invalid MSE information technology (IT) systems and gain backdoor
32 can lead to network operators making wrong decisions and access to the interconnected EMS or SCADA systems;
33 the changes in real-time electricity market prices can benefit internet service providers (ISP) and telecommunications are
34 only the attacker. [65]. With FDIAs, the power grid can other sources they can attack. Some electric utility providers
35 suffer economic attacks, load redistribution attack, or energy are dependent on corporate IT systems and this is how
36 deceiving attack. An economic attack is a type of FDIA their interconnected SCADA systems greatly intensify the
37 which can affect operations of the deregulated electricity vulnerability of the electric power grid [37]. Similarly,
38 market which is comprised of two markets: the day-ahead AMI systems can be attacked. AMI includes smart meters,
39 market and the real-time market. An attacker can manipulate customer gateways, AMI communication network and head-
40 market prices for power and get monetary gains. A load end; AMI is considered a fundamental technology of smart
41 redistribution attack is an attack which can affect power grid grids for making two-way communications along with
42 operation by attacking the security-constrained economic various other functions. However, there have been several
43 dispatch (SCED). The purpose of SCED is to minimize potential vulnerabilities with AMI specifically the insertion
44 the total system operation cost; however, when the raw of malicious software and disabling of metering systems [38].
45 measurements are manipulated by an attacker, the SCED will Cyber attacks and intrusions can occur multiple times
46 result in an overload of the lines that will remain unnoticed from a single origin and spread to different areas. A typical
47 by the system operator and ultimately causes large physical example is electric vehicle charge stations (EVCS) [43,
48 damages to the power grid. An energy deceiving attack 68–72]: when a consumer charges its EV at multiple stations,
49 affects the distributed energy routing process; essentially it is likely that malware can spread due to vehicle-to-
50 this is a scheme to determine the optimal energy routes for infrastructure and EVSE communications. Essentially, an
51 load demand or generation. When measured data has been attack on an EV may spread to the power grid infrastructure
52 tampered, it can cause erroneous energy demand or supply starting from the EVSE and all the way up to the
53 messages to initiate [13]. Overall, cyber attacks can impact utility systems [40]. The integration of transportation and
54 four main aspects of the bulk power systems which are SE, power systems may leave many open doors for hackers,
55 AGC, voltage control, and energy market. FDIAs deceive especially in the interconnected environment, i.e., the EV
56 the system operators to believe that the current operating infrastructure, including EVs, EVSE, meters and other
57 conditions are secure both physically and economically when roadside infrastructures and when deeply integrated with
58
VOLUME xx, 20xx 5
59 For Review Only
60
1
2 etc.) and mission-critical services (e.g., health sector, defense
3 bases, etc.). The electric sector’s approach to the protection
4 of the grid critical infrastructure is generally known as
5 “defense-in-depth”, which contains prevention, preparation,
6 response, and recovery for an inclusive range of credible
7 hazards to electric grid operations. Resilience in power grids
8 entails accurate threat detection, infrastructure vulnerability
9 monitoring, and timely response and recovery (see Figure 4).
10 Both “long-term” and “short-term” strategies for enhancing
11 the grid resilience against extreme conditions have been
12 addressed in the literature. In the former, enhancing the grid
13 structural resilience is primarily the focus of concern and
14 suggestions are toward deployment of the “grid hardening”
15 plans through reinforcement, preventive maintenance of the
16 critical assets, vegetation management, efficient allocation
17 of flexible energy resources (e.g., storage units), etc. In
18 the latter, improving the operational resilience is targeted
19 Figure 4: Power system transition states following a critical
through fast emergency response and remedial actions,
20 HILP disturbance: a proactive response and recovery strategy
defensive islanding, use of the micro-grids, etc.
21 that can be implemented in an automated manner can
The IT employed in industrial control systems (ICS)
22 minimize or bypass the preparation time when the event hits
is cyber-vulnerable in general and can potentially impose
23 the power grid, thereby effectively boosting its resilience.
direct impacts on the physical power grids. CPS will be
24 the core component of many critical infrastructures, yet
25 vulnerable to random failures and cyber attacks. Hence, it
26 critical infrastructure systems [73–75]. is critical to design, develop, and implement ICS and CPS
27 with resilient cyber defense systems [12], i.e., integrating
28 III. POWER GRID RESILIENCE TO CYBER ATTACKS robust intrusion detection systems (IDS) to ensure the power
29 The concept of resilience has become a well-researched grid resilience with countermeasures being taken effectively
30 topic in recent years as it mainly drives the swift detection [39]. Energy theft is an important concern relating to smart
31 and effective mitigation of the power grid against high- grid implementation; while the implementation of AMI
32 impact low-probability (HILP) events [76–78]. The word is used to mitigate energy theft, penetration tests have
33 “resilience” is originated from the Latin word “resilire”, uncovered several vulnerabilities with smart meters [15, 17].
34 reflecting “the ability to rebound” [32]. Power system Deregulation of the electric power industry has unbounded
35 resilience in the face of the devastating natural-driven HILP generation and transmission systems which, in turn, allows
36 events has been studied widely in the literature [76–97]. The for a broad range of participants to make decisions in the
37 past research defines, quantifies, and categorizes the concept power sector. This is critical as an attack on the SCADA
38 of resilience in many different ways. For instance, the systems can disrupt and damage critical infrastructural
39 National Infrastructure Advisory Council (NIAC) proposed a operations, contaminate the ecological environment, cause
40 universal definition of infrastructure resilience in 2010: “the major economic losses and, and even more dangerously,
41 ability to reduce the magnitude and/or duration of disruptive claim human lives [102]. In presence of all these challenges
42 events. The effectiveness of a resilient infrastructure depends and vulnerabilities and the intensified number of access
43 on its ability to anticipate, absorb, adapt to and/or rapidly points and functionalities to tamper with [103] in highly-
44 recover from a potentially disruptive event” [98]. In a complex cyber-physical power grids, new strategies are
45 similar attempt, [99] defines resilience as the system’s needed to secure the entire network against malicious cyber
46 ability to withstand the main interruption within acceptable intrusions [3, 97, 104–106].
47 degradation parameters and to recover within an acceptable The potential for achieving power system resilience
48 time and composite risks and costs. An alternative definition depends highly on how preventive and corrective
49 of resilience is the ability to prepare for and adapt to changing maintenance strategies are planned and implemented
50 conditions, featured with robustness and fast recovery [100]. component-wise [107–112] and system-wide [113–127] as
51 The last but not the least interpretation of resilience could be well as where and how the security measures and systems
52 the system’s ability to prepare and plan for, absorb, recover are deployed. Incorporating data and cyber infrastructure
53 from, and more successfully adapt to adverse events [101]. to the power grid exposes the system to many cyber
54 The power grid is required to supply the electric power security threats. The smart grids of the future with
55 continuously and reliably to end-users in general and critical massive renewable resources and an expanded uncertainty
56 lifeline infrastructures (e.g., water networks, oil and gas set [128–130] will inherit not only the vulnerabilities of
57 systems, communication systems, transportation networks, advanced communication systems but also the vulnerabilities
58
6 VOLUME xx, 20xx
59 For Review Only
60
1
2 of the legacy power system. Security mechanisms should NIDS detects discrepancies and intrusions which violate
3 be designed into the power grid with the goal of reducing the predefined security rules by using a specification-based
4 vulnerabilities and mitigating their consequences [131]. algorithm. To detect energy theft, another common challenge
5 Anomaly detection and root-cause analysis are essential for in power systems, [17] uses normal and malicious data of
6 building resilient CPS since the grid may not know how to consumer consumption patterns and a consumption pattern-
7 counteract the damage if it does not know what caused the based energy theft detector (CPBETD). This tool combined
8 damage. Accurately detecting anomalies and isolating their with the application of a Support Vector Machine (SVM)
9 causes is important for applying appropriate proactive and anomaly detector allows the algorithm to use silhouette plots
10 preventive measures [57], [41], [66]. to identify different distributions in the dataset and relies on
11 distribution transformer meters to detect nontechnical loss
12 IV. CYBER ATTACK PROTECTION (NTL) at the transformer level. In order to detect cyber
13 Most methods for detecting cyber intrusions rely on outdated intrusions in the system, it is essential to classify it for
14 techniques that are originated from the IT domain and identification. Effective techniques to classify cyber attacks
15 adopted in smart grids in an insufficient manner. Typically, or anomalies are using SVMs and a variety of machine
16 the inherited techniques from power experts mainly focus learning algorithms.
17 on existing types of attacks, e.g., load redistribution [132], Detecting intrusions through the entire sector of the power
18 distributed DoS [133], etc. Real-time cyber vulnerability network is challenging; in [135], a proposal of grouping
19 assessment in power systems brings new challenges due to network buses and designing filters for detection and
20 the fact that the conventional techniques for cyber intrusion isolation of faults addresses a feasible detection mechanism.
21 detection in dynamic power systems are computationally In addition to grouping network buses, [135] suggests using
22 demanding to be applied in real-time. the swing equation to model the power network which can
23 Fundamentally, there are two types of attack detection and be used in tandem with grouping power buses. Investigating
24 identification strategies widely researched in the literature: system models and security requirements of AMIs to present
25 static and dynamic. Dynamic detection and identification an attack tree based threat model for AMI has shown an
26 outperform its static counterpart while possibly using improvement in the detection accuracy and detection speed
27 fewer measurements. With a comprehensive assessment of of intrusions in [15].
28 the limitations in both static and dynamic detection and While cyber attacks may become prominent in the future,
29 identification techniques, [22] proposes a provably-valid there are normal fault contingencies which occur in the
30 dynamic detection and identification procedure borrowing system on a daily basis driven by environmental stressors
31 tools from the geometric control theory domains: the tools and equipment failures. The system needs to be able to
32 are comprised of geometrically designed residual filters. differentiate the difference between an intrusion attack and
33 Cyber attack detection can be performed using relevant a natural discrepancy. In [136], a devised algorithm is
34 and high-fidelity data. Spotting slight anomalies in PMU implemented to accurately detect and locate faults in power
35 data helps identify unobservable cyber attacks which can systems in addition to identifying bad data using weighted
36 not be detected by existing technologies. In [134], a least absolute value (WLAV). WLAV has the ability to reject
37 convex optimization-based decomposition approach utilizes bad data to reduce dimensionality. A Bayesian framework
38 the low-ranking property of PMU data to formulate an can also be utilized to unify different approaches of network
39 unobservable cyber attack identification problem as a matrix detection based on random diffusions and algorithms which
40 decomposition problem where the observed data matrix is the are based on network’s spectral properties [137]. This
41 sum of the low-ranking PMU data and a linear projection of algorithm detects threat networks using partial observations
42 a column-spare matrix. The majority of the existing attack which can be optimal in the Neyman-Pearson sense and
43 detection methods use measurements at one-time instance prepares the system for cyber intrusion attacks should they
44 and only explore the spatial correlations whereas the convex- are launched in the future. A data-driven algorithm for online
45 optimization decomposition method in [134] exploits the power grid topology change identification with PMUs is
46 temporal correlations as well and can identify unobservable suggested in [58], where the proposed machine learning
47 cyber-data attacks even when the system is dealing with the algorithm can differentiate the various types of faults in
48 aftermath of disturbances. power grids and the topology switching actions initiated by
49 Strategies to detect cyber intrusions are plentiful and the system operators or attackers.
50 endless since there is an expanded set of cyber attack
51 surfaces and vectors to be able to manipulate the grid A. FDIA DETECTION
52 towards an intruder’s favor. In [19], a new network- The FDIA problem is viewed as a matrix separation problem
53 based cyber intrusion detection system (NIDS) uses multi- and two methods which are presently employed to solve
54 cast messages in substation automation systems (SASs) to this problem are nuclear norm minimization and low rank
55 monitor anomalies and malicious activities of multi-cast matrix factorization. These methods can recover lost or
56 messages which are based on IEC 61850, generic object- missing data in addition to detecting malicious attacks in
57 oriented substation event (GOOSE) and sample value (SV). the power grid. FDIA happens when an attacker injects false
58
VOLUME xx, 20xx 7
59 For Review Only
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17 Figure 5: FDI attacks on state estimation in a power grids.
18
19
20 data, usually on a communication line between sensors and from FDIAs to some extent, smart intruders may be still
21 the control center with the intent to deceive the network able to damage the PMU (or RTU) measurements in power
22 operator and even disturb the SE processes [8, 13]. Reference grids and bypass the bad data detection (BDD) mechanisms
23 [138] presents an approach using observer nodes to detect in SCADA systems and wide-area measurement system
24 and isolate cyber attacks on network nodes and those on (WAMS) platforms [14, 144–154] (see Figure 5). This
25 the communication links between the nodes. In order to can be accomplished through manipulated measurements
26 minimize the computational complexity, observer nodes and injecting artificially generated data to the basic
27 are reduced, while the observability of the system is not measurements in power grids [155–158]. An FDIA detection
28 compromised. A perturbation-based approach is employed mechanism in smart meters is modeled in [159]. Correlation
29 in [44] for detecting both fault-induced and maliciously- between the power system components and detection
30 injected bad data in the power grid. This method probes the methods against smart grid intrusions is proposed in [160].
31 system by applying known perturbations and measuring the An efficient approach to protect the power system from
32 values elsewhere to find unexpected responses in terms of FDIA is by implementing precautions in advance [161–164].
33 measurement values. [139] presents a mechanism for false Robust SE algorithms against FDIA based on Markov chain
34 data detection which notices the intrinsic low dimensionality theory and Euclidean distance metric are introduced in
35 of temporal measurements in power grids as well as the [165]. [26] modeled the FDIAs with multiple adversaries
36 sparse nature of the FDIAs. Several research efforts discussed against one defender implemented in the smart grid. A game
37 methods of building and detecting such an attack. Successful theoretic approach is used in [30] to study the interactions
38 implementation of FDIAs commonly requires full knowledge between the defender and the attacker in CPS. DoS attacks,
39 of the network topology. [140] proposed a form of an random attacks, and FDIA intrusions are detected in [166]
40 attack without having complete information of the network using Kalman filter by estimating the variables of the
41 topology. This can be done by using the kernel-independent state processes and feeding them to either the χ2 detector
42 component analysis to map the restricted data into a new or euclidean detector. In order to detect the injected bad
43 Jacobian matrix, through which the undetectable attack is data by PMUs, [167] introduced a distributed host-based
44 modeled [140]. [141] proposed an extreme learning machine collaborative detection method using a conjunctive rule based
45 (ELM) technique based on one-class-one-network (OCON) majority voting algorithm to detect such an attack.
46 framework to detect any cyber threat on the AC state
47 estimation. FDIA attacks are detected using Kullback-Leibler B. PHASOR MEASUREMENT UNIT PROTECTION
48 Distance in [142], where the accuracy of the detection In order for protocols and measurements to be true, exact,
49 mechanism is influenced by the predefined thresholds. A and valid at all times with robustness against any external
50 novel false data detection technique based on the separation changes, they need to be protected in smart grids. Protecting
51 of nominal power grid states and anomalies is discussed in a set of basic key measurements and having PMU based
52 [139]. [143] used an algorithm to ensure shorter decision protection mechanisms or secure PMU equipment [7, 13, 58,
53 time and a more promising FDIA detection accuracy by 134] can retain the fidelity of the measured data and accurate
54 tracking the unfamiliar parameters and process multiple state estimates in a wide variety of smart grid applications
55 measurements at the same time. using such measurements. When a set of measurements is
56 protected, an attacker can not inject unobservable attacks
57 Even though these techniques can prevent the system without hacking into the protected units [134] and allowing
58
8 VOLUME xx, 20xx
59 For Review Only
60
1
2 themselves to be noticed. A distributed intrusion detection
3 system can be deployed for smart grids to pinpoint cyber
4 intrusions. This system contains an analyzing module (AM)
5
M ag
and an intelligent module which communicates between
6 three different cyber layers of home area network (HAN),
7 the neighborhood area network (NAN) and the wide-area
8 network (WAN) [38, 168]. Scale
9 It is estimated that in order to achieve a full power Tim e (m sec)
10 system observability, one typically needs to install PMUs (a)

