CHAPTER-4: RECONFIGURATION OF

UNBALANCED RADIAL DISTRIBUTION

NETWORKS IN REGULATED
ENVIRONMENTS ALONG WITH DG AND
CAPACITOR INTEGRATION
4.1 Introduction

Electric power systems are widespread, normally interconnected and impressively more
complex. Practically, in distribution networks the loadings at the nodes are unbalanced,
which causes losses of the system resulting in capacity limit violation. Hence power quality
becomes crumbled and electricity cost is also increased (Raju and Bijwe (2008)). There are
a few operational plans in power distribution systems and one of these is the system
reconfiguration. Reconfiguration of any network not only reduces the active power loss
but also releases the system from overloads as there are change in operating conditions
(Siti et al. (2007)).

The tie-line switches of a power distribution system are closed and sectionalizing switches
are opened during the process of network reconfiguration to obtain a network topology so
that the goals are achieved (Das (2006)). The circuit topology gets changed due to operation
of switches to get a new network configuration, reducing losses of the system and working
costs (Imran et al. (2014)). Network reconfiguration is a noteworthy distribution system
administrator control, which is utilized for different purposes (Capitanescu et al. (2015)).
Contingent upon the present loading conditions, reconfiguration may get to be vital so as
to take out overloads on a particular system segments, for example, transformers and this
is known as load balancing. The system load balancing index is utilized to focus the loading
status of the system and greatest system loading limit (Lantharthong and
Rugthaicharoencheep (2012)). The network reconfiguration issue in studies has been
detailed as a single target optimization issue with uniformity and imbalance constraints
(Ahmadi and Martí (2015)).

Page | 35
In conventional power systems, network reconfiguration adds to the accompanying
advantages, which a few researchers had tested using smart routines, like decreasing
system power losses (Amanulla et al. (2012)), effective restoring system (Gholami et al.
(2015)), enhancing system reliability (Kavousi-Fard and Akbari-Zadeh (2013)), and
balancing load (Mirhosein et al. (2014)). The primary contrast between these networks is
the presence of numerous vitality merchants with different conditions for offering energy
in a focused cost (Kargar et al. (2008)). These days, distribution networks that are being
affected by expanding the addition of distributed generations (DGs), which are constantly
utilized as a part of the transmission network. There could be found considerable reduction
in real power losses with DGs integration in the network, which results in saving of energy
and fuel cost. Also, there will be overall improvement in terms of technical and economic
benefits in the energy market operation (Kumar and Gao (2010)). Hereafter, DG got to be
one of the applicable parameters in the assessment of network reconfiguration.
Incorporation of DG in any distribution network influences the operation of this
distribution network in different courses, which needs a research-based exploration.

In this chapter, the focused work on optimal network reconfiguration of unbalanced

distribution network in regulated environment along with integration of DG and capacitor
is presented, so that the power losses are reduced and the voltage profiles are improved.
Role of network reconfiguration and integration of DG and capacitor has been discussed
in Section 1.2 (Chapter 1) and literature survey is detailed in Section 2.3 (Chapter 2). In
this work fuzzy firefly algorithm is applied for optimal network reconfiguration and is a
different technique than those existing in literature. Also, the DG and capacitor have been
integrated in the reconfigured network to reduce the system losses.

The formulation of the problem proposed herein considers the accompanying viewpoints

in regulated environments:

 System losses before network reconfiguration.

 Network reconfiguration using the fuzzy-firefly algorithm.
 Losses computed after reconfiguration.
 Integration of DG and capacitor in reconfigured network for further loss reduction.

Page | 36
4.2 Proposed method for network reconfiguration of unbalanced
distribution networks
Most of the network reconfiguration issues for reduction of power loss and improvement
of voltage profile concentrate on balanced distribution networks. In this research work, a
new methodology is proposed for network reconfiguration of unbalanced distribution
networks by using fuzzy-firefly (FF) algorithm for reduction of losses of the system and
improving the voltage profiles under acceptable limits. Initially, the load-flow analysis is
done on unbalanced radial distribution networks. The metaheuristic algorithm called firefly
algorithm, which has been motivated by the flashing manner of fireflies with fuzzy
dependent objective function in terms of real power loss, bus voltage and load equalizing
index to get a suitable solution for the optimal network reconfiguration. This proposed
methodology for network reconfiguration of unbalanced distribution networks using Fuzzy
Firefly algorithm produces better results in terms of power loss reduction as well as
maintaining a voltage profile. The steps of the suggested method are shown in Figure 4.1.

Figure 4.1: Steps of the proposed method

From Figure 4.1, an unbalanced distribution network is considered in which the magnitude
of voltage, currents and power losses at all the nodes are computed by using load-flow
analysis. In the next step a fuzzy-firefly (FF) algorithm is designed for optimal network
reconfiguration in which the objective function is computed based on fuzzy membership
functions. Finally, an optimized unbalanced distribution network with reduced power
losses as well as improved voltage profile will be obtained from this work. In the next
section, detailed explanation of the proposed load-flow analysis of unbalanced distribution
network is presented.

Page | 37
4.3 Fuzzy – firefly (FF) algorithm

In this proposed method, a Fuzzy-firefly algorithm is designed for optimal network

reconfiguration. Objective functions have dissimilar dimensions. For easier contrast, a
fuzzy multi-objective approach is used. In the fuzzy sphere, a membership function is
extracted for each target, which signifies the degree of fuzzy satisfaction of the objective.
The membership value for each target is a real number between 0 and 1 and in this segment
it is squared up by using the trapezoidal fuzzy membership function. In this research work,
minimization of power losses, minimizing the bus voltage deviation and load balancing in
the feeders are deliberated as the objectives and fuzzified as described below.

