Centralized, decentralized, and

distributed control for Energy 1

Internet
Hajir Pourbabak, Tao Chen, Wencong Su
Department of Electrical and Computer Engineering, University of Michigan-Dearborn,
Dearborn, MI, United States

Chapter Outline

1.1 Introduction 3
1.1.1 Smart grid versus Energy Internet 4
1.1.2 The role of microgrids in the structure of the Energy Internet 7
1.1.3 Data acquisition in the legacy power system and Energy Internet network 8
1.2 Energy management approaches in energy networks 10
1.2.1 Centralized control 10
1.2.2 Decentralized control 12
1.2.3 Distributed control 13
1.3 Characteristics of communication networks of Energy Internet network 15
1.4 Conclusion and future research 15
References 16

The great economic revolutions in history occur when new energy regimes converge
with new communications revolutions.
Jeremy Rifkin.

1.1 Introduction
The data Internet (also known as simply the internet) is a network of interconnected
networks including, local, private, and public computer networks. The Internet pro-
vides various agents with an opportunity to share data in the information space (World
Wide Web) via this complicated network of networks. The Energy Internet can be
considered as a dual of the internet. The (electrical) energy, in the place of information,
is shared among various agents in the Energy Internet networks [1]. In other words, the
Energy Internet is an Internet-type network of all the components of a power system,
which closely interact with others by sharing both energy and information. Agents or
components of this network consist of different prosumers and consumers that have the

The Energy Internet. https://doi.org/10.1016/B978-0-08-102207-8.00001-1

Copyright © 2019 Elsevier Ltd. All rights reserved.
4 The Energy Internet

ability to make decisions by themselves. Microgrids, distributed generations, smart

grids, private or governmental energy networks, and any community of prosumers
and consumers can be a part of this enormous network as agents. The Energy
Internet is also known as the second generation of the power grid because it is
equipped with advanced sensing and measurement technology, as well as latest
control and monitoring technology [2]. The Energy Internet also takes advantage
of an advanced communication network to reach a higher level of safety and reli-
ability and increase the economic and efficient operation of the power system. In
addition, the integration of an advanced communication network and smart devices
into the power system enables system operators to embed plug-and-play character-
istics and intelligent energy management [3]. Fig. 1.1 is an Internet network used to
indicate the concept of the Energy Internet network.

1.1.1 Smart grid versus Energy Internet

At first glance, there is a slight confusion between the definitions of the smart grid
and Energy Internet because both of them use high technologies and modern methods.
The definition of the smart grid itself is confusing, as there are somewhat different
definitions because it is not possible to present a singular definition of the smart

Energy transmission lines

Difference level of energy
Consumption or generation nodes
Communication lines
Communication nodes

Figure 1.1 A snapshot of the Energy Internet.

Created by Harryarts - freepik.com, customized by authors.
Centralized, decentralized, and distributed control for Energy Internet 5

grid that puts all the various components and concepts together. S. Rahman [4] intro-
duces an adequate definition of the smart grid according to the US Department of
Energy’s modern grid initiative, “an intelligent or a smart grid integrates advanced
sensing technologies, control methods, and integrated communications into the current
electricity grid.” Thus, it provides an opportunity for consumers to have an active role in
the electricity market, accommodate various types of energy sources to support system
demand, improve energy efficiency, and enhance system security (self-healing) [5].
N. Hatziargyriou et al. extracted different features for smart grids from various
definitions [6,7]. Thus, the smart grid is:
• intelligent/smart because overloads can be determined/predicted to avoid potential outages
by rerouting power and doing other preventive measurements.
• efficient because the peak shaving technique is used to reduce electrical power consumption
during periods of maximum demand on the power utility.
• able to easily accommodate new energy sources and energy storage technologies.
• a good platform for a competitive electricity market.
• quality focused due to its technological capabilities to deliver high-quality energy.
• a resilient network if it uses a new method of controlling and monitoring, such as distributed
methods.
• green because it is an excellent opportunity to slow the advance of global climate change [8].

