Multi Objective Reconfiguration of Distribution Networks with the Objective to Reduce Losses and Improving Power Quality Parameters

Transcription

1 Applied mathematics in Engineering, Management and Technology 2 (5) 214: Multi Ojective Reconfiguration of Distriution Networks with the Ojective to Reduce Losses and Improving Power Quality Parameters Hadi Kianersi 1, Hamid Asadi 2, Mostafa Karimi Darani 3, Mehran Jalalifar 4 1,2,3,4 Department of Electrical Engineering, Fereydan Branch, Islamic Azad University, Isfahan, Iran hadi_kianersi@yahoo.com Astract In this paper, multi-ojective distriution network reconfiguration with the ojectives of minimizing the actual losses and improving power quality parameters is carried out. For this purpose, inary genetic algorithm(bga) optimization procedures with the help of weighted sum method used. Along with these algorithms, graph theory and the adjacency matrix of the graph corresponds to the distriution network, to investigate the constraints and ojective functions such as the reconfiguration of the radial transmission and distriution networks can e used all over again. All the algorithms and methods on a standard 33-us IEEE distriution network have een studied and implemented. Keywords: Reconfiguration, Power Quality, Binary genetic algorithm (BGA) 1- Introduction Reconfiguration was first discussed in 1975 in order to reduce active power losses in distriution networks [1] was proposed. In this paper, first of all keys in the distriution package is not intended to replace the various ranches, the proper response is reconfiguration. The main disadvantage of this study can e assumed to e active all over again and ignoring the network constraints such as voltage and limits the current through each ranch could e mentioned. In [2] ranch switching methods for solving reconfiguration selecting two to two ranches, one for each ring opening and the closing have een selected to have een developed y. The disadvantages of this method can e too time solving reconfiguration noted. In [3] studied a network reconfiguration of a seasonal or daily (function of time) and at various times has een made. In [4] an innovative approach to automatically remove unwanted keys to choose which leads to non- radial networks are presented. In this paper, only the total radioactive waste prolem has een selected as the target network. In [5] and [6] for the first time, the ant algorithm to rearrange distriuted systems with the goal of reducing the losses can e utilized. This algorithm is very suitale solutions for prolems such as traveling salesman prolem and provides a sequential order. The excellent results of the ant algorithm to solve the prolem reconfiguration has to offer. In [7], the ant algorithm for parallel search feeders to minimize losses and to improve voltage levels are used. Than the enefits of parallel search to reduce the computation time can e noted. In this paper, the network is much less ale to find gloal minimum. In [8] an innovative approach to prolem solving migration reconfiguration using ant colony algorithm is presented. Login nonlinear loads in the power distriution network can have a negative impact on the quality issue. One way to improve the power quality parameters are reconfiguration [9]. Therefore, one goal of reconfiguration can improve power quality parameters, such as voltage harmonics. In this paper, multi ojected reconfiguration of distriution networks in order to reduce losses and improve power quality parameters using intelligent methods is studied. Also, using other definitions and properties of the adjacency matrix of graph theory, graph constraints and ojective functions are evaluated. 177

