IEEE/CAA JOURNAL OF AUTOMATICA SINICA
Networked Control Approach for Distributed
Generation Systems
Magdi S. Mahmoud and Mohamed Saif Ur Rahman
Abstract—Microgrid has emerged as an answer to growing
demand for distributed generation (DG) in power systems. It
contains several DG units including microalternator, photovoltaic
system and wind generation. It turns out that sustained operation
relies on the stability of these constituent systems. In this paper, a
microgrid consisting of microalternator and photovoltaic system
is modeled as a networked control system of systems (SoS)
subjected to packet dropouts and delays. Next, an observer-
based controller is designed to stabilize the system in presence
of the aforementioned communication constraints and simulation
results are provided to support the control design methodology.
Index Terms—Networked control; Microgrid systems;
Observer-based controller; Microalternator; Photovoltaic
Systems
I. I NTRODUCTION
E CONOMIC challenges, technological advancements and
environmental impacts are now demanding distributed
generation in place of the conventional centralized gener-
ation [1]. Power operation companies are now confronted
with unprecedented difficulties in terms of meeting the load
requirements, consumer satisfaction and environmental con-
siderations. Thus, distributed generation has received good
attention because of its potential to alleviate pressure from
the main transmission system by supplying a few local loads
[2]. The waste heat generated from the fuel to electricity
conversion is exploited by the distributed generation system
with the help of microturbines, reciprocating engines and fuel
cells to provide heat and power to the customers. Adding to
the system distributed energy sources (DES) like photovoltaic
(PV) panels, wind turbines, energy storage devices such as
batteries and capacitors, generators and controllable loads. The
distributed generators can extract energy from other renewable
nearby systems and provide momentous contributions to fu-
ture energy generation and distribution. Another noteworthy
feature is that the carbon emission is reduced to a large
extent satisfying the commitment of many nations concerning
decrease of carbon footprints [3]. However, the distributed
This article has been accepted for publication in a future issue of this
journal, but has not been fully edited. Content may change prior to final
publication.
Recommended by Associate Editor Li Chengdong. (Corresponding author:
Magdi S. Mahmoud.)
Citation: M. S. Mahmoud and M. S. U. Rahman, “Networked control
approach for distributed generation systems,” IEEE/CAA J. of Autom. Sinica,
pp. 1−17, 2017. DOI: 10.1109/JAS.2017.7510688.
M. S. Mahmoud and M. S. U. Rahman are with Systems Engineering
Department, King Fahd University of Petroleum and Minerals, P. O. Box
5067, Dhahran 31261, Saudi Arabia (e-mail: msmahmoud@kfupm.edu.sa,
saif.rahman02@gmail.com).
Color versions of one or more of the figures in this paper are available
online at http://ieeexplore.ieee.org.
1
generation faces technical issues regarding its connection to
the intermittent renewable generation and feeble areas of the
distribution network. Further, owing to the distinct behavior
of the distributed generation unlike the conventional load,
alteration in power flow results in problems. To counter the
irregular behavior and increasing penetration of the distributed
generation, the microgrid was introduced.
The microgrid has paved its way into distributed generation
and looks promising for future prospects. It has the ability to
respond to changes in the load, while decreasing feeder losses
and improving local reliability. Basically designed to cater the
heat and power requirements of local customers, it can serve as
uninterruptible power supply for critical loads. Several control
strategies for the microgrid have been proposed in the literature
including PI controllers in [4−12]. Robust H∞ control is
presented in [13] and [14] for the control of two distributed
generation units. An optimal controller is presented for con-
trolling the frequency and voltage fluctuations during islanded
mode in [15]. The inverters connected to these systems operate
under imbalance conditions due to sensitive loads. This leads
to switching harmonics, voltage and frequency variations in the
microgrid system and disturbs the stability of the system. Thus
modeling and control problems associated with microgrids
including multiple distributed generation units are typically
complex and interconnected in nature.
One efficient way to resolve the foregoing microgrid prob-
lem is to cast the integrated dynamic model into the system of
systems (SoS) framework for better operation and organized
control. The concept of (SoS) has opened up a new school
of thought in systems engineering. System of systems has
emerged as a hot topic for research over the past few years.
Although still in the infant stages, the concept of system of
systems has managed to achieve widespread acclaim. Being
restricted to defense and IT, at one point of time, system
of systems has now entered a plethora of domains [16−17].
It is worth noting that a microgrid is a complex system
comprising of a variety of systems which are nonlinear in
nature and possess strong cross-coupling between them. Hence
viewing the microgrid from an intelligent system of systems
perspective is need of the hour. Moreover an efficient control
methodology based on system of systems has to be established
in order to overcome the challenges posed by the microgrid.
The concept of System of systems is now widespread and
has entered several domains including defense, IT, health care,
manufacturing, energy and space stations and exploration to
name a few.
Recent progress in analysis and design issues of networked
control systems is reported in [18−22]. Additional related
2 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

work are published in [23]. Networked control of system is discussed.

of systems has been introduced in [24]. A control system
consisting of a real time network in its feedback can be termed II. M ODELING OF THE M ICRO A LTERNATOR -PV S YSTEM
as a networked control system (NCS) [25]. As mentioned in To model the microgrid system consisting of microalternator
[24] that the need to design a SoS control system which can and photovoltaic system, we consider the separate modeling
tolerate packet loss and delays is one of the prime challenges of microalternator and photovoltaic system initially. After
in SoS networked control, we considered a network which is modeling both these individual systems, we integrate them
subjected to both delays and non-stationary packet dropouts into a microgrid system which is connected to a load and the
and the controller stabilizes the system in the presence of main grid. Two sets of such systems are considered eventually
these communication issues. The controller design for such which form a system of systems structure for the microgrid
a networked control is presented in [26]. A further extension system.
is made in the controller design by introducing a distinct X̂
matrix in the linear matrix inequality (LMI) that is used to A. Micro Alternator
obtain the controller gain matrices and a realistic approach is
adopted to compute the gains. The swing equation of the alternator can be written as two
The main objective of this paper is to formulate a system first order differential equations [27]:
of systems framework including a communication network dδ dω 1
= ω0 (ω − 1), = (Pm − Pe ) (1)
for the distributed generation units of a microgrid and conse- dt dt 2H
quently design a feedback controller to guarantee closed-loop where δ and ω are rotor angle and rotor speed. Pm and Pe
stabilization in the presence of communication constraints, are mechanical power input and electrical power output of the
network delays and packet dropouts. The considered microgrid generator. The internal voltage e0q is given by
system comprised two sets of microalternator and photovoltaic
de0q 1
(PV) panels as distributed generation units, which are even- = 0 0
dt 0 [Ef d − eq − (xd − xd )itd ]
Tdo
(2)
tually connected to a load and the main grid. To achieve
our objective, the work reported herein combines two parts: where xd , x0d and Tdo 0
are the d-axis synchronous resistance,
a detailed modeling and analysis part leading to establish transient reactance and open circuit field constants respec-
a stabilizing controller and real-time implementation part to tively. e0q is the voltage behind the transient reactance along
illustrate the performance. This latter part is achieved at the q-axis. An IEEE type ST is used for the voltage regulator
SOS research lab at KFUPM using pilot-scale lab equipment excitation.
to build an experimental microgrid including up to 12 DG dEf d 1
units. The simulation tools are Matlab-Simulink environment, = [KA (Vtref − Vt ) − (Ef d − Ef do )] (3)
dt TA
particularly the products Simscape Power Systems Examples
and RAPSim an open source simulation software for micro- where Ef d is the field voltage along d-axis, KA and TA are
grids. the gain and time constants of the exciter. Fig. 1 shows a
The main contributions of this work are: microalternator connected to the main grid.
1) A generalized approach to modeling microgrids with
multiple distribution generation units is developed based on
system of systems (SoS) framework. This is a novel approach
in the context of renewable energy systems.
2) The microgrid system incorporated communication net-
work to enable processing control signals and power measure-
ments. Fig. 1. Microalternator connected to grid.
3) An improved output feedback networked controller is
constructed to stabilize the microgrid system in presence of The terminal voltage of the alternator is given as
delays and non stationary packet dropouts. Vt = Vs + (rt + jxt )it
The paper is organized as follows. A brief introduction of
microgrids and system of systems is presented in section I. In d-q terms it becomes
Detailed modeling of both the microalternator and photovoltaic Vd + jVq = Vsd + jVsq
(PV) system is presented in section II. A combined framework
+ (rt + jxt )(itd + jitq )
of microalternator and PV system connected to a load and 0 0
the main grid is proposed. A state space representation of the xq itq + j(eq − xd itd ) = (Vsd + rt itd − xt itq )
combined system is presented. Further the detailed networked + j(Vsq + rt itq + xt itd )
control system, its stability analysis and controller design is
The real part can be written as
explained in section III and IV with relevant theorems. The
controller design is implemented on the proposed framework Vsd = (xq + xt )itq − rt itd (4)
and simulation results are shown for two sets of PV and
and the imaginary part can be written as
microalternator systems. Finally conclusions are drawn and the
effectiveness of the proposed framework and control strategy e0q − Vsq = rt itq + (x0d + xt )itd (5)
MAHMOUD AND RAHMAN: NETWORKED CONTROL APPROACH FOR DISTRIBUTED GENERATION SYSTEMS 3

