On the Adaptive Protection of Microgrids: A Review on How

to Mitigate Cyber Attacks and Communication Failures

Abstract— One main challenge in the practical implementation attack rendering the communication link either insecure or
of a microgrid is the design of an adequate protection scheme in severed [6]-[7].
both grid connected and islanded modes. Conventional This paper aims to present an analysis of various protection
overcurrent protection schemes face selectivity and sensitivity schemes based on an adaptive protection technique that is
issues during grid and microgrid faults since the fault current level deployed with communication links between Intelligent
is different in both cases for the same relay. Various approaches
Electronic Devices (IEDs). Different types of attacks that may
have been implemented in the past to deal with this problem, yet
the most promising ones are the implementation of adaptive impact the communication network is described followed by a
protection techniques abiding by the IEC 61850 communication discussion over how various energy storage (ES) and hybrid ES
standard. This paper presents a critical review of existing adaptive (HESS) solutions could be used as a secondary measure in
protection schemes, the technical challenges for the use of classical assisting the system to fill the required fault current, even in the
protection techniques and the need for an adaptive, smart event of a communication failure.
protection system. However, the risk of communication link The paper is organized as follows: Section II addresses
failures and cyber security threats still remain a challenge in protection issues in a microgrid. Section III reviews existing
implementing a reliable adaptive protection scheme. A adaptive protection techniques. Section IV introduces the
contingency is needed where a communication issue prevents the
different types of attacks that could affect the communication
relay from adjusting to a lower current level during islanded
mode. An adaptive protection scheme is proposed that utilizes network. Section V discusses how ES or HESS could be used
energy storage (ES) and hybrid ES (HESS) already available in the to inject the required fault current given a relay remains in grid
network as a mechanism to source the higher fault current. Four connected mode and Section VI concludes the paper.
common grid ES and HESS are reviewed for their suitability in
feeding the fault while some solutions are proposed. II. MAIN PROTECTION CHALLENGES IN MICROGRIDS
Index Terms— Protection Scheme, Microgrids, Cyber Attacks, There are several challenges that should be taken into
Energy Storage, Hybrid Energy Storage Systems. consideration in designing a proper protection scheme for a
microgrid. The following subsections discuss some of the major
I. INTRODUCTION issues one may face in the proper operation of a microgrid
protection scheme.
Distributed energy resources (DERs) connected to a
microgrid and its integration with the main grid are the A. Selectivity and Sensitivity Issues
targets in achieving optimum operation of electric power
Two main characteristics that should be found in microgrid
system networks [1]. Reducing greenhouse gases generated by
system protection devices are selectivity and sensitivity
the conventional energy resources while increasing the
features. First, selectivity refers to the ability of the protection
reliability and the power quality for consumers who require
system to locate and classify a fault correctly.
uninterruptible power supplies are some of its main advantages.
Overcurrent (OC) relays should determine whether the fault
However, along with these benefits, microgrids introduce some
is internal or external to its zone. The protection algorithm
major technical challenges in terms of protection schemes.
should be able to distinguish between the main grid and local
First, dynamic changes in the configuration of the system
microgrid faults [8]-[9]. During main grid faults, the microgrid
requires monitoring the status of all distributed generation (DG)
needs to be capable of islanding to protect equipment in the
while automatically adjusting their relay settings [2].
system [10]-[11]. During microgrid faults, the function of the
The second problem is the introduction of bidirectional
protection scheme becomes more complicated as it is required
power flow. The large penetration of DERs based on an
to disconnect the smallest part of the faulty section from the
interfaced converter as shown in Fig. 1 will reduce the capacity
system [12]-[13]. It should be noted that the importance of the
of the short circuit current when the microgrid is disconnected
selectivity feature of the relays is to operate under various faults
from the grid (islanded mode) [3]-[4]. Unfortunately, this
while isolating the faulty part.
complication requires dynamic toggling of relay settings
Secondly, sensitivity refers to the fact that OC relays should
whether the relay is in grid connected or islanded mode, an
be able to detect the fault conditions in the system. The fault
operation which also requires a fast and reliable communication
should be cleared as quickly as possible to maximize safety
network [5]. Given this need, a question arises of how the
while minimizing system instability and damage to equipment
microgrid can remain protected even if it encounters a cyber-
[14]. The OC relay should quickly arrive at a decision and
This work was partially supported by grants for the US Department of circuit breakers (CB) must function rapidly [15]. The sensitivity
Energy and the Office of Naval Research. The authors are with the Energy of the relays must be adjusted such that a high redundancy can
Systems Research Laboratory, Department of Electrical and Computer be achieved without affecting selectivity of the protection
Engineering, Florida International University, Miami, FL 33174 (e-mail:
[email protected]).
system [16]-[18].
Fig. 1. The Configuration of a Typical Microgrid.

B. Reducing the Short Circuit Fault Current level with the connection or disconnection of DERs affecting
The high penetration of DERs based on power electronic the reach of the OC relay settings [29]-[30].
inverters interfaced with the microgrid has a great impact on the D. Dynamic Changes in the Microgrid Architecture
protection scheme. These systems limit the short circuit current
Dynamic changes of the microgrid configuration should be
during islanded mode of operation, as they are equipped with a
taken into consideration when designing the protection scheme.
fault limited current (FCL) that prevents high OC levels to flow
Some changes include the disconnection of generation units,
during the fault period [19]-[21]. A significant difference exists
shutting down some loads during peak hours, or exporting of
in the amount of short circuit current that occurs when in grid
power to the main grid during excess generation periods for
connected and islanded modes of operation. In grid connected,
optimum and economic operations [25]. Adaptive protection is
fault currents of a much higher magnitude (5-10 times the full
required to change the relay settings according to the current
load current) are available to help conventional OC relays to
microgrid configuration. Prior knowledge of every state in the
activate under abnormal operations. However, in islanded
microgrid as well as online monitoring and calculation of the
mode, the fault current reduces to only about 3 times the full
short circuit fault current is needed for proper operation. As
load current [12]. Furthermore, a large integration of connected
previously discussed, this requires the application of a fast,
converter-based DERs further reduces the fault current to only
reliable, and robust communication system [31].
