Optimal DG Placement in Distribution Network Using Big ... - IEEE Xplore

4 downloads 11590 Views 253KB Size Report
importance due to growing environmental and economic issues. This paper focuses on optimal placement of DG generators in distribution networks to minimize ...
2016 Second International Conference on Science Technology Engineering And Management ( ICONSTEM )

Optimal DG Placement in Distribution Network Using Big Bang-Big Crunch Algorithm Ms. A. Vinoliya, B.E., (M.E.,)

Dr. T. D. Sudhakar, M.E., Ph.D., Professor Department of Electrical and Electronics Engineering

Student Department of Electrical and Electronics Engineering St. Joseph’s college of Engineering Chennai 600119, India [email protected]

St. Joseph’s College of Engineering Chennai 600 119, India [email protected]

improve the quality of power and relieve the overload [4] and [5]. Types of loss minimization are explained in [6]. Network reconfiguration allows the system to supply with less power losses on the system lines. Network reconfiguration increase the reliability of power voltage on the user side and it is essential for optimization problem for best location of sectionalizing and tie switches. Network reconfiguration is efficient electric transmission; the improved voltage stability is achieved without any addition cost and increases the network reliability [7]. Optimal allocation of capacitor done with the appropriate location where the capacitor to be placed, type of capacitor, size of the capacitor and total number of capacitor that should be placed so the total cost is minimized without violating the operational constraints. The capacitors are widely used in distribution systems to reduce energy losses of distribution apparatus and to maintain a voltage profile within limits. The capacitor placement for low voltage systems determines capacitor type size, location and control scheme. The fundamental function of power capacitor is to provide needed reactive power compensation. Distributed generation (DG) will rise in the future of the power system. DG reduces line losses [8].Many researchers have used certain methods for finding the optimal DG placement. There are no clear restrictions on location of DG units in the distribution network, as there are no geographical limitations as in the case of substations. Hence, there only limitations arise from electrical requirements. The optimization method also is used to find DG location and minimize the loss. The optimization method can be classified into analytical, numerical and heuristic. Many algorithm used in optimization method like GA [9],[10], Artificial Bee Colony [11], PSO [12],[13] etc. Big bang big crunch algorithm is best method for DG location to minimize the loss. Big Bang-Big Crunch (BB-BC) optimization method was introduced by Erol [14],[15]. The method was successfully

Abstract— In recent years DG installations are gaining much importance due to growing environmental and economic issues. This paper focuses on optimal placement of DG generators in distribution networks to minimize the overall system losses. In the proposed work the identifying the DG locations can be found out using the Big Bang Big Crunch algorithm. Based on these identified locations the DG’s are placed and load flow is run to identify the total losses of the network. The proposed method is applied to 33 – Bus distribution network and the load flow used is forward sweeper method. The results obtained for the network using the proposed methodology are tabulated. Key terms— Big bang-big crunch, distributed generators, energy loss, and optimization.

I.

INTRODUCTION

The distribution system plays a significant position in the power system that connects bulk power with the consumers. The effective distribution network will cope up with increasing demand for domestic, industrial and commercial load. There is a large range of terminologies used for distribution system and several technologies ranging from traditional to nontraditional which it used for DG application. The distribution system is separated as primary distribution network and secondary distribution network [1]. In recent year electrical energy is repeatedly lost due to resistance in power system networks [2]. Moreover, distribution systems are very known for a higher R/X ratio and voltage drops can result in substantial power losses along distribution feeders. So Reconfiguration methods are used for loss reduction in distribution systems. Placement of DG into distribution systems has been rising rapidly for loss minimization [3]. Major causes are constraints on building new transmission and distribution lines, DG’s are placed in the consumer side to supply reliable power to the consumer,

556 978-1-5090-1706-5/16$31.00©2016 IEEE

2016 Second International Conference on Science Technology Engineering And Management ( ICONSTEM )

applied to nonlinear multidimensional functions and showed good convergence speed.

II.

STEP 2: Calculate the fitness function values of candidate solution done by the load flow. The control variable values taken by different candidates are in the system data and load flow is done. The total line loss corresponding to different candidates are calculated by load flow. STEP 3: Determine the centre of mass which produce the global best fitness using equation. The candidates are arranged in the ascending order (fitness) and the first candidate will consider to be the candidate with best fitness (minimum loss). The Centre of mass is denoted by Xc and is calculated according to the following equation.