Substituting x1 = (x0d + xt ) and x2 = (xq + xt ) and solving

for itd and itq , we obtain
−rt Vsd + (e0q − Vsq )(xq + xt )
itd =
rt2 + (x0d + xt )(xq + xt )
V̄sd = Vsd [(x0d + xt )(xq + xt )]
+ rt (e0q − Vsq )(xq + xt )
V̄sd Fig. 2. Model of PV cell.
itq = (6)
(xq + xt )[rt2 + (x0d + xt )(xq + xt )]
Hence, the terminal voltage and power output of the alternator
are given as (Vpv + Ipv Rs )
Ipv = Iph − ID − (14)
Rsh
Vt = (Vd2 + Vq2 )1/2 = ((xq itq )2 + (e0q − x0d itd )2 )1/2
Pe = Vd itd + Vq itq = (e0q itq ) + (xq − x0d )itd itq (7) The diode current ID is given as
(Vpv +Rs Ipv )
By perturbing the above nonlinear equations around a ID = Is (e nVT
− 1) (15)
normal operating point and letting [δ, ω, e0q , Ef d ] as the states,
we obtain the linearized state equations: where Ipv is the cell current, Vpv is the cell voltage, Is is the
∆δ̇ = ω0 ∆ω reverse saturation current(depends on temperature), n is the
1 ideality factor, Rs is the series resistance, Rsh is the shunt
∆ω̇ = [−∆Pe ] resistance, VT (= kT
2H q ) is the thermal voltage, k is Boltzmann
1 constant, T is working temperature of the cell and q is the
∆e˙0q = 0 0
0 [∆Ef d − ∆eq − (xd − xd )∆itd ] charge of the electron.
Tdo
KA 1 The solar irradiation and the working temperature of the
∆E˙f d = ∆Vt − ∆Ef d (8) cell determines the photo current. At given cell temperature
TA TE
T , the photo current can be expressed as
along with the generator output current:
∆it = ∆itd + j∆itq (9) Iph = [Isc + a(T − Tref )]G (16)

The first-order difference of itd can be obtained from (6) as where Isc is the short current of the cell at 250 C and 1kW/m2 ,
−rt x2 a is the temperature coefficient of Isc , Tref is the reference
∆itd = 2 ∆Vsd + 2 (∆e0q − ∆Vsq ) (10) temperature of the cell and G is the irradiation in kW/m2 .
rt + x1 x2 rt + x1 x2
The reverse saturation current also depends on temperature
Similarly obtaining the change in the q-axis generator output and can be expressed as
current as
∆Vsd T 3 −qE g 1
( − 1 )
∆itq = Is = Isref ( ) n e nk T Tref (17)
x1 Tref
rt
+ 0 (∆e0q − ∆Vsq ) (11) where Isref is the saturation current of the cell, Eg is the
xd + xt
energy band gap of the semiconductor used in the solar cell.
Further algebraic manipulation yields The characteristic equation of the cell from the approximate
Vdo Vqo model is given as
∆Vt = ∆Vd + ∆Vq
Vto Vto q(
Vpv +Ipv Rs
)
Vdo Vqo Ipv = Isc − Is [e nVT
− 1] (18)
= (xq ∆itq ) + (∆e0q − x0d ∆itd ) (12)
Vto Vto
Because of the fact that the power generated by a solar cell is
∆Pe = e0qo ∆itq + itqo ∆e0q
low, multiple solar cells are connected in series and parallel to
+ (xq − x0d )[itdo ∆itq + itqo ∆itd ] (13) generate power in the range of watts. Thus the characteristic
equation of the photovoltaic array consisting of Ns series and
B. Photovoltaic System Np parallel modules can be derived from the PV cell equation
A mathematical expression describing the I-V characteris- mentioned above
tics of a solar cell has been studied extensively. An equivalent (
Vpv
+
Ipv Rs
)/nVT
model of a solar cell shown in Fig. 2 includes a photo Ipv = Np Iph − Np Is [e Ns Np
− 1] (19)
diode, a shunt resistor depicting leakage current and a series
which is a nonlinear current voltage relationship of a PV array.
resistor representing an internal resistance to current flow. The
Alternatively, it can be rewritten as
difference between photovoltaic current Iph and sum of normal
diode current ID and leakage current in the shunt resistor Ish Np Iph − Ipv Ipv Rs
gives the net output current from the PV cell. VT = Ns [ln( + 1)nVT − ] (20)
Np Is Np
4 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

This can be iteratively solved using Newton-Raphson algo- DC Link Capacitor Model
rithm [28] of the form The DC link capacitor functions as an energy storage and
f (xn ) filter for the DC voltage.
xn+1 = xn − 0 , By applying Kirchhoff current law (KCL) at the DC-link
f (xn )
continues xn+1 − xn
node, the dynamics of the DC link capacitor can be obtained
−→ | | ≤ Es (21) as
xn
where Es is a prescribed tolerance and xn is the nth solution. dVdcp 1
= (Idc1 − Idc2 ) (24)
Letting dt Cdc
xn = VT,n , where Idc1 = (1 − dc )Ipv and Idc2 is the input current to
Np Iph − Ipv − Np Is the inverter, which is derived later in this section. Vdc is the
f (VT ) = VT − Ns [nVT ∗ ln( ) voltage across capacitor Cdc .
Np Is
Ipv R Inverter model
+ ] The inverter is responsible for the conversion of the PV
Np
array DC output and giving it to the grid at an appropriate fre-
Since f 0 (VT ) = 1, the recursive formula (20) can be written quency. A voltage gain model of a voltage source inverter(VSI)
as operating in PWM mode is considered as shown in Fig. 4.
Np Iph − Ipv − Np Is
Vn+1 = Ns [nVT ∗ ln( )
Np Is
Ipv R
+ ] (22)
Np

C. Power Conditioning Unit of PV System

The power conditioning unit (PCU) consists of devices
needed to connect the PV array to the microgrid [29]. The
significant components of the PCU are
· DC/DC Converter
· DC link Capacitor
· Inverter
Fig. 4. Inverter model.
· Output filter circuit
DC/DC Converter Model The power on DC side of the inverter is given as
The primary function of the DC/DC converter is to either
increase or decrease the DC output voltage. Out of the various Pdc = Vdcp Idc2 (25)
topologies of DC/DC converters, buck and boost converters are
the more fundamental ones. Because the PV output voltage The instantaneous active power on the AC side of the inverter
has be to stepped up, a boost converter is used in this case. A is expressed as
typical converter configuration is shown in Fig. 3. The boost ∗
Pac = Re[Vp Ipf ] (26)
converter steps up the DC voltage level. It consists of an
inductor, a diode and a power electronic switch. where Vp and Ipf are the inverter output voltage and inverter
output current respectively. In d-q terms, Vp and Ipf can be
expressed as

Vp = Vpd + jVpq , Ipf = Ipf d + jIpf q

Little algebra yields the power relations

Pac = Vpd Ipf d + Vpq Ipf q ,

Pdc = Vpd Ipf d + Vpq Ipf q (27)
Fig. 3. DC/DC Converter Configuration. During its operation in PWM mode and referring to Fig. 4, the
output voltage of the inverter can be written as
The dynamics of the converter can be expressed as
Vpv = Ldc I˙pv + (1 − dc )Vdcp Vp = mp + Vdcp ψp (28)
1 where mp is the modulation index and ψp is the phase angle
−→ I˙pv = (Vpv − (1 − dc )Vdcp ) (23)
Ldc of the inverter. In d-q terms, this becomes
where Ldc is the inductance of the converter and dc is the duty
cycle defined as the ratio of ON period to the switching time Vpd = mp ∗ Vdcp ∗ cos(ψp + θ),
period (T ). Vpq = mp ∗ Vdcp ∗ sin(ψp + θ) (29)
MAHMOUD AND RAHMAN: NETWORKED CONTROL APPROACH FOR DISTRIBUTED GENERATION SYSTEMS 5

