Article in Press

+Model
JESIT 170 1–17 ARTICLE IN PRESS

Available online at www.sciencedirect.com
ScienceDirect
Journal of Electrical Systems and Information Technology xxx (2017) xxx–xxx
1 Optimal renewable resources placement in distribution networks by

2 combined power loss index and whale optimization algorithms
3 Q1 P. Dinakara Prasad Reddy a,∗ , V.C. Veera Reddy b , T. Gowri Manohar a
4
a Department of EEE, SV University, Tirupati, India
5
b AITTS, Tirupati, India
Received 5 November 2016; received in revised form 12 January 2017; accepted 22 May 2017
7 Abstract
8 Distributed generator (DG) resources are the emerging micro-generating technologies such as fuel cells, micro turbines, IC
9 engines. They also make use of renewable energy sources such as PV arrays and wind turbines. DG units have low emission rates
10 and are environment friendly and economical. Power loss reductions, voltage profile improvement and increasing reliability are
11 some advantages of DG units. The above benefits can be achieved by optimal placement of DGs. Optimal DG locations are obtained
12 from power loss index method. A novel meta heuristic algorithm called whale optimization algorithm (WOA) is used to determine
13 the optimal DG-unit’s size in this paper. WOA modeled based on the unique hunting behavior of humpback whales. The WOA
14 algorithm is tested on IEEE 15-bus, 33-bus, 69-bus, 85-bus and 118-bus test systems. The results obtained by the proposed WOA
15 algorithm was compared with different types of DGs and other evolutionary algorithms. When compared with other algorithms the
16 WOA algorithm gives better results. From the analysis best results have been achieved from type III DG operating at 0.9 pf.
17 © 2017 Electronics Research Institute (ERI). Production and hosting by Elsevier B.V. This is an open access article under the CC
18 BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
19 Keywords: Whale optimization algorithm; Power loss index method; Distributed generation placement; Radial distribution system; Loss reduction
20
21 1. Introduction
22 Distribution system is that part of the power system which connects the high voltage transmission system to low
23 voltage consumers. 70% of the total losses are occurring in the primary and secondary distribution system, while the
24 remaining 30% in transmission and sub transmission lines. Distribution losses are 15.5% of the generation capacity
25 whereas the target level is 7.5%. Therefore the primary and secondary distribution system must be properly planned
26 to ensure losses within the tolerable limits.
∗ Corresponding author.
E-mail address: pdinakarprasad@gmail.com (P. Dinakara Prasad Reddy).
Peer review under the responsibility of Electronics Research Institute (ERI).
http://dx.doi.org/10.1016/j.jesit.2017.05.006
2314-7172/© 2017 Electronics Research Institute (ERI). Production and hosting by Elsevier B.V. This is an open access article under the CC
BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Please cite this article in press as: Dinakara Prasad Reddy, P., et al., Optimal renewable resources placement in distribu-
tion networks by combined power loss index and whale optimization algorithms. J. Electr. Syst. Inform. Technol. (2017),
+Model
2 P. Dinakara Prasad Reddy et al. / Journal of Electrical Systems and Information Technology xxx (2017) xxx–xxx
27 By installing DG units at appropriate positions the losses can be minimized. Photovoltaic (PV) energy, wind turbines,
28 and other distributed generation plants are typically situated in remote areas, requiring the operation systems that are
29 fully integrated into transmission and distribution network. The aim of the DG is to integrate all generation plants to
30 reduce the loss, cost and greenhouse gas emission. The DG units can also improves the profile of voltage, power factor
31 and stability. Thus, DG locations and sizing have an important role in distribution systems.
32 Different types of distributed generations and their definitions are have been discussed in Ackermann et al. (2001).
33 An analytical approach based on exact loss formula has been presented in Acharya et al. (2006) and Duong Quoc
34 et al. (2010) to find the optimal size and location of DG however voltage constraint has not been considered. The
35 distribution system operation uncertainties including daily time varying load, stochastic DG power production, network
36 configuration, and voltage control devices operation are all taken into account in Su (2010). Naik et al. (2013) proposed
37 sensitivity based simultaneous optimal placement of capacitors and DG. In this paper analytical approach is used for
38 sizing. The authors Abu-Mouti and El-Hawary (2011) proposed Artificial Bee Colony Algorithm to find the optimal
39 allocation and sizing of distributed generation. Distributed generation uncertainties (Zangiabadi et al., 2011) have been
40 taken in account for the placement of DG.
41 Alonso et al. (2012), Rahim et al. (2012), Doagou-Mojarrad et al. (2013b), Hosseini et al. (2013), Das et al. (2016)
42 and Doagou-Mojarrad et al. (2013a) proposed evolutionary algorithms for the placement of distributed generation.
43 Nekooei et al. (2013) planned multi-objective harmony search was used for optimal placement of DGs in distribution
44 systems
45 A novel combined genetic algorithm (GA)/particle swarm optimization (PSO) is presented in Moradi (2012) for
46 optimal location and sizing of DG on distribution systems. The non-optimal placement of DG can increase the system
47 losses and thus make the voltage profile lower than the allowable limit. With the proper size of DG gives the positive
48 benefits in the distribution systems. These benefits include improved voltage profile, reduce the losses, increase the
49 distribution capacity, and improve the reliability of system (Rahim et al., 2013; Ameli et al., 2014).
50 DG installations which consider sizing and allocation using Embedded Meta Evolutionary-Firefly Algorithm
51 (EMEFA) was proposed in Rahim et al. (2013). This method mainly focused on the population size on loss. Optimal
52 simultaneous siting and sizing of DGs and capacitors considering reconfiguration by in smart automated distribution
53 systems was proposed by Esmaeilian and Fadaeinedjad (2015) and Golshannavaz (2014). Dynamic load conditions
54 have been taken in Gampa and Das (2015). Big bang big crunch method was implemented for the placement of DG
55 in Hegazy et al. (2014). Authors Murty and Kumar (2014) uses mesh distribution system analysis for the placement
56 of distributed generation with time varying load model. Assessment of maximum distributed generation penetration
57 levels in low voltage networks using a probabilistic approach was discussed in Kolenc et al. (2015). The backtracking
58 search optimization algorithm (BSOA) was used in DS planning with multi type DGs in El-Fergany (2015), BSOA was
59 used to study the impact of various load models on DG placement and sizing. Prakash and Lakshminarayana (2016)
60 proposed whale optimization algorithm for capacitor placement in distribution network. Multiple DG and capacitor
61 placement was proposed in Zeinalzadeh et al. (2015), Rahmani-andebili (2016), Khodabakhshian and Andishgar (2016)
62 and Kaur et al. (2014). In distribution network voltage profile improvement and voltage stability issues have been taken
63 as objectives in Sultana et al. (2016), Singh and Parida (2016) and Aman et al. (2012). Kaveh and Ghazaan (2016)
64 proposed whale optimization algorithm for sizing of skeletal structures.
65 In optimization algorithm literature there is no optimization algorithm that logically proves No-Free-Lunch (NFL)
66 theorem for solving all optimization problems. But whale optimization algorithm (Mirjalili, 2016) proves that it can
67 be used for all optimization problems. A novel nature inspired Meta heuristic optimization algorithm called whale
68 optimization algorithm is used to find the optimal DG size in this paper. To the best knowledge of authors WOA
69 algorithm has not been used in literature of DG placement. WOA algorithm has been modeled based on the unique
70 hunting behavior of humpback whales. The WOA is used to determine the optimal size of DGs at different power
71 factors to reduce the power losses of the distribution system as much as possible and enhancing the voltage profile of
72 the system. IEEE-15, 33-bus, 69-bus, 85-bus and 118-bus systems are examined as test cases with different types of
73 DG units for the objective function.
74 DG types can be characterized (Reddy et al., 2016) as
75 Type I: Injects real power. It operates at unity pf. PV cells, micro turbines, fuel cells are good examples of type I
76 DG.
+Model
P. Dinakara Prasad Reddy et al. / Journal of Electrical Systems and Information Technology xxx (2017) xxx–xxx 3
77 Type II: Injects reactive power. kVAR compensator, synchronous compensator, capacitors, etc. fall under type II DG
78 Type III: Injects both real and reactive power, e.g. synchronous machines (cogeneration, gas turbine, etc.).
79 Type IV: Consuming reactive power but injecting real power, e.g. induction generators in wind farms.
80 2. Problem formulation
81 2.1. Objective function
82 In distribution system more losses are there due to low voltage compared to transmission system. Copper losses
83 predominant in distribution system, this can be calculated as follows