1.2-1.5 times the full load current [22]-[24]. Usually,
conventional OC relays are set to operate at 1.2-8 times full load
current. Accordingly, the time-current coordination of OC III. EXISTING TECHNIQUES FOR ADAPTIVE PROTECTION
relays and OC devices with extremely inverse characteristics Adaptive protection refers to an online process which
are disturbed. modifies the preferred protective responses and correlates them
to a change in system conditions or requirements in a timely
C. The Direction of Power Flow manner through control or signaling [8]. Adaptive protection
Distribution systems become active systems due to the consists of a smart control unit that monitors the grid
integration of DERs in the system [25]. The microgrid feeds parameters including a smart fault diagnostic system to detect
local loads and can also support power to the grid in the case of the fault, the fault type, identify the faulted zone, and a smart
excess generation. This operation changes the flow of the power relaying system which protects the system against faults.
from unidirectional to bidirectional. Bi-directional power flow Adaptive protection of the microgrid can be realized with the
has effects on the amplitude and direction of the fault current use of microprocessor-based directional OC relays (DOCR).
thereby effecting coordination of the protective relays [19],[26]. DOCR have the possibility to choose different tripping curves
In this case, the protection system in a typical distribution to suit a particular system. An adaptive protection scheme can
system is designed accordingly to a time or current coordination be implemented with centralized or decentralized control
principle, in which the relay closest to the fault operates first approaches, but each requires a different communication
and in the event of a failure, a backup protective relay operates architecture [32].
after a specific delay time (~200 ms). The centralized control architecture communicates with each
However, the effectiveness of this protection requires a radial IED and adjusts their settings accordance with the configuration
grid connection, which is no longer the case when DGs are of the present microgrid as shown in Fig.1. Centralized
connected to the network. This results in either changing or communication architecture is supported by various
completely losing coordination between protective devices communication protocols including Modbus, DNP3, IEC
[27]-[28]. The different adverse effects of DG connections on 60870-5-101/104, IEC 61850 and can be implemented with a
the distribution network protection include false tripping of serial/bus communication over a PLC (power-line carrier) or
feeders, blinding of protection, increase or decrease in the fault via an Ethernet network [33].
 The decentralized control architecture depends on sending IV. CYBER PHYSICAL ATTACKS
and receiving data between IEDs. Each IED can change its Security attacks in microgrid communication networks are
settings according to the data that it receives from other IEDs. intended to interrupt messages that are exchanged between
Presently, the industry is focused on the IEC 61850 standard as IEDs. The attacks can be implemented using common network
the protocol for decentralized communication [34]-[35]. The security attacks or by exploiting a Generic Object Oriented
microgrid protection scheme requires an adaptive, smart, and Substation Event (GOOSE) and Sampled Value (SV) messages
upgradable protection algorithm. A survey of existing key as shown in Fig. 2.
adaptive protection schemes are presented in this section.
In [16], a novel adaptive protection technique is proposed A. Network Security Attacks
using extensive communication and digital relays. Settings of This attack is performed through the communication network
the relays are changed according to the microgrid configuration. to damage, change, or access user data without permission [41].
The technique is further extended by using numerical The following are several types of network security attacks:
directional relays with a directional interlock capability for
selective operation. Additional adaptive protection schemes 1) Denial of Service
were introduced in [36]-[37] with advanced communication Denial of Service (DoS) is the process of preventing the
technologies for updating the relay settings in accordance with authorized user of a service to access that service [41]. Main
the microgrid configuration. The authors proposed a central types of DoS that may affect the communication between
protection unit to be equipped with the microgrid to change the protective devices include SYN (synchronization) floods and
tripping characteristics of the relays with every interruption call buffer overflow. For the first type, the attacker sends a
for the connection and disconnection of DGs. Reference [17] continuously spoofed SYN request to the target IED to distort
studied the fault current of an inverter-based microgrid and the connection between the authorized user and IEDs [42]. This
proposed an adaptive protection algorithm for the fault current. type of attack can be orchestrated by running several protocol
This technique depended on comparing the impedance of the services like FTP, HTTP and Telnet on the IED at the same time
grid and microgrid using the voltage and current fault [43]-[44]. For the second type, the attacker may transmit
components in a real time manner. In this way, the relay malicious code to an IED and write oversized data to cause a
updated its settings by observing the changing impedance of the buffer overflow. This attack is possible as a result of the
utility and microgrid. Refrence [28] suggested a system with a vulnerability of IEDs and the unavailability of security
high penetration of DGs that loosened the coordination of the measures for IEDs to detect the malicious code [42].
protective relays due to bidirectional power flow in the system.
2) Password Cracking Attempts
A centralized control center communicated online with modern
CBs in the system to send a trip signal under the fault condition. This type of attack is defined as an attempt to gain access to
Fault contribution from the sources was obtained online an IED, another device, or a system by guessing the password
through a Thevenin equivalent impedance. Under abnormal [45]. Attackers who are able to access an IED can send a false
conditions, a change in the Thevenin impedance of the tripping signal to the CB that can cause it to disconnect a
adjoining fault bus source was detected. Thus, the relay healthy part in the system. This can be accomplished in two
determined the fault and tripped the appropriate CBs. In [38], ways: a brute force attack and a dictionary attack. A brute force
the authors used numerical relays to detect the fault current is performed by arranging all possible combinations of
under grid connected and islanded modes. Different trip passwords and trying these combinations one-by-one to achieve
characteristics settings were calculated in both modes and saved the correct password but can take a long time. In a dictionary
in the relays. When the microgrid switched into islanded mode, attack, the password is simply guessed. This type may take less
the relay automatically changed a group of settings that time than the latter. In [43] and [44], hackers were able to utilize
matched this mode. FTP, HTTP and Telnet services already running on the IED to
Previously in [4], a supercapacitor with a DC-DC converter crack the password.
was added to each DER with the same rated power that would
add to the cost of the protection system, especially given Cyber Attacks
supercapacitors were not originally used during normal
operation. In [23], a technique is presented to protect the
microgrid based on a voltage measurement of the fault detection Network Goose & SMV
Security Attacks Messages
modules, but it may not be accurate to determine the faulty part
of the system and may also require time. This would have direct Denial of Modification
Service Attack
impacts on selective protection of the microgrid. Reference [39] Password
Cracking DoS
introduces communication failures as a critical issue which is Attack
now being widely presented in literature. The operation of the SYN Buffer Replay
Eavesdropping
Flood Overflow Attack
relay depends on the voltage drop during the fault that may
Brute
require many loads to disconnect from the system. Finally, in Force
Dictionary
[40] different techniques that could be used to protect the
microgrid are reviewed. In [40], a direction relation to our work Switch Port
ARP CAM
is identified as it highlights the importance of communication Stealing
in protection schemes which cannot be carried out in a cyber-
Fig. 2. Types of Cyber Attacks.
attack and identifies the need for another solution.