Finally, we obtain the expression for Idc2 as where Cpf is the filter capacitor. Once again in d-q terms, we
get
Idc2 = (Ipf d mp cos(ψp + θ) dVcpd 1
+ Ipf q mp sin(ψp + θ)) (30) = (Ipf d − Ipd ) + ω0 ωVcpq ,
dt Cpf
dVcpq 1
LC Filter and Coupling Inductance Model = (Ipf q − Ipq ) + ω0 ωVcpd (38)
The purpose of using a low pass filter is attenuation of dt Cpf
switching frequency ripple of the output voltage of an inverter. Linearized Model of the Photovoltaic System
The filter is a T section of an RL circuit shunted by a capacitor. A linearized model of the photovoltaic system includes a
While the inductor blocks high frequency harmonics, the small signal model of characteristic equation of the PV array
capacitor stops low frequency harmonics. Collectively, they and power conditioning unit. From (28) the characteristic
block most of the harmonics, thereby reducing ripples from equation of the PV arrays is
going through the system [30]. By applying Kirchhoff current Np Iph − Ipv
Vpv = Ns [ln( + 1)nVT
law (KVL) around the PV inverter and filter capacitor, we Np Is
obtain a nonlinear relation as: Ipv Rs
− ]
dIpf Np
Vp = Ipf Rpf + Lpf + Vcp
dt Making first-order changes of (20) with VT = Vpv , we get
+ (Ipf − Ip )Rpdr (31)
∆Vpv = Kpv ∆Ipv ,
where Rpf is the filter resistance, Lpf is the filter inductance, nVT
Kpv = −Ns [
Rpdr is the damping resistance and Vcp is the capacitor Np Iph − Ipvo + Np Is
voltage. Expressing in d-q frame, then (31) can be written Rs
+ ] (39)
as Np
dIpf d −ω0 Rpf By linearization each of the components of the power con-
= Ipf d + ω0 ωIpf q
dt Lpf ditioning unit, a small signal model can be readily obtained.
ω0 mp Vdcp cos(ψp + θ) The linearized state equations are given as [29]:
+ 1
Lpf ∆V̇dcp = − [−Ipf d0 mp sin(ψp + θ)∆ψp
ω0 Vcpd Cdcp
− − ω0 Rpdr Ipcd (32) + mp cos(ψp + θ)∆Ipf d
Lpf
dIpf q −ω0 Rpf + Ipf d0 cos(ψp + θ)∆mp
= Ipf q − ω0 ωIpf d
dt Lpf + Ipf q0 mp cos(ψp + θ)∆ψp + mp sin(ψp
ω0 mp Vdcp sin(ψp + θ) ω0 Vcpq + θ)∆Ipf q + Ipf q0 sin(ψp + θ)∆mp
+ −
Lpf Lpf − (1 − dc )∆Ipv ] (40)
− ω0 Rpdr Ipcq (33)
−ω0 Rpf
By coupling the transmission line between microgrid and PV ∆I˙pf d = ∆Ipf d
Lpf
filter capacitor, we obtain a nonlinear equation as + ω0 (∆Ipf q + Ipf q0 ∆ω)
dIp ω0
Vcp = Ip Rp + Lp + Vs + [mp cos(ψp + θ)∆Vdcp
dt Lpf
− (Ipf − Ip )Rpdr (34) − mp Vdcp0 sin(ψp + θ)∆ψp
ω0 ∆Vcpd
Rewriting in d-q terms, we have + Vdcp0 cos(ψp + θ)∆mp ] −
Lpf
dIpd −ω0 Rp ω0 Rpdr
= Ipd + ω0 ωIpq − (∆Ipf d − ∆Ipq ) (41)
dt Lp Lpf
ω0
+ (Vcpd − Vsd ) + ω0 Rpdr Ipcd (35) −ω0 Rpf
Lp ∆I˙pf q = ∆Ipf q
dIpq −ω0 Rp Lpf
= Ipq + ω0 ωIpd − ω0 (∆Ipf d + Ipf d0 ∆ω)
dt Lp
ω0
ω0 + [mp sin(ψp + θ)∆Vdcp
+ (Vcpq − Vsq ) + ω0 Rpdr Ipcq (36) Lpf
Lp
+ mp Vdcp0 cos(ψp + θ)∆ψp
where Ip is the coupling current, Rp is the coupling resistance ω0 ∆Vcpq
and Lp is the coupling inductance. The voltage across the + Vdcp0 sin(ψp + θ)∆mp ] −
Lpf
capacitor is given as ω0 Rpdr
− (∆Ipf q − ∆Ipq ) (42)
dVcp Lpf
Cpf = (Ipf − Ip ) (37)
dt
6 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

Vs . Microalternator and PV system dynamics have already

Rp been explained above. Both the systems had components of
∆I˙pd = − ∆Ipd + ω0 (∆Ipq Vs along d-q axes. Now to get a closed form representation of
Lp
+ Ipq0 ∆ω) the combined state model, Vsd and Vsq have to be expressed
1 in terms of the selected states. For this we apply KCL at the
+ (∆Vcpd − ∆Vsd ) common bus Vs , which gives
Lp
ω0 Rpdr
+ (∆Ipf d − ∆Ipd ) (43)
Lp
Rp
∆I˙pq = − ∆Ipq
Lp
+ ω0 (∆Ipd + Ipd0 ∆ω)
1
+ (∆Vcpq − ∆Vsd )
Lp
ω0 Rpdr
+ (∆Ipf q − ∆Ipq ) (44)
Lp
The voltage across the filter capacitor
Fig. 5. Combined system.
∆V̇cpd = ω0 (∆Vcpq + Vcpq0 ∆ω)
1
+ (∆Ipf d − ∆Ipd ) (45) It + Ip = Ib + Il
Cpf
∆V̇cpq = ω0 (∆Vcpd + Vcpd0 ∆ω) where It , Ip , Ib and Il are microalternator output current,
1 PV output current, grid current and load current respectively.
+ (∆Ipf q − ∆Ipq ) (46)
Cpf Considering the d-q components of these currents, we get

III. I NTEGRATED M ICROALTERNATOR -PV S YSTEM Itd + Ipd = Ibd + Ild (47)
M ODELING Itq + Ipq = Ibq + Ilq (48)
In the sequel, we consider a system where the microalterna-
tor and the PV system are the supply sources for a combined Further, we express the non-state currents (Itd , Itq , Ibd , Ibq , Ild , Ilq )
system. This eventually means that we have a microgrid as functions of Vsd and Vsq . However, the microalternator
with two distributed generation units. The modeling proceeds output currents Itd and Itq are already previously presented
done in two parts. Initially both subsystems are modeled in terms of Vsd and Vsq .
individually as presented in the earlier sections. An integrated Load Current
modeling approach is presented here in this section from a At the microgrid, the load is modeled as admittance Y =
novel view. g − jb. The load current is Il = Vs Y
The microalternator-PV system is modeled to represent a
Ild + jIlq = (Vsd + jVsq )(g − jb)
system of systems, where two different subsystems are oper-
ating independently to achieve a common goal. There exists Equating real and imaginary parts, we get
an integration between the two subsystems which is clearly
evident from 5. It shows that both systems are connected Ild = gVsd + bVsq , Ilq = gVsq − bVsd (49)
through a common bus having voltage Vs . Eventually, the
integrated system is connected to the main grid and a load. Grid Current
In the end, the combined model consisting of both systems The main grid current Ib is given as
is considered for the networked control system (NCS) design
V s − Vb
implementation. For this reason, the state matrices shown Ib =
are for the overall system. However, the A matrix for the rb + jxb
alternator was given just to give an idea about the four Vsd + jVsq − (Vb sinδ + jVb cosδ)
Ibd + jIbq =
states of alternator that are taken into account for the overall rb + jxb
system. It is to be noted that the overall system matrices Equating real and imaginary parts
incorporate states and dynamics from both systems forming
an integrated microalternator-PV system. Hence, A, B and C (Vsq − Vb sinδ)rb + (Vsq − Vb cosδ)xb
Ibd = (50)
matrices are given only for the overall system. The A matrix rb2 + x2b
of the microalternator is given just to describe which states (Vsq − Vb cosδ)rb + (Vsq − Vb sinδ)xb
are considered in the main system. Ibq = (51)
rb2 + x2b
Fig. 5 shows a microgrid system with a microalternator and
PV generator along with a load connected to the main grid. All Standard manipulations yield the bus voltage components Vsd
of these are connected through a common bus having voltage and Vsq as
MAHMOUD AND RAHMAN: NETWORKED CONTROL APPROACH FOR DISTRIBUTED GENERATION SYSTEMS 7

