Published by : International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
Vol. 9 Issue 07, July-2020
Optimal Allocation and Size Selection of
Dispersed Generation in Radial Distribution
System
Aditya Prasad Padhy1 Manisha Bhatt2
Department of Electrical Engineering Department of Electrical Engineering
Kalinga University, Kalinga University,
Raipur, Chhattisgarh, India Raipur, chhattisgarh, India
Abstract—This paper contributes an analytical technique for the substituted by the available methods for load flow of
allocation and size selection of dispersed generations (DGs) in distribution system as backward forward [4] and direct
radial distributed system (DS)). The proposed technique is approach methods [5]. These methods have the advantages of
computationally expeditious as compared to other existing quick convergence and less computation time.
techniques. The prime objective of this method is to improve
voltage profile at each node and reduce total active power loss in
In conventional system, renewable energy invasion has
33- bus radial distribution system (RDS). Voltage stability increased in earlier years due to environmental concerns. The
indicator (VSI) is used to find heavily loaded bus in the system. conventional generation resources such as thermal, hydro,
After finding the heavily loaded bus, it is set as optimal location nuclear, etc. are being penetrated by small-scale generations
for placement of DG. At this optimal location, the size of DG is like photovoltaic (PV), wind, and fuel cell. Although these
determined by using continuous increment of step size (CISS) DGs have very less capacity (Range: 1 KW to 50MW )
technique. The effectiveness of the proposed technique is compared to other conventional resources, availability of
validated from simulated results, by comparing voltage profile, these resources minimize power loss, improves voltage
branch current and power losses with the presence and absence profile [6]. Therefore, DGs are required for optimal sizing
of DG.
and proper location.
Keywords: Radial distribution system, dispersed generation,
optimal allocation, voltage stability indicator, continous increment Many researchers have found proper solution of optimal
of step size. sizing and location issues of DGs in DS. Genetic algorithm
(GA) is implemented to solve the metaheurstic optimization
I. INTRODUCTION problem for location and size selection of DG [7]. The
technique used in [8] minimizes the fitness function using
In recent era, Distribution system (DS) is added as an integral GA for power loss minimization. In general, constant power
part of power distribution system [1]. The indispensible models are considered in the DS. But in [9], different load
growth of power demand has increased the role and models i.e. residential, commercial and industrial are carried
importance of DS. To encounter the load demand effectively, out in RDS while optimally allocating and sizing DGs using
several changes occur in DS resulting in establishment of GA. A hybridized GA and simulated annealing (SA)
complex distribution structure. One of the most important algorithm is implemented to compute optimal location of DG
changes in DS structure is invasion of renewable energy in in [10]. A mixed GA-particle swarm optimization (GA-PSO)
the system. These invasions are known as dispersed algorithm is mentioned in [11] to find out the optimal
generations (DGs). The involvement of DGs in the system location and sizing of DG. In [12, 13], binary PSO and fuzzy
affect DS such as: minimization of power loss, improvement embedded GA techniques are proposed for optimal DS
of node voltage, and reliability [2]. Thus, DS in accordance planning in presence of DGs. The problem of optimal DS
with DG is the most prominent part of research domain. planning with DGs in [14] is treated by means of multi-
objective PSO (MOPSO). A MOPSO approach is employed
Most of the DS structures are radial in nature in which power to deal with problem considering different load models in
flows in a single way from the distribution substation to the [15]. Artificial bee colony (ABC) algorithm is employed to
consumer [3]. Radial distribution system (RDS) has high determine the optimal allocation, size and power factor in
R / X ratio, radial structure, and unbalance conditions of order to minimize active power loss [16]. Bacterial foraging
load where R and X are the resistance and reactance of line algorithm (BFA) is implemented to find the optimal size of
respectively. High R / X ratio leads to high power loss and DG [17]. In [18], backtracking search algorithm (BSA) is
more voltage drops. In RDS nodes, sudden voltage drops used to assign DGs in DS. An immune algorithm (IA) is
occurs under critical loading conditions. In transmission formulated to solve the optimal DG planning problem in
system, load flow by traditional methodologies has serious smart grid [19]. Modified teaching-learning based algorithm
convergence problems due to the high R / X ratio of the (MTLBA) is proposed in [20] to compute the optimal
RDS. For proper RDS planning, an efficient load flow allocation and size of the DG units in RDS.