PROBLEM STATEMENT

The optimization problem under study can be stated as follows: Minimize the power loss:

Subject to, 0 ≤ P DG

Nf  P loss , f f =1

NP 1 xi  f ( xi) X c = i=1 NP 1  f ( xi) i=1

max ≤ P DG

V min ≤ V ≤ V max

Pg min ≤ Pg ≤ Pg max

Where

xi

Q g min ≤ Q g ≤ Q g max

f ( x i ) is a fitness function value of this point. NP is the population size in Big Bang phase. STEP 4: Generate new candidates around the centre of mass by adding or subtracting a normal random number. It should be insure that the control variables are within their limits otherwise adjust the values of r and α. r α ( X max − X min ) X new = X c + t

The real and reactive powers of the DG unit are constrained between maximum and minimum limit. Where, f is feeder number. Nf is the total number of feeders. Ploss , f is the power loss in the feeder. P DG max P DG

is the power of DG. is the maximum power of DG.

In general, optimal placement of DGs can be derived by considering other factors, such as economic and geographic considerations. In this work, the power loss is taken into account and placement of DGs is found by considering objective. However, any other objectives can be consider as as problem and optimized by the proposed algorithm.

Where r is a random number. α is a parameter which limiting the size of the search space.

X max and X min are the upper and lower limits.

t is the iteration step. STEP 5: Repeat steps 2-4 until optimal point shas been achieved. Due to complexity, unbalance, and offshoot of the feeder branches, the function should be optimized either as the power loss or the energy loss which it can not be formulated as a function of DGs location. when considering, the locations of DGs as one of the system variables, is not continuous as it should be integer numbered in a random way. So the current optimization problem indulges many local minimum locations that may trap the traditional BB-BC algorithm.

BIG BANG-BIG CRUNCH ALGORITHM In Big Bang phase, solutions are distributed over the search area and in the Big Crunch phase, randomly distributed particles are drawn in a specific fashion. The Big Bang-Big Crunch optimization creates random points in the phase and a single point in the Big Crunch phase is collected after a number sequential Big Bangs and Big Crunches.[10] BIG BANG BIG CRUNCH ALGORITHM INVOLVES THE STEPS SHOWN BELOW

III.

STEP 1: Form an initial generation of NP candidates in a random manner in the limits of search space. Each candidate is consider to be a vector for all control variables.

PROPOSED SUPERVISED BIG BANG BIG CRUNCH

This proposed method gives an effective approach for calculating the sensitivities of active and reactive power loss and voltage magnitudes for active/reactive power

557 978-1-5090-1706-5/16$31.00©2016 IEEE

is the ith candidate in the search space.

2016 Second International Conference on Science Technology Engineering And Management ( ICONSTEM )

injection in the radial distribution system and it also gives a detailed account on the big bang-big crunch optimization technique of the work being carried out in the distribution side of the power system network. [10] Algorithm steps Step 1: Read the system data Step 2: Generate randomly the initial values of the system i.e DGs active power, DGs locations. Step 3: Calculate the active power loss analogous to all initial DG locations and powers by running the balanced load flow. Step 4: Select the best DG location and power which achieve minimum active power loss. Step 5: Call the DG location corresponding to the best DG power obtained from step 4 from the guidance table. Step 6: Calculate the active power loss corresponding to the DG power determined from step 4 and the recalled DG location then compare this power loss pl re with the power

FIG 1: single line diagam of 33-bus

read system data

loss determined from step 4 ( pl re ). Step 7: Check the power loss at the recalled location, if its value is lower than the power loss determined from step 4 set the recalled location as the best location, otherwise set the best location to that obtained from step 4. Step 8: Update the DG locations and powers using (1) and (2).Keeping the best DG location and power as a one of the new system variables and round the DG locations to the nearest integer. The new DG locations and powers upper and lower bounded. The square of the iteration step is used to quickly convergence:

loc new = loc best +

Calculate initial values for all DG location and power

Load flow analysis for DG

Calculate power loss

max loc × Randn it 2

max p × Randn Pgnew = Pgbest + it 2

i=1

Where, and P gnew are the new candidates DG locations and active powers. and max p are the maximum value of DG max loc

loc new

Find best location to place DG and power

locations and active powers. Step 9: Repeat steps 4)–8) until the convergence criteria is achieved, the convergence is considered achieved when more than 50% of the DG locations and DG active powers are converged to a certain value. In the supervised BB-BC the optimal location and power loss is calculated. when comparing the traditional BBBCwith Supervised BB-BC is to prove the efficency and robustness ,convergence speed is improved.

Call the location corresponding to the best DG power Load flow analysis for DG at best relocation and power

558 978-1-5090-1706-5/16$31.00©2016 IEEE

i=i+1

2016 Second International Conference on Science Technology Engineering And Management ( ICONSTEM )

plre ≤ pl prev

yes

no

locbest = loc prev

locbest = locre

Update the DG location and powers subject to constraints FIG 3: The active power loss versus iteration of 33-bus feeder Yes

exit

Convergence

V.

no

Optimal placement of DG is crucial for increasing the total real power loss reduction in the distribution system with active power compensation. In this paper the optimal location of DG’s in a distribution network are identified using the Big Bang Big Crunch algorithm. The proposed method is applied to 33 – Bus distribution network and from the results it is observed that there is an improvement in loss reduction and voltage improvement.