where
1 1
Vsd = [zb z1 Ipd + zb x2 e0q Θ0 = Vb [ (z1 Θ3 (rb cosδ0 − xb sinδ0 )
Θ1 Den
+ Vb z1 (rb sinδ + xb cosδ) − Θ2 Vsq ] (52) − Θ1 x2 z1 (rb sinδ0 + xb cosδ0 ))]
Vsq = Θ4 Ipd + Θ5 e0q Similarly, we also obtain
+ Θ6 Ipq + Θ7 Vb (53) 1
∆Vsd = [zb z1 ∆Ipd + zb x2 ∆e0q
where Θ1
+ Vb z1 (rb cosδ0 − xb sinδ0 )∆δ (55)
Θ1 = [gzb z1 + z1 rb + rt zb ], 0
− Θ2 (Θ4 ∆Ipd + Θ5 ∆eq + Θ6 ∆Ipq + Θ0 ∆δ)]
Θ2 = [bzb z1 + xb z1 + x2 zb ]
Finally, the values of ∆Vsd and ∆vsq are substituted in the
Θ3 = [xb x2 z1 + bzb x2 z1 + z1 zb − rt2 zb ], linearized equations of the individual component models
Θn = Θ2 Θ3 + Θ21 x2 to obtain the closed-form equation. Selecting the states as
1 [δ, ω, e0q , Ef d , Ipv , Vdcp , Ipf d , Ipf q , Ipd , Ipq , Vcpd , Vcpq ]T and
Θ4 = Θ3 zb z1 ,
Θn the control input as [mp , ψp ], we consequently obtain the
1 state matrices for the combined microalternator-PV system.
Θ5 = (Θ3 zb x2 + zb x2 rt Θ1 ) where
Θn
1 −(xd − x0d )itd1 −(xd − x0d )itd3
Θ6 = (Θ1 x2 z1 zb ), A3,1 = , A3,3 =
Θn Tdo 0 0
Tdo
1 A9,9 = −kp (Req1 + Θ11 ),
Θ7 = [z1 Θ3 (rb sinδ + xb cosδ)
Θn −Θ2 Θ6
+ Θ1 x2 z1 (rb cosδ − xb sinδ)] A9,10 = −ω0 − kp ( )
Θ1
Linearized Model of the Combined System A10,9 = −ω0 − kp Θ4 , A10,10 = −kp (Req1 + Θ6 )
A linearized model of the combined system is obtained zb z1 Θ2 Θ4
Θ11 = ( − )
by expressing the linearized microgrid voltage components Θ1 Θ1
(∆Vsd , ∆Vsq ) along d-q axes in terms of the selected states.
First-order analysis shows that IV. N ETWORKED C ONTROL S YSTEM M ODELING
Consider the microgrid with two sets of microalternator and
∆Vsq = Θ4 ∆Ipd + Θ5 ∆e0q PV system representing a system of systems (SoS)-NCS with
+ Θ6 ∆Ipq + Θ0 ∆δ (54) random communication delays, where the sensor is clock

———————————————————————————————————————————————————–
 
0 ω0 0 0 0 0 0 0 0 0 0 0
 − P2H
e1
0 − P2H
e3
0 0 0 Pe1 Pe2 0 0 0 0 
 A(3,1) 0 A(3,3) 1
0 0 eq1 eq2 0 0 0 0 
 0
Tdo 
 KA KA −1 
 V
TA t1
0 V
TA t3
0 0 0 0 0 0 0 0 
 TA
kpv 
 0 ω0 Ipf q0 0 0 (dc −1)
0 0 0 0 0 0 
 Ldc Ldc 
 0 ω0 Ipf d0 0 0 (1−dc )
0 mp kpd1 mp kpd2 0 0 0 0 
A= Cdcp 
 −kp D1 ω0 Ipd0 −kp C1 0 kpf 0 −kpf Req ω0 kpf Rd 0 −kpf 0 
 
 −kp G −ω0 Ipq0 −kp C1 0 0 0 −ω0 −kpf Req 0 kpf Rd 0 −kpf 
 0 ω0 Vcpq0 0 0 0 0 kp Rd 0 A(9,9) A(9,10) kp 0

 
 0 −ω0 Vcpq0 0 0 0 0 0 kp Rd A(10,9) A(10,10) 0 kp 
 ω0 
 0 0 0 0 0 0 Cpf
0 − Cωpf
0
0 0 ω0 
ω0
0 0 0 0 0 0 0 Cpf
0 − Cωpf
0
−ω0 0

 
0 0
 0 0 
 0 0 
 
 0 0 
 
 kpf Vdcp0 cos(ψp ) −kpf Vdcp0 mp sin(ψp ) 
 kpf Vdcp0 sin(ψp ) kpf Vdcp0 mp cos(ψp ) 
 
B= 0 0 
 
 0 0 
 0 0 
 
 0 0 
 
 0 0 
(Ipf d0 cos(ψP )+Ipf q0 sin(ψP )) −mp (Ipf d0 sin(ψP )+Ipf q0 cos(ψP ))
Cdcp Cdcp
8 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

driven and the controller and the actuator are event driven. The of computer communication management that experiences
discrete-time linear time-invariant plant model is as follows: different time-dependent operational modes, we let
xp (k + 1) = Axp + Bup , yp = Cxp (56)
P rob{δ(k) = 1} = pk
where xp (k) ∈ <n is the plant’s state vector and up (k) ∈ <m
and yp (k) ∈ <p are the plant’s control input and output
vectors, respectively. A, B, and C are known as real matrices where pk assumes discrete values, see Table II. Two particular
with appropriate dimensions. classes can be considered:
TABLE I TABLE II
PARAMETERS OF THE M ICROGRID S YSTEM PATTERN OF pk

Parameter Value (p.u) Parameter Value (p.u) pk q1 q2 ··· qn−1 qn

H 3 Pe1 0.02 P rob(pk = q) r1 r2 ··· rn−1 rn
M 2.5 Pe3 0.15
ω0 377 Req 0.1 Class 1: pk has the probability mass function in Table II
KA 15 eq1 0.04 where qr − qr−1 = constant for r = 2, ..., n. This covers a
TA 0.03 eq2 0.01 wide range of cases including
xd 1.3 mp 100 1) If there is no information about the likelihood of different
xq 0.47 kp = kpf 1885 values, we use the uniform discrete distribution, ri = 1/n, i =
x0d 0.3 kpd1 0.08 1, 2, ..., n,
0
Tdo 7 kpd2 0.9
2) If it is suspected that pk follows a symmetric triangle
itd1 0.28 kpv 0.01
distribution, we use the following function:
itd2 0.08 dc 0.6
i) If n is even, ri = a + jd, j = 0, 1, ..., n/2 and
Ldc 0.05
ri = a + (n − j)d, j = 0, 1, ..., n/2 + 1, n/2 + 2, ..., n,
Parameter Value (p.u)
where na + dn(n − 1)/4 = 1,
Pe2 0.1
ii) If n is odd, ri = a + jd, j = 0, 1, ..., (n − 1)/2 and
Cdcp 1.1
ri = a + (n − j)d, j = 0, 1, ..., (n + 1)/2, (n + 2)/2, ..., n,
Cpf 0.2
where na + dn(n − 1)2 /4 = 1
Req1 0.2
Rd 0.1 3) If it is suspected that ri is a decreasing linear function, we
ipf q0 0.67
use ri = a − jd, j = 0, 1, ..., n where na − dn(n − 1)/2 = 1
ipf d0 0.13 4) If it is suspected that ri is an increasing linear func-
rt 0.1 tion, we use ri = a − (n − j)d, j = 0, 1, ..., n where
rb 0.15 na − dn(n − 1)/2 = 1
xt 0.2 TABLE III
xb 0.265 PATTERN OF sk

Remark 1: We note that the developed renewable energy sk s1 s2 ··· sn−1 sn

model is formulated in linearized continuous-time form as P rob(sk = t) t1 t2 ··· tn−1 tn
it emerges from physical considerations. For all practical
purposes, we seek a discrete-time system of the type of (56). Class 2: pk = X/n, n > 0 and 0 ≤ X ≤ n is a random
As will be mentioned in the simulation results section, the variable that follows the Binomial distribution B(q, n), q > 0,
microalternator-PV system containing A, B and C matrices are that is
discretized at a sampling time of 0.01 seconds. This is selected
after a series of computer simulation experiments based on
the closeness of the step-response of the linearized continuous P rob(pk = (ax + b)/n) =
µ ¶
model and its discretized counterpart. n
q x (1 − q)n−x , b > 0,
For a more general case, we assume that the measurement x
after passing through the network exhibits a randomly varying x = 0, 1, 2, ..., n, an + b < n
communication delay and is described by [26]
½
yp (k − τkm ), δ(k) = 1 Remark 2: It is significant to note that the case P rob{δ(k) =
yc (k) = (57)
yp (k), δ(k) = 0 1} = δ̄, where δ̄ is a constant value, is widely used in majority
where τkm stands for measurement delay, the occurrence of of results on NCS. In this paper, we focus on nonstationary
which satisfies the Bernoulli distribution, and δ(k) is Bernoulli dropouts.
distributed white sequence exhibiting the occurrence of mes- When the full state information is not available and the time
sage (packet) dropouts. In order to capture the current practice delay occurs on the actuation side, it is desirable to design the
MAHMOUD AND RAHMAN: NETWORKED CONTROL APPROACH FOR DISTRIBUTED GENERATION SYSTEMS 9

