Sunteți pe pagina 1din 12

International Journal of Software Engineering and Its Applications

Vol. 5 No. 1, January, 2011

Genetic Algorithm for Solving Optimal Power Flow


Problem with UPFC

Vijayakumar Krishnasamy
SRM University
kvijay_srm@rediffmail.com

Abstract
This paper concerns the Optimal Power Flow [OPF] in multi machine Power System with
UPFC using Genetic Algorithm [GA]. The objective is to minimize the cost of the power
system, to enhance the power flow in transmission lines and to maintain the voltages at the
buses using UPFC. Using the proposed method, the optimal cost and real power losses of the
power system with UPFC is achieved by developing a simple Genetic Algorithm and the
location and rating of UPFC is also achieved by Newton Raphson’s load flow method. IEEE
9 bus system has been studied to show the effectiveness.

Keywords: Genetic Algorithm (GA), Optimal Power Flow (OPF), Newton Raphson’s
Power Flow, Unified Power Flow Controller (UPFC).

1. Introduction
In present days with the deregulation of electricity market, the traditional practices of
power system have been completely changed. Better utilization of the existing power system
resources to increase capabilities by installing FACTS devices [2] with economic cost
becomes essential.
The parameters such as transmission line impedances, terminal voltages and voltage angle
can be controlled by FACTS devices in an efficient way. The benefits brought about FACTS
include improvement of system dynamic behavior and enhancement of system reliability.
However their main function is to control of power as ordered [4].
A few research works [1, 5] were done on the impact of FACTS devices on improving
static performance of the power system. There is also a great need for studying the impact of
FACTS devices on optimal power flow. The investment costs of FACTS devices and their
impact on the power generation cost are also reported [8]. Many researches were made on the
optimal choice and the location of FACTS devices [3, 6].
The objective of this paper is to develop an algorithm to simultaneously find the real
power allocation of generators and to choose the best location of UPFC such that overall cost
function which includes the generation cost of power plants and investment cost of UPFC is
minimized. The combinatorial analysis is solved using Genetic Algorithm and Newton
Raphson’s load flow method.
This paper is organized as follows: following the introduction the static model of UPFC is
described in section 2. In section 3, the cost function used in this paper is given. In section 4,
optimal power flow with UPFC is presented. In section 5 Genetic Algorithm for optimal
power flow with UPFC is discussed. In section 6 the results are discussed. Finally the
conclusions are deduced.
2. Static Model of UPFC

39
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

Among the various FACTS devices, Unified Power Flow Controller (UPFC) is considered
in this paper. The detailed model of UPFC is discussed below.
The UPFC is a combination of shunt and series controller. It has three controllable
parameters namely, the magnitude of the boosting injected voltage (UT), phase of this voltage
(ØT) and the exciting transformer reactive current (Iq)
When an UPFC is installed in the power system as depicted in figure 1 i.e. the exciting transformer
of UPFC is directly connected to bus l. In figure.1, Zlm and Ylm denote the parameters of transmission
line 1-m. Y1 and Ym denote the respective shunt admittance for bus l and bus m.

Figure 1. UPFC connected between node l and m, with exciting transformer at node l
When the UPFC is placed in the transmission line connected between node l and m, the
load flow equations can be expressed as follows [8].

PGi − PLi = ∑ U iU j (Gij cos δ ij + Bij sin δ ij )


j∈i
(1)

QGi − QLi = ∑U iU j (Gij sin δ ij − Bij cos δ ij )


j∈i (2)
i = 1, 2 . . . . n; but i ≠ l, m

PGl − PLl = ∑ U lU j (Glj cos δ lj + Bil sin δ lj ) + ∆Pl


j∈i
(3)

QGl − QLl = ∑ U lU j (Glj sin δ lj + Blj cos δ lj ) + ∆Ql


j∈i
(4)

PGm − PLm = ∑UmUj (Gmj cosδmj + Bmj sinδmj) +∆Pm


j∈i
(5)

QGm − Q Lm = ∑ U mU j (G mj sin δ mj − Bmj cos δ mj ) + ∆Qm


j∈i (6)
Where n is the total number of nodes of the power system:
PGi, QGi, PLi and QLi (∀ i) are the respective real and reactive power of generator and load
of node i.
Ui δi (∀ i, j) are respective magnitude and phase angle of the voltage of node i.
jεi signifies that bus j is connected to bus i;
δij = δi-δj (∀ i, j)

40
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

Gij and Bij (∀ i, j) are the respective real part and imaginary part of Yij which represents the
elements of the network admittance matrix. Here Yij is exactly the same as that of the
network without UPFC.
∆Pl, ∆Ql, ∆Pm and ∆Qm are the modified items due to the added UPFC.

