INTRODUCTION

Transcription

1 OPTIMAL PLACEMENT AND SIZING OF A DISTRIBUTED GENERATOR IN A POWER DISTRIBUTION SYSTEM USING DIFFERENTIAL EVOLUTION M. Abbagana 1, G. A. Bakare 2, and I. Mustapha 3 1 & 3 Department of Electrical and Electronics Engineering, University of Maiduguri, Nigeria 2 Electrical Engineering Programme, Abubakar Tafawa Balewa University, Bauchi Nigeria. moduabbagana@yahoo.com; Bakare_03@yahoo.com; ibrahimmassa@yahoo.com Abstract Centralized generation of electricity based on bulk power planning pose many economical and environmental challenges, the best alternatives to overcome these challenges is to introduce distributed and dispersed generation, which can be conveniently located closer to load centers. There have been many studies, to define the optimal locations of distributed generation. In this paper, Differential Evolution approach is used to find the optimal location and size of a Distributed Generation (DG) unit. The DG sources are added to the network to mainly reduce the power losses and improve the voltage profile by supplying a net amount of power. The feasibility and effectiveness of the tool has been demonstrated on IEEE 33 bus radial distribution system consisting of 32 sections. MATPOWER and MATLAB software were used for simulation. The result revealed that the system losses have been reduced by percent for the installation of one DG. The nodes violating the voltage limits reduced to 3 from 18 and the sum of square of voltage error dropped to from p.u. Keywords: Distributed Generation, Real Power, Loses, Reactive Power, Optimization, Voltage, Nodes. INTRODUCTION Centralized generation of electricity based on bulk power planning methodology require large facilities, including land and personnel needed to operate and high capital cost. Moreover, since these big power stations cannot be constructed closer to load centers for some obvious reasons, there was a need for long extra high voltage or ultra high voltage transmission line, including transmission sub stations. Similar to power station, these transmission lines and sub stations need an ample amount of money in design, construction, operation and maintenance. The lengthy structure of the transmission line makes them vulnerable to natural hazards such as heavy wind, rain storms and lightning. These natural hazards, in some case become the major reason for partial or full black out of the power system triggered by some line outages. These conditions added with economical and environmental pressures have in the recent past, been changing the generation approaches of traditional electric power utilities. Some of the economical and environment factors associated with large power plants are environmental impacts, transmission right of way problems, high investment and long term planning, land requirement for power plant construction and resettlement, to overcome the economical and environmental factors associated with large power plants, one of the best alternatives for change in the traditional way of generation and delivery arrangement is to introduce distributed and dispersed generation, which can be conveniently located closer to load centers (Mithulananthan et al, 2004). Distributed generation is not a new concept. If one looks back on the evolution of the electric power industry, electricity was introduced as an alternative for steam, hydraulics, direct heating and cooling which were produced near the point of consumption in a small scale. The main idea behind the Distributed Generation is that generation is small scale, which can bee easily placed closer to the point of consumption. Various advantages and the climate of the current electricity business, strongly favour the application of DGs. However, there are many issues that need to be considered before allowing the Distributed generators to operate in power systems, in large numbers. Co-Published By: Human Resource Management Academic Research Society 536