n
84 Ploss = Ii2 Ri (2.1)
i
85 where Ii is current, Ri is resistance and n is number of buses. Objective taken in this paper is real power loss minimization.
86 2.2. Constraints
87 The constraints are

88 • Voltage constraints
89 0.95 ≤ Vi ≤ 1.05 (2.2)
90 • Power balance constraints

N
91 P+ PDG = Pd + Ploss (2.3)
k=1
92 • Upper and lower limits of DG

93 60 ≤ PDG ≤ 3000 (2.4)
94 where the limits are in kW, kVAR and KVA for Type I, II and III DG, respectively.
95 3. Power loss index (PLI) method
96 Power losses and their loss reductions are calculated at each bus by using Eqs. (3.1)–(3.3). Calculate loss reductions
97 at each bus by compensating total reactive load at each bus. Then these are normalized into the range [0, 1]. From these
98 values minimum and maximum loss reductions are obtained.
99 Real and reactive power loss in kth line are given by
(P 2 [j] + Q2 [j]) ∗ Rk
100 PL[j] = (3.1)
(V [j])2
(P 2 [j] + Q2 [j]) ∗ Xk
101 QL[j] = (3.2)
(V [j])2
102 Power loss index (PLI) represent the loss reduction of that particular bus with respect to the maximum and minimum
103 loss reductions in the system so that when the DG units are placed on the buses with high power loss index, maximum
104 loss reduction can be anticipated. This index can be represented by the following equation
LR(b) − LR(min)
105 PLI(b) = (3.3)
LR(max) − LR(min)
106 where net active and reactive power beyond bus j are P, Q. LR(b) is the loss reduction at bus b. LR(min) and LR(max)
107 are the minimum and maximum loss reduction values.
+Model
108 Select the buses with highest PLI values as the best locations. The proposed PLI method is applied to IEEE 15, 33,
109 69, 85 and 118-bus systems. For DG placement the locations are 15, 30, 61, 54 and 113-bus for IEEE 15, 33, 69, 85
110 and 118 bus test systems, respectively.
111 4. Whale optimization algorithm (WOA)
112 Recently a new optimization algorithm called whale optimization algorithm (Mirjalili, 2016) has been introduced
113 to meta heuristic algorithm by Mirjalili and Lewis. The whales are considered to be as highly intelligent animals with
114 motion. The WOA algorithm is inspired by the unique hunting behavior of humpback whales. Usually the humpback
115 whales prefer to hunt krills or small fishes which are close to the surface of sea. Humpback whales uses a special unique
116 hunting method called bubble net feeding method. In this method they swim around the prey and creates a distinctive
117 bubbles along a circle or 9-shaped path. The mathematical model of WOA is described in the following sections
118 1. Encircling prey.

