Professional Documents
Culture Documents
A New Approach For Optimum Simultaneous multi-DG Distributed PDF
A New Approach For Optimum Simultaneous multi-DG Distributed PDF
Energy
journal homepage: www.elsevier.com/locate/energy
a r t i c l e i n f o a b s t r a c t
Article history: This paper presents a new approach for optimum simultaneous multi-DG (distributed generation)
Received 7 February 2013 placement and sizing based on maximization of system loadability without violating the system con-
Received in revised form straints. DG penetration level, line limit and voltage magnitudes are considered as system constraints.
13 December 2013
HPSO (hybrid particle swarm optimization) algorithm is also proposed in this paper to find the optimum
Accepted 16 December 2013
Available online 18 January 2014
solution considering maximization of system loadability and the corresponding minimum power losses.
The proposed method is tested on standard 16-bus, 33-bus and 69-bus radial distribution test systems.
This paper will also compare the proposed method with existing Ettehadi method and present the
Keywords:
Distributed generation
effectiveness of the proposed method in terms of reduction in power system losses, maximization of
Reactive power system loadability and voltage quality improvement.
System loadability Ó 2013 Elsevier Ltd. All rights reserved.
Hybrid particle swarm optimization (HPSO)
1. Introduction distribution system to meet the active power demand (MW) and
get the maximum profit. For example in United States the per-
Power utilities are facing major challenges as the demand of centage of nonutility generators in distribution system has
power system is growing exponentially. The existing transmission increased from 40 GW to more than 150 GW in ten years from 1990
line infra-structure is unable to support such a huge power de- to 2000 [9].
mand. The present need is either to invest in transmission system In deregulated operation of distribution system, the DNO
to increase the capacity or provide the consumer demand locally by (distribution network operator) is responsible in providing good
DG (distributed generation). Electric power generation integrated quality service to consumer and maintain the security and effi-
within the distribution systems is known as “distributed” or ciently utilize the existing infrastructure under different un-
“dispersed” generation. The DG source can be a traditional com- certainties including load changes, available generation and
bustion generator (such as diesel reciprocating generator and nat- operating schedule of DG Units, particularly in case of intermit-
ural gas-turbine) and non-traditional generator including fuel cell, tent energy source (e.g. wind and solar). Different artificial
storage device and renewable energy source (such as wind turbine intelligent based techniques have been proposed in literature for
and photovoltaic) [1,2]. smooth operation of distribution system under various opera-
DG has many advantages over centralized power generation tional scenarios [10e15]. In some countries, the DNOs are allowed
including power losses reduction, voltage profile improvement, to invest DG Units based on its interests and requirements.
system stability improvement, pollutant emission reduction, However, in some other countries, the DNOs are not allowed to
relieving transmission and distribution congestion and others [3e own DG Units [10].
7]. The optimum placement of DG is necessary to achieve the Researchers have solved the problem of optimum DG place-
maximum benefits with less investment cost [8]. After deregulation ment problem on the basis of minimization of power losses [16e
of power system, non-utility companies are investing in 20] and voltage stability [21e27] approach. Authors have also
considered DG placement as a multi-objective function, consid-
* Corresponding author. Tel.: þ60 3 7967 5348. ering reduction in power losses, improved voltage regulations and
E-mail address: mohsinaman@gmail.com (M.M. Aman). voltage stability in fitness function [28]. It has also been seen in
0360-5442/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.energy.2013.12.037
M.M. Aman et al. / Energy 66 (2014) 202e215 203
Ref. [29] that the DG placement also improves the reliability and
voltage profile of the system. DG with reactive power control also
helps in better network voltage profile and lower power losses.
τ1
In this paper DG will be considered as a PQ (activeereactive where t is number of iterations, d is a total number of initial par-
power) source of energy (Type 4). Following definitions are used ticles, c1 and c2 are acceleration constants, set at 2.0, r1 and r2 are
in this paper to see the potential effect of DG on distribution system. random numbers, w is inertia weight given by Eqn. (4).
wmax wmin
1. DG penetration level is defined as the ratio of the total DG w ¼ wmax þ t (4)
power generation (SDG) over the system demand (SLoad) as given itermax
by Eqn. (1).
where,
SDG ðkVAÞ wmax and wmin are 0.9 and 0.4 respectively. HPSO parameter
DG penetration levelð%Þ ¼ 100 (1)
SLoad ðkVAÞ settings (c1, c2, wmin, wmax) are defined in accordance with
standard PSO parameters [17].
Table 1
DG placement evaluation indices [22,42].
KMML KMML
kVA margin to maximum loadability improvement (KMMLI) KMMLI% ¼ ð0Þ
KMMLð0Þ
ðDGÞ
100 Li is active load at bus i pu.
Pnbus
Voltage profile improvement (VPI) VPI% ¼ ðV
i ¼ 1 ið0Þ
ViðDGÞ Þ2 lmax is maximum loadability of the system.
Number of bus violating voltage limit (NBVV) NBVV ¼ 0 KMML is kVA margin to maximum loadability
of the system.
for i ¼ 1:nbus
if Vi < 0.95 or Vi > 1.05
NBVV ¼ NBVV þ 1
end
end
M.M. Aman et al. / Energy 66 (2014) 202e215 205
In the proposed method of optimum DG placement and sizing, 4.1. Fitness function calculations
HPSO is used to find the fitness function, given by Eqn. (5).