3) Eavesdropping Attacks
Eavesdropping attack is an attempt to steal packets that are
being transmitted through the system. This type of attack is
launched from within the Local Area Network (LAN) and can
target FTP, HTTP and Telnet services because messages for
these protocols are not encrypted [43]. There are three types of
this attack. The first is known as Address Resolution Protocol
(ARP) cache poisoning. ARP is a communication protocol that
converts an IP address into incorrect MAC address (i.e. the
Fig. 3. GOOSE Poisoning Attack.
MAC address of the attacker), the switch will end up
forwarding all packets addressed to the attacker and enables the overwhelmed and is unable to respond to a legitimate request.
attacker to capture those packets. The second type is to cause GOOSE poisoning attacks are another way to perform DoS as
Content Addressable Memory (CAM) table flooding. This is shown in Fig. 3. The aim of these attacks is to get the subscriber
done by filling in the CAM table of the switch with fake entries. to accept GOOSE messages with a higher sequence number
When it is full, packets addressed to a MAC address that are not than those sent by the publisher. As a result, all GOOSE
in the CAM table will be broadcasted to the whole network messages from the publisher will be considered invalid by the
allowing attackers to capture them. Finally, the third type is subscribers leaving the subscribers to only accept GOOSE
switch port stealing where fake frames are sent to the switch messages from the attacker. There are three variables of
that contain the target host MAC address causing the switch to GOOSE poisoning attacks including high status number
modify the CAM table. This allows the MAC address to attacks, high rate flooding attacks, and semantic attacks. For the
connect to the interface that leads to the attacker [46]. high status number attack, the attacker sends a single spoofed
GOOSE frame with a very high status number to a GOOSE
B. Attacks on GOOSE and SV Messages
subscriber by the attacker. Secondly, in a high rate flooding
There are two multicast message protocols used in IEC attack, the attacker multicasts a range of spoofed GOOSE
61850 which include GOOSE and SMV messages. The main messages with higher status numbers. The spoofed GOOSE
function of the GOOSE message is to send a trip signal to the frames employ a status number higher than the expected status
CB to isolate the faulty part from the system. SV messages are number of a GOOSE subscriber. In a semantic attack, the
used to send voltage and current values from merging units to attacker will fix the status number in the GOOSE message and
the protective devices. Both messages use a switched Ethernet determine the rate of status change. Then, the attacker will
network for communication purposes. These messages must be spoof GOOSE messages that are higher than the detected rate
transmitted within 4 ms. The following section discusses of status change [47].
several attacks that exploit GOOSE and SMV messages.
3) GOOSE and SV Replay Attacks
1) GOOSE and SV Modification Attacks In this attack, GOOSE messages are captured and kept by the
In this type, the attacker changes the message that is attacker. Then, the attacker sends a message to trip the CB
exchanged between the protective devices without allowing the under normal operation which may result in an undesirable
publisher (the sender of GOOSE messages) or subscriber (the action. For a SV message replay attack, the attacker can also
receiver of GOOSE messages) to notice. Two types can be capture a SV packet containing certain values of power and
found in this attack. For the first type, the attacker captures the current and send it to another protective device in the substation
GOOSE message and modifies it with another message that several times. SV packets with the same power and voltage
enables the attacker to control a CB. For a SV packet, an values circulating inside the system can lead to an unplanned
attacker sends a fabricated analog value to a control center in outage [48].
the system that leads to gain control of IEDs and cause a power
outage [46]. The second type of attack is performed using a V. ENERGY STORAGE AS A CONTINGENCY
malware script [45]. The malware can capture the message that
is exchanged between IEDs altering and re-injecting GOOSE The previous section revealed many communication
message packets into the IEC 61850 network. In order to enable vulnerabilities which can be exploited by hackers. However,
the malware to perform its job, it must be installed on a clearly not all communication outages are related to cyber-
attacks. The loss of adequate communication to a relay can be
computer inside the network. Attackers exploit the weakness in
catastrophic to a microgrid protection scheme where the most
GOOSE where encryption and digital signatures cannot be
difficult aspect is handling a dramatically increased fault
applied due an IEC 61850 requirement that any operation which
is communicated using a GOOSE message must take effect current magnitude when moving from islanded to grid
within 4 ms. Without encryption and digital signatures, a connected mode. In the event of a cyber-physical attack or loss
of communication, a protection scheme would be paralyzed
transmitted packet can be easily captured, modified, and
where relays will no longer be capable in adjusting OC limits.
retransmitted into the network.
Assuming a cyber-physical attack or communication loss has
2) GOOSE and SV Denial of Service Attacks taken place, relays will now require 5-10 times the full load
current to trip in islanded mode. In this case, ES devices already
These attacks prevent IEDs to respond to legitimate
present in the system for other purposes could now also be
messages made by other IEDs. This can be done by sending a
tasked in assisting to increase the fault current magnitude.
large number of messages to the target IED, where it becomes
 TABLE I should be very large to limit the peak array current. Finally,
COMPARISON OF FOUR PROMINENT ENERGY STORAGE RESOURCES
their shelf and cycle life are highly limited.