The Energy Internet is a newly developed environment of energy systems. Fig. 1.2
shows an evolution timeline of energy systems. Jeremy Rifkin [9] believes “The power
grid would be transformed into an info-energy net, allowing millions of people who
produce their own energy to share surpluses peer-to-peer.” Based on this definition,
hundreds of millions of distributed energy resources (DERs) will eventually produce
electricity everywhere and share it with each other through a network of Energy Internet
such as sharing data through information internet. The Energy Internet integrates smart
power grids, advanced distributed control systems, smart devices, smart communication
systems, etc., to provide interactive flexibility and efficient energy management.
Therefore, according to Refs. [1,10e12], the following are some features that
distinguish Energy Internet from the smart grid:
1. As mentioned earlier, the Energy Internet is a kind of duality of the Internet i.e., all agents,
including prosumers and consumers, are able to sell/generate and buy/consume energy.
2. A smart plug-and-play interface is a fundamental requirement of Energy Internet networks.
An intelligent communication interface supports plug-and-play characteristic to detect the
connection/disconnection of any device as soon as a plug-and-play happens.
3. The Energy Internet requires efficient management of energy supply and demand in the
power grid. Huang et al. specify this management as “status monitoring and data collection
of all devices as well as providing control references to each device” [10]. Energy routers, as
a dual of packet routers, are responsible for dynamically adjusting the energy distribution in
the grid by rerouting energy flows in transmission and distribution networks.
4. A regional centralized control method is the dominant control method for smart grids,
whereas the Energy Internet is based on multiagent (or intelligent agents) approaches.
This type of control provides more flexibility for consumers and prosumers to have an active
role in the energy system. Customers will be able to choose various services satisfying their
budget and preferences.
 6
Local investor–owned The legacy centralized Smart grids Energy internet
electric utility power system

The investor-owned electric utility Separated power grids connected to An intelligent or a smart grid The power grid will be transformed
can be considered as the very first each other through long transmission integrates advanced sensing into an info-energy net, allowing
version of the microgrid, such as, the lines, the electric utilities were technologies, control methods, and millions of people who produce their
first investor-owned electric utility moving from separate and integrated communications into the own energy to share surpluses peer-
on pearl street in lower manhattan independent systems to a highly current electricity grid. to-peer, thanks to its use of the
centralized and regulated one. Internet concept

Figure 1.2 The evolution of grids.

The Energy Internet

Centralized, decentralized, and distributed control for Energy Internet 7

In sum, the Energy Internet is an upgraded version of the smart grid, accommoda-
ting all types of distributed energy with great flexibility in energy sharing. The
major characteristics of an Energy Internet are its openness, robustness, reliability,
and competitive environment for the whole procedure of energy generation and
consumption.

1.1.2 The role of microgrids in the structure of the

Energy Internet
As mentioned earlier, microgrids are one of the parts of the Energy Internet network,
which act as agents. Microgrids have considerable potential to be the main element of
infrastructure of the Energy Internet networks, as they are a promising technology that
can increase the reliability and profitability of an energy supply to end consumers [13].
A microgrid, technically, is a low-voltage distribution network that is located down-
stream of a distribution system and connected to the distribution grid through a point
of common coupling.
Microgrids are small-scale power systems that have DER, distributed storage, and
local loads, which are able to work in islanding mode in the electrical distribution sys-
tem [14,15]. According to Ref. [10], Energy Internet networks should have three main
features: smart plug-and-play characteristics, intelligent energy management, and
distributed grid intelligence (see Fig. 1.3).
A microgrid, as an extension of distributed generation, can be easily integrated with
a two-way communication network, smart devices and metering, energy storage,
energy monitoring and management system, and load management tools to serve
the Energy Internet as the main element of its infrastructure.
The attractive features of a microgrid that enable it to be an unignorable component
of future energy management systems (EMSs) can be listed as in [16]:
1. A microgrid comprises various distributed power sources such as solar, wind, fossil fuels,
and biomass, fuel cells, internal combustion engines, and energy storage units.
2. A microgrid can play the role of a controllable power supply or a flexible load because of its
greater flexibility in the use of various power sources, its cost-efficient scheduling, and smart
management.

Rest of the power

system

Figure 1.3 A simple representation of a microgrid structure.

8 The Energy Internet

3. A microgrid can operate separately as an island to guarantee uninterrupted power supply to

users inside of the island until achieving the safe grideconnected mode.
4. A microgrid provides efficient energy utilization for users and satisfies their energy require-
ments as long as it provides great system reliability.

1.1.3 Data acquisition in the legacy power system and Energy

Internet network
Supervisory control and data acquisition (SCADA) systems have been widely used to
monitor and control plants and equipment in industry since the 1980s. SCADA is an
advanced automation control system, which is centrally responsible for the manage-
ment of power system energy, wide-area data gathering, and the operation of entire
electrical power systems [17,18]. Fig. 1.4 shows a general diagram of SCADA and
simply depicts how SCADA can access wide-area data. SCADA is a well stabilized
system and widely use in the legacy power system.
For a successful mission, various parameters should be monitored, such as
nodal voltages, line power flows, active/reactive power, system demand, frequency,

Power system

All gathered data of

power system, e.g.,
status information,
analog values

Manual
Sensors
inputs
Direct control signals
Operator control commands

RTU/PLS
Remote terminal unit
and programmable logic
control send data to
SCADA system

SCADA

Supervisory control
and data acquisition
(SCADA)