To find the maximum loadability or loading factor (l) of the
f ¼ Maxflmax g (5)
system, the active and reactive load is increased on all buses, using
where, Eqn. (7), with equal loading factor of 0.01, till the divergence is
observed in load flow analysis.
f is the fitness function and lmax is maximum loadability or
loading factor of the system. Pnew ¼ P0 Loading FactorðlÞ (7a)
2 3 Yes
Fitness Sorting Criteria Lmb==Lmb(max)
zfflffl}|fflffl{ zfflfflffl}|fflfflffl{
6 7
6 x11 x12 x13 x14 lmax1 Plosses1 7 No
6 7
6 x21 : : : lmax2 Plosses2 7 P=Po+Lmb x Po
6 7
6 7 Q=Qo+Lmb x Qo
K-Matrix ¼ 6 : : : : lmax3 Plosses3 7 (6)
6 7
6 : : : : lmax4 Plosses4 7
6 7 It=0
6 7
4 : : : : : : 5 Accuracy=1
xn1 : : : lmaxn Plossesn
No
Due to above two reasons, a modified form of original PSO is Accuracy>ep
proposed and referred as hybrid particle swarm optimization
Lmb=Lmb+0.01
Yes
(HPSO). The HPSO algorithm is inspired from Artificial Bee Colony
algorithm, in terms of scout bees’ behaviour, thus it is referred as It=It+1 ; Vpr=V
‘hybrid’. The detail of ABC (Artificial Bee Colony) algorithm can be
seen in Ref. [47]. Here it is noticed that HPSO differs from multi- Yes
It==Cutoff Break
objective approach, multi-objective approach is commonly used
in solving two independent quantities (e.g. cost and minimum No
power losses). HPSO is particularly useful when two quantities are
Forward Loop:
highly dependent on each other and the improvement in one Compute I using Eq. (A-1)
quantity will also result in better solution of other quantity (e.g. Backward Loop:
loadability, voltage profile or minimum power losses). The appli- Compute V using Eq. (A-2)
cation of HPSO in solving multi-DG Unit placement will be
explained in next section.
Compute Accuracy Using Eq. (A-3)
4. Proposed algorithm for optimum DG placement and sizing Stored Lmb and Vmin
4.2. Problem formulation amperage limits and bus voltage constraints are also introduced in
problem formulation, given in Eqns. (11) and (12) respectively.
HPSO method is used to find the optimum fitness based on
maximization of system loadability and minimum power losses. i
IDG i
< ILimit i ¼ 1 to nbr (11)
The initial particle (xi) is a 1 2 vector (PDG1, PDG1), 1 4 vector
(PDG1, PDG1, PDG2, PDG2) or 1 6 vector (PDG1, PDG1, PDG2, PDG2, PDG3,
PDG3), representing random DG positions (PDG) and DG sizes (PDG). k 1:05
0:95 Vbus k ¼ 1 to nbus (12)
Following constraints, given in Eqns. (8)e(10), related to DG
placement are considered in problem formulation. where
X
n X
Size of DG : 0 PDG Pload ðn ¼ No: of DG unitsÞ nbr is total number of branches, nbus is total number of buses and
k¼1 Vbus is bus voltage.
(8)
The amount of reactive power from DG is given by Eqn. (13).
Position of DG : 2 DG position nbuses (9)
QDG ¼ PDG tan cos1 ðpower factorÞ ; (13)
Position of DG : PDG1 sPDG2 sPDG3 (10)
In case of fixed DG penetration, the amount of active and reac-
The placement of DG may increase the line current in some of tive power from a single DG Unit can be calculated using Eqns. (14a)
the branches and bus voltages on some of the buses. Thus the line and (14b) respectively.
Start
It=It+1
Initialize random particles,
following the constrains
No given in Eqns. (8)-(12)
Is It==It_max?
Yes
Sort K-matrix:
If K[fitness(p)]==gbest(fitness) then construct K-matrix(new)
K-matrix(new)=[p fitness(p) losses(p)]
Sort K-matrix(new):
Sort K-matrix(new) on the basis of losses(p)
such that losses(p1)<losses(p2)<losses(p3)….losses(pn)
Cycle=Cycle+1
Is Cycle==Max_Cycle?
No
Yes
Select the first best particles of K-matrix(new) as optimum result
End
Fig. 4. Flow Chart of Proposed Algorithm for optimum simultaneous multi DG unit Placement based on maximization of system loadability using HPSO.
M.M. Aman et al. / Energy 66 (2014) 202e215 207
Table 2a
Maximum loadability calculations under different DG penetration and operating power factor for 16-bus radial distribution test system.
80 60 40 20 80 60 40 20 80 60 40 20
1 9.63 9.14 8.62 8.1 9.41 8.96 8.5 8.03 9.32 8.88 8.45 8.01
0.95 9.69 9.18 8.66 8.12 9.52 9.04 8.56 8.06 9.44 8.98 8.51 8.04
0.9 9.64 9.15 8.64 8.11 9.51 9.03 8.55 8.06 9.43 8.98 8.51 8.04
0.85 9.58 9.11 8.61 8.1 9.47 9 8.53 8.05 9.4 8.95 8.5 8.03
0.8 9.51 9.05 8.58 8.08 9.42 8.97 8.51 8.04 9.36 8.92 8.48 8.02
0.75 9.43 9 8.54 8.07 9.36 8.93 8.48 8.03 9.3 8.88 8.45 8.01
Highlighted result shows the maximum loadability of the system at different level of DG penetration.
PF ¼ power factor.
Table 2b
Maximum loadability calculations under different DG penetration and operating power factor for 33-bus radial distribution test system.