4.3.1 Real power loss membership function ( Pi )

Mathematically, the real power losses in the network can be expressed by Equation (4.1)
Nbranch Nbranch P 2
 Qm ( k 1) 2 
 
2
PLoss  I ( jj ) Rm ( jj )   m ( k 1)  Rm ( jj ) (4.1)
jj 1 jj 1  V
2

 m ( k 1) 
In Equation (4.1), jj is the branch between nodes k and k + 1, where k is the sending-end
node and k + 1 is the receiving-end node. I(jj) is the current through the branch-jj. Rm (jj) is
the resistance of the branch (jj) for phase m. Pm (k+1) and Qm(k+1) have already been defined
in Chapter 3 [vide Equation (3.13) and (3.14) respectively]. The magnitude of voltage at
each bus should remain within its allowable intervals. On the other hand, the current in
each branch should satisfy the branch current limitations as presented by Equation (4.2)
and Equation (4.3) respectively.
Vmin  Vm ( k 1)  Vmax (4.2)

I ( jj )  I ( jj ),max (4.3)

Where Vm(k+1) is the voltage at bus (k+1) of phase ‘m’ and Nbranch is the total number of
branches in the system. Vmin and Vmax are the acceptable minimum and maximum voltages
such that Vmin = 0.95 and Vmax =1.05. I(jj),max is the maximum value of current through the
branch-jj ,which is I(jj),max =1. The power loss minimization index is expressed by
Equation (4.4).

Page | 38
 PLi
XPLi  (4.4)
PL0

In Equation (4.4), PL and PL are the respective real power losses before reconfiguration
0 i

of the network and after reconfiguration in i th system. The degree of fuzzy satisfaction of
power loss objective function can be computed by using the membership function as
described in fuzzy domain. The membership function in the fuzzy domain helps to compute
the degree of fuzzy satisfaction of power loss objective function and expressed by Equation
(4.5) (Saffar et al. (2011)).

 1 XPLi  XPLmin
 XP  XP

PLi   Lmax Li
XPLmin  XPLi  XPLmax (4.5)
 XPLmax  XPLmin
 0 XPLi  XPLmax

In Equation (4.5), XPL min

and XPL max
represent the respective lower and upper limits of XPL i

index. The best and the worst system configuration for real power losses are deliberated to
compute these two limits. PL will have a minimum value of the power loss for the best
i

system configuration where as if the power loss is equal to that of at the initial
configuration, the worst system configuration will be anticipated.

4.3.2 Maximum bus voltage membership function:

The index of bus voltage XVi is described by Equation (4.6), which is for the purpose of

minimizing the bus voltage deviation.

XVi  max1  Vmin and 1  Vmax  (4.6)

In Equation (4.6) V min and Vmax are the respective minimum and maximum value of bus

voltage. The membership function of maximum bus voltage deviation index (Sedighizadeh
et al. (2013), Sedighizadeh et al. (2015)) is expressed by Equation (4.7).
 1 XVi  XVmin
 XV  XV
Vi   max i
XVmin  XVi  XVmax (4.7)
XV
 max  XVmin
 0 XVi  XVmax

Page | 39
In Equation (4.7) the respective lower limit and upper limit of XVi index are XVmin and
XVmax. To compute the XVmin and XVmax, the best and worst system configurations are
required for respective minimum and maximum bus voltage deviation.

4.3.3 Load balancing index (LBI) membership function:

An appropriate parameter is required to be defined at first, which will indicate what portion
of the branches is loaded. This is needed for the purpose of load balancing (Saffar et al.
(2011)). This portion is termed as the line usage index for the jjth branch. If I (jj) is the
current through the branch-jj and I(jj),max is its maximum current carrying capacity, the line
usage index for jjth branch is defined by Equation (4.8).
I jj
Line Usage Index = (4.8)
I ( jj ) max
The load balancing index for all the branches of the system is to be computed, which is
defined by Equation (4.9).
 I I2 I3 IN 
Y  1   (4.9)
 I (1) max I (2) max I (3) max I ( Nbr )max 

LBI  Var Y  (4.10)

In Equation (4.10) Var denotes the variance operation. If the LBI has the minimum value,
i.e., smaller one, successful achievement of load balancing is done. In the next phase, the
index of XBi for load balancing is expressed by Equation (4.11).

LBI i
XBi  (4.11)
LBI 0
In Equation (4.11), LBI0 represents the load balancing of the base network, i.e., before
network reconfiguration, which is computed in the initial power-flow. LBI i in Equation
(4.11) is the ith radial system’s load balancing after network reconfiguration. Membership
function for feeder load balancing index is defined by Equation (4.12).

 1 XBi  XBmin
 XB  XB
I i   max i
XBmin  XBi  XBmax (4.12)
XB
 max  XBmin
 0 XBi  XBmax

Page | 40
In Equation (4.12), the respective lower and upper limits of XBi are XB min and XBmax . The

best and the worst system configuration is considered for feeder load balancing so that
XB min and XBmax are computed.

4.3.4 Formation of objective function

The objective function is designed for the two purposes. The objective function is designed
to obtain the finest as well as the most matched network configuration so that the
operational limits such as voltage and current constraints are satisfied and load islanding
will also be presented (Sedighizadeh et al. (2013), Sedighizadeh et al. (2015)). The
objective function is given by Equation (4.13).

 Fi    PL  Vi   Ii 
1
3
i
(4.13)

The best well matched configuration will be acquired when the fitness function in Equation
(4.13) will be maximized during the optimization process. Fi will be allocated a value
zero when any membership function of each objective attains the value of zero. The correct
information regarding the achievement of ideal state, i.e., a value of 1 is provided by this
objective function. The objective function given in Equation (4.13) is utilized as a fitness
function. The algorithm used for optimization purpose is explained in the next section.

