Distribution System Performance Enhancement through Optimal DG Placement using PSO
Navleenpreet Kaur1, Rajni Bala2
1Research Scholar, 2Assistant Professor
B.B.S.B.E.C, Fatehgarh Sahib, Punjab, India
[email protected], [email protected]
Abstract – Optimal allocation of distributed generators (DG) plays a significant role in improving the performance of the radial distribution system. The positive and negative impact of the distributed generators will depend on the optimal location and size of distributed generators. Therefore this paper focuses on selecting optimal size and location for distributed generator in the radial distribution systems (RDS) to minimizes losses and improve voltage profile. Backward Forward Sweep methodology for the load flow analysis of the distribution system is used to determine the power losses in the system and Particle Swarm Optimization (PSO) is used to solve the optimization problem. A 33-bus system has been taken as the test system.
Keywords: Distributed Generators (DG), Particle Swarm Optimization (PSO), Radial Distribution system (RDS).
Electrical distribution network systems include distribution feeder, which are arranged in mesh or radial pattern. The radial distribution system has high resistance to reactance ratio. Because of its high resistance to reactance ratio there will be increase in power losses and reduction in the magnitude of bus voltage. Therefore, to increase the performance of distribution system, introduction of DG’s takes place in the network 1-2.
Distributed Generation refers to any electric power production technology that is integrated within distribution systems or on the customer site of the meter. Optimal allocation of distributed generators helps in minimization of power losses and improvement of voltage stability in power systems. To achieve the most from DG installation, DG has to be optimally placed and sized. Improper installation of DG may degrade the performance of the distribution system. This has made the problem of optimal allocation of distributed generation to be an interesting research topic for many researchers.
There are various methods used for solving the problem of DG allocation for reduction of power losses in the distribution system.. These methods are: Analytical 3, Improved Analytical (IA) 4, Dynamic Programming 5, Mixed Integer Linear Programming (MILP) 6. These methods are used to solve only linear optimization problems. But the DG location and sizing problem is not a linear optimization problem, it is a discrete nonlinear optimization problem. Because of nonlinear optimization problem, these techniques are not that much effective in solving optimal allocation problem in the distribution system. So it is essential to consider search algorithms importance in solving above mentioned problem. Recently many search algorithms, i.e. Genetic Algorithm (GA) 7, Particle Swarm Optimization (PSO) 8, Combined Genetic and Particle Swarm Optimization 9, Cuckoo search algorithm (CSA) 10, Tabu Search (TS) 11, Simulated Annealing (SA) 12, Artificial Bee Colony (ABC) 13 are used to solve DG allocation problem effectively with reduced power losses as an main objective.
To select appropriate location and to calculate DG size for minimum real power losses Naresh Acharya et al presented a heuristic method in 14. But more computational efforts are required to solve the problem. At each bus the optimal value of DG is calculated for minimum system losses. Placing the calculated DG size for each and every bus of the system, corresponding system losses are calculated and compared to decide the appropriate location.
PSO was introduced by Kennedy, J. and R. Eberhart 15. AlRashidi M.R. and El Hawary M.E. 16 have noted the advantages of PSO technique over other optimization techniques.
This paper is aimed at reducing power loss and improving voltage profile of distribution systems using DG sources. Optimal placement of DG is obtained by considering the voltage profile. Here Particle Swarm Optimization (PSO) is employed to obtain the optimal sizing of DG. The analysis has been carried out for IEEE 33 radial bus test system.
II. PROBLEM FORMULATION
The main objective of this paper is to obtain power flow solution of IEEE 33 bus radial distribution system using backward and forward sweep method. From the load flow analysis most sensitive bus is identified for DG placement. Particle Swarm Optimization (PSO) is used to find the optimal size of DG. The purpose of DG placement is to minimize the total active power loss and improving the voltage profile.