11 at around one-third of the network buses; nevertheless,
12 it is recognized that this is difficult and costly to
13 achieve in the near future [169]. Therefore, one will have
14
M ag
to estimate the state of the system with a hybrid of
15 both PMUs and conventional measurements. This practice
16 essentially leads to careful selections of PMU placement
17 strategies in the power grid in order to minimize the SE Scale
18 errors. [169] optimizes PMU placement to increase the Tim e (m sec)
19 SE accuracy using an algorithm that is related to key (b)

20 property and submodularity which contributes to efficient
21 Figure 6: Simulation results in a IEEE 30-Bus system from
greedy algorithms. An optimal PMU placement problem
22 [58], where (a) Transmission Line 2-4 is switched-off at t =
is interpreted as an optimal experiment design problem
23 30ms only, (b) Transmission Line 2-5 is switched-off at t =
with a class of optimality criteria. In particular, the greedy
24 30ms only. Both features are extracted at Bus 6.
PMU placement algorithm achieves at least 63 % of the
25 optimal total variance reduction for typical power systems.
26 Performing a vulnerability assessment is critical to ensure
27 that power infrastructure cyber security is systematically
28 evaluated. This proposed framework provides a measure
29 to quantify system vulnerability and a planning tool to
30 assist system analysts to identify bottlenecks in the system
31 where improvements are most effective [64]. Similarly,
32 a novel vulnerability measure is introduced in [170] to
33 compare and prioritize different grid topologies against
34 FDIAs with incomplete information of the grid’s topology.
35 This measure can potentially help build power grids that are
36 less vulnerable against practical FDIAs when the attacker Local PDC Local PDC Local PDC
37 has limited information and launches an imperfect attack.
38 In [3], discussions on how optimal placement of PMUs Cor por ate PDC
39 throughout the power network may lead to very accurate Data Stor age
Syn ch r oph asor Dat a Di m en si on al i t y Redu ct i on
40 SE are provided. PMUs also provides advanced mechanisms
41 in detecting stealthy attacks. Rerouting the topology of the Ear l y Even t Det ect i on
42 power grid intensifies the complexity of the grid topology Data Stor age
43 and is used as a defense mechanism against FDIAs which are

