See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/260113379

Modeling and Control of Microgrid: An Overview

Article in Journal of the Franklin Institute · May 2014

DOI: 10.1016/j.jfranklin.2014.01.016

CITATIONS READS
263 16,093

3 authors, including:

Magdi S Mahmoud Mohammed Abido

King Fahd University of Petroleum and Minerals King Fahd University of Petroleum and Minerals
673 PUBLICATIONS 12,461 CITATIONS 355 PUBLICATIONS 17,050 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Mohammed Abido on 02 December 2017.

The user has requested enhancement of the downloaded file.

SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 1

Modeling and Control of Microgrid: An overview

M. S. Mahmoud , S. Azher Hussain , M A. Abido

Abstract—A Microgrid (MG) is a building block of future the growth in market. Government organizations should ensure
smart grid, it can be defined as a network of low voltage power that these regulatory policies should include guidelines and
generating units, storage devices and loads. System of Systems schemes to implement microgrid technologies
(SoS) is another concept involving large scale integration of
various systems. In this paper, we provide an overview of recent Control of the voltage and frequency during islanded op-
developments in modeling and control methods of microgrid as eration of DGs is also a major challenge. A method for
well as presenting the reason towards incorporating MG into the intentionally islanding a single DG to feed a local load was
existing grid. Various SoS control strategies when applied to MG proposed in [6]. A much more complex and challenging task is
are discussed. to operate more than one DG on the island. With more than one
Index Terms—Distributed Generation, Microgrid, SoS, De- DG on the island, it is necessary to regulate the voltage during
centralized Control, Multi-Level Control, Networked Control microgrid operation, which could be achieved by using a
System.
voltage versus reactive power droop controller [7]. There needs
to be an algorithm that should complete the resynchronization
I. I NTRODUCTION process once the grid is restored. A supervisory control mecha-
The burden on the transmission network is increasing at an nism will monitor the overall process and provide information
unexpected pace due to the increasing demand of power. Since to the local controller to respond accordingly.
updates to the transmission network are economically chal- The concept of Power grid is based on the technology
lenging, microgrids have evolved to become an economically introduced around 120 years ago. It is facing lot of issues
viable alternative. In microgrids, generating units are com- in keeping up with modern challenges. One of the main
missioned within the scope of the conventional distribution challeges is to guarantee electricity supply to customers and
network so that power can directly flow from the generators maintaining long-term energy security. Therefore increased
to the load without having to pass through the transmission reliability/efficiency is very much needed in todayâĂŹs world
network. The other advantage of using such an architecture where the demand of electricity is ever-growing. Microgrids
is that loads can be served even if the transmission network (MG) incorporate various distributed generator (DG) units
is down due to a fault, increasing the overall reliability of into the utility grid and solves and solves many problems of
the system. A Microgrid is generally known as the system existing power systems. It is also the vital building block of
consisting of small distributed generating stations along with the future Smart Grid [7].
the loads which is capable of going into islanded operation at On another front, the concept of System of systems (SoS)is
times of need [1]. gaining rapid interest in the field of research and can also
Among the many benefits of having a microgrid, one is that possibly lead to a new branch of engineering known as SoS
it facilitates distributed generation (DG) and high penetration Engineering. This branch is related to systems engineering
of renewable energy sources [2]-[4]. They increase power and deals with the optimization of a network of interacting
quality and reliability of electric supply. A microgrid having systems. Any large scale integrated system or any complex
renewable energy sources will help to alleviate some of the system can be viewed as an example of SoS [8], [9]. A
environmental issues related to burning fossil fuels. There is microgrid can be viewed as a System of System (SoS).
extensive literature on the various challenges posed by micro- In this paper, motivation towards development of MG and an
grids. Despite having some benefits of microgrid architecture overview will be presented on the two key aspects,modeling
in the grid environment, there are some challenges related to and control, of MG. Recent developments in these two key
this also. Implementation is an issue. Microgrid protection is aspects will be presented. A better control strategy, by viewing
also considered one of the most important challenges facing MG as a special case of SoS, will be discussed.
the implementation of microgrids. Once a microgrid is formed,
it is important to assure that the loads, lines, and DGs on II. D ISTRIBUTED G ENERATION : A PPLICATIONS AND
the island are protected because conventional unidirectional I SSUES
power flow protection method is no longer viable [5]. Solid
regulatory base is another issue related to microgrids. It is The existing grid has small number of producers, long
known tha energy related industries established policies and distribution ways and high maintenance cost, it is also difficult
’solid regulatory base’ in place which became important for to achieve load balancing. Moreover, the depleting fossil fuels
and the adverse effect on environment by its consumption
ManuscriptMsM-Azhar-Abido-JFI.tex has gain multi-national interest in reducing the excess use of
Systems Engineering Department, KFUPM, P. O. 5067, Dhahran 31261, nonrenewable energy resources and many nations are keeping
Saudi Arabia, email: msmahmoud@kfupm.edu.sa
Electrical Engineering Department, KFUPM, P. O. 985, Dhahran 31261, tap on CO2 emissions [10].
Saudi Arabia The main concerns with the existing centralized power
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 2

system grid are summarized below [11] Distributed Energy Resources (DER), integration of DG into
the utility grid is possible. Power electronic devices improve
• Increasing demand and lack of high reliability the flexibility and adaptability of the system by converting the
• No scope of expansion on power system expansion power from source to a fixed frequency AC power. They also
• Limitations of centralized power system planning provide various ancillary services to the grid [19]-[20].
• Risks of volatile bulk power markets
• Security Threats
• Limited power quality
III. M ICROGRID : D EFINITION AND A PPLICATIONS
• Environmental effect (Release of CO2 , Nuclear waste
etc.)
All the above issues urges the need to incorporate Dis- A microgrid can be defined as, ’A network of low voltage
tributed Generating (DG) units into the existing power systems power generating units, storage devices and loads capable of
[12]. The concept of DG is of early 1990’s, it has multiple supplying a local area such as suburban area, an industry or
advantages for both source and consumers [13]. In literature, any commercial area with electric power and heat’. The com-
there exists various definitions of DG which are summarized in ponents of Microgrid are interfaced through quick response
[14]. DG is defined as , "Generation of electricity by facilities power electronics and presents itself as a single entity and
that are sufficiently smaller than central generating plants so therefore can be connected to traditional power grid or can
as to allow interconnection at nearly any point in the power also be operated in stand-alone mode as a self-sustained power
system"[15]. system [7].
DG units are the emerging micro-generating technologies As stated in [17], "The heart of the microgrid concept is
such as microturbines, fuel cells, Internal Combustion (IC) the notion of a flexible, yet controllable interface between
engines. It also make use of renewable energy sources such the microgrid and the wider power system.". Microgrid acts
as Photo Voltaic (PV) arrays and wind turbines. The DG as a Good Citizen, that is, ideal conventional load behavior
units have low emission rates, environment friendly and are towards the grid which is less troublesome than distributed
economical.The introduction of DG units should reduce the generation system. It also has environmental benefits because
pressure on central power grid principally but in technically it uses renewable energy sources.
speaking, penetration of distributed generation into the power Different countries around the world adopts various topolo-
grid creates a new class of issues different from those found gies and structure basing on their priorities on functionality
in traditional power sources. Random applications of DG offered by microgrid. The research on microgrid is more active
units will cause as many issues as it may solve [16]. Some in US, Canada, Europe and Japan. Several demonstration
of the problems are discussed below projects and laboratory facilities are developed and a lot of re-
search is in progress concerning various issues in the microgrid
• First of all, DG units operates close to the distribution [21]-[22]. Various objectives which can be achieved by the use
voltage level of 480V as it is geographically located near of microgrid are listed below, ride through capability provided
the loads and provides a DC or variable frequency AC by energy-storage is a common objective of microgrid.
output and hence requires power electronic devices in
• Reliability of power supply
order to interface with the power-grid/load. The power
• Reduction of environmental impact of electric supply
electronic interface leads to development of new control
• Reduction of investment in plant, equipment and cost
strategies [17].
• Increase of energy efficiency Stable
• The output of renewable energy systems fluctuates with
• Ensure diversity of energy supply
conditions of weather which is also a debatable issue
• Power supply to a remote site
when DG units are connected to power grid [18].
• Ride-through capability provided by energy storage
• The existing power grid follows a multi-level flow of
power from transmission to distribution network, any The future Smart grid is expected to be a well organized
change in power flow causes problems because DG units plug-and-play integration of microgrids connected via dedi-
behavior is different than a conventional load [10]. cated highways for exchange of command, data and power.
• Finally, the initial energy balance for a new load is taken The emerging standards, research, development and demon-
care by the power stored in the generator inertia and stration are also discussed in [23].
the micro generating units are inertia less. This lack of
inertia is the major problem leading to power imbalances
between the generation and load. There are also number
of barriers in form of technical, business and regulatory IV. M ICROGRID : C OMPONENTS AND F ORMATION
issues when it comes to connecting DG’s to electrical
grid [17]. A generalized structure of microgrid is shown in fig (1). The
To overcome these issues and to utilize the potential of microgrid can be connected to the utility grid through single
distributed generation, the concept of MicroGrid was intro- Point of Common Coupling (PCC). The isolating device is
duced in [17]. Using power electronic devices in addition with used to isolate the microgrid from the utility grid.
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 3

