ISSN (Print): 0974-6846

ISSN (Online): 0974-5645

Indian Journal of Science and Technology, Vol 11(23), DOI: 10.17485/ijst/2018/v11i23/119654, June 2018

Application of SSPSO and SPSO to Optimal Power

Flow Solutions along with SSC Device
D. Ganga Bhavani*, N. Rama Narayana and R. Madhusudan
Department of Electrical and Electronics Engineering, Sir C R Reddy College of Engineering,
1

Eluru – 534007, Andhra Pradesh, India;

[email protected], [email protected], [email protected]
2
Department of Electrical and Electronics Engineering, Sir C R Reddy College of Engineering, Eluru – 534007,

Abstract
Objectives: In this study, optimization techniques SSPSO and SPSO are used to minimize the total generation cost and
active power loss. Methods/Analysis: The power flow analysis gives the information of magnitude of voltages, phase
angles and complex power flow at each bus. Optimal Power Flow (OPF) solutions with Static Series Compensator (SSC)
FACTS device in the power system to increase the total power transfer. Fast Relative Electrical Distance (F-RED) method
based on Simple-Voltage Stability Index (H-index) is used to find the maximum generation supplied to the load. Findings:
Comparison was made on SSPSO and SPSO methods and SPSO with SSC gives better results. The H-index is used to judge
the power system is stable or unstable. F-RED gives the information, the amount of generators contribution to satisfy the
loads. Novelty/ Improvement: SSPSO gives better solution results than SPSO, the methods are applied on IEEE-30 bus
system. H- INDEX and F-RED are calculated for the IEEE-5 bus system.

Keywords: OPF, SPSO, SSC, SSPSO

1. Introduction Nowadays, the new algorithms have evolved as a solu-

tion for complex problems. Along with conventional
Optimal power flow is considered as backbone tool in methods intelligent methods like SSPSO and SPSO gives
the power systems. With the increase in demand leads to greater convergence characteristics and gives the best
increase in generation that requires increase in the ther- global minimum solution.
mal capacity. For this reason, the problem of Optimal With the addition of SSC device on the power sys-
Power Flow (OPF)1 is yet to be studied by minimizing tem forenhancement of voltage stability index by SPSO
the cost, losses andhas becomes the source of concern to based power flow solutions was examined in this paper.
secure operation of the power system. The voltage stabil- The proposed SPSO with SSC device was tested on
ity limits2,3 is also considered for secure operation of the IEEE-30 bus6. The results are observed by considering
system. SSC cost also. Comparison analysis was done on the total
The F-RED4 concept is used to know how much generation cost and active power losses by optimization
amount of generation is generated from each generator to methods SPSO and SSPSO for IEEE-30 Bus.
satisfy the load. FACTS are used in power system for con-
trolling power and to increase the transmission capacity
2. Simple-voltage Stability Index
near to their thermal limits. FACTS also improve the sys-
tem stability and security, reduce the reactive power flow, (H-index)
thereby permitting greater active power flow5. Different The stability of the power system is represented in terms of
types of FACTS controllers are: a) Series, b) Shunt, voltage stability studies. This can be represented with the
c) combination of Series and Shunt. help of voltage stability index. The scalar magnitude of this

*Author for correspondence

 Application of SSPSO and SPSO to Optimal Power Flow Solutions along with SSC Device

index can be helpful for monitoring the system parameters. attained by applying the voltage either at midpoint or in
This index-based instability measures and captures unique series with the line or in phase quadrature with the cur-
system behavior in terms of number and interprets them to rent flow in the transmission line7.
give the idea of distance to instability. Normal power sys- Static Series Compensator: Series compensators are
tems have H-index value lies in between 0 and 1. For stable used to improve the power flow in the line where it is
operation of the system the H-index should be near to zero. connected.The series capacitive compensation is used
The H-index can be computed with the formula: to reduce the overall transmission line impedance from
g sending end to receiving end. SSC also minimizes the pos-
Vn
H = 1 − ∑ Fij , m = g + 1...., n1 (1) sibility of voltage collapse. Representation of SSC is shown
n =1 Vm in Figure 1; the active and reactive power flow equations
where, H is the simple-voltage stability index when SSC is included between bus ‘m’ and ‘n’ are:

Vn – is the voltage at the bus n Pmn = Vm2 g mm − VmVn ( g mm cos θmn + bmn sinθmn )