44 undetectable via conventional means [63]. References [63,
45 171–178] suggest that leveraging defensive circuit breakers Figure 7: Early event detection mechanisms in power grids.
46 and simultaneously applying grid re-configuration practices
47 can enhance the overall network efficiency, reliability, and
48 security. This is achieved at minimum cost and by harnessing features corresponding to different topology changes are
49 the network built-in flexibility only. Nevertheless, additions extracted as shown in Figure 6 which were used to detect and
50 of circuit breakers may not be a viable security measure classify the associated line switching actions characterized
51 if the attacker has compromised a large set of sensor through commutation jamming and/or FDIA scenarios.
52 nodes and knows a large portion of the grid topology. Implementing different techniques to reduce the number
53 In [58], advanced wavelet transform and machine learning of simulations and achieve a quicker SE allows for early
54 analytics are embedded in existing PMUs, devices with PMU event detection. This provides an opportunity for the network
55 functionalities, or as a stand-alone sensor in power grids that operators to be prepared for the potential adversarial cyber
56 can detect the malicious changes in network topology by an attacks since there will be additional time saved for the
57 attacker (unwanted line switching operations). The waveform optimal response to be deployed (see Figure 7). Using a
58
VOLUME xx, 20xx 9
59 For Review Only
60
1
2 Principal Component Analysis (PCA) based dimensionality the optimal feature and rule is extracted based on the
3 reduction of PMU data allows for raw data blocks to simplistic method [186], NNge which is a nearest-neighbor-
4 be processed quicker, thereby realizing an early detection like algorithm that classifies samples by comparing them
5 of cyber disruptions [179, 180]. Similarly, [134] uses to those which already have been observed and comparing
6 an unobservable cyber attack identification as a matrix the new examples to their surrounding data points [187],
7 decomposition problem which contains a sum of low-ranked Random Forests which is an ensemble of tree predictors
8 matrices with a linear projection of a column-sparse matrix. where each tree casts a vote for the most popular class on the
9 Since low-dimensional structure of PMU data matrix is input of a new instance [188]. In [189], an extended version
10 recently observed, the matrix decomposition problem has of deep belief network (DBN) called conditional DBN
11 attracted more attention and has wide spread applications (CDBN) was proposed to analyze the sequential PMU data in
12 such as internet monitoring, medical imaging and image real-time and detect the existence of information corruption
13 processing [134]. In [181], a similar technique is proposed using auto-regressive (AR) data modeling scheme. In [190],
14 which reduces the simulation run-time by incorporating the efficiency of the DL-based cyber-physical approach for
15 Importance Sampling which is used to speed up simulations FDIA detection is demonstrated. The proposed approach
16 several orders of magnitudes compared to the standard addresses both cyber (e.g., information corruption) and
17 simulation practices. This essentially increases the efficiency physical disruptions. Reference [191] used a scenario-
18 of simulations associated with Markovian models on highly based sparse cyber-attack model with incomplete network
19 dependable dynamic systems. information to detect the possibility of data manipulation.
20 In this paper, the results demonstrated that the proposed
21 C. DETECTION USING MACHINE LEARNING approach not only requires less assumption on system
22 Machine learning and artificial intelligence techniques are topologies and attack types, but also verifies the high
23 more recently proposed and applied in power systems to detection accuracy of the adopted DL. Reference [192]
24 identify disturbances and detect cyber attacks even through compared the performance of three different DL approaches:
25 deception [62]. Recent advancement in deep learning (DL), (i) gradient boosting machines (GBM), (ii) generalized
26 a subcategory of machine learning that uses artificial neural linear modelings (GLM), and (iii) distributed random
27 networks to extract accurate features from raw data, brings forests (DRF). The numerical results justified that DL-based
28 about new solutions for data-driven attack detectors. In fact, approaches can accurately detect FDIA scenarios against
29 DL approaches use feature learning techniques to extract SE algorithms. Reference [8] proposed two DL techniques
30 novel features (aka signatures) in an unsupervised, self- for FDIA detection in smart grids. The first model uses
31 guided manner. Given a set of measurement data, with raw the multivariate Gaussian semi-supervised learning while the
32 features as the input, DL tries to crate and refine a set of second model uses a measurement-based deviation analysis
33 algorithms to reproduce the same data set as the output. The algorithm. Both models are used to identify anomalies in
34 generated algorithms try to minimize the difference between transmission networks. In [193], a new detection framework
35 the input and the output so that the original data can be was proposed to develop a density ratio estimation (DRE)
36 recovered directly from the generated features [97]. technique: an efficient countermeasure against cyber-attacks.
37 A machine-learned framework is created in [182] Reference [194] proposes a DL-based model for FDIA
38 and refined with unsupervised feature learning to detect detection in smart meter data utilizing a state vector estimator
39 different types of cyber attacks in power systems. Stacked (SVE) and a DL-based identification (DLBI) algorithm. The
40 autoencoder-based unsupervised feature learning is proposed model uses the historical data and tries to recognize a pattern
41 to capture useful and rich patterns hidden in the data to identify FDIA scenarios in real-time.
42 to recognize the cyber attack, and achieve competitive
43 results compared with detectors relying on detailed system V. IMPACT MITIGATION AND RESTORATION
44 information and human expertise. In [41], research was In industrial applications, strengthening industrial control
45 done to combine SVM with a variety of machine learning systems (ICS) will protect different classes of infrastructure
46 algorithms to find the most promising algorithm which such as utilities and oil and gas facilities. The ICS is
47 can detect an adversarial intrusion. A robust spam filtering strengthened by designing an intrusion detection system
48 method is introduced in [183] using a hybrid method for contained in the cyber layer with a controller at the
49 rule-based processing and back-propagation neural network. physical layer dynamic system [12]. Having a resilient smart
50 In [184], different types of deep learning mechanisms, e.g. grid entails both accurate and swift attack detection and
51 ANN, decision trees, etc., are tested to assess the cyber timely response and recovery. This goal can be achieved
52 security of a particular IEEE test system. Reference [185] by having distributed control agents that facilitate both
53 introduced a new model for malicious code detection using a attack detection and system recovery through iterative
54 new hybrid DL model. A decision support tool is proposed local processing and message transmission. These control
55 in [62] which enables power system operators to classify agents are distributed across the grid, thereby characterizing
56 various types of attacks. In this paper, different types of distributed intelligence mechanisms [46, 58, 59, 104, 105,
57 classification algorithms are considered, e.g., OneR in which 195]. Limitations by a lack of information about cyber
58
10 VOLUME xx, 20xx
59 For Review Only
60
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16 Figure 8: System architecture which supports EVSEs.
17
18
19 1
MultiClass Three-Class Binary • Anomaly-based IDS detects malicious activities with
20 0.9 regards to deviations from statistically normal behavior
0.8 in the system.
21 0.7
22 • Specification-based IDS also recognizes intrusions with
Precision
0.6
23 0.5 regards to deviations from normal behaviors of the
0.4
24 0.3
system. However, instead of statistical measurements,
25 0.2 normal behaviors are characterized based on manually
0.1
26 0
extracted specifications of the system.
27 OneR Nearest
Neighbor
Random Naive Bayes
Forest
SVM JRip Adaboost+JRip
Characterization of irreducible attacks or observable
28 attacks with the compromise of two power injection meters
29 Figure 9: Precision results of various classifiers [62]. is performed in [7] with the use of an efficiently designed
30 algorithm to group all observable attacks. In addition,
31 the deployment of secure PMUs is approached as a
32 attacks can be partially removed by future research and countermeasure against unobservable attacks. When cyber
33 development of the advanced testbeds for comprehensive attacks occur, parts of the system will be compromised
34 testing and evaluations. Test beds are extremely useful tools and it is important to isolate them quickly while ensuring
35 for thorough evaluation of mitigation and economic strategies a sufficient supply of power (through available equipment)
36 in response to cyber vulnerabilities [46, 53, 64]. to the system load points and mission-critical systems
37 Modeling vulnerabilities in power grids are critical for and services [40, 85, 88]. Reference [40] claims that
38 its survival under adversarial attacks. In order to create a EVs mobility contributes to attack propagation. Therefore,
39 network topology model of vulnerabilities, device visibility when an attack spreads via EVs, a mixed-integer linear
40 and device vulnerability need to be defined and quantified. programming (MILP) optimization problem is suggested that
41 The concept of device visibility path, with the use of minimizes the risk of attack propagation while considering
42 a small Prolong application to assess the vulnerability the EV loads, EV threat levels and demand profile in power
43 level of a hypothetical target device, can help map the distribution system (see Figure 8). In such a CPS ecosystem
44 cyber vulnerabilities within a system, thereby enabling of EVs, isolating the compromised systems will mitigate the
45 opportunities to fortify the network security where needed effect of a malware or worm while continuing to supply the