80 60 40 20 80 60 40 20 80 60 40 20
1 4.3 4.18 4.04 3.84 4.69 4.43 4.15 3.84 4.66 4.43 4.16 3.84
0.95 4.43 4.28 4.11 3.88 4.99 4.65 4.28 3.89 4.9 4.6 4.28 3.91
0.9 4.45 4.3 4.12 3.89 5.05 4.69 4.31 3.89 4.95 4.64 4.3 3.92
0.85 4.45 4.3 4.12 3.89 5.07 4.71 4.31 3.89 4.97 4.65 4.31 3.92
0.8 4.45 4.3 4.12 3.89 5.07 4.7 4.31 3.89 4.97 4.65 4.31 3.92
0.75 4.43 4.29 4.11 3.88 5.05 4.69 4.3 3.88 4.95 4.64 4.3 3.91
Highlighted result shows the maximum loadability of the system at different level of DG penetration.
PF ¼ power factor.
Table 2c
Maximum loadability calculations under different DG penetration and operating power factor for 69-bus radial distribution test system.
80 60 40 20 80 60 40 20 80 60 40 20
1 4.63 4.3 3.96 3.6 4.65 4.32 3.97 3.6 4.65 4.32 3.96 3.6
0.95 4.88 4.48 4.08 3.66 4.91 4.5 4.08 3.66 4.91 4.5 3.84 3.66
0.9 4.92 4.51 4.1 3.67 4.96 4.53 4.1 3.67 4.96 4.53 4.1 3.67
0.85 4.93 4.52 4.1 3.67 4.97 4.54 4.1 3.67 4.97 4.53 4.1 3.67
0.8 4.91 4.51 4.1 3.67 4.96 4.53 4.09 3.67 4.95 4.52 4.1 3.66
0.75 4.89 4.49 4.08 3.66 4.93 4.51 4.08 3.66 4.93 4.5 4.08 3.66
Highlighted result shows the maximum loadability of the system at different level of DG penetration.
PF ¼ power factor.
(14c)
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u nbus Table 3
uX X
nbus
16-Bus, 33-bus and 69-bus radial distribution test system’s details.
SLoad ¼ t Pi2 þ Qi2 (14d)
i¼2 i¼2 Test Total Active Total reactive Total apparent System power
system load in load in load in factor
The complete flow chart of proposed algorithm of multi-DG Unit kW (P) kVAR (Q) kVAS ¼ Sqrt(P2 þ Q2) (PF ¼ P/S)
placement is shown in Fig. 4. 16-Bus 28 700 5900 29 300.17 0.9795
system
33-Bus 3715 2300 4369.35 0.8502
4.3. Selection of DG operating power factor system
69-Bus 3801.9 2694.1 4659.67 0.8159
In this section, the maximization of system loadability is system
computed for 16-bus, 33-bus and 69-bus radial distribution test where Sqrt ¼ square root; PF ¼ power factor.
208 M.M. Aman et al. / Energy 66 (2014) 202e215
Fig. 5. Single line diagram of modified 16-bus radial distribution test systems.
By comparing, Tables 2 and 3, it can be seen that the most op- Case 2 (with 40% DG penetration): In case 2, the DG pene-
timum (maximum loadability) results are achieved when the DG tration is considered 40% of total load (SLoad). The DG constraints
operating power factor is nearer to the system power factor. Thus given in Eqns. (9) and (10) and line amperage constraint given in
0.95 operating power-factor of DG is considered in case of 16-bus Eqn. (12) are introduced in problem formulation. Case 2 is
test system, and 0.85 power-factor of DG is considered in case of considered in order to compare the proposed method with [22]
33-bus and 69-bus test systems. From Table 2, it can also be noticed method where author has considered 40% DG penetration level
that higher DG penetration will result in higher system loadability, and have applied the proposed method on 33-bus radial distri-
thus the amount of DG penetration level can be decided on eco- bution test system.
nomic grounds. The applications of proposed multi-DG Unit placement on
different radial distribution system are as follows.
5. Application of proposed DG placement algorithm on radial
distribution system 5.1. 16-Bus radial distribution test system
In this section, the proposed algorithm of optimum multi-DG 16-Bus system is a standard 12.66 kV network, having 3 feeders
Unit placement will be applied on 3-phase, 12.66 kV standard 16- [49]. The system details are given in appendix Table A-1. In order to
bus [49], 33-bus [50] and 69-bus [51] radial distribution test sys- solve the 3 feeder system, the system has been transformed to a
tems. The system details including bus and line data are given in single feeder, 15-bus radial distribution test system, as shown in
Appendix A. In the proposed, DG will be considered as a ePQ load Fig. 5. When the proposed algorithm of DG Unit placement is
and the DG will provide active and reactive power. The DG oper- applied on modified 16-bus radial distribution test system, results
ating power factor is set at 0.95 for 16-bus system and 0.85 for 33- presented in Table 4 are obtained. The base load of modified 16-bus
bus and 69-bus radial distribution system. Following two case system is 29 300.17 kVA.
studies are presented: From Table 4, following points can be concluded:
Case 1 (with maximum DG penetration): In case 1, DG pene-
tration is considered as maximum as possible, without violating DG 1. Using the proposed method of optimum DG placement, system
constraints, line amperage constraint and bus voltage constraint, maximum loadability, kVA margin to maximum loadability
given in Eqns. (8)e(12). In case 1, it is highly possible, the bus (KMML) and voltage quality has been significantly improved.
voltages and line current may cross the maximum limit, thus line However the power losses have been increased with maximum
amperage constraint and bus voltage constraint are introduced in DG penetration and reduced with limited 40% DG penetration.
problem formulation. Thus the amount of DG penetration must be decided on
Table 4
Application of proposed algorithm for optimum single and multi-DG Units placement on 16-bus test system.