2 Given the choices, where would the DG be placed in the system to enjoy maximum technical benefits such as low losses, higher reliability, increase in load ability, and better voltage profile. Climate change has thrust energy production to the top of the political agenda. The developed and developing world is currently dominated by centralized electricity generating system, which is the embodiment of technological inertia, performing little better today than it did in the 1970s. This centralized system is wasteful and environmentally damaging (NPO, 2007). Nigeria, like every other developing country, most of its power plants are more than years old and will need to be replaced in the nearest decade or so, offering an opportunity to move towards a more suitable system which protects the climate and provide future generations with secure energy. The trends, globally is towards Distributed Generation DG. This paper proposed a Differential Evolution technique for the optimum placement and sizing of a DG in a distribution network the technique will minimize real power losses and improve system voltage profile. Problem Formulation This paper intend to explore decision making techniques to determine the optimum sitting and sizing of distributed generation in an established distribution power network. The decision making technique is to be based on meta-heuristics optimization technique employing differential evolution. A real power loss and voltage profile analysis is to be evaluated for the system with and without DG. The real power loss reduction in a distribution system is required for efficient power system operation. The loss in the system can be calculated using eqn. (1) (Witchit and Ongasakul, 2007), called the exact loss formula (Elgerd et al, 1971) given the system operating conditions. Mathematically, the objective function can be written as (Edward et al, 2004) (1) Where, (2) are net real and reactivee power injection in bus i, respectively is the resistance between buses i and j and are the voltage and angle at bus i respectively Subject to power balance constraints: (4) (3) Co-Published By: Human Resource Management Academic Research Society 537

3 (5) Where, is the real power loss in the system is the real power generation of DG at bus is the power demand at bus is the current between buses and Related Works A survey of the literature shows that there is no consensus in the definition of DG (Pepermas et al., 2003; Zareipour et al., 2004; Mahat et al, 2006; Sedighizdeh and Rezazadeh, For the integration of distributed generation into distribution network, several literatures have proposed the use of different optimization techniques. The optimum placement and sizing is done to achieve different objectives. In (Benemar et al., 2006), evolutionary programming with the objective of maximizing the reduction on the load supply costs was used. In (Witchit and Ongasakul, 2006), Particles Swarm Optimization (PSO) was used for optimal placement of multi-dgs, with the aim of minimizing the total real power loss. Similarly, (Jahanbani et al., 2007) proposed a PSO technique with the same objective as above. The improvement in the voltage profile with this technique was presented. (Siano et al., 2007) proposed the combination of Genetic Algorithm (GA) and Optimal Power Flow (OPF) to efficiently site and size a predefined number of DGs. This differs with other proposed methods that only define the optimal locations and capacities of DG as a means of ensuring that the maximum amount of DG can be connected to existing and future networks. Other literature sources on GA optimization technique with the aim of reduction of losses and improve voltage profile are proposed in (Sedighizaeh and Rezazadeh, 2008; Deependra et al., 2007; Mithulanathan et al., 2004; Hasesen et al., 2005). In (Devi and subranmanyam, 2007), the use of Fuzzy Logic for optimal DG unit placement for loss reduction was proposed. The use of analytical approach was presented in (Mahat et al., 2006). In (Kumar and Goswani, 2009) a Genetic Algorithm based approach was used for optimal allocation of distributed generations in power systems for voltage sensitive loads. (Ajay et al., 2008) used analytical approach for sizing of DG unit operated at optimal power factor to reduce losses in radial distribution. DE has been applied in a number of engineering problems. In power engineering DE has been used to solve generation planning problems (Kannan et al., 2003); capacitor placement problems (Chiou et al., 2004); distribution network reconfiguration problems, (Chiou, et al., 2005); and inductionn motor identification problems, (Ursem and Vadstrup, 2003), etc. In this particular research, the application of DE for optimal placement and sizing of DG in a power distribution system was carried out. Modelling of DG Units DGs can be divided into two parts from the energy source view point. One is non renewable energy including cogeneration, fuel cells and micro turbine systems and the other is renewable energy including photovoltaic, wind, geothermal, biomass and so on. A constraint for DG source, similar to central generation, is active power constraint. It can be formulated as: (6) Co-Published By: Human Resource Management Academic Research Society 538

