## Dr. K. Balamurugan

Abstract: At present, there is a rising load demand for electric power which can lead to major problems, such as voltage collapse, blackout, grid outage. The most important constraint in the power systems is a huge transmission capability joint through a voltage shape surrounded by the boundary at which consumers function. To overcome from this problem reactive power planning are carried out. The reactive power scheduling problem is modelled as a solo purpose optimization solution where the active power losses to be reduced.

VAR scheduling is the best possible distribution of reactive control devices like FACTS Devices (SVC) considering location and size. Firefly Algorithm finds the global finest solution in the course of two processes: Attractiveness and Brightness. In the proposed problem, FA has been implemented to find the best possible distribution of SVC devices for reducing the real power losses. A case study with an IEEE fourteen& thirty bus systems are carried out to demonstrate the effectiveness of the proposed technique.

Index Terms: Reactive power planning, FACTS Devices, Line Power Flow, SVC optimal locations, Active power loss, Firefly Algorithm.

## I. INTRODUCTION

Reactive power scheduling is the best composite optimization solution of power system network. It is essential for powerful and dynamic management of reactive power production by the all reactive power devices having in the power system. The main causes of reactive power are synchronous machines; tap varying transformers, static condensers, and FACTS devices. Reactive power planning is an optimization solution mostly associated by the deceasing of real power loss. At the present suppose FACTS controller such as SVCs are carried concurrently by the conventional reactive power controller there in the system, not only the transmission loss minimizes but also reasonable development of the voltage outline is observed throughout the power system. Therefore to determine the reactive power scheduling by the whole sources, as a result to formulate a optimization solution. An optimal Reactive Power Planning (RPP) of power system is discussed in [1]. The optimal placement multi-type FACTS controllers and their parameters are described in [2]. Modeling of SVC devices is studied in [3]. Loss minimization equations with equality and inequality constraints were described in [4]. To increase the load ability of power systems were discussed in [5]. Location of various types of FACTS controllers using Genetic Algorithm are in [6]. Background studies of reactive power planning for transmission loss reduction are mentioned in [7]. Obtained result for optimal reactive power flow using BBO technique [8]. Reactive power scheduling problem are solved to maintain the voltage profile in the system using various techniques like BBO, and cuckoo search algorithm [8-9]. Comparative analysis between GA and EP for the most favorable generator bus voltages, tap varying transformer, VAR supply, power production and power losses are carried out [10]. In the present work, Firefly algorithm is used for placing the SVC controllers in the system in order to minimize the transmission losses.

### II. STATIC VAR COMPENSATOR (SVC)

SVC is a shunt related static VAR compensation whose output is in tune to transfer capacitive or inductive current so as to sustain or manage particular parameters of the electrical power system. Fundamentally an SVC consists of a mixture of fixed capacitors or reactors, Thyristor Switched Capacitors (TSC) and Thyristor Controlled Reactors (TCR) joined in parallel with the power system network.

Fig. 1. Schematic Diagram of Static VAR Compensator

## A. Modeling of SVC

SVCs are placed in buses, and modeling of the SVC is done to inject or absorb reactive power in the bus. The change in reactive power at bus-i with SVC can be represented as in equation (1).

?Qi = QF (1)

In the initial case it absorbs reactive power while in the subsequent the reactive power is injected. The SVC is modeled with two ideal switching elements in parallel: a capacitance and an inductance. The rate of reactive power injected or absorbed, Qc, is restricted between lowest and highest choice as given in equation (2).

?50 < Qc < 100 MVar (2)

### III. PRABLEM FORMULATION

The objective of the proposed work was to minimize the transmission loss of the power network using SVC. Problem in raising transmission loss as well as voltage profile is the key concern with the large power demand. So, when the power system demand is rising steadily, it needs reactive power support to keep the voltage stability. Therefore the core objective of the proposed problem was to minimize the active power loss which is given in Equation (3) and to reducing voltage fluctuation at weak buses.

(3)

### A. Equality Constraints

Nodal active and reactive power balance are given in equation (4) and (5).

(4)

(5)

## Where

Pgi , Pdi are the real power production and load demand at its transmission lines. Qgi, Qdi are the reactive power production and load demand at its transmission lines. G and B are the conductance and susceptance of the transmission line respectively.

