19390122, 2015, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/sec.751 by Motilal Nehru National Institu, Wiley Online Library on [08/11/2022].

See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
SECURITY AND COMMUNICATION NETWORKS
Security Comm. Networks 2015; 8:2431–2439
Published online 12 March 2013 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/sec.751

SPECIAL ISSUE PAPER

Physical layer security in wireless smart grid

Xin Wang1*, Qilian Liang1, Jiasong Mu2, Wei Wang2 and Baoju Zhang2
1
Department of Electrical Engineering, University of Texas, Arlington, TX, 76019-0016, U.S.A.
2
College of Physics and Electronic Information, Tianjin Normal University, Tianjin, 300387, China

ABSTRACT
The smart grid system is composed of the power infrastructure and communication infrastructure and thus is characterized by
the flow of electric power and information, respectively. Although there is no doubt that the wireless communication
architecture will play a significant role in smart grid, the wireless network introduces additional vulnerabilities, given the scale
of potential threats. Therefore, the physical layer security issue is of first priority in the study of smart grid and has already
attracted substantial attention in the industry and academia. In this paper, we aimed to present a general overview of the physical
layer security in wireless smart grid and cover the effective countermeasures proposed in the literature of smart grid to date. We
first investigate the security challenges from malicious attacks. Specifically, two typical forms of malicious attack in smart grid,
namely, jamming and bad data injecting, are studied. In addition, the related countermeasures against these malicious attacks
are illustrated. Further, we analyze the state of the art of the privacy issues in smart grid. The private information and privacy
concerns are introduced, and then the effective solutions to privacy security are provided. Finally, voltage regulation, a security
topic that has been hardly studied in the wireless smart grid domain, is presented. We expect that the work presented here will
advance the research on smart grid security. Copyright © 2013 John Wiley & Sons, Ltd.

KEYWORDS
smart grid; big data; physical layer security; malicious attacks; privacy issues; voltage regulation

*Correspondence
Xin Wang, Department of Electrical Engineering, University of Texas, Arlington, TX, 76019-0016, U.S.A.
E-mail: [email protected]

1. INTRODUCTIONS AND grid and establishes the two-way information flow between
MOTIVATIONS customers and utilities [1,2]. For instance, smart meters,
which are basic components in smart grid, are capable of
Although the existing electrical grid has served well in collecting and delivering power consumption information
providing the necessary power supply of electricity, the to a remote intelligent control center more efficiently than
growing demand, fast depletion of primary energy resources, conventional meters. Smart meters measure the load profile
unreliability and negative impact on the environment must and demand, record historical information and act as the
be taken into account in a vision of the future. This vision meter as gateway or meter as device. With the aid of smart
is being realized with the smart grid, which expands the meters, the utilities (intelligent control center) are able to
current capabilities of the electric grid’s generation, trans- monitor the peak load and control various appliances and
mission and distribution systems to provide an infrastructure industry electricity consumption for optimal power genera-
capable of handling future requirements for distributed tion and distribution.
generation, renewable energy sources, electric vehicles and However, the amount of data generated by smart meters
the demand-side management of electricity (Figure 1). and intelligent sensors in the smart grid will experience
Smart grid is composed of the power infrastructure and explosive growth in the next few years. According to the
communication infrastructure and thus is characterized by recent report of SBI Research (Rockville, MD), the volume
the flow of electric power and information, respectively. of smart grid data managed by utilities over the next few
The power infrastructure generates and distributes electricity years is going to surge from 10 780 TB of new data in
power to consumers; the communication infrastructure 2010 to over 75 200 TB in 2015 [3]. Pervasive sensing
controls the power infrastructure and enables the intelligent and computing across the smart grid are generating hetero-
operation of smart grid. Therefore, an efficient and reliable geneous data at unprecedented scale and complexity.
communication architecture plays a significant role in smart These tremendous data, which fall into the region of “big

Copyright © 2013 John Wiley & Sons, Ltd. 2431

 19390122, 2015, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/sec.751 by Motilal Nehru National Institu, Wiley Online Library on [08/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Physical layer security in wireless smart grid X. Wang et al.

