2012 IEEE International Power Engineering and Optimization Conference (PEOCO2012), Melaka, Malaysia: 6-7 June 2012
Optimal Allocation and Sizing of Distributed Generation in Distribution System via Firefly Algorithm M. H. Sulaiman, M. W. Mustafa, A. Azmi, O. Aliman, S. R. Abdul Rahim
Therefore, the factors of the best location and sizing are among the important issues in the implementation of distributed generation in the distribution system. With optimal placement and sizing of distributed generation, maximum potential benefits could be obtained especially to improve the performance and to reduce the system losses [9-11]. To date, the application of artificial intelligence and optimization techniques become the choice of many researchers to determine the optimal allocation of DG. The using of evolutionary programming (EP) in optimal allocation of DG has been proposed in [3]. The authors use the sensitivity indices as the tools to predict the placement of EG at a particular bus. The optimal allocation problem using ant colony optimization (ACO) is proposed in [4]. A cost based model to find the optimal size and location of DG sources which using a minimization of DG investment cost and total operation cost of the system are presented. ACO also has been applied to solve optimization problem of voltage and reactive power control with considering the distributed generators [12]. The incorporation of particle swarm optimization (PSO) for DG sizing and location is proposed in [5]. The authors emphasize on improvement of voltage profile, total harmonic distortion and losses in their approach to determine the location and size of DG. The incorporation of genetic algorithm (GA) in this problem also has been proposed [1, 2, 13]. In this paper, a new swarm intelligence approach that utilizes Firefly Algorithm (FA) to determine the optimal DGunit’s size and location in order to minimize the total system real power loss is proposed. By using FA, the size and location of DG can be determined simultaneously. This paper is organized as follows. The concept of FA is presented in Section 2. The application of FA into optimal allocation of DG is discussed in Section 3. In Section 4, the case study including discussion is presented. Finally, conclusion is stated in Section 5.
Abstract-- This paper presents an application of Firefly Algorithm (FA) in determining the optimal location and size of Distributed Generation (DG) in distribution power networks. FA is a meta-heuristic algorithm which is inspired by the flashing behavior of fireflies. The primary purpose of firefly’s flash is to act as a signal system to attract other fireflies. In this paper, IEEE 69-bus distribution test system is used to show the effectiveness of the FA. Comparison with other method is also given. Index Terms--distributed generation, firefly algorithm, genetic algorithm.
I. INTRODUCTION
I
N recent years, distributed generation (DG) installation has shown an increasing amount of growth in the distribution networks throughout the world due to the rise of promotion towards utilization of renewable energy resources and development of co-generation plants. As DG is bound to the effect of power flow of the system associated distribution network, the losses of such networks will in turn to be affected as well. Recent studies have shown that DG can help in reducing the system losses and significantly improving the magnitude of the voltage profile [1-5]. This depends on several factors including the location of DG, the relative magnitude between the generation and the total connected load local to the generator and the topology of the network under consideration. With proper planning, the integration of distributed generations in a distribution system would lead to enhancement in the network performance in terms of voltage profile improvement, reduction in line losses and improve reliability and power quality [6-8]. As a result, the demand required from the grid could be reduced, thus cutting the need to strengthen the feeders connecting the network to the grid.
II. FIREFLY ALGORITHM M. H. Sulaiman and O. Aliman are with the Faculty of Electrical & Electronics Engineering, Universiti Malaysia Pahang, 26600 Pekan, Pahang. (phone: +609-4246140; fax: +6049-4242032; e-mail:
[email protected],
[email protected]). M. W. Mustafa is with the Faculty of Electrical Engineering, Universiti Teknologi Malaysia, Johor 81310, Malaysia. (e-mail:
[email protected]). A. Azmi and S. R. Abdul Rahim are with the School of Electrical System Engineering, Universiti Malaysia Perlis (UniMAP), Perlis 01000, Malaysia. (e-mail:
[email protected],
[email protected]).