Master operator of
power system for final
decision

Figure 1.4 General system configuration of supervisory control and data acquisition (SCADA).
Centralized, decentralized, and distributed control for Energy Internet 9

switch statuses, and system topology [19]. Digital and analog parameters and data are
usually gathered by a remote terminal unit [20]. Then, they are transmitted to the
central monitoring/control station. This huge amount of data imposes a heavy compu-
tational load on both the monitoring/control system and communication network [21]
and then introduces its own technical problems. For example, transmission and anal-
ysis of such broad information requires a complicated communication infrastructure.
Thus, the SCADA system may not easily support plug-and-play characteristics of
distributed generations/storage and scalability.
Furthermore, for any economic analysis, more financial information is required.
Any access to private financial information of various prosumers and consumers by
a third party, such as SCADA, can easily violate their and the system’s overall privacy.
This type of data acquisition and sharing makes the legacy power system extremely
vulnerable. In sum, the SCADA system suffers from a heavy computational burden,
vulnerable privacy, and a single point of failure. In Ref. [22] Y. Yan et al. mention
the SCADA-based power systems (monitoring and control systems) may only be
restricted to transmission systems, and the SCADA are not suitable for larger-scale
monitoring and control of the entire electrical grid.
As is well known, information sharing and collection play a vital role in Energy
Internet networks because most of the functionalities of energy networks depend
on wide-area data collection and sharing. Consequently, however, the privacy of
prosumers and consumers can at the severe risk as agents share information. Load
monitoring is a common method for determining energy consumption for any
devices/units in a power system. There are two important methods for load
monitoring: intrusive and nonintrusive monitoring methods [23]. Intrusive and
nonintrusive monitoring methods are, respectively, referred to as the distributed
sensing and single point sensing methods. Accurate data can be obtained through
intrusive load monitoring (ILM) by connecting power meters to each appliance in
a unit, but this method suffers from some drawbacks, such as high cost, complicated
sensor configuration, and installation complexity [24]. In addition, this method is
barely trusted in the Energy Internet environment because it collects each individual
appliance’s energy consumption.
An immediate alternative method for ILM is nonintrusive load monitoring
(NILM), which only uses a single meter per user. NILM is an alternative and effec-
tive method for discovering the energy consumption of individual appliances based
on analysis of the aggregate load measured by the main power meter in a building. It
would seem to better protect privacy because it does not require to violate the private
information of an individual when measuring the power consumption of different
appliances [23]. However, a study by C. Hui et al. [25] shows that the analysis of
electrical data gathered by NILM can intrude on privacy because the economic
behavior of users can be inferred by some analytical study. Thus, nonintrusive
load monitors and smart meter data may reveal precise user information. On the other
hand, battery-based load hiding (BLH), a practical and cost-effective solution,
ensures the protection of prosumers and consumers’ privacy against information
leakage by third parties or neighbors [26].
10 The Energy Internet

In BLH method, a battery is installed in the consumer or prosumer’s side. This

battery is charged and discharged in a strategic way to hide accurate data that can
be obtained by the analysis of consumption data [27]
This process removes the detailed load profile information needed by NILM algo-
rithms to detect a user’s behavior. This method has some limitations because it
completely depends on a battery, whereas the alternative methods can be categorized
as nonbattery. The nonbattery methods can be grouped as cryptology [28,29] and
differential privacy [30].

1.2 Energy management approaches in energy

networks
In the 1980s and 1990s, the computer industry gradually used the distributed
computing approach in place of the centralized computing approach, thanks to the
worldwide Internet platform. This alteration in the structure of computer network
infrastructure led to computation cost reductions, avoidance of a single point of failure,
and data privacy improvement [10]. As the penetration of distributed energy resources
is continuously growing, altering power systems in the same way as the computer
network can revolutionize their future [21]. The Energy Internet, as a technological
revolution, can apply a new approach to the energy delivery and management of power
systems.
In this section, all three major control methodsdcentralized, decentralized and
distributed control methodsdwill be investigated in detail for figuring out the pros
and cons of applying them to Energy Internet.

1.2.1 Centralized control

In a centralized method, all agents, i.e., prosumers and consumers, will independently
communicate and directly interact only with a central operator. This center should be
able to monitor, gather, and analyze real-time data and provide all components with
appropriate control signals as long as it records events in a log file. Fig. 1.5 shows a
simple schematic of the communication structure of a power system. It is worth
mentioning that the centralized structure is only applied to the communication
network, and the energy system is a pool-based system.
For almost the last two decades, the centralized control approaches were evolving,
thanks to the high technology of communication networks and powerful computers.
Optimization of microgrid operation is one of the well-known problems that have
been frequently addressed by various centralized control approaches. The work of
A.G. Tsikalakis et al. is one of the earliest methods that proposed a centralized controller
to maximize the value of microgrid and the optimization of its operation during intercon-
nected operation, i.e., the production of local generators and energy exchanges with the
distribution network are maximized [31]. Unit Commitment (UC) is an optimization
problem that is responsible to determine the least cost of the operation commitment
of generation units [32]. UC problem is one the most complex optimization problem
Centralized, decentralized, and distributed control for Energy Internet 11