Figure 1. Overview of the smart grid.

data,Ӡ place considerable pressure on the communications fundamental issues, that is, malicious attacks, privacy
infrastructure of the smart grid. It is worth mentioning that issues and voltage regulation, will serve as the criteria to
because of the nature of the power grid, a wired data investigate smart grid security.
communication network based on power line communica-
tion (PLC) is apparently a suitable candidate; however, in • Malicious attacks: malicious individuals can attack the
smart grid, wireless communication is more preferred by communication system and thus cause various damages
many application scenarios [5], such as the following: to the smart grid or even result in a large-area blackout.
(i) when many parameters need to be monitored, PLC • Privacy issues: wireless communication architecture
can result in a costly and complicated system architecture; provides a window into the actives within homes,
(ii) PLC cannot easily bypass transformers in a power exposing once-private activities to anyone with access
distribution network; and (iii) wired communications to electricity usage information.
cannot provide peer-to-peer communications among elec- • Voltage regulation: the introduction of distributed
tric devices in a flexible manner. Further, among many energy resources (DERs) makes the traditional voltage
proposals for the communication infrastructure, wireless regulation and power control methods no longer valid.
communication is a promising one because of its low cost,
large coverage and fast deployment [6,7]. The remainder of this paper is organized as follows. First,
Although there is no doubt that the wireless com- we formulate the data communication and wireless network
munication network provides significant benefits over model in Section 2. The challenges from malicious attacks
wired communication architecture, the wireless network are studied in Section 3. We further investigate the privacy
introduces additional vulnerabilities. For example, the issue in smart grid in Section 4. The analysis of the voltage
malicious individuals can attack the wireless communica- regulation is presented in Section 5. Finally, in Section 6,
tion system, cause various damages to the smart grid or we present some concluding remarks.
even result in a large-area blackout [8–11]. Therefore, the
physical layer security is of first priority in the study of
smart grid and has already attracted substantial attention
in the industry and academia. The goal of this paper is to 2. DATA COMMUNICATION AND
present a general overview of the critical physical layer WIRELESS NETWORK
security issues in smart grid and cover the effective ARCHITECTURE
countermeasures proposed in the literature of smart grid
to date. To this end, in this survey, the following three As we have discussed, the integration of the information and
communication infrastructures into smart grid introduces
† new security-related and privacy-related challenges. To
In the BIGDATA solicitation [4], the phase “big data” refers to large,
diverse, complex, longitudinal and/or distributed data sets generated from help understand the potential threats in smart grid, in this
instruments, sensors, Internet transactions, email, video, click streams section, we formulate the data communication and wireless
and/or all other digital sources available today and in the future. architecture in smart grid.

2432 Security Comm. Networks 2015; 8:2431–2439 © 2013 John Wiley & Sons, Ltd.
DOI: 10.1002/sec
 19390122, 2015, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/sec.751 by Motilal Nehru National Institu, Wiley Online Library on [08/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
X. Wang et al. Physical layer security in wireless smart grid

The smart grid is usually deployed in a considerably personally identifiable information, which could com-
large geographical field, and accordingly, the communica- promise consumer privacy.‡
tion infrastructure of the smart grid has to cover the entire
region to connect all the wired or wireless nodes. As 2.4. Wireless sensor network
shown in Figure 2, the smart grid is characterized by the
two-way flow of electric power and information. Different A wireless sensor network (WSN) is composed of a
data communication and network architectures extend number of wireless sensors used to monitor and control
across the whole smart grid from the home area to the wide the operational characteristic behavior of smart grid
area [2,3]. devices so that any outage or disturbance can be prevented.
These sensors are deployed at multiple sites in smart grid
2.1. Home area network (e.g. at the transformer, a distribution substation or home)
to measure different system parameters. WSN is required
Home area networks communicate with various smart to provide secure and continuous information transfer
devices, for example, smart TV and plug-in hybrid electric among sensors using wireless links.
vehicles, to provide energy efficiency management and
demand response. One objective of smart grid is to
reduce the energy consumption in the household. To this 3. MALICIOUS ATTACKS
end, the home energy management system is developed for
efficient energy management at consumer side. Clearly, in In this section, we provide a deep analysis of malicious
Figure 2, the home energy management system provides attacks and potential threats to the physical layer of
the capability of monitoring and controlling different wireless communication in smart grid. A physical layer
electrical appliances. The HAN employs short-range or attack is defined as the malicious behavior disturbing the
local area transmission techniques, such as PLC, Zigbee, legitimate communication of a wireless network [13].
Bluetooth and Wi-Fi. Therefore, the power management
services and smart meters could be monitored and controlled 3.1. Features of malicious attacks
in a real-time fashion.
The nature of malicious attacks on smart grids can be very
different from that on traditional communication networks
2.2. Wide-area network such as the Internet. The objective of an adversary may not
be just gaining unauthorized information; an adversary could
Neighborhood area networks connect multiple HANs to cripple the power grid by attacking the data management
local access points. Wide-area networks (WANs) provide system, which collects the data from remote meters and
the communication links between the neighborhood area produces estimates of the system states at intervals of
networks and the utility systems to transfer information. roughly 15 min. If an adversary is able to hack into the power
WAN is a monitoring and controlling system that can grid at the control center (e.g., injecting false messages into
simultaneously acquire a variety of information (phase of the network), the data management system at the control
voltage, current and load profile) to provide effective center may be misled by the state estimator, potentially
security and fault tolerance for the entire power grid. With making erroneous decisions on contingency analysis,
the main focus on power generation, transmission and dispatch or even billing. In summary, some features of
distribution, WANs consist of an intelligent control center, malicious attacks are as follows [13]:
phasor measurement unit (PMU) and phasor data collector
to measure the electrical waves so that the state and 1. Malicious attacks are potential hazards that can
performance of the power grid could be analyzed. materialize as attacks, that is, threats cause damage,
when executed as offensive actions.
2.3. Advanced metering infrastructure 2. Malicious attacks are the actual implementation of
threats and thus are an effective cause of damages.
Advanced metering infrastructure (AMI) is probably the 3. Malicious threats are selective: the more the target
most commonly known technology in smart grid to can produce disruptive effects, the more it is likely
support data communications architecture between the to be attacked; the more the target is protected, the
smart meter and data management system [12]. AMI less likely it will be attacked.
allows the utility companies to remotely measure, collect 4. Malicious attacks are carried through as processes,
and analyze the usage statistics from smart meters. Addi- where the actions of the attackers and defenders
tionally, AMI provides an interface to other parts of the
smart grid. AMI is used to transfer real-time data including ‡
According to a privacy impact analysis performed for the National
fault and outage to the electrical power control center.
Institute of Standards and Technology (NIST), “The data item collected
The information gathered by the smart meter drives the from the Distributed Energy Resources (DERs) and smart meters will
demand-response function and enables the majority of reveal different types of information about residential consumers
smart grid applications. However, these data will include and activities.”