119 2. Bubble net hunting method.
120 3. Search the prey.
121 4.1. Encircling prey
122 WOA algorithm assumes that the current best candidate solution is the target prey. The other search agents try to
123 updates their positions towards best search agent. The behavior is represented by the following equations.
124 + 1) = X∗ (t) − A.

X(t D
(4.1)

125 = C.
D
X∗ (t) − X(t) (4.2)
126 = 2.a.
A r − a (4.3)
127 = 2.r
C (4.4)
128 where X∗ is the position of the best solution and it can be updated if better solution is obtained. X
is the position vector
129 C
and t indicates the current iteration. A, are coefficient vectors. a is linearly decreased from 2 to 0. r is a random
130 vector [0,1].
131 4.2. Bubble net hunting method
132 In this hunting method two approaches are there
133 4.2.1. Shrinking encircling prey

134 is a random value in the interval [−a, a]. Where A
Here A is decreased from 2 to 0. Here A position is setting down
135 at random values in between [−1, 1]. The new position of A is obtained between original position and position of the
136 current best agent. Fig. 1 shows the possible positions from (X, Y) towards (X *, Y*) that can be achieved by 0 ≤ A ≤ 1
137 in a 2D space represented by Eq. (4.3).
138 4.2.2. Spiral position updating

139 A spiral equation is used between position of whales and prey to mimic the helix shaped movement as follows
140 + 1) = D.
X(t ebl . cos(2πl) + X∗ (4.5)

141 where D = X∗ − X(t)
represents the distance between whale to the prey (best solution). b is constant, l is a random
142 number in [−1, 1]
+Model
Fig. 1. Bubble net search shrinking encircling mechanism.
Fig. 2. Bubble net search spiral updating position mechanism.
143 In hunting whales swims around the prey in shrinking circle and along a spiral shaped path simultaneously. Here
144 50% probability is taken between either shrinking encircling or the spiral model to update the position of whales.

D
X∗ (t) − A. ifp < 0.5
145 + 1) =
X(t (4.6)
ebl . cos(2πl) + X∗ ifp ≥ 0.5
D.
146 P is random number [0,1]. Fig. 2 shows the spiral updating position approach represented by Eq. (4.6).
147 4.3. Search for prey
148 To make the algorithm global optimizer when greater than 1 or less than 1 the search agent is updating according
149 to a randomly chosen search agent instead of the best search agent.