978-1-4673-0662-1/12/$31.00 ©2012 IEEE
Firefly Algorithm (FA) is invented by Xin-She Yang [14] for solving multimodal optimization problem. The development of FA is based on flashing behavior of fireflies. There are about two thousand firefly species where the flashes often unique for a particular species. The flashing light is produced by a process of bioluminescence where the exact
84
2012 IEEE International Power Engineering and Optimization Conference (PEOCO2012), Melaka, Malaysia: 6-7 June 2012
functions of such signaling systems are still on debating. Nevertheless, two fundamental functions of such flashes are to attract mating partners (communication) and to attract potential prey. For simplicity, the following three ideal rules are introduced in FA development [14]: 1) all fireflies are unisex so that one firefly will be attracted to other fireflies regardless of their sex, 2) attractiveness is proportional to their brightness, thus for any two flashing fireflies, the less brighter one will move towards the brighter one, 3) the brightness of a firefly is affected by the landscape of the objective function. For maximization problem, the brightness can simply be proportional to the value of the objective or fitness function. The basic steps of the FA can be summarized as the pseudo code which is depicted in Fig.1 [14].
allocation of DG is shown in Fig. 2. Start Initialize location of fireflies Iteration =1
Insert variable x into load flow data
Run load flow
Firefly Algorithm……………………………………………….. Objective function f(x), x = (x1,…,xd)T Generate initial population of fireflies xi ( i=1, 2…, n) Light intensity Ii at xi is determined by f(xi) Define light absorption coefficient γ while ( t < MaxGeneration) for i = 1: n all n fireflies for j = 1: i all n fireflies if (Ij > Ii), More firefly i towards j in d-dimension; end if Attractiveness varies with distance r via exp[-γr] Evaluate new solutions and update light intensity end for j end for i Rank the fireflies and find the current best end while Post process results and visualization……………………………..
Objective function evaluation Ranking fireflies by their light intensity/objective Find the current best solution Move all fireflies to the better locations (Updating fireflies)
Fig. 1. Pseudo code of the FA.
III. FA FOR OPTIMAL ALLOCATION OF DG In this section, FA is proposed to find the optimal location and size of DG unit in the distribution system to minimize the total loss in the system. By minimizing the loss, the voltage profile at each bus is also expected to be improved. The variables of the optimal allocation of DG are coded as follows:
x = [x1L , x1S ,...x nL , x nS ]
No
Yes
(1)
Print results
where L indicates the location of DG, S indicates the size of DG and n is the number of DG unit that need to be installed in the system. These variables then are included in the data of load flow and the load flow is executed to obtain the total loss of the system. The procedure to obtain the optimal allocation of DG requires load flow to be run iteratively. After obtaining the best location and size of DG simultaneously, the procedure is stopped. The objective function, f(x) is the results of the total loss of the system, PLoss to be minimized, as follows:
⎞ ⎛ line f ( x) = min⎜⎜ ∑ PLoss ⎟⎟ ⎠ ⎝ j =1
Iteration maximum
End Fig. 2. Flow of optimal allocation of DG using FA IV. RESULTS AND DISCUSSION The proposed method has been tested on IEEE 69-bus radial distribution system. Basically, the algorithm is modified from [15] to suit the DG allocation problem. This test system can be obtained in [16]. Fig. 3 shows the test system with a total real and reactive power demand of 3802.19 kW and 2694.60 kVAr respectively. The FA properties in this simulation are set as follow: • Number of fireflies: 20 • Maximum iteration: 30
(2)
where line is number of transmission lines in the distribution system. The process of incorporating the FA into optimal
85
2012 IEEE International Power Engineering and Optimization Conference (PEOCO2012), Melaka, Malaysia: 6-7 June 2012
• • • • •
Number of DG unit: 1 and 2 DG size: 0.01 MW< PDG < 2.5 MW Alpha (scaling parameter): 0.25 Minimum value of beta (attractiveness): 0.2 Gamma (absorption coefficient): 1
Fig. 4. Objective function, f(x) vs. iteration for Case 2
Fig. 3. IEEE 69-bus radial distribution system The value of alpha, beta and gamma is similar with [15]. This is due to ensure that there is not much modification from the ideal FA. To study the impact of DG installation on the system performance, the following three cases are considered:
Fig. 5. Objective function, f(x) vs. iteration for Case 3
Case 1: Calculate the distribution network losses and minimum voltage before DG installation. Caser 2: Repeat Case 1 with the 1 DG included once its optimal location and size are determined. Case 3: Repeat case 1 with the 2 DGs included once the optimal locations and sizes are determined.
TABLE I RESULTS FOR CASES 1, 2 AND 3
Real power losses (MW) Minimum bus voltage (p.u) DG location and DG size (MW)
Since this test system is moderate in size, it is adequate to install up to two units of DG. Figs. 4 and 5 show the best objective function, f(x) versus iteration for cases 2 and 3 respectively out of ten times simulation. It can be seen that the minimum values of losses are 0.0832 MW and 0.0747 MW for cases 2 and 3 respectively. It also can be noted that the minimum bus voltage for cases 2 and 3 are similar, which is at bus 27 compared to case 1, which is at bus 65. The comparison studies of these 3 cases are tabulated in Table 1.