4.3.5 Optimal reconfiguration by firefly algorithm

The firefly algorithm is a nature enlivened algorithm for optimization (Yang (2010)) and
it is by all accounts more proficient than different algorithms like particle swarm
optimization (PSO). As in nature, the firefly is pulled into a flashing light created from
alternate ones. This is to be mated by the same species or lethally luring by alternate
species. One firefly tends to move toward the brightest flashing light and the brightness is
shifted by the separation. Three idealized rules are used in the firefly algorithm. The main
is that all fireflies are unisex accordingly they are pulled into the next depending just on
the brightness. The attractiveness of a firefly can be figured as one takes over, which is
expressed by Equation (4.14).

  r    0e  r
2
(4.14)

Page | 41
The next one is that the brightness varies by distance, which can be computed by using
Equation (4.15).
rmn  xm  xn (4.15)

Between two fireflies, the smaller brightness moves to the brighter one. The movement of
firefly ‘m’ attracted to firefly ‘n’ is computed by Equation (4.16).
xm'  xm   0 e   rmn  xn  xm     rand  0.5 
2
(4.16)

Where,  is intensity or attractiveness of the firefly’s flashing light,

 0 is attractiveness at zero distance,
rmn is the distance between two fireflies m and n (m and n are firefly numbers),

x is location of the firefly,

x’ is new location of the firefly,
 is light absorption coefficient of a given medium,
 is randomization parameter,   0 ,1 and
rand is uniformly distributed random number n [0, 1].

The third is that the brightness is exaggerated by the objective function. Consequently, in
this proposed method fuzzy based fitness function is formed, which is given by Equation
(4.13). In this method, the firefly algorithm is implemented to reconfigure any unbalanced
distribution network. A required solution is a set of the switching status. Though, in a
certain system, the number of feeders is fixed and so does the number of open switches.
Consequently, a solution set to deliberate here contains only switch numbers to be open,
which is encoded as the location of fireflies.

The fuzzy firefly algorithm is used to decide the optimal network reconfiguration issue.
The process of searching an optimal network reconfiguration of the distribution network
utilizing firefly algorithm starts with parameter readings and network data. The firefly
population number is the solution set number. Each and every set comprises of different
open switches and denotes to a resultant network configuration. The open switches in each
and every set are initially produced randomly, which are the initial location of fireflies. The
firefly brightness at each location is computed by inverse of the fuzzy based objective
function. The most promising one is reserved as local optimum for initial placement.

Page | 42
Regard it as global optimum, if that is the best solution ever found. Each and every firefly
is combined randomly and then it is actuated to the other new location due to the
attractiveness among each other. After the relocation, the algorithm initializes the next
iteration. The firefly brightness at the new location is computed to obtain the new local
optimum, which needs to relate with the global optimum of the current iteration. The
brighter one is reserved as a new global solution. This procedure is carried on until the
maximum iteration is reached. The procedure of incorporating the algorithm to solve
network reconfiguration is shown beneath.

Initialization: The algorithm starts with encoding parameters like set of

branches (switches), where each switch has two possible states either ‘0’
for tie switch (Opened Switch) or ‘1’ for sectionalizing switch (closed
switch), set of buses, base configuration and firefly parameters.
Process: Run the load flow analysis for each reconfigured pattern and
compute the fuzzy based objective function.
Termination: The process continues until maximum number of iterations
reached or until no improvement of objective function detected after
specified number of iterations.
Print: Display the result.

The objective function is a combination of real power loss, maximum bus voltage
deviation, and load balancing index, which is given in Equation (4.13). In firefly algorithm,
the opposite of the target function is deliberated as brightness, which signifies that the
lower losses are the most promising. The solution candidate with the corresponding
brightest at the final iteration is considered to be a global solution. In this proposed method,
the fireflies move with,   1 ,   0.75 and   0.25 . These sets of parameter values gave
better results. The number of firefly population is 100 and the maximum number of
iterations is selected as 150. The flow chart of the suggested method for optimal network
reconfiguration is shown in Figure 4.2.

Page | 43
In this process of reconfiguration to find out the system parameters load-flow solution
suggested in Chapter 3 is utilized and gave us promising results.

4.4 Cost of energy loss

The monetary value of energy losses (Murty and Kumar (2015)) (CL): The annual price of
energy loss is made by

CL = (Net active power loss) × (Ec ×T) $ (4.17)

Where E c =Energy Rate = 0.06 $/kW h and T = time duration = 8760 h.

4.5 DG and capacitor placement in unbalanced distribution network

The optimum location of DG and capacitor in unbalanced distribution network is decided

by Equation (4.18) (Hong (2014)), which presents the mathematical expression of “Loss
Sensitivity Factor” (LSF) of three phase unbalanced distribution network. LSF is arranged
in descending order and the highest value is considered for placement of DG and capacitor.
dPLoss
LSF   2 Kbr  n I br  a Rbr  .a cos( I  br  a   I  n  )
d I 
+2 Kbr  n I br  b Rbr  .b cos( I  br  b   I  n  )
+2 Kbr  n I br  c Rbr  .c cos( I  br  c   I  n  )
(4.18)

In Equation (4.18), K, I, br, n,  and R are the voltage ratio, current, branch number, bus
number, phase and resistance.

Page | 44
 Start

Read System Data

Initialize variables as open switch numbers and populate

initial location of Firefly

Reconfigure the system in correspondence of firefly

population

Determine Load flow solution for each reconfigured pattern

Compute the fuzzy based objective functions

Pick inverse of Objective function as brightness

Pick the solution with the brightest as local optimum (for

current iteration)

Pick the brightest of all iterations as global optimum

Iteration maximum?

Compute distance, attractiveness and movement to obtain

new location of firefly

Optimal Reconfiguration in terms of switching statuses

Stop

Figure 4.2: Flowchart of unbalanced distribution network reconfiguration

using fuzzy-firefly algorithm

Page | 45
The DG (Type-I i.e., it delivers only active power) is placed at first to the most sensitive
node and its size is determined by Bacteria foraging optimization algorithm (BFOA). The
details of BFOA are available in Das et al. (2009). The objective function in this case is
given in Equation (4.19).