### B. Inequality Constraints

A voltage magnitude constraint is given in equation (6).

(6)

Nodal reactive power generation constraints for SVC is given in equation (7)

(7)

## IV. FIREFLY ALGORITHM

Firefly Algorithm is an environment stimulated, Meta heuristic optimization algorithm which is concerned on the collective blinking actions of fireflies. This Algorithm was first initiated by Xin-She Yang [11] in 2010. It is analogous to another optimization algorithm such as particle swarm optimization and artificial bee colony algorithm. Firefly Algorithm is depends on the three idealized process. 1st entire fireflies were unisex thus single firefly will be good look to another firefly in spite of their sex. 2nd good look is related to their glow intensity, therefore for some two blinking fireflies, the lesser brighter fireflies will be travel in the direction of the brighter fireflies. The attractiveness is proportional to the glow intensity and they together reduce as their distance raise. If there is no brighter firefly than a specific firefly, it will travel erratically. 3rd The glow intensity of a firefly is exaggerated or calculated by the background of the fitness function. For a maximum fitness solution, the glow intensity can basically be related to the amount of the fitness solution through dissimilar set of initial points. However, in every situation, the algorithm determined for the best possible solution.

## A. Flowchart

Fig. 2. Flowchart for firefly algorithm

The glow intensity of the mth firefly, LIm is shown in equation (8).

### LIm = Fitness (xm) (8)

The good looking between the mth and nth firefly, ?m,n is shown in equation (9).

(9)

Where rm,n is cartestian space between mth and nth firefly is given in equation (10).

=|| ||= (10)

The location of firefly is restructured in every iterative movement. If the glow strength of nth firefly is bigger than the glow strength of the mth firefly, then mth firefly travels in the direction of the nth firefly and its movement at kth iteration is defined by the equation (11).

(11)

### V. CASE STUDY AND SIMULATED OUTPUT

### A. Case Study (IEEE 14 Bus System)

In order to prove its effectiveness, the FA algorithm is implemented to best location of SVC controllers on the IEEE fourteen bus systems. The fourteen bus system contains 5 generator buses and 11 loads buses. The one line picture for IEEE fourteen bus system is shown in Fig. 3.

Fig. 3. One Line Picture for IEEE Fourteen Bus System

## Without Facts Devices

In this case, the simulated output for conventional method without considering the FACTS Devices is shown in Table I it prevails that active power losses are 13.593MW for IEEE 14 bus systems. From load flow using Newton-Raphson method.

Table I: Simulation Output for conventional method

Test Systems Total Active Power Losses

### IEEE-14 Bus System 13.593MW

## With One SVC

The simulated result for the case of one SVC is connected in bus 7 using Firefly algorithm is shown in Table II. From the table it is inferred that the overall real power loss is decreased to 13.4184 MW as compared with without FACTS device loss (13.593MW) Convergence graph for placing one SVC is shown in Fig. 4.

Table II: Simulated output with one SVC

Analysis Location Q setting (Mvar) Active power loss (MW)

### Without Facts devices – – 13.593

### With one SVC Bus No. 7 -1.009 13.4184

### Fig. 4. Output waveform for one SVC

## With Two SVC

The simulated result for the case of two SVC is connected in buses 7 & 9 using Firefly algorithm is shown in Table III. From the table it is observed that the overall real power loss is decreased to 13.3544 MW as compared with without FACTS device loss (13.593MW) Convergence graph for placing two SVCs is shown in Fig. 5.

Table III: Simulated output with one SVC

Analysis Location Q setting (Mvar) Active power loss (MW)

### Without Facts devices – – 13.593

### With two SVC Bus No. 7

### Bus No. 9 12.250, 21.981 13.3544

### Fig. 5. Output waveform for Two SVC

## With Three SVC

The simulated result for the case of three SVC is connected in buses 9, 5 & 7 using Firefly algorithm is shown in Table IV. From the table it is observed that the total real power loss is decreased to 13.3121MW as compared with without FACTS device loss (13.593MW) Convergence graph for placing three SVCs is shown in Fig. 6.