Case 1 0.2249
Case 2 0.0832
Case 3 0.0747
0.9092 @ bus 65
0.9683 @ bus 27
0.9791 @ bus 27
-
Bus 61 with 1.8753
Bus 61 with 1,7496 Bus 67 with 0.9269
From Table 1, it can be seen that by installing the DG unit at bus 61 with the size of 1.6 MW for case 2 has improved about 60% compared to the case 1. For case 3, the result is better compared to case 2 in term of loss reduction. Table 2 shows the comparison voltage profile at each bus in the system for these 3 cases. From this table, it shows that the voltage profile is improved from the base case.
86
2012 IEEE International Power Engineering and Optimization Conference (PEOCO2012), Melaka, Malaysia: 6-7 June 2012
TABLE II VOLTAGE PROFILES FOR CASES 1, 2 AND 3 Bus
Case 1
Case 2
Case 3
42
0.999
-0.028
0.999
-0.026
0.999
-0.026
43
0.999
-0.029
0.999
-0.027
0.999
-0.026
|V|
Angle(⁰)
|V|
Angle(⁰)
|V|
Angle(⁰)
44
0.999
-0.029
0.999
-0.027
0.999
-0.026
1
1
0
1
0
1
0
45
0.998
-0.031
0.998
-0.029
0.998
-0.028
2
1.000
-0.001
1.000
0.000
1.000
0.000
46
0.998
-0.031
0.998
-0.029
0.998
-0.028
3
1.000
-0.002
1.000
-0.001
1.000
0.000
47
1.000
-0.007
1.000
-0.003
1.000
-0.002
0.999
-0.052
0.999
-0.048
0.999
-0.046
4
1.000
-0.006
1.000
-0.002
1.000
0.000
48
5
0.999
-0.018
0.999
0.007
1.000
0.017
49
0.995
-0.191
0.995
-0.187
0.995
-0.186
6
0.990
0.050
0.995
0.201
0.997
0.265
50
0.994
-0.211
0.994
-0.207
0.994
-0.205
7
0.981
0.121
0.991
0.405
0.995
0.524
51
0.979
0.139
0.990
0.455
0.994
0.587
8
0.979
0.139
0.990
0.455
0.994
0.586
52
0.979
0.139
0.990
0.455
0.994
0.587
9
0.977
0.147
0.989
0.481
0.994
0.619
53
0.975
0.169
0.989
0.564
0.993
0.697
10
0.972
0.232
0.984
0.564
0.994
0.792
54
0.971
0.195
0.988
0.660
0.992
0.788
11
0.971
0.251
0.983
0.582
0.994
0.831
55
0.967
0.230
0.987
0.795
0.991
0.916
0.963
0.265
0.986
0.928
0.990
1.041
12
0.968
0.304
0.980
0.634
0.991
0.881
56
13
0.965
0.350
0.977
0.679
0.988
0.925
57
0.940
0.662
0.984
1.666
0.987
1.749
14
0.962
0.396
0.974
0.724
0.985
0.970
58
0.929
0.864
0.983
2.030
0.985
2.100
15
0.960
0.442
0.971
0.769
0.982
1.013
59
0.925
0.945
0.983
2.172
0.985
2.236
16
0.959
0.451
0.971
0.777
0.982
1.022
60
0.920
1.050
0.983
2.345
0.984
2.403
17
0.958
0.465
0.970
0.791
0.981
1.035
61
0.912
1.119
0.982
2.585
0.983
2.630
18
0.958
0.465
0.970
0.791
0.981
1.035
62
0.912
1.122
0.982
2.587
0.983
2.632
19
0.958
0.474
0.970
0.799
0.980
1.043
63
0.912
1.125
0.981
2.590
0.982
2.636
0.910
1.143
0.980
2.606
0.981
2.651
20
0.957
0.479
0.969
0.805
0.980
1.049
64
21
0.957
0.488
0.969
0.813
0.980
1.057
65
0.909
1.148
0.979
2.610
0.980
2.656
22
0.957
0.488
0.969
0.814
0.980
1.057
66
0.971
0.252
0.983
0.583
0.995
0.852
23
0.957
0.490
0.969
0.815
0.979
1.058
67
0.971
0.252
0.983
0.583
0.995
0.853
24
0.957
0.493
0.969
0.818
0.979
1.061
68
0.968
0.310
0.980
0.639
0.990
0.887
25
0.956
0.496
0.968
0.821
0.979
1.064
69
0.968
0.310
0.980
0.639
0.990
0.887
26
0.956
0.497
0.968
0.822
0.979
1.065
27
0.956
0.497
0.968
0.822
0.979
1.066
28
1.000
-0.003
1.000
-0.001
1.000
0.000
29
1.000
-0.005
1.000
-0.004
1.000
-0.003
30
1.000
-0.003
1.000
-0.001
1.000
-0.001
31
1.000
-0.003
1.000
-0.001
1.000
0.000
32
1.000
-0.001
1.000
0.001
1.000
0.002
33
0.999
0.004
0.999
0.005
0.999
0.006
34
0.999
0.010
0.999
0.011
0.999
0.012
35
0.999
0.011
0.999
0.012
0.999
0.013
36
1.000
-0.003
1.000
-0.001
1.000
-0.001
37
1.000
-0.009
1.000
-0.008
1.000
-0.007
38
1.000
-0.012
1.000
-0.010
1.000
-0.009
39
1.000
-0.012
1.000
-0.011
1.000
-0.010
40
1.000
-0.012
1.000
-0.011
1.000
-0.010
41
0.999
-0.023
0.999