Vm – is the voltage at the bus m − VmVse ( g mn cos (θm − θ se ) + bmn sin (θm − θ se ) ) (4)
All the variable sinside the sigma in the above equa-
Qmn = −Vm2bmm − VmVn ( g mn sinθmn − bmn cos θmn )
tion are complex quantities. The value Fij is calculated
from the admittance matrix. − VmVse ( g mn sin (θm − θ se ) − bmn cos (θm − θ se ) ) (5)
The P and Q equations from bus ‘n’ to bus ‘m’ are:
3. Fast-relative Electrical Pnm = Vn2 g nm − VmVn ( g mn cos θnm + bmn sinθnm )
Distance + VnVse ( g mn cos (θn − θ se ) + bmn sin (θm − θ se ) ) (6)
The process of F-RED computes the relative distance
between the load and generation buses. Is calculated by Qnm = −Vn2bnn − VmVn ( g mn sinθnm − bmn cos θnm )
using the transmission system parameters. Using this + VnVse ( g mn sin (θn − θ se ) − bmn cos(θn − θ se ) ) (7)
approach, the requirement of generation increase or
decrease can be identified. The load shared by different Here 1
generators acts as function of their relative distance from g mn + jbmn = , g mm = g mn ,
Z se
each of the loads. The reactive power contributed by vari-
ous sources is also evaluated using this approach. =
bmm b=
mn , g nn g=mn , bnn bmn
[F-RED] = M - [DLG ] (2) Operating limits of the SSC:
where, M is the unity matrix of size L × G, G is the PE = Re ( Vse Inm
*
) = 0 or
no. of generator buses and L is the no. of load buses.
−VmVse ( g mn cos (θm − θ se ) + bmn sin (θm − θ se ) )
where, [ DLG ] is also known as desired generation pro-
portions matrix + VnVse ( g mn cos (θn − θ se ) + bmn sin (θn − θ se ) ) t = 0  (8)
[ DLG ] =abs {[FLG ] }(3)
The P and Q limitations are
[ DLG ] gives the information about the location of the specified
load buses with reference to the generator buses, is widely Pnm − Pnm = 0 (9)
called as F-RED. The elements of [FLG ] matrix is complex, Qnm − Qnmspecified
= 0 (10)

the columns of [FLG ] represents the no.of generator buses
Here
and rows of [FLG ] represents the no.of load buses.
specified
Pnm and Qnm
specified
are specified “P” and “Q”

4.  Facts Controllers The voltage injected by SSC isVse ∠θ se and the limits
are
FACTS represent a family of power electronic based Vsemin ≤ Vse ≤ Vsemax
devices which are able to improve the power system con-
θ semin ≤ θ se ≤ θ semax
trollability and stability. The power controllability can be

2 Vol 11 (23) | June 2018 | www.indjst.org Indian Journal of Science and Technology
 D. Ganga Bhavani, N. Rama Narayana and R. Madhusudan

where, 2. Voltage values at every bus:

Vse =0.04 p.u, Vsemin = 0.001, Vsemax =0.2
Vmmin ≤ Vm ≤ Vmmax , m = 1...N (16)
θ se =87.13°, θ min
se =90°, θ max
se =180°
3. Transformer tap limits:

Tmmin ≤ Tm ≤ Tmmax , m = 1...NT (17)

4. Reactive power supply due to capacitor banks:

min max
QCm ≤ QCm ≤ QCm , m = 1...CS (18)

5. Transmission lines loading limits:

Sm ≤ Smmax , m = 1...nl (19)

6. Simple voltage stability index limits:

Ljm ≤ Ljmmax , m = 1...NL (20)