Distributed Generation (DG) Unit frequency is controlled by main grid. Depending on load
of the main grid, it will either supply or absorb power and
DC/AC Micro-turbine act as either controllable load or controllable source. If any
DC/DC Fuel cell
fault or disturbance occurs in the main grid, microgrid has an
ability to disconnect and operate autonomously. This ability of
PCC DC/AC IC engine microgrid increases the quality of power to its local customer
Isolating by providing local voltage control. In this mode of operation,
Device
To Grid DC/AC Wind turbine the points to be noted are
• The frequency and voltage magnitude are controlled by
DC/DC PV Array
utility grid
• DG units supply the total or a part of the load
DC/DC Energy Storage Unit
Islanding of microgrid can be due to unplanned faulty events
Central Controller discussed in [26] and can also be due to planned actions
like maintenance etc. The microgrid controls the voltage and
DC/AC Local Load frequency in autonomous mode by continuously adjusting the
output active and reactive power. This is very common mode
Fig. 1. Generalized microgrid structure of operation. In this mode, it supplies a local load which is
closely located geographically. The local load can be a small
The Distribution Generation (DG) unit is responsible for village, a university, an industry or a commercial building
generation of electricity. It consists of rotating type and etc. The main issues which the microgrid should address
inverter type generating devices. Rotating type includes IC in this mode is the management of voltage and frequency,
engines, gas turbines, microalternator etc. whereas the inverter Quality of Power (QoP), balancing between load and supply,
type includes photovoltaic, fuel cells and wind turbines etc. communication among its components etc. In this mode of
Both rotating and inverter type requires power electronic operation, the points to be noted are
converters for their interface. The power range of DG unit • The DG units control the frequency and voltage magni-
components is small-scale ranging from 4KW-10000KW [11]. tude
Energy storage unit is essential to balance the flow of • It supplies active and reactive power to the load
power at the onset of islanding mode of operation. It is
also used to control the flow of power to and from the VI. M ICROGRID : OVERVIEW OF M ODELING
main grid. They help in improving the quality of power and A microgrid integration of various units. Basically, it con-
assist in voltage control. Batteries, flywheels, super-capacitors, sists of DG unit, energy-storage unit, controller unit and
superconducting magnetic energy storage etc. can be used conventional load. The DG unit again compromises of var-
to store the energy. All these devices again require power ious micro-generating devices. Therefore, microgrid modeling
electronic devices for their interface [11]. varies from one configuration to other depending on the
There has to be a Control System for the safe operation components used. Various approaches for the modeling and
of microgrid in various modes of its operation. This system control of microgrid can be found in the literature [27]. We
can be based on a central controller or distributed controller. will discuss the different models available in the literature.
The selection of controller depends mainly on the operation A small signal dynamic analysis of an autonomous hybrid
mode of microgrid and its requirements [24].Various control system is performed in [28]. The configuration of the system
strategies will be discussed in this paper in other sections. is shown in fig (2).
The purpose of microgrid is not obtained until the customer
is served with nominal voltage and frequency by a stable
system during all the modes of operation [24].

V. MG: M ODES OF O PERATION

Microgrid can operate autonomously and can also be con-
nected to the utility/main grid. In case any fault occurs
while operating in grid connected mode, microgrid has an
ability to disconnect itself from grid and operate independently
supplying its local load [25]. Therefore, the microgrid modes
of operation can be classified into grid connected, islanded,
transition between grid-connected mode to the islanded mode
and vice-versa [26].In any mode of operation, the heat gener-
ated by some of the micro-sources can be used to supply the
heat demand of the local load.
When functioning in parallel with the utility (main) grid,
it acts as a Model or Good citizen and the voltage and Fig. 2. Configuration of a microgrid system [28]
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 4

The dynamics of all the DG units are approximated by a first

order linear model with a time constant and a gain factor while
the network is neglected [28], [29]. The transfer functions of
various components are obtained and time domain analysis is
performed by considering various components at each time.
The transfer functions of various components are given as
follows and fig (3) shows the configuration in one of the cases.

KW T G
W ind T urbine : GW T G (s) =
1 + sTW T G
Fig. 4. Model of MG in [30], [32], [33]
KP V
P V System :
1 + sTP V The dynamic model of Fig. 4 is represented by the following
KF C nonlinear equations
F uel Cell :
1 + sTF C ditd Rt vd vt d
= ωitq − itd − +
KDEG dt Lt Lt Lt
Diesel Engine Generator :
1 + sTDEG dvd 1 1 iL d
= itd − vd −
KAE dt C RC C
Aqua Electrolyser :
1 + sTAE iL d vq Rl
= ωiLq + − iLd
Ksto dt L L
Storage System :
1 + sTsto it q Rt vtq
= −ωitd − itq +
dt Lt Lt
K WTG iL q Rl
1 + S WTG = −ωiLd − iLq
dt L
+ PWTG
K WTG +
PW
1 + S WTG + ωCvd = itq − iLq
K WTG
After performing the linearization, whose details can be seen
1 + S WTG
1- Kn in [30], [32], [33] , the state space matrices are obtained as
K AE mentioned below. The overall test system was simulated in the
1 + S AE PAE PFESS K FESS Matlab Simulink and ATPDraw environment.
Kn
K FC 1 + S FESS
1 + S FC −R − L1t
 
+ +
+
- PS - 1 Lt
t
ω0 0
Fuel Cells
D f
K FC D + MS
 
+ + - +  
1 + S FC +
PS*
 ω0
 − RLl −2ω0 Rl Cω0
L − ω0 
R
K BESS
A=



Diesel Generator PBESS 1 + S BESS
 0
 ω0 − RLl 1
L − ω02 C 

K DEG  
1 + S DEG PDEG 1
0 − C1 1
− RC
C

BT = 1
   
Lt 0 0 0 C= 0 0 0 1
Fig. 3. Block diagram of a microgrid

X T = itd
 
Since a MG is a power generating unit, it can be represented itq iLd vd
by a DC source. This concept of modeling a MG with an RLC
load in islanded mode is proposed in [30]–[31]. As shown in This concept of representing MG as a combination of
Fig. 4, MG is represented by a DC source connected to a DC source with VSC is also presented in [34], [35]. This
voltage-sourced converter (VSC). The MG is connected to the paper models the islanded operation of MG consisting of two
grid by means of a R-L filter, step-up transformer and a circuit parallel DG units. Again the local load is a passive RLC
breaker. The circuit breaker is open when the MG is islanded. network located at the PCC. The schematic diagram of such
The load which is passive RLC type is connected on the high an arrangement is shown in Fig. 5. The MG structure is used
voltage side of the transformer. A control system is used to for the application of decentralized control and hence there is
control the VSC. a separate controller for each DG unit.
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 5

Fig. 5. Radial configuration of DG’s used in [34]

By applying KVL and KCL and further application of frame

transformation, gives the below dynamic equations governing
the MG

dVdq 1 1 1 1 Fig. 7. State-space model of MG in [36]

=− Vdq − it,dq1 − IL,dq + It,dq2
dt RC C C C
By combining individual sub-modules, the overall state-
It,dq1 1 Rt1 1 space model of MG is given below. The subscript represents
+ jω0 It,dq1 = − Vdq − It,dq1 + Vt,dq1 states of inverter, network and load. Detailed derivation and
dt Lt1 Lt1 Lt1
information on the state-space matrices can be found in [36].
IL,dq 1 Rl
+ jω0 IL,dq = Vdq − iL,dq 
∆ẋInv
  
dt L L ∆xInv
 ∆i̇networkdq  = Amg  ∆inetworkdq  (1)
it,dq2 1 Rt 2 1 ∆i̇loaddq ∆iloaddq
+ jω0 It,dq2 = − − it,dq2 + vt,dq2
dt Lt2 Lt 2 Lt 2 A small signal dynamic model of MG which includes
The model is then represented in state space which is synchronous generator based DG and power electronically
simulated using Matlab SimPowerSystems toolbox [34]. interfaced DG is presented in [39]. Fig. 8 shows the single
line diagram of the MG. DG1 is a synchronous machine
A microgrid consisting of only inverter based DG’s is
(diesel or gas-turbine generator) with excitation and governor
modeled in [36] [37]. Typical structure of such microgrid is
control system whereas DG2 is a dis-patchable source (micro-
shown in Fig. 6. The modeling approach considered the full
turbine or wind generator etc) equipped with a Voltage-Source
dynamic model of the complete network rather than algebraic
Converter (VSC). The system parameters are given in [26]
equations.