Security Comm. Networks 2015; 8:2431–2439 © 2013 John Wiley & Sons, Ltd. 2433
DOI: 10.1002/sec
 19390122, 2015, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/sec.751 by Motilal Nehru National Institu, Wiley Online Library on [08/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Physical layer security in wireless smart grid X. Wang et al.

Figure 2. Data communication and wireless network architecture.

follow and affect each other. In other words, an could distort the data signal, which aggravates the signal-
attack is a chain of mutually dependent offensive to-interference and noise ratio value at the receiver,
and defensive events. In most cases, the hostile part and the node cannot recover the information out of the
of an attack will comprehend several steps, probably damaged packets. The goal of jamming is to block the
organized in successive phases. information flow.

3.2. Types of malicious attacks

3.2.3. Injecting
An injecting attack inserts the formatted bad data into
The level of attack threats, for a given component, depends
the wireless network, with the purpose of biasing the
on the attitudes, decisions and interaction between attacks
power system state information. If successful, such attacks
and defenders at a given point in time and space.
may mislead the control center to take erroneous actions
Here, we classify the malicious attacks into three groups
or, at the minimum, make the control center distrust the
according to the objectives and behaviors: eavesdropping,
state estimation. Moreover, by taking advantage of the
jamming and injecting [9].
configuration information of the power system, the attacker
can inject malicious measurements that will mislead the
3.2.1. Eavesdropping
state estimation process without being detected by any of
Because a wireless signal propagates over open space,
the existing techniques for bad data detection.
any network node within radio range is able to capture the
In the following, we mainly focus on the current studies
signal. Further, as legitimate nodes follow the communi-
on the “initiative” challenges from the jamming and
cation standards, the neighboring nodes could obtain
injecting malicious attacks.
meaningful information from the captured signal. Therefore,
an unauthorized node eavesdrops on the data transmission
and accesses credential information, which violates the 3.3. Countermeasures against jamming
confidentiality requirements. Clearly, this kind of attack is attacks
not easily detected because the adversary does not expose
its activity. First, in contrast to other attacks such as bad-data-injecting
attacks, the jamming attack is much simple as it only
3.2.2. Jamming requires brute force interference and does not need to hack
As the most typical form of malicious attack, a jamming into the computer and communication systems in the smart
attack uses high-power interference to suppress the infor- grid. The traditional security methodologies, such as
mation signal [14,15]. This affects a legitimate node in public/private key and hash function, cannot combat the
two ways. First, when the node performs the carrier sense jamming attacks in the physical layer. As pointed out
before transmitting a packet, the channel is always “busy.” in [16], the jamming attack can only block but cannot
This defers its transmission, and the node eventually change the information; the worst case is that the infor-
gives up communication. Second, the node may fail to mation flow in the power market is completely blocked
receive the packets. Suppose it is receiving packets from and the power market degenerates to the traditional one
a legitimate communication partner. The noise signal without communication infrastructures.