Energy Energy Power Energy Response Self Lifespan
Storage Density Density Cost Time Discharge Max. Service 2) Lithium Ion Batteries
Type (Wh/L) (W/kg) ($/kWh) Rate Cycles (years)
LI battery usage has surged in recent years not only in
Lead Acid
85 180 8.50 Slow 3-20%/mo 1,500 3-12 portable electronics, but also in large scale EVs and even grid
Battery
Lithium Ion
storage [55]. Advanced LI battery management systems have
463 295 250.00 Medium 6-8%/mo 4,200 5-20 been demonstrated in microgrid applications for both islanded
Battery
Supercapacitor 10 3,500 1,000.00 Very Fast 1-2%/day 1,000,000 10-35 and grid-connected modes to provide voltage and frequency
Flywheel 8 5,000 500.00 Fast 3-40%/hr 10,000,000 20-35
support [56]-[57]. Since LI battery management is more
complex than that of the LA, studies have looked at the best
In this section, four types of common grid ES devices are method to control these schemes [58]-[59].
reviewed, compared, and discussed for their suitability to inject Operation is significantly different from that of LA, where
the required fault current. It is assumed that the microgrid under energy is stored inside its electrodes utilizing the electrolyte as
study already has one or multiple ES devices connected to the simply a transfer layer [60]. This combined with a smaller cell
network, hence no cost would be added to the system. Thus, construction allow them to respond faster to a pulsed load
their participation in supporting the required fault current would which in turn, would also make them suitable in feeding a fault
be classified as another capability. Next, the concept of a HESS current. LI batteries offer a significant improvement in their
is presented as well as some examples of their usage in grid capability to source high current without the same trade-offs in
applications. Following a review, some potential single ES and lifespan as seen with the LA battery. Their lifespans are much
HESS solutions are discussed as well as considerations and longer, but similar to the LA battery, are still limited by
trade-offs in deploying them. excessive operational currents and particularly temperature
[61]. Battery management system cost for LI batteries is
A. Common Grid Energy Storage Types
increased as a result of required crucial cell balancing and
The following subsections will review the operation of four thermal control for safety. Finally, a number of companies have
common grid ES types devices: Lead Acid (LA) batteries, been working to reduce the cost of grid-scale LI battery arrays
Lithium Ion (LI) batteries, Supercapacitors (SC), and Flywheel as their cost is still at a premium as compared to LA.
(FW) ES. Table I depicts a comparison of each ES type as well
as the categories considered. The operation of each type, their 3) Supercapacitors
strengths and weaknesses, and examples of previous grid The SC provides a significant increase in the speed of
applications are discussed. response versus any electrochemical battery. Although their
construction is chemical in nature, no reaction takes place. This
1) Lead Acid Batteries
enables them to respond extremely fast to a demand while even
The LA battery has maintained a strong hold in the market as under heavy current, their lifetime is virtually unaffected [62].
a result of its simplicity in design and inexpensive materials Composed of two porous electrodes divided by a separator
[49]. Despite the fact that emerging electric vehicles (EV) have soaked in a solvent electrolyte, their construction enables a
moved onto LI types, the LA market remains strong as it is still much higher charge density versus the traditional capacitor as a
the most common starter battery and battery in uninterruptible result of an increased surface area [63]. Their usage has been
power supplies (UPS) [50]. Furthermore, some of the studied in mobile shipboard applications where weight is a
drawbacks in deploying them in EVs are relieved when placing concern, but also in some grid applications with multiple
them in a stationary grid application. While real estate may still renewable energy resources where they provide short term
remain a premium, concerns with weight can be alleviated. The storage to supply the deficiency power [64]-[65].
LA battery has been demonstrated as a dependable resource in Unfortunately, their low energy density can require an
stationary grid applications to smoothen the energy harvested enormous capacitance which may not prove to be practical as a
by renewables, but has been more universally accepted in sole ES device. Although their response time in feeding the
restoring system frequency and voltage following an outage required power would be superb, sourcing the total energy
[51]-[53]. Despite its strengths in reliability and low cost, LA required to fill a fault current would come at a massive cost
has a relatively low energy and power density and suffers from [66]. Its energy density is only around 10% of that of the LA
a number of drawbacks. LA are not ideal in sourcing high battery and 2% of a LI battery and their self-discharge rate is
frequency pulsed loads as a result of its large double layer relatively high. Furthermore, of all ES included in this survey,
capacitance, a weakness which would also reduce their they have the highest cost.
suitability to solely supply a fault current [54].
Its operation utilizes a sulfuric acid electrolyte as a fuel 4) Flywheel Energy Storage
primarily governed by changes in the concentration, an Like the LA battery, FW ES has been synonymous with
inefficient process which leads to a number of lifespan industrial UPS systems, but for a very different purpose. FW
concerns. First, their operational current is severely limited, as ES provides some of the highest power density in our study and
an increase from a conservative 20-hour discharge rate (C/20) are primarily purposed to support to pulsed loads which would
would result in reduced usable capacity and increased ageing. initially make them appear to be a great candidate to inject a
Ergo, in order to fill a fault current, the size of the LA array
fault current [67]. Analogous to an electromechanical battery, Fault
Breaker
the FW stores kinetic energy in a high inertia rotating mass Current
Tripped
Reached
where an electric machine operates simultaneously as a motor (5-10)I pu
during charging and generator during discharging. The power
output is a function of a square of the speed allowing them to

LA
provide extremely high power density. This has made them an

LI
excellent solution for maintaining power quality [68]-[69].