150 = C.
D Xrand
−X (4.7)
+Model
Fig. 3. Flowchart of proposed whale optimization algorithm.
151 + 1) = Xrand
X(t D
− A. (4.8)
152 is the random whales in current iteration. The symbol ‘||’ denotes the absolute values. Fig. 3 shows flowchart of
Xrand
153 the proposed algorithm.
+Model
Fig. 4. Single line diagram of 15-bus system.
154 4.4. Implementation of WOA
155 The detailed algorithm is as follows.
156 Step 1: Read line and load data of the system and solve the feeder line flow for the system using load flow method.
157 In this paper branch current load flow method is used.
158 Step 2: Find the best DG locations using the power loss index method.
159 Step 3: Initialize the population/solutions and itmax = 50, number of DG locations d = 1 for, dgmin = 60,dgmax =3000.
Step 4: Generate the population of DG sizes randomly using equation
160
161 population = (dgmax − dgmin ) × rand() + dgmin
162 where dgmin and dgmax are minimum and maximum limits of DG sizes.
163 Step 5: Determine active power loss for generated population by performing load flow.
164 Step 6: Select the DG value with low loss as current best solution.
165 Step 7: Update the position of the whales using Eqs. (4.1) and (4.8).
166 Step 8: Determine the losses for updated population by performing load flow.
167 Step 9: Replace the current best solution with the updated values if obtained losses are less than the current best
168 solution. Otherwise go back to step 7.
169 Step 10: If maximum number of iterations is reached then print the results.
170 5. Simulation results
171 To evaluate the performance of WOA algorithm in the application of DG planning problem IEEE 15-bus, 33-bus,
172 69-bus, 85-bus and 118-bus test systems are simulated using Matlab software. The WOA is used to obtain the optimal
173 size of DG.
174 5.1. IEEE 15-bus system
175 The single line diagram of IEEE 15-bus distribution system (Baran and Wu, 1989) is shown in Fig. 4. The system
176 voltage is 11 kV and total system active and reactive loads are 1226.400 kW and 1251.178 kVAr, respectively. This test
177 system consists of 15 buses and 14 branches. For 15-bus system without installation of DG real, reactive power losses
178 are 61.7933 kW and 57.2969 kVAR respectively.
+Model
Table 1
Results of 15 bus system.
Without DG With Type I With Type II With Type III With Type III
DG (kW) DG (kVAR) DG (kVA) at DG (kVA) at
0.9 pf lag upf pf
DG location – 15 15 15 15
DG size – 673.868 681.688 910.503 673.868
TLP (kW) 61.7933 42.8186 42.2364 28.0484 42.8186
TLQ(kVAR) 57.2969 37.9705 37.3813 22.9392 37.9705
Vmin 0.9445 0.9596 0.9594 0.9705 0.9596
1
Without DG
Type I DG
Type II DG
0.99 Type III DG @ 0.9pf
Type III @ unity pf
Voltages at Buses in per unit
0.98
0.97
0.96
0.95
0.94
0 5 10 15
Bus Number
Fig. 5. Voltage profile of 15-bus system.
179 Table 1 shows the real, reactive power losses and minimum voltages after the placement of different types of DGs.
180 The optimal location for 15 bus test system is 15. The minimum voltage is more in case of type III DG operating at
181 0.9 pf. The losses are also lower with DG type III operating at 0.9 pf when compared to other types of DGs which is
182 shown in Table 1. It is observed from the results that the DG size obtained is higher at lagging power factor compared
183 to the size obtained at unity power factor however the losses are found lower with DGs at lagging power factor rather
184 than DGs at unity power factor. This is due the reason of reactive power available locally for the loads and thereby
185 decrease in the reactive power available from substation. The voltage profile also improves with DGs at lagging power
186 factor and it is observed from Fig. 5. The minimum voltage obtained for the system is better compared to the voltage
187 obtained with DGs at unity power factor. Thus, it is essential to consider the reactive power available from DGs for
188 its size calculations and its impact on losses reduction and voltage profile improvement. The results obtained with
189 consideration of reactive power are better than the results obtained with DGs at unity power factor.
190 5.2. IEEE 33-bus system
191 The single line diagram of IEEE 33-bus distribution system (Baran and Wu, 1989) is shown in Fig. 6. The system
192 voltage is 12.66 kV and total system active and reactive loads are 3715 kW and 2300 kVAr, respectively. This test
193 system consists of 33 buses and 32 branches. For 33-bus system without installation of DG real, reactive power losses
194 are 210.99 kW and 143.0324 kVAR, respectively.
196 Tables 3 and 4 show comparison of results with type III DG operating at 0.9 pf and unity pf respectively. The optimal
197 location for 33 bus system is 30. The minimum voltage is more in case of type III DG operating at 0.9 pf. From Table 2
+Model
Table 2
0.9 pf lag upf pf
DG size – 1542.67 1258 1940.33 1542.67
TLP (kW) 210.9974 125.161 151.379 78.4337 125.161
TLQ (kVAR) 143.032 89.2867 103.82 58.969 89.2867
Vmin 0.9038 0.9272 0.9165 0.9386 0.9272
Table 3
Comparison of results with DG operating at 0.9 pf.
With DG (Murty and Kumar, 2014)
CPLS method IV method VSI method Proposed method
DG location 8 30 16 30
DG size 2100 1950 1200 1940.33
TLP (kW) 84.5 78.4 112.8 78.4
TLQ (kVAR) 62.1 58.9 77.4 58.9
Vmin 0.9534 0.9391 0.9378 0.9386
198 it is inferred that by using DG type III operating at 0.9 pf the losses are reduced more when compared to other type of
199 DGs. It is observed from the results that the DG size obtained is higher at lagging power factor compared to the size
200 obtained at unity power factor, however, the losses are found lower with DGs at lagging power factor rather than DGs
201 at unity power factor. This is due the reason of reactive power available locally for the loads and thereby decrease in the