Objective function (OF) = Minimize (Real power loss) (4.19)

Next LSF is found once again. The capacitor is placed to the most sensitive node and its
size is determined by Bacteria foraging optimization algorithm (BFOA). Here DG is placed
at first because the available LSF is related to resistance of the network. The proposed LSF-
BFOA method for placement of DG and capacitor is to be compared with LSF-GA method.
The details of Genetic algorithm (GA) are available in Yadav and Prajapati (2012).

4.6 Simulation results and discussions

The proposed method, i.e., load-flow and network reconfiguration are implemented in
MATLAB 2013a working platform (Windows 8, Intel Core i3-3210, 3.19 GHz). The
parameters used for simulation are shown in Table 4.1.

Table 4.1: The parameters used for simulation

Parameters Specifications

Substation voltage 1 0 p.u.

Maximum branch voltage 1.05

Minimum branch voltage 0.95

Maximum branch current 1.0

Alpha 0.25

Beta 1.0

Gamma 0.75

Firefly population 100

Maximum number of iterations 150

Page | 46
The effectiveness of the proposed Fuzzy-Firefly algorithm for network reconfiguration of
unbalanced distribution networks is tested rigorously on 19-node and 25-node unbalanced
distribution networks as discussed below.

4.6.1 19-node unbalanced distribution network

Figure 4.3 shows the initial configuration of 19-node unbalanced radial distribution
network having 11 kV and 1 MVA as base values. The network data is available in
Vulasala et al.(2009). The magnitude of voltage in p.u. of each node and voltage angle
(deg.) before network reconfiguration for each phase is shown in Table 4.2.

Figure 4.3: 19-node unbalanced distribution network

Page | 47
Table 4.2: Magnitude of voltage in p.u. of each node and voltage angle (deg.) before
network reconfiguration for 19-node unbalanced distribution network
Node Voltage (p.u.) Voltage angle (deg.)
Number Phase A Phase B Phase C Phase A Phase B Phase C
1 1.0000 1.0000 1.0000 0.00 -120.00 120.00

2 0.9875 0.9891 0.988 0.01 -119.98 120.05

3 0.9854 0.9887 0.9863 0.00 -119.98 120.06

4 0.9824 0.9839 0.9830 0.03 -119.97 120.06

5 0.9820 0.9837 0.9828 0.03 -119.97 120.07

6 0.9793 0.9808 0.9801 0.04 -119.96 120.07

7 0.9786 0.9803 0.9796 0.04 -119.96 120.08

8 0.9728 0.9738 0.9735 0.06 -119.94 120.08

9 0.9659 0.9660 0.9657 0.08 -119.91 120.09

10 0.9563 0.9555 0.9550 0.09 -119.86 120.09

11 0.9550 0.9543 0.9533 0.10 -119.86 120.10

12 0.9548 0.9538 0.9536 0.11 -119.87 120.10

13 0.9544 0.9534 0.9521 0.10 -119.85 120.11

14 0.9545 0.9539 0.9528 0.10 -119.86 120.11

15 0.9527 0.9512 0.9513 0.11 -119.83 120.12

16 0.9534 0.9515 0.9522 0.13 -119.86 120.10

17 0.9537 0.9534 0.9523 0.10 -119.90 120.11

18 0.9538 0.9532 0.9521 0.10 -119.82 120.10

19 0.9514 0.9494 0.9501 0.13 -119.86 120.10

Page | 48
In the initial configuration of the network, the tie-line switches 19 and 20 are opened. After
closing these two tie line switches, the proposed method for network reconfiguration is
implemented and Figure 4.4 is obtained after opening the switches 2 and 4. The magnitude
of voltage in p.u. of each node and voltage angle (deg.) after network reconfiguration for
each phase is shown in Table 4.3. The real and reactive power losses of each branch at
three phases before and after network reconfiguration are shown in Figure 4.5 and
Figure 4.6 respectively. The respective voltage profile of 19-node unbalanced distribution
network before and after network reconfiguration at three phases is shown in Figure 4.7.
The results obtained for 19-node unbalanced distribution network before and after network
reconfiguration using the suggested method is shown in Table 4.4. Table 4.4 shows that
there is improvement of minimum voltage for each phase and the real power loss and
reactive power loss of each phase are considerably reduced. The energy cost is also reduced
drastically after network reconfiguration.

Figure 4.4: Reconfigured 19-node unbalanced distribution network in regulated

environment

Page | 49
Table 4.3: Magnitude of voltage in p.u. of each node and voltage angle (deg.) after
network reconfiguration for 19-node unbalanced distribution network
Node Voltage (p.u.) Voltage angle (deg.)
Number Phase A Phase B Phase C Phase A Phase B Phase C
1 1.0000 1.0000 1.0000 0.00 -120.00 120.00

2 0.9917 0.9922 0.9982 0.02 -119.97 120.06

3 0.9831 0.9847 0.9893 0.00 -119.93 120.10

4 0.9904 0.9902 0.9958 0.05 -119.96 120.06

5 0.9836 0.9848 0.9893 0.04 -119.96 120.06

6 0.9887 0.9890 0.9941 0.04 -119.97 120.07

7 0.9838 0.9887 0.9936 0.07 -119.97 120.07

8 0.9867 0.9871 0.9937 0.05 -119.97 120.07

9 0.9856 0.9864 0.9924 0.10 -119.96 120.06

10 0.9851 0.9861 0.9915 0.09 -119.96 120.06

11 0.9845 0.9859 0.9912 0.13 -119.86 120.13

12 0.9848 0.9860 0.9914 0.10 -119.94 120.07

13 0.9841 0.9855 0.9902 0.14 -119.85 120.13

14 0.9840 0.9850 0.9909 0.12 -119.86 120.13

15 0.9844 0.9858 0.9896 0.10 -119.94 120.07

16 0.9844 0.9857 0.9907 0.15 -119.93 120.06

17 0.9835 0.9848 0.9899 0.10 -119.89 120.12

18 0.9825 0.9844 0.9891 0.12 -119.86 120.13

19 0.9840 0.9851 0.9895 0.10 -119.95 120.07

Page | 50
 (a) Real power loss at phase A

(b) Real power loss at phase B

(c) Real power loss at phase C

Figure 4.5: Real power losses at three phases of 19-node unbalanced
distribution network