2 Applied mathematics in Engineering, Management and Technology Introduction of ojective functions The reconfiguration of the studied networks using weighted sum method to optimize multi-purpose, three functions is used for the purpose stated in the previous chapters. The relations (1) to (3) these relationships are re-created. F = V, (1) F = THD (2) F =Ploss (3) In these equations: i { The total numer of network uses} Constraints in the ojective function in (2) are oserved. However, a separate radial us network and lack of reconfiguration are major constraints. V V V I min I i,max max (4) I : The current of -th ranch R :The resistance of -th ranch N : The total numer of network ranches l R I : The maximum current through -th ranch V i : The magnitude voltage of i-th use The first and second ojective functions in order to improve power quality parameters and the third goal is to reduce active power losses are defined. The purpose of this paper is that each ojective functions to e minimized simultaneously. But given that these three functions have different dimension and thus cannot e directly these three functions together, he would ecome a target function. To resolve this prolem, the method of weighting coefficients is proposed. 3- The method of weighting coefficients For reconfiguration prolem solving with weighted coefficients of the ojective function is suggested: =., (5) In this formula w1, W2 and w3 are the weighting coefficients and C is the penalty factor. This approach to solving the prolem of dimension per unit order to provide an ojective function F3. This will solve the prolem dimension. However, with the proper choice of weighting coefficients w1 to w3 affect any of these functions, we determine the target. This usually allows the operator to choose the coefficients of the ojective function, which is more important to look at the highlight. Given the constraints that the rearrangement of the distriution networks, such as the preservation of the radial system serves all loads placed on each us voltage and current of each ranch, there is a range of acceptale answers must violate the provisions of can e removed from the set of solutions. Therefore, the parameters in each of the ojective functions is considered to e as follows: 178

3 Applied mathematics in Engineering, Management and Technology 214 ( mesh) + E Numer( isolated ) C = D Numer (6) Equation (6), and the coefficients are penalized for incorrect choices. Penalty factor for the numer of meshes penalty factor for the numer of uses isolated network and system requirements. Put the parameters in the ojective functions and possile solutions improale een removed from the search space and the algorithm quickly converges to the optimum solution is etter. It is worth noting constraints and lack of us radial network isolation system using graph theory are confirmed. 4- SIMULATION RESULTS: In this section, the ojective inary genetic algorithm (BGA) are implemented [1]. The system used a 33-us distriution system with rated voltage kv is [11]. Single-line diagram of this system is depicted in the figure 1. Also assumed to e non-linear loads in Tale 1 are in the network. Fig1. 33 uses test system Bus numer Taale1. Nonlinear loads Harmonics Before performing the simulation assumes that the keys are open lines It also has loads of ass that are sensitive to voltage dips and harmonics are: (12, 2, 24, 28). In this case, the ojective values of the quality parameters are as follows: it1 = V, = (7) 2= =.296 (8) The implementation of this algorithm in networks with BGA due to mesh networks due to the closure of all five keys, a gene on chromosome 5 have een considered. Each chromosome [Key1, Key2, Key3, Key4, Key5] is 179

4 Applied mathematics in Engineering, Management and Technology 214 defined. The program repeats on chromosome numer 2 and numer 2 is selected y the algorithm. Cross over rate against.9, the mutation ratio.5 is selected. In the first case, the coefficients for the intended purpose and penalty functions are: and w3=1. Implementation results in Tale 2 and Figures (2) and (3) are summarized fitness iter Fig2. Convergene characteristics of BGA Taale2. The results of reconfiguration (the first case) Fitness value after reconfiguration F 2 = Selected ranches Fitness value efore reconfiguration Improveme nt percent F 3 = ,7,22, ,2 F 1 = Single-line diagram of the system after rearrangement is depicted in the figure 3. This figure shows the radial and interconnected networks remains after rearrangement. The radial limits are eing adhered systems. And the optimization method converges to the right answer. Fig3. 33 us test System after reconfiguration(at the first case) In this section, three factors were considered equal to 1. But convenience is not possile to adjust these coefficients. Secondly usually farmers tend to have a unch of good answers to choose from among the est answer rather than simulation program where decision. Continuing to operate the system assumes that to reduce 18