The formulae of modification can then be written as


δm −φT ) −Bsin(δm −φT )]+GFUT2 +2UlUTGF cos(δl −φT )
∆Pl =−UmUT[Gcos(
(7)
∆Ql = U mU T [G F sin(δ l − φT ) − B F cos(δ l − φT )] − U l I q
(8)
∆Pm = −U mU T [G cos(δ m − φT ) + B sin(δ m − φT )] (9)
∆Qm = −U mU T [G sin(δ m − φT ) − B cos(δ m − φT )] (10)
Where G + jB = 1/Z1m; GF = g1m + G;
BF = b1m + B; Ylm = glm + jblm;
The injected voltage of UPFC has a maximum voltage magnitude of 0.1Vm ,where Vm is
the rated voltage of the transmission line where UPFC is installed. The angle of the UPFC can
be varied from -180 to +180 degrees.

3. Cost function
The objective of this paper is to find simultaneously the optimal generation, optimal
location and rating of UPFC so as to minimize the overall cost function, which comprises of
generation cost and investment costs of UPFC.

3.1 Generation cost function


The generation cost function is represented by a quadratic polynomial as follows:
C2 (PG) = α0+α1PG+α2PG2 (11)
Where PG is the output of the generator (MW), and α0, α1 and α2 are cost coefficients.

3.2 Investment costs function of FACTS devices


Based on the Siemens AG Database, the cost functions or TCSC and UPFC are developed
The cost functions for UPFC and TCSC are:
C1UPFC = 0.0003S2 – 0.2691S + 188.22 (US$ / kVar) (12)
C1TCSC = 0.0015S2– 0.7130S+153.75 (US$ / kVar) (13)
Where C1UPFC and C1TCSC are in US$/kVar and S is the operating range of the FACTS
devices in MVar.
The cost function for TCSC and UPFC are shown in figure 2

41
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

Figure 2. Cost Function of UPFC

4. Optimal Power Flow with UPFC


The formulation of the optimal allocation of UPFC can be expressed as follows
Minimize CTotal = C1 (f) + C2 (PG) (14)
Subjected to E (f, g) = 0 (15)
B1 (f) < b1, B2 (g) < b2 (16)
Where
CTotal : the overall cost objective function which includes the average investment costs of
FACTS devices C1 (f) and the generation cost C2(PG).
E (f.g): the conventional power flow equations.
B1 (f) and B2 (g) are the inequality constraints for FACTS devices and the conventional
power flow respectively.
f and PG are vectors that represent the variables of FACTS devices and the active power
outputs of the generators.
g represents the operating state of the power system.
The unit for generation cost is US$/Hour and for the investment cost of FACTS devices
are US$. They must be unified into US$/Hour. Normally the FACTS devices will be in
service for many years. However only a part of its life time is employed to regulate the power
flow. In this paper three years is employed to evaluate the cost function. Therefore the
average value of the investment costs are calculated as follows
C 1 (f) = C (f)/ {8760 x 3} (17)
As mentioned above, power system parameters can be changed using UPFC. These
different parameters derive different results on the objective function. Also, the variation of
UPFC locations and ratings also influences on the objective function. Therefore, the
conventional optimization methods are not easy to find the optimal location of UPFC, and
control parameters simultaneously. To solve this problem, genetic algorithm is employed
here.
5. Algorithm for Optimal Power Flow with UPFC

42
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

GAs are global search techniques based on the mechanism of natural selection and
genetics. Without any prior knowledge of the objective function they can search several
possible solutions simultaneously. GAs are best suited for complex problems. Moreover it
produces high quality solution.
GA start with random generation of initial population and then the selection, crossover and
mutation operations are carried out until best population is found. Gas are simple and
practical algorithm and easy to be implemented in power system.

5.1. Encoding
The objective is to find simultaneously the optimal generation and optimal location of
UPFC subjected to equality and inequality constraints. Therefore the configuration of UPFC
is encoded by three parameters: active power outputs of generator, location and rating of
FACTS devices. The first value of each string corresponds to the active power outputs of
generator, second value represents the location the third value represents the rated value of
each UPFC. This value ranges between -1 and +1. The real value of each UPFC is then
converted accordingly.

5.2. Initial Population


For each population string, the first value represents a set of generators real power output
which is randomly selected. The second value of each string represents the location of UPFC
in the transmission line which is also randomly selected among the existing number of
transmission lines in the system. The third value represents the rating of the UPFC again
randomly selected between -1 and +1.
The above operation is repeated for the entire initial population.