No. of DG Units DG size in kVA / (DG position) Active power Reactive power Maximum KMML QLI VPI NBVV lV
losses (kW) losses (kVAR) system
loadability
Base case
Without DG e 511.40 590.33 7.55 192 209.12 28.1496 e 0 1.05
Proposed method (case 1)
Single DG Unit 21859.62 / (8) 315.02 407.66 9.13 238 210.39 28.8557 0.0895 0 1.33
Two DG Units 7921.98 / (7) 492.59 585.49 9.4 246 121.43 29.0211 0.1131 0 1.34
22058.09 / (8)
Three DG Units 11506.39 / (11) 536.56 635.06 9.64 253 153.4 29.0695 0.1025 0 1.71
14430.99 / (9)
9654.92 / (8)
Proposed method (case 2)
Single DG Unit 11720.1 / (08) 176.54 227.58 8.4 216 821.26 28.5396 0.0438 0 1.30
Two DG Units 11720.1 (1/2) / (11) 200.63 248.69 8.55 221 216.29 28.5723 0.0445 0 1.45
11720.1 (1/2) / (10)
Three DG Units 11720.1 (1/3) / (11) 178.36 225.17 8.50 219 751.28 28.5623 0.0418 0 1.45
11720.1 (1/3) / (08)
11720.1 (1/3) / (10)
M.M. Aman et al. / Energy 66 (2014) 202e215 209
70
60
50
Percentage Improvement
40
30
20
10
0
PLR RLR KMMLI SLI QLII PLR RLR KMMLI SLI QLII
-10
Proposed Method (With maximum DG Proposed Method (With 40% DG penetration)
penetration)
-20
Fig. 6. Comparison of DG Evaluation indices in presence of DG Unit(s) for modified 16-bus radial distribution system.
economic as well as on technical grounds. The DG placement 2. To show the advantage of the proposed method of DG place-
evaluation indices defined in Table 1 are used to develop the ment, the loading margin factor (l) is also calculated from 0 up
effectiveness of proposed method in quantitative form and the to the voltage limit (lV), satisfying the line and bus voltage
obtained results are shown in Fig. 6. constraints also. The higher value of lV will allow the operator to
add more load with the same existing infrastructure. It was
found that the base case can carry only 30 765.18 kVA (lV ¼ 1.05)
without violating the voltage and line constraints. However the
proposed method with 40% DG penetration can carry up to
38 969.23 kVA (lV ¼ 1.33), 39 262.23 kVA (lV ¼ 1.34) and
50 103.29 kVA (lV ¼ 1.71) in case of single, two and three DG
Units respectively. These results have been summarized in Fig. 7.
Similar results are achieved in case 2.
Table 5
Application of proposed algorithm for optimum single and multi-DG Units placement on 33-bus test system (case 1).
No. of DG Units DG size in kVA / (DG position) Active power Reactive power Maximum KMML QLI VPI NBVV
losses (kW) losses (kW) system
loadability
Base case
Without DG e 210.99 143.01 3.41 10 530.40 3.5228 e 21
Proposed method
Single DG Unit 3623.9 / (8) 131.85 113.03 4.31 14 462.55 3.7213 0.1757 0
Two DG Units 1313.9 / (16) 87.65 73.32 5.00 17 477.40 3.7609 0.0700 0
2212.3 / (22)
Three DG Units 444.0 / (29) 84.16 68.02 5.04 17 652.18 3.7713 0.0643 0
1364.1 / (15)
1973.0 / (31)
presented in Table 5 are obtained. The base load of 33-bus system is losses have been reduced to 131.90 kW with optimum DG
4369.35 kVA. The 33-bus system details are given in appendix placement at bus-8 and DG size of 3625.2 kVA. In last cycle
Table A-2. (cycle 3), the best result with maximum loadability (lmax ¼ 4.31)
From Table 5, following points can be concluded: and corresponding minimum power losses (131.85 kW) is
80
70
Percentage Improvement
60
50
40
30
20
10
0
PLR RLR KMMLI SLI QLII
Single DG Unit Two DG Units Three DG Units
4.32 80
Percentage Improvement
70
Maximum Loadability of the System
4.31 60
Cycle 1 (8, 3623.1 kVA) 50
4.3 λmax=4.31
Plosses=132.08 kW 40
4.29 30
( λmax)
Table 6
Comparison of proposed algorithm with Ettehadi method [22] for optimum single and multi-DG Units placement with 40% DG penetration level (case 2).
No. of DG Units DG size in kVA / (DG position) Active power Reactive power System loadability KMML QLI VPI lV NBVV
losses (kW) losses (kVAR) margin
Ettehadi method
Single DG Unit 1747.7 / (33) 148.46 116.29 3.73 11 928.33 3.6141 0.1022 0.8 9
Two DG Units 1747.7 (1/2) / (33) 103.38 75.25 4.14 13 719.76 3.6290 0.0451 1.2 0
1747.7 (1/2) / (18)
Three DG Unit 1747.7 (1/3) / (33) 99.00 70.75 4.07 13 413.91 3.6257 0.0318 1.15 0
1747.7 (1/3) / (18)
1747.7 (1/3) / (32)
Proposed method
Single DG Unit 1747.7 / (14) 112.80 81.92 4.12 13 632.37 3.6562 0.2168 1.0 3
Two DG Units 1747.7 (1/2) / (17) 52.72 38.15 4.31 14 462.55 3.6529 0.0587 1.35 0
1747.7 (1/2) / (31)
Three DG Unit 1747.7 (1/3) / (14) 59.59 41.38 4.31 14 462.55 3.6559 0.0973 1.25 0
1747.7 (1/3) / (18)
1747.7 (1/3) / (32)
212 M.M. Aman et al. / Energy 66 (2014) 202e215
36 37 38 39 40 41 42 43 44 45 46
66 67
51 52
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Substation
Main
68 69
47 48 49 50
53 54 55 56 57 58 59 60 61 62 63 64 65
28 29 30 31 32 33 34 35
Table 7
Application of proposed algorithm for optimum single and multi-DG Units placement on 69-bus test system.