FW
They have also been tested for their usage in both wind and I pu

SC
solar applications as a mechanism to quickly store and expend
energy [70]. FW also carry a unique capability unlike other ES: t0 t1 t2 t3 t4 tf
the ability to supply inertia to the system [71]. Unfortunately, ~ 80 ms
the FW has a very high self-discharge rate as a result of friction
Fig. 4. Controlled Hybrid Energy Storage to Feed the Fault.
losses and like the SC, a very low energy density [72]. Although
their cost is half that of the SC, their response time is still very 1) Single Source Solutions
fast, limited only by the initial inertia required to start moving For a microgrid with only one type of ES, tasking it to assist
the rotating mass. However, their energy density is very low, in feeding the fault is more complicated. First, a system which
some 20% below the SC which would further reduce their only contains a SC or FW could require a very large and
capability in solely injecting the fault current. expensive infrastructure to be feasible. The two most suitable
devices in this case are the LI and LA batteries, as they provide
5) Hybrid Energy Storage Systems the greatest balance. The LA battery has a much lower cost, but
Although some ES devices such as the LI or LA battery can its response time is much slower than that of the LI battery
provide a rather balanced contribution of energy versus power which could prolong the fault period resulting in equipment or
density, cost and lifetime aspects could jeopardize their sole infrastructure damage. Furthermore, a large array would be
integration with the grid. Similarly, the integration of a SC or needed or an excessive current magnitude would damage the
FW alone could provide excellent voltage and frequency batteries over time. The LI battery has the best response time of
support for a small outage, but would fail to fill energy demands the 2 battery types and would require a smaller array as a result
over longer periods without a huge system. For this reason, of being able to operate at higher current. Despite a higher cost,
HESS could provide a more balanced solution in terms of not LI batteries already present in the system would be the most
only power and energy density, but also cost, lifespan, and self- feasible sole source for the fault current.
discharge. HESS have emerged in an effort to utilize the
strengths of multiple ES devices in a way that is not only more 2) Availability of all Energy Storage Devices
efficient, but potentially cost and lifetime effective. Recent In the event that all resources in Table I are available on the
HESS for microgrid applications have focused on the network, a combination of the strengths from all ES devices
collaboration of batteries with SCs, evaluating their could be used to feed the fault current in a dynamically switched
combination with both LA and LI [73]-[74]. Minimizing losses SC-FW-LI-LA HESS. Fig. 4 breaks down the initial concept
while ensuring an optimum power split between the two sources which involves contribution from each ES to the surge pulse
has spawned research in this topic [75]. A great deal of work broken into 4 switching steps. An ES type is first connected at
has focused specifically on pulsed load management through the start of the fault (𝑡0 ), where each of the remaining 3 devices
the deployment of HESS systems [76]-[77]. A pulsed load are sequentially added in parallel until the final required fault
typically contains a very high power demand over a relatively current level has been reached at 𝑡4 tripping the breaker (𝑡𝑓 ).
short period of time. Unfortunately, reaching the required level The proposed sequence begins with the deployment of the SC
of power and filling the energy demand over such a short period at 𝑡0 as it responds quickest to the demand initially injecting
of time can be challenging. Examples can be found in a number high power until reaching 𝑡1 , where the FW assists in
of Naval weapons platforms, the starter current for a vehicle, increasing the power even faster to get within the range of the
and the case addressed in this paper: a fault current. fault current level while concurrently contributing inertia to the
system. Since the SC and FW have low susceptibility to lifespan
B. Potential Solutions to Source the Fault Current
issues, the selection of 𝑡1 would primarily rely upon the
As previously discussed, each ES device has its own availability of each to provide the energy to reach 𝑡2. Since the
respective strengths and weaknesses and few would be suitable energy stored in both is relatively small, the LI battery would
to solely source a fault current. A general example of a fault be added at 𝑡2 to assist in maintaining the energy required to fill
current is shown in Fig. 4, where a high in-rush current is the pulse. Since the LI battery provides a good balance of both
needed for approximately 80 ms to reach the requiring between power and energy, it can assist in feeding the required energy
5 and 10 times the base load current Ipu . A multitude of while simultaneously increasing the power until the LA battery
solutions could assist in injecting the required fault current, is added at 𝑡4 to reach the fault current level. A trade-off can
however, they are limited by both the available infrastructure as once again be made between the availability of LA and LI
well as the proximity of each ES device from the CB. In the capacities, but now involving lifespan as well. The cost of LI
following subsections, 2 scenarios are presented representing array may force the operator to require more support from LA,
microgrids which already have a single ES element or the shifting the location of 𝑡3 . However, the severe consequences
availability of all sources in Table I on the network located a in overusing LA should be considered as their lifespans would
reasonable distance from the location of the CB. take the highest toll.
 VI. CONCLUSION [11] N. Schaefer, T. Degner, A. Shustov, T. Keil and J. Jaeger, "Adaptive
protection system for distribution networks with distributed energy
This paper presents the technical issues that should be taken resources," in 10th IET International Conference on Developments in
into consideration when designing an adequate microgrid Power System Protection (DPSP 2010). Managing the Change,
Manchester, pp. 1-5, 2010.
protection scheme. Adaptive protection is required to protect a [12] S. Chowdhury, S.P. Chowdhury, P. Crossley, “Microgrids and active
microgrid under grid connected and islanded modes of distribution networks”, in IET Renewable Energy Series 6, The Institution
operation. Most of the proposed adaptive protection methods of Engineering and Technology, London, United Kingdom, 2009.
[13] R. Lasseter, et al., “Integration of distributed energy resources”, in The
depend on exchanging information amongst IEDs via of a
CERTS Microgrid Concept, Lawrence Berkeley National Lab, 2002.
communication link. Therefore, protection against cyber- [14] H. Laaksonen, K. Kauhaniemi, S. Voima,” Protection system for future
attacks is crucial to ensure a safe communication link. Different LV Microgrids”, in CIRED 21st International Conference on Electricity
types of cyber-attacks were introduced that may impact the Distribution, Frankfurt, 6–9June, 2011.
[15] J. Tang, Y. Gong, N. Schulz, M. Steurer and P. G. McLaren, "Implemen-
communication network keeping in mind these as well as a total tation of a Ship-Wide Area Differential Protection Scheme," in IEEE
loss of communication could be catastrophic to a microgrid Transactions on Industry Applications, vol. 44, no. 6, pp. 1864-1871,
protection scheme. Utilizing ES devices is discussed as a Nov.-Dec. 2008.