2434 Security Comm. Networks 2015; 8:2431–2439 © 2013 John Wiley & Sons, Ltd.
DOI: 10.1002/sec
 19390122, 2015, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/sec.751 by Motilal Nehru National Institu, Wiley Online Library on [08/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
X. Wang et al. Physical layer security in wireless smart grid

To combat jamming-based attacks, various defense formulates the malicious attack by simply jamming the
countermeasures can be taken in both the communication pricing signal, which could block or release the informa-
system design and the marketing policy. For the com- tion flow in the power market. To efficiently counter this
munication aspect, it is effective to carry out a frequency- kind of attack, the authors borrowed the idea from random
hopping spread spectrum (FHSS) over a wide frequency access in communication systems, such as Aloha and carrier
band [17]. FHSS is a spread-spectrum method of trans- sense multiple-access protocols, in which different transmit-
mitting radio signals by rapidly switching a carrier among ters take random back-offs to avoid the collision incurred by
many frequency channels, using a shared algorithm known simultaneous transmission. The basic idea is each power
to both the transmitter and the receiver, such that the attack consumer chooses a random time to change its power
cannot catch the wireless transmission. FHSS brings response. The simulation results show that the attack strategy
forward many advantages: validly manipulates the power market although the counter-
measure can effectively combat such an attack.
• It minimizes unauthorized interception and jamming
of radio transmission between the nodes. 3.4. Countermeasures against injecting
• The signal-to-noise ratio required for a carrier, relative attacks
to the background, decreases as a wider range of
frequencies is used for transmission. We consider the covert attacks by injecting malicious data
• It deals effectively with the multipath fading effects. to the communications system with the goal of biasing the
power system state estimation.
One of the main drawbacks of frequency hopping is First, Lie et al. [22] presented a false data injection
the required wide bandwidth and the low spectrum attack that will mislead the state estimation process without
efficiency [15]. being detected by any of the existing techniques for bad
Direct sequence spread spectrum transmissions are measurement detection. The authors considered two realis-
performed by multiplying the transmitted data and a tic attack scenarios: random false data injection attacks, in
pseudonoise (PN) digital signal [17,18]. The PN digital which the attacker aims to find any attack vector as long as
signal is a pseudorandom sequence of 1 and 1 values, it can lead to a wrong estimation of state variables, and
at a frequency (chip rate) much higher than the original targeted false data injection attacks, in which the attacker
signal. Therefore, the original signal is replaced with a aims to find an attack vector that can inject arbitrary errors
wide-bandwidth signal with the spectrum equivalent of into certain state variables. The simulations on IEEE test
the noise. The aforementioned FHSS advantages also systems demonstrate the success of these attacks.
apply to the direct sequence spread spectrum strategy. Malicious attacks against power systems are investigated in
Moreover, multiplying the PN sequence makes it difficult [23,24], in which an adversary controls a set of smart meters
for the attack to descramble the transmitted radiofrequency and is able to inject malicious data to those smart meters with
carrier and recover the original signal. the goal of biasing the power system state estimation. In the
Besides, uncoordinated frequency hopping [19], ultra weak-attack regime, the authors proposed a detector based
wide band [20] and directional transmission technologies on a generalized likelihood ratio test and further studied a
could be applied to combat the jamming attacks in the detector using a convex regularization of the optimization
smart grid. Because of space limitation, in this paper, we problem, based on L1 norm minimization. Similarly, in [25],
will not further illustrate these strategies with details. the authors implemented a space–time block coding over
Further, in [21], the authors studied the denial-of-service multiple antennas at the smart meter to enhance real-time com-
jamming and anti-jamming in multichannel wireless com- munication between the smart meter and the control center
munication systems from a remote sensor to the control (Figure 3). The reliability challenges from the link failures
center in the smart grid. The authors proposed a heuristic are also considered. It is verified that more antennas will give
zero-sum stochastic game to model the competition between more diversity gains and better security performance.
the jammer and the sensor, in which the actions are depen- The potential economic impact of false data injection
dent on the system state. A quadratic reward function has attacks is presented in [26,27]. Recently, the study of
been defined for the sensor, which facilitates the linear bad-data-injecting attack focuses on the improvement of
quadratic Gaussian control at the controller. Note that unlike robustness using phasor units (PMUs) [28,29]. For example,
the traditional communication networks, in the smart grid, Chen and Abur [28] employed PMUs to transform critical
the reward is measured by stability, instead of throughput measurements into redundant measurements such that the
or delay, which are typical metrics in traditional networks. bad measurements can be detected by the measurement
This work concludes that it is necessary to take the system residual testing and the system is still observable.
state-aware anti-jamming actions for the remote sensors.
Note that the previous anti-jamming strategies are
discussed from the communication perspective. For the 4. PRIVACY ISSUES
marketing side, Mpitziopoulos et al. [16] proposed a
random back-off scheme for the load adjustment to avoid Although deploying the smart grid has enormous social
the impulsive impact on market price and power load. It and technical benefits, several privacy concerns arise. In

Security Comm. Networks 2015; 8:2431–2439 © 2013 John Wiley & Sons, Ltd. 2435
DOI: 10.1002/sec
 19390122, 2015, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/sec.751 by Motilal Nehru National Institu, Wiley Online Library on [08/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Physical layer security in wireless smart grid X. Wang et al.