Fig. 8. Small signal of synchronous generator model [39]

Fig. 6. Inverter based MG structure in [36] To obtain the linearized mathematical model of the above
system the following steps are followed
The approach of modeling was divided in three modules • The ordinary differential equations (ODE) of DG units
namely inverter, network and loads. The inverter model com- including network components are developed in their
promises of dynamics of controller, output filter and coupling respective local dq0 reference frames
inductor. The state equations of network and load are rep- • The obtained equations are transformed to the global dq0
resented on one of the inverters reference frame which is as- frame of MG
sumed to be common reference. Then using the transformation • Linearized about a nominal operating point and arranged
technique [38], all the other inverters are transformed to this in the state space form
common frame. Each sub-module is modeled in state-space The dynamic model of DG1 in its local reference dq0 frame
form and combined together on this common reference frame. is obtained from [40] and the dynamic model of DG2 can
Block diagram of state space model of the MG is shown in be found in [41], [42]. The electrical network modeling is
Fig. 7. carried out on the basis outlined in [43] and can be found in
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 6

[42]. The overall system model is represented by the block fuel-cell (SOFC) is described in detail with the values of each
diagram in Fig. 9. The small signal model was validated in parameters in [47], [48]. A gas turbine (GAST) was used
the PSCAD/EMTDC environment. for the primary unit of microturbine. The dynamic model of
GAST is adopted from [47]. A fifth-order induction generator
connected directly to the network serves as a wind generator.
This model was available in Matlab Simulink toolboxes. An
empirical model for the PV generator based on experimental
results was adopted from [49].
Flywheels and batteries are used for the modelling of
storage devices. They were modeled as a constant dc voltage
sources and were coupled to the electrical network using power
electronic interface.
The inverter modelling can be derived as per two control
strategies, PQ inverter control modeling [50] and Voltage
Source Inverter Control (VSI) model [51], [52]. Inverters are
modeled based only on their control functions for the purpose
of analyzing the dynamic behavior of MG [26], [53], [54],
[55], [56]. Two types of loads were considered, one is constant
impedance load and other is motor load.
An LV test network was built in Matlab/Simulink SimPow-
erSystemsenvironment. The implementation of this network is
Fig. 9. Block diagram of MG model [39] shown in Fig. 11 whose detailed description can be found in
[57] and [58].
The same MG structure is taken into account in [26] and
the stability analysis for various transient conditions such
as energizing load, transition from grid-connected mode to
islanded mode and vice-versa is performed. An operational
architecture developed within EU R&D microgrids projects
[44], [45] is adopted in [46]. This concept is shown in Fig.
10. It is a multi-level type control and management scheme
supported by a communication infrastructure.

PV

Flywheel
DC AC
AC DC
MV LV MC
LC

MGCC LC
AC DC
Battery
LC
CHP
LC

MC AC DC
Fuel Cell
AC DC

LC

Microturbine

Fig. 10. MG structure in [46]

Fig. 11. Implementation of MG on LV network

The head of this multi-level control system is MicroGrid
central controller (MGCC) installed at the MV/LV substation
and centrally controls the MG. Load controllers (LC) and mi- A low voltage MG with three unbalanced phases
crosource controller (MC) forms the second level of Hierarchy was proposed in [59], Fig. 12 shows the Matlab
and exchange information with the MGCC. LC acts as an Simulinkimplementation of the structure used. Inverters
interface to controllable loads and MC controls the active and droop controls were used to interface DG’s and loads were
reactive power of each microsource. Both LC and MC receives modeled as constant power. Simulations were carried out for
their set-points from MGCC. both grid connected and isolated modes. This model has an
The dynamic modeling of each DG components was picked advantage of modeling small unbalanced networks but lacks
from different literature. The dynamic model of solid oxide the analytic details required for stability analysis.
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 7

to stabilize the operation of microgrid. When designing a

controller, operation mode of MG plays a vital role. Therefore,
after modelling the key aspect of the microgrid is control. In
this section we will discuss the various control paradigms.

A. Microgrid Control: Grid-connected Mode

In grid connected mode, microgrid acts as a controllable
load/source. It should not actively regulate the voltage at
the point of common coupling (PCC). Its main function is
to satisfy its load requirements with good citizen behavior
towards main grid. The balance between generation and de-
Fig. 12. MG network and Matlab/Simulink structure used in [59] mand, control of the parameters of the system is taken care
by the utility grid. The voltage and frequency reference of the
A new method to form the system matrices of large MG’s microgrid is also set by the main grid. Therefore the main
in islanded mode is discussed in [60]. The MG under con- task of a DG unit is to control the output real power (P) and
sideration has DG’s power electronically interfaced and hence reactive power (Q) [68]-[69]. The P, Q generated by a DG
the dynamics are similar to that shown in [36]. There were can be controlled either by current-based or by voltage-based
two types of DG’s, one was PQ regulated and other was Vf power flow control [70].
regulated which are introduced in [39].
The proposed modeling approach is based on four deïňAned ˛
complex vectors. These vectors allow for complex-valued B. Power flow control by current regulation
system matrices to be formed in a quite automated way. The control scheme for power flow control through current
Moreover, a convenient partition of the system matrices is regulation is illustrated in Fig. 13. It is desired to control both
proposed, which in turn allows fast and easy modiïňAcations.
˛ the real and reactive power. The real power control loop is used
Additionally, a multivariable methodology is proposed to si- to obtain the synchronous frame d-axis reference current and
multaneously determine the control system gains in an optimal reactive power control loop is used to obtain the q-axis current.
sense. The synchronous d-q frame current can then be controlled in
An alternative approach is based on the "hub model" for a closed loop manner [69].
microgrids [61] in which the couplings between an integrated
electricity and natural gas system to yield optimal operation
are modeled by energy hubs. It turns out that this concept
serves as interface between the loads and the transmission
infrastructures and supports the application of distributed con-
trol schemes. Similarly, hybrid modeling control techniques
are applied to a two generator power system connected to the
grid and the plant consists of a solar field and a secondary
power source formed by an electrolyzer, hydrogen tank and
fuel cell stack. It is shown that the system has essentially
hybrid dynamics, as it can operate in four distinct modes,
depending on the power circuit configuration and the fuel cell
stack state [62]. With focus on bulk power flow of microgrids, Fig. 13. PQ power control through output current regulation[70]
research investigations are reported in [63], [64] in which an
optimal design of an electrical microgrid and sizing of its com-
ponents is sought to balance capital investment with expected C. Power flow control by voltage regulation
operational cost while meeting performance requirements. In
The other method to control the power flow is based on
[65], an comprehensive review on current control technology is
the output voltage of DG. Therefore it is know as power flow
presented with emphasis on challenges of microgrid controls.
control through voltage regulation. It can be shown that real
The impact of frequency and voltage regulation on the optimal
power (P) flow is proportional to the voltage phase angle
design of an autonomous military microgrid, comprised of a
(δ) and reactive power (Q) flow is proportional to voltage
solar panel and vehicles as power sources, with each vehicle
difference (Vl - Vg ), where Vl is DG voltage and Vg is PCC
incorporating a battery and generator, is developed in [66],
voltage and δ is the phase angle difference between these two
[67]
voltages. Therefore the flow of P can be regulated using δ
and flow of Q can be regulated using Vl - Vg . This scheme
VII. M ICROGRID - OVERVIEW OF C ONTROL is illustrated in Fig. 14. To improve the accuracy of reactive
The control strategies for microgrid depends on the mode power control, integral control can be included into the reactive
of its operation. The aim of control technique should be power controller [68], [71], [72].
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 8

2) Microsource Controller : It includes Generator Agent,

Schedule Agent, Bid Agent
3) Load Controller : It includes Load Agent, Status Agent,
Switch Agent
The effectiveness and applicability of the introduced soft-
ware has been evaluated on a laboratory environment.

E. MAS based distributed control

A distributed control approach based on Multi Agent Sys-
tem (MAS) for microgrids is proposed in [74], where advan-
tages of MAS technology is utilized for controlling microgrids.
As shown in Fig. 16, a fully decentralized approach is adopted
with 3 distinguished control level.

Fig. 14. PQ power control through output voltage regulation [70]

Power control through voltage regulation is more sensitive

than current regulation to the line impedance between the DG
and the PCC.

D. Agent based control

Microgrid management system was developed using agent
based technology in [73]. Microgrid agents were developed on
JADE (Java Agent Development Framework). The proposed
system has several functionalities like SCADA system, selling
bids managing system, load shifting system etc. The software
architecture of the management system is shown in Fig. 15.

AMS DF

Pulling DB Fig. 16. Control levels of MAS environment

DB
Agent Agent

Shifting Curtail
Agent Agent Distribution Network Operator (DNO) and Market Operator
Central (MO)are at medium voltage level and does not belong to
Agent
microgrid. DNO refers to the operational functions of the
system and is responsible for technical operation of one or
Microgrids Agent Platform
more microgrids whereas one or more MO are responsible for
Micro-source Controller Micro-source Controller market functions of the area.
Normal Controller Load Controller
Sch Status Microgrid Central Controller (MGCC) is the main interface
Agent
Agent between DNO/MO and the microgrid. Its main function is to
Bid Gen Switch
Agent Agent Agent
Load
Agent
optimize the operation of microgrid and coordinate the local
controllers.
XML-PRC XML-PRC On the lower level, Load Controllers (LC) control the DG,
Bid production, storage and some of the local loads.
Using MAS technology, model of the system is obtained
in detail where every agent uses the the exact piece of
DC AC Load
information it needs, leaving the technical details for the
agents that are below it in the organization chart. The paper
Fig. 15. Architecture of management system proposes three types of agent. Control Agent which controls
physical units of the system directly. Management Agents
Microgrid agent platform consists of following components which manage the microgrid and takes the decision. Ancillary
1) Microgrid Central Controller (MGCC) : It includes Agents which performs tasks like communication and storage
Pulling Agent, Database Agent, Control Agent, Shifting of data. The proposed MAS platform is depicted in the Fig.
Agent, Curtailment Agent 17.
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 9

G. Microgrid control : autonomous/islanded mode

In the autonomous or islanded mode of operation, microgrid
supplies its local load and is not connected to the utility grid.
The main challenges in this mode are
1) Voltage and frequency control
Fig. 17. Types of MAS agents 2) Balance between supply and demand
3) Power quality
In [74], internal operation of the microgrid and its participa- 4) Issues relating to Microsources
tion in energy market was focused particularly. The algorithm 5) Communication among microgrid components
is proposed so that the every DER or controllable load decides Lot of research has been done on control of microgrid in
what is best for it. Autonomous/Islanded operation [78] which will be discussed
in this section. The two main control strategies PQ and VSI
F. H∞ control control is discussed first, detailed description and explanation
on these two controls can be found in [55].
A new power balancing method based on H∞ control theory When connected to grid, all the DG’ss can operate in PQ
is proposed in [75]. The power fluctuations were considered as control mode because the voltage and frequency is dictated by
the disturbances added to the MG. Fig. (18) shows the block the utility grid but at least one DG has to follow VSI control
diagram of MG structure. in islanded mode since the voltage and frequency reference is
set by the MG.