202 reactive power available from substation. The voltage profile also improves with DGs at lagging power factor and it is
203 observed from the Fig. 7. The minimum voltage obtained for the system is better compared to the voltage obtained with
204 DGs at unity power factor. Thus, it is essential to consider the reactive power available from DGs for its size calculations
+Model
Table 4
Comparison of results with DG operating at unity pf.
CPLS method IV method VSI method Proposed method
DG location 8 30 16 30
DG size 1800 1550 1000 1542.67
TLP (kW) 118.1 125.2 136.8 125.2
TLQ(kVAR) 82.9 89.3 92.7 89.3
Vmin 0.9433 0.9275 0.9318 0.9272
0.99
0.98
0.97
0.96
0.95
0.94
0.93
Without DG
0.92 Type I DG
Type II DG
Type III DG @ 0.9 pf
0.91 Type III DG @ unity pf
0.9
0 5 10 15 20 25 30
Fig. 7. Voltage profile 33-bus system.
205 and its impact on losses reduction and voltage profile improvement. The results obtained with consideration of reactive
206 power are better than the results obtained with DGs at unity power factor. When comparison (Murty and Kumar, 2014)
207 with combined power loss sensitivity based(CPLS) and voltage sensitivity index(VSI) methods proposed method gives
208 better results as shown in Tables 3 and 4.
209 5.3. IEEE 69 bus system
210 The IEEE 69-bus distribution system with 12.66-kV base voltage (Baran and Wu, 1989) is shown in Fig. 8 is
211 employed in this paper. It consists of one slack bus and 68 load buses. The total real and reactive power demand is
212 3802.190 kW and 2694.600 kVAR, respectively. This test system consists of 69 buses and 68 branches. For 69-bus
213 system without installation of DG real, reactive power losses are 225.023 kW and 102.1763 kVAR respectively.
215 The optimal location for 69 bus system is 61. The minimum voltage is more in case of type III DG operating at 0.9 pf.
216 From Table 5 it is inferred that by using DG type III operating at 0.9 pf the losses are reduced more when compared
217 to other type of DGs. From the results it is observed that the DG size is higher at lagging power factor compared to
218 the size obtained at unity power factor however the losses are found lower with DGs at lagging power factor rather
219 than DGs at unity power factor. This is because of reactive power available locally for the loads and thereby decrease
220 in the reactive power available from substation. The voltage profile also improves with DGs at lagging power factor
221 and it is observed from the Fig. 9. The minimum voltage obtained for the system is better compared to the voltage
222 obtained with DGs at unity power factor. Thus, it is essential to consider the reactive power available from DGs for
223 its size calculations and its impact on losses reduction and voltage profile improvement. The results obtained with
224 consideration of reactive power are better than the results obtained with DGs at unity power factor. When comparison
+Model
Table 5
0.9 pf lag upf pf
DG size – 1872.82 1329.99 2217.39 1872.82
TLP(kW) 225.023 83.2279 152.064 27.9649 83.2279
TLQ (kVAR) 102.176 40.5381 70.5143 16.4606 40.5381
Vmin 0.9092 0.9683 0.9307 0.9724 0.9683
1.02
1
Voltage at buses in per unit
0.98
0.96
0.94
WIthout DG
Type I DG
0.92 Type II DG
Type III DG @ 0.9 pf
Type III DG @ unity pf
0.9
0 10 20 30 40 50 60
Bus Number
+Model
Table 6
Comparison of results with DG operating at 0.9 pf.
MN method CPLS method IV method VSI method Proposed method
DG location bus 61 61 61 65 61
DG size 2236.704 2200 2200 1750 2217.387
TLP (kW) 27.4 27.9 27.9 65.5 27.9
TLQ (kVAR) 16.2 16.5 16.5 35.6 16.5
Vmin 0.9728 0.9726 0.9726 0.9693 0.9724
Table 7
Comparison of results with DG operating at unity pf.
MN method CPLS method IV method VSI method Proposed method
DG location 61 61 61 65 61
DG size 1832.536 1850 1850 1450 1872.823
TLP (kW) 83.2 83.2 83.2 112.0 83.2
TLQ (kVAR) 40.6 40.5 40.5 55.1 40.5
Vmin 0.9685 0.9685 0.9685 0.966 0.9683
Table 8
0.9 pf lag upf pf
DG location bus – 54 54 54 54
DG size – 910.075 837.285 1243.18 910.075
TLP (kW) 315.7 227.105 232.136 162.329 227.105
TLQ (kVAR) 198.356 138.249 142.145 94.0364 138.249
Vmin 0.8714 0.9101 0.9023 0.9242 0.9101
225 (Murty and Kumar, 2014) with voltage sensitivity index(VSI) method proposed method gives better results as shown
226 in Tables 6 and 7.
228 The IEEE 85-bus distribution system with 12.66-kV base voltage (Baran and Wu, 1989) is shown in Fig. 10 is
229 employed in this paper. It consists of one slack bus and 84 load buses. This test system consists of 85 buses and 84
230 branches. For 85-bus system without installation of DG real, reactive power losses are 315.6996 kW and 198.3560 kVAR
231 respectively.
233 The optimal location for 85 bus test system is 54. The minimum voltage is more in case of type III DG operating at
234 0.9 pf. It is observed from the results that the DG size obtained is higher at lagging power factor compared to the size
235 obtained at unity power factor however the losses are found lower with DGs at lagging power factor rather than DGs
236 at unity power factor. This is due the reason of reactive power available locally for the loads and thereby decrease in
237 the reactive power available from substation.
238 Voltage profiles of the IEEE 85-bus system with and without placement of different types of DGs are shown in
239 Fig. 11. From figure it is clear that the type III DG operating at 0.9 pf have better voltage profile improvement.
+Model
1
Without DG
Type I DG
Type II DG
0.98 Type III DG @ 0.9 pf
Type III DG @ unity pf
Voltage at Buses in per unit
0.96
0.94
0.92
0.9
0.88
0.86
0 10 20 30 40 50 60 70 80
Bus Number
+Model
Table 9
0.9 pf lag upf pf
DG location – 113 113 113 113