Figure 3. Bad-data-injecting attacks to smart meters [25].

this section, we will give a brief overview of current For customer-side components such as smart meter,
studies on privacy issues in the smart grid. the wired or wireless connection used to transfer
meter data might also expose the valuable and private
4.1. Private information information of a consumer’s habits and behaviors.
Certain activities, such as watching television, have
Personal private information is any recorded information that detectable power consumption signatures.
can identify an individual directly or indirectly [30–32]. 3. Third, obtaining near-real-time data regarding energy
Taking into consideration the smart grid context, any type consumption may infer whether a residence or facility
of energy use data that link to private information should is occupied, where people are in the structure, what
be secured and monitored in a proper way. The US NIST they are doing and so on.
guidelines [31] have provided a list of private information 4. Fourth, personal lifestyle information derived from
that may be available through the smart grid as follows [33]: energy use data could be valuable to some vendors
or parties. For instance, vendors may use this infor-
• Name: responsible for the account. mation for targeted marketing, which could not be
• Address: location to which service is being taken. welcomed by those targets.
• Account number: unique identifier for the account.
• Meter reading: kW, kWh consumption recorded at 15- to 4.3. Solutions to privacy security
60-min intervals during the current billing cycle.
• Current bill: current amount due on the account. To secure the privacy issue in smart gird, in [34], the authors
• Billing history: past meter readings and bills, including suggested that the customers would possibly deploy a
history of late payments/failure to pay, if any. separate measurement device at home to better monitor their
• HAN: in-home electrical appliances. power usage. However, redundant wireless power metering
• Lifestyle: when the home is occupied and it is unoccu- introduces a potential confidentiality leak because an eaves-
pied, when occupants are awake and when they are dropper could easily discover the instantaneous power usage
asleep, how many various appliances are used and so on. of the customer. Therefore, the authors developed a coding
• DER: the presence of on-site generation and/or method to guarantee information-theoretic confidentiality
storage devices, operational status, net supply to or of the redundant meter data. The key idea is to compress
consumption from the grid and usage patterns. the redundant measurement to a rate below its entropy, so
• Meter IP: the IP address for the meter, if applicable. that it cannot be recovered from the encode bits and the infor-
• Service provider: identity of the party supplying this mation is theoretically secure regardless of the computational
account, relevant only in retail access markets. power of an eavesdropper.
Note that many previous studies on user privacy have
4.2. Privacy concerns been conducted in a centralized manner, where owners have
the ability to adjust the data from a global perspective [35].
In smart grid, energy consumption data obtained by a third However, these models are not exactly suited to privacy
party may disclose private information. According to the problems of smart meters, which are set up in a decentralized
study of NIST [33,31], four typical areas of privacy environment. In [36], a “smart metering privacy model” is
concern in smart grid are presented as follows: introduced to measure the degree of privacy that a smart
metering application can provide. To protect the privacy of
1. First of all, fraud should be considered, especially the users in a completely decentralized setting, Wang et al.
when energy consumption is attributed to a different [36] proposed a novel protocol, allowing individual meters
location. The metering system (physical recording or to report the true electric consumption reading with a prede-
electronically metering systems) should not allow termined probability. Specifically, smart meters report sam-
any personal abuse or modify the collected data [31]. ples from the Gaussian mixture models with K Gaussian
2. Second, data in the smart meter and HAN could components, in which the actual smart meter readings are
reveal certain activities of home smart appliances. mixed with faked readings from K 1 predetermined