Page | 51
 (a) Reactive power loss at phase A

(b) Reactive power loss at phase B

(c) Reactive power loss at phase C

Figure 4.6: Reactive power losses at three phases of unbalanced 19-node unbalanced
distribution network

Page | 52
 (a)Voltage Profile at phase A

(b) Voltage Profile at phase B

(c) Voltage Profile at phase C

Figure 4.7 Voltage Profiles at three phases of 19-node unbalanced distribution network

Page | 53
Table 4.4: Results of 19-node unbalanced distribution network before and after network
reconfiguration
Before Reconfiguration After Reconfiguration
Parameters
Phase A Phase B Phase C Phase A Phase B Phase C
Real Power Loss
4.85 4.56 4.78 0.41 1.02 1.3
(kW)
Reactive Power
1.99 1.9 1.97 0.155 0.169 0.14
Loss (kVAr)
Magnitude of
0.9514 0.9494 0.9501 0.9825 0.9844 0.9891
Minimum Voltage
(19) (19) (19) (18) (18) (18)
(p.u.)
Switches Opened 19, 20 2, 4
% Real Power
- - - 91.55 77.63 72.80
Loss Reduction
% Reactive Power
- - - 92.21 91.11 92.89
Loss Reduction
Load Balancing
0.0329944 0.1035784
Index (LBI)
Energy Cost
7458.264 1434.89
($/kWh)

4.6.2 25-node unbalanced distribution network

Among a few unbalanced studies that have been performed in reconfiguration problem, a
4.16 kV, 30 MVA, 25-node unbalanced distribution network is considered. It has 3 tie
switches and its total demand is 3.239 + j2.393 MVA. The initial configuration is shown
in Figure 4.8. The fuzzy based objective function, which is the combination of real power
loss, bus voltage deviation and load balancing, is considered as fitness function for firefly
algorithm. In the initial configuration of this network, the tie-line switches 25, 26 and 27
are opened. After closing these tie line switches, the suggested method for network
reconfiguration is used and Figure 4.9 is obtained after opening the switches 6, 15, and 17.
The magnitude of voltage in p.u. of each node and voltage angle (deg.) before network
reconfiguration for each phase is shown in Table 4.5. The magnitude of voltage in p.u. of
each node and voltage angle (deg.) after network reconfiguration for each phase is shown
in Table 4.6.

Page | 54
 Figure 4.8: 25-node unbalanced distribution network

Figure 4.9: Reconfigured unbalanced 25-node unbalanced distribution network in

regulated environment

Page | 55
Table 4.5: Magnitude of voltage in p.u. of each node and voltage angle (deg.) before
reconfiguration for 25-node unbalanced distribution network
Node Voltage (p.u.) Voltage angle (deg.)
Number Phase A Phase B Phase C Phase A Phase B Phase C

1 1.0000 1.0000 1.0000 0.00 -120.00 120.00

2 0.9702 0.9711 0.9755 -0.57 -120.41 119.31
3 0.9632 0.9644 0.9698 -0.70 -120.52 119.15
4 0.9598 0.9613 0.9674 -0.77 -120.57 119.08
5 0.9587 0.9603 0.9664 -0.76 -120.57 119.08
6 0.9550 0.9559 0.9615 -0.56 -120.36 119.29
7 0.9419 0.9428 0.9492 -0.55 -120.30 119.27
8 0.9529 0.9538 0.9596 -0.56 -120.35 119.29
9 0.9359 0.9367 0.9438 -0.55 -120.28 119.26
10 0.9315 0.9319 0.9395 -0.55 -120.26 119.25
11 0.9294 0.9296 0.9376 -0.55 -120.26 119.25
12 0.9285 0.9287 0.9369 -0.55 -120.25 119.26
13 0.9287 0.9287 0.9373 -0.55 -120.26 119.25
14 0.9359 0.9370 0.9434 -0.55 -120.27 119.26
15 0.9338 0.9349 0.9414 -0.55 -120.27 119.25
16 0.9408 0.9418 0.9483 -0.55 -120.30 119.27
17 0.9347 0.9360 0.9420 -0.55 -120.27 119.26
18 0.9573 0.9586 0.9643 -0.70 -120.50 119.15
19 0.9524 0.9544 0.9600 -0.69 -120.49 119.16
20 0.9548 0.9563 0.9620 -0.70 -120.49 119.15
21 0.9537 0.9549 0.9605 -0.69 -120.49 119.16
22 0.9518 0.9525 0.9585 -0.69 -120.48 119.17
23 0.9565 0.9584 0.9648 -0.76 -120.57 119.08
24 0.9544 0.9565 0.9631 -0.76 -120.57 119.07
25 0.952 0.9547 0.9612 -0.76 -120.57 119.08