5.3. Decoding
The parameters of the initial population are then decoded to actual values.
Then for a given load demand the Newton Raphson’s power flow is performed [8]

 H N   ∆δ   ∆P 
 J L  ∆V / V  =  ∆Q  (18)
    
After convergence the voltage magnitude and phase angles of the bus voltages are known. Using these
the real power loss PL is calculated using the equation (19)
n n

∑∑ (P α P + Q α Q
i =1 j=1
i ij j i ij j - Piβ ijQj + Pjβ ijQi)
PL = (19)
rij cos(δ i - δ j )
vi . vj
αij = (20)

43
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

r ij sin (δ i - δ j )

vi . vj
βij = (21)
Where
rjk are the real components of the elements of the bus impedance matrix
n is the number of buses
Pi is the real power at bus i
Qi is the reactive power at the bus i
Vi is the magnitude of the voltage at bus i

5.4. Fitness function


After encoding, the objective function (fitness) is evaluated for each individual of the
population. The fitness is a measure of quality which is used to compare different solutions.
In this paper fitness is defined as follows
1
N
Fitness = CTotal + W ( ∑ PGi − PDi − PLi ) (22)
i
Since the Gas can only find the maximum value of the objective, so inverse function is
selected to convert the objective function into a maximum one.
Then reproduction, crossover and mutation are applied successively to generate the
offspring.

5.5. Reproduction
Reproduction is a process where the individual is selected to move to a new generation
according to their fitness. The biased roulette wheel selection is employed. The probability
of an individual’s reproduction is proportional to its part on the biased roulette wheel.

5.6. Crossover
The main objective of crossover is to reorganize the information of two different
individuals and produce a new one. A single point crossover is applied and probability of
crossover is selected as 1.0

5.7. Mutation
Mutation is used to introduce some sort of artificial diversification in the population to
avoid premature convergence to local optimum.
The above-mentioned operations of selection, crossover and mutation are repeated until the
best individual is found.

44
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

6. Results
A VC++ coding is developed for Genetic Algorithm. In order to verify the effectiveness of
the proposed method IEEE 9 bus system is used. Different operating conditions are
considered for finding the optimal location and rating UPFC. It has also been found that by
using UPFC the power flow in the transmission line is increased, the bus voltages are also
improved and the total real power losses in the system is reduced
The total population size is selected as 150, the mutation probability as 0.01 and crossover
probability as 1.0.

Case 1
For the normal loading of IEEE 9 bus system it has been found that No UPFC is required.
The generators outputs are 202MW, 31.6MW, 86.4MW, 204MW, 19 MW and 75MW
respectively.

Case 2
When the loading at bus 5 is increased by three times and removing the generation at bus
3, it has been found that UPFC is selected in the transmission line 4 and VAR compensation
required is 156.02 MVAR. The generators outputs are 245.3 MW, 263.6 MW, 0MW,
250MW, 223MW and 183MW respectively.
Different loading conditions are considered and the results are shown in Table.1

Table 1. Optimal Location & Rating of UPFC


Bus Loading Compensatio Location Remarks

(MVAR)

5 Normal Loading -- -- --

5 Twice Normal loading 10.166 Line 3 Line 3 connects buses 5 & 6

5 3 Times Normal loading 156.02 Line 4 Line 4 connects buses 3 & 6

& without generation at

bus 3

9 3 Times normal loading 96.038 Line 2 Line 2 connects buses 4 & 5

9 Twice normal loading 47.52 Line 6 Line 6 connects buses 7 & 8

with generation at bus

45
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

Case 3
The comparison of line flows in each of the lines and voltages at each bus without and with
UPFC for one of the ΦT value under twice the normal loading condition in the 5th bus is given
in the Table 2. and Table 3. When the UPFC is embedded between the 5th and 6th bus the
voltage profile at bus 5 has been improved from 0.9741 to 1.0100.

Case 4
The optimal real power generation for minimum loss are given in Table 4 and the
minimum loss is found out with and without UPFC under normal loading and at the twice the
normal loading at 5th bus. The results prove that with UPFC, the real power losses can be
reduced considerably by choosing appropriate control parameters and location of UPFC. In
this case UPFC is connected between Buses 5 and 6.

Table 2. Line flows with & with out UPFC


Line Start End Complex Line Flows

No. Bus Bus Without UPFC With UPFC Between 5 & 6

1 1 4 4.637 0.0373 1.626 -0.127

2 4 5 1.104 -0.001 1.1468 -0.211

3 5 6 -0.71 -0.403 0.8423 -0.160

4 3 6 0.850 -0.341 0.85 -0.194

5 6 7 0.113 -0.172 0.169 0.2217

6 7 8 -0.88 -0.089 -0.831 -0.146

7 8 2 -1.63 0.4139 -1.63 0.4043

8 8 9 0.737 -0.248 0.7928 -0.291

9 9 4 -0.52 -0.211 -0.476 -0.265

The graphical comparison of the outputs for line flows in various lines and voltage
magnitudes at all the buses with and without UPFC for the increasing the loadings at bus 5 by
two times are given in the figure 3 and 4