Transmission level

Substation
Substation
Substationn

Communication link
Transmission line
Figure 1.5 Centralized control schematic.

that could be extended over some period of time [33]. Independent system operator as a
third party handles wholesale electricity markets to find their day-ahead market sched-
ules based on centralized unit commitment [34]. The EMS is a sophisticated administra-
tive controlling system in the power system, which also is benefiting from the
centralized control methods to monitor, control, and optimize the performance of the
generation and transmission system. D.E. Olivares et al. introduce a centralized control
architecture and mathematical formulation of the microgrid’s energy management
problem. Their proposed centralized EMS for an isolated microgrids features a detailed
three-phase (unbalanced) model of the system [35]. A novel EMS for a microgrid based
on a rolling horizon strategy is developed by Palma-Behnke [36]. This EMS provides
the online set points for generation units while minimizing the operational cost and
considering the forecast of renewable resources and loads.
Unfortunately, the penetration of distributed generation, distributed energy storage,
renewable energy sources, and prosumers/consumers is continuously growing,
meaning the centralized algorithms are no longer effective [37,38] because they are
incapable of operating, monitoring, and controlling future power systems, which
includes tremendous numbers of agents [39,40].
Although the centralized methods are mature and established approaches for control
of many systems in recent decades, they are not a practical solution for Energy Internet
systems. The reasons behind of this are:
1. Heavy computation burden is a technical barrier for a centralized control as the number of
agents increases to hundreds of thousands [18].
12 The Energy Internet

2. Centralized methods are not easy to expand and are not appropriate for smart grids as they
need to expand very fast [41].
3. Because of the single point of failure of one center-based control systems, these approaches
are suitable only for relatively small-scale systems. Thus, a small number of users is affected
in the case of the failure of a center [42].
4. The centralized algorithms are not well designed to support plug-and-play functionalities of a
large number of participants [43].
5. Finally, the centralized approaches need a high level of connectivity because each agent
should directly interact with the center.

1.2.2 Decentralized control

The two terms, decentralized and distributed methods, are often used in place of each
other, but there is a slight difference between these two terms. In this section, the
concept of decentralized methods and their pros/cons are reviewed.
In a decentralized control method, each agent or group of agents is controlled by it-
self or a leader, respectively. Fig. 1.6 shows a sample structure of the decentralized
approach. The decisions are made based on local measurements, such as voltage and
frequency values, and there is a limited number of local connections. It stands to reason,
then, that decentralized control methods do not require a high level of connectivity
because this approach is mainly based on the local measurements of parameters [44].

Micro grid unit

Local measurements
Communication network

Decentralized controller

Figure 1.6 Decentralized control schematic.

Centralized, decentralized, and distributed control for Energy Internet 13

Figure 1.7 Distributed control schematic.

Furthermore, the decentralized control approach does not need to go through the whole
decision-making process of the entire system via one center; therefore, it is not imposed
by a high computation load. The global optimization, stability, or reliability of the entire
system cannot be assured, due to the lack of communication links and information
sharing among agents. However, this feature also enables decentralization with a higher
level of privacy protection.
Another strong point of decentralized methods is their robustness against a single
point of failure. A system equipped with a decentralized method has a massive redun-
dancy in the number of controllers because, in contrast to the centralized method, there
are some leaders/controllers in a decentralized system. For example, if some leaders
lose their connection with other agents or an agent fails to operate, the entire system
can still remain stable (Fig. 1.7).

1.2.3 Distributed control

The difficulties of both the centralized and decentralized methods can be overcome by
the distributed control approaches. In this section, the concept of the distributed control
approaches and their difference compared with decentralized control approaches are
studied.
In contrast to the system being controlled by a decentralized method, in which
agents use local measurements, agents of a system equipped with a distributed control
method are allowed to share their information with neighbors. In other words, agents in
a system with distributed methods not only use local measurements, but also are able to
send and receive required information. Therefore, this type of control system can
reach the global optimization, reliability, and stability such as centralized control
methods [45,46].
The security of the communication networks of the Energy Internet is the most
important factor, which depends on the privacy of agents. Privacy technologies
and encryption standards/algorithms are well matured now and provide energy sys-
tems with one of highest levels of security [47]. However, sharing the private and
14 The Energy Internet