No. of DG DG size in kVA / (DG position) Active power Reactive power Maximum KMML QLI VPI lV NBVV
losses losses system
loadability
Base case
Without DG e 224.95 102.15 3.22 10 344.47 3.6189 e 0.6 9
Proposed method with maximum DG penetration (case 1)
Single DG Unit 3684.7 / (61) 87.13 38.34 4.91 18 219.32 3.8713 0.1459 1.75 0
Two DG Units 3685.1 / (61) 86.68 37.17 4.91 18 219.32 3.8719 0.1459 1.75 0
547.6 / (48)
Three DG Units 3652.5 / (61) 87.00 38.18 4.91 18 219.32 3.8715 0.1459 1.75 0
32.2 / (63)
152.9 / (46)
Proposed method with 40% DG penetration (case 2)
Single DG Unit 1863.9 / (64) 41.94 24.09 4.1 14 444.98 3.7582 0.0981 1.45 0
Two DG Units 1863.9 (1/2) / (64) 29.98 18.01 4.1 14 444.98 3.7573 0.0968 1.45 0
1863.9 (1/2) / (61)
Three DG Units 1863.9 (1/3) / (62) 28.78 17.40 4.1 14 444.98 3.7569 0.0967 1.45 0
1863.9 (1/3) / (64)
1863.9 (1/3) / (61)
100
90
Percentage Improvement
80
70
60
50
40
30
20
10
0
PLR RLR KMMLI SLI QLII PLR RLR KMMLI SLI QLII
Proposed Method (With Maximum DG Proposed Method (With 40% DG penetration)
penetration)
Fig. 15. Comparison of DG Evaluation indices in presence of DG unit(s) with maximum DG penetration (Case 1) and 40% DG penetration (Case 2) for 69-bus radial distribution
system.
M.M. Aman et al. / Energy 66 (2014) 202e215 213
0.98
6. Conclusion
0.96
0.95 This paper has presented a new approach for optimum
0.94
simultaneous multi-distributed generation (DG) Units’ placement
Base case (Without DG)
Single DG (61)
and sizing on the basis of maximization of system loadability.
0.92
Two DGs (61 & 48) Hybrid particle swarm optimization (HPSO) algorithm is also
0.9
Three DGs (61,63 & 46) proposed to solve the single objective and multi-constraints
10 20 30 40 50 60 69 problem. HPSO will help in getting the best result considering
Bus No.
loadability as well as minimum power losses, when two or more
Fig. 16. Voltage profile curve of 69-bus radial system (case 1). DG combinations will result in the same maximum loadability.
The application of proposed algorithm on 16-bus, 33-bus and 69-
bus test system shows that with 40% DG penetration only, power
3. From Table 7, it can also be seen that the loading margin factor losses can be reduced up to 60e75% and maximum loadability of
considering voltage and line limit (lV) for base case (when no the system can increase up to 15e40% and overall voltage quality
DG is present) is only 2795.80 kVA (lV ¼ 0.6). However the significantly improves. The proposed method was also compared
proposed method can carry up to 8154.42 kVA (lV ¼ 1.75) and with Ettehadi method and it was found that the proposed method
6756.52 kVA (lV ¼ 1.45) in case of maximum DG penetration works better than the existing method in terms of maximization
and 40% DG penetration respectively, in all scenario of DG of system loadability, reduction in power system losses, improving
placement (single DG Unit, two DG Units and three DG Units). the KVA margin to maximum loadability (KMML) and improve-
These results are also presented in Fig. 17. ment in voltage quality.
4. From above results it can also be observed that in both cases
either maximum DG penetration or 40% DG penetration, the
performance of one, two and multi-DG Unit placement is Acknowledgements
found approximately the same in terms of power loss reduction,
maximum loadability improvement and voltage quality This work was supported by the Bright Spark Programme and
improvement. Thus in present case of 69-bus test system, op- the Institute of Research Management and Monitoring Fund-IPPP
timum number of DG placement must be ‘one’ in order to get (Grant Code: PV144/2012A) of University of Malaya and HIR/
the maximum benefit with minimum installation and operation MOHE Research Grant (Grant Code: D000004-16001) of Ministry of
cost of DG. Higher Education, Malaysia.
where S is the connected load, V is the voltage at ith bus in nth bus
system.
However, in forward sweep, the receiving end voltage (Vr) is
calculated using the difference of sending end voltage (Vs) and line
drop, as given by Eqn. (A-2).
Table A-3
h iD 69-Bus test system data [51].