technique is required. Hence, these methodologies are
IJERTV9IS070036 www.ijert.org 145
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by : International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
Vol. 9 Issue 07, July-2020
Although numerous techniques are available for optimal tb
location and sizing of DG, but traditional techniques are still Total power loss PTL = PL ( , + 1) (7)
a current interest. Thus, this article presented a voltage =1
stability indicator (VSI) which indicates the heavily loaded where, “tb” represents total number of buses.
bus. This bus signifies the optimal location of DG placement.
After proper allocation of DG, the size of DG is found out by III. PROPOSED METHODOLOGY
continuous increment of step size (CISS) technique. CISS
technique is applied to monitor active and reactive power loss In distribution system, the node voltage plays a important role
in the IEEE-33bus system. to maintain good voltage regulation. Ideally, voltage
regulation should be nearly equal to zero, but due to line
The paper is organized as follows: In section 2, presents load resistance and reactance there are slight drop in voltage in
flow equations and power loss computation for 33 bus RDS. different nodes of the RDS. To improve voltage profile and
Proposed methodology for optimal allocation and size reduce line losses, different DGs are used at different power
selection of DG are explained in Section 3, Section 4 factor [21] according to the requirement as shown in table
discusses the 33 bus test system and simulation results. A. Voltage stability node indicator for dg placement
Finally, conclusion is outlined in section 5.
To identify the most sensitive or heavily loaded bus in the
II. PROBLEM FORMULATION system, voltage stability node indicator (VSNI) is used.
From figure 1 branch current is given by
A. Radial distribution load flow
In [5] author proposed a simplified and efficient technique ( P )2 + ( Q )2
( I ) = +1 +1
2
for radial distribution system (RDS). A simplified RDS (8)
(V +1 ) 2
shown in Fig. 1.
The equivalent load current injection, corresponding to power
injection equations Active and reactive power loss of the branch between two
I = ( P +1 + jQ +1 V +1 ) = 1,2,3, .
* nodes is computed by using following equation
(1)
For the branch connecting buses 1 to 4, the branch currents ( P )2 + ( Q )2
calculated using KCL as PL = r +1 +1
(9)
G1 = I 2 + I 3 + I 4 ( +1 )
V
2
G2 = I 3 + I 4 (2)
G3 = I 4 TABLE-1 DG SCENARIO
Branch injection Matrix (BIM) becomes Types of Power factor Power injection Examples
G1 1 1 1 I 2 DG (P.F) ability
G = 0 1 1 I
2 3 (3) A P.FDG=0 Only reactive power Synchronous
G3 0 0 1 I 4 compensator
B P.FDG=1 Only active power Photovoltaic
system
Bus 1 Bus 2 Bus 3 Bus 4
V1 V2 V3 V4
C 0<P.FDG<1 Both Active and Synchronous
G1 G2 G3 reactive power generator
D 0<P.FDG<1 Active power and Wind turbine
consuming reactive
Substation P2 + jQ2 P3 + jQ3 P4 + jQ4 power
PLoad2 + jQLoad2 PLoad3 + jQLoad3 PLoad4 + jQLoad4
Fig. 1. Single line diagram of radial distribution system ( P )2 + ( Q )2
In general form, PL = x +1 +1
(10)
(V +1 ) 2
G = BIM I (4)
Node voltages at different buses calculated by using KVL
From the above (9) and (10) the developed VSNI is given by
V +1 = V − G Z , +1 (5)
Power loss in between the buses and + 1 is calculated as
PL ( , + 1) = ( I ) * R
2
(6) VSNI =
4
(( P
+1 ) + ( Q +1 )
2 2
) 1 (11)
The total power loss of the RDS is computed by the (V +1 )2
summation of all the branch losses, which is represented by
IJERTV9IS070036 www.ijert.org 146
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by : International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
Vol. 9 Issue 07, July-2020
From (11) it can be cleared that 0 VSNI 1 . Further, when
the value of the VSNI approaches to 1.0, the system will Start
become its unstable. Similarly, the values away from zero
indicates improved stability of the system.