46 [37]. A model-based IDS for home area networks (HAN) services to the customers.
47 is suggested in [131] by identifying the security challenges In order to mitigate the detrimental consequences of an
48 in HAN first and determining next how a Bayesian network adversarial cyber attack, one first step is to identify the
49 intrusion detection system can be used in future HANs. attack itself. Classifying the attack and giving it an identity
50 In order to determine the IDS requirements for HAN, allows operators to understand what they are dealing with. In
51 examination of the existing types of IDS is needed; there are [62], different machine learning classifying algorithms were
52 signature IDS, anomaly-based IDS, and specification-based tested in order to determine the viability of using machine
53 IDS, as in the following: learning as a decision support for system operators; the
54 • Signature-based IDS usually has a database of results demonstrated in Figure 9 show that it is a viable
55 predetermined attack patterns, known as signatures, and approach but more research is needed for deployment in an
56 detects the intrusions by comparing the system behavior operational environment and practical settings.
57 with these signatures.
58
VOLUME xx, 20xx 11
59 For Review Only
60
1
2 VI. CONCLUSION [7] A. Giani et al. “Smart grid data integrity attacks:
3 This paper offers a detailed and comprehensive description characterizations and countermeasures π”. In: 2011
4 of the links between the adversarial cyber attacks and IEEE International Conference on Smart Grid
5 power grid resilience, off-the-shelf cyber intrusion detection Communications (SmartGridComm). IEEE. 2011,
6 techniques, and what systems are or could be in place to pp. 232–237.
7 protect the smart power grids against malicious cyber attacks. [8] M. Esmalifalak et al. “Detecting stealthy false data
8 The mechanisms through which cyber attacks can impact the injection using machine learning in smart grid”.
9 bulk power grid are reviewed to understand where and how to In: IEEE Systems Journal vol. 11, no. 3 (2014),
10 enhance and reinforce countermeasures to mitigate the attack pp. 1644–1652.
11 consequences. Although there is a variety of cyber detection [9] A. Teixeira et al. “Cyber security analysis of state
12 and protection methods already in place, this review estimators in electric power systems”. In: 49th IEEE
13 highlighted the importance of considering cyber attacks in Conference on Decision and Control (CDC). Dec.
14 planning for resilience in power grids: strategies that entail 2010, pp. 5991–5998. DOI: 10 . 1109 / CDC . 2010 .
15 both grid hardening practices for structural resilience as well 5717318.
16 as procedures for operational resilience; this is due to the [10] A. Estebsari et al. “Techno-economic impacts
17 recently more-frequent realization of emerging threats with of automatic undervoltage load shedding under
18 no or very few similarities to those formerly-experienced emergency”. In: Electric Power Systems Research
19 incidents. While there might be found additional methods vol. 131 (2016), pp. 168–177. ISSN: 0378-7796. DOI:
20 for detection, protection and mitigation against cyber attacks https://doi.org/10.1016/j.epsr.2015.10.016. URL:
21 than those listed in this review and there will certainly http : / / www. sciencedirect . com / science / article / pii /
22 be new schemes and measures in the future, this survey S0378779615003120.
23 aimed to collect the state-of-the-art already-investigated or [11] A. Ashok et al. “Experimental evaluation of cyber
24 implemented solutions to provide a basis for future research attacks on Automatic Generation Control using a
25 and developments. Implementation of these various methods CPS Security Testbed”. In: 2015 IEEE Power Energy
26 on testbeds and real-world environments will finally allow Society General Meeting. July 2015, pp. 1–5. DOI:
27 for improvements in monitoring, protection, mitigation, and 10.1109/PESGM.2015.7286615.
28 resilience of the smart power grids against the looming [12] Y. Yuan et al. “Resilient control of cyber-physical
29 threats of cyber adversaries. systems against denial-of-service attacks”. In: 2013
30 6th International Symposium on Resilient Control
31 References Systems (ISRCS). IEEE. 2013, pp. 54–59.
32 [1] Y. Mo et al. “Cyber–physical security of a smart grid [13] G. Liang et al. “A Review of False Data Injection
33 infrastructure”. In: Proceedings of the IEEE vol. 100, Attacks Against Modern Power Systems”. In: IEEE
34 no. 1 (2011), pp. 195–209. Transactions on Smart Grid vol. 8, no. 4 (July 2017),
35 [2] A. Sanjab and W. Saad. “Power System Analysis: pp. 1630–1638. ISSN: 1949-3053. DOI: 10 . 1109 /
36 Competitive Markets, Demand Management, and TSG.2015.2495133.
37 Security”. In: Handbook of Dynamic Game Theory [14] O. Kosut et al. “Malicious data attacks on
38 (2017), pp. 1–38. smart grid state estimation: Attack strategies and
39 [3] S. Cui et al. “Coordinated data-injection attack countermeasures”. In: 2010 First IEEE International
40 and detection in the smart grid: A detailed Conference on Smart Grid Communications. IEEE.
41 look at enriching detection solutions”. In: IEEE 2010, pp. 220–225.
42 Signal Processing Magazine vol. 29, no. 5 (2012), [15] R. Jiang et al. “Energy-theft detection issues for
43 pp. 106–115. advanced metering infrastructure in smart grid”. In:
44 [4] J. Pacheco and S. Hariri. “IoT security framework Tsinghua Science and Technology vol. 19, no. 2 (Apr.
45 for smart cyber infrastructures”. In: 2016 IEEE 2014), pp. 105–120. ISSN: 1007-0214. DOI: 10.1109/
46 1st International Workshops on Foundations and TST.2014.6787363.
47 Applications of Self* Systems (FAS* W). IEEE. 2016, [16] S. McLaughlin et al. “AMIDS: A multi-sensor
48 pp. 242–247. energy theft detection framework for advanced
49 [5] A. Ashok and M. Govindarasu. “Cyber attacks on metering infrastructures.” In: SmartGridComm.
50 power system state estimation through topology 2012, pp. 354–359.
51 errors”. In: 2012 IEEE Power and Energy Society [17] P. Jokar; N. Arianpoo, and V. C. Leung. “Electricity
52 General Meeting. July 2012, pp. 1–8. DOI: 10.1109/ theft detection in AMI using customers’ consumption
53 PESGM.2012.6345091. patterns”. In: IEEE Transactions on Smart Grid
54 [6] J. Depoy et al. “Risk assessment for physical vol. 7, no. 1 (2015), pp. 216–226.
55 and cyber attacks on critical infrastructures”. In: [18] D. U. Case. “Analysis of the cyber attack on the
56 MILCOM 2005-2005 IEEE Military Communications Ukrainian power grid”. In: Electricity Information
57 Conference. IEEE. 2005, pp. 1961–1969. Sharing and Analysis Center (E-ISAC) (2016).
58
12 VOLUME xx, 20xx
59 For Review Only
60
1
2 [19] J. Hong; C.-C. Liu, and M. Govindarasu. “Detection [31] F. Pasqualetti; F. Dörfler, and F. Bullo. “Attack
3 of cyber intrusions using network-based multicast detection and identification in cyber-physical
4 messages for substation automation”. In: ISGT 2014. systems”. In: IEEE transactions on automatic control
5 IEEE. 2014, pp. 1–5. vol. 58, no. 11 (2013), pp. 2715–2729.
6 [20] A. Sanjab; W. Saad, and T. Başar. “Graph-Theoretic [32] A. Gholami; F. Aminifar, and M. Shahidehpour.
7 Framework for Unified Analysis of Observability and “Front lines against the darkness: Enhancing the
8 Data Injection Attacks in the Smart Grid”. In: arXiv resilience of the electricity grid through microgrid
9 preprint arXiv:1801.08951 (2018). facilities”. In: IEEE Electrification Magazine vol. 4,
10 [21] Q. Zhu and T. Başar. “Robust and resilient control no. 1 (2016), pp. 18–24.
11 design for cyber-physical systems with an application [33] S. M. Amin and B. F. Wollenberg. “Toward a smart
12 to power systems”. In: 2011 50th IEEE Conference grid: power delivery for the 21st century”. In: IEEE
13 on Decision and Control and European Control power and energy magazine vol. 3, no. 5 (2005),
14 Conference. IEEE. 2011, pp. 4066–4071. pp. 34–41.
15 [22] F. Pasqualetti; F. Dörfler, and F. Bullo. “Cyber- [34] C. Kwon; W. Liu, and I. Hwang. “Security analysis
16 physical attacks in power networks: Models, for cyber-physical systems against stealthy deception
17 fundamental limitations and monitor design”. In: attacks”. In: 2013 American control conference.
18 2011 50th IEEE Conference on Decision and Control IEEE. 2013, pp. 3344–3349.
19 and European Control Conference. IEEE. 2011, [35] C.-W. Ten; G. Manimaran, and C.-C. Liu.
20 pp. 2195–2201. “Cybersecurity for critical infrastructures: Attack
21 [23] A. Ferdowsi et al. “Game theory for secure critical and defense modeling”. In: IEEE Transactions on
22 interdependent gas-power-water infrastructure”. Systems, Man, and Cybernetics-Part A: Systems and
23 In: 2017 Resilience Week (RWS). IEEE. 2017, Humans vol. 40, no. 4 (2010), pp. 853–865.
24 pp. 184–190. [36] S. Zonouz et al. “SOCCA: A security-oriented
25 [24] G. El Rahi et al. “Prospect theory for enhanced smart cyber-physical contingency analysis in power
26 grid resilience using distributed energy storage”. infrastructures”. In: IEEE Transactions on Smart
27 In: 2016 54th Annual Allerton Conference on Grid vol. 5, no. 1 (2013), pp. 3–13.
28 Communication, Control, and Computing (Allerton). [37] D. C. de Leon et al. “Modeling complex control
29 IEEE. 2016, pp. 248–255. systems to identify remotely accessible devices
30 [25] Y. Liu; P. Ning, and M. K. Reiter. “False data vulnerable to cyber attack”. In: ACM Workshop
31 injection attacks against state estimation in electric on Scientific Aspects of Cyber Terrorism,(SACT).
32 power grids”. In: ACM Transactions on Information Citeseer. 2002.
33 and System Security (TISSEC) vol. 14, no. 1 (2011), [38] Y. Zhang et al. “Distributed intrusion detection
34 p. 13. system in a multi-layer network architecture of smart
35 [26] A. Sanjab and W. Saad. “Data injection attacks grids”. In: IEEE Transactions on Smart Grid vol. 2,
36 on smart grids with multiple adversaries: A game- no. 4 (2011), pp. 796–808.
37 theoretic perspective”. In: IEEE Transactions on [39] S. Han et al. “Intrusion Detection in Cyber-
38 Smart Grid vol. 7, no. 4 (2016), pp. 2038–2049. Physical Systems: Techniques and Challenges”. In:
39 [27] L. Xie; Y. Mo, and B. Sinopoli. “Integrity data attacks IEEE Systems Journal vol. 8, no. 4 (Dec. 2014),
40 in power market operations”. In: IEEE Transactions pp. 1052–1062. ISSN: 1932-8184. DOI: 10 . 1109 /
41 on Smart Grid vol. 2, no. 4 (2011), pp. 659–666. JSYST.2013.2257594.