[16] A. Oudalav and Antonio Fidigatti ,“Adaptive Network Protection in
backup mechanism following a communication failure. A brief Microgrids,” in International Journal of Distributed Energy Resources,
analysis for each type of ES devices was proposed to investigate vol.5, pp. 201-225, 2009.
the ability of lead acid and lithium ion batteries, [17] Y. Han, X. Hu and D. Zhang, "Study of adaptive fault current algorithm
for microgrid dominated by inverter based distributed generators," in The
supercapacitors, flywheel energy storage, or a hybrid system to 2nd International Symposium on Power Electronics for Distributed
feed the fault current. Generally, future microgrid protection Generation Systems, Hefei, China, pp. 852-854, 2010.
schemes are most likely to be a combination of traditional [18] M. Dewadasa, A. Ghosh and G. Ledwich, "Protection of microgrids using
differential relays," AUPEC, Brisbane, QLD, pp. 1-6, 2011.
effective protection schemes and new adaptive communication- [19] Martin Geidl, “Protection of Power Systems with Distributed Generation:
based protection schemes using pattern recognition approaches, State of the Art,” Swiss Federal Institute of Technology (ETH) Zurich,
synchrophasors, fault current limiters, but in the event of a 20th July 2005.
[20] P. Mahat, Z. Chen, B. Bak-Jensen and C. L. Bak, "A Simple Adaptive
communication failure could include ES or HESS as a
Overcurrent Protection of Distribution Systems With Distributed
contingency to meet the required fault current to trip a relay. Generation," in IEEE Transactions on Smart Grid, vol. 2, no. 3, pp. 428-
437, Sept. 2011.
REFERENCE [21] C. A. Plet, M. Graovac, T. C. Green and R. Iravani, "Fault response of
grid-connected inverter dominated networks," in IEEE PES General
[1] M. A. Haj-ahmed and M. S. Illindala, "Investigation of Protection Meeting, pp. 1-8, Minneapolis, MN, 2010.
Schemes for Flexible Distribution of Energy and Storage Resources in an [22] H. Laaksonen, K. Kauhaniemi, “Fault type and location detection in
Industrial Microgrid," in IEEE Transactions on Industry Applications, islanded Microgrid with different control methods based converters”, in
vol. 51, no. 3, pp. 2071-2080, May-June 2015. CIRED 19th International Conference on Electricity Distribution, pp. 1–
[2] O. V. Gnana Swathika and S. Hemamalini, "Prims-Aided Dijkstra 4, 21–24 May, Vienna, 2007.
Algorithm for Adaptive Protection in Microgrids," in IEEE Journal of [23] T. Loix, T. Wijnhoven and G. Deconinck, "Protection of microgrids with
Emerging and Selected Topics in Power Electronics, vol. 4, no. 4, pp. a high penetration of inverter-coupled energy sources," in CIGRE/IEEE
1279-1286, Dec. 2016. PES Joint Symposium Integration of Wide-Scale Renewable Resources
[3] H. F. Habib; T. Yossef; M. Cintuglu; O. Mohammed, "A Multi-Agent Into the Power Delivery System, Calgary, AB, pp. 1-6, 2009.
Based Technique for Fault Location, Isolation, and Service Restoration," [24] D. Klapp and H. T. Vollkommer, "Application of an Intelligent Static
in IEEE Transactions on Industry Applications , vol. no.99, pp.1-1, Feb. Switch to the Point of Common Coupling to Satisfy IEEE 1547
2017. Compliance," in IEEE Power Engineering Society General Meeting,
[4] K. O. Oureilidis and C. S. Demoulias, "A Fault Clearing Method in Tampa, FL, pp. 1-4, 2007.
Converter-Dominated Microgrids With Conventional Protection Means," [25] M. H. Cintuglu, T. Ma and O. A. Mohammed, "Protection of Autonomous
in IEEE Transactions on Power Electronics, vol. 31, no. 6, pp. 4628- Microgrids Using Agent-Based Distributed Communication," in IEEE
4640, June 2016. Transactions on Power Delivery, vol. 32, no. 1, pp. 351-360, Feb. 2017.
[5] M. A. Haj-ahmed and M. S. Illindala, "The Influence of Inverter-Based [26] N. K. Choudhary, S. R. Mohanty and R. K. Singh, "A review on Microgrid
DGs and Their Controllers on Distribution Network Protection," in IEEE protection," in International Electrical Engineering Congress (IEECON),
Transactions on Industry Applications., vol. 50, no. 4, pp. 2928-2937, Chonburi, pp. 1-4, 2014.
July-Aug. 2014. [27] A. A. Kalage and N. D. Ghawghawe, "A protection scheme for optimum
[6] Hany F. Habib, A.A.S. Mohamed, Mohamad El Hariri, Osama A. coordination of directional overcurrent relays in presence of distributed
Mohammed, “Utilizing supercapacitors for resiliency enhancements and generation," in International Conference on Electrical, Electronics,
adaptive microgrid protection against communication failures,” in Signals, Communication and Optimization (EESCO), Visakhapatnam, pp.
Electric Power Systems Research, Vol. 145, pp. 223-233, April 2017. 1-5, 2015.
[7] H. Wan, K. K. Li and K. P. Wong, "An Adaptive Multiagent Approach to [28] S. M. Brahma and A. A. Girgis, "Development of adaptive protection
Protection Relay Coordination With Distributed Generators in Industrial scheme for distribution systems with high penetration of distributed
Power Distribution System," in IEEE Transactions on Industry generation," in IEEE Transactions on Power Delivery, vol. 19, no. 1, pp.
Applications, vol. 46, no. 5, pp. 2118-2124, Sept.-Oct. 2010. 56-63, Jan. 2004.
[8] W.E. Feero, D.C. Dawson, J. Stevens, “Protection Issues of the Microgrid [29] A. K. Tiwari, S. R. Mohanty and R. K. Singh, "Review on protection is-
Concept”, White Paper by the Concortium for Electric Reliability sues with penetration of distributed generation in distribution system," in
Technology Solutions, March 2002. International Electrical Engineering Congress (IEECON), Chonburi, pp.
[9] G. Buigues, A. Dysko, V. Valverde, I. Zamora and E. Fernandez, 1-4, 2014.
“Microgrid Protection: Technical Challenges and Existing Techniques,” [30] M. Baran and I. El-Markabi, "Adaptive over current protection for distri-
in International Conference on Renewable Energies and Power Quality, bution feeders with distributed generators," in IEEE PES Power Systems
Bilbao (Spain), 20th to 22th March, 2013. Conference and Exposition, pp. 715-719 vol.2. 2004.