DG size – 2704 2198 3357 2704
TLP (kW) 1298 1092.46 1156 986.69 1092.46
TLQ(kVAR) 976 862.69 899.55 805.1222 862.69
241 The IEEE 118-bus distribution system with 11-kV base voltage (Zhang et al., 2007) is shown in Fig. 12 is employed
242 in this paper. It consists of one slack bus and 117 load buses. This test system consists of 118 buses and 117 branches.
243 For 118-bus system without installation of DG real, reactive power losses are 1298 kW and 976 kVAR respectively.
244 Table 9 shows the real, reactive power losses after the placement of different types of DGs. The optimal location for
245 118 bus test system is 113. It is observed from the results that the DG size obtained is higher at lagging power factor
246 compared to the size obtained at unity power factor however the losses are found lower with DGs at lagging power
247 factor rather than DGs at unity power factor. This is due the reason of reactive power available locally for the loads
248 and thereby decrease in the reactive power available from substation.
+Model
Fig. 13. Convergence characteristics.
249 Fig. 13 shows convergence characteristics of IEEE 15, 33, 69, 85 and 118 buses with 0.9 pf.The characteristics
250 shows that the WOA algorithm converged faster. Hence WOA algorithm is efficient, robust and capable of handling
251 mixed integer nonlinear optimization problems.
252 6. Conclusions
253 A novel nature-inspired algorithm called Whale optimization algorithm (WOA) is used to determine the optimal
254 DG-unit’s size in this paper. WOA modeled based on the unique hunting behavior of humpback whales. Reduction in
255 system real power losses and improvement in voltage profile of the system buses are the main objectives in this paper.
256 The proposed optimization technique has been applied on typical IEEE 15, 33, 69, 85 and 118-bus radial distribution
257 systems with different types of DGs and compared with other algorithms. Better results has been achieved with WOA
258 algorithm when compared other algorithms. The simulation results indicated that the overall impact of the DG units on
259 voltage profile is positive and proportionate reduction in power losses is achieved. It can be interfered that best results
260 can be achieved with type III DG operating at 0.9 pf. Because it generates both real power and reactive power. The
261 results show that the WOA algorithm is efficient, robust, and capable of handling mixed integer nonlinear optimization
262 problems.
263Q2 Uncited references
264 ChithraDevi et al. (2016), Devi (2014), Qian et al. (2011) and Murthy and Kumar (2013).
265 Acknowledgment
266 The authors are thankful to the authorities of Sri Venkateswara University, Tirupati 517502, India, for providing all
267 the facilities to do research work.
268 References
269 Abu-Mouti, F.S., El-Hawary, M.E., 2011. Optimal distributed generation allocation and sizing in distribution systems via artificial bee colony
270 algorithm. IEEE Trans. Power Deliv. 26, 2090–2101, 95.
+Model
271 Acharya, N., Mahat, P., Mithulananthan, N., 2006. An analytical approach for dg allocation in primary distribution network. Int. J. Electr. Power
272 Energy Syst. 28, 669–678.
273 Ackermann, T., Andersson, G., Soder, L., 2001. Distributed generation: a definition. Electr. Power Syst. Res. 57, 195–204, 519.
274 Alonso, M., Amaris, H., Alvarez-Ortega, C., 2012. Integration of renewable energy sources in smart grids by means of evolutionary optimization
275 algorithms. Expert Syst. Appl. 39, 5513–5522.
276 Aman, M., Jasmon, G., Mokhlis, H., Bakar, A., 2012. Optimal placement and sizing of a dg based on a new power stability index and line losses.
277 Int. J. Electr. Power Energy Syst. 43, 1296–1304.
278 Ameli, A., Bahrami, S., Khazaeli, F., Haghifam, M.R., 2014. A multiobjective particle swarm optimization for sizing and placement of dgs from dg
279 owner’s and distribution company’s viewpoints. IEEE Trans. Power Deliv. 29, 1831–1840.
280 Baran, M.E., Wu, F.F., 1989. Optimal sizing of capacitors placed on a radial-distribution system. IEEE Trans. Power Deliv. 4, 735–743.
281 ChithraDevi, S., Lakshminarasimman, L., Balamurugan, R., 2016. Stud krill herd algorithm for multiple dg placement and sizing in a radial
282 distribution system. Eng. Sci. Technol., Int. J.
283 Das, B., Mukherjee, V., Das, D., 2016. Dg placement in radial distribution network by symbiotic organisms search algorithm for real power loss
284 minimization. Appl. Soft Comput. 49, 920–936.
285 Devi, S., G.M., 2014. Application of modified bacterial foraging optimization algorithm for optimal placement and sizing of distributed generation.
286 Expert Syst. Appl. 41, 2772–2781.
287 Doagou-Mojarrad, H., Gharehpetian, G., Rastegar, H., Olamaei, J., 2013a. Optimal placement and sizing of {DG} (distributed generation) units in
288 distribution networks by novel hybrid evolutionary algorithm. Energy 54, 129–138.
289 Doagou-Mojarrad, H., Gharehpetian, G.B., Rastegar, H., Olamaei, J., 2013b. Optimal placement and sizing of dg (distributed generation) units in
290 distribution networks by novel hybrid evolutionary algorithm. Energy 54, 129–138, 13.
291 Duong Quoc, H., Mithulananthan, N., Bansal, R.C., 2010. Analytical expressions for dg allocation in primary distribution networks. IEEE Trans.
292 Power Deliv. Trans. Energy Convers. 25, 814–820.
293 El-Fergany, A., 2015. Study impact of various load models on dg placement and sizing using backtracking search algorithm. Appl. Soft Comput.
294 30, 803–811.
295 Esmaeilian, H.R., Fadaeinedjad, R., 2015. Energy loss minimization in distribution systems utilizing an enhanced reconfiguration method integrating
296 distributed generation. IEEE Syst. J. 9, 1430–1439.
297 Gampa, S.R., Das, D., 2015. Optimum placement and sizing of dgs considering average hourly variations of load. Int. J. Electr. Power Energy Syst.
298 66, 25–40.
299 Golshannavaz, S., 2014. Optimal simultaneous siting and sizing of dgs and capacitors considering reconfiguration in smart automated distribution
300 systems. J. Intell. Fuzzy Syst. 27, 1719–1729.
301 Hegazy, Y.G., Othman, M.M., El-Khattam, W., Abdelaziz, A.Y., 2014. Optimal sizing and siting of distributed generators using big bang big crunch