H. PQ and VSI control

The aim of PQ control is to provide constant active and re-
active power at a desired power factor [46], [60]. The reference
values of power is defined by a local controller or centrally
from the MGCC. This scheme can be implemented as a current
controlled voltage source or voltage controlled current source
as discussed earlier in section VII,A-(1). Current or voltage
components in direct (Id or Vd ) and in quadrature (Iq or Vd )
with inverter terminal voltage are computed based on method
given in [50].
Fig. 18. Structure of MG [75]

where Gge and Gbt are the first order transfer function
representation of gas turbine [76] and Kge and Kbt are
controller gains for gas turbine and battery, respectively. F
is the low/high pass filter gain.
Using the Robust Control Toolbox of MATLAB, the con-
troller gains Kge and Kbt were determined as standard con-
trollers. Fig. 19 shows the block diagram of these controller
designs, where W with subscript 1, 2 and 3 are weighing func-
tions for tracking performance, gain margin of the microgrid
system, and robustness for power fluctuations respectively.

Fig. 20. PQ Control Scheme [46]

Fig. 20 shows the control block for this strategy using

current control. The direct component (Id ) of the current
is used for the control of active power and the quadrature
Fig. 19. Design of H∞ controllers [75] component (Iq ) is used for the reactive power control.
Voltage source inverter (VSI) matches the behavior of a
In [77], technical challenges and stability of DG’s when synchronous machine controlling the voltage and frequency
connected into the distribution system is detailed. Since high on the ac system [53], [79], [50]. It acts as a voltage source
penetration of the DG’s can be considered as a microgrid the whose output voltage’s magnitude and frequency is controlled
same technical challenges can be assumed correct for a grid through Vf droop characteristics which are shown in fig. (21).
connected microgrid. Hence, this method is also known as Droop Method.
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 10

Microsource controller based on droop characteristics is

shown in Fig. 23. P,Q and V are calculated and then the droops
are implemented in two separate blocks. Then the controller
generates the voltage at desired magnitude and frequency at
the inverter terminals.

Fig. 21. V and f characteristics

Inverter Current Ereq
Reactive Power (Q)
In VSI control, voltage is related to reactive power (V − Q) Calculation
Q-V Droop
Q
whereas frequency/phase shift is related to active power (f − I
P ) by the below shown equations. N
Input V
Voltage
Voltage (E) Dynamic E
f = f0 − Kp ∆P Calculation E Controller R
|V| T
V = V0 − Kq ∆Q E
R
Inverter or Line Real Power (P) P Vs fq
Current Calculation P Calculation
where Kp and Kq are respective slopes of droop δVreq
characteristics and f0 and V0 are the idle values of frequency
and voltage [51]. The microsources will change their output Fig. 23. Microsource controller using droops
by ∆P (or ∆Q) when the frequency (or voltage) changes by
∆f (or ∆v) from the nominal values f0 (or v0 ).

A VSI model is shown in fig (22) [51], [52]. The active and
reactive powers are computed using the VSI terminal voltage.
The output voltage frequency f is determined by the active
power droops and magnitude V is determined by reactive
power droops. The output voltage are the reference signals
which control the VSI switching sequence.

J. New Q − V̇ droop control

Fig. 22. Model of a VSI

In islanded mode, the reactive power sharing is highly
The key point to be observed here is VSI uses the local dependent on impedance of power line. Due to the different
measurements at its terminals and reacts to any system distur- distances among DERs interface converters (DICs), the equiva-
bances quickly and hence does not require any communication lent transmission line impedance could be unequal [51]. Pf and
infrastructure [53], [79]. But there will be a communication QV droop characteristics are used in DER interface converters
infrastructure within the DG for the optimal management for power sharing operations [82], [81]. The Pf droop control
[46].A complete review of Vf control strategies can be found provides an accurate real power sharing among the DIC’s but
in [80] and validation of Vf control in both grid and islanded the problem arises in QV droop control. Because of these
mode is performed in [24]. unequal impedance load sharing performance of QV control
can be affected. Various control methods addressing this issue
I. Autonomous control has been proposed [83], [84], [85] but with some constraints.
The concept of utilizing VQ and Pf droops for controlling
the microgrid is also proposed in the [12], [16], [81]. They Therefore a new droop control method for the islanded
propose an autonomous control for the peer-to-peer and plug- operation of MG is proposed in [86] to overcome the effect
and-play model of the microgrid components. The concept of of line impedance on the reactive power flow, this method is
peer-to-peer allows the continuous operation of microgrid even known as Q-V̇ droop control method where V̇ represents the
with loss of any component/DG because there are no master rate of change of voltage. By regulating the voltage with V̇ ,
controller or central storage unit. The concept of plug-and- the reactive power sharing can be made independent of the line
play ensures that any component can be added at any point in impedance. The operation principle of the proposed method is
the system without re-engineering the controls. shown in fig (24).
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 11

Droop Controller P-f allows a DG to control its real & reactive power components
Vx,abc Vsx,qd V e x ,qd Px Droo fx
*
Px,inst p
contr
in grid-connected mode and to control the voltage & frequency
Ve*x,qd
abc/ dqs abc/ dqs wf ol
Sine + in autonomous mode.
P&Q s + wf wav PI qd e / dq s
Calcula e -
ix,abc tion wf
Q-V
Droop
abc/ dqe abc / dq s contro Vex,qd
s + wf l Vx
*

s e i s * x , qd
i x,qd i x ,qd Qx,inst Qx

Current
Regulat
Vsx,qd or
&PWM
isx,qd Modulat
or

*
Vax
*
D Qx Vx
*
* Qex +
V 1 +
ax +
KresQRx 1 +
nx * Fig. 26. Block diagram for the frequency control [31]
*
- s - - s - Vx
V
x Qx
Va
The proposed VSC frequency control is similar to that of a
Q-V*droop Control
V*restoration synchronous machine and is shown by the block diagram in
fig (26). In grid-connected mode, the frequency at PCC (ωpcc )
is equal to the grid frequency and hence have no impact on
Fig. 24. Block diagram of DIC with QV̇ droop control [86]
the system dynamics and only the output real power reference
has to be set. In autonomous mode, is same as VSC frequency
K. Control design based on transfer function determined by the droop characteristics.
Controller based on the transfer function of the plant is
designed in [30], [33], this is adopted from the classical
feedback control approach presented in [87].

Fig. 27. Block diagram for the Voltage control [31]

Fig. 25. Control strategy [30], [33] The block diagram in fig (27) shows the proposed voltage
control scheme. During grid-connected mode, output reactive
For the islanded mode of operation, fig(25) shows the power of VSC at PCC is set while in autonomous mode,
structure of controller. Reference angle is provided by a three the DG has to supply the load with reactive power, which is
phase PLL. The q component of load voltage is set to zero achieved by setting E2 to zero and only E1 remains effective.
and d component is regulated to the desired peak value.
Regulation of vd is done by comparing with reference signal IX. S YSTEM OF S YSTEMS - I NTRODUCTION
and the error is applied to the controller. The controller then To understand the concept of system of systems (SoS) or
provides inputs to the gating signal generator of the VSC cyberphysical system (CPS), let us consider an airplane which
(Fig. 4). More details can be found in [30], [33]. is an example of large scale complex system. Various parts
of airplane are operated by different systems but the plane
Instead of using frequency droops, an internal oscillator is flies only when all its systems operate collectively and does
used to design a multivariable controller in [32]. The function not fly if they operate individually. Therefore, a SoS is the
of this oscillator is to control the frequency in open loop way. large-scale integration of many systems that combine their
The robust servomechanism controller was designed using the capabilities together to form a more complex system offering
parameter optimization methods [88], [89] in addition with more functionalities than the individual sum of the constituent
non-conservative robustness constraint [90]. systems [91].
SoS is inherently multidisciplinary. The synthesis of multi
VIII. MG - C ONTROL IN B OTH M ODES systems requires the study of their interdependency because
A novel control strategy which can be applied to MG in each effects the other. This will result in different problems
both the modes is introduced in [31]. This scheme of control ranging from modeling to control. Therefore almost all the
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 12

key issues of systems engineering have to be revised. The

methodology of developing models, tools and control of SoS System of Systems
is typically referred to as system of systems engineering. SoS
methodology finds large number of applications in the defense
sector but recently it is also being used in auto transportation,
space exploration, search and rescue and many other non-
defense areas [92]. Controller:1 Controller:n
For SoS engineering branch, all aspects of systems engineer-
ing have to be revised. But the main problem arises in the two Fig. 28. Decentralized SoS control scheme
key aspects which are modeling and control. The challenge in
modeling point of view lies in the indirect effects, the cause
and effect need not be directly related, any change here can
produce effect over there because of their interdependency.It is
clear that these systems are very large and posses unpredictable
behavior which is difficult to model. SoS models available so B. MG - Decentralized control
far are still immature and there should be focus on additional
development. These models are quite complex and needs a
multidisciplinary approach [92], [93], [94], [95], [96]. A robust decentralized control strategy is proposed in [34]
Control is perhaps the most critical challenge facing SoS for the islanded operation of a MG. The proposed model of
designers. Due to the difficulty or impossibility of developing MG is shown in fig (5). It is shown that the MG can be
a comprehensive SoS model, either analytically or through represented by an interconnected composite system consisting
simulation, SoS control remains an open problem and is, of of two subsystems [99] and each subsystem can be controlled
course, different for each application domain. Moreover, real- using the local controllers.
time control which is required in almost all application do- The dynamic model of the MG is decomposed into two
mains of interdependent systems poses an especially difficult subsystem as follows
problem [97], [98].
• Master Subsystem
• Slave subsystem