B. Optimal size selection using Continuous increment of step Read load and line data
size (CISS)
Once the optimal location of DG is fixed, then DG size
change from 0 to 1 p.u. of the total load. As the DG size
increases, total active power is injected to the system
increases. This injected power minimizes the total active
Run Load flow
power loss. A parabolic curve is formed between DG size and
total active power loss. This curve indicates, first the losses of
the system decreases till it reaches the optimum point.
Thereafter, system losses suddenly increases. Thus, step size
plays factor while selecting DG size. In present scenario, the Compute VSNI for each bus and choose
step size chosen is 0.1 Mw. For exact DG size, step size must
maximum value
be as small as possible, but the simulation time increases.
19 20 21 22
Fix the location of DG at Maximum value of
VSNI bus
Sub-
Station
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Initially set the DG value PDG =PLoad
and rerun the load flow compute PTL
Save(k)th
Save(k)thiteration powerloss
iteration power loss
as as
26 27 28 29 30 31 32 33 P
PTLTL (k)th
(k)th
23 24 25
Fig. 2. Single line diagram of 33-bus radial distribution system
If No
PTL (k)th < PTL(k+1)th
Yes
Set PDG as optimal size
Stop
Fig. 3. Flow chart for proposed methodology
IJERTV9IS070036 www.ijert.org 147
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by : International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
Vol. 9 Issue 07, July-2020
IV. RESULT AND DISCUSSIONS
The Fig. 3 shows the single line diagram of 33 bus radial Table 2 VSNI table
distribution system. The detailed methodology of the Bus Real Power Reactive power Voltage
proposed method is described by flow chart as shown in Fig. No Demand Demand Stability
2. The proposed flow chart described in two steps. At First, (Kw) (Kvar) Node
optimal location of DGs are calculated by using VSNI. The Indicator
tabulated values of VSIN are shown in table 1. The table 1 (VSI)
indicates the active and reactive power consumed by different 1 0 0 0
buses and their corresponding VSNI values. In next step, the
2 100 60 0.0316
optimal size of the DGs are determined by using continuous
increment of step size (CISS) technique with a step size of 3 90 40 0.0304
0.1Mw. 4 120 80 0.0361
5 60 30 0.0259
A. Case A (DG placement at bus no: 25) 6 60 20 0.0265
In this case DGs are connected at bus no 25. From the Fig. 4 7 200 100 0.0501
it is noticed that, active power loss of the system increases, 8 200 100 0.0547
when the DG size increases beyond an optimum point. 9 60 20 0.0286
Variation of Active Power Loss with DG size ) 10 60 20 0.0290
183.5
Active Power loss 11 45 30 0.0251
183
12 60 35 0.0291
182.5
13 60 35 0.0319
182 14 120 80 0.0430
Ploss in (kw)
181.5 15 60 10 0.0296
181
16 60 20 0.0300
17 60 20 0.0312
180.5
18 90 40 0.0372
180
19 90 40 0.0367
179.5
20 90 40 0.0324
179
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4
21 90 40 0.0307
DG size in (Mw)
22 90 40 0.0313
Fig. 4. Power loss curve with step increment of DG size at bus no. 25 23 90 50 0.0308
Optimal point is calculated based on values of voltage 24 420 200 0.0694
stability node indicator, which is shown in Table. 2 25 420 200 0.0704
Line No Vs Line Loss 26 60 25 0.0261
60
NDG 27 60 25 0.0273
DG
28 60 20 0.0280
50
29 120 70 0.0411
40
30 200 600 0.0631
31 150 70 0.0473
Line Loss(Kw)
30 32 210 100 0.0546
33 60 40 0.0296
20
B. Case B (DG placement at bus no 24)
10
From the VSNI table, it is clear that bus no 24 is the most
critical or weakest bus of the system. This weakest bus
0
0 5 10 15 20 25 30 35 identified as the optimal DG allocation point. Further, DG
Line No
size is computed by using CISS technique.