[10] P. Mahat, et al., "Islanding Operation of Distribution System with [31] S. J. Lee, B. W. Min, K. H. Chung, M. S. Choi, S. H. Hyun and S. H.
Distributed Generations", in Proceedings of the Danish PhD Seminar on Kang, "An adaptive optimal protection of a distribution system using a
detailed Modelling and Validation of Electrical Components and Systems, multi-agent system," in Eighth IEE International Conference on Devel-
pp. 23-26, 2010. opments in Power System Protection, pp. 611-614 Vol.2, 2004.
[32] M. A. Zamani, T. S. Sidhu and A. Yazdani, "A Protection Strategy and [56] G. Diaz, C. Gonzalez-Moran, J. Gomez-Aleixandre, A. Diez, “Scheduling
Microprocessor-Based Relay for Low-Voltage Microgrids," in IEEE of droop coefficients for frequency and voltage regulation in isolated
Transactions on Power Delivery, vol. 26, no. 3, pp. 1873-1883, July 2011. microgrids”, in IEEE Trans. Power Syst., vol.25, pp.489–496, 2010.
[33] N. Hatziargyriou, “Microgrids: Architectures and Control”, John Wiley [57] Z. Miao, L. Xu, V. R. Disfani, L. Fan, “An SOC-Based Battery
& Sons,West Sussex, UK, 2013. Management System for Microgrids”, in IEEE Trans. on Smart Grid, vol.
[34] I. Almutairy, "A review of coordination strategies and techniques for 5, no. 2, pp. 966–973, Mar. 2014.
overcoming challenges to microgrid protection," in Saudi Arabia Smart [58] J.-Y. Kim, J.-H. Jeon, S.-K. Kim, C. Cho, J. H. Park, H.-M. Kim, K.-Y.
Grid (SASG), Jeddah, Saudi Arabia, pp. 1-4, 2016. Nam, “Cooperative Control Strategy of Energy Storage System and
[35] H. J. Laaksonen, "Protection Principles for Future Microgrids," in IEEE Microsources for Stabilizing the Microgrid during Islanded Operation”,
Transactions on Power Electronics, vol. 25, no. 12, pp. 2910-2918, Dec. in IEEE Trans. on Power Electron., vol.25, pp. 3037–3048, 2010.
2010. [59] H. Karimi, H. Nikkhajoei, R. Iravani, “Control of an Electronically-
[36] Taha Selim Ustun, Cagil Ozansoy and Aladin Zayegh, “Fault current Coupled Distributed Resource Unit Subsequent to an Islanding Event”, in
coefficient and time delay assignment for microgrid protection system IEEE Trans. on Power Delivery, vol.23, pp. 493–501, 2008.
with central protection unit”, in IEEE Transactions on power systems, vol. [60] M. Doyle, J. Newman, A. S. Gozdz, C. N. Schmutz, and J.-M. Tarascon,
28, no. 2, pp. 598-606, 2013. “Comparison of Modeling Predictions with Experimental Data from
[37] T. S. Ustun and R. H. Khan, "Multi-terminal Hybrid Protection of Plastic Lithium Ion Cells,” in Journal of the Electrochemical Society, vol.
Microgrids Over Wireless Communications Network," in IEEE 143, no. 6, pp. 1890–1903, Jun. 1996.
Transactions on Smart Grid, vol. 6, no. 5, pp. 2493-2500, Sept. 2015. [61] B. Sood, M. Osterman, and M. Pecht, “Health monitoring of lithium-ion
[38] M. Khederzadeh, "Adaptive setting of protective relays in microgrids in batteries,” in IEEE Symposium on Product Compliance Engineering
grid-connected and autonomous operation," in 11th IET International (ISPCE), pp. 1–6, 2013.
Conference on Developments in Power Systems Protection (DPSP), [62] “Supercapacitor Information – Battery University.” [Online]. Available:
Birmingham, UK, pp. 1-4, 2012. http://batteryuniversity.com/learn/article/whats_the_role_of_the_superca
[39] E. Sortomme, S. S. Venkata and J. Mitra, "Microgrid Protection Using pacitor. [Accessed: 10-Feb-2017].
Communication-Assisted Digital Relays," in IEEE Transactions on [63] F. Béguin, E. Frąckowiak, and A. Burke, “Testing of Electrochemical
Power Delivery, vol. 25, no. 4, pp. 2789-2796, Oct. 2010. Capacitors”, in Supercapacitors, Eds. Wiley-VCH Verlag GmbH & Co.
[40] S. Mirsaeidi, D. M. Said, M. W. Mustafa, M. H. Habibuddin, and Kimia KGaA, pp. 437–471, 2013.
Ghaffari, "An analytical literature review of the available techniques for [64] J. P. Kelley, D. A. Wetz, J. A. Reed, I. J. Cohen, G. K. Turner, W. Lee,
the protection of micro-grids," in Int. J. Elect. Power Energy Syst., vol. “The impact of power quality when high power pulsed DC and continuous
58, pp. 300-306, Jun. 2014. AC loads are simultaneously operated on a MicroGrid testbed”, in IEEE
[41] Understanding Denial-of-Service Attacks | US-CERT.” [Online]. Electric Ship Tech. Symp., pp. 6–12, 2013.