Max Vik Vik1 tolerance i ¼ 1; 2; 3; .n (A-3)
From To P (kW) Q (kW) R (ohms) X (ohms) Imax
where, k is no. of iteration and n is total number of buses. 1 2 0 0 0.0005 0.0012 400
2 3 0 0 0.0005 0.0012 400
3 4 0 0 0.0015 0.0036 400
4 5 0 0 0.0251 0.0294 400
5 6 2.6 2.2 0.366 0.1864 400
Table A-1 6 7 40.4 30 0.381 0.1941 400
16-Bus test system data [49]. 7 8 75 54 0.0922 0.047 400
8 9 30 22 0.0493 0.0251 400
From To P (kW) Q (kVAR) R (ohms) X (ohms) Imax
9 10 28 19 0.819 0.2707 400
1 2 0 0 1.0E-10 1.0E-10 2500 10 11 145 104 0.1872 0.0619 200
2 3 2000 1600 0.1202 0.1603 1400 11 12 145 104 0.7114 0.2351 200
3 4 3000 400 0.1282 0.1763 500 12 13 8 5 1.03 0.34 200
3 5 2000 400 0.1442 0.2885 500 13 14 8 5.5 1.044 0.345 200
5 6 1500 1200 0.0641 0.0641 500 14 15 0 0 1.058 0.3496 200
2 7 4000 2700 0.1763 0.1763 1400 15 16 45.5 30 0.1966 0.065 200
7 8 5000 1800 0.1282 0.1763 1000 16 17 60 35 0.3744 0.1238 200
7 9 1000 900 0.1763 0.1763 500 17 18 60 35 0.0047 0.0016 200
8 10 600 500 0.1763 0.1763 500 18 19 0 0 0.3276 0.1083 200
8 11 4500 1700 0.1282 0.1763 500 19 20 1 0.6 0.2106 0.069 200
2 12 1000 900 0.1763 0.1763 1400 20 21 114 81 0.3416 0.1129 200
12 13 1000 1100 0.1442 0.1923 500 21 22 5 3.5 0.014 0.0046 200
12 14 1000 900 0.1282 0.1763 500 22 23 0 0 0.1591 0.0526 200
14 15 2100 800 0.0641 0.0641 500 23 24 28 20 0.3463 0.1145 200
24 25 0 0 0.7488 0.2475 200
Imax is assumed maximum line current limit in Amperes. SB ¼ 100 MVA;
25 26 14 10 0.3089 0.1021 200
VB ¼ 12.66 kV.
26 27 14 10 0.1732 0.0572 200
3 28 26 18.6 0.0044 0.0108 200
28 29 26 18.6 0.064 0.1565 200
29 30 0 0 0.3978 0.1315 200
30 31 0 0 0.0702 0.0232 200
31 32 0 0 0.351 0.116 200
Table A-2 32 33 14 10 0.839 0.2816 200
33-Bus test system data [50]. 33 34 19.5 14 1.708 0.5646 200
34 35 6 4 1.474 0.4873 200
From To P (kW) Q (kVAR) R (ohms) X (ohms) Imax
3 36 26 18.55 0.0044 0.0108 200
1 2 100 60 0.0922 0.0470 400 36 37 26 18.55 0.064 0.1565 200
2 3 90 40 0.4930 0.2510 400 37 38 0 0 0.1053 0.123 200
3 4 120 80 0.3661 0.1864 400 38 39 24 17 0.0304 0.0355 200
4 5 60 30 0.3811 0.1941 400 39 40 24 17 0.0018 0.0021 200
5 6 60 20 0.8190 0.7070 400 40 41 1.2 1 0.7283 0.8509 200
6 7 200 100 0.1872 0.6188 300 41 42 0 0 0.31 0.3623 200
7 8 200 100 1.7117 1.2357 300 42 43 6 4.3 0.041 0.0478 200
8 9 60 20 1.0299 0.7400 200 43 44 0 0 0.0092 0.0116 200
9 10 60 20 1.0440 0.7400 200 44 45 39.22 26.3 0.1089 0.1373 200
10 11 45 30 0.1967 0.0651 200 45 46 39.22 26.3 0.0009 0.0012 200
11 12 60 35 0.3744 0.1237 200 4 47 0 0 0.0034 0.0084 300
12 13 60 35 1.4680 1.1549 200 47 48 79 56.4 0.0851 0.2083 300
13 14 120 80 0.5416 0.7129 200 48 49 384.7 274.5 0.2898 0.7091 300
14 15 60 10 0.5909 0.5260 200 49 50 384.7 274.5 0.0822 0.2011 300
15 16 60 20 0.7462 0.5449 200 8 51 40.5 28.3 0.0928 0.0473 200
16 17 60 20 1.2889 1.7210 200 51 52 3.6 2.7 0.331 0.1114 200
17 18 90 40 0.7320 0.5739 200 9 53 4.35 3.5 0.174 0.0886 300
2 19 90 40 0.1640 0.1564 200 53 54 26.4 19 0.203 0.1034 300
19 20 90 40 1.5042 1.3555 200 54 55 24 17.2 0.2842 0.1447 300
20 21 90 40 0.4095 0.4784 200 55 56 0 0 0.2813 0.1433 300
21 22 90 40 0.7089 0.9373 200 56 57 0 0 1.59 0.5337 300
3 23 90 50 0.4512 0.3084 200 57 58 0 0 0.7837 0.263 300
23 24 420 200 0.8980 0.7091 200 58 59 100 72 0.3042 0.1006 300
24 25 420 200 0.8959 0.7010 200 59 60 0 0 0.3861 0.1172 300
6 26 60 25 0.2031 0.1034 300 60 61 1244 888 0.5075 0.2585 300
26 27 60 25 0.2842 0.1447 300 61 62 32 23 0.0974 0.0496 300
27 28 60 20 1.0589 0.9338 300 62 63 0 0 0.145 0.0738 300
28 29 120 70 0.8043 0.7006 200 63 64 227 162 0.7105 0.3619 300
29 30 200 600 0.5074 0.2585 200 64 65 59 42 1.041 0.5302 300
30 31 150 70 0.9745 0.9629 200 11 66 18 13 0.2012 0.0611 200
31 32 210 100 0.3105 0.3619 200 66 67 18 13 0.0047 0.0014 200
32 33 60 40 0.3411 0.5302 200 12 68 28 20 0.7394 0.2444 200
68 69 28 20 0.0047 0.0016 200
Imax is maximum line current limit in Amperes, given in Ref. [54]. SB ¼ 100 MVA;
VB ¼ 12.66 kV. Imax is maximum line current limit in Amperes, SB ¼ 100 MVA, VB ¼ 12.66 kV.