confidential information may provide third parties or other agents with the opportunity
to intrude on privacy. New algorithms and protocols of distributed control methods
[21,48e50] preserve the privacy of each agent and the entire system by sharing
minimal pieces of information. It is worth mentioning that none of the agents share
information with a center as a third party.
As discussed in the introduction section, the Energy Internet, which is the integra-
tion of an advanced communication network and smart devices into the power system,
provides the system with plug-and-play characteristics [3]. This ability enables the
system to not be influenced by the dynamic topology of the Energy Internet network.
Additionally, it would be easy to extend as new agents arbitrarily connect to the
network. The energy infrastructure, including 10,000 power plants, 131 million
customers, and 157,000 miles of transmission lines, is one of the most complex
infrastructures ever built by the humans. Based on what is discussed here, the best
method to manage such a system is a distributed multiagents-based approach. A
mature multiagent environment with the right set of protocols that allows all agents
to locally/globally interact with each other can overcome the drawbacks of the central-
ized control approach [1].
In sum, distributed methods/algorithms for multiagent systems, as one of the great
revolutions in the energy industry, are very effective tools for energy management of
the Energy Internet because:
• Energy Internet networks are easy to expand and support scalability.
• Computational cost is distributed among multiagents over the Energy Internet network.
• Energy Internet would not be affected by a single point of failure.
• Energy Internet would not be affected by the dynamic topology.
• Energy Internet supports a plug-and-play characteristics.

Distributed control approaches have a wide variety of applications in the power

system management, such as power flow control, demand management, reliability,
and stability study. H. Dagdougui et al. proposed a distributed control strategy for a
network of smart microgrids. In this method, agents, i.e., smart microgrid, can share
information about their internal load and generation. The objective is to minimize
the cost of energy storage and exchanged power among smart microgrids when the
internal load is being supported. One important feature of the proposed approach is
that the cooperation among the agents was achieved with no direct knowledge of
the others [51]. In Ref. [52], a two-level cooperative optimization multiagent system
is designed for distributed energy resources economic dispatch. In this chapter, a
multiagent-based optimal microgrid control using a fully distributed diffusion strategy
solves an economic dispatch problem. The lower level implements an adaptive droop
scheme based on online no-load frequency adjustments. There is a peer-to-peer
communication among the agents, which simultaneously performs resource optimiza-
tion while regulating the system frequency.
In a nutshell, distributed optimization techniques could be classified into main two
groups: based on Lagrangian decomposition and based on KarusheKuhneTucker
(KKT) necessary conditions [53]. Lagrangian decompositionebased techniques can
also categorized in some famous group of methods such as Dual decomposition [54]
Centralized, decentralized, and distributed control for Energy Internet 15

and Alternating Direction Method of Multipliers [55]. Some novel cooperative distrib-
uted algorithms that solve the constrained nonlinear optimization problem using KKT
conditions and consensus networks are discussed in Refs. [56,57].

1.3 Characteristics of communication networks of

Energy Internet network
A mature communication network is a vital part of the Energy Internet and effectively
improves the performance of future smart grids. The communication network should
be changed and developed as fast as the development of the Energy Internet and its
requirements [2,47], such as stability, reliability, and profitability of the entire system.
The major features of communication networks are:
1. Two-way and pervasive communication: The next generation power grid should be sup-
ported by advanced two-way and pervasive communications to bring efficiency, reliability,
and safety for the entire system.
2. Wide bandwidth: The communication infrastructure should be able to support the increasing
number of agents and help them share their information seamlessly.
3. Cyber security and privacy: Privacy and information security, including confidentiality,
entity authentication, authorization, validation, etc., are essential for every competitive
system. Therefore, communication channels should have some level of security to ensure
that there is no leakage of private information to third parties, and no preferences are given
to one or more agents in a competitive environment [21].
4. Interoperability: Interoperability among the various agents of an Energy Internet network is
the ability of the agents to cooperate and share information to perform tasks without consider-
able effects on their operations. In other words, agents of a multiagent system should have
integration, effective cooperation, and two-way communications to reach their common
goals. The National Institute for Standards and Technology is working on protocols and
standards for information management and communication standards to provide a compre-
hensive interoperability for Energy Internet systems.
5. Scalability: Scalability is an essential attribute of a network, system, or process that can
support and accommodate increasing number of users, amount of workload, or more compli-
cated process. In an energy system, the communication network should have the ability to
easily support new users, new devices, and new control methods [58].
6. Smart plug-and-play: An intelligent communication interface should support a smart
plug-and-play characteristic to detect connection/disconnection of any device as soon as
plug-and-play happens.
7. Self-healing: One of the approaches for enhancing the system reliability is a self-healing
ability. The communication network plays an important role to provide the power system
with this self-healing ability.

1.4 Conclusion and future research

The Energy Internet as an upgraded version of the smart grids accommodates all types
of distributed energy with great flexibility in energy sharing and includes important
features such as openness, robustness, reliability, and a flexible environment.
16 The Energy Internet