following observer-based controller [26]: 1, α(k) = 0), (δ(k) = 0, α(k) = 0), (δ(k) = 0, α(k) =
1)}:
· ¸ · ¸
Observer : A 0 A + BK −BK
A1 = , A2 = ,
0 A 0 A
x̂(k + 1) = Ax̂ + Bup (k) + L(yc (k) − ŷc (k)) · ¸
½ A + BK −BK
C x̂(k), δ(k) = 0 A3 = ,
ŷc (k) = (58) 0 A − LC
C x̂(k − τkm ), δ(k) = 1 · ¸
Controller : A 0
A4 = ,
0 A − LC
uc (k) = K x̂(k) · ¸ · ¸
½ BK −BK 0 0
uc (k), α(k) = 0 B1 = , B2 = ,
up = (59) 0 0 0 0
uc (k − τka ), α(k) = 1 · ¸ · ¸
0 0 BK −BK
B3 = , B4 = ,
0 0 0 0
where x̂(k) ∈ <n is the estimate of the system (5), ŷc (k) ∈ <p · ¸ · ¸
is the observer output, and L ∈ <n×p and K ∈ <m×n 0 0 0 0
C1 = , C2 = ,
are the observer and controller gains, respectively, and τka is 0 −LC 0 −LC
· ¸ · ¸
the actuation delay. The stochastic variable α(k), mutually 0 0 0 0
C3 = , C4 = (64)
independent of δ, is also a Bernoulli distributed white sequence 0 0 0 0
with Remark 3: It is remarked for simulation processing that we
can express (61)-(62) in the form
P rob{α(k) = 1} = sk
xp (k + 1)= sk [Axp (k) + BKxp (k − τkα )
− BKe(k − τkα )]
where sk assumes discrete values. By similarity, a particular
class is that sk has some probability mass function as in Table + (1 − sk )[(A + BK)xp (k)
III, where sr − sr−1 = constant for r = 2, . . . , n. − BKe(k)] (65)
In this paper, we assume that τka and τkm are time-varying e(k + 1) = pk [Ae(k) − LCe(k − τkm )]
and have the following bounded condition: + (1 − pk )[(A − LC)e(k)] (66)
where the values of the random variables pk , sk are generated
in the manner discussed earlier.
−
τm ≤ τkm ≤ τm
+
, τa− ≤ τka ≤ τa+ (60) Remark 4: It is important to note from (64) that
· ¸
A + BK −BK
Aj + Bj + Cj = , j = 1, .., 4 (67)
0 A − LC
Define the estimation error by e(k) = xp (k) − x̂(k). Then,
it yields The interpretation of this result is that Aj + Bj + Cj represents
the fundamental matrix of the delayed system (63), which must
 be independent of the mode of operation. This will help in

 Axp (k) + BKxp (k − τkα ) simplifying the control design algorithm.


 −BKe(k − τkα ), The aim of this paper is to design an observer based
xp (k + 1) = α(k) = 1, (61) feedback stabilizing controller in the form of (58) and (59)



 (A + BK)xp (k) − BKe(k), such that the closed loop system (63) is exponentially stable

α(k) = 0, in the mean square sense. Our approach is based on the
e(k + 1) = xp (k + 1) − x̂(k + 1) concepts of switched time-delay systems [31]. For simplicity


 Ae(k) − LCe(k − τkm ), in exposition, we introduce

δ(k) = 1,
= (62) σ1 (k) = P rob{δ(k) = 1, α(k) = 1}, σ̂1 = IE[σ1 ]

 (A − LC)e(k),
 σ2 (k) = P rob{δ(k) = 1, α(k) = 0}, σ̂2 = IE[σ2 ]
δ(k) = 0
σ3 (k) = P rob{δ(k) = 0, α(k) = 0}, σ̂3 = IE[σ3 ]
σ4 (k) = P rob{δ(k) = 0, α(k) = 1},
In terms of ξ(k) = [xTp (k) eT (k)]T , system (61) and (62)
can be cast into the form: σ̂4 = IE[σ4 ] (68)
where IE[σi ] is the expected value of σi , i = 1, .., 4. Since we
ξ(k + 1) = Aj ξ(k) + Bj ξ(k − τkm ) assume that δ(k) and α(k) are independent random variables,
then it follows from (68) that
+ Cj ξ(k − τka ) (63)
σ̂1 = IE[pk ]IE[sk ], σ̂2 = IE[pk ]IE[1 − sk ]
σ̂3 = IE[1 − pk ]IE[1 − sk ],
where {Aj , Bj , Cj , j = 1, ..., 4} and j is an index identifying
one of the following pairs {(δ(k) = 1, α(k) = 1), (δ(k) = σ̂4 = IE[1 − pk ]IE[sk ] (69)
10 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

V. D ESIGN R ESULTS where

In this section, we will thoroughly investigate the stability
+ −
and controller synthesis problems for the closed-loop system Ψj = −P + σ̂j (τm − τm + τa+ − τa− + 2)Qj
(63). First, let us deal with the stability analysis problem + R1 + R1T + R2 + R2T
and derive a sufficient condition under which the closed- Φj1 = (Aj + Bj + Cj )T σ̂j P (Aj + Bj + Cj )
loop system (63) with the given controller (58) and (59) is
exponentially stable in the mean square sense. Extending on Φj2 = (Aj + Bj + Cj )T σ̂j P Bj
the basis of [32], the following Lyapunov function candidate Φj3 = (Aj + Bj + Cj )T σ̂j P Cj , Φj5 = BTj P Cj
is constructed to establish the main theorem: Φj4 = BTj σ̂j P Bj , Φj6 = CTj σ̂j P Cj
X5
V (ξ(k)) = Vi (ξ(k)) (70) Proof 1: Defining y(k) = x(k + 1) − x(k), one has
i=1
X4
k−1
X
V1 (ξ(k)) = ξ T (k)P ξ(k), P > 0
ξ(k − τkm ) = ξ(k) − y(i) (75)
j=1
i=k−τkm
4
X k−1
X k−1
X
V2 (ξ(k)) = ξ T (i)Qj ξ(i), Qj = QTj > 0
ξ(k − τka ) = ξ(k) − y(i) (76)
j=1 i=k−τkm
i=k−τka

4
X k−1
X
V3 (ξ(k)) = ξ T (i)Qj ξ(i) (71) Then the system (12) can be transformed into
j=1 i=k−τka
−
−τm +1
ξ(k + 1) = (Aj + Bj + Cj )ξ(k) − Bj λ(k) − Cj γ(k) (77)
4
X X k−1
X
V4 (ξ(k)) = ξ T (i)Qj ξ(i)
j=1 + i=k+`−1
where
`=−τm +2
−τa− +1 k−1 k−1
4
X X k−1
X X X
V5 (ξ(k)) = ξ T (i)Qj ξ(i) λ(k) = y(i), γ(k) = y(i).
j=1 `=−τa+ +2 i=k+`−1 i=k−τkm i=k−τka

It is not difficult to show that there exist real scalars µ > 0

and υ > 0 such that Evaluating the difference of V1 (ξ(k)) along the solution of
system (77), we have
µkξk2 ≤ V (ξ(k)) ≤ υkξ(k)k2 (72)
Remark 5: By carefully considering Remark 4 in view of IE[∆V1 (ξ(k))] = IE[V1 (ξ(k + 1))] − V1 (ξ(k)) =
X4 ·
model (12), it is justified to select matrix P to be the same for
all operational modes, hence independent of j, while keeping ξ T (k)[Φj1 − P ]ξ(k) − 2ξ T (k)Φj2 λ(k) −
j=1
matrix Qj dependent on mode j.
We now present the analysis result for system (12) to be 2ξ T (k)Φj3 γ(k) + λT (k)Φj4 λ(k) +
¸
exponentially stable.
2λT (k)Φj5 γ(k) + γ T (k)Φj6 γ(k) (78)
Theorem 1: Let the controller and observer gain matrices K
and L be given. The closed-loop system (12) is exponentially
stable if there exist matrices 0 < P, 0 < QTj = Qj , j = A straightforward computation gives
1, .., 4 and matrices Ri , Si , and Mi , i = 1, 2, such that the
following matrix inequality holds [26] 4
X · k
X
· ¸ IE[∆V2 (ξ(k)] = σ̂j ξ T (i)Qj ξ(i) −
Λ1j Λ2j
Λj = < 0 (73) j=1 m
i=k+1−τk+1
• Λ3j
k−1
X
Λ1j = ξ T (i)Qj ξ(i) =
  i=k−τkm
Ψj + Φj1 −R1 + S1T −R2 + S2T ·
4
X k
X
 • −S1 − S1T − σ̂j Qj 0 
σ̂j ξ T (i)Qj ξ(i) −
• • −S2 − S2T − σ̂j Qj
  j=1 m
i=k+1−τk+1
−R1 + M1T − Φj2 −R2 + M2T − Φj3 k−1
Λ2j =  −S1 − M1T 0  X
T ξ(k − τkm )Qj ξ(k − τkm ) + ξ T (i)Qj ξ(i) −
0 −S2 − M2
· T
¸ m
i=k+1−τk+1
−M1 − M1 + Φj4 Φj5 k−1 ¸
Λ3j = X
• −M2 − M2T + Φj6 ξ(i)Qj ξ(i) (79)
(74) i=k+1−τkm
MAHMOUD AND RAHMAN: NETWORKED CONTROL APPROACH FOR DISTRIBUTED GENERATION SYSTEMS 11