4 The reactive power output of DG units is also important and must be considered. Small and medium sized DG units mostly use asynchronous generators that are not capable of providing reactive power. Several options are available to solve this problem. On the other hand, DG units with a power electronic interface are sometimes capable of delivering a certain amount of reactive power (Pepermans et al, 2003). These interfaces or power converters can generate and inject reactive power (Q) to the network, but ratings of elements increase. The reactive power generation of DG units which use synchronous generators, depends on reactive power control strategy. There are two control strategies for this group. Constant Q/ constant power factor mode, Voltage regulated mode. Considering this point, the bus connected to the DG can be modeled as PQ or PV bus, depending on control strategy. DG Type 1 Certain type of DGs like photovoltaic will produce real power only. To find the optimal DG size at bus i, when it supplies only real power, the necessary condition for minimum loss is given by: (7) From equation (7), we obtain the following relationship: (8) Equation (8) gives the optimal DG size for each bus so as to minimize the total real power loss. Any size of DG other than placed at bus, will lead to a higher loss. This loss however is function of loss coefficient.when DG is installed in the system, the values of loss coefficients will change as it depends on the state variable voltage and angle. DG Type 2 For synchronous condenser DG, it provides only reactive power to improve voltage profile. To determine the optimal DG placement, we again differentiate the loss equation on either side with respect to. The optimal DG size for every bus in the system is given by equation (9) (9) DG Type 3 Here we consider that the DG will supply real power and in turn will absorb reactive power. In the case of the wind turbines, induction generator is used to produce real power and the reactive power will be consumed in the process (Ermis et al, 1992). The amount of reactive power they require is an ever increasing function of the active power output. The reactivee power consumed by the DG wind generation in simple form can be given as in equation (10), (Mahat et al., 2006) (10) Co-Published By: Human Resource Management Academic Research Society 539

5 The loss equation willl be modified. After following the similar methodology of the two types, optimal DG size can be found by solving (11) [ ] + ( ) = 0 (11) Equation (12) gives the amount of real power that a DG should produce when located at bus i, so as to obtain the minimum system loss whereas the amount of reactive power that it consumes can be calculated from equation (12). Load and Feeder Model The distribution feeder model adopted is shown in Fig.1, as suggested in (Siano et al., 2007), which allows the installation of loads and generation in all buses. Each branch has the following properties: origin bus, destiny bus, impedance per unit length, apparent power installed, and load power factor. The model chosen is the constant power one. There can be a load ( ) and a power generation ( ) in any bus. The substation is the feeder swing bus, while all the others, including those where generators are found, are PQ buses, with active and reactive powers specified and voltage to be determined. The reactive power generated by the unit installed at the bus must be such that: (12) + + Fig 1: Feeder model Realization of DG Based DG Placement And Sizing The optimal placement and sizing of distributed generators in a power distribution system can be achieved using the following procedure: Step i: At the initialization stage, relevant DE parameters such as maximum generation, number of control, D, population size, np, scaling factor for mutation, F, and cross over rate, CR, are defined. Also, power distribution system data required for computation process are actualized from the database. Co-Published By: Human Resource Management Academic Research Society 540

6 Step ii: Run the base case Newton Raphson load flow using MATPOWER package version 3.0 to determine the initial bus voltage, and active power losses respectively. Step iii: Each control device of the possible location and the active power are treated as parameters for optimization. Then randomly generate initial population comprising the parameters within the parameter space. The objective function for each vector of the population is computed using equation (13); (13) Where Step iv: Update the generation count. Step v: Perform mutation, cross over, selection and evaluation of the objective function as described in iii. Step vi: If the generation count is less than the preset maximum number of generations, go to step IV otherwise. Step vii: With the optimal size and location of DGs, run the final load flow to obtain the final voltage profile and the corresponding system active power loss SIMULATION RESULTS AND DISCUSSIONS In order to see the best location of DG in the distribution system with the view of minimizing the total real power losses, the differential evolution algorithm was used. Also, an IEEE 33 bus radial distribution system consisting of 32 sections shown in Fig. 2 is used in order to demonstrate the effectiveness and feasibility of the techniques. Fig 2: Single line diagram of the 33-Bus Radial Distribution System The DE parameters were varied according to the scenarios in Table 1 to see which scenario is the best. The algorithm reached a stable (optimum) solution with 30 iterations as depicted in Figures 3 8 below, and the corresponding detailed outputs are given in Table 2. Table 1: Different DE Parameters Setting. Co-Published By: Human Resource Management Academic Research Society 541