This new type of energy ecosystem provides a great opportunity for renewable energy to
get more shares of energy sources to guarantee a better future for human beings. Now,
technological achievements in the fields of communication, the computer/Internet
industry, and economics are coupling together to affect our world more than ever. In
the future, young researchers can focus on how to accommodate the knowledge of
communications and Internet into smart grid to continue this scientific revolution.
• Information-flow protection on the network of Energy Internet:
After the emergence of the smart grids, the privacy protection became more important
than ever because of the massive and ceaseless flow of private information through the
communication system. As the number of agents is increasing to hundreds of thousands,
the huge amount of two-way transmission of data among the various agents is necessary.
Data must be kept secure with the minimum loss and latency for the sake of effectively con-
trolling and monitoring. If an intelligent entity wants to survive in this Energy Internet envi-
ronment, it requires keeping its information as secure as possible. Cryptography is a practice
concerned with the enhancement of secret communication in the presence of adversaries.
Encryption, as one of the subfields of cryptography, protects the stored or transferred data
across the communications networks to shield confidential data against strangers. There
are two different point of views for security: conditional security and unconditional security.
Both of them could be targeted as future research trends.
• Flexible and open source EMS:
An efficient, flexible, and open source of energy management system, which is allowing
millions of people to produce and consume energy, is required. This system will integrate
smart power grids, advanced distributed control systems, smart devices, smart communica-
tion systems, etc., to provide all agents in Energy Internet system with an opportunity to
easily produce/consume energy and follow their own plug-and-play characteristic. This
energy management system must be open source to all energy industry player; thus, they
can adapt it based on their preferences.

References
[1] L.H. Tsoukalas, R. Gao, From smart grids to an energy internet: assumptions, architectures
and requirements, in: 2008 Third International Conference on Electric Utility Deregulation
and Restructuring and Power Technologies, IEEE, April 2008, pp. 94e98.
[2] Z. Hu, L. Lu, G. Liu, J. Yi, L. Zhaoc, Research on adaptability evaluation method of new
communication technology applied to energy internet communication network, in: 2017
IEEE International Conference on Energy Internet (ICEI), IEEE, April 2017, pp. 250e255.
[3] Q. Tang, W. Xu, L. Shen, X. Ye, Y. Xiang, J. Liu, Y. Niu, J. Hong, Urban energy internet:
concept and key technology, in: 2016 3rd International Conference on Systems and
Informatics (ICSAI), IEEE, November 2016, pp. 318e323.
[4] S. Rahman, Smart grid expectations [in my view], IEEE Power and Energy Magazine 7
(September 2009) 88, 84e85.
[5] J. Zhu, Operation of smart grid, in: Optimization of Power System Operation, John Wiley &
Sons, Inc., Hoboken, NJ, USA, January 2015, pp. 579e628.
[6] V. Agarwal, L.H. Tsoukalas, Smart grids: importance of power quality, in: Lecture Notes
of the Institute for Computer Sciences, Social-informatics and Telecommunications
Engineering, vol. 54, LNICST, 2011, pp. 136e143.
Centralized, decentralized, and distributed control for Energy Internet 17

[7] N. Hatziargyriou, A. Dimeas, T. Tomtsi, A. Weidlich, Energy-efficient Computing and

Networking, Vol. 54 of Lecture Notes of the Institute for Computer Sciences, Social
Informatics and Telecommunications Engineering, Springer Berlin Heidelberg, Berlin,
Heidelberg, 2011.
[8] S. Borlase, Smart Grids: Infrastruture, Technology, and Solutions, CRC Press, 2013,
p. 559.
[9] J. Rifkin, The Third Industrial Revolution: How Lateral Power Is Transforming Energy,
the Economy, and the World, 2013.
[10] A.Q. Huang, M.L. Crow, G.T. Heydt, J.P. Zheng, S.J. Dale, The future renewable electric
energy delivery and management (FREEDM) system: the energy internet, in: Proceedings
of the IEEE, vol. 99, January 2011, pp. 133e148.
[11] Y. Xu, J. Zhang, W. Wang, A. Juneja, S. Bhattacharya, Energy router: architectures and
functionalities toward energy internet, in: 2011 IEEE International Conference on Smart
Grid Communications (SmartGridComm), IEEE, October 2011, pp. 31e36.
[12] K. Wang, J. Yu, Y. Yu, Y. Qian, D. Zeng, S. Guo, Y. Xiang, J. Wu, A Survey on Energy
Internet: Architecture, Approach, and Emerging Technologies, 2017.
[13] E. Foruzan, S. Asgarpoor, J.M. Bradley, Hybrid system modeling and supervisory control
of a microgrid, in: 2016 North American Power Symposium (NAPS), IEEE, September
2016, pp. 1e6, 1.
[14] B. Kroposki, R. Lasseter, T. Ise, S. Morozumi, S. Papathanassiou, N. Hatziargyriou, Making
microgrids work, in: IEEE Power and Energy Magazine, vol. 6, May 2008, pp. 40e53.
[15] A. Dehghan Banadaki, F. Doost Mohammadi, A. Feliachi, State space modeling of inverter
based microgrids considering distributed secondary voltage control, in: 2017 North
American Power Symposium (NAPS), IEEE, Morgantown, WV, USA, 2017, pp. 1e6.
[16] P. Peng, Microgrids: a bright future, Huawei Communications (63) (2011) 6e8.
[17] M. Higgs, Electrical SCADA systems from the operator’s perspective, in: IEE Seminar
Condition Monitoring for Rail Transport Systems, IEE, 1998, pp. 3/1e3/4.
[18] H. Pourbabak, T. Chen, B. Zhang, W. Su, Control and energy management system in
microgrids (Chapter 3), in: Clean Energy Microgrids, Institution of Engineering and
Technology, 2017, pp. 109e133.
[19] E. Vaahedi, Power system monitoring, in: Practical Power System Operation, John Wiley
& Sons, Hoboken, NJ, March 2014, pp. 6e20.
[20] W.N.S.E.W. Jusoh, M.R.A. Ghani, M.A.M. Hanafiah, S.H. Raman, Remote Terminal Unit
(RTU) hardware design and development for distribution automation system, in: 2014
IEEE Innovative Smart Grid Technologies - Asia, ISGT ASIA 2014, IEEE, May 2014,
pp. 572e576.
[21] H. Pourbabak, J. Luo, T. Chen, W. Su, A novel consensus-based distributed algorithm for
economic dispatch based on local estimation of power mismatch, IEEE Transactions on
Smart Grid (2017) 1.
[22] G. Taylor, M. Irving, P. Hobson, C. Huang, P. Kyberd, R. Taylor, Distributed monitoring
and control of future power systems via grid computing, in: 2006 IEEE Power Engineering
Society General Meeting, IEEE, 2006, p. 5.
[23] E. Aladesanmi, K. Folly, Overview of non-intrusive load monitoring and identification
techniques, IFAC-PapersOnLine 48 (30) (2015) 415e420.
[24] A. Zoha, A. Gluhak, M.A. Imran, S. Rajasegarar, Non-intrusive Load Monitoring
Approaches for Disaggregated Energy Sensing: A Survey, December 2012.
[25] H. Cao, S. Liu, R. Zhao, H. Gu, J. Bao, L. Zhu, A privacy preserving model for energy
internet base on differential privacy, in: 2017 IEEE International Conference on Energy
Internet (ICEI), IEEE, April 2017, pp. 204e209.
18 The Energy Internet