Table IV: Simulated output with three SVC

Analysis Location Q setting (Mvar) Active power loss(MW)

### Without Facts devices – – 13.593

### With three SVC Bus No. 9

## Bus No. 5

Bus No.7 20.743, 10.657, -18.099 13.3121

### Fig. 6. Output waveform for Three SVC

### B. Case Study (IEEE 30 Bus System)

In order to prove its effectiveness, the FA algorithm is implemented to best location of SVC controllers on the IEEE thirty bus systems. The thirty bus system contains 6 generator buses and 22 load buses. The one line picture of the IEEE thirty bus test systems is shown in Fig. 7.

Fig. 7. One Line Picture for IEEE Thirty Bus System

## Without FACTS Devices

In this case, the simulated output for conventional method without considering the FACTS Devices is shown in Table 5 it prevails that active power losses is 17.579 for IEEE 30 bus system. From power flow solution using NR method, the voltage level of 21st, 7th, 17th and 15th buses are decreased as compared to the other buses.

Table V: Simulation Output for conventional method

Test Systems Total Active Power Losses

### IEEE-30 Bus System 17.579MW

## With One SVC

The simulated result for the case of two SVC is connected in bus 21 using Firefly algorithm is shown in Table VI. From the table it is observed that the total real power loss is decreased to 17.3509MW as compared with without FACTS device loss (17.652MW). Convergence graph for placing one SVC is shown in Fig. 8

Table VI: Simulated output with one SVC

Analysis Location Q setting (Mvar) Active power loss(MW)

### Without Facts devices – – 17.652

With one SVC Bus No. 21 20.170 17.3509

### Fig. 8. Output waveform for one SVC

## With Two SVC

The simulated result for the case of two SVC is connected in bus 21 & 4 using Firefly algorithm is shown in Table VII. From the table it is observed that the total real power loss is decreased to 17.2610MW as compared with without FACTS device loss (17.652MW). Convergence graph for placing two SVC is shown in Fig. 9.

Table VII: Simulated output with two SVC

Analysis Location Q setting (Mvar) Active power loss(MW)

### Without Facts devices – – 17.652

### With two SVC Bus No.21

### Bus No. 4 18.232, 31.132 17.2610

### Fig. 9. Output waveform for two SVC

## With Three SVC

The simulated result for the case of two SVC is connected in buses 3, 19 & 21 using Firefly algorithm is shown in Table VIII. From the table it is observed that the total real power loss is reduced to 17.2549MW as compared with without FACTS device loss (17.652MW). Convergence graph for placing two SVC is shown in Fig. 10.

Table VIII: Simulated output with three SVC

### Analysis Location Q setting

(Mvar) Active power loss (MW)

### Without Facts devices – – 17.652

### With three SVC Bus No. 3

## Bus No. 19

## Bus No. 21 23.141,

6.865, 17.393 17.2549

Fig. 10. Output waveform for three SVC

C. Comparative Analysis of Voltage Magnitude

Table IX: Simulated Output for Voltage Magnitude

(IEEE 14 bus system)

Bus No Voltage Magnitude (Before installation of SVC) Voltage Magnitude (after installation of 3 SVC)

1 1.0600 1.0600

2 1.0450 1.0450

3 1.0100 1.0100

4 1.0132 1.0161

5 1.0166 1.0212

6 1.0700 1.0700

7 1.0457 1.0438

8 1.0800 1.0800

9 1.0305 1.0440

10 1.0299 1.0411

11 1.0461 1.0518

12 1.0533 1.0543

13 1.0466 1.0486

14 1.0193 1.0279

In this condition, the result of voltage magnitude is compared with the before and after installation of three SVCs in IEEE 14 and 30 bus system is shown in Table IX, X.

Table X: Simulated Output for Voltage Magnitude

(IEEE 30 bus system)

Bus No Voltage Magnitude (Before installation of SVC) Voltage Magnitude (after installation of 3 SVC)