X. S O S C ONTROL - A PPLICATION TO MG Since the two subsystems are controllable and observable,
it is shown that the composite system is stabilizable by using
only local controllers i.e., decentralized control strategy can
The key issue of SoS, which is control, faces a main chal- be applied.
lenge of developing a comprehensive SoS model, analytically
or by simulation. Availability of a proper model is necessary For the master subsystem, an H∞ controller was designed
to design a controller. If a proper mathematical model is to meet the robust characteristics [100]. This control strategy
available then there are several available control strategies. fulfills the voltage and frequency requirements of the load.
Also control strategy for each system is not only dependent The configuration of H∞ control is shown in Fig. 29. Matlab
on its own sensory information but also on the communication LMI toolbox is used to synthesize H∞ controller [101].
links among its neighboring systems or components, this is
another difficulty which rises from the control point of view.
Control of SoS, which is different for each application domain,
is still an open research area. In this section we will discuss
several potential control strategies.

A. Decentralized control

Another control lacking real-time consideration is decentral- Fig. 29. H∞ Control for Master subsystem [34]
ized control [92]. In this scheme of control, SoS is assumed
to be having multiple input and output variables. The control
design aims at assigning proper inputs for proper controller
which can observe a set of outputs. Thus there are multiple A simple PI controller was designed for the slave subsystem
controllers, each one controls a particular operation of SoS. using the conventional dq current control method [102], [103].
As it can be seen in Fig. 28, this scheme avoids storage of This is depicted in Fig. 30. The overall model and its con-
data. trollers were simulated in Matlab/SimPowerSystems toolbox.
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 13

Fig. 32. Inverter Control scheme in [105]

Fig. 30. dq Current control method for slave subsystem [34] Droop based inverter control scheme is adopted [106], [36].
The controller of each individual inverter consists of three
parts, i.e., the power controller, voltage controller, and current
controller. Only the stability of power controller is considered
whereas the voltage and current controllers are based on
A fundamental concepts of Power Management System traditional PI controllers. A fully decentralized communica-
(PMS) and robust decentralized control strategy for the is- tion is considered which implies any inverter only needs to
landed MG is proposed in [104]. The schematic diagram of communicate with its immediate neighbors to calculate the
proposed control is illustrated in fig (31). It consists of a total power generation of all DG units. Stability analysis with
PMS, Local Controller (LC) for each DER and MG frequency and without incorporating communication delay is presented.
control and synchronization scheme. A novel decentralized controller for load sharing among
parallel connected inverters in an islanded MG is proposed
in [107]. The controller has 3 nested loops
1) Inner loop
• Regulates the output voltage of inverter
• Voltage gain is responsible for good output voltage
tracking
2) Resistive output impedance loop
• Reduces the impact of line impedance unbalance
• Used to fix the output impedance of the inverter in
terms of magnitude and phase
• The output impedance presented to harmonic com-
ponents can be fixed
Fig. 31. PMS and control scheme in [104]
3) P/Q sharing outer loop
• Used to obtain proper P/Q sharing
• Droop/boost control scheme is used

A low bandwidth communication system is used to supply

the instantaneous values of real/reactive power of each DER C. Multilevel control
unit and load to the PMS. The PMS determines the set points
of real, reactive power and voltage for the PC buses and As discussed earlier SoS is integration of large-scale sys-
transmits to LC’s which measures the magnitude of voltage tems and large-scale systems can be decomposed into subsys-
at its PC bus provides voltage tracking based on the received tems. MultiLevel control assumes SoS is characterized by N
reference set point. finite set of subsystems coordinated by system coordinator as
shown in Fig 33. By employing any optimal control method
A decentralized inverter control based on wireless commu- the subsystems can be optimized and repeatedly performing
nication is proposed in [105]. wireless communication is used the modeling, the interactions between the coordinator and
to enhance the stability of droop based decentralized inverter subsystems can be converged to an optimal solution. In liter-
control. A wireless network is developed so that each inverter ature, multi-level control is obtained by classical steady state
can communicate with a certain set of inverters. Fig. 32 shows approach but lot issues has to be dealt while its implementation
the block diagram on inverter control. in real time [92], [93], [94], [95], [96].
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 14

external centralized controller to correct the errors produced

System of Systems by the primary control. Both controls are depicted in Fig 35.

Coordinator Systems Layer

Subsytems
Layer
Fig. 35. Primary and secondary control
Fig. 33. Multilevel SoS control
The tertiary control level comes into play mainly when the
MG interacts with the utility grid. The key points related to
D. MG - Multilevel control
Tertiary control are listed below
In this scheme of control there are three main levels namely 1) Controls the power flow between MG and the utility grid
primary control, secondary control and tertiary control as 2) Send the frequency and voltage references to the sec-
shown in fig (33) [108], [109]. ondary control
3) Can perform islanding detection or voltage harmonic
reduction
4) Can also improve the quality of power at PCC
The proposed multilevel control scheme allows the system
to integrate more and more MG’s and with this scheme of
control microgrids can operate in both grid connected and
islanded mode.

E. Networked Control Systems

In modern control systems, we find more and more ap-
plication of networks owing to impressive advancements in
network technology. One such example is Networked Control
System (NCS). In NCS, the feedback channel is closed using
Fig. 34. Structure of multilevel control scheme a real time communication network and all the data among
the components of system is exchanged through this com-
The key points related to Primary Control are listed below munication network [110]. In [111], NCS is properly defined
as, "Network Control Systems (NCS) are spatially distributed
1) Used to share load between converters
systems in which the communication between sensors, actu-
2) Improves the system performance and stability
ators and controllers occurs through a shared band limited
3) Regulate the output frequency and voltage magnitude
digital communication network". This definition explains that
4) Droop-control method is often used
the components of NCS are distributed and may operate
5) Can also include virtual impedance control loop to
asynchronously to reach some overall objective [112].
provide proper output impedance
One of main issues in NCS is the transmission delays and
The key points related to Secondary Control are listed below packet dropouts, therefore the challenge in NCS for SoS is to
1) Restores the f and V to nominal values whenever load develop an SoS distributed control system which can overcome
change occurs these issues. As mentioned in [113], these communication
2) Removes any steady-state error introduced by the droop infractions can be compensated by
control 1) Adjusting control power and controlling distances be-
3) During transition from islanded to grid-connected mode, tween systems (power control)
this control can perform synchronization to the main grid 2) Trading off modulation, coding, and antenna diversity
before interconnection versus throughput (adaptive modulation coding)
4) Make use of low bandwidth communication 3) The (non-wireless) intra-feedback (on-board hardware)
5) More global responsibilities loop of the autonomous control within Si is lower
The primary control loop makes use of only local output latency than the inter-wireless distributed control loop
voltage and current to perform calculations of droop control between Si and Sj or the inter-wireless System of
method whereas the secondary control level consists of an systems controller and the Si controller
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 15

Another way to check on these communication is to design from DG units using the communication system, average them
a wireless network control system (WNCS) taking account of and generates the proper control signal for the primary control
all the aspects of the ad hoc network [114]. In this design, level. The schematic diagram of this proposed control scheme
the distributed control will generate two components at each can be seen in Fig. 39 and the detailed explanation can be
sampling period, one is local controller which is classical or found in [115].
modern control and the other is correction component of the
controller which compensates for the ad hoc network quality Controller realized by
Synchronous Signal
of service (QoS) parameters. As shown in Fig. 36, with the a PC

combination of a local controller, correction component and Ethercat Inverter Module with Local Controller
Frame
adaptive sampler the stability and robustness will be enhanced.

1 2 3 ...

Central
Controllert
c

Computer t ca t sc

Local
Controller

Fig. 38. Control of parallel multi-inverter system [116]

In another approach, a real time network is used for the

Fig. 36. NCS for SoS with sampler
control of parallel multi-inverter system [116]. Microgrid
makes use of this type of inverter connections to deliver
F. MG - Networked Control energy.
In previous section multilevel control scheme for MG is The considered system is shown in Fig. 38. It consists of a
discussed. The primary and tertiary control levels are de- central controller, communication network and inverters with
centralized and centralized respectively because one aims their local controllers. The control strategy is as follows, local
at the control of DG and other at global optimization of controllers sends the voltage and current measurements to the
MG. Conventionally secondary control is implemented in the central controller via network frame. A closed loop control
MGCC but recently a new distributed control scheme for inside the central controller produces the satisfactory PWM
the secondary known as networked control system (NCS) is duty ratio for each inverter module and sends it back to local
proposed in [115] which is shown in Fig. 37. controller via another network frame. This is a centralized
control strategy where central controller has all the central
information.
A PID controller is used to achieve better inverter per-
formance [117] and D-partition method is used to determine
the stability region of PI controller [118]. It was found that
network induced delay brings about a considerable effect on
closed loop control of a single inverter. The practical model
implementing the NCS is shown in Fig. 40.