[26] D. Egarter, C. Prokop, W. Elmenreich, Load hiding of household’s power demand, in:
2014 IEEE International Conference on Smart Grid Communications, SmartGridComm
2014, IEEE, November 2015, pp. 854e859.
[27] W. Yang, N. Li, Y. Qi, W. Qardaji, S. McLaughlin, P. McDaniel, Minimizing private data
disclosures in the smart grid, in: Proceedings of the 2012 ACM Conference on Computer
and Communications Security - CCS ’12, ACM Press, New York, New York, USA, 2012,
p. 415.
[28] S. Finster, I. Baumgart, Elderberry: a peer-to-peer, privacy-aware smart metering protocol,
in: Proceedings - IEEE INFOCOM, IEEE, April 2013, pp. 3411e3416.
[29] C.-K. Chu, J.K. Liu, J.W. Wong, Y. Zhao, J. Zhou, Privacy-preserving smart metering with
regional statistics and personal enquiry services, in: Proceedings of the 8th ACM SIGSAC
Symposium on Information, Computer and Communications Security - ASIA CCS ’13,
2013, p. 369.
[30] C. Dwork, A. Roth, The algorithmic foundations of differential privacy, Foundations and
Trends R in Theoretical Computer Science 9 (3e4) (2013) 211e407.
[31] A.G. Tsikalakis, N.D. Hatziargyriou, Centralized control for optimizing microgrids
operation, IEEE Transactions on Energy Conversion vol. 23 (March 2008) 241e248.
[32] B. Saravanan, S. Das, S. Sikri, D.P. Kothari, A solution to the unit commitment problem-a
review, Frontiers in Energy 7 (June 2013) 223e236.
[33] A.J. Wood, B.F. Wollenberg, G.B. Sheble, Power Generation, Operation, and Control (2014).
https://www.worldcat.org/title/power-generation-operation-and-control/oclc/886509477.
[34] R. Sioshansi, R. O’Neill, S.S. Oren, Economic consequences of alternative solution
methods for centralized unit commitment in day-ahead electricity markets, IEEE Trans-
actions on Power Systems 23 (May 2008) 344e352.
[35] D.E. Olivares, C.A. Canizares, M. Kazerani, A centralized energy management system for
isolated microgrids, IEEE Transactions on Smart Grid 5 (July 2014) 1864e1875.
[36] R. Palma-Behnke, C. Benavides, F. Lanas, B. Severino, L. Reyes, J. Llanos, D. Saez,
A microgrid energy management system based on the rolling horizon strategy, IEEE
Transactions on Smart Grid 4 (June 2013) 996e1006.
[37] H. Pourbabak, A. Kazemi, A new technique for islanding detection using voltage phase
angle of inverter-based DGs, International Journal of Electrical Power and Energy Systems
57 (2014) 198e205.
[38] M. Warnier, S. Dulman, Y. Koç, E. Pauwels, Distributed monitoring for the prevention of
cascading failures in operational power grids, International Journal of Critical Infrastruc-
ture Protection 17 (June 2017) 15e27.
[39] H. Pourbabak, T. Chen, W. Su, Consensus-based distributed control for economic oper-
ation of distribution grid with multiple consumers and prosumers, in: 2016 IEEE Power
and Energy Society General Meeting (PESGM), vol. 2016, IEEE, Boston, MA, July 2016,
pp. 1e5. Novem.
[40] R. Mudumbai, S. Dasgupta, B.B. Cho, Distributed control for optimal economic dispatch
of a network of heterogeneous power generators, in: IEEE Transactions on Power Systems,
vol. 27, November 2012, pp. 1750e1760.
[41] A.X. Sun, D.T. Phan, S. Ghosh, Fully decentralized AC optimal power flow algorithms, in:
IEEE Power and Energy Society General Meeting, 2013.
[42] W. Ren, R.W. Beard, Distributed Consensus in Multi-vehicle Cooperative Control,
Springer, London.
[43] W. Su, J. Wang, K. Zhang, A.Q. Huang, Model predictive control-based power dispatch
for distribution system considering plug-in electric vehicle uncertainty, Electric Power
Systems Research vol. 106 (January 2014) 29e35.
Centralized, decentralized, and distributed control for Energy Internet 19