III. PROPOSED WORK
A. Load Flow Analysis
The real power loss reduction in a distribution system is required for efficient power system operation. The loss I2R loss (PLoss) in a distribution system having n number of branches is given by equation (1)
Ploss = ?_(i=1)^n?I(i) ^2 R(i) (1)
where I(i) is the magnitude of the branch current and R(i) is the resistance of the ith branch respectively. The branch current can be obtained from the load flow solution 17.
B. Sensitive Node Identification
From Load Flow solution obtained by using Backward and Forward Sweep Method, voltage magnitude is computed at each and every bus of the system. The bus having minimum voltage magnitude is considered to be the most sensitive node for the placement of DG.
C. Particle Swarm Optimization (PSO)
For finding the best solution for any given circumstances, the optimization technique is used. For example, if a company required to improve its rating, technological and managerial plans have to be taken many times. Here, maximization of profits and minimization of the spending efforts are the basic aim of the plans. Optimization is referred to as both minimizing and maximizing the tasks. Due to this reason, optimization became very important in many fields. Basically, in order to solve the problems of optimization, PSO technique is inspired by the animal’s activity. In PSO, the meaning of swarm is population; particle represents each member of the population. Each particle searches through the entire space by randomly moving in different directions and remembers the previous best solutions of that particle and also positions of its neighbor particles. Particles of a swarm adapt their position and velocity dynamically by communicating best positions of all the particles with each other. In this way, all the particles in the swarm try to shift in the direction of better positions until the swarm reaches an optimal solution. PSO technique is becoming very popular. Because of its easy implementation and its ability to obtain fast convergence and also it uses only basic mathematics and it does not involve any derivative or gradient information.
The Basic Model of PSO Algorithm
Kennedy and Eberhart introduced a solution to non-linear and complex optimization problem by examining the behavior of flock of birds and established the concept of optimizing the function using swarm of particles. Consider a function of n dimension which is defined by:
f(x1, x2, x3…..xn) = f(x)
where xi is the optimizing variable, which represents the set of variables for a given function f(x). Here, the goal is to get an optimum value x* so that the function f(x*) can become either a maximum value or a minimum value.
The Particle Swarm Optimization (PSO) technique is a parallel search technique which utilizes multi-agents (swarm of particles). Each agent in the swarm represents a solution. All agents go through entire search space and updates its position and velocity based on their own experience and on experience of other agents. Suppose xit denote the agent or particle ‘i’ position vector search space at time step t, then each agent position is updated in the search space by
xit+1 = xit + Vit+1
where, Vit is the particle velocity vector which is used to update the own experience and other particles experience and also drives the optimization process. Thus, in PSO technique, all agents are randomly initialized and fitness value is computed by updating the personal best (best value of each agent) and global best (best value of all agents in the entire swarm). The loop starts by assuming initial values of position of the particles as personal best and then updates every particle position by using the updated velocity. When the stopping criterion is met, loop will be ended
Advantages and Disadvantages of PSO
PSO technique is a powerful technique for solving the non-linear optimization problems. It has its own advantages and disadvantages.
The advantages and disadvantages of PSO are discussed below:
Advantages of the PSO algorithm:
PSO technique is a gradient-free technique.
It is applied both in scientific research and engineering problems as the implementation of this algorithm is easy.
Compared to other optimization techniques, this algorithm has less impact of parameters to the optimal solution as it has only less number of parameters.
Optimum value can be obtained easily within short time.
Compared to other optimization techniques, this algorithm has less dependence on set of initial values.
Simple concept is involved here.
Disadvantages of the PSO algorithm:
Here the speed and direction may be degraded as this technique suffers from partial optimism.
Non-coordinate system exit problem may occur.
D. Implementation of PSO Algorithm to determine the size of DG Algorithm:
The PSO-based approach for finding sizes of DGs at selected locations to minimize the real power loss is as follows:
Considering Objective Function:
f = Min (Total Real Power Loss)
Where the total real power loss is given by the expression:
Ploss = ?_(i=1)^n?I(i) ^2 R(i)
Step 1. Choose the parameters that are to be optimized by using PSO. Here the parameters are real power that are injected through DG into distributed system i.e size of DG in order to minimize the losses and improve voltage.