XI. C OMPARATIVE A NALYSIS

After discussing the control techniques, it is worth perform-
ing the comparative analysing of the control techniques which
are most commonly used. In this section, we will classify the
control techniques considering vital aspects for the purpose of
simplification and better understanding. Control strategies for
MG are very vast and detailed comparison of each techniques
Fig. 37. Networked controlled microgrid [115] with another is out of scope of this paper.
MG control (depending on architecture) can be generally
This strategy is proposed for power electronically based classified into two main streams namely Centralized, Dis-
MG’s. The primary and secondary control are implemented in tributed (or Decentralized). Multi-Level control is also most
DG unit. The primary control which is generally droop control widely used one but again depending on architecture of
is already discussed in section VII. The secondary control has control levels, even this control techniques falls in former
frequency, voltage and reactive power controls in a distributed mentioned categories. For instance in multi-level control, the
manner. The secondary control gathers all the measurements secondary control level can be single (Centralized) [109] or
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 16

Fig. 39. Distributed secondary control [115]

Central Controller

iref + - Ethernet Communication

Current
Reference Disturbance
He-13 1i
ide
+
+ Z Z
Uref + +
Zd
G
Voltage Reference Zload

Ethernet Communication
iref + -
Current
Reference
Disturbance
He-13 ide
+
+ Z Z
Uref + + 2
i
Zd
G
Voltage Reference

iref + - Ethernet Communication

Current Reference
Disturbance
ide
He-13
+
+ Z Z
Uref + +
Zd 3i
G
Voltage Reference

Fig. 40. NCS controlled parallel multi-inverter model [116]

SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 17

it can be implemented in a distributed way [115]. Fig. (41) are presented in both grid-connected and autonomous mode.
shows the performance of the controller when implemented in The control techniques of microgrid available in literature for
centralized and decentralized manner. The time delay in the various modes of operation are also discussed. The microgrid
communication network is taken in account here. It is obvious can be viewed as a special case of SoS. It can be concluded
from the figure that, the control implemented in distributed that using networked control system, a better control of
way is better. microgrid can be obtained.

ACKNOWLEDGMENTS
The authors would like to thank the Associate Editor and
the reviewers for their constructive comments on our initial
submission. This work is supported by the deanship for
scientific research (DSR) at KFUPM through group research
project RG-1316-1.

R EFERENCES
[1] IEEE Standards Coordinating Committee 21, "IEEE Std.1547.4-2011:
IEEE Guide for Design, Operation, and Integration of Distributed
Resource Island Systems with Electric Power Systems", Jul. 2011.
[2] A. Llria, O. Curea, J. Jimenez. H. Camblong., "Survey on Microgrids:
Unplanned islanding and related inverter control techniques," Int. J.
Renewable Energy, vol.36, pp. 2052–2061, 2011.
[3] I. Bae, J. Kim, "Reliability Evaluation of Customers in a Microgrid",
Fig. 41. Performance analysis of centralized and distributed controllers IEEE Trans. Power Systems, vol. 23, no. 3, pp. 1416–1422, Aug. 2008.
[4] P. Barker, D. Herman, "Technical And Economic Feasibility of
Microgrid- Based Power Systems", Proc. Seventh EPRI Distributed
Depending on the power sharing, the methods available in Resources Conference and Exhibition Dallas, TX March 20–22, 2002.
literature can be broadly divided in two groups, one using com- [5] R. Lasseter, A. Akhil, C. Marnay, J. Stephens, J. Dagle, R. Gut-
munication medium and another is non-communication based. tromson, A. Sakis Meliopoulous, R.Yinger, and J. Eto, "Integration of
Distributed Energy Resources-The CERTS MicroGrid Concept", Rep.
Because of thier inherent advantages, only communication- LBNL-50829, April 2002.
less control which is also known as Droop-based Control is [6] H. Zeineldin, M. I. Marei, E. F. El-Saadany, and M. M. A. Salama,
used. It was initially proposed in [79] and since then there "Safe controlled islanding of inverter based distributed generation",
IEEE 35th Annual Power Electronics Specialists Conference, PESC
were many variations performed keeping the basic idea droop 04, vol. 4, pp. 2515–2520, June 2004.
control [71], [10], [86]. One such comparision is shown in Fig [7] R. H. Lasseter, "MicroGrids", IEEE Power Engineering Society Winter
(42). A modified droop control technique, designed to improve Meeting, vol. 1, pp. 305–308, 2002.
[8] G. Alex, B. Sauser, and J. Boardman. "System-of-systems engineering
the reactive power sharing among the DG units, is proposed management: a review of modern history and a path forward," IEEE
and analysed with the conventional one [86]. Systems Journal, vol 2, no. 4, pp. 484–499, 2008.
[9] C. Keating, R. Rogers, R. Unal, D. Dryer, A. Sousa-Poza, R. Safford,
W. Peterson, and G. Rabadi, "System of systems engineering," Engi-
neering Management Journal, vol. 15, pp. 36–43, 2003.
[10] A .Arulampalam, et al. "Control of power electronic interfaces in
distributed generation microgrids." Int. J. Electronics, vol. 91, no. 9,
pp. 503–523, 2004.
[11] R. H. Lasseter et al.,"White paper on integration of distributed en-
ergy resources-the CERTS Microgrid concept" [Online], Available :
http://certs.lbl.gov/pdf/50829-app.pdf.
[12] R. H. Lasseter, "Microgrids and distributed generation." Journal of
Energy Engineering, vol. 133, no. 3, pp. 144–149, 2007.
[13] W. El-Khattam and M. M. A. Salama. "Distributed generation tech-
nologies, definitions and benefits." Electric power systems research,
vol. 71, no. 2, pp. 119–128, 2004.
[14] G. Pepermans, , et al. "Distributed generation: definition, benefits and
issues." Energy Policy, vol. 33, no. 6, pp. 787–798, 2005.
[15] H. Zareipour, K. Bhattacharya, and C. A. Canizares. "Distributed
generation: Current status and challenges." Proc. 36th Annual North
American Power Symposium (NAPS), 2004.
[16] R. H. Lasseter and P. Paigi, "Microgrid: A conceptual solution." IEEE
35th Annual Power Electronics Specialists Conference, PESC 04, vol.
Fig. 42. Performance analysis of primary controllers 6, 2004.
[17] B. Lasseter,"Microgrids [distributed power generation]", IEEE Power
Engineering Society Winter Meeting, Columbus, Ohio, vol. 1, pp.
146âĂŞ-149, Feb 2001.
XII. C ONCLUSIONS [18] S. Morozumi, "Micro-grid demonstration projects in Japan." IEEE
The role of Microgrid in penetration of DG’s in the present Power Conversion Conference-Nagoya, PCC’07, 2007.
[19] J. M. Carrasco, et al. "Power-electronic systems for the grid integra-
utility network is discussed. Modeling of microgrid is a key as- tion of renewable energy sources: A survey." IEEE Trans. Industrial
pect and the recent developments in the modeling of microgrid Electronics, vol. 53, no. 4, pp. 1002–1016, 2006.
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 18