Fig. 5. Line loss with and without placement of DG at bus no. 25
IJERTV9IS070036 www.ijert.org 148
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by : International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
Vol. 9 Issue 07, July-2020
Variation of Active Power Loss with DG size ) Variation of Active Power Loss with DG size )
183 170
Active Power loss Active Power loss
182.5 165
182 160
181.5 155
Ploss in (Kw)
Ploss in (Kw)
181 150
180.5 145
180 140
179.5 135
179 130
0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4
DG size in (Mw)
125
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 2.2
DG size in (Mw)
Fig. 6. Power loss curve with step increment of DG size at bus no. 24
Fig. 8. Power loss curve with step increment of DG size at bus no. 30
Line No Vs Line Loss
60 Line No Vs Line Loss
NDG 60
DG NDG
50 DG
50
40
40
Line Loss(Kw)
Line Loss(Kw)
30
30
20
20
10
10
0
0 5 10 15 20 25 30 35
Line No 0
0 5 10 15 20 25 30 35
Line No
Fig. 7. Line loss with and without placement of DG at bus no. 24 Fig. 9. Line loss with and without placement of DG at bus no. 30
Bus No Vs Bus Voltage
C. Case C (DG placement at bus no: 30) 1
To examine the efficacy of the proposed technique, it is NDG
0.99 DG at bus 25
applied to 12.66 kV, 3.72 Mw and 2.3 Mvar RDS consisting DG at bus 24
of 33 buses. Further, after application of load flow the real 0.98 DG at bus 30
power losses incurred by the system before placement of DGs 0.97
is 210 kW. To reduce the total power loss of the system DGs
Bus voltage(P.U)
are placed at different locations as shown in Table. 3. The 0.96
optimal allocation point is computed by using (11). After 0.95
computing VSIN at all buses, the three locations i.e. bus no
0.94
25, 24 and 30 identified as the most critical bus. To determine
the optimum size, CISS technique is applied on all the critical 0.93
buses.
0.92
From Fig. 4, 6 and 8, shows the curve between DG size and
total power loss. It is basically follow a parabolic curve, first 0.91
the total loss decreases and then increases. While selecting 0.9
DG size, appropriate step size should be chosen carefully. 0 5 10 15 20 25 30 35
Bus No
DG size should not exceed the optimum value, above
optimum value the total active power loss of the system Fig. 10. Comparison of voltage scenarios at different optimal locations.
increases. As a result, system operates in poor voltage
regulation and efficiency.
IJERTV9IS070036 www.ijert.org 149
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
Published by : International Journal of Engineering Research & Technology (IJERT)
http://www.ijert.org ISSN: 2278-0181
Vol. 9 Issue 07, July-2020
Table 3 Active power loss scenario after placement of DG at [12] A. Soroudi and M. Afrasiab, "Binary PSO-based dynamic multi-
objective model for distributed generation planning under
different optimal locations
uncertainty," IET renewable power generation, vol. 6, pp. 67-78,
Active power Active power loss Optimal DG size 2012.
loss before after DG Locations (Bus (Mw) [13] K. Vinothkumar and M. Selvan, "Fuzzy embedded genetic algorithm
DG no) method for distributed generation planning," Electric Power
Components and Systems, vol. 39, pp. 346-366, 2011.
210.7929 179.2400 25 0.920 [14] S. Ganguly, N. Sahoo, and D. Das, "Multi-objective particle swarm
optimization based on fuzzy-Pareto-dominance for possibilistic
210.7929 175.3113 24 0.925 planning of electrical distribution systems incorporating distributed
210.7929 125.1074 30 1.325 generation," Fuzzy Sets and Systems, vol. 213, pp. 47-73, 2013.
[15] A. El-Zonkoly, "Optimal placement of multi-distributed generation
units including different load models using particle swarm
From Fig. 5, 7 and 9 it clear that after placement of DG at optimisation," IET generation, transmission & distribution, vol. 5,
pp. 760-771, 2011.
optimal locations, the total active power loss reduced in 33- [16] F. S. Abu-Mouti and M. El-Hawary, "Optimal distributed generation
bus radial distribution system. Finally, Fig. 10 shows the allocation and sizing in distribution systems via artificial bee colony
comparison of voltage profile at different optimal locations. algorithm," IEEE transactions on power delivery, vol. 26, pp. 2090-
2101, 2011.