Step 2. Choose the size of swarm.
Step 3. Generate the random values for DG size.
Step 4. Run the load flow and obtain the voltage profile and losses of the system.
Step 5. Also obtain the location of the DG to be placed by using voltage magnitude.
Step 6. Assume the fitness function as the real loss as we need to find the optimal DG size that minimize the losses to a maximum extent.
Step 7. Randomly initialize the position and velocity of swarm.
Step 8. By placing different sizes of DG in the obtained location, compute and store the fitness function of all particles in the swarm.
Step 9. Assume the initial randomly generated sizes of DG as pbest.
Step 10. Iterate through all the values of fitness function and the particle with minimum loss is considered as the gbest.
Step 11. Initialize the acceleration coefficients as c1=1 and c2=1.
Step 12. Initialize the loop and iteration count. For each particle calculate and update the velocity and position from (2) and (3).
Vit+1 = wVit + c1 r1 (pbesti – xit) + c2r2 (gbest – xit) (2)
xit+1 = xit + Vit+1 (3)
Step 13. Run the load flow after placing DG and obtain the new fitness function for each particle. If the new fitness value for any particle is better than previous pbest value then pbest value for that particle is set to present fitness value. Similarly gbest value is identified from the latest pbest values.
Step 14. If it reaches maximum iteration count then terminate the loop and plot the results. Otherwise increment the iteration count and go to step 12.
Step 15. gbest value gives the size of DG
IV. RESULTS AND DISCUSSION
The proposed method for the DG placement and sizing is demonstrated on IEEE 33 bus system. The Base voltage for this system is 12.66 kV. The single line diagram of 33-bus system is shown in Fig. 2.
Fig. 2- Single Line Diagram of 33 bus system
It has 33 buses and 32 branches. The total real and reactive power loads of the system are 3715 kW and 2300 kVAr.
Load flow results are shown in Table 1.
Bus System Real Power Loss (kW) Reactive Power Loss (kVRr)
33 208.281 139.303
Table I – Load flow Results
The optimal location for placement of DG is identified based on minimum voltage magnitude.
Fig. 3 Voltage magnitude profile
The bus having minimum voltage is the most sensitive bus which is selected to install DG in the system. From Fig. 3 it is observed that bus 18 has minimum voltage.
By optimally allocating the DG in the distribution system, the active and reactive power losses in the system are reduced and the system voltage profile is improved. The voltage profiles of the system before placement of DG and after placement of DG using PSO are shown in Fig. 4.
Figure 4 Voltage Profile Comparison
Using the PSO algorithm, the optimal size of DG unit is determined. Table II represents the real and reactive power loss reduction after placing DG using PSO.
(without DG) PSO
DG location – 18
DG size (MW) – 0.529
Real power loss (kW) 208.28 150.4
Reactive power loss (kVar) 139.3 115.17
Table II – Power Loss Reduction
In view of ever increasing load demand in the power sector, DG is playing a very vital role to improve the system performance by reducing the real power losses and improving the voltage profile. Optimal capacity and location of DG are very significant in the application of DG for loss minimization and voltage profile improvement in electric power system. Improper placement of DG in the system operation will lead to the negative effect on system operation. This paper presents new population based heuristic methods i.e. PSO algorithms is used to place the optimal sizing of DG at suitable location. The optimal location for placing the DG is identified with comparatively less voltage profile. IEEE-33 bus system is examined and the results obtained. The power loss results are tabulated and the voltage profile improvement is shown graphically. Reduction in active power losses and improvement in voltage profile can be observed.
1 T. Ackermann, G. Anderson and L. Soder., (2001). “Distributed Generation: a definition”, Electrical Power System Research, Vol. 57, pp. 195-204.