FIG 2 : Flow chart of supervise BB-BC IV. ANALYSIS AND RESULTS The proposed method implemented in MATLAB and tested on 33-bus feeder, and results are obtained to evaluate its effectiveness, whose single line diagram in diagram is shown above. The voltage profile of the system obtained using load flow solution. In this proposed system voltage and real power loss is calculated. Then location of DG is found by using BBBC algorithm. The real power loss is reduced when DG is generated by using BB-BC algorithm.

References [1] D. Singh and R. K. Misra, ‘‘Effect of load models in distributed generation planning,’’ IEEE Trans. Power Syst., vol. 22, no. 4, pp. 2204---2212, Nov. 2007. [2] J. Federico, V. Gonzalez, and C. Lyra, “Learning classifiers shape reactive power to decrease losses in power distribution networks,” in Proc. IEEE Power Eng. Soc. General Meet., Jun. 2005, vol. 1, pp. 557–562. [3] N. S. Rau, and Y. H. Wan, Optimum location of resources in distributed planning, IEEE Transaction Power System, vol. 9, no. 4, Nov. 1994, pp. 20142020.

TABLE II Optimal location of DG Bus no

location

14

1

17

2

26

3

31

4

Size of DG(KVA) 2751.502514 KVA 1318.570346 KVA 2476.927085 KVA 2071.487409 KVA

[4] T. Ackermann, G. Andersson, and L. Soder, “Distributedgeneration:A definition,” Elect. Power Syst. Res., vol. 57, pp. 195–204, 2001.

Power loss(KW)

[5] G. Celli and F. Pilo, “Optimal distributed generation allocation in MV distribution networks,” in Proc. 22nd IEEE Power Eng. Soc. Int. Conf. Power Industry Computer Applicat., 2001. PICA 2001. Innovative Computing for Power – Electric Energy Meets the Market, May 20–24, 2001, pp. 81–86. [6] N. Acharya, P. Mahat, and N. Mithulananthan, “An analytical approach for DG allocation in primary distribution network,” Int. J. Elect. Power Energy Syst., vol. 28, no. 10, pp. 669–678, Dec. 2006. [7] M. E. Baran and F. F. Wu, ‘‘ Network reconfiguration in distribution systems for loss reduction and load balancing’’, IEEE Transactions on Power Delivery, Vol. 4, No 2 Apr 1989, pp. 1401-1407,. [8] T. Griffin, K. Tomsovic, D. Secrest, and A. Low, ‘‘Placement of dispersed rd generations systems for reduced losses,’’ in Proc. IEEE 33 Annu. Hawaii Int. Conf. Syst. Sciences (HICSS), Big Island, HI, 2000, pp. 1446---1454. [9] T. Gözel and M. H. Hocaoglu, “An analytical method for the sizing. [10] G. P. Harrison, A. Piccolo, P. Siano, and A. R. Wallace, “Hybrid GA and OPF evaluation of network capacity for distributed generation connections,”Elect. Power Syst. Res., vol. 78, no. 3, pp. 392–398, 2008. [11] F. S. Abu-Mouti and M. E. El-Hawary, “Optimal distributed generationallocation and sizing in distribution systems via artificial bee

60.779245 kW

559 978-1-5090-1706-5/16$31.00©2016 IEEE

CONCLUSION

2016 Second International Conference on Science Technology Engineering And Management ( ICONSTEM )

algorithm. In: Proceedings of international power and energy conference, Singapore; 2010. p.1083–7. [14] K. Osman and I. Eksin, “A new optimization method: Big Bang- Big Crunch,” Adv. Eng. Softw., vol. 37, no. 2, pp. 106–111, 2006. [15] K. Osman and I. Eksin, “A new optimization method: Big Bang- Big.

colony algorithm,” IEEE Trans. Power Del., vol. 26, no. 4, pp. 2090– 2101,Oct. 2011. [12] Kennedy J and Eberhart R, ‘‘Particle Swarm Optimizer,’’ IEEE International Conference on Neural Networks (Perth Australia), IEEE Service Center Piscataway, NJ, IV, pp1942- 1948, 1995. [13] Sedighi M, Igderi A, Parastar A. Sitting and sizing of distributed generation in distribution network to improve of several parameters by PSO

560 978-1-5090-1706-5/16$31.00©2016 IEEE