V. CONCLUSION [17] M. Kowsalya, "Optimal size and siting of multiple distributed
generators in distribution system using bacterial foraging
In this paper, a mixed analytical method is presented on the optimization," Swarm and Evolutionary computation, vol. 15, pp.
impact of dispersed generation in RDS. In this method, VSNI 58-65, 2014.
and CISS were used to determine the exact location and [18] A. El-Fergany, "Optimal allocation of multi-type distributed
compute the size of DGs respectively. Proposed method was generators using backtracking search optimization algorithm,"
International Journal of Electrical Power & Energy Systems, vol. 64,
tested for 33- bus RDS to reduce active power losses and pp. 1197-1205, 2015.
improve voltage profile at each node. The results showed that [19] M. Junjie, W. Yulong, and L. Yang, "Size and location of distributed
clearly indicates that case C (allocation of DG at bus no 30) is generation in distribution system based on immune algorithm,"
found to be most appropriate location in minimizing the total Systems Engineering Procedia, vol. 4, pp. 124-132, 2012.
active power loss and improving node voltage as compared to [20] J. A. M. García and A. J. G. Mena, "Optimal distributed generation
location and size using a modified teaching–learning based
the other cases considered. optimization algorithm," International journal of electrical power &
REFERENCES energy systems, vol. 50, pp. 65-75, 2013.
[21] D. Q. Hung and N. Mithulananthan, "Multiple distributed generator
placement in primary distribution networks for loss reduction," IEEE
[1] W. H. Kersting, Distribution system modeling and analysis: CRC
Transactions on industrial electronics, vol. 60, pp. 1700-1708, 2011.
press, 2006.
[2] K. Zou, A. P. Agalgaonkar, K. M. Muttaqi, and S. Perera,
"Distribution system planning with incorporating DG reactive
capability and system uncertainties," IEEE Transactions on
Sustainable Energy, vol. 3, pp. 112-123, 2012.
[3] P. P. Barker and R. W. De Mello, "Determining the impact of
distributed generation on power systems. I. Radial distribution
systems," in Power Engineering Society Summer Meeting, 2000.
IEEE, 2000, pp. 1645-1656.
[4] E. Bompard, E. Carpaneto, G. Chicco, and R. Napoli, "Convergence
of the backward/forward sweep method for the load-flow analysis of
radial distribution systems," International Journal of Electrical
Power & Energy Systems, vol. 22, pp. 521-530, 2000.
[5] J.-H. Teng, "A direct approach for distribution system load flow
solutions," IEEE Transactions on power delivery, vol. 18, pp. 882-
887, 2003.
[6] T. Ackermann, G. Andersson, and L. Söder, "Distributed generation:
a definition1," Electric power systems research, vol. 57, pp. 195-
204, 2001.
[7] G. Celli, E. Ghiani, S. Mocci, and F. Pilo, "A multiobjective
evolutionary algorithm for the sizing and siting of distributed
generation," IEEE Transactions on power systems, vol. 20, pp. 750-
757, 2005.
[8] C. L. Borges and D. M. Falcao, "Optimal distributed generation
allocation for reliability, losses, and voltage improvement,"
International Journal of Electrical Power & Energy Systems, vol. 28,
pp. 413-420, 2006.
[9] D. Singh, D. Singh, and K. Verma, "Multiobjective optimization for
DG planning with load models," IEEE transactions on power
systems, vol. 24, pp. 427-436, 2009.
[10] M. Gandomkar, M. Vakilian, and M. Ehsan, "A combination of
genetic algorithm and simulated annealing for optimal DG allocation
in distribution networks," in Electrical and Computer Engineering,
2005. Canadian Conference on, 2005, pp. 645-648.
[11] M. H. Moradi and M. Abedini, "A combination of genetic algorithm
and particle swarm optimization for optimal DG location and sizing
in distribution systems," International Journal of Electrical Power &
Energy Systems, vol. 34, pp. 66-74, 2012.
IJERTV9IS070036 www.ijert.org 150
(This work is licensed under a Creative Commons Attribution 4.0 International License.)
