2009 IEEE/ASME International Conference on Advanced Intelligent Mechatronics
Suntec Convention and Exhibition Center
Singapore, July 14-17, 2009
Optimal Location and Capacity of STATCOM for Voltage stability
Enhancement using ACO plus GA
1
Whei-Min Lin, 1Kai-Hung Lu, 2Cong-Hui Huang, 3Ting-Chia Ou, and 1Yuan-Hui Li
Abstract—Voltage security is a crucial issue in power
systems especially under heavily loaded condition. In the new
scheme of restructuring, voltage stability problem becomes
even more serious. To solve the problem, we integrate reactive
power compensation concept by Static Synchronous
Compensator (STATCOM) with Equivalent – current Injection
(ECI). We derive a new STATCOM with ECI model. This
paper shows the application of Ant Colony Optimization (ACO)
plus Genetic Algorithms (GA) for optimal capacity and location
of a new STATCOM with ECI model in a power system. Finally
simulation shows the optimal location and capacity of new
STATCOM with ECI model to enhance power system voltage
stability by using GACO. The proposed method demonstrates
the improvement of voltage stability margin.
Keywords: Genetic Algorithms, Ant Colony Optimization,
voltage stability, Static Synchronous Compensator
(STATCOM), Equivalent -Current-Injection (ECI)
I. INTRODUCTION
P ower systems components mainly consist of generators,
transmission lines, transformers, switches, active or
passive compensators and loads. Power system networks are
complex systems that are nonlinear, non-stationary, and
prone to disturbances and faults. Reinforcement of a power
system can be accomplished by improving the voltage
profile, increasing the transmission capacity and others.
Flexible AC Transmission System (FACTS) devices are an
alternate solution to address some of those problems [5].
The FACTS devices can be categorized into three types,
such as series controllers, shunt controllers and combined
series-shunt controllers. In principle, the series controllers
inject voltage in series with the line and the shunt controllers
inject current into the system at the point of connection. The
combined series-shunt controllers inject current into the
system with the shunt part of the controllers and voltage in
series in the line with the series part of the controllers. Table
1 list the representative kind of FACTS devices with
controlled parameters.
In the case of voltage support, shunt FACTS devices,
such as STATCOM and SVC are typically used. This study
is focused on the steady state performance of multiple
1
The Authors are with the Department of Electrical Engineering, National
Sun Yat-Sen University, Kaohsiung 80424, Taiwan.
(e-mail: D953010011@student.nsysu.edu.tw)
2
Cong-Hui Huang is with the Department of Automation and Control
Engineering, Far East University, Tainan County 744, Taiwan.
3 reactive power and voltage magnitude of the system can be
Ting-Chia Ou is with the Institute of Nuclear Energy Research Atomic
Energy Council, Taoyuan 325, Taiwan.
978-1-4244-2853-3/09/$25.00 ©2009 IEEE
Authorized licensed use limited to: Rajalakshmi Engineering College. Downloaded on August 06,2010 at 08:25:52 UTC from IEEE Xplore. Restrictions apply.
STATCOM devices in the power system. Particularly,it is
desired to determine their optimal location and capacity.
Traditional optimization methods such as mixed integer
linear and non linear programming have been investigated to
address this issue; however difficulties arise due to multiple
local minima and overwhelming computational effort. In
order to overcome these problems, Evolutionary
Computation Techniques have been employed to solve the
optimal allocation of FACTS devices. This study, which
uses the Genetic algorithms (GA), has been tested for
finding the optimal location and capacity, with promising
results [3].
This paper applied the ability of the GA operated after
can promote the Ant Colony Optimization (ACO) efficiency.
The objective of GA is to improve the searching quality of
ants by optimizing themselves to generate a better result,
because the ants produced randomly by pheromone process
are not necessary better. This method can not only enhance
the neighborhood search, but can also search the optimum
solution quickly to advance convergence. [1] [4].
The load flow analysis (commonly called load flow or
power flow) is the basic tool for investigating power system
state variables, and it is very important part of the system
supervisory, planning and optimal operation. The unbalance
three-phase load flows based on the
Equivalent-Current-Inject (ECI) were applied successfully to
the distribution system. It is unable to apply the ECI model
to the high voltage transmission systems, because of the
voltage – controlled buses (PV Bus). In this a new power
flow approach based on ECI model and Cartesian
coordination is presented. PV Bus model were developed,
and according to the network characteristics, the decoupled
models were also proposed.
This paper introduces the application of GACO for
optimal location and capacity of a new STATCOM with ECI
model in the power system. It is organized as follows:
Section II STATCOM with ECI. Section III presents the
basic concepts of GA. In section IV the objective function to
be optimized is described. In section V simulation results are
presented. In section VI conclusions and future work are
given.
II. STATCOM with ECI MODEL
A. STATCOM Basic Concept
STATCOM is a second generation FACTS device used
for shunt reactive power compensation. The principle of
STATCOM is the reactive power compensation where the
adjusted such as shown in Fig. 1. It consists of three paths:
1915
shunt (coupling) transformer, voltage source converter B. STATCOM with ECI
(VSC), and capacitor. The reactive power is distributed in The STATCOM can act as on equivalent voltage source
the power system by the converter control [6~7]. series reactance. Voltage source can transform the current
source by way of Norton Theorem as shown in Fig. 2.