2 El-Samahy and E. El-Saadany, (2005). “The effect of DG on power quality in a deregulated environment”, Proc. IEEE Power Eng. Soc. Gen. Meet., Vol. 3, pp. 2969–2976.
5 Wang, Caisheng, and M. Hashem Nehrir, (2004). “Analytical approaches for optimal placement of distributed generation sources in power systems.” Power Systems, IEEE Transactions, Vol. 19, No. 4, pp. 2068-2076.
6 Hung, Duong Quoc, and Nadarajah Mithulananthan. (2013). “Multiple distributed generator placement in primary distribution networks for loss reduction”, Industrial Electronics, IEEE Transactions, Vol. 60, No. 4, pp. 1700-1708.
4 Khalesi, N., N. Rezaei, and M-R. Haghifam, (2011). “DG allocation with application of dynamic programming for loss reduction and reliability improvement.” International Journal of Electrical Power & Energy Systems, Vol. 33, No. 2, pp. 288-295.
3 Borghetti, A., (2012). “A mixed-integer linear programming approach for the computation of the minimum-losses radial configuration of electrical distribution networks.” Power Systems, IEEE Transactions, Vol. 27, No. 3: pp. 1264-1273.
7 López-Lezama, Jesús María, Javier Contreras, and Antonio Padilha-Feltrin. (2012). “Location and contract pricing of distributed generation using a genetic algorithm”, International Journal of Electrical Power & Energy Systems, Vol. 36, No. 1, pp. 117-126.
8 El-Zonkoly, A. M. (2011). “Optimal placement of multi-distributed generation units including different load models using particle swarm optimization”, Swarm and Evolutionary Computation, Vol. 1, No. 1, pp. 50-59.
9 Moradi, M. H., and M. Abedini. (2012). “A combination of genetic algorithm and particle swarm optimization for optimal DG location and sizing in distribution systems”, International Journal of Electrical Power & Energy Systems, Vol 34, No. 1, pp. 66-74.
10 Sudabattula, S. and Kowsalya M., 2016. “Optimal allocation of wind based distributed generators in distribution system using Cuckoo Search Algorithm”, 2nd International Conference on Intelligent Computing, Communication & Convergence, Bhubaneswar, Odisha, India, Vol. 92, pp. 298-304.
11 Nara, Koichi, et al. (2001). “Application of tabu search to optimal placement of distributed generators”, Power Engineering Society Winter Meeting,. IEEE. Vol. 2. IEEE, 2001.
12 Injeti, Satish Kumar, and N. Prema Kumar. (2013). “A novel approach to identify optimal access point and capacity of multiple DGs in a small, medium and large scale radial distribution systems”, International Journal of Electrical Power & Energy Systems, Vol. 45, No. 1, pp. 142-151.
13 Abu-Mouti, Fahad S., and M. E. El-Hawary. (2011). “Optimal distributed generation allocation and sizing in distribution systems via artificial bee colony algorithm”, Power Delivery, IEEE Transactions, Vol. 26, No. , pp. 2090-2101.
14 Naresh Acharya, Pukar Mahat, N. Mithulanathan, (2006), “An analytical approach for DG allocation in primary distribution network”, Electric Power and Energy Systems, Vol. 28, pp. 669-678.
15 Kennedy, J. and R. Eberhart, (1995). “Particle Swarm Optimization”, IEEE Proceedings of the International Conference on Neural Networks, Perth, Australia. pp. 1942-1948.
16 AlRashidi, M.R. and M.E. El-Hawary, (2009). “A Survey of Particle Swarm Optimization Applications in Electric Power Systems”, IEEE Transactions on Evolutionary Computation, Vol. 13, No. 4, pp. 913-918.
17 A. D. Rana, J. B. Darji, Mosam Pandya. (2014), “Backward / Forward Sweep Load Flow Algorithm for Radial Distribution System” International Journal for Scientific Research & Development| Vol. 2, No. 01, pp. 398-400.