[20] F. Blaabjerg, Z. Chen, and S. B. Kjaer, "Power electronics as efficient [43] Parniani, M., and M. R. Iravani. "Computer analysis of small-
interface in dispersed power generation systems." IEEE Trans. Power signal stability of power systems including network dynamics." IEE
Electronics, 19.5 (2004): 1184-1194. Proceedings-Generation, Transmission and Distribution 142.6 (1995):
[21] N. Hatziargyriou et al.,"Microgrids:An Overview of Ongoing Research, 613-617.
Development, and Demonstration Projects." IEEE power & energy [44] European Research Project MicroGrids. [Online]. Available:
magazine, July-August 2007. http://microgrids.power.ece.ntua.gr/.
[22] US Department of Energy Electricity Distribution [45] J. A. P. Lopes, et al., Management of micro-
Programme,"Advanced Distribution Technologies and grids. presented at JIEEC2003. [Online]. Available:
Operating Concepts- Microgrids" [Online], Available: http://microgrids.power.ece.ntua.gr/documents/Microgrids _
http://www.electricdistribution.ctc.com/MicroGrids.htm. management _ jieec2003.pdf.
[23] H. Farhangi, "The path of the smart grid,"Power and Energy Magazine, [46] Lopes, JA Pecas, C. L. Moreira, and A. G. Madureira. "Defining control
IEEE 8.1 (2010): 18-28. strategies for microgrids islanded operation." Power Systems, IEEE
[24] S. N. Bhaskara, M. Rasheduzzaman, and B. H. Chowdhury, Transactions on 21.2 (2006): 916-924.
"Laboratory-based Microgrid Setup for Validating Frequency and Volt- [47] Zhu, Y., and K. Tomsovic. "Development of models for analyzing the
age Control in Islanded and Grid-Connected Modes." Green Technolo- load-following performance of microturbines and fuel cells." Electric
gies Conference, 2012 IEEE. IEEE, 2012. Power Systems Research 62.1 (2002): 1-11.
[25] F. D. Kanellos, A. I. Tsouchnikas, and N. D. Hatziargyriou, "Micro-grid [48] Padulles, Joel, G. W. Ault, and J. R. McDonald. "An integrated SOFC
simulation during grid-connected and islanded modes of operation." plant dynamic model for power systems simulation." Journal of power
International Conference on Power Systems Transients. Vol. 6. 2005. sources 86.1 (2000): 495-500.
[26] Katiraei, F., M. R. Iravani, and P. W. Lehn, "Micro-grid autonomous [49] N. Hatziagyriou , et al.,"Modeling of micro-sources for security stud-
operation during and subsequent to islanding process." Power Delivery, ies.", CIGRE session Paris 2004.
IEEE Transactions on 20.1 (2005): 248-257. [50] Georgakis, D., et al. "Operation of a prototype microgrid system
[27] A. Dobakhshari, S. Salehi, S. Azizi and A. M. Ranjbar, "Control of based on micro-sources quipped with fast-acting power electronics
microgrids: Aspects and prospects." Networking, Sensing and Control interfaces." Power Electronics Specialists Conference, 2004. PESC 04.
(ICNSC), 2011 IEEE International Conference on. IEEE, 2011. 2004 IEEE 35th Annual. Vol. 4. IEEE, 2004.
[28] D. J. Lee and L. Wang, "Small-signal stability analysis of an au- [51] Engler, Alfred. "Applicability of droops in low voltage grids." Int.
tonomous hybrid renewable energy power generation/energy storage Journal of Distributed Energy Resources 1.1 (2005): 1-6.
system part I: time-domain simulations." Energy Conversion, IEEE [52] Lopes, JA PeÃğas, C. L. Moreira, and F. O. Resende. "Microgrids
Transactions on 23.1 (2008): 311-320. black start and islanded operation." 15th Power Systems Computation
[29] T. Senjyu, et al., "A hybrid power system using alternative energy Conference (PSCC), Liege. 2005.
facilities in isolated island." Energy Conversion, IEEE Transactions on [53] Lasseter, Robert, and P. Piagi. "Providing premium power through
20.2 (2005): 406-414. distributed resources." System Sciences, 2000. Proceedings of the 33rd
[30] Karimi, Houshang, Hassan Nikkhajoei, and R. Iravani, "Control of Annual Hawaii International Conference on. IEEE, 2000.
an electronically-coupled distributed resource unit subsequent to an [54] R. Caldon, et al.,"Analysis of dynamic performance of dispersed
islanding event." Power Delivery, IEEE Transactions on 23.1 (2008): generation connected through inverters to distribution networks.",
493-501. 17th International Conference on Electricity Distribution - CIRED,
Barcelona (May 2003).
[31] F. Gao and M. R. Iravani, "A control strategy for a distributed
generation unit in grid-connected and autonomous modes of operation." [55] Barsali, Stefano, et al. "Control techniques of dispersed generators
Power Delivery, IEEE Transactions on 23.2 (2008): 850-859. to improve the continuity of electricity supply." Power Engineering
Society Winter Meeting, 2002. IEEE. Vol. 2. IEEE, 2002.
[32] H. Karimi, E. J. Davison, and R. Iravani, "Multivariable servomecha-
[56] El-Sharkh, M. Y., et al. "Analysis of active and reactive power control
nism controller for autonomous operation of a distributed generation
of a stand-alone PEM fuel cell power plant." Power Systems, IEEE
unit: Design and performance evaluation." Power Systems, IEEE
Transactions on 19.4 (2004): 2022-2028.
Transactions on 25.2 (2010): 853-865.
[57] S. Papathanassiou, et al., Study-Case LV Network. [Online]. Available:
[33] Popov, Marjan, et al. "Modeling, control and islanding detection of
http://microgrids.power.ece.ntua.gr/documents/Study-Case%20LV-
microgrids with passive loads." Power Electronics and Motion Control
Network.pdf.
Conference (EPE/PEMC), 2010 14th International. IEEE, 2010.
[58] Papathanassiou, Stavros, Nikos Hatziargyriou, and Kai Strunz. "A
[34] M. Babazadeh and H. Karimi, "Robust decentralized control for is- benchmark low voltage microgrid network." Proceedings of the CIGRE
landed operation of a microgrid." Power and Energy Society General Symposium: Power Systems with Dispersed Generation. 2005.
Meeting, 2011 IEEE. IEEE, 2011.
[59] Soultanis, Nikos L., Stavros A. Papathanasiou, and Nikos D. Hatziar-
[35] R. Moradi, H. Karimi, and M. K. Ghartemani, "Robust decentralized gyriou. "A stability algorithm for the dynamic analysis of inverter dom-
control for islanded operation of two radially connected DG systems." inated unbalanced LV microgrids." Power Systems, IEEE Transactions
Industrial Electronics (ISIE), 2010 IEEE International Symposium on. on 22.1 (2007): 294-304.
IEEE, 2010.
[60] Diaz, GuzmÃan, ˛ et al. "Complex-Valued State Matrices for Simple
[36] N. Pogaku, M. Prodanovic, and T. C. Green, "Modeling, analysis and Representation of Large Autonomous Microgrids Supplied by <formula
testing of autonomous operation of an inverter-based microgrid." Power type>." Power Systems, IEEE Transactions on 24.4 (2009): 1720-1730.
Electronics, IEEE Transactions on 22.2 (2007): 613-625. [61] M. Arnold, R. R. Negenborn, G. Andersson and B. D. Schutter,
[37] M. A. Hassan and M. A. Abido, "Optimal design of microgrids in "Multi-area predictive control for combined electricity and natural
autonomous and grid-connected modes using particle swarm optimiza- gas systems". Proceedings of the European Control Conference 2009,
tion" IEEE Trans. Power Electronics, vol. 26, no. 3, pp. 755–769, Budapest, Hungary, pp. 1408âĂŞ-1413, Aug. 2009.
March 2011. [62] del Real, A. J., Arce, A., and Bordons, C., "Hybrid model predictive
[38] J. M. Uudrill, "Dynamic stability calculations for an arbitrary num- control of a two-generator power plant integrating photovoltaic panels
ber of interconnected synchronous machines." Power Apparatus and and a fuel cell". Proceedings of the 46th IEEE Conference on Decision
Systems, IEEE Transactions on 3 (1968): 835-844. and Control, New Orleans, LA, USA, Dec. 12-14, pp. 5447–5452,
[39] F. Katiraei, M. R. Iravani, and P. W. Lehn, "Small-signal dynamic 2007.
model of a micro-grid including conventional and electronically inter- [63] Whitefoot, J. W., Mechtenberg, A. R., Peters, D. L., and Papalambros,
faced distributed resources." Generation, Transmission & Distribution, P. Y., "Optimal component sizing and forward-looking dispatch of an
IET 1.3 (2007): 369-378. electrical microgrid for energy storage planning". In Proceedings of
[40] Krause, Paul C., Oleg Wasynczuk, and Scott D. Sudhoff. "Analysis of the ASME 2011 International Design Engineering Technical Confer-
electric machinery and drive systems." (2002). ences & Computers and Information in Engineering Conference, Paper
[41] Schauder, C., and H. Mehta. "Vector analysis and control of advanced number DETC2011-48513.
static VAR compensators." Generation, Transmission and Distribution, [64] Peters, D.L., Mechtenberg, A.R., Whitefoot, J.W., and Papalambros,
IEE Proceedings C. Vol. 140. No. 4. IET, 1993. P.Y., "Model Predictive Control of a Microgrid with Plug-in Vehicles:
[42] Faridaddin Katiraei, "Dynamic analysis and control of distributed en- Error Modeling and the Role of Prediction Horizon," ASME Dynamic
ergy resources in a micro-grid.", Ph. D. Thesis Dissertation, University Systems and Controls Conference, Arlington, VA, Oct. 31 - Nov. 2,
of Toronto, 2005. DSCC2011-6030, 2011.
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 19