In view of Therefore, for any appropriately dimensioned matrices Ri , Si

k−1 k−τkm and Mi , i = 1, 2, we have the following equations:
X X
ξ T (i)Qj ξ(i) = ξ T (i)Qj ξ(i) + 2[ξ T (k)R1 + ξ T (k − τkm )S1 + λT (k)M1 ] ×
m
i=k+1−τk+1 m
i=k+1−τk+1
[ξ(k) − ξ(k − τkm ) − τ (k)] = 0 (87)
k−1
X 2[ξ T (k)R2 + ξ T (k − τka )S2 + γ T (k)M2 ]
ξ T (i)Qj ξ(i) ≤
i=k+1−τkm [ξ(k) − ξ(k − τka ) − γ(k)] = 0 (88)
k−1
X On combining (78)−(88), we reach
ξ T (i)Qj ξ(i) + 4 ·
X
i=k+1−τkm
IE[∆V (ξ(k))] ≤ ξ T (k)Ψj ξ(k) +
−
k−τm
X j=1
ξ T (i)Qj ξ(i) (80) 4
X
+
i=k+1−τm ξ T (k)(−2R1 + 2S1T )ξ(k − τkm ) +
j=1
We readily obtain
ξ T (k)(−2R2 + 2S2T )ξ(k − τka ) +
4
X ·
IE[∆V2 (ξ(k))] ≤ σ̂j ξ T (k)Qj ξ(k) − ξ T (k)(−2R1 + 2M1T − 2Φj2 )λ(k) +
j=1 ξ T (k − τkm )(−S1 − S1T − σ̂j Qj )ξ(k − τkm ) +
ξ T (k − τkm )Qj ξ(k − τkm ) + ξ T (k − τkm )(−2S1 − 2M1T )λ(k) +
k−τm
X
−
¸ ξ T (k − τka )(−S2 − S2T − σ̂j Qj )ξ(k − τka ) +
ξ T (i)Qj ξ(i) (81) ξ T (k − τka )(−2S2 − 2M2T )γ(k) +
+
i=k+1−τm
λT (k)(−M1 − M1T + Φj4 )λ(k) +
Following parallel procedure, we get γ T (k)(−M2 − M2T + Φj5 )γ(k) +
X4 · λT (k)Φj6 γ(k) + +
IE[∆V3 (ξ(k))] ≤ σ̂j ξ T (k)Qj ξ(k) − ¸
j=1 ξ T (k)(−2R2 + 2M2T − 2Φj3 )γ(k) =
T
ξ (k − τk )Qj ξ(k − τka )
a
+ 4 · ¸
X
k−τa−
X ¸ T e
ζ (k)Λj ζ(k) (89)
T
ξ (i)Qj ξ(i) (82) j=1
i=k+1−τa+ where
£ ¤T
Finally, ζ(k) = ζ1T (k) ζ2T (k)
£ ¤T
IE[∆V4 (ξ(k))] = ζ1 (k) = ξ T (k) ξ T (k − τkm ) ξ T (k − τka )
4 · −τX
−
m +1
£ T ¤T
X ζ2 (k) = λ (k) γ T (k)
σ̂j [ξ T (k)Qj ξ(k) −
j=1 `=−τ + m +2
e j corresponds to Λj in (74) by Schur complements. If
and Λ
¸
Λj < 0, j = 1, .., 4 holds, then
ξ T (k + ` − 1)Qj ξ(k + ` − 1)] = 4 · ¸
X
4 · IE[V (ξ(k + 1)) − V (ξ(k))] = T e
ζ (k)Λj ζ(k)
X
σ̂j (τ + m − τ − m )ξ T (k)Qj ξ(k) − j=1
j=1 4 ·
X ¸
k−τ − ¸ ≤ e min (Λ
−Λ e j )ζ T (k)ζ(k)
Xm
T j=1
ξ (i)Qj ξ(i) (83)
4 ·
X ¸
i=k+1−τ + m
IE[∆V5 (ξ(k))] = <− βj ζ T (k)ζ(k) (90)
X4 · j=1

σ̂j (τ + a − τ − a )ξ T (k)Qj ξ(k) − where

j=1
£ ¤
0 < βj < min λmin (Λj ), max{λmax (P ), λmax (Qj )}
k−τ
Xa
− ¸
ξ T (i)Qj ξ(i) (84) Inequality (90) implies that IE[V (ξ(k + 1)) − V (ξ(k))] <
i=k+1−τ + a
−φV (ξ(k)), 0 < φ < 1. In the manner of [33], we get
υ λ
It follows from (75) and (76) that: ||ξ(k)||2 ≤ ||ξ(0)||2 (1 − φ)k +
κ µφ
ξ(k) − ξ(k − τkm ) − λ(k) = 0 (85) Therefore, it can be verified that the closed-loop system (12)
ξ(k) − ξ(k − τka ) − γ(k) = 0 (86) is exponentially stable. This completes the proof.
12 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

A solution to the problem of the observer-based stabilizing b = P −1 , invoking Schur complements, we write
Setting X
controller design is provided by the following theorem: matrix Λj in (96) equivalently as
+ −
Theorem 2: Let the delay bounds τm , τm , τa+ , τa− be
given. Evaluate the quantities σ̂j , j = 1, .., 4. Then, the  
e 1j
Λ e 2j
Λ Ωj
closed-loop system (12) is exponentially stable if there exist  
 • e 3j
Λ 0  < 0 (98)
matrices 0 < X1 , X2 , Y1 , Y2 , Z1 , 0 < Ξj , j = 1, .., 4
• • −Xb
and matrices Πi , Υi and Γi , i = 1, 2, such that the following
matrix inequality holds for j = 1, .., 4:
  Applying the congruence transformation
b 1j Λ
Λ b 2j bj
Ω
 • Λ3j 0  < 0 (91)
• • −σ̂j X b b X,
Tj = diag[X, b X,
b X,
b X,
b I]
· ¸
b= X1 X2 to matrix inequality in (98) and manipulating using (92) and
X (92)
X2T X2
b j = −X̂ + σ̂j (τm
Ψ +
− τm−
+ τa+ − τa− + 2)Ξj + b j X,
b Πj = XR
b j X,
b Υj = XS
b j X,
b
Ξj = XQ
Π1 + ΠT1 + Π2 + ΠT2 b j X,
b Z 1 = X X2 .
−1
  Γj = XM 1
−Π1 + ΓT1 −Π2 + ΓT2
b 2j =  −Υ1 − ΓT1
Λ 0 
0 −Υ2 − ΓT2 we readily obtain matrix inequality (91) subject to (94).
· ¸ Remark 6: Expanding the microgrid to accomodate addi-
b −Γ1 − ΓT1 0
Λ3j = tional number of distributed generation units will eventually
• −Γ2 − ΓT2
h i lead to increase the static computational burden as well.
bj = Ω
Ω b 1j 0 0 −Ω b 4j −Ω b 5j However using linear matrix inequalities as efficient numerical
· T T ¸ tool makes the job of determining the feedback gains becomes
b Y1 B − Y1T B T Z1 0
Ω4j = , j = 1, 4 moderate. Moreover, it is executed offline and performed only
0 0
· T
¸ once.
b 5j = 0 −Y2T , j = 1, 2
Ω Remark 7: The selection of X b as given by (92) has
0 −Y2
the advantage of converting the solution of bilinear matrix
b
Ω4j = 0, j = 2, 3, Ω b 5j = 0, j = 3, 4 (93) inequalities to that of seeking the feasibility of linear matrix
  inequalities and hence avoiding iterative procedures. It should
bj
Ψ −Π1 + ΥT1 −Π2 + ΥT2 be noted that the LMI (91) depends on the average dropout
b 1j =  • −Υ1 − ΥT − σ̂j Ξj
Λ 0 
1 patterns identified by (69), which is quite useful in illustrating
T
• • −Υ2 − Υ2 − σ̂j Ξj different operating conditions of the communications network.
· T T T T T T T
¸
b 1j = X1 A + Y1 B T− Y1 B Z1 X2 AT − Y2T
Ω Remark 8: It is remarked that the implementation of The-
X2 A X2 A − Y2 orem 2 is on-line in nature as it requires calling random
∀j (94) generators to pick-up numbers corresponding to the scalars
σ̂1 , ..., σ̂4 and to evaluate the probabilities in model (65) and
where the gain matrices are given by
(66) to compute the state and error trajectories. This represents
K = Y1 X1−1 , L = Y2 X2−1 C † (95) a salient feature not shared by other methods for networked
control design under unreliable communication links.
Proof 2: Define
£ ¤T Remark 9: In this paper, the X b matrix considered in
Ωj = (Aj + Bj + Cj ) 0 −Bj −Cj Theorem 2 is distinct from the one in [26]. The X b matrix
then matrix inequality (73) can be expressed as incorporates two different elements X1 and X2 unlike the
Xb matrix in [26] which contains the same element X. This
Λj e + Ωj P ΩTj < 0
= Λ (96) extension results in a distinct LMI term Ω cj given in (41).
" #
e 1j Λ
Λ e 2j Further, the expression for calculating the controller and
ej
Λ = < 0 observer gains (K and L) is also different from [26] This is
• e 3j
Λ
  done to present a more realistic approach in determining the
Ψj −R1 + S1T −R2 + S2T controller and observer gains.
e 1j
Λ =  • −S1 − S1 − QjT
0 
Remark 10: The networked control system model consid-
• • −S2 − S2T − Qj ered in this paper is distinct from the one presented in [26].
 