2436 Security Comm. Networks 2015; 8:2431–2439 © 2013 John Wiley & Sons, Ltd.
DOI: 10.1002/sec
 19390122, 2015, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/sec.751 by Motilal Nehru National Institu, Wiley Online Library on [08/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
X. Wang et al. Physical layer security in wireless smart grid

distributions so that the appliance usage information of 5. VOLTAGE REGULATION IN

individual users is protected. Similarly, Li et al. [37] pro- SMART GRID
posed a distributed incremental data aggregation approach
for all level meters based on the spanning tree topology (each In this section, we will address the voltage regulation issue, a
node collects data from its children, aggregates them with its topic that has been hardly investigated in the wireless smart
own data and sends the intermediate result to the parent grid domain. Currently, supplying isolated locations with
node). To protect the user and neighborhood privacy in this electricity comes with an increased cost to the distributed
data aggregation, the authors [37] employed a homomorphic network operators, as the majority of the energy that is
encryption to secure the data all the way from home meters to destined for customers is wasted in the form of heat before
the control center. any useful energy reaches the consumer [39]. In smart grid,
Further, Kalogridis et al. [38] proposed a heuristic DERs, such as wind turbines and solar panels, provide a
approach to hide the information contained in the con- cheaper and more efficient solution that can deliver the
sumption data by moderating the home load signatures, energy from points closer to the consumer than the distributed
after mixing utility energy with energy provided by a network operator’s centralized power grid [40] (Figure 1).
rechargeable battery. From the privacy security perspective, Maintaining adequate operating voltage at all the customer
hiding or obscuring the load signature could prevent the delivery points is very critical to the whole smart grid.
appliance usage events from being detected. Further, three
privacy measurement metrics, that is, relative entropy, 5.1. Voltage regulation concerns
cluster classification and regression, are introduced. As
mentioned in [38], user privacy requirement may conflict Voltage regulation in traditional distribution systems is
with the other requirements, such as the cost-saving require- mainly provided by onload tap changer, line regulators
ment. Therefore, it is reasonable to strike a balance between and switched shunt capacitors at the substations and
these system requirements. feeders, on the assumption that the current always flows
from the transmission to high-voltage/middle-voltage
substations and then to the middle-voltage feeders. However,
the introduction of DERs makes this assumption no longer
valid [41]. DERs alter the flows, which, in turn, alter feeder
voltage profiles and affect the voltage regulation and power
control in distributed systems [42].
Hence, the integration of DERs into microgrid makes
power supply more reliable, whereas it also poses a challenge
to voltage regulation (Figure 4). Without reasonable coordi-
nations, the voltage is not stable after introducing the DERs.

5.2. Solutions to voltage regulation

Therefore, accurately monitoring the real-time voltage

level and coordinating DERs with the traditional voltage
regulators are imperative in microgrid.
These challenges motivate the application of WSNs into
smart grid. To maintain the voltage level and coordinate
Figure 4. Impact of distributed energy resources (DERs) on volt- DERs with the traditional voltage regulators, in smart grid,
age regulation. a group of voltage sensors (Figure 5) are deployed to

Figure 5. Application of wireless sensor networks (WSNs) in smart grid to regulate the voltage [43].

Security Comm. Networks 2015; 8:2431–2439 © 2013 John Wiley & Sons, Ltd. 2437
DOI: 10.1002/sec
 19390122, 2015, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/sec.751 by Motilal Nehru National Institu, Wiley Online Library on [08/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Physical layer security in wireless smart grid X. Wang et al.

