IEEE Electrification - December 2020 - 102

network over large geographical areas. The strong
dependence on communication networks makes MGs
vulnerable to cyberattacks. Therefore, it is crucial to
study the cyberattacks of MG systems as well as their
solutions. According to the different strategies performed by the attacker, various types of attacks can
be defined. Some of the most common cyberattacks
are: denial-of-service (DoS), false data injection (FDI),
and replay.
A DoS attack is a kind of attempt to make the system
resources unavailable. From a technology point of view,
attackers can fill buffers of user domains or kernel
domains, jamming the shared network medium to prevent measurements and actuator data from reaching
their destinations.
An FDI attack is a type of cyberattack in which data
integrity is modified among different cyberparts. For
instance, in MG systems, adversaries may launch attacks
through hacking remote terminal units, such as the sensors in substations. It is worth mentioning that FDI attacks
in different scenarios can also be called deception attacks or
malicious attacks.

A replay attack is a natural strategy in which valid
data transmission is fraudulently repeated or delayed.
For instance, attackers can replicate the data recorded
from the compromised sensors or actuators at a particular time.

Cyberattack Detections
For systems without sufficient security protection strategies, malicious attacks may induce damage to power supplies and, thus, lead to significant societal detriments or
the loss of human lives. Different from system faults
whose characteristics are generally known in advance and
act in a straight way to the system, cyberattacks usually
occur randomly and are performed in a coordinated fashion to prevent them from being detected. This leads to the
requirement of increasing the sensitivity of detection
schemes to cyberattacks. Taking the security issue into
consideration, the design and analysis of attack detection
schemes for MGs have been recognized as more attractive
and something that must be designed for both the network layer and the physical layer. Figure 3 presents a summary of various attack detection and mitigation methods.

Physical
Layer

Network
Layer

Attack Detection
Methodologies
* Pattern Detection
* Anomaly Detection
* Third-Party Detection

Principles
* Compare Data Packets With Attack Database
* Compare Network With Normal Performance
* Make Authentications to the Data Packets

Limitations
* Cannot Detect New Attacks
* Increase Communication Burden
* Increase Communication Burden

Methodologies
Principles
* State Estimation
* Estimate System States
* Statistics
* Monitor the Statistics of Measurements
* Observer/Filter
* Monitor the System With Residual Signals
* Data Based
* Build a System Model by Historical Data
* Model Decomposition
* Portion the System Into Several Subsystems
* Disturbance Decoupling * Eliminate the Effect From Disturbances

Limitations
* Cannot Detect Intelligent Attacks
* Cannot Detect Intelligent Attacks
* Difficult to Design
* Heavy Training Burden
* Undesirable in LargeScale System
* Difficult to Design

Network
Layer

Methodologies
* Rate Limiting
* Firewall
* Topology Resilience

Physical
Layer

Attack Mitigation
Principles
* Limit the Packet Transmitting Rate
* Compare the Data Packets With Blacklist
* Design Multiple Network Resources

Methodologies
Principles
* Time Delay
* Design the Controller
* Model Prediction
Based on Delay System
* Estimate States to Compensate
* Switching System
* Secure State Estimation the Packet Drops
* Design the Controller Based
on Switching System
* Make a State Estimation
From Corrupted Sensors

Figure 3. A summary of attack detection and mitigation methods.

102

I E E E E l e c t r i f i cati o n M agaz ine / DECEMBER 2020

Limitations
* Lower System Dynamics
* Cannot Deal With New Attacks
* Increase System Cost
Limitations
* Only Effective on DoS Attack
* Only Effective on DoS Attack
* Only Effective on DoS Attack
* Rely on the Number of
Good Sensors
IEEE Electrification - December 2020

Table of Contents for the Digital Edition of IEEE Electrification - December 2020