42 [28] A. Sanjab and W. Saad. “Smart grid data [40] S. Mousavian et al. “A risk-based optimization model
43 injection attacks: To defend or not?” In: 2015 for electric vehicle infrastructure response to cyber
44 IEEE International Conference on Smart Grid attacks”. In: IEEE Transactions on Smart Grid vol. 9,
45 Communications (SmartGridComm). IEEE. 2015, no. 6 (2017), pp. 6160–6169.
46 pp. 380–385. [41] N. Chand et al. “A comparative analysis of
47 [29] Y. W. Law; T. Alpcan, and M. Palaniswami. SVM and its stacking with other classification
48 “Security games for risk minimization in automatic algorithm for intrusion detection”. In: 2016
49 generation control”. In: IEEE Transactions on Power International Conference on Advances in Computing,
50 Systems vol. 30, no. 1 (2014), pp. 223–232. Communication, & Automation (ICACCA)(Spring).
51 [30] A. Sanjab and W. Saad. “On bounded rationality IEEE. 2016, pp. 1–6.
52 in cyber-physical systems security: Game-theoretic [42] J.-W. Kang; I.-Y. Joo, and D.-H. Choi. “False data
53 analysis with application to smart grid protection”. injection attacks on contingency analysis: Attack
54 In: 2016 Joint Workshop on Cyber-Physical Security strategies and impact assessment”. In: IEEE Access
55 and Resilience in Smart Grids (CPSR-SG). IEEE. vol. 6 (2018), pp. 8841–8851.
56 2016, pp. 1–6. [43] B. Wang et al. “Electrical Safety Considerations in
57 Large Electric Vehicle Charging Stations”. In: IEEE
58
VOLUME xx, 20xx 13
59 For Review Only
60
1
2 Transactions on Industry Applications vol. 55, no. 6 [54] T. Becejac; P. Dehghanian, and M. Kezunovic.
3 (2019), pp. 6603–6612. “Probabilistic Assessment of PMU Integrity for
4 [44] K. L. Morrow et al. “Topology perturbation Planning of Periodic Maintenance and Testing”. In:
5 for detecting malicious data injection”. In: 2012 International Conference on Probabilistic Methods
6 45th Hawaii International Conference on System Applied to Power Systems (PMAPS). 2016, pp. 1–6.
7 Sciences. IEEE. 2012, pp. 2104–2113. [55] M. Kezunovic; P. Dehghanian, and J. Sztipanovits.
8 [45] B. Wang et al. “New Reward and Penalty Scheme “An Incremental System-of-Systems Integration
9 for Electric Distribution Utilities Employing Load- Modelling of Cyber-Physical Electric Power
10 Based Reliability Indices”. In: IET Generation, Systems”. In: Grid of the Future Symposium, CIGRE
11 Transmission & Distribution vol. 12, no. 15 (2018), US National Committee. CIGRE. 2016, pp. 1–6.
12 pp. 3647–3654. [56] M. H. Rezaeian Koochi et al. “A Synchrophasor-
13 [46] T. Becejac and P. Dehghanian. “PMU Multilevel based Decision Tree Approach for Identification
14 End-to-End Testing to Assess Synchrophasor of Most Coherent Generating Units”. In: The
15 Measurements during Faults”. In: IEEE Power and 44th Annual Conference of the IEEE Industrial
16 Energy Technology Systems Journal vol. 6, no. 1 Electronics Society (IECON). 2018, pp. 1–6.
17 (Mar. 2019), pp. 71–80. [57] S. Krishnamurthy; S. Sarkar, and A. Tewari.
18 [47] A. Razi-Kazemi and P. Dehghanian. “A Practical “Scalable anomaly detection and isolation in cyber-
19 Approach to Optimal RTU Placement in Power physical systems using bayesian networks”. In:
20 Distribution Systems Incorporating Fuzzy Sets Proceedings of asme dynamical systems and control
21 Theory”. In: International Journal of Electrical conference, san antonio, tx, usa. 2014.
22 Power and Energy Systems vol. 37, no. 1 (2012), [58] S. Wang; P. Dehghanian, and B. Zhang. “A Data-
23 pp. 31–42. Driven Algorithm for Online Power Grid Topology
24 [48] P. Dehghanian; A. Razi-Kazemi, and M. Fotuhi- Change Identification with PMUs”. In: IEEE Power
25 Firuzabad. “Optimal RTU Placement in Power and Energy Society (PES) General Meeting. 2019,
26 Distribution Systems Using a Novel Method pp. 1–5.
27 Based on Analytical Hierarchical Process (AHP)”. [59] S. Wang; P. Dehghanian, and Y. Gu. “A Novel
28 In: The 10th International IEEE Conference on Multi-Resolution Wavelet Transform for Online
29 Environmental and Electrical Engineering (EEEIC). Power Grid Waveform Classification”. In: The
30 2011, pp. 1–6. 1st IEEE International Conference on Smart
31 [49] M. Moeini-Aghtaie; P. Dehghanian, and S. H. Grid Synchronized Measurements and Analytics
32 Hosseini. “Optimal Distributed Generation Placement (SGSMA). 2019, pp. 1–6.
33 in a Restructured Environment via a Multi-Objective [60] T. Becejac; P. Dehghanian, and M. Kezunovic.
34 Optimization Approach”. In: 16th Conference on “Analysis of PMU Algorithm Errors During Fault
35 Electric Power Distribution Networks (EPDC). 2011, Transients and Out-of-Step Disturbances”. In: IEEE
36 pp. 1–6. Power and Energy Society (PES) Transmission &
37 [50] A. Razi-Kazemi; P. Dehghanian, and G. Karami. “A Distribution (T&D) Conference and Exposition Latin
38 Probabilistic Approach for Remote Terminal Unit America. 2016, pp. 1–6.
39 Placement in Power Distribution Systems”. In: The [61] T. Becejac; P. Dehghanian, and M. Kezunovic.
40 33rd IEEE International Telecommunications Energy “Impact of PMU Errors on the Synchrophasor-based
41 Conference (INTELEC). 2011, pp. 1–6. Fault Location Algorithms”. In: 48th North American
42 [51] P. Dehghanian; A. Razi-Kazemi, and G. Karami. Power Symposium (NAPS). 2016, pp. 1–6.
43 “Incorporating Experts Knowledge in RTU Placement [62] R. C. B. Hink et al. “Machine learning for power
44 Procedure Using Fuzzy Sets Theory- A Practical system disturbance and cyber-attack discrimination”.
45 Approach”. In: The 33rd IEEE International In: 2014 7th international symposium on resilient
46 Telecommunications Energy Conference (INTELEC). control systems (ISRCS). IEEE. 2014, pp. 1–8.
47 2011, pp. 1–6. [63] K. Ly et al. “Approximate Power Grid Protection
48 [52] M. Shojaei et al. “A New Look on the Automation of Against False Data Injection Attacks”. In: 2017 IEEE
49 Medium Voltage Substations in Power Distribution 15th Intl Conf on Dependable, Autonomic and Secure
50 Systems”. In: 17th Conference on Electric Power Computing, 15th Intl Conf on Pervasive Intelligence
51 Distribution Networks (EPDC). 2012, pp. 1–6. and Computing, 3rd Intl Conf on Big Data
52 [53] M. Kezunovic et al. “Life-Cycle Management Intelligence and Computing and Cyber Science and
53 Tools for Synchrophasor Systems: Why We Need Technology Congress(DASC/PiCom/DataCom/Cyber-
54 Them and What They Should Entail”. In: The SciTech). Nov. 2017, pp. 527–533. DOI: 10 . 1109 /
55 2016 IFAC CIGRE/CIRED Workshop on Control of DASC-PICom-DataCom-CyberSciTec.2017.97.
56 Transmission and Distribution Smart Grids. CIGRE. [64] C. Ten; C. Liu, and G. Manimaran. “Vulnerability
57 2016, pp. 1–6. assessment of cybersecurity for SCADA systems”.
58
14 VOLUME xx, 20xx
59 For Review Only
60
1
2 In: IEEE Transactions on Power Systems vol. 23, [77] P. Dehghanian; S. Aslan, and P. Dehghanian.
3 no. 4 (2008), pp. 1836–1846. “Maintaining electric system safety through an
4 [65] L. Jia; R. J. Thomas, and L. Tong. “On the enhanced network resilience”. In: IEEE Transactions
5 nonlinearity effects on malicious data attack on on Industry Applications vol. 54, no. 5 (2018),
6 power system”. In: 2012 IEEE Power and Energy pp. 4927–4937.
7 Society General Meeting. July 2012, pp. 1–8. DOI: [78] B. Zhang; P. Dehghanian, and M. Kezunovic.
8 10.1109/PESGM.2012.6345685. “Optimal allocation of PV generation and battery
9 [66] K. Chatterjee; V. Padmini, and S. A. Khaparde. storage for enhanced resilience”. In: IEEE Transactions
10 “Review of cyber attacks on power system on Smart Grid vol. 10, no. 1 (2017), pp. 535–545.
11 operations”. In: 2017 IEEE Region 10 Symposium [79] M. Nazemi et al. “Energy Storage Planning
12 (TENSYMP). July 2017, pp. 1–6. DOI: 10 . 1109 / for Enhanced Resilience of Power Distribution
13 TENCONSpring.2017.8070085. Networks against Earthquakes”. In: IEEE Transactions
14 [67] P. Chen et al. “Detection of false data injection on Sustainable Energy vol. 11, no. 2 (Apr. 2020),
15 attacks in smart-grid systems”. In: IEEE Communications pp. 795–806. ISSN: 1949-3029. DOI: 10.1109/TSTE.
16 Magazine vol. 53, no. 2 (Feb. 2015), pp. 206–213. 2019.2907613.
17 ISSN : 0163-6804. DOI : 10 . 1109 / MCOM . 2015 . [80] M. Amirioun et al. “Metrics and quantitative
18 7045410. framework for assessing microgrid resilience
19 [68] M. Moeini-Aghtaie et al. “PHEV’s Centralized/Decentralized against windstorms”. In: International Journal of
20 Charging Control Mechanisms: Requirements and Electrical Power & Energy Systems vol. 104 (2019),
21 Impacts”. In: The 45th North American Power pp. 716–723.
22 Symposium (NAPS). 2013, pp. 1–6. [81] S. Wang et al. “Resilience-Assured Protective
23 [69] B. Wang; P. Dehghanian, and D. Zhao. “Chance- Control of DC/AC Inverters under Unbalanced and
24 Constrained Energy Management System for Power Fault Scenarios”. In: The 10th IEEE Power and
25 Grids with High Proliferation of Renewables and Energy Society (PES) Conference on Innovative
26 Electric Vehicles”. In: IEEE Transactions on Smart Smart Grid Technologies-North America (ISGT-NA).
27 Grid (2020), pp. 1–13. 2019, pp. 1–5.
28 [70] B. Wang et al. “Aggregated Electric Vehicle Load [82] Z. Yang; P. Dehghanian, and M. Nazemi. “Enhancing
29 Modeling in Large-Scale Electric Power Systems”. Seismic Resilience of Electric Power Distribution
30 In: IEEE Transactions on Industry Applications Systems with Mobile Power Sources”. In: IEEE
31 (2020), pp. 1–14. Industry Applications Society (IAS) Annual Meeting.
32 [71] P. Jamborsalamati et al. “Enhancing Power Grid 2019, pp. 1–7.
33 Resilience through An IEC61850-based EV-Assisted [83] S. Wang et al. “A Machine Learning Approach
34 Load Restoration”. In: IEEE Transactions on to Detection of Geomagnetically-Induced Currents
35 Industrial Informatics vol. 16, no. 3 (Mar. 2020), in Power Grids”. In: IEEE Industry Applications
36 pp. 1799–1810. Society (IAS) Annual Meeting. 2019, pp. 1–7.
37 [72] B. Wang et al. “Adaptive Operation Strategies [84] M. Babakmehr et al. “Sparse Representation-
38 for Electric Vehicle Charging Stations”. In: IEEE based Classification of Geomagnetically Induced
39 Industry Applications Society (IAS) Annual Meeting. Currents”. In: Clemson University Power Systems
40 2019, pp. 1–7. Conference. 2020, pp. 1–6.
41 [73] M. A. Saffari et al. “Robust/Stochastic Optimization [85] Z. Yang et al. “Toward Resilient Solar-Integrated
42 of Energy Arbitrage in Smart Microgrids using Distribution Grids: Harnessing the Mobility of Power
43 Electric Vehicles”. In: Electric Power Systems Sources”. In: IEEE Power and Energy Society (PES)
44 Research vol. 174 (Sept. 2019), pp. 1–14. Transmission and Distribution (T&D) Conference