Page | 56
Table 4.6: Magnitude of voltage in p.u. of each node and voltage angle (deg.) of 25-node
reconfigured unbalanced distribution network
Node Voltage (p.u.) Voltage angle (deg.)
Number Phase A Phase B Phase C Phase A Phase B Phase C
1 1.0000 1.0000 1.0000 0.00 -120.00 120.00
2 0.9805 0.9825 0.9880 -0.56 -120.41 119.32
3 0.9798 0.9813 0.9825 -0.71 -120.53 119.14
4 0.9786 0.9801 0.9805 -0.80 -120.61 119.04
5 0.9601 0.9759 0.9730 -0.80 -120.60 119.03
6 0.9797 0.9815 0.9871 -0.56 -120.36 119.29
7 0.9775 0.9788 0.9838 -0.56 -120.32 119.28
8 0.9790 0.9801 0.9850 -0.55 -120.32 119.28
9 0.9755 0.9763 0.9788 -0.56 -120.3 119.27
10 0.9699 0.9760 0.9755 -0.56 -120.29 119.26
11 0.9619 0.9759 0.9738 -0.56 -120.29 119.26
12 0.9587 0.9758 0.9730 -0.79 -120.59 119.05
13 0.9581 0.9756 0.9728 -0.56 -120.29 119.26
14 0.9734 0.9758 0.9835 -0.56 -120.31 119.27
15 0.9718 0.9755 0.9761 -0.54 -120.32 119.28
16 0.9686 0.9757 0.9742 -0.56 -120.32 119.27
17 0.9685 0.9756 0.9830 -0.55 -120.30 119.27
18 0.9676 0.9763 0.9738 -0.71 -120.51 119.14
19 0.9623 0.9752 0.9736 -0.70 -120.50 119.14
20 0.9631 0.9761 0.9737 -0.71 -120.50 119.13
21 0.9634 0.9762 0.9735 -0.71 -120.51 119.14
22 0.9628 0.9760 0.9733 -0.80 -120.59 119.03
23 0.9697 0.9799 0.9797 -0.80 -120.60 119.03
24 0.9696 0.9762 0.9736 -0.80 -120.60 119.03
25 0.9687 0.9760 0.9731 -0.79 -120.60 119.04

Page | 57
The real and reactive power losses of each branch at three phases before and after
reconfiguration of 25-node unbalanced distribution network is shown in Figure 4.10 and
Figure 4.11 respectively. The respective voltage profile of 25-node unbalanced distribution
network before and after network reconfiguration at three phases is shown in Figure 4.12.

(a) Real power loss at phase A

(b) Real power loss at phase B

(c) Real power loss at phase C

Figure 4.10: Real power losses at three phases of unbalanced 25-node unbalanced
distribution network before and after network reconfiguration in
regulated environment

Page | 58
 (a) Reactive power loss at phase A

(b) Reactive power loss at phase B

(c) Reactive power loss at phase C

Figure 4.11: Reactive power losses at three phases of unbalanced 25-node unbalanced
distribution network before and after network reconfiguration in regulated
environment

Page | 59
 (a) Voltage profile at phase A

(b) Voltage profile at phase B

(c) Voltage profile at phase C

Figure 4.12: Voltage profiles at three phases of unbalanced 25-node unbalanced

distribution network before and after network reconfiguration in
regulated environment

Page | 60
The results obtained for of 25-node unbalanced distribution network before and after
network reconfiguration using the proposed method are shown in Table 4.7.

Table 4.7: Results of 25-node unbalanced distribution network before and after network
reconfiguration
Before Reconfiguration After Reconfiguration
Parameters
Phase A Phase B Phase C Phase A Phase B Phase C

Real Power Loss

52.92 54.44 43.97 14.95 15.03 9.75
(kW)
Reactive Power
60.43 52.32 55.87 17.30 15.39 9.54
Loss (kVAr)
Magnitude of
0.9285 0.9287 0.9369 0.9581 0.9756 0.9728
Minimum Voltage
(12) (12) (12) (13) (13) (13)
(p.u.)
Switches Opened 25, 26, 27 6, 15, 17
% Real Power
- - - 71.75 72.39 77.85
Loss Reduction
% Reactive Power
- - - 71.37 70.58 82.92
Loss Reduction
Load Balancing
0.1009658 0.0454214
Index (LBI)
Energy Cost
79539.048 20882.088
($/kWh)

Table 4.7 shows that there is improvement of minimum voltage for each phase and the real
power loss and reactive power loss for each phase are considerably reduced. The energy
cost is also reduced drastically after network reconfiguration.

4.6.3 Performance comparison

In the proposed method, firefly algorithm has been used. In order to prove or justify its
superiority, the results (real power and reactive power losses, minimum voltage of each
phase along with its node number, CPU time, number of power-flows and energy cost)
obtained by the proposed method have been compared with that of obtained by other
optimization algorithms like “Genetic Algorithm (GA), Particle Swarm Optimization
(PSO), Artificial Bee Colony (ABC) and GA-PSO algorithm” using the same objective
function and proposed load-flow method. The outcomes have been presented in Table 4.8
for 25-node and 19-node unbalanced distribution networks (UDNs).

Page | 61
 Table 4.8: Performance comparison of suggested technique of network reconfiguration with other algorithms

Real Reactive
Magnitude of CPU No. of Energy
Open Power Power Value of
Methods UDNs Minimum Voltage Time Power- Loss
Switches Loss Loss fitness
(p.u.) (sec.) Flows ($/kWh)
(kW) (kVAr) function
0.9581 (13), 0.9756
Fuzzy- 25-node 6, 15, 17 39.73 42.23 0.52 21 20882.088
(13), 0.9728 (13) 0.925
Firefly
0.9825 (18), 0.9844
Algorithm 19-node 2, 4 2.73 0.464 0.35 16 1434.89
(18), 0.9891 (18) 0.942
17, 22, 0.9408 (25), 0.9487
25-node 91.36 105.41 2.110 98 48018.816
Genetic 24 (25), 0.9523 (25) 2.917
Algorithm 0.9624 (16), 0.9503
19-node 9, 14 12.05 4.95 1.586 84 6333.48
(16), 0.9509 (16) 2.723
15, 17, 0.9468 (12), 0.9568
25-node 71.14 81.98 2.01 44 37391.184
ABC 22 (12), 0.9591 (12) 2.674
Algorithm 0.9685 (18), 0.9588
19-node 17, 18 9.59 3.39 1.29 35 5040.504
(18), 0.9606 (18) 2.701
15, 17, 0.9489 (13), 0.9618
25-node 62.14 63.21 1.421 31 32660.784
PSO 20 (13), 0.9641 (13) 2.225
Algorithm 0.9729 (18), 0.9624
19-node 2, 14 5.89 2.22 1.12 26 3095.784
(18), 0.9689 (18) 2.187
0.9521 (25), 0.9699
25-node 6, 17, 24 45.32 49.147 0.99 27 23820.192
GA-PSO (25), 0.9697 (25) 1.688
Algorithm 0.9790 (16), 0.9715
19-node 10, 11 4.0 1.11 0.59 18 2102.4
(16), 0.9803 (16) 1.680

Page | 62
Table 4.8 shows that the firefly algorithm has better performance than other algorithms like
Genetic Algorithm, ABC algorithm, PSO algorithm and GA-PSO algorithm.