monitor the real-time voltage statement and detect the load REFERENCES
fluctuation teXin12. WSNs, characterized by monitoring,
automatic control and two-way communications, could 1. Niyato D, Hossain E, Rashid MM, Bhargava VK.
effectively monitor the real-time voltage statement, detect Wireless sensor networks with energy harvesting tech-
load fluctuation and track power flow [6,7]. When the nologies: a game-theoretic approach to optimal energy
manually deployed sensors detect the voltage fluctuates, they management. IEEE Wireless Communications 2007;
will transmit the real-time voltage measurement to the 14(4):90–96.
control center, which could immediately respond to the 2. Niyato D, Wang P. Cooperative transmission for meter
voltage fluctuation by turning on/off the related DERs and data collection in smart grid. IEEE Communications
voltage regulators to stabilize the power supply and thus Magazine 2012; 50(4):90–97.
avoid significant voltage perturbations. To achieve proper 3. Yu R, Zhang Y, Yuen C, Xie S, Guizani M. Cognitive
voltage regulation, in [43], two fundamental issues, that is,
radio based hierarchical communications infrastructure
voltage estimation accuracy and power efficiency, will serve
for smart grid. IEEE Network 2011; 25(5):6–14.
as the criteria for sensor selection. Wang and Liang [43] first
proposed an opportunistic sensor selection scheme under 4. Core Techniques and Technologies for Advancing Big
equal power allocation, and then they addressed the sensor Data Science and Engineering (BIGDATA) solicitation.
selection scheme under optimal power allocation and derived Available at: http://www.nsf.gov/pubs/2012/nsf12499/
a reminiscent “water-filling” solution. nsf12499.htm
Further, to coordinate the collaboration of multiple 5. Wang X, Yi P. Security framework for wireless
DERs in voltage control, Mao and Li [44] proposed a communications in smart distribution grid. IEEE
sample-contention mechanism in WiMAX to prioritize Transactions on Smart Grid 2011; 2(4):809–818.
voltage state reports. Note that the sparseness of large 6. Zaballos A, Vallejo A, Selga J. Heterogeneous
voltage derivation presents the challenge of efficient utiliza- communication architecture for the smart grid. IEEE
tion of communication resources. Therefore, it is unwise to
Network 2011; 25(5):30–37.
allocate a large number of reporting time slots on a regular
7. Hongrae K, Jahns TM. Current control for AC motor
basis. An area central controller located at the WiMAX base
station combines newly updated reports with prediction drives using a single DC-link current sensor and
estimations as current complete voltage state and applies measurement voltage vectors. IEEE Transactions on
optimal control accordingly. This proposed scheme is Industry Applications 2006; 42(6):1539–1547.
tailored for quality-of-service communication requirements 8. Fadlullah ZM, Shen X, Nozaki Y. An early warning
of the power system with special attention to the sparseness system against malicious activities for smart grid
of voltage deviation, so that it is able to provide communica- communications. IEEE Network 2011; 25(5):50–11.
tion resource-efficient service for voltage state reporting in 9. Lee E, Gerla M, Oh SY. Physical layer security in
smart grid. wireless smart grid. IEEE Communications Magazine
2012; 50(8):46–52.
10. Khurana H, Hadley M, Ning L, Frincke DA. Smart-grid
6. CONCLUSIONS security issues. IEEE Security and Privacy 2010; 8(1):
81–85.
The physical layer security has become a critical topic in
smart grid, given the scale of potential threats. Our goal 11. He D, Chen C, Bu J, Chan S, Zhang Y, Guizanni M.
is to provide a general overview of the physical layer Secure service provision in smart grid communications.
security in wireless smart grid. To help understand the IEEE Communications Magazine 2012; 50(8):53–61.
potential threats in smart grid, we formulated the data 12. Flick T, Morehouse J. Securing the Smart Grid: Next Gen-
communication and wireless network model. In this eration Power Grid Security. Syngress: Amsterdam, 2011.
survey, three fundamental issues, that is, malicious attacks, 13. Bompard E, Gao C, Napoli R, Russo A, Masera M,
privacy issues and voltage regulation, are chosen as the Stefanini A. Risk assessment of malicious attacks
criteria to investigate the smart grid security. We first against power systems. IEEE Transactions on Systems,
discussed the security challenges from malicious attacks. Man and Cybernetics, Part A: Systems and Humans
The features and types of malicious attacks in smart grid 2009; 39(5):1074–1085.
are summarized. Specifically, two typical forms of 14. Su H, Qiu M, Wang H. Secure wireless communica-
malicious attack, namely, jamming and bad data injecting,
tion system for smart grid with rechargeable electric
are studied. In addition, the related countermeasures
against these malicious attacks are illustrated. Further, we vehicles. IEEE Communications Magazine 2012; 50(8):
analyzed the state of the art of the privacy issues in smart 62–68.
grid. The private information and privacy concerns are 15. Mpitziopoulos A, Gavalas D, Konstantopoulos C,
introduced, and then the effective solutions to privacy Pantziou G. A survey on jamming attacks and counter-
security are provided. Finally, presented is the discussion measures in WSNs. IEEE Communications Surveys
of voltage regulation in smart grid. and Tutorials 2009; 11(4):42–56.

2438 Security Comm. Networks 2015; 8:2431–2439 © 2013 John Wiley & Sons, Ltd.
DOI: 10.1002/sec
 19390122, 2015, 14, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/sec.751 by Motilal Nehru National Institu, Wiley Online Library on [08/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
X. Wang et al. Physical layer security in wireless smart grid