−R1 + M1T −R2 + M2T The model represents a typical System of systems networked
e 2j
Λ =  −S1 − M1T 0  control model consisting of a sensor suite and an ad-hoc
T
0 −S2 − M2 network for multiple subsystems. Moreover the illustrated
· ¸
−M1 − M1 T
0 example for the proposed control strategy is exclusively a
e 3j
Λ = (97)
• −M2 − M2T microgrid system of systems.
MAHMOUD AND RAHMAN: NETWORKED CONTROL APPROACH FOR DISTRIBUTED GENERATION SYSTEMS 13

VI. S IMULATION R ESULTS have different parameter values.

In this section, the microgrid system which is modeled as a
networked control system of systems is simulated in the SOS
research lab at KFUPM using pilot-scale lab equipment to
build an experimental microgrid including up to 12 DG units.
The simulation tools are Matlab-Simulink environment, par-
ticularly the products Simscape Power Systems Examples and
RAPSim an open source simulation software for microgrids.
A single-line diagram of the microgrid under consideration is
depicted in Fig. 6.
The output feedback controller design is implemented which
stabilizes the system in presence of packet dropouts and
delays. The parameter values given in Table I are substituted in
the state matrices A and B presented earlier in this paper. Both
the sets of PV-microalternator system are similar in nature but
The controller and observer gains can be obtained by using
the relation K = Y1 X1−1 and L = Y2 X2−1 C † , X1 , X2 , Y1 , Y2
can be found by solving the LMI (91) in Theorem 2. The
values of ᾱ and δ̄ are set at 0.1 each. The measurement delay
is varied from 1 to 5 and the actuation delay is varied from
3 to 9. The system is discretized at a sampling time of 0.01
seconds. This is selected after a series of computer simulation
experiments based on the closeness of the step-response of the
linearized continuous model and its discretized counterpart.
The NCS is modeled in such a way that delays are accom-
modated in both the measurement as well as the actuation
channel. The delays are generated by employing random
number generators. The number obtained from the Uniform
Random Number Generator is compared to a variable proba-
bility pk explained earlier in the paper. If the random generated

———————————————————————————————————————————————————–
£ ¤
K= K1 K2
· ¸
−0.0003 −0.0002 0.0001 −0.0002 0.0000 0.0001
K1 =
−0.0001 −0.0016 −0.0002 0.0004 −0.0003 −0.0001
· ¸
−0.0112 0.0001 0.0000 −0.0006 −0.0003 0.0000
K2 =
−0.0003 −0.0001 −0.0004 −0.0001 −0.0002 −0.0067
 
−0.0076 −0.0172
 −0.0172 −0.0013 
 
 0.0041 0.0005 
 
 −0.0244 −0.0002 
 
 0.0121 −0.0042 
 
 0.0061 0.0004 
L=  

 −0.0069 −0.0170 
 −0.0489 0.0010 
 
 −0.0039 −0.0009 
 
 −0.0002 0.0005 
 
 0.0156 0.0008 
−0.0015 0.0221

£ ¤
K= K1 K2
· ¸
−0.0007 −0.0100 0.0028 0.0004 0.0003 −0.0419
K1 =
−0.0005 −0.0003 0.0000 0.0020 −0.0002 −0.0090
· ¸
0.0000 −0.0001 0.0016 0.0021 0.0023 0.0000
K2 =
0.0000 0.0000 −0.0012 0.0011 0.0000 −0.0047
 
−0.0152 −0.0036
 −0.0030 −0.0003 
 
 0.0016 −0.0025 
 
 0.0000 0.0005 
 
 −0.0116 0.0010 
 
 0.0008 0.0002 
L=  −0.0003 −0.0001 

 
 0.0011 −0.0126 
 
 0.0000 0.0003 
 
 0.0000 −0.0124 
 
 0.0166 −0.0008 
−0.0003 0.0003
14 IEEE/CAA JOURNAL OF AUTOMATICA SINICA

number is less than pk , then the output of the comparator is

high meaning a delay has occurred. Otherwise, if the output of
the comparator is low, there is no delay. This is accomplished
by using the variable functional delay block in Simulink.
The gain matrices obtained for both sets are shown above.
The figures shown above represent the state response of both
sets of microalternator-PV system named as set 1 and set 2.
Each state represents a typical dynamics of the system.

Fig. 9. Internal voltage of microalternator along q-axis.

Fig. 6. Single-line diagram of the microgrid under consideration.

Fig. 10. Field voltage of microalternator along d-axis.

Fig. 7. Rotor angle of the microalternator. Fig. 11. Photovoltaic cell current.

