> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 1

( A , C) should have full column rank for the system to be

observable and more information one can read in [12].

A. Iterative observer method for real-time harmonic

estimation in power distributions grids

A novel iterative observer approach using real-time

simulation, the harmonic of electrical current, is estimated in
the power distribution network and determined in [13].
Figure.4. Block diagram of the non-linear burdens inside power distribution
Because the harmonics (distorted sine waves) are produced by
system, (modified from [13]).
the increased number of the non-linear load.
This research employs two iterative observers. On the one The matrices A , B are constant within the distribution
hand, the continuous-time observer is responsible of
system representation, and C j is also a matrix with a size of
estimating the branch current and node voltage on the power
distribution network. On the other hand, the discrete-time j ∈ [ 1 , m ], in which m is a set of measurements with a delay
observer is designed and used only if the state variables profile τ j ∈[0 ≤ τ 1 ≤ τ 2 ≤ ⋯ ≤ τ m ] . The corresponding values
change from analog to digital signals through a software for the discrete-time observer are similar for j ∈ [ 1 , m ].
converter in LabVIEW/MatLab. The latter observer is relying
on Routh-Hurwitz criterion to design this observer and
estimating the branch current measurement with or without B. Luenberger observers for mixed load conditions
harmonic disturbances in the proposed distribution system.
Figure 4 presents the harmonics of electrical current along The design and implementation of the Luenberger observers
with its electrical burden in parallel mode. Harmonic means within a microgrid control model in [14] under three scenarios
sinusoidal current having frequencies integer multiple of the for improving the power quality of the distributed energy
fundamental frequency (e.g., in USA the fundamental resources (DERs).
frequency is 60Hz). Second and third harmonic involve The power quality will decrease if there are increased non-
frequencies of 120 and 180 Hz, respectively on the linear burdens. The observers shall estimate the electrical
distribution system. It is worth noting that the distribution currents for each phase of the DER since the currents are
system could be severely harmed by these harmonics, since considered the state variables of the state-space equation of the
they would overheat the conductors and produce a significant microgrid system model.
overshoot in current and voltage in a power grid. Figure 5 shows that estimated electrical currents by the
The iterative performance of the continuous and discrete observer are due to virtual impedance through one close loop
observers, which comply with the below state-space equation, system linked by a sine pulse width modulation (SPWM).
yields the estimated harmonics: This SPWM will operate for switching sequence of the
diodes, in the sense that three phase currents are subjected to
ẋ (t)= Ax (t)+ Bu(t)+ Ew(t ) the observability analysis. Because negative sequence currents
m ( β n) are given by the integration of imbalanced burdens to the
y (t)=∑ C j x (t−τ j) microgrid system and harmonics of electrical current is
j=1 created by nonlinear loads.
( 5) As well as positive sequence current (α p) is originated by
one-directional power flow simulation and then these
x ( k +1 )= Ax ( k )+ Bu ( k ) + Ew(k ) parameters are subject to the αβ coordinate system:
m
y (k )=∑ C i x (k−τ i)
i=1
(6 )
where (t) is time and (k ) is the discrete data or integer
values, x (t) are the state variables of the system, including all
branch current and node voltages in the electrical distribution
system. The u ( t ) is the system input and y (t ) is a set of
measurements based on the mix between branch current and
node voltage. The w (t) indicates the harmonic disturbance
along with matrix E that is unknown during harmonic
estimation.
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 2

Figure.5. Control system configuration based on the block diagram Transformer (CT/PT). In the fault-events, for instance, when
(modified from [14]).
the CT/PT is saturated due to a high core magnetization

[ ]
current in the primary side of the CT/PT, the secondary
m
current side would receive false information to operate the
∑ ( ihα +ihα )
[]
microgrid protection system. That is why sliding-mode
iα h=1
p n

i αβ = = m
observer were designed to estimate the values for isolation,
iβ detection, correction, and estimation of failure occurring in
∑ ( ihβ +ihβ )
p n sensors mounted on a microgrid.
h=1
Figure 6 depicts the Simulink diagram model of a microgrid
(7 ) that covers DC voltage reference, a three-phase inverter along
with its space vector pulse width modulation (SVPWM), two
where the first and last harmonics are denoted by the CT protection devices, and an external AC referential grid.
subscripts h and m , respectively. Since their state variables Moreover, a non-linear differential equation of the modeled
are involving positive ( p) and negative (n ) sequences, microgrid can be represented by deterministic linear equation
therefore will be two state-space equations of the system. In such that

[]
spite of having not input control matrix, it does not affect the
1

[][ ][ ]
observability analysis. Thus, the state-space descriptions for
this model are given by:
i̇i a11 0 a13 i i Lf
˙i = 0 a22 a23 ∙ i + 0 v i +…
o o
ẋ α =A x a a31 a32 a33 v o 1
v˙o
y α =i α =C x α Lf
( 8)

[ ]
0
…+ −1/ Lc v g
ẋ β= A x β R d / Lc
y β=i β=C x β
( 11)
( 9) were,
A requirement of observability analysis is that the rank of
the observability matrix shall be the account of the harmonic −r f −1 1 rf R d
a 11= ; a 13 = ; a 31= − ;
by positive and negative sequence (i.e., 2m). As can be Lf Lf Cf Lf
represented in the next equation:

rank [ C CA ⋯ C A
(10 )
2 m−1 T
] =2m a 32=−
( 1 rc R d
Lf
−
Lc )
; a 33 =−
Rd Rd
+
Lf Lc
;
( )
where, 2 m is the number of state variables that describe the Can be (11) adjusted by an observer equation, hence it
yields:
size of square matrix A and is given by (4). Then, the system
is observable and can design two observers (see above figure).
They attempt to estimate the electrical current and voltage in x˙c = Ac x c + Bc u c + E c f + Dc ζ
their operating sequence, which are feed by Clarke ( 12 )
transformation of the three-phase electrical current from
microgrid system and thereby the observers are providing the where y c ∈ R
p ×1
, and I is an identity matrix.
estimated measurements to the virtual impedance and the
droop control model.

C.Sensor fault-tolerant of a microgrid using sliding-mode

observer
y c =[ 0 I ]∙
[]
xs
xh
( 13 )
By applying a sliding-mode observer, sensor fault
and,
diagnostics and fault-permissive control for voltage source
converters inner microgrid system were developed in [15]. It
means of modeling a robust control model from the microgrid,
which is tacking to the reliability of its protection system,
considering the precise operation of the current and potential
xc=
[] [] [
xs
xh
; uc =
u A = As
w
; c
0
A h C s − Ah
;
]
> REPLACE THIS LINE WITH YOUR PAPER IDENTIFICATION NUMBER (DOUBLE-CLICK HERE TO EDIT) < 3

Bc =
[ Bs Bg
0 0 ]
; Ec =
0
[ ]
A h Es
; and D c =
0
Ah Ds
;
[ ]
where x c and Bc are the internal variables of the microgrid
nc ×1 2m× 1 p ×1
Simulink diagram. The x c ∈ R , uc ∈ R , xh ∈ R ,
q×q q ×q nc × qc
Ah Ds =Do ∈ R , Ah E s=Eo ∈ R , D c ∈ R , and
n ≥ p ≥ q were analyzed when p=4 , q=4, and n+ p=10.
Given (12), the authors of this research have been using the
fault detect observer with appropriate dimension such that the
next condition holds:

Figure.6. Simulink diagram model of a microgrid system (modified from

[15]).