Figure 1.  Equivalent circuit of SSC.
iii)  FACTS device constraints
Vsemin ≤ Vse ≤ Vsemax
5.  Optimal Power Flows
θ semin ≤ θ se ≤ θ semax
The SPSO technique with SSC device is applied to opti-
2)  Minimization of total active power loss is
mize the fuel cost of generation and to improve the
system performance by maintaining active power and ∑P = ∑P − ∑P
L G D (21)
voltage constraints. Mathematically, the OPF objectives
are functioning to minimize generation cost and reduce
active power losses of the power system. 6. Simplified-particle Swarm
The objective functions are: Optimization
1) Minimization of generation fuel cost Simplified-Particle Swarm Optimization was developed
ng
through the imitation of bird flocking or school of fish.
F = ∑(am PGm
2
+ bm PGm + cm ) (11)
m =1
It is an evolutionary process, used to obtain the optimal
To minimize of above objective function subjected to solution for different problems8–10.
both exact equality and inequality constraints are: The idea of SPSO consists of changing velocity of each
particle towards local-best and global-best positions at
i)  Exact-equality constraints:
n1
each iteration. The velocity term is multiplied by a ran-
PGm − PDm − ∑ Vm Vn Ymn cos(θmn − δ m + δ n ) = 0 (12) dom term, will generate random values towards local best
n =1 and global best positions. By using method the global
n1
best values are obtained very faster way. The other advan-
QGm − QDm + ∑ Vm Vn Ymn sin(θmn − δ m + δ n ) = 0 (13)
n =1 tage using this method is parameters are few in number
(Figure 2) the above The figure represents the searching of
where, PGm and QGm are the real and reactive powers particles towards its best position. Each particle modifies
at bus ‘m’. The load demand at the bus ‘m’ is represented its position with changing its velocity using the formula:
by PDm and, QDm and the values of Ymn and θmn are taken vm
n +1
= w ∗vmn + c ’1rand()1∗ ( pbest m − smn )
from Y-bus.
+ c ’2 rand()2∗ ( gbest − smn ) (22)
ii)  Exact-Inequality constraints are:
where, vmn is velocity of particle ‘m’ at iteration ‘n’,
1. Generators power limits:
c and c2’ are constants, which varies the velocity of a
’
1
min max
PGm ≤ PGm ≤ PGm , m = 1,..., N (14) particle towards pbest (local best) and gbest (global best),
min max
rand1 () and rand2 () are the numbers generated randomly
QGm ≤ QGm ≤ QGm , m = 1...N (15) between zero and one. smn is current position of particle ‘m’

Vol 11 (23) | June 2018 | www.indjst.org Indian Journal of Science and Technology 3
 Application of SSPSO and SPSO to Optimal Power Flow Solutions along with SSC Device

at iteration ‘n ’, pbest m is the local best of ‘m’ and gbest 7. Supervisory-sliced Particle
is the best value among the local bests of all the particles. Swarm Optimization
The weight function in the above equation can be deter-
mined using the below formula. A new optimization method Sliced Particle Swarm
Optimization (SSPSO) is introduces the slicing of search
w = wmax − ((wmax − wmin ) / (itermax ))∗ iter (23)
space into rectangular portions. This SSPSO gives the
where, wmax is the final weight, wmin is the initial weight complete solution in terms of reduction in the calcula-
as these limits controls exploration and exploitation of the tions and tracking of optimal value in each search space.
search space, ‘ itermax ’ is the total iterations and ‘iter’ is the In SSPSO, the algorithm split the search space into no.
present iteration. The positionupdating can be done by of rectangular slotted sections i.e. from outer rectangular
using the formula: slot with high search space towards inner rectangular slot
smn+1 = smn + vmn+1 (24) with less search space. The complete solution is obtained
by searching all the slices.The comparison among all the
The above procedure is represented in step-by-step slices gives S-best (pbest) and the comparison of S-best of
manner is shown in Figure 3. each slices leads to gbest .
Momentum factor (mc) used in the equation (27)
with position change gives the convergence and particle
velocity changes according position.
The position and velocity updating equations for each
particle in each slice are:
vmn +1 = w ∗vmn + c ’1rand1 ()∗ ( pbest m − smn )
+ c ’2 rand2 ()∗ ( gbest − smn ) (25)

w=
(w final )
− winitial (Titer − iter )
+ winitial (26)
Titer
Figure 2.  Concept of modification of a searching where, w final and winitial are the pre-defined highest
point by SPSO.
and lowest weighting function values, respectively. Larger
weighting value gives a global exploration and smaller
weightingvalue gives a local exploration.
The current position (new position) can be modified
by using the following equation:

smn +1 = (1 − mc)smn + mc.vmn +1 (27)

MC =momentum factor (o<mc<1)

The boundaries are:
Vmin = Xmin (28)

Vmax = Xmax (29)

SSPSO Algorithm
1.  Splitting the search space and divide the particles in
each sliced search space.
2.  Initialize the total population.
3.  For trials = 1: 30
For iterations = 1 : maximum iterations
For K = 1 : number of particles
Figure 3.  Pictorial representation of SPSO method. For L = 1 : number of dimensions

4 Vol 11 (23) | June 2018 | www.indjst.org Indian Journal of Science and Technology
 D. Ganga Bhavani, N. Rama Narayana and R. Madhusudan