-0.022
0.999
-0.021
In order to show the effectiveness of the proposed method in solving optimal allocation of DG, the comparison with GA [1] has been conducted. The GA properties are set as follows: • • • •
Population: 20 Maximum iteration: 30 Crossover probability: 0.9 Mutation probability: 0.1
Fig. 6 shows the best objective function, f(x) versus iteration for 1 DG out of ten times simulation using GA. It can be seen that the result is similar with the FA, where the location for installing the DG is at bus 61. Nevertheless, the size of this simulation is varied compared to FA even the objective function or loss obtained is similar, which is 0.0832 MW. Thus it can be said that FA can gives good optimal result in determining the DG allocation, as good as GA.
87
2012 IEEE International Power Engineering and Optimization Conference (PEOCO2012), Melaka, Malaysia: 6-7 June 2012
[10]
[11]
[12]
[13]
[14]
Fig. 6. Objective function, f(x) vs. iteration using GA for 1 DG installed at bus 61, size 1.8673 MW
[15]
V. CONCLUSION
[16]
In this paper, a new swarm based Firefly Algorithm has been presented to solve the optimal DG allocation and sizing problem. The effectiveness of FA was demonstrated and tested. The results show that incorporating the DG in the distribution system can reduce the total line power losses and improve the voltage profile of the system. The comparison with GA also has been conducted to see the performance of FA where it is as good as GA in solving the optimal allocation problem.
VII. BIOGRAPHIES Mohd Herwan Sulaiman (M’12) obtained his B. Eng. (Hons) in Electrical-Electronics, M. Eng. (Electrical-Power) and PhD from Universiti Teknologi Malaysia (UTM) in 2002, 2007 and 2011 respectively. He is currently a lecturer at Faculty of Electrical & Electronics Engineering, Universiti Malaysia Pahang (UMP). His research interests are power system deregulation/ optimization and application of AI to power system studies.
VI. REFERENCES [1] [2]
[3] [4] [5]
[6]
[7] [8] [9]
Tech Conference Proceedings, IEEE Bologna, Vol. 2, pp. 1- 5, Jun 2003. C. Wang and M. H. Nehrir, “Analytical approaches for optimal placement of distributed generation sources in power systems”, IEEE Transactions on Power Systems, Vol. 19, Issue: 4. pp. 2068 – 2076, Nov. 2004. M. H. Moradi and M. Abedeni, “Optimal multi-distributed generation location and capacity by Genetic Algorithms”, 4th International Power Engineering and Optimization Conference (PEOCO) , pp. 400-404, Jun 2010. T. Niknam, "A new approach based on ant colony optimization for daily Volt/Var control in distribution networks considering distributed generators," Energy Conversion and Management, vol. 49, pp. 34173424, 2008 Y. Alinejad-Beromi, M. Sedighizadeh, M. R. Bayat, and M. E. Khodayar, "Using genetic alghoritm for distributed generation allocation to reduce losses and improve voltage profile," in Universities Power Engineering Conference, 2007. UPEC 2007. 42nd International, 2007, pp. 954-959. X.-S. Yang, "Firefly algorithms for multimodal optimization," Stochastic Algorithms: Foundation and Applications SAGA 2009, vol. 5792, pp. 169-178, 2009 Firefly Algorithm for constrained optimization by Xin-She Yang. Retrieved on January 1st, 2012 [Online]. Available: http://www.mathworks.com/matlabcentral/fileexchange/29693-fireflyalgorithm. M. E. Baran and F. F. Wu, "Optimal capacitor placement on radial distribution systems," Power Delivery, IEEE Transactions on, vol. 4, pp. 725-734, 1989.