45 [74] M. Moeini-Aghtaie et al. “Optimized Probabilistic and Exposition. 2020, pp. 1–5.
46 PHEV Demand Management in the Context of [86] M. Nazemi and P. Dehghanian. “Seismic-Resilient
47 Energy Hubs”. In: IEEE Transactions on Power Bulk Power Grids: Hazard Characterization,
48 Delivery vol. 30, no. 2 (2015), pp. 996–1006. Modeling, and Mitigation”. In: IEEE Transactions
49 [75] M. S. Misaghian et al. “Electric Vehicles Contributions on Engineering Management (2020), pp. 1–17.
50 to Voltage Improvement and Loss Reduction in [87] S. Wang et al. “A Machine Learning Approach
51 Microgrids”. In: North American Power Symposium to Detection of Geomagnetically-Induced Currents
52 (NAPS). 2018, pp. 1–6. in Power Grids”. In: IEEE Transactions on
53 [76] P. Dehghanian et al. “Predictive risk analytics Industry Applications vol. 56, no. 2 (Mar. 2020),
54 for weather-resilient operation of electric power pp. 1098–1106.
55 systems”. In: IEEE Transactions on Sustainable [88] Z. Yang; P. Dehghanian, and M. Nazemi. “Seismic-
56 Energy vol. 10, no. 1 (2019), pp. 3–15. Resilient Electric Power Distribution Systems:
57 Harnessing the Mobility of Power Sources”. In:
58
VOLUME xx, 20xx 15
59 For Review Only
60
1
2 IEEE Transactions on Industry Applications (2020), for Sustainable Development vol. 55, no. 2 (2013),
3 pp. 1–10. pp. 25–29.
4 [89] M. Nazemi et al. “Multivariate Uncertainty [102] B. Zhu and S. Sastry. “SCADA-specific intrusion
5 Characterization for Resilience Planning in Electric detection/prevention systems: a survey and taxonomy”.
6 Power Systems”. In: IEEE/IAS 56th Industrial and In: Proceedings of the 1st workshop on secure control
7 Commercial Power Systems (I&CPS) Technical systems (SCS). Vol. 11. 2010, p. 7.
8 Conference. 2020, pp. 1–7. [103] A. Giani et al. “The VIKING project: An initiative
9 [90] B. Zhang; P. Dehghanian, and M. Kezunovic. on resilient control of power networks”. In: 2009
10 “Simulation of Weather Impacts on the Wholesale 2nd International Symposium on Resilient Control
11 Electricity Market”. In: 10th International Conference Systems. Aug. 2009, pp. 31–35. DOI: 10 . 1109 /
12 on Deregulated Electricity Market Issues in South ISRCS.2009.5251361.
13 Eastern Europe (DEMSEE). 2015, pp. 1–6. [104] S. Wang; P. Dehghanian, and Y. Gu. “A Novel
14 [91] T. Dokic et al. “Risk Assessment of a Transmission Multi-Resolution Wavelet Transform for Online
15 Line Insulation Breakdown due to Lightning and Power Grid Waveform Classification”. In: The
16 Severe Weather”. In: The 49th Hawaii International 1st IEEE International Conference on Smart
17 Conference on System Science (HICSS). 2016, Grid Synchronized Measurements and Analytics
18 pp. 1–8. (SGSMA). 2019, pp. 1–6.
19 [92] P. Dehghanian; S. Aslan, and P. Dehghanian. [105] S. Wang; L. Li, and P. Dehghanian. “Power
20 “Quantifying Power System Resiliency Improvement Grid Online Surveillance through PMU-Embedded
21 using Network Reconfiguration”. In: IEEE 60th Convolutional Neural Networks”. In: IEEE Industry
22 International Midwest Symposium on Circuits and Applications Society (IAS) Annual Meeting. 2019,
23 Systems (MWSCAS). 2017, pp. 1–5. pp. 1–7.
24 [93] J. Su et al. “Distributed Wind Power Resources [106] S. Wang; P. Dehghanian, and L. Li. “Power
25 for Enhanced Power Grid Resilience”. In: The 51st Grid Online Surveillance through PMU-Embedded
26 North American Power Symposium (NAPS). 2019, Convolutional Neural Networks”. In: IEEE Transactions
27 pp. 1–6. on Industry Applications vol. 56, no. 2 (Mar. 2020),
28 [94] D. Wang et al. “Power Grid Resilience to pp. 1146–1155.
29 Electromagnetic (EMP) Disturbances: A Literature [107] P. Dehghanian; Y. Guan, and M. Kezunovic. “Real-
30 Review”. In: The 51st North American Power Time Life-Cycle Assessment of High Voltage Circuit
31 Symposium (NAPS). 2019, pp. 1–6. Breakers for Maintenance using Online Condition
32 [95] M. Babakmehr et al. “Artificial Intelligence-Based Monitoring Data”. In: IEEE Transactions on Industry
33 Cyber-Physical Event Classification for Islanding Applications vol. 55, no. 2 (2019), pp. 1135–1146.
34 Detection in Power Inverters”. In: IEEE Journal of [108] P. Dehghanian et al. “Security-Based Circuit Breaker
35 Emerging and Selected Topics in Power Electronics Maintenance Management”. In: IEEE Power and
36 (2020), pp. 1–11. Energy Society (PES) General Meeting. 2013,
37 [96] S. Wang et al. “Advanced Control Solutions for pp. 1–5.
38 Enhanced Resilience of Modern Power-Electronic- [109] Y. Guan et al. “Assessing Circuit Breaker Life
39 Interfaced Distribution Systems”. In: Journal of Cycle using Condition-based Data”. In: IEEE Power
40 Modern Power Systems and Clean Energy vol. 7, and Energy Society (PES) General Meeting. 2013,
41 no. 4 (July 2019), pp. 716–730. pp. 1–5.
42 [97] B. Shinde et al. “Real-Time Detection of Critical [110] P. Dehghanian and M. Kezunovic. “Cost/Benefit
43 Generators in Power Systems: A Deep Learning Analysis for Circuit Breaker Maintenance Planning
44 HCP Approach”. In: The 4th IEEE Texas Power and and Scheduling”. In: The 45th North American
45 Energy Conference (TPEC). 2020, pp. 1–6. Power Symposium (NAPS). 2013, pp. 1–6.
46 [98] A. Berkeley; M. Wallace, and C. COO. “A [111] P. Dehghanian; Y. Guan, and M. Kezunovic. “Real-
47 framework for establishing critical infrastructure Time Life-Cycle Assessment of Circuit Breakers
48 resilience goals”. In: Final Report and Recommendations for Maintenance using Online Condition Monitoring
49 by the Council, National Infrastructure Advisory Data”. In: IEEE/IAS 54th Industrial and Commercial
50 Council (2010). Power Systems (I&CPS) Technical Conference.
51 [99] Y. Y. Haimes. “On the definition of resilience in 2018, pp. 1–8.
52 systems”. In: Risk Analysis: An International Journal [112] P. Dehghanian; T. Popovic, and M. Kezunovic.
53 vol. 29, no. 4 (2009), pp. 498–501. “Circuit Breaker Operational Health Assessment via
54 [100] W. House. Critical infrastructure security and Condition Monitoring Data”. In: The 46th North
55 resilience. White House, 2013. American Power Symposium. 2014, pp. 1–6.
56 [101] S. L. Cutter et al. “Disaster resilience: A national [113] S. Moradi et al. “A Mathematical Framework for
57 imperative”. In: Environment: Science and Policy Reliability-Centered Asset Management Implementation
58
16 VOLUME xx, 20xx
59 For Review Only
60
1
2 in Microgrids”. In: International Transactions on Conference and Exhibition on Electricity Distribution
3 Electrical Energy Systems (2018). (CIRED). 2011, pp. 1–4.
4 [114] H. Mirsaeedi et al. “Long-Term Maintenance [125] P. Dehghanian and M. Fotuhi-Firuzabad. “A
5 Scheduling and Budgeting in Electricity Distribution Reliability-Oriented Outlook on the Critical Components
6 Systems Equipped with Automatic Switches”. In: of Power Distribution Systems”. In: The 9th
7 IEEE Transactions on Industrial Informatics vol. 14, IET International Conference on Advances in
8 no. 5 (2018), pp. 1909–1919. Power System Control, Operation, and Management
9 [115] M. Asghari Gharakheili; M. Fotuhi-Firuzabad, and (APSCOM). 2011, pp. 1–6.
10 P. Dehghanian. “A New Multi-Attribute Support [126] P. Dehghanian et al. “A Practical Application of the
11 Tool for Identifying Critical Components in Power Delphi Method in Maintenance-Targeted Resource
12 Transmission Systems”. In: IEEE Systems Journal Allocation of Distribution Utilities”. In: The 13th
13 vol. 12, no. 1 (2018), pp. 316–327. International Conference on Probabilistic Methods
14 [116] F. Pourahmadi; M. Fotuhi-Firuzabad, and P. Applied to Power Systems (PMAPS). 2014, pp. 1–6.
15 Dehghanian. “Application of Game Theory in [127] F. Pourahmadi; M. Fotuhi-Firuzabad, and P.
16 Reliability Centered Maintenance of Electric Power Dehghanian. “Identification of Critical Components
17 Systems”. In: IEEE Transactions on Industry in Power Systems: A Game Theory Application”.
18 Applications vol. 53, no. 2 (2017), pp. 936–946. In: IEEE Industry Application Society (IAS) Annual
19 [117] F. Pourahmadi; M. Fotuhi-Firuzabad, and P. Meeting. 2016, pp. 1–6.
20 Dehghanian. “Identification of Critical Generating [128] J. Lai et al. “Broadcast Gossip Algorithms for
21 Units for Maintenance: A Game Theory Approach”. Distributed Peer-to-Peer Control in AC Microgrids”.
22 In: IET Generation, Transmission & Distribution In: IEEE Transactions on Industry Applications
23 vol. 10, no. 12 (2016), pp. 2942–2952. vol. 55, no. 3 (May 2019), pp. 2241–2251.
24 [118] H. Sabouhi et al. “Identifying Critical Components of [129] F. Pourahmadi et al. “Dynamic Uncertainty Set
25 Combined Cycle Power Plants for Implementation of Characterization for Bulk Power Grid Flexibility
26 Reliability Centered Maintenance”. In: IEEE CSEE Assessment”. In: IEEE Systems Journal vol. 14, no. 1
27 Journal of Power and Energy Systems vol. 2, no. 2 (Mar. 2020), pp. 718–728.
28 (2016), pp. 87–97. [130] M. Khoshjahan et al. “Harnessing Ramp Capability
29 [119] H. Sabouhi et al. “Reliability Modeling and of Spinning Reserve Services for Enhanced Power
30 Availability Analysis of Combined Cycle Power System Flexibility”. In: IEEE Transactions on
31 Plants”. In: International Journal of Electrical Power Industry Applications vol. 55, no. 6 (Nov. 2019),
32 and Energy Systems vol. 79 (2016), pp. 108–119. pp. 7103–7112.
33 [120] R. Ghorani et al. “Identifying Critical Component [131] P. Jokar. “Model-based intrusion detection for home
34 for Reliability Centered Maintenance Management area networks in smart grids”. In: University of
35 of Deregulated Power Systems”. In: IET Generation, Bristol, Bristol (2012), pp. 1–19.
36 Transmission, and Distribution vol. 9, no. 9 (2015), [132] D. I. Dogaru and I. Dumitrache. “Robustness of
37 pp. 828–837. Power Systems in the Context of Cyber Attacks”.
38 [121] P. Dehghanian et al. “A Comprehensive Scheme for In: 2017 21st International Conference on Control
39 Reliability Centered Maintenance Implementation Systems and Computer Science (CSCS). IEEE. 2017,
40 in Power Distribution Systems- Part II: Numerical pp. 506–512.
41 Analysis”. In: IEEE Transactions on Power Delivery [133] I. Dumitrache and D. I. Dogaru. “Smart grid
42 vol. 28, no. 2 (2013), pp. 771–778. overview: infrastructure, cyber-physical security and
43 [122] P. Dehghanian et al. “A Comprehensive Scheme for challenges”. In: 2015 20th International Conference
44 Reliability Centered Maintenance Implementation in on Control Systems and Computer Science. IEEE.
45 Power Distribution Systems- Part I: Methodology”. 2015, pp. 693–699.
46 In: IEEE Transactions on Power Delivery vol. 28, [134] M. Wang et al. “Identification of “unobservable”