7 Scenario Number of population members (np) Iteration maximum (itermax) DE-Step size (F) Cross over probability constant (CR) Table 2: Best Placement, Size and Power Losses Scenario Best location 1 NODE 12 2 NODE 12 3 NODE 12 4 NODE 12 5 NODE 12 6 NODE 12 DG size (MW) Initial power loss(kw) Final power loss(kw) %Power loss reduction Table 3: Ssve and Number of Nodes Violating Limits Scenario Initial Initial Final Final Ssve (p.u) No of nodes Violating limits Ssve (p.u) No of nodes Violating limits According to the outputs of the six scenarios, which are presented in table 2 and 3, the initial power loss of the test system which is kw reduced to kw which is 47.39% percent of the initial loss. The nodes that violate the voltage limit dropped from 18 to 3 signifying the voltage profile has fall within the maximum and minimum limits. The sum of square of voltage error also reduced to from p.u. The corresponding DG size is MW to be located at node 12. Compared with the remaining five scenarios, scenario 5 is the best in terms of the power loss. For the sum of square of voltage error and the number of nodes violating voltage error, it is the same for all the scenarios. The convergence characteristics and the voltage profile before and after allocation of DG for the above scenarios are shown in the figures Fig 2b: Convergence Characteristics for Case 1 Co-Published By: Human Resource Management Academic Research Society 542

8 Fig 2b: Convergence characteristics for case 1, Fig 3: Convergence Characteristics for Case 1, Scenario 2 of the 33bus Test System Fig 4: Voage Profile for Case1, Scenario 2 of the 33bus System Co-Published By: Human Resource Management Academic Research Society 543

9 Fig 5: Convergence Characteristics for Case 1, Scenario 3 of the 33bus Test Fig 6: Voltage for Case1, Scenario 3 of the 33bus System Fig 7: Convergence Characteristics For Case 1, Scenario 4 of the 33bus Test System Co-Published By: Human Resource Management Academic Research Society 544

10 Fig 8: Voltage Profile For Case1, Scenario 4 of the 33bus System Co-Published By: Human Resource Management Academic Research Society 545

11 Fig 10: Convergence characteristics for case 1, scenario 5 of the 33bus test system Fig 11: Voltage profile for case1, scenario5of the 33 bus system Co-Published By: Human Resource Management Academic Research Society 546

12 Fig 12: Convergence characteristics for case 1, scenario 6 of the 33bus test system Fig 13: Voltage profile for case1, scenario 6 of the 33 bus system CONCLUSIONS An extensive review of the DG technologies and their placement and sizing in a power distribution system using differential evolution with the view to reduce real power loss and improvement of voltage profile was carried out. The advantages and disadvantages of the differential evolution algorithm have been reviewed. The feasibility and effectiveness of the developed tool has been demonstrated on IEEE 33 bus radial distribution system consisting of 32 sections. The study revealed that the proper placement and size of DG units can have a significant impact on system loss reduction and voltage profile improvement. It also revealed how improper choice of size would lead to higher losses REFERENCES Ajay-D-Vimal Raj, P., Senthilkumar, S., Raja, J., Ravichandran, S. and Palanivelu, T.G. (2008). Optimization of Distributed Generation Capacity for Line Loss Reduction and Voltage Profile Improvement Using PSO. Journal of Electrical Engineering (ELEKTIKA), Faculty of Electrical Engineering UTM, 10: Benemar, A. and Joao, M.C. (2006). Optimal Placement of Distributed Generators Networks Using Evolutionary Programming. Proceedings of IEEE PES Transmission and Distribution Conference and Expansion. Latin American, Venezuela. Co-Published By: Human Resource Management Academic Research Society 547