5 Applied mathematics in Engineering, Management and Technology 214 losses are more important than any other ojective function. In this case, for example, the weighting coefficients can e considered as follows: w =w =1, w =1 By choosing the weighting coefficients, and the simulation results otained are as follows. The Tale 3 shows a loss of aout 87.9% drop. However, they have had less improvement in the ojective function. By increasing the weighting coefficients associated with the third ojective function, the ojective function compared to others in the ojective function is reduced further. This topic shows how to change the weighting coefficients in the final solution is an effective optimization algorithm. This makes it very difficult to choose the coefficients. Single-line diagram of the system, however, is depicted in Figure 4. This figure shows the radial and interconnected networks remains after rearrangement. The radial limits are eing adhered systems. And the optimization method converges to the right answer. Taale3. The results of reconfiguration (the second case) Fitness value Selected Fitness value Improvement after ranche efore percent reconfiguration s reconfiguration F 2 = F 3 = F 1 = Fig4. 33 us test System after reconfiguration (at the second case) In the third case, the weighting coefficients is selected as follow: w =w =1, w =1 By choosing the weighting coefficients and the simulation results are otained in Tale 4. Taale4. The results of reconfiguration (the third case) Fitness Fitness value value Selected Improveme after efore ranches nt percent reconfiguration reconfigur ation F 2 = F 3 = F 1 =

6 Applied mathematics in Engineering, Management and Technology 214 The Tale 4 shows that the ojective function is aout 44.27% drop compared to the previous two cases have een reduced. In fact, increasing the weight coefficients of the ojective function, the ojective function in the optimization prolem has een further reduced. Single-line diagram of the system after rearrangement in this case is depicted in Figure 5. This figure shows the radial and interconnected networks remains after rearrangement. The radial limits are eing adhered systems. And the optimization method converges to the right answer. Fig5. 33 us test System after reconfiguration (at the third case) Aove discussion shows that the choice of weighting coefficients in solving multi-ojective optimization prolems are of particular importance. And changes in these coefficients will e significant changes to optimize the output response. 5- Conclusions In this paper, multi-ojective distriution network rearrangement aims to reduce losses and improve power quality parameters were measured.for this purpose, a multi-ojective prolem into a single ojective using weighting coefficients were converted and analyzed using inary genetic algorithm.to demonstrate the effectiveness of the proposed method 33 us IEEE test system was used. References [1] A. Merlin and H. Back, "Search for a minimal loss operating spanning tree configuration for an uran power distriution system", Proceedings of the 5th Power System Computation Conference (PSOC), Camridge, pp. 1-18, [2]S. Civanlar, J.J Grainger, H. Yin and H. Lee, "Distriution feeder reconfiguration for loss reduction", IEEE Transactions on Power Delivery, Vol. 3, No. 3, pp , [3] R.P Broadwater, A.H. Khan, H.E Shalaan and R.E. Lee, "Time varying load analysis to reduce distriution losses through reconfiguration", IEEE Transactions on Power Delivery, Vol. 8, No. 1, pp , [4] R.J Safri, N.M.A Salama and A.Y Chickani, "Distriution system reconfiguration for loss reduction: a new algorithm ased on a set of quantified heuristic rules", Proceedings of Electrical and Computer Engineering, Vol. 1, Canada, pp , [5] E. Carpaneto and G. Chicco, Ant colony search ased minimum losses reconfiguration of distriution systems, IEEE MELECON, pp , 24. [6] A. Ahuja and A. Pahwa, using ant colony optimization for loss minimization in distriution networks, Proceedings of the 37 th Annual North American on Power symposium, pp , 25. [7] C. T. Su, C. F. Chung, Distriution network reconfiguration for loss reduction y ant colony search algorithm, Electric Power System Research, Elsevier, pp , 25. [8] Y. Wu, C. Lee, study of reconfiguration for distriution system with distriuted generation, IEEE Transaction on Power Delivery, Vol. 25, No. 3, pp , 21. [9] Saeed Jazei, Behrooz Vahidi, "Reconfiguration of distriution networks to mitigate utilities power quality Disturances", Electric Power Systems Research, pp. 9-17,

7 Applied mathematics in Engineering, Management and Technology 214 [1]Golshan, M.E.H. and Arefifar, S.A., "Distriuted Generation, Reactive Sources and Network-Configuration Planning for Power and Energy-Loss Reduction", IEE Proceedings Generation, Transmission and Distriution, Vol. 153, No. 2, pp ,