47 no. 2 (2013), pp. 761–770. cyber data attacks on power grids”. In: 2014
48 [123] P. Dehghanian et al. “Critical Component Identification IEEE International Conference on Smart Grid
49 in Reliability Centered Asset Management of Communications (SmartGridComm). Nov. 2014,
50 Distribution Power Systems via Fuzzy AHP”. pp. 830–835. DOI: 10.1109/SmartGridComm.2014.
51 In: IEEE Systems Journal vol. 6, no. 4 (2012), 7007751.
52 pp. 593–602. [135] H. Nishino and H. Ishii. “Distributed detection of
53 [124] P. Dehghanian; M. Fotuhi-Firuzabad, and A. Razi- cyber attacks and faults for power systems”. In:
54 Kazemi. “An Approach for Critical Component IFAC Proceedings Volumes vol. 47, no. 3 (2014),
55 Identification in Reliability-Centered Maintenance pp. 11932–11937.
56 of Power Distribution Systems Based on Analytical [136] F. Vosgerau et al. “Power system state and topology
57 Hierarchical Process”. In: The 21st International coestimation”. In: 2010 IREP Symposium Bulk
58
VOLUME xx, 20xx 17
59 For Review Only
60
1
2 Power System Dynamics and Control - VIII (IREP). Transactions on Power Delivery vol. 32, no. 1 (2016),
3 2010, pp. 1–6. pp. 504–514.
4 [137] S. T. Smith et al. “Bayesian Discovery of [150] C. Murphy and A. Keane. “Local and remote
5 Threat Networks”. In: IEEE Transactions on estimations using fitted polynomials in distribution
6 Signal Processing vol. 62, no. 20 (Oct. 2014), systems”. In: IEEE Transactions on Power Systems
7 pp. 5324–5338. ISSN: 1053-587X. DOI: 10 . 1109 / vol. 32, no. 4 (2016), pp. 3185–3194.
8 TSP.2014.2336613. [151] S. Bi and Y. J. A. Zhang. “Graph-based cyber security
9 [138] A. Teixeira; H. Sandberg, and K. H. Johansson. analysis of state estimation in smart power grid”. In:
10 “Networked control systems under cyber attacks with IEEE Communications Magazine (2017).
11 applications to power networks”. In: Proceedings of [152] J. Zhao et al. “Short-term state forecasting-aided
12 the 2010 American Control Conference. June 2010, method for detection of smart grid general false data
13 pp. 3690–3696. DOI: 10.1109/ACC.2010.5530638. injection attacks”. In: IEEE Transactions on Smart
14 [139] L. Liu et al. “Detecting false data injection attacks Grid vol. 8, no. 4 (2015), pp. 1580–1590.
15 on power grid by sparse optimization”. In: IEEE [153] Y. Weng et al. “Robust data-driven state estimation
16 Transactions on Smart Grid vol. 5, no. 2 (2014), for smart grid”. In: IEEE Transactions on Smart Grid
17 pp. 612–621. vol. 8, no. 4 (2016), pp. 1956–1967.
18 [140] Y. Li and Y. Wang. “False data injection attacks with [154] S. Maharjan et al. “Improved sample value
19 incomplete network topology information in smart adjustment for synchrophasor estimation at off-
20 grid”. In: IEEE Access vol. 7 (2018), pp. 3656–3664. nominal power system conditions”. In: IEEE
21 [141] D. Xue; X. Jing, and H. Liu. “Detection of False Transactions on Power Delivery vol. 32, no. 1 (2016),
22 Data Injection Attacks in Smart Grid Utilizing ELM- pp. 33–44.
23 Based OCON Framework”. In: IEEE Access vol. 7 [155] M. A. Rahman and H. Mohsenian-Rad. “False
24 (2019), pp. 31762–31773. data injection attacks with incomplete information
25 [142] G. Chaojun; P. Jirutitijaroen, and M. Motani. against smart power grids”. In: 2012 IEEE Global
26 “Detecting false data injection attacks in ac state Communications Conference (GLOBECOM). IEEE.
27 estimation”. In: IEEE Transactions on Smart Grid 2012, pp. 3153–3158.
28 vol. 6, no. 5 (2015), pp. 2476–2483. [156] O. Kosut et al. “Malicious data attacks on the smart
29 [143] Y. Huang et al. “Real-time detection of false grid”. In: IEEE Transactions on Smart Grid vol. 2,
30 data injection in smart grid networks: An adaptive no. 4 (2011), pp. 645–658.
31 CUSUM method and analysis”. In: IEEE Systems [157] X. Liu et al. “Modeling of local false data injection
32 Journal vol. 10, no. 2 (2014), pp. 532–543. attacks with reduced network information”. In: IEEE
33 [144] L. Xie; Y. Mo, and B. Sinopoli. “False data Transactions on Smart Grid vol. 6, no. 4 (2015),
34 injection attacks in electricity markets”. In: 2010 pp. 1686–1696.
35 First IEEE International Conference on Smart Grid [158] Z.-H. Yu and W.-L. Chin. “Blind false data injection
36 Communications. IEEE. 2010, pp. 226–231. attack using PCA approximation method in smart
37 [145] O. Vuković and G. Dán. “Security of fully grid”. In: IEEE Transactions on Smart Grid vol. 6,
38 distributed power system state estimation: Detection no. 3 (2015), pp. 1219–1226.
39 and mitigation of data integrity attacks”. In: IEEE [159] X. Liu et al. “A collaborative intrusion detection
40 Journal on Selected Areas in Communications mechanism against false data injection attack
41 vol. 32, no. 7 (2014), pp. 1500–1508. in advanced metering infrastructure”. In: IEEE
42 [146] Y. Yamaguchi et al. “Cyber security analysis of Transactions on Smart Grid vol. 6, no. 5 (2015),
43 power networks by hypergraph cut algorithms”. In: pp. 2435–2443.
44 IEEE Transactions on Smart Grid vol. 6, no. 5 [160] P.-Y. Chen et al. “Detection of false data
45 (2015), pp. 2189–2199. injection attacks in smart-grid systems”. In: IEEE
46 [147] Y. Zhang et al. “Power system reliability evaluation Communications Magazine vol. 53, no. 2 (2015).
47 with SCADA cybersecurity considerations”. In: [161] S. Bi and Y. J. Zhang. “Graphical methods for
48 IEEE Transactions on Smart Grid vol. 6, no. 4 defense against false-data injection attacks on power
49 (2015), pp. 1707–1721. system state estimation”. In: IEEE Transactions on
50 [148] X. Liu and Z. Li. “Local load redistribution Smart Grid vol. 5, no. 3 (2014), pp. 1216–1227.
51 attacks in power systems with incomplete network [162] S. Bi and Y. J. Zhang. “Using covert topological
52 information”. In: IEEE Transactions on Smart Grid information for defense against malicious attacks on
53 vol. 5, no. 4 (2014), pp. 1665–1676. DC state estimation”. In: IEEE Journal on Selected
54 [149] S. Choi and A. S. Meliopoulos. “Effective real- Areas in Communications vol. 32, no. 7 (2014),
55 time operation and protection scheme of microgrids pp. 1471–1485.
56 using distributed dynamic state estimation”. In: IEEE [163] J. M. Hendrickx et al. “Efficient computations of
57 a security index for false data attacks in power
58
18 VOLUME xx, 20xx
59 For Review Only
60
1
2 networks”. In: IEEE Transactions on Automatic Intelligent Systems Applications to Power Systems
3 Control vol. 59, no. 12 (2014), pp. 3194–3208. (ISAP). 2015, pp. 1–6.
4 [164] H. Sedghi and E. Jonckheere. “Statistical structure [176] P. Dehghanian and M. Kezunovic. “Probabilistic
5 learning to ensure data integrity in smart grid”. In: Impact of Transmission Line Switching on Power
6 IEEE Transactions on Smart Grid vol. 6, no. 4 System Operating States”. In: IEEE Power and
7 (2015), pp. 1924–1933. Energy Society (PES) Transmission and Distribution
8 [165] H. Karimipour and V. Dinavahi. “Robust massively (T&D) Conference and Exposition. 2016, pp. 1–6.
9 parallel dynamic state estimation of power systems [177] M. Alhazmi et al. “Power Grid Optimal Topology
10 against cyber-attack”. In: IEEE Access vol. 6 (2017), Control Considering Correlations of System
11 pp. 2984–2995. Uncertainties”. In: IEEE Transactions on Industry
12 [166] K. Manandhar et al. “Detection of faults and attacks Applications vol. 55, no. 6 (Nov. 2019), pp. 5594–5604.
13 including false data injection attack in smart grid [178] M. Nazemi; P. Dehghanian, and M. Lejeune.
14 using Kalman filter”. In: IEEE transactions on “A Mixed-Integer Distributionally Robust Chance-
15 control of network systems vol. 1, no. 4 (2014), Constrained Model for Optimal Topology Control in
16 pp. 370–379. Power Grids with Uncertain Renewables”. In: 13th
17 [167] B. Li et al. “Distributed host-based collaborative IEEE Power and Energy Society (PES) PowerTech
18 detection for false data injection attacks in smart grid Conference. 2019, pp. 1–6.
19 cyber-physical system”. In: Journal of Parallel and [179] L. Xie; Y. Chen, and P. R. Kumar. “Dimensionality
20 Distributed Computing vol. 103 (2017), pp. 32–41. Reduction of Synchrophasor Data for Early
21 [168] B. Wang et al. “Electrical Safety Considerations in Event Detection: Linearized Analysis”. In: IEEE
22 Large-Scale Electric Vehicle Charging Stations”. In: Transactions on Power Systems vol. 29, no. 6 (Nov.
23 IEEE Transactions on Industry Applications (2019), 2014), pp. 2784–2794. ISSN: 0885-8950. DOI: 10 .
24 pp. 1–1. ISSN: 0093-9994. DOI: 10.1109/TIA.2019. 1109/TPWRS.2014.2316476.
25 2936474. [180] M. Wang et al. “A low-rank matrix approach
26 [169] Q. Li; R. Negi, and M. D. Ilić. “Phasor measurement for the analysis of large amounts of power
27 units placement for power system state estimation: A system synchrophasor data”. In: 2015 48th Hawaii
28 greedy approach”. In: 2011 IEEE Power and Energy International Conference on System Sciences. IEEE.
29 Society General Meeting. July 2011, pp. 1–8. DOI: 2015, pp. 2637–2644.
30 10.1109/PES.2011.6039076. [181] A. Goyal et al. “A unified framework for simulating
31 [170] M. A. Rahman and H. Mohsenian-Rad. “False Markovian models of highly dependable systems”.
32 data injection attacks with incomplete information In: IEEE Transactions on Computers vol. 41, no. 1
33 against smart power grids”. In: 2012 IEEE Global (Jan. 1992), pp. 36–51. ISSN: 0018-9340. DOI: 10 .
34 Communications Conference (GLOBECOM). Dec. 1109/12.123381.
35 2012, pp. 3153–3158. DOI: 10 . 1109 / GLOCOM . [182] D. Wilson et al. “Deep learning-aided cyber-attack
36 2012.6503599. detection in power transmission systems”. In: 2018
37 [171] P. Dehghanian and M. Kezunovic. “Probabilistic IEEE Power & Energy Society General Meeting
38 Decision Making for the Bulk Power System Optimal (PESGM). IEEE. 2018, pp. 1–5.
39 Topology Control”. In: IEEE Transactions on Smart [183] C.-H. Wu. “Behavior-based spam detection using a
40 Grid vol. 7, no. 4 (2016), pp. 2071–2081. hybrid method of rule-based techniques and neural
41 [172] P. Dehghanian et al. “Flexible Implementation of networks”. In: Expert Systems with Applications
42 Power System Corrective Topology Control”. In: vol. 36, no. 3 (2009), pp. 4321–4330.
43 Electric Power System Research vol. 128 (2015), [184] N. V. Tomin et al. “Machine learning techniques
44 pp. 79–89. for power system security assessment”. In: IFAC-
45 [173] M. Alhazmi et al. “Power Grid Optimal Topology PapersOnLine vol. 49, no. 27 (2016), pp. 445–450.
46 Control Considering Correlations of System [185] Y. Li; R. Ma, and R. Jiao. “A hybrid malicious