The suggested method has also been compared with the methods of (Vulasala et al.(2009),
Subrahmanyam and Radhakrishna (2009), Sedighizadeh et al. (2013), Taher and Karimi
(2014), Sedighizadeh (2015)) as presented in Table 4.9.

Table 4.9: Performance Comparison of suggested technique of network

reconfiguration with Existing Methods (Vulasala et al.(2009),
Subrahmanyam and Radhakrishna (2009), Sedighizadeh et al. (2013),
Taher and Karimi (2014) and Sedighizadeh (2015))
Magnitude
Real Reactive
of Energy
Bus Open Power Power
Methods Minimum Cost
System Switches Loss Loss
Voltage ($/kWh)
(kW) (kVAr)
(p.u.)
0.9581 (13)
25-node
6, 15, 17 39.78 42.27 0.9756 (13) 20908.368
System
Proposed 0.9728 (13)
Method 0.9825 (18)
19-node
2, 4 2.73 0.464 0.9844 (18) 1434.888
System
0.9891 (18)
25-node 0.9389
System 15,17,22 133.5 157.61 0.9401 70167.5
Vulasala
0.9460
et al.
0.9703
(2009) 19-node
10,11 8.22 3.53 0.9694 4320.43
System
0.9690
0.9389,
25-node
Subrahma 22, 17, 15 136.13 - 0.9401, 71549.928
System
nyam and 0.9459
Radhakris 0.97043,
19-node
hna (2009) 10, 11 8.42 - 0.96954, 4425.552
System
0.96724
Sedighiza 0.93889,
25-node
deh et al. 15, 17, 22 133.49 - 0.94015, 70162.344
system
(2013) 0.94596
Taher and
25-node 120.28
Karimi 11, 14, 21 - - 63220.69
System 29
(2014)
Sedighiza 25-node
25, 17, 15 94.179 - - 49500.482
deh (2015) system

Page | 63
The effectiveness of the proposed method is compared with the existing method
(Vulasala et al. (2009)) in terms of real and reactive power losses, minimum voltage
magnitudes at three phases and energy cost. The suggested method is compared with the
existing methods (Subrahmanyam and Radhakrishna (2009)], Sedighizadeh et al. (2013))
in terms of real power loss, minimum voltage magnitudes at three phases and energy cost.
The suggested method is set side by side with the methods (Taher and Karimi (2014),
Sedighizadeh (2015)) in terms of the real power loss and energy cost. The real and reactive
power losses obtained by the proposed method are lower compared to that of obtained by
the method proposed by Vulasala et al. (2009) for 25-node and 19-node unbalanced
distribution networks. The real power loss obtained by the proposed method is lower
compared to that of obtained by the method proposed by Subrahmanyam and Radhakrishna
(2009) for the above two networks and the magnitude of minimum voltages at three phases
obtained by the proposed method compared to that of obtained by the methods of Vulasala
et al. (2009), Subrahmanyam and Radhakrishna (2009) for these two networks has been
improved. The energy cost obtained by the proposed method is lower compared to that of
obtained by the methods of Vulasala et al. (2009), Subrahmanyam and Radhakrishna
(2009) for the above two networks.

The loss reduction obtained by the proposed method is higher compared to that of obtained
by the method proposed by Sedighizadeh et al. (2013) for 25-node unbalanced distribution
network. The magnitude of minimum voltage of each phase obtained by the proposed
method compared to that of obtained by the method given by Sedighizadeh et al. (2013)
for this network has also been improved. The energy cost obtained by the proposed method
is lower compared to that of obtained by the method Sedighizadeh et al. (2013) for this
network. The proposed method also gives better results in terms of loss reduction and
reduction of energy cost compared to that of obtained by the methods proposed by Taher
and Karimi (2014) and Sedighizadeh (2015) for 25-node unbalanced distribution network.

From Table 4.8 and Table 4.9, it is evident that the proposed method produces an efficient
network configuration after network reconfiguration of an unbalanced distribution
network, which has lower real and reactive power losses, improved voltage profiles and
lower energy cost.

Page | 64
4.6.4 Integration of DG and capacitor in 25-node unbalanced distribution
network in regulated environment
The detailed results for the base case and after network reconfiguration of 25-node
unbalanced distribution network are discussed in section 4.5.2. In the same distribution
network, the DG and capacitor have been integrated and the results are compared before
and after integration. The location DG and capacitor are obtained by the LSF (Loss
sensitivity factor). Their sizes are determined by Genetic Algorithm (GA), Particle Swarm
Optimization (PSO) algorithm, Artificial Bee Colony (ABC) Algorithm and Bacterial
Foraging Optimization algorithm (BFOA ). The outcomes in terms of DG size (kW),
capacitor size (kVAr), Real power loss (kW) and Reactive power loss (kVAr) have been
depicted in Figure 4.13, Figure 4.14, Figure 4.15 and Figure 4.16 respectively. The BFOA
gives the better results. The DG and capacitor are integrated into the 25-node unbalanced
reconfigured network shown in Figure 4.9 using the LSF and BOFA to reduce the losses.
The comparison of results before and after the integration of DG and capacitor in the
reconfigured network are presented in Table 4.10. The results of voltage levels before and
after integration of DG and capacitor in the 25-node unbalanced reconfigured distribution
network are presented in Table 4.11.