46
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

Table 3. Voltages (Twice Normal Loading)

No UPFC With UPFC Between 5 & 6

Bus Voltage Angle Voltage Angle

No. (p.u) (radians) (p.u) (radians)

1 1.0000 0.0000 1.0000 0.0000

2 1.0000 0.0762 1.0000 0.093

3 1.0000 -0.0267 1.0000 0.0007

4 1.0023 -0.0942 1.0117 -0.0927

5 0.9741 -0.1957 1.0100 -0.1978

6 1.0212 -0.0755 1.0126 -0.0485

7 1.0155 -0.0861 1.0115 -0.0652

8 1.0204 -0.0238 1.0198 -0.0071

9 0.9927 -0.1391 0.9985 -0.1319

LINEFLOW VS LINE
1.8
1.6
1.4
POWERFLOW

1.2
1
0.8
0.6
0.4
0.2
0
1,4 4,5 5,6 3,6 6,7 7,8 8,2 8,9 9,4
LINE
WITHOUT UPFC
WITH UPFC

Figure 3. Line flow in various lines

47
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

VOLTAGE VS ANGLE
1.04
1.02

VOLTAGE
1
0.98
0.96
0.94
1 2 3 4 5 6 7 8 9
ANGLE WITHOUT UPFC
WITH UPFC

Figure 4. Voltage magnitude at various buses

Table 4. Real Power Loss


Loading Without UPFC With UPFC

Generation Losses Generation Losses

(MW) (MW) (MW) (MW)

Under normal Pg1=141.7 3.72 Pg1=167.1 2.7422

loading Pg2=107.8 Pg2=69.13

Pg3=69.13 Pg3=82.39

Twice normal Pg1=187.8 6.55 Pg1=223.6 5.62

loading at bus 5 Pg2=66.86 Pg2=83.92

Pg3=156.8 Pg3=102.7

7. Conclusions
In this paper, a genetic algorithm based optimal power flow is proposed to determine the
optimal location and rating of the UPFC in power systems and also to simultaneously
determine the active power generation for different loading condition. The overall system cost
function which includes generation cost of power plants and the investment costs of UPFC is
employed to evaluate the power system performance.
It has been verified that by using UPFC the power flow in transmission lines are improved,
the voltage magnitudes are increased and also the real power losses in the system are
minimized.
The proposed algorithm is an effective and practical method for the location of UPFC.

48
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

References
[1] Chung T.S and Li Y. Z, A Hybrid GA for OPF with consideration of FACTS Devices, IEEE Power
Engineering Review, 2001, pp. 47-57.
[2] Duan X , Chen J, Peng F, Luo Y , Huang Y, Power Flow Control with FACTS Devices, IEEE Trans. Power
Systems, 2000, pp. 1585-1589.
[3] Gerbex. S, Cherkaoui R and Germond A.J , Optimal Location of Multiple Type FACTS Devices in a Power
System by means of Genetic Algorithm, IEEE Trans. Power Systems, Vol. 16, 2001, pp. 537-544.
[4] Gyugyi L, Schauder C.D, Williams S. L, et al, The Unified Power Flow Controller : A New Approach to
Power Transmission Control, IEEE Trans. Power Delivery, Vol. 10, No. 2, 1995, pp. 1085 - 1097.
[5] Lie T.T and Deng W, Optimal Flexible AC Transmission Systems (FACTS) Devices Allocation, Electrical
Power & Energy Systems, Vol. 19, No. 2, 1997, pp. 125-134.
[6] Patemi P, Vitet S, Bena M, and Yokoyama A, Optimal Location of Phase Shifters in French Network by
Genetic Algorithm, IEEE Trans. Power Systems, Vol. 14, 1999, pp. 37-42.
[7] Gyugyi L, Schauder C.D, Williams S. L, et al, Optimal Placement of FACTS Controller in Power System by
a Genetic Based Algorithm, in : IEEE International Conference on Power Electronics and Drive and Systems,
1999, Honkong .
[8] Wanliang F, Ngan H. W, Extension of Newton Raphson’s Load Flow techniques to cover Multi Unified
Power Flow Controllers , in : Proceedings of the 4th International Conferences on Advances in Power System
Control Operation and Management, 1995, Honkong.

Author

Vijayakumar krishnasamy completed B.E(Electrical and Electronics


Engineering) and M.E (Power Systems) at Annamalai University. Currently
working as Assistant professor in SRM University. Research areas of interest
are Power system optimization, Deregulation, operation & control, FACTS
and so on.

49
International Journal of Software Engineering and Its Applications
Vol. 5 No. 1, January, 2011

50

S-ar putea să vă placă și