16. Hossain E, Han Z, Poor HV. Smart Grid Communica- 31. US NIST. Guidelines for smart grid cyber security
tions and Networking. Cambridge University Press: (vol. 1 to 3). Available at: http://csrc.nist.gov/publications/
Cambridge, 2012. PubsNISTIRs.html
17. Pickholtz RL, Schilling DL, Milstein LB. Theory 32. Cavoukian A, Polonetsky J, Wolf C. SmartPrivacy for
of spread spectrum communications—a tutorial. IEEE the smart grid: embedding privacy into the design of
Transactions on Communications 1982; 20(5):855–884. electricity conservation. Identity in the Information
18. DSSS Wikipedia. http://en.wikipedia.org/wiki/Direct- Society 2010; 3(2):275–294.
sequence-spread-spectrum 33. Liu J, Xiao Y, Li S, Liang W, Chen P. Cyber security
19. Popper C, Strasser M, Capkun S. Anti-jamming and privacy issues in smart grids. IEEE Communications
broadcast communication using uncoordinated spread Surveys and Tutorials 2012; 14(4):981–997.
spectrum techniques. IEEE Journal on Selected Areas 34. Varodayan DP, Gao GX. Redundant metering for
in Communications 2010; 28(5):703–715. integrity with information-theoretic confidentiality.
20. Oppermann I, Stoica L, Rabbachin A, Shelby Z, Proc. 1st IEEE SmartGridComm 2010, Gaithersburg,
Haapola J. UWB wireless sensor networks: UWEN— MD, Oct. 2010; 345–349.
a practical example. IEEE Communications Magazine 35. Fienberg S. Privacy and confidentiality in an e-commerce
2004; 42(12):27–32. world: data mining, data warehousing, matching and
21. Li H, Lai L, Qiu RC. A denial-of-service jamming disclosure limitation. Statistical Science 2006; 21(2):
game for remote state monitoring in smart grid. 2011 143–154.
45th Annual Conference on Information Sciences and 36. Wang S, Cui L, Que J, et al.. A randomized response
Systems, Mar. 2011; 1–6. model for privacy preserving smart metering. IEEE
22. Liu Y, Ning P, Reiter MK. False data injection Transactions on Smart Grid 2012; 3(3):1317–1324.
attacks against state estimation in electric power 37. Li F, Luo B, Liu P. Secure information aggregation for
grids. Proceedings of the 16th ACM conference on smart grids using homomorphic encryption. IEEE
Computer and communications security, Nov. 2009; SmartGridComm 2010, Gaithersburg, MD, Oct. 2010;
21–32. 327–332.
23. Kosut O, Jia L, Thomas RJ, Tong L. Malicious data 38. Kalogridis G, Efthymiou C, Denic SZ, Lewis TA,
attacks on the smart grid. IEEE Transactions on Smart Cepeda R. Privacy for smart meters: towards undetectable
Grid 2011; 2(4):645–658. appliance load signatures. IEEE SmartGridComm 2010,
24. Kosut O, Jia L, Thomas RJ, Tong L. Malicious data Gaithersburg, MD, Oct. 2010; 232–237.
attacks on smart grid state estimation: attack strategies 39. Qiu M, Gao W, Chen M, Niu J, Zhang L. Energy
and countermeasures. IEEE SmartGridComm 2010, efficient security algorithm for power grid wide are
Oct. 2010; 220–225. monitoring system. IEEE Transactions on Smart Grid
25. Wang X, Zhang L, Deokar A, Liang Q. Enhanced 2011; 2(4):715–723.
security and reliability with MIMO communications 40. Bouhafs F, Mackay M, Merabti M. Links to the future:
for smart grid. Security and Communication Networks communication requirements and challenges in the
2012; (published online). smart grid. IEEE Power and Energy Magazine 2012;
26. Xie L, Mo Y, Sinopoli B. Integrity data attacks in 10(9):24–32.
power market operations. IEEE Transactions on Smart 41. Farag H, Saadany E, Seethapathy R. A two ways com-
Grid 2011; 2(4):659–666. munication-based distributed control for voltage regu-
27. Xie L, Mo Y, Sinopoli B. False data injection attacks lation in smart distribution feeders. IEEE Transactions
in electricity markets. IEEE SmartGridComm 2010, on Smart Grid 2012; 3(1):271–481.
Oct. 2010; 226–231. 42. Viawan F, Karlsson D. Voltage and reactive power
28. Chen J, Abur A. Placement of PMUs to enable bad control in systems with synchronous machine-based
data detection in state estimation. IEEE Transactions distributed generation. IEEE Transactions on Power
on Power Systems 2006; 21(4):1608–1615. Delivery 2008; 23(2):1079–1087.
29. Zhao L, Abur A. Multi area state estimation using 43. Wang X, Liang Q. Stabilizing the power supply in
synchronized phasor measurements. IEEE Transactions microgrid using sensor selection. Globecom 2012,
on Power Systems 2005; 20(2):611–617. Anaheim, CA, Dec. 2012.
30. Bohli JM, Sorge C, Ugus O. A privacy model for 44. Mao R, Li H An efficient multiple access scheme for
smart metering. ICC Workshops 2010, Cape Town, voltage control in smart grid using WiMAX. ICC
May 2010; 1–5. 2012, Ottawa, ON, Jun. 2012.

Security Comm. Networks 2015; 8:2431–2439 © 2013 John Wiley & Sons, Ltd. 2439
DOI: 10.1002/sec