302 method. In: Power Engineering Conference (UPEC), 2014 49th International Universities, pp. 1–6.
303 Hosseini, S.A., Madahi, S.S.K., Razavi, F., Karami, M., Ghadimi, A.A., 2013. Optimal sizing and siting distributed generation resources using a
304 multiobjective algorithm. Turk. J. Electr. Eng. Comput. Sci. 21, 825–850, 4.
305 Kaur, S., Kumbhar, G., Sharma, J., 2014. A minlp technique for optimal placement of multiple dg units in distribution systems. Int. J. Electr. Power
306 Energy Syst. 63, 609–617.
307 Kaveh, A., Ghazaan, M.I., 2016. Enhanced whale optimization algorithm for sizing optimization of skeletal structures. Mech. Based Des. Struct.
308 Mach. 0, 1–18, arXiv:http://dx.doi.org/10.1080/15397734.2016.1213639.
309 Khodabakhshian, A., Andishgar, M.H., 2016. Simultaneous placement and sizing of dgs and shunt capacitors in distribution systems by using IMDE
310 algorithm. Int. J. Electr. Power Energy Syst. 82, 599–607.
311 Kolenc, M., Papic, I., Blazic, B., 2015. Assessment of maximum distributed generation penetration levels in low voltage networks using a probabilistic
312 approach. Int. J. Electr. Power Energy Syst. 64, 505–515.
313 Mirjalili, S., 2016. The whale optimization algorithm. Adv. Eng. Softw. 95, 51–67.
314 Moradi, M.H., A.M., 2012. A combination of genetic algorithm and particle swarm optimization for optimal dg location and sizing in distribution
315 systems. Int. J. Electr. Power Energy Syst. 33, 66–74.
316 Murthy, V.V.S.N., Kumar, A., 2013. Comparison of optimal dg allocation methods in radial distribution systems based on sensitivity approaches.
317 Int. J. Electr. Power Energy Syst. 53, 450–467.
318 Murty, V.V.S.N., Kumar, A., 2014. Mesh distribution system analysis in presence of distributed generation with time varying load model. Int. J.
319 Electr. Power Energy Syst. 62, 836–854.
320 Naik, S.G., Khatod, D.K., Sharma, M.P., 2013. Optimal allocation of combined dg and capacitor for real power loss minimization in distribution
321 networks. Int. J. Electr. Power Energy Syst. 53, 967–973.
322 Nekooei, K., Farsangi, M.M., Nezamabadi-Pour, H., Lee, K.Y., 2013. An improved multi-objective harmony search for optimal placement of dgs
323 in distribution systems. IEEE Trans. Smart Grid 4, 557–567.
324 Prakash, D.B., Lakshminarayana, C., 2016. Optimal siting of capacitors in radial distribution network using whale optimization algorithm. Alexandria
325 Eng. J.
326 Qian, K., Zhou, C., Allan, M., Yuan, Y., 2011. Effect of load models on assessment of energy losses in distribution generation planning. Electr.
327 Power Res. 2, 1243–1250.
328 Rahim, S.R.A., Musirin, I., Othman, M.M., Hussain, M.H., Sulaiman, M.H., Azmi, A., 2013. Effect of population size for dg installation using
329 EMEFA. In: Power Engineering and Optimization Conference (PEOCO), 2013 IEEE 7th International, pp. 746–751.
330 Rahim, S.R.A., Musirin, I., Sulaiman, M.H., Hussain, M.H., Azmi, A., 2012. Assessing the performance of dg in distribution network. In: Power
331 Engineering and Optimization Conference (PEDCO) Melaka, Malaysia, 2012 IEEE International, pp. 436–441.
332 Rahmani-andebili, M., 2016. Simultaneous placement of dg and capacitor in distribution network. Electr. Power Syst. Res. 131, 1–10.
+Model
333 Reddy, P.D.P., Reddy, V.V., Manohar, T.G., 2016. Application of flower pollination algorithm for optimal placement and sizing of distributed
334 generation in distribution systems. J. Electr. Syst. Inform. Technol. 3, 14–22.
335 Singh, A., Parida, S., 2016. Novel sensitivity factors for dg placement based on loss reduction and voltage improvement. Int. J. Electr. Power Energy
336 Syst. 74, 453–456.
337 Su, C.L., 2010. Stochastic evaluation of voltages in distribution networks with distributed generation using detailed distribution operation models.
338 IEEE Trans. Power Syst. 25, 786–795.
339 Sultana, U., Khairuddin, A.B., Aman, M., Mokhtar, A., Zareen, N., 2016. A review of optimum dg placement based on minimization of power
340 losses and voltage stability enhancement of distribution system. Renew. Sustain. Energy Rev. 63, 363–378.
341 Zangiabadi, M., Feuillet, R., Lesani, H., Hadj-Said, N., Kvaloy, J.T., 2011. Assessing the performance and benefits of customer distributed generation
342 developers under uncertainties. Energy 36, 1703–1712.
343 Zeinalzadeh, A., Mohammadi, Y., Moradi, M.H., 2015. Optimal multi objective placement and sizing of multiple dgs and shunt capacitor banks
344 simultaneously considering load uncertainty via MOPSO approach. Int. J. Electr. Power Energy Syst. 67, 336–349.
345 Zhang, D., Fu, Z., Zhang, L., 2007. An improved TS algorithm for loss-minimum reconfiguration in large-scale distribution systems. Electr. Power
346 Syst. Res. 77, 685–694, http://dx.doi.org/10.1016/j.epsr.2006.06.005, pii: S0378779606001477.
347 P. Dinakara Prasad Reddy received his B.Tech degree in Electrical and Electronics Engineering from JNT University
348 Hyderabad, India, in 2007. He received his M.Tech. degree in 2011 from Sri Venkateswara University and is currently a
349 Ph.D. candidate in the Department of Electrical and Electronics Engineering at the JNT University, kakinada. His research
350 interests are in the areas of Soft computing techniques, Artificial neural networks, FACTS, Nature inspired algorithms,
351 Economic load dispatch, Renewable energy sources.
352 V.C. Veera Reddy received his B.Tech degree in Electrical and Electronics Engineering from JNT University, India, in
353 1979. He received his M.Tech, Ph.D. degrees in 1981 and 1999 from Sri Venkateswara University, Tirupati, India. He has
354 32 years of experience in teaching. His current research areas include power quality issues, FACTS technology, distributed
355 energy resources interface and control and application of artificial intelligent techniques on power systems.
356 T. Gowri Manohar received his B.Tech, M.Tech and Ph.D. degrees from S.V. University, Tirupati, India, in 1996, 1998 and
357 2007 respectively. Currently he is working as professor in department of EEE, SV University, Tirupati. His current research
358 areas include power quality issues, FACTS technology, distributed energy resources and application of artificial intelligent
359 techniques on power systems.