M.M. Aman et al. / Energy 66 (2014) 202e215 215
References [28] Moradi MH, Abedini M. A combination of genetic algorithm and particle
swarm optimization for optimal DG location and sizing in distribution sys-
tems. Int J Electr Power Energy Syst 2012;34(1):66e74.
[1] Ackermann T, Andersson G, Söder L. Distributed generation: a definition.
[29] Borges CLT, Falcão DM. Optimal distributed generation allocation for reli-
Electr Power Syst Res 2001;57(3):195e204.
ability, losses, and voltage improvement. Int J Electr Power Energy Syst
[2] Chicco G, Mancarella P. Distributed multi-generation: a comprehensive view.
2006;28(6):413e20.
Renew Sustain Energy Rev 2009;13(3):535e51.
[30] Quezada VM, Abbad JR, Roman TGS. Assessment of energy distribution losses
[3] Chiradeja P, Ramakumar R. An approach to quantify the technical benefits of
for increasing penetration of distributed generation. Power Syst IEEE Trans
distributed generation. IEEE Trans Energy Convers 2004;19(4):764e73.
2006;21(2):533e40.
[4] Bayod-Rújula AA. Future development of the electricity systems with
[31] Hemdan NGA, Kurrat M. Distributed generation location and capacity effect
distributed generation. Energy 2009;34(3):377e83.
on voltage stability of distribution networks. In: Conference distributed
[5] Zangiabadi M, Feuillet R, Lesani H, Hadj-Said N, Kvaløy JT. Assessing the
generation location and capacity effect on voltage stability of distribution
performance and benefits of customer distributed generation developers
networks. p. 1e5.
under uncertainties. Energy 2011;36(3):1703e12.
[32] Wan H, McCalley JD, Vittal V. Increasing thermal rating by risk analysis. IEEE
[6] Soroudi A, Ehsan M. A distribution network expansion planning model
Trans Power Syst 1999;14(3):815e28.
considering distributed generation options and techno-economical issues.
[33] Prada RB, Souza LJ. Voltage stability and thermal limit: constraints on the
Energy 2010;35(8):3364e74.
maximum loading of electrical energy distribution feeders. IEE Proc Gener
[7] Bracco S, Dentici G, Siri S. Economic and environmental optimization model
Transm Distrib 1998;145(5):573e7.
for the design and the operation of a combined heat and power distributed
[34] Milano F. Continuation power flow analysis. Power system modelling and
generation system in an urban area. Energy 2013;55(0):1014e24.
scripting; 2010. pp. 103e30.
[8] Celli G, Ghiani E, Mocci S, Pilo F. A multiobjective evolutionary algorithm for
[35] Mahmud MA, Hossain MJ, Pota HR, Nasiruzzaman ABM. Voltage control of
the sizing and siting of distributed generation. IEEE Trans Power Syst
distribution networks with distributed generation using reactive power
2005;20(2):750e7.
compensation. In: Conference voltage control of distribution networks with
[9] Puttgen HB, Macgregor PR, Lambert FC. Distributed generation: semantic hype
distributed generation using reactive power compensation. p. 985e90.
or the dawn of a new era? IEEE Power Energy Mag 2003;1(1):22e9.
[36] Turitsyn K, Sulc P, Backhaus S, Chertkov M. Options for control of reactive
[10] Soroudi A, Ehsan M, Caire R, Hadjsaid N. Possibilistic evaluation of distributed
power by distributed photovoltaic generators. Proc IEEE 2011;99(6):1063e73.
generations impacts on distribution networks. IEEE Trans Power Syst
[37] Ellis A, Nelson R, Von Engeln E, Walling R, MacDowell J, Casey L, et al.
2011;26(4):2293e301.
Reactive power performance requirements for wind and solar plants. In:
[11] Soroudi A, Ehsan M, Caire R, Hadjsaid N. Hybrid immune-genetic algorithm
Conference reactive power performance requirements for wind and solar
method for benefit maximisation of distribution network operators and
plants. IEEE. p. 1e8.
distributed generation owners in a deregulated environment. IET Gener
[38] Xu L, Cartwright P. Direct active and reactive power control of DFIG for wind
Transm Distrib 2011;5(9):961e72.
energy generation. IEEE Trans Energy Convers 2006;21(3):750e8.
[12] Soroudi A, Caire R, Hadjsaid N, Ehsan M. Probabilistic dynamic multi-objective
[39] Tang Y, Xu L. A flexible active and reactive power control strategy for a var-
model for renewable and non-renewable distributed generation planning. IET
iable speed constant frequency generating system. IEEE Trans Power Electron
Gener Transm Distrib 2011;5(11):1173e82.
1995;10(4):472e8.