4. Evaluate the objective function & fitness of each Table 1. Optimal parameters for SPSO
particle Parameter SPSO
5.  Update the position according to formula (27). Population size 20
6.  Bound the position of particle within the boundaries
Number of iteration 200
of the slice.
Cognitive constant, c1 2
7.  Update the velocity using the formula (25).
8.  Bound the velocity of particles within each slice using Social constant, c2 2
equations (28) and (29). Inertia weight, w 0.3-0.9
9.  Find S*-best and fitness of the particle.
10. Check for initialization of each slice is completed, if Table 2. Optimal parameters for SSPSO
No go to steps 2-11 otherwise, end. Parameter SSPSO
11. Find g-best fitness, if yes go to 14, otherwise go to Population size 40
step 4.
Number of iteration 150
12. End the algorithm.
Cognitive constant, c1 2
The selected mean of S∗1, S∗2, S∗3, S∗4 represents
Social constant, c2 2
the S∗-best value of each slice and the selected mean ‘S∗’
represents the global best value of each slice. The fitness Inertia weight, w 0.3-0.9
value of SSPSO for each slice is much better than SPSO. momentum factor (mc) 0.3

Table 3. Simple-Voltage Stability Index for the IEEE-5

8. Simulation Results and bus system
Discussions Load Bus H-index
The proposed SPSO & SSPSO algorithms for solving opti- 3 0.0637
mal power flows is tested on the IEEE-30 bus test system. 4 0.0671
Simple-Voltage Stability Index and Fast-Relative Electrical 5 0.0772
Distance are calculated on the IEEE-5 bus. The power
flow control ability of the power system can be improved Table 4. F-RED values for the IEEE -5 bus system
with the addition ofSSC device. The improvement in
Load Bus Load (MW) G1 (%) G2 (%)
H-index is also observed by adding series compensator
L3 45 0.7443 0.2490
in OPF problem with SPSO algorithm. SSC with SPSO is
done on the IEEE-30 bus test system. L4 40 0.7952 0.1978
The Tables 1, 2 gives the parameter settings for SPSO L5 60 0.9315 0.0629
and SSPSO algorithms. Table 3 represents the H-index
values at each load bus. The columns correspond to the Cost Characteristics of the IEEE-30 Bus by SPSO and
generator buses numbers and rows correspond to the load SSPSO
bus numbers (Table 4).
Figure 4 shows the cost characteristics of the IEEE-30 bus
The load at bus-3 is 45 MW. To meet this load G1 con-
by taking number of iterations on X-axis and total genera-
tributes 74.43% of 45 MW (i.e 45∗0.7443=33.4935 MW)
tion cost on Y-axis for the first graph. Second graph is the
and G2 contributes 24.49% of 45 MW (i. e 45∗0.2490
fitness of the cost function which is obtained by taking
= 11.205). The load at bus-4 is 40 MW, for this load G1
inverse of the generation cost.
gives 79.52% of 40 MW and G2 gives 19.78% of 40 MW.
Similarly the load bus at 5 is 60 MW, to meet this load Figure 5 s the cost characteristics of IEEE-30 bus by
G1 contributes 93.15% of 60 MW and G2 contributes using SSPSO algorithm. The generation cost for above
6.29% of 60 MW. (G1-generator 1, G2-generator 2). system is 800.30($/hr) and the loss is 0.0895 p.u.

Vol 11 (23) | June 2018 | www.indjst.org Indian Journal of Science and Technology 5
 Application of SSPSO and SPSO to Optimal Power Flow Solutions along with SSC Device

Table 5. Comparison of SPSO and SSPSO for the

IEEE-30 bus system
Objective Function SPSO algorithm SSPSO
algorithm
Total cost ($/hr) 800.77 800.30
Active power loss (p.u) 0.0972 0.0895

The characteristics shows that combined characteris-

tics of SSPSO and SPSO optimization methods. It is clear
that SSPSO gives better results compared to SPSO
method.

Figure 4.  Cost characteristics of the IEEE-30 using

SPSO algorithm.

Figure 6.  Comparison of Cost characteristics for the

IEEE-30 bus system using SSPSO and SPSO.

Results for SPSO with SSC for the IEEE-30 Bwus

Table 6 describes the voltage stability indices of the IEEE-
30 bus with the addition of SSC between the buses 9-30.
The H-index values are improved with the addition of
SSC compared to without SSC (Figure 7).
The Figure 7 shows that the generation cost is the total
generation cost plus SSC device cost and second graph
shows that inverse of cost function i.e. fitness of the cost
Figure 5.  Cost characteristics of the IEEE-30 using SSPSO
algorithm. function.

