Page 150 - Advances in Renewable Energies and Power Technologies

1. Introduction 123

The current position x i (t), the particle velocity vi(t), and the global best position G i (t)
are used to determine the new position of the particle by calculating the velocity as
shown in the following equation:

v i ðt þ 1Þ¼ uðtÞvðtÞ þ c 1 r 1 ðP i ðtÞ x i ðtÞÞ þ c 2 r 2 ðG i ðtÞ x i ðtÞÞ
|fflfflfflffl{zfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflffl} |fflfflfflfflfflfflfflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflfflfflfflfflfflfflffl}
Inertial parameters Personal best velocity components Global best velocity component
(4.5)
where u(t) is the inertia weight factor that controls the exploration of the search
space. The value of u(t) can be chosen as a constant value equal to 0.5 or as a var-
iable value to accelerate the GP obtained [50]. c 1 and c 2 are the acceleration con-
stants, which provide a weight for the individual and social components GP,
respectively. c 1 is called self-confidence, range: 1.5e2; and c 2 is called swarm con-
fidence, range: 2e2.5 [51].
In using PSO in MPPT of PV system, the duty ratio has been used as particle po-
sition and the increment change in duty ratio as the velocity. The fitness function is
the model for the PV system.
A modified PSO technique called Deterministic PSO (DPSO) [52,53] has be
used to improve the performance of MPPT under partial shading component. In
this technique, the random numbers in the acceleration factor of the conventional
PSO velocity equation have been removed. Additionally, the change in the velocity
values has been restricted to be within a certain value. This research used the expe-
rience about PeV curve under partial shading conditions to impose the limits of the
maximum velocity of each particle and the maximum allowable duty ratio depend-
ing on the resistance of the load and the internal impedance of the PV system, which
add a complexity to the proposed technique and may skip GP out of this limits. Also,
the preset values need previous experience about the size and configuration of PV
system and it may not be suitable for any future extension of the size of the PV sys-
tem without a modification of the optimization parameter. Also, in this paper hill
climbing (HC-MPPT) will be used in uniform distribution of insulation on the PV
system but DPSO will be used in case of partial shading conditions. Authors used
three branches in parallel and four PV modules in series. They used the current in
each branch to distinguish the shading condition depending on the difference in
the current between each branch. But shading can exist in each branch in the
same pattern, so the criteria will not catch partial shading. The only way to catch
shading is to have current and voltage sensors for each PV cell, which may increase
the cost of PV system considerably.
Another PSO research [54] proposed a great idea of using some particles to
locate the local peak (LP) and some other particles to look for GP. The particle
search around the LP will ensure the smooth search of the MPP. Meanwhile, the
particle search for GP will ensure the system catches the GP but these particles
maycause thesystemto workinsomeplacesawayfrommaximum,which can
cause instability of the system and cause unpredicted values away from real GP.
This technique needs more consideration about the system stability; also, this