[13] Mohammadi-Ivatloo B, Zareipour H, Amjady N, Ehsan M. Application of
[40] Liew SN, Strbac G. Maximising penetration of wind generation in existing
information-gap decision theory to risk-constrained self-scheduling of Gen-
distribution networks. IEE Proc Gener Transm Distrib 2002;149(3):256e62.
Cos. Power Syst IEEE Trans 2013;28(2):1093e102.
[41] Ghosh S, Ghoshal SP, Ghosh S. Optimal sizing and placement of distributed
[14] Soroudi A, Ehsan M. IGDT based robust decision making tool for DNOs in load
generation in a network system. Int J Electr Power Energy Syst 2010;32(8):
procurement under severe uncertainty. IEEE Trans Smart Grid 2013;4(2):886e95.
849e56.
[15] Soroudi A. Robust optimization based self scheduling of hydro-thermal Genco
[42] Abou El-Ela AA, Allam SM, MM S. Maximal optimal benefits of distributed
in smart grids. Energy 2013;61(0):262e71.
generation using genetic algorithms. Electr Power Syst Res 2010;80:869e77.
[16] Duong Quoc H, Mithulananthan N. Multiple distributed generator placement
[43] Doagou-Mojarrad H, Gharehpetian GB, Rastegar H, Olamaei J. Optimal place-
in primary distribution networks for loss reduction. IEEE Trans Ind Electron
ment and sizing of DG (distributed generation) units in distribution networks
2013;60(4):1700e8.
by novel hybrid evolutionary algorithm. Energy 2013;54(0):129e38.
[17] Nguyen Cong H, Mithulananthan N, Bansal RC. Location and sizing of
[44] Moravej Z, Akhlaghi A. A novel approach based on cuckoo search for DG
distributed generation units for loadability enhancement in primary feeder.
allocation in distribution network. Int J Electr Power Energy Syst 2013;44(1):
IEEE Syst J 2013;7(4):797e806.
672e9.
[18] Caisheng W, Nehrir MH. Analytical approaches for optimal placement of
[45] Venkatesh B, Ranjan R, Gooi H. Optimal reconfiguration of radial distribu-
distributed generation sources in power systems. Power Syst IEEE Trans
tion systems to maximize loadability. IEEE Trans Power Syst 2004;19(1):
2004;19(4):2068e76.
260e6.
[19] Acharya N, Mahat P, Mithulananthan N. An analytical approach for DG allo-
[46] Kennedy J, Eberhart R. Particle swarm optimization. Conference Particle
cation in primary distribution network. Int J Electr Power Energy Syst
Swarm Optim;4. p. 1942e8.
2006;28(10):669e78.
[47] Karaboga D. An idea based on honey bee swarm for numerical optimization.
[20] Gözel T, Hocaoglu MH. An analytical method for the sizing and siting of distrib-
Techn Rep TR06. Erciyes: Erciyes Univ Press; 2005.
uted generators in radial systems. Electr Power Syst Res 2009;79(6):912e8.
[48] Gözel T, Eminoglu U, Hocaoglu MH. A tool for voltage stability and optimi-
[21] Hedayati H, Nabaviniaki S, Akbarimajd A. A method for placement of DG units
zation (VS&OP) in radial distribution systems using matlab graphical user
in distribution networks. IEEE Trans Power Deliv 2008;23(3):1620e8.
interface (GUI). Simul Model Pract Theory 2008;16(5):505e18.
[22] Ettehadi M, Ghasemi H, Vaez-Zadeh S. Voltage stability-based DG placement
[49] Zhu JZ. Optimal reconfiguration of electrical distribution network using the
in distribution networks. IEEE Trans Power Deliv 2013;28(1):171e8.
refined genetic algorithm. Electr Power Syst Res 2002;62(1):37e42.
[23] Aman MM, Jasmon GB, Mokhlis H, Bakar AHA. Optimal placement and sizing
[50] Baran ME, Wu FF. Network reconfiguration in distribution systems for loss
of a DG based on a new power stability index and line losses. Int J Electr Power
reduction and load balancing. IEEE Trans Power Deliv 1989;4(2):1401e7.
Energy Syst 2012;43(1):1296e304.
[51] Savier J, Das D. Impact of network reconfiguration on loss allocation of radial
[24] Al Abri RS, El-Saadany EF, Atwa YM. Optimal placement and sizing method to
distribution systems. IEEE Trans Power Deliv 2007;22(4):2473e80.
improve the voltage stability margin in a distribution system using distrib-
[52] Thukaram D, Wijekoon Banda H, Jerome J. A robust three phase power flow
uted generation. Power Syst IEEE Trans 2013;28(1):326e34.
algorithm for radial distribution systems. Electr Power Syst Res 1999;50(3):
[25] Alonso M, Amaris H. Voltage stability in distribution networks with DG. In:
227e36.
IEEE PowerTech conference 2009. pp. 1e6. Bucharest.
[53] Jen-Hao T. A direct approach for distribution system load flow solutions. IEEE
[26] Aman MM, Jasmon GB, Bakar AHA, Mokhlis H. A new approach for optimum
Trans Power Deliv 2003;18(3):882e7.
DG placement and sizing based on voltage stability maximization and mini-
[54] Aman MM, Jasmon GB, Bakar AHA, Mokhlis H. Optimum network reconfigu-
mization of power losses. Energy Convers Manag 2013;70(0):202e10.
ration based on maximization of system loadability using continuation power
[27] Mithulananthan N, Oo T. Distributed generator placement to maximize the
flow theorem. Int J Electr Power Energy Syst 2014;54(0):123e33.
loadability of a distribution system. Int J Electr Eng Educ 2006;43(2):107e18.