1 1.0600 1.0600

2 1.0430 1.0430

3 1.0271 1.0414

4 1.0196 1.0288

5 1.0100 1.0100

6 1.0134 1.0183

7 1.0043 1.0072

8 1.0100 1.0100

9 1.0235 1.0360

10 1.0074 1.0304

11 1.0820 1.0820

12 1.0259 1.0382

13 1.0710 1.0710

14 1.0099 1.0267

15 1.0048 1.0254

16 1.0107 1.0276

17 1.0029 1.0242

18 0.9931 1.0214

19 0.9895 1.0222

20 0.9931 1.0235

21 0.9932 1.0245

22 0.9993 1.0228

23 0.9933 1.0231

24 0.9883 1.0123

25 1.0015 1.0189

26 0.9836 1.0012

27 1.0185 1.0315

28 1.0095 1.0143

29 0.9986 1.0118

30 0.9871 1.0004

## VI. CONCLUSION

In this proposed work, the Firefly algorithm has to be effectively applied to determine the Reactive Power Planning solution for reduction of real power loss. This concept has been implemented and analysed on IEEE 14 & 30 bus systems to demonstrate its effectiveness. It has been concluded that the FA have the capability to minimize the real power loss considerably without exceeding the any constraints limits. Moreover, FA possesses excellent convergence characteristics and robustness. Therefore, from the simulation results it may be concluded that the FA is superior in terms of solution quality, computational efficiency and robustness for solving Reactive Power Planning problems.

## REFERENCES

1. Biplab Bhattacharyya, Sanjay Kumar, Reactive power planning with FACTS devices using gravitational search algorithm. Elsevier, Feb. 2015, pp. 1 7.

2. Z. Lu, M. S. Li, W. J. Tang, Q. H. , Optimal Location of FACTS Devices by Bacterial Swarming Algorithm for Reactive Power Planning, IEEE Congress on Evolutionary Computation, 2007.

3. B Chandra Sekhar, N Visali, Optimal Placement of SVC with Cost Effective Function Using Particle Swarm Optimization, International Journal of Emerging Trends in Engineering Research,, Vol. 1, No.2, October 2013, pp. 1 7.

4. Biplab Bhattacharyya, Vikash Kumar Gupta And S.Das, Evolutionary Programming for Reactive Power Planning Using FACTS Devices, Wseas Transactions On Power Systems, Vol. 9, 2014.

5. A. B.Bhattacharyya, B. S.K.Goswami, Optimal Placement of FACTS Devices byGenetic Algorithm for the Increased Load Ability of a Power System, World Academy of Science, Engineering and Technology, Vol. 5, 2011.

6. Biplab Bhattacharyya, Vikash Kumar Gupta, Sanjay Kumar, Reactive Power Optimization with SVC & TCSC using Genetic Algorithm, Power Engineering And Electrical Engineering, Vol. 12, 2014.

7. A. M. Ramly, N. Aminudin, I. Musirin, D. Johari, N. Hashim, Reactive Power Planning for Transmission Loss Minimization, The 5th International Power Engineering and Optimization Conference, June 2011.

8. Aniruddha Bhattacharya, Pranab Kumar Chattopadhyay, Solution of Optimal Reactive Power Flow using Biogeography-Based Optimization, International Journal of Electrical, Computer, Energetic, Electronic and Communication Engineering, Vol. 4, No. 3, 2010.

9. K.R.Vadivelu And Dr.G.V. Marutheswar, Optimal Reactive Power Planning for LossReduction and Improvement of Voltage Profile Using Cuckoo Search Algorithm, International Electrical Engineering Journal (IEEJ), Vol. 6, No. 2, 2015, pp. 1780 1786.

10. S.K.Nandha Kumar, and P.Renuga, Reactive Power Planning using Real GA Comparison with Evolutionary Programming, ACEEE International Journal on Electrical and Power Engineering, Vol. 1, No. 1, Jan 2010.

11. Xin-She Yang, Firefly Algorithm, L?evy Flights and Global Optimization, Research and Development in Intelligent Systems, XXVI (EdsM. Bramer, R. Ellis, M. Petridis), Springer London, 2010, pp. 209-218.

## AUTHORS PROFILE

Dr. K. Balamurugan received M.E and MBA from Thiagarajar College of Engineering and Madurai Kamaraj University, Madurai, India in 2002 and 1999. He is an Associate Professor in the department of EEE, Dr. Mahalingam College of Engineering and Technology, Pollachi, India. He has more than 18 years of teaching experience. He has completed PhD on Optimal location of FACTS devices in deregulated power system from SASTRA University, Thanjavur, India in 2016. He published twenty four International Journal paper, one National Journal paper and presented a paper on one National Conference and seven International Conference. He is a Fellow of ISTE (India).