47 Uncertainties”. In: IEEE/IAS 55th Industrial and code detection method based on deep learning”.
48 Commercial Power Systems (I&CPS) Technical In: International Journal of Security and Its
49 Conference. 2019, pp. 1–7. Applications vol. 9, no. 5 (2015), pp. 205–216.
50 [174] M. Kezunovic et al. “Reliable Implementation of [186] R. C. Holte. “Very simple classification rules perform
51 Robust Adaptive Topology Control”. In: The 47th well on most commonly used datasets”. In: Machine
52 Hawaii International Conference on System Science learning vol. 11, no. 1 (1993), pp. 63–90.
53 (HICSS). 2014, pp. 1–10. [187] B. Martin. “Instance-based learning: nearest
54 [175] P. Dehghanian and M. Kezunovic. “Impact neighbor with generalization [thesis”. In: (1995).
55 Assessment of Power System Topology Control [188] L. Breiman. “Random forests”. In: Machine learning
56 on System Reliability”. In: IEEE Conference on vol. 45, no. 1 (2001), pp. 5–32.
57
58
VOLUME xx, 20xx 19
59 For Review Only
60
1
2 [189] G. W. Taylor. Composable, distributed-state models MOHANNAD ALHAZMI received the B.Sc. and
3 for high-dimensional time series. University of M.Sc. degrees in electrical engineering from Umm
4 Toronto Toronto, 2009. Al-Qura University, Saudi Arabia in 2013 and
The George Washington University, Washington
5 [190] J. Wei and G. J. Mendis. “A deep learning- D.C., USA, in 2017, respectively. He is currently
6 based cyber-physical strategy to mitigate false data pursuing his Ph.D. degree in the Department
7 injection attack in smart grids”. In: 2016 Joint of Electrical and Computer Engineering at The
8 Workshop on Cyber-Physical Security and Resilience George Washington University. His research
interests include power system control, reliability
9 in Smart Grids (CPSR-SG). IEEE. 2016, pp. 1–6. and resiliency of power grids and critical
10 [191] H. Wang et al. “Deep learning-based interval infrastructure, cyber security and smart electricity grid applications.
11 state estimation of AC smart grids against sparse
12 cyber attacks”. In: IEEE Transactions on Industrial
13 Informatics vol. 14, no. 11 (2018), pp. 4766–4778.
14 [192] M. Ashrafuzzaman et al. “Detecting stealthy false
15 data injection attacks in power grids using deep MOSTAFA NAZEMI received the B.Sc. degree
16 learning”. In: 2018 14th International Wireless from K. N. Toosi University of Technology,
Tehran, Iran, in 2015, and the M.Sc. degree
17 Communications & Mobile Computing Conference from Sharif University of Technology, Tehran,
18 (IWCMC). IEEE. 2018, pp. 219–225. Iran, in 2017, in electrical engineering and in
19 [193] Y. Chakhchoukh et al. “Statistical outlier detection energy systems engineering, respectively. He is
20 for diagnosis of cyber attacks in power state currently pursuing the Ph.D. degree in electrical
engineering at the Department of Electrical
21 estimation”. In: 2016 IEEE Power and Energy and Computer Engineering, George Washington
22 Society General Meeting (PESGM). IEEE. 2016, University, Washington, D.C., USA. His research
23 pp. 1–5. interests include power system resilience, power system planning and
24 [194] Y. He; G. J. Mendis, and J. Wei. “Real-time detection operation, energy optimizations, and smart electricity grid applications.
25 of false data injection attacks in smart grid: A deep
26 learning-based intelligent mechanism”. In: IEEE
27 Transactions on Smart Grid vol. 8, no. 5 (2017),
28 pp. 2505–2516.
ABOUZAR ESTEBSARI received his Ph.D.
29 [195] A. Tajer et al. “Distributed joint cyber attack degree in electrical engineering from Polytechnic
30 detection and state recovery in smart grids”. In: University of Turin (Politecnico di Torino), Italy,
31 2011 IEEE International Conference on Smart Grid in 2015. He is currently an Assistant Professor
32 Communications (SmartGridComm). IEEE. 2011, with the Department of Energy, at the Polytechnic
University of Turin, Italy. In summer 2016, he was
33 pp. 202–207. a visiting scholar at the Institute for Automation of
34 Complex Power Systems, E.ON. Energy Research
35 Center at RWTH-Aachen University, in Germany.
36 His main research interests include power system
security of supply and resiliency, smart grid applications, and digital real-
37 TIAN NGUYEN is currently pursuing B.Sc. time simulation.
38 degree in electrical engineering at the Department
39 of Electrical and Computer Engineering, The
40 George Washington University, Washington, D.C.,
USA. His research interests include power
41 grid reliability, energy efficiency, and renewable
42 integration. PAYMAN DEHGHANIAN (S11, M17) is
43 an Assistant Professor at the Department of
44 Electrical and Computer Engineering in George
Washington University, Washington, D.C., USA.
45 He received the B.Sc., M.Sc., and Ph.D. degrees
46 all in Electrical Engineering respectively from
47 University of Tehran, Tehran, Iran, in 2009,
48 SHIYUAN WANG (S’18) received the B.Eng Sharif University of Technology, Tehran, Iran, in
2011, and Texas A&M University, Texas, USA
49 degree in mechanical engineering from University
in 2017. His research interests include power
50 of Science and Technology Beijing, China, in
system protection and control, power system reliability and resiliency, asset
2012; the M.Sc. degree in electrical engineering
51 from The George Washington University, management, and smart electricity grid applications.
52 Washington, DC., USA, in 2014. He is currently Dr. Dehghanian is the recipient of the 2013 IEEE Iran Section Best M.Sc.
Thesis Award in Electrical Engineering, the 2014 and 2015 IEEE Region
53 pursuing the Ph.D. degree at the Department
5 Outstanding Professional Achievement Awards, and the 2015 IEEE-HKN
of Electrical and Computer Engineering, George
54 Outstanding Young Professional Award.
Washington University, Washington, D.C., USA.
55 His research interests include power system
56 reliability and resiliency, smart grid and renewable energy, power grid
57 harmonic analysis, and application of signal processing in energy analytics.
58
20 VOLUME xx, 20xx
59 For Review Only
60
1
2
3 Tian Nguyen is currently pursuing B.Sc. degree in electrical engineering at the
4 Department of Electrical and Computer Engineering, The George Washington
5 University, Washington, D.C., USA. His research interests include power grid
6
reliability, energy efficiency, and renewable integration.
7
8
9
10
11
12
13
14
15 Shiyuan Wang (S’18) received the B.Eng degree in mechanical engineering from
16 University of Science and Technology Beijing, China, in 2012; the M.Sc. degree
17 in electrical engineering from The George Washington University, Washington,
18 DC., USA, in 2014. He is currently pursuing the Ph.D. degree at the Department
19
of Electrical and Computer Engineering, George Washington University,
20
Washington, D.C., USA. His research interests include power system reliability
21
22
and resiliency, smart grid and renewable energy, power grid harmonic analysis,
23 and application of signal processing in energy analytics.
24
25
26
Mohannad Alhazmi received the B.Sc. and M.Sc. degrees in electrical
27
engineering from Umm Al-Qura University, Saudi Arabia in 2013 and The
28
29
George Washington University, Washington D.C., USA, in 2017, respectively.
30 He is currently pursuing his Ph.D. degree in the Department of Electrical and
31 Computer Engineering at The George Washington University. His research
32 interests include power system control, reliability and resiliency of power grids
33 and critical infrastructure, cyber security and smart electricity grid applications.
34
35
36
37
38 Mostafa Nazemi (S’18) received the B.Sc. degree from K. N. Toosi University
39 of Technology, Tehran, Iran, in 2015, and the M.Sc. degree from Sharif University
40 of Technology, Tehran, Iran, in 2017, in electrical engineering and in energy
41 systems engineering, respectively. He is currently pursuing the Ph.D. degree in
42 electrical engineering at the Department of Electrical and Computer Engineering,
43 George Washington University, Washington, D.C., USA. His research interests
44 include power system resilience, power system planning and operation, energy
45 optimizations, and smart electricity grid applications.
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 For Review Only
1
2
3 Abouzar Estebsari received his Ph.D. degree in electrical engineering from
4 Polytechnic University of Turin (Politecnico di Torino), Italy, in 2015. He is
5 currently an Assistant Professor with the Department of Energy, at the
6
Polytechnic University of Turin, Italy. In summer 2016, he was a visiting scholar
7
at the Institute for Automation of Complex Power Systems, E.ON. Energy
8
9
Research Center at RWTH-Aachen University, in Germany. His main research
10 interests include power system security of supply and resiliency, smart grid
11 applications, and digital real-time simulation.
12
13
14
15 Payman Dehghanian (S11, M17) is an Assistant Professor at the Department of
16 Electrical and Computer Engineering in George Washington University,
17 Washington, D.C., USA. He received the B.Sc., M.Sc., and Ph.D. degrees all in
18 Electrical Engineering respectively from University of Tehran, Tehran, Iran, in
19
2009, Sharif University of Technology, Tehran, Iran, in 2011, and Texas A\&M
20
University, Texas, USA in 2017. His research interests include power system
21
22
protection and control, power system reliability and resiliency, asset management,
23 and smart electricity grid applications.
24 Dr. Dehghanian is the recipient of the 2013 IEEE Iran Section Best M.Sc.
25 Thesis Award in Electrical Engineering, the 2014 and 2015 IEEE Region 5
26 Outstanding Professional Achievement Awards, and the 2015 IEEE-HKN Outstanding Young
27 Professional Award.
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60 For Review Only

Access-2020-20869 Proof Hi

Uploaded by

Copyright:

Available Formats

You might also like

Access-2020-20869 Proof Hi

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Access-2020-20869 Proof Hi

Uploaded by

Copyright:

Available Formats

IEEE Access

Electric Power Grid Resilience to Cyber Adversaries: State of

Journal: IEEE Access

Manuscript Type: Regular Manuscript

Date Submitted by the

Keywords: <b>Please choose

Cyber Physical Systems (CPS), Cyber Attacks, Intrusion Detection

For Review Only

16 Corresponding author: Payman Dehghanian (e-mail: payman@gwu.edu).

10 system observability, one typically needs to install PMUs (a)

19 SE accuracy using an algorithm that is related to key (b)

43 and is used as a defense mechanism against FDIAs which are

You might also like