g ij + jbij Ij Vj

Zvr = gvr + jbvr

Vi ∠θi
Fig. 2 Transmission line equivalent model shunt STATCOM equivalent
Vs ∠θ s circuit diagram

According to Fig. 2 with ECI model inferential reasoning

as follows: Bus i and Bus j the beginnings and ends injection
current show in (3) and (4)
I ivr = I i + I vr
= (Vi − V j )( gij + jbij ) + Vi jbc +Vi ( g vr + jbvr ) (3)
Fig. 1 STATCOM connected to power system

where Vi∠θi : are the bus voltage and its phase angle of I j = (V j − Vi )( gij + jbij ) + jbcV j (4)
power system.
Vs∠θs : is the STATCOM voltage and its phase angle.
where V i = V i c o s θ i + j V i s in θ i

The SATCOM active P and reactive power Q are shown V j = V j cos θ j + j V j sin θ j
in (1) and (2).
VsVi
P= sin δ (1) Polar to Cartesian coordinate transformation is used in (5)
Xs
2 VV e + jf = V cos θ + j V sin θ (5)
V s i
Q= s − cos δ (2)
Xs Xs
where e : is the bus voltage real-part
where Xs : is coupling transformer equivalent reactance f : is the bus voltage imaginary-part
δ : θ s − θi Therefore I ivr , I j can be rewritten as show in (6) and (7)

The STATCOM is a combination of a voltage sourced

converter and an inductive reactance and shunt connected to Iivr = {gij (ei − e j ) − bij ( fi − f j ) − bc fi + gvr ei − bvr fi }
power system. The convert supplies leading current to the + j{gij ( fi − f j ) + bij (ei − e j ) + bc ei + gvr fi + bvr ei } (6)
AC system if the converter output voltage Vi is made to lead
the corresponding AC system voltage VS. Then it supplies
reactive power to the AC system by capacitive operation.
Conversely, the converter absorbs lagging current from the I j = {gij (e j − ei ) − bij ( f j − fi ) − bc f j }
AC system; if the converter output voltage Vi is made to lag
the AC system voltage VS then it absorbs reactive power to + j{bij (e j − ei ) + gij ( f j − fi ) + bce j } (7)
the AC system by inductive operation. If the output voltage
is equal to the AC system voltage, the reactive power
exchanges. Then Iivr, Ij divides into the real part, imaginary part and
respectively to differential state variable e i , e j , fi , f j .