M. H. Sulaiman, O. Aliman, and S. R. A. Rahim, "Optimal allocation of EG in distribution system using genetic algorithm technique," Journal of Energy and Power Engineering, vol. 4, pp. 56-63, 2010. M. H. Sulaiman, O. Aliman, and S. R. A. Rahim, "Optimal embedded generation allocation in distribution system employing real coded genetic algorithm," in International Conference on Power Systems Engineering (ICPSE 2010), Penang, Malaysia, 2010. T. K. A. Rahman, S. R. A. Rahim, and I. Musirin, "Optimal allocation and sizing of embedded generators," in Power and Energy Conference, 2004. PECon 2004. Proceedings. National, 2004, pp. 288-294. H. Falaghi and M. R. Haghifam, "ACO Based Algorithm for Distributed Generation Sources Allocation and Sizing in Distribution Systems," in Power Tech, 2007 IEEE Lausanne, 2007, pp. 555-560. Y. Alinejad-Beromi, M. Sedighizadeh, and M. Sadighi, "A particle swarm optimization for sitting and sizing of Distributed Generation in distribution network to improve voltage profile and reduce THD and losses," in Universities Power Engineering Conference, 2008. UPEC 2008. 43rd International, 2008, pp. 1-5 Ding Xiaoqun; Wu Jiahong; Zhao Feng, “Optimal location and capacity of distributed generation based on scenario probability”, International Conference on Sustainable Power Generation and Supply, 2009. SUPERGEN '09, pp:1-5, 2009. S. Conti, S. Raiti and G. Tina. “Small-scale EG effect on voltage profile: An analytical method”, IEEE Proc. Generation,Trans,Distb.Vol 150,No1, pp.78-86, Jan 2003. Ding Xu and A.A. Girgis; “Optimal load shedding strategy in power systems with distributed generation”, Power Engineering Society Winter Meeting, IEEE, Vol. 2, pp.788-793, Jan-Feb 2001. C.L.T Borges and D.M Falcao, “Impact of distributed generation allocation and sizing on reliability, losses and voltage profile”, Power
Mohd Wazir Mustafa received his B. Eng. Degree (1988), M. Sc. (1993) amd PhD (1997) from University of Strathclyde. He is currently an Associate Professor and Deputy Dean at Faculty of Electrical Engineering, Universiti Teknologi Malaysia. He is a member of Institution of Engineers, Malaysia (IEM) and a member of IEEE. His research interest includes power system stability, FACTS, wireless power transmission and power system distribution automation.
Azralmukmin bin Azmi obtained his B. Eng. (Hons) (Elecrical System Engineering) from Universiti Malaysia Perlis (UniMAP) in 2007 and M. Eng. (Electrical-Power) from Universiti Teknologi Malaysia (UTM) in 2009. Currently he is a lecturer at School of Electrical System Engineering, University Malaysia Perlis. His research interests included high voltage, power system and renewable energy. .
88
2012 IEEE International Power Engineering and Optimization Conference (PEOCO2012), Melaka, Malaysia: 6-7 June 2012
Omar Aliman (M’07) was born in Johor, Malaysia, on Dec., 1973. He received his Diploma in Electrical Power Engineering from University Teknologi Malaysia, graduated his BSc. from Hanyang University, Korea and MEng. from Universiti Teknologi Malaysia. Currently, he is a PhD candidate at Universiti Teknologi MARA. His ten-years industrial experience included in the area of EPCC Projects in Industrial Plants, Oil & Gas, Power Plants, R&D Commercialization and Consultation projects, local and overseas. He joined local universities as an academician-cum-researcher for the last six years. Currently, he works at Faculty of Electrical & Electronics Engineering, Universiti Malaysia Pahang. His special field of interest includes power distribution, distributed generation and arc-flash analysis study.
Siti Rafidah Binti Abdul Rahim received her Master in Electrical Engineering from Universiti Teknologi MARA (UiTM) in 2006, Bachelor of Electrical Engineering (Hons) from Universiti Teknologi MARA in 2003 and Diploma in Electrical Engineering (Power) from Universiti Teknologi Malaysia (UTM) in 1999. She is currently a lecturer at School of Electrical Systems Engineering, Universiti Malaysia Perlis (UniMAP). Her research interest includes power system optimization, distributed generation (DG) and artificial intelligence techniques in power systems.
89