13 Chiou.J.P, Shang C.F and Su C. (2004) Ant Direction Hybrid Differential Evolution for Solving Large Capacitor Placement Problems. Deependra, S., Devender, S. and Verma, K.S. (2007). GA Based Optimal Sizing And Placement Of Distributed Generation For Loss Minimization. Proceedings of World Academy of Science, Engineering and Technology Vol. 26, Pp Edward, M.P., Wills, H.L., and Takahashi, M. (2004). Distributed Generation in Developing Countries. distribution_abb.pdf, Accessed June Elgerd I.O (1971), Electric energy system theory: an introduction McGrawhill. Ermis M., H. B. Eratn, M. Demirekler, B. M.Saribatir, Y. Uctung. (1992) Various Induction Generator Scheme forwind Power Electricity Generation, Electric Power Systems Research, vol.23, pp71-83,1992. Hasesen, E., Esplnoza, M., Pluymers, B., Goethals, I., Thong, V.V., Driesen, J. Belmans, R. and De Moor, B. (2005). Optimal Placement and Sizing of Distributed Generator Units Using Genetic Optimization Algorithms. Electrical Power Quality and Utilization, Journal, 11: Jahanbani, A.A., Kashefi, A.K., Pourmousavi, S.A., Hosseinian, S.H. and Abedi, M. (2007). Sitting and Sizing of Distributed Generation for Loss Reduction. Proceeding of International Conference on Power Systems (ICPS), Bangalore, India. Kannan S.K, Sharf D,J and Ganu L.U (2003) Generation planning using Differential Evolution. Mithulananthan, N., Than, O. and Le Van P. (2004). Distributed Generator Placement in Power Distribution System Using Genetic Algorithm to Reduce Losses. Thammasat Int. Journal of Science, Technology, 9: Mahat, P., Ongaskul, W. and Mithulananthan, N. (2006). Optimal Placement of Wind Turbine DG in Primary Distributions System for Real Loss Reduction. Proceeding of International Conference on Energy for Suitable Development: Issues and Prospects for Asia. Phuket. Thailand. NPO (2007); No2 nuclearpower.org.uk. Decentralized Energy Power for the 21 st Century (Briefing). One Sky/Energetic Solution Conference: OSESC, (2004). Status of Renewable Energy in Nigeria: 21 27:5 of 5 12 of 12. Pepermans G., Driesen J., Haeseldonckx D., D haeseleer W., Belmans R. (2003). Distributed Generation: Definition, Benefits and Issues. Working Paper Series No , Katholieke University Leuven Energy Institute, Website at Kumar Singh and S. K. Goswami (2009).A genetic Algorithm Based Approach for Optimal Allocation of Distributed Generations in Power Systems for Voltage sensitive Loads Co-Published By: Human Resource Management Academic Research Society 548

14 Sedighizadeh, M. and Rezazadeh, A. (2008). Using Genetic Algorithm For Distributed Generation Allocation to Reduce Losses and Improve Voltage Profile. Proceedings of World Academy of Science, Engineering and Technology, Vol. 27, Pp Siano, P., Harrison, G.P., Piccolo, A and Wallace, A.R. (2007). Strategic Placement of Distributed Generation Capacity. Proceeding of 19 th International Conference on Electricity Distribution (CIRED), Vienna. Ursem, R. K. and Vadstrup G. H (2003) Parameter Identification of Induction Motors Using Differential Evolution. EVALife, Dept. of Computer Science. Witchit, K. and Ongasakul, W. (2006). Optimal Placement of Distributed Generation Using Particle Swarm Optimization. Proceedings of the Australasian Universities Power Engineering Conference (AUPEC), Melboume, Victoria, Australia. Zareipour, H., Bhattacharya, K., and Anizares, C.A. (2004). Distributed Generation: Current Status and Challenges. Proceedings of 30 th Annual North American Power Symposium Co-Published By: Human Resource Management Academic Research Society 549