1916

Authorized licensed use limited to: Rajalakshmi Engineering College. Downloaded on August 06,2010 at 08:25:52 UTC from IEEE Xplore. Restrictions apply.
We can obtain the new admittance matrix and equivalent
injection current with the relations shown in (8).
new
I = YmatrixV (8)

∆Iivr,r   gij + gvr −gij −(bij + bc + bvr ) bij   ∆ei 

 ∆I   −g g b −(b
 j,r  =  ij ij ij ij c   j 
∆Iivr,i  (bij + bvr + bc ) −bij gij + gvr −gij   ∆fi 
  
∆I gij  ∆f j 
 j,i   −bij (bij + bc ) −gij
III. Ant Colony Optimization plus Genetic Algorithm
The vast literature on metaheuristics tells us that a
promising approach to obtaining high-quality solutions is to
couple a local search algorithm with a mechanism to
generate initial solutions. In fact, in most search procedures,
the better the solutions quality returned, the higher the
computation time required. Such a coupling of solution
construction of ant colony system with local search of GA is
a promising approach for unit commitment problems. GA
has been successfully applied to a wide range of applications,
mainly in solving combination optimization problems. GA
improves a solution constructed by an ant. In the following,
we study how the performance of the ACO algorithms
improved when coupled with GA.
A common form of hybrid algorithm is the combination of
the local search and global search. ACO is good at global
search but slow to converge; GA is good at fine tuning but
often falls into local optima. In this paper, the hybrid
approach complements the properties of ACO and GA
search heuristic methods. Ant colony optimization is used to
perform global search to escape from local optima; Genetic
algorithms are used to conduct fine tuning. The procedure of
GACO is as follow:
Procedure: ACO plus GA
1. Input test data;
2. Set the parameters of GACO ;
3. Generated solutions based on the level of pheromone;
4. fitness evaluation;
5. Calculate selection probability to select a pair of ant to be
parents;
6. Apply a crossover and mutation to generate offspring;
7. Evaluate offspring ;
8. Select the better solution from parents and offspring;
9. Update the pheromon ;
10. If a prespecified stopping condition is satisfied, stop the
run; otherwise, return to step 3.
+ b )
 ∆e 
 

Fig 2 Hybrid Structure of Genetic Algorithms Combined with Ant Colony
Optimization
TABLE II : GACO PARAMETERS
The population of ants 30
Iteration times 100
Initial pheromone trail 30,00
The relative influence of the pheromone trail 1
The relative influence of the heuristic information 2
The pheromone evaporation rate 0.085
The chromosome crossover rate 0.5
The chromosome mutation rate 0.08
IV. OBJECTIVE FUNCTION
In such a power network, it is desirable to keep the
voltage deviations between ±5% to avoid voltage collapses
during faulty conditions. In general, if the load requirements
increase, the voltages at the corresponding buses may drop
below 0.95p.u. and consequently an additional voltage
support is needed at that particular bus. In this study, the
voltage support will be provided by a STATCOM with ECI
model, and its optimal location and capacity will be
determined by using GACO.
For instance, the IEEE 30-Bus system in Fig. 4 has 5
generators buses where voltage is regulated by the generator
AVRs. These generator buses do not need a STATCOM and
are omitted from the GACO search process. Also
considering the topology of the system, the bus numbers are
limited to the range from 1 to 30.
STATCOM selection to install the location principle
1. Because STATCOM are expensive, therefore the
minimum device installed is searched for economic
efficiency reasons.
2. Generator buses where voltages are regulated by the
generator do not need STATCOM installation.
3. Each bus is limited to the installation of one device.
Installing more does not represent a significant effect.
4. If the bus voltage is above 0.95 p.u., then STATCOM
is not installed.

According to the above discussion, candidates Bus are

shown in Table II.

1917

Authorized licensed use limited to: Rajalakshmi Engineering College. Downloaded on August 06,2010 at 08:25:52 UTC from IEEE Xplore. Restrictions apply.
 TABLE II where
CANDIDATE BUS ai, bi, ci: fuel cost coefficients of thermal plant
Bus 15 18 19 20 21 22 23 Vi : voltage magnitude at bus
Bus 24 25 26 27 28 29 30 ei , f i : Real and imaginary part of voltage Vi at bus i
PGi : Dispatchable active power at bus i
(i, j ) : Transmission line connecting buses i and j
S ij2 , S 2ji : Apparent power of transmission line (i,j) or (j,i)
S L2,ij , S L2, ji : Apparent power limit of transmission line (i,j) or
(j,i) , where S L2,ij = S L2, ji .
Vsm: The STATCOM voltage Vs at candidate bus m
θsm: The phase angle of STATCOM at candidate bus m

The final implementation of the GACO algorithm is

illustrated in the flow chart shown in Fig. 3.