Comparison between SPSO and SSPSO Table 6. H-index for the IEEE-30 bus system using
Table 5 shows that the objective functions total genera- SPSO along with SSC
tion cost and active power losses are compared by using Load bus H-index Load bus H-index
SPSO and SSPSO. The total generation cost and active 7 0.0229 19 0.0987
power loss are minimized by using SSPSO compared to 8 0.0150 20 0.0946
SPSO (Figure 6). 9 0.0413 21 0.0871

6 Vol 11 (23) | June 2018 | www.indjst.org Indian Journal of Science and Technology
 D. Ganga Bhavani, N. Rama Narayana and R. Madhusudan

Load bus H-index Load bus H-index 9. Conclusion

10 0.0748 22 0.0867
In this study, we implemented SPSO and SSPSO optimi-
11 0.176 23 0.1012
zation techniques for the solution of power flows under
12 0.0502 24 0.0987 steady conditions and also improvement in the power
13 0.0156 25 0.0955 system was observed with the addition of SSC device on
14 0.0717 26 0.1096 the IEEE-30 bus.H-indexand F-RED are also calculated
15 0.0823 27 0.0868 on the IEEE-5 bus test systems. The SPSO and SSPSO
16 0.0667 28 0.0208 algorithms give reliable and accurate optimal power flow
17 0.0775 29 0.1160 solutions.
18 0.947 30 0.1357
10. References
1. Venikov VA, Stroev VA, Idelchick VI, Tarasov VI.
Estimation of electric power system steady state stabil-
ity in load flow calculations. IEEE Transactions on Power
Apparatus and Systems. 1975 Jun; 94(3):1034–40. Crossref.
2. Perez-Londono S, Rodriguez LF, Olivar G. A Simplified
Voltage Stability Index (SVSI). Electrical Power and Energy
Systems. 2014 Dec; 63(2014):806–13.
3. Subramani C, Dash SS, Kumar MJ, Bhaskar MA. Stability
index based voltage collapse prediction and contingency
analysis. Journal of Electrical Engineering and Technology.
2009; 4(4):438–42. Crossref.
4. Thukaram D, Vyjayanthi C. Relative electrical distance
concept for evaluation of network reactive power and loss
contributions in a deregulated system. IET Institution of
Engineering and Technology Generation, Transmission
and Distribution. 2009; 3(11):1000–19. Crossref.
5. Li N, Xu Y, Chen H. FACTS-based power flow control
Figure 7.  Cost characteristics for the IEEE-30 bus system
in interconnected power systems. IEEE Transactions on
Using SPSO with SSC.
Power Systems. 2000 Feb; 15(1):257–62. Crossref.
6. Hingaroni NG, Gyugyi L. Understanding FACTS: Concepts
Table 7 shows the results that are obtained by using and technology of flexible AC transmission systems. IEEE
SPSO algorithm along with SSC for the IEEE-30 by con- Press, New-York; 2000. p. 1–452.
sidering the objective functions as total generation cost 7. Kennedy J, Eberhart M. Particle swarm optimization.
and active power loss. The results are obtained by tak- Proceedings of IEEE International Conference on Neural
ing all the exact-equality and inequality constraints. Here Networks; 1995 Dec. Crossref.
generation cost represents total generation cost along with 8. Abido MA. Optimal power flow using particle swarm
the SSC device cost. The active power loss is reduced with optimization. Electrical Power Energy System. 2002;
the addition of SSC compared to without SSC. H-indexof 24(7):563–71. Crossref.
9. Hayeghi H, Shayanfar HA, Shojaei A. An improved PSO
the system is also improved with the inclusion of SSC.
based solution for the optimal power flow problems.
University of Pitesti – Electronics and Computers Science
Table 7. SPSO with SSC results for the IEEE-30 bus Scientific Bulletin. 2008; 8(2):99–105.
system 10. Garg H, Pattnaik SS, Devi S, Bakwad KM, Panigrahiz
Objective function SPSO with SSC BK, Das SK. Sliced Particle Swarm Optimization (SPSO):
Total generation cost ($/hr) 1012.8 A computationally efficient optimization technique. World
Active power loss (p.u) 0.1138 Congress on Nature and Biologically Inspired Computing
(NABIC); 2009. p. 61–6. Crossref.
H- index 0.1588

Vol 11 (23) | June 2018 | www.indjst.org Indian Journal of Science and Technology 7