[65] Zamora, R., and Srivastava, A., "Controls for microgrids with storage: [88] Davison, E., and I. Ferguson. "The design of controllers for the multi-
Review, challenges, and research needs". Renewable and Sustainable variable robust servomechanism problem using parameter optimization
Energy Reviews, vol. 14, no. 7, pp. 2009-2018, 2010. methods." IEEE Trans. Automatic Control, vol. 26, no. 1, pp. 93–110,
[66] Ersal, T., Ahn, C., Hiskens, I. A., Peng, H. Stefanopoulou, A. and Stein, 1981.
J. L., "On the effect of DC Source Voltage on Inverter-Based Frequency [89] Davison, D. E., and E. J. Davison. "Optimal transient response shaping
and Voltage Regulation in a Military Microgrid." Proceedings of the of the servomechanism problem." Journal of optimization theory and
American Control Conference, Fairmont Queen Elizabeth, MontrÃl’al, applications 115.3 (2002): 491-515.
Canada, June 27-June 29, pp. 2965–2971, 2012. [90] Qiu, L., and E. J. Davison. "New perturbation bounds for the robust
[67] Ersal, T., Ahn, C., Peters, D. L., Whitefoot, J. W., Mechtenberg, A. stability of linear state space models." Decision and Control, 1986 25th
R., Hiskens, I. A., Peng, H., Stefanopoulou, A. G., Papalambros, IEEE Conference on. Vol. 25. IEEE, 1986.
P. Y. and Stein, J. L., "Coupling Between Component Sizing and [91] Maier, Mark W. "Architecting principles for systems-of-systems."
Regulation Capability in Microgrids", IEEE Trans. Smart Grid, DOI Systems Engineering 1.4 (1998): 267-284.
10.1109/TSG.2013.2260363. [92] Hipel, Keith W., et al. "The future of systems, man, and cybernetics:
[68] Y. W. Li, et al.,"Design, Analysis and Real-Time Testing of Controllers application domains and research methods." Systems, Man, and Cyber-
for Multi-Bus Microgrid System", IEEE Trans. Power Electron., vol. netics, Part C: Applications and Reviews, IEEE Transactions on 37.5
19, pp. 1195-1204, Sept. 2004. (2007): 726-743.
[69] Katiraei, F., and M. R. Iravani. "Power management strategies for a [93] M. S. Mahmoud, “Resilient Control of Uncertain Dynamical Systems,
microgrid with multiple distributed generation units." Power Systems, Springer-Verlag, Heidelberg, 2004.
IEEE Transactions on 21.4 (2006): 1821-1831. [94] M. S. Mahmoud, Decentralized Control and Filtering in Intercon-
nected Dynamical Systems, CRC Press, New York, 2010.
[70] Peng, Fang Z., Yun Wei Li, and Leon M. Tolbert. "Control and protec-
[95] M. S. Mahmoud, Decentralized Systems with Design Constraints,
tion of power electronics interfaced distributed generation systems in a
Springer-Verlag, UK, 2011.
customer-driven microgrid." Power & Energy Society General Meeting,
[96] M. S. Mahmoud, Distributed Control and Filtering for Industrial
2009. PES’09. IEEE. IEEE, 2009.
Systems, IET Press, UK, December 2012
[71] Marwali, Mohammad N., Jin-Woo Jung, and Ali Keyhani. "Control of
[97] M. Jamshidi, System of Systems Engineering- Principles and Applica-
distributed generation systems-Part II: Load sharing control." Power
tion, Taylor Francis CRC Publishers, Boca Raton, FL, USA, 2008.
Electronics, IEEE Transactions on 19.6 (2004): 1551-1561.
[98] M. Jamshidi, System of Systems Engineering- Innovations for the 21
[72] Dai, Min, et al. "Power flow control of a single distributed generation Century, Wiley & sons Inc., NY, 2009.
unit." Power Electronics, IEEE Transactions on 23.1 (2008): 343-352. [99] Davison, E., "The robust decentralized control of a general servomech-
[73] Oyarzabal, J., et al. "Agent based micro grid management system." anism problem." IEEE Trans. Automatic Control, 21.1 (1976): 14-24.
Future Power Systems, 2005 International Conference on. IEEE, 2005. [100] K. Zhou, et al., Essentials of Robust Control, Prentice Hall, 1997.
[74] Dimeas, Aris L., and Nikos D. Hatziargyriou. "Operation of a multia- [101] S. Skogestad and I. Postlethwaite, Multivariable Feedback Control:
gent system for microgrid control." Power Systems, IEEE Transactions Analysis and Design Vol. 2. New York: Wiley, 2007.
on 20.3 (2005): 1447-1455. [102] Song, Hong-seok, and Kwanghee Nam. "Dual current control scheme
[75] Nagahara, M., et al. "H-infinity control of microgrids involving gas for PWM converter under unbalanced input voltage conditions." Indus-
turbine engines and batteries." Proc. of 51th IEEE Conference on trial Electronics, IEEE Transactions on 46.5 (1999): 953-959.
Decision and Control. 2012. [103] Yazdani, Amirnaser, and Reza Iravani, "A unified dynamic model
[76] Yee, Soon Kiat, Jovica V. Milanovic, and F. Michael Hughes. and control for the voltage-sourced converter under unbalanced grid
"Overview and comparative analysis of gas turbine models for system conditions." Power Delivery, IEEE Transactions on 21.3 (2006): 1620-
stability studies." IEEE Trans. Power Systems, vol. 23, no. 1, pp. 108– 1629.
118, 2008. [104] Etemadi, Amir H., Edward J. Davison, and Reza Iravani, "A decen-
[77] Nazari, M. Honarvar, and M. Ilic. "Technical challenges in modernizing tralized robust control strategy for multi-DER microgridsâĂŤPart I:
distribution electric power systems with large number of distributed Fundamental concepts." (2012): 1-1.
generators." Proc. IEEE Power Tech, 2009. [105] H. Liang; B. J. Choi, W. Zhuang and X. Shen, "Stability Enhancement
[78] Pedrasa, Michael Angelo, and Ted Spooner. "A survey of techniques of Decentralized Inverter Control Through Wireless Communications
used to control microgrid generation and storage during island oper- in Microgrids," Smart Grid, IEEE Transactions on , vol.4, no.1,
ation." Proceedings of the Australian Universities Power Engineering pp.321,331, March 2013.
Conference. 2006. [106] Y. A. R. I. Mohamed and E. F. El-Saadany. "Adaptive decentralized
[79] Chandorkar, Mukul C., Deepakraj M. Divan, and Rambabu Adapa. droop controller to preserve power sharing stability of paralleled in-
"Control of parallel connected inverters in standalone ac supply sys- verters in distributed generation microgrids." Power Electronics, IEEE
tems." IEEE Trans. Industry Applications, vol. 29, no. 1, pp. 136–143, Transactions on 23.6 (2008): 2806-2816.
1993. [107] Guerrero, Josep M., et al, "Decentralized control for parallel operation
[80] Baudoin, Sylvain, Ionel Vechiu, and Haritza Camblong. "A review of of distributed generation inverters using resistive output impedance."
voltage and frequency control strategies for islanded microgrid." The Industrial Electronics, IEEE Transactions on 54.2 (2007): 994-1004.
16th Int. IEEE Conference System Theory, Control and Computing [108] Vasquez, Juan C., et al, "Hierarchical control of intelligent microgrids."
(ICSTCC), 2012. Industrial Electronics Magazine, IEEE 4.4 (2010): 23-29.
[109] Guerrero, Josep M., et al, "Hierarchical control of droop-controlled
[81] Piagi, Paolo, and Robert H. Lasseter. "Autonomous control of micro-
AC and DC microgridsâĂŤA general approach toward standardization."
grids." Proc. IEEE Power Engineering Society General Meeting, 2006.
Industrial Electronics, IEEE Transactions on 58.1 (2011): 158-172.
[82] Lasseter, R. H., et al. "CERTS microgrid laboratory test bed." IEEE
[110] W. Zhang, M. S. Branicky, and S. M. Phillips, "Stability of networked
Trans. Power Delivery, vol. 26, no. 1, pp. 325–332, 2011.
control systems." Control Systems, IEEE 21.1 (2001): 84-99.
[83] Sao, Charles K., and Peter W. Lehn. "Autonomous load sharing of [111] J. P. Hespanha, P. Naghshtabrizi, and Y. Xu, "A survey of recent results
voltage source converters." IEEE Trans. Power Delivery, vol. 20, no. in networked control systems." Proceedings of the IEEE 95.1 (2007):
2, pp. 1009–1016, 2005. 138-162.
[84] Li, Yun Wei, and Ching-Nan Kao. "An accurate power control strategy [112] J. Baillieul and P. J. Antsaklis, "Control and communication challenges
for power-electronics-interfaced distributed generation units operating in networked real-time systems." Proceedings of the IEEE 95.1 (2007):
in a low-voltage multi-bus microgrid." IEEE Trans. Power Electronics, 9-28.
vol. 24, no. 12, pp. 2977–2988, 2009. [113] B. Kelley, M. Jamshidi and D. Akopian, "MRI Development of a Test-
[85] Guerrero, Josep M., et al. "Output impedance design of parallel- bed Swarm of Autonomous Unmanned Vehicles Ad-hoc Networked
connected UPS inverters with wireless load-sharing control." IEEE Control Via Secure 4G Wireless." NSF MRI Proposal (2009).
Trans. Industrial Electronics, vol. 52, no. 4, pp. 1126–1135, 2005. [114] J. Colandairaj, G. W. Irwin, and W. G. Scanlon, "Wireless networked
[86] Lee, C-.T.; Chu, C-.C.; Cheng, P-.T., "A new droop control method control systems with QoS-based sampling." Control Theory & Appli-
for the autonomous operation of distributed energy resource interface cations, IET 1.1 (2007): 430-438.
converters," IEEE Trans. Power Electronics, vol. 28, no. 4, pp. 1980– [115] Q. Shafiee, J. C. Vasquez, and J. M. Guerrero, "Distributed sec-
1993, April 2013. ondary control for islanded MicroGrids-A networked control systems
[87] Franklin, Gene F., J. David Powell, and Abbas Emami-Naeini. Feed- approach." IECON 2012-38th Annual Conference on IEEE Industrial
back Control of Dynamics Systems. Prentice Hall Inc, New York, 1986. Electronics Society. IEEE, 2012.
SUBMITTED TO J. THE FRANKLIN INSTITUTE: REVISED 20

[116] W. Zheng, H. Ma and X. He, "Modeling, analysis, and implementa-

tion of real time network controlled parallel multi-inverter systems."
Power Electronics and Motion Control Conference (IPEMC), 2012 7th
International. Vol. 2. IEEE, 2012.
[117] l. Peng, Research on Control Technique for PWM Inverters Based
on State-Space Theory, Ph. D. dissertation, Hua zhong University of
Science & Technology Wuhan, P. R. China, September, 2004.
[118] O. L. Lin, G. Dan-ying, Z. Wei-dong, "Optimal tuning method of PID
controller based on gain margin and phase margin", Control Theory &
Applications, vol. 24, no.5, pp. 837-840, Oct. 2007.

View publication stats