Figure 4.13: DG sizes (kW) by GA, PSO, ABC and BFOA at Phase A, Phase B
and Phase C

Page | 65
Figure 4.14: Capacitor sizes (kVAr) by GA, PSO, ABC and BFOA at Phase A,
Phase B and Phase C

Figure 4.15: Real power losses at Phase A, Phase B and Phase C

Page | 66
 Figure 4.16: Reactive power losses at Phase A, Phase B and Phase C

Table 4.10: Comparison results before and after Integration of DG and capacitor in
25-node unbalanced reconfigured distribution network

Before Integration After Integration

Parameters Phase Phase
A B C A B C

Real Power Loss (kW) 14.95 15.03 9.75 4.89 4.96 3.20
Reactive Power Loss
17.30 15.39 9.54 6.01 5.45 3.42
(kVAr)
Minimum Voltage 0.9581 0.9756 0.9728 0.9813 0.990 0.989
(p.u.) (13) (13) (13) (13) 1 (13) 1 (13)
Switches Opened 6, 15, 17 -

DG Location - - - 7 7 7

DG Size (kW) - - - 190 180 150

Capacitor Location - - - 15 15 15

Capacitor Size (kVAr) - - - 200 200 200

Energy Cost ($/kWh) 20882.088 11063.88

Page | 67
Table 4.11: Voltage levels before and after integration of DG and capacitor in
25-node unbalanced reconfigured distribution network
Voltage Levels (p.u.)
Node Before Integration After Integration
No.
Phase A Phase B Phase C Phase A Phase B Phase C
1 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000
2 0.9805 0.9825 0.9880 0.9994 0.9963 0.9946
3 0.9798 0.9813 0.9825 0.9993 0.9959 0.9931
4 0.9786 0.9801 0.9805 0.9989 0.9947 0.9919
5 0.9601 0.9759 0.9730 0.9853 0.9908 0.9898
6 0.9797 0.9815 0.9871 0.9992 0.9962 0.9945
7 0.9775 0.9788 0.9838 0.9989 0.9939 0.9930
8 0.9790 0.9801 0.9850 0.9990 0.9948 0.9943
9 0.9755 0.9763 0.9788 0.9988 0.9933 0.9915
10 0.9699 0.9760 0.9755 0.9972 0.9923 0.9914
11 0.9619 0.9759 0.9738 0.9883 0.9910 0.9912
12 0.9587 0.9758 0.9730 0.9825 0.9905 0.9896
13 0.9581 0.9756 0.9728 0.9813 0.9901 0.9891
14 0.9734 0.9758 0.9835 0.9984 0.9923 0.9929
15 0.9718 0.9755 0.9761 0.9980 0.9917 0.9913
16 0.9686 0.9757 0.9742 0.9953 0.9921 0.9899
17 0.9685 0.9756 0.9830 0.9949 0.9918 0.9909
18 0.9676 0.9763 0.9738 0.9940 0.9925 0.9917
19 0.9623 0.9752 0.9736 0.9891 0.9903 0.9910
20 0.9631 0.9761 0.9737 0.9907 0.9921 0.9915
21 0.9634 0.9762 0.9735 0.9915 0.9924 0.9914
22 0.9628 0.9760 0.9733 0.9901 0.9919 0.9907
23 0.9697 0.9799 0.9797 0.9969 0.9947 0.9918
24 0.9696 0.9762 0.9736 0.9966 0.9932 0.9908
25 0.9687 0.9760 0.9731 0.9956 0.9927 0.9898

Page | 68
Figure 4.17 shows the representation of voltage levels of the Phases A, B and C
respectively before and after integration of DG and capacitor in the reconfigured network.

(a) Voltage level at Phase A

(b) Voltage level at Phase B

(c) Voltage level at Phase C

Figure 4.17: Voltage levels of 25-node unbalanced reconfigured network before

and after integration of DG and capacitor in regulated environment

Page | 69
4.7 Conclusion
A Fuzzy Firefly optimization algorithm has been implemented for multi-objective
reconfiguration of three-phase unbalanced radial distribution networks. The minimization
of total power losses and bus voltage deviation as well as balancing the load in the feeders,
are the main objectives of this method. To acquire the optimal solution for the multi-
objective fitness function, initially each objective is transferred into the fuzzy domain at
first utilizing the membership function and then the resultant overall fuzzified function is
considered as a fitness function, which is maximized during the optimization process. The
Fuzzy-Firefly algorithm for network reconfiguration has been successfully tested on
25-node and 19-node unbalanced distribution networks. Simulation results show that the
fuzzy based firefly algorithm has produced better outcomes in terms of reduction of real
and reactive power losses, improvement of the magnitude of minimum voltage for each
phase and reduction of energy cost in these two unbalanced distribution networks
compared to that of obtained by the other optimization existing algorithms like GA, ABC,
PSO and GA-PSO using the same objective function and proposed load-flow method. This
Fuzzy-Firefly algorithm takes less CPU time and least number of power-flows compared
to that of obtained by the other optimization algorithms like GA, ABC, PSO and GA-PSO
algorithms. The proposed method also gives better results in terms of loss reduction,
improvement of the magnitude of minimum voltage of each phase and reduction of energy
cost compared to that of obtained by the methods Vulasala et al. (2009) and Subrahmanyam
and Radhakrishna (2009) for 25-node and 19-node unbalanced distribution networks and
that obtained by the method of Sedighizadeh et al. (2013) for 25-node unbalanced
distribution network respectively. The suggested method also yields better performance in
terms of loss reduction and reduction of energy cost compared to that of obtained by the
methods Taher and Karimi (2014) and Sedighizadeh (2015). In this work the DG and
capacitor are placed optimally in reconfigured network in the regulated environment using
LSF and BFOA. The results in terms of voltage level and other parameters are compared
for before and after integration of DG and capacitor. The results obtained are effective and
DG and capacitor placement for loss minimization along with reconfiguration comes out
to be a decent choice.

Page | 70