V. SIMULATION RESULTS

A 30-Bus test system as shown in Fig. 4 is used for this

paper. The test system consists of 5 generators and 24 PQ
bus (or load bus). The simulation is run in MATLAB.
The problem to be addressed consists of finding the
optimal location (bus number) and power rating (MVA) of
STATCOM with ECI model. In this case the GA is able to
find different options for both location and capacity of the
STATCOM with the ECI model.
The solution found by GACO, in terms of bus location
and capacity for each STATCOM unit, is shown in Table III.

Fig. 3 Flowchart of GACO optimization

The objective function to minimize the system operation

cost is presented in (9)

NG
Min f(PGi ) = ∑ ai PGi + bi PGi + c (9)
2

i =1
subject to
 - I spec
r + I cal
r =0
 spec cal
 - I i + I i =0 (10)
- P + P + P cal = 0
 G 2 Load 2
 - VG + V cal = 0

S 2 ≤ S 2L ,ij
 ij Fig. 4 The 30-bus test system
2
 S ji2 ≤ S L ,ij
 (11) TABLE III
PGi ≤ PGi ≤ PGi
 V 2 ≤ (e 2 + f 2 ) ≤ V 2 SOLUTION FOUND BY GACO
 i i i i
STATCOM Location Capacity

VSm ≤ VSm ≤ VSm Unit (Bus number) (MVA)
 1 18 51.613
θ Sm ≤ θ Sm ≤ θ Sm
2 27 24.165

1918

Authorized licensed use limited to: Rajalakshmi Engineering College. Downloaded on August 06,2010 at 08:25:52 UTC from IEEE Xplore. Restrictions apply.
 TABLE V
The best solution is found by inserting STATCOM at ITERATIONS AND CPU TIMES BY GACO
GACO GA EP
bus 18 and bus 27. The power flow results, the voltage Tine(s) Max_time(s) 869 712 1169
comparison with and without STATCOM is show in Fig. 5 Min_time(s) 95 51 122
and Table IV. Table V show the comparison of GACO, GA Avg_time(s) 420 583 506
and EP. The system without the STATCOM has 17 buses CPUtime/count 18.4 18.33 16.96
with voltage below 0.95 p.u.. Once the STATCOM units are Cost Worst 774.7703 774.8686 775.5145
($/h) Best 774.7575 774.7575 774.7575
connected to buses 18 and 27 the voltage is improved. Average 774.7580 774.7600 774.7732
Count Max_count 44 39 69
Min_count 4 2 6
Voltage Comparison Avg_count 23.14 31.82 29.86
1.2

1
Additionally, In order to study the effect of the load
conditions in the optimal solution found by the GACO,
0.8
simulations are carried out by changing the load in each load
0.6 center in a range from 30% to 90%.
V
0.4 The Fig. 6 and Fig. 7 are the with and without
0.2
STATCOM voltage comparison for different load. In Fig. 6
and Fig. 7 is registered again a voltage improvement.
0
10 14 15 16 17 18 19 20 21 22 23 24 25 26 27 29 30

Bus
Voltage Comparison
without STATCOM
1.2
with STATCOM

1
Fig.5 Voltage comparison

0.8
TABLE IV
BUS VOLTAGE FROM POWER FLOW RESULT V 0.6
Bus Voltage p.u. w/o Voltage p.u. with
number STATCOM units STATCOM units 0.4

10 0.93760 0.99402
0.2
14 0.92612 0.98576
15 0.91577 0.99541 0
16 0.93683 0.98307 14 15 18 19 20 21 22 23 24 25 26 27 29 30
without STATCOM
17 0.92725 0.98056
Bsu with STATCOM
18 0.89363 1.05500
19 0.88846 1.01710 Fig. 6 Voltage comparison for 30% load
20 0.89860 1.00930
21 0.88873 0.96099
Voltage Comparison
22 0.89516 0.96705
23 0.89802 0.98157 1.2