Available: https://www.us-cert.gov/ncas/tips/ST04-015. [Accessed: 08- [65] P. Thounthong, A. Luksanasakul, P. Koseeyaporn, and B. Davat,
Mar-2017]. “Intelligent Model-Based Control of a Standalone Photovoltaic/Fuel Cell
[42] K. Choi, X. Chen, S. Li, M. Kim, K. Chae, JC. Na, “Intrusion Detection Power Plant With Supercapacitor Energy Storage,” in IEEE Transactions
of NSM Based DoS Attacks Using Data Mining in Smart Grid,” Energies, on Sustainable Energy, vol. 4, no. 1, pp. 240–249, Jan. 2013.
pp. 4091-4109, 19 Oct. 2012. [66] B. Vural and C. S. Edrington, “Ultra-capacitor based pulse power
[43] U.-K. Premaratne, J. Samarabandu, T.-S. Sidhu, R. Beresh, and J.-C. Tan, management in electrical ships”, in IEEE Transportation Electrification
“An Intrusion Detection System for IEC61850 Automated Substations,” Conference and Expo (ITEC), 2012.
in IEEE Trans. Power Del., vol. 25, no. 4, pp. 2376-2383, Oct. 2010. [67] J. McGroarty, J. Schmeller, R. Hockney and M. Polimeno, "Flywheel
[44] Microsoft TechNet Library, “Common Types of Network Attacks”. energy storage systems for electric start and an all-electric ship," in IEEE
Accessed on 20 September 2014, [Online]. Available: Electric Ship Technologies Symposium, PP. 400-406, 2005.
http://technet.microsoft.com/en-us/library/cc959354.aspx. [68] L. Zhang, L. Li, W. Cui, and S. Li, “Study on improvement of micro-
[45] Brute-Force Password Cracker.” [Online]. Available: grid’s power quality based on APF and FESS,” in IEEE Innovative Smart
http://www.oxid.it/ca_um/topics/brute-force_password_cracker.htm. Grid Technologies - Asia (ISGT Asia), pp. 1–6, 2012.
[Accessed: 08-Mar-2017]. [69] C. Xie and C. Zhang, “Research on the Ship Electric Propulsion System
[46] M. T. A. Rashid, S. Yussof, Y. Yusoff and R. Ismail, "A review of security Network Power Quality with Flywheel Energy Storage,” in Power and
attacks on IEC61850 substation automation system network," Energy Engineering Conference, Asia-Pacific, pp. 1–3, 2010.
Proceedings of the 6th International Conference on Information [70] G. O. Suvire and P. E. Mercado, “Improvement of power quality in wind
Technology and Multimedia, Putrajaya, pp. 5-10, 2014. energy applications using a DSTATCOM coupled with a Flywheel
[47] N. Kush, E. Ahmed, M. Branagan, E. Foo, “Poisoned GOOSE: Exploiting Energy Storage System,” in Power Electronics Conference. COBEP ’09.
the GOOSE Protocol,” Proceedings of the Twelfth Australasian Brazilian, pp. 58–64, 2009.
Information Security Conference (AISC 2014). Vol. 149, pp. 17-22, 2014. [71] S. Samineni, B. K. Johnson, H. L. Hess, and J. D. Law, “Modeling and
[48] S. Malladi, J. Alves-Foss, and R. B. Heckendorn, “On Preventing Replay analysis of a flywheel energy storage system for Voltage sag correction,”
Attacks on Security Protocols,” [Online]. Available: https://oai.dtic.mil. in IEEE Trans. Ind. Appl., vol. 42, no. 1, pp. 42–52, Jan. 2006.
[Accessed: 08-Mar-2017]. [72] M. Subkhan and M. Komori, “New Concept for Flywheel Energy Storage
[49] D. Linden, T. B. Reddy, “Lead Acid Batteries” in Handbook of Batteries, System Using SMB and PMB,” in IEEE Trans. Appl. Supercond., vol. 21,
3rd ed. New York, NY: McGraw-Hill, ch, 23, 2002. no. 3, pp. 1485–1488, Jun. 2011.
[50] A. Cooper and P. Moseley, “Progress in the Development of Lead-Acid [73] Lahbib, A. Lahyani, A. Sari, P. Venet, “Performance analysis of a lead-
Batteries for Hybrid Electric Vehicles,” in IEEE Vehicle Power and acid battery/supercapacitors hybrid and a battery stand-alone under pulsed
Propulsion Conference,. VPPC ’06, pp. 1–6, 2006. loads”, in Intl. Conf. on Green Energy, pp. 273–278, 2014.
[51] Michael Koller, Theodor Borsche, Andreas Ulbig, Göran Andersson, [74] M. Ibrahim, S. Jemei, G. Wimmer, and D. Hissel, “Nonlinear
“Review of grid applications with the Zurich 1 MW battery energy storage autoregressive neural network in an energy management strategy for
system”, in Electric Power Systems Research, pp.128-135, 2015. battery/ultra-capacitor hybrid electrical vehicles,” in Electric Power
[52] Maninder Singh, Luiz A.C. Lopes, Nayeem A. Ninad, “Grid forming Systems Research, vol. 136, pp. 262–269, Jul. 2016
Battery Energy Storage System (BESS) for a highly unbalanced hybrid [75] P. J. Kollmeyer, L. W. Juang, and T. M. Jahns, “Loss optimization and
mini-grid”, in Electric Power Systems Research, , pp.126-133, 2015. ultracapacitor pack sizing for vehicles with battery/ultracapacitor hybrid
[53] P. Denholm, E. Ela, B. Kirby, M. Milligan, “The Role of Energy Storage energy storage,” in IEEE Transportation Electrification Conference and
with Renewable Electricity Generation”, NREL. Golden, CO. NREL/TP- Expo (ITEC), pp. 1–8, 2014.
6A2-47187. Jan. 2010. [76] B. M. Huhman, J. M. Neri, D. A. Wetz, “Design of a battery intermediate
[54] C. Pillot, “Micro hybrid, HEV, P-HEV and EV market 2012 #x2013;2025 storage system for rep-rated pulsed power loads”, in IEEE Electric Ship
impact on the battery business,” in Electric Vehicle Symposium and Technologies Symp. (ESTS), 2013.
Exhibition (EVS27), pp. 1–6, 2013. [77] G. L. Kusic, J. M. Heinzel, D. J. Hoffman, “Monitoring pulsed power on
[55] E. Wesoff, “Energy Storage at Grid Scale: A123 Gets Li-Ion to Market,” ship electrical systems”, in IEEE Electric Ship Technologies Symposium
08-Jun-2011. [Online]. Available: (ESTS), pp. 13–20, 2013.
https://www.greentechmedia.com/articles/read/Energy-Storage-at-Grid-
Scale-A123-Gets-Li-Ion-to-Market. [Accessed: 10-Feb-2017].