Article in Press

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Article in Press

Uploaded by

Copyright:

Available Formats

+Model

JESIT 170 1–17 ARTICLE IN PRESS

1 Optimal renewable resources placement in distribution networks by

81 2.1. Objective function

87 The constraints are

92 • Upper and lower limits of DG

95 3. Power loss index (PLI) method

111 4. Whale optimization algorithm (WOA)

118 1. Encircling prey.

121 4.1. Encircling prey

124 + 1) = X∗ (t) − A.

131 4.2. Bubble net hunting method

132 In this hunting method two approaches are there

133 4.2.1. Shrinking encircling prey

138 4.2.2. Spiral position updating

Fig. 1. Bubble net search shrinking encircling mechanism.

Fig. 2. Bubble net search spiral updating position mechanism.

147 4.3. Search for prey

Fig. 3. Flowchart of proposed whale optimization algorithm.

Fig. 4. Single line diagram of 15-bus system.

154 4.4. Implementation of WOA

155 The detailed algorithm is as follows.

161 population = (dgmax − dgmin ) × rand() + dgmin

170 5. Simulation results

174 5.1. IEEE 15-bus system

Fig. 5. Voltage profile of 15-bus system.

190 5.2. IEEE 33-bus system

Fig. 6. Single line diagram of 33-bus system.

CPLS method IV method VSI method Proposed method

CPLS method IV method VSI method Proposed method

Fig. 7. Voltage profile 33-bus system.

209 5.3. IEEE 69 bus system

Fig. 8. Single line diagram of 69-bus system.

Fig. 9. Voltage profile of 69-bus system.

MN method CPLS method IV method VSI method Proposed method

MN method CPLS method IV method VSI method Proposed method

227 5.4. IEEE 85 bus system

Fig. 10. Single line diagram of 85-bus system.

Fig. 11. Voltage profile of 85-bus system.

Fig. 12. Single line diagram of 118-bus system.

DG location – 113 113 113 113

240 5.5. IEEE 118 bus system

Fig. 13. Convergence characteristics.

263Q2 Uncited references

You might also like