Fig. 8. Rotor speed of the microalternator. Fig. 12. Voltage across DC-link capacitor.
MAHMOUD AND RAHMAN: NETWORKED CONTROL APPROACH FOR DISTRIBUTED GENERATION SYSTEMS
Fig. 13. Inverter output current along d-axis.
Fig. 14. Inverter output current along q-axis.
Fig. 15. Coupling line current of the filter along d-axis.
Fig. 16. Coupling line current of the filter along q-axis.
15
Fig. 17. Capacitor voltage of the filter along d-axis.
Fig. 18. Capacitor voltage of the filter along q-axis.
The simulation results elucidate the effectiveness of the pro-
posed control methodology. Two sets of the microalternator-
PV system are considered which have distinct parameter
values. As evident from the graphs, the output feedback
controller stabilizes the system in presence of delays and
packet dropouts. It can be observed from the state-wise re-
sponse that each state of both the sets possessing different
values is stabilized by the controller an a reasonably less
time with minimum overshoot and very less oscillations.
Thus, the simulation results obtained significantly exhibit the
controller responses for different parameter values. Moreover,
stabilization in presence of communication issues such as
packet dropouts and delays further explain the effectiveness
of the controller. It is also to be noted that the state response
of both sets of microalternator-PV system is presented in one
graph to demonstrate the ability of the controller to stabilize
the system irrespective of the parameter values.
VII. C ONCLUSIONS
A networked control system of systems methodology for the
stabilization of a microgrid system with two sets of microal-
ternator and photovoltaic systems is presented in this paper.
The microalternator and photovoltaic system are modeled in
detail. The combination of both systems connected to the main
grid and a load is also modeled. The networked control system
based on this model is subjected to delays and non stationary
packet dropouts. An improved output feedback controller is
proposed, which stabilizes the system in presence of the
aforementioned communication constraints. Using pilot-scale
16
lab equipment to build an experimental microgrid, simulation
of both sets of microalternator-pv systems are attained using
Matlab-Simulink environment, particularly the products Sim-
scape Power Systems Examples. From the presented results,
the effectiveness of the controller is demonstrated in solving
a major issue in microgrids arising from stability of its
constituent systems.
Future work would go along different avenues. One avenue
would be to expand the model by including battery storage
units and examine the transition between islanded and grid-
connected modes of operation. Another avenue is to cast the
model into a stochastic set-up by including effects of load
daily as well as solar radiation variations. A final avenue
is to treat the model from a multi-agent perspective and
investigate various protocols. Research work along these ideas
is underway.
ACKNOWLEDGMENTS
This work is supported by the deanship for scientific re-
search (DSR) at KFUPM through distinguished professorship
research project IN-141003.
R EFERENCES
[1] R. H. Lasseter, “Microgrids and distributed generation”, Journal of
Energy Engineering, American Society of Civil Engineers, Vol. 133, No.
3, 2007, pp. 144−149.
[2] A. Arulampalam, M. Barnes, A. Engler, A. Goodwin and N. Jenk-
ins, “Control of power electronic interfaces in distributed generation
microgrids”, International Journal of Electronics, Vol. 91, 2004, pp.
503−523.
[3] N. Hatziargyriou, H. Asano, R. Iravani and C. Marnay, “Microgrids: An
overview of ongoing research, development and demonstration projects”,
IEEE Power and Energy Magazine, Vol. 5, 2007, pp. 78−94.
[4] J. A. P. Lopes, C. L. Moreira and A. G. Madureira, “Defining control
strategies for analyzing microgrids islanded operation”, Proc. IEEE
Russia Power Tech, 2005, pp. 1−7.
[5] M. H. Ashourian , M. A. A. Zin , A. S. Mokhtar, S. J. Mirazimi and
Z. Muda, “Controlling and modeling power-electronic interface DERs
in islanding mode operation microgrid”, Proc. IEEE Symposium on
Industrial Electronics and Applications (ISIEA), 2011, pp. 161−166.
[6] X. Zhang, J. Liu, T. Liu and L. Zhou, “A novel power distribution
strategy for parallel inverters in islanded mode microgrid”, Proc. The
Twenty-Fifth Annual IEEE Applied Power Electronics Conference and
Exposition (APEC), 2010, pp. 2116−2120.
[7] P. K. Ray, S. R. Mohanty and N. Kishor, “Dynamic modeling and
control of renewable energy based hybrid system for large band wind
speed variation”, Proc. IEEE PES Innovative Smart Grid Technologies
Conference Europe (ISGT Europe), 2010, pp. 1−10.
[8] N. S. Jayalakshmi and D. N. Gaonkar, “Performance study of isolated
hybrid power system with multiple generation and energy storage units”,
Proc. International Conference on Power and Energy Systems (ICPS),,
2011, pp. 1−5.
[9] F. Katiraei and M. R. Iravani, “Power management strategies for a
microgrid with multiple distributed generation units”, IEEE Trans.
Power Systems,, Vol. 21, 2006, pp. 1821−1831.
IEEE/CAA JOURNAL OF AUTOMATICA SINICA
[10] D. J. Lee and Li Wang, “Small-signal stability analysis of an autonomous
hybrid renewable energy power generation/energy storage system part
I: time-domain simulations”, IEEE Trans. Energy Conversion, Vol. 23,
No. 1, 2008, pp. 311−320.
[11] A. M. O. Haruni , A. Gargoom, M. E. Haque and M. Negnevitsky,
“Dynamic operation and control of a hybrid wind-diesel stand alone
power system”, Proc. The Twenty-Fifth Annual IEEE Applied Power
Electronics Conference and Exposition (APEC), 2010, pp. 162−169.
[12] G. K. Kasal and B. Singh, “Voltage and frequency controllers for an
asynchronous generator-based isolated wind energy conversion system”,
IEEE Trans. Energy Conversion,, Vol. 26, No. 2, 2011, pp. 402−416.
[13] M. Babazadeh and H. Karimi, “Robust decentralized control for islanded
operation of a microgrid”, Proc. IEEE Power and Energy Society
General Meeting, 2011, pp. 1−8.
[14] R. Moradi, H. Karimi and M. Karimi-Ghartemani, “Robust decentralized
control for islanded operation of two radially connected DG systems”,
Proc. IEEE International. Symposium on Industrial Electronics (ISIE),
2010, pp. 2272−2277.
[15] S. M. A. Shabestary, M. Saeedmanesh, A. Rahimi-Kian and E. Jalal-
abadi, “Real-time frequency and voltage control of an islanded mode
microgrid”, Proc. The 2nd Iranian Conference on Smart Grids (ICSG),,
2012, pp. 1−6.
[16] M. Henshaw, C. Siemieniuch, M. Sinclair, S. Henson, V. Barot, M.
O. Jamshidi, D. Delaurentis, C. Ncube, S. L . Lim, and H. Dogan,
“Systems of systems engineering: A research imperative”, Proc. IEEE
International Conference on System Science and Engineering, Hungary,
2013, pp. 389−394.
[17] M. O. Jamshidi, System of systems Engineering: Principles and Appli-
cations, Taylor Francis CRC Publishers, Boca Raton, USA, 2008.
[18] M. S. Mahmoud, “Networked control systems analysis and design: an
overview”, The Arabian Journal of Science and Engineering: Invited
Paper, vol. 41, no. 3, pp. 711−758, 2016.
[19] Y. Q. Xia, Y. L. Gao, L. P. Yan and M. Y. Fu, “Recent progress
in networked control systems–A survey,” Int. J. Automation and
Computing, vol. 12, no. 4, pp. 343−367, 2015.
[20] L. X. Zhang, H. J. Gao and O. Kaynak, “Network–induced constraints
in networked control systems–A survey,” IEEE Trans. Industrial Infor-
matics, vol. 9, no. 1, pp. 403–416, 2013.
[21] E. Garcia and P. J. Antsaklis, “Model-based event-triggered control for
systems with quantization and time-varying network delays,” IEEE
Trans. Automatic Control, vol. 58, no. 2, pp. 422–434, 2013.
[22] D. E. Quevedo, D. Nesic, “Robust stability of packetized predictive
control of nonlinear systems with disturbances and Markovian packet
losses,” Automatica, vol. 48, no. 8, pp. 1803–1811, 2012.
[23] H. R. Karimi, “Robust H∞ filter design for uncertain linear systems
over network with network-induced delays and output quantization,”
Modeling, Identification and Control, vol. 30, no. 1, pp. 27−37, 2009.
[24] M. O. Jamshidi, “Control of system of systems”, Proc. 7th IEEE
International Conference on Industrial Informatics, 2009, pp. 1−16.
[25] W. Zhang, M. S. Branicky and S. M. Phillips, “Stability of networked
control systems”, IEEE Control Systems Magazine, Vol. 21, No. 1,
2001, pp. 84−99.
MAHMOUD AND RAHMAN: NETWORKED CONTROL APPROACH FOR DISTRIBUTED GENERATION SYSTEMS
[26] M. S. Mahmoud, S. Z. Selim, P. Shi and M. H. Baig “New results on
networked control systems with non-stationary packet dropouts”, IET
Control Theory Appl., Vol. 6, Issue 15, 2012, pp 2442−2452.
[27] Y. N. Yu Electric power system dynamics, Academic Press, New York,
1983.
[28] V. Quaschning and R. Hanitsch, “Numerical simulation of current-
voltage characteristics of photovoltaic systems with shaded solar cells”,
Solar Energy, Vol. 56, 1996, pp 513−520.
[29] M. T. Hussain, “Modeling and control of a microgrid including pho-
tovoltaic and wind generation”, M.Sc Thesis, King Fahd University of
Petroleum and Minerals, Saudi Arabia, 2012.
[30] N. Mahamad, C. M. Hadzer and S. Masri, “Application of LC filter in
harmonics reduction”, Proc.National Power and Energy Conference,
Malaysia, 2004.
[31] M. S. Mahmoud, ” Switched Time-Delay Systems, Springer-Verlag, New
York, 2010.
[32] M. S. Mahmoud and Y. Xia, “Robust stability and stabilization of a class
of nonlinear switched discrete-time systems with time-varying delays”,
J. Optimization Theory and Applications, Vol. 143, 2009, pp. 329−355.
[33] F. W. Yang, Z. D. Wang, Y. S. Hung, and M. Gani, “H∞ control for
networked systems with random communication delays”, IEEE Trans.
Autom. Control, Vol. 51, No. 3, 2006, pp. 511−518.
17
Magdi S. Mahmoud MagdiSadek Mahmoud ob-
tained B.Sc.(Honors) in Communication Engineer-
ing, M.Sc. in Electronic Engineering and Ph.D. in
Systems Engineering, all from Cairo University in
1968, 1972 and 1974, respectively. He has been
Professor of Engineering since 1984. He is now a
Distinguished Professor at KFUPM, Saudi Arabia.
He was on the faculty at different universities world-
wide including Egypt (CU, AUC), Kuwait (KU),
UAE (UAEU), UK (UMIST), USA (Pitt, Case West-
ern), Singapore (Nanyang) and Australia (Adelaide).
He lectured in Venezuela (Caracas), Germany (Hanover), UK ((Kent), USA
(UoSA), Canada (Montreal) and China (BIT, Yanshan). He is the principal
author of thirty-seven (37) books, inclusive book-chapters and the author/co-
author of more than 550 peer-reviewed papers. He is the recipient of two
national, one regional and several university prizes for outstanding research
in engineering and applied mathematics. He is a fellow of the IEE, a senior
member of the IEEE, the CEI (UK), and a registered consultant engineer
of information engineering and systems (Egypt). He is currently actively
engaged in teaching and research in the development of modern methodologies
to distributed control and filtering, networked-control systems, triggering
mechanisms in dynamical systems, fault-tolerant systems and information
technology.
Mohammed Saif Ur Rahman Mohammed Saif Ur
Rahman was born in Hyderabad, India in 1989. He
received his M.Sc. degree in Control and Instru-
mentation from King Fahd University of Petroleum
and Minerals, Saudi Arabia in 2015 and B.E. in In-
strumentation and Control from Osmania University,
India in 2011, where he was awarded a silver medal
for standing second in the university. He is currently
working as a Visiting Graduate Researcher with the
department of Electrical Engineering, The Petroleum
Institute, Abu Dhabi, United Arab Emirates. His
research interests include Control and Automation, Microgrids, Networked
Control and application of Control theory to Power Electronics.