[44] T. Senjyu, R. Kuninaka, N. Urasaki, H. Fujita, T. Funabashi, M. Ieee, S. Member, Power

system stabilization based on robust centralized and decentralized controllers, in: 2005
International Power Engineering Conference, vol. 2, 2005, pp. 905e910.
[45] H. Pourbabak, S. Xu, T. Chen, Z. Liang, W. Su, Distributed control algorithm for optimal
power allocation of EV parking lots, in: 2017 IEEE Power and Energy Society General
Meeting, IEEE, Chicago, IL, 2017, pp. 1e5.
[46] S. Xu, H. Pourbabak, W. Su, Distributed cooperative control for economic operation of
multiple plug-in electric vehicle parking decks, International Transactions on Electrical
Energy Systems (2017) e2348.
[47] Y. Yan, Y. Qian, H. Sharif, D. Tipper, A survey on smart grid communication infra-
structures: motivations, requirements and challenges, IEEE Communications Surveys and
Tutorials 15 (1) (2013) 5e20.
[48] G. Hug, S. Kar, C. Wu, ConsensusþInnovations approach for distributed multiagent
coordination in a microgrid, IEEE Transactions on Smart Grid 6 (4) (2015) 1893e1903.
[49] N. Rahbari-Asr, U. Ojha, Z. Zhang, M.Y. Chow, Incremental welfare consensus algorithm
for cooperative distributed generation/demand response in smart grid, IEEE Transactions
on Smart Grid 5 (6) (2014) 2836e2845.
[50] R. Deng, Z. Yang, F. Hou, M.-Y. Chow, J. Chen, Distributed real-time demand response in
multiselleremultibuyer smart distribution grid, IEEE Transactions on Power Systems 30
(September 2015) 2364e2374.
[51] H. Dagdougui, R. Sacile, Decentralized control of the power flows in a network of smart
microgrids modeled as a team of cooperative agents, IEEE Transactions on Control
Systems Technology 22 (2) (2014) 510e519.
[52] R. De Azevedo, M.H. Cintuglu, T. Ma, O.A. Mohammed, Multiagent-based optimal
microgrid control using fully distributed diffusion strategy, IEEE Transactions on Smart
Grid 8 (July 2017) 1997e2008.
[53] D.K. Molzahn, F. Dorfler, H. Sandberg, S.H. Low, S. Chakrabarti, R. Baldick, J. Lavaei,
A Survey of Distributed Optimization and Control Algorithms for Electric Power Systems,
nov 2017.
[54] B. Yang, M. Johansson, Distributed optimization and games: a tutorial overview, in:
Lecture Notes in Control and Information Sciences, vol. 406, 2010, pp. 109e148.
[55] T. Erseghe, Distributed optimal power flow using ADMM, IEEE Transactions on Power
Systems 29 (5) (2014) 2370e2380.
[56] N. Rahbari-Asr, M.Y. Chow, Cooperative distributed demand management for community
charging of PHEV/PEVs based on KKT conditions and consensus networks, IEEE
Transactions on Industrial Informatics 10 (3) (2014) 1907e1916.
[57] P. Yi, Y. Hong, F. Liu, Distributed gradient algorithm for constrained optimization with
application to load sharing in power systems, Systems and Control Letters 83 (2015)
45e52.
[58] A.B. Bondi, Characteristics of scalability and their impact on performance, in: Proceedings
of the Second International Workshop on Software and Performance - WOSP ’00, ACM
Press, New York, New York, USA, 2000, pp. 195e203.