24 0.89414 0.98194 1
25 0.89148 1.01800
0.8
26 0.85369 0.98311
27 0.90785 1.05700 0.6
V
29 0.85377 1.00760
0.4
30 0.82140 0.97714
0.2

0
9 10 12 14 15 16 17 18 19 20 21 22 23 24 25 26 27 29 30
Bus without STATCOM
with STATCOM

Fig. 7 Voltage comparison for 90% load

1919

Authorized licensed use limited to: Rajalakshmi Engineering College. Downloaded on August 06,2010 at 08:25:52 UTC from IEEE Xplore. Restrictions apply.
 The results obtained by the different load conditions are VI. CONCLUSIONS
shown in Table V and Table VI. Fig. 8 and Fig. 9 illustrate
the relationship between the different load conditions. From The paper has demonstrated the application of GACO
Table VI and Table VII, the capacity of the STATCOM does for location and capacity of STATCOM with ECI model in a
change under different load condition. power system considering at each bus the voltage constraints.
The study is carried out for the aim of voltage security
TABLE VI margin enhancement. Simulation results through an IEEE
LOCATION AND CAPACITY OF FIRST STATCOM FOR DIFFRTRNT 30bus validate the efficiency of the optimal location and
LOAD CONDITIONS capacity of STATCOM with the ECI model. The result
Load (%) Location (Bus) Capacity (MVA) shows the significantly voltage stability enhancement.
30 18 41.611
60 18 51.613
90 18 65.221 REFERENCES
[1] S. Sundhararajan, A. Pahwa, “Optimal selection of capacitors for
radial distribution systems using a genetic algorithm”, IEEE
Transactions on Power System, Vol. 9, Aug. 1994, pp. 1499-1507.
[2] Sung-Hwan Song, Jung-Uk Lim, and sung-Il Moon, ”FACTS
Operation Scheme for Enhancement of Power System Security”,
Power Tech Conference Proceedings, Vol. 3, June 2003, pp. 36 – 41.
[3] Y.D. Valle, J.C. Hernandez, G.K. Venayagamoorthy, R.G. Harley,
“Multiple STATCOM Allocation and Sizing Using Particle Swarm
Optimization”, Power Systems Conference and Exposition, Oct. 2006,
pp. 1884-1891.
[4] Marco Dorigo and Thomas Stutzle, "Ant Colony Optimization", 2004
Massachusetts Institute of Technology.
[5] Y. del Valle, J.C. Hernandez, G.K. Venayagamoorthy, and R.G. Harley,
“Optimal STATCOM Sizing and Placement Using Particle Swarn
Optimization”, Transmission & Distribution Conference, Latin
America, Aug. 2006, pp.1-6.
[6] N. Boonpirom, K. Paitoonwattanakij, ”Static Voltage Stability
Enhancement using FACTS”, The 7th International Power Engineering
Conference, Vol. 2, Dec 2005, pp. 711 – 715.
Fig. 8 STATCOM capacity for different load condition [7] A Karami, M. Rashidinejad, and A.A Gharaveisi, ”Optimal Location
of STATCOM for Voltage Security Enhancement via Artificial
Intelligent”, IEEE International Conference on Industrial Technology,
TABLE VII Dec 2006, pp. 2704 – 2708.
LOCATION AND CAPACITY OF SECOND STATCOM FOR [8] Whei-Min Lin, Cong-Hui Huang, Tung-Sheng Zhan, “A Hybrid
Current-Power Optimal Power Flow Technique”, IEEE Transactions
DIFFRTRNT LOAD CONDITIONS
on Power System, Vol. 23, Feb. 2008, pp. 177-185
Load (%) Location (Bus) Capacity (MVA)
30 27 22.729
60 27 24.165
90 27 25.265

Fig. 9 STATCOM capacity for different load condition

1920

Authorized licensed use limited to: Rajalakshmi Engineering College. Downloaded on August 06,2010 at 08:25:52 UTC